The American Journal of Emergency Medicine
Volume 23, Issue 4, Pages 423-588 (July 2005)
1. Masthead • MISCELLANEOUS Page A1
2. Editorial Board • EDITORIAL BOARD Page A2
3. Table of Contents • CONTENTS LIST Pages A3-A5
4. Predictors of medication refill–seeking behavior in the ED • ARTICLE Pages 423-428 Adam H. Miller, Gregory L. Larkin and Claudie H. Jimenez
5. Pulse oximetry in the adult ED patient with sickle cell • ARTICLE Pages 429-432 Bernard L. Lopez, Jacob F. Cogen, Leemu Kerkula, Theodore Corbin and Pamela Flenders
6. Ultrasound image transmission via camera phones for overreading • ARTICLE Pages 433-438 Michael Blaivas, Matthew Lyon and Sandeep Duggal
7. The inadvertent administration of anticoagulants to ED patients ultimately diagnosed with thoracic aortic dissection • ARTICLE Pages 439-442 Daniel P. Davis, Karun Grossman, Danielle C. Kiggins, Gary M. Vilke and Theodore C. Chan
8. Real-time paramedic compared with blinded physician identification of ST- segment elevation myocardial infarction: results of an observational study • ARTICLE Pages 443-448 James A. Feldman, Kathryn Brinsfield, Sheilah Bernard, Daniel White and Thomas Maciejko
9. Predictive value of C-reactive protein at different cutoff levels in acute appendicitis • ARTICLE Pages 449-453 Han-Ping Wu, Ching-Yuang Lin, Chin-Fu Chang, Yu-Jun Chang and Chin-Yi Huang
10. Pediatric prescription pick-up rates after ED visits • ARTICLE Pages 454-458 Eric H. Kajioka, Erick M. Itoman, M. Lily Li, Deborah A. Taira, Gaylyn G. Li and Loren G. Yamamoto
11. Does the presence or absence of sonographically identified cardiac activity predict resuscitation outcomes of cardiac arrest patients? • ARTICLE Pages 459-462 Philip Salen, Larry Melniker, Carolyn Chooljian, John S. Rose, Janet Alteveer, James Reed and Michael Heller
12. The effect of pesticide spraying on the rate and severity of ED asthma • ARTICLE Pages 463-467 Bonnie Chin-Yet O'Sullivan, John Lafleur, Kirsa Fridal, Stephen Hormozdi, Steve Schwartz, Mark Belt and Madelon Finkel
13. Changes in cardiac troponin T measurements are associated with adverse cardiac events in patients with chronic kidney disease • ARTICLE Pages 468-473 Jin H. Han, Christopher J. Lindsell, Richard J. Ryan and W. Brian Gibler
14. Moderate-to-severe blood pressure elevation at ED entry: Hypertension or normotension? • ARTICLE Pages 474-479 Thomas Dieterle, Macé M. Schuurmans, Werner Strobel, Edouard J. Battegay and Benedict Martina
15. Comparing 2 methods of emergent zipper release • ARTICLE Pages 480-482 Nobuaki Inoue, Steven C. Crook and Loren G. Yamamoto 16. The interrater variation of ED abdominal examination findings in patients with acute abdominal pain, • ARTICLE Pages 483-487 Jesse Pines, Lori Uscher Pines, Annara Hall, John Hunter, Rajagopalan Srinivasan and Chris Ghaemmaghami
17. Biphasic extrathoracic cuirass ventilation for resuscitation • ARTICLE Pages 488-491 Ilan Gur, Ephraim Bar-Yishay and Ron Ben-Abraham
18. Effect of a social services intervention among 911 repeat users • ARTICLE Pages 492-496 Steven J. Weiss, Amy A. Ernst, Margaret Ong, Ray Jones, Debra Morrow, Rosemary Milch, Katie O'Neil, Jay Glass and Todd Nick
19. The sensitivity of room-air pulse oximetry in the detection of hypercapnia • ARTICLE Pages 497-500 Michael D. Witting, Sam Hsu and Carlos Andres Granja
20. Correlation between chest x-ray and B-type natriuretic peptide in congestive heart failure • ARTICLE Pages 501-503 Chad Aleman, Shu B. Chan and Mary Frances Kordick
21. Sudden cardiac death in athletes: a guide for emergency physicians • ARTICLE Pages 504-509 Carl A. Germann and Andrew D. Perron
22. Electrocardiographic ST-segment elevation in the trauma patient: acute myocardial infarction vs myocardial contusion • ARTICLE Pages 510-516 Claire U. Plautz, Andrew D. Perron and William J. Brady
23. Diagnosis of periorbital gas on ocular ultrasound after facial trauma • ARTICLE Pages 517-520 S. Timothy McIlrath, Michael Blaivas and Matthew Lyon
24. Purple urine bag syndrome • ARTICLE Pages 521-524 Federico Vallejo-Manzur, Eduardo Mireles-Cabodevila and Joseph Varon 25. Chest radiograph screening for severe acute respiratory syndrome in the ED • ARTICLE Pages 525-530 Stewart Siu-Wa Chan, Paulina Siu-Kuen Mak, Kwok Kuen Shing, Po Nin Chan, Wing Hung Ng and Timothy Hudson Rainer
26. Clinical predictors of bleeding esophageal varices in the ED • ARTICLE Pages 531-535 Siang-Hiong Goh, Wee-Ping Tan and Shu-Woan Lee
27. The epidemiology and early clinical features of West Nile virus infection • ARTICLE Pages 536-543 Jacek M. Mazurek, Kim Winpisinger, Barbara J. Mattson, Rosemary Duffy and Ronald L. Moolenaar
28. Emergency diagnosis of Fournier's gangrene with bedside ultrasound • SHORT COMMUNICATION Pages 544-547 Daniel Morrison, Michael Blaivas and Matthew Lyon
29. Open endotracheal tubes may be safely left on an ED airway cart for 48 hours • ARTICLE Pages 548-551 W. Douglas Browder, Marcus P. Cromer and Francis L. Counselman
30. Use of gum elastic bougie for prehospital difficult intubation • ARTICLE Pages 552-555 Patricia Jabre, Xavier Combes, Bertrand Leroux, Emanuelle Aaron, Harold Auger, Alain Margenet and Gilles Dhonneur
31. Emergency medicine and older adults: continuing challenges and opportunities • ARTICLE Pages 556-560 John G. Schumacher
32. Chronic digoxin toxicity and significantly elevated BNP levels in the presence of mild heart failure • ARTICLE Pages 561-562 Kenneth Heinrich, Heather M. Prendergast and Timothy Erickson
33. Speed bump–induced spinal column injury • ARTICLE Pages 563-564 Sahin Aslan, Ozgur Karcioglu, Yavuz Katirci, Hayati Kandiş, Naci Ezirmik and Ozlem Bilir 34. Multiple organ failure as initial presentation of pheochromytoma • ARTICLE Pages 565-566 Alexandre Herbland, Nam Bui, Anne Rullier, Frederic Vargas, Didier Gruson and Gilles Hilbert
35. Carbon dioxide asphyxiation caused by special-effect dry ice in an election campaign • ARTICLE Pages 567-568 Cheng-Chun Hsieh, Chung-Liang Shih, Cheng-Chung Fang, Wen-Jone Chen and Chien-Chang Lee
36. Emergency department procedural sedation formularies • CORRESPONDENCE Pages 569-570 Alfred Sacchetti, Russell H. Harris and Dighton Packard
37. Dieulafoy's lesion presenting as near-syncope with massive gastrointestinal hemorrhage • CORRESPONDENCE Pages 570-571 Timothy Mader and Eric Beauvois
38. Effects of diabetes on the ED presentation of acute intermittent porphyria • CORRESPONDENCE Pages 571-572 Huan-Wen Chen, Deng-Huang Su, Tzuu-Shuh Jou and Jia-Horng Kao
39. Training emergency physicians to perform out-of-hospital ultrasonography • CORRESPONDENCE Page 572 Frédéric Lapostolle, Tomislav Petrovic, Jean Catineau, Gille Lenoir and Frédéric Adnet
40. Spontaneous pneumothorax presenting as epigastric pain • CORRESPONDENCE Pages 572-574 Ryoko Ogawa, Yusuke Yamamoto, Norihiro Haraguchi, Hiroaki Satoh and Kiyohisa Sekizawa
41. A cross-sectional ED survey of infantile subclinical methemoglobinemia • CORRESPONDENCE Pages 574-576 Robert J. Freishtat, James M. Chamberlain, Christina M.S. Johns, Stephen J. Teach, Cynthia Ronzio, Melissa M. Murphy-Smith and Neelam Gor
42. Common causes of hemiballism • CORRESPONDENCE Pages 576-578 Hahn-Shick Lee, Seung-Whan Kim, In-Sool Yoo and Sung-Pil Chung
43. Group A streptococcal necrotizing fasciitis: reducing the risk of unwarranted litigation • CORRESPONDENCE Pages 578-579 Zuber D. Mulla
44. Finally, a paper documenting the delay • CORRESPONDENCE Page 579 Klementyna Breyer, S. Eliza Halcomb, Nicole C. Bouchard, Robert S. Hoffman, Paul Dohrenwend and Mary Ann Howland
45. Oligoanalgesia in ED patients with isolated extremity injury without documented fracture • CORRESPONDENCE Page 580 Jesse M. Pines and Andrew D. Perron
46. Slow supraventricular tachycardia in a patient on β-blocker therapy • CORRESPONDENCE Pages 581-582 Steve Allison and Sean O. Henderson
47. Lactic acidosis caused by nucleoside analogues • CORRESPONDENCE Pages 582-583 Joseph R. Shiber
48. Detrusor hyporeflexia presents as an early manifestation in encephalitis • CORRESPONDENCE Pages 583-585 Wei Hsi Chen, Yi Fen Kao and Ming Chin Hsu
49. One-year experience with aripiprazole exposures • CORRESPONDENCE Pages 585-586 Frank LoVecchio, David Watts and Joe Winchell
50. Three-year experience with ziprasidone exposures • CORRESPONDENCE Pages 586-587 Frank LoVecchio, David Watts and P. Eckholdt
51. Erratum • ERRATUM Page 588
The American Journal of Emergency Medicine
THE AMERICAN JOURNAL OF EMERGENCY MEDICINE (ISSN including photocopy, recording, or any information storage and 0735-6757) is published seven times a year (January, March, May, retrieval system, without permission in writing from the Pub- July, September, October, November) (For Post Office use only: lisher. Printed in the United States of America. Volume 23 issue 4 of 7) by Elsevier, 1600 John F. Kennedy PERMISSIONS: Permissions may be sought directly from Blvd, Suite 1800, Philadelphia, PA 19103-2899. Periodicals postage Elsevier’s Rights Department in Philadelphia, PA, USA: phone paid at Orlando, FL 32887-4800 and additional mailing offices. 215-239-3600, e-mail [email protected]. Requests POSTMASTER: Send address changes to THE AMERICAN JOUR- may also be completed on-line via the Elsevier homepage (www. NAL OF EMERGENCY MEDICINE, Elsevier, 6277 Sea Harbor Dr, Or- elsevier.com/locate/permissions). lando, FL 32887-4800. REPRINTS: For 100 or more copies of an article in this Manuscripts, correspondence, and editorial material should publication, please contact the Commercial Reprints Depart- be sent to The Editor, THE AMERICAN JOURNAL OF EMERGENCY ment, Elsevier Science Inc., 360 Park Avenue South, New MEDICINE, P.O. Box 1494, West Bethesda, MD 20827-1494. York, New York 10010-1710. Tel. (212) 633-3813 Fax: For enquiries relating to the submission of articles (in- (212) 633-3820 e-mail: [email protected] cluding electronic submission where available) please visit ADVERTISING: Display Advertising and related corres- Elsevier’s Author Gateway at http://authors.elsevier.com. The pondence should be addressed to Inez Herrero, 360 Park Avenue, Author Gateway also provides the facility to track accepted South, New York, NY 10010. Tel: 212-633-3122 / Fax: 212-633- articles and set up e-mail alerts to inform you of when 3820; E-mail: [email protected]. Classified & Recruitment an article’s status has changed, as well as detailed artwork Advertising and related correspondence should be addressed to guidelines, copyright information, frequently asked questions Simone Imbert, 360 Park Avenue, South, New York, NY 10010. Tel: and more. 212-462-1908 / Fax: 212-633-3820; E-mail: [email protected] Contact details for questions arising after acceptance of an The ideas and opinions expressed in THE AMERICAN JOURNAL OF article, especially those relating to proofs, are provided after EMERGENCY MEDICINE do not necessarily reflect those of the Editor registration of an article for publication. or the Publisher. Publication of an advertisement or other product mention in THE AMERICAN JOURNAL OF EMERGENCY MEDICINE should YEARLY SUBSCRIPTION RATES: United States and pos- not be construed as an endorsement of the product or the manufac- sessions: individual, $203.00; institution, $301.00; student and resi- turer’s claims. Readers are encouraged to contact the manufacturer dent, $99.00; single issue, $59.00. All other countries: individual, with any questions about the features or limitations of the products $302.00; institution, $393.00; student and resident, $152.00; single mentioned. The Publisher does not assume any responsibility for issue, $59.00. For all areas outside the United States and possessions, any injury and/or damage to persons or property arising out of or there is no additional charge for surface delivery. To receive student/ related to any use of the material contained in this periodical. The resident rate, orders must be accompanied by name of affiliated reader is advised to check the appropriate medical literature and the institution, date of term, and the signature of program/residency product information currently provided by the manufacturer of coordinator on institution letterhead. Orders will be billed at indi- each drug to be administered to verify the dosage, the method and vidual rate until proof of status is received. duration of administration, or contraindications. It is the responsi- Prices are subject to change without notice. Current prices bility of the treating physician or other health care professional, are in effect for back volumes and back issues. Single issues, relying on independent experience and knowledge of the patient, to both current and back, exist in limited quantities and are offered determine drug dosages and the best treatment for the patient. for sale subject to availability. Back issues sold in conjunction The appearance of the code at the bottom of the first page with a subscription are on a prorated basis. of an article in this journal indicates the copyright owner’s Checks should be made payable to Elsevier and sent to consent that copies of the article may be made for personal or THE AMERICAN JOURNAL OF EMERGENCY MEDICINE, W.B. Saun- internal use, or for the personal or internal use of specific ders, Periodicals Department, PO Box 628239, Orlando, FL clients, for those registered with the Copyright Clearance Cen- 32862-8239. ter, Inc (222 Rosewood Drive, Danvers, MA 01923; (978) Correspondence regarding subscriptions or change of address 750-8400; www.copyright.com). This consent is given on should be directed to THE AMERICAN JOURNAL OF EMERGENCY the condition that the copier pay the stated per-copy fee for MEDICINE, Elsevier, Periodicals Department, 6277 Sea Harbor Dr, that article through the Copyright Clearance Center, Inc for Orlando, FL 32887-4800. Telephone number, (800) 654-2452; out- copying beyond that permitted by Sections 107 or 108 of the side the United States and Canada, (407) 345-4000. Changes of US Copyright Law. This consent does not extend to other address should be sent preferably 60 days before the new address kinds of copying, such as copying for general distribution, for becomes effective. Missing issues will be replaced free of charge advertising or promotional purposes, for creating new collec- if the Publisher is notified at the above address within 2 months of tive works, or for resale. Absence of the code indicates that the publication of the issue for US and Canadian subscribers and within material may not be processed through the Copyright 4 months for subscribers from all other countries. Clearance Center, Inc. © 2005 Elsevier Inc. All rights reserved. No part of this THE AMERICAN JOURNAL OF EMERGENCY MEDICINE is indexed publication may be reproduced or transmitted in any form or by and abstracted in Index Medicus, EMBASE/Excerpta Medica, any means now or hereafter known, electronic or mechanical, Current Concepts/Clinical Medicine, ISI/BIOMED, and BIOSIS.
Elsevier, John F. Kennedy Blvd, Philadelphia, PA 19103-2899. Director, Journals Production Production Editor Michael T. Miller Karen Stover Executive Publisher Publisher Christine Rullo Theresa Monturano The American Journal of Emergency Medicine
EDITOR J. Douglas White, MD, MPH, MBA, Medical College of Virginia/VCU, Richmond
EDITORIAL BOARD William J. Brady, MD, University of Virginia, Charles J. McCabe, MD, Massachusetts General Charlottesville Hospital, Boston C. Gene Cayten, MD, MPH, New York Medical College, Richard M. Nowak, MD, MBA, Henry Ford Hospital, Valhalla Detroit Neal E. Flomenbaum, MD, Cornell Medical Center, New York Joseph P. Ornato, MD, Medical College of Virginia, Richmond Glenn C. Hamilton, MD, Wright State University, Dayton Norman A. Paradis, MD, University of Colorado, Denver Oregon Health Sciences Jonathan Jui, MD, MPH, Bruce M. Thompson, MD, Henry Ford Hospital, University, Portland Detroit Gabor D. Kelen, MD, Johns Hopkins University, Baltimore Howard A. Werman, MD, Ohio State University, Columbus Toby L. Litovitz, MD, Georgetown University, Loren Yamamoto, MD, MPH, MBA, University of Hawaii, Washington, DC Honolulu
EDITORIAL CONSULTANTS 2004–2005
Neal R. Abarbanell, MD Daniel DeBehnke, MD Thomas E. Kearney, PharmD Robert D. Powers, MD J. Abbott, MD E. de la Hunta, MD Bruce Klein, MD John Prescott, MD Thomas Abramo, MD Robert Derlet, MD Wendy Klein-Schwartz, PharmD Kenneth J. Rhee, MD J. Allegra, MD Charlotte Dills, MD Michael Kobernick, MD Peter Richman, MD Michael Altieri, MD J. Ward Donovan, MD Michael Kontos, MD Raymond Roberge, MD James Amsterdam, MD Steven C. Dronen, MD Alan Kornblut, MD Steven Rothrock, MD Thomas Aufdeheide, MD Richard F. Edlich, MD Edward Krenzelok, PharmD Richard F. Salluzzo, MD Charles Bagwell, MD Martin Eichelberger, MD Paul Ladenson, MD Arthur B. Sanders, MD Joel Bantfield, MD Robert Elenbaas, PharmD Jerrold Leikin, MD Daniel T. Schelble, MD Christopher Banton, MD Charles L. Emerman, MD M. Andrew Levitt, DO Marc Bayer, MD Amy Ernst, MD Larry Lewis, MD Seth Schonwald, MD Gresham C. Bayne, MD Jay L. Falk, MD Louis Ling, MD S. Selbst, MD Michael Beeson, MD Peter Ferrera, MD Jeffrey Love, MD Michael Shannon, MD, MPH Neal Benowitz, MD Francis M. Fesmire, MD Robert A. Lowe, MD, MPH Suzanne Shepherd, MD Louis S. Binder, MD R.P. Fischer, MD Sharon Mace, MD Robert Shesser, MD K. Blaho, MD Garrett Foulke, MD Celeste Madden, MD Robert Simon, MD Michael Blaivas, MD Howard Freed, MD Thomas G. Martin, MD Adam Singer, MD James Brady, MD Robert E. Fromm, Jr, MD John Marx, MD Edward Sloan, MD Barry E. Brenner, MD, PhD Paul Gennis, MD Charles McCabe, MD Corey M. Slovis, MD Jeffrey Brent, MD Daniel Goodenberger, MD Margaret McCarron, MD Marty Smilkstein, MD Judith C. Brillman, MD Louis Graff, MD Patrick McKinney, MD Richard Sobel, MD Bryan Browne, MD Douglas Greer, MD Richard McPherson, MD Thomas O. Stair, MD William Burdick, MD David Guss, MD Robert McNamara, MD Ian Stall, MD Keith Burkhart, MD Lester Haddad, Jr, MD Scott Melanson, MD J. Stephan Stapczynski, MD Sean Bush, MD Alan H. Hall, MD David Milzman, MD Janet Talbot-Stern, MD Michael Callaham, MD Michael Handrigan, MD Howard Mofenson, MD Daniel Tandberg, MD E. Martin Caravati, MD S. Hargarten, MD Antonio Muniz, MD Donna Cardon, MD Elliott M. Harris, MD Marc S. Nelson, MD Thomas Terndrup, MD James Chamberlain, MD Ann L. Harwood-Nuss, MD Eric Noji, MD Stephen Thomas, MD Howard Champion, MD Michael B. Heller, MD Daniel J. O’Brien, MD Alexander Trott, MD William Chiang, MD Philip L. Henneman, MD J. Ohlshaker, MD Verena Valley, MD Carey Chisholm, MD Robert Hockberger, MD Harold H. Osborn, MD Ron Waldrop, MD S. Colucciello, MD Judd Hollander, MD Edward Otten, MD Daniel B. Walsh, MD Marco Coppola, DO Mark Hostetler, MD David T. Overton, MD William Watson, PharmD William Cordell, MD John Stephen Huff, MD Robert Palmer, PhD Robert L. Wears, MD Francis Counselman, MD Kenneth V. Iserson, MD Edward Panacek, MD Richard Weisman, PharmD Cameron Crandall, MD Edward J. Jackson, MD Paul M. Paris, MD Richard Wolfe, MD David Crippen, MD Andy Jagoda, MD Paul Pearigen, MD Allan B. Wolfson, MD Barbara Crouch, PharmD D. Jehle, MD Paul Pentel, MD Keith Wrenn, MD T. Forcht Dagi, MD James Jones, MD Andrew D. Perron, MD Donald Yealy, MD Daniel F. Danzl, MD Jeffrey Jones, MD Michael Plewa, MD Tighe Zimmers, MD Richard Dart, MD Marshall Kapp, JD David Plummer, MD Leslie Zun, MD Charles Davis, MD Steven Karch, MD James Pointer, MD The American Journal of Emergency Medicine Contents VOLUME 23 • NUMBER 4 • JULY 2005
ORIGINAL CONTRIBUTIONS Predictors of Medication Refill– Seeking Behavior in the ED Adam H. Miller, Gregory L. Larkin, 423 and Claudie H. Jimenez Pulse Oximetry in the Adult ED Patient With Sickle Cell Bernard L. Lopez, Jacob F. Cogen, 429 Leemu Kerkula, Theodore Corbin, and Pamela Flenders Ultrasound Image Transmission Via Camera Phones for Overreading Michael Blaivas, 433 Matthew Lyon, and Sandeep Duggal The Inadvertent Administration of Anticoagulants to ED Patients Ultimately Diagnosed 439 With Thoracic Aortic Dissection Daniel P. Davis, Karun Grossman, Danielle C. Kiggins, Gary M. Vilke, and Theodore C. Chan Real-Time Paramedic Compared With Blinded Physician Identification of ST-Segment 443 Elevation Myocardial Infarction: Results of An Observational Study James A. Feldman, Kathryn Brinsfield, Sheilah Bernard, Daniel White, and Thomas Maciejko Predictive Value of C-Reactive Protein at Different Cutoff Levels in Acute Appendicitis 449 Han-Ping Wu, Ching-Yuang Lin, Chin-Fu Chang, Yu-Jun Chang, and Chin-Yi Huang Pediatric Prescription Pick-Up Rates After ED Visits Eric H. Kajioka, Erick M. Itoman, 454 M. Lily Li, Deborah A. Taira, Gaylyn G. Li, and Loren G. Yamamoto Does the Presence or Absence of Sonographically Identified Cardiac Activity Predict 459 Resuscitation Outcomes of Cardiac Arrest Patients? Philip Salen, Larry Melniker, Carolyn Chooljian, John S. Rose, Janet Alteveer, James Reed, and Michael Heller The Effect of Pesticide Spraying on the Rate and Severity of ED Asthma 463 Bonnie Chin-Yet O’Sullivan, John Lafleur, Kirsa Fridal, Stephen Hormozdi, Steve Schwartz, Mark Belt, and Madelon Finkel Changes in Cardiac Troponin T Measurements are Associated With Adverse Cardiac 468 Events in Patients With Chronic Kidney Disease Jin H. Han, Christopher J. Lindsell, Richard J. Ryan, and W. Brian Gibler Moderate-To-Severe Blood Pressure Elevation at ED Entry Hypertension or Normotension? 474 Thomas Dieterle, Macé M. Schuurmans, Werner Strobel, Edouard J. Battegay, and Benedict Martina BRIEF REPORTS Comparing 2 Methods of Emergent Zipper Release Nobuaki Inoue, Steven C. Crook, and 480 Loren G. Yamamoto The Interrater Variation of ED Abdominal Examination Findings in Patients With Acute 483 Abdominal Pain Jesse Pines, Lori Uscher Pines, Annara Hall, John Hunter, Rajagopalan Srinivasan, and Chris Ghaemmaghami Biphasic Extrathoracic Cuirass Ventilation for Resuscitation Ilan Gur, Ephraim Bar-Yishay, 488 and Ron Ben-Abraham
(Continued on next page) Contents continued Effect of a Social Services Intervention Among 911 Repeat Users Steven J. Weiss, 492 Amy A. Ernst, Margaret Ong, Ray Jones, Debra Morrow, Rosemary Milch, Katie O’Neil, Jay Glass, and Todd Nick The Sensitivity of Room-Air Pulse Oximetry in the Detection of Hypercapnia Michael D. Witting, 497 Sam Hsu, and Carlos Andres Granja Correlation Between Chest X-Ray and B-Type Natriuretic Peptide in Congestive Heart Failure 501 Chad Aleman, Shu B. Chan, and Mary Frances Kordick REVIEWS Sudden Cardiac Death in Athletes: A Guide for Emergency Physicians Carl A. Germann 504 and Andrew D. Perron DIAGNOSTICS Electrocardiographic ST-Segment Elevation in the Trauma Patient: Acute Myocardial 510 Infarction vs Myocardial Contusion Claire U. Plautz, Andrew D. Perron, and William J. Brady Diagnosis of Periorbital Gas on Ocular Ultrasound After Facial Trauma S. Timothy McIlrath, 517 Michael Blaivas, and Matthew Lyon Purple Urine Bag Syndrome Federico Vallejo-Manzur, Eduardo Mireles-Cabodevila, and 521 Joseph Varon Chest Radiograph Screening for Severe Acute Respiratory Syndrome in the ED 525 Stewart Siu-Wa Chan, Paulina Siu-Kuen Mak, Kwok Kuen Shing, Po Nin Chan, Wing Hung Ng, and Timothy Hudson Rainer Clinical Predictors of Bleeding Esophageal Varices in the ED Siang-Hiong Goh, 531 Wee-Ping Tan, and Shu-Woan Lee The Epidemiology and Early Clinical Features of West Nile Virus Infection 536 Jacek M. Mazurek, Kim Winpisinger, Barbara J. Mattson, Rosemary Duffy, and Ronald L. Moolenaar Emergency Diagnosis of Fournier’s Gangrene With Bedside Ultrasound Daniel Morrison, 544 Michael Blaivas, and Matthew Lyon THERAPEUTICS Open Endotracheal Tubes May Be Safely Left on an ED Airway Cart for 48 Hours 548 W. Douglas Browder, Marcus P. Cromer, and Francis L. Counselman Use of Gum Elastic Bougie for Prehospital Difficult Intubation Patricia Jabre, 552 Xavier Combes, Bertrand Leroux, Emanuelle Aaron, Harold Auger, Alain Margenet, and Gilles Dhonneur CONTROVERSIES Emergency Medicine and Older Adults: Continuing Challenges and Opportunities 556 John G. Schumacher CASE REPORTS Chronic Digoxin Toxicity and Significantly Elevated BNP Levels in the Presence of 561 Mild Heart Failure Kenneth Heinrich, Heather M. Prendergast, and Timothy Erickson Speed Bump–Induced Spinal Column Injury Sahin Aslan, Ozgur Karcioglu, 563 Yavuz Katirci, Hayati Kandis¸, Naci Ezirmik, and Ozlem Bilir Multiple Organ Failure as Initial Presentation of Pheochromytoma Alexandre Herbland, 565 Nam Bui, Anne Rullier, Frederic Vargas, Didier Gruson, and Gilles Hilbert Carbon Dioxide Asphyxiation Caused by Special-Effect Dry Ice in an Election Campaign 567 Cheng-Chun Hsieh, Chung-Liang Shih, Cheng-Chung Fang, Wen-Jone Chen, and Chien-Chang Lee
(Continued on next page) Contents continued CORRESPONDENCE Emergency Department Procedural Sedation Formularies Alfred Sacchetti, Russell H. Harris, 569 and Dighton Packard Dieulafoy’s Lesion Presenting as Near-Syncope With Massive Gastrointestinal Hemorrhage 570 Timothy Mader and Eric Beauvois Effects of Diabetes on the ED Presentation of Acute Intermittent Porphyria Huan-Wen Chen, 571 Deng-Huang Su, Tzuu-Shuh Jou, and Jia-Horng Kao Training Emergency Physicians to Perform Out-of-Hospital Ultrasonography 572 Frédéric Lapostolle, Tomislav Petrovic, Jean Catineau, Gille Lenoir, and Frédéric Adnet Spontaneous Pneumothorax Presenting as Epigastric Pain Ryoko Ogawa, Yusuke Yamamoto, 572 Norihiro Haraguchi, Hiroaki Satoh, and Kiyohisa Sekizawa A Cross-Sectional ED Survey of Infantile Subclinical Methemoglobinemia Robert J. Freishtat, 574 James M. Chamberlain, Christina M.S. Johns, Cynthia Ronzio, Stephen J. Teach, and Melissa M. Murphy-Smith Common Causes of Hemiballism Hahn-Shick Lee, Seung-Whan Kim, In-Sool Yoo, and 576 Sung-Pil Chung Group A Streptococcal Necrotizing Fasciitis: Reducing the Risk of Unwarranted Litigation 578 Zuber D. Mulla Finally, A Paper Documenting the Delay Klementyna Breyer, S. Eliza Halcomb, 579 Nicole C. Bouchard, Robert S. Hoffman, Paul Dohrenwend, and Mary Ann Howland Oligoanalgesia in ED Patients With Isolated Extremity Injury Without Documented Fracture 580 Jesse M. Pines and Andrew D. Perron Slow Supraventricular Tachycardia in a Patient on β-blocker Therapy Steve Allison and 581 Sean O. Henderson Lactic Acidosis Caused by Nucleoside Analogues Joseph R. Shiber 582 Detrusor Hyporeflexia Presents as an Early Manifestation in Encephalitis Wei Hsi Chen, 583 Yi Fen Kao, and Ming Chin Hsu One-Year Experience With Aripiprazole Exposures Frank LoVecchio, David Watts, 585 and Joe Winchell Three-Year Experience With Ziprasidone Exposures Frank LoVecchio, David Watts, 586 and P. Eckholdt American Journal of Emergency Medicine (2005) 23, 423–428
www.elsevier.com/locate/ajem
Original Contributions Predictors of medication refill–seeking behavior in the ED
Adam H. Miller MD, MSc*, Gregory L. Larkin MD, MSPH, Claudie H. Jimenez MD, MSc
Division of Emergency Medicine, University of Texas Southwestern Medical Center and the Parkland Health and Hospital System, Dallas, TX 75390-8579, USA
Received 2 January 2005; accepted 8 January 2005
Abstract Objective: Determine predictors of medication refill–seeking behavior in ED patients with chronic illness. Methods/Design: Prospective cross-sectional ED survey conducted for 6 weeks. Setting: Public hospital ED (N140000 visits per year). Subjects: ED patients (N18 years) taking chronic medications for congestive heart failure, diabetes, and/ or hypertension. Results: Of 1168 patients surveyed, 344 (29%) presented to the ED secondary to running out of medications and requiring a medication refill. Univariate predictors included age younger than 50 years, non-Hispanic ethnicity, low income (bUS$5000 per year), self-pay payor status, and being told to call a primary care physician before medication would be refilled. Lack of knowledge about refill or pharmacy numbers on the medication bottle resulted in patients being more than twice as likely to be in the ED for a medication refill (odds ratio 2.4 [1.6, 3.6] and 2.0 [1.3, 2.9], respectively). Conclusion: Presenting for medication refills is common in ED patients with chronic illness. D 2005 Elsevier Inc. All rights reserved.
1. Introduction access [3-6], and the inverse trend between prescription drug costs and insurance benefits [7]. EDs have seen an ever-increasing volume of users, Total prescription drug expenditures in the United States fueling an overcrowding crisis in EDs across the nation [1]. have more than doubled in the past decade to more than In the year 2000, ED visits were at an all-time high of 108 US$160 billion, and prescription drug spending per capita million [2]. Contributing to the overcrowding problem, a has nearly tripled over the same period to approximately small group of patients (3-4%) present to the ED for US$500 per person per year [8,9]. Prescription drug costs medication refills, and some studies suggest that this group currently represent the fastest growing component of all accounts for a disproportionate number (12%-20%) of total health care spending in the United States (15.3%) [8,10] ED visits [3,4]. These frequent ED users reflect the urban and have led consumers to Mexico, Canada, the World social problems of poverty, chronic illness, health care Wide Web, and even the EDs in search of affordable medications. Although numerous authors have investigated the overall reasons for presenting to the ED for medical care T Corresponding author. Department of Emergency Medicine, Depart- ment of Surgery, University of Texas Southwestern Medical Center, Dallas, [3,4,11-16], few have measured the associated risks of Texas 75390-8579, USA. Tel.: +1 214 648 2965. relying on the ED as a pharmacy of last resort for routine E-mail address: [email protected] (A.H. Miller). medications. Therefore, we sought to evaluate the preva-
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2005.01.009 424 A.H. Miller et al. lence and risks associated with ED-based refill-seeking were chosen a priori and entered into the multivariate behavior, a growing practice among ED patients with analysis if they were significant at the P = .15 level or better, chronic illness. and they were retained, provided P b .05.
2. Methods 3. Results The local institutional review board approved this study 3.1. General demographics and the attendant consent process. Adult patients (age N18 years) presenting to a large public ED (serving over Research assistants identified 1168 consecutive patients 140000 patients per year) who, by initial history taking, who were eligible for the study. No patient refused. Upon carried the diagnosis of congestive heart failure (CHF), audit of the ED electronic record, there were no statistical diabetes mellitus (DM), and/or hypertension (HTN) were differences in sex, ethnicity, age, or insurance status eligible for enrollment. Patients were excluded if they were compared with the survey data obtained. Of the 1168 pregnant, unable to speak in Spanish or English, had acute trauma, or were otherwise in extremis. Table 1 Survey percentages of 1168 patients with CHF, DM, From June 1 to July 15, 2002, trained research assis- and HTN who sought care in our ED based on ethnicity, tants were instructed to come in to the ED during ran- household income, insurance status, and other social factors dom time blocks to identify eligible patients for the study. b After consent was obtained, researchers administered the Demographic % Of all % Of the average patients ED patientQ 39-question survey in English or Spanish. The survey n study nationwide instrument was reviewed by an independent panel of health policy experts and psychologists and found to have good face Sex (men) 41.7% 45.9% Age N50 y 59.3% 33.0% validity using a modified Delphi technique for measuring the N b Q Weight 225 lb 20.9% construct of medication refill–seeking behavior. Ethnicity For analysis and weighting, the 3 chronic disease states African American 51% 22.6% were further defined as follows: patients who carried an White 20% 74.2% antecedent diagnosis of CHF were categorized into the Hispanic 27% bCHF groupQ even if they also had DM or HTN. Patients Other 2% 3.3% with DM alone or DM in combination with HTN were Insurance status categorized into the DM group. Patients with HTN alone Private insurance 6% 38.9% were categorized into the HTN group. Confirmation of Medicaid managed care 33% 19.7% disease status and the chronic use of medication for these Medicare 19% 15.4% disease states were obtained when patients voluntarily Medicaid 9% No insurance 33% 14.5% demonstrated their medication bottles to the investigators. Household income (US$/y) Patients were also categorized by demographics and b5000 51% insurance status. Patients with public assistance from the 5000-10000 23% county government were classified as having bMedicaid 10000-30000 23% managed care.Q Privately insured patients (with or without a 30000 3% managed care component) were categorized as bprivateQ even Unemployed 8.7% if they had other types of insurance. Patients without private Social factors insurance or Medicaid managed care were categorized into Lacking social support 28% 3 remaining groups: Medicaid, Medicare, or buninsured.Q (lives alone) Noncitizen 6.6% 2.1. Data analysis Non–English-speaking 25% Own home 59% Descriptive statistics, frequency counts, and means were Education compared using 2-tailed t tests or nonparametric (Mann- bHigh school education 46.5% Whitney U or Kruskal-Wallis) tests as appropriate. Differ- High school education 53.5% ences of proportion for categorical variables were compared Some college 22.1% for the 3 chronic disease states using standard v2 statistics Medical history and Fisher exact test, P b .05 being significant and multiple DM 49.2% comparisons being controlled for with Bonferroni correc- HPN 79.3% CHF 19.7% tion, when appropriate ( P b .01). Bivariate comparisons Two or more conditions 38.7% were made using simple odds ratios (ORs) with 95% b Q confidence intervals (CIs), and multivariate analyses were Also included are the national percentages of the average ED patient for age, sex, ethnicity, and insurance status (column 2) [48]. conducted using logistic regression; candidate variables Predictors of medication refill–seeking behavior in the ED 425
patients surveyed, 20% were categorized into the CHF Table 2 Univariate predictors of 1168 patients with chronic group, 39% DM, and 41% HTN alone. Of the 229 patients disease (eg, CHF, DM, and HTN) who were bout of medi- cationQ when presenting to the ED categorized as CHF, 13% reported a diagnosis of CHF alone, 5% also had DM, and 34% also had HTN, whereas Covariate Out of OR (95% CI) 48% had all 3" chronic diseases. Of the 449 patients with medication (%) DM, 55% also had HTN. The average age was 54 years (FSD 13). The most common type of insurance in the Demographic overall cohort was Medicaid managed care (32%), whereas Men 28.4 0.97 (0.75, 1.3) 6% had private insurance, and 33% had no insurance. Most Age N50 y 26.3 0.73 (0.56, 0.94)T Weight N225 lb 25.1 0.78 (0.57, 1.1) patients (51%) were African American, and the next largest Ethnicity ethnic group was Hispanic (27%), whereas 24% of the White 26.2 0.84 (0.61, 1.17)T sample preferred to communicate in Spanish (Table 1). African American 31.8 1.3 (1.0, 1.7)T Hispanic 24.9 0.76 (0.57, 1.0) 3.2. Univariate and multivariate associations Household income (US$/y) Of the 344 (29%, 95% CI 26.3%-32%) patients b5000 32.1 1.4 (1.1, 1.7)T Having to pay NUS$10 29.9 1.2 (0.77, 1.8) presenting to the ED who ran out of medications, 51% to see PCP had been out for 10 days or more. Only 22% were out for Social factors 1 day, and 27% were out from 2 to 9 days. Significant Low social supporta 25.5 0.78 (0.58, 1.0) univariate predictors of bbeing in the ED for a medication Noncitizen 22.4 0.69 (0.4, 1.2) refillQ included age younger than 50 years, non-Hispanic, Unemployed 29.8 1.2 (0.89, 1.7) African-American ethnicity, low income (bUS$5000 per Own home 27 0.81 (0.62, 1.0) year), and being uninsured. In addition, patients who had Have own physician 92 0.54 (0.78, 1.1) diminished awareness of documented refill information ED reliant for most 31.5 1.2 (0.96, 1.6) were also at risk. Specifically, patients who did not know of care where refill number is on their medication bottle were Uninsured 32.5 1.49 (1.13, 1.9)TT nearly 2.4 times (OR 2.4 [1.6, 3.6]) as likely to be in the bHigh school education 27.9 0.93 (0.72, 1.2) Some college 28.6 0.99 (0.73, 1.4) ED for a medication refill. Similarly, patients unaware of a Medical history pharmacy phone number on the bottle were twice as likely Diabetes 28 0.92 (0.71, 1.2) (OR 2.0 [1.3, 2.9]) to need a medication refill. Conversely, Hypertension 29.2 1.1 (0.8, 1.5) patients who had called the pharmacy for a refill in the past Congestive heart failure 31.4 1.2 (0.85, 1.6) were only approximately 25% as likely to be in the ED for Two or more conditions 30.0 1.11 (0.86, 1.4) a medication refill. As these 2 factors are not independent, MD exposure in previous they were combined into a third variable, buninformed 12 months about refill or prescription numbersQ which was subse- See physician every 30.9 1.2 (0.9, 1.5) quently entered into the multivariate logistic regression 4 weeks model. Nearly 9% of all patients (n = 100) were told that See physician every 29.6 1.1 (0.83, 1.5) they must call a primary care physician (PCP) before 6 months Admitted to hospital 30.7 1.2 (0.45,1.6) obtaining a refill, and they were more than 9 times as likely in prior 12 months (OR 9.2 [5.7, 14.8]) to come to the ED in need of Know where pharmacy 55.4 0.51 (0.34, 0.75)TTT medication refills (Table 2). number isb Called pharmacy 28.6 24 (0.16, 0.34)TTT for refill Table 3 Multivariate logistic regression model predicting Prescription factors those patients with chronic disease (eg, CHF, HTN, and DM), b Q Told to see PCP 75 9.2 (5.7, 14.8)TTT who present to the ED out of medication for refill Covariate Odds ratio (95% CI) Attempted to call 59.3 0.92 (0.62, 1.4) Uninsured 2.1 (1.5, 2.95)T PCP/clinic Admitted in previous 12 months 1.5 (1.1, 2.0)TT Rx filled at county 33.3 1.4 (1.0, 1.8)T Uninformed at refill or prescription 5.05 (3.7, 6.9)T hospital number Refills listed on bottle 37.7 0.31 (0.2, 0.49)TTT Told to see PCP at refill 6.2 (3.7, 10.4)T a Lives alone. ED reliance/insurance status 1.8 (1.2, 2.6)TT b If refill and number are on a prescription bottle. The model is significant; Wald statistic = 183.4, Nagelkerke R2 = 0.255. T Fisher exact test P = .05. Fisher exact test P = .05. TT Fisher exact test P = .01. T Fisher exact test P = .001. TTT Fisher exact test P = .001. TT Fisher exact test P = .01. 426 A.H. Miller et al.
In a multivariate logistic regression model controlling for our study, the univariate and the multivariate model both age, ethnicity, employment status, income, PCP contact, and revealed that if the patient knew where the refill number was social support, 5 candidate variables remained significant on the bottle and called the pharmacy for a refill, they were predictors of being out of medication and presenting to the unlikely to present to the ED for a medication refill. This ED (Table 3). Lack of insurance, being admitted in the finding is indirectly supported by Mandelberg and Kohn [3], previous year, being uninformed about refill or prescription who found that access to a telephone was itself associated number, being told to see their PCP about a refill, and the with lower overall ED use. interaction between ED reliance and insurance status were PCP access was contributory as well, because patients all significant in the multivariate model. who were told to see their PCP before getting a refill were significantly more likely to present for an ED refill. (Table 2). The most recent annual report from the Centers 4. Discussion for Medicare and Medicaid Services suggests that health care costs in the United States reached US$1.6 trillion in This is 1 of the first studies to describe the prevalence of 2002 which is a 9.3% increase from 2001. The prescription refill-seeking behavior in the ED. Others have examined drug costs rose to 15.3% in 2002, higher than any other medication nonadherence phenomena in other settings and component of the health care dollar [8,10]. discovered similar associations with being uninsured and Indeed, refill compliance problems are, in part, caused by younger age with medication compliance [17-35]. Previous the inverse relationship between drug costs and declining work has shown that patients do not follow their prescribed Medicare, Medicaid, and managed care reimbursements. regimen of medication because of a perceived lack of Increasing co-pay and out-of-pocket costs for prescriptions susceptibility to illness, severity of illness, benefits of [20,34], coupled with poor economic conditions [5,11,36-38, treatment, and access to treatment [36-38]. However, we 43,44] and the expediency of provider access are powerful could find no other studies demonstrating the strong incentives to refill seeking behavior in the ED. association between showing up in the ED for a medication Didlake et al [32] studied how poor compliance with refill and modifiable risk factors such as knowledge about immunosuppressant medications in renal transplant patients, common refill practices. in part caused by the high cost of the medication as well as The relationship between medication adherence and discontinuation of immunosuppressant medication by Medi- medical outcomes in those with chronic disease (eg, DM care after the first year from transplantation (1992), is a [21,22], HTN [23-26,28-31,39], CHF [17], organ transplant leading cause of transplant rejection [33,34]. In addition, [32-35]) is well documented. Indeed, when noncompliance hospital readmission rates per patient among noncompliant occurs, hospital admissions increase, and less favorable renal transplant patients (5.9) was higher than compliant outcomes ensue [32,33,40]. The potential economic value to renal transplant patients (2.5); the mean length of stay was be gained in more favorable health outcomes from 11.3 versus 8.6 days, and mean costs were US$28542 uninterrupted insurance coverage and medication adherence versus US$12885, respectively [34]. is estimated to be between US$65 and US$130 billion each Soumerai et al [40] reported that the limitation of year [41]. reimbursement to 3 medications, for example, insulin, Although not tested here, patient knowledge and thiazide, and furosemide, resulted in a decline in the use of perceptions can be altered by interventions, which link the medication and an increase in nursing-home admissions communication to improved compliance [23,31,42]. The so- in those patients older than 60 years. Monane et al [17] called bESFT modelQ for communication and compliance examined digoxin therapy for CHF in the face of decreased described by Betancourt et al [23] addresses 4 domains: reimbursement and demonstrated that digoxin was not used (1) the explanatory model (domain E) whereby the physician regularly throughout a 1-year follow-up (111/365 days). Only must first understand the patient’s conceptualization of his or 10% of patients obtained enough refill to have sufficient her illness; (2) providers must then address the patient’s digoxin for the whole study period. ability to afford or physically obtain a prescribed medication Our study confirms previous findings as both being described as social risk for noncompliance (domain S); uninsured and a frequent visitor were both statistically (3) fears and concerns about the medication (domain F) significantly associated with presenting to the ED for a dangers including a variety of sociocultural conceptualiza- medication refill. We found that social issues were tions and folk beliefs; and (4) therapeutic contracting and associated with patients presenting to the ED for medi- playback (domain T) for patients with linguistic or contextual cation refill: lack of social support (living alone), low barriers to communication ensuring that patients understand income, and being a noncitizen, unemployed, and un- the prescribed regimen and can repeat or bplay backQ medica- insured. However, there was no significant association tion instructions. between having a PCP or not, when presenting to the ED Cramer [42] also suggests that if the provider delivers for a medication refill. (Table 2) In a multivariate model, specific feedback about missed doses, timing, and other we found that uninsured patients who had been admitted adherence behaviors, overall compliance might improve. In to the hospital in the previous 12 months and who were Predictors of medication refill–seeking behavior in the ED 427 heavily reliant on the ED for most of their care were As a single site study, our results may not be significantly more likely than other chronically ill patients generalizable to other rural or nonurban settings, although to present to the ED for medication refills (Table 3). we had a 0% refusal rate and randomly sampled from an Increased dependence on the ED bsafety netQ as the ethnically diverse, large county hospital ED population. source of routine [7,30,45-47] care directly impacts hospital Future work in other ED and clinic setting would be admission rates and the evolving scope and role of the ED important to confirm the alarmingly high prevalence of for the treatment and maintenance of chronic health refill-seeking behavior seen in this study. problems [36-38]. The old practice paradigm of bfewer prescriptions in the ED and follow-up with primary careQ is not going to work in a population that is increasingly reliant 6. Conclusion on the ED to fill gaps in routine care. New strategies are needed. Pharmacist refill algorithms, for example, have In ED patients with chronic disease, coming to the ED demonstrated decreased cost, unscheduled physician visits, for medication refills is common. Significant predictors of urgent care visits, emergency room visits, and hospital refill-seeking behavior include age younger than 50 years, b admissions days [45]. bBridging prescriptions,Q once taboo, African-American ethnicity, low income ( US$5000 per may give way to novel bprescription deskQ strategies and year), unfamiliarity with common medication refill practi- advanced educational and triage interventions to mitigate ces, and being told to call their primary care provider for a the ED refill phenomenon. refill. While the gaps in knowledge and awareness regarding medication refill practices may be remedial to intervention with high-risk patients, further study will be required to 5. Limitations document the ultimate impact of ED prescription writing on patient-centered outcomes. Scientifically unvalidated instruments and self-report data are fraught with some limitations and reporting biases. Although most of the survey instrument questions have References face validity when subsequently reviewed by independent health policy experts and psychologists, there may be other [1] Derlet RW, Richards JR. Overcrowding in the nation’s emergency unmeasured risk and protective factors that could have departments: complex causes and disturbing effects. Ann Emerg Med 2000;35:63-8. accounted for prescription refill–seeking behavior in the [2] McCraig LF, Nghi L. National hospital ambulatory care survey: 2000 ED. The candidate predictors offered here were obtained emergency department summary advanced data from vital and health cross-sectionally and do not account for long-term trends in statistics. Advanced Data 2002;326. behavior, but provide an important first step into under- [3] Mandelberg JH, Kohn MS. Epidemiologic analysis of an urban, public standing this phenomenon. Although reliance on self-report emergency department’s frequent users. Acad Emerg Med 2000;7: 637-46. data is an important second limitation, the direction of [4] Murphy AW, Leonard C, Plunkett PK, et al. Characteristics of attenders these biases would tend toward patients underreporting of and their attendances at an urban accident and emergency department using the ED for medication refills, and the actual over a one year period. J Accid Emerg Med 1999;16:425-7. prevalence of this behavior may be even greater than the [5] Rask KJ, Williams MV, McHagny SE, Parker RM, Baker DW. 29% measured here. Ambulatory health care use by patients in a public hospital emergency department. J Gen Intern Med 1998;13:614-20. The sample was predominately of low-income patients [6] McCusker J, Healey E, Bellavance F, Connolly B. Predictors of repeat who were already in the ED, which may lend itself to emergency department visits by elders. Acad Emerg Med 1997; adherence differences when compared with a wealthier 4:581-8. population who primarily receive their care in private [7] Institute of Medicine, IOM Board. The future of Emergency care physician offices or private hospitals. However, our data in the United States Health system/ Consequences of uninsurance. h tt p: // ww w.i om .e d u/ pr o je c t. a sp ? id =1 60 7 a nd h tt p : // w ww.i om .e d u/ were obtained in 1 of the largest EDs in the country with a file.asp?id=17732. highly diverse population that represents many EDs across [8] Centers for Medicare and Medicaid Services news. Healthcare the nation. The demographics from our study (Table 1, Spending reaches $1.6 trillion in 2002. http://cms.hhs.gov/media/ column 2) can be compared with those of the baverage ED press/release.asp?counter=935. patientQ (Table 1, column 3) [48]. [9] Carey J, Barrett A. Hill McGraw: New York. http://Bussinessweek. com/magazine/content/01_50/b3761004.htm. The inclusion criteria consisted of the 3 chronic [10] Appelby J. Heathcare spending’s rapid growth continues. http:// illnesses that represent a sample of problems commonly USAToday.com/news/health/2004-01-09-healthcosts_x.htm. seen in the ED, however, other presenting diseases may [11] Lucas RH, Sanford SM. An analysis of frequent users of emergency have different refill-seeking characteristics than the care at an urban university hospital. Ann Emerg Med 1998;32:563-8. reported study group. The selected chronic diseases may [12] Cramer JA. Relationship between medication compliance and medical outcomes. Am J Health Syst Pharm 1995;52(Suppl 3):S27-9. bias toward a more severely ill population, although we [13] Ullman R, Block JA, Stratmann WC. An emergency room’s patients: carefully excluded the sickest patients and those in need of their characteristics and utilization of hospital services. Med Care critical care. 1975;13:1011-9. 428 A.H. Miller et al.
[14] Bamberger JD, Unick J, Klein P, et al. Helping the urban poor stay with [31] Inui TS, Yourtee EL, Williamson JW. Improved outcomes in hyper- antiretroviral drug therapy. Am J Public Health 2000;90:699-701. tension after physician tutorials. Ann Intern Med 1976;84:646-51. [15] Mannon JM. Defining and treating bproblem patientsQ in a hospital [32] Didlake RH, Dreyfus K, Kerman RH, et al. Patients noncompliance: a emergency room. Med Care 1976;14:1004-13. major cause of late graft failure in cyclosporine-treated renal patients. [16] Sun BC, Burstin HR, Brennan TA. Predictors and outcomes of Transplant Proc 1988;20:63-9. frequent emergency department users. Acad Emerg Med 2003; [33] Rovelli M, Palmeri D, Vossler E, et al. Noncompliance in organ 10:320-8. transplant recipients. Transplant Proc 1989;21:833-4. [17] Monane M, Bohn RL, Gurwitz JH, et al. Noncompliance with [34] Swanson M, Hull D, Bartus S, et al. Economic impact of congestive heart failure therapy in the elderly. Arch Intern Med noncompliance in kidney transplant recipients. Transplant Proc 1994;154:433-7. 1992;24:2722. [18] Becker MH, Maiman LA. Sociobehavioral determinants of com- [35] Schweizer RT, Rovelli M, Palmeri D, et al. Noncompliance in organ pliance with health and medical care recommendations. Med Care transplant recipients. Transplantation 1990;49:374-7. 1975;13:10-24. [36] Cross LA. Pressure on the emergency department: the expanding right [19] DiMatteo MR, Sherbourne CD, Hays RD, et al. Physicians’ to medical care. Ann Emerg Med 1992;21:1266. characteristics influence patients’ adherence to medical treatment: [37] Access to emergency medical care: emergency physicians and results from the medical outcome study. Health Psychol 1993; uncompensated care. Ann Emerg Med 1987;16:1302. 12:93-102. [38] Medicaid Access Study Group. Access of Medicaid recipients to [20] Sherbourne CD, Hays RD, Ordway L, et al. Antecedents of adherence outpatient care. N Engl J Med 1994;330:142. to medical recommendations: results from the Medical Outcomes [39] Brown CM, Segal R. The effects of health and treatment perceptions on Study. J Behav Med 1992;15:447-68. the use of prescribed medication and home remedies among African [21] Jack Jr L, Liburd L, Vinicor F, et al. Influence of the environmental American and white American hypertensives. Soc Sci Med 1996;43: context on diabetes self-management: a rationale for developing a 903-17. new research paradigm in diabetes education. Diabetes Educ 1999; [40] Soumerai SB, Ross-Degnan D, Avorn J, et al. Effects of Medicaid 25:775-80. drug-payment limits on admission to hospitals and nursing homes. [22] Avry DV, Toobert D, Wilson W, et al. Patient’s perspective on factors N Engl J Med 1991;325:1072-7. contributing to nonadherence to diabetes regimens. Diabetes Care [41] Institute of Medicine. IOM Board. The future of Emergency care in 1986;9:168-72. the United States Health system/Consequences of uninsurance. http:// [23] Betancourt JR, Carrillo JE, Green AR. Hypertension in multicultural www.iom.edu/file.asp?id=12313. and minority populations: linking communication to compliance. Curr [42] Cramer JA. Quality of life and compliance. In: Trimble MR, Dodson Hypertens Rep 1999;1:482-8. WE, editors. Epilepsy and quality of life. New York7 Raven; 1994. [24] Ahluwalia JS, McNagny SE, Kanuru NK. A randomized trial to im- p. 49. prove follow-up care in severe uncontrolled hypertensives at an inner- [43] Petersen LA, Burstin HR, O’Neil AC, Orav EJ, Brennan TA. Non- city walk-in clinic. J Health Care Poor Underserved 1996;7:377-89. urgent emergency department visits: the effect of having a regular [25] Shea S, Misra D, Erlich MH, et al. Correlates of nonadherence to doctor. Med Care 1998;36:1249-55. hypertension treatment in an inner-city minority population. Am J [44] Purdie FR, Honigman B, Rosen P. The chronic emergency department Public Health 1992;82:1607-12. patient. Ann Emerg Med 1981;10:298-301. [26] Rose LE, Kim MT, Dennison CR, et al. The contexts of adherence for [45] Eggers PW, Greenberg LG. Racial and ethnic differences in African Americans with high blood pressure. J Adv Nurs 2000;32: hospitalization rates among aged Medicare beneficiaries, 1998. Health 587-94. Care Financ Rev 2000;21:91-105. [27] Johnson KH, Bazargan M, Bing EG. Alcohol consumption and [46] Hansagi H, Olsson M, Sjoberg S, Tomson Y, Goransson S. compliance among inner-city minority patients with type-2 diabetes Frequent use of the hospital emergency department is indicative mellitus. Arch Fam Med 2000;9:964-70. of high use of other health care services. Ann Emerg Med 2001; [28] Rudd P. Clinicians and patients with hypertension: unsettled issues 37:561-7. about adherence. Am Heart J 1995;130:572-9. [47] Cassidy IB, Keith MR, Coffey EL, Noyes MA. Impact of [29] Bailey JE, Lee MD, Somes GW, et al. Risk factors for antihyperten- pharmacist-operated general medicine chronic cares refill clinics sives medication refill failure by patients under Medicaid managed on practitioner time and quality of care. Ann Pharmacother 1996; care. Clin Ther 1996;18:1252-62. 30(7-8):745-55. [30] Monane M, Bohn RL, Gurwitz JH, et al. Compliance with [48] McCaig LF, Burt CW. National Hospital Ambulatory Medical Care antihypertensive therapy among elderly Medicaid enrollees: the roles Survey: 2002 emergency department summary. Adv Data of age, gender, and race. Am J Public Health 1996;86:1805-8. 2004;18(340):1-34. American Journal of Emergency Medicine (2005) 23, 429–432
www.elsevier.com/locate/ajem
Pulse oximetry in the adult ED patient with sickle cell
Bernard L. Lopez MD, MS*, Jacob F. Cogen, Leemu Kerkula, Theodore Corbin MD, Pamela Flenders MSN
Department of Emergency Medicine, Jefferson Medical College, Philadelphia, PA 19107, USA
Received 14 August 2004; revised 1 September 2004; accepted 13 September 2004
Abstract We performed a retrospective chart review of adult patients with sickle cell presenting with vasoocclusive crisis to determine the association between pulse oximetry and emergency department (ED) disposition. Subjects were divided into a NORMAL (pulse oximetry z95%) and a LOW (pulse oximetry b95%) group. Two hundred ten consecutive charts showed no significant difference between NORMAL (n = 163) and LOW (n = 47) groups regarding admission or discharge from the ED ( P N .05). A higher percentage in the LOW group received chest radiographs (40.9% vs 29%, P b .05), suggesting that pulse oximetry may have influenced ordering of this test. No significant differences in historical and physical exam characteristics were found. The pulse oximetry level does not appear to be associated with a particular ED disposition in adult sickle cell anemia. D 2005 Elsevier Inc. All rights reserved.
1. Introduction syndrome, a condition of new pulmonary infiltrates, fever, chest pain, and hypoxia. Acute chest syndrome is the Sickle cell anemia (SCA) is the most common inherited leading cause of death and second leading cause of blood disorder in the United States, affecting about 1 in 600 hospitalization in the patient with SCA [3]. For this reason, African Americans [1]. Uncomplicated vasoocclusive crisis the ability to accurately evaluate the pulmonary status of (VOC) represents the most common emergency department patients with SCA is important. (ED) presentation and is the result of microvascular Pulse oximetry is a noninvasive and commonly used test sludging and thrombosis causing tissue hypoxia and in the ED for detecting hypoxemia. Although it has been ischemia. The emergency physician has 2 main duties found useful in many conditions, its usefulness in SCA, regarding a patient with SCA: to provide analgesic therapy especially in the ED, has not been fully determined. The and to evaluate for the presence of coexisting and few studies available, typically comparing pulse oximetry potentially life-threatening illness. The patient with SCA to arterial blood gas and/or measured oxygen saturation, is at risk for a number of disease states. They are at high risk show conflicting results regarding the accuracy of pulse for infection with encapsulated organisms and are thus more oximetry in determining oxygenation in the sickle cell susceptible to pneumonia [2]. More ominous is acute chest population [4-6]. Further, the vast majority of the studies have been done in children and in either an inpatient or an outpatient setting; the ED literature is deficient. In the only Presented at the ACEP Research Forum, October 2003. T Corresponding author. Tel.: +1 215 955 6844; fax: +1 215 923 6225. ED study done to date, we found that pulse oximetry E-mail address: [email protected] (B.L. Lopez). underestimated arterial oxygenation in 13 adult ED subjects
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.09.035 430 B.L. Lopez et al. presenting with uncomplicated VOC [7]. No studies have Table 2 Chest radiographs examined its actual use in the ED setting. The purpose of b 2 this study was to examine the association between pulse NORMAL (n = 47) LOW (n = 19) P .05, v oximetry and ED disposition in VOC through a retrospec- Chest pain (n = 44) Chest pain (n = 15) tive chart review. We hypothesize that a categorical range Cough (n = 2) Cough (n = 2) of pulse oximetry reading is not associated with differences Shortness of Shortness of in the disposition of the adult patient with VOC presenting breath (n = 1) breath (n = 2) to the ED. For the purposes of comparison, subjects were divided into 2 groups—a NORMAL group (pulse oximetry z95%) and a 2. Methods LOW group (pulse oximetry b95%). Ninety-five percent was chosen as the cutoff based upon the fact that it This study was approved by the Institutional Review represents a Pao of 75 mm Hg, considered by many to be Board. We conducted a retrospective chart review of all 2 the lower end of normal. The groups were compared for subjects who presented to an urban university ED from differences in historical/physical examination data, labora- March through October 2003 with either an initial tory testing, radiographs, and final ED disposition. Cate- complaint of or a final ED diagnosis of VOC. Included gorical data were analyzed using v2 testing; continuous data were those 18 years or older with at least one chief were analyzed with the t test. Continuous data were complaint of VOC and who had a pulse oximetry reading presented as mean F SD. obtained in the ED. Subjects with sickle cell heterozygous SS, homozygous SC, and b thalassemia disease as docu- mented by prior hemoglobin electrophoresis in our SCA 3. Results center were eligible. Excluded were those younger than 18 years. Two hundred ten consecutive charts representing all To assure capture of consecutive patients with VOC, patients with sickle cell who presented to the ED met subjects were identified retrospectively on a daily basis. The inclusion/exclusion criteria and were reviewed. All subjects research nurse or assistant reviewed both the daily ED enrolled were patients of our Sickle Cell Center. The mean computer log as well as the ED charts for the preceding age of the group was 34.9 F 1.1 years; 114 were men and 24 hours for sickle cell disease and VOC. In addition, the 96 were women. names of commonly treated patients with sickle cell In the NORMAL group (n = 163), the mean pulse (identified from a list of patients with SCA treated at our oximetry was 97.87% F 1.7%; the LOW group (n = 47) institution—we are a National Institutes of Health desig- was 91.55% F 2.44%. There was no significant difference nated regional sickle cell center with a regular cohort of between NORMAL and LOW groups with regards to 150 patients) were searched. Charts were reviewed to patient disposition (Table 1, P N .05). There was no determine if the subject presented with typical (defined by significant difference in historical and physical exam the subject as pain of typical severity and location based characteristics between groups. upon prior episodes) or atypical (different severity and/or A higher percentage of subjects in the LOW group had a location or presence of a coexistent infectious illness) VOC. chest radiograph done when compared to the NORMAL Because of our designation as an NIH Regional Sickle Cell group (19 [47%] vs 47 [29%], P b .05) (Table 2). All 47 in Center, we routinely document if a patient with sickle cell the NORMAL group and all but one in the LOW group has typical (defined by the patient based upon their past were read as bno new infiltrate.Q The one subject with the experiences on severity and location) or atypical pain new infiltrate presented with atypical VOC accompanied by (defined as pain of different severity, location, or duration). chest pain and shortness of breath had a temperature of Because of our documentation practices with regards to 100.58F, a heart rate of 123, a pulse oximetry reading of patients with sickle cell, if an item was not recorded, it was 86% (the lowest reading of our cohort), and was admitted considered normal. with a final ED diagnosis of sickle cell crisis with possible acute chest syndrome. The most common blood tests ordered in the evaluation of the patient with sickle cell in Table 1 Disposition by pulse oximetry category the ED are the complete blood count and reticulocyte count. NORMAL LOW In our study, 87 (53.3%) in the NORMAL and 26 (55.3%) (z95%) (b95%) in the LOW group had a complete blood count/reticulocyte N Admitted (n = 88) 67 21 count drawn ( P .05), illustrating that no significant Discharged (n = 120) 94 26 association with pulse oximetry level and blood testing was Left during treatment (n = 2) 2 0 noted. There was a nonsignificant trend toward lower hemoglobin levels in the LOW group (8.75 F 2 g/dL P N .05, v2. NORMAL vs 7.92 F 2.2 g/dL LOW, P = .07; t test). Pulse oximetry in the adult ED patient with sickle cell 431
Most of the patients (196/210) presented to the ED with shift of the oxyhemoglobin dissociation curve [12] com- typical VOC. The other 14 presented with atypical VOC monly found in sickle cell disease may also alter the alone or with a coexistent illness. Interestingly, 11 of these estimation of oxygenation. were categorized in the NORMAL pulse oximetry group. The literature on the use of pulse oximetry in SCA is not definitive. Fitzgerald and Johnson [4] measured initial pulse oximetry readings in pediatric ED and intensive care unit 4. Discussion patients in a children’s hospital and found correlation with cooximeter-measured saturation in patients with sickle cell. Our study suggests that categorical ranges of pulse They found that pulse oximetry underestimated arterial O2 oximetry do not have a significant association with the saturation but stated that the bias found was clinically disposition of adult patients with SCA in the ED. It would insignificant and concluded that pulse oximetry could be seem intuitive that a low pulse oximetry reading would be used as a reliable estimate. Blaidsell et al [6] simultaneously suggestive of and associated with either more severe disease measured pulse oximetry and cooximetry in outpatient sickle or the presence of a coexistent illness (and hence, a higher cell children. They found that pulse oximetry in this setting amount of admissions to the hospital). Conversely, it would had poor specificity in detecting hypoxemia and commented seem likewise intuitive that a high pulse oximetry reading that bmany more patients would have been considered might suggest either a less severe presentation or an absence hypoxic based on pulse oximetry measurements alone than of a coexistent illness (and thus, a higher likelihood of being were hypoxic enough to require intervention.Q The usefulness discharged from the ED). We found neither to be the case. In of pulse oximetry in the adult ED patient with sickle cell is our study, a low pulse oximetry reading was not associated unclear. In our previous study, we found, similar to the with higher admission rate, nor was a high pulse associated Blaidsell study, that pulse oximetry underestimated oxygen- with a higher likelihood of being discharged from the ED. In ation in acute VOC. In contrast to these other studies, this addition, an atypical presentation (with or without coexis- study aimed not to determine correlation of pulse oximetry to tent illness) was not associated with lower pulse oximetry as oxygenation, but rather, attempted to examine pulse oximetry 11 of the 14 had normal readings. These findings help to readings as a predictor of disposition in the clinical setting. confirm our observation over the years that the pulse Our study has a number of limitations due mostly to its oximetry reading in adult VOC may not be a reliable retrospective nature. We are unable to determine the true indicator of ED disposition. effect of the pulse oximetry reading on test ordering, Interestingly, there was a higher percentage of subjects in decision-making, and disposition. Rather, we could only the LOW group that received chest radiographs. Given the make the associations or lack thereof between oxygen potential for a pulmonary process as a cause for hypoxia, saturation and the various outcomes. Inadequate documen- this is not surprising and might suggest that the pulse tation may have affected the conclusions. This study may oximetry reading may have influenced diagnostic testing. also reflect a selection bias of a subgroup of patients with However, when viewed in another manner, there was a sickle cell who frequent the ED, and thus our findings may significant number in the LOW group that did not have a not be generalizable to all adult ED patients. Despite these, chest radiograph done, suggesting that pulse oximetry was and given the paucity of literature in this setting, our study not always a determinant. Only one patient in our study had appears to shed more light on the actual use of pulse an abnormal chest radiograph—this patient had the lowest oximetry in the adult patient with sickle cell. pulse oximetry reading of our cohort and had presenting In conclusion, we found that the determination of a complaints beyond her typical sickle cell VOC. Laboratory normal or abnormal pulse oximetry reading in the adult testing did not appear to be affected by pulse oximetry ED patient with SCA was not associated with a particular reading as there were equal percentages of patients in the patient disposition. NORMAL and LOW groups having a complete blood count/reticulocyte count drawn. Pulse oximeters measure arterial hemoglobin saturation and pulse rate by positioning a pulsatile arterial vascular bed References between red and infrared light sources and a detector, and detects the differential absorption of light by oxy- and [1] Steinberg MH. Management of sickle cell disease. N Engl J Med 1999;340(13):1021-30. deoxyhemoglobin. A microprocessor then converts this [2] Ballas SK, Park CH. Severe hypoxemia secondary to acute sternal detected light into an arterial blood saturation reading infarction in sickle cell anemia. J Nucl Med 1991;32:1617-8. [8,9]. Although it should also work well in SCA, certain [3] Castro O, Brambilla DJ, Thorington B, et al. The acute chest characteristics of the disease (such as peripheral arteriolar syndrome in sickle cell disease: incidence and risk factors. The damage from multiple acute episodes of vascular occlusion cooperative study of sickle cell disease. Blood 1994;84:643-9. [4] Fitzgerald RK, Johnson A. Pulse oximetry in sickle cell anemia. Crit over the years, the increased presence of met- and Care Med 2001;29:1803-6. carboxyhemoglobin) [10,11] may prevent it from being a [5] Needleman JP, Setty BN, Varlotta L, Dampier C, Allen JL. reliable predictor of hypoxemia. In addition, the rightward Measurement of hemoglobin saturation by oxygen in children and 432 B.L. Lopez et al.
adolescents with sickle cell disease. Pediatr Pulmonol 1999;28: [9] Scnapp LM, Cohen NH. Pulse oximetry: uses and abuses. Chest 423-8. 1990;98:1244-50. [6] Blaidsell CJ, Goodman G, Clark K, Casella JF, Loughlin GM. Pulse [10] Craft JA, Alessandrini E, Kenney LB, Klein B, Bray G, Luban NLC, oximetry is a poor predictor of hypoxemia in stable children with et al. Comparison of oxygenation measurements in pediatric patients sickle cell disease. Arch Pediatr Adolesc Med 2000;154:900-3. during sickle cell crisis. J Pediatr 1994;124:93-5. [7] Comber JT, Lopez BL. Evaluation of pulse oximetry in sickle cell [11] Ortiz FO, Aldrich TK, Nagel RL, et al. Accuracy of pulse oximetry in anemia patients presenting to the emergency department in acute sickle cell disease. Am J Respir Crit Care Med 1999;159:447-51. vasoocclusive crisis. Am J Emerg Med 1996;14:16-8. [12] Rackoff WR, Kunkel N, Silber JH, et al. Pulse oximetry and factors [8] Tremper KK, Barker SJ. Pulse oximetry. Anesthesiology 1989;70: associated with hemoglobin oxygen desaturation in children with 98-108. sickle cell disease. Blood 1993;81:3422-7. American Journal of Emergency Medicine (2005) 23, 433–438
www.elsevier.com/locate/ajem
Ultrasound image transmission via camera phones for overreading
Michael Blaivas MD, RDMS*, Matthew Lyon MD, RDMS, Sandeep Duggal MD, RDMS
Department of Emergency Medicine, Medical College of Georgia, Augusta, GA 30912-4007, USA
Received 29 July 2004; revised 8 September 2004; accepted 10 September 2004
Abstract Emergency physicians using ultrasound frequently encounter unfamiliar findings during routine ultrasound examination. This is especially common for less experienced practitioners. Objective: To compare high-resolution thermal printer ultrasound images and images recorded and transmitted via commercial camera cell phones. Methods: This was a study comparing randomly selected images of actual ultrasound examinations performed in an academic level I ED with hospital-based emergency ultrasound credentialing. Two hospital credentialed emergency sonologists with extensive experience were asked to review 50 randomly selected images from actual patients as seen on a camera cell phone screen after being captured from a high-resolution thermal printout and sent to a similar phone. Reviewers recorded initial impression of the image and identified structures, measurements, and pathology. After hearing a brief clinical vignette, reviewer rated the images for image quality, detail, resolution, as previously defined, on a 10-point Likert scale. This process was then repeated with the original thermal printouts. Data were analyzed using descriptive statistics, agreement analysis, and the Student t test. Results: Reviewers showed good interrater agreement for pathology and structure detection between phone and thermal printer images. There was no statistically significant difference in image quality, resolution, and detail between phone images and thermal printer images. However, there was statistically significantly increase in confidence in diagnosis for reviewers when using thermal printer images as compared with phone images, P = .003 and P = .02. Several phone images were felt to be suboptimal, and there was moderate agreement on these between reviewers. Conclusions: Ultrasound pictures recorded by one phone and then sent to another yielded images that showed no statistically significant differences from traditional high-resolution thermal printouts in image quality, detail, and resolution. Measurements were too small to be read on the camera phones, and reviewers had statistically significantly lower confidence in their diagnosis when using the camera phones to review images. D 2005 Elsevier Inc. All rights reserved.
1. Introduction
T Corresponding author. Tel.: +1 706 721 4467. Ultrasound use by emergency physicians can has many E-mail address: [email protected] (M. Blaivas). benefits for patient care, satisfaction, department length of
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.09.037 434 M. Blaivas et al. stay, and group finances [1-4]. However, the speed of its 2 camera-equipped cellular phones. The study was approved spread is limited to some degree by the lack of experienced by the institutional review board. This study took place at a emergency sonologists. Emergency sonologists may be level 1 trauma center ED with an accredited emergency trained in residency or fellowships, and this training often medicine residency program and an active emergency continues once in practice. Others work to obtain ultrasound ultrasound program. Emergency ultrasound credentialing skills when already in practice through continuing education is available through the hospital, allowing performance and and self-education efforts. Unfortunately, novice ultrasound interpretation of ultrasound examinations. Currently, the users are unfamiliar with the pitfalls of emergency department is in transition to a fully credentialed faculty. ultrasound and may feel uncomfortable performing ultra- Ultrasound services are available from the department of sound examinations without oversight from a more experi- radiology during business hours and are not available during enced emergency sonologist. Novice emergency sonologists most evenings and nights as well as weekends. are often able to obtain adequate images long before they Emergency ultrasound education is mandatory for all are familiar with the range of pathology and normal faculty and residents, and biannual courses, monthly anatomy, thus limiting their ability to accurately diagnose lectures, and quality review sessions, as well as individual or rule out typical pathology. instruction, are available. Diagnostic ultrasound services Such limitations are usually not a factor when the ED has provided by the ED include pelvic, focused cardiac, lower a more experienced emergency sonologist, such as a extremity venous, biliary, renal, abdominal aortic, trauma, department designated ultrasound director, who is able to procedural, and superficial structure ultrasound. Creden- preside over the ultrasound examinations by the novice user. tialed faculty issue a written ultrasound report based on Such review can also take place after the ultrasound several preprinted forms. The report contains examination examination has been completed by reviewing thermal findings and diagnosis, as well as high-resolution thermal printer images on a report and thus providing feedback to print images on the side of the report. Typically, 3 images the novice user [5]. This occurs commonly with attending are added to the report unless more are called for to illustrate radiologist overreading of resident scans 1 or 2 days after a pathology. The report then becomes a part of the medical night or weekend shift in many academic institutions [6]. record and is scanned into an electronic medical record However, although this delayed review may work for system accessible throughout the medical campus. general radiology, the problems with having a delay in the Fifty images were randomly selected from approximately expert reading are magnified in emergency ultrasound. In 200 ultrasound examinations during a 1-month period. Only most cases, emergency ultrasound is used to evaluate time- 1 image was randomly picked from each examination. sensitive emergency conditions, and a misinterpreted Studies performed by either of the 2 reviewers were ultrasound can be disastrous. Ideally, oversight should be disallowed to avoid unfair bias. All of the studies selected performed in real-time or at least before the completion of had thermal printed images to document study results. the ultrasound examination so that additional images could Randomly selected images were then graded by a hospital still be obtained and the diagnosis clarified or corrected. credentialed sonologist with fellowship training and ap- Radiologists have explored the use of telemedicine proximately 10 years’ experience. systems allowing them to review images and render a diag- The premise of this study was to duplicate a scenario nosis from remote locations such as the physicians home at where an emergency ultrasound provider has performed an night [7-9]. However, the cost of such transmission systems ultrasound examination and has questions about the findings can be prohibitive to most EDs [10]. The solution may lie in or is required to have study conclusions confirmed by the the rapidly developing information technology that is emergency ultrasound director or his/her designate. The available to the general population and is often marketed to provider then takes images of the same thermal ultrasound young cell phone users in a recreational or entertainment prints used for examination documentation in the medical fashion. Specifically, the cameras available on many new record with a camera-equipped cellular phone and transmits cellular phones may be of use for transmitting images for the images to the designated reviewer. The designated image rapid review [11]. We sought to compare the quality of reviewer can be anywhere that a cellular connection is ultrasound images of normal anatomy and pathology from available. Depending on the particular cellular service, this actual patients as represented on standard high-resolution can mean in much of North America and the rest of the thermal paper with those photographed by one camera cell industrialized world, thus not limiting the reviewer to an phone and sent to another one for review on the screen. internet connection and computer for reviewing. We used a commonly available camera phone produced by Motorola (Schaumburg, IL). The phone, a V400, contains 2. Methods a digital camera capable of capturing a 640 Â 480 pixel image in color, Fig. 1. The image can then be transmitted to This was a study with comparison of ultrasound images another phone or e-mail account. Images were captured in printed on a standard high-resolution thermal printer the ED with overhead fluorescent lighting to simulate the compared with images photographed with and sent between less-than-ideal lighting conditions that are likely to be Camera phone ultrasound 435
education, and research and publication were asked to evaluate the images. For each phone ultrasound image, the reviewers were initially asked if they could tell what the image was without any clinical data. A limited clinical vignette was then read to each of the 2 reviewers describing patient complaint and area of the body being scanned by the less experienced sonologist. For example, the reviewers may be told that a 45-year-old woman with 2 days of right upper quadrant pain that has now worsened was being scanned for gallbladder pathology. Reviewers performed the evaluation in isolation from each other and could not compare notes or discuss the patient or image. Each reviewer was allowed to manipulate the camera phone as they wished, tilting and angling the LCD screen to maximize image quality. Based on previous studies that compared ultrasound images, each one was graded for the following [12,13]: (1) total image quality, defined as an overall assessment encompassing contrast of solid and fluid-filled structures, and absence of noise; (2) image resolution, defined as the sharpness and crispness of the image and a lack of haziness/ blurriness; (3) image detail, which was defined as the clarity of organ outlines and ease with which boundaries of structures are seen and how well they are defined. Each reviewer ranked their impression on a 10-point Likert scale with 1 being the worst and 10 the best. Reviewers were also asked if any pathology was seen in the image, if any measurements were present and what they were, and if a diagnosis could be given and to list major visible structures. Finally, each reviewer was asked if the image being viewed was either suboptimal for review or contained image artifacts other than expected from ultrasound. Two weeks later, this review process was repeated for the thermal printer images, or originals, which were viewed in random order, Fig. 2. Emergency physicians (EPs) filled out the standardized data Fig. 1 A cell phone with an image of a heart in the subxiphoid form, which was then entered into a computerized database. view is being held after the ultrasound image was captured and sent Statistical analysis included descriptive statistics and a paired between phones. Note in this figure that the image on the screen is too distant to be seen clearly. Some glare is seen on the right side of the cell phone screen. present in real-life use of this application. Image upload speed is approximately 14000 bytes per second, and download speed is approximately 9000 bytes per second. Patient names, present on all formally issued reports, were covered with a black sheet of paper. The images were captured and stored in the camera phone at one time. Images were specifically captured to fit the entire screen on the phone. No glare-eliminating glass, special lighting, lenses, or filters were allowed. The camera operator was allowed to angle the phone and cup his hands over the camera lens to optimize the images. Each image was then sent to a second identical phone. The time to send 1 image and 3 images was recorded and averaged. One month after image selection, 2 credentialed emer- Fig. 2 A thermal print image used for comparison during the gency sonologists with extensive ultrasound experience second phase of the study of the cardiac view seen in Fig. 1 on the including study performance, interpretation, ultrasound phone screen. 436 M. Blaivas et al.
of the cell phone image had good agreement between the 2 Table 1 Distribution of emergency ultrasound study types reviewers with j = 0.55. Agreement for initial impression of captured the hard copy image between reviewers was very good with Study type No. of images and j = 0.83. Agreement for initial impression by each reviewer percent of total between phone and printer images was good with j = 0.76 Pelvic 14 (28%) and 0.69. Agreement for presence of pathology on phone Biliary 13 (26%) images between the 2 reviewers was very good at j = 0.79. Renal 4 (8%) Agreement for presence of pathology between the 2 Trauma 4 (8%) reviewers for printer images was also very good at j = 0.8. Lower extremity r/o deep vein thrombosis 5 (10%) Ocular 2 (4%) Agreement for each reviewer between phone and paper Cardiac 3 (6%) images was good with j = 0.59 and 0.67. Abdominal aorta 4 (8%) There was a statistical difference between Likert scale Peritonsillar abscess localization 1 (2%) ranking for confidence in cell phone image diagnosis between the reviewers with P = .01. There is no statistical difference between the Likert scale ranking for confidence t test with 95% confidence intervals. Data were analyzed of diagnosis between the 2 reviewers for printer images, using statistical calculators from a commercially available P = .29. However, statistically significant differences software package, StatsDirect. (Cheshire, United Kingdom). existed for each reviewer’s confidence between phone and hard copy images, P = .003 and P = .02. The differences between phone and printer image quality scores were not 3. Results statistically significant for the 2 reviewers, P = .56 and P = .12. The differences between phone and printer image A total of 50 images were randomly selected from 200 quality were not significant for each reviewer, P = .47 and studies. Of the 50 randomly chosen, 26 contained pathology. P = .36. The differences between phone and printer image Distribution of ultrasound study types represented by the resolution scores were not statistically significant for the 50 images is shown in Table 1. Five attempts were made at 2 reviewers, P = .45 and P = .45. The differences between transmitting 1 image of 44 kilobytes in size from one phone phone and printer image resolution was not significant for to another. The median time was 1 minute 38 seconds with a each reviewer, P = .35 and P = .41. The differences between range of 1 minute 35 seconds to 2 minutes 43 seconds. phone and printer image detail scores were not statistically Similarly, 5 attempts were made to transmit 3 images at a significant for the 2 reviewers, P = .18 and P = .43. The time. This is a typical number of hard copy images attached differences between phone and printer image detail was not to a formal medical record report in our ED. The median significant for each reviewer, P = .31 and P = .5. time was 3 minutes 12 seconds with a range of 3 minutes 4 seconds to 3 minutes 18 seconds. No transmission failures were noted; however, failure can occur if the cellular signal 4. Discussion is lost during transmission. Results for image quality scores and confidence in Information technology invaded the house of medicine diagnosis by the reviewers are given in Table 2. There approximately a decade ago, but it was not until the last was moderate agreement among the reviewers about few years that most physicians felt its impact in their daily suboptimal images with a j = 0.47. All of the 6 suboptimal practice. Although initially relegated to limited computer images were produced by the camera phone. None of the systems in hospital accounting and supply departments, hard copy images were felt to be suboptimal. No measure- modern information technology is likely impacting many of ments were missed on any image. However, cell phone us at this time. Patient medical records are frequently image measurements were too small for reviewers to read in available for rapid review on computer screens as well as all cases. Reviewers showed good agreement in listing laboratory results for previous visits to the ED. Patients are identifiable structures with j = 0.67 and 0.6 between cell being registered at bedside in their room with remotely phone and hard copy images, respectively. Initial impression linked computers. Specially designed computer programs
Table 2 10-Point Likert scale scores for the reviewers Phone Thermal print Phone image Thermal Phone image Thermal Phone image Thermal confidence confidence quality print image resolution print image detail print detail quality resolution Reviewer 1 6.5 7.8 7.8 7.5 7.9 7.8 7.7 7.8 Reviewer 2 7.3 8.0 7.9 7.8 7.9 7.8 7.9 7.9 Camera phone ultrasound 437 are helping triage patients into appropriate categories. (110000-pixel). This resolution is considerably less than Others write portions of patients’ charts automatically from what was used in our study. triage data and even recognize the human voice for Our study results indicate that, although images are not dictation. It appears that many of the information technol- the same on a camera cell phone as a high-resolution ogy innovations implemented in EDs have been driven by thermal prints, reviewers were still able to accurately the necessity to be more efficient and provide better care to interpret the images. In fact, there was no statistically our patients. Similarly, emergency ultrasound’s evolution significant difference between image quality, resolution, has been driven by the desire to improve patient service and detail scores between the 2 image types on a 10-point and care. Likert scale. On the other hand, the reviewers were less Emergency ultrasound is a relatively new field in clinical confident in the interpretation of the image from cell medicine. Borne out of the need for immediate diagnosis for phones as compared with thermal prints suggesting that the many emergency conditions, emergency ultrasound fills a 2 are not identical in quality. There was good agreement by void caused by dwindling resources including ultrasound reviewers for identification of important structures between technologists. However, despite the proven utility and image types, as well as among themselves. Another accuracy of emergency ultrasound for many applications, important drawback to our cell phone images was the emergency ultrasound is a technical skill that must be inability of reviewers to read measurements which appeared acquired over time. Traditionally, supervision in sonography in too small a print on the image. However, this problem is involves direct review of the sonologist’s images with easily solved in the real-world application by reading all verification of the findings, clarification of uncertainties, measurements off when discussing the case with a cell and the provision of ongoing feedback for improvement phone reviewer. [14]. In emergency ultrasound, this supervision includes This type of system can have a significant impact in the hands-on instruction, review of previously performed practice of emergency medicine. As emergency ultrasound ultrasound examinations, and didactic teaching. Typically, becomes more widespread, real-time feedback on pathology this service is provided by the expert in the department, and diagnosis can be provided easily by those more usually the emergency ultrasound director. experienced in the field. This system can reduce medical Teleradiology is used in many locations around the errors by encouraging the use of emergency ultrasound world to review ultrasound examinations performed by which has been shown to prevent unnecessary treatment providers unable to render adequate interpretation of the [16]. In addition, the novice user can be provided with real- acquired images. Typically, static or dynamic images of time quality assurance from the director of ultrasound and examinations performed at remote locations are transmitted provide ongoing education when there is diagnostic electronically via network connections to the radiologist for uncertainty. This is not only of utility in the review of interpretation. However, the systems and system support for images generated by attending physicians but may be a teleradiology can be cost-prohibitive for emergency ultra- great way to overread ultrasound examinations performed sound. Estimates for the cost of acquiring the equipment by residents. only for transmitting medical images of sufficient quality to The camera cell phone offers other benefits to its users. perform medical diagnostics vary in the literature, depend- The phone can send the pictures to several phone and e-mail ing on the bandwidth used, from US$25000 to US$55000 addresses simultaneously. Therefore, the ED could have an per site, in addition to transmission and maintenance costs bEDQAQ address for the pictures sent. Hence, a database is [10,11]. However with the recent advances in mobile phone built simultaneously with the quality assurance process. systems and cellular phone networks, it may be possible to This database can also be used for comparison review on transmit high-quality medical images over cellular phone subsequent ultrasound examinations not only those done in connections. Currently, newest generation cellular camera the ED, but also on follow-up examinations done by phones have a resolution of 640 Â 480 pixels. Prior studies radiology. This is advantageous as some emergency have shown that the required resolution for a medical image ultrasound examinations may require follow-up evaluations to be diagnostic depends on the image being interpreted. at some point in the future. Mammography, for instance, requires higher resolution for The initial investment for the equipment used was diagnostic purposes than computed tomography which US$200 to US$300 per camera phone, and the communi- typically uses a 512 Â 512 matrix at an 8-bit gray level. cation charge for the image was approximately US$0.25 per However, compared with analog images, 80% of viewers image. Image transfer took approximately 1.5 minutes on found images with resolution as low a s 530 Â 400 to be average for a single image and 3.5 for 3 images. In addition, indistinguishable [15]. In a study by Yamada et al [11], the confidentiality of the image data and patient information using images obtained by a cell phone camera, computed is protected by the server in the same manner as is e-mail tomography scans and magnetic resonance imaging between personal computers. However, if this method of scans were of sufficiently high quality for interpretation. security for data transfer is unacceptable, patient names and In this study, the cell phone camera used had a display personal information are easily left off of the images. Hence, size of 31 Â 30 mm and only 120 Â 128 pixel resolution there is no personal information associated with the image. 438 M. Blaivas et al.
This study has several limitations including the potential [4] Hilty WM, Hudson PA, Levitt MA, Hall JB. Real-time ultrasound- bias of ultrasound image reviewers against the images on guided femoral vein catheterization during cardiopulmonary resusci- the cellular phones that are smaller and more difficult to tation. Ann Emerg Med 1997;29:331-6. [5] Tayal V, Blaivas M, et al. American College of Emergency Physicians. interpret. Although actual ultrasound images were used ACEP emergency ultrasound guidelines—2001. Ann Emerg Med from diagnostic ultrasound examinations on ED patients, 2001;38:470-81. we did not actually evaluate rendering diagnosis from a [6] Lowe RA, Abbuhl SB, Baumritter A, Brensinger C, Propert K, cellular phone on real-time patients. This is a subject for Horii S, et al. Radiology services in emergency medicine residency further study now that such review appears possible. Fifty- programs: a national survey. Acad Emerg Med 2002;9:587-94. [7] McCredie M, Bell J, Lee A, Rogers J. Differences in patterns of care four percent of images randomly chosen contained some of prostate cancer, New South Wales. Aust N Z J Surg 1996;66: kind of pathology. This may not represent a common 727-30. distribution. The spectrum of studies included may not [8] Tually P, Walker J, Cowell S. The effect of nuclear medicine represent the spectrum found at other facilities using telediagnosis on diagnostic pathways and management in rural and ultrasound in the ED. remote regions of Western Australia. J Telemed Telecare 2001;7 (Suppl 2):S2:50-S3. Real-time image transfer of diagnostic emergency [9] Calder LD, Maclean JR, Bayliss AP, Gilbert FJ, Grant AM. The ultrasound examinations is possible using existing cell diagnostic performance of a PC-based teleradiology link. Clin Radiol phone camera equipment and technology. This system is 1999;54:659-64. inexpensive compared with traditional teleradiology sys- [10] Malone FD, Athanassiou A, Nores J, D’Alton ME. Effect of ISDN tems. Ultrasound pictures recorded by one phone and then bandwidth on image quality for telemedicine transmission of obstetric ultrasonography. Telemed J 1998;42:161-5. sent to another yielded images that showed no statistically [11] Yamada M, Watarai H, Andou T, Sakai N. Emergency image transfer significant differences with traditional high-resolution ther- system through a mobile telephone in Japan: technical note. mal printouts in image quality, detail, and resolution. Neurosurgery 2003;52:986-8. Measurements were too small to be read on the camera [12] Blaivas M, DeBehnke D, Sierzenski PR, Phelan MB. Tissue harmonic phones, and reviewers had statistically significantly lower imaging improves organ visualization in trauma ultrasound when compared with standard ultrasound mode. Acad Emerg Med 2002;9: confidence in their diagnosis when using the camera phones 48-53. to review images. [13] Shapiro RS, Wagreish J, Parsons RB, Stancato-Pasik A, Yeh HC, Lao R. Tissue harmonic imaging sonography: evaluation of image quality compared with conventional sonography. AJR Am J Roent- References genol 1998;171:1203-6. [14] Johnson MA, Davis P, McEwan AJ, Jhangri GS, Warshawski R, [1] Durston W, Carl ML, Guerra W. Patient satisfaction and diagnostic Gargum A, et al. Preliminary findings from a teleultrasound study in accuracy with ultrasound by emergency physicians. Am J Emerg Med Alberta. Telemed J 1998;4:267-76. 1999;17:642-6. [15] Schellingerhout D, Chew F, Mullins M, Gonzalez G. Projected digital [2] Blaivas M, Sierzenski P, Plecque D, Lambert M. Do emergency radiologic images for teaching. Acad Radiol 2002;9:157-62. physicians save time when locating a live intrauterine pregnancy with [16] Blaivas M, DeBehnke D, Phelan MB. Potential errors in the diagnosis bedside ultrasonography? Acad Emerg Med 2000;7:988-93. of pericardial effusion on trauma ultrasound for penetrating injuries. [3] American College of Emergency Physicians. Emergency ultrasound Acad Emerg Med 2000;7:1261-6. coding and reimbursement. ACEP website 2002. American Journal of Emergency Medicine (2005) 23, 439–442
www.elsevier.com/locate/ajem
The inadvertent administration of anticoagulants to ED patients ultimately diagnosed with thoracic aortic dissection
Daniel P. Davis MDa,*, Karun Grossman MDb, Danielle C. Kigginsa, Gary M. Vilke MDa, Theodore C. Chan MDa aDepartment of Emergency Medicine, University of California, San Diego, CA 92103-8676, USA bSchool of Medicine, University of California, San Diego, CA 92103-8676, USA
Received 23 September 2004; accepted 13 October 2004
Abstract Objectives: Aortic dissection (AD) may present similarly to acute coronary syndrome or pulmonary embolus; however, anticoagulation may be detrimental to patients with AD. Methods: Clinical data were abstracted from medical records of emergency department (ED) patients with nontraumatic AD. Patients administered with anticoagulants were compared with non–anti- coagulated patients with regard to presenting symptoms, chest radiograph and electrocardiogram (ECG) findings, and outcome. Results: A total of 44 ED patients with nontraumatic AD was identified over a 4-year period; anticoagulants were administered to 9 (21%). Anticoagulated patients had a higher incidence of chest pain without back pain (78% vs 23%; P = .002) and ST elevations or depressions on ECG (89% vs 6%; P b .001) and were less likely to have a widened mediastinum on chest radiograph (0% vs 67%; P b .001). Two ED anticoagulated patients died, one required a second surgery for bleeding complications, and another suffered a stroke after reversal of anticoagulation. Conclusions: There is a clinically significant incidence of anticoagulation administration to ED patients ultimately diagnosed with AD, especially in the presence of ambiguous ECG and radiographic findings. D 2005 Elsevier Inc. All rights reserved.
1. Introduction (ED), making chest pain the second most common ED complaint [1]. Although most protocols are oriented toward Each year, nearly 4.6 million patients present with the rapid identification and treatment of acute coronary nontraumatic chest pain in the emergency department syndrome, other potentially life-threatening diagnoses in- cluding pulmonary embolus and aortic dissection (AD) must be considered. Although anticoagulation of patients with Presented at the ACEP National Meeting 2003, Boston, Mass. either acute coronary syndrome or pulmonary embolus can T Corresponding author. be life saving, the administration of anticoagulants to E-mail address: [email protected] (D.P. Davis). patients later diagnosed with AD may have catastrophic
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.10.010 440 D.P. Davis et al. consequences such as aortic wall rupture and bleeding StatsDirect (StatsDirect Software Inc, Ashwell, UK). Sta- complications during surgery. tistical significance was assumed for P b .05. Current screening standards for AD are neither sensitive nor specific, with postmortem diagnoses made in up to 30% of cases. A widened mediastinum on plain chest 3. Results radiograph is commonly used as a screening strategy; however, chest radiographs are prospectively interpreted as A total of 44 ED patients was identified with non- suggestive of the diagnosis in only 25% of AD cases [2,3]. traumatic AD over the study period; 9 of them (20.5%) Retrospective review by chest radiologists improved the received anticoagulation (6 heparin/aspirin, 1 fibrinolytics, sensitivity to 48% [4]. This leads to the potential for 2 aspirin). Table 1 defines the predominant presenting inadvertent administration of anticoagulation to patients symptom as chest pain (63%), followed by back pain (43%) with AD. We review our experience with ED patients and shortness of breath (34%). About half of all patients ultimately diagnosed with nontraumatic AD, focusing on (57%) presented with a systolic blood pressure of at least the reasons and consequences of anticoagulation in this 140 mm Hg or greater. Demographic data, past medical patient group. history, and diagnostic studies are also included in Table 1. The 9 anticoagulated patients were more likely than those not receiving anticoagulation to have chest pain without 2. Methods 2.1. Design Table 1 Clinical presentation for all patients Parameter % 95% CI We performed a retrospective cohort analysis using medical records from 1997 to 2000. Approval was obtained Demographics from our institutional investigational review board. Male sex (%) 77 59-89 Age (y) 59 54-65 2.2. Setting Past medical history Hypertension 86 70-95 This study was conducted in a university-affiliated urban Marfan syndrome 3 0-15 ED with an annual census of approximately 45000 visits. Prior aortic disease 14 5-30 2.3. Subjects Presenting symptoms Chest pain 63 45-79 Patients were identified using a hospital discharge Back pain 43 26-61 diagnosis of nontraumatic AD. Patients were excluded for Syncope 11 3-27 death immediately upon arrival in the ED or incomplete Shortness of breath 34 19-52 medical records. Neurological 14 5-30 Systolic blood pressure 2.4. Data collection z140 mm Hg 57 39-74 100-139 mm Hg 20 8-37 The following data were abstracted from the ED records: b100 mm Hg 16 7-30 age, sex, past medical history, presenting chief complaint, Not specified 7 1-19 triage vital signs, and ED course including medications Chest radiograph administered. Diagnostic studies, formal chest radiograph Wide mediastinum 55 39-70 and electrocardiogram (ECG) interpretations, hospital Tortuous aorta 9 3-22 course, complications, and final outcome were obtained Infiltrate/atelectasis 9 3-22 from the hospital medical records. Cardiomegaly 7 1-19 2.5. Data analysis Normal 20 10-35 Electrocardiogram Data regarding demographics, past medical history, and ST elevation or depression 20 10-35 chief complaint were analyzed descriptively. Aortic dissec- Nonspecific T-wave abnormalities 20 10-35 tion patients with ED administration of anticoagulants were Sinus bradycardia 9 3-22 identified and compared with all other AD patients with Sinus tachycardia 7 1-19 regard to patient demographics, presenting complaint, vital Normal 23 11-38 signs, chest radiograph and ECG results, hospital course, Diagnostic imaging modality and outcome. Anticoagulation was defined as one of the Chest computed tomography 68 52-81 following: fibrinolytic therapy, heparin, or antiplatelet Echocardiogram 27 15-43 2 therapy. v Analysis and t testingwereusedforall Aortogram 5 1-15 comparisons. Statistical calculations were performed using Anticoagulation in patients with aortic dissection 441
Table 2 Clinical data for patients with nontraumatic AD (comparing those receiving vs those not receiving anticoagulation in the ED) Parameter ED anticoagulation No ED anticoagulation P Chief complaint Chest pain 8/9 (89) 23/35 (66) .174 Back pain 1/9 (11) 15/35 (43) .077 Chest pain without back pain 7/9 (78) 8/35 (23) .002 Chest radiograph findings Widened mediastinum 0 (0) 22/33 (67) b.001 Electrocardiogram findings ST elevation or depression 8/9 (89) 2/31 (6) b.001 back pain, more likely to have an ST-segment elevation or use of anticoagulation therapy, which appeared to result in a depression, and less likely to have a widened mediastinum measurable incidence of complications. (Table 2). Several complications that were directly attribut- The timely diagnosis of AD appears to prevent compli- able to the administration of anticoagulation including a cations and improve prognosis, with an estimated increase stroke upon reversal of anticoagulation in one patient and in the death rate of 1% each hour that the disease goes multiple operations from bleeding complications in another undiagnosed [5,6]. The American College of Emergency were observed. Two other patients in the anticoagulation Physicians published a clinical policy for patients presenting group died from uncontrolled hemorrhage. Complications to the ED with chest pain, recommending the presence of for all patients are displayed in Table 3. particular symptoms along with the results of diagnostic studies to determine a causative etiology [1]. In light of the lack of sensitivity for clinical evaluation, plain chest 4. Discussion radiographs, or ECG in differentiating patients with AD from those with acute coronary syndrome, it is not We present our experience with the ED diagnosis of AD surprising that a significant percentage of patients with over a 4-year period. We observed a relatively high AD are inappropriately treated with anticoagulation therapy. incidence of inadvertent anticoagulation administration in The extent of this clinical problem has not been studied our patient group. The potential reasons for this include an previously. ambiguous clinical presentation (including chest pain The classic presentation of AD includes chest or back without back pain), the presence of ST-segment abnormal- pain, often described as btearing,Qbripping,Q or bknifelikeQ ities, and the absence of a widened mediastinum on chest [7]. In our study, patients with AD were more likely to radiograph. This likely resulted in the initiation of appro- complain of chest pain (63%) than back pain (43%), with priate therapy for acute coronary syndrome, including the previous investigations supporting these figures. The occurrence of chest pain on presentation has been described Table 3 Complications in anticoagulated and non–antico- in 69% to 75%, with back pain in 33% to 64% [8-11]. The agulated patients in the ED variability in the location of pain makes this difficult, Age Sex Anticoagulant Complication although our physicians appeared to use the presence of 46 M Fibrinolytics Died of bleeding chest pain without back pain as more consistent with acute complications before coronary syndrome, especially with a provocative ECG and operative repair normal mediastinal width. 61 M Heparin/ASA Died of bleeding The use of plain chest radiography to screen for AD can complications before be problematic. A widened mediastinum can be observed in operative repair up to 75% of patients with AD but can easily be confused 71 F Heparin/ASA Required repeat thoracotomy with widening in the ascending aorta, aortic arch, or the for bleeding complications descending portion of the thoracic aorta resulting from 59 M Heparin/ASA Suffered a stroke after aortic tortuosity [7]. The sensitivity of plain chest radiog- reversal of anticoagulation raphy for detecting AD is even lower if the real-time ED 69 M ASA Embolic stroke reading is used, with Armstrong et al [8] reporting a 68 M None Cardiac arrest during surgery sensitivity value of 64% for the treating physician’s 67 F None Cardiac arrest during surgery interpretation as compared with 72% for that of radiologists 81 M None Pericardial tamponade trained in aortic pathology. We used the final attending 20 M None Congestive heart failure radiologist’s interpretation in this study, with a wide mediastinum observed in 55% of all AD patients but none ASA indicates acetylsalicylic acid. of those administered with anticoagulation. More subtle 442 D.P. Davis et al. abnormalities may increase sensitivity but may not be 5. Conclusions detected reliably by the treating physician [12]. Despite its limited sensitivity and specificity, the use of chest radiog- The inadvertent administration of anticoagulation to ED raphy as a screening test for AD has continued because of its patients with a clinical presentation suggesting acute relatively low cost when compared with other modalities coronary syndrome is a problem in the management of such as chest computed tomography, transesophageal AD. Potential solutions to increase sensitivity to the echocardiography, and aortography [13]. diagnosis of AD likely come at the price of decreased Abnormalities in the ECG have long been described in specificity and delays to definitive therapy in patients with association with AD but are neither sensitive nor specific. acute coronary syndrome or pulmonary embolus, both of Nonspecific ECG abnormalities have been reported in up to which are far more common than AD. 80% of patients with AD [3,8,10,11,13]. The presence of ST-segment abnormalities can be especially problematic in patients with AD given the potential for misinterpretation of References the presentation as part of an acute coronary syndrome. The relatively high percentage of AD patients (1 in 5) [1] American College of Emergency Physicians. Clinical policy for the initial approach to adults presenting with a chief complaint of chest receiving anticoagulation is alarming, especially considering pain, with no history of trauma. Ann Emerg Med 1995;25:274-99. the potential to adversely affect outcome. On the other hand, [2] Gregario MC, Baumgartner FJ, Omari BO. The presenting chest these patients appear to have been treated appropriately, roentgenogram in acute type A aortic dissection: a multidisciplinary considering the presence of chest pain, ST-segment abnormal- study. Am Surg 2002;68(1):6-10. ities, andnormalmediastinal width onchestradiograph, which [3] Meszaros I, Morocz J, Szlavi J, Schmidt J, Tornoci L, Nagy L, et al. Epidemiology and clinicopathology of aortic dissection. Chest 2000; would lead clinicians to suspect the alternative and more 117(5):1271-8. common diagnoses of an acute coronary syndrome or [4] Luker GD, Glazer HS, Eagar G, Gutierrez FR, Sagel SS. Aortic pulmonary embolus. Excessive concern about administering dissection: effect of prospective chest radiographic diagnosis on delay anticoagulation to AD patients could result in delays for time- to definitive diagnosis. Radiology 1994;193(3):813-9. critical therapies in patients with acute coronary syndrome or [5] Anagnostopoulos CE. Acute aortic dissection. Baltimore (Md)7 University Park Press; 1975. pulmonary embolus. This is especially true given the lack of [6] Lindsay J, Hurst JW. Clinical features and prognosis in dissecting specificity for AD observed with plain chest radiography aneurysm of the aorta: a reappraisal. Circulation 1967;53:122-54. [12,14]. The optimal diagnostic and therapeutic approach in [7] Bourland MD. Aortic dissection. In: Barkin R, Rosen P, editors. patients with a clinical presentation that is suspicious for AD Emergency medicine: concepts and clinical practice. 4th ed. St Louis and an abnormal mediastinum on plain chest radiograph is not (Mo)7 Mosby; 1998. [8] Armstrong WF, Bach DS, Carey LM, Froehlich J, Lowell M, Kazerooni as clear and warrants additional research. EA. Clinical and echocardiographic findings in patients with suspected Limitations to this analysis include the retrospective acute aortic dissection. Am Heart J 1998;136(6):1051-60. design and relatively small sample size, although we were [9] Pate JW, Richardson RL, Eastridge CE. Acute aortic dissections. able to demonstrate statistically significant differences Am Surg 1976;42(6):395-404. between the groups that provide a plausible explanation [10] Hagan PG, Nienaber CA, Isselbacher EM, et al. The International Registry of Acute Aortic Dissection (IRAD): new insights into an old for the high rate of inadvertent anticoagulation in patients disease. JAMA 2000;283(7):897-903. with AD. In addition, we used the final cardiology and [11] von Kodolitsch Y, Schwartz AG, Nienaber CA. Clinical presentation radiology interpretations of ECG and chest radiographs of acute aortic dissection. Arch Intern Med 2000;160(19):2977-82. rather than the ED physician’s interpretation; if the bwet [12] Klompas M. Does this patient have an acute thoracic aortic dissection? readQ were used instead, the results would likely have been JAMA 2002;287(17):2262-72. [13] Vu FH, Young N, Soo YS. Imaging of thoracic aortic dissection. even more disparate because nonradiologists are less likely Australas Radiol 1994;38(3):170-5. to detect subtle abnormalities in the mediastinal silhouette. [14] Chen K, Varon J, Wenker OC, Judge DK, Fromm Jr RE, Sternbach Finally, a limited amount of clinical data was available from GL. Acute thoracic aortic dissection: the basics. J Emerg Med 1997; ED records, limiting the data available for this analysis. 15(6):859-67. American Journal of Emergency Medicine (2005) 23, 443–448
www.elsevier.com/locate/ajem
Real-time paramedic compared with blinded physician identification of ST-segment elevation myocardial infarction: results of an observational studyB
James A. Feldman MDa,*, Kathryn Brinsfield MDa,c, Sheilah Bernard MDb, Daniel White EMT-Pc, Thomas Maciejko EMT-Pc aDepartment of Emergency Medicine, Boston Medical Center, Boston University School of Medicine, Boston, MA 02118, USA bDepartment of Cardiology, Boston Medical Center, Boston University School of Medicine, Boston, MA 02118, USA cBoston Emergency Medical Services, Boston, MA 02118, USA
Received 4 October 2004; accepted 13 October 2004
Abstract The aim of the study were to determine if paramedics can accurately identify ST-segment elevation myocardial infarction (STEMI) on prehospital 12-lead (PHTL) electrocardiogram and to compare paramedic with blinded physician identification of STEMI. Paramedics identified definite STEMI, or possible acute myocardial infarction but not definite, and nondiagnostic. Two blinded readers (cardiologist and emergency physician) independently categorized each PHTL. A third reviewer assigned final diagnoses and determined whether the PHTL met STEMI criteria. One hundred sixty-six PHTL were acquired over an 8-month period. Fifteen were excluded from analysis. Sixty-two percent of the patients (94/151) were male, mean age was 61.1 years (F14.8 SD, range 20-92 years), and 81% had chest pain. Twenty-five patients (16.6%; 95% confidence interval [CI], 11%-23.5%) had confirmed STEMI and 16 (10.6%) had confirmed non-STEMI acute myocardial infarction. Paramedic sensitivity was 0.80 (95% CI, 0.64-0.96); specificity was 0.97 (95% CI, 0.94-1.00) with positive likelihood ratio of 25.2 and negative likelihood ratio of 0.21. Overall accuracy was similar for paramedic and physician reviewers (0.94, 0.93, 0.95). Highly trained paramedics in an urban emergency medical services system can identify patients with STEMI as accurately as blinded physician reviewers. D 2005 Elsevier Inc. All rights reserved.
1. Introduction
The use of prehospital 12-lead (PHTL) electrocardio- gram (ECG) provides more rapid diagnosis and improves Presented at the Society for Academic Emergency Medicine’s annual treatment for patients with acute myocardial infarction meeting, Chicago, Ill, May 1998. (AMI). The American Heart Association’s ECC-CPR B This study was supported by the Boston Emergency Physicians Research Fund. Guidelines recommended the use of diagnostic 12-lead T Corresponding author. Tel.: +1 617 414 5972; fax: +1 617 414 5975. ECGs in urban and suburban emergency medical services E-mail address: [email protected] (J.A. Feldman). (EMS) systems as a class I (evidence supports) recom-
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.10.011 444 J.A. Feldman et al. mendationin2000[1]. Research has shown that 2.1. Phase 1: training and certification prehospital ECGs may identify patients who are candi- dates for reperfusion therapy and improve the accuracy of All department field paramedics completed a 6-hour the evaluation of patients with prehospital chest pain training course including practical and written examinations. [1-3], may be accurately acquired by paramedics [3,4], The training course emphasized pathophysiology of the and will usually have minimal effect on scene times [5-7]. acute coronary syndrome, proper acquisition of a 12-lead In addition, transmission of the ECG by cellular modem ECG, and accurate ECG interpretation with emphasis on to the ED has been shown to reduce time delay for diagnostic changes associated with AMI and recognition of treatment of AMI [8,9] and provide accurate information STEMI patterns. Paramedics were trained to identify for field thrombolysis decisions to ED physicians common sources of misclassification such as bundle-branch [8,10,11]. block, ventricular aneurysms, pericarditis, and the early The American Heart Association guidelines and most repolarization variant. The goal of the training program was PHTL ECG studies suggest that a prehospital ECG program to maximize specificity of interpretations (minimize false- include the acquisition of a diagnostic 12-lead ECG with positive STEMI identification). We instructed paramedics to transmission to a receiving hospital for physician review identify patients with STEMI only when they were quite [1,3,12] noting that approximately 85% of obtained tracings certain of the ECG diagnosis. can be successfully transmitted [1]. Few investigators have 2.2. Phase 2: patient enrollment explored whether transmission of 12-lead ECGs, incurring additional costs as well as potential for transmission failure, Paramedics obtained 12-lead ECGs on patients with a is required when highly trained paramedics provide inter- clinical suspicion of AMI from May to December 1997. pretations to receiving facilities [13]. Electrocardiograms were acquired using the LIFEPAK During the formative years of prehospital emergency 10 with serial 12-lead adapter and LIFEPAK 12 (both medical care, routine telemetry transmission to a base manufactured by Medtronic Physio-Control, Redmond, physician was considered essential until the information WA). Included were patients presenting with chest pain, provided beyond paramedic interpretation was directly syncope, weakness, dyspnea, or other symptoms considered examined [14-18]. Some have suggested that computerized by the field provider to be strongly suggestive of acute interpretation could improve the sensitivity and specificity cardiac ischemia. Paramedics were instructed to exclude of prehospital 12-lead ECG review [16,19,20]. However, patients with left bundle-branch block and pacemaker computer interpretations have been demonstrated to be less artifact from further analysis. Prehospital 12-lead interpreta- sensitive for anterior and inferior AMI as compared to tion was conducted in real time and recorded on a data cardiologists’ review [21]. collection form. Electrocardiograms were assigned to 1 of 3 We hypothesized that highly trained paramedics in a large categories: definite STEMI (certainty of diagnosis), possible urban EMS system could identify patients with ST-segment AMI or ischemia (to be used even if STEMI thought elevation myocardial infarction (STEMI) with sufficient probable, but not certain), and nondiagnostic. accuracy that transmission of 12-lead ECGs for physician All ALS transfers with PHTL were identified by daily trip review would be unnecessary. Furthermore, in EMS systems sheet review. A board-certified emergency physician and a where patient information and an ECG are provided to the board-certified cardiologist independently reviewed each base physician [16,22], it is unclear how much additional PHTL without clinical data available. We included all ECGs information is contributed by physician review. Previous with multiple leads submitted by paramedics with a data studies have suggested that clinical information can influ- sheet completed (intent to perform 12-lead ECG). These ence ECG interpretations for both experienced and novice included 4 tracings without complete 12 leads recorded ECG readers [23]. Thus, we wished to examine the where providers believed clear-cut ST segment changes were performance of paramedic ECG interpretation in real time evident before completion of the serial 12-lead recording as compared with blinded physician ECG review. (LIFELPAK 10) used at the inception of the project. Data were collected upon ED presentation, during hospitalization, and at discharge. These included socio- 2. Methods demographic information, initial and follow-up clinical features, ECGs, cardiac biomarker, and other diagnostic This prospective observational study was carried out tests. Time to treatment by either thrombolysis or time to between May and December 1997 at Boston EMS, an urban arrival at the cardiac catheterization laboratory was third-party 2-tiered advanced life support (ALS)/basic life recorded. A third physician blinded to both paramedic and support (BLS) system providing emergency services to the physician PHTL interpretations assigned final diagnoses city of Boston. Boston EMS responds to more than 100000 using the World Health Organization (WHO) criteria for calls to 911 with more than 9000 ALS transports per year. AMI diagnoses. The WHO criteria for assignment of There were 41 field ALS providers at the time of the diagnoses of AMI [24-26] required at least 2 of the 12-lead project. following criteria: (1) a clinical history of ischemia-type Paramedic identification of STEMI 445 chest discomfort; (2) changes on serial ECGs; and (3) a rise Table 2 Overall performance of paramedic interpretation and fall in serum cardiac markers. Cardiac troponins were not measured at the time of this study. Patients with aborted Paramedic and confirmed STEMI AMI were considered to have had a STEMI if they had Yes No Total ischemic symptoms, diagnostic ST elevation on the PHTL, Yes 20 4 24 and evidence of reperfusion with thrombolysis or angio- No 5 122 127 plasty without characteristic biomarker changes. Total 25 126 151 The third reviewer then classified the PHTL as either STEMI or not based upon the PHTL demonstrating 0.1 mV or higher in 2 contiguous limb leads or 0.2 mV or higher in 4 patients as having STEMI who did not meet diagnostic 2 contiguous precordial leads. All discordant final diagnoses criteria for confirmed STEMI. Of these misclassifications, and the final determination as to whether the PHTL met 1 was considered a misread, 2 patients had bAMI mimicsQ STEMI criteria were resolved by consensus. (1 patient with early repolarization and prior history of MI Data were entered into a database (Microsoft Access and 1 patient with left ventricular hypertrophy and anterior 2000, Microsoft Corporation, Redmond, Wash) and ana- septal ST elevations). The fourth patient had confirmed lyzed using SAS 8.2 (SAS Institute Inc, Cary, NC). We acute coronary syndrome with unstable angina and had calculated sensitivity, specificity, negative and positive coronary artery bypass graft within 24 hours, but did not predictive values, likelihood ratio (LR), and overall accuracy satisfy the WHO criteria for AMI. The overall performance with 95% confidence intervals (CIs) where appropriate. We of paramedic interpretation (Table 2) was a sensitivity of compared readers’ analyses with McNemar test and Cohen 80%, specificity of 97%, positive predictive value of 83%, j. The institutional review board at Boston Medical Center, negative predictive value of 96%, accuracy of 94%, and the Medical Control Base Station for Boston EMS, approved positive LR of 25.2 for recognition of STEMI (95% CI, 9.4- this study. Patient consent was not required for 12-lead 67.4) (Table 3). acquisition because this technology was implemented as a The 2 physician reviewers were blinded to any patient new standard of care and no intervention was planned based characteristics, clinical information, or follow-up data. The upon the paramedic interpretation of the PHTL. physician and paramedic interpretations are compared in Table 3. The overall accuracy of paramedic classifications performed in real time with the incorporation of clinical data 3. Results (94%; 95% CI, 90%-98%) was comparable to blinded emergency medicine (93%; 95% CI, 89%-97%) and A total of 166 patients had PHTL ECGs acquired by cardiology review (95%; 95% CI, 92%-99%). There was prehospital providers over an 8-month period. Of these, good agreement among all 3 paired readings (paramedic– 15 subjects were withdrawn for the following reasons: 14 emergency physician, paramedic-cardiology, emergency could not be assigned confirmed diagnoses (7 with no physician–cardiology) with j of 0.73, 0.67, and 0.79, follow-up data available, 3 left receiving sites AMA, 2 with respectively (Table 4). missing PHTL, 2 with missing case report forms [1 with confirmed STEMI, 1 confirmed non-STEMI]) and 1 because the PHTL available was too poor in quality for 4. Discussion reviewers to code. Sixty-two percent of the patients (94/151) F were male with a mean age of 61.1 years ( 14.8 SD, range Our study demonstrated that paramedics can acquire 20-92 years), and most (123 [81%]) had a chief complaint of prehospital 12-lead ECGs and independently identify those chest pain. The prehospital 12-lead population is summa- patients with STEMI without transmission or review by a rized in Table 1. Base Hospital physician. The overall accuracy of paramedic Of the 151 patients, 25 had confirmed STEMI (preva- interpretation is comparable to the results of other studies lence, 16.6%; 95% CI, 11%-23.5%). Two patients with that have examined clinician (paramedic, emergency phy- STEMI did not report chest pain as a complaint, but had sician, cardiologist) performance on static test cases or PHTL obtained for suspected AMI. scenarios [10,27,28] and consistent with other prehospital The paramedics correctly identified 20 of the 25 patients systems [13]. The high specificity of paramedic interpreta- with confirmed STEMI. The paramedics categorized tions and very high positive LR observed indicate that paramedic classification of STEMI has strong diagnostic Table 1 Characteristics of the 12-lead study group (n = 151) value (LR N10) and that few false positives would be expected in our EMS system. Furthermore, paramedics Age (y) 61.1 (14.8, 20-92) Male sex 94 (62) interpreting ECGs in a clinical setting are as accurate as Chief complaint—chest pain 123 (81) physicians who classified tracings without the benefit of clinical data. It is possible that prior studies that have Values are expressed as mean (SD, range) or n (%). required physician review may have been confounded by 446 J.A. Feldman et al.
Table 3 Estimates (95% CI) Paramedic EM MD Cardiology MD Sensitivity 0.80 (0.64-0.96) 0.80 (0.64-0.96) 0.92 (0.81-1.00) Specificity 0.97 (0.94-1.00) 0.96 (0.93-0.99) 0.96 (0.93-0.99) positive predictive value 0.83 (0.68-0.98) 0.80 (0.64-0.96) 0.82 (0.68-0.96) Negative predictive value 0.96 (0.93-0.99) 0.96 (0.93-0.99) 0.98 (0.96-1.00) Overall accuracy 0.94 (0.90-0.98) 0.93 (0.89-0.97) 0.95 (0.92-0.99) Likelihood ratio Positive 25.2 (9.4-67.4) 20.16 23.18 Negative 0.21 0.21 0.08 additional clinical data provided along with the transmitted reperfusion teams [38]. In these systems there may be some 12-lead ECG. cost to false positives; however, there should not be Like any diagnostic test, prehospital ECG interpretation additional risks. There can be costs to facilities that are can involve a trade-off between sensitivity and specificity. bypassed (lost patient volume and revenue) and to receiving The purpose of prehospital STEMI identification affects the PCI centers (unnecessary activation of PCI teams, especially desired or optimal performance goals. In the case of if not on site 24/7). In such a system, high specificity is still initiating thrombolysis in the prehospital setting where there desirable from the system’s perspective, although not as may be important risks to treatment of false positive high as would be required for field thrombolysis. patients, extremely high specificity is most appropriate. In We conducted this observational study with the goal of such systems, paramedic interpretation, computer and base achieving a very high specificity and positive predictive value station review may be indicated to maximize specificity for prehospital identification of STEMI. We trained providers [10,29]. It may also be appropriate to modify ECG criteria to identify STEMI only with high ECG certainty and to obtain to maximize specificity by increasing the cut-point for ST PHTL only in patients with high clinical suspicion of AMI elevation [30,31], especially in the septal leads by requiring (high prevalence of disease). Thus, we did not mandate that a threshold of more than 0.2 mV or more than 0.3 mV for all prehospital patients with symptoms suggestive of possible identification of STEMI [32]. Increasing the ST-elevation acute coronary syndrome or all chest pain patients have a requirement has a reported sensitivity and specificity with PHTL. Paramedic judgment rather than strict protocol computer analysis of 52% and 98% and cardiologist online guidelines directed the decision to obtain 12-lead ECGs. review of 62% and 98% [33]. During this initial phase, concerns were raised that obtaining Another well-described role for PHTL is to aide in the 12-lead ECGs on all patients with symptoms suggestive of triage decision or inhospital treatment preparation for acute cardiac ischemia, including those thought to be at very patients with suspected STEMI [6,9,34-37].Insucha low risk of STEMI, might adversely affect the operation of setting, it is appropriate to apply PTHL screening to a the EMS system. Therefore, paramedics were instructed to much broader population of patients with a lower preva- obtain PHTLs only on patients felt to have a reasonable lence of disease and to accept a lower specificity and a suspicion of AMI, rather than observe a strict complaint- higher sensitivity for STEMI classification. In this config- driven protocol such as requiring that a PHTL be obtained on uration, the decision to initiate reperfusion therapy is made all patients over a certain age with a complaint of chest pain by physicians in the hospital. The risks of thrombolytic regardless of the clinical characteristics of the complaint. therapy or the costs associated with activating a percutane- The prevalence of STEMI in our sample (16.6%; 95% ous coronary intervention (PCI) team for false-positive CI, 11%-23.5%) was higher than was observed in earlier patients are hospital-based. prehospital thombolysis studies. The STEMI prevalence Finally, in EMS systems where mechanical reperfusion rates reported in the earlier MITI project (4.2%) and (PCI) is the preferred treatment strategy for patients with Milwaukee feasibility study reported (8.6%) differ because STEMI, paramedic PHTL interpretation can be used to disease prevalence will vary with the population included triage patients to PCI-capable centers and to activate for PHTL [10,22,39]. Patients were not excluded from 12-lead ECG based upon factors that might affect reperfu- Table 4 Interobserver agreement sion decisions such as duration of symptoms, contra- Reader McNemar P kappa 95% CI indications to thrombolytic therapy, or clinically compromised condition. In fact, patients with evident EM-cardiology 1.0 .32 (NS) 0.79 (0.67-0.92) contraindications to thrombolysis or shock, exclusions for Paramedic- 1.14 .28 (NS) 0.67 (0.52-0.83) cardiology PHTL in other studies, were eligible for PHTL based upon Paramedic-EM 0.09 .76 (NS) 0.73 (0.58-0.88) the assumption that prehospital identification of STEMI may play an even more important role in determining appropriate EM indicates emergency medicine. triage and transportation decisions for these patients. Paramedic identification of STEMI 447
The results of our study confirm the hypothesis that the catheterization centers, provision of prehospital thrombol- additional costs required to transmit PHTL ECGs, especially ysis, or mobilization of interventional teams. Independent in large urban systems, may not be necessary. Furthermore, paramedic ECG interpretation should aide in the timely a system reliant on accurate paramedic interpretation is not delivery of reperfusion therapy for patients with STEMI. subject to equipment or transmission problems. Subsequent studies must examine the impact of paramedic-identified STEMI on hospital treatment practices, including the Acknowledgments accuracy of prehospital diversion of selected patients with AMI such as those with high-risk features, STEMI, or The authors wish to recognize the efforts of (1) all of the contraindications to thrombolytic therapy. Such a triage paramedics who provide care for the citizens of Boston scheme is based upon the reliability of the prehospital whose interest and enthusiasm made this project possible; clinical evaluation and accurate interpretation of the PHTL. (2) Ms Clara Safi, ANP, for her tireless efforts on behalf of Such a system requires highly trained ALS providers and the 12-lead Project; (3) the Data Coordinating Center at the rigorous medical control and review. Boston University School of Public Health and Ms Jeena There were a number of limitations to our study. Because Easow, MPH, for their assistance with the statistical of the small sample size, small changes in ECG classi- analysis; and (4) Supriya Mehta PhD, MHS, for her fications could have a large effect on reader performance thoughtful comments on the manuscript. (paramedic and physician). A larger study could provide more robust estimates of accuracy and could possibly detect clinically important differences between ECG readers. This References study was conducted with experienced urban system para- medics. The results may not generalize to all EMS providers [1] Part 7: The era of reperfusion : section 1: acute coronary syndromes or EMS systems. Indeed, one of the most important (acute myocardial infarction). Circulation 2000;102(Suppl I):172-203. [2] Aufderheide TP, Hendley GE, Thakur RK, et al. The diagnostic elements of this project is a stable force of highly trained, impact of prehospital 12-lead electrocardiography. Ann Emerg Med motivated, and experienced ALS providers. Other authors 1990;19:1280-7. have noted the need for rigorous training and performance [3] Kudenchuk PJ, Maynard C, Cobb LA, et al. Utility of the prehospital review of a 12-lead ECG program [12]. electrocardiogram in diagnosing acute coronary syndromes: the Furthermore, our system was predicated on the assump- Myocardial Infarction Triage and Intervention (MITI) project. J Am Coll Cardiol 1998;32:17-27. tion that there would be real direct costs to false-positive [4] Aufderheide TP, Hendley GE, Woo J, et al. A prospective evaluation paramedic STEMI diagnoses and that this required mini- of prehospital 12-lead ECG application in chest pain patients. mizing such misclassifications. Other systems could be J Electrocardiol 1992;24:8-13. envisioned where patients with suspected but less clear-cut [5] Brown Jr JL. An eight-month evaluation of prehospital 12-lead STEMI could be reevaluated without direct costs in terms of electrocardiogram monitoring in Baltimore county. Md Med J 1997; (Suppl):64-6. activation of specialized personnel. These systems could [6] Canto JG, Rogers WJ, Bowlby LJ, et al. The prehospital electrocar- include sites where PCI facilities are staffed for uniform diogram in acute myocardial infarction: is its full potential being 24-hour operation analogous to level I trauma centers. In realized? National Registry of Myocardial Infarction 2 Investigators. such a setting, the sensitivity-specificity trade-off for J Am Coll Cardiol 1997;29:498-505. STEMI might be different. In our system, we believed that [7] Myers RB. Prehospital management of acute myocardial infarction: electrocardiogram acquisition and interpretation, and thrombolysis by mobilization of cardiac catheterization teams for patients prehospital care providers. Can J Cardiol 1998;14:1231-40. with STEMI or pre-arrival preparation of an expensive [8] Kereiakes DJ, Gibler WB, Martin LH, et al. Relative importance of thrombolytic agent, although facilitated by very early emergency medical system transport and the prehospital electrocar- prehospital identification, would be problematic should diogram on reducing hospital time delay to therapy for acute teams be activated for an important number of bfalse myocardial infarction: a preliminary report from the Cincinnati Heart Q Project. Am Heart J 1992;123:835-40. alarms. Finally, the focus of this project was prehospital [9] Ljosland M, Weydahl PG, Stumberg S. Prehospital ECG reduces the identification of patients with STEMI. Further changes in delay of thrombolysis in acute myocardial infarction. Tidsskr Nor the treatment of patients with non-STEMI may modify the Laegeforen 2000;120:2247-9. role of prehospital ECG interpretation and prehospital triage [10] Aufderheide TP, Keelan MH, Hendley GE, et al. Milwaukee prehospital of patients with non-STEMI AMI. chest pain project-phase I: feasibility and accuracy of prehospital thrombolytic candidate selection. Am J Cardiol 1992;69:991-6. [11] Gibler WB, Kereiakes DJ, Dean EN, et al. Prehospital diagnosis and treatment of acute myocardial infarction: a north-south perspective. 5. Conclusions The Cincinnati Heart Project and the Nashville Prehospital TPA Trial. Am Heart J 1991;121:1-11. Experienced, highly trained urban paramedics can [12] Aufderheide TP, Kereiakes DJ, Weaver WD, et al. Planning, implementation, and process monitoring for prehospital 12-lead accurately and specifically identify STEMI. Paramedics ECG diagnostic programs. Prehosp Disaster Med 1996;11:162-71. without online physician review may be sufficiently [13] Foster DB, Dufendach JH, Barkdoll CM, et al. Prehospital recognition accurate to allow triage of selected patients with AMI to of AMI using independent nurse/paramedic 12-lead ECG evaluation: 448 J.A. Feldman et al.
impact on in-hospital times to thrombolysis in a rural community and non-AMI syndromes by emergency physicians. Acad Emerg Med hospital. Am J Emerg Med 1994;12:25-31. 2001;8:349-60. [14] Lambrew CT. The experience in telemetry of the electrocardiogram to [28] Salerno SM, Alguire PC, Waxman HS. Competency in interpretation a base hospital. Heart Lung 1974;3:756-64. of 12-lead electrocardiograms: a summary and appraisal of published [15] Pozen MW, Fried DD, Voigt GG. Studies of ambulance patients with evidence. Ann Intern Med 2003;138:751-60. ischemic heart disease. 11. Selection of patients for ambulance [29] Brady WJ, Perron AD, Chan T. Electrocardiographic ST-segment telemetry. Am J Public Health 1977;67:532-5. elevation: correct identification of acute myocardial infarction (AMI) [16] Pozen MW, Fried DD, Smith S, et al. Studies of ambulance patients and non-AMI syndromes by emergency physicians. Acad Emerg Med with ischemic heart disease. 1. The outcome of pre-hospital life- 2001;8:349-60. threatening arrhythmias in patients receiving electrocardiographic [30] Lamfers EJ, Schut A, Hertzberger DP, et al. Prehospital versus telemetry and therapeutic interventions. Am J Public Health 1977;67: hospital fibrinolytic therapy using automated versus cardiologist 527-31. electrocardiographic diagnosis of myocardial infarction: abortion of [17] Pozen MW, Berezin MM, Kulp RC. Cost and utility considerations in myocardial infarction and unjustified fibrinolytic therapy [see implementing ambulance telemetry. Heart Lung 1980;9:866-72. comment, review, 24 refs]. Am Heart J 2004;147:509-15. [18] Woodwark GM, Gillespie IA. Monitoring of ambulance patients by [31] Otto LA, Aufderheide TP. Evaluation of ST segment elevation criteria radio telemetry. CMAJ 1970;102:1277-9. for the prehospital electrocardiographic diagnosis of acute myocardial [19] Kudenchuk PJ, Ho MT, Weaver WD, et al. Accuracy of computer- infarction. Ann Emerg Med 1994;23:17-24. interpreted electrocardiography in selecting patients for thrombolytic [32] Lamfers EJ, Schut A, Hertzberger DP, et al. Prehospital versus therapy. MITI Project Investigators. J Am Coll Cardiol 1991; hospital fibrinolytic therapy using automated versus cardiologist 17:1486-91. electrocardiographic diagnosis of myocardial infarction: abortion of [20] Lamfers EJ, Schut A, Hertzberger DP, et al. Prehospital versus myocardial infarction and unjustified fibrinolytic therapy [see hospital fibrinolytic therapy using automated versus cardiologist comment, review, 24 refs]. Am Heart J 2004;147:509-15. electrocardiographic diagnosis of myocardial infarction: abortion of [33] Lamfers EJ, Schut A, Hertzberger DP, et al. Prehospital versus myocardial infarction and unjustified fibrinolytic therapy [see hospital fibrinolytic therapy using automated versus cardiologist comment, review, 24 refs]. Am Heart J 2004;147:509-15. electrocardiographic diagnosis of myocardial infarction: abortion of [21] Willems J, Abreu-Lima C, Arnaud P, et al. The diagnostic myocardial infarction and unjustified fibrinolytic therapy [see performance of computer programs for the interpretation of electro- comment, review, 24 refs]. Am Heart J 2004;147:509-15. cardiograms. N Engl J Med 1991;325:1767-73. [34] Patel RJ, Vilke GM, Chan TC. The prehospital electrocardiogram. [22] Weaver WD, Cerqueira M, Hallstrom AP, et al. Prehospital-initiated J Emerg Med 2001;21:35-9. vs hospital-initiated thrombolytic therapy. The myocardial infarction [35] Purvis GM, Weiss SJ, Gaffney FA. Prehospital ECG monitoring of triage and intervention trial. JAMA 1993;270:1211-6. chest pain patients. Am J Emerg Med 1999;17:604-7. [23] Hatala R, Norman GR, Brooks LR. Impact of a clinical scenario on [36] Wall T, Albright J, Livingston B, et al. Prehospital ECG transmission accuracy of electrocardiogram interpretation. J Gen Intern Med speeds reperfusion for patients with acute myocardial infarction. N C 1999;14:126-9. Med J 2000;61:104-8. [24] Anonymous. WHO MONICA Project. MONICA Manual, revised [37] Welsh RC, Ornato J, Armstrong PW. Prehospital management of acute edition. 1990. Geneva, Switzerland, WHO Cardiovascular Disease ST-elevation myocardial infarction: a time for reappraisal in North Unit. Ref type: Report. America [review, 32 refs]. Am Heart J 2003;145:1-8. [25] Anonymous. Myocardial infarction redefined—a consensus docu- [38] Moyer P, Feldman J, Levine J, et al. Implications of the mechanical ment of the Joint European Society of Cardiology/American College (PCI) vs thrombolytic controversy for ST segment elevation of Cardiology Committee for the Redefinition of Myocardial myocardial infarction on the organization of emergency medical infarction. JACC 2000;36:959-69. services: the Boston EMS Experience. Crit Pathw Cardiol 2004; [26] Gillum RF, Fortmann SP, Prineas RJ, et al. International diagnostic 3:53-61. criteria for acute myocardial infarction and acute stroke. Am Heart J [39] Weaver WD, Eisenberg MS, Martin JS, et al. Myocardial infarction 1984;108:150-8. triage and intervention project-phase I: patient characteristics and [27] Brady WJ, Perron AD, Chan T. Electrocardiographic ST-segment feasibility of prehospital initiation of thrombolytic therapy. J Am Coll elevation: correct identification of acute myocardial infarction (AMI) Cardiol 1990;15:925-31. American Journal of Emergency Medicine (2005) 23, 449–453
www.elsevier.com/locate/ajem
Predictive value of C-reactive protein at different cutoff levels in acute appendicitis
Han-Ping Wu MDa,e, Ching-Yuang Lin MD, PhDb,e,*, Chin-Fu Chang MDa, Yu-Jun Changc, Chin-Yi Huangd aDepartment of Emergency Medicine, Changhua Christian Hospital, Changhua 500, Taiwan bDepartment of Pediatrics, Changhua Christian Hospital, Changhua 500, Taiwan cLaboratory of Epidemiology and Biostatistics, Changhua Christian Hospital, Changhua 500, Taiwan dEvidence Base Medicine Center, Changhua Christian Hospital, Changhua 500, Taiwan eInstitute of Medical Research, Chang Jung Christian University, Changhua 500, Taiwan
Received 21 July 2004; revised 16 September 2004; accepted 13 October 2004
Abstract Determining the different cutoff values of C-reactive protein (CRP) on the basis of how long the patient’s symptoms were present can be used to early predict acute appendicitis. We analyzed retrospectively from 2001 to 2004 the hospital records of 568 patients who underwent appendectomies for suspected appendicitis. Receiver operating characteristic analysis has shown that CRP measurement can increase the diagnostic accuracy in acute appendicitis. The cutoff values of CRP concentration taken as the first, second, and third days after onset of symptoms that distinguish acute appendicitis from other acute abdominal diseases were 1.5, 4.0, and 10.5 mg/dL, respectively; the values that distinguish perforated appendicitis from other acute abdominal diseases were 3.3 mg/dL (first day), 8.5 mg/dL (second day), and 12.0 mg/dL (third day). The different cutoff values of CRP concentration may serve as a useful predictive parameter in the early diagnosis of acute appendicitis on the first 3 days after the onset of symptoms. D 2005 Elsevier Inc. All rights reserved.
1. Introduction the right iliac fossa, and signs of local peritonitis; diagnostic accuracy based on these symptoms ranges from 70% to 80% Acute appendicitis is the most common abdominal [2,3,5-7]. However, in some cases, the clinical presentations surgical emergency [1-6]. In the emergency department, a are atypical and diagnostic errors are made. The diagnostic great number of patients with right lower quadrant pain are difficulties lead surgeons to perform unnecessary laparoto- evaluated by clinicians to rule out appendicitis everyday. mies whereas misdiagnosis can lead to perforation and Classically, the diagnosis of acute appendicitis is based on a abscess formation [3,7,8]. Perforation of an inflamed brief history of abdominal pain, nausea, migration of pain to appendix occurs in 15% to 25% of patients treated surgically for suspected acute appendicitis, with the highest T Corresponding author. Children’s Hospital, Changhua Christian rates encountered in young children and elderly patients Hospital, Changhua 500, Taiwan. [2-7,10]. In general, 2 important facts remain: a normal E-mail address: [email protected] (C.-Y. Lin). appendix found during appendectomy represents a misdi-
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.10.013 450 H.-P. Wu et al. agnosis, and a delayed diagnosis of acute appendicitis may pyrexia, migration of pain, and tenderness over the right lead to perforation and peritonitis. Thus, improving accu- lower quadrant of the abdomen. The patients whose racy is desirable both for earlier diagnosis and for avoiding symptoms and signs lasted more than 3 days were excluded unnecessary appendectomies. Comparatively new techni- from our analysis. Also, patients with clinically suspected ques, including nuclear medicine scans, ultrasonography, acute appendicitis who were treated by nonsurgical methods and computed tomographic scanning, have been used were excluded. Therefore, 542 patients of suspected increasingly to evaluate patients with suspected appendicitis appendicitis were studied in this series. The ultimate [1,4-6,8,9]. However, some authors have suggested that diagnosis of acute appendicitis was based on histological these tests are not easy to apply in primary healthcare examination of the excised appendix. settings and simply increase costs without significantly increasing diagnostic accuracy [1,3,4,10]. 2.2. Method In contrast, preoperative laboratory tests can be per- In the hospital charts, the following data were recorded formed easily in primary healthcare settings and often aid on admission: age, sex, body temperature, time of onset of primary clinicians with decision making about patients with symptoms, and time of admission. On admission, the blood clinically suspected acute appendicitis. Among these tests, samples were obtained from all patients and the CRP C-reactive protein (CRP) concentration is the most widely concentration was recorded. The concentration of CRP in estimated acute phase protein and is considered to be a good serum was measured by immunoturbidimetry (Beckman predictor of acute appendicitis [6,7,9-16]. To our knowl- Coulter, Fullerton, Calif). In all patients with suspected edge, however, the cutoff values of CRP concentration in appendicitis, the correlation between the CRP levels and the distinguishing acute appendicitis from other acute abdom- period from the onset of symptoms to admission was inal disorders have not been determined. In this study, we analyzed statistically based on the different cutoff levels of aimed to determine the different cutoff values of CRP based CRP obtained from data gathered at the time of admission. on how long the patient’s symptoms were present to improve diagnostic accuracy of acute appendicitis. 2.3. Statistical analysis The methods of statistical analysis were the t test, the Mann-Whitney U test, and the receiver operating charac- 2. Methods teristic (ROC) curve. Possibility levels of less than .05 were 2.1. Patients population taken as significant. The informative value of each biological marker was determined, and the cutoff points From 2001 to 2004, we analyzed retrospectively the were defined by ROC analysis. The area under the curve hospital records of 568 consecutive patients who underwent (AUC), calculated using the trapezoidal rule, was consid- appendectomies for clinically suspected acute appendicitis ered a global measure of the diagnostic value of the para- at the Changhua Christian Hospital, Taiwan. The term meter. An optimal test result gives a value of 1.0, and a clinically suspected acute appendicitis was reserved for useless test result gives a value of 0.5. Positive and negative cases with clinical symptoms and signs that included likelihood ratios (LRs) (positive LR = ratio of the fraction of abdominal pain, anorexia, nausea, vomiting, diarrhea, true positives and false positives, and negative LR = ratio of
Table 1 The mean CRP levels of patients with simple appendicitis, perforated appendicitis, and normal appendices on the first 3 days after onset of symptoms Daya Normalb Appendicitis P Mean F SD (No.) 95% CI Mean F SD (No.) 95% CI 1 1.41 F 2.29 (37) 0.67-2.15 2.31 F 3.45 (180) 1.80-2.81 .056 2 2.99 F 2.92 (40) 1.75-3.56 7.09 F 6.06 (145) 6.10-8.07 b.001 3 4.95 F 4.95 (23) 2.93-6.97 15.47 F 8.58 (117) 13.92-17.03 b.001
Daya Appendicitis P Simple Perforated Mean F SD (No.) 95% CI Mean F SD (No.) 95% CI 1 2.05 F 3.35 (167) 1.54-2.56 5.59 F 3.10 (13) 3.90-7.27 b.001 2 6.14 F 5.75 (116) 5.09-7.19 10.85 F 5.87 (29) 8.74-13.01 b.001 3 11.81 F 6.45 (67) 10.26-13.35 20.38 F 8.67 (50) 17.98-22.79 b.001 a The period from the onset of symptoms to admission. b Normal appendixes. C-reactive protein in appendicitis 451
ABC Day 1 Day 2 Day 3 1.00 1.00 1.00
.75 .75 .75
.50 .50 .50 Sensitivity Sensitivity Sensitivity .25 .25 .25
0.00 0.00 0.00 0.00 .25 .50 .75 1.00 0.00 .25 .50 .75 1.00 0.00 .25 .50 .75 1.00 1 - Specificity 1 - Specificity 1 - Specificity DEFDay 1 Day 2 Day 3 1.00 1.00 1.00
.75 .75 .75
.50 .50 .50 Sensitivity Sensitivity Sensitivity .25 .25 .25
0.00 0.00 0.00 0.00 .25 .50 .75 1.00 0.00 .25 .50 .75 1.00 0.00 .25 .50 .75 1.00 1 - Specificity 1 - Specificity 1 - Specificity
Fig. 1 Receiver operating characteristic curves for CRP in distinguishing between patients with appendicitis (both nonperforated and perforated) and patients with normal appendices (A-C), and between patients with perforated appendicitis and patients with normal appendices (D-F). the fraction of false negatives and true negatives, respec- Thus, our diagnostic rate of acute appendicitis was 81.5% tively) were calculated for the best cutoff values. and normal appendicectomy rate was 18.5%. The mean CRP levels of patients with simple appendi- citis, perforated appendicitis, and normal appendices were 3. Results analyzed on the first 3 days after onset of symptoms (Table The patient comprised 331 males (61.1%) and 211 1). The mean CRP levels were statistically significantly females (38.9%) with a mean age of 22.3 F 19.4 years higher in patients with appendicitis than those with normal (range, 0-89 years). Of those, 350 had histologically proven appendices ( P b .001) and also significantly higher in simple acute appendicitis (nonperforated), 92 had perforat- patients with perforated appendicitis than simple appendi- ed, and 100 had normal appendices. Among those with citis ( P b .001) during the first 3 days. The ability of CRP normal appendices, only 6 patients who underwent laparot- concentration on admission to predict appendicitis was omies actually showed evidence indicating the need for analyzed by ROC curves (Fig. 1A-F). All CRP cutoff values surgical treatment, including perforation of diverticulum, had an AUC significantly greater than 0.5. On the basis of intestinal perforation, and ruptured tubo-ovarian abscess. sensitivity, specificity, positive likelihood ratio (LR+), and
Table 2 Sensitivity, specificity, LR+, and LRÀ at different cutoff levels of CRP in distinguishing appendicitis from normal appendices on the first 3 days after onset of symptoms Groupa CRP (mg/dL) Sensitivity (%) Specificity (%) LR+ LRÀ AUC (95% CI) Group I Day 1 1.5 0.38 0.81 2.00 0.77 0.60 (0.50-0.70) Day 2 4.0 0.63 0.78 2.88 0.45 0.77 (0.68-0.84) Day 3 10.5 0.72 0.83 4.13 0.34 0.88 (0.80-0.95)
Group II Day 1 3.3 0.77 0.89 7.12 0.26 0.90 (0.81-0.99) Day 2 8.5 0.70 0.95 13.79 0.33 0.92 (0.85-0.98) Day 3 12.0 0.90 0.96 20.70 0.10 0.96 (0.92-1.00) a Comparison of the patients with appendicitis (both nonperforated and perforated) and the patients with normal appendices (group I); comparison of the patients with perforated appendicitis and the patients with normal appendices (group II). 452 H.-P. Wu et al. negative likelihood ratio (LRÀ), the predictive values at of acute appendicitis on the first 3 days after onset of different cutoff levels of CRP in diagnosing appendicitis on symptoms. Clinically, the probability of acute appendicitis the first 3 days after onset of symptoms were determined, increased when the CRP level is greater than 1.5 mg/dL on respectively (Table 2). The cutoff levels of CRP in the first day, greater than 4.0 mg/dL on the second day, and distinguishing appendicitis from other acute abdominal greater than 10.5 mg/dL on the third day after onset of diseases (group 1) were 1.5 mg/dL on the first, 4.0 mg/dL symptoms. However, we found that the AUC value on the on the second, and 10.5 mg/dL on the third day after onset first day was not as statistically significant as that on the of symptoms. However, on the first day after onset of other 2 days. Thus, the predictive value of CRP in symptoms, the AUC (0.60) was smaller than that on the diagnosing appendicitis on the first day after onset of other 2 days (0.77 and 0.88). Thus, the diagnostic value of symptoms decreased. Kinetic studies have reported that CRP on the first day decreased. The cutoff levels of CRP in serial levels of CRP show an increase after 12 to 24 hours of distinguishing perforated appendicitis from other acute symptoms [11-15]. When symptoms of acute appendicitis abdominal diseases (group 2) were 3.3 mg/dL on the first, proceed rapidly, a patient’s levels of CRP may not increase on admission. Therefore, this explains why we found 8.0 mg/dL on the second, and 12.0 mg/dL on the third day negative CRP values in some patients with appendicitis on after onset of symptoms. The AUCs of the ROC curves admission, especially on the first day after onset of were all favorable on the first 3 days after onset of symptoms. Importantly, in further study, we need additional symptoms (0.90, 0.92, and 0.96). investigation to estimate whether other inflammatory markers can detect early acute appendicitis on the first day 4. Discussion after onset of symptoms. In our study, we noticed that the mean CRP level in Acute appendicitis is the most common problem requir- patients with perforated appendicitis was much greater than ing emergency abdominal surgery [1-6]. Early diagnosis of in patients with simple appendicitis. Furthermore, we also appendicitis can prevent perforation, abscess formation, and found that the ROC curve analysis of the different cutoff postoperative complications, and can decrease cost by values of CRP could be used to distinguish between shortening hospitalizations. However, despite intensive perforated appendicitis and other acute abdominal diseases. research and discussion, accurate diagnosis of acute Therefore, we calculated the cutoff values of CRP concen- appendicitis is still difficult. Although peritoneal aspiration tration on the first 3 days after onset of symptoms to see cytology, ultrasonography, and computed tomographic whether they indicated perforated appendicitis. We found that scanning have all been applied to increase diagnostic once the CRP level was greater than 3.3 mg/dL on the first accuracy in patients with clinically suspected acute appen- day, greater than 8.5 mg/dL on the second day, and greater dicitis [1,3,4,7,9], these procedures are not available in all than 12.0 mg/dL on the third day after onset of symptoms, the healthcare settings and the cost of performing them remains probability of perforated appendicitis significantly increased. high. In contrast, some serum inflammatory markers today The role of CRP did serve as a strong predictive parameter in are fast, economical, and universally available, and have the differential diagnosis of perforated appendicitis. been used for many years to improve the preoperative In the emergency department, once the diagnosis of diagnosis of acute appendicitis. C-reactive protein, an acute appendicitis is made, consequent surgical intervention will phase protein, can be used as a screening device for occult be indicated. However, the type of operation (open or inflammation, as a marker of disease activity, and as a laparoscopic appendectomy) is often based on the surgeon’s diagnostic tool [11]. Numerous studies have concluded that experience and whether the appendix is perforated or not elevated CRP levels aid in the diagnosis of acute appendi- [17,18]. Some studies have reported higher complication citis [6,7,9-16]. Specifically, some authors claim that CRP rates using laparoscopy for perforated cases [18-21]. can predict perforated appendicitis more effectively than Preoperative diagnosis of perforation, therefore, becomes simple appendicitis [11]. The diagnostic value of CRP important and essential. Indeed, we believe that the high concentration in predicting appendicitis appears to be predictive values of these different cutoff levels of CRP in reliable and widely applied to patients with suspected diagnosing appendicitis can aid primary clinicians in appendicitis [1,7,8,13]. Nevertheless, the application of the detecting acute appendicitis before the appendix perforates. cutoff values of CRP concentration in distinguishing acute Clinically, the much higher CRP values found in the appendicitis from other acute abdominal diseases based on perforated appendicitis cases may be of use to surgeons in how long that patient’s symptoms were present has never planning their operative approach. In our study, we have been reported. In this study, 542 patients with clinically determined CRP levels using different cutoff values based suspected acute appendicitis were analyzed and the results on how long the patient’s symptoms were present to indicated that the cutoff values of CRP could predict improve diagnostic accuracy of simple and perforated significantly the presence or absence of acute appendicitis. appendicitis. C-reactive protein will be helpful in predicting According to the results of ROC analysis, we determined appendicitis earlier and reducing the rate of complications the cutoff values of CRP concentration for early diagnosis caused by delay in diagnosis. C-reactive protein in appendicitis 453
Nevertheless, there was a limitation in this retrospective [7] Tompson MM, Underwood MJ, Dookeran KA, et al. Role of study. We did not report CRP serum values for patients with sequential leukocyte counts and C-reactive protein measurement in other clinical diagnoses who did not undergo laparotomy for acute appendicitis. Br J Surg 1992;70:822-4. [8] Korner H, Sondenaa K, Soreide JA. Perforated and non-perforated suspected appendicitis during the study period. In our acute appendicitis—one disease or two entities? Eur J Surg 2001;167: ongoing prospective study, these patients will be included. 525-30. The results will be more useful if they show a significant [9] Wu HP, Chang CF, Lin CY. Predictive inflammatory parameters in the difference in cutoff CRP levels between patients who did diagnosis of acute appendicitis in children. Acta Paediatr Taiwan not undergo laparotomies and those who had either 2003;44:227-31. [10] Paajanen H, Mansikka A, Laato M, et al. Novel serum inflamma- appendicitis or perforated appendicitis. tory markers in acute appendicitis. Scand J Clin Lab Invest 2002; In conclusion, the different cutoff values of CRP 62:579-84. concentration may serve as a predictive parameter in early [11] Clyne B, Olshaker JS. The C-reactive protein. J Emerg Med 1999; diagnosis of acute appendicitis on the first 3 days after 17:1019-25. onset of symptoms. The predictive value of CRP is [12] Gronroos JM, Gronroos P. Diagnosis of acute appendicitis. Radiology 2001;219:297-8. especially effective in the patients with perforated appen- [13] Oosterhuis WP, Zwinderman AH, Teeuwen M. C-reactive protein in dicitis. We therefore propose the addition of CRP mea- the diagnosis of acute appendicitis. Eur J Surg 1993;159:115-9. surement as a routine laboratory test in patients suspected [14] Gurleyik E, Gurleyik G, Unalmiser S. Accuracy of serum C- of having acute appendicitis. reactive protein measurements in diagnosis of acute appendicitis compared with surgeon’s clinical impression. Dis Colon Rectum 1995;38:1270-4. [15] Gronroos JM, Forsstron JJ, Irjala K, et al. Phospholipase A2, C- References reactive protein, and white blood cell count in the diagnosis of acute appendicitis. Clin Chem 1994;40:1757-60. [1] Andersson RE. Meta-analysis of the clinical and laboratory diagnosis [16] Paajanen H, Mansikka A, Laato M, et al. Are serum inflammatory of appendicitis. Br J Surg 2004;91:28-37. markers age dependent in acute appendicitis? J Am Coll Surg 1997; [2] Korner H, Soreide JA, Sondenaa K. Diagnostic accuracy of 164:303-8. inflammatory markers in patients operated on for suspected acute [17] Guller U, Jain N, Peterson ED, et al. Laparoscopic appendectomy in appendicitis: a receiver operating characteristic curve analysis. Eur J the elderly. Surgery 2004;135:479-88. Surg 1999;165:679-85. [18] Meter DE, Guzzetta PC, Barber RG, et al. Perforated appendicitis in [3] Ng KC, Lai SW. Clinical analysis of the related factors in acute children: is there a best treatment? J Pediatr Surg 2003;38:1520-4. appendicitis. Yale J Biol Med 2002;75:41-5. [19] Krisher SL, Browne A, Dibbins A, et al. Intra-abdominal abscess after [4] Rothrock SG, Pagane J. Acute appendicitis in children: emergency laparoscopic appendectomy for perforated appendicitis. Arch Surg department diagnosis and management. Ann Emerg Med 2001;36: 2001;136:438-41. 3-51. [20] Varlet F, Tardieu D, Limnone B, et al. Laparoscopic versus open [5] Mattei PA, Stevenson RJ, Ziegler MM. Appendicitis. In: Wyllie R, appendectomy in children—comparative study of 403 cases. Eur J Hyams JS, editors. Pediatric gastrointestinal disease. Philadelphia: Pediatr Surg 1994;4:333-7. WB Saunders; 1999. p. 466-472. [21] Horwitz JR, Custer MD, May BH, et al. Should laparoscopic [6] Hallan S, Asberg A, Edna TH. Additional value of biochemical tests appendectomy be avoided for complicated appendicitis in children? in suspected acute appendicitis. Eur J Surg 1997;163:5333-8. J Pediatr Surg 1997;32:1601-3. American Journal of Emergency Medicine (2005) 23, 454–458
www.elsevier.com/locate/ajem
Pediatric prescription pick-up rates after ED visitsB
Eric H. Kajioka MS, MDa, Erick M. Itoman MDa, M. Lily Lia, Deborah A. Taira ScDb,d, Gaylyn G. Li MDc,d, Loren G. Yamamoto MD, MPH, MBAa,e,* aDepartment of Pediatrics, John A. Burns School of Medicine, University of Hawaii, Honolulu, HI 96826, USA bDepartment of Clinical Epidemiology, John A. Burns School of Medicine, University of Hawaii, Honolulu, HI 96822, USA cDepartment of Obstetrics/Gynecology, John A. Burns School of Medicine, University of Hawaii, Honolulu, HI 96826, USA dCase Management Department, Hawaii Medical Service Association1, Honolulu, HI 96814, USA eEmergency Department, Kapiolani Medical Center for Women and Children, Honolulu, HI 96826, USA
Received 23 June 2004; accepted 14 October 2004
Abstract Objective: To determine the compliance rate in filling outpatient medication prescriptions written upon discharge from the emergency department (ED). Methods: Emergency department records of children during a 3-month period were examined along with pharmacy claim data obtained in cooperation with the largest insurance carrier in the community (private and Medicaid). Pharmacy claim data were used to validate the prescription pick-up date. Results: Overall, 65% of high-urgency prescriptions were filled. The prescription pick-up rate in the 0-to 3-year age group (75%) was significantly higher than in the rest of the cohort (55%) ( P b .001). Children with private insurance were more likely to fill their prescriptions (68%) compared to children with Medicaid insurance (57%) ( P = .03). Conclusion: This study demonstrates that filling a prescription after discharge from an ED represents a substantial barrier to medication compliance. D 2005 Elsevier Inc. All rights reserved.
1. Introduction
Patients discharged from the emergency department (ED) are commonly prescribed drugs for the management and treatment of disease. Disease prognosis and drug B An abstract of this study has been presented at the American College efficacy are dependent upon patient compliance with of Emergency Physicians Research Forum, Boston, MA, October 2003, and prescribed medications. Compliance with drug therapy is the Emergency Medicine Section of the American Academy of Pediatrics essential as many ED patients do not obtain follow-up National Conference and Exhibition, New Orleans, LA, October 2003. evaluations [1-4]. Repeat ED visits and subsequent hospital T Corresponding author. Department of Pediatrics, Honolulu, HI admissions have also been associated with prescription drug 96826, USA. Tel.: +1 808 983 8387; fax: +1 808 945 1570. E-mail address: [email protected] (L.G. Yamamoto). noncompliance [5,6]. 1 An independent licensee of the Blue Cross and Blue Shield The first requirement for outpatient medication compli- Association. ance is to actually obtain the medication. Prescription
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.10.015 Prescription pick-up rates 455 pick-up rates for patients discharged from EDs have been provide the patient with antibiotic samples, which would found to range between 44% and 92% [7-14]. Although delay or eliminate the need to fill the prescription. previous studies have used patient surveys to determine Albuterol exceptions were as follows: (a) children on this, few have confirmed that patients obtained their oral albuterol at home and prescribed aerosolized albuterol medication using actual insurance claim data. In addition, from ED were placed in the high-urgency group because there is little information on prescription pick-up rates in this change is most often due to wheezing which is pediatric populations. refractory to oral albuterol, resulting in respiratory difficulty. The purpose of this study is to estimate pediatric (b) Children on aerosolized albuterol at home and pre- prescription pick-up rates after an ED visit and to assess scribed oral albuterol from ED were placed in the low- factors associated with unfilled prescriptions. urgency medication group. All children receiving prescriptions that did not fall under the above criteria were placed in the high-urgency medica- 2. Methods tion group. For the patients meeting the inclusion criteria, pharmacy This is a retrospective study examining ED medical claim information was linked, using the insurer’s member records at a children’s hospital. Pharmacy claim information identification numbers, to determine whether and when the was provided by the largest insurance provider in the state, prescriptions given in the ED were filled. Prescription pick- which provides Medicaid and private insurance to approx- up information (date of filled prescription) was obtained by imately half of the state’s population. Daily ED records were examining medical insurance pharmacy claim data submitted reviewed to identify pediatric patients (age under 18 years) to the health insurer. Presumably, an insurance reimburse- covered by this insurer who were discharged from the ED ment claim would be submitted by the pharmacy if it filled with a high-urgency drug prescription during a 3-month the prescription because without this claim the pharmacy period (April 1, 2001-June 30, 2001). would not be paid. A filled prescription was defined as High- and low-urgency drug prescriptions were deter- having the prescription filled at a pharmacy and recorded as mined to study only those prescription which needed to be filled in the health insurance provider database (ie, compli- filled soon (preferably, immediately after discharge from ant). Absence of the prescription claim in the health the ED). A low-urgency prescription does not mean that insurance provider database was determined to be an unfilled the drug is of low importance, but it does mean that filling prescription (ie, noncompliant). The date of discharge from this prescription is either optional or it can be delayed the ED was compared to the date the prescription was filled. without significant medical consequence. High- and low- It is possible that the patient paid for their prescription urgency medications were determined according to the without using their medical insurance coverage, but this is following criteria: (1) children prescribed over-the-counter unlikely as this insurance provider participates with nearly medications were placed in the low-urgency medication all the pharmacies in the state. Other demographic informa- group. Private insurance does not cover over-the-counter tion was obtained from the ED records of the patient. medications, so the study methodology would not be able The primary outcome of the study was the proportion of to determine when these medications were filled. (2) high-urgency prescriptions written in the pediatric ED that Children prescribed a medication they were currently were filled as determined by insurance claim data. taking in the same form (ie, a refill of medication) were The v2, Fisher exact test, and odds ratios with 95% placed in the low-urgency medication group. (3) Children confidence intervals (CIs) were used to compare categorical given a new medication in the same drug class as a variables. Statistical analysis was performed using SPSS previous new prescription were placed in the low-urgency software, version 12.0 (SPSS Inc, Chicago, Ill). medication group (eg, on amoxicillin is changed to This study was approved by the institutional review cephalexin). This was considered low urgency because in board of this medical center. many instances the reason for a change is an allergic reaction, adverse effects, or findings suggesting the need for broader antibiotic coverage. These reasons were felt to 3. Results be less urgent, as the patient was already partially treated. (4) Children administered a dose of medication in the ED There were 6354 total records reviewed during the 3- before release, with the exception of albuterol (eg, IM/IV month study period, with 1829 excluded as adult patients ceftriaxone given in the ED and then prescribed oral (age z18). Of the remaining 4525 pediatric patient records, antibiotics), were placed in the low-urgency medication 2123 were covered by the participating health plan (47%). group. This was considered low urgency as the parenteral Of those remaining records, 1334 were eliminated because dose in the ED is sufficient for 12 to 24 hours, thus of hospital admission or no medication prescribed upon obtaining the outpatient prescription is less urgent. Primary discharge. Of the remaining 789 patient records, 403 were care physicians might choose to administer a second prescribed high-urgency medications and are further de- parenteral antibiotic dose during office follow-up or scribed below. 456 E.H. Kajioka et al.
The mean age of these 403 children was 5.9 years, with prescription filling and ethnicity, prescription of multiple 88% younger than 13 years old. The mean age in the medications, time of discharge from the ED, type of compliant group was 5.4 F 4.9 years, compared to the mean medication prescribed, or diagnosis. age in the noncompliant group which was 6.8 F 5.0 years. Prescription pick-up rates in the different subgroups are summarized in Table 1. Overall compliance with prescrip- 4. Discussion tion filling of high-urgency medications was 65% (260/ 403). Of those who picked up their medications, 83% (216/ Children represent a unique population for prescription 260) were filled on the same day of discharge from the ED, filling as compliance is dependent on the reliability of their and 93% (243/260) were filled within 1 day after discharge. parents or guardians [14-17] In addition, other factors such Ten patients filled their prescriptions 2 days after discharge, as insurance type, income level, age of the child, and and 7 patients filled their prescriptions more than 2 days (3- ethnicity have been shown to be important factors affecting 20 days) after discharge. compliance [13,14,18-20]. Although most published studies Fifty-one children received multiple medications, of on prescription filling for patients discharged from the ED which 71% (36/51) filled their prescriptions. There were have focused on adults [7-11], our study is one of the few to no children with multiple medications who filled partial assess children. prescriptions. Two hundred one children were discharged This medical insurance carrier has many different private between 9 pm and 9 am, of which 62% (125/201) filled insurance plans, all of which have an outpatient copayment their prescription. The most common prescriptions were (variable amount) for medications. Some patients who are anti-infectives (61%) and respiratory drugs (h-agonists/ coveredby2ormoreinsuranceplanswillhaveno corticosteroids/anticholinergics) (28%). copayment, but this is dependent on the specifics of the Prescription pick-up rates in the different age groups 2 insurance plans. All medicaid plans have no medication were as follows: 0 to 3 years, 75% (141/188); 4 to 12 years, copayments. The pick-up rates were lower in the medicaid 56% (93/165); 13 to 17 years, 52% (26/50). The prescrip- group, despite the absence of copayments. tion pick-up rate in the 0- to 3-year age group was The prescription filling rate of 65% in the present study is significantly higher than in the rest of the cohort (4-17 low compared to the 77% to 88% shown in 4 previous years, 55% pick-up rate) ( P b .001, odds ratio 2.42, 95% CI studies on prescription filling after discharge from EDs in 1.58-3.70). Males were more likely to fill their prescriptions the United States [7-9,14]. These studies assessed compli- (69%) compared to females (57%) ( P = .01, odds ratio 1.70, ance based on telephone interviews of patients, and 95% CI 1.12-2.58). Children with private insurance were prescription fill rates were not verified by pharmacy claim more likely to fill their prescriptions (68%) compared to data. In addition, a percentage of the patients initially children with Medicaid insurance (57%) ( P = .03, odds ratio surveyed in each study could not be contacted for study 1.63, 95% CI 1.07-2.49). No association was found between completion. For these reasons, these findings cannot be completely accurate for prescription fill rates. One study which used pharmacy claim data to assess prescription fill Table 1 Prescription pick-up rates among subgroups rates found a 74% compliance rate for filling a prescription Number Prescription of azithromycin [10]. Another study using pharmacy claim in group pick-up rate (%) data to assess oral corticosteroid prescription filling found a Overall total 403 65 44% prescription fill rate [13]. However, both studies Age groupT examined specific types of prescriptions and, thus, cannot 0-3 y 188 75 be applied to other prescription types or prescription fill 4-17 y 215 55 rates as a whole. Studies in other countries examining SexT patients discharged from the ED found prescription fill rates Male 242 69 Female 161 57 of 75% of 92% [11,12]. Our present study using insurance InsuranceT claim data accurately assesses high-urgency prescription Private 263 68 fill rates for children discharged from this ED during this Medicaid 140 57 study period. Ethnicity The finding in our present study that children with Hawaiian/part Hawaiian 133 64 private insurance were more likely to fill prescriptions than Samoan/other Pacific Islander 30 67 children with Medicaid insurance also confirms the findings Asian 115 64 of previous studies and further emphasizes improving Caucasian 38 68 compliance strategies for children with Medicaid insurance. Mixed ethnicity 55 71 When examining children discharged from the ED in the Others/unknown 32 50 United States, Wang et al [14] found children with Medicaid * Statistically significant difference (see text for P values, odds or no insurance to have significantly higher noncompliance ratios, and 95% CIs). rates of 41% and 60%, respectively, compared to those with Prescription pick-up rates 457 private insurance. Cooper and Hickson [13] examined from the inpatient pharmacy of a not-for-profit hospital, children with Medicaid insurance and found that less than many insurance plans will not reimburse the hospital for this half (44%) were compliant with oral corticosteroid pre- medication. The rationale for this possibly originates from scription filling. Poor compliance by children with Medicaid case law which restricts the use of pharmaceuticals also suggests that socioeconomic reasons such as insuffi- purchased by not-for-profit hospitals at reduced institutional cient funds, lack of transportation, and attitudes toward rates [24] which creates an unfair competitive advantage health care can play a major role in prescription filling after over smaller community pharmacies. However, in most discharge from the ED [3,7,8,10,12,4,21]. The higher rates instances, patients fill their prescriptions at large chain of noncompliant prescription filling could also contribute to pharmacies which might have similar purchasing power the higher rates of ED visits and hospitalizations seen in arrangements. Regulatory/statutory restrictions only serve to pediatric Medicaid patients [22,23]. inhibit our ability to improve medication compliance. In a study by Ginde et al [10], patients given a This study clearly shows that there is a substantial prescription to be filled free of charge at a 24-hour percentage of patients who do not obtain high-urgency pharmacy located in close proximity to the ED filled their medications. Even for those who did obtain their medi- prescription 74% of the time, compared to 100% who were cations, going to a community pharmacy is simply given their medication in the ED. This serves to reinforce inconvenient and burdensome. After a comprehensive the point that a major factor in completion of a medication evaluation in an ED, the family of an ill child must now course is actually obtaining the medication. Going to a go to a pharmacy where they must wait for their prescription pharmacy to fill the prescription reduces potential compli- to be filled. Presumably, the patient is waiting in the car or a ance by 26% despite the absence of financial barriers. potentially contagious patient is in the pharmacy with its Our finding that the younger children (0-3 years) were other customers. Calling or faxing a prescription to a more likely to have their prescriptions filled than older pharmacy or using automated prescription forwarding children (11-17 years) is consistent with the literature which systems could potentially reduce patient waiting time, but shows that adolescent patients are at higher risk for it still requires the patient’s family to drive to a pharmacy. noncompliance with drug therapy [13,18,19]. Anecdotally, patients inform us that even if the prescrip- Limitations of this study include limiting the patient base tion is called in, it is not ready when they get to the from one insurance carrier. However, we were fortunate to pharmacy. The pharmacy potentially benefits financially if obtain cooperation with this major insurance carrier to patients wait for their prescription, as family members provide this information for us. Certain factors involved in browse the store while waiting and purchase other items (eg, selection of insurance carriers could possibly select for snacks, drinks, household items) which adds to the revenue certain patient types and influence the outcome of the data. of the pharmacy. In addition, expensive antibiotic suspen- However, this particular insurance carrier covers more sions must be committed to use once water is added to the patients than any other carrier in the state, and it covers powder-containing bottle. If the patient decides to fill the private insurance and Medicaid insurance plans. The prescription elsewhere or the prescription is not picked up at selection criterion of high-urgency medications is also all, the expensive antibiotic must be discarded and the another limitation. Although we elected to study only pharmacy must absorb the cost of this. It is understandable urgent medications, the inclusion of low-urgency medica- that a pharmacy would be reluctant to commit and prepare tions could have altered the study outcome. It is likely that the expensive antibiotic suspension until the patient actually including low-urgency medications would have a lower rate arrives in the pharmacy. of prescription medication pick-ups as these would be less In summary, the prescription pick-up rate for children urgent. Thus, the inclusion of low-urgency medications discharged from the ED is approximately 65%. Age and would most likely have lowered the overall prescription insurance type play a significant role, with younger children pick-up rates. Another limitation is that we do not know and children with private insurance more likely to fill whether the patients actually took their medications as prescriptions than older children and children with Medic- directed, so our estimate of bcomplianceQ is only a measure aid, respectively. This study demonstrates that filling a of acquiring the medication and not actually taking it. prescription after discharge from an ED represents a An obvious intervention to improve compliance would substantial barrier to medication compliance. be to dispense the medication from the hospital pharmacy so that the medication is given to the patient upon discharge from the ED. This is often done for patients with limited References resources (lack of a means to obtain the medication), such as lack of funds or lack of transportation. Yet, this is not done [1] Barlas D, Homan CS, Rakowski J, Houck M, Thode Jr HC. How well routinely for patients with insurance. Regulatory and do patients obtain short-term follow-up after discharge from the emergency department? Ann Emerg Med 1999;34(5):610-4. statutory limitations might restrict the dispensing of [2] Scarfone RJ, Joffe MD, Wiley II JF, Loiselle JM, Cook RT. medications from the hospital’s inpatient pharmacy. For Noncompliance with scheduled revisits to a pediatric emergency example, if a prescription is dispensed for outpatient use, department. Arch Pediatr Adolesc Med 1996;150(9):948-53. 458 E.H. Kajioka et al.
[3] Zorc JJ, Scarfone RJ, Li Y, et al. Scheduled follow-up after a pediatric [14] Wang NE, Gisondi MA, Golzari M, van der Vlugt TM, Tuuli M. emergency department visit for asthma: a randomized trial. Pediatrics Socioeconomic disparities are negatively associated with pediatric 2003;111(3):495-502. emergency department aftercare compliance. Acad Emerg Med [4] Magnusson AR, Hedges JR, Vanko M, McCarten K, Moorhead JC. 2003;10(11):1278-84. Follow-up compliance after emergency department evaluation. Ann [15] Matsui DM. Drug compliance in pediatrics. Clinical and research Emerg Med 1993;22(3):560-7. issues. Pediatr Clin North Am 1997;44(1):1-14. [5] Olshaker JS, Barish RA, Naradzay JF, Jerrard DA, Safir E, Campbell [16] Mattar ME, Markello J, Yaffe SJ. Pharmaceutic factors affecting L. Prescription noncompliance: contribution to emergency department pediatric compliance. Pediatrics 1975;55(1):101-8. visits and cost. J Emerg Med 1999;17(5):909-12. [17] Nelson EW, Van Cleve S, Swartz MK, Kessen W, McCarthy PL. [6] Keith KD, Bocka JJ, Kobernick MS, Krome RL, Ross MA. Improving the use of early follow-up care after emergency department Emergency department revisits. Ann Emerg Med 1989;18(9):964-8. visits. A randomized trial. Am J Dis Child 1991;145(4):440-4. [7] Saunders CE. Patient compliance in filling prescriptions after [18] Kyngas HA, Kroll T, Duffy ME. Compliance in adolescents with discharge from the emergency department. Am J Emerg Med chronic diseases: a review. J Adolesc Health 2000;26(6):379-88. 1987;5(4):283-6. [19] Liptak GS. Enhancing patient compliance in pediatrics. Pediatr Rev [8] Thomas EJ, Burstin HR, O’Neil AC, Orav EJ, Brennan TA. Patient 1996;17(4):128-34. noncompliance with medical advice after the emergency department [20] La Greca AM. Issues in adherence with pediatric regimens. J Pediatr visit. Ann Emerg Med 1996;27(1):49-55. Psychol 1990;15(4):423-36. [9] Tackitt RD, Winship III HW. Effects of an outpatient pharmacy on the [21] Soliday E, Hoeksel R. Health beliefs and pediatric emergency acquisition of prescription medications by emergency room patients. department after-care adherence. Ann Behav Med 2000;22(4): Hosp Pharm Aug 1975;10(8):333-4, 338-9. 299-306. [10] Ginde AA, Von Harz BC, Turnbow D, Lewis LM. The effect of ED [22] Shatin D, Levin R, Ireys HT, Haller V. Health care utilization by prescription dispensing on patient compliance. Am J Emerg Med children with chronic illnesses: a comparison of medicaid and 2003;21(4):313-5. employer-insured managed care. Pediatrics 1998;102(4):E44. [11] Freeman CP, Guly HR. Do accident and emergency patients [23] Finkelstein JA, Barton MB, Donahue JG, Algatt-Bergstrom P, collect their prescribed medication? Arch Emerg Med 1985;2(1): Markson LE, Platt R. Comparing asthma care for Medicaid and 41-3. non-Medicaid children in a health maintenance organization. Arch [12] Matsui D, Joubert GI, Dykxhoorn S, Rieder MJ. Compliance with Pediatr Adolesc Med 2000;154(6):563-8. prescription filling in the pediatric emergency department. Arch [24] U.S. Supreme Court. ABBOTT LABS. v. PORTLAND RETAIL Pediatr Adolesc Med 2000;154(2):195-8. DRUGGISTS, 425 U.S. 1 (1976). 425 U.S. 1 ABBOTT LABORA- [13] Cooper WO, Hickson GB. Corticosteroid prescription filling for TORIES ET AL. v. PORTLAND RETAIL DRUGGISTS ASSN., INC. children covered by Medicaid following an emergency department CERTIORARI TO THE UNITED STATES COURT OF APPEALS visit or a hospitalization for asthma. Arch Pediatr Adolesc Med FOR THE NINTH CIRCUIT. No. 74-1274. Argued December 16, 2001;155(10):1111-5. 1975. Decided March 24, 1976. American Journal of Emergency Medicine (2005) 23, 459–462
www.elsevier.com/locate/ajem
Does the presence or absence of sonographically identified cardiac activity predict resuscitation outcomes of cardiac arrest patients?
Philip Salen MDa,*, Larry Melniker MD, MSb, Carolyn Chooljian MDc, John S. Rose MDd, Janet Alteveer MDe, James Reed PhDa, Michael Heller MDa aDepartment of Emergency Medicine, St. Lukes Hospital, St. Lukes Hospital EM Residency, Bethlehem, PA 18015, USA bDepartment of Emergency Medicine, New York Methodist Hospital, NY 11215, USA cDepartment of Emergency Medicine, University of California, UCSF Fresno, Emergency Medicine Residency, University Medical Center, Fresno, CA 93727, USA dDepartment of Emergency Medicine, University of California, Davis Medical Center, Sacramento, CA, USA eDepartment of Emergency Medicine, University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School, Camden, NJ
Received 29 October 2004; accepted 18 November 2004
Abstract This study evaluated the ability of cardiac sonography performed by emergency physicians to predict resuscitation outcomes of cardiac arrest patients. A convenience sample of cardiac arrest patients prospectively underwent bedside cardiac sonography at 4 emergency medicine residency–affiliated EDs as part of the Sonography Outcomes Assessment Program. Cardiac arrest patients in pulseless electrical activity (PEA) and asystole underwent transthoracic cardiac ultrasound B-mode examinations during their resuscitations to assess for the presence or absence of cardiac kinetic activity. Several end points were analyzed as potential predictors of resuscitations: presenting cardiac rhythms, the presence of sonographically detected cardiac activity, prehospital resuscitation time intervals, and ED resuscitation time intervals. Of 70 enrolled subjects, 36 were in asystole and 34 in PEA. Patients presenting without evidence of cardiac kinetic activity did not have return of spontaneous circulation (ROSC) regardless of their cardiac rhythm, asystole, or PEA. Of the 34 subjects presenting with PEA, 11 had sonographic evidence of cardiac kinetic activity, 8 had ROSC with subsequent admission to the hospital, and 1 had survived to hospital discharge with scores of 1 on the Glasgow-Pittsburgh Cerebral Performance scale and 1 in the Overall Performance category. The presence of sonographically identified cardiac kinetic motion was associated with ROSC. Time interval durations of cardiac resuscitative efforts in the prehospital environment and in the ED were not accurate predictors of ROSC for this cohort. Cardiac kinetic activity, or lack thereof, identified by transthoracic B-mode ultrasound may aid physicians’ decision making regarding the care of cardiac arrest patients with PEA or asystole. D 2005 Elsevier Inc. All rights reserved.
Presented at the ACEP Scientific Assembly in Boston, MA, October 2003. * Corresponding author.
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.11.007 460 P. Salen et al.
1. Introduction initiated and transport cardiac arrest victims to the nearest ED; no change in transport policy was instituted for this Cardiac arrest results in the absence of signs of circulation, study. The institutional review board granted an exemption including the absence of a pulse. The pulse check has been of informed consent with the caveat that there were the bgold standardQ usually relied on by clinicians to explicit instructions in the study protocol that cardiac US determine if the heart is or is not beating [1]. Absence of a results do not influence the standard American Heart pulse has been used to indicate the need to place automated Association decision-making guidelines in the care of the external defibrillator leads and initiate chest compressions. cardiac arrest subjects. Pulseless subjects underwent an Unfortunately, the pulse check is neither a rapid nor a reliable initial cardiac US examination on presentation and means for detecting the presence or absence of circulation for sequential examinations every 3 to 5 minutes during the both laypersons and healthcare providers [1]. Multiple resuscitation at the time during which carotid pulse checks studies have concluded that as a diagnostic test to assess were carried out to determine if there was cardiac kinetic for perfusion during suspected cardiac arrest, the pulse check activity. Sonographic evidence of cardiac kinetic activity has serious limitations in specificity, sensitivity, and accuracy was defined as any detected motion within the heart: atrial, [2-4]. The absence of a palpable pulse does not always reflect valvular, or ventricular. Emergency physician sonographers cardiac standstill during cardiac resuscitation as inefficient used the transthoracic, subxiphoid, or the parasternal long- cardiac contractions occur because of many factors, includ- axis B-mode–view approaches to visualize the heart with ing cardiac tamponade, pulmonary embolism, and pneumo- 3.5-MHz curvilinear or sector probes. The end points of thorax [5]. Because some causes of inefficient cardiac the resuscitations were death, return of spontaneous contractions are more reversible than others, (ie, hypo/ circulation (ROSC), admission to the hospital, survival to hyperkalemia, hypovolemia, hypothermia, acidosis, cardiac discharge from the hospital, and functional status after tamponade), a tool that would allow for early identification of discharge. Return of spontaneous circulation was defined groups of pulseless patients with more or less likelihood of as resumption of a palpable pulse and blood pressure, even resuscitation would be welcome [1,6]. Ultrasound (US) is if transient. Recorded outcome variables were demographic well suited to bedside ED use in the care of critically ill data, the presence of PEA (evidence of electrical rhythm) patients and provides rapid, noninvasive, potentially useful or asystole, results of cardiac US examinations (evidence information regarding the patient’s clinical condition. The or absence of US-detected cardiac kinetic activity), importance of transesophageal cardiac US has been recog- whether the cardiac arrest occurred in the prehospital nized in the American Heart Association guidelines for environment or the ED, estimated duration of prehospital differentiating patients with true electromechanical dissoci- cardiac arrest resuscitation, duration of ED cardiac arrest ation (EMD; pulseless patients with cardiac electrical activity resuscitations, cardiac US evidence of pericardial effusion, but no cardiac contractility) from patients with pseudo-EMD and resuscitation outcomes. Whenever applicable for our (pulseless patients with cardiac electrical activity and cardiac inhospital study design, data were collected according to contractility), a group that should be aggressively treated the recommended guidelines for uniform reporting of data [1,6]. Recent reports have suggested that transthoracic from out-of-hospital cardiac arrest, the Utstein Style [9]. cardiac sonography might assist in identifying patients with All data were contemporaneously noted on a standardized potentially treatable conditions when pulses are not palpable, data collection forms along with a still image of the but its role in the setting of cardiac arrest is unclear [7,8]. This cardiac US. Statistical analysis included descriptive statis- multicenter trial examined the use of cardiac sonography as tics, 95% confidence intervals based on proportions, an adjunctive diagnostic aid and as a predictor of resuscita- univariate logistic regression, and receiver operating tion outcomes in cardiac arrest patients with pulseless characteristic curve analysis; a was set at .05, 2-tailed. electrical activity (PEA) and asystole.
3. Results 2. Methods During a 12-month period, 70 subjects were enrolled, This institutional review board–approved multicenter each center with at least 10 subjects. Demographically, the trial, carried out at 4 academic EDs with volumes of at least cohort consisted of 61% males and 39% females with an age 40000 visits per year and served by extensive emergency range of 16 to 94 years. Resuscitations were begun in the medical services systems, all with both basic life support and prehospital environment for 67 of 70 cardiac arrest subjects; advanced life support units but with advanced life support the prehospital resuscitation intervals ranged from 5 to units transporting cardiac arrest patients. These study sites 77 minutes. Resuscitation duration intervals for 60 of the prospectively enrolled a convenience sample of nontrauma 70 subjects lasted less than 12 minutes in the ED. Within the pulseless adult subjects in cardiac arrest as part of the asystole cohort, 31 of 36 resuscitation intervals lasted Sonography Outcomes Assessment Program. In all 4 centers, 12 minutes or less in the ED; for the PEA cohort, 29 of the policy and practice was to continue resuscitation once 34 lasted 12 minutes or less. As a point of comparison to the Sonographically identified cardiac activity prediction of cardiac arrest resuscitation 461
Table 1 + Sonographic + Sonographic À Sonographic À Sonographic cardiac activity cardiac activity cardiac activity cardiac activity Rhythm PEA Asystole PEA Asystole ROSC + 8a 000 ROSC À 302336 a Only 1 of 8 survived to hospital discharge with scores of 1 on the Glasgow-Pittsburgh Cerebral Performance scale (good cerebral performance) and 1 in the Overall Performance category (capable of normal life) [9]. predictive abilities of cardiac sonography, a logistic regres- positives [US identification of cardiac activity with no sion model was conducted to control for cardiac arrest ROSC] + false negatives [US identification of no cardiac resuscitation intervals both in the prehospital and ED activity with ROSC] + true negatives) of cardiac US to environments to correlate resuscitation times with ROSC. predict ROSC vs death in our cardiac arrest cohort The univariate logistic regression model for prehospital and was 0.95 (95% CI, 0.92-1.0). The accuracy of cardiac ED resuscitation intervals showed correlation for 4 (52%) of US to predict survival to hospital discharge vs death was the 7 patients that had ROSC and subsequent admission to 0.87 (95% CI, 0.79-0.95). the hospital with an area under the curve of 0.26 (95% CI, 0-0.52; P = NS). Therefore, duration of resuscitation time intervals of our cardiac arrest cohort was not significantly 4. Discussion associated with ROSC. The presenting rhythms before US were asystole in 36 The resuscitation of ED patients with PEA or asystole and PEA in 34 (Table 1). Serial cardiac USs, more than a may entail considerable time and effort with only a very single US examination, were carried out in 53 of the small chance of success. Unfortunately, there is no subjects. Single cardiac US examinations were carried out consensus as to when the resuscitation should be terminated in 17 of the subject: 9 for asystole subjects and 8 for PEA or continued. Time to cardiopulmonary resuscitation, time subjects. Eight subjects presenting with PEA had ROSC to defibrillation, comorbid diseases, precardiac arrest and survived long enough for admission to the intensive condition, initial arrest rhythm, and duration of resuscitative care unit; all subjects with asystole died. Fifty-nine effort have been reported to have prognostic importance for subjects had no evidence of sonographically identifiable cardiac arrest patients [10]. A time-efficient, reliable means cardiac kinetic activity and their resuscitations were of determining those patients who are potentially salvage- uniformly unsuccessful. Subjects with asystole on present- able from cardiopulmonary resuscitation from those who are ing rhythm strip uniformly lacked sonographic evidence not would be of great clinical use. myocardial contractions and did not have ROSC. Twenty- Although there are few studies on the use of transthoracic three subjects (68%) presenting with PEA had no sono- B-mode cardiac US to predict the outcome of cardiac arrest, graphic evidence of cardiac activity (true EMD) and did this study is consistent with prior data that showed that not have ROSC. The presence of sonographically identi- patients with sonographically identified cardiac standstill do fied cardiac kinetic activity (pseudo-EMD) during the not have ROSC regardless of the initial cardiac electrical resuscitation was identified in 11 PEA subjects (32%). rhythm [7,8]. Cardiac sonography confirmed lack of cardiac Eight of this pseudo-EMD subgroup had sonographic kinetic motion in our entire asystole cohort. It also evidence of cardiac activity and also had ROSC with distinguished between pseudo-EMD and true EMD in our 6 surviving for less than 12 hours after intensive care PEA cohort. Our entire true-EMD cohort like our asystole admission. Two subjects, who were noted to have cardiac cohort did not have ROSC, which is consistent with activity both in their presenting sonographic examination previously published data regarding true EMD [5,7]. Our and all subsequent examinations, survived a more than data further support previously published data that sono- 24-hour stay in the intensive care unit. One of these graphically identifiable cardiac kinetic activity is associated subjects survived to hospital discharge after hemodialysis with ROSC [8,10,11]. Specifically, our study suggests that for renal failure–induced hyperkalemia with scores of 1 on sonographic evidence of cardiac kinetic motion during the Glasgow-Pittsburgh Cerebral Performance scale (good cardiopulmonary resuscitation is a potential predictor of cerebral performance) and 1 in the Overall Performance ROSC (accuracy = 0.95, 95% CI, 0.92-1.0). This is evidenced category (capable of normal life). None of the subjects in by the fact that 8 of 11 subjects in our PEA cohort with cardiac the PEA cohort were noted to have a pericardial effusion kinetic motion on US examination had ROSC. The sono- or tamponade as a potential cause of their cardiac arrest. graphic determination of cardiac kinetic motion was also an The accuracy (true positives [US identification of cardiac accurate predictor of survival to hospital discharge vs death activity with ROSC] + true negatives [US confirmation of (accuracy = 0.87, 95% CI, 0.79-0.95); 1 subject survived to no cardiac activity and no ROSC] / true positives + false hospital discharge with scores of 1 on the Glasgow-Pittsburgh 462 P. Salen et al.
Cerebral Performance and Overall Performance scales [9]. 5. Conclusion Other physician investigators studying pseudo-EMD both invasively and with US have noted pseudo-EMD as These data support the idea that transthoracic B-mode proportion of the total PEA population to range from 42% cardiac sonography can provide physicians involved in the to 86% [5,11-13]. It is notable that our pseudo-EMD resuscitation of pulseless patients with asystole and PEAwith population, which was primarily a population that had their information that may help in making pertinent management resuscitations started in the prehospital environment, only decisions regarding cardiopulmonary resuscitation. constituted 23% of our PEA cohort. Previously suggested parameters for the prediction of ROSC, including prehospital resuscitation time intervals and References ED resuscitation time intervals, were not reliable predictors for this cohort [14]. For this data set, there was no correlation [1] Stapleton ER, Aufderheide TP, Hazinski MF, Cummins RO. Adult between ROSC and time duration of resuscitation for patients CPR. BLS Healthcare Providers 2001:75. with PEA. These data suggest that duration of resuscitation [2] Ochoa FJ, Ramalle-Gomara E, Carpintero JM, Garcia A, Saralegui I. Competence of health professionals to check the carotid pulse. may be a more useful predictor in the setting of cardiac arrests Resuscitation 1998;37:173-5. secondary to ventricular tachyarrhythmias than for subjects [3] Flesche CW, Breuer S, Mandel LP, Beivik H, Tarnow J. The ability with PEA and asystole. of health professionals to check the carotid pulse. Circulation 1994; Important limitations of our study are notable. The total 90:I-288. sample size of 70 was relatively modest and limits the [4] Eberle B, Dick WF, Schneider T, Wisser G, Doetsch S, Tzanova I. Checking the carotid pulse: diagnostic accuracy of first responders in confidence of our speculations that the presence or absence patients with and without a pulse. Resuscitation 1996;33:107-16. of sonographically detected cardiac activity can predict [5] Paradis NA, Martin GB, Goetting MG, Rivers EP, Feingold M, Nowak positive or negative outcomes. The predictive accuracy of RM. Aortic pressure during human cardiac arrest: identification of cardiac standstill on US appears to be reliable; however, a pseudo-electromechanical dissociation. Chest 1992;101(1):123-8. large sample size of subjects with pseudo-EMD would be [6] Sanders AB, Kern KB, Berg RA. Searching for a predictive rule for terminating cardiopulmonary resuscitation. Acad Emerg Med 2001;8: required to confidently associate ROSC with the sono- 654-7. graphic detection of cardiac kinetic motion. Although the [7] Blaivas M, Fox JC. Outcome in cardiac arrest patients found to have multicenter nature of our trial may enhance the universality cardiac standstill on the bedside emergency department echocardio- of our results, the fact that our subjects were a convenience gram. Acad Emerg Med 2001;8:616-21. rather than a consecutive sample of cardiac arrest patients [8] Salen P, O’Connor R, Sierzenski P, et al. Can cardiac sonography and oceanography be used independently and in combination to predict should be noted. This may be partly explained by (1) the resuscitation outcomes? Acad Emerg Med 2001;8:610-5. availability of the US machine and (2) the availability of [9] Cummins RO, Chamberlain DA, Abramson NS, et al. Recommended the study participant physician sonographers within their guidelines for uniform reporting of data from out-of-hospital cardiac respective departments. Further 17 of 70 subjects did not arrest: the Utstein Style. Circulation 1991;84(2):960-75. receive sequential sonographic examinations after initial [10] Part 6: Advanced Cardiovascular Life Support: Section 7 Algo- rithm Approach to ACLS Emergencies. August 2000. Circulation cardiac US examination, or the sequential examinations 2000:I-150-2 [Suppl I]. were not recorded on the data sheets. This is partially [11] Bocka JJ, Overton DT, Hauser A. Electromechanical dissociation in mitigated by the fact that of the 70 subjects, 60 had human beings; an echocardiographic evaluation. Ann Emerg Med resuscitations lasting less than 12 minutes. The sono- 1988;17(5):450-2. graphic cardiac examinations were not videotaped and [12] Mayron R, Gaudio FE, Plummer D, Asinger R, Elsperger J. Echocardiography performed by emergency physicians: impact on were not reviewed for accuracy. Finally, although there diagnosis and therapy. Ann Emerg Med 1988;17(2):73-7. were instructions in the study protocol that cardiac US [13] Tayal VS, Kline JA. Emergency echocardiography to detect pericar- results do not influence the standard American Heart dial effusions in patients in PEA and near PEA states. Resuscitation Association decision-making guidelines, the study was not 2003;59:315-8. blinded in that the sonographers were aware of the [14] Part 2: Ethical aspects of CPR and ECC. August 2000. Circulation 2000:I 14-9 [Suppl I]. progress and results of the resuscitative efforts. American Journal of Emergency Medicine (2005) 23, 463–467
www.elsevier.com/locate/ajem
The effect of pesticide spraying on the rate and severity of ED asthma
Bonnie Chin-Yet O’Sullivan MDa, John Lafleur MDa, Kirsa Fridal MDa, Stephen Hormozdi MDa, Steve Schwartz MDb, Mark Belt MDa, Madelon Finkel PhDc,* aDepartment of Emergency Medicine, Lincoln Medical and Mental Health Center, Bronx, NY 10451, USA bDepartment of Emergency Medicine, North Shore Medical Center, Long Island, NY 11030, USA cDepartment of Public Health, Weill Medical College, Cornell University, New York, NY 10021, USA
Received 26 August 2004; accepted 7 December 2004
Abstract We report on the incidence of emergency department (ED) asthma presentations and admissions to the Lincoln Hospital, located in the South Bronx of New York City, during the 1999 eradication program of the mosquito vector for West Nile virus. Spraying of Malathion and Resmethrin occurred in the hospital’s geographic area over 4 days in September 1999. During that time, 1318 pediatric and adult patients were seen in the ED for asthma-related symptoms. Of these, 222 (16.8%) were hospitalized. Emergency department visits, during days when spraying occurred, were compared with visits during days when no spraying occurred. Comparisons were made with previous years as a reference point. Findings showed that the spraying of insecticides did not increase the rate or severity of asthma presentations as measured by the Lincoln Hospital’s ED asthma census or hospital admissions for asthma. D 2005 Elsevier Inc. All rights reserved.
1. Background tion by veterinarians and wildlife specialists found pathological evidence of a Flavivirus later identified as The first known outbreak of the West Nile virus in the West Nile virus [1,2]. western hemisphere occurred in the summer of 1999 in an The DOH, after careful consideration of the risks of area in northern Queens, a borough of New York City. The spraying to the general public, responded by initiating a New York City Department of Health (DOH) received a citywide spraying campaign in an effort to control the report of an unusual cluster of cases of meningoenceph- outbreak. This decision was based on its assessment that alitis associated with muscle weakness suggestive of an the potential for risks to the general population was low. The arboviral cause. An environmental investigation was spraying during the fall of 1999 occurred under emergency undertaken and, simultaneously, an independent investiga- conditions and thus was considered exempt from the State and City Environmental Quality Review Acts. Ultralow volume (0.6-3.0 oz/acre quantities) Malathion was sprayed * Corresponding author. Tel.: +1 212 746 1257; fax: +1 212 746 8544. from helicopters and Resmethrin was sprayed from trucks E-mail address: [email protected] (M. Finkel). (0.11-0.34 oz/acre) [3].
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.12.017 464 B.C.-Y. O’Sullivan et al.
2. Insecticide toxicity likely to have adverse health consequences to those exposed. Indeed it was thought that there would be less Eradication of mosquito vectors may be accomplished by significant risks from spraying than there would be risks by spraying of insecticides, although these have known toxicity transmission of the West Nile virus from mosquito bites to human beings depending on the dosage and extent of [15]. It was assumed that any adverse effect of pesticide exposure [4,5]. Malathion is an organophosphate insecticide spraying would be evident by an increase in visits for that can be absorbed through virtually any route [6]. asthma to the Lincoln Hospital’s adult and pediatric EDs. Organophosphates exert their toxicity primarily by the inhibition of acetylcholinesterase, resulting in excess accu- mulation of acetylcholine at the synapse. Organophosphate 4. Study area toxicity compromises the airway via a muscarinic effect, causing bronchorrhea and bronchospasm. Onset of symp- The densely populated South Bronx area of New York toms is most rapid after inhalation and least rapid after City has one of the highest rates of hospitalization and death percutaneous absorption. Most patients become symptom- from asthma [16-18] Hospitalization for asthma among atic within 12 hours of exposure. Routine occupational those younger than 14 years in this area was approximately exposure to Malathion regularly causes wheezing even in 23 per 1000 in 1997. The New York State, by contrast, had patients without asthma [7,8]. about one tenth the hospital admissions for asthma in this Resmethrin is a man-made pyrethroid insecticide found to age group (2.4 per 1000) [19]. The ED at the Lincoln be effective against adult mosquitoes including Culex, the Hospital sees approximately 14500 asthma visits per year, species most commonly associated with the West Nile virus. resulting in 1500 annual hospital admissions for this disease Pyrethroids, synthetic analogs of natural pyrethrins, are and representing about 5% of all asthma hospitalizations in active extracts from Chrysanthemum cinerariaefolium [9]. New York City [20]. The ED sees a full spectrum of The major site of action of all pyrethroids has been shown to asthmatic symptoms, not just the most acute presentations, be a voltage-dependent sodium channel. Poisoning syn- because most asthmatic patients tend to use the ED for both dromes manifest as reflex hyperexcitability, fine tremor, primary and emergency care. salivation, choreoathtosis, and seizures [10]. In general, During the mosquito eradication program in 1999 in this pyrethroids are much less acutely toxic to rats and presum- area, 5349 gallons of Adulticide were used in the West Nile ably to human beings than are the natural pyrethrins. eradication effort, and Malathion made up 4561 gallons, or Although the mechanism of toxicity of pyrethroids is not about 80%, of that [21]. Malathion was sprayed from known to induce bronchospasm, wheezing in response to helicopters and Resmethrin from trucks. The South Bronx, tetramethrin (a pyrethroid) has been reported [11].In which contains a large majority of the Lincoln Hospital’s addition, there is a well-documented bronchospastic reaction patient population, was sprayed on 4 days in September that occurs with acute exposures to pyrethrins in sensitized 1999: on the 11th, 12th, 13th, and again on the 20th. patients, as well as several case reports of fatality from Spraying occurred during evening hours and continued past bronchospasm associated with a pyrethrin shampoo [12,13]. midnight into the early hours of the following morning. The effect of spraying of insecticides on the health of the exposed population was of concern to the DOH. In particular, the risks of exacerbating airway disease among 5. Methods vulnerable populations, notably among asthmatics, needed to be taken into account. Only one study in the literature A retrospective review of all pediatric and adult asthma examined the risk posed to asthmatics by insecticide visits and asthma hospital admissions for August and spraying over urban areas and findings showed neither an September of 1999 was undertaken. Included were all increase in shortness of breath after spraying of Malathion in patients with a primary diagnosis of asthma from the ED. Santa Clara County, California, nor increased emergency To assess the usual patterns of asthma presentations to the department (ED) use for asthma-like symptoms [14]. ED, data were also collected on all adult and pediatric asthma ED visits for August and September of 1997 and 1998. Asthma presentations to and admission rates from the adult 3. Rationale for study and pediatric EDs were noted for the 11th, 12th, 13th, and 20th of September, the days the city conducted spraying in the This study seeks to ascertain adverse risks of insecticide South Bronx. These data were compared with the other days spraying in an area that has one of the highest rates of of the month when no spraying occurred. To account for the asthma in New York City. This study seeks to answer the fact that spraying was conducted before and after midnight on question bTo what extent did the spraying of insecticides the spraying days, we conducted the same analysis for the over the South Bronx in 1999 contribute to an increase in 12th, 13th, 14th, and 21st of September as well. A third asthmatic symptoms as measured by ED presentations?Q It analysis compared the 5 days after each spraying episode was hypothesized that the spraying of insecticides was not with the days of nonspraying to capture any delayed Pesticide spraying on the rate and severity of ED asthma 465
Fig. 1 Emergency department asthma visits (adult and pediatric). presentation to the ED. Two-tailed Student t tests with a set at of 1318 patients presented to the Lincoln Hospital ED with a .05 were used to calculate statistical significance. primary diagnosis of asthma exacerbation (753 or 57.1% adult patients and 565 or 42.9% pediatric patients). Of those patients, 222 (16.8%) were admitted (102 or 45.9% adult 6. Results patients and 120 or 54.1% pediatric patients). When rates of admission for asthma in September 1999 were compared Fig. 1 shows asthma presentations to the adult and with those in September 1997 and September 1998, no pediatric EDs for August and September of 1997, 1998, and significant differences were seen. Twenty-four percent of 1999. The expected seasonal increase in asthma was pediatric patients seen in the ED for asthma were admitted consistent for each year. During September 1999, a total in 1997 vs 22% in 1998 and 27% in 1999. Among adult
Fig. 2 Comparison of daily visits and admission for spraying vs nonspraying days. 466 B.C.-Y. O’Sullivan et al.
Fig. 3 Adult asthma visits (3-day running average). patients seen in the ED for asthma, approximately 16% were on the nonspraying days and 13.0% on the spraying days. The admitted to the hospital during all 3 periods. percentage of pediatric admissions on nonspraying and The data show that in 1999, the adult ED evaluated spraying days was 20.7% and 25.4%, respectively. approximately 25.1 patients with asthma per day on the A 5-day running average of patient ED asthma visits was nonspraying days and 25 patients per day on the spraying calculated to account for possible delayed effects of days. In the pediatric ED, 19.3 asthma patients were spraying; no significant difference between spraying and evaluated per day on the nonspraying days and 15.8 nonspraying days was seen. Figs. 3 and 4 show adult and patients on the spraying days. These differences are not pediatric ED visits during September 1999. There were no statistically significant. statistically significant differences between spraying and Data collected on the 12th, 13th, 14th, and 21st of nonspraying days in September 1999. September show that an average of 22 asthma patients per day was evaluated in the adult ED whereas 13 patients per day were evaluated in the pediatric ED. Fig. 2 illustrates the 7. Discussion average number of ED asthma visits and hospital admissions per day for the adult and pediatric EDs in September 1999. Since the first outbreak of the West Nile virus in New The admission rate was used as an indicator for the severity of York City in 1999, the virus has spread across most of illness. The percentage of total adults with asthma who were the United States and has become enzootic and endemic admitted to the hospital during September 1999 was 13.6% in the central and eastern parts of the country and
Fig. 4 Pediatric asthma visits (3-day running average). Pesticide spraying on the rate and severity of ED asthma 467 increasingly in the West; there have been more than 200 References known deaths in the United States from this arbovirus, primarily among the elderly individuals [22]. Transmis- [1] Nash D, Mostashari F, et al. The outbreak of West Nile virus infection sion of the West Nile virus has been found to occur in the New York City area in 1999. N Engl J Med 2001;344(24): through organ donation, blood transfusions, and breast 1807-14. [2] Lanciotti RS, Roehrig JT, Deubel V, et al. Origin of the West Nile feeding [23]; intrauterine transmission also has been virus responsible for an outbreak of encephalitis in the northeast recently reported [24]. United States. Science 1999;286:2333-7. Certainly, the rapid spread of this dangerous disease is a [3] Novello AC. West Nile virus in New York State: the 1999 outbreak cause for concern, and public health measures, including and response plan for 2000. Viral Immunol 2000;13(4):463-7. eradication of its mosquito vector, seem warranted. Yet, in [4] O’Malley M. Clinical evaluation of pesticide exposure and poison- ings. Lancet 1997;349:1161-6. light of the very high burden of asthma morbidity and [5] Al-Saleh. Pesticides. J Environ Pathol Toxicol Oncol 1994;13(3): mortality in the South Bronx, the spraying of insecticides in 152-4. this inner city, an urban population with a high prevalence [6] Aaron CK, Howland MA. Insecticides: organophosphates and of asthma, was a cause for concern given the potential for carbamates. In: Goldfrank LR, Flomenbaum NE, Lewin NA, et al, 7 exacerbation of the disease. Findings from this study editors. Goldfranks Toxicological Emergencies. 5th ed. New York McGraw Hill; 1998. p. 1105 - 14. showed no acute effect on asthma patients after insecticide [7] Deschamps D, Questel F, Baud FJ, et al. Persistent asthma after acute spraying. The New York City DOH reported on citywide inhalation of organophosphate insecticide. Lancet 1994;344:1712. statistics for hospital-based treatment of asthma during the [8] Newton JG, Breslin AB. Asthmatic reactions to a commonly used mosquito eradication of September 1999 and found no aerosol insect killer. Med J Aust 1983;1:378-80. significant increases in hospital admissions or acute [9] Howland MA. Insecticides: chlorinated hydrocarbons, pyrethrins, and DEET. Goldfrank. 5th edition. p. 1121-26. outpatient asthma care during the period of interest [25]. [10] Ray DE, Forshaw PJ. Pyrethroid insecticides: poisoning syndromes, Although this study showed no increase in asthma-like synergies, and therapy. Clin Toxicol 2000;38(2):95-101. symptoms as a result of spraying, we acknowledge [11] Vandenplas O, Delwiche JP, Auverdin J, et al. Asthma to tetramethrin. limitations that may have influenced the results. Our Allergy 2000;55(4):417-8. methodology does not allow us to assess individual [12] Wagner SL. Fatal asthma in a child after use of an animal shampoo containing pyrethrin. West J Med 2000;173(2):86-7. exposures to insecticides. Also, because of the small number [13] Wax PM, Hoffman RS. Fatality associated with inhalation of a of spraying days and a changing baseline caused by pyrethrin. J Toxicol Clin Toxicol 1994;32(4):457-60. seasonal variability, subtle acute effects of spraying on [14] Kahn E, Berlin M, Deane M, et al. Assessment of acute health effects asthmatic airways could have been missed. Possible from the Medfly Eradication Project in Santa Clara County, confounders include other factors relating to air quality California. Arch Environ Health 1992;47(4):279-84. [15] Lopez W, Miller J. The legal context of mosquito control for West such as ozone and particulate matter and epidemics of viral Nile virus in New York City. J Law Med Ethics 2002;30(3 suppl): upper airway infections. Finally, the epidemiology of 135-8. asthma is complicated, multifactorial, and incompletely [16] Claudio L. New asthma efforts in the Bronx. NIEHS News. Environ understood. The long-term effects on individuals with Health Perspect 1996;104(10):1028-9. asthma or on those at risk for developing asthma of [17] De Palo VA, Mayo PH, Friedman P, Rosen MJ. Demographic influences on asthma hospital admission rates in New York City. exposure to chemicals such as Malathion and Resmethrin Chest 1994;106(2):447-51. are difficult if not impossible to predict. [18] Crain EF, Weiss KB, Bijur PE, et al. An estimate of the prevalence of Given the high concentration of asthma patients in the asthma and wheezing among inner-city children. Pediatrics Lincoln Hospital population and the high volume of 1994;94(3):356-62. asthma patients treated at the ED, it is unlikely that a [19] Asthma facts. New York (NY): New York City Department of Health; 1999. significant number of patients in the South Bronx would [20] Adult mosquito control programs FEIS. vol. 3c. New York: develop clinically significant symptoms of bronchospasm Department of Health; 2001. p. 116. without there being a corresponding increase in emergency [21] Adult mosquito control programs FEIS, vol. 2. Department of Health;, asthma presentations to the Lincoln Hospital. We feel 2001. p. 24. confident that the sample of patients presenting to the [22] Chow CC, Montgomery SP, O’Leary DR, et al. Provisional Surveillance Summary of the West Nile virus epidemic—United Lincoln Hospital with asthma during the study period is States, January-November 2002. JAMA 2003;289(3):293-5. representative of the experience of individuals with asthma [23] Ongan A, Boulton ML, Somsel P, et al. Possible West Nile virus in the South Bronx during September 1999. Therefore, we transmission to an infant through breast feeding—Michigan, 2002. are confident that the data presented accurately portray the JAMA 2002;288(16):1976-7. negligible effect of insecticide spraying on this study [24] Nguyen Q, Morrow C, Novick L, et al. Intrauterine West Nile virus infection—New York, 2002. JAMA 2003;289(3):295-6. population as reflected by those seeking care for asthma [25] Adult mosquito control programs FEIS. vol. 3c. Department of during the spraying. Health;, 2001. p. 104. American Journal of Emergency Medicine (2005) 23, 468–473
www.elsevier.com/locate/ajem
Changes in cardiac troponin T measurements are associated with adverse cardiac events in patients with chronic kidney disease
Jin H. Han MDa,*, Christopher J. Lindsell PhDb, Richard J. Ryan MDa,W.BrianGiblerMDa aDepartment of Emergency Medicine, University of Cincinnati Medical Center, Box 670769, Cincinnati, OH 45267-0769, USA bThe Institute for the Study of Health, University of Cincinnati Medical Center, Cincinnati, OH, USA
Received 21 December 2004; accepted 21 January 2005
Abstract The purpose of this study was to determine whether long-term and short-term changes in cardiac troponin T (cTnT) were associated with adverse cardiac events (ACEs) in patients with chronic kidney disease. Long-term changes were defined as changes in cTnT between ED visits, and short- term changes were defined as changes between 2 consecutive serial cTnT measurements within an ED visit. A retrospective chart review of patients with chronic kidney disease with suspected acute coronary syndromes presenting to the ED between December 1999 and November 2003 was conducted. The primary outcome variable was an ACE which was a composite endpoint consisting of a discharge diagnosis of acute myocardial infarction, unstable angina, revascularization, cardiac dysrhythmias, all-cause mortality, or congestive heart failure exacerbation. The primary predictor of ACE abstracted from the charts was the initial cTnT measured during each ED presentation. There were 90 patients with 397 visits enrolled in the study. Using a mixed-models analysis of variance, cTnT was higher in the ACE group than in the non-ACE group (difference in log cTnT = 0.054, 95% CI 0.006-0.101) after adjusting for age, race, sex, dialysis status, and smoking history. No other variables were found to be associated with cTnT. To evaluate the clinical significance of acute changes in cTnT, a secondary analysis was performed on 64 patients with an initial cTnT measurement above 0.10 ng/mL. For in-hospital and 30-day ACE, a short-term increase in cTnT of 0.11 ng/mL had a positive likelihood ratio of 13.3 and 11.9, respectively. Long-term and short-term increases in cTnT are associated with an ACE. D 2005 Elsevier Inc. All rights reserved.
1. Introduction
Patients with chronic kidney disease who present to the ED with symptoms suggestive of acute coronary syndromes (ACS) are a diagnostic challenge. In this population, an Poster presentation at the Society of Academic Emergency Medicine, Orlando, Fla, May 17, 2004. elevated cardiac troponin T (cTnT) is often thought to be * Corresponding author. Tel.: +1 513 558 7917. nonspecific; elevated cTnT has been observed in both E-mail address: [email protected] (J. Han). dialyzed and nondialyzed patients with chronic kidney
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2005.01.006 Changes in cTnT measurements are associated with ACES in patients with chronic kidney disease 469 disease in the absence of any obvious myocardial necrosis than 2.0 mg/dL (estimated creatinine clearance b30 mL/min) [1,2]. As a result, it is unclear what impact elevated cTnT for at least 6 months before enrollment. Patients were should have on clinical decision making within the excluded if they had received a kidney transplant, died emergency setting in patients with chronic kidney disease secondary to trauma, or had terminal cancer. Terminal cancer and suspected ACS. was defined as metastatic disease requiring ongoing chemo- In patients without chronic kidney disease, an elevated therapy, radiation therapy, or palliative care. An interim cTnT is both indicative of myocardial infarction and analysis was conducted with 90 patients; because our prognostic for poor cardiovascular outcomes [3,4]. These primary hypothesis was adequately answered, no additional results are difficult to extrapolate to the patient with chronic patients were enrolled. The final sample included patients kidney disease and suspected ACS; although an elevated with consecutive medical record numbers for ED visits cTnT in an asymptomatic patient with chronic kidney disease between December 1999 and November 2003. is prognostic for long-term mortality [5], elevated cTnT is To test the secondary hypothesis, we included patients less sensitive and specific to ACS in patients with chronic from this cohort with an initial cTnT level greater than kidney disease than in patients with normal renal function [6]. 0.10 ng/mL and 2 consecutive cTnT markers within a In patients with chronic kidney disease, measuring 12-hour period. This represents the patient population in changes in cTnT may be more useful in diagnosing ACS which there is greatest uncertainty as to how to interpret the than interpreting a single absolute measurement. In patients cTnT; 0.10 ng/mL is the upper limit of normal at our with normal renal function, a typical rise and fall in cTnT is institution, and values below this would be considered observed in patients with an acute myocardial infarction [7]. negative whether the patient had known chronic kidney Theoretically, these changes should be observed in patients disease or not. If the patient had multiple visits meeting with chronic kidney disease. Such changes could be inclusion and exclusion criteria, the first ED visit was used. detected in 2 ways: (1) by having a prior, non–cardiac All cTnT measurements were performed using the event–related measure of cTnT against which to compare a Elecsys 2010 (Roche Diagnostics Inc, Indianapolis, Ind), suspected cardiac event–related cTnT, or (2) by measuring which uses a third-generation cTnT immunoassay. In our cTnT serially over several hours. Therefore, we sought to laboratory, the maximum coefficient of variation measured determine whether long-term and short-term changes in during this period was 9.5% for a cTnT value of 0.09 ng/mL cTnT levels were useful in diagnosing adverse cardiac and 5.0% for a cTnT value of 0.42 ng/mL. Patients were events (ACEs). Because rapid diagnosis of ACS can drive followed using medical record review and the Social initial therapy and improve patient outcomes, our study was Security Death Index for adverse events during hospitaliza- designed primarily to test the hypothesis that comparing a tion. Chart review used predefined data definitions and a single baseline ED measurement to a prior cTnT measure- standardized case report form. Data were entered into a ment known not to be related to ACS would improve customized Microsoft Access database (Microsoft Corp, diagnostic accuracy compared to interpretation of the Redmond, Wash), which included secondary data checks to absolute value. Our secondary hypothesis was that short- ensure all data values were within range and consistent with term changes in cTnT would be predictive of ACE. clinical course. The primary outcome variable was an ACE, which was a composite endpoint consisting of a discharge diagnosis of acute myocardial infarction, unstable angina, 2. Methods revascularization, cardiac dysrhythmias, all-cause mortality, or congestive heart failure (CHF) exacerbation. Diagnoses This was a retrospective cohort study with data obtained were made by the primary admitting team if the patient was using chart review methodology. This study was approved admitted, or the emergency physician if the patient was by the institutional review board and was conducted at an discharged home from the ED. The diagnosis of acute urban, academic ED which has approximately 88000 myocardial infarction and unstable angina was usually patient visits per year. Patients were initially identified by based on evidence obtained from electrocardiograms, ICD-9 code indicative of chronic kidney disease (585, 586, cardiac catheterization, or noninvasive cardiac imaging. 403.01, 403.11, 403.91, 404.02, 404.03, 404.12, 404.13, Baseline data were summarized using means and 404.92, 404.93). Medical record numbers were obtained standard deviations for continuous variables, and frequen- from electronic hospital databases. These patients were then cies with proportions for categorical variables. Data were screened to determine whether they had a cTnT measured compared between adverse event and nonevent groups using during an ED visit. a v2 test or a Student t test where appropriate. Patients were selected based on the primary hypothesis. To determine the clinical significance of long-term They were included if they had presented to the ED with changes in cTnT, a mixed-models analysis of variance was symptoms suggestive of ACS, had at least 2 ED presenta- performed to evaluate the relationship between changing tions with suspected ACS presentations within a 12-month levels of cTnT within an individual and an ACE. A common period, had cTnT drawn during their ED visits, and had statistical approach to binary outcome data is to use logistic chronic kidney disease defined as a serum creatinine greater regression modeling the change in cTnT, the rate of change 470 J. Han et al. in cTnT, the percentage change in cTnT, or some other Table 2 Twelve-lead ECG, serum creatinine, and cTnT for derived variable as a predictor variable. This would involve each patient group assumptions as to the nature of the relationship between changing cTnT and outcome, and would not allow us to All Visits Visits visits without with consider all of the information contained in the dataset; ACE ACE patients made a variable number of visits to the ED at variable intervals. A mixed-models analysis of variance is an ECG Findings alternative statistical model, which takes full advantage of all Normal 13 (3.6) 12 (4.2) 1 (1.2) ST Elevation 3 (0.8) 2 (0.7) 1 (1.2) data, as well as minimizing assumptions about how the ST Depression 8 (2.2) 1 (0.4) 7 (8.5) manner of changing cTnT is related to outcomes. Using this T-wave 40 (10.9) 33 (11.6) 7 (8.5) statistical method, cTnT was considered the dependent inversion variable and the occurrence of an adverse event was an Hyperacute 1 (0.3) 0 (0.0) 1 (1.2) independent variable. This allowed us to consider the fixed T wave effects, such as sex and race, while at the same time LBBB 2 (0.5) 1 (0.4) 1 (1.2) considering the random effect of within-individual variation RBBB 5 (1.4) 2 (0.7) 3 (3.7) not related to a measured or controllable difference. Thus, we Hemiblock 5 (1.4) 1 (0.4) 4 (4.9) could include multiple visits for patients with different A-fib/flutter 5 (1.4) 1 (0.4) 4 (4.9) outcomes at each visit and with visits variably spaced over SVT 5 (1.4) 5 (1.8) 0 (0.0) time. The interpretation of this model is such that the Sinus 4 (1.1) 4 (1.4) 0 (0.0) bradycardia parameter estimate for the adverse event predictor variable is Nonspecific 112 (30.6) 92 (32.4) 20 (24.4) an estimate of the difference in cTnT between a visit with an ST changes ACE and a visit without an ACE for the same individual. An Unchanged 163 (44.5) 130 (45.8) 33 (40.2) unadjusted model and a model adjusted for demographics cTnT 0.14 (0.21) 0.12 (0.13) 0.23 (0.41) and medical history were computed. Because of positive Creatinine 6.82 (4.15) 7.32 (4.20) 4.58 (3.18) skew and kurtosis, a log transformation was used. Laboratory values are given as mean and SD and ECG findings as To evaluate the diagnostic accuracy of short-term changes frequency and percentage in cTnT, the change in cTnT between the first and second LBBB indicates left bundle branch block; RBBB, right bundle branch measurement during a single ED visit was computed. block; SVT, supraventricular tachycardia. Receiver-operating-characteristic curves were constructed, and the area under the curve was calculated using the nonparametric approach, and interval likelihood ratios were calculated. Sensitivity and specificity were assessed for the breakpoints in interval likelihood ratios. The prognostic value of acute changes in cTnT to predict in-hospital, 30-day, Table 1 Characteristics of the patient cohort, data are given and 6-month ACEs was tested in a similar manner. Thirty- as frequency and percentage except for age, which is given as day and 6-month ACEs were based on medical record and mean and SD Social Security Death Index review. Data analyses were Demographics performed using SPSS version 12.0 (SPPS Inc, Chicago, Ill) Age (y) 55.9 (13.4) and SAS version 8.2 (SAS Institute, Cary, NC). African American 75 (83.3) White 14 (15.6) Unknown 1 (1.1) 3. Results Male 47 (52.2) Female 43 (47.8) 3.1. Long-term changes in cTnT History Diabetes 52 (58.4) An interim analysis of 90 patients with 397 ED visits was HTN 81 (90.0) performed; because our primary hypothesis was answered, Hypercholesterolemia 11 (12.2) additional patients were not enrolled. The minimum number Prior CAD 31 (34.4) of ED visits per patient was 2 and the maximum was 11. CHF 30 (33.7) Patient demographics, past history, and dialysis status are PVD 14 (15.6) listed in Table 1. The mean (SD) age was 56 (13.4) years, Family history of CAD 28 (31.1) 83% were black and 48% were females. Thirty-seven Smoker 47 (52.2) percent of the patients were on hemodialysis and 3% of Dialysis the patients were on peritoneal dialysis at the start of the None 54 (60.0) study. The mean (SD) serum creatinine for the index visit Hemodialysis 33 (36.7) Peritoneal dialysis 3 (3.3) was 6.8 (4.2) mg/dL. Of the patients who were not on dialysis, 34 were placed on long-term dialysis during the HTN indicates hypertension; PVD, peripheral vascular disease. study period. Changes in cTnT measurements are associated with ACES in patients with chronic kidney disease 471
Table 2 describes laboratory values and ECG findings for Table 4 Changes in cTnT given in means and SDs seen all ED visits stratified by whether an adverse event was in patients with and without an event during hospitalization, evident or not. Eighty-three ACEs were recorded. Mean 30 days, and 6 months (SD) cTnT for visits without an ACE was 0.12 (0.13) ng/mL, Patients with no event Patients with an event whereas for visits with an ACE was 0.23 (0.41) ng/mL. Unadjusted cTnT was higher for visits associated with an In-hospital 0.04 (0.21) 0.47 (1.05) ACE than for visits that were not associated with an ACE 30 d 0.04 (0.21) 0.45 (1.03) (mean difference in log cTnT = 0.057, 95% CI 0.010-0.104). 6 mo 0.05 (0.24) 0.35 (0.92) Adjusted for demographics, medical history, and dialysis as given in Table 1, the mean difference in log cTnT was 0.054 The mean (SD) time between cTnT levels was 6.9 (95% CI 0.006-0.101). (2.3) hours. Table 4 shows the mean changes in cTnT in patients with and without ACEs during hospitalization, at 3.2. Short-term changes in cTnT 30-days, and at 6 months. The mean difference between Sixty-four patients had 2 cTnTs drawn within a 12- patients with and without an in-hospital ACE in the change hour period with an initial cTnT measurement greater than in cTnT was 0.43 (95% CI À0.01 to 0.87). For patients with 0.10 ng/mL. This cohort is described in Table 3. Thirty-nine and without 30-day ACEs, the difference was 0.40 (95% CI percent of the patients were on hemodialysis and 4% were on À0.02 to 0.83). For 6-month ACEs, the difference between peritoneal dialysis. Patients with a 6-month ACE tended to groups was 0.30 (95% CI À0.04 to 0.63). In all cases, there be older, African American, diabetic, and have previous was a tendency toward a greater increase in cTnT in the history of coronary artery disease (CAD) and CHF. ACE group than in the non-ACE group. The areas under the receiver-operating-characteristic curves for changes in cardiac troponin T were 0.63 (95% CI 0.48-0.78), 0.58 (95% CI 0.43-0.73), and 0.60 (95% CI Table 3 Characteristics and outcomes of the patient cohort for subanalysis looking at acute changes in cTnT. Data are 0.45-0.74) for in-hospital, 30-day, and 6-month events, given as frequency and percentage except for age, which is respectively. An increase in cTnT of 0.11 ng/mL was 36% given as mean and SD sensitive and 97% specific to in-hospital ACEs, with a All Patients Patients positive likelihood ratio of 13.3. For 30-day ACEs, an patients with no with a increase of 0.11 ng/mL had a sensitivity of 35%, a 6-mo 6-mo specificity of 97%, and a positive likelihood ratio of 11.9. event event For 6-month ACEs, an increase of 0.11 ng/mL had a Demographics sensitivity of 27%, a specificity of 96%, and a positive Age 54.9 (13.7) 48.1 (11.1) 61.0 (13.2) likelihood ratio of 7.2. African American 51 (79.7) 21 (77.8) 27 (79.4) White 12 (18.8) 5 (18.5) 7 (20.6) Unknown 1 (1.6) 1 (3.7) 0 (0.0) 4. Discussion Male 37 (57.8) 17 (63.0) 17 (50.0) Female 27 (42.2) 10 (37.0) 17 (50.0) Interpretation of absolute cTnT measurements in patients History with chronic kidney disease is problematic. These elevations Diabetes 39 (61.9) 14 (53.8) 25 (73.5) have been thought to be nonspecific [8] and have been HTN 55 (85.9) 23 (85.2) 29 (85.3) attributed to a reexpression of fetal cTnT in skeletal muscle Hypercholesterolemia 9 (14.1) 1 (3.7) 7 (20.6) [9]. Although the development of second- and third- Prior CAD 26 (40.6) 7 (25.9) 19 (55.9) Prior ACS 27 (42.2) 8 (29.6) 19 (55.9) generation cTnT immunoassays are believed to have CHF 24 (37.5) 6 (22.2) 17 (50.0) eliminated this problem, cTnT elevations have still been PVD 14 (21.9) 6 (22.2) 8 (23.5) reported in up to 20% of asymptomatic patients on Family history 17 (26.6) 7 (25.9) 9 (26.5) hemodialysis [2,10]. Similar elevations have been observed of CAD in patients with chronic kidney disease who are not on Smoker 30 (46.9) 13 (48.1) 16 (47.1) dialysis [1]. Our study suggests that measuring changes in Dialysis cTnT between ED visits and serial cTnT measurements None 36 (56.3) 13 (48.1) 22 (64.7) within an ED visit may improve our ability to diagnose Hemodialysis 25 (39.1) 11 (40.7) 12 (35.3) ACEs in patients with chronic kidney disease. Peritoneal dialysis 3 (4.7) 3 (11.1) 0 (0.0) The majority of studies investigating the relationship Adverse cardiac events between elevated cTnT and chronic kidney disease have In-hospital event 25 (40.3) 25 (73.5) 30-d event 26 (42.6) 26 (78.8) observed long-term outcomes in asymptomatic patients 6-mo event 34 (55.7) 34 (100.0) over a period of several years. However, little is known Baseline cTnT 0.37 (0.78) 0.22 (0.18) 0.48 (1.04) about how to interpret a single elevated cTnT in patients with chronic kidney disease and suspected ACS. Aviles 472 J. Han et al. et al [11] analyzed data from GUSTO-IV, a double-blind ED visits and in-hospital events. Analysis pertaining to our placebo-controlled trial evaluating the use of abciximab. secondary hypothesis was not sufficiently powered to detect Although these investigators found that subjects with a a significant association between acute changes in cTnT and creatinine clearance less 58.4 mL/min and a cTnT greater adverse events. However, the strong trend toward a positive than 0.10 ng/dL had an adjusted odds ratio of 2.5 for death association suggests a strong likelihood for a relationship and myocardial infarction within 30 days, only 11 of the that could be evaluated by a more thorough study of acute 7000 patients had end-stage renal disease, and the study changes in cTnT in patients with chronic kidney disease sample was a high-risk patient group [11]. Van Lente et al [6] being evaluated for ACS. evaluated absolute measurements of cTnT in a case-matched The limitations of a retrospective chart review are well prospective study of 51 patients with chronic kidney disease documented. Chart review methodology is prone to bias. To and suspected ACS. The diagnostic accuracy for in-hospital limit bias in this study, patients were selected based upon and 6-month adverse cardiac outcomes (myocardial infarc- strict adherence to inclusion and exclusion criteria. The tion, revascularization, positive angiography, CHF, and primary outcome variable was based upon the judgment of death) was significantly lower for patients with chronic the treating physicians and objective evidence of a cardiac kidney disease than for patients with normal renal function. event, and not chart reviewer opinion. Because not all For in-hospital outcomes, the optimal cTnT cutoff for patients underwent cardiac catheterization or noninvasive patients with chronic kidney disease was 0.50 ng/mL, with stress testing, there is a possibility of workup bias and a sensitivity of 29% and a specificity of 87% [6]. missed patients. There are limited data available concerning the signifi- Cardiac troponin T measurements were not blinded to the cance of long-term changes in cTnT for detecting ACEs. In treating physician, which may have also biased the results. asymptomatic patients on hemodialysis, Ooi et al [12] However, at our institution elevated cardiac troponin levels reported that patients with cTnT levels increasing by more are routinely considered bnonspecificQ in patients with than 60% over a period of several weeks had a relative risk chronic kidney disease, and ancillary studies such as cardiac of 2.0 for death. Wayand et al [13] measured cTnT levels at catheterization or stress cardiac imaging are usually 0, 4 weeks, 3 months, 6 months, and 1 year in asymptomatic performed to confirm ACS diagnosis. patients on dialysis. Patients with an ACE tended to have increasing cTnT levels [13]. Our data extend these findings by showing that for a 6. Conclusion patient with chronic kidney disease presenting to the ED with suspected ACS, an increased cTnT compared to a prior Changes in cTnT, both in the long-term and acutely, may non–ACS-related measure should not be ignored. Routinely be more useful than considering a single absolute value; a measuring a non–event-related cTnT may enhance the single elevated cTnT might be misconstrued as nonspecific emergency physician’s ability to interpret elevated cTnT in if interpreted individually. Our results suggest that increases this difficult patient population. in cTnT measured between ED visits and through serial Similarly, measuring acute changes in cTnT also cardiac biomarker measurements indicate higher risk in improves diagnostic accuracy. A typical rise and fall in patients with chronic kidney disease and suspected ACS. A cardiac biomarkers indicates myocardial necrosis in patients prospective study should be performed to validate these with normal renal function [7]. Our study found that an findings and better define the diagnostic cutoffs for both increase in cTnT of 0.11 ng/dL over several hours had 36% long-term and short-term changes in cTnT. sensitivity and 97% specificity for predicting in-hospital ACEs, indicating that the typical rise and fall of cardiac biomarkers seen in myocardial necrosis may indeed be References generalizable to the chronic kidney disease population. This [1] Lowbeer C, Stenvinkel P, Pecoits-Filho R, et al. Elevated cardiac analysis purposefully included patients with initial cTnT troponin T in predialysis patients is associated with inflammation and values above the diagnostic cutoff because it is in this predicts mortality. J Intern Med 2003;253(2):153-60. patient population that cTnT interpretation is problematic. [2] Apple FS, Murakami MM, Pearce LA, Herzog CA. Predictive value Only 25% of this cohort experienced an ACE, reflecting the of cardiac troponin I and T for subsequent death in end-stage renal poor specificity when using a single absolute value. disease. Circulation 2002;106(23):2941-5. [3] Hamm CW, Goldmann BU, Heeschen C, Kreymann G, Berger J, Meinertz T. Emergency room triage of patients with acute chest pain by means of rapid testing for cardiac troponin T or troponin I. N Engl 5. Limitations J Med 1997;337(23):1648-53. [4] Antman EM, Tanasijevic MJ, Thompson B, et al. Cardiac-specific This study has several limitations; we used retrospective troponin I levels to predict the risk of mortality in patients with acute coronary syndromes. N Engl J Med 1996;335(18):1342-9. chart review methodology with a relatively small sample [5] Dierkes J, Domrose U, Westphal S, et al. Cardiac troponin T predicts size. The study was stopped when a statistically significant mortality in patients with end-stage renal disease. Circulation 2000; association was found between increasing cTnT levels and 102(16):1964-9. Changes in cTnT measurements are associated with ACES in patients with chronic kidney disease 473
[6] Van Lente F, McErlean ES, DeLuca SA, Peacock WF, Rao JS, troponin T expression in skeletal muscle. Clin Chem 1997;43(6 Pt 1): Nissen SE. Ability of troponins to predict adverse outcomes in patients 976-82. with renal insufficiency and suspected acute coronary syndromes: a [10] Muller-Bardorff M, Hallermayer K, Schroder A, et al. Improved case-matched study. J Am Coll Cardiol 1999;33(2):471-8. troponin T ELISA specific for cardiac troponin T isoform: assay [7] Alpert JS, Thygesen K, Antman E, Bassand JP. Myocardial infarction development and analytical and clinical validation. Clin Chem 1997; redefined—a consensus document of The Joint European Society of 43(3):458-66. Cardiology/American College of Cardiology Committee for the [11] Aviles RJ, Askari AT, Lindahl B, et al. Troponin T levels in patients redefinition of myocardial infarction. J Am Coll Cardiol 2000;36(3): with acute coronary syndromes, with or without renal dysfunction. 959-69. N Engl J Med 2002;346(26):2047-52. [8] Li D, Keffer J, Corry K, Vazquez M, Jialal I. Nonspecific elevation of [12] Ooi DS, Zimmerman D, Graham J, Wells GA. Cardiac troponin T troponin T levels in patients with chronic renal failure. Clin Biochem predicts long-term outcomes in hemodialysis patients. Clin Chem 1995;28(4):474-7. 2001;47(3):412-7. [9] McLaurin MD, Apple FS, Voss EM, Herzog CA, Sharkey SW. [13] Wayand D, Baum H, Schatzle G, Scharf J, Neumeier D. Cardiac Cardiac troponin I, cardiac troponin T, and creatine kinase MB in troponin T and I in end-stage renal failure. Clin Chem 2000;46(9): dialysis patients without ischemic heart disease: evidence of cardiac 1345-50. American Journal of Emergency Medicine (2005) 23, 474–479
www.elsevier.com/locate/ajem
Moderate-to-severe blood pressure elevation at ED entry Hypertension or normotension?
Thomas Dieterle MD, Mace´ M. Schuurmans MD, Werner Strobel MD, Edouard J. Battegay MD, Benedict Martina MD*
Medical Emergency and Outpatient Department, University Hospital Basel, CH-4031 Basel, Switzerland
Received 19 November 2004; revised 25 January 2005; accepted 1 February 2005
Abstract Purpose: It is controversial whether arterial hypertension (AHT) can be diagnosed in the emergency department (ED). We sought to prospectively investigate the natural time course of blood pressure (BP) to define an optimal period for AHT screening in ED patients with an elevated initial BP. Procedures: Patients with a BP greater than 160/100 mm Hg upon ED admission underwent repeated BP measurements every 5 minutes for 2 hours using an automated device. Arterial hypertension was confirmed using 12-hour ambulatory BP measurement or repeated office BP measurement according to the Joint National Committee VII guidelines by the primary care physician after discharge from the hospital. Main Findings: Systolic BP decreased significantly during the first 10 to 20 minutes of ED stay in hypertensive and normotensive patients without further significant changes thereafter. Diastolic BP remained stable in both hypertensive and normotensive patients. Discrimination between hypertensive and normotensive patients was best between minutes 60 and 80 after ED admission. An average BP of 165/105 mm Hg or higher during this period strongly suggests AHT whereas a BP of less than 130/80 mm Hg excludes AHT with high sensitivity. Conclusions: Screening for AHT in the ED is possible with high specificity and sensitivity. Blood pressure measurements between minutes 60 and 80 after entry into the ED yield the highest diagnostic value. D 2005 Elsevier Inc. All rights reserved.
1. Introduction Zampaglione et al [1] reported symptomatic high BP (ie, hypertensive urgencies and emergencies) in 27.5% of Elevated arterial blood pressure (BP) in an emer- patients in an ED. A retrospective study found an incidence gency department (ED) setting is a frequent observation. of asymptomatic hypertension in the ED of 2.3% of the patients after excluding those with cardiovascular, renal, or central nervous dysfunction [2]. High BP in this setting T Corresponding author. Department of Medicine, University Hospital Basel, Petersgraben 4, CH-4031 Basel, Switzerland. Tel.: +41 61 265 2525; might be transient and be caused by a physiological fax: +41 61 265 4604. reaction to acute disease, pain, and stress [3] or simply E-mail address: [email protected] (B. Martina). be caused by a white coat effect [4]. It is therefore
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2005.02.046 Screening for hypertension in the ED 475 controversial whether elevated BP in the ED reflects 2.3. Selection of patients hypertension and whether the ED is an appropriate site for screening for high BP. Patients with unknown or presently untreated AHT However, several studies have shown that elevated BP in admitted to the medical ED for any reason were eligible the ED might be a useful indicator for arterial hypertension for enrollment into the study. Systolic BP at the time of (AHT) [5]. Even a single elevated BP value in the ED admission to the ED had to be 160 mm Hg or higher and indicates AHT in about 25% of patients [6]. Both studies diastolic BP had to be 100 mm Hg or higher. support the notion that elevated BP in the ED may be a Predefined criteria for noninclusion were life-threatening useful indicator for AHT. conditions or conditions needing emergent BP control such It is known that BP assessment by a physician or a nurse as hypertensive encephalopathy, suspected intracerebral elicits an alerting reaction in patients [7]. Office BP values hemorrhage or ischemic stroke, acute myocardial infarction, taken by physicians or nurses were markedly higher than the acute pulmonary edema, or aortic dissection. Further criteria previsit BP values but dropped considerably during the first for noninclusion were the inability of a patient to give 10 minutes of the office visit [7]. It might be speculated that informed consent and the need for immediate transfer to the stress caused by an ED visit might lead to an even higher other departments that made consecutive BP measurements BP increase compared with BP values outside the hospital in the ED impossible. and that repetitive BP measurements during the ED stay might better reflect a patient’s true BP. Surprisingly, neither 2.4. Methods of measurement prospective data on the spontaneous time course of initially 2.4.1. Blood pressure measurement in the ED elevated BP in an ED setting nor data on the diagnostic The first BP measurement was conducted in supine value of repeated BP measurements in the ED for the position directly after entry into the ED using a standard screening for AHT are available in the literature. mercury sphygmomanometer. If eligible for the study, Therefore, the aims of this study were to evaluate patients were informed by the study physician about the prospectively the time course of initially elevated BP by study’s rationale and procedures. After having obtained serial BP measurements in patients without current antihy- written informed consent, the following BP measurements pertensive treatment admitted to a medical ED and to were conducted in supine position with a validated automat- prospectively define the diagnostic value of repetitive BP ed ambulatory BP measurement (ABPM) device (Profilomat measurements in the ED for the screening for AHT. II, Disetronic, Switzerland) programmed to 5-minute inter- vals over a period of 2 hours. Maximal delay between initial BP measurement and ABPM was 15 minutes. Diagnostic 2. Methods procedures such as drawing of blood samples, electrocar- 2.1. Study design diogram, and chest x-ray were usually done within the first hour after admission as ordered by the treating ED physician. The natural time course of BP in medical ED patients presenting with moderate-to-severe BP elevations but 2.4.2. Confirmation of AHT without signs of target organ damage during the first hours Definite diagnosis or exclusion of AHT was made after of an ED stay is unknown. It is also unknown whether it is discharge from the hospital by a 12-hour daytime ABPM with possible to diagnose AHT in the ED by repeated measure- 20-minute intervals between measurements. Cutoff values for ments of BP. We assessed these questions in a prospective 12-hour daytime mean BP were defined as 135 mm Hg for study in a convenience sample of patients presenting to the systolic BP and 85 mm Hg for diastolic BP. Ambulatory BP medical ED at any time of the day during a 3-month period. measurement was performed within 1 week after discharge The study was approved by the local ethical committee. from the hospital. If ABPM could not be performed (n = 15), repeated BP values obtained by the family physician accord- 2.2. Setting ing to the Joint National Committee (JNC) VI [3] and JNC VII [8] guidelines were used to determine the presence of AHT. The study was performed at the medical ED of the University Hospital Basel. This ED is the only institution 2.4.3. Selection of best time interval for diagnosis or in town providing 24-hour emergency care for a popula- exclusion of AHT tion of 300000 people. Twelve thousand medical patients Obtaining BP values at distinct time points is often not are treated per year. Fifty-five percent of all patients are feasible in an ED. Moreover, JNC [3,8] and World Health treated as outpatients and 45% require inpatient treatment. Organization guidelines [9] recommend to average 2 or more Medical care is primarily provided by ED physicians (ie, BP readings for each BP measurement. Therefore, time house staff with 2-3 years of experience in internal medicine intervals instead of time points for the screening for AHT in on their regular ED rotation). Supervision is provided the ED should be defined. To get an estimate of the average 24 hours a day by attending physicians, board certified in BP in a given time interval, BP values obtained with the internal medicine. ABPM device were averaged around distinct time points. 476 T. Dieterle et al.
For example, the mean BP at 15 minutes was calculated as Table 2 Classification of diagnoses according to subspecialty the average of the BP readings at 10 to 20 minutes after entry at entry into the ED into the ED. Similar calculations were done for each 15- minute interval up to 120 minutes after entry into the ED. Classification of emergency No. of patients Averaged BP readings from hypertensive patients were Neurology 12 compared with those from patients diagnosed as normoten- Pneumology 9 sive after discharge from the hospital to define the best time Cardiology 7 interval for the screening for AHT in the ED setting. Gastroenterology 5 Rheumatology 2 Sensitivities and specificities of different cutoff values Hematology 2 for the screening for AHT at different time intervals were Angiology 1 computed. From these data, receiver operator characteristic Endocrinology 1 curves for systolic and diastolic BP measurements were Infectious diseases 1 constructed and the area under the curve was calculated. Nephrology 1
2.5. Statistics percent had normotension. Normotensive and hypertensive All statistical analyses were performed using an SPSS patients did not differ concerning sex distribution, risk factor, (SPSS, Chicago, IL) for Windows version 11.0 statistical or pain status. No differences were observed for medical package. For comparison of dichotomous variables and history and physical examination (data not shown). comparison of continuous variables, v2 analysis and the Student’s t test, respectively, were used. Repeated-measures 3.2. Blood pressure at entry into the ED analysis of variance was used for within- and between- group comparisons of consecutive BP measurements. Post Mean systolic BP at the time of entry into the ED was hoc comparisons were done using the Fisher protected 176 F 14 mm Hg and diastolic BP was 99 F 11 mm Hg t test. A P value less than .05 was considered to be (n = 41 patients). Systolic BP upon entry into the ED in statistically significant. patients with proven hypertension (n = 26) during follow-up was 180 F 14 mm Hg (n = 26) and did not differ from that in normotensive patients (171 F 14 mm Hg; n = 15). Also, 3. Results no significant differences between hypertensive and normo- tensive patients were found for diastolic BP (102 F11 vs 3.1. Patient characteristics 94 F 9 mm Hg). All patients had systolic BP values greater than 140 mm Hg at the time of entry into the ED. Five Forty-five ED patients with BP values greater than 160/ patients had diastolic BP values lower than 90 mm Hg at 100 mm Hg were included into the study. Four patients were this time point, 2 of them were subsequently proven to be not willing to perform the confirmatory follow-up ABPM and hypertensive during follow-up examinations. data on office BP measurements were not available in these patients. Data of 41 patients (91%) were included into the 3.3. Time course of BP in the ED analysis. Patient characteristics and classification of diagno- ses upon entry into the ED are given in Tables 1 and 2.At The time course of BP in hypertensive compared with entry into the ED, 22 patients (54%) had BP values of 160/ normotensive patients is given in Fig. 1. Systolic BP 100 mm Hg or higher and 19 patients (46%) had BP values of decreased significantly during the first 10 to 20 minutes of 180/110 mm Hg or higher. Arterial hypertension was ED stay in hypertensive and normotensive patients. In confirmed by ABPM or repeated office BP measurement both groups, no further statistically significant changes according to JNC guidelines in 26 patients (61%). Thirty-nine were observed between consecutive time intervals. Com-
Table 1 Patient characteristics All patients Hypertensive patients Normotensive patients P No. of patients 41 26 15 Age (y) 60.1 F 19.9 61.5 F 20.3 57.6 F 19.7 NS Sex (male/female) 28/13 17/9 11/4 NS Body mass index (kg/m2) 27.5 F 4.9 27.5 F 5.0 27.4 F 5.1 NS Blood pressure at entry into the ED (mm Hg) 176 F 14/99 F 11 180 F 14/102 F 11 171 F 14/94 F 9NS Smoker 13 9 4 NS Diabetes 6 5 1 NS Pain 14 8 6 NS Data are given as mean F SD where applicable. Screening for hypertension in the ED 477
A areas under the curve are given in Fig. 2 for each time 220 interval. Including the time points 60 and 65 minutes and thus averaging BP measurements for the time points 60 to 200 80 minutes increased the area under the curve further from 0.73 to 0.8 for systolic BP and from 0.71 to 0.76 for 180 # diastolic BP (Fig. 2). Arterial hypertension can be predicted with a specificity 160 ** ** ** ** ** ** ** ** greater than 90% if the mean BP taken between minutes 60 ** # # ## ## # ** 140 ** ** ** ** and 80 after entry into the ED is 165/105 mm Hg or higher. ** ** On the other hand, AHT can be excluded with a sensitivity 120 greater than 90% if the BP is lower than 130/80 mm Hg.
Systolic blood pressure (mmHg) 100 5 0 5 0 5 0 -3 -5 -6 -8 -9 12 Entry 10-20 25 40 55 70 85 5- 4. Discussion 100-110 11 Time (min) An elevated arterial BP in an ED is frequent [1,2], but it B remains unclear whether patients with an elevated arterial 140 BPs actually suffer from AHT with persistent high BP or whether the hypertensive values are merely related to the 120 ED admission and will normalize soon thereafter. Although several studies show that elevated BP in the ED predicts AHT in a high percentage of patients [5,6], elevated BP 100 values generally are considered to be associated with pain # ## ## and apprehension with ED visits and to be of low value for the diagnosis of AHT. Furthermore, there is controversy 80 as to whether or not initially elevated BP should be lowered immediately.
Diastolic blood pressure (mmHg) 60 Surprisingly, neither the natural time course of initially
y 0 5 0 5 0 5 elevated BP nor the diagnostic value of repeated BP tr -2 -3 -5 -6 -8 -9 En 10 25 40 55 70 85 100-110 115-120 measurements in an ED has been systematically assessed Time (min) and therefore the optimal time point for BP measurement in this setting allowing to differentiate between hypertension Fig. 1 Time course of systolic BP (A) and diastolic BP (B) during and normotension is unknown. ED stay (filled squares, hypertensive patients; open squares, In this study, we addressed these points by determining F b normotensive patients). Data are given as mean SD. *P .05 prospectively the time course of BP in patients with an and **P b .01 vs BP value at entry into the ED. #P b .05 and ##P b elevated BP at entry into a medical ED. Patients presenting .01 normotensive vs hypertensive patients. with an initial BP of 160/100 mm Hg or higher without signs parisons between hypertensive and normotensive patients of acute target organ damage were included into the study. revealed significant BP differences between minutes 25 We chose this limit because JNC VI guidelines recommen- and 95, with highly significant differences between ded a referral of these patients for further evaluation within 1 minutes 55 and 80. week [3] and the new JNC VII guidelines recommend an Diastolic BP remained stable between consecutive time initial 2-drug combination treatment [8]. intervals in both hypertensive and normotensive patients. At admission, BP did not differ between hypertensive However, intergroup comparisons showed significant differ- and normotensive patients and no clinical features that could ences in diastolic BP between minutes 40 and 80 after entry identify hypertensive patients at this time point were found. into the ED. Highly significant systolic and diastolic BP differences Of 15 normotensive and 26 hypertensive patients, 7 and between normotensive and hypertensive patients were found 5, respectively, reached normal BP values lower than 140/90 between minutes 55 and 80 after entry into the ED. In mm Hg at this time point ( P b .1). No adverse events addition, there was a trend indicating that more normoten- occurred during the ED stay. sive than hypertensive patients reached normotensive BP values during this time interval. The time interval between 3.4. Screening for AHT in the ED minutes 60 and 80 after entry into the ED yielded the best diagnostic accuracy for the detection or exclusion of Receiver operator characteristic curves were constructed hypertension. In patients presenting an average BP of 165/ for each time interval, revealing best diagnostic accuracy 105 mm Hg or higher during this time interval, AHT can be between minutes 70 and 80 after entry into the ED. The assumed with a specificity greater than 90%. Arterial 478 T. Dieterle et al.
A B 1.0 1.0
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2 Area under the ROC curve Area under the ROC curve
0.0 0.0 y 0 5 0 5 0 0 5 0 0 ry 0 5 0 5 0 0 5 0 0 tr -2 -3 -5 -6 -8 -8 -9 11 12 t -2 -3 -5 -6 -8 -8 -9 11 12 En 10 25 40 55 70 60 85 0- 5- En 10 25 40 55 70 60 85 0- 5- 10 11 10 11 Time (min) Time (min)
C D 1.0 1.0 130 120 110 70 80 0.8 140 0.8 90 0.6 0.6 150
0.4 160 0.4 100 Sensitivity Sensitivity
170 0.2 0.2 110 180 AUC = 0.803, AUC = 0.760, p<0.01 120 190 p<0.05 0.0 0.0 0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 1-Specificity 1-Specificity
Fig. 2 Diagnostic accuracy of averaged BP measurements for different periods. A and B, Comparison of diagnostic accuracy at different periods for systolic (A) and diastolic (B) BP measurements. The areas under the receiver operator characteristic curve are given together with the 95% CIs. C and D, Receiver operator characteristic curve showing the sensitivity and specificity of different BP cutoff values for the screening for AHT by averaged BP values taken between minutes 60 and 80 after entry into the ED for systolic (C) and diastolic (D) BP measurements. hypertension may be excluded with a sensitivity greater than We could not address the influence of medication, 90% when BP values lower than 130/80 mm Hg are especially pain medication, or interventions in the ED on documented in this time interval. the time course of BP in all patients. It is likely, however, that Our data show and confirm the view of others that the such interventions can influence the diagnostic value of BP ED may be a useful site for the screening for AHT [5,6]. measurements in the ED. A further limitation is the Moreover, the data extend previous findings showing that noninclusion of patients with a mild initial BP elevation. even diagnosis or exclusion of AHT in the ED is possible Nothing can be said about this population but it is likely that with high accuracy if repeated BP measurements are taken these patients will more often be diagnosed as normotensive. 60 to 80 minutes after admission to the ED. Finally, time of admission to the ED might play an important Several aspects may have influenced the results of this role because daytime BP usually is higher than nighttime BP. study. Because of the inclusion of a convenience sample, we Therefore, the time course of BP in the ED might be different could not provide data on the prevalence of AHT in medical during the day compared with that during nighttime. ED patients. Although normotensive and hypertensive In summary, we conclude from our findings that a patients did not differ significantly at admission, small substantial percentage of patients presenting to the ED with differences that might help to identify hypertensive patients high BP reaches normalized BP values without therapy cannot be completely excluded because of the relatively within 2 hours after admission. Despite the limited sample small number of patients. size and despite the likely effect of medications and Screening for hypertension in the ED 479 interventions on BP, it was possible to define a period during [2] Chiang WK, Jamshahi B. Asymptomatic hypertension in the ED. Am J which screening for hypertension is possible with relatively Emerg Med 1998;16:701-14. high accuracy. Elevated BP values in untreated ED patients [3] Joint National Committee (JNC) on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: The Sixth Report of the with a persistent elevated BP 1 hour after admission are Joint National Committee on Prevention, Detection, Evaluation, likely to be caused by definite AHT. Further evaluation of and Treatment of High Blood Pressure. Arch Intern Med 1997;157: these patients has to be ensured. A considerable part of ED 2413-46. patients with an initial BP elevation is finally normotensive. [4] Pierdomenico SD, Mezzetti A. White coat hypertension in patients with Repeated measurements of BP between minutes 60 and 80 newly diagnosed hypertension: evaluation of prevalence by ambulatory monitoring and impact on cost of health care. Eur Heart J 1995;16: after ED entry allow best discrimination between normoten- 692-7. sive and hypertensive patients. [5] Chernow SM, Iserson KV, Criss E. Use of the emergency department for hypertension screening: a prospective study. Ann Emerg Med 1987; 6:180-2. Acknowledgments [6] Slater RN, Da Cruz DJ, Jarrett LN. Detection of hypertension in accident and emergency departments. Arch Emerg Med 1987;4:7-9. We thank Mrs Verena Brenneisen and the ED team for [7] Mancia G, Casadei R, Gropelli A, Parati G, Zanchetti A. Effect of logistic support in conducting this study. We appreciate the stress on diagnosis of hypertension. Hypertension 1991;17(Suppl. III): help of M. Trindler, MD, P. Schlegel, MD, and C. III-56-62. [8] Chobanian AV, Bakris GL, Black HR, et al. The Seventh Report of the Kaufmann, MD in the enrollment of patients. Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. JAMA 2003;289:2560-72. References [9] World Health Organization, International Society of Hypertension Writing Group. World Health Organization (WHO)/International Society of Hypertension (ISH) statement on management of hyperten- [1] Zampaglione B, Pascale C, Marchisio M, Cavallo-Perin P. Hyperten- sion. J Hypertens 21 (2003) 1983-92. sive urgencies and emergencies. Prevalence and clinical presentation. Hypertension 1996;27:144-7. American Journal of Emergency Medicine (2005) 23, 480–482
www.elsevier.com/locate/ajem
Brief Reports Comparing 2 methods of emergent zipper release
Nobuaki Inoue MD, Steven C. Crook MD, Loren G. Yamamoto MD, MPH, MBA*
Department of Pediatrics, University of Hawaii John A. Burns School of Medicine, Honolulu, HI 96826, USA Emergency Department, Kapiolani Medical Center for Women and Children, Honolulu, HI 96826, USA
Received 15 July 2004; accepted 21 July 2004
Abstract Background: There are several types of emergent zipper release methods described. The standard method can be difficult. The purpose of this study is to determine if an alternate method of zipper release can be easier to accomplish. Methods: Subjects were provided with zippers and were taught 2 methods of emergent zipper release using a standard method (cutting the median bar of the actuator) and an alternate method (cutting the closed teeth of the zipper). The elapsed times to successful zipper release for both methods were measured. Results: Mean zipper release times were faster for the alternate method (10.5 seconds) compared with the standard method (75.8 seconds) ( P b .001). Conclusion: The alternate method of zipper release is faster and easier than the standard method of zipper release; however, the optimal procedure is also dependent on the location of the entrapped tissue relative to the zipper actuator and the type of zipper. D 2005 Elsevier Inc. All rights reserved.
1. Introduction from the rear [5]. The purpose of this study is to compare the 2 methods of emergent zipper release. Penile zipper injury is usually caused by entrapment of the penile tissue (foreskin, shaft, or glans) in the actuator or the teeth of the zipper. The most commonly described 2. Methods method of zipper release is to sever the median bar of the actuator (the mobile portion of the zipper) [1-5]. This bar Commercially available zippers and diagonal cutting holds the upper and lower halves of the actuator together. pliers were prepared for this study. Forty practitioners This standard method can be difficult to accomplish because (pediatricians, residents, and medical students) volunteered the metal bar is strong and penile tissue entrapment within to participate in this study. After obtaining signed informed the zipper actuator can limit access to the metal bar. An consent, the participants were instructed on how to release alternate method of releasing the zipper is to cut the closed the zipper using the 2 different methods: (1) standard teeth of the zipper, permitting the unzipping of the zipper (cutting the median bar of the zipper); (2) alternate method (simply cutting the closed zipper teeth at any position to unzip the remaining zipper). Elapsed time to complete the T Corresponding author. Department of Pediatrics, University of Hawaii John A. Burns School of Medicine, Honolulu, HI 96826, USA. zipper release was recorded for each method. The sequence Tel.: +1 808 983 8387; fax: +1 808 945 1570. of the methods was not randomized. The standard method E-mail address: [email protected] (L.G. Yamamoto). was always done first. If one took more than 5 minutes to
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2005.02.023 Comparing 2 methods of emergent zipper release 481
Table 1 Mean zipper release times F SD (range) n Standard method time (s) Alternate method time (s) Difference Pa (standard vs alternate method) All 40 75.8 F 95.3 (4-300) 10.5 F 6.5 (3-30) 65.3 F 92.4 .001 Males 20 42.6 F 60.7 (4-226) 9.0 F 5.3 (3-20) 33.6 F 60.2 .022 Females 20 109.1 F 112.4 (4-300) 12.0 F 7.4 (3-30) 97.1 F 108.5 .001 Pb (males vs females) .025 NS a Paired t test (standard vs alternate method). b Conventional t test (males vs females). release the zipper, then the time was recorded as 5 minutes must be severed, is fairly strong. It can be severed more (300 seconds) because prompt zipper release is less likely at easily if the diagonal cutting pliers are very large. Severing this point. Study subjects were also asked after completing this bar is substantially more difficult if a small diagonal both methods which method was easier to release the zipper. cutter is used. Our study only used large diagonal cutters. This explains why females took longer time in the standard method compared to males. There was no significant time 3. Results difference between males and females for the alternate method, probably because less strength is required for the Of 40 participants, there were 20 males and 20 females. alternate method. The mean times to complete the procedure are summarized This study used zippers that are very inexpensive. This in Table 1. Release times were faster for the alternate type of zipper has a small metal actuator and plastic zipper method compared with the standard method for the group as teeth, which is most commonly found on small coin purses a whole and for each sex. The magnitude of the difference and dress pants. Larger all-metal zippers with metal teeth was greater for females. The standard method of zipper (usually brass) are more commonly found on jeans and release showed a large statistically significant time differ- casual shorts. The actuator on these heavy-duty zippers is ence between males and females (males accomplished the larger and heavier. In preliminary trials, we broke several release faster). The alternate method did not show a large diagonal cutters trying to cut the median bar of the significant time difference between males and females. actuator. Thus, our study results are limited to the standard There were 4 subjects who could not complete the light-duty zipper more commonly found on dress pants. If a standard method within 5 minutes and all 4 subjects were patient entraps penile tissue within a heavy-duty all-metal female. There were 38 subjects who preferred the alter- zipper, it will be very difficult to release the zipper using the native method as faster and easier method and only 2 sub- standard method. The alternate method is easier in this case. jects who preferred the standard method rather than the If the standard method is used, an extra large heavy-duty alternate method. diagonal cutter will be required. Other types of zippers exist, such as those with very large 4. Discussion plastic or metal teeth, most commonly on large sports bags and luggage. It is unlikely that penile tissue will become Penile injuries (trauma) are relatively uncommon, but entrapped in these, but hair or other tissue of medical con- when it occurs, zipper entrapment is involved in 3% to 14% cern could become caught in these zippers, requiring release. cases [6-8]. This injury usually occurs between 3 and 6 Penile tissue entrapments could occur at several locations years of age and many of the victims were not wearing within a zipper: (1) within the actuator adjacent the open underpants [4,5]. side of the zipper against the median bar; (2) within the There are variety types of methods for releasing the actuator on the closed side of the zipper; (3) within the zipper. These include forcibly unzipping the zipper [1,3,9], enclosed zipper without involvement of the actuator. cutting off the actuator of the zipper with the standard For location 1 above, the standard method of zipper method [1-5], or cutting the zipper itself (the alternative removal would be impeded by the entrapped penile tissue method) [5]. Circumcising the penile foreskin entrapped in because the tissue would be blocking access to the median the zipper actuator has been described [1-3,5]. Soaking the bar. If the alternate method is attempted, it would release the penile tissue in mineral oil for 10 minutes followed by zipper from the wrong side of the actuator and this method gentle traction has also been described [5,10]. Cutting the will not necessarily succeed in releasing the tissue from the actuator bar with a mini hacksaw has also been described zipper. Both methods are problematic here. Local anesthesia [11]. Although the standard zipper release method was more (infiltrated or topical) would likely be beneficial to facilitate commonly discussed in textbook references, this study moving the actuator to a more favorable position permitting shows that the alternative method is faster and preferred. one of the methods to be used. The standard method of zipper release requires substan- For location 2 above, it is likely that the standard method tial strength. The median bar of the zipper actuator, which would succeed, but as the median bar of the actuator is 482 N. Inoue et al. severed, it is likely to briefly put additional pressure on the [3] Saraf P, Rabinowitz R. Zipper injury of the foreskin. Am J Dis Child entrapped tissue. The alternate method would also succeed if 1982;136:557-8. the actuator could be moved away from the entrapped tissue. [4] Gausche M. Releasing penile foreskin trapped in a zipper. Pediatr Rev 1993;14:40-1. Local anesthesia might be beneficial for this location as well. [5] Santamaria JP. Chapter 21. Office-based emergencies. Section 5.9 For location 3 above, the alternate method of zipper management of penile zipper injury. In: Gausche-Hill M, Fuschs S, removal would clearly succeed and because this method is Yamamoto L, editors. APLS: The pediatric emergency medicine easier, it should be the preferred method of zipper release. resource. 4th ed. American Academy of Pediatrics and American In conclusion, the alternate method of zipper release is College of Emergency Physicians. Sudbury (MA): Jones and Bartlett Publishers; 2004. p. 667-8. faster and easier than the standard method of zipper release; [6] Ezell WM, Smith I, McCarthy RP, et al. Mechanical traumatic injury however, the optimal procedure is also dependent on the to the genitalia in children. J Urol 1969;102:788-92. location of the entrapped tissue relative to the zipper [7] El-Bahnasawy MS, El-Sherbiny MT. Paediatric penile trauma. BJU actuator and the type of zipper. Int 2002;90:92-6. [8] Yip A, Ng SK, Wong WC, et al. Injury to the prepuce. Br J Urol 1989;63:535-8. References [9] Watson CCM. Zipper injuries. Clin Pediatr 1971;10:188. [10] Kanegaya JT, Shonfeld N. Penile zipper entrapment: a simple and less [1] Flowerdew R, Fishman IJ, Churchill BM. Management of penile threatening approach using mineral oil. Pediatr Emerg Care 1993; zipper injury. J Urol 1997;117:671. 9:90-1. [2] Oosterlinck W. Unbloody management of penile zipper injury. Eur [11] Strait RT. A novel method for removal penile zipper entrapment. Urol 1981;7:365-6. Pediatr Emerg Care 1999;15:412-4. American Journal of Emergency Medicine (2005) 23, 483–487
www.elsevier.com/locate/ajem
The interrater variation of ED abdominal examination findings in patients with acute abdominal painB,BB
Jesse Pines MD, MBAa,*, Lori Uscher Pines MSb, Annara Hallc, John Hunterd, Rajagopalan Srinivasan PhDe, Chris Ghaemmaghami MDf aDepartment of Emergency Medicine, University of Pennsylvania, Ground Silverstein, Philadelphia, PA 19104, USA bThe Johns Hopkins Bloomberg School of Public Health, Baltimore, MD 21205, USA cCollege of Arts and Sciences, University of Virginia, Charlottesville, VA 22905, USA dSchool of Engineering, University of Virginia, Charlottesville, VA 22911, USA eINC Research, Statistics Department, Charlottesville, VA 22911, USA fDepartment of Emergency Medicine, UVA Health Sciences Center, Charlottesville, VA 22905, USA
Received 6 March 2004; revised 2 September 2004; accepted 29 September 2004
Abstract Objective: The physical examination of the abdomen is crucial to emergency department (ED) management of patients with abdominal pain. We sought to determine the interrater variation between attending and resident physicians in detecting abdominal exam findings. Methods: Research enrollers surveyed attending and resident physicians on abdominal exam findings in the ED in patients with abdominal pain. Strength of agreement was calculated using the j statistic. Results: A convenience sample of 122 surveys was completed. Calculated j results are in parentheses. There was almost perfect agreement on the presence of masses and substantial agreement on the need for imaging studies. There was moderate agreement on guarding, distension, tenderness, and need for laboratory tests and surgical consultation. For 88 (72%) patients with tenderness, substantial agreement was calculated for epigastric tenderness, moderate agreement on right upper quadrant, supraumbilical, suprapubic, left lower quadrant, right lower quadrant tenderness, and fair agreement on left upper quadrant tenderness. Sixty-one (50%) patients received pain medicine in the ED. Among those, there was fair agreement on a presence of a surgical abdomen. Upper level resident physicians noted a higher level of agreement with the attending physician for tenderness than junior resident physicians.
B This work was presented at the Plenary Session of the SAEM mid-Atlantic meeting in Washington, DC, in March 2003 and also as an oral presentation for the SAEM national assembly in Boston in May 2003. BB Dr Pines developed and conceived the idea for this study, prepared and submitted the IRL proposal, and prepared the first draft and implanted the revisions into the manuscript. Ms Uscher was involved in the extensive manuscript editing, and was involved in helping Dr Pines develop the study idea. Mr Hall and Mr Hunter attended development meetings for this project and were in charge of data entry and initial data analysis. Dr Srinivasan attended development meetings and performed the statistical analysis for this project. Dr Ghaemmaghami was involved with the conception and development, and helped edit the manuscript. He oversaw the project. T Corresponding author. Tel.: +1 215 662 4050. E-mail address: [email protected] (J. Pines).
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.09.034 484 J. Pines et al.
Conclusions: There was moderate agreement between resident and attending physicians for most of the findings in patients with abdominal pain. Recognition that selected findings are more variable than others should encourage careful confirmation of resident physicians’ assessments in teaching settings. D 2005 Elsevier Inc. All rights reserved.
1. Introduction Virginia Health Sciences Center, Charlottesville, Va, from September 2002 to December 2002, 7 days a week, during Abdominal pain is a common chief complaint in patients the hours of 8 am to 12 pm. The institutional review board presenting to the emergency department (ED). According to reviewed the study and declared it exempt from informed 1999 statistics, about 3.4 million patients with acute consent because no unique patient or physician identifiers abdominal pain sought care in EDs in the United States appeared on the questionnaire forms. [1]. Acute abdominal pain can be caused by a large variety of pathology, ranging from the benign to the acutely life 2.2. Study setting and population threatening [2,3]. Overall, about 20% to 25% of patients University of Virginia Health Sciences Center is a presenting with acute abdominal pain are found to have a suburban hospital located in Charlottesville, Va, with serious condition requiring hospital admission [4]. approximately 60000 patient visits per year. Undergraduate The approach to diagnosing and treating acute abdominal enrollers asked both resident and attending physicians to fill pain in ED routinely includes a complete history and physical out a questionnaire regarding the results of their abdominal examination and may include laboratory tests, x-rays, examination. computed tomography scans, and/or other ancillary diagnos- A convenience sample of patients presenting with the tic testing. Sometimes, a surgeon, a gynecologist, a gastro- chief complaint of abdominal pain were enrolled in this enterologist, or another specialist is asked to render an study. Minors (younger than 18 years), prisoners, and opinion about the diagnosis and treatment of patient in the ED pregnant patients were excluded from this study. In addition, [5]. On initial assessment, the emergency physician attempts the treated physicians filled out the questionnaires at their to sort out patients who have immediately life-threatening convenience. Occasionally, physicians found themselves too diseases from those who may have more benign causes for busy to fill out the questionnaires. their abdominal pain [6]. However, even with the best clinical judgment, laboratory testing, and imaging, the actual 2.3. Study protocol diagnosis of a patients’ abdominal pain remains ambiguous in more than 40% of cases [7]. This is discouraging given the Resident and attending physicians filled out question- fact that the physical examination often plays a pivotal role in naires detailing the results of their abdominal examination determining the urgency of the problem [7]. in the ED. Each study patient was examined by both an In both academic and community practice, multiple attending and resident physician. Questionnaires were health care providers of various skill and experience levels paired to include the resident and attending interpretation examine patients. Physical exam findings of resident of the physical examination. Specifically, the surveys physicians are often used to determine the diagnostic included yes/no questions on the presence of the modalities and patient treatment of patients with abdominal following: tenderness, distension, masses, rebound, guard- pain. This requires confidence that the physician in ing, and the presence of a bsurgical abdomen.Q If training’s assessment of the abdominal examination is tenderness was marked yes, surveys detailed the location accurate. of the tenderness: left upper quadrant (LUQ), epigastric, Because of the highly subjective nature of the physical right upper quadrant (RUQ), left lower quadrant (LLQ), examination presents the opportunity for discrepancy right lower quadrant (RLQ), supraumbilical, and supra- between attending and resident assessments, we sought to pubic. Surveys also included an assessment of whether the determine the interrater variation between attending and patient needed laboratory tests, imaging studies, or resident physicians on detecting abdominal exam findings in surgical consultation. There was also a question about patients with the chief complaint of abdominal pain. whether or not the patient received pain medicine in the ED. If yes was checked in either the resident or attending survey, the assumption was that the patient received pain 2. Methods medicine in the ED. 2.1. Study design 2.4. Data analysis A prospective observational study using a convenience Interrater variability for 122 pairs was calculated using sample of nonconsecutive patients was performed surveying the j statistic. Significance of the association was deter- resident and attending physicians at the University of mined as a confidence interval (CI) that did not cross 0 or 1 Inter-rater variation of the abdominal exam 485
(j was significantly different from 0). Subgroup analysis Table 2 was done for patients who were found to have any tenderness, among those who received pain medicine, and j 95% CI among different resident training levels. Specific j’s were j Values for individual physical examination findings (n = 122) interpreted using the following parameters. A j less than 0 Masses palpated 0.82 0.59-1 was interpreted as no agreement between raters, 0.00 to 0.19 Guarding 0.49 0.31-0.68 as poor agreement, 0.20 to 0.39 as fair agreement, 0.40 to Tenderness 0.42 0.23-0.61 Distention 0.42 0.16-0.68 0.59 as moderate agreement, 0.60 to 0.79 as substantial Normal bowel sounds 0.36 0.1-0.61 agreement, and 0.80 to 1.00 as almost perfect agreement [8]. Surgical abdomen 0.27 À0.05-0.61 (see Table 1) j Values for perceived need for ED management (n = 122) Labs 0.46 0.28-0.64 3. Results Imaging studies 0.6 0.46-0.75 Surgical consultation 0.55 0.36-0.74 A total of 122 patients were enrolled in this study over a 3-month period, and data were collected from resident and resident physicians noted a higher level of agreement with attending physicians’ questionnaires. The interrater varia- the attending physician (PGY-3: j = 0.51; 95% CI, 0.09- tion and 95% CIs for 122 patients is detailed in Table 2. 0.92; PGY-2: j = 0.49; 95% CI, 0.23-0.75) than PGY-1 There was almost perfect agreement on the presence of resident physicians (j = 0.45; 95% CI, 0.13-0.77). In masses (j = 0.82; 95% CI, 0.56-1.00). There was moderate distinguishing the need for imaging studies, there did not agreement on the presence of guarding (j = 0.49; 95% CI, seem to be a learning curve among PGY-years (PGY-1: j = 0.31-0.68), distension (j = 0.42; 95% CI, 0.23-0.61), and 0.59, 95% CI, 0.36-0.83; PGY-2: j = 0.68; 95% CI, 0.47- tenderness (j = 0.42; 95% CI, 0.16-0.68). Also, there was 0.90; PGY-3: j = 0.52; 95% CI, 0.19-0.85). Other subgroup moderate agreement on the need for laboratory tests (j = comparisons for individual PGY-years were not significantly 0.46; 95% CI, 0.28-0.64) and surgical consultation (j = different from 0 ( P N .05). 0.55; 95% CI, 0.36-0.74) and substantial agreement on the need for imaging studies (j = 0.60; 95% CI, 0.46-0.75). Tenderness was noted by at least 1 examiner in 88 (72%) 4. Discussion of patients. There was substantial agreement on the location of epigastric tenderness (j = 0.69; 95% CI, 0.54-0.84) and The practice of medicine in the 21st century involves a moderate agreement on RUQ (j = 0.57; 95% CI, 0.40- team approach. In most hospitals, multiple health care 0.75), supraumbilical (j = 0.56; 95% CI, 0.34-0.78), providers examine patients and render opinions on the suprapubic (j = 0.44; 95% CI, 0.23-0.64), LLQ (j = diagnosis, treatment, and need for laboratory testing and 0.43; 95% CI, 0.24-0.64), and RLQ tenderness (j = 0.40; imaging studies. Therefore, practitioners often trust that 95% CI, 0.20-0.59). There was fair agreement on the another’s examination or evaluation of a patient is accurate. presence of LUQ tenderness (j = 0.39; 95% CI, 0.16-0.62). In some instances, the patient may be examined by the Sixty-one (50%) out of the 122 patients received pain following persons: the triage nurse, ED nurse, ED resident, medicine. For those patients who received pain medicine, medical student, attending physician, consulting resident there was fair agreement on the presence of a surgical physician, and consulting attending physician. Across this abdomen (j = 0.32). j was not significant for tenderness or group of providers, there is a wide range of experience and rebound. Post-graduate year (PGY) and PGY-3 resident expertise in the evaluation of the abdomen. Although this is physicians noted a higher level of agreement with the true in academic practice, the same may not be true in attending physician for tenderness (PGY-2: j = 0.49; 95% community practice. CI, 0.18-0.81; PGY-3: j = 0.49; 95% CI, 0.11-0.86) than The physical examination of the abdomen, the vital PGY-1 resident physicians (j = 0.33; 95% CI, 0.01-0.63). signs, and the degree of subjective discomfort the patient is Similarly, on the need for laboratory tests, PGY-2 and PGY-3 feeling often guides the diagnosis and treatment of patients presenting with abdominal pain. Often a judgment is Table 1 Interpretation of j values [8] rendered and a degree of concern is determined at the time j Interpretation of the physical examination. Location of tenderness, distension, and the presence or absence of bowel sounds b0 No agreement also guides diagnostic test ordering, consultation, and 0.0-0.19 Poor agreement 0.20-0.39 Fair agreement ultimate treatment. 0.40-0.59 Moderate agreement This study found highly variable interrater agreement— 0.60-0.79 Substantial agreement from fair to almost perfect in the presence and/or absence of 0.80-1.00 Almost perfect agreement specific features of the abdominal examination between resident and attending physicians in a real-time ED setting. 486 J. Pines et al.
This corroborates what other studies have indicated about ment of pain in the ED. A recent publication found that 43% the physical examination; what we do/how we treat patients of patients with acute abdominal pain wait too long to with abdominal pain is also highly variable depending upon receive analgesia in an ED population [21]. Our results the physician [9]. For example, in some instances, neither showed that the administration of pain medicine in the ED specialization nor experience has been shown to affect the actually increased the interrater agreement slightly; howev- accuracy in diagnosing a common clinical condition such as er, there were large CIs for both values. anemia on physical examination [10,11]. Although studies The interrater agreement on the need for laboratory tests, show that there is poor intraexaminer reliability for imaging studies, and surgical consultation was moderate to bimanual pelvic examinations in ED patients [12], the level substantial. Laboratory tests and imaging studies are ordered of interobserver variability for abdominal examination in a for a myriad of reasons; however, the need for laboratory university ED setting had never been studied. A recent study tests and imaging studies often points to diagnostic found a poor interrater agreement among family practi- uncertainty and/or concern that a patient may have a serious tioners in predicting whether a child had a streptococcal condition. throat infection based upon examining the throat [13].A For more than two thirds of the patients who had similarly designed study on ED patients found a poor tenderness associated with their abdominal pain, there was interobserver reliability between senior ED resident physi- substantial agreement on the presence of epigastric tender- cians and attending physicians on the results of a pelvic ness. There was moderate agreement on tenderness in other examination, which was between only 17% and 33% for areas, such as the RUQ, LUQ, RLQ, LLQ, and suprapubic positive findings [11]. Even among the same rater, Levy and supraumbilical regions. Regional tenderness suggests a [14] found an intrarater reproducibility of only 46% for the specific diagnosis, and moderate interrater agreement in, for determination of the shoulder laxity (grading) examination example, RLQ tenderness may indicate a variable rate of among orthopedic surgeons when the participant was suspicion for appendicitis or gallbladder disease (in the case blinded to the previous results of his/her own examination. of RUQ tenderness.) Actually, most studies reviewed by these authors found that Looking at subgroups of resident physicians reveals a j’s for many subjective and sometimes objective findings trend toward a learning curve for a few measures such as the often showed only fair to moderate agreement among presence of tenderness and the need for laboratory studies. examiners. Where j was greater than 0 ( P V 0.05), upper level resident Our study found only moderate agreement on the physicians tended to agree more frequently with attending presence of both rebound and guarding. Worrisome signs physicians than junior (PGY-1) resident physicians. These such as the presence of rebound and guarding suggest that results, however, were not significantly different as the CIs the patient has developed peritoneal irritation and may have overlapped significantly. an acutely life-threatening condition [15]. Presence of these Although there were very few masses palpated in this signs is often associated with a high level of concern and study (n = 5), there was substantial agreement in this area. In certainly warrants further evaluation of the patient’s fact, there was almost perfect agreement on the presence of condition, whether it be laboratory tests, imaging studies, masses. This may be due to an a priori awareness of the or specialty consultation. presence of a mass. Intraabdominal masses are often There was only fair agreement on the presence of a difficult to palpate, particularly in obese patients. However, surgical abdomen. This may suggest that the classic exam a recent study did find that there is a high rate or interrater for a surgical abdomen may not be as stereotypical as was agreement (77%) in the evaluation of patient with an thought. The fair level of interrater agreement between abdominal aortic aneurysm [22]. resident and attending physicians may suggest that the determination of whether a patient needs surgery is indeed highly subjective. And paradoxically, the level of interrater 5. Limitations and future questions agreement was slightly higher for patients who received pain medicine in the ED. Multiple studies have supported This study has several limitations. First, the abdominal the contention that the use of analgesia in patients with acute examination can change over time as the underlying abdominal pain does not affect ultimate diagnosis and pathology may progress or regress making serial examina- management [16-19]. This is consistent with the new tions vary. Also, examination results can vary depending guideline teachings in Cope’s acute abdomen to not upon the use of pharmaceuticals and fluids. The data points withhold analgesia until patients are examined by the recorded in this study were usually separated in time, surgeon [15]. However, a recent survey by Graber et al sometimes by hours; thus, disagreement between raters may [20] in 1999 found that 67% of general surgeons believed be partially due to a very different objective experience. The that pain medicines interfere with diagnostic accuracy, so medical therapy in the interim between physician examina- attitudes regarding this practice are still controversial. tions was not taken into account. Exactly half of the patients enrolled in this study received This study is also a convenience sample and was only pain medicine in the ED, which highlights the undertreat- collected during times when enrollers were available and Inter-rater variation of the abdominal exam 487 physicians could fill out the study forms. Another problem References with the methodology was that the enrollers did not record adequately when there was a refusal to fill out a study [1] McCraig LF, Burt CW. National Hospital Ambulatory Medical Care form. Therefore, it was impossible to calculate a response Survey: 1999 emergency department summary. Advance data from rate for this study. There also may be a significant recall vital and health statistics (DHHS publication no. (PHS) 2001-1250 01- 0357) No. 320, Hyattsville, MD, National Center for Health Statistics. and reporting bias in the way the data were collected [2] Tintanelli JE, Kelen GD, Stapczynski JS. Emergency Medicine: A because attending physicians may have already heard the Comprehensive Study Guide. 5th ed. 2003. p. 499. resident physicians’ presentation before they examined the [3] Cope, Silen W. Early diagnosis of the acute abdomen. 19th ed. New patient. In addition, many physicians were too busy to fill York7 Oxford University Press; 1995. out the survey. Enrollers did not record when they were too [4] Graff LG, Radford MJ, Werne C. Probability of appendicitis before and after observation. Ann Emerg Med 1991;20:503-7. busy so it is impossible from our data set to determine a [5] Graff L, Robinson D. Abdominal pain and emergency department response rate. evaluation. Emerg Med Clin North Am 2001;19(1):123-36. An inherent flaw in this study is the use and interpre- [6] Howard PA. High-risk presentations in emergency medicine. Emerg tation of the j statistic. There is a lot of disagreement among Med Clin North Am 2003;21(1):123-36. statisticians on how to interpret j. When quantifying actual [7] Diethelm AG, Stanley RJ, Robbin ML. The acute abdomen. In: Sabiston DC, editor. Textbook of Surgery. 15th ed. Philadelphia7 WB levels of agreement, j’s calculation uses the proportion of Saunders; 1997. p. 825-46. chance/expected agreement. However, because this concept [8] Landis JR, Koch GG. The measurement of observer agreement for is only relevant if raters are truly independent (which by categorical data. Biometrics 1997;33:159-74. definition is impossible to achieve), j is not truly a chance- [9] Kamin RA, Nowicki TA, Courtney DS, et al. Pearls and pitfalls in the corrected measure of agreement. In addition, many statis- emergency department evaluation of abdominal pain. Emerg Med Clin North Am 2003;21(1):61-72. ticians have written about j’s paradoxes. There are [10] Wallace DE. The influence of experience and specialization on the occasions when j may be low although there are high reliability of a common clinical sign. Ann R Coll Surg Engl levels of agreement [23]. 2000;82(5):336-8. The results of this project suggest the inherent subjec- [11] Muellner T. Inter- and intra-tester comparison of the Rolimeter knee tivity of the physical examination and corroborate the results tester: effect of tester’s experience and the examination technique. Knee Surg Sports Traumatol Arthrosc 2001;9(5):302-6. of previous researchers. Future investigators may consider [12] Close RJ, Sachs CJ, Dyne PL. Reliability of bimanual pelvic measuring the interrater variation of other common ED examinations performed in emergency departments. West J Med decision points in patient management to characterize which 2001;175(4):240-5. management decisions may require either objective confir- [13] Donner-Banzhoff N. Clinical findings in patients presenting with sore mation, more careful examination, or focus in a resident throat: a study on inter-observer reliability. Fam Pract 2002;19(5): 466-8. teaching curriculum. [14] Levy AS. Intra- and inter-observer reproducibility of the shoulder laxity examination. Am J Sports Med 1999;27(4):460-3. [15] Silen W. Cope’s early diagnosis of the acute abdomen. 19th ed. New 6. Conclusions York (NY)7 Oxford University Press; 1996. [16] Attard AR, Corlett MJ, Kidner NJ, et al. Safety of early pain relief for This study demonstrates that for the emergency physi- acute abdominal pain. BMJ 1992;305:554-6. cian in a real-time ED setting, the abdominal examination is [17] Pace S, Burke TF. Intravenous morphine for early pain relief in patients a highly subjective diagnostic experience. Some findings with acute abdominal pain. Acad Emerg Med 1996;3:1086-92. [18] LoVecchio F, Oster N, Sturmann K, et al. The use of analgesics in show higher interrater variation than others. This finding patients with acute abdominal pain. J Emerg Med 1997;15:775-9. raises concern because the interpretation of the physical [19] Vermeulen B, Morabia A, Unger PF, et al. Acute appendicitis: examination is often used to determine the diagnosis, influence of early pain relief on the accuracy of clinical and US treatment, and ultimate disposition of ED patients. Aca- findings in the decision to operate—a randomized trial. Radiology demic emergency physicians must carefully evaluate resi- 1999;210:639-43. [20] Graber MA, Ely JW, Clarke S. Informed consent and general dent assessments of the physical examination, particularly surgeons’ attitudes toward the use of pain medication in the acute in high-risk areas, such as the assessment of acute abdomen. Am J Emerg Med 1999;17:113-6. abdominal pain. [21] Tait IS, Ionescu MV, Cuschieri A. Do patients with acute abdominal pain wait unduly long for analgesia? J R Coll Surg Edinb 1999; 44:181-4. Acknowledgments [22] Fink HA, Lederle FA, Roth CS, et al. The accuracy of physical examination to detect abdominal aortic aneurysm. Arch Intern Med 2000;160:833-6. I would like to acknowledge Marat Ivanov for his [23] http://ourworld.compuserve.com/homepages/jsuebersax/kappa.htm— involvement in data analysis. search done July 28, 2003. American Journal of Emergency Medicine (2005) 23, 488–491
www.elsevier.com/locate/ajem
Biphasic extrathoracic cuirass ventilation for resuscitation
Ilan Gur MDa, Ephraim Bar-Yishay PhDb, Ron Ben-Abraham MD, DEA, MHAc,* aNeonatal Intensive Care Unit, Bikur Holim Hospital, Jerusalem, Israel bPulmonary Function Laboratory, Hadassah University Hospital, Jerusalem, the Hadassah Faculty of Medicine, Hebrew University, Jerusalem, Israel cDepartment of Anesthesiology and Critical Care Medicine, Tel-Aviv Sourasky Medical Center, the Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
Received 13 October 2004; accepted 24 October 2004
Abstract Purposes: The MRTX portable lightweight respirator (MRTX) provides noninvasive respiratory support using biphasic extrathoracic ventilation via a cuirass fitted around the patient’s chest. Methods: MRTX was applied with or without full protective gear, on adult volunteers simulating nerve agent (NA) victims by nonmedical caregivers. Assessment was made based on scores for correct positioning of the cuirass, quality of seal, and rapid ness. Results: For the unprotected and protected personnel, the respective median (F95% confidence interval) scores for correct positioning of the cuirass were 2 (1.4-1.9) and 1 (1.2-1.8) (n = 15 per group, P = NS); quality of seal scores were 2 (1.5-2.0) and 2 (1.3-1.8) ( P = NS); and mean (FSD) time required for instituting mechanical ventilation was 90.5 F 10.9 and 100.3 F 7.9 seconds ( P b .05). The respirator was activated at first attempt 11 times in the group of 15 without protective gear and 8 times in the group of 15 with protective gear ( P = NS). Discussion: Biphasic cuirass ventilation is an easily learned and rapidly applied method suitable for use by nonmedical personnel, even when wearing cumbersome protective gear. D 2005 Elsevier Inc. All rights reserved.
1. Introduction NA derives from their powerful ability to irreversibly inhibit acetylcholinesterase, which regulates the activity of Chemical warfare using nerve agents (NA) can cause both the nicotinic and muscarinic synapses. The resultant high mortality rates, as demonstrated in the terrorist attacks accumulation of acetylcholine at the neuro-effector junc- against urban populations in Japan in 1994 and 1995 and tions leads to disruption of normal synaptic transmission in in the military use of NA by Iraq against Iran and the Iraqi both the peripheral and central cholinergic systems, leading Kurdish population during the 1980s. The high toxicity of to the clinical manifestations of severe cholinergic crisis [1]. The immediate cause of death is usually rapid progressive respiratory failure brought about by NA’s T Corresponding author. poisonous and depressive action on the central respiratory E-mail address: [email protected] (R. Ben-Abraham). center and on the neuromuscular junction and airways.
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.10.004 Extrathoracic ventilation of victims of nerve agent poisoning 489
Irritation of both the upper and lower airways, their 2. Methods blockade by secretion, toxic pulmonary edema, and severe bronchoconstriction are all parts of the effects of the NA- 2.1. Participants induced severe parasympathetic overstimulation that An institutional review board approval for the study was increases airway resistance and poses mechanical obstruc- obtained as well as verbal consent from each of the tion to ventilation. participants. Two groups of 15 nonmedically trained Initial resuscitative efforts for NA-poisoned victims personnel participated in this prospective comparative study. focus upon immediate respiratory support to prevent All were members of nonmedical extrication teams with instant death from hypoxia. Current therapeutic protocols similar professional skills. Members were randomly stress the need for urgent laryngoscopy and intubation, assigned to either one of the 2 groups. They were asked with concomitant provision of positive pressure venti- to connect the MRTX respirator to adult volunteers (age lation until signs of muscle paralysis disappear [1]. In the range, 25-40 years; weight, 70-100 kg) who simulated a event of a sudden and unexpected need to resuscitate NA-poisoned patient and who was lying fully dressed on a multiple victims who are highly diverse in age and size stretcher. One group wore ordinary clothing while treating and in an atmosphere of chaos, there will inevitably be a the bpatientQ and the other group wore complete antichem- shortage in personnel who are well trained in airway ical warfare gear, which included a full-body plastic suit, management. Moreover, the cumbersome protective gear, rubber gloves, and a gas mask with an attached filter. Both including thick rubber gloves, worn by the rescue teams groups were given 15 minutes to read the instruction manual will inevitably present difficulties in performing a for operating the MRTX respirator as well as a short laryngoscopy. demonstration on its use. The MRTX respirator (MRTX; Medivent, London UK) is a lightweight, easy to operate, portable respirator which 2.2. Study protocol can provide proper artificial ventilation by means of external high-frequency oscillation (EHFO) via a cuirass The ease and efficacy of operation of the MRTX by both that is tightly fitted around the patient’s chest. The groups was prospectively and comparatively assessed by the cuirass consists of a clear, flexible plastic enclosure following score parameters which were validated in surrounding the chest and abdomen. Its borders are preliminary experiments: covered by a soft foam rubber, which creates an airtight seal around the patient. By choosing the appropriate 1. Scoring for correct positioning of the cuirass around cuirass size, the apparatus is capable of ventilating a wide the simulated NA-poisoned victim’s chest using a range of different-sized subjects, from infants to the scale of 0 to 2 (0 = cuirass positioned on the abdomen, obese adult. The cuirass is connected to a computerized 1 = cuirass partially positioned on the chest, 2 = power unit by a wide-bore tube and the respiratory cuirass correctly positioned around the chest). parameters are controlled by a feedback mechanism bet- 2. Scoring for air leakage using a scale of 0 to 2 (0 = ween the two. The power unit works by creating cyclic audible and tactile leak of air around the rim of the pressure changes inside the cuirass. The negative pressure cuirass, 1 = tactile leak only, 2 = no air leak [ie, an (vacuum) creates chest expansion-inhalation. The positive adequately tight fit of the cuirass to the body]). pressure creates chest compression-exhalation. Thus, both 3. Recording the number of times the respirator was inspiratory and expiratory phases are actively controlled, actually activated on the first attempt. and the chest is oscillated around a variable negative 4. Measurement of the time needed for instituting baseline pressure. The system was found to be effective respiratory assistance. This included removing the in a variety of clinical settings, with pressures of À25 to cuirass from its box, fixing it onto the patient’s chest, connecting it to the power unit, switching on +15 cm H2O, inspiratory/expiratory (I/E) ratios of 1/1 to 1/3, and frequencies of 60 to 150 cpm [2-4].Inan the respirator, and inspecting that chest inflation experimental study in cats intoxicated with organophos- is symmetrical. phate [5], EHFO was reported to be capable in keeping 2.3. The MRTX portable respirator animals alive until spontaneous recovery from the poi- soning occurred. In addition, excessive bronchial sec- The MRTX consists of a flexible, lightweight plastic retions, typical of organophosphate intoxication, drained cuirass that covers the anterior part of the chest and upper spontaneously through the mouth making airway protec- abdomen. It has a wide foam rim, which creates an airtight tion and suction unnecessary. seal, and a back plate, which is secured by straps. The We hypothesized in the present preliminary study that cuirass is connected by wide-bore flexible tubing to a proper respiratory assistance using the portable MRTX mobile and microprocessor-controlled power unit. The respirator can be provided to adult volunteers simulating power unit contains a turbine that generates an oscillatory NA victims by a nonmedical rescue team, even when pressure that operates over a wide range of frequencies wearing protective gear. (ie, 6-1200 cpm) and pressures (À50 to +50 cm H2O). The 490 I. Gur et al. frequency, I/E cuirass chamber pressures, and the I/E ratio tion for mass casualties from NA poisoning [7], but there are are automatically set and controlled by negative feedback none on the use of any type of negative pressure ventilation loop [6] which makes it possible to achieve high frequen- in such a scenario. cies. This is in contrast to regular negative pressure The MRTX respirator is a portable, relatively small ventilation, which depends on passive recoil of the chest (14 Â 14 Â 18 cm), lightweight (3 kg including battery), during the expiratory phase, thus, limiting the attainable battery-operated rugged respirator. It is a modern version of frequencies. the Hayek Oscillator that has long been used for adminis- tering respiratory support for both children and adults [2-4]. 2.4. Statistical analysis Its small dimensions and lightweight permit it to be easily Timing events are presented as means F SD with carried by rescue personnel even under difficult conditions differences among groups analyzed using the 2-tailed (ie, the wearing of cumbersome protective gear, the need to unpaired Student t test. All scores are presented as medians carry other equipment, etc), making it easily deployed and and 95% confidence intervals with differences analyzed capable of operation both indoors and outdoors. The power using the nonparametric independent 2-group comparisons supply can suffice for a period up to 3 hours. Mann-Whitney U test. v2 Tests were used to compare Mass casualties can be anticipated from the exposure of a categorical values. P b .05 was considered significant. population to nonconventional warfare agents. The safest mode of airway protection in NA poisoning is probably endotracheal intubation [1], but this may not be the case for 3. Results intubation performed by physicians who are unskilled in the procedure and who must perform it in the prehospital setting Slightly better scores were achieved by the 15 operators under chaotic conditions. The application of biphasic cuirass- not wearing protective gear as compared with the group of based negative pressure ventilation using the MRTX respi- 15 operators with protective gear, but differences did not rator may spare the need for laryngoscopy and intubation, reach significance. For the unprotected and protected both exacting procedures that may well be hampered by the personnel, the respective median (F95% confidence interval) impaired vision and manual dexterity of protected medical scores for correct positioning of the cuirass were 2 (1.4-1.9) personnel. In addition, cross-contamination of the acute and 1 (1.2-1.8) (n = 15 per group, P = NS). For quality of seal, rescue team can cause ocular signs and symptoms, such as the scores were 2 (1.5- 2.0) and 2 (1.3-1.8) ( P = NS). severe miosis, dim vision, and persistent rhinorrhea which, Mean (FSD) time required for instituting mechanical although not lethal, can compromise even the most proficient ventilation was 90 5 F 10.9 and 100.3 F 7.9 seconds ( P b physician’s ability to function [1]. Another potential benefit .05). The respirator was activated at first attempt 11 times in for obviating laryngoscopy is sparing the use of muscle the group of 15 without protective gear and 8 times in the relaxants, which might act unpredictably in cases of group of 15 with protective gear ( P = NS). Initial underlying systemic inhibition of acetylcholinesterase [8]. inadequate placement of the cuirass and failure to switch Because systemic toxicity of NA is invariably associated on the respirator were rapidly corrected and respiratory with severe bronchorrhea [8], repeated active drainage of support could be speedily provided in all cases. bronchial secretions is a major component of the respiratory care of the intoxicated patient. Clearance of secretions was shown to be enhanced by using biphasic cuirass-based 4. Discussion negative pressure EHFO because of reduction in sputum viscosity and enhancement of ciliary’s clearance [5,9]. In the The results of the present study, although descriptive in prehospital setting where massive numbers of casualties nature, indicate that the cuirass can be rapidly and easily need to be treated and when medical personnel is expected applied to apneic victims by gear-protected caregivers and, to be too limited to be available for performing airway if necessary, biphasic external cuirass-based ventilation can suctioning for every patient that needs it, periodic emer- be instituted by the portable MRTX respirator. Thus, this gency ventilation using the MRTX respirator might serve as method of ventilation might serve as an effective tool in case a useful solution. of need for urgent resuscitative respiratory support to adult The EHFO is a relatively new modality of noninvasive victims of NA poisoning. Nonmedical personnel, with no ventilation which controls the inspiratory and expiratory previous experience with patient ventilation, could efficient- phases of respiration, both of which are active [10]. The peak- ly operate the device and respiratory support could be inspiratory chamber pressure and the end-expiratory chamber rapidly instituted, even when wearing bulky protective gear. pressure are controllable as well. The inspiratory pressure Its simplicity of use and the need for fewer medical staff will always be negative to expand the chest and inflate the members compared with a hand-operated respirator (ie, self- lungs, but the expiratory pressure may be negative, atmo- inflating bag) are additional benefits. There are earlier spheric, or positive to produce an end-expiratory lung reports on resuscitative respiratory support having been volume above functional residual capacity. Tidal volume is given by intubation followed by positive pressure ventila- determined by the pressure difference between the expiratory Extrathoracic ventilation of victims of nerve agent poisoning 491 and peak inspiratory chamber pressure and frequency: ble respirator is easily learned and rapidly applied by increasing this pressure difference increases tidal volume, nonmedical personnel, even those wearing cumbersome, whereas increasing the frequency reduces it. Because minute full-body antichemical protective gear. This method, when ventilation is the product of frequency and tidal volume, used for urgent respiratory support, saves the need for minute ventilation increases as long as the increment in laryngoscopy and aids in secretion removal. Our encourag- frequency is higher than the decrease in tidal volume [5]. ing preliminary findings warrant further assessment as for With both the inspiratory and expiratory phases being its beneficial role in resuscitation of mass casualties from controlled, high respiratory rates (up to 1200 oscillations NA poisoning. per minute) may be achieved, in contrast to normal negative pressure ventilation (which does not have an active compo- nent of expiration). Optimal carbon dioxide removal is References achieved with a rate of 90 per minute in adults with normal lungs. An I/E of 1/1 is optimal, although changing the I/E [1] Ben Abraham R, Rudick V, Weinbroum AA. Practical guidelines ratio may be needed for optimal ventilation in patients with for acute care of victims of bioterrorism: conventional injuries and concomitant nerve agent intoxication. Anesthesiology 2002;97: inspiratory or expiratory obstruction [5]. 989-1004. Trials have proved the efficacy of EHFO in ventilating [2] Petros AJ, Fernando SS, Shenoy VS, et al. The Hayek Oscillator. normal and sick lungs [3,4]. Potential preservation of Nomograms for tidal volume and minute ventilation using external cardiac output by EHFO compared with conventional high frequency oscillation. Anaesthesia 1995;50:601-6. positive pressure ventilation [3] might be preferred in light [3] Penny DJ, Hayek Z, Redington AN. The effects of positive and negative extrathoracic pressure ventilation on pulmonary blood flow of the negative inotropic effects induced by NA [8]. after the total cavopulmonary shunt procedure. Int J Cardiol Although EHFO can actively aid in secretion clearance in 1991;30:128-30. a forceful manner, a fact that probably reduces the chances [4] Shekerdemian LS, Schulze-Neick I, Redington AN, et al. Negative for aspiration [5], adequate separation of the digestive and pressure ventilation as haemodynamic rescue following surgery for respiratory tracts is not maintained with this method. Hence, congenital heart disease. Intensive Care Med 2000;26:93-6. [5] Hayek Z, Sohar E. External high frequency oscillation—concept and it would be prudent to exchange the extrathoracic ventila- practice. Intensive Care World 1993;10:36-40. tion that had been provided by the cuirass by the placement [6] Dilkes MG, Hill AC, McKelvie P, et al. The Hayek Oscillator: a new of an endotracheal tube as soon as possible. This is method of ventilation in microlaryngeal surgery. Ann Otol Rhinol especially true when respiratory failure is prolonged and Laryngol 1993;102:455-80. support is needed for an extended period. [7] Morita H, Yanagisawa N, Nakajima T, et al. Sarin poisoning in Matsumoto, Japan. Lancet 1995;346:290-3. Our study is based on a descriptive model. For obvious [8] Weinbroum AA, Rudick V, Paret G, et al. Anaesthesia and critical care reasons, experimental studies using models of NA intoxi- considerations in nerve agent warfare trauma casualties. Resuscitation cation are difficult to conduct even in animals because of the 2000;47:113-23. extreme lethal potency of these agents and the need for [9] King M, Phillips DM, Gross D, et al. Enhanced tracheal mucus special well-equipped laboratory and personnel familiar clearance with high frequency chest wall compression. Am Rev Respir Dis 1983;128:511-5. with the use of NA. [10] al-Saady NM, Fernando SS, Petros AJ, et al. External high frequency In conclusion, the methodology of instituting biphasic oscillation in normal subjects and in patients with acute respiratory extrathoracic cuirass-based ventilation via an MRTX porta- failure. Anaesthesia 1995;50:1031-5. American Journal of Emergency Medicine (2005) 23, 492–496
www.elsevier.com/locate/ajem
Effect of a social services intervention among 911 repeat users
Steven J. Weiss MDa,*, Amy A. Ernst MDa, Margaret Ong RNb, Ray Jones EMTb, Debra Morrow MSWc, Rosemary Milchc, Katie O’Neilc, Jay Glass EMTb, Todd Nick PhDd aDepartment of Emergency Medicine, University of New Mexico, Albuquerque, NM 87131, USA bSacramento City Fire/EMS, Sacramento, CA 95816, USA cDepartment of Health and Human Services (DHHS), Sacramento, CA 95823, USA dCenter for Epidemiology & Biostatistics, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH 45229, USA
Received 5 October 2004; accepted 1 November 2004
Abstract Objective: To determine whether emergency medical services (EMS) 911 frequent users would benefit from social services intervention. Methods: The design was a descriptive prospective subject evaluation. All nonhomeless frequent EMS users (N3Â in 1 month) were identified monthly from December 2 to May 3 and contacted by 2 social workers. Information extracted from their contact with the subjects included demographics, ability to enter a social services intervention, and reason for transport. Results: Eighty-four patients were eligible for inclusion in the study. Seventy-four patients were unable to enter a social services intervention for the following reasons: not home (2Â) (26%), not at address (19%), refused (13%), unable to complete Mini-Mental Status Exam (10%), deceased (6%), hospitalized (5%), safety issues (4%), and others (10%). The reasons for frequent EMS use were cardiac (24%), asthma/chronic obstructive pulmonary disease (25%), seizures (14%), dialysis problems, alcohol problems, and diabetes-related problems (b10% each). Conclusion: Among all patients, the primary reasons for transport were cardiac, asthma/chronic obstructive pulmonary disease, and seizures. Only 12% of patients contacted could enter a social services intervention. On the basis of the small cohort of patients that were able to enter a social services interventions, more targeted interventions are warranted. D 2005 Elsevier Inc. All rights reserved.
1. Introduction
Frequent users of the medical system represent a burden T Corresponding author. Tel.: +1 505 272 5062; fax: +1 505 272 6503. to the system, to society, and to their own care [1,2].We E-mail address: [email protected] (S.J. Weiss). recently showed that a small group of patients were
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.11.003 Social Services and EMS 493 accounting for a large number of ambulance transports [2]. After signing a consent form, all subjects completed the No previous studies have evaluated the emergency medical SF-36. The intervention consisted of traditional social services (EMS) 911 calls and the problems among frequent services management including an intake questionnaire, users. People who call 911 repeatedly within a short period assessment of the patient’s specific problems, and referral not only overburden the existing systems, but also may be in to specific community services. All resources of the need of extra help or ancillary services. The rationale of this Department of Health and Human Services (DHHS) were study was to determine if these repeat callers could be made available. After 30 days, the subjects completed contacted for a social services intervention and whether such another QOL questionnaire. Outcome variables included an intervention could improve the quality of the subject’s number and types of interventions, past vs future ambulance life. The EMS system is not presently able to assist with transports, and changes in the QOL as measured by the alternative referrals to resources that could help these SF-36. The investigator extracted numbers and types of individuals. Social services interventions have been proven interventions from the social services records for the sub- to be cost effective and helpful in other emergency settings jects. At the end of the 30-day period, all subjects in both [3-9]. We devised this study to determine if a social service arms of the study were offered social services follow-up. intervention was feasible. Changes in QOL results were compared using an The hypothesis of this study was that a social services independent samples t test. Results for ambulance use were intervention is feasible for frequent EMS users and is an compared using v2 and relative risks. The University and efficient use of the social worker’s time. City Department of Health Services Institutional Review Board both approved all phases of this study.
2. Methods 4. Results The study population included all nonhomeless men and women 18 years and older who have had at least 3 calls to Eighty-four patients were included in the study. The 911 in the previous month and can be found by our average age was 49 F 2 years with 4 patients younger than representatives within 2 attempts. These data were obtained 18 years and 17 patients older than 65 years. The reasons for from 911 dispatch records available through the County/ frequent EMS use were cardiac (24%), asthma/chronic City EMS. The City Department of Social Services was obstructive pulmonary disease (COPD) (25%), seizures consulted and 2 social workers contacted each frequent user. (14%), dialysis problems, alcohol problems, and diabetes- Descriptive information was extracted from charts and related problems (b10% each). In 25 cases, all transports social services paperwork regarding problems and age. were for different complaints. The number of calls per month ranged from 2 to 8 with an average of 2.9. These results are summarized in Table 1. 3. Targeted intervention pilot study Seventy-four of the 84 patients identified were unable to participate in the intervention. A total of 10 patients (12%) If patients were willing to undergo a social services were enrolled in the intervention over the 6-month course of intervention, the social workers continued their care for a the study. Reasons patients were excluded from the period of 1 month. Reasons that patients were not able to intervention were the following: not home (2) (26%), not have an interaction with social services included if they at address (19%), refused (13%), unable to complete MMSE were unwilling or unable to consent, if they could not be (10%), deceased (6%), hospitalized (5%), safety issues found or were unavailable, if their Mini-Mental Status (4%), and others (10%). Reasons patients were unable to Exam (MMSE) [1,9-11] score was less than 22, if they take part in a social services intervention are summarized were incarcerated at the time they were entered in the study, if they did not speak English, if there was a safety issue to the social workers, or if the patients could not Table 1 Specific problems that were the recurring cause of complete the forms. transport by 911 during the study period We then randomized the available patients to either a social services intervention or observation. We tracked their Diagnosis N % progress using 2 outcome variables pre- and poststudy: (1) the Cardiac 20 24 frequency of EMS calls and (2) the subject’s quality of life Asthma/COPD 21 25 (QOL) as measured by the SF-36. The SF-36 is a nationally Seizures 12 14 Dialysis problems 7 8 recognized validated scale for QOL determination that is split Alcohol problems 6 7 into physical component scale and a mental component scale Diabetes-related problems 8 9 [1,10,12-29] in injured patients [30,31] and in geriatric No contributing factors 17 13 populations [32]. Both the physical and mental scales of the Some patients had enough transports to fit into more than one category. SF-36 scale have national averages of 44 F 10. 494 S.J. Weiss et al.
interested in help, with little ability to intervene in their Table 2 Reasons for exclusion of the subjects from the social lives. After many discussions with our social workers, we services intervention have concluded from the fact that only 12% of the frequent Reasons N % EMS users are interested in being helped, and that social Not home (2Â)2226services support needs to be directed more critically to the Not at address 16 19 needed problems or the specific EMS patients. This study Refused 11 13 was concluded with the belief that the social workers need Unable to complete MMSE 8 9 to focus on a need-related group rather that on general Deceased 5 6 frequent users. Future studies will look at the focused ability Hospitalized 4 5 Safety issues to social workers 3 3 of social services to aid with patients with specific problems Others 5 6 such as falls, drug/alcohol problems, seizures, chest pain, Total excluded 74 88 diabetes, asthma, dialysis, and cardiac disease. Percentages are out of the 84 patients that took part in the study. The use of social services in conjunction with EMS is novel. This is the first in a series of studies evaluating the most practical way to include social services in the pre- in Table 2. There were no deviations from the study hospital environment. Working with all nonhomeless repeat- as planned. ers was too broad and may have led to futile efforts by our social workers. By picking referrals carefully, we may see 5. Targeted intervention pilot study results much more prominent changes in the patient outcomes. Table 3 shows suggested social services interventions that (10 patients) could be accomplished in a diagnosis-focused system. Change in the SF-36 results for the physical score was Even for many of the patients amenable to social services, À3.3 F 4.6 in the control group and +4.0 F 4.1 in the our social workers were unable to make significant inroads intervention group. Changes in the physical component into the system problems that they were facing. Often, the score were significantly different (diff = 7.1 F 2.7. P = .02; subjects had exhausted social services systems and were still 95% CI, 0.8-13.4) between the 2 groups. Change in the not improving. Again, this suggests that social services sup- mental component scale of the SF-36 was +13.1 F 20.6 in port needs to be focused on patients with solvable problems. the control group and +1.0 F 18.7 in the intervention group. ED social services have been depicted as one of the Changes in the mental component scale were not signifi- significant types of interventions, according to the Society cantly different (diff = 12.0, P = .4; 95% CI, À16.6 to 40.7). for Academic Emergency Medicine (SAEM) preventative There were physical score improvements in 4/5 patients in task force. The theoretical basis for this is found in the social the intervention group and none of the patients in the control services literature. A social worker within the medical sys- group. There were mental score improvements in 4/5 of tem can use Crisis Intervention and Task Centered theories both groups. The number of EMS transports was equivalent to provide optimum care in brief patient care settings. Payne between the 2 groups before the study at a rate of 3.2 [33] describes the Crisis Intervention model as an action to transports per person per month for both the control and interrupt a series of events, which lead to a disruption in intervention groups. The 6-month transport rate after the people’s normal functions, whereas Task Centered models intervention phase was 0.8 transports per person for the focus in the defined categories of a problem. According to a control group and 5.0 transports per person for the study by Keehn et al [9], the greatest decline of emergency intervention group. This represented a significant increase department recidivism occurred when social workers used in the intervention group’s use of EMS in the months after a proactive approach after the Task Centered model. Both the study period (RR = 1.76, 1.17-2.65, P b .01). models, although somewhat different, are attempting to improve a patient’s capacity to deal with the problems vital for overall life satisfaction [33]. Harrington [34] supports the 6. Discussion use of the Crisis Intervention model by stating that bthe emergency room social worker exemplifies in a crisis- We found that an attempt to help all nonhomeless 911 oriented context, to make a difference in real terms in repeat users with a social services intervention was not an situations of patient and family need—physical, environ- efficient use of social services time. This is the most crucial mental, and emotional needsQ. finding in the study. The social services team made serious On the basis of limited pilot study data, if patients were attempts to enlist all patients but were unable to include amenable to social services intervention, there was a 88% of all the patients they contacted. For an interesting set significant improvement in the physical portion but not of reasons, most frequent users of EMS were either the mental portion of the QOL. This was unexpected by us unavailable or unconcerned in the intervention. This because emotional support is more of a social workers’ suggests that the broad use of social services would lead focus. We believe that it may be a reflection of usage and to a lot of time spent trying to find patients and get them knowledge of resources. Many of the clients were house- Social Services and EMS 495
Table 3 Social services intervention suggestions 1. Falls Conduct a home safety assessment, including the following: assess the flooring for cracks, carpet, and steps. Do pull bars and hand rails need to be installed? Discuss the option of using a walker or cane. Help the patient with obtaining LIFE-ALERT, a program that provides an apparatus worn around the neck for easy accessibility to get immediate medical assistance if they are not near a phone. Assist with arranging full medical examination including optometrist, magnetic resonance imaging, and a yearly physical examination. 2. Drug and alcohol Aid patient in finding appropriate services, eg, narcotics anonymous, alcoholic anonymous, complications residential or outpatient detoxification support. Address psychosocial issues and stressors that may have occurred because of their substance abuse issue. Address medical complications that are a result of substance abuse such as, syncope seizures, contracted diseases from intravenous drug use, liver damage, and dental complications. 3. Epilepsy Help patient make the appropriate doctors appointments such as a neurologist as well as their primary doctor. Assist with obtaining the appropriate medications and make sure that the medications the patient has currently are not expired. Come up with a plan for patient and family/friend support if a seizure was to occur in a public area to keep the patient safe at all times. 4. Chest pain Find reasons the patient is calling 911 to differentiate whether the call is a medical or psychological issue, such as anxiety. If the patient’s chest pain is caused by cardiac complications, do they need to see a cardiologist and make a medical appointment? If the patient’s issue is psychiatric in nature identify the triggers for anxiety and discuss with client referrals for appropriate long-term counseling and a doctor for a medical evaluation. Help the patient with obtaining LIFE-ALERT. 5. Diabetes (both insulin and Assess the patient’s knowledge of illness. noninsulin dependent) Assist the client in obtaining doctors appointments for evaluation of medications and appropriate management skills. Provide education materials such as literature or videos. Do they need a public health nurse referral to help remind tem of their diabetes regimen? Is a nutritionist needed for evaluation of diet? bound and cared for by family members who may or may be enrolled. We found that the final results were closer not have proper medical training. Therefore, the intervention to 10%. may have improved the ability of the patient to physically Another limitation was bias among patients. Many of the manage his or her own care or the ability of the caregiver to patients were weary of being labeled as frequent users of the physically manage the patient’s care. system and were resistant to taking part in a study that We were surprised to find that patients receiving the tagged them as such. The low recruitment for a social intervention uniformly used EMS more in the follow-up services intervention was not a deficit in patient entry but a months of the study. We would explain this as being related to statement about the problems encountered by frequent users increased attention from the medical establishment leading to of the EMS system. The systems in place are not adequate an increased awareness of available resources. However, this for this marginal group of people. effect was on the basis of a small number of patients and needs to be reproduced in larger studies. Preventive services are an important part of the future of 8. Conclusion EMS. Consensus articles have been written but only a few studies have been completed. This direction for expansion Most frequent users were nonelderly adults. Three of scope of practice for EMS personnel is practical, realistic, categories of problems comprised the great majority of the and cost effective. cases (cardiac, asthma/COPD, and seizures). Use of a social services intervention on all frequent users of EMS is not an efficient use of their time as only 12% of patients contacted 7. Limitations can enter a meaningful interaction. In cases with meaningful interactions, these interactions do seem to improve physical The length of the study was a limitation. We originally QOL and increase EMS use. A more focused use of social estimated that 20% to 25% of patients contacted could services may be more efficient. 496 S.J. Weiss et al.
References validation of a new specific health status questionnaire: final report from the Italian QPD project involving 4000 patients. Am J Gastroenterol 1999;94:730-8. [1] Thumboo J, Fong KY, Ng TP, Leong KH, Feng PH, Thio ST, et al. [18] Anderson C, Laubscher S, Burns R. Validation of the Short Form 36 Validation of the MOS SF-36 for quality of life assessment of patients (SF-36) health survey questionnaire among stroke patients. Stroke with systemic lupus erythematosus in Singapore. J Rheumatol 1996;27:1812-6. 1999;26:97-102. [19] Marquis P, Fayol C, Joire JE. Clinical validation of a quality of life [2] Weiss SJ, Ernst AA, Miller P, Russell S. Repeat EMS transports questionnaire in angina pectoris patients. Eur Heart J 1995;16: among elderly emergency department patients. Prehosp Emerg Care 1554-60. 2002;6:6-10. [3] Pierce LL, Steiner V, Govoni AL. In-home online support for [20] Akashiba T, Kawahara S, Akahoshi T, Omori C, Saito O, Majima T, caregivers of survivors of stroke: a feasibility study. Comput Inform et al. Relationship between quality of life and mood or depression in Nurs 2002;20:157-64. patients with severe obstructive sleep apnea syndrome. Chest 2002; [4] D’Onofrio G, Degutis LC. Preventive care in the emergency 122:861-5. department: screening and brief intervention for alcohol problems in [21] Abresch RT, Carter GT, Jensen MP, Kilmer DD. Assessment of pain the emergency department: a systematic review. Acad Emerg Med and health-related quality of life in slowly progressive neuromuscular 2002;9:627-38. disease. Am J Hospice Palliat Care 2002;19:39-48. [5] Aminzadeh F, Dalziel WB. Older adults in the emergency department: [22] Birbeck GL, Hays RD, Cui X, Vickrey BG. Seizure reduction and a systematic review of patterns of use, adverse outcomes, and quality of life improvements in people with epilepsy. Epilepsia 2002; effectiveness of interventions. Ann Emerg Med 2002;39:238-47. 43:535-8. [6] Guo S, Biegel DE, Johnsen JA, Dyches H. Assessing the impact of [23] Ekman I, Fagerberg B, Lundman B. Health-related quality of life community-based mobile crisis services on preventing hospitalization. and sense of coherence among elderly patients with severe chronic Psychiatr Serv 2001;52:223-8. heart failure in comparison with healthy controls. Heart Lung 2002; [7] Babcock Irvin C, Wyer PC, Gerson LW. Preventive care in the 31:94-101. emergency department: Part II. Clinical preventive services—an [24] Rodger AJ, Jolley D, Thompson SC, Lanigan A, Crofts N. The impact emergency medicine evidence-based review. Society for Academic of diagnosis of hepatitis C virus on quality of life. Hepatology 1999; Emergency Medicine Public Health and Education Task Force 30:1299-301. Preventive Services Work Group. Acad Emerg Med 2000;7: [25] Rydman RJ, Isola ML, Roberts RR, Zalenski RJ, McDermott MF, 1042-54. Murphy DG, et al. Emergency department observation unit versus hos- [8] Naughton BJ, Moran MB, Feinglass J, Falconer J, Williams ME. pital inpatient care for a chronic asthmatic population: a randomized Reducing hospital costs for the geriatric patient admitted from the trial of health status outcome and cost. Med Care 1998;36:599-609. emergency department: a randomized trial. J Am Geriatr Soc 1994;42: [26] Shyu YI, Yip PK. Factor structure and explanatory variables of the 1045-9. Mini-Mental State Examination (MMSE) for elderly persons in [9] Keehn DS, Roglitz C, Bowden ML. Impact of social work on Taiwan. J Formos Med Assoc 2001;100:676-83. recidivism and non-medical complaints in the emergency department. [27] Agrell B, Dehlin O. Mini mental state examination in geriatric stroke Soc Work Health Care 1994;20:65-75. patients. Validity, differences between subgroups of patients, and [10] Belasco AG, Sesso R. Burden and quality of life of caregivers for relationships to somatic and mental variables. Aging (Milano) 2000; hemodialysis patients. Am J Kidney Dis 2002;39:805-12. 12:439-44. [11] Smith WR, Culley L, Plorde M, Murray JA, Hearne T, Goldberg P, [28] Grigoletto F, Zappala G, Anderson DW, Lebowitz BD. Norms for the et al. Emergency medical services telephone referral program: an Mini-Mental State Examination in a healthy population. Neurology alternative approach to nonurgent 911 calls. Prehosp Emerg Care 1999;53:315-20. 2001;5:174-80. [29] Braekhus A, Laake K, Engedal K. The Mini-Mental State Examina- [12] McCarthy WJ, Zhou Y, Hser YI, Collins C. To smoke or not to smoke: tion: identifying the most efficient variables for detecting cognitive impact on disability, quality of life, and illicit drug use in baseline impairment in the elderly. J Am Geriatr Soc 1992;40:1139-45. polydrug users. J Addict Dis 2002;21:35-54. [30] Lips P, Cooper C, Agnusdei D, Caulin F, Egger P, Johnell O, et al. [13] Failde I, Ramos I. Validity and reliability of the SF-36 health survey Quality of life in patients with vertebral fractures: validation of the questionnaire in patients with coronary artery disease. J Clin Quality of Life Questionnaire of the European Foundation for Epidemiol 2000;53:359-65. Osteoporosis (QUALEFFO). Working Party for Quality of Life of [14] Chang JA, Curtis JR, Patrick DL, Raghu G. Assessment of health- the European Foundation for Osteoporosis. Osteoporos Int 1999;10: related quality of life in patients with interstitial lung disease. Chest 150-60. 1999;116:1175-82. [31] Kopjar B. The SF-36 health survey: a valid measure of changes in [15] Berry C, McMurray J. A review of quality-of-life evaluations in health status after injury. Inj Prev 1996;2:135-9. patients with congestive heart failure. Pharmacoeconomics 1999;16: [32] Doraiswamy PM, Khan ZM, Donahue RM, Richard NE. The 247-71. spectrum of quality-of-life impairments in recurrent geriatric depres- [16] Bushnell DM, Martin ML. Quality of life and Parkinson’s disease: sion. J Gerontol A Biol Sci Med Sci 2002;57:M134-7. translation and validation of the US Parkinson’s Disease Question- [33] Payne M. Modern Social Work Theory. 2nd ed. Chicago, Ill: Lyceum naire (PDQ-39). Qual Life Res 1999;8:345-50. Books; 1997. [17] Bamfi F, Olivieri A, Arpinelli F, De Carli G, Recchia G, Gandolfi L, [34] Harrington DV. The E.R. social worker: cost effective, crisis-oriented et al. Measuring quality of life in dyspeptic patients: development and discharge planning—and more. Disch Plann Update 1991;4:8-11. American Journal of Emergency Medicine (2005) 23, 497–500
www.elsevier.com/locate/ajem
The sensitivity of room-air pulse oximetry in the detection of hypercapniaB
Michael D. Witting MD, MS*, Sam Hsu MD, Carlos Andres Granja MD1
Division of Emergency Medicine, University of Maryland School of Medicine, Baltimore, MD 21201, USA
Received 13 October 2003; accepted 7 December 2004
Abstract Objective: To estimate the sensitivity of room-air pulse oximetry in the detection of mode- rate hypercapnia. Methods: In this retrospective case-control study, charts were reviewed from patients with and without moderate hypercapnia (Paco2 N50 mm Hg), as determined by analysis of arterial blood gas samples obtained in the ED. Test characteristics (sensitivity, specificity, and likelihood ratios [LR]) for room-air pulse oximetry V96% to detect hypercapnia were calculated, as were confidence intervals. Results: A total of 349 charts were eligible for abstraction—92 cases and 257 controls. A room-air pulse oximetry reading V96% detected 88 of 92 cases of hypercapnia. Test characteristics were as follows (with 95% confidence interval): sensitivity, 0.96 (0.89-0.99); specificity, 0.39 (0.33-0.45), LR of a room-air pulse oximetry value N96%, 0.1 (0.04-0.3); and LR of a room-air pulse oximetry value V96%, 1.6 (1.4-1.7). Conclusion: Room-air pulse oximetry detects moderate hypercapnia with high sensitivity. D 2005 Elsevier Inc. All rights reserved.
1. Introduction arterial blood gas (ABG) sample, but this process is expensive, painful, labor-intensive, and time consuming. Respiratory status, one of the cardinal determinations in The advent of pulse oximetry, allowing real-time assessment every ED evaluation, has 2 components: oxygenation, of oxygenation, has revolutionized clinical practice. One assessed by the arterial blood oxygen (Pao2) level, and would expect a real-time test for hypercapnia to also have a ventilation, assessed by the arterial carbon dioxide (Paco2) great impact on clinical practice. However, although level. Either level can be measured accurately using an noninvasive methods of measuring CO2 levels have been invented [1,2], they require technology that is unavailable in most EDs and other outpatient settings. B Financial Support: University of Maryland Statewide Health Many reports, including one from the American Society Network, Other Tobacco-Related Diseases Research Grant, Baltimore, of Anesthesiologists, have indicated that pulse oximetry MD. cannot be used to assess ventilation, these claims being based T Corresponding author. Tel.: +1 443 310 2002. E-mail address: [email protected] (M.D. Witting). entirely on failure to detect hypoventilation in patients 1 Dr Granja is currently at the Hospital of Saint Raphael, New receiving supplemental oxygen [1,3-8]. However, in patients Haven, Conn. breathing room air, the alveolar gas equation imposes an
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.12.006 498 M.D. Witting et al. upper limit on the pulse oximetry value for any given Paco2 order for broom airQ oxygen saturation measurement, (4) value [9]. In fact, a room-air pulse oximetry cutpoint of 96% nurse’s note indicating broom airQ or the absence of oxygen detected 12 of 12 cases of hypercapnia in a small validation therapy. If 2 values at the same hierarchical level were set [10]. In this study, we estimate the sensitivity of room-air present, the one obtained closer to the time of ABG sampling pulse oximetry V96% to detect moderate hypercapnia was selected. Differences in selection of oximetry value (Paco2 N50 mm Hg) in a larger number of patients. positions were reconciled during regular training sessions. The assistant abstracted data from eligible charts using a standardized abstraction form. The following variables were 2. Materials and methods collected: age, the times of room-air pulse oximetry measurement and ABG sampling, whether the ABG sample This was a retrospective case-control study of patients was taken on supplemental oxygen, and whether improve- presenting to the ED of a university-affiliated community ment was noted during the ED visit. hospital. Pulse oximeters used in this ED were all from 2.3. Analysis Nellcor (Pleasanton, Calif) or Datascope (Montvale, NJ); all had a pulsed plethysmographic display, and all had specified A2Â 2 contingency table was constructed for hyper- accuracy within 2%. The study design was approved by the capnia status (V50 mm Hg, N50 mm Hg) versus room-air institutional review board. pulse oximetry test result. A positive room-air pulse oximetry test was defined prospectively as a value V96%, 2.1. Chart selection and a negative test was defined as a value N96%. This table A study manual was maintained throughout the data was used to calculate sensitivity, specificity, and likelihood collection period. Two groups were defined: a hypercapnia ratio (LR) values. Confidence intervals (CIs) for proportions were calculated using formulas described by Fleiss [11]. LRs group (Paco2 N50 mm Hg) and a control group (Paco2 V50 and CIs were calculated as described by Simel et al [12]. mm Hg). Lists of patients with Paco2 N50 mm Hg and with The sample size was chosen to allow, with a power of 0.8, Paco2 V50 mm Hg based on ABG sampling done in the ED were obtained from the hospital laboratory. Two unmatched adequate precision to exclude 0.9 from the 95% CI about a control patients were selected for each case. A trained sensitivity of 0.98. This required a minimum of 77 cases. assistant reviewed these lists and determined which patients met preliminary inclusion criteria: arrival by means other than ambulance and ABG obtained within 8 hours of ED 3. Results arrival. The assistant prepared charts that met preliminary Of 520 charts selected for secondary review, 171 were inclusion for chart review by eliminating all identifying excluded by one or both reviewers for the following information, masking all pulse oximetry and ABG values, reasons: no available pulse oximetry value, 30; all oximetry assigning each chart a study identification number, and values on supplemental oxygen, 71; arrival by ambulance assorting cases and controls in random order. (not detected by preliminary screen), 11; narcotic adminis- In the chart review, 2 board-certified emergency physi- cians independently reviewed the ED charts for eligibility according to the following exclusion criteria: arrival by ambulance; coma; absence of a documented room-air pulse oximetry value; potential decline in ventilation between pulse oximetry measurement and ABG sampling, such as when noted by decreased alertness, increased fatigue, narcotic administration, or ED intubation; a diagnosis of carboxyhemoglobinemia or methemoglobinemia. During periodic training sessions, these attending physicians compared charts, keeping only those with independent agreement for later abstraction. Percent agreement regarding eligibility was calculated. 2.2. Data abstraction During the independent review, each attending physician circled the position of the best masked room-air pulse oximetry value. The following hierarchy was used in Fig. 1 Scatterplot of room-air pulse oximetry values versus determining the best value: (1) physician note indicating Paco2 values. Graph includes 342 of 349 patients; 7 patients with b Q room air oxygen saturation, (2) triage value if the patient room-air pulse oximetry values from 52% to 74% and Paco2 was not on home oxygen, (3) nurse’s note in response to values from 57 to 85 mm Hg are not shown. Room-air pulse oximetry to detect hypercapnia 499
supplemental oxygen [1,3-8]. However, the common-sense Table 1 2 Â 2 Contingency table for room-air pulse oximetry notion that patients breathing room air will develop result versus moderate hypercapnia, from ABG result hypoxemia when they hypoventilate is based on a physical co N co N Pa 2 50 mm Hg Pa 2 50 mm Hg Total law—the alveolar gas equation [10]. This report provides Sat V96% 88 158 246 evidence that room-air pulse oximetry, as measured by Sat N96% 4 99 103 emergency personnel, can detect moderate hypercapnia 92 257 349 (Paco2 N50 mm Hg) with high sensitivity. An earlier report found that room-air pulse oximetry tration, 33; venous ABG sample, 7; intubated, 9; other V96% identified 12 of 12 patients with hypercapnia [10].In decline in status, 6; pediatric age, 2; and carbon monoxide that study, unlike the current one, pulse oximetry measure- exposure, 2. The remaining 349 charts met eligibility criteria ment and ABG sampling occurred nearly simultaneously, according to both reviewers. Agreement with regard to and most oximetry measurements were taken by the study’s eligibility of individual charts was 95%. principal investigator. The data from this study also suggest Of the 349 eligible charts, 92 were cases and 257 were that sensitivity to detect hypercapnia is quite high (96%) and controls. Cases had median age of 61 (interquartile range, that a room-air pulse oximetry value N96% provides strong 46-73) years versus 51 (interquartile range, 41-65) years for evidence against hypercapnia, with a LR [À] value of 0.1. controls. The triage saturation measurement was selected in One feature of this study is that pulse oximetry was 93% of charts. The lag between saturation and ABG measured outside a study protocol. Those who were sampling was b2 hours in 151 (43%), 2 to 5 hours in 156 measuring oxygen were not specially trained for the study. (45%), and 5 to 8 hours in 42 (12%), with a similar This is representative of most clinical settings. Pulse distribution between cases and controls. Interval improve- oximetry measurement is not taught routinely in medical ment was noted in 40 (43%) cases versus 57 (22%) controls. school, and we have noted that resident physicians and ABG values were determined on supplemental oxygen in 30 nurses occasionally record an erroneous oximetry measure- (57%) of 53 cases versus 27 (17%) of 156 controls where ment based on a low-amplitude tracing. It is possible that an this information was available. erroneous oximetry reading was recorded initially for some Fig. 1 shows a scatterplot of abstracted room-air pulse of the 4 false-negative patients, because oxygen was oximetry values versus Paco2 values.Asshown,the eventually given to 3 of them, 2 of whom were significant frequency of hypercapnia was inversely related to the outliers in the scatterplot. Improving measurement accuracy room-air pulse oximetry value. For each high oximetry may lead to even higher sensitivity than we observed. value (N96%), aside from a few outlying values, the In addition to limiting the use of ABG sampling, room-air distribution of Paco2 values is skewed toward lower values. pulse oximetry hypercapnia screening has a potential use in This is consistent with the concept that at high room-air conscious sedation. The goal of conscious sedation is pulse oximetry values, the alveolar gas equation imposes an achieving a depth of anesthesia at which patients can still upper bound on Paco2 values. respond to stimulation while maintaining airway protection The association between the prospectively defined room- and ventilatory reflexes [8,13]. In an emergency setting, air pulse oximetry test (V96%) and moderate hypercapnia where patients frequently have full stomachs, conscious (Paco2 N50 mm Hg from ABG testing) is shown in Table 1. sedation is the preferred depth of anesthesia. In these patients, The following test characteristics are noted (with 95% CI): failure to detect hypoventilation may cause excessive sensitivity, 0.96 (0.89-0.99); specificity, 0.39 (0.33-0.45); sedation and undue risk. If an end-tidal CO2 monitor is not LR [À], 0.1 (0.04-0.3); and LR [+], 1.6 (1.4-1.7). available, a room-air pulse oximetry drop to 95% or below Of the 4 false-negative cases, 3 were on supplemental can alert the treating physician to the onset of hypoventila- oxygen at the time of ABG sampling, and oxygen status was tion, cautioning against inducing deeper sedation. undocumented in the other. Two of the 3 patients (97%, 70 mm Hg; 99%, 80 mm Hg) were significant outliers. All 4.1. Limitations and future questions 21 patients with ABG noted to be measured on room air had a room-air pulse oximetry value N96%. This study was limited by a time lag between pulse oximetry measurement and ABG sampling, during which a patient’s hypercapnia status may have changed. Interval 4. Discussion worsening or improvement could have affected our results. It is not clear whether the clinicians recognized or recorded Chapters on virtually every emergency medicine topic these changes; hypercapnia is virtually never documented in emphasize the primary importance of the bABCs.Q Avoid- an ED setting for patients breathing room air. In ED ance of hypercapnia requires that the first 2, bairwayQ and practice, clinicians tend to respond to low oximetry values bbreathing,Q are intact. Many authors have emphasized the by giving supplemental oxygen and implementing measures limitations of pulse oximetry in detecting hypercapnia, to improve ventilation, not by obtaining a room-air ABG based solely on its poor record in patients breathing measurement. Inducing hypercapnia for study purposes 500 M.D. Witting et al. would be unethical, so we felt that this study design was the References safest, most efficient way to evaluate the potential of room- air oximetry to screen for hypercapnia in an ED setting. [1] Nelson DB, Freeman ML, Silvis SE, et al. A randomized controlled An additional limitation of this study is that it was a trial of transcutaneous carbon dioxide monitoring during ERCP. retrospective chart review. Retrospective chart reviews Gastrointest Endosc 2000;51:288-95. [2] Ward WR, Yealy DM. End-tidal carbon dioxide monitoring in often suffer from missing data, and such reviews are emergency medicine, part 1: basic principles. Acad Emerg Med prone to bias unless specific methodologic steps are taken 1998;5:628-36. [14]. The methods we used to limit bias were formal [3] Stoneham MD. Uses and limitations of pulse oximetry. Br J Hosp inclusion criteria, blinding reviewers to pulse oximetry Med 1995;54:35-41. values and case/control status, a standardized abstraction [4] Place B. Pulse oximetry: benefits and limitations. Nurs Times 2000; 96:42. form, prospectively determined selection of room-air pulse [5] Lowton K. Pulse oximeters for the detection of hypoxaemia. Prof oximetry values, maintenance of a study manual, and Nurse 1999;14:343-7. periodic training [14]. [6] Hutton P, Clutton-Brock T. The benefits and pitfalls of pulse oximetry. Despite the limitations of this study’s design, it allowed BMJ 1993;307:457-8. inclusion of many more patients with hypercapnia than the [7] Davidson JA, Hosie HE. Limitations of pulse oximetry: respiratory insufficiency—a failure of detection. BMJ 1993;307:372-3. earlier study and produced a more precise sensitivity [8] Gross JB, Bailey PL, Connis RT, et al. Practice guidelines for sedation estimate. The use of room-air oximetry as a screening test and analgesia by non-anesthesiologists. Anesthesiology 2002;96: for hypercapnia requires further study. One way to do this 1004-17. [9] Rose CC, Wolfson AB. Respiratory physiology. Emerg Med Clin noninvasively would be to substitute end-tidal CO2 mea- surement for formal ABG measurement. End-tidal CO can North Am 1989;7:187-204. 2 [10] Witting MD, Lueck CH. The ability of pulse oximetry to screen for be measured simultaneously with room-air pulse oximetry, hypoxemia and hypercapnia in patients breathing room air. J Emerg either during conscious sedation or in a population with a Med 2001;20:341-8. high prevalence of hypercapnia, such as a clinic for patients [11] Fleiss J. Statistical methods for rates and proportions. 2nd ed. New with chronic obstructive pulmonary disease. York (NY)7 John Wiley & Sons; 1981. p. 13-5. [12] Simel DL, Samsa GP, Matchar DB. Likelihood ratios with confidence: sample size estimation for diagnostic test studies. J Clin Epidemiol Acknowledgments 1991;44:763-70. [13] Proudfoot J. Analgesia, anesthesia, and conscious sedation. Emerg Med Clin North Am 1995;13:357-79. The authors thank Dale Woolridge, MD, for his help with [14] Gilbert EH, Lowenstein SR, Koziol-McLain J, et al. Chart reviews in early preparation and study design. They also thank Linda emergency medicine research: where are the methods? Ann Emerg Kesselring, MS, ELS, for editorial support. Med 1996;27:305-8. American Journal of Emergency Medicine (2005) 23, 501–503
www.elsevier.com/locate/ajem
Correlation between chest x-ray and B-type natriuretic peptide in congestive heart failure
Chad Aleman MD, Shu B. Chan MD, MS*, Mary Frances Kordick MBA, PhD, RN
Resurrection Emergency Medicine Residency Program, Resurrection Medical Center, Chicago, IL 60631, USA
Received 27 November 2004; accepted 13 December 2004
Abstract Introduction: B-Type natriuretic peptide (BNP) values greater than 400 pg/mL have high positive likelihood ratio (N10) for the diagnosis of clinical congestive heart failure (CHF). However, in patients with CHF, it is not known what correlation, if any, exists for the BNP levels above 400 pg/mL and the findings on the initial chest x-ray (CXR). Methods: Retrospective review of emergency department patients with CHF and initial BNP greater than 400 pg/mL. Descriptive statistics were analyzed and logistic regression was performed. Results: Fifty-four patients mean age of 81.7 (SD, 8.2), 64.8% women. The mean BNP was 1493 pg/mL (SD, 1106). Only 68.5% had a finding of CHF on CXR. Logistic regression showed no correlation (Wald P = .568; R2 = 0.8%). Conclusion: In patients with clinical CHF, there is no correlation between very high BNP levels (N400 pg/mL) and CXR readings. Clinicians should not be surprised to find patients with very high BNP levels but negative CXR. D 2005 Elsevier Inc. All rights reserved.
1. Introduction diagnosis and treatment in a cost-saving manner is the next step in heart failure management [4,5]. The diagnosis and treatment of heart failure continues to Recent studies have proved the accuracy of BNP levels be an extraordinary cost burden to the health care system in the diagnosis of CHF in the emergency department (ED) [1]. Rapid bedside B-type natriuretic peptide (BNP) testing [6]. Elevated BNP levels were also shown to be more has made the diagnosis of congestive heart failure (CHF) accurate than chest x-ray (CXR) findings in the diagnosis of more rapid and more accurate for the clinician [2,3]. CHF [7]. An algorithmic approach to the diagnosis of Figuring out how to integrate new BNP testing into the patients presenting with dyspnea has been proposed. In this algorithm, BNP levels greater than 400 pg/mL should be treated as if CHF is bvery likelyQ [8]. The proposal is based on the fact that BNP levels greater than 400 pg/mL display a Presented at: American College of Emergency Physician Scientific Assembly Research Forum October 17 to 18, 2004, San Francisco, Calif. positive likelihood ratio greater than 10 [9]. T Corresponding author. Tel.: +1 773 594 7816; fax: +1 773 594 7805. This retrospective chart review had the objective of E-mail address: [email protected] (S.B. Chan). determining the correlation between very elevated BNP
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.12.009 502 C. Aleman et al.
Fig. 1 Included/excluded cases. levels and CXR findings confirming the diagnosis of CHF. the 65 cases reviewed, 6 were excluded because the final We hypothesize that very high BNP levels would, in fact, diagnosis was not available, and 5 were excluded because predict a positive confirmatory CXR. of a final diagnosis other than CHF leaving a total of 54 patients. The mean age was 81.7 years (SD, 8.2). Women 2. Methods comprise 64.8% of the cases. Dyspnea was the chief complaint in 94.4%, with the remaining cases presenting We designed a retrospective 6-month chart review of with a chief complaint of pedal edema. There was a history patients with a final diagnosis of CHF or pulmonary edema of CHF in 59.3% of the patients. The mean BNP level was N and a very high BNP levels ( 400 pg/mL). Only patients 1493 pg/mL (range, 411-4910). Eighteen patients (33.3%) with ED chief complaints of dyspnea or leg edema were had CXR readings of pulmonary edema. Eleven patients included. Exclusion criteria included renal failure (serum (20.4%) had CXR readings of CHF. Eight patients (14.8%) N creatinine 2.5 mg/dL), final diagnosis other than CHF, and had CXR readings of both CHF and pulmonary edema. those in which a final diagnosis was not available. Combining those 3 groups revealed a cumulative total of 37 The setting of our study is a suburban/urban community patients (68.5%) with CXR readings consistent with active hospital with a very large elderly population. Our ED is heart failure. Other x-ray readings are shown in Table 1. staffed by residents and attending physicians. Inpatient On logistic regression, BNP levels and CXR readings services are staffed by residents in some cases and attending were not correlated. The coefficient for BNP was 1.000 physicians with final diagnosis left to the attending (95% CI, 1.000-1.001). The Wald P value was .568 and the physicians. The attending radiologists performed all CXR R2 was only 0.8%. readings. Both ED and inpatient records were reviewed so that only cases of actual clinical CHF were included in the study. Internal review board approval was obtained before 4. Discussion the study. The data collected included age, sex, history of CHF, BNP levels, CXR readings, ED diagnosis, and final Because of its high prevalence in the US population, discharge diagnosis. heart failure is a significant financial burden on the health Descriptive statistics were analyzed and logistic regres- care system as a whole [1]. Recent advances in diagnosis sion was performed to correlate BNP levels with CXR findings. All statistical calculations were performed on SPSS Ver 11.5 (SPSS, Inc, Chicago, Ill, 2002). Table 1 CXR not read as CHF or pulmonary edema CXR reading No. (%) Cardiomegaly 6 (11.1%) 3. Results Cardiomegaly with effusion 4 (7.4%) Cardiomegaly with infiltrates 3 (5.6%) A total of 82 cases were reviewed (Fig. 1). Seventeen Effusion only 2 (3.7%) cases were excluded before chart review based on the fact Infiltrates only 2 (3.7%) that the patients had renal failure on laboratory analysis. Of BNP__CXR 503 and treatment of heart failure have offered hope of trimming be surprised to find patients with very high BNP levels but some of those costs [5]. Physical and radiographic signs of negative CXR. heart failure are unreliable [10,11]. Recently, elevated BNP levels have proven to be a more reliable indicator of heart failure than CXR findings [7]. Our study evaluates whether References there is a correlation between elevated BNP values greater than 400 pg/mL and CXR findings. [1] Balinsky W, Muennig P. The costs and outcomes of multifaceted interventions designed to improve the care of congestive heart failure The results of our study show that there is virtually no in the inpatient setting: a review of the literature. Med Care Res Rev correlation between very high BNP levels and CXR 2003;60(3):275-93. findings. One secondary result of note was that of the [2] Wieczorek SJ, Wu AH, Christenson R, Krishnaswamy P, Gottlieb S, 5 patients excluded for final diagnosis other than heart Rosano T, et al. A rapid B-type natriuretic peptide assay accurately failure; none would have likely had their diagnosis and diagnoses left ventricular dysfunction and heart failure: a multicenter evaluation. Am Heart J 2002;144(5):834-9. management altered by CXR findings. The one patient with [3] McCullough PA, Nowak RM, McCord J, Hollander JE, Herrmann HC, bronchiectasis had known prior bronchiectasis. Our study Steg PG, et al. B-Type natriuretic peptide and clinical judgment in numbers, however, are far too small to comment on the emergency diagnosis of heart failure: analysis from Breathing Not ability of BNP levels to exclude alternative diagnoses. Properly (BNP) Multinational Study. Circulation 2002;106(4):416-22. [4] Valli N, Gobinet A, Bordenave L. Review of 10 years of the clinical 4.1. Study limitations and future questions use of brain natriuretic peptide in cardiology. J Lab Clin Med 1999; 134(5):437-44. Primary limitations of our study include the age of our [5] Ogawa K, Oida A, Sugimura H, Kaneko N, Nogi N, Hasumi M, et al. patient population, the high prevalence of cardiovascular Clinical significance of blood brain natriuretic peptide level measure- ment in the detection of heart disease in untreated outpatients: disease in our institution, and the relatively small study size. comparison of electrocardiography, chest radiography and echocardi- The age of our patient population may not accurately reflect ography. Circ J 2002;66(2):122-6. the average age in most emergency rooms. The high [6] Maisel AS, Krishnaswamy P, Nowak RM, McCord J, Hollander JE, prevalence of cardiovascular disease may be the primary Duc P, et al. Rapid measurement of B-type natriuretic peptide in the reason that there were no more patients excluded because of emergency diagnosis of heart failure. N Engl J Med 2002;347(3): 161-7. alternate diagnoses. [7] Dao Q, Krishnaswamy P, Kazanegra R, Harrison A, Amirnovin R, Our study indicates that there is no correlation between Lenert L, et al. Utility of B-type natriuretic peptide in the diagnosis of very high BNP levels and CXR findings. As previously congestive heart failure in an urgent-care setting. J Am Coll Cardiol stated, BNP has proven more reliable than CXR findings in 2001;37(2):379-85. the diagnosis of heart failure. Further studies may be needed [8] Maisel A. B-Type natriuretic peptide measurements in diagnosing congestive heart failure in the dyspneic emergency department patient. to determine whether CXR results do in fact alter the Rev Cardiovasc Med 2002;3(Suppl 4):S10-7. diagnosis and the management of heart failure patients with [9] Gibler WB, Blomkalns AL, Collins SP. Evaluation of chest pain and very high BNP levels. Clinicians should not be surprised heart failure in the emergency department: impact of multimarker when patients have very high BNP levels, yet display strategies and B-type natriuretic peptide. Rev Cardiovasc Med insignificant findings on CXR. In this situation, it may be 2003;4(Suppl 4):S47-S55. [10] Chakko S, Woska D, Martinez H, de Marchena E, Futterman L, prudent to treat according to BNP results. Kessler KM, et al. Clinical, radiographic, and hemodynamic correlations in chronic congestive heart failure: conflicting results 4.2. Summary may lead to inappropriate care. Am J Med 1991;90(3):353-9. [11] Thomas JT, Kelly RF, Thomas SJ, Stamos TD, Albasha K, Parrillo JE, In patients with clinical CHF, there is no correlation et al. Utility of history, physical examination, electrocardiogram, and between very high BNP levels (N400 pg/mL) and a finding chest radiograph for differentiating normal from decreased systolic of CHF or pulmonary edema on CXR. Clinicians should not function in patients with heart failure. Am J Med 2002;112(6):437-45. American Journal of Emergency Medicine (2005) 23, 504–509
www.elsevier.com/locate/ajem
Reviews Sudden cardiac death in athletes: a guide for emergency physicians
Carl A. Germann MD*, Andrew D. Perron MD
Department of Emergency Medicine, Maine Medical Center, Portland, ME 04102, USA
Received 4 September 2004; revised 4 September 2004; accepted 12 September 2004
Abstract A conditioned athlete is usually regarded as a member of the healthiest segment of society, and exercise itself is looked upon as a means to improve health. Although extremely uncommon, sudden cardiac death (SCD) in young athletes is a devastating medical event to all involved (patient, family, community, team, and caregivers). Most etiologies of SCD in athletes result in the same final common denominator (cardiac arrest) on presentation to an emergency physician. There are, however, certain historic, physical examination, and electrocardiographic features of many of these disease processes that emergency physicians should have a working knowledge of to try to identify them before they result in SCD. This review examines the clinical presentation, diagnostic techniques, and management options applicable to emergency practitioners. D 2005 Elsevier Inc. All rights reserved.
1. Introduction Conditioned athletes are regarded to constitute the healthiest segment of the society while exercise itself is The first proposed case of sudden cardiac death (SCD) in considered a means of improving personal health and an athlete involved Pheidippides in 490 bc. This trained decreasing the chance of cardiovascular disease. Although runner was reported to have collapsed and died after running relatively uncommon, the sudden death of an athlete can to Athens from Marathon with the news of the military have a tragic and devastating impact on the family, defeat of Persia [1]. More recent accounts of athletes dying community, and medical staff involved. For many years, suddenly include Loyola Marymount basketball player the lack of understanding regarding the causes of such Hank Gathers (1990), marathon runner Jim Fixx (1984), catastrophes was reflected by vague descriptions of these US Olympic volleyball player Flo Hyman (1986), former events such as sudden death syndrome [2]. Over the past professional basketball players Pete Maravich (1980) and decade, however, a greater comprehension of the under- Reggie Lewis (1993), and Olympic figure skater Sergei lying cardiovascular processes of SCD in athletes has Grinkov (1995). All these professional athletes suffered been achieved. from presumed SCD. The role of emergency physicians (EPs) in this setting is primarily to attempt resuscitation and stabilization of the patients. However, it is helpful for EPs to be aware of the T Corresponding author. Tel.: +1 207 842 7015; fax: +1 207 842 7054. demographics, risk factors, etiologies, and current preven- E-mail address: [email protected] (C.A. Germann). tive strategies regarding SCD.
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.09.036 Sudden cardiac death in athletes: a guide for emergency physicians 505
2. Definition African Americans constitute only about 12% of the US population, this subset represents a large portion (N40%) of There is no universally accepted definition for SCD. One individuals experiencing SCD. This fact is probably a result popular definition is that of a nontraumatic and unexpected of a disproportional level of participation in certain sudden cardiac arrest that occurs within 6 hours of a competitive sports. previously normal state of health [3]. Other definitions limit the time frame in relation to sport participation and symptoms anywhere from within 1 to 24 hours [4-6]. 4. Causes of SCD in athletes The term athlete implies involvement in a regimented exercise program with participation in a team or individual Sudden death in young athletes is predominantly caused sport. To better differentiate the etiology of SCD, athletes by structural non–atherosclerotic heart disease [17]. Sudden are often classified as byoungQ or bold.Q Although open for cardiac death is caused by the combination of exercise and debate, for the purpose of this review, the term young underlying heart disease rather than by exercise alone, athlete will refer to those aged younger than 35 years which usually leads to the final common pathway of lethal because most studies in this arena have defined it as such. arrhythmia [11,18-20]. By far, the most common reason for SCD is HCM, accounting for at least 36% of deaths [15]. The second most frequent cause of death, encountered in 17% to 19% of deaths, is congenital coronary anomalies 3. Epidemiology [11,15]. The next most common cause of SCD in young Sudden cardiac death in young athletes is rare. The exact athletes is idiopathic left ventricular hypertrophy (ILVH), prevalence is not known because there is no national accounting for 9% to 10% of cases [11,15]. Infrequent database to track the death of athletes. The largest available causes include aortic rupture, arrhythmogenic right ventric- studies estimate the incidence among high school and ular dysplasia, aortic valve stenosis, prolonged QT syn- collegiate athletes to be between 1 per 100000 and 1 per drome, mitral valve prolapse (MVP), commotio cordis, 300000, respectively, each year [7-10]. An estimated Wolff-Parkinson-White (WPW) syndrome, and atheroscle- number of 50 to 100 cases occurs annually in the United rotic coronary artery disease. States [7,9,11]. On the other hand, in mature athletes (those older than 35 The overall estimate of incidence may be low given that years), coronary artery disease is found in most cases of fatal arrhythmias associated with subtle structural abnor- SCD [3,5,21,22]. In fact, atherosclerotic heart disease is malities or a normal heart would be difficult or impossible to found to be the cause of SCD in anywhere from 73% to 95% identify on postmortem examination. Also, often it is a of this patient population [5,6,10,23,24]. coroner’s priority to exclude foul play rather than establish a 4.1. Hypertrophic cardiomyopathy precise cardiovascular diagnosis [12]. Likewise, lack of a national registry for SCD in athletes might lead to an By and large, HCM is the predominant cause of sudden underestimation. There is often reliance on media coverage death in young athletes, occurring in more than one third of rather than on voluntary or mandatory reporting for all cases [15]. The estimated prevalence of HCM in the identifying cases of SCD. general population is about 1 in 500 [11], although many Despite the overall low incidence of SCD, it represents cases may go undetected during a patient’s lifetime. 30% of all nontraumatic deaths, with one third of victims The diagnostic criteria for HCM include a left ventricle younger than 65 years [13]. For young athletes, SCD that is hypertrophied but not dilated in the absence of other remains a significant cause of nontraumatic exercise–related cardiac or systemic diseases that would produce left death. The incidence of cardiovascular collapse as the cause ventricular hypertrophy [23]. This pathological hypertrophy of athletic fatalities in high school and college athletes contributes to decreased ventricular compliance and diastol- outnumbers death caused by trauma by nearly 2 to 1 [11,14]. ic dysfunction with impaired filling. The left ventricular Most (N80%) cases of SCD in young athletes occur outflow obstruction and possibly small lumen intramural either during or immediately after strenuous exercise [15]. vessels may potentiate myocardial ischemia. Arrhythmias This suggests that physical activity may be a trigger for arise in the context of an electrically unstable myocardial cardiac arrhythmias in those with certain cardiac disease. substrate caused by cardiac muscle cell disorganization, Athletes suffering from SCD participate in a large variety of replacement fibrosis, or myocardial ischemia [24]. sports, most frequently basketball and football (about 68%) Hypertrophic cardiomyopathy is genetically transmitted [15]. Most cases of SCD occur in men with a ratio ranging as an autosomal dominant condition with variable expres- from 5:1 to 9:1 [8,15]. This may be because of the lower sion because it may not manifest itself until adolescence participation rate of women in competitive sports in or young adulthood [23,25]. If a family history of HCM addition to the higher incidence of hypertrophic cardiomy- is confirmed, it is recommended that all first-degree opathy (HCM) among men [15]. A considerable proportion relatives be examined with echocardiography [23,26].It of SCD occurs in African American athletes [16]. Although may be difficult, however, to distinguish mild HCM from 506 C.A. Germann, A.D. Perron the normal cardiac hypertrophy that occurs in highly 4.3. Idiopathic left ventricular hypertrophy trained athletes [27]. Unfortunately, often the first clinical manifestation of This condition accounts for approximately 10% of all HCM is sudden death. In a series of SCD occurring SCD in young athletes [11,15]. Idiopathic left ventricular during exercise in patients with HCM, only 35.7% had a hypertrophy is a symmetric, concentric hypertrophy where previous cardiac evaluation for risk factors that included the left ventricular mass exceeds that of physiological syncope, family history, murmurs, fatigue, or ventricular hypertrophy. Unlike HCM, ILVH is not associated with tachycardia on exercise testing [17]. In another study, 21% genetic transmission and there is no cellular disarray ofathleteswhodiedfromHCMhadsymptomsof microscopically. However, it is uncertain whether some cardiovascular disease before their death [15]. Symptoms cases of ILVH represent mild morphological expressions of may include chest pain, exertional dyspnea, lightheaded- HCM, unusual instances of exercise-induced cardiac hyper- b Q ness, or syncope. trophy ( athlete’s heart syndrome ) with nonbenign con- Hypertrophic cardiomyopathy should be suspected in sequences, or possibly examples of undetected right any patient in whom a harsh systolic ejection murmur is ventricular dysplasia with mild left ventricular hypertrophy heard on examination. The characteristic murmur increases [27,30-32]. The precise mechanism of death is poorly in intensity with maneuvers that decrease venous return established but thought to be similar to that of HCM. such as Valsalva’s maneuver. Examination may also reveal the presence of a fourth heart sound or a rapid initial upstroke of the carotid pulse. These patients will also 5. Less common causes of SCD demonstrate signs of left ventricular enlargement electro- cardiographically and radiographically. If HCM is suspected 5.1. Myocarditis clinically, the diagnosis should be confirmed by echocardi- This inflammatory condition of the myocardium is ography. If seen in the ED, patients should be proscribed frequently the result of a viral infection most commonly from exertion or exercise until they can have an echocar- caused by either coxsackievirus or echovirus. Characteristic diogram and cardiology consultation. symptoms of myocarditis include a prodromal viral illness followed by progressive exercise intolerance and congestive 4.2. Coronary artery abnormalities symptoms of dyspnea, cough, and orthopnea. On physical examination, an audible S3 gallop or soft apical murmur as Second in frequency to HCM as a cause of SCD in this well as signs of congestive heart failure may be present. The population is congenital coronary abnormalities. These electrocardiogram (ECG) may show low-voltage com- congenital abnormalities are also infrequently diagnosed plexes, diffuse ST-segment and T-wave changes, and sinus during life. The most common coronary anomaly associated tachycardia. Sudden cardiac death may occur in the with SCD in athletes is a left main artery arising from the presence of either active or healed myocarditis [9]. Thus, right sinus of Valsalva [28,29]. Coronary anomalies a convalescent period of at least 6 months is recommended presumably lead to myocardial hypoperfusion during before a return to sports [9,33]. exercise; however, the precise mechanism is often unknown. Possible reasons for ischemia in anomalous coronary artery 5.2. Mitral valve prolapse include a slit-like ostium that narrows with aortic dilatation during exercises, an acute-angled takeoff, and compression Idiopathic MVP is the most common valvular disorder, of the artery as it passes between the aorta and pulmonary occurring in approximately 5% of the population [28,34]. trunk [3,9,15,23]. Other coronary anomalies include origin Although sudden death from MVP has been reported, it is of the right coronary artery from the left sinus of Valsalva or extremely rare and most athletes with MVP are totally from the pulmonary artery, presence of a single coronary asymptomatic. Physical examination of patients with MVP artery, and coronary aneurysm [3,8,9,15]. may reveal a midsystolic click and a late systolic murmur. If Some individuals with coronary artery anomalies may an athlete with known MVP develops syncope, exertional demonstrate symptoms such as syncope or angina before chest pain, or moderate-to-severe mitral regurgitation, it is an event. Among athletes who died of this disorder, 31% recommended that athletic participation be restricted [33]. were found to have symptoms before death [15]. A high 5.3. Aortic rupture index of suspicion in conjunction with an echocardiogram may help suggest the presence of an anomaly in order for Aortic rupture makes up 5% to 7% of sudden death in angiography to be obtained and to confirm this often young athletes according to investigations by Maron et al surgically correctable abnormality [6]. It should be noted [3,15]. Half of these cases occur in athletes with the Marfan that recent research have shown that both magnetic syndrome, which is an autosomal dominant disorder with a resonance imaging and computed tomographic coronary prevalence in the general US population of 1 per 10000 angiography have demonstrated promise in detecting [9,35]. In the Marfan syndrome, the aortic media are coronary artery anomalies [30,31]. deficient in the number of elastic fibers, which leads to Sudden cardiac death in athletes: a guide for emergency physicians 507 weakening of the aortic wall and predisposes the individual interval of more than 440 milliseconds. The mechanism of to aortic dissection and death. A diagnosis of Marfan SCD is the development of wide, polymorphic tachycardia syndrome is based on clinical features as well as family (ie, torsades de pointes) [11,39]. These potential fatal history. Common manifestations involve the skeletal, rhythms can be provoked by exercise-related tachycardia; ophthalmologic, and cardiovascular systems [35]. Skeletal thus, anyone with this condition should be restricted from features include tall stature with arm span greater than competitive sports. height, arachnodactyly, hyperextensible joints, scoliosis, a high-arched palate, and pectus excavatum. Ophthalmologic 5.7. Commotio cordis examination may show myopia and ocular lens subluxation, Commotio cordis is an electrophysiological event and echocardiography may reveal a dilated aortic root or caused by precordial chest impact that occurs in individuals MVP. Patients with suggestive physical features or a free from structural cardiac disease. The mechanism of family history of Marfan syndrome may undergo clinical SCD appears to be ventricular fibrillation that is produced and echocardiographic evaluation or DNA testing to when the chest impact is delivered within a narrow, confirm the diagnosis. electrically vulnerable period of the cardiac cycle [40]. Specifically, the susceptible time is during repolarization, 5.4. Arrhythmogenic right ventricular dysplasia just before the peak of the T wave. Resuscitation of these Arrhythmogenic right ventricular dysplasia was found to victims is possible with prompt cardiopulmonary resusci- be the most common cause of sudden death in young tation and defibrillation. A US commotio cordis registry athletes in the Veneto region of Northern Italy [36]. has reported a 10% survival rate, with 2.8% experiencing However, this disease appears to be much less prevalent full recovery [41]. in the United States, occurring in approximately 3% of SCD 5.8. Brugada syndrome cases [15]. Right ventricular dysplasia is an autosomal dominant condition leading to fibrosis and fatty infiltration The Brugada syndrome, described as a syndrome of SCD of the right ventricle. caused by unpredictable episodes of recurrent ventricular This process results in the thinning and dilatation of the tachycardia, is also postulated to be a possible etiology of right ventricular wall and leads to recurrent and intractable death in this patient population. It is characterized as a ventricular tachyarrhythmias. Diagnosis of this disease often syndrome of SCD in individuals with a structurally normal proves difficult. If echocardiography does not demonstrate heart and no evidence of atherosclerotic coronary artery right ventricular dilation and dysfunction, a magnetic disease. Patients are noted (if identified before their event) resonance imaging can be diagnostic in demonstrating fatty to have a distinct set of ECG abnormalities that include a infiltrates of the myocardium [11,37]. complete or incomplete right bundle branch block with ST- segment elevation in the right precordial leads. Originally 5.5. Wolff-Parkinson-White syndrome thought to be a disease primarily of men from Southeast Asia, it is now identified in women, children, and non– The WPW syndrome is an abnormality of the cardiac Asian ethnic groups. Patients may present with self- conduction system whereby an additional electrical pathway terminating episodes of ventricular tachycardia during can lead to tachycardia and, rarely, SCD [11,37,38]. This exertion. If the diagnosis is supported by ECG findings, conduction abnormality is relatively rare, affecting only referral should be made for EP testing and consideration of 0.15% to 0.2% of the general population, and has a very automated implantable cardiac defibrillator placement. small risk of sudden death (b0.1%) [11]. These individuals may have various symptoms including palpitations, synco- 5.9. Atherosclerotic coronary artery disease pe, and lightheadedness. The mechanism of SCD is the development of atrial fibrillation with rapid atrioventricular Acquired coronary artery disease is occasionally respon- conduction via the bypass tract, resulting in rapid ventricular sible for sudden death in young athletes. Atherosclerotic response and subsequent ventricular fibrillation [38]. The coronary artery disease is thought to be responsible for resting ECG may demonstrate an initial slurred upstroke of approximately 2% of SCD cases in young athletes [15]. the QRS complex (delta wave), short PR interval, and wide However, atherosclerotic disease is responsible for the vast QRS complex. majority of SCD cases in mature athletes (N35) [6,23].
5.6. Long QT syndrome 6. Medications The long QT syndrome is an abnormality of the electrical conduction system characterized by prolonged Sudden cardiac death has been attributed to proarrhyth- repolarization of the ventricle with an associated high risk mic drugs such as epinephrine, ephedrine, cocaine, and of SCD [11]. This syndrome can be congenital, pharmaco- related sympathetic medications. Deaths have also been genic, or metabolic. It is often defined as a corrected QT linked to performance-enhancing agents such as erythro- 508 C.A. Germann, A.D. Perron poietin and anabolic steroids [5,23]. Anabolic steroids may 8. Conclusion cause cardiac hypertrophy, myocardial fibrosis, and accel- erated atherosclerosis that could promote cardiac necrosis Sudden cardiac death rarely occurs in athletes with a and ischemia. structurally normal heart. Therefore, an assessment for structural heart disease is essential in evaluating athletes with possible cardiac signs or symptoms. Because certain 7. Emergency department management SCD patients will lack a history of symptoms and physical findings, not all patients at risk will be identified. However, The role of EPs is primarily to attempt resuscitation of by knowing how to guide history and physical examination, patients experiencing SCD events. Although few athletes potential diagnostic clues may help identify most patients at are electrocardiographically monitored at the time of death, risk for SCD. Once cardiac arrest has occurred in patients, most sudden deaths are thought to be caused by ventricular most of them may initially appear the same and death may arrhythmias in the setting of an underlying heart disease. In not be preventable. In many cases, however, warning one population of 16 athletes who survived cardiac arrest, 2 symptoms, a suggestive family history, clinical findings, were found to have long QT syndrome, 5 had WPW or underlying conduction abnormalities will be present. syndrome, 8 had ventricular tachycardia or ventricular fibrillation, and 1 athlete had atrial fibrillation with heart block [42,43]. Treatment of these rhythms is established by References advanced cardiac life support based on American College of Cardiology/American Heart Association recommendations [1] Van Camp SP. Sudden death. Clin Sports Med 1992;11:273-89. and continued data from ongoing studies. Individuals [2] Maron BJ. Triggers for sudden cardiac death in the athlete. Cardiol experiencing successful resuscitation require a complete Clin 1996;14(May):195-210. cardiac workup including echocardiogram, cardiac catheter- [3] Maron BJ, Epstein SE, Roberts WC. Causes of sudden death in ization, and electrophysiological testing. competitive athletes. J Am Coll Cardiol 1986;7:204-14. [4] Corrado D, Thiene G, Nava A, et al. Sudden death in young Emergency physicians should maintain a high index of competitive athletes: clinicopathologic correlations in 22 cases. Am J suspicion when syncope or near-syncope occurs in the Med 1990;89:588-96. presence of exercise. Exertional syncope is a distinct red [5] Reisdorff EJ, Prodinger RJ. Sudden cardiac death in the athlete. flag and all patients with this symptom should be Emerg Med Clin North Am 1998;16:281-94. excluded from participation until they have a complete [6] Futterman LG, Myerburg R. Sudden death in athletes: an update. Sports Med 1998;26:335-50. cardiac evaluation. Likewise, the evaluation of athletes is [7] Ades PA. Preventing sudden death: cardiovascular screening of young similar to that of other patients with possible symptoms athletes. Phys Sportsmed 1992;20:75-89. and signs of arrhythmic disorders. Symptoms of arrhyth- [8] Van Camp SP, Bloor CM, Mueller FO, et al. Nontraumatic sports mias in athletes are similar to those in nonathletes and death in high school and college athletes. Med Sci Sports Exerc range from brief palpitations to syncope and resuscitated 1995;27:641-7. [9] Drezner JA. Sudden cardiac death in young athletes. Postgrad Med sudden death [42]. The key elements of the evaluation 2000;108(Oct):37-50. include the severity of the symptoms, the presence of [10] Ragosta M, Crabtree J, Sturner WQ, et al. Death during recreational structural heart disease, and the family medical history. In exercise in the State of Rhode Island. Med Sci Sports Exerc one study, 18% of athletes with SCD were known to have 1984;16:339-42. had symptoms judged to be cardiovascular in origin [11] Basilico FC. Cardiovascular disease in athletes. Am J Sports Med 1999;27:108-21. during the 36 months preceding death [15]. These symp- [12] Firoozu S, Sharma S, McKenna WJ. Risk of competitive sport in toms included chest pain, syncope, exertional dyspnea, and young athletes with heart disease. Heart 2003;89:710-4. dizziness. A family history of congenital heart disease, [13] Lown B. Sudden cardiac death: the major challenge confronting syncope, or unexpected death at an early age may indicate contemporary cardiology. Am J Cardiol 1979;43:313-28. an underlying cardiac abnormality. The physical examina- [14] Cantu RC. Congenital cardiovascular disease: the major cause of athletic death in high school and college. Med Sci Sports Exerc tion may elicit evidence of Marfan syndrome, aortic 1992;24:279-80. stenosis, MVP, HCM, or ILVH. When findings of the [15] Maron BJ, Shirani J, Poliac LC, et al. Sudden death in young physical or historic evaluation suggest a syndrome associ- competitive athletes. Clinical, demographic, and pathological profiles. ated with high risk for SCD, further diagnostic testing is JAMA 1996;276:199-204. necessary. This will include an ECG and usually an [16] Maron BJ, Carney KP, Lever HM. Relationship of race to sudden cardiac death in competitive athletes with hypertrophic cardiomyop- echocardiogram. Likewise, long-term ECG monitoring with athy. J Am Coll Cardiol 2003;41:974-80. a Holter monitor can be useful in patients who present with [17] Maron BJ, Roberts WC, McAllister HA, et al. Sudden death in young frequent or reproducible symptoms. Investigative work and athletes. Circulation 1980;62:218-29. further risk stratification will usually be performed or [18] Kohl HW, Powell KE, Gordon NF, et al. Physical activity, physical followed up by a primary care physician. Participation in fitness, and sudden cardiac death. Epidemiol Rev 1992;14:37-58. [19] Myerburg RJ, Kessler KM, Castellanos A. Sudden cardiac death: high-risk sports activity should remain forbidden until this structure, function, and time-dependence of risk. Circulation workup is complete. 1992;85(Suppl. 1):2-10. Sudden cardiac death in athletes: a guide for emergency physicians 509
[20] Pelliccia A, Maron BJ, Culasso F, et al. Athlete’s heart in women: screening, and standards for disqualification. Card Electrophysiol echocardiographic characterization of highly trained elite female Rev 2002;6:100-3. athletes. JAMA 1996;276:211-5. [33] Spirito P, Maron BJ, Bonow RO, et al. Severe functional limitation in [21] Virmani R, Robinowitz M, McAllister Jr HA. Nontraumatic death patients with hypertrophic cardiomyopathy and only mild localized in joggers: a series of 30 patients at autopsy. Am J Med 1982; left ventricular hypertrophy. J Am Coll Cardiol 1986;8:537-44. 72:874-81. [34] Thiene G, Nava A, Corrado D, et al. Right ventricular cardiomyopathy [22] Waller BF, Roberts WC. Sudden death while running in conditioned and sudden death in young people. N Engl J Med 1988;318:129-33. runners aged 40 years or over. Am J Cardiol 1980;45:1282-300. [35] Maron BJ, Mitchell JH. Recommendations for determining eligibility [23] Rich BSE. Sudden death screening. Med Clin North Am 1994; for competition in athletes with cardiovascular abnormalities. 26th 78:267-88. Bethesda Conference. J Am Coll Cardiol 1994;24:845-99. [24] Northcote RJ, Ballantyne D. Sudden cardiac death in sport. BMJ [36] Jeresaty R. Mitral valve prolapse: definition and implications in 1983;287:1357-9. athletes. J Am Coll Cardiol 1986;7:231-6. [25] Maron BJ, Gaffney FA, Jeresaty RM. Task force III: hypertrophic [37] Braunwald E, Zipes DP, Libby P, editors. Heart disease: a textbook cardiomyopathy, other myopericardial diseases and mitral valve of cardiovascular medicine. 6th ed. Philadelphia (Pa)7 Saunders; prolapse. J Am Coll Cardiol 1985;6:1215-7. 1996. p. 2000-3. [26] McCaffrey FM, Braden DS, Strong WB. Sudden cardiac death in [38] Marcus FI, Fontaine G. Arrhythmogenic right ventricular dysplasia/ young athletes. Am J Dis Child 1991;145:177-83. cardiomyopathy: a review. Pacing Clin Electrophysiol 1995;18: [27] Maron BJ, Pelliccia A, Spirito P. Cardiac disease in young trained 1298-314. athletes: insights into methods for distinguishing athlete’s heart from [39] Gallagher JJ, Gilbert M, Svenson RH, et al. Wolff-Parkinson-White structural heart disease with particular emphasis on hypertrophic syndrome: the problem, evaluation, and surgical correction. Circula- cardiomyopathy. Circulation 1995;91:1596-601. tion 1975;51:767-85. [28] Wight JN, Salem D. Sudden cardiac death and the dathlete’s heartT. [40] Klein GJ, Bashore TM, Sellers TD, et al. Ventricular fibrillation in the Arch Intern Med 1995;155:1473-80. Wolff-Parkinson-White syndrome. N Engl J Med 1979;301:1080-5. [29] Burke AP, Farb A, Virmiani R. Causes of sudden death in athletes. [41] Tan HL, Hou CJ, Lauer MR, et al. Electrophysiological mechanisms Cardiol Clin 1992;10:303-17. of the long QT interval syndromes and torsades de pointes. Ann Intern [30] Schoepf UJ, Becker CR, Ohnesorge BM, et al. CT of coronary artery Med 1995;122:701-14. disease. Radiology 2004;232(1):18-37. [42] Perron AD, Brady WJ, Erling BF. Commotio cordis: an underappre- [31] van Geuns RJ, Wielopolski PA, de Bruin HG, et al. Magnetic ciated cause of sudden cardiac death in young patients: assessment resonance imaging of the coronary arteries: techniques and results. and management in the ED. Am J Emerg Med 2001;19:406-9. Prog Cardiovasc Dis 1999;42(2):157-66. [43] Maron BJ, Link MS, Wang PJ, et al. Clinical profile of commotio [32] Maron BJ. The young competitive athlete with cardiovascular cordis: an underappreciated cause of sudden death in the young during abnormalities: causes of sudden death, detection by preparation sports and other activities. J Cardiovasc Electrophysiol 1999;10:114-20. American Journal of Emergency Medicine (2005) 23, 510–516
www.elsevier.com/locate/ajem
Diagnostics Electrocardiographic ST-segment elevation in the trauma patient: acute myocardial infarction vs myocardial contusion
Claire U. Plautz MDa, Andrew D. Perron MDb, William J. Brady MDa,* aDepartment of Emergency Medicine, University of Virginia School of Medicine, Charlottesville, VA 22908, USA bDepartment of Emergency Medicine, Maine Medical Center, Portland, ME 04102, USA
Received 5 March 2004; accepted 6 March 2004
Abstract The diagnosis of myocardial contusion in the setting of blunt trauma engenders much discussion and controversy—partly because of the lack of a gold standard for its identification other than histologic findings at autopsy. Furthermore, blunt cardiac trauma represents a spectrum of disorders ranging from transient electrocardiographic change to sudden death from myocardial rupture; hence, no single terminology exists to define such a wide range of scenarios. Here, we present 2 cases of electrocardiographic ST-segment elevation after high-speed motor vehicle crashes resulting in numerous injuries, including blunt chest trauma. Both patients demonstrated electrocardiographic ST-segment elevation, resulting from myocardial contusion and acute myocardial infarction. D 2005 Published by Elsevier Inc.
1. Introduction obtained by the clinician suspecting the potential for significant myocardial injury, should be interpreted within Blunt chest trauma may affect all of the thoracic organs, the context of the clinical situation. It must be stressed that including the heart, with sequelae ranging from subclinical the ECG is not the ultimate diagnostic tool in the evaluation myocardial contusion to potentially life-threatening hemor- of the patient with possible traumatic myocardial injury. In rhage. These diverse injuries are classified into direct and fact, the 12-lead ECG is rarely used in the vast majority of indirect causes, as it relates to the mechanism of trauma and trauma patients. the physiologic response of the cardiovascular system. The The diagnosis of myocardial contusion engenders much electrocardiographic manifestations related to the insult can discussion—this is partly because of the lack of a gold be equally diverse, ranging from a normal electrocardio- standard for its identification other than histologic findings gram (ECG) or subtle, inconsequential changes, to sudden at autopsy. Furthermore, blunt cardiac trauma represents a cardiac death with ventricular fibrillation. The ECG, spectrum of disorders ranging from minor, transient electrocardiographic change to sudden death from myocar- dial rupture; hence, no single terminology exists to define such a wide range of scenarios (Table 1). Not surprisingly, * Corresponding author. Tel.: +1 804 982 2529; fax: +1 804 924 2877. incidence rates among trauma victims range from E-mail address: [email protected] (W.J. Brady). bnegligibleQ to 76% [1,2]. Although moderate speeds have
0735-6757/$ – see front matter D 2005 Published by Elsevier Inc. doi:10.1016/j.ajem.2004.03.014 ST-segment elevation in the blunt trauma patient 511
Table 1 Clinical and electrocardiographic findings in blunt cardiac injury (A) Prospective investigations
(B) Retrospective investigations
been implicated in the mechanism, these injuries often result after low-impact chest trauma, usually secondary to from high-speed deceleration incidents. Other causes ventricular fibrillation. The blow is thought to arrive at the include sports injuries, falls, blast injuries, crush trauma, vulnerable period of cardiac repolarization, causing the and direct blows to the chest. Mechanisms proposed to well-studied R-on-T phenomenon [3], and producing explain its evolution include sudden impaction and direct ventricular fibrillation. compression of the heart against the sternum and vertebrae, Myocardial contusion has been used to define injuries as well as intrathoracic transmission of increased intra- including transient and persistent dysrhythmias, traumatic abdominal pressure. myocardial infarctions, cardiac rupture resulting in pericar- Blunt cardiac trauma encompasses a wide range of dial tamponade, mechanical complications including valve clinical entities. The most acute form, termed myocardial rupture or septal perforation, pericardial injuries, and concussion, occurs when blunt injury to the anterior chest damage to coronary arteries [4-10]. Histologic findings in stuns the myocardium and produces a transient dysrhythmia, the myocardium are often seen at autopsy, typically hypotension, or loss of consciousness. Within this class of involving subendocardial hemorrhage associated with focal disorders, commotio cordis refers to sudden cardiac death areas of edema, interstitial hemorrhage, and cytolysis.
Fig. 1 Wide complex tachycardia consistent with monomorphic ventricular tachycardia. 512 C.U. Plautz et al.
Fig. 2 Twelve-lead ECG demonstrating normal sinus rhythm with PAC and ST segment elevation in leads V1 to V4 consistent with either acute myocardial infarction or myocardial contusion.
Finally, traumatic myocardial infarctions are known. Reports of this electrocardiographic finding is similar to the of direct injury to coronary vasculature causing intimal tears, nontrauma chest pain presentation, particularly concerning vessel rupture, and traumatic occlusion have attempted to AMI and myocardial contusion. explain the pathophysiology. More importantly, many of the cases of acute myocardial infarction (AMI) after blunt chest trauma have shown no injury to the coronary vessels at all, 2. Case presentations but occur as a result of severe, direct injury to myocardium 2.1. Case 1 itself. Clinical outcomes have varied from excellent to fatal, often determined by the presence of other injuries. A 62-year-old man with a history of hypertension Short of autopsy, there is no gold standard for the presented to the ED after a motor vehicle crash. The patient diagnosis of myocardial contusion. The ECG, serum cardiac was the restrained driver of a small truck that suffered a marker, and echocardiogram have been used in conjunction head-on collision with a large tree at approximately 45 mph. to identify these cases [11]. Other tests, including radionu- Although there was significant damage to the front of the clide scanning and computed tomography, have been used, vehicle, the patient was able to self-extricate. The patient but their utility has not been established [1,11]. was transported by emergency medical services (EMS) to Although various algorithms for the diagnosis and the ED. On arrival, the patient noted chest pain. Examina- management of myocardial contusions exist, no single tion revealed an alert man with moderate distress; a strong method has been chosen as a bbest practice.Q There remains odor of alcohol was present. The chest revealed bilateral no gold standard for the diagnosis of blunt myocardial breath sounds with tenderness in the mid-sternal area; the injury. In the ED, diagnostic modalities are most important abdomen was tender in the right upper quadrant. While in attempting to differentiate low-risk patients who can be assessing the patient, a 20-beat run of ventricular tachycar- discharged safely home from those who may develop dia was noted (Fig. 1). A lidocaine bolus with infusion was cardiac complications—the former being exceedingly more started. A 12-lead ECG revealed prominent ST-segment common than the latter. Among those diagnostic modalities elevation in the right to mid-precordial leads (Fig. 2). Rapid studied are use of the history and physical examination, the consultation was made to both trauma surgery and ECG, serial cardiac markers, and the echocardiogram. cardiology while appropriate radiographs and computed This report will review the electrocardiographic findings tomographic scans were performed. A bedside echocardio- of traumatic myocardial injury. We present 2 cases of blunt gram revealed hypokinesis of the right ventricle and the trauma patients presenting with chest pain and electrocar- anterior wall of the left ventricle. The radiographic trauma diographic ST-segment elevation. The differential diagnosis evaluation was unremarkable. The patient was then taken
Fig. 3 Sinus tachycardia with early ST segment elevation and prominent T waves. ST-segment elevation in the blunt trauma patient 513
Fig. 4 Twelve-lead ECG demonstrating sinus rhythm with significant ST-segment and T-wave changes. Prominent T waves are seen in leads V2 to V4. ST-segment elevation is also seen in this distribution. ST-segment depression is also noted in the inferior leads. emergently to the cardiac catheterization laboratory where 2.2. Case 2 normal coronary arteries with normal flow were noted. The patient was transferred to the intensive care unit and A 54-year-old man with a past history of angina monitored. No further dysrhythmias were noted. The presented to the ED via EMS with substernal chest pain patient’s ECG demonstrated normalization of the ST- after a single motor vehicle crash. The automobile struck a segment changes over the next 72 hours. Creatinine telephone pole at approximately 55 mph. The EMS phosphokinase values were elevated with positive MM evaluation revealed evidence of alcohol intoxication, a and MB fractions. The patient was discharged with a tender chest, and an electrocardiographic rhythm strip (V diagnosis of myocardial contusion. lead) demonstrating sinus tachycardia with prominent T- wave and ST-segment elevation (Fig. 3). Examination in the
Fig. 6 Inverted T waves in leads V1 to V5. In addition, a Fig. 5 ST-segment depression in leads V2 to V6 with variable prominent R wave in leads V1 and V2 is seen likely consistent T-wave abnormalities (biphasic to inverted). with incomplete right bundle dysfunction. 514 C.U. Plautz et al.
Fig. 7 Complete right bundle branch block.
ED revealed an intoxicated man complaining of chest pain cluding nodal tachycardias (Fig. 9), atrial fibrillation and with chest tenderness. A 12-lead ECG demonstrated flutter, ventricular tachycardia (Fig. 2), and ventricular prominent T waves and early ST-segment elevation in leads fibrillation (Fig. 10). V1 to V4 along with ST-segment depression in the inferior The electrocardiographic manifestations of cardiac trau- leads (Fig. 4). An expedited radiographic trauma evaluation ma are diverse and very nonspecific. Blunt myocardial was undertaken while both surgery and cardiology con- injury is seen not infrequently in the patient with blunt sultations were requested. The patient was taken to the thoracic injury; one large study reported a 54% incidence of catheterization laboratory. Left-ventricle catheterization electrocardiographic abnormality in patients with blunt revealed 90% proximal left anterior descending artery chest injury. Such injury may present electrocardiographi- occlusion with thrombus. The ventriculogram showed cally in many ways. The vast majority of electrocardio- hypokinesis of the anterior wall of the left ventricle; cardiac graphic abnormalities were nonspecific with nonspecific ST- serum markers were elevated, confirming the diagnosis of segment–T-wave changes predominating; the remainder of acute myocardial infarction. The ultimate cause of the findings included conduction abnormalities, axis deviation, infarction was never determined—whether it resulted from and dysrhythmias [13]. The most frequently encountered the blunt injury to the coronary anatomy or a spontaneous electrocardiographic pattern included sinus tachycardia with acute coronary event—although the patient had noted new- nonspecific ST-segment–T-wave changes—extremely non- onset intermittent chest discomfort (consistent with angina) specific findings which are seen in many trauma victims, over the preceding 2 weeks. both with and without cardiac injury. Although less frequent yet more suggestive of myocardial injury, conduction system abnormality is another electrocardiographic sign of 3. Discussion cardiac contusion, manifested most often by right bundle branch block followed by nonspecific intraventricular The ECG has long been used as an initial diagnostic conduction delay. Significant ST-segment changes, includ- study in blunt trauma patients with chest pain. Maenza et al ing both elevation and depression, as well as T-wave [1] conducted a large review of prospective and retrospec- tive studies, concluding that abnormal electrocardiographic findings correlated directly with complications requiring treatment; these investigators noted that a normal ECG was associated with clinically insignificant injuries. Dysrhyth- mias were the most common manifestation, with sinus tachycardia (Fig. 3) present in up to 70% of patients with confirmed myocardial contusion [12]. Other electrocardio- graphic findings included nonspecific ST-segment–T-wave changes, ST-segment elevation and depression (Fig. 5), T- wave abnormalities (Fig. 6), atrioventricular and bundle branch blocks (Figs. 7 and 8), bradycardia, premature Fig. 8 Second-degree, high-grade atrioventricular block in a ventricular contractions, supraventricular tachycardias in- hypotensive blunt trauma patient with myocardial injury. ST-segment elevation in the blunt trauma patient 515
Fig. 9 (A, B) Narrow complex tachycardias consistent with junctional (nodal) tachycardia. abnormalities are also seen in blunt trauma patients with troponin does not definitively diagnose cardiac contusion, cardiac injury. The entire range of dysrhythmia is encoun- a normal troponin in a hemodynamically stable patient tered; in the majority of such presentations, sinus tachycar- 6 hours after injury reliably excludes significant blunt dia is the only dysrhythmia encountered followed by cardiac injury [15]. Furthermore, when used in conjunction frequent premature ventricular contractions—these 2 dys- with a normal admission ECG, the combined negative rhythmias are very nonspecific and are more often predictive value approaches 100% [16]. As with the ECG, encountered in trauma patients without significant blunt the decision to measure cardiac serum markers must be cardiac trauma. Blunt thoracic trauma may cause direct based upon the individual clinical presentation; it is injury to the coronary arteries with spasm, dissection, or anticipated that such testing will be rarely required. thrombosis; these events would all present with ST- Transthoracic echocardiograms have also been used to segment–T-wave changes typical of an acute coronary evaluate the presence and extent of cardiac injury. However, event. its use as a screening tool is greatly limited by its lack of In the majority of penetrating thoracic trauma presenta- immediate availability in the ED, high cost, and operator tions, the ECG as a test has low priority in the initial expertise. Although use of transthoracic echocardiograms in stabilization and evaluation phase. If performed in this patients with known myocardial contusions is widely situation, the ECG may reveal ST-segment changes sugges- known, its utility in risk stratifying blunt chest trauma tive of acute coronary syndrome. A retrospective review of patients with possible cardiac injury has been consistently penetrating cardiac trauma revealed multiple patterns in the inconclusive [1]. Combined with the aforementioned presenting ECGs, including repolarization changes in 35%, diagnostic screening tests, however, the TEE becomes very pericarditis in 27% of patients, and changes consistent with useful as a secondary study for evaluating the extent of AMI in 10% [14]. Penetrating pericardial injury may also myocardial damage. manifest changes consistent with pericardial effusion and Dysrhythmias remain the most common complication of cardiac tamponade (tachycardia, low voltage, or electrical myocardial contusion. These range from the frequently alternans). benign sinus tachycardia, atrial fibrillation, and premature Other diagnostic studies, such as serum markers and ventricular contractions to the more ominous atrioventricu- echocardiography, are of value in the trauma patient. lar and intraventricular blocks and ventricular dysrhythmias. Cardiac serum marker measurements have a significant role These dysrhythmias may be transient or long-term. It is in the diagnosis of cardiac contusion. Although earlier thought that injuries to the anterior left side of the chest studies have previously discounted its benefit, emergence of predominantly induce ventricular rhythm disturbances, the highly sensitive and specific troponin within the last whereas injuries to the right side of the chest produce decade has suggested that this test is more useful than sinoatrial or atrioventricular conduction disorders [17]. creatinine kinase measurements. Although an elevated Given its anatomical position, the right ventricle appears more susceptible to blunt injury, in some reports twice as frequently injured, than the left ventricle [2]. Although relatively uncommon, myocardial infarction can result from blunt chest trauma. Retrospective reviews show an incidence of approximately 5% among patients Fig. 10 Ventricular fibrillation. with significant blunt chest injury [1]. Multiple mechanisms 516 C.U. Plautz et al. have been reported, including complete occlusion [3] Perron AD, Brady WJ, Erling BF. Commodio cordis: an underappre- [10,12,17,18] or direct injury to coronary vasculature [5]. ciated cause of sudden cardiac death in young patients: assessment Meluzin reports the occurrence of myocardial infarction and management in the ED. Am J Emerg Med 2001;19:406-9. [4] Curzen N, Brett S, Fox K. Concrete induced cardiac contusion. Heart diagnosed 3 weeks after a multitrauma patient’s discharge 1997;78:313-5. from the hospital, with cardiac catheterization later showing [5] Meluzin J, Groch L, Toman J, Hornacek I, Fischerova B. Rupture of a ruptured coronary artery as culprit [5]. More common, the coronary artery after blunt nonpenetrating chest wall trauma however, is the finding of normal coronaries in patients with detected by color Doppler echocardiography: a case report. J Am Soc ST-segment elevation after blunt chest trauma, supporting Echocardiogr 2000;13:1043-6. [6] Cloutier RL, Mehr MF, Lin RJ, Tanel RE. Junctional ectopic the diagnosis of myocardial contusion [8,9,17]. tachycardia in association with blunt abdominal trauma. Ann Emerg Treatment of a patient with suspected myocardial contu- Med 2002;40:308-12. sion remains supportive [19,20]. In cases when myocardial [7] Moulin A, Bason-Mitchell M, Henderson SO. Cardiac contusion infarction is suspected (ie, ST-segment elevation is ob- diagnosed by cardiac troponin [correspondence]. Am J Emerg Med served), coronary angiography with potential angioplasty is 2002;20:382-3. [8] Tsoukas A, Andreades A, Zacharogiannis C, Kifnidis K, Karlis P, et al. the preferred method of evaluation and management. Myocardial contusion presented as acute myocardial infarction after Thrombolytic agents are contraindicated in the setting of chest trauma. Echocardiography 2001;18:167-70. acute trauma [12]. Dysrhythmias are treated with appropriate [9] Atalar E, Acil T, Aytemir K, Ozer N, Ovunc K, et al. Acute anterior medications. As yet, there is no role for prophylactic myocardial infarction following a mild nonpenetrating chest trauma— medications to suppress dysrhythmias; however, measures a case report. Angiology 2001;52:279-82. [10] Goktekin O, Unalir A, Gorenek B, Kudaiberdieva G, Cavusoglu Y, should be taken to treat and prevent conditions that increase et al. Traumatic total occlusion of left main coronary artery caused by myocardial irritability (ie, metabolic acidosis or electrolyte blunt chest trauma. J Invasive Cardiol 2002;14:463-5. imbalance) [12]. When cardiogenic shock is present, judi- [11] Greenberg MD, Rosen CL. Evaluation of the patient with blunt chest cious intravenous fluids should be administered, followed trauma: an evidence-based approach. Emerg Med Clin North Am by appropriate vasopressors and inotropes as indicated. 1999;17:41-62. [12] Marx J, Hockberger R, Walls R, editors. Rosen’s emergency medicine—concepts and clinical practice. 6th ed. St Louis (Mo): Mosby, Inc; 2004. p. 528-36. 4. Conclusion [13] Helling TS, Duke P, Beggs CW, Crouse LJ. A prospective evaluation of 68 patients suffering blunt chest trauma for evidence of cardiac With the absence of a gold standard for diagnosis, injury. J Trauma 1989;29:961-5 [with discussion 1989;29:965-6]. [14] Duque HA, Florez LE, Moreno A, Jurado H, Jaramillo CJ, et al. myocardial contusion remains an elusive entity. Admission Penetrating cardiac trauma: follow-up study including electrocardiog- electrocardiogram and cardiac enzyme measurements have raphy, echocardiography, and functional test. World J Surg 1999; been shown to correlate with the presence of cardiac injury, 23:1254-7. and those with normal results on admission may be [15] Collins JN, Cole FJ, Weireter LJ, Riblet JL, Britt LD. The usefulness discharged safely to home in the absence of comorbid of serum troponin levels in evaluating cardiac injury. Am Surg 2001;67:821-5 [with related discussion 2001;67:825-6]. injury or hemodynamic instability. Myocardial infarctions [16] Salim A, Velmahos GC, Jindal A, Chan L, Vassiliu P, et al. Clinically can occur after blunt chest trauma, and immediate treatment significant blunt cardiac trauma: role of serum troponin levels in the catheterization laboratory is advocated rather than combined with electrocardiographic findings. J Trauma 2001; anticoagulating the patient who is at high risk for exsangui- 50:237-43. nation. Finally, ST-segment elevation has many etiologies, [17] Tintinalli JE, Krome RL, Ruiz E, editors. Emergency medicine: a comprehensive study guide. New York (NY)7 McGraw-Hill, Inc; and the emergency physician must be astute in recognizing 2003. p. 1690-3. acute injury in the appropriate clinical setting [21]. [18] Gunnar WP, Martin M, Smith RF, Manglano R, Resnick DJ, et al. The utility of cardiac evaluation in the hemodynamically stable patient with suspected myocardial contusion. Am Surg 1991;57:373-7. [19] Lindstaedt M, Germing A, Lawo T, von Dryander S, Jaeger D, et al. References Acute and long-term clinical significance of myocardial contusion following blunt thoracic trauma: results of a prospective study. J [1] Maenza RL, Seaberg D, D’Amico F. A meta-analysis of blunt cardiac Trauma 2002;52:479-85. trauma: ending myocardial confusion. Am J Emerg Med 1996;14: [20] Orliaguet G, Ferjani M, Riou B. The heart in blunt trauma. 237-41. Anesthesiology 2001;95:544-8. [2] Weiss RL, Brier JA, O’Connor W, Ross S, Brathwaite CM. The [21] Brady WJ, Aufderheide TP, Chan T, Perron AD. Electrocardiographic usefulness of transesophageal echocardiography in diagnosis cardiac diagnosis of acute myocardial infarction. Emerg Med Clin North Am contusions. Chest 1996;109:73-7. 2001;19:295-320. American Journal of Emergency Medicine (2005) 23, 517–520
www.elsevier.com/locate/ajem
Diagnosis of periorbital gas on ocular ultrasound after facial trauma
S. Timothy McIlrath MD, Michael Blaivas MD, RDMS*, Matthew Lyon MD, RDMS
Department of Emergency Medicine, Medical College of Georgia, Augusta, GA 30912-4007, USA
Received 14 October 2004; accepted 14 October 2004
Abstract Ocular trauma can occur from isolated facial trauma or in major blunt trauma such as motor vehicle accidents or falls. Despite the etiology of the injury, a thorough evaluation is important but may often be difficult if severe swelling is present. Recently, emergency ultrasound has seen the use of ocular ultrasound to evaluate visual changes and trauma. Literature suggests that unsuspected and difficult to diagnose pathology may be easily detected on ultrasound of the orbit. We present 3 cases of isolated facial trauma in which routine evaluation with ocular ultrasound led to the discovery of periorbital air with one patient having air insufflating the upper lid of the affected side. D 2005 Elsevier Inc. All rights reserved.
1. Introduction suspected, CT may not be ordered and unsuspected facial injuries could be missed. The eye can be particularly Facial trauma is seen as a result of multiple etiologies in difficult to examine as the patient may experience signifi- the emergency department. Drivers involved in motor cant pain and swelling. Previous work has shown that vehicle crashes who chose not to restrain themselves are emergency physicians are able to accurately diagnose ocular more likely to suffer head and subsequent facial injury. pathology with ultrasound with little or no physical However, a new safety restraint introduced into most cars evaluation of the eye being required initially [1]. over the last decade, the airbag, itself is responsible for We present 3 cases of periorbital air diagnosed on facial injury in persons sitting too close to the dash. bedside ultrasound. In all cases, manipulation of the lids was Although brain injury is frequently of greatest concern for difficult and the extent of injury found on ultrasound was the evaluating emergency physician, injury to the facial not suspected clinically. Computed tomography, not initially bones, orbits, or eyes requires immediate attention as well. planned for in these cases, was ordered and confirmed Typically, head and facial injury is now evaluated using orbital fractures. computed tomography (CT). This has largely replaced both the wait-and-see approach to head injury of the past as well as plain films to evaluate for occult skull and facial bone 2. Case 1 fractures. However, when significant head injury is not MH is a 24-year-old woman who was dancing in a club and was punched in the left eye by an unseen attacker T Corresponding author. Tel.: +1 706 721 4467. 8 hours before presentation. The patient suffered no loss of E-mail address: [email protected] (M. Blaivas). consciousness and continued to dance. When she returned
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.10.001 518 S.T. McIlrath et al.
the medial wall of the left orbit through the lamina papyracea. After ophthalmology consultation, the patient was started on prophylactic antibiotics and managed conservatively with clinic follow-up.
3. Case 2
WT is a 21-year-old man who was kicked in the left eye 2 hours before presentation. He denied loss of conscious- ness but did complain of generalized headache as well as left eye pain and blurry vision. He recalled having a bloody nose. Presenting vital signs were normal. Initial physical exam was remarkable for moderate orbital swelling and conjunctival injection in the affected eye. Visual acuity was 20/40 in the left vs 20/20 in the right. A moderate hyphema was present in the left eye. Pupils were equal and reactive and extraocular muscles were grossly intact. Fundoscopic Fig. 1 Note the significant swelling of the left upper eyelid. exam was normal. Bedside ultrasound examination demonstrated no evi- home and fell asleep, she noticed increasing swelling and dence of retinal detachment; however, retroorbital gas was ecchymosis of the left eye but had no complaints of visual seen (Fig. 3). Head and orbital CT confirmed left lamina deficits. Upon awakening 5 hours later, she noticed severe papyracea fracture with ethmoid air cell opacification and orbital swelling and thought she noticed increased swelling extensive left orbital emphysema and proptosis. The patient of the upper eyelid after sneezing. On presentation, the was seen by ophthalmology in the emergency department. patient had normal and stable vital signs. The left lid was He was instructed to avoid heavy lifting or nose blowing swollen and ecchymotic. Retraction of the upper lid was and to sleep with the head of the bed elevated. Eye drops difficult without significant manipulation because of the were prescribed as well as prophylactic oral antibiotics. He swelling (Fig. 1). A bedside ultrasound examination was was seen in follow-up in ophthalmology clinic for several performed to evaluate for globe injury and traumatic retinal weeks and was eventually discharged. detachment. The globe could only be visualized through the inferior aspect of the lower lid because of soft tissue gas in the upper lid of the patient’s eye (Fig. 2). The globe itself 4. Case 3 appeared intact and no evidence of retinal detachment was noted. A facial CT was ordered, which revealed fractures of CN is a 24-year-old man who was a restrained driver involved in a head on motor vehicle crash at moderate
Fig. 2 The eye is seen through the lower lid. Air in the middle of Fig. 3 The eye is seen with the lens anterior, at the top of the the swollen upper lid is seen as bright reflectors (arrows) and casts a image. The singe arrow points to a bright area that represents the large dirty shadow blocking all useful information deep to it. A air. Dirty shadowing is seen behind the air (arrows) similar to that indicates anterior chamber; G, globe; arrow, lens. seen from bowel gas in abdominal ultrasound. Periorbital gas on ultrasound 519 speed. His airbag deployed striking him in the chest and years, emergency physicians have used ultrasound as an forehead. The patient complained of abrasions to the effective tool in the evaluation of the smaller structures of forehead and blurry vision in the left eye but suffered no the eye in the emergency setting, screening for such loss of consciousness. Epistaxis, which occurred after conditions as vitreous hemorrhage, retinal detachment, impact, had stopped during transport to the emergency central retinal artery/vein occlusion, and globe rupture department. The patient’s vital signs were normal and stable. [1]. Ultrasound is also a useful modality in the evaluation Physical examination showed multiple abrasions to the of orbital trauma. The bony orbit is especially well suited to forehead and swelling of the left upper lid as well as dried ultrasound imaging. The markedly different acoustic blood at both nares. Visual acuity in the left eye was 20/50 impedances of bone and soft tissue in the orbit make the as compared to 20/20 in the right eye. Ultrasound was used orbital margins highly visible [4]. Described first in 1981, at the patient’s bedside to evaluate for possible traumatic Ord et al [5] applied B-mode scanning ultrasound in the retinal detachment. evaluation of orbital wall defects. Since then, the use of Ultrasound examination revealed no evidence of retinal ultrasound as a reliable method in the diagnosis of orbital detachment, but posterior to the globe, distinct evidence of fractures is well described [3,6-9]. The findings of Forrest gas was seen. Although not clinically suspected, subsequent et al [6] suggest that high resolution ultrasound can CT showed multiple facial fractures. Facial trauma service diagnose orbital wall fractures and orbital emphysema with was consulted. He was prescribed prophylactic antibiotics 94% correlation with axial and coronal CT images. In and discharged to home with clinic follow-up on the comparisons of ultrasound in the detection of orbital following day. The patient went on to have several fractures fractures vs CT as the standard, recent studies report surgically corrected approximately 2 weeks later but did not sensitivity from 85% to 92% and specificity and positive return for further follow-up. predictive value of 100% [4,6]. Newer studies describe the use of curvilinear transducers, formally used for intraoral applications, for improved fitting and scanning of the 5. Discussion anatomic structures of the orbital rims [10,11]. In the cases above, routine bedside ultrasound proved a Ocular trauma is frequently evaluated in the emergency useful adjunct in diagnosing facial fractures when clinical setting and is often the result of isolated facial trauma or suspicion was low. In case 1, the presence of severe blunt trauma such as from motor vehicle accidents or falls. eyelid swelling made it difficult to satisfactorily examine Although early recognition and treatment of orbital the patient. The detection of orbital emphysema on injuries is important, they are often difficult to evaluate bedside ultrasound led to further CT evaluation. The air and occult fractures may still be missed despite a thorough was not immediately under the skin but rather toward the physical examination. Clinical findings of orbital fractures inner portion of the swollen lid. This caused us to miss such as decreased eye motility, diplopia, and palpable the presence of air on mild palpation of the eyelid. The crepitus due to subcutaneous emphysema may not always patient in case 2 had been kicked in the eye and had a be present or detectable. Furthermore, severe swelling of relatively benign eye examination. Until the presence of the eye or facial tissues may further complicate initial retroorbital air was detected on ultrasound examination, examination [2]. CT examination had not been planned for. Likewise, the The role of CT scanning in the diagnosis and manage- patient in case 3 had a fairly normal eye exam and the ment of orbital injuries is well documented. Thin-section decision to obtain a CT was made only after the finding coronal CT is typically regarded as the bgold standardQ of retroorbital air on ultrasound. In the above cases, CT imaging method of midface fractures and orbital trauma. was not initially planned because of low clinical suspicion But even CT is not perfect and has resulted in false based upon physical examination and mechanisms of negatives and false positives in studies investigating orbital injury. Routine bedside ultrasound evaluation of the eye, trauma using perioperative findings as a reference method however, demonstrated the presence of periorbital air that [3]. Coronal CT evaluation also requires adequate patient prompted further evaluation by CT imaging that revealed positioning that may not be possible given coexisting orbital wall fractures. injuries such as an unstable cervical fracture [4]. Further- In the future, ultrasound may become more widely used more, in the less traumatic scenario, orbital injury may not in the screening, diagnosis, and treatment of various ocular be suspected and therefore CT scan may not be ordered at diseases and injuries in the emergency department. It all. The reasonable clinician may opt to forgo the extra already holds some advantages to CT in that it is less expense and radiation dose that a CT may present to the expensive and may be conducted at the patient’s bedside. patient if the physical examination or mechanism of injury Ultrasound is a widely available diagnostic modality fails to raise adequate suspicion for orbital injury. without the associated radiation load of CT. Its use may Ultrasonography has been used by ophthalmologists for be justified as an adjunct to physical examination when decades in the diagnosis, evaluation, and treatment of clinical suspicion of orbital wall fracture is low and CT may various ocular diseases and injuries. However, in recent not otherwise be practical or clinically warranted. 520 S.T. McIlrath et al.
References [6] Forrest CR, Lata AC, Marcuzzi DW, et al. The role of ultrasound in the diagnosis or orbital fractures. Plast Reconstr Surg 1993;92:28-34. [1] Blaivas M, Theodoro D, Sierzenski P. A study of bedside ocular [7] Hirai T, Manders EK, Nagamoto K, et al. Ultrasonic observation of ultrasonography in the emergency department. Acad Emerg Med facial bone fractures: report of cases. J Oral Maxillofac Surg 2002;9:791-9. 1996;54:776-9. [2] Rosen P, editor. Emergency medicine concepts and clinical practice, [8] Jenkins CM, Thuau H. Ultrasound imaging assessment of fractures of 5th ed. St Louis (Mo)7 Mosby; 2002. p. 908-10. the orbital floor. Clin Radiol 1997;52:708-11. [3] Jank S, Emshoff R, Etzelsdorfer M, et al. Ultrasound versus computed [9] McCann PJ, Brocklebank LM, Ayoub AF. Assessment of zygomatico- tomography in the of orbital floor fractures. J Oral Maxillofac Surg orbital complex fractures using ultrasonography. Br J Oral Maxillofac 2004;62:150-4. Surg 2000;38:525-9. [4] Lata AC, Marcuzzi DW, Forrest CR. Comparison of real-time [10] Jank S, Emshoff R, Strohl H, et al. Effectiveness of ultrasonography in ultrasonography and coronal computed tomography in the diagnosis determining medial and lateral orbital wall fractures with a curved- of orbital fractures. Can Assoc Radiol J 1993;44:371-6. array scanner. J Oral Maxillofac Surg 2004;62:451-5. [5] Ord RA, May ML, Duncan JG, et al. Computerized tomography and [11] Jank S, Emshoff R, Etzelsdorfer M, et al. The diagnostic value of B-scan ultrasonography in the diagnosis of fractures of the medial ultrasonography in the detection of orbital floor fractures with a orbital wall. Plast Reconstr Surg 1981;67:281-8. curved array transducer. Int J Oral Maxillofac Surg 2004;33:13-8. American Journal of Emergency Medicine (2005) 23, 521–524
www.elsevier.com/locate/ajem
Purple urine bag syndrome
Federico Vallejo-Manzur MDa, Eduardo Mireles-Cabodevila MDb, Joseph Varon MDc,* aUniversity of Massachusetts, Worcester, MA 01605, USA bMaryland General Hospital, Baltimore, MD 21201, USA cThe University of Texas Health Science Center, St. Lukes Episcopal Hospital, Houston, TX 77030, USA
Received 15 October 2004; accepted 15 October 2004
Abstract The purple urine bag syndrome (PUBS) is a rare condition associated with chronic urinary catheterization. It is characterized by the purple discoloration of the urine, collecting bag, and tubing. A number of factors are involved, but not always present, in its development including female sex, urinary tract infection, constipation, indicanuria, and alkaline urine. Despite multiple theories that involve the complex tryptophan metabolism to the tubing dye, the cause remains elusive. The syndrome resolves usually after treatment of urinary tract infection or changing of the collecting bag. We present a case of a patient with purple urine bag syndrome and a pertinent literature review. D 2005 Elsevier Inc. All rights reserved.
1. Introduction 2. Case report
The purple urine bag syndrome (PUBS) is a term that We report on a 72-year-old gentleman who is bedrid- has been used to describe the purple discoloration of the den, chronically debilitated, and a nursing home resident collecting bag and tubing. The PUBS occurs predominantly with a past medical history of chronic renal failure, in elderly women who are bedridden, chronically cathe- cerebral vascular accident, hypertension, advanced Parkin- terized, and constipated and has also been described in son’s disease, and the requirement for a chronic indwell- patients with ileal diversions [1-3]. This condition appears ing urinary catheter. The patient was brought from a hours or days after catheterization and has been related to nursing home to the emergency department because he indicanuria and urinary tract infection caused by indican was found to be lethargic and had recent purple discolora- (indoxyl sulfate) degrading bacteria [1-8]. tion of his urine. A review of his medications failed to Many hypotheses as to the etiology of this phenomenon disclose any pharmacological agent that could cause urine have been formulated; most of them have not been proven discoloration but included an alkalinizing agent, sodium to apply to all the patients reported in the literature [9-12]. citrate and citric acid (Bicitra), to manage his chronic We report a case of PUBS with absence of some of the renal tubular acidosis. Upon arrival to the emergency typical features previously described. department, purple discoloration of the urine bag and tubing was found (Fig. 1). On physical examination, the patient was found to be T Corresponding author. Tel.: +1 713 669 1670; fax: +1 713 839 1467. hypotensive with a blood pressure of 70/50 mm Hg, E-mail address: [email protected] (J. Varon). afebrile, and lethargic and the urinary bladder was found
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.10.006 522 F. Vallejo-Manzur et al.
(indican) and is then excreted to the urine, which, if exposed to the air, becomes oxidized into indigo blue. Months later, Sammons et al [2] and Payne and Grant [3] recognized the basis for the long-standing theory that has prevailed. They described that this phenomenon is infre- quent but not entirely unobserved and found this phenom- enon in 5 elderly female patients with indwelling catheters of whom one was incontinent and the rest were constipated and all had indigo blue in the urine. Indicanuria was again suggested and the presence of blue discoloration happened with the addition of an oxidizing agent such as sodium hypochlorite. These authors also considered that alkaline urine favors the deposition of indigo; chronic urinary tract infection with Pseudomonas aeuruginosa and/or Proteus spp was present in all the patients [2,3]. The PUBS was associated with the blue diaper syndrome Fig. 1 Purple discoloration of the urinary catheter plastic tubing described by Drummond et al [13] in 1964 and has been and bag. frequently cited as an explanation of the metabolism and to be distended above the umbilicus despite the presence of intestinal transport of tryptophan as the cause of indicanuria a urinary catheter. that was correlated with blue discoloration of the diaper His urinary analysis revealed a pH of 9.0, too-numerous- [2,5-9]. to-count white blood cells per high power field, positive Hildreth and Cass [5] described 5 cases of asymptomatic leukocyte esterase, negative nitrites, red blood cells 20 to discoloration of blue collecting bags in children with ileal 25 per high power field, and marked bacteria. His urine diversion, all related with urinary tract infections (only culture grew more of 100000 colonies of Escherichia coli. mentioned E coli in one patient), of whom only 2 were Urine indican and serum tryptophan were tested and there tested and had indicanuria. Hinberg et al [6] described the was no indicanuria (not detected in qualitative method), and process by which the purple discoloration of the urine was the serum tryptophan level was 23 lmol/L (20-95 lmol/L). secondary to highly alkaline urine in contact with the plastic Ultrasound of the urinary bladder and kidneys revealed in the bags, after which the dye used to eliminate the yellow findings suggesting chronic bladder outlet obstruction, with tint of the bag discolors the urine. In a similar observation, presence of echogenic sediment in a distended urinary Pankau [7] placed indoxyl sulfate (indican) in a plastic bladder and hydronephrosis. After changing the bladder tubing for 72 hours, after which it became oxidized and catheter, frank pus drained from the bladder. The second discolored the tubing to purple and the solution became collecting bag also turned purple. colorless. Years later, Wagner and Joyner [8] described the A diagnosis of severe sepsis secondary to pneumonia and process by which tryptophan is converted to indole by urinary tract infection was made; the patient was treated mixing enteric indole-producing bacteria with an enriched with piperacillin/tazobactam for the first 3 days and with tryptophan media in the presence of liver enzyme extracts levofloxacin for 14 days. After receiving antibiotic treat- obtaining indigo blue pigment. ment, his urine bag did not become purple again. The patient improved clinically, returning to his baseline mental status, and was transferred back to the nursing home after 8 days of hospitalization (Fig. 2).
3. Discussion
The purplish or bluish discoloration of urine collecting bags was first described by Barlow and Dickson [1] in 1978 when they reported the classic discoloration of the disposable bags of children with spina bifida and urinary diversions. These authors assumed that the discoloration was secondary to the presence of indigo blue as identified by spectrophotometric analysis [1]. Their article described the presence of indigo to the metabolism of tryptophan in the setting of intestinal obstruction and constipation that undergoes bacterial decomposition to indoxyl sulfate Fig. 2 Foley catheter and tubing after removal. Purple urine bag syndrome 523
increase even more if an infectious process is added [26,27]. Table 1 Bacteria isolated in urine of some patients with PUBS In our patient, the renal loss of bicarbonate in addition to the infectious process created highly alkaline urine. Alcaligenes spp Klebsiella ozaenae Since its description, the cause of PUBS has not been Bacteroides spp Klebsiella pneumoniae consistently proven, it remains to be elusive. Ultimately, the Citrobacter diversus Morganella morganii Citrobacter freundi Proteus mirabilisi PUBS appears to be a benign condition. It has not been Citrobacter koseri Proteus vulgaris demonstrated to have any implication other than the Enterobacter aerogenes Providencia rettgeri possibility of a urinary tract infection and has not been Enterobacter agglomerans Providencia stuartii proven to change the prognosis of patients. Clinicians Enterobacter cloacae Pseudomonas aeuruginosa dealing with patients who have indwelling bladder catheters Enterococcus avium Pseudomonas spp need to be aware of this condition. Enterococcus faecalis Serratia marcescens Escherichia coli Staphylococcus epidermidis Klebsiella oxytoca Streptococcus agalactiae Acknowledgment
More recently, Stott et al [9] described 3 elderly female The authors wish to acknowledge Mabrook Shehata, MD patients with PUBS, all of whom had alkaline urine, for his help in data gathering. concluding that the bacterial pathogens had no relation with the PUBS, that indicanuria is not a prerequisite, and that References the blue diaper syndrome has possibly a different etiology than PUBS. Moreover, these investigators proposed that [1] Barlow GB, Dickson JAS. Purple urine bags. Lancet 1978;28:220-1. the structure of the dye suggests a steroidal or bile acid [2] Sammons HG, Skinner C, Fields J. Purple urine bags. Lancet conjugate. 1978;1:502. [3] Payne B, Grant A. Purple urine bags. Lancet 1978;1:502. The PUBS has been associated with Providencia stuartii, [4] Fain-Ghironu N, Le Gondiec P, Schaeffer M. Purple urine bag Klebsiella pneumoniae,andEnterobacter agglomerans, syndrome. Presse Med 2003;32:985-7. bacteria with the ability to produce indigo [10,11]. [5] Hildreth TA, Cass AS. Blue collection bag after ileal diversion. Nobukuni et al [12] described that indicanuria is not a Urology 1978;11:168-9. requirement for PUBS. Other investigators have shown that [6] Hinberg I, Katz L, Weber F, Dodge P. Purple urine? Nursing 1982;1:124. the activity of bacterial indoxyl sulfatase, which catalyzes [7] Pankau EF. Purple urine bags. J Urol 1983;130:372. the conversion of indican into indigo blue, was present only [8] Wagner GE, Joyner PA. Investigating the indigo drainage bag. Am J in strong alkaline urine [13-15]. Nurs 1984;84:180. Nakayama and Kanmatsuse [16] reported lower serum [9] Stott A, Khan M, Roberts C, Galpin IJ. Purple urine bag syndrome. levels of tryptophan and valine in patients with PUBS, Ann Clin Biochem 1987;24:185-8. [10] Dealler SF, Belfield PW, Bedford M, Whitley AJ, Mulley GP. Purple suggesting abnormal absorption of amino acids secondary to urine bags. J Urol 1989;142:769-70. colonic dismotility and intestinal bacterial overgrowth. [11] Dealler SF, Hawkey PM, Millar M. Enzymatic degradation of urinary As noted before, multiple bacteria have been isolated in indoxyl sulfate by Providencia stuartii and Klebsiella pneumoniae the urine in several case reports of PUBS, some with known causes the purple urine bag syndrome. J Clin Microbiol 1988; sulfatase and phosphatase activities but the rest without it. 26:2152-6. [12] Nobukuni K, Kawahara S, Nagare H, Fujita Y. Study on purple The presence of alkaline urine is a common characteristic in pigmentation in five cases with purple urine bag syndrome. most cases [3,5,8,9,17-24] (Table 1). Kansenshogaku Zasshi 1995;69:1269-71. It is interesting to note that PUBS usually occurs in [13] Drummond K, Alfred AF, Ulstrom RA, Good RA. The blue diaper patients in geriatric wards and that in most of these patients, syndrome: familial hypercalcemia with nephrocalcinosis and indica- common occurrences include alkaline urine and female nuria. Am J Med 1964;37:928-48. [14] Umeki S. Purple urine bag syndrome (PUBS) associated with strong predominance. No specific cause has been found. The alkaline urine. Kansenshogaku Zasshi 1993;67:1172-7. PUBS has not been observed in short-term care facilities [15] Arnold WA. King George III’s urine and indigo blue. Lancet and has been associated with bhospital infectionQ transmitted 1996;347:1811-3. by staff. There is no causative relationship between PUBS [16] Nakayama T, Kanmatsuse K. Serum levels of amino acid in patients and bacteria strains, but bacterial counts are significantly with purple urine bag syndrome. Nippon Jinzo Gakkai Shi 1997;39:470-3. higher in PUBS patients. Indican is present in the urine of [17] Al-Jubouri MA, Vardhan MS. A case of purple urine bag syndrome healthy people, and the measurement can be hindered by associated with Providencia rettgeri. J Clin Pathol 2001;54:412. sampling errors [24,25]. [18] Ishida T, Ogura S, Kawakami Y. Five cases of purple urine bag Of interest is the presence of this syndrome in patients syndrome in a geriatric ward. Nippon Ronen Igakkai Zasshi with ileal diversion and urostomies having, in different 1999;36:826-9. [19] Matsuo H, Ishibashi T, Araki C, Sakamaki H, Mazume H, Ueki Y, degrees, the presence of hyperchloremic metabolic acidosis et al. Report of three cases of purple urine bag syndrome which secondary to loss of bicarbonate through the ileal or sigmoid occurred with a combination of both E coli and M morganii. pouch. This evidently increases the urinary pH, which will Kansenshogaku Zasshi 1993;67:487-90. 524 F. Vallejo-Manzur et al.
[20] Coquard A, Martin E, Jego A, Capet C, Chassagne PH, Doucet J, et al. [24] Mantani N, Ochiai H, Imanishi N, Kogure T, Terasawa K, Tamura J. Purple urine bags: a geriatric presentation of lower urinary tract A case-control study of purple urine bag syndrome in geriatric wards. infection. J Am Geriatr Soc 1999;47:1481-2. J Infect Chemother 2003;9:53-7. [21] Lin H-H, Li S-J, Su K-B, Wu L-S. Purple urine bag syndrome: a [25] Kirkland JL, Vargas E, Lye M. Indican excretion in the elderly. case report and review of the literature. J Intern Med Taiwan 2002; Postgrad Med J 1983;59:717-9. 13:209-12. [26] Lueng AKH, Tan IKS. Critical care and the urologic patient. Crit Care [22] Ihama Y, Hokama A. Purple urine bag syndrome. Urology 2002; Clin 2003;19:1-10. 60:910. [27] Ollapallil J, Irukulla S, Gunawardena I. Purple urine bag syndrome. [23] Fain-Ghironi N, Le Giondaec P, Schaeffer M. Purple urine bag ANZ J Surg 2002;72:309-10. syndrome. Presse Med 2003;32:985-7. American Journal of Emergency Medicine (2005) 23, 525–530
www.elsevier.com/locate/ajem
Chest radiograph screening for severe acute respiratory syndrome in the ED
Stewart Siu-Wa Chan MBBS(Syd), FRCSEd, FHKAMa,* Paulina Siu-Kuen Mak BNa, Kwok Kuen Shing MBChB, FRCRb, Po Nin Chan MBChB, FRCRb, Wing Hung Ng MBChB, FRCRb, Timothy Hudson Rainer MD, BSc, MBBCh, MRCP, FHKAMa aAccident and Emergency Medicine Academic Unit, Prince of Wales Hospital, The Chinese, University of Hong Kong, Shatin, Hong Kong bDepartment of Diagnostic Radiology and Organ Imaging, Prince of Wales Hospital, The Chinese, University of Hong Kong, Shatin, Hong Kong
Received 21 July 2004; accepted 18 October 2004
Abstract The purpose of the study was to evaluate the use of chest radiography for the screening of severe acute respiratory syndrome (SARS). We retrospectively analyzed all patients who attended an Emergency Department SARS screening clinic during the outbreak in Hong Kong, from March 10 to June 5, 2003. Patients with clinical and epidemiologic suspicion of SARS were evaluated by serial chest radiography. All radiographs were reported by consensus from 2 radiologists, blinded to the clinical records. The prevalence of SARS was 13.3% among 1328 patients included. The initial radiograph had sensitivity 50.3%, specificity 95.0%, positive likelihood ratio 10.06, negative likelihood ratio 0.52, positive predictive value 61.5%, and negative predictive value 92.3% for diagnosing SARS. Serial chest radiography had sensitivity 94.4%, specificity 93.9%, positive likelihood ratio 15.48, negative likelihood ratio 0.06, positive predictive value 71.4%, and negative predictive value 99.0%. The initial chest radiograph has poor sensitivity, and serial radiographs are required to rule out SARS. D 2005 Elsevier Inc. All rights reserved.
1. Introduction from Beijing and 2 were from Anhui Province, located in east-central China. In January 2004, 4 confirmed cases were The resurgence of severe acute respiratory syndrome reported in Guangdong Province, southern China. The likely (SARS) has been a topic of much interest. During April 22 reemergence of this disease in Hong Kong has pressurized to 29, 2004, the Chinese Ministry of Health reported a total frontline health care workers, especially those working in of 9 new cases of SARS in China; 7 of the patients were EDs. The ED has to assume the challenging role of administering a rapid and accurate screen for SARS, at the same time of carrying out appropriate containment and T Corresponding author. Tel.: +852 2632 2219; fax: +852 2337 3226. infection control measures. Because of the high infectivity E-mail address: [email protected] (S.S.-W. Chan). and high mortality nature of SARS, prompt diagnosis is
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.10.014 526 S.S.-W. Chan et al. essential. However, early clinical features of SARS are often 2. Methods nonspecific, and investigations have shown that screening at the ED is aided by laboratory and radiographic investiga- 2.1. Study design tions [1-6]. Various symptom scores and associated predic- This was a retrospective cohort study of the use of ED tion rules, applicable to the ED setting, have recently been plain chest radiography in the diagnosis of SARS. All derived and validated, and reported in emergency medicine patient information remained confidential. The chief exec- literature [2,4,5]. This study mainly focuses on the use of utives and associated institutional review boards of the chest radiography as a screening tool. Hospital Authority of Hong Kong and of the Prince of It is recognized that chest radiography plays an important Wales Hospital approved the collection of clinical data for role in the diagnosis of SARS [1,3,4,7-22]. At fever onset, surveillance, analysis, research, and reporting. almost 80% of patients with SARS had abnormal chest radiographs [9]. However, the sensitivity and specificity of 2.2. Study setting and population chest radiography before fever onset and at early stages of the disease is not known. Chest radiographic abnormality The study was performed on all patients who attended may precede lower respiratory tract symptoms in some the SARS screening clinic based in the ED of a 1400-bed patients with SARS [3,22]. For suspicious cases of SARS university teaching hospital, during the outbreak in Hong with initially normal chest films, repeating chest radiogra- Kong from March 10 to June 5, 2003. Patients during the phy every 1 to 3 days is likely to increase the yield [20]. early part of this period were triaged for assessment at the However, as yet, this subject has also not seen investigated screening clinic (a designated section of the ED) if they had in detail. respiratory tract symptoms, influenza-like illness, or fever, During the outbreak of SARS in Hong Kong that started together with a contact history. During the later stages, when in March 2003, the ED of Prince of Wales Hospital operated the outbreak had become more widespread in Hong Kong, a screening center to evaluate patients with epidemiologic patients with the above-mentioned symptoms, even without and clinical suspicion of SARS. The authors retrospectively a definite contact history, were triaged for assessment at the analyzed all cases that were screened at the ED and clinic. Patients were followed up at the clinic every 1 to 3 followed up at the SARS review clinic. These cases days until either hospitalization was required or until follow- included patients who were severely ill at presentation, up was no longer necessary because of satisfactory as well as those who presented to the ED at an early stage resolution of symptoms. Serial frontal chest radiography of the illness. The prevalence of SARS in this setting was performed at follow-up visits. was 13.3%. The objectives of this study were: 2.3. Data collection Patients’ medical records from the ED were retrieved 1. To calculate the sensitivity, specificity, predictive and reviewed by a research nurse who extrapolated clinical values, and likelihood ratios of chest radiography in and demographic details into a database. The record of predicting SARS, as used in an ED-based screening patients’ body temperature during attendance was included clinic during a community outbreak; in the collation. 2. To study the duration in days (i) since onset of symptoms and (ii) since initial presentation for plain 2.4. Interpretation of radiographs chest radiography to develop abnormal changes consistent with SARS; Every ED chest film was retrospectively read by 2 of 3 3. To report the yield and performance of chest radiog- radiologists (W-HN, P-NC, and K-KS) who were blinded to raphy from screening patients who presented (i) all clinical records. The radiologists had knowledge of the without fever (temperature not exceeding 37.58C) dates when the radiographs were taken and were allowed to and (ii) with temperature not exceeding 388C. evaluate serial chest films of each patient chronologically.
Table 1 The performance of the initial chest radiograph in diagnosing SARS SARS Non-SARS Total number of patients Test (chest radiography) True positive (A) = 88 False positive (B) = 55 With positive radiograph (A + B) = 143 positive Test (chest radiography) False negative (C) = 87 True negative (D) = 1043 With negative radiograph (C + D) = 1130 negative Total number of patients With SARS (A + C) = 175 Non-SARS (B + D) = 1098 (A + B + C + D) = 1273 Sensitivity = A/A + C = 50.3%. Specificity = D/B + D = 95.0%. Chest radiography for SARS screening 527
Table 2 The performance of serial chest radiography in diagnosing SARS SARS Non-SARS Total number of patients Test (chest radiography) True positive (A) = 167 False positive (B) = 67 With positive radiograph (A + B) = 234 positive Test (chest radiography) False negative (C) = 10 True negative (D) = 1034 With negative radiograph (C + D) = 1044 negative Total number of patients With SARS (A + C) = 177 Non-SARS (B + D) = 1101 (A + B + C + D) = 1278 Sensitivity = A/A + C = 94.4%. Specificity = D/B + D = 93.9%.
The presence of airspace opacification was regarded as test The reports on the initial chest radiograph were positive for SARS. Using a structured pro forma, the 2 indeterminate in 55 patients (4.1%). After excluding these radiologists reported in consensus whether each radiograph from the analysis, the initial chest radiograph had a was normal, suspicious, or positive. After the review of sensitivity of 50.3% (95% confidence interval [CI], 42.6- serial chest radiographs of each patient, the overall 57.9), specificity of 95.0% (95% CI, 93.5-96.2), positive impression was also reported. likelihood ratio of 10.06, negative likelihood ratio of 0.52, positive predictive value of 61.5%, negative predictive value 2.5. Outcome measure of 92.3%, and an accuracy of 88.8% for predicting SARS The final diagnosis of SARS was based on the official (Table 1). list of patients recorded in the Hong Kong Department of The reports on serial chest radiography were indetermi- Health’s SARS registry. This list had been prepared by nate in 50 patients (3.7%). After excluding these from the public health experts in Hong Kong, according to the analysis, serial chest radiography had a sensitivity of 94.4% recommendations of the World Health Organization on (95% CI, 89.9-97.3), specificity of 93.9% (95% CI, 92.3- interpretation of laboratory results for the diagnosis of 95.3), positive likelihood ratio of 15.48, negative likelihood SARS. The SARS-associated coronavirus (SARS-CoV) ratio of 0.060, positive predictive value of 71.4%, negative antibody status of these patients (immunoglobulin G levels predictive value of 99.0% , and an accuracy of 93.9% for measured by immunofluorescence assay) was also docu- predicting SARS (Table 2). mented for reference. 2.6. Data interpretation and analysis
Descriptive statistics were generated. The performance of Table 3 The performance of serial chest radiography in chest radiography was assessed by analyzing (i) the chest identifying SARS in patients who presented with a temperature radiograph at the initial presentation and (ii) the overall (i) V37.58C and (ii) V388C conclusion after review of serial chest radiographs, in Presenting Presenting correlation with the final diagnosis. Reports that were temperature temperature suspicious were regarded as indeterminate and excluded V37.58C V388C from the analyses. Data were analyzed using Statview for Disease 9.1% 10.8% Windows version 5.0 Statistical Analysis Software (Abacus prevalence Concepts, SAS Institute, Cary, NC) and MedCalc version 7.0 Sensitivity (%) 90.7% (95% CI, 92.8% (95% CI, (MedCalc, Belgium). 81.7%-96.1%) 86.3%-96.8%) Specificity (%) 96.0% (95% CI, 95.6% (95% CI, 94.4%-97.3%) 94.1%-96.9%) 3. Results Positive 22.79 21.27 likelihood There were a total of 1378 consecutive patients during ratio the study period. Fifty patients (3.6%) were excluded Negative 0.010 0.08 because their chest radiographs were unavailable for likelihood radiologists’ review. Of 1328 cases included for analysis, ratio 38.7% were males and 61.3% females. The mean (SD) age Positive 69.4 72.0 predictive was 38.4 (17.0) years. There were 177 cases with final value (%) diagnosis of SARS, and the prevalence in this setting was Negative 99.0 99.1 thus calculated to be 13.3%. The SARS-CoV antibody status predictive was positive in 171 patients (96.6%) of those diagnosed value (%) with SARS in our study. 528 S.S.-W. Chan et al.
For patients who have SARS with positive radiograph the patients’ clinical status. As far as we are aware, our (n = 167), the mean (SD) duration from onset of symptoms to study was the first to investigate the performance of chest positive radiograph was 5.8 (5.4) days. Seventy-nine radiography as an independent diagnostic test. (47.3%) of these patients had negative or indeterminate Our study found that the sensitivity (50.3%) of the initial chest radiographs at initial presentation, and the mean (SD) chest radiograph was much lower than most authorities duration from initial presentation to positive chest radio- would expect. This result has important practical implica- graph was 5.4 (6.3) days. tions. Clinicians ought to be careful not to be falsely Of 829 patients who initially presented with a temper- reassured by an initially negative chest film. The false- ature of 37.58C or lower, serial chest radiographs identified negative rate is high in this setting. Previous anecdotal 98 (11.8%) positive cases. Of these, 68 (69.4%) were finally reports suggested that the chest radiograph could be normal diagnosed with SARS. Of 1028 patients who initially at first presentation [25]. Our results confirmed this presented with a temperature of 388C or lower, serial possibility, and we report the incidence of its occurrence radiographs identified 143 (13.9%) positive cases. Of these, at our ED screening clinic (Table 1). In contrast, the 103 (72.0%) were finally diagnosed with SARS. The sensitivity of serial chest radiography in our study was performance of serial chest radiography in identifying 94.4%, which is a marked improvement from the former. SARS in these 2 categories of cases is shown in Table 3. Only by close follow-ups with serial chest radiography and clinical reassessment can the condition be more confidently excluded. Hence, our study lends evidence to the statement 4. Discussion that chest radiography is highly sensitive for SARS, but only in the context of interpretation of serial radiographs, The resurgence of SARS and its propensity to spread instead of the initial screening chest film. The negative rapidly across international borders continue to cause predictive value (the probability that the patient is disease- concern for public health authorities worldwide. Recent free if the test is negative) of serial chest radiography was cases linked to exposures at research laboratories have 99.0%. The negative likelihood ratio of 0.06 also shows that prompted concerns particularly to countries with laborato- it is a strong negative predictor. Serial chest radiography is ries working with live SARS-CoV. The US Centers for thus extremely useful in ruling out SARS in the setting of a Disease Control and Prevention continues to update its community outbreak. guidelines and recommendations for the clinical evaluation It is also important to be aware of the specificity of chest of possible SARS-CoV disease [23]. Much has already been radiography for predicting SARS in this setting. The initial published about the etiology, diagnosis, clinical features, radiographic appearance of SARS is not readily distinguish- and treatment of this illness. Several authors investigated the able from pneumonia caused by other etiologic agents [8]. early predictors of SARS in the ED setting [1,3,22], whereas Earlier reports suggest that the changes associated with others have reported the derivation and validation of SARS more commonly involve the lower zone and peri- symptom scores or prediction rules [2,4,5].Thechest pheral lung fields, and that features of cavitation, lymph- radiographic features of SARS have also been extensively adenopathy, and pleural effusion are rare [8,12,16-19]. studied, and the most consistent initial sign is reported to be However, the initial appearance and distribution of the airspace opacities or shadowing [7-22]. Radiographic airspace disease is also reported to be highly variable, and evidence is also an essential element of the case definition may be normal, unifocal, multifocal, bilateral, or even [24]. Most authorities would regard the chest radiograph as extensive [8,12-14,18,19]. Moreover, the cause of airspace a very sensitive screening test for diagnosing SARS. opacification is not limited to infection. It may also be However, unfortunately, most studies were based on patients caused by the presence of fluid, such as in pulmonary who were already hospitalized, with a high probability of edema. Because patients with clinical and radiographic the diagnosis [7,9,12-16,18-20]. Very few looked at the use evidence of SARS require hospitalization with full isolation of chest radiography as an independent screening test in the measures, a low specificity (high false-positive rate) at the ED setting, and few were able to report its performance screening level would likely result in rapid overwhelming of objectively. Only by studies directed at the screening level the hospital system. Few of the earlier studies were able to can full details about btrue negativesQ and bfalse negativesQ report on the specificity of chest radiography because few of the test be captured. For those investigators who did were conducted at the screening level, with a database study at the screening setting, their studies were limited by inclusive of the btrue-negativeQ cases. The specificity is the fact that chest radiographs were read by ED physicians dependent upon the prevalence of other causes of pneumo- who were not blinded to clinical details [1-4]. Therefore, nia during the study period. Nevertheless, our study despite the use of multivariate analyses to identify indepen- demonstrated that, despite the concerns mentioned above, dent associations, their results need to be interpreted with the specificity of chest radiography under our setting was the understanding that any merit identified for the accuracy excellent, being above 90% for both the initial radiograph of chest radiography might have been exaggerated because and serial chest radiography. The positive likelihood ratio radiographs were interpreted with a beneficial knowledge of for serial chest radiography was 15.48, showing that it is a Chest radiography for SARS screening 529 strong positive predictor, capable of generating conclusive useful as a screening tool in ruling out SARS. However, the changes in post-test probability. chest radiograph at initial presentation performs poorly as a Furthermore, our study revealed that there was consid- screen, and practitioners are cautioned against overreliance erable yield in screening afebrile patients, or patients with on it to rule out SARS. Conversely, both the initial low-grade fever, with serial chest radiography. The 2 radiograph and serial radiography are highly specific for temperature cutoffs of 37.58C and 388C were chosen for SARS in our setting and are therefore excellent in ruling in analyses because these values were most often taken as the disease. Radiographic changes of SARS may precede inclusion criteria for triage to bfever clinicsQ or bSARS fever, and the use of chest radiography for screening should screening clinicsQ in many institutions around the world not be limited to febrile patients. during the outbreak. Certain recently published prediction rules or novel scoring systems for SARS screening are only applicable to febrile patients with temperatures above 388C References [2,4,5,26]. Our findings show that relying only on these prediction rules would result in a substantial proportion of [1] Wong WN, Sek ACH, Lau RFL, et al. Early clinical predictors of missed cases (those with temperature not exceeding 388C) severe acute respiratory syndrome in the emergency department. Can that are identifiable by chest radiography. Our findings J Emerg Med 2003;6(1) (Fast-track article). Available: www.caep.ca/ provide further evidence to support the current Centers for 004.cjem-jcmu/004-00.cjem/vol-6.2004/v61-erO1.wong.htm#main Disease Control and Prevention recommendation that, when [accessed 2004 Jan 25]. [2] Chen S-Y, Su C-P, Ma MH-M, et al. Predictive model of diagnosing person-to-person transmission is occurring in the world, a probable cases of severe acute respiratory syndrome in febrile patients patient with contact history who develops respiratory with exposure risk. Ann Emerg Med 2004;43:1-5. symptoms, even if afebrile, should promptly undergo [3] Chen S-Y, Chiang W-C, Ma MH-M, et al. Sequential symptomatic screening inclusive of chest radiography [23]. analysis in probable severe acute respiratory syndrome cases. Ann Emerg Med 2004;43:27-33. [4] Wang T-L, Jang T-N, Huang C-H, et al. Establishing a clinical decision rule of severe acute respiratory syndrome at the emergency 5. Limitations department. Ann Emerg Med 2004;43:17-22. [5] Su C-P, Chiang W-C, Ma MH-M, et al. Validation of a novel severe Our study was limited in that it was retrospective in acute respiratory syndrome scoring system. Ann Emerg Med 2004; design, and 3.6% of patients were excluded from the study 43:34-42. [6] Rainer TH, Cameron PA, Smit DV. Should differential white cell because of the unavailability of radiographs. Nevertheless, counts be reported as percentages or absolute count in patients with the recall bias generally associated with retrospective studies severe acute respiratory syndrome? Hong Kong J Emerg Med 2004; did not affect data collection for the core parameters (the 11:12-5. radiologists’ report and the final diagnosis) of our study. The [7] Wong KT, Antonio GE, Hui DSC, et al. Severe acute respiratory patients’ body temperatures at presentation were also syndrome: radiographic appearances and pattern of progression in 138 patients. Radiology 2003;228:401-6. objectively charted. Our results may not be reproducible at [8] Antonio GE, Wong KT, Chu WCW, et al. Imaging in severe acute other settings for a number of reasons. First, the prevalence respiratory syndrome (SARS). Clin Radiol 2003;58:825-32. of SARS at other settings may be different. The prevalence [9] Hui DS, Sung JJ. Severe acute respiratory syndrome. Chest 2003; of other causes of pneumonia may also have geographic or 124(1):12-5. seasonal variation. A notable example is the increased [10] Goh JS, Tsou IY, Kaw GJ. Severe acute respiratory syndrome (SARS): imaging findings during the acute and recovery phases of disease. prevalence of influenza pneumonia during the winter period. J Thorac Imaging 2003;18(3):195-9. This would theoretically affect the specificity of chest [11] Nicolaou S. Radiologic manifestations of severe acute respiratory radiography, a concern already discussed above. Second, the syndrome. N Engl J Med 2003;348(20):2006. radiologists who interpreted our radiographs were all [12] Grinblat L, Shulman H, Glickman A, et al. Severe acute respiratory specialists, Fellows of the Royal College of Radiologists. syndrome: radiographic review of 40 probable cases in Toronto, Canada. Radiology 2003;228:802-9. By contrast, in practice, chest radiographs in most ED [13] Ma W, Chen GF, Li TS, et al. Analysis of chest X-ray manifestations settings are usually read by ED physicians. We do not in 118 patients with severe acute respiratory syndrome. Chinese Crit believe, however, that this would cause significant difficulty Care Med 2003;15(6):338-42 [in Chinese]. in generalizing our results. Finally, our result applies to an [14] Liu JX, Tang XP, Jiang SF, et al. The chest X-ray image features of ED screening setting during a community outbreak of patients with severe SARS: a preliminary study. Chin Med J 2003; 116(7):968-71. SARS, with a disease prevalence of 13.3%. It may not be [15] Du X, Zhu Q, Hong N, et al. The investigation of clinical images of extrapolated to other settings. 88 hospital staff with SARS. J Peking Univ (Health Sciences) 2003; 35(Suppl):34-7. [16] Ma J, Li N, Que C, et al. Dynamic observation of the features of chest radiograph in SARS patients. J Peking Univ (Health Sciences) 6. Conclusions 2003;35(Suppl):38-40. [17] Zhang XZ, Xue AH. The role of radiological imaging in diagnosis In conclusion, our study shows that, in the setting of a and treatment of severe acute respiratory syndrome. Chin Med J community outbreak, serial chest radiography is extremely 2003;116(6):831-3. 530 S.S.-W. Chan et al.
[18] Wang R, Sun H, Song L, et al. Plain radiograph and CT features of [23] Centers for Disease Control and Prevention. Clinical guidance on the 112 patients with SARS in acute stage. J Peking Univ (Health identification and evaluation of possible SARS-CoV disease among Sciences) 2003;35(Suppl):29-33. persons presenting with community-acquired illness. Available: [19] Zhao DW, Ma DQ, Wang W, et al. Early X-ray and CT appearances of http://www.cdc.gov/ncidod/sars/clinicalguidance.htm [accessed 2004 severe acute respiratory syndrome: an analysis of 28 cases. Chin Med May 31]. J 2003;116(6):823-6. [24] World Health Organization. Case definitions for surveillance of [20] Lu PX, Zhou BP, Chen XC, et al. Chest X-ray imaging of patients severe acute respiratory syndrome (SARS). Geneva7 The Organi- with SARS. Chin Med J 2003;116(7):972-5. zation; 2003. Available: www.who.int/csr/sars/casedefinition/en [21] Kaw GJL, Tan DYL, Leo YS, et al. Chest radiographic findings of a [accessed 2004 Jan 24]. case of severe acute respiratory syndrome (SARS) in Singapore. [25] Wu EB, Sung JJY. Haemorrhagic-fever–like changes and normal Singapore Med J 2003;44(4):201-4. chest radiograph in a doctor with SARS. Lancet 2003;361: [22] Rainer TH, Cameron PA, Smit D, et al. Evaluation of WHO criteria 1520-1. for identifying patients with severe acute respiratory syndrome out of [26] Chan SS. Screening for severe acute respiratory syndrome in the hospital: prospective observational study. BMJ 2003;326:1354-8. emergency department (Letter). Ann Emerg Med 2004;44:424-5. American Journal of Emergency Medicine (2005) 23, 531–535
www.elsevier.com/locate/ajem
Clinical predictors of bleeding esophageal varices in the ED
Siang-Hiong Goh MBBS*, Wee-Ping Tan MBBS, Shu-Woan Lee MBBS
Accident and Emergency Department, Changi General Hospital, Singapore 529889, Republic of Singapore
Received 23 September 2004; accepted 23 October 2004
Abstract Objectives: Some authors have found that thrombocytopenia (b118000/mm3), splenomegaly, and ascites are useful predictors of large esophageal varices in cirrhotic patients. We decide to see whether these factors could also be used to predict bleeding esophageal varices in patients known to have chronic liver disease in the ED. Methods: A case record review was done of all patients admitted to the ED of Changi General Hospital with upper gastrointestinal bleeding from esophageal varices from October 1999 to April 2004. The criteria of thrombocytopenia, splenomegaly, and ascites were applied retrospectively to these patients to see how accurately they performed in predicting bleeding esophageal varices. Results: Only 55% of patients had thrombocytopenia, whereas 45% had splenomegaly, and 27.5% had ascites. Combining thrombocytopenia with the presence of either ascites or splenomegaly did not improve the yield (only 40%), and only 6 patients had all 3 criteria. Twelve patients with bleeding varices did not have any of the criteria. Conclusions: Thrombocytopenia, splenomegaly, or ascites is an unreliable predictor of bleeding esophageal varices. Urgent or emergent endoscopy is still advocated to accurately diagnose bleeding esophageal varices. D 2005 Elsevier Inc. All rights reserved.
1. Introduction bleeding [1]. In view of the high hospital mortality rate of 15% to 40% [1], it had been advocated that emergency It is not uncommon to have patients with chronic liver physicians quickly place such patients on some kind of disease presenting to the emergency department (ED) with vasoactive treatment (somatostatin, octreotide, or terlipres- acute upper gastrointestinal (GI) bleeding. Often, the sin [2-5]) while in the ED, in tandem with rapid approach taken by the emergency physician is to assume resuscitation with fluids and blood products. This is to be that the bleeding originated from esophageal varices and to followed by urgent endoscopy, in which a number of manage the patient as such until proven otherwise. This is therapeutic options (sclerotherapy [1], variceal multiband because bleeding esophageal varices have a higher morbid- ligation [6-8], use of tissue-glue adhesives [6,8]) can be ity and mortality rate than any other source of upper GI used to stop further esophageal variceal bleeding. However, 20% to 50% of patients with symptoms and signs of chronic liver disease who develop upper GI T Corresponding author. Tel.: +65 8501687; fax: +65 260 3756. bleeding—despite findings such as spider angiomas, E-mail address: [email protected] (S.-H. Goh). asterixis, and ascites—do so from sources other than
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.10.005 532 S.-H. Goh et al. esophageal varices [1]. For instance, in the study of Josen 3. Results et al [9], only 27% of the patients in the liver disease group with varices (cirrhotics) had frank variceal hemor- Based on the search criteria outlined above, a total of rhage, whereas 57% bled from hemorrhagic gastritis. 44 patients had been admitted to the ED of Changi General Conversely, in approximately 50% of patients with known Hospital with upper GI bleeding from esophageal varices varices who present with upper GI bleeding, the bleeding between July 1999 and May 2004. The presence of bleeding is from a source other than the varices [10]. Our own varices was confirmed in all cases by endoscopy. Four experience in Changi General Hospital showed that for the patients were excluded from the study, as 1 had a years 2001 to 2003, only 27% of chronic liver disease splenectomy before for hypersplenism, 2 had transjugular patients who presented to the ED with upper GI bleeding intrahepatic portosystemic shunts done before, and another did so from esophageal varices. Therefore, starting had coexisting hepatocellular carcinoma. Of the remaining vasoactive treatment for all patients with chronic liver 40 patients studied, 19 had presented to our hospital for the disease presenting with hematemesis and/or melena (on the first time. Thirteen (32.5%) were female. The ages of the assumption that bleeding is from esophageal varices) may 40 patients ranged from 29 to 86 years, with a mean age of not be cost-effective. 55.8 years. Fifteen patients (37.5%) had alcoholism as the Recently, several authors have found that thrombocyto- etiology of the cirrhosis, 11 (27.5%) patients had hepatitis penia, splenomegaly, and ascites are useful predictors for B, 4 (10%) had hepatitis C, 1 had Wilson’s disease, 3 large esophageal varices in cirrhotic patients [11-18].We patients had dual pathology (alcoholism and hepatitis B or therefore decided to see whether these factors could also be C), with 6 cases being cryptogenic. used to predict bleeding esophageal varices in patients Of these 40 patients, the range of platelet counts at the known to have chronic liver disease and presenting with time of presentation to the ED ranged from 38000 to 3 3 hematemesis and/or melena to the ED. 300000/mm , with a mean of 129525/mm , and a median of 117500/mm3. Eighteen patients had platelet levels equal to or higher than 118000/mm3, both at the time of bleeding 2. Materials and methods in the ED, or in the preceding 3 months at the specialist clinic (although 9 patients had no previous clinic visits). An electronic search was made of all patients admitted to the ED of Changi General Hospital with an admission diagnosis of GI bleeding (ICD-9 code 578). This search was made from our hospital’s medical record database Table 1 Distribution of platelet levels below 113000/mm3 in and began from July 1999 to May 2004. These patients 23 patients were then further selected for the presence of bleeding S/N ED platelet levels Platelet levels (per mm3) esophageal varices (ICD-9CM code of 4650). Patients (per mm3) at the obtained from previous thus chosen for study must have the diagnosis of bleeding time of bleeding specialist clinic or ED visits esophageal varices confirmed on endoscopy. episode in the preceding 3 months Patients thus identified had their case records reviewed 1 38000 No preceding visit for their initial platelet levels taken at the time of pre- 2 42000 44000 sentation to the ED with upper GI bleeding. If the patient 3 46000 No preceding visit had previous attendances at the specialist clinic or ED 4 61000 No preceding visit within the 3 months preceding the bleeding episode, the 5 67000 No preceding visit 6 67000 65000 baseline platelet levels (if any) was also traced. We 7 74000 No preceding visit identified all those in the 2 groups with a platelet level 3 8 82000 90000 below 118000/mm . 9 90000 74000 We also checked for the presence of ascites and 10 90000 62000 splenomegaly for each patient. This can be from either 11 91000 No preceding visit clinical examination of the abdomen (as detected by the 12 91000 82000 attending physicians) or with the aid of imaging studies 13 93000 94000 (ultrasonography or computed tomography [CT] scan) 14 93000 No preceding visit performed during the same period of hospitalization. The 15 95000 89000 causes of their cirrhosis were also determined. The use of 16 108000 No preceding visit diuretics in patients was also noted, as this could affect the 17 109000 72000 18 110000 No preceding visit volume of ascites and its clinical detection. 19 113000 75000 Patients who had splenectomy, splenorenal shunts, or 20 116000 80000 transjugular intrahepatic portosystemic shunts done before 21 117000 130000 were excluded from the study, as were patients with 22 121000 56000 coexisting hepatocellular carcinoma and renal failure. Predicting variceal bleeding 533
Table 2 Numbers of patients with the criteria of Finally, only 6 patients with bleeding had all 3 criteria thrombocytopenia, splenomegaly, and ascites of thrombocytopenia, splenomegaly, and ascites, whereas 12 patients had none of the 3 criteria (Table 2). Patients with bleeding N=40 esophageal varices Number with 22 55.0% 4. Discussion thrombocytopenia present Number with 18 45.0% Several authors [11-18] have put forward papers suggest- splenomegaly present (11 with imaging) ing that thrombocytopenia, splenomegaly, and ascites are Number with ascites present 11 27.5% useful predictors for large esophageal varices in cirrhotic (5 with imaging) patients. This is intuitive as the triad is the result of portal Number with thrombocytopenia 10 40.0% and 1 other hypertension in cirrhotic patients. These criteria were criteria present (either developed for the purpose of improving yield and cost- ascites or splenomegaly) effectiveness when selecting patients for screening endos- Number with all 3 criteria 6 15.0% copy and prophylactic therapy to decrease the incidence of Number without any 12 30.0% bleeding. A brief summary of the salient points of these thrombocytopenia, studies is listed in Table 3, along with the references. splenomegaly or ascites Interestingly, the authors found that the Child-Pugh score was not a reliable predictive criterion for esophageal varices. We used 40 patients who had confirmed bleeding Twenty-two patients had platelet counts below 118000/ esophageal varices by endoscopy (taking endoscopy as the mm3, either at the time of bleeding in the ED or in the bgold standardQ) and applied the criteria of thrombocytope- preceding 3 months at the specialist clinic (although 10 of nia, splenomegaly, and ascites to this group to see how these patients had no previous clinic visits; see Table 1). As accurately they could predict bleeding varices. such, the criterion was positive in only slightly more than half Levels of platelet counts recommended for the screening of the patients (55%; see Table 2). of esophageal varices by various authors [11-15] varied Splenomegaly was detected on physical examination in from 150000/mm3 (proposed by Freeman et al [15])to 7 patients, and imaging studies (either ultrasonography or 68000/mm3 (proposed by Madhotra et al [12]). However, CT scan) showed the presence of splenomegaly in an Freeman et al [15] limited their study population to only additional 11 patients. In total then, 45% (18) of patients had alcoholic cirrhotic patients. For the purpose of our study, splenomegaly (Table 2). therefore, we applied the next highest platelet cutoff level of Ascites was detected on physical examination in 118000/mm3 (Table 3), which was proposed by Thomo- 6 patients, and imaging studies (either ultrasonography or poulos et al [11], whose study population included all forms CT scan) showed the presence of ascites in an additional of liver cirrhosis. We included platelet levels done both at 5 patients. In total then, 27.5% (11) of patients had the time of bleeding in the ED and any platelet levels done ascites. We would like to point out that 5 of 15 specialist within the preceding 3 months (usually by the relevant clinic follow-up patients who had no ascites were on discipline in the specialist outpatient clinics). Using this spironolactone treatment at the time of the bleeding criterion, only 55% of patients with bleeding esophageal episode (Table 2). varices fulfilled it.
Table 3 Summary of various papers citing clinical predictors for esophageal varices in cirrhotic patients Author Year Number of Factor studied by authors patients studied Thrombocytopenia (cutoff level) Splenomegaly Ascites Thomopoulos et al [11] 2003 184 Yes (118000/mm3)Yesa Yesa Madhotra et al [12] 2001 184 Yes (68000/mm3)Yesb Yesb Chalasani et al [13] 1999 346 Yes (88000/mm3)Yesb Zaman et al [14] 1999 98 Yes (88000/mm3)Yesb Freeman et al [15]—study limited 1999 65 Yes (150000/mm3) to alcoholic cirrhotic patients only Barcia et al [16] 1998 95 Yesc Alacorn et al [17] 1998 79 Yes (not specified) Yesb Yesb Lavergne et al [18] 1997 52 Yesa a Detected on ultrasonography. b Detected by either clinical examination or imaging (ultrasound or CT scan). c Detected on ultrasonography or CT scan. 534 S.-H. Goh et al.
The criterion of splenomegaly was advocated in papers 5. Conclusion by Thomopoulos et al [11], Madhotra et al [12], Chalasani et al [13], Zaman et al [14], and Alacorn et al [17]. The last Some authors have suggested that thrombocytopenia, 4 authors used both clinical examination and imaging splenomegaly, and ascites are useful predictors for large studies to detect the presence of splenomegaly, whereas esophageal varices in cirrhotic patients and can be used for Thomopoulos et al [11] used imaging studies. This is the purpose of improving yield and cost-effectiveness when understandable as clinical examination alone to detect a electively selecting patients for screening endoscopy and mild-to-moderately enlarged spleen is often unreliable and prophylactic therapy to decrease the incidence of first-time subject to many variables such as the experience of the variceal bleeding. clinician, position of the patient (lying on the right as However, these 3 factors fail to translate into reliable compared to lying supine), the presence of ascites, and the predictors of bleeding esophageal varices in chronic liver abdominal girth of a patient. In our series, only 7 patients disease patients presenting with hematemesis and/or melena had a palpably enlarged spleen on clinical examination at to the ED. Urgent or emergent endoscopy is still advocated the time of the variceal bleeding, whereas imaging done for such patients to accurately diagnose bleeding esophageal within the same period of hospitalization revealed an varices and institute appropriate treatment. additional 11 patients with splenomegaly. All in all, only 18 patients (40.0%) fulfilled the criteria of splenomegaly. In the same vein, the criterion of ascites was advocated References in papers by Thomopoulos et al [11], Madhotra et al [12], Barcia et al [16], Alacorn et al [17], and Lavergne et al [1] McQuaid KR. Alimentary tract. In: Tierney Jr LM, McPhee SJ, [18]. Madhotra et al [12] and Alacorn et al [17] used both Papadakis MA, editors. Current medical diagnosis and treatment, clinical examination and imaging studies to confirm the 43rd ed. New York (NY)7 Lange Medical Books/McGraw-Hill; 2004. presence of splenomegaly, whereas Thomopoulos et al [11], p. 515-622. [2] Gow PJ, Chapman RW. Modern management of oesophageal varices. Barcia et al [16], and Lavergne et al [18] used both imaging Postgrad Med J 2001;77:75-81. studies. The detection of free fluid in the abdomen by [3] de Franchis R. Somatostatin, somatostatin analogues and other clinical examination depends on relatively crude techniques vasoactive drugs in the treatment of bleeding oesophageal varices. (percussion for shifting dullness, and fluid thrill) [19].At Dig Liver Dis 2004;36(Suppl 1):S93-S100. least 500 to 1000 mL is required to produce abnormal [4] Goulis J, Burroughs AK. Role of vasoactive drugs in the treat- ment of bleeding oesophageal varices. Digestion 1999;60(Suppl 3): physical signs, whereas ultrasound can detect as little as 50 25-34. mL [19]. In our series, only 6 patients had ascites on [5] D’Amico G. The role of vasoactive drugs in the treatment of clinical examination at the time of the variceal bleeding, oesophageal varices. Expert Opin Pharmacother 2004;5:349-60. whereas imaging done within the same period of hospital- [6] Helmy A, Hayes PC. Current endoscopic therapeutic options in the ization revealed an additional 5 patients with ascites. All in management of variceal bleeding. Aliment Pharmacol Ther 2001; 15:575-94. all, only 11 patients (27.5%) fulfilled the criteria of ascites. [7] Nevens F, Rutgeerts P. Variceal band ligation in the management of The use of spironolactone would have reduced the number bleeding oesophageal varices: an overview. Dig Liver Dis 2001; of patients with ascites. (Some might argue against using 33:284-7. imaging to detect splenomegaly and ascites in patients with [8] Lata J, Hulek P, Vanasek T. Management of acute variceal bleeding. bleeding varices. However, an increasing number of Dig Dis 2003;21:6-15. [9] Josen AS, Giuliani E, Voorhees Jr AB, et al. Immediate endoscopic emergency physicians are using emergency ultrasonogra- diagnosis of upper gastrointestinal bleeding. Its accuracy and value in phy for trauma to detect fluid in the traumatized abdomen. relation to associated pathology. Arch Surg 1976;111:980-6. In the future, the experience thus gained can facilitate the [10] Erdman CW, Mensching JJ. Gastrointestinal bleeding. In: Stone CK, bedside detection of splenomegaly and ascites in patients in Humphries RL, editors. Current emergency diagnosis and treatment, 7 the ED.) 5th ed. New York (NY) Lange Medical Books/McGraw-Hill; 2004. p. 283-95. Combining thrombocytopenia with either the presence of [11] Thomopoulos KC, Labropoulou-Karatza C, Mimidis KP, et al. Non- ascites and splenomegaly did not improve the yield (only invasive predictors of the presence of large oesophageal varices in 40%), and only 6 patients had all 3 criteria. We would also patients with cirrhosis. Dig Liver Dis 2003;35:473-8. have missed 12 patients with bleeding varices, none of whom [12] Madhotra R, Mulcahy HEI, Willner I, et al. Prediction of had any of the criteria. As the percentages were low, statistical esophageal varices in patients with cirrhosis. J Clin Gastroenterol 2002;34:81-5. analysis for tests of significance was not attempted. [13] Chalasani N, Imperiale TF, Ismail A, et al. Predictors of large One limitation was the small number of patients (40) in esophageal varices in patients with cirrhosis. Am J Gastroenterol this study, which might lead to a type II error, but this 1999;94:3285-91. represented all the patients seen in our hospital in the last [14] Zaman A, Hapke R, Flora K, et al. Factors predicting the presence of 4 years with confirmed bleeding esophageal varices by esophageal or gastric varices in patients with advanced liver disease. Am J Gastroenterol 1999;94:3292-6. endoscopy. Perhaps another study with a larger number of [15] Freeman JG, Darlow S, Cole AT. Platelet count as a predictor for the patients could be done, although we doubt whether any presence of oesophageal varices in alcoholic cirrhotic patients. differences will arise. Gastroenterology 1999;116:A1211 [abstract]. Predicting variceal bleeding 535
[16] Barcia HX, Robalino GA, Molina EG, et al. Clinical predictors [18] Lavergne J, Molina E, Reddy KR, et al. Ascites predicts the presence of large varices in cirrhotic patients. Gastrointest Endosc 1998; of high grade varices by screening endoscopy. Gastrointest Endosc 47:A78. 1997;45:A649. [17] Alacorn F, Burke CA, Larive B. Esophageal varices in patients with [19] Gray D, Toghill P. The gastrointestinal tract. An introduction to the cirrhosis: is screening endoscopy necessary for everyone? Am J signs and symptoms of clinical medicine. London (UK)7 Arnold; 2001. Gastroenterol 1998;93:1673 [abstract]. p. 213-25. American Journal of Emergency Medicine (2005) 23, 536–543
www.elsevier.com/locate/ajem
The epidemiology and early clinical features of West Nile virus infection
Jacek M. Mazurek MD, PhDa,b,*,1, Kim Winpisinger MSb, Barbara J. Mattson MAEd, MSTEb, Rosemary Duffy DDS, MPHc, Ronald L. Moolenaar MD, MPHd aEpidemic Intelligence Service, Epidemiology Program Office, Centers for Disease Control and Prevention, Atlanta, GA 30333, USA bOhio Department of Health, Colombus, OH 43215, USA cNational Center for Chronic Diseases Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, GA 30333, USA dEpidemiology Program Office, Centers for Disease Control and Prevention, Atlanta, GA 30333, USA
Received 9 November 2004; accepted 28 November 2004
Abstract We studied early clinical features of the West Nile virus (WNV) infection. Case patients were Ohio residents who reported to the Ohio Department of Health from August 14 to December 31, 2002, with a positive serum or cerebrospinal fluid for anti–WNV IgM. Of 441 WNV cases, medical records of 224 (85.5%) hospitalized patients were available for review. Most frequent symptoms were fever at a temperature of 38.08C or higher (n = 155; 69.2%), headache (n = 114; 50.9%), and mental status changes (n = 113; 50.4%). At least one neurological symptom, one gastrointestinal symptom, and one respiratory symptom was present in 186 (83.0%), 119 (53.1%), and 46 (20.5%) patients, respectively. Using multivariate logistic regression and controlling for age, we found that the initial diagnosis of encephalitis ( P = .001) or reporting abdominal pain ( P b .001) was associated with death. Because initial symptoms of WNV infection are not specific, physicians should maintain a high index of suspicion during the epidemic season, particularly in elderly patients with compatible symptoms. This is a US government work. There are no restrictions on its use. Published by Elsevier Inc.
1. Introduction
The West Nile virus (WNV) was first identified in 1937 T Corresponding author. NIOSH, DRDS, Surveillance Branch Mailstop in the West Nile district of Uganda [1]. Initially, WNV HG 900.2 Morgantown, WV 26505, USA. Tel.: +1 304 285 5983; fax: +1 infections were characterized by mild disease with low- 304 285 6111. grade fever, malaise, anorexia, nausea and vomiting, eye E-mail address: [email protected] (J.M. Mazurek). 1 Current address: National Institute for Occupational Safety and pain, headache, myalgia, rash, and lymphadenopathy [2-4]. Health, Centers for Disease Control and Prevention, Morgantown, WV In recent years, an increase in the frequency and clinical 26505, USA. severity of WNV infection has been observed [5-9].
0735-6757/$ – see front matter. Published by Elsevier Inc. doi:10.1016/j.ajem.2004.11.005 Early symptoms of WNV infection 537
The first case of WNV in North America occurred in the 2. Methods United States in New York City in 1999 [7]. From that year to the present, WNV has spread rapidly westward to most states 2.1. Case definition [7-9]. In Ohio (2000 population, 11.3 million), the first WNV We defined a probable human WNV case patient as an activity was observed in 2001 in crows and blue jays. In 2002, Ohio resident with a positive serum or CSF anti–WNV IgM all of Ohio’s 88 counties reported WNV activity in birds, test result who was reported to the ODH between January 1 mosquitoes, horses, or human beings. The first bird tested and December 31, 2002. A confirmed case patient was a positive on May 19, the first mosquito pool on May 28, and probable case patient who tested positive for WNV infection the first horse on June 6. The first infected human being in at the CDC by PRNT (with or without serum- or CSF- Ohio became ill on July 27 and was reported on August 14. A positive anti–WNV IgM test results) or who had CSF test total of 441 WNV human cases was reported to the Ohio results that were positive for anti–WNV IgM and negative Department of Health (ODH) for 2002. for anti–SLE IgM by ELISA. Probable and confirmed case In Ohio, WNV encephalitis is a reportable disease [10]. patients who presented to Ohio EDs and were subsequently Local health departments report cases to the ODH. In 2002, admitted to a hospital were included in the analysis. serum or cerebrospinal fluid (CSF) samples taken from hospitalized patients with suspected WNV encephalitis or 2.2. Case ascertainment meningitis were sent to the ODH laboratory for anti–WNV IgM testing by enzyme-linked immunosorbent assay We reviewed the serological WNV test results and (ELISA). The ODH laboratory provided free arboviral information from WNV surveillance of human beings as testing for WNV, St Louis encephalitis (SLE), and La reported to the ODH for 2002 and created a list of patients Crosse encephalitis for patients hospitalized with a diagno- who tested positive for WNV at the ODH, a private, or the sis of viral encephalitis, viral meningoencephalitis, or viral CDC laboratory. Sublists of WNV patients were provided to (aseptic) meningitis. Patients with mild illness or those who the local health departments in whose jurisdiction the were exposed but remained asymptomatic were referred to patients resided and to the infection control practitioners at private laboratories for testing. Laboratories that identify the hospital where the patients were evaluated. Local health WNV infection are required to report it to the ODH departments and hospital staff ensured that case patient laboratory and to the local health department within whose medical charts were available for review. jurisdiction an infected patient resides. Human WNV Twenty-eight abstractors from the ODH, local health infections were deemed confirmed at the ODH laboratory departments, and hospital infection control departments by ELISA testing on CSF when the anti–WNV IgM result reviewed ED medical charts by using a standard data was positive and the anti–SLE IgM result was negative [11]. collection form and collected demographic data, clinical data, CSF laboratory results, information on initial diagno- In addition, some cases were confirmed at the Centers for ses, potential risk factors, disposition, and final diagnosis. Disease Control and Prevention (CDC) by plaque reduction Abstractors received detailed instruction on how to collect neutralization test (PRNT) [12]. The ODH requested this the data. If the desired information was not available, we additional testing in the initial phase of the WNV outbreak, reviewed the physicians’ and nurses’ notes. Test results when it was thought that the patient might have been other than WNV serology performed on CSF were obtained exposed to other closely related viruses (eg, SLE virus), from the charts or directly from the hospitals’ laboratories. which might result in a false-positive laboratory test for WNV. The CDC reported PRNT test results to the ODH. 2.3. Statistical analysis The 1999 and 2000 outbreaks in New York and New Jersey and the WNV surveillance system implemented in We used Epi Info (CDC, Atlanta, GA) and SAS statistical conjunction with the CDC provided an opportunity to software (SAS, Cary, NC) to analyze collected data [16,17]. describe in detail the clinical characteristics of hospitalized Simple proportions and medians were generated to describe human cases [13-15]. However, the frequency of various the demographic characteristics and the frequencies of initial signs and symptoms associated with early WNV infection diagnoses, signs or symptoms reported on admission, or has not been specifically addressed. Because another WNV underlying conditions. We used the Wilcoxon test to outbreak is possible, knowledge of early symptoms could compare means of continuous data and the Fisher exact enable physicians to more rapidly diagnose patients with test to determine if any initial diagnosis, sign or symptom WNV infection. This paper describes the epidemiology of reported on admission, or underlying condition was WNV in Ohio, the initial signs and symptoms of the WNV associated with death; a 2-tailed P value less than .05 was infection of patients who presented to Ohio hospital accepted as significant. We calculated 95% confidence emergency departments (EDs) and were subsequently intervals (CIs) of odds ratios (ORs) using exact method. hospitalized in 2002, initial diagnoses of cases noted within A multiple logistic regression analysis was used to 24 hours after hospital admission, and the final diagnoses. In evaluate the association of initial diagnosis, signs and addition, we evaluated the association of individual risk symptoms, and underlying conditions reported on admission factors with death from WNV infection. with death while controlling for age. Variables were 538 J.M. Mazurek et al.
and negative for anti–SLE IgM. Thirty-one (7.0%) cases were fatal. The 441 patients had a median age of 61 years (range, 2-98 years); 227 (51.5%) were women. The median age of the 31 decedents was 75 years (range, 32-98 years); 9 (29.0%) were women. Onset of symptoms of WNV infection among reported patients occurred during the period of July 27 to October 19, 2002 (Fig. 1). Cases occurred among residents of 56 (63.6%) of Ohio’s 88 counties, with 219 (49.4%) cases reported from the most populated county in Ohio, Cuyahoga County (2000 population, 1.3 million). The attack rate during the period of January 1 to December 31, 2002, was 3.9 per 100000 population for the state, compared with 15.6 per 100000 for Cuyahoga County. Of the 441 WNV patients reported in Ohio, 262 (59.4%) Fig. 1 Human West Nile virus infections reported from Ohio by had medical charts available for immediate review. Of these, onset date (n = 441). 224 (85.5%) were hospitalized, 28 (10.7%) were ED-only patients, and 10 (3.8%) were outpatients. considered for inclusion in the multivariate model if they Of the 224 hospitalized case patients, 113 (50.4%) were were associated with death at a statistical significance level women and 179 (79.9%) were white (Table 1). Their ages of P b .25 in the univariate analysis and were reported by ranged from 11 to 98 years (median, 67 years). Over half of more than 5% of the study population. Using all-subsets the patients were aged 65 years or older (n = 129; 57.6%). regression, we then identified several models all of whose Fourteen (6.3%) case patients died. The median age of these variables remained significantly associated with death at a 14 decedents was 75 years (range, 62-98 years); 7 (50.0%) significance level of P = .05. were women. The highest proportion of deaths (1/3 [33.3%]) among hospitalized patients occurred among persons aged 90 to 99 years (Fig. 2). The median time 3. Results from onset to death was 13 days (range, 2-44 days). Case patients were hospitalized in 47 hospitals in 27 counties. During the period of January 1 to December 31, 2002, a The admission diagnoses of aseptic meningitis, enceph- total of 441 probable cases with positive test results for alitis, and meningoencephalitis were made in 68 (30.4%), WNV in CSF or serum (IgM ELISA) was reported to the 21 (9.4%), and 9 (4.0%) of the hospitalized patients, ODH. Of these, 140 (31.7%) cases were tested and respectively (Table 2). Frequent initial diagnoses included confirmed at the CDC by PRNT or at the ODH laboratory bfever of unknown originQ (n = 39; 17.4%), viral infections by ELISA CSF testing that was positive for anti–WNV IgM (n = 39; 17.4%), urinary tract infection (n = 33; 14.7%), and
Table 1 Number and percentage of hospitalized WNV-infected patients by selected demographic characteristics—Ohio, 2002 Characteristics Nonfatalities [n (%)] Fatalities [n (%)] Total [n (%)] Crude OR 95% Confidence limits P n = 210 (93.7%) n = 14 (6.3%) n = 224 (100%) Lower Upper Age group (y) 10-19 5 (2.4) 0 (0.0) 5 (2.2) 0.00 0.00 26.67 1.000 20-29 7 (3.3) 0 (0.0) 7 (3.1) 0.00 0.00 18.17 1.000 30-39 14 (6.7) 0 (0.0) 14 (6.3) 0.00 0.00 8.60 1.000 40-49 31 (14.8) 0 (0.0) 31 (13.8) 0.00 0.00 3.78 .511 50-59 25 (11.9) 0 (0.0) 25 (11.2) 0.00 0.00 4.71 .538 60-69 41 (19.5) 2 (14.3) 43 (19.2) 1.00 – – Ref. 70-79 58 (27.6) 5 (35.7) 63 (28.1) 1.39 0.19 16.06 1.000 80-89 27 (12.9) 6 (42.9) 33 (14.7) 4.06 0.59 44.92 .118 90-99 2 (1.0) 1 (7.1) 3 (1.3) 9.75 0.11 256.84 .195 Sex Women 106 (50.5) 7 (50.0) 113 (50.4) 0.98 0.28 3.40 1.000 Men 104 (49.5) 7 (50.0) 111 (49.6) 1.00 – – Ref. Race White 166 (79.0) 13 (92.9) 179 (79.9) 1.00 – – Ref. Black 30 (14.3) 1 (7.1) 31 (13.8) 0.43 0.01 3.05 .698 Other/Unknown 14 (6.7) 0 (0.0) 14 (6.3) 0.00 0.00 3.41 .604 Early symptoms of WNV infection 539
of cancer were the least reported. None of the patients were reported to be HIV positive or to take immunosup- pressive medications. In 65 (29.0%) case patients, CSF was collected on or before admission. The findings reflected changes charac- teristic of viral infection: mild leukocytosis (median, 128 cells/mm3; range, 0-725 cells/mm3; normal values [18], 0-5 cells/mm3), high protein (median, 72 mg/dL; range, 23-700 mg/dL; normal values, 15-45 mg/dL), and normal glucose (median, 61 mg/dL; range, 22-224 mg/dL; normal values, 50-80 mg/dL). Twenty-four (36.9%) patients had a predominance of neutrophils (N50%; normal values, 0%-8%) in the CSF. The disease onset ranged between July 27 and October 2, 2002. Patients were admitted during the period of August 8 to October 17, 2002. The mean and median times from symptom onset to admission were 4.3 and 3 days, Fig. 2 Rate of hospitalizations and percentage of deaths among respectively (range, 0-32 days). The mean and median hospitalized patients by age group (n = 224). durations of hospitalization were 9.5 and 7 days, respec- dehydration (n = 32; 14.3%). In 12 (5.4%) cases, the tively (range, 1-65 days). There was no difference in time admission diagnosis was WNV infection. from symptom onset to admission ( P = .242) and duration The signs and symptoms most frequently recorded at of hospitalization ( P = .153) between those who subse- admission are reported in Table 3. At least one neurological quently died and those who survived. Hospitalization longer symptom (headache, stiff neck, photophobia, muscle weak- than 7 days was associated only with hypertension (OR = ness, paresthesias, seizures, or mental status changes), one 2.4; 95% CI, 1.1-4.8; P = .021). gastrointestinal symptom (nausea, vomiting, or diarrhea), The final diagnoses of the 224 hospitalized patients, as and one respiratory symptom (cough or shortness of breath) noted in the medical charts at the time of data collection, was present in 186 (83.0%), 119 (53.1%), and 46 (20.5%) included viral encephalitis (n = 77; 34.4%), aseptic patients, respectively. Rarely reported symptoms included meningitis (n = 64; 28.6%), WNV infection (n = 55; slurred speech, sore throat, blurred vision, lymphadenopa- 24.6%), viral meningoencephalitis (n = 21; 9.4%), others thy, and paresthesias. (n = 6; 2.7%), and Guillain-Barre´ syndrome (n = 1; 0.4%). Some of the underlying medical conditions present in Hospitalized patients were discharged home (n = 126; case patients at the time of admission are reported in 56.3%), to a skilled nursing facility (n = 66; 29.5%), or to Table 3. Diabetes and hypertension were the most the patient caretaker’s home (n = 5; 2.2%). Nine patients prominent conditions reported, and asthma and history (4.0%) were transferred to another acute care hospital.
Table 2 Admission diagnoses of hospitalized WNV-infected patients Admission diagnosis Nonfatalities [n (%)] Fatalities [n (%)] Total [n (%)] Crude OR 95% Confidence limits P n = 210 (93.7%) n = 14 (6.3%) n = 224 (100%) Lower Upper Aseptic meningitis 66 (31.4) 2 (14.3) 68 (30.4) 0.36 0.04 1.71 .237 Fever 33 (15.7) 6 (42.9) 39 (17.4) 1.76 0.24 77.60 1.000 Viral infection 38 (18.1) 1 (7.1) 39 (17.4) 0.35 0.01 2.46 .473 Urinary tract infection 31 (14.8) 2 (14.3) 33 (14.7) 0.96 0.10 4.65 1.000 Dehydration 28 (13.3) 4 (28.6) 32 (14.3) 2.60 0.55 9.77 .121 Mental status changes 23 (11.0) 2 (14.3) 25 (11.2) 1.36 0.14 6.69 .660 Pneumonia 21 (10.0) 1 (7.1) 22 (9.8) 0.69 0.02 5.08 1.000 Encephalitis 16 (7.6) 5 (35.7) 21 (9.4) 6.74 1.56 25.42 .005 Sepsis 20 (9.5) 0 (0.0) 20 (8.9) 0.00 0.00 2.45 .620 WNV infection 12 (5.7) 0 (0.0) 12 (5.4) 0.00 0.00 4.44 1.000 Cerebrovascular 9 (4.3) 1 (7.1) 10 (4.5) 1.72 0.04 14.18 .483 accident/stroke Meningoencephalitis 8 (3.8) 1 (7.1) 9 (4.0) 1.94 0.04 16.47 .447 Cephalgia 9 (4.3) 0 (0.0) 9 (4.0) 0.00 0.00 6.23 1.000 Parkinson’s disease 3 (1.4) 0 (0.0) 3 (1.4) 0.00 0.00 26.78 1.000 Rhabdomyolysis 3 (1.4) 0 (0.0) 3 (1.4) 0.00 0.00 26.78 1.000 540 J.M. Mazurek et al.
Table 3 Clinical characteristics of nonfatal and fatal hospitalized WNV-infected patients by signs, symptoms, and underlying conditions recorded on admission Characteristic Nonfatalities [n (%)] Fatalities [n (%)] Total [n (%)] Crude OR 95% Confidence limits P n = 210 (93.7%) n = 14 (6.3%) n = 224 (100%) Lower Upper Signs and symptoms Fever z 388C(z100.48F) 145 (69.0) 10 (71.4) 155 (69.2) 1.12 0.31 5.08 1.000 Headache 113 (53.8) 1 (7.1) 114 (50.9) 0.07 0.00 0.46 .001 Mental status changes 100 (47.6) 13 (92.9) 113 (50.4) 14.30 2.06 613.11 .001 Nausea 76 (36.2) 6 (42.9) 82 (36.6) 1.32 0.36 4.53 .775 Vomiting 73 (34.8) 7 (50.0) 80 (35.7) 1.88 0.54 6.52 .262 Chills 66 (31.4) 6 (42.9) 72 (32.1) 1.64 0.45 5.61 .386 Muscle weakness 63 (30.0) 6 (42.9) 69 (30.8) 1.75 0.48 6.01 .372 Confusion 55 (26.2) 3 (21.4) 58 (25.9) 0.77 0.13 3.06 1.000 Fatigue 49 (23.3) 4 (28.6) 53 (23.7) 1.31 0.29 4.81 .745 Lethargy 41 (19.5) 8 (57.1) 49 (21.9) 5.50 1.56 20.15 .003 Decreased appetite 43 (20.5) 5 (35.7) 48 (21.4) 2.16 0.54 7.59 .186 Diarrhea 35 (16.7) 2 (14.3) 37 (16.5) 0.83 0.09 4.00 1.000 Myalgia 33 (15.7) 1 (7.1) 34 (15.2) 0.41 0.01 2.94 .700 Malaise 29 (13.8) 2 (14.3) 31 (13.8) 1.04 0.11 5.05 1.000 Abdominal pain 20 (9.5) 6 (42.9) 28 (12.5) 7.13 1.82 25.85 .002 Neck stiffness 28 (13.3) 0 (0.0) 28 (12.5) 0.00 0.00 1.65 .226 Skin rash 27 (12.9) 1 (7.1) 28 (12.5) 0.52 0.01 3.75 1.000 Shortness of breath 24 (11.4) 3 (21.4) 27 (12.1) 2.11 0.35 8.77 .386 Cough 24 (11.4) 2 (14.3) 26 (11.6) 1.29 0.13 6.35 .670 Dizziness 21 (10.0) 1 (7.1) 22 (9.8) 0.69 0.02 5.08 1.000 Increased sleepiness 18 (8.6) 3 (21.4) 21 (9.4) 2.91 0.47 12.40 .132 Balance problems 16 (7.6) 2 (14.3) 18 (8.0) 2.02 0.20 10.34 .312 Photophobia 15 (7.1) 0 (0.0) 15 (6.7) 0.00 0.00 3.42 .606 Back pain 10 (4.8) 2 (14.3) 12 (5.4) 3.33 0.32 18.36 .167 Join pain (arthralgia) 11 (5.2) 0 (0.0) 11 (4.9) 0.00 0.00 4.91 1.000 Tremor 9 (4.3) 1 (7.1) 10 (4.5) 1.72 0.04 14.18 .483 Weight loss 9 (4.3) 1 (7.1) 10 (4.5) 1.72 0.04 14.18 .483 Slurred speech 5 (2.4) 3 (21.4) 8 (3.6) 11.18 1.50 65.04 .009 Neck pain 7 (3.3) 0 (0.0) 7 (3.1) 0.00 0.00 8.45 1.000 Sore throat 6 (2.9) 0 (0.0) 6 (2.7) 0.00 0.00 10.25 1.000 Seizures 4 (1.9) 1 (7.1) 5 (2.2) 3.96 0.07 43.61 .274 Blurred vision 4 (1.9) 0 (0.0) 4 (1.8) 0.00 0.00 17.54 1.000 Coma 2 (1.0) 1 (7.1) 3 (1.3) 8.00 0.13 159.83 .177 Numbness 2 (1.0) 1 (7.1) 3 (1.3) 8.00 0.13 159.83 .177 Flaccid paralysis 1 (0.5) 0 (0.0) 1 (0.4) 0.00 0.00 285.00 1.000 Lympadenopathy 1 (0.5) 0 (0.0) 1 (0.4) 0.00 0.00 285.00 1.000 Paresthesias 1 (0.5) 0 (0.0) 1 (0.4) 0.00 0.00 285.00 1.000
Underlying condition Diabetes 38 (18.1) 4 (28.6) 42 (18.8) 1.81 0.39 6.69 .305 Hypertension 33 (15.7) 5 (35.7) 38 (17.0) 2.98 0.73 10.60 .067 Chronic obstructive 11 (5.2) 1 (7.1) 12 (5.4) 1.39 0.03 11.05 .548 pulmonary disease Dementia 12 (5.7) 0 (0.0) 12 (5.4) 0.00 0.00 4.44 1.000 Coronary artery disease 7 (3.3) 0 (0.0) 7 (3.1) 0.00 0.00 8.45 1.000 Alcoholism 5 (2.4) 0 (0.0) 5 (2.2) 0.00 0.00 12.97 1.000 Asthma 2 (1.0) 0 (0.0) 2 (0.9) 0.00 0.00 53.44 1.000 Cancer 2 (1.0) 0 (0.0) 2 (0.9) 0.00 0.00 53.44 1.000 Immunosuppression 1 (0.5) 0 (0.0) 1 (0.4) 0.00 0.00 285.00 1.000
In the univariate analysis, case patients who subsequently diagnosed with encephalitis (Table 2), and to report the died were more likely than survivors to be aged 68 years or presence of the following symptoms: mental status changes older (OR = 6.9; 95% CI, 1.5-31.4; P = .005), to be initially (referred here to a constellation of symptoms including Early symptoms of WNV infection 541 altered mental status, confusion, lethargy, increased sleep- We found that an initial diagnosis of encephalitis or iness, slurred speech, or memory loss), lethargy, slurred reporting abdominal pain at admission placed patients at speech, and abdominal pain (Table 3). Survivors were more greater risk for death. West Nile virus encephalitis has been likely than those who died to report headache (OR = 15.1; previously documented as a cause of death [22-26], but 95% CI, 1.9-117.9). The presence of at least one there are no previous reports discussing the significance of neurological symptom, one gastrointestinal symptom, and abdominal pain. Three categories of abdominal pain are one respiratory symptom at admission was not associated recognized: visceral, somatoparietal, and referred pain [27]. with death, nor was any underlying condition. Pain sensations are conveyed from neuroreceptors through 2 We performed multivariable logistic regression analysis types of afferent nerve fibers in which cell bodies are and evaluated 3 models within each group—initial diagnosis, located in the dorsal root ganglia of spinal afferent nerves sign and symptoms, and underlying condition—with the [27]. A lesion in the dorsal root ganglion can result in acute following predictor variables included in the models, respec- autonomic and sensory neuropathy [28,29]. The reduced tively: (1) age, diagnosis of meningitis, dehydration, and sensory nerve action caused by WNV infection has been encephalitis; (2) age, presence of headache, mental status demonstrated [30]. In addition, pathological data showed changes, lethargy, decreased appetite, abdominal pain, neck that WNV causes dorsal root ganglia lesions. The lesions stiffness, increased sleepiness, and back pain; and (3) age and might be severe enough to produce sensory deficits and hypertension. The final models included (1) age and diagnosis correlate with the severity of the poliomyelitis [31,32]. of encephalitis and (2) age and presence of abdominal pain. The case fatality rate among hospitalized patients in Ohio After controlling for age, hospitalized patients who subse- was 6.3%, similar to that reported in Romania (4.3%) but quently died were more likely to be initially diagnosed with lower than that reported in New York and New Jersey encephalitis (OR = 9.4; 95% CI, 2.4-36.7; P = .001) and (11.0%) and in Israel (14.1%) [6,15,19,20]. The reasons for reported abdominal pain (OR = 10.5; 95% CI, 2.8-39.0; P b the observed differences are not apparent. Among them may .001) compared with case patients who survived. have been differences in the length of the follow-up periods of hospitalized patients, the criteria for WNV-related death, completeness of the ascertainment of encephalitis cases, 4. Discussion changes in WNV virulence [19,21], the immunological background in a population, the prevalence of comorbid In this paper, we described the early clinical features of conditions, age distribution of the population, and others. 224 WNV hospitalized patients during an outbreak of WNV This study had limitations. First, in Ohio, only hospital- infection in Ohio in 2002. We found that at least one ized patients presenting with neurological symptoms of neurological symptom was present in most patients (83%), infection were tested for WNV without charge. Nonhospi- but an initial diagnosis of encephalitis, meningitis, or talized patients with WNV infection were not studied. meningoencephalitis was made in only 90 (34.4%) case Consequently, the full spectrum of symptoms of WNV patients. Despite an ongoing outbreak, the WNV infection illness, especially mild disease, could not be discovered, and was initially suspected in only 12 (5.4%) case patients. The the total burden of disease might have been underestimated. small number of cases initially diagnosed with WNV Second, the information on symptoms was obtained from infection might reflect the nonspecific presentation of the ED medical charts, which vary in completeness and early phase of infection and the fact that this was the first accuracy of clinical details. We were not able to validate year that human WNV infection was present in Ohio. the data provided by each hospital. In addition, 28 recorders The clinical presentation of disease in Ohio was similar conducted chart reviews. Although we gave them specific to that reported from the 1996 outbreak in Romania and the instructions on how to complete the survey, interabstractor 2000 outbreaks in New York and New Jersey and in Israel agreement was not determined and differences in complete- [6,15,19,20], with frequent reports of fever, headache, and ness and quality of data likely occurred. Also, for analysis mental status changes. The clinical findings from the New purposes, we assumed that signs and symptoms not York outbreak showed that neurological and gastrointestinal recorded on the chart were not present, resulting in potential symptoms were present in about 89% and 58% of misclassification. Moreover, because of financial, time, and hospitalized patients, respectively. We observed similar legal constraints, we were able to collect data on only about frequencies of neurological and gastrointestinal symptoms 60% of the reported WNV cases from 27 (48%) of 56 in 83% and 53% of the hospitalized case patients, counties. West Nile virus cases in other counties might differ respectively (Table 2). Similarities in the frequencies of from those that we analyzed. symptoms might be caused by the fact that the WNV strain Our study demonstrated that WNV infection begins with circulating in the United States was genetically similar to the nonspecific symptoms, making early clinical diagnosis WNV strain that was responsible for the 2000 outbreak in challenging. Diagnosis is based mainly on serological tests Israel [21]. Although frequent, the presence of at least one for the presence of anti–WNV IgM, which can be neurological or one gastrointestinal symptom at admission demonstrated either in serum or in CSF. Nevertheless, the was not associated with death. results have to be interpreted with caution because cross- 542 J.M. Mazurek et al. reactivity with other members of the genus Flavivirus is [2] Goldblum N, Sterk VV, Paderski B. West Nile fever; the clinical possible. Because no effective therapy is known, patients features of the disease and the isolation of West Nile virus from the blood of nine human cases. Am J Hyg 1954;59:89-103. with WNV infection can be offered only supportive care. [3] Marberg K, Goldblum N, Sterk VV, et al. The natural history of West Physicians should maintain a high index of suspicion for Nile fever: I. Clinical observations during an epidemic in Israel. Am J WNV infection during the epidemic season, particularly Hyg 1956;64:259-69. when evaluating elderly patients with neurological or [4] Hubalek Z, Halouzka J. West Nile fever—a reemerging mosquito- gastrointestinal symptoms. borne viral disease in Europe. Emerg Infect Dis 1999;5:643-50. [5] Platonov AE, Shipulin GA, Shipulina OY, et al. Outbreak of West Nile virus infection, Volgograd Region, Russia, 1999. Emerg Infect Dis 2001;7:128-32. Acknowledgments [6] Tsai TF, Popovici F, Cernescu C, et al. West Nile encephalitis epidemic in southeastern Romania. Lancet 1998;352:767-71. [7] Centers for Disease Control and Prevention. Outbreak of West Nile– We thank Virginia Abel (Akron City Hospital, Summa like viral encephalitis—New York, 1999. MMWR Morb Mortal Wkly Health System), Marlene Becker (Massillon Community Rep 1999;48:845-9. Hospital), Shari Botts (the Fort Hamilton Hospital), Jane [8] Centers for Disease Control and Prevention. Human West Nile virus Carmean (Clinton Memorial Hospital), Sandra Hensey surveillance—Connecticut, New Jersey, and New York, 2000. (Medical College of Ohio), Joan Kimmet (Mercy Hospital MMWR Morb Mortal Wkly Rep 2001;50:265-8. [9] Centers for Disease Control and Prevention. West Nile virus of Willard), Toni Leshinski (Hillcrest Hospital), Carolyn activity—United States, 2001. MMWR Morb Mortal Wkly Rep May (Kettering Medical Center), Patricia A. Nelson (Mercy 2002;51:497-501. Medical Center), Joan Pugnale (Aultman Health Founda- [10] Ohio Administrative Code. Chapter 3701-3-04. Communicable tion), Charla Ulrich (St Vincent Hospital), Sharon L. Wells diseases and ailments, laboratory results reporting. Available at: (Miami Valley Hospital), Kathy Zegarski (Southview http://www.odh.state.oh.us/Rules/Final/Chap3/FR3_lst.htm [Accessed March 16, 2004]. Hospital), Vickie Davis (Coshocton City Health Depart- [11] Martin DA, Muth DA, Brown T, et al. Standardization of ment), Judy Fisher (Paulding Co Health Department), Patty immunoglobulin M capture enzyme-linked immunosorbent assays Fruth (Defiance Co General Health District), Ann Glidewell for routine diagnosis of arboviral infections. J Clin Microbiol (Darke Co General Health District), Jo Holt (Clark Co 2000;38:1823-6. Health Department), Susan Irvine (Hamilton City Health [12] Beaty BJ, Calisher CH, Shope RE. Arboviruses. In: Schmidt NH, Emmons RW, editors. Diagnostic procedures for viral, rickettsial and Department), Cindy Jones (Auglaize Co Health Depart- chlamydial infections. 6th ed. Washington (DC)7 American Public ment), Amy Jones (Wood Co Health District), Joan Health Association Inc;, 1989. p. 797-856. Kowalczyk (Elyria City Department of Health), Selene [13] West Nile Virus Workgroup. Combating West Nile virus. The plan for Layton (Trumbull Co Health Department), Sheryl Long the state of Ohio. May 2003. Available at: http://www.odh.state.oh.us/ (Stark Co Combined General Health District), Kathy Martin ODHPrograms/ZOODIS/WNV/Pubs/WNVStPlan.PDF [Accessed March 16, 2004]. (Fairfield Department of Health), Linda Mehl (Guernsey Co [14] Centers for Disease Control and Prevention. Case definitions for Health Department), Pam Pflum (Henry Co Health Depart- infectious conditions under public health surveillance. MMWR Morb ment), Jacquelyn Phillips (Middletown City Health Depart- Mortal Wkly Rep 1997;46:12-3. ment), Helen Rogers (Wayne Co Health Department), [15] Weiss D, Carr D, Kellachan J, et al. West Nile Virus Outbreak Cynthia Rose (Fulton Co Health Department), Sharon Response Working Group. Clinical findings of West Nile virus infection in hospitalized patients, New York and New Jersey, 2000. Schaeffer (Erie Co General Health District), Judy Louden Emerg Infect Dis 2001;7:654-8. and Peggy Scherer (Geauga Co General Health District), [16] Dean AG, Dean JA, Culombies D, et al. Epi Info, version 6.04: a word Heidi Scaife (Cuyahoga Co Board of Health), Janine processing, database, and statistical program for public health on IBM Trottier (Lorain Co General Health District), Rose-Ann compatible microcomputers. Atlanta (GA)7 Centers for Disease Warth (Canton City Health Department), and Barbara Control and Prevention; 1995. [17] SAS Institute Inc. The SAS system for Windows, release 8.02. Cary Wilhelm (Findlay City Health Department) for their help (NC)7 SAS Institute Inc; 1999-2001. in accessing and abstracting medical charts. [18] Brunzel NA. Fundamentals of urine and body fluid analysis. 2nd ed. We are grateful to Dr Amy Bode (Division of Vector- Philadelphia (PA)7 WB Saunders Co;, 2004. p. 337. Borne Infectious Diseases, National Center for Infectious [19] Chowers MY, Lang R, Nassar F, et al. Clinical characteristics of the Diseases) and Dr Matthew Zack (Division of Adult and West Nile fever outbreak, Israel, 2000. Emerg Infect Dis 2001; 7:675-8. Community Health, National Center for Chronic Disease [20] Weinberger M, Pitlik SD, Gandacu D, et al. West Nile fever Prevention and Health Promotion, Centers for Disease outbreak, Israel, 2000: epidemiologic aspects. Emerg Infect Dis Control and Prevention) for their thoughtful comments. 2001;7:686-91. [21] Lanciotti RS, Roehrig JT, Deubel V, et al. Origin of the West Nile virus responsible for an outbreak of encephalitis in the northeastern United States. Science 1999;286:2333-7. References [22] Shieh WJ, Guarner J, Layton M, et al. The role of pathology in an investigation of an outbreak of West Nile encephalitis in New York, [1] Smithburn KC, Hughes TP, Burke AW, et al. A neurotropic virus 1999. Emerg Infect Dis 2000;6:370-2. isolated from the blood of a native of Uganda. Am J Trop Med Hyg [23] Sampson BA, Ambrosi C, Charlot A, et al. The pathology of human 1940;20:471-92. West Nile virus infection. Hum Pathol 2000;31:527-31. Early symptoms of WNV infection 543
[24] Kelley TW, Prayson RA, Ruiz AI, et al. The neuropathology of West [29] Satake M, Nakagawa Y, Yamashita S, et al. Subacute autonomic and Nile virus meningoencephalitis. A report of two cases and review of sensory ganglionopathy: a postmortem case. J Neurol Neurosurg the literature. Am J Clin Pathol 2003;119:749-53. Psychiatry 1998;64:561. [25] Bouffard JP, Riudavets MA, Holman R, et al. Neuropathology of the [30] Doron SI, Dashe JF, Adelman LS, et al. Histopathologically proven brain and spinal cord in human West Nile virus infection. Clin poliomyelitis with quadriplegia and loss of brainstem function due to Neuropathol 2004;23:59-61. West Nile virus infection. Clin Infect Dis 2003;37:e74-7 [Electronic [26] Smith RD, Konoplev S, DeCourten-Myers G, et al. West Nile virus publication 2003 Aug 11]. encephalitis with myositis and orchitis. Hum Pathol 2004;35:254-8. [31] Fratkin JD, Leis AA, Stokic DS, et al. Spinal cord neuropathology in [27] Glasgow RE, Mulvihill SJ. Abdominal pain, including the acute human West Nile virus infection. Arch Pathol Lab Med 2004; abdomen. In: Sleissenger MH, Fordtran JS, editors. Gastrointestinal 128:533-7. disease. 7th ed. Philadelphia (PA)7 WB Saunders Co;,2002. p. 71-83. [32] Kelley TW, Prayson RA, Isada CM. Spinal cord disease in West Nile [28] Colan RV, Snead III OC, Oh SJ, et al. Acute autonomic and sensory virus infection. N Engl J Med 2003;348:564-6. neuropathy. Ann Neurol 1980;8:441-4. American Journal of Emergency Medicine (2005) 23, 544–547
www.elsevier.com/locate/ajem
Emergency diagnosis of Fournier’s gangrene with bedside ultrasound
Daniel Morrison MD, Michael Blaivas MD*, Matthew Lyon MD
Department of Emergency Medicine, Medical College of Georgia, Augusta, GA 30912-4007, USA
Received 12 December 2004; accepted 12 December 2004
Abstract FournierTs gangrene is a life-threatening infection of the scrotal skin. Although originally thought to be an idiopathic process, FournierTs gangrene has been shown to have a predilection for patients with diabetes as well as chronic alcohol abuse; however, it can also affect patients with nonobvious immune compromise. Because of potential complications, it is important to diagnose the disease process as early as possible. Ultrasound has been previously described to aid in the diagnosis of FournierTs gangrene. In patients with low to moderate suspicion of FournierTs gangrene, it may provide a rapid and reliable diagnosis and differentiate the pathological process from mimicking entities such as scrotal edema or cellulitis. We present 6 cases of FournierTs gangrene diagnosed in the ED at the patient’s bedside using ultrasound. None of the patients had a history of diabetes, and 5 had sources of infection determined. D 2005 Elsevier Inc. All rights reserved.
1. Introduction to be more frequent in men with impaired immune function such as those with diabetes mellitus. FournierTs gangrene was originally described in 1883 FournierTs gangrene may be easily diagnosed solely on and was thought to be an idiopathic process. J.A. Fournier, a clinical grounds when advanced, and morbidity and Parisian venereologist, noted the disease process to be a mortality are quite high. However, early detection is rapidly progressing fasciitis of the scrotum and penis that essential to decreasing the morbidity and mortality of this was without a known source [1]. Since then, FournierTs life-threatening disease. Because FournierTs gangrene is a gangrene has been identified as having a polymic- disease on the far end of a spectrum of infection severity, in robial origin with causative agents including Escherichia many cases, it is difficult to distinguish mild FournierTs coli, Proteus, Pseudomonas, Streptococcus, Staphlococcus, gangrene from scrotal cellulitis. This distinction, however, is Bacteriodes, Clostridium, Salmonella, Klebsiella, Entero- important because the treatment and ultimate outcome differ coccus, Peptostreptococcus, Corynebacterium, and Fuso- between these entities. Any delay in diagnosis of FournierTs bacterium [2-9]. FournierTs gangrene is commonly thought gangrene such as misidentification as simple cellulitis or idiopathic scrotal edema can significantly delay treatment and lead to an increased morbidity and mortality. We present T Corresponding author. Tel.: +1 706 721 4467. 6 cases of FournierTs gangrene diagnosed at the bedside E-mail address: [email protected] (M. Blaivas). with ultrasound in patients without diabetes or other known
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.12.010 Bedside US Fournier’s diagnosis 545 immunologic disturbance. All presented with complaints of scrotal pain, swelling, and redness.
2. Case 1
J.T., a 44-year-old man with a history of hypertension, presented to the ED with complaints of scrotal swelling. The patient had been seen 5 days before arrival in an urgent care facility and diagnosed with and treated for a urinary tract infection. The patient stated that he noticed scrotal swelling 2 days prior, but it did not alarm him until the day of evaluation. He denied nausea, vomiting, or fever at home but noted burning sensation upon urination. The patient was Fig. 2 A focal collection (arrows) of gas is seen in the scrotum. febrile in the ED with a temperature of 101.58F. His remaining vital signs were unremarkable. The physical breath. He had mild to moderate lower abdominal pain and a examination was significant for diffuse scrotal swelling and significantly swollen scrotum with diffuse erythema. Rectal bilateral testicular tenderness (Fig. 1). The epididymides examination revealed an enlarged and tender prostate. were enlarged and tender bilaterally. A bedside ultrasound Bedside ultrasound examination showed a large collection revealed diffuse scrotal edema and increased flow on color of soft tissue gas in the scrotal wall. Abdominal and pelvic Doppler to both testicles with mildly enlarged epididymal computed tomography showed gas tracking into the lower heads bilaterally. In addition, a small collection of soft tissue abdominal wall. The patient was admitted to the surgical gas was noted in the scrotal midline (Fig. 2). Laboratory service with intravenous antibiotics. The patient underwent values included an elevated white blood cell count of extensive surgical debridement but did not survive to 14700 cells per microliter and presence of urinary tract discharge from the hospital. infection. Surgical consultation was obtained, and the patient was admitted to the hospital for intravenous antibiotics and surgical debridement. The patient was 4. Case 3 eventually discharged with moderate loss of tissue. T.K., a 48-year-old man with no significant medical history, presented to the ED with a complaint of scrotal 3. Case 2 swelling and discomfort for 2 days. The patient stated that a similar event occurred approximately 2 years ago and was I.O., a 56-year-old man with no significant medical self-limited. Physical examination revealed an afebrile man history, presented to the ED with complaints of lower with unremarkable vital signs. His scrotum was diffusely abdominal pain for 1 day. The patient was obviously swollen; however, there was no obvious erythema. The right intoxicated and stated that he was drinking to ease the testicle was tender. Laboratory values were significant for a discomfort. His blood pressure was 100/55 mm Hg, heart normal urine analysis and a low white blood count at 3200 rate was 95 beats per minute, respiratory rate of 18 breaths cells per microliter. A bedside ultrasound examination per minute, and a temperature of 95.58F. Physical exami- revealed diffuse scrotal edema with soft tissue gas on the nation revealed a thin patient with a smell of alcohol on his right side of the patient’s scrotum. The patient was admitted to the surgical service and was taken to the operating room for debridement. He left the hospital approximately 2 weeks later with moderate tissue loss.
5. Case 4
G.O., a 58-year-old man with a history of hypertension and mild chronic obstructive pulmonary disease, presented to the ED with a complaint of lower abdominal pain for several days. He had been seen 5 days previously and treated for presumptive diverticulitis by his primary care physician. The patient’s vital signs were unremarkable except for a temperature of 100.68F. The physical exami- nation was significant for mild suprapubic tenderness and an Fig. 1 The patient’s swollen scrotum is shown. erythematous swollen scrotum. The scrotum, as well as both 546 D. Morrison et al.
7. Case 6
L.B., a 44-year-old man with no significant medical history except for a recently treated sexually transmitted disease, presented to the ED with a complaint of scrotal pain for 5 days. The patient had been seen in 3 EDs over the previous 2 weeks and had been treated for gonorrhea and urinary tract infection. He denied any fever, nausea, or vomiting. On physical examination, the patient had unre- markable vital signs and was afebrile. Physical examination was normal except for a tender and enlarged prostate and an edematous, mildly erythematous scrotum which was dif- fusely tender. Bilateral inguinal lymph node enlargement was noted as well, but there was no abdominal tenderness or rebound. Initially thought to be consistent with idiopathic scrotal edema, the laboratory values showed a normal urinalysis and complete blood count. A bedside ultrasound Fig. 3 A plain x-ray of the pelvis and scrotum showing scrotal examination was performed to evaluate for the presence of gas (arrows). epididymitis or orchitis. However, the scan showed diffuse testicles, was diffusely tender. No focal areas of mass or scrotal edema with areas of subcutaneous gas. No focal swelling were noted. The patient alleged that this was his abscess was noted. The soft tissue edema and gas were baseline and continued to complain of lower abdominal pain traced posteriorly toward the patient’s rectum. The general only. A bedside ultrasound examination revealed bilateral surgery service was consulted regarding the patient and epididymo-orchitis and a focal area of abscess in the left admitted him with a diagnosis of FournierTs gangrene for hemiscrotum. On further evaluation, a small area of soft surgical debridement. The patient had significant tissue loss tissue gas was noted in the left hemiscrotum. Abdominal but survived to hospital discharge. and pelvic computed tomography looking for a link to possible diverticulitis or abdominal abscess confirmed a small area of scrotal soft tissue gas and focal scrotal abscess, 8. Discussion but no intra-abdominal abscess was located. The patient was T admitted for surgical debridement and antibiotics. The There have been an estimated 750 cases of Fournier s patient recovered with limited tissue loss and was dis- gangrene described in the literature since 1883 [10]. Once charged from the hospital. thought to be an idiopathic process, studies have revealed that the majority of cases have a primary source. These
6. Case 5
P.J., a 40-year-old male with a history of renal stones and hypertension, presented to the ED with a complaint of testicular pain and swelling for 12 hours. The patient denied any scrotal trauma or risk factors for sexually transmitted diseases. He was afebrile and had unremarkable vital signs. His physical examination was significant for diffuse scrotal edema and mild erythema. The patient was uncomfortable with scrotal manipulation. Although he indicated that the right hemiscrotum was most painful, he could not lateralize the discomfort on palpation. A pelvic x-ray was read out as having no evidence of subcutaneous gas. A bedside ultrasound was then performed for evaluation of possible orchitis or epididymitis. The ultrasound showed diffuse scrotal edema with a small area of soft tissue gas in the posterior midline scrotum. The patient was seen by surgical consultants and admitted for antibiotics and surgical debridement. The patient had a protracted hospital course with complicating infections, not directly related to his Fig. 4 Two loops of bowel (arrows) are seen in the scrotum with primary illness, to which he eventually succumbed. occasional specs of bowel gas. Bedside US Fournier’s diagnosis 547 sources include rectal abscesses, urinary sources such as scrotal contents, does not subject a patient to radiation, does strictures or catheters, testicular infections (epididymo- not require intravenous access or contrast, and although orchitis), and skin infections (hidradenitis, abscesses of the operator-dependent, can be easily replicated. Furthermore, scrotal wall, and infected varicella pustules). In addition, ultrasound is able to evaluate for other possible diagnosis blunt trauma to the penis, postsurgical complications, and such as torsion or epididymitis. injection of intravenous drugs into the superficial penile As in the cases presented, FournierTs gangrene can have vein have been associated with FournierTs gangrene a variety of presentations, and in fact, this diagnosis is [2,5,8,11]. The disease process typically affects men in often not considered when the presentation is in the early their 50s, and whereas it was originally thought to have a stages which can lead to the delay in diagnosis. Frequently, higher incidence in patients with diabetes, recent research the infection suspected is epididymitis or orchitis; however, has indicated a stronger association with alcohol abuse sometimes, only the scrotum is tender and not the testicles [2,5,8,11,12]. or epididymides. In all the cases presented, subcutaneous Patients with FournierTs gangrene typically have a scrotal gas noted on the ultrasound examination yielded the delayed presentation with the interval between symptoms diagnosis of FournierTs gangrene. Several of these cases and diagnosis ranging from 1 to 30 days [2,5,8,12]. The were early in the disease progression, and the amount of disease process can have significant morbidity, including gas was small. Therefore, a high index of suspicion must prolonged hospitalizations, loss of testicles, and need for be used when evaluating an edematous scrotum for the colostomies or cystostomies, and mortality which can reach presence of small pockets of gas. Like previously published 25% [2-5,8,11,12]. As a result, it is important to differentiate cases, the patients in this series had often seen other FournierTs gangrene from other presentations of acute scrotal physicians before the presentation to the ED and frequently complaints such as scrotal cellulitis, idiopathic scrotal carried a different diagnosis. The use of ultrasound by an edema, testicular torsion, testicular fracture, hemorrhage or experienced sonologist, in conjunction with a thorough necrosis of a testicular tumor, strangulated or incarcerated history and physical examination, was integral in making hernia, and other inflammatory or infectious processes. the diagnosis of FournierTs gangrene and commencing the Findings on physical examination typically include appropriate treatment. scrotal swelling and tenderness. In addition, crepitus can be identified in up to 64% of patients because of the subcutaneous air produced by the growing bacteria [8]. The radiographic detection of subcutaneous air is more sensitive References than that of physical examination with up to 89% of [1] Fournier JA. Gangrene foudroyante de la verge. Medecin Pratique patients demonstrating what is sometimes described as a 1883;4:589-97. bhoneycomb scrotumQ (Fig. 3) [7,8,11]. Ultrasonographic [2] Baskin LS, Carroll PR, Cattolica EV, McAninch JW. Necrotising soft findings typically include marked thickening of the scrotal tissue infections of the perineum and genitalia: bacteriology, treatment skin, as well as discrete focal regions of high-amplitude and risk assessment. Br J Urol 1990;65:524-9. echoes with posterior acoustic shadowing indicative of [3] Hejase MJ, Simonin JE, Bihrle R, Coogan C. Genital Fournier’s T gangrene: experience with 38 patients. Urology 1996;47:734-9. subcutaneous gas which is pathognomonic for Fournier s [4] Lamb RC, Juler GL. Fournier’s gangrene of the scrotum: a poorly gangrene (Fig. 2). Hernias in the scrotum are another defined syndrome or a misnomer? Arch Surg 1983;118:38-40. pathological process which may also demonstrate gas on [5] Benziri E, Fabiani P, Migliori G, et al. Gangrene of the perineum. ultrasound examination; however, this process is easily Urology 1996;47:935-9. differentiated from the scrotal subcutaneous gas seen in [6] Begley MG, Skawker TH, Robertson CN, et al. Fournier gangrene: T diagnosis with scrotal US. Radiology 1988;169:387-9. Fournier s gangrene as the air of a hernia is within bowel and [7] Biyani CS, Mayor PE, Powell CS. Case report: Fournier’s gangrene— contained within the scrotum (Fig. 4). Additional ultrasono- roentgenographic and sonographic findings. Clin Radiol 1995;50: graphic signs seen in FournierTs gangrene include perites- 728-9. ticular fluid and possible signs of epididymo-orchitis [6,9]. [8] Carroll PR, Cattolica EV, Turzan CW, et al. Necrotizing soft-tissue Ultrasound is the radiographic method of choice if infections of the perineum and genitalia: etiology and early reconstruction. West J Med 1996;144:174-8. history and physical examination are not diagnostic in [9] Kane CJ, Nash P, McAninch JW. Ultrasonographic appearance of suspected FournierTs gangrene. Subcutaneous air on plain necrotizing gangrene: aid in early diagnosis. Urology 1996;48:142-4. films is not universally identified, and computed tomogra- [10] Vick R, Carson CC. Fournier’s disease. Urol Clin North Am phy does not visualize the scrotal structures as well as 1999;26:841-9. ultrasound. Furthermore, computed tomography requires [11] Clayton MD, Fowler JE, Sharifi R, et al. Causes, presentation and T survival of fifty-seven patients with necrotizing fasciitis of the male intravenous contrast, and patients with Fournier s gangrene genitalia. Surg Gynecol Obstet 1990;170:49-55. often have renal impairment [2,8,11]. Testicular ultrasound [12] BaYog˘lu M, Gfl O, Yildirgan I˙. Fournier’s gangrene: review of fifteen can be performed at the patient’s bedside, can evaluate the cases. Am Surg 1997;63:1019-21. American Journal of Emergency Medicine (2005) 23, 548–551
www.elsevier.com/locate/ajem
Therapeutics Open endotracheal tubes may be safely left on an ED airway cart for 48 hours
W. Douglas Browder MD, Marcus P. Cromer MD, Francis L. Counselman MD*
Department of Emergency Medicine, Eastern Virginia Medical School and Emergency Physicians of Tidewater, Norfolk, VA 23507, USA
Received 19 August 2004; accepted 19 October 2004
Abstract Objective: To determine if endotracheal tubes (ETTs) that are opened, prepared, and stored in an ED airway cart are prone to bacterial contamination. Methods: A prospective study conducted in the ED of a level 1 trauma center. A study group of 50 endotracheal tubes were opened, preloaded with a stylet, the cuff checked for integrity by air inflation, and then stored in an ED airway cart. The study group was subdivided into 5 groups of 10 ETTs each, cultured at different time intervals. The ETTs were cultured at 6, 12, 24, 36, and 48 hours for the presence of bacterial contamination. The control group consisted of 10 ETTs that were left in their sterile packing in the ED airway cart and then removed at 48 hours and cultured. Results: In the study group, 7 (14%) ETTs resulted in a positive culture; 43 ETTs cultured negative for bacteria. In the control group, 2 (20%) ETTs cultured positive, the remainder was all negative. There was no statistically significant difference between the 2 groups ( P = .63). If only clinically significant bacteria are considered, defined as a culture with 5 or more colony-forming units, 3 ETTs in the study group cultured positive; there were no clinically significant bacteria cultured in the control group. Once again, there was no statistically significant difference between the 2 groups, with a P value of .57. Conclusion: It appears that opening, preparing, and storing ETTs in an ED airway cart for up to 48 hours does not increase the risk of bacterial contamination of the ETTs. D 2005 Elsevier Inc. All rights reserved.
1. Introduction molding the ETT into the preferred shape. When this is done, the sterile packing is violated. While intubation is not In the ED, it is not uncommon for emergency medicine a sterile procedure, this preparation could potentially lead to (EM) physicians to bset upQ the airway cart for the day. bacterial contamination. Although the majority of these Often, this involves preparing endotracheal tubes (ETTs) by ETTs are used in a timely manner, on occasion, they may be placing a stylet, checking the integrity of the cuff, and left in the cart for longer periods before use. The potential for contamination of ETTs in the ED has not been studied. A thorough Medline search performed to T Corresponding author. Tel.: +1 757 668 3397; fax: +1 757 668 2885. identify similar studies yielded no match. However, the E-mail address: [email protected] (F.L. Counselman). contamination of other medical devices has been docu-
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.10.012 Open endotracheal tubes on ED airway cart 549 mented in the literature [1-5]. Jones et al [2] demonstrated Table 2 Clinically significant positive bacterial cultures of that Staphylococcus species could be cultured from the ETTs at designated time intervals diaphragm of stethoscopes of ED workers. Moreover, the number of colonies correlated with the cleaning practices of Time of culture (h) the providers. While Staphylococcus species are by far the 612243648 most common organisms isolated, studies have also No. of positive cultures 01200 demonstrated colonization by gram-negative organisms, (study groups) such as Acinetobacter and Enterobacter [4]. Other research- No. of positive cultures 00000 ers have demonstrated that health care equipment, such as (control group) otoscopes, can carry pathogenic bacteria [3]. Our study was designed to determine the incidence of procedure was the packing opened more than half-way bacterial contamination of ETTs after preparation and down the ETT. The 3 physicians did not directly touch the storage in an airway cart in a level 1 trauma center ED. distal portion of the ETT at anytime; this was intended to We used a model similar to the one used with stethoscopes prevent direct hand contact with the portion of the ETT that to determine if the preparation and storage of ETTs in enters the airway. an airway cart lead to colonization with bacteria over The 50 study ETTs were then placed into the airway cart time [2]. in the trauma bay of the ED at Sentara Norfolk General This study was considered exempt from review by the Hospital in the same area as all of the other ETTs used for Eastern Virginia Medical School institutional review board. intubation. The study ETTs were clearly marked bnot for patient useQ on the outside packaging. The airway cart was used in its normal manner (ie, ETTs prepared, removed for 2. Methods use, etc) throughout the study period. From the study group, 1 set of 10 ETTs were removed This study was conducted in the trauma bay of the ED in from the cart and cultured at each of the following time Sentara Norfolk General Hospital from April 13 to April 15, intervals: 6, 12, 24, 36, and 48 hours. Each ETT was cultured 2004. This hospital is a tertiary-care, level 1 trauma center, by rolling the distal cuff portion across the blood agar, and with an annual ED volume of approximately 48000 patients. then a sterile swab was used to sample the distal portion of It is the primary training site for our fully accredited post- the tube and rolled onto the agar. Great care was taken by the graduated year (PGY)-1 through PGY-3 EM residency authors to perform this in a sterile manner. The cultures were program. Airway management for all ED patients is hand-carried to the laboratory and incubated at 378C. All performed by the EM residents and faculty. cultures were examined for the presence of microbes and the A total of 60 new sterile Hi-lo Evac ETTs (Nellcor/Tyco number of colony-forming units (CFUs) for a total of Healthcare), all adult size 7.5 mm, donated by Nellcor/Tyco 48 hours. This was performed by the Sentara Norfolk Gen- Healthcare, were used for the study. These ETTs are the eral Microbiology laboratory personnel. Laboratory person- same type used in our hospital ED, intensive care units, and nel were blinded to which group of ETTs was being cultured. operating rooms. A total of 50 ETTs were used in the study The control group consisted of 10 ETTs that were placed group; this was subdivided into 5 groups of 10 ETTs each. in the airway cart at the same time as the study ETTs The 3 EM physicians participating in this study washed (0 hour), but were not opened. The ETTs in the control their hands before handling all ETTs. The sterile package group were removed at 48 hours and cultured in the same was opened at the adapter end of the ETT, a 10-mL syringe manner as the study ETTs. was placed onto the cuff port and inflated to test cuff The study group and the control group were then integrity. A rigid stylet was then inserted into the ETT compared for the presence of bacteria and the number of lumen. Finally, the provider molded the ETT by adjusting it CFUs. Statistical analysis was performed using the Fisher (while still in its packaging), into the preferred shape, which exact test to determine if there was a statistically significant for our study was the bhockey stick.Q At no time during the difference in the presence of bacteria on ETTs from the study groups versus the control group. The 2 groups were also Table 1 All positive bacterial cultures of ETTs at designated compared regarding the presence of clinically significant time intervals bacteria, defined as those cultures with 5 or more CFUs. A statistically significant difference was considered to be Time of culture (h) present if P b .05. 612243648 No. of positive cultures 03400 (study groups) 3. Results No. of positive cultures 00002 (control group) In the study group, 7 (14%, 95% CI 0.08-0.28) ETTs cultured bacteria; the remaining 43 were negative. Two 550 W.D. Browder et al.
(20%; 95% CI:0.04-0.51) of the 10 control-group ETTs 4. Discussion cultured bacteria. Using the Fisher exact test to compare the 2 groups, there was no statistically significant difference Nosocomial infections are a significant source of ( P = .63). See Table 1. morbidity and mortality in hospitalized patients [6]. If only clinically significant bacteria cultures are consid- Approximately 2 million nosocomial infections occur in ered, there were 3 (6%; 95% CI 0.0.2-0.51) positive cultures the United States each year [7]. Population-based surveil- in the study group and none in the control group. Once lance studies of nosocomial infections in US hospitals again, using the Fisher exact test, there was no statistically indicate a 5% attack rate [8]. For many larger institutions, significant difference between the 2 groups ( P = .57). See the nosocomial infection rate is probably closer to 10% [8]. Table 2. Several studies have examined bacterial contamination of Although there was a trend for an increase in the hospital and medical equipment. Jones et al [2] found that incidence of bacterial contamination of ETTs with increas- the stethoscopes used in ED practice were often contami- ing time in the airway cart, this difference did not reach nated with staphylococci, and therefore a potential vector of statistical significance. infection. Other studies have demonstrated that contaminat- Table 3 lists the bacteria and number of CFUs cultured ed blood pressure cuffs [9], electronic thermometers [5], and for each contaminated ETT. white coats [10] have resulted in hospital endemics [2]. Recently, an unpublished study presented at the annual meeting of the American Society for Microbiology in New Orleans in May 2004 demonstrated that neckties worn by Table 3 Bacteria cultured and number of CFUs for each physicians are a potential carrier of bacterial pathogens. In contaminated ETT this study, the neckties of 42 physicians, medical students, Cultured Cultured Cultured and physician assistants were swabbed and cultured. Nearly at 12 h at 24 h at 48 h 50% of the neckties tested positive for bacteria, including Study ETT no. 12 ETT no. 22 staphylococcus; no methicillin-resistant strains were cul- group Gram (+) cocci Gram (+) cocci tured. Although this study was widely disseminated in the 2 CFUs 1 CFU national press, it has yet to undergo the rigorous process of Gram (+) cocci ETT no. 25 peer review necessary for publication. 2 CFUs Gram (+) cocci Bacterial contamination of ETTs could be a potential ETT no. 16 5 CFUsa cause or contributing factor to ventilator-associated pneu- Gram (+) cocci Gram (+) cocci monia (VAP). VAP is a cause of significant morbidity and a 10 CFUs 1 CFU mortality in critically ill patients [11]. It has been postulated Gram (+) rods Gram (À) rods to start by colonization of the ETT and entry of the 3 CFUs 1 CFU pathogenic bacteria into the lower respiratory tract [12]. The ETT no. 17 ETT no. 28 Gram (+) rods Gram (+) cocci contribution of initial contamination of the airway device to 1 CFU 1 CFU the development of VAP has not been well studied. Gram-variable This study was designed to determine if the common rods practice of preparing ETTs for future use and storing them in 2 CFUs an airway cart in the ED is a potential cause of bacterial ETT no. 30 contamination. Our results show no statistically significant Gram (+) cocci increase in the incidence of bacterial contamination of ETTs a 14 CFUs prepared and stored in an airway cart for up to 48 hours. Gram (+) cocci a Although the study size was limited, we conclude that it is 5 CFUs safe to continue this practice and use ETTs that have been Gram ( ) rods À prepared less than 48 hours prior. 2 CFUs Similar to other studies examining the bacterial contam- Control ETT no. 52 ination of medical equipment, our study has several Group Gram (+) cocci limitations. We cannot prove that simply culturing bacteria 2 CFUs on medical equipment leads to actual infection and disease. Gram (+) rods For our particular study, we did not show that ETTs 1 CFU contaminated with bacteria lead to VAP. It is not an ETT no. 58 unreasonable assumption, however, that such contamination Gram-variable probably plays some role in nosocomial infections. Second, cocci because of the study design, we were not able to make a 1 CFU specific identification of the bacteria cultured from the a Clinically significant bacterial contamination, defined as 5 or ETTs. We could have determined the degree of risk for more CFUs. significant nosocomial infection (ie, methicillin-resistant Open endotracheal tubes on ED airway cart 551
Staphylococcus aureus vs Staphylococcus epidermidis) with References identification of specific bacteria. However, this lack of specific bacterial identification in no way changes the [1] Guinto CH, Bottone EJ, Raffalli JT, Montecalvo MA, Wormser GP. results or conclusion of our study. In addition, there is a Evaluation of dedicated stethoscopes as a potential source of possibility that bacterial contamination occurred at the time nosocomial pathogens. Am J Infect Control 2002;30(8):499-502. [2] Jones JS, Hoerle D, Riekse R. Stethoscopes: a potential vector of of plating the cultures, rather than from the preparation and infection? Ann Emerg Med 1995;26(3):296-9. storage of the ETTs in the ED airway cart. Great care was [3] Cohen HA, et al. Stethoscope and otoscopes—a potential vector of taken, however, by the authors to prevent this occurrence. infection? Fam Pract 1997;14:446-9. Finally, this study was conducted in only 1 ED; it is [4] Nunez S, Moreno A, Green K, Villar J. The stethoscope in the possible that the results would be different in another emergency department: a vector of infection? Epidemiol Infect 2000; 124(2):233-7. hospital ED. It would be interesting to repeat this study [5] Livornese L, Dias S, Samel C, et al. Hospital-acquired infection with using multiple institutions. vancomycin-resistant Enterococcus faecium transmitted by electronic thermometers. Ann Intern Med 1992;117:112-6. [6] Weinstein RA. Nosocomial infection update. Emerg Infect Dis 1998; 5. Conclusion 4(3):416-20. [7] Centers for Disease Control and Prevention. Monitoring hospital- The preparation of ETTs for use in emergent intubation acquired infections to promote patient safety — US 1990-1999. and their storage in an ED airway cart for up to 48 hours in MMWR Morb Mortal Wkly Rep 2000;49:149-53. the ED are not associated with an increased incidence of [8] Wenzel RP, Edmond MB. The impact of hospital-acquired blood- stream infections. Emerg Infect Dis 2001;7(2):174-7. bacterial contamination. [9] Layton MC, Perez M, Heald P, et al. An outbreak of mupirocin- resistant Staphylococcus aureus on a dermatology ward associated with an environmental reservoir. Infect Control Hosp Epidemiol 1993; Acknowledgments 14:369-75. [10] Wong D, Nye K, Hollis P. Microbial flora on doctors’ white coats. The authors thank Tim P. Boyle, RRT, and Robert E. BMJ 1991;303:1602-4. St John, MSN, RN, RRC, of Nellcor/Tyco Healthcare, [11] Rello J, Sonora R, Jubert P, Artigas A, Rue M, Valles J. Pneumonia in Pleasanton, CA for generously donating the ETTs used in intubated patients: role of respiratory airway care. Am J Respir Crit Care Med 1996;154:111-5. our study. They also thank Dr Maria Frontini from Eastern [12] Fleming CA, Balaguera HU, Craven DE. Risk factors for nosocomial Virginia Medical School Epidemiology and Biostatistics pneumonia. Focus on prophylaxis. Med Clin North Am 2001;85(6): Core for her expertise and assistance in the statistical 1545-63. analysis of our data. American Journal of Emergency Medicine (2005) 23, 552–555
www.elsevier.com/locate/ajem
Use of gum elastic bougie for prehospital difficult intubationB
Patricia Jabre MDa, Xavier Combes MDa,*, Bertrand Leroux MDa, Emanuelle Aaron MDa, Harold Auger MDa, Alain Margenet MDa, Gilles Dhonneur MDb aService d’Aide Me´dicale Urgente of Henri Mondor Hospital, Cre´teil 94100, France bService d’Anesthe´sic Re´animation, Hoˆpital Jean Verdier, Bobigny, France
Received 11 August 2004; revised 18 November 2004; accepted 8 December 2004
Abstract The objective of this study was to assess effectiveness of gum elastic bougie (GEB) in case of difficult intubation occurring in the prehospital settings. After manikin training to GEB handling, physicians were recommended to use GEB as first alternative technique in case of difficult intubation. Intubating conditions and details of patients requiring GEB-assisted laryngoscopy were recorded over 30 months. Among the 1442 extrahospital intubations performed, 41 patients (3%) required GEB. Gum elastic bougie allowed successful intubation in 33 cases (78%) and 8 patients sustained a second alternative technique. One patient was never intubated, another 1 required rescue cricothyroidotomy. Twenty-four (60%) GEB patients had associated factors for difficult intubation such as reduced or limited cervical spine mobility, morbid obesity, cervicofacial trauma, and ears, nose, and throat neoplasia. The success rate of GEB was 75% and 94%, respectively, depending on whether associated factors for difficult intubation are present or not. No adverse events associated to GEB use were noted. D 2005 Elsevier Inc. All rights reserved.
1. Introduction a 60-cm-long tracheal tube introducer fabricated from a braided polyester base with a resin coating and a smooth The gum elastic bougie (GEB) is an old tool widely used angled distal tip. During direct laryngoscopy, when glottis is in European countries in case of difficult tracheal intubation not visualized, GEB is passed blindly behind the epiglottis. occurring in the operating room [1-4]. Gum elastic bougie is Progression of the bougie within the trachea is confirmed by clicks felt (due to distal tip of the bougie stumbling over tracheal rings) and distal holdup (bronchus tree) limiting Presented as an abstract at the annual meeting of the Socie´te´ Franc¸aise insertion at 30 to 40 cm from dental arcades. The tracheal d’Anesthe´sie et de Re´animation, Paris, France, September 2003. tube is then railroaded over the GEB while direct Support was provided solely from departmental sources. laryngoscopy is maintained. Several studies performed at B This work was performed in the Service d’Aide Me´dicale Urgente of induction of general anesthesia have reported its effective- Henri-Mondor University Hospital and Paris XII School of Medicine. ness when optimal laryngeal landmarks could not be T Corresponding author. Service d’anesthe´sie re´animation, Hoˆpital Henri-Mondor, 94100 Cre´teil cedex, France. Tel.: +33 01 49 81 21 11; obtained under direct laryngoscopy [2,5,6]. This device fax: +33 01 49 81 23 34. has been exported outside the operating room and has E-mail address: [email protected] (X. Combes). become a cornerstone for difficult airway management in
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.12.005 Gum elastic bougie for prehospital airway management 553 some EDs [7,8]. Recently, we have validated the efficiency extension in the absence of cervical immobilization) and of a simple algorithm to manage unanticipated difficult external laryngeal manipulation (backward, upward, and airway occurring in anesthetized patients using GEB as first right pressure). One intubation attempt under direct laryn- alternative technique in case of difficult laryngoscopy [9]. goscopy was defined by an advancement of the tracheal tube We have demonstrated 80% success rate of GEB. Because toward the glottis followed by its removal from oral cavity. airway management of prehospital emergency patients If 2 successive insertions of GEB failed at trachea resembles that of unanticipated difficult airway anesthetized intubation, the difficult airway management algorithm patients, we have hypothesized that GEB might be of recommended second-step alternative techniques of intuba- similar interest out of the hospital. We have conducted an tion depending on physicians’ preference and experience. A observational study to determine the success rate and safety call for help to a senior anesthesiologist was mandatory in of GEB as first-line alternative technique in case of difficult case of difficult ventilation scenario or if GEB-assisted intubation occurring in prehospital intensive care medicine. intubation failed. Before starting the study, all emergency care providers of the SMUR received an oral formation on the study’s design, 2. Methods query form completion, and the algorithm of airway management. Cormack and Leane laryngeal view classifi- This monocenter study was performed in the suburbs of cation drawing was photocopied on each data sheet (grade Paris (Val de Marne, population 1300000) by our local 1, whole larynx visible; grade 2, partial view of vocal cords; prehospital emergency medical unit (SMUR) between grade 3, only epiglottis visible; grade 4, no part of larynx February 2001 and July 2003, according to French ethic visible). All anesthetized (rapid sequence) or cardiac arrest laws. Our SMUR is equipped with 5 mobile intensive care patients in the care of our SMUR who requested rescue units (MICUs) and its annual activity is about 10000 medical intubation and for whom tracheal access failed after 2 direct emergency out-of-hospital interventions. laryngoscopy attempts were included with the query form A MICU unit is composed of a vehicle driver, a nurse available in all vehicles systematically completed immedi- anesthesiologist, and either a senior emergency physician ately after the airway management process. (N90% of cases) or a senior anesthesiologist. In France, Circumstances and clinical conditions of the airway extrahospital tracheal intubation is always performed by or management, characteristics of difficult laryngoscopy under the control of a physician. patients, intubation difficulty score, and details of the Forty-five intubators (23 emergency physicians, 5 anes- intubating process were recorded (Table 1). The success rate thesiologists, and 17 nurses specialized in anesthesia) were of GEB was calculated [10]. involved in this study. Before joining the MICU team, spe- Airway care providers had to document the following cialized nurses were requested to have an experience of parameters for all patients: estimated or known height and more than 4 years of intubation in operating room weight, history of ears, nose, and throat (ENT) disease, conditions and emergency physicians had at least 3 years objective cervical mobility, cervical immobilization, and of prehospital and inhospital emergency medicine experi- history of maxillofacial disease. ence (with a mean of 30 intubations per year). During the The relative position of the intubator and the patient and study period, a senior anesthesiologist was always on call in occurrence of complications, such as macroscopic inhala- case of severe airway management difficulties encountered by the emergency physicians. This senior anesthesiologist was motorized to rapidly assist airway management on the scene of difficult intubation. Table 1 Intubation difficulty scale from Adnet et al [10] Before introducing GEB in the prehospital emergency Parameter Score medicine, the medical team of our SMUR adhered to an No. of attempt N1N1 intense theoretical and practical formation (training manikin, No. of operators N1N2 Laerdal) with GEB handling. After completion of this initial No. of alternative techniques N3 formation, GEB (Eschmann tracheal tube introducer; Sims Cormack grade 1 N4 Portex, Hythe, UK) was available in all medical vehicles of Lifting force required the SMUR. The standard airway equipment on vehicles com- Normal N5 =0 prises laryngoscope with Macintosh blades (cold light) of Increased N5 =1 different sizes, a GEB, a retrograde intubation kit, a cricothy- Laryngeal pressure rotomy kit, and an intubating laryngeal mask airway (ILMA). Not applied N6 =0 Applied N =1 As integral part of our local prehospital difficult airway 6 Vocal cord mobility management, GEB was recommended as the first-line Abduction N7 =0 alternative technique in case of difficult intubation defined Adduction N7 =1 by a number of failed tracheal access attempts under direct IDS = sum of scores N1-N7 laryngoscopy (N2) with optimal head position (head 554 P. Jabre et al. tion, dental trauma, arterial oxygen desaturation, and The success rate of GEB was 75% and 94%, respectively, hemodynamic instability, were recorded. depending on whether AFDI are present or not. Most failed GEB-assisted intubation occurred in patients suffering from ENT neoplasia, with 95% GEB success rate in other patients 3. Results with AFDI. All but 1 difficult laryngoscopy patients free from AFDI (n = 17) were intubated with GEB demonstrat- During the study period, 1442 patients required intuba- ing a 95% (16/17) success rate. Clinical conditions, tion. Six hundred forty were cardiac arrest patients and characteristics of difficult laryngoscopy patients, and details 802 had a spontaneous cardiac activity the moment of airway of the intubating process are reported in Table 2. No specific management. Of the 802 patients with persistent cardiac complication associated with GEB was noted. activity, 771 (95%) received succinylcholine. Among the 1442 emergency cases, 186 were trauma patients, the remaining (1266) were medical patients. 4. Discussion Forty-one (3%) patients required GEB-assisted tracheal intubation challenge. Difficulties at laryngoscopy were We observed about 3% difficult intubation requiring encountered by 28 of 45 potential intubator of our SMUR more than 2 laryngoscopic attempts in our emergency out- emergency team. Three airway care providers were of-hospital patients. We demonstrated the effectiveness and concerned 3 times, 7 twice, and 18 once. Median (SD) safety of the GEB proposed as the first-line alternative intubation difficulty score and Cormack grade for these technique in difficult laryngoscopy patients; 80% of these 41 patients were 8 (2.2) and 4 (0.2), respectively. Twenty- patients were rapidly intubated with this device. eight patients (68% of GEB patients and 2% of all patients) We have chosen GEB to facilitate tracheal intubation of were classified as Cormack grade 4. Gum elastic bougie emergency difficult airway patients because of its simplicity allowed rapid successful intubation in 33 cases (78%); 24 at of use, convenience, efficiency, and low cost. In addition, we the first attempt, and 9 at the second. Rescue oxygenation have observed on manikin a very short learning process of was mandatory for 2 difficult mask ventilation patients after this pocket-sized ready-to-use airway device. Although failed GEB challenge. Cricothyroidotomy was performed in popular in the United States, we preferred to challenge 1 patient because of an allergic laryngeal angioedema and the GEB rather than the stylet because it was demonstrated in resuscitation process was conducted while the second patient hospital studies to be systematically more efficient at was ventilated then intubated through the ILMA. assisting intubation [5,11]. Moreover, a recent report has A potential difficult airway management was anticipated demonstrated more than 95% success rate of tracheal in 24 (60%) of GEB patients that demonstrated associated intubation associated with the use of the GEB in case of factors for difficult intubation (AFDI) such as reduce unpredicted difficult laryngoscopy, which contrasts with our cervical spine mobility (n = 7), morbid obesity (n = 7), results [12]. ENT neoplasia (n = 7), cervicofacial trauma (n = 2), and Several factors can explain the lower efficiency of GEB to allergic angioedema (n = 1). Seven patients with failed GEB facilitate tracheal intubation in our emergency out-of-hospital that were intubated with a second-step alternative technique difficult laryngoscopy patients. First, most of operating room (ILMA [n = 1], retrograde intubation [n = 1], blind nasal studies that have demonstrated the interest of GEB included intubation [n = 1], or by the rescue senior anesthesiologist selected patients. Anticipated difficult airway and emergency under direct laryngoscopy [n = 4]) were among the patients were excluded from these trials. None of the patients 24 patients with AFDI. included in these studies had factors that were shown to increase the risk of difficult intubation such as 60% (24/41) of the difficult laryngoscopy patients that we have managed. Table 2 Characteristics of patients with or without AFDI Interestingly, we demonstrated a 94% (16/17) success rate of Patients with Patient without the bougie in difficult laryngoscopy patients that were free evident AFDI AFDI (n = 17) (n = 17) from AFDI. All but 1 (7/8) bougie failure was (n = 24) observed which occurred in difficult laryngoscopy patients H/F 17/7 12/5 with AFDI. In the 24 difficult laryngoscopy patients with Age 52 F 15 59 F 13 AFDI, we demonstrated a 75% success rate of bougie. Cormack 2/3/4 3/4/17 1/6/10 Surprisingly, both groups of difficult laryngoscopy patients, F F IDS 8 282 with (n = 24) or without (n = 17) AFDI, were similar in mean Failure GEB 7 1 Cormack score (3.5 and 3.6, respectively), suggesting that AFDI AFDI are determining factors for more difficult GEB-assisted Morbid obesity 7 Reduced cervical mobility 7 intubation. Among difficult laryngoscopy patients with Upper airway distortion 8 AFDI, most failed bougie-assisted intubation (3/7) occurred Maxillofacial trauma 2 in patients suffering from ENT neoplasia, with 95% (16/17) success rate of bougie in morbidly obese patients and those Gum elastic bougie for prehospital airway management 555 with reduced cervical spine mobility and cervicofacial as the first strategy in case of difficult intubation because trauma. Our results confirm a recent prospective study this simple device is very easy to use. Although efficiency conducted in similar conditions. Gum elastic bougie effi- of GEB has not been evaluated in non medical personal, our ciency was challenged in elective anesthetized patients as a educational program of medical student and paramedics has first alternative technique in case of unexpected difficult showed us that the learning curve of GEB was extremely laryngoscopy. The authors demonstrated that GEB allowed short on the dummy and clinical acquisition of the technique rapid tracheal intubation in 80% of difficult laryngoscopy of GEB-assisted tracheal intubation was immediate. Be- patients [9]. We believe that unexpected difficult airway cause of these reasons we recommend GEB handling to be management of elective anesthetized patients is somehow taught to any care provider performing laryngoscopy. comparable to that of prehospital anesthetized or cardiac In summary, our prehospital study outlines the major arrest patients. interest and safety of GEB to assist tracheal intubation of The second factor that may have influenced overall GEB emergency patients. We have demonstrated that this simple success rate is related to environmental intubation circum- device allowed rapid intubation of nearly 80% of prehospi- stances. Indeed the interaction between the patient and the tal patients with difficult direct laryngoscopy. operator body positions at intubation seems to be of particular importance. The lying position of the patient on the ground, such as we experienced in almost half of our References difficult laryngoscopy patients, was demonstrated to be an independent risk factor for prehospital difficult intubation [1] Dogra S, Falconer R, Latto IP. Successful difficult intubation. [13]. This specific interplay between the operator and the Tracheal tube placement over a gum-elastic bougie. Anaesthesia patient, directly linked to prehospital emergency medicine, 1990;45:774-6. [2] Kidd JF, Dyson A, Latto IP. Successful difficult intubation. Use of the might have induced additional difficulties at laryngoscopy. gum elastic bougie. Anaesthesia 1988;43:437-8. The third factor that could explain a lower success rate than [3] Nolan JP, Wilson ME. Orotracheal intubation in patients with potential previously reported with GEB may be the high incidence of cervical spine injuries. An indication for the gum elastic bougie. Cormack grade 4 reported in our study. Gum elastic bougie Anaesthesia 1993;48:630-3. is classically attended to facilitate intubation in difficult [4] Nolan JP, Wilson ME. An evaluation of the gum elastic bougie. Intubation times and incidence of sore throat. Anaesthesia 1992;47: laryngoscopy Cormack grade 2 and 3 patients. In our 878-81. prehospital population, 27 of 41 patients were classified as [5] Gataure PS, Vaughan RS, Latto IP. Simulated difficult intubation. Cormack grade 4 at the first laryngoscopy, and blind (with Comparison of the gum elastic bougie and the stylet. Anaesthesia 1996; regard to laryngeal landmarks) tracheal intubation was 51:935-8. attempted. Finally, the last factor that has probably [6] Koay CK. Difficult tracheal intubation—analysis and management in 37 cases. Singapore Med J 1998;39:112-4. influenced our results concerns the experience of the [7] Morton T, Brady S, Clancy M. Difficult airway equipment in English physicians that performed initial laryngoscopy. In the emergency departments. Anaesthesia 2000;55:485-8. present study, most of the physicians in charge of the [8] Moscati R, Jehle D, Christiansen G, D’Aprix T, Radford J, Connery intubation were emergency physicians whose experience in C, et al. Endotracheal tube introducer for failed intubations: a variant difficult airway management is certainly less accurate than of the gum elastic bougie. Ann Emerg Med 2000;36:52-6. [9] Combes X, Suen P, Dumerat M, Duvaldestin P, Dhonneur G. Validation that of experienced anesthesiologist. Surprisingly, when a of an intubation algorithm for unanticipated difficult tracheal intubation rescue senior anesthesiologist was requested (4/41), its occurring in operating room. Anesthesiology 2004;100:1146-50. evaluation of the Cormack score was systematically lowered [10] Adnet F, Borron SW, Racine SX, Clemessy JL, Fournier JL, Plaisance P, by 1 grade as compared with that scored by the emergency et al. The intubation difficulty scale (IDS): proposal and evaluation of a physician. This observation may explain 4 difficult laryn- new score characterizing the complexity of endotracheal intubation. Anesthesiology 1997;87:1290-7. goscopy patients which had experienced failed GEB [11] Noguchi T, Koga K, Shiga Y, Shigematsu A. The gum elastic bougie attempts that were finally intubated under standard laryn- eases tracheal intubation while applying cricoid pressure compared to a goscopy by the senior rescue anesthetist. The results and stylet. Can J Anaesth 2003;50:712-7. success rate we report may not be completely exportable to [12] Latto IP, Stacey M, Mecklenburgh J, Vaughan RS. Survey of the use other system of prehospital care using less educated care of the gum elastic bougie in clinical practice. Anaesthesia 2002;57: 379-84. providers. We are aware that specialized physician aboard [13] Adnet F, Cydulka RK, Lapandry C. Emergency tracheal intubation of emergency ambulance is relatively uncommon out of patients lying supine on the ground: influence of operator body European countries. However, we recommend using GEB position. Can J Anaesth 1998;45:266-9. American Journal of Emergency Medicine (2005) 23, 556–560
www.elsevier.com/locate/ajem
Controversies Emergency medicine and older adults: continuing challenges and opportunitiesB
John G. Schumacher MA, PhD*
Department of Sociology and Anthropology, University of Maryland, Baltimore County, Baltimore, MD 21250, USA
Received 11 December 2004; accepted 22 December 2004
1. Introduction nationwide [4-6]. At the same time, the size of the older ED population represents just one dimension of the issue, The treatment of older adults in the emergency depart- because we also know that older adults (65 years and older) ment (ED) has been a part of emergency medicine (EM) represent a highly heterogeneous population in terms of since its inception. Nationwide, older adults represent an health status, function, and presenting diagnosis. important and growing patient population representing about 18% of ED admissions annually [1]. Although geriatric EM has attracted limited clinical and research attention, there continues to be significant challenges, needs, 2. Initial efforts in geriatric emergency as well as opportunities in pursuit of providing high-quality medicine care to older adults. Clinical and scholarly attention to geriatric EM has been Over the next 25 years, the US population of older adults sporadic during EM’s history. The most recent effort was a (65 years and older) will double from 34 million in 2000 to 1992 collaboration between the John Hartford Foundation more than 69 million by the year 2030 [2]. Because these and the Society for Academic Emergency Medicine individuals have already been born and are aging, there is (SAEM). It was designed to develop a comprehensive little doubt in the veracity of these projections. This educational program for training EM residents and physi- dramatic aging of the US population will significantly cians. Led by Sanders, the group published a series of affect a range of social institutions from education to health journal articles and a textbook and developed an educational care including hospital EDs. Currently, EDs nationally training manual distributed by SAEM [7-9]. In addition, a report an estimated 18.5 million older ED admissions model of geriatric EM was proposed to guide the clinical annually [1,3]. Despite the tremendous projected growth approach to treating older adults in the ED [7]. of older adults, there are very limited, if any, projections Findings from Sanders and colleagues documented specifically regarding the numbers of older patients EDs can higher ED use rates among older adults compared with anticipate serving. Nonetheless, it is reasonable to assume younger patients [7-9]. Clinically, older adults were also any increases in the older adults will likely exacerbate the shown to present with: nonstandard disease presentations, current overcrowding conditions experienced by many ED altered laboratory values, multiple comorbid diseases, extensive medical histories, communication problems, and potentially altered mental status [7]. In some cases, classic B This research was supported by a University of Maryland, Baltimore County, Faculty Research Fellowship, 2003. symptoms for diagnosis were absent in the older ED patient T Tel.: +1 410 455 3184; fax: +1 410 455 1154. [10,11]. In addition, 30% to 50% of older adults ED patients E-mail address: [email protected]. were hospitalized, in sharp contrast to just a 12%
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.12.011 Emergency medicine and older adults 557 hospitalization rate for younger patients [7]. These issues management. Fourth, cognitively impaired older adults can make health-care delivery to this population much more may present to the ED for evaluation solely because of challenging and time consuming. The lack of training and the unavailability, inadequacy, or failure of other support educational programs in geriatric EM was documented as services, which inappropriately ties up resources [16,17]. part of this project. These challenges are illustrated in the case of a 79-year-old The work of Sanders [7] provided a solid foundation for female patient transferred to a large county ED from a local the development of geriatric EM. Unfortunately, since this nursing home for evaluation of increased temperature and work of more than a decade ago, there has been only modest increasing lower extremity edema. She had a history of evidence of geriatric EM research or new scholarly moderate dementia and was not oriented to person or place. literature. Aminzadeh and Dalziel [12] recently provided a She was unaccompanied and transported via private ambu- useful literature review on older adults in the ED. Their lance. In the ED, she resisted lying down and continuously review reiterates the distinct pattern of service use and care attempted to get down from the cart in an attempt to exit the needs required by older adults. They also note the area. After a few minutes of persistent struggle and inadequacy of current approaches to serving this population combativeness with several ED nurses, the patient was and call for more research and intervention projects targeting loosely tied to the cart with a bedsheet across her chest. In at-risk older adults. addition, the cart was strategically placed near the nursing Overall, it appears that clinical and scholarly work in station so she could be closely observed. Within minutes she geriatric EM appears to have lost momentum since the used a bHoudiniQ maneuver to release herself and climbed efforts of Sanders [7-9]. The early work characterized the down from the cart. The staff returned her to the cart; population and provided a model, however, the pace of however, she immediately set about the task of struggling to work has diminished despite rapidly increasing older adults free herself, oblivious to the pleadings of the staff. Again, she ED populations. The research has been dominated by freed herself and the frustrated staff applied soft Velcro descriptive studies of use rates and issues such as acuity restraints. Her examination continued, with some difficulty, levels, repeat visits, and 6-month mortality rates [4,12]. as she was combative and uncooperative. Her care required Research considering the impact that the older adult additional staff support during every phase of the examina- population has on the practice of EM is much more limited tion. Eventually, the patient was diagnosed with an acute in scope [13]. urinary tract infection and was transferred to a medical ward for continued therapy. Appropriate management of cognitive impairment in the 3. Continuing and emerging challenges ED requires substantial staff time, training, resources, and commitment. Despite the increasing prevalence of cognitive Despite the absence of sustained scholarly attention impairment, it is widely recognized that the training, regarding older patients, emergency physicians (EPs) con- policies, and resources for handling this population remain tinue to face challenges in providing care to this population. inadequate [17]. The realities of clinical practice and the existing literature on older adults suggest at least 4 areas that would benefit from 3.2. Assisted-living facilities and emergency renewed attention by EM professionals. departments 3.1. Managing cognitive impairment in the Assisted-living facilities (ALFs) are a residential envi- emergency department ronment broadly guided by a social model of care designed Currently, 4.5 million older adult are cognitively to maximize the independence of older adults by providing impaired, and estimates suggest that by 2050, these numbers assistance and services to maintain an individual’s ability to could be as high as 16 million [14]. Cognitively impaired age-in-place. Although regulations vary by state, in general, older adults are a specific patient population that will these facilities generally house 3 or more unrelated adults continue to demand new and innovative patient care (although some purpose-built facilities have over 200 units); strategies from EM on several levels. First, EPs need the provide shelter (room), food (board), and 24-hour supervi- skills and instruments to accurately diagnose and screen this sion/protective oversight and personal care services; and can population. Recent findings suggest EPs fail to identify respond to unscheduled needs for assistance. There are as nearly 50% of older patients with cognitive impairment many as 35000 AL facilities nationwide [18], and the total [15]. Second, cognitively impaired patients are a heteroge- number of residents is as high as 1.5 million. ALFs are neous population presenting with a wide range of diagnoses, considered an intermediate between home and nursing home impairment levels, comorbidities, and medical histories. care [19-22]. Third, cognitively impaired patients in the ED also exhibit The rapid expansion and saturation by the ALF industry a range of challenging behavior patterns including non- presents a number of challenges to EDs. First, ALFs responsiveness, combativeness, and/or wandering. Such typically do not employ trained medical staff and conse- characteristics complicate rapid diagnosis and patient quently rely on community physicians and EDs for medical 558 J.G. Schumacher care. Second, older adults typically move into an ALF that older adults required more time and resources than other because of some health, functional, or cognitive decline that patients and that their training in geriatric EM was not compromises their ability to live independently and sufficient [8]. EPs have also been shown to highly increases their likelihood of ED use. Third, residents of overestimate the percentage of older patients they treat in ALFs may lack family involvement or other caregivers to the ED. In one study, EPs estimated the mean percent- provide health histories and support, which may make age of older adults they treated was near 40%, which emergency care more challenging. On a structural level, the is more than double the actual national average of 18% presence of multiple large ALFs (N100 units) in a (J Schumacher, unpublished manuscript, 2004). Such catchment area has the potential to overwhelm the resources overestimation may indicate feelings of burden, particularly, of an ED particularly during a public health epidemic if these patients also are considered to be unsatisfying and/ (eg, influenza outbreak). or time-consuming cases. Despite a decade-old call for In response to the rise of ALFs as a living environment for research into factors associated with occupational stress older adults, EDs may consider taking proactive steps to among EPs, few studies examine this issue and none has address some of the challenges ALF present. EDs may want focused on the impact of older adult patients [25]. to assess the number and type of ALFs and nursing homes that fall into their catchment area. Based on this evaluation, 3.4. Geriatric emergency medicine education EDs may encourage large ALFs to develop primary medical and training coverage for their residents to avoid overreliance on area The EM training curriculum is overfilled and consistently EDs. In addition, EDs could encourage ALFs to use challenged to incorporate new and emerging information. consistent medical records in the event their residents are Geriatric EM content represents another curriculum pres- transferred to an ED. An ED may also wish to initiate a sure. To date, it has not been a formal part of the residency dialogue with area ALFs to educate them about the training with core competencies or explicit skills testing. appropriate use of EDs and the services that can and cannot However, with the projected demographic transition fueled be offered. by Baby Boomers, geriatric instruction may rapidly need to 3.3. Emergency physician satisfaction treating become a core part of training. In a recent survey, EM residents reported lower levels of confidence treating older older patients adults (N65 years) compared with adults (age 18-64 years) Examining EPs attitudes about their patients is important (J. Schumacher, unpublished manuscript, 2004). One because patient encounters have been identified as a key quarter of these residents indicated that their training source of professional satisfaction and dissatisfaction across experience bavoidedQ instruction in geriatric EM. At the specialties [23]. EP practice dissatisfaction has been same time, three quarters agreed that more training in associated with abusive and demanding patients, anger on geriatric EM is necessary to provide better care to older the part of patients or relatives, and difficult or violent adults. These results suggest that efforts by professional patients [24]. Considering these dimensions of EP satisfac- organizations to develop and disseminate educational tion and dissatisfaction, older adult patients admitted from programs in geriatric medicine are warranted. long-term care settings or presenting with cognitive impairment overlays are at risk of being considered particularly unsatisfying patients. Research comparing 4. Innovations and future directions of satisfaction between EPs and internists reported EPs were geriatric emergency medicine less satisfied than internists regarding current and preferred conditions of their practice. A source of this gap was Several challenges have been described above regarding suggested to be the discontinuity for EPs between being older adults and EM. In response to these challenges, the trained to treat urgent, emergently ill patients while following subsections suggest 3 thematic areas to renew frequently treating patients with nonurgent problems [25]. attention to issues of older adults in the ED. Issues of practice satisfaction/dissatisfaction among EPs are 4.1. Emergency department case finding, important because low levels of satisfaction are predictive of assessment, and referral projects decisions to leave the field [24]. In addition, leaving the field is also associated with work-related stress and burden. In response to the complexity of older adult cases EM has been described as a high-stress work environ- presenting to the ED, several innovative case finding, ment due to its unpredictability and the need for rapid action assessment, and community referral projects have been and decision making [25]. Increasing numbers of encounters implemented. The case finding and assessment projects are with older ED patients has the potential to add to physician designed to screen older patients to identify at-risk older stress because these patients may be more time-consuming, adults who would benefit from comprehensive geriatric perceived as less satisfying, and EPs report less confidence assessment by a geriatric nurse specialist [26-28]. Early in their treatment [8]. Research indicates that EPs agreed identification of these at-risk patients also facilitates their Emergency medicine and older adults 559 management because appropriate ED resources can be SAEM) and national health agencies and organizations directed earlier in the admission. Some programs extend the (eg, Agency for Healthcare Research and Quality, the screening function into the prehospital setting by training National Institute on Aging) is needed to promote geriatric paramedics in intermediate care and assessment of older EM as an educational and research priority. Despite the patients [29]. These programs are frequently coordinated growing older population, a lack of geriatric EM educational with a referral network comprised of a coalition of opportunities persists, particularly on the national level community-based service providers. Evaluations of the (eg, few national conferences). In addition, funded research programs find high levels of support from staff; however, opportunities in the area have been limited or nonexistent. improvements in reported outcomes such as repeat ED The professional EM societies need to provide critical visits, hospitalizations, nursing home admissions, and costs leadership and advocacy in establishing priorities, convening have been modest [30]. Such programs are one way to workgroups, developing position statements, setting and begin to address the challenges of cognitive impairment, funding research goals, and fostering the growth of the area ALF, and EP provider burden by providing EPs with of geriatric EM. information and resources to improve their management of this patient population. 5. Conclusions 4.2. Evaluating the structure and process of emergency department care Older adults are a heterogeneous and rapidly growing population in the ED. Responding to their unique care needs Responsive ED management requires continuous evalu- requires a sustained, focused effort by the EM profession. ation of their structure and processes of care to appropriately Consistent clinical and scholarly themes suggest the need accommodate the needs of older patients. Structurally, the for innovation, increased geriatric education, and systematic typical design/layout of EDs with many open bays, curtain research about older adults. The challenges facing EM dividers, and extraneous noise may interfere with the related to older adults will continue building particularly in clinical examination of older patients with hearing loss. the areas of cognitive impairment and ALFs. EM has the Tracking these patients to quieter examination rooms may opportunity to prepare and respond to these challenges if it be good clinical practice as well as a priority in future ED takes the initiative to address the issues associated with architectural design and renovations. ED observation units providing quality care to older adults in the ED. or 23-hour beds represent another structural resource in the appropriate clinical evaluation of older patients [31]. Processes of care issues in the ED include evaluating References patient flow and supportive services. Older adults waiting for hours, without food/fluids, and sitting in hard chairs are poor [1] American Hospital Association. Hospital Statistics. Chicago (Ill)7 processes that may lead to the patients’ diminished physi- American Hospital Association; 2003. ological capacity. Cognitively impaired older adults may [2] US Bureau of the Census. Current population survey. Washington 7 require innovative support services and treatment protocols (DC) US Government Printing Office; 2003. [3] McCaig L, Burt C. National Hospital Ambulatory Medical Care that accommodate their needs. Some EDs have partnered Survey: 2002 Emergency Department summary. Adv Data with local Alzheimer disease associations to develop 2004;340:1-34. collaborative arrangements providing staffing or volunteer [4] Strange G, Chen E. Use of emergency departments by elder patients: support for cognitively impaired patients. In addition, EDs a five-year follow-up study. Acad Emerg Med 1998;5:1157-62. could work more closely with nursing facilities to develop [5] Grumbach K, Keane D, Bindman A. Primary care and public emergency department overcrowding. Am J Public Health 1993; procedures for handling cognitively impaired patients. Many 83:372-8. nursing homes have developed bspecial care unitsQ specif- [6] Derlet R, Richards J, Kravitz R. Frequent overcrowding in U.S. ically designed to accommodate cognitively impaired emergency departments. Acad Emerg Med 2001;8:151-5. patients. The techniques and treatment protocols developed [7] Sanders AB. Emergency care of the elder person. St Louis (Mo)7 for use in these special care units may be successfully Beverly Cracom; 1996. p. 1-21. [8] McNamara R, Rousseau E, Sanders A. Geriatric emergency medicine: applied in the ED setting. One thing remains clear—the a survey of practicing emergency physicians. Ann Emerg Med population of cognitively impaired older adults presenting to 1992;21:796-801. EDs will increase in the future. An organized response by [9] Sanders AB. Care of the elderly in emergency departments: EM would benefit the patients, families, and ED staff. conclusions and recommendations. Ann Emerg Med 1992;21:830-4. [10] Eliastam M. Elderly patients in the emergency department. Ann 4.3. Focusing national efforts regarding geriatric Emerg Med 1989;18:1222-9. [11] Kamin R, Nowicki T. Pearls and pitfalls in the emergency department emergency medicine evaluation of abdominal pain. Emerg Med Clin North Am 2003;21:61-72. Educating EPs at all training levels about geriatric EM [12] Aminzadeh F, Dalziel W. Older adults in the emergency department: remains a critical need. Leadership by professional organ- a systematic review of patterns of use, adverse outcomes, and izations (eg, American College of Emergency Physicians, effectiveness of interventions. Ann Emerg Med 2002;39:238-47. 560 J.G. Schumacher
[13] McDonald A, Abrahams ST. Social emergencies in the elderly. Emerg [22] National Center for Assisted Living. Fact and trends: the assisted Med Clin North Am 1990;8:443-59. living source book. Washington (DC)7 National Center for Assisted [14] Herbert L, Scherr P, Bienias J, et al. Alzheimer disease in the U.S. Living; 1998. population: prevalence estimates using the 2000 census. Arch Neurol [23] Richardson A, Burke R. Occupational stress and job satisfaction 2003;60:1119-22. among physicians: sex differences. Soc Sci Med 1991;33:1179-87. [15] Hustey F, Meldon S, Smith M, et al. The effect of mental status [24] Doan-Wiggins L, Zun L, Cooper M, et al. Practice satisfaction, screening on the care of elderly emergency department patients. Ann occupational stress, and attrition of emergency room physicians. Acad Emerg Med 2003;41:678-84. Emerg Med 1995;2:556-63. [16] Cole M, Dendukuri N, McCusker J, et al. An empirical study of [25] Murphy J, Jacobson S. Satisfaction with practices: emergency different diagnostic criteria for delirium among elderly medical physicians versus internists. Ann Emerg Med 1987;16:277-83. inpatients. J Neuropsychiatry Clin Neurosci 2003;15:200-7. [26] Mion LC, Palmer RM, Anetzberger GJ, et al. Establishing a case- [17] Chiovenda P, Vincentelli G, Alegiani F. Cognitive impairment in finding and referral system for at-risk older individuals in the elderly ED patients: need for multidimensional assessment for better emergency department setting: the Signet model. J Am Geriatr Soc management after discharge. Am J Emerg Med 2002;20:332-5. 2001;49:1379-86. [18] Assisted Living Work Group. A report to the U.S. Senate Special [27] Mion L, Palmer R, Meldon S, et al. Case finding and referral model Committee on Aging: assuring quality in assisted living guidelines for for emergency department elders: a randomized clinical trial. Ann federal and state policy, state regulations, and operations. Washington Emerg Med 2003;41:57-68. (DC)7 AAHSA; 2003. [28] Basic D, Conforti D. Standardised assessment of older patients by a [19] Kane R, Wilson KB, Clemmer E. Assisted living in the U.S.: a new nurse in an emergency department. Aust Health Rev 2002;25:50-8. paradigm for residential care for frail elders? Washington (DC)7 Public [29] Mason S, Wardrope J. Developing a community paramedic practi- Policy Institute, AARP; 1993. tioner intermediate care support scheme for older people with minor [20] Mollica R. State policy and regulations. In: Zimmerman S, Sloane P, conditions. Emerg Med J 2003;20:196-8. Eckert K, editors. Assisted living and residential care in transition. [30] Moss JE, Flower CL, Houghton LM. A multidisciplinary care Baltimore (Md)7 Johns Hopkins University Press; 2002. p. 9-33. coordination team improves emergency department discharge plan- [21] General Accounting Office, United States Congress. Assisted living: ning practice. Med J Aust 2002;177:435-9. quality of care and consumer protection issues in four states. Report to [31] Ross M, Compton S. The use and effectiveness of an emergency Congressional Requesters. Washington (DC)7 General Accounting department observation unit for elderly patients. Ann Emerg Med Office; 1999. 2003;41:668-77. American Journal of Emergency Medicine (2005) 23, 561–562
www.elsevier.com/locate/ajem
Case Reports Chronic digoxin toxicity and significantly elevated BNP levels in the presence of mild heart failure
Kenneth Heinrich MDa,*, Heather M. Prendergast MDb, Timothy Erickson MDb aEmergency Medicine Residency Program, University of Illinois Medical Center, Chicago, IL 60612, USA bDepartment of Emergency Medicine, University of Illinois Medical Center, Chicago, IL 60612, USA
Received 15 October 2004; accepted 15 October 2004
We present a case of patient with chronic digoxin toxicity rubs or gallops. The remainder of the examination was with minimal evidence of heart failure, but a significantly unremarkable except for mild bilateral lower extremity elevated B-type natriuretic peptide (BNP) levels. pitting edema. Peripheral pulses were normal. A 78-year-old woman was brought to an ED by her Diagnostic studies included an electrocardiogram, which daughter with complaints of nausea, increasing dyspnea on demonstrated a sinus bradycardia with a left bundle-branch exertion, and lower extremity edema for approximately block and no other ST-T wave changes. Portable chest 2 weeks. The patient denied chest pain, fevers, chills, radiograph showed cardiomegaly, surgical clips, and no vomiting, diarrhea, or abdominal pain. frank infiltrates. Initial laboratory data included hyper- The patient’s medical history was significant for moder- kalemia with a potassium of 7.0 mmol/L, elevated blood ate to severe congestive heart failure, coronary artery urea nitrogen/creatinine over baseline at 105/2.5 mg/dL, a disease status after bypass graft in 1994, diabetes mellitus, metabolic acidosis with respiratory compensation, an unre- hypertension, hypothyroidism, dementia, depression, dys- markable complete blood count, normal cardiac enzymes, pepsia, and osteoarthritis. The patient was taking a and an elevated digoxin level at 4.2 ng/mL. The BNP level significant number of medications including 0.125 mg of was reported as greater than 1300 pg/mL (the upper limit of digoxin daily, with which she was reportedly compliant. the assay at the time). Physical examination revealed a thin elderly woman in Given the apparent chronic digoxin toxicity from which moderate distress. Vital signs were significant for a the patient was symptomatic, the decision was made in con- bradycardia with a rate of 50 beats per minute and tachypnea sultation with a cardiologist to administer Digibind. Shortly with a rate of 30 breaths per minute. The patient was afebrile, after administration, the patient experienced a brief run of had a normal blood pressure, and normal room air oxygen ventricular tachycardia which was spontaneously terminated saturation by pulse oximetry. Head, eyes, ears, neck, and before any antiarrhythmic agents could be given.The patient throat examination was unremarkable except for mild jugular was then transferred to the cardiac intensive care unit. venous distension. Lung examination revealed mild crackles The patient was started on intravenous (IV) rehydration at the bases, tachypnea without retractions, and good air in the cardiac intensive care unit for her acute renal exchange. Cardiac examination revealed a bradycardia, insufficiency and received treatment for hyperkalemia. which was regular, and a systolic murmur. There were no The patient proceeded to have a complicated hospital course demonstrating signs of fluid overload despite later attempts at diuresis. She developed multiple right upper extremity T Corresponding author. deep venous thromboses secondary to an indwelling
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.10.009 562 K. Heinrich et al. peripherally inserted central catheter line. A transthoracic 5c-monophosphate) levels in CHF patients after therapeu- echocardiogram showed severely depressed left ventricular tic IV followed by oral administration of digoxin [2]. The function, increased left ventricular end-diastolic pressure, population for this study consisted of 25 patients, and moderate pulmonary hypertension, and elevated right atrial again, the authors demonstrated a statistically significant pressure. This represented a worsening of findings when com- increase in BNP levels in digoxin patients versus placebo pared with a transthoracic echocardiogram done 7 months (130 vs 227 both 3 hours after IV and 3 days after oral prior, which showed moderate to severe left ventricular administration). dysfunction. Because of the extremely poor prognosis, As the role of BNP as a marker for congestive heart end-of-life issues were discussed, and she elected to sign a failure continues to evolve, it will become increasingly bdo not resuscitateQ/bdo not intubateQ order. The patient died important to identify factors which could affect an accurate 10 days after this admission. interpretation [3]. This case highlights a possible confound- As the echocardiogram was not performed at the time of ing variable of digitalis toxicity when interpreting the results the initial ED evaluation, it is unclear whether objective of BNP levels during the evaluation of patient with signs findings consistent with worsening congestive heart failure and symptoms suggestive of a heart failure exacerbation. were present on arrival. Givenpreliminaryresearchsuggestinganindependent A review of the medical literature reveals no previous effect of digitalis administration on serum BNP concentra- case reports of digitalis toxicity (acute or chronic) and tion, a future study examining BNP levels in patients elevated BNP levels. However, there have been a few presenting with digitalis toxicity would be useful. Caution studies demonstrating a direct effect of digitalis on BNP. should be exercised in interpretation of BNP levels in One such project was completed at the Shiga University of patients with heart failure on digoxin. Medical Science in Otsu, Japan [1]. This study measured BNP and atrial natriuretic peptide levels in patients with CHF after therapeutic administration of digitalis. Although References this was a small population (total of 13 patients), the study showed an increase from 628 to 689 pg/mL at 1 hour after [1] Kobusiak-Prokopowicz M, Swidnicka-Szuszkowska B, Mysiak A. receiving IV digitalis versus placebo. This increase was Effect of digoxin on ANP, BNP, and cGMP in patients with chronic despite a decrease in pulmonary capillary wedge pressure, congestive heart failure. Pol Arch Med Wew 2001;105(6):475-82. which the authors concluded suggests a possible direct [2] Tsutamoto T, Wada A, et al. Digitalis increases brain natriuretic peptide in patients with severe congestive heart failure. Am Heart J 1997; action of digitalis on cardiac natriuretic peptides. 134(5 Pt 1):910-6. A second study published in 2001 measured BNP, [3] Adams KF, Gheorghiade VS, Gheorghiade M. B-type natriuretic atrial natriuretic peptide, and cGMP (cyclic guanosine 3c, peptide: from bench to bedside. Am Heart J 2003;145:S34-46. American Journal of Emergency Medicine (2005) 23, 563–564
www.elsevier.com/locate/ajem
Speed bump–induced spinal column injury
Sahin Aslan MDa,*, Ozgur Karcioglu MDb, Yavuz Katirci MDa, Hayati KandisS MDa, Naci Ezirmik MDc, Ozlem Bilir MDa aDepartment of Emergency Medicine, Atatu¨rk University, School of Medicine, 25090 Erzurum, Turkey bDepartment of Emergency Medicine, Dokuz Eylul University, School of Medicine, 35340 Inciralti, Izmir, Turkey cDepartment of Orthopaedic Surgery, Atatu¨rk University, School of Medicine, 25090 Erzurum, Turkey
Received 28 December 2004; accepted 29 December 2004
Abstract Introduction: Compression fracture of the vertebral body is common, especially in older adults. Injuries to the spinal column are one of the most frequent injuries by accidents and falls from heights. Vertebral fracture associated with minor trauma, however, is a rare occasion. Case Report: Five cases were injured in the inner city buses after passing onto speed bumps are presented. On presentation, four patients complained of severe pain in the thoracolumbar region, while in the other patient, physical examination revealed pain and tenderness on the neck. No neurologic deficit was noted except for one patient with tenderness on thoracic spines. Examination of the thoracolumbar X-ray and computed tomography displayed compression fractures in four patients. Other laboratory data obtained on admission were within normal limits. Posterior instrumentation was applied to three patients. All patients recovered well except for the one with cervical fracture. Conclusion: Drivers should be strongly warned and educated on the potential hazards of traversing past such bumps in roads too fast and such barriers should be built regarding tested standards. D 2005 Elsevier Inc. All rights reserved.
1. Compression fractures due to speed bumps death toll involving especially pedestrians and are placed predominantly on minor roads, which are both crowded by Authorities apply a wide array of precautions to cut down pedestrians and where vehicles are likely fast cruising. the relatively high rate of in-city motor vehicle accidents. Spinal column injury is an unintended result of these safety For instance, speed bumps are built on roads in a number of measures, as illustrated by the following 5 cases. countries to limit intraurban speed of vehicles at a point well Two patients (cousins), 50 and 57 years old, walked into below the legal speed limits [1]. These are designated to the university-based emergency department (ED) complain- eliminate motor vehicle accident and diminish the resultant ing of severe pain in the thoracolumbar region, which appeared during the bus ride they had nearly an hour before. The bus has passed a speed bump over the speed T Corresponding author. Tel.: +90 442 3166333/1461; fax: +90 442 3166340. limit in the city, which caused the bus to jump up high E-mail addresses: [email protected], [email protected] injuring the passengers at the back seats. The speed bump (S. Aslan). was nearly 15 cm in height, which was beyond the standard
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.12.015 564 S. Aslan et al. measure. Eyewitnesses told that the bus went over this magnetic resonance imaging disclosed minimal posterior barrier without slowing (~50 km/h). dislocation at C3 and C4. The patient was discharged with The men were not wearing safety belts during the event. cervical collar. The patient was still complaining of neck pain Both patients were placed onto spine board. On arrival, the in her follow-up visits at 1, 3, and 6 months. victims’ vital signs were normal. A complete physical In developed countries, traffic accidents appear to be the examination revealed pain and tenderness in the thoraco- most common cause of spinal fractures and spinal cord lumbar region; however, no neurological deficit was noted. injuries, whereas in developing countries, the most com- Examination of other organ systems was unremarkable. mon mechanism is falling from heights. Osteoporosis, Examination of the thoracolumbar plain radiographs dis- malignancy, inflammatory diseases, other metabolic bone played compression fractures in both patients. Thoracolum- diseases, and hematological diseases are known risk factors bar computed tomography images confirmed the for the development of spinal compression fractures [2,3]. presumptive diagnoses in the plain x-rays. Laboratory data A high percentage of lumbosacral fractures occur in (including serine calcium to rule out metastatic disease) individuals younger than 30 years. Nearly 60% of patients obtained on admission were within normal limits. The have serious disabling deficits. The most common types of patients denied any history of osteoporosis, degenerative acute vertebral fractures are compression fractures or bone disease, and malnutrition. The patients were admitted vertebral endplate fractures caused by sudden axial loading, to the hospital for operative repair. Posterior instrumentation transverse process avulsion by the origin of the psoas was performed in both patients in the same day. Both muscle, spinous process avulsions, and acute fracture of the patients were discharged within 15 days after an unevent- pars interarticularis from hyperextension [4,5]. Sudden ful follow-up. axial loading, which our patients were exposed to, could Within nearly 5 months after the first 2 cases, 2 other be the underlying mechanism in the resultant thoracolum- patients, 48 and 43 years old, were brought to ED because of bar compression fractures seen in the present cases. severe pain in the thoracolumbar region, which appeared Speed bumps to enforce 50-km/h speed limit in the city during an urban bus ride. Both patients complained of severe are built in areas with intense pedestrian population, such as back pain. It was learned that the bus passed a speed bump schools, parks, and residential areas, where cruising at a without slowing, which caused the bus to jump up high. higher speed could be too dangerous to afford. The women were not wearing safety belts during the In bus routes, the speed bumps should be 7.5 to 10 cm event. Both patients were placed onto spine board. On high and a warning sign should be placed 10 m before the arrival, the victims’ vital signs were normal, and physical bump [1]. The speed bumps could be effective if only the examination revealed pain and thoracolumbar tenderness. drivers obeyed the rules. However, if the speed bumps are One of these patients was noted to have motor neurological not built with respect to the standards and if the drivers do deficit, including hypoesthesia and 3/5 motor deficit in the not obey the rules, these bumps can be harmful, causing right leg. Examination of other organ systems was property damage, injuries, and even death. According to unremarkable. Thoracolumbar plain radiographs displayed statistics, in year 2002, 400000 traffic accidents caused compression fractures in both patients. Thoracolumbar 5000 deaths and 110000 injuries in Turkey. magnetic resonance imaging confirmed the presumptive In conclusion, drivers should be strongly warned and diagnoses in the plain x-rays. Laboratory data (including educated against the potential hazards of traversing too fast serine calcium to rule out metastatic disease) obtained on on speed bumps. Another point to mention is that such admission was within normal limits. The patients denied any barriers should be constructed regarding tested standards. history of osteoporosis, degenerative bone disease, or The pros and cons of various types of precautions to limit malnutrition. The patients were admitted to the hospital speed of motor vehicles should be taken into account by for operative repair. Posterior instrumentation was per- the local authorities before deciding which one should formed in one of these patients in the same day. This patient be chosen. was discharged after 13 days of follow-up, which was uneventful. The other patient wore a corset and was suggested to rest. Follow-up was uneventful and full References recovery was achieved. A 28-year-old woman was brought to ED because of [1] Speed control humps built on the roads. Institution of Turkish Standards. TS 6283/ December 1988 [in Turkish]. severe pain in the neck region. She reported that 30 minutes [2] National Institute of Health, Consensus Conference SD. Osteoporosis. ago, the bus she was traveling in passed a speed bump with JAMA 1984;252:799-802. its usual speed. The unexpected event caused a sudden back [3] Toth E. Vertebral deformities and the underlying diseases. Orv Hetil and forth movement of her neck. On arrival, the patient’s 2003;144(40):1955-63. vital signs were normal and a cervical collar was immedi- [4] Gertzbein SD, Court-Brown CM. Flexion-distraction injuries of the lumbar spine: mechanism of injury and classification. Clin Orthop ately applied. On examination, her neck was tender, with no 1988;227:52-60. apparent neurological deficit. There was no fracture on [5] Triantafyllou SJ, Gertzbein SD. Flexion-distraction injuries of the lateral and anterior-posterior cervical x-rays. Cervical thoracolumbar spine. Orthopaedics 1992;15:357-64. American Journal of Emergency Medicine (2005) 23, 565–566
www.elsevier.com/locate/ajem
Multiple organ failure as initial presentation of pheochromytoma
Alexandre Herbland MDa,*, Nam Bui MDa, Anne Rullier MDb, Frederic Vargas MDa, Didier Gruson MD, PhDa, Gilles Hilbert MD, PhD (Professor)a aIntensive Care Unit 2, Department of Pulmonary and Critical Care Medicine, Pellegrin University hospital, Bordeaux, France bDepartment of Pathology, Pellegrin University Hospital, Bordeaux, France
Received 31 December 2004; accepted 31 December 2004
A 50-year-old woman presented to the emergency MB, 144 U/L (NR b 6). She had a severe lactic acidosis À department complaining of recent cephalalgias with sudden (ie, pH, 7.17; HCO3 , 8.8 mmol/L; lactate, 8.6 mmol/L) but onset of palpitations, abdominal, and thoracic pain followed without hypoxemia. C-reactive protein level was 9 mg/L by nausea and vomiting. She had no medical history, but (NR b 5); urea nitrogen, 7 mmol/L; creatinine, 172 lmol/L; she reported a case of rhinitis treated for 2 days with a protein, 92 g/L; amylase, 423 IU/L; and lipase, 190 IU/L. nonsteroidal anti-inflammatory agent. On examination, Her leukocytes count was 18 Â 109/L. The clinical and she was afebrile, with cold peripheries and discreet biologic findings suggested a viral infection with acute cyanosis, and abdominal palpation was normal and painless. myocarditis, pancreatitis, and pneumonia. Because of severe There were no signs of heart failure, blood pressure was dehydration, the patient received intravenous rehydration 164/114 mm Hg, and heart rate was 150 beats per minute, and an antivomiting agent and was rapidly transferred to the but she was oliguric. Electrocardiogram showed a sinusal intensive care unit (ICU). tachycardia, an ST-segment elevation, Q waves in leads V1 On admission to ICU, she presented worsening dys- and V2, and an ST-segment depression of more than 2 mm pnea, extended marbling, and anuria despite a blood in leads V4 to V6. Chest radiograph showed bilateral pressure of 124/84 mm Hg. The new biologic findings interstitial infiltrates with right lower lobe consolidation. confirmed rapid evolution toward multiorgan-failure syn- Echocardiography showed global alteration of left ventric- drome with shock, rhabdomyolysis, hepatic cytolysis ular systolic function and grade II mitral regurgitation. (alanine aminotransferase, 9490 IU/L), acute pancreatitis Troponin Ic on admission was 66.8 lg/L (NR b 0.5); (amylase, 3050 IU/L), anuric renal failure (creatinine, creatine kinase, 634 U/L (NR b 175); and creatine kinase– 343 lmol/L), and intravascular disseminated coagulation (fibrin degradation products, 160 lg/mL; prothrombin time, 30%). Less than 24 hours after her admission at the hospital and a few hours after the initial treatments, the T Corresponding author. Service de re´animation polyvalente, Centre patient’s clinical condition deteriorated and she was hospitalier Pierre-Zobda-Quitman, BP 632, 97261 Fort de France Cedex, promptly intubated and ventilated. Continuous venovenous Martinique, French West Indies. Tel.: +33 6 96 95 63 85; fax: +33 5 96 75 84 76. ultrafiltration was rapidly initiated. E-mail addresses: [email protected], The echocardiographist noted an abnormal image of the [email protected] (A. Herbland). liver under diaphragmatic view. An abdominal ultrasound
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2004.12.001 566 A. Herbland et al. scan examination made by an experienced physician revealed a 7 Â 6 Â 6-cm hepatic mass in segment VI, suggesting a collected abscess. The patient was finally treated as a probable case of severe septic shock of hepatic and pulmonary origin, and after collection of blood and urine specimens and tracheal suctions for culture, antimicrobial therapy was administered. Despite multiple organ failure, the hemodynamic state was stable, without any need for vasopressor or inotropic support. Immediately after intubation, she had an atrioventricular block followed by asystole. The patient died despite a long resucitation, including external and intraventricular artificial cardiac pacing. Autopsy showed a normal macroscopic liver with no mass (Fig. 1). A 9 Â 7-cm (216-g), right retroperitoneal, hemorrhagic and necrotic mass involving the adrenal gland was found. Histological examination diagnosed a typical encapsulated pheochromocytoma with extensive necrotic Fig. 2 Microscopic aspects of the adrenal mass: at the periphery changes and no signs of malignancy (Fig. 2). Systematic of the tumour ie at distance of the areas of haemorrhage and sampling of organs also revealed in the left heart ventricle necrosis, pheochromocytoma was composed of large monotonous a subacute inflammatory infiltrate with numerous eosino- amphophilic cells with round nuclei. philic polymorphonuclear cells associated or not with unicellular necrotic muscular cells corresponding to adren- ergic myocarditis. suggesting pheochromocytoma. However, the lack of Pheochromocytoma is a rare cause of hypertension and hypertension, the report of a recent rhinitis, and the is known to have a potentially fulminant course [1,2], erroneous echographic diagnosis of liver abscess were especially after massive hemorrhagic and necrotic changes. probably the most confounding elements, leading to a The majority of patients report headache, sweating, false diagnosis. palpitations, and pallor. Other less frequent symptoms Surgical management is the appropriate treatment of are nervousness, tremor, nausea, fatigue, acute abdominal pheochromocytoma, together with preoperative medical pain, visual disturbances, weight loss, and dyspnea [3].In management using a- and b-blocking agents to limit the our case, initial symptoms (cephalalgias, palpitations, adverse effects of cathecholamine excess, which is mediated tachycardia, nausea and vomiting, and abdominal and by a- and b-adrenergic receptors [3]. Earlier diagnosis chest pain) could have been suggestive elements of would have led us to consider this treatment despite shock, etiologic orientation. Second, the continuous presence of leading to adrenalectomy. sustained blood pressure despite shock is another element Pheochromocytoma is often lethal when unsuspected [4]. Although uncommon, this condition should be considered early in patients with unexplained shock, multiorgan-failure syndrome, and preserved blood pressure because it is be potentially curable by surgical removal [5].
References
[1] Kolhe N, Stoves J, Richardson D, Davison AM, Gilbey S. Hyperten- sion due to phaeochromocytoma—an unusual cause of multiorgan failure. Nephrol Dial Transplant 2001;16:2100-4. [2] Kizer JR, Koniaris LS, Edelman JD, St John Sutton MG. Pheochro- mocytoma crisis, cardiomyopathy, and hemodynamic collapse. Chest 2000;118:1221-3. [3] Werbel SS, Ober JP. Pheochromocytoma: update on diagnosis, localization and management. Med Clin North Am 1995;79:131-53. [4] Sutton MG, Sheps SG, Lie JT. Prevalence of clinically unsuspected pheochromocytoma. Review of a 50-year autopsy series. Mayo Clin Proc 1981;56:354-60. [5] Newell KA, Prinz RA, Pickleman J, Braithwaite S, Brooks M, Karson Fig. 1 Macroscopic aspects of the adrenal mass with extensive TH, et al. Pheochromocytoma multisystem crisis. A surgical emergen- necrotic and hemorrhagic changes. cy. Arch Surg 1988;123:956-9. American Journal of Emergency Medicine (2005) 23, 567–568
www.elsevier.com/locate/ajem
Carbon dioxide asphyxiation caused by special-effect dry ice in an election campaign
Cheng-Chun Hsieh MD, Chung-Liang Shih MD, Cheng-Chung Fang MD, Wen-Jone Chen MD, PhD, Chien-Chang Lee MD, MSc*
Department of Emergency Medicine, National Taiwan University Hospital, Taipei 100, Taiwan
Received 31 January 2005; accepted 1 February 2005
During an election campaign held in a mansion hall, a Neurologic examination revealed bilateral 4-mm-sized, previous healthy 33-year-old man was discovered uncon- fixed, nonreactive pupils, and an elevated muscle tone with scious in a wooden coverless box. The young man was a a decorticate posture. Some skin wounds with intact or worker of a public relations company and was assigned to ruptured clear blisters were noted on his right elbow (Fig. 1). hide in a 7-ft-long, 2-ft-wide and 2.5-ft-deep topless wooden The location of the lesions was compatible with the site of box. According to the designed scenario, he was made to the outlet of the dry ice pipe. Witness reported that there were help rise a big Taiwan-map slogan out of the box, symbolic no signs of struggle before the accident. The victim had no of bTaiwan on the rise.Q He lay at the bottom of the box and history of psychiatric disorders, recent personal crises, or a 3-in pipe was pouring dry ice near his right arm. Because medical illnesses. Later investigation of the scene of the of the fog, the inner side of the box soon became invisible. accident also revealed a normal functioning air-conditioning At least 20 minutes had elapsed before the young man was and ventilation system. discovered cyanotic and unconscious. Without palpable Series of laboratory results on ED admission were within pulse and spontaneous respiration, the victim was resusci- reference range (Table 1). Electrocardiography and chest tated with mouth-to-mouth artificial ventilation and cardiac radiograph were unremarkable. Computed tomography massage by a surgeon at the scene. Emergency medical showed diffuse brain edema. His consciousness level technicians arrived 9 minutes later, when weak pulse was improved to E1M3VT 20 minutes later. Under the impres- felt and irregular spontaneous respiration was restored. sion of hypoxic encephalopathy, he was cared in the Laryngeal mask airway was established and he was sent to neurologic intensive care unit. Two months later, he still our ED for further evaluation and management. remained vegetative. On arrival, physical examination revealed a comatose Dry ice is widely used in various public activities as consciousness with a Glasgow Coma Scale of E1M1V1. special-effect fog. However, even in open space, its safety is still questionable. The molecular weight of dry ice is 44, which means it is heavier than air and easily accumulates in the bottom of a box, pool, or building. According to the No specific funding was established by any author or institution for Occupational Safety and Health Administration and Amer- this work. T Corresponding author. Tel.: +886 2 23123456x5926; fax: +886 2 ican Conference of Governmental Industrial Hygienists, the 23223150. exposure limit of carbon dioxide is 5000 parts per million 3 E-mail address: [email protected] (C.-C. Lee). (9000 mg/m ) or 0.5% for an 8-hour time-weighted
0735-6757/$ – see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2005.02.045 568 C.-C. Hsieh et al. average [1]. The Occupational Safety and Health admin- Table 1 Laboratory results of our patient on admission istration considers a 10% carbon dioxide concentration pH 7.468 potentially lethal [2]. Pao (mm Hg) 95.1 From a toxicologic standpoint, carbon dioxide simply 2 Paco2 (mm Hg) 35.2 resembles asphyxants [3]. However, it has direct toxic effect À HCO3 (mmHg) 22.3 in high concentration despite normal atmospheric oxygen Creatinine (mg/dL) 1.3 concentrations [4]. It causes hypercapnia, increased heart Blood urea nitrogen (mg/dL) 12.5 rate, cardiac output, mean pulmonary artery pressure, and Prothrombin time (s) 12 pulmonary vascular resistance [5,6] and therefore imposes Creatinine kinase (U/L) 185 excess load on the myocardium [3]. The predominant Creatinine kinase–MB (U/L) 31.7 mechanism for hypercapnia is decreased hypoxic ventilatory Troponin-I (ng/mL) 0.007 drive [7]; this aggravates the hypoxia and eventually causes brain damage. The chest radiograph of the victim was normal, Dry ice should be used cautiously in public activities. Its compatible with previous observations [3,8,9].Infact, odorless, nonirritant characteristics may cause asphyxia reports of fatal cases of carbon dioxide exposure [8,9] silently even in an open space, if regional accumulation is indicate that the lungs were normal on postmortem possible. Once carbon dioxide intoxication occurred, the examination. The rate of excretion of carbon dioxide is victim should be evacuated and resuscitated immediately. very rapid with hyperventilation. Blood partial pressure of The general public should be made aware of the potential carbon dioxide returns to normal within a few minutes [10]. hazard of dry ice and the silent nature of asphyxiation, This explained the normal arterial gas profile after intensive which is the most effective way to prevent the next tragedy. resuscitation minutes later. Contradictory to general belief, respiratory arrest usually occurs without any warning signs like cough, strenuous References breathing, or struggle, preventing early recognition and rescue. In our case, the evaporating dry ice yielded carbon [1] National Institute for Occupational Safety and Health. The registry of dioxide, accumulated in the bottom of the box, and toxic effects of chemical substances: carbon dioxide RTECS displaced much of the oxygen; thus, unconsciousness FF6400000, CAS 124-38-9 [Centers for Disease Control and Preven- developed silently. Without timely resuscitation, prolonged tion Web site]; 2004. hypoxia may lead to hypoxic encephalopathy and even [2] Occupational Safety and Health Administration. Asphyxiation hazard when filling stationary low pressure carbon dioxide supply death. Besides, loss of consciousness predisposed the victim systems; 1996. to the frostbite because he had no reactions to pain or [3] Halpern P, Raskin Y, Sorkine P, et al. Exposure to extremely high coldness upon dry ice contact. concentrations of carbon dioxide: a clinical description of a mass casualty incident. Ann Emerg Med 2004;43:196-9. [4] Ikeda N, Takahashi H, Umetsu K, et al. The course of respiration and circulation in death by carbon dioxide poisoning. Forensic Sci Int 1989;41:93-9. [5] Kiely DG, Cargill RI, Lipworth BJ. Effects of hypercapnia on hemodynamic, inotropic, lusitropic, and electrophysiologic indices in humans. Chest 1996;109:1215-21. [6] Mas A, Saura P, Joseph D, et al. Effect of acute moderate changes in
Paco2 on global hemodynamics and gastric perfusion. Crit Care Med 2000;28:360-5. [7] Gomersall CD, Joynt GM, Freebairn RC, et al. Oxygen therapy for hypercapnic patients with chronic obstructive pulmonary disease and acute respiratory failure: a randomized, controlled pilot study. Crit Care Med 2002;30(1):113-6. [8] Gill JR, Ely SF, Hua Z. Environmental gas displacement: three accidental deaths in the workplace. Am J Forensic Med Pathol 2002; 23:26-30. [9] Guillemin MP, Horisberger B. Fatal intoxication due to an unexpected presence of carbon dioxide. Ann Occup Hyg 1994;38:951-7. [10] Halpern P, Neufeld MY, Sade K, et al. Middle cerebral artery flow Fig. 1 Frost bite wound with ruptured clear blisters on the velocity decreases and electroencephalogram (EEG) changes occur as patientTs right elbow. acute hypercapnia reverses. Intensive Care Med 2003;29:1650-5. American Journal of Emergency Medicine (2005) 23, 569–587
www.elsevier.com/locate/ajem
Correspondence
Emergency department procedural sedation Table 1 Emergency physician medication access formularies Drug Unrestricted Adult Child RSI Not (%) only only only available To the Editor, (%) (%) (%) (%) Choral 51 0 13 0 36 The ideal procedural sedation agent for any patient may hydrate be less a function of pharmacology and more a function of Droperidol 42 8 0 1 49 physician access. This study was undertaken to profile ED Etomidate 46 4 0 17 33 procedural sedation formularies and define which agents are Fentanyl 62 12 0 3 23 available for use by emergency physicians (EPs) within their Ketamine 43 3 9 4 41 departments. Relationships between medication availability Methohexital 18 3 0 3 76 and ED characteristics were also examined. Midazolam 86 1 0 3 10 A survey was conducted of emergency physicians Nitrous oxide 8 1 1 0 90 attending a national meeting of community ED directors Pentobarbital 28 1 0 7 64 Propofol 34 8 0 13 45 regarding their access to various procedural sedation Remifentanil 7 1 1 3 88 medications. The specific information collected in the survey included ED annual census, medications available Abbreviation: RSI, rapid sequence intubation. to emergency physicians, ED procedural sedation policy development, and credentialing of EPs for specific levels number of medications with unrestricted access and ED of sedation. Medications available only for endotracheal census ( P b .002) and EP credentialing ( P b .002). intubations were differentiated from those used for non– To our knowledge, this report represents the first endotracheal intubations procedures. Statistical analysis published multi-site study on ED procedural sedation was through v2 and regression modeling. formularies. Our findings demonstrate a marked degree of A total of 76 completed surveys were returned out of a variability between institutions with regard to emergency possible 155 hospital sites (49%). Annual ED censuses physician access to these agents. It is surprising to note the ranged from 6000 to 111000 visits with a median of 28200 number of community hospitals with unrestricted access to visits. Procedural sedation policies were created solely by agents such as propofol (34%), etomidate (46%), ketamine emergency medicine in 9 (12%) sites, anesthesia only in 22 (43%), and droperidol (42%), considering the recent (29%) sites, and by both in 40 (53%) sites. Five sites (6%) controversies surrounding these agents [1-3]. The finding had no written or an unknown policy. Emergency physicians that almost half the sites have access to at least one of the were credentialed for deep sedation at 16 (21%) sites and agents indicates the successful relationship that emergency moderate sedation at 48 (63%) sites. Twelve sites (16%) did and anesthesia departments can develop. not specify the level of credentialing. The percentage of Unfortunately, no useful correlation between ED departments with described access to medications are listed characteristics and procedural sedation formularies could in Table 1. be determined from this study. Future prospective studies The median number of unrestricted medications per of actual patient outcomes will be needed to assess EP’s site was 4, whereas 3 sites had unrestricted access to all performance with these agents. 11 medications, and 8 sites had unrestricted access to no medications. References There were no statistically significant differences be- tween access to individual medications and ED census, [1] American Society of Anesthesiologists. AANA-ASA joint state- policy development, or credentialing for moderate or deep ment regarding propofol administration. http://www.asahq.org/news/ sedation. A significant correlation was noted between propofolstatement.htm.
0735-6757/$ – see front matter n 2005 Elsevier Inc. All rights reserved. 570 Correspondence
[2] Shale JH, Shale CM, Mastin WD. Safety of droperidol in behavioural normoactive bowel sounds. There was no distention emergencies. Expert Opin Drug Saf 2004;3:369-78. observed, no tympany elicited on percussion, no masses [3] van Zwieten K, Mullins ME, Jang T. Droperidol and the black box palpated, and no hepatosplenomegaly appreciated. The warning. Ann Emerg Med 2004;43:139-40. rectal examination was initially deferred as the clinician was acquainted with the patient and her family. Alfred Sacchetti, MD After IV rehydration, the patient was no longer ortho- Russell H. Harris, MD static and felt much better, although still quite tired as she Our Lady of Lourdes Medical Center ambulated to the rest room. Laboratory results revealed the Camden, NJ 08103, USA following: hematocrit of 30.3; blood urea nitrogen and creatinine of 38 and 0.9, respectively. A nasogastric tube Dighton Packard, MD was inserted and 350 mL of maroon blood was aspirated. Baylor Medical Center Gastroenterology consultation was requested and the Dallas, TX 75246, USA patient underwent immediate endoscopy at which time she was found to have a Dieulafoy’s lesion in the gastric doi:10.1016/j.ajem.2004.12.004 mucosa. The patient was treated with sclerotherapy and on follow-up has been asymptomatic for 2 years. Dieulafoy’s lesion is an arterial malformation in the Dieulafoy’s lesion presenting as near-syncope with submucosal layer that can lead to GI hemorrhage. Although massive gastrointestinal hemorrhage the majority of these lesions are found in the proximal stomach within 6 cm of the gastroesophageal junction [1], To the Editor, there are documented cases of Dieulafoy’s lesion elsewhere in the GI tract as well [2-4]. The site of bleeding is usually a Although upper gastrointestinal (GI) bleeding has an 2- to 5-mm mucosal defect, and clinical presentation in one overall annual incidence of 100 per 100000 in the general study showed 28% of patients developed hematemasis population, bleeding secondary to Dieulafoy’s lesion is rare alone, 51% with hematemasis and melena, and 18% with and likely underdiagnosed. This vascular malformation can melena alone [5]. Hemorrhage, shock, and death can occur present with massive and catastrophic hemorrhage that can with severe bleeding. Initial stabilization is the same as that be fatal if not managed aggressively. We present the case of for other causes of GI bleeding, with 2 large bore peripheral an unusual presentation of upper GI bleeding secondary to IV lines and infusion of crystalloid. Although the initial Dieulafoy’s lesion. management of Dieulafoy’s lesion historically has been A 45-year-old woman presented to our ED by ambulance surgical, endoscopic treatment is now the preferred means after feeling fatigue, malaise, and indigestion all day while of management [6-9]. at work as a nurse in a pediatricians’ office. She stated she This case report presents an interesting presentation of a felt hot, sweaty, and lightheaded most of the day. Upon rare cause of GI bleeding. Emergency physicians must arriving home at the end of a very busy day, she bcollapsedQ remain cognizant that GI bleeding has a spectrum of into a chair and had diminished level of responsiveness but possible presentations and etiologies, some of which have did not lose consciousness. She had been lowered to the the potential to be extremely surreptitious but rapidly floor by her husband who instructed his son to call 9-1-1. progressive and life-threatening. She was found by paramedics to be awake but poorly responsive with a blood pressure of 90/54 and a heart rate of 80. She became more conscious during transport and arrived References in the ED alert and oriented. On initial questioning, she states she ate what she thought was bad fish the night before [1] Veldhuyzen van Zanten SJ, Bartelsman JF, Schipper ME, Tytgat GN. Recurrent massive haematemesis from Dieulafoy’s vascular malforma- and attributed her symptoms to food poisoning. She denied tions: a review of 101 cases. Gut 1986;27:213-22. nausea, vomiting, diarrhea, or abdominal pain. She had a [2] Iglasias SS, Roses LL, Ramirez AG, Seco AL, Blance ES. Ileal past medical history significant for migraine headaches, Dieulafoy’s lesion. Gastrointest Endosc 2004;59:266. depression, and lactose intolerance, and took only Prozac [3] Womack N. Blood flow through the stomach and duodenum. Am J daily and Lactaid as needed. The patient’s social history was Surg 1969;117:771-80. [4] Fujimaru T, Akahoshi K, Matsuzaka H, Sumita Y, Kubokawa M. significant for occasional tobacco and alcohol use, but she Bleeding rectal Dieulafoy’s lesion. Gastointest Endosc 2003;57:922. denied drug use. [5] Reilly III HF, al-Kawas FH. Dieulafoy’s lesion. Diagnosis and On physical examination, she was noted to be pale but management. Dig Dis Sci 1991;36:1702-7. was awake, alert, oriented, and conversant. Her vital signs [6] Norton ID, Peterson BT, Sorbi D, Balm RK, Alexander GL, Gastout are recorded as follows: blood pressure, 103/64; heart rate, CJ. Management and long-term prognosis of Dieulafoy’s lesion. Gastrointest Endosc 1999;50:762-7. 84; respiratory rate, 18; O2 saturation, 100%. She was noted [7] Skok P. Endoscopic hemostasis in exulceratio simplex-Dieulafoy’s to be orthostatic, but the capillary refill was less than disease hemorrhage: a review of 25 cases. Endoscopy 1998;30: 2 seconds. The patient’s abdomen was soft, nontender with 590-4. Correspondence 571
[8] Parra-Blancso A, Takahashi H, Mendez Jerez PV, Kojima T, Aksoz K, his HbA1c level increased to 9.2% in December 2002. In Kirihara K, et al. Endoscopic management of Dieulafoy’s lesions of the November 2003, he was admitted with abdominal pain and stomach: a case study of 26 patients. Endoscopy 1997;29:834-9. dark urine. His symptoms subsided after parenteral use of [9] Yamaguchi Y, Yamamato T, Katsumi N, Hashimoto T, Morozumi K, glucose with negative Watson-Schwartz test. Meanwhile, Sugiyama M, et al. Endoscopic treatment of hemorrhagic gastric ulcer his porphobilinogen deaminase activity was 44.6 in patients aged 80 years or more. Hepatogastroenterology 2001;48: d À1d À1 1195-8. nmol h mL red blood cell, which was the lower limit of the age-matched healthy Taiwanese subjects (44.2-55.4 nmold hÀ1d mLÀ1 red blood cells) [2]. His HbA1c was 9.6% Timothy Mader, MD during this admission, and there was no more AIP attack Eric Beauvois, MD until September 2004. Baystate Medical Center/ Although diabetes is shown to decrease the risk of AIP Tufts University School of Medicine attacks [3,4], 3 patients with diabetes antedating AIP for 3 Springfield, MA 01199, USA to 20 years have been reported [5]. Usually, AIP affects E-mail address: [email protected] persons in their 30s [6]; however, patients having diabetes- antedated AIP may experience relevant symptoms after 50 doi:10.1016/j.ajem.2004.12.002 years, as is shown in our case. Of particular note is that our patient had fewer AIP attacks when his glycemic control Effects of diabetes on the ED presentation of acute worsened in recent years. Perhaps, a higher blood glucose intermittent porphyria level could suppress aminolevulinic acid synthase and then ameliorate the frequency and/or severity of AIP attacks, yet To the Editor, it still cannot elicit a complete remission of AIP in some patients. On the other hand, these discrepant observations Whether diabetes could prevent attacks of acute inter- between previous reports may be reasoned by the different mittent porphyria (AIP) remains incompletely understood. ethnic populations studied. Further large comparative Lithner described 16 AIP patients who had complete studies with prospective design are therefore needed to remission of AIP attacks after onset of diabetes [1]; examine this interesting and important issue. In conclusion, however, a correlation between blood glucose level and although diabetes may have beneficial effects on patients frequency/severity of AIP attacks was not mentioned. Can with AIP, AIP should still be considered as a differential AIP be excluded in a diabetic patient who visits ED for diagnosis in a diabetic patient with severe abdominal pain severe abdominal pain? AIP is a very rare hereditary at the ED. metabolic disorder in Taiwan; herein, we reported a patient who had diabetes antedated AIP. A 68-year-old Taiwanese man took herbal medications References for the treatment of hyperglycemia in 1980s. He had visited our ED because of severe abdominal pain with vomiting and [1] Lithner F. Beneficial effect of diabetes on acute intermittent porphyria. dark urine on 30 occasions since August 1992 (when he was Diabetes Care 2002;25:797-8. 56 years old) to December 1999. AIP was diagnosed [2] Lee FY, Hsiao KJ, Tsai YT, et al. Determination of erythrocyte hydroxymethylbilane synthase activity and its application for study of according to the typical symptoms and positivity of Watson- acute intermittent porphyria. J Formos Med Assoc 1988;87:1029-35. Schwartz test for 4 times. No significant change was found [3] Andersson C, Bylesjo I, Lithner F. Effects of diabetes mellitus on patients on radiographic, panendoscopic, or laboratory examinations with acute intermittent porphyria. J Intern Med 1999;245:193-7. except for hyperglycemia (190-250 mg/dL). He has received [4] Yalouris AG, Raptis SA. Effect of diabetes on porphyric attacks. BMJ glibizide 2.5 mg BID since June 1993, and his glycosylated 1987;295:1237-8. [5] Sterling K, Silver M, Richetts HT. Development of porphyria in hemoglobin (HbA1c) level was 8.3% (upper limit of diabetes mellitus. Arch Intern Med 1949;6:965-75. normal, 6%) in November 1998. The number of AIP attacks [6] Thadani H, Deacon A, Peters T. Diagnosis and management of declined to less than 3 times a year since 2000 (Table 1) and porphyria. BMJ 2000;320:1647-51.
Table 1 The relation of serum glucose and episodes of AIP Year 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Episodes 3 4 0 0 2 0 8 11 1 3 0 1 0 HbA1c (%) 8.3 9.2 9.9 Sugar (mg/dL)a 200-250 200 250-360 260-300 220 a Serum glucose level was recorded upon arrival at the ED. 572 Correspondence
Huan-Wen Chen, MD validated the following training protocol associating at least Department of Internal Medicine 8 hours of initial tutorial and at least 25 examinations Far Eastern Memorial Hospital including patients’ follow-up. According to our results, Taipei 100, Taiwan performance reached with such training was similar to the performance reached by hospital emergency physicians. The Deng-Huang Su, MD results we reported are consistent with those of Jang et al Far Eastern Polyclinic whereas ultrasonography indications are different. In both Taipei 100, Taiwan studies, the minimal training seemed to involve 25 exami- nations or less. Despite the lack of strong evidence, as Jang et Tzuu-Shuh Jou, MD, PhD al, we concluded that the recommendations of the American Jia-Horng Kao, MD, PhD College of Emergency Physicians were pertinent as the Department of Internal Medicine suggested number of examinations required (25 for each National Taiwan University Hospital and indication) is similar to our results [3]. Such attitude seems to National Taiwan University College of Medicine be adequate for emergency physicians willing to reach a high Taipei 100, Taiwan level of performance. It should be noted that, in particular Email address: [email protected] indications, a much shorter tutorial has been proposed with satisfactory results [4,5]. This point illustrates the need to doi:10.1016/j.ajem.2004.12.003 validate tutorials and practical training according to a specific indication, as Jang et al and we undertook. Training emergency physicians to perform out-of-hospital ultrasonography References To the Editor, [1] Jang T, Aubin C, Naunheim R. Am J Emerg Med 2004;22:439-43. [2] Robinson NA, Clancy MJ. Should UK emergency physicians We read with interest the paper from Jang et al [1] on undertake diagnostic ultrasound examinations? J Accid Emerg Med training for ultrasonography in emergency cases. Training is 1999;16:248-9. a crucial matter when ultrasonographic examination is [3] ACEP Board of Directors. ACEP emergency ultrasound guidelines. performed by a nonradiologist operator [2]. Our experience Available at: www.acep.org [accessed September 24, 2004]. [4] Marshburn TH, Legome E, Sargsyan A, Li SM, Noble VA, et al. could contribute to this issue as we recently evaluated Goal-directed ultrasound in the detection of long-bone fractures. J initial training for emergency physicians performing out-of- Trauma 2004;57:329-32. hospital ultrasonography. [5] Kimura BJ, Amundson SA, Willis CL, Gilpin EA, DeMaria AN. A 4-hour tutorial on focused abdominal sonography for Usefulness of a hand-held ultrasound device for bedside examination of trauma included theoretical lecture, demonstration, and left ventricular function. Am J Cardiol 2002;90:1038-9. practice. Eight experimented emergency physicians attended 1 to 6 tutorials. Then, ultrasonography was routinely used Fre´de´ric Lapostolle, MD by emergency physicians in out-of-hospital settings. Tomislav Petrovic, MD Patients were followed up to validate out-of-hospital Jean Catineau, MD diagnosis. Indeed, it should be noted that, because of out- Gille Lenoir, MD of-hospital settings, patients’ follow-up was indispensable Fre´de´ric Adnet, MD, PhD for the physicians to validate their out-of-hospital diagnosis. SAMU 93, EA 3409, CHU Avicenne Two hundred twenty-two examinations were analyzed to 125, rue de Stalingrad evaluate initial training. Diagnostic performance (sensitivity, 93009 Bobigny, France specificity, and positive and negative predictive values) E-mail address: [email protected] increased for emergency physicians after attending 2 tutorials (Table 1). Diagnostic performance according to the number doi:10.1016/j.ajem.2005.01.008 of examinations performed using a receiver operator curve showed a cutoff point of 25 examinations. Then, we Spontaneous pneumothorax presenting as epigastric pain
Table 1 Comparison of diagnostic performance according to To the Editor, attended formations Sensitivity Specificity PPV NPV Spontaneous pneumothorax is a common disease en- countered in hospital practice. It usually occurs in young b 2 formations 20 98 67 83 men and presents with sudden onset of respiratory distress N2 formations 92 98 89 99 and chest pain. We report on a patient with sponta- PPV indicates positive predictive value; NPV, negative predictive value. neous pneumothorax, who presented with epigastric pain, Correspondence 573
in tall thin boys and men between the ages of 10 and 30 years [1-4]. It results from rupture of subpleural bullae and air leaks into pleural cavity. Practically, all patients have ipsilateral pleural chest pain or acute dyspnea because reduction of lung volume results in respiratory distress [1-4]. Other minor symptoms include cough and general malaise [5]. However, our patient presented with epigastric pain, mimicking acute peptic ulcer disease or pancreatitis and, surprisingly, without respiratory symptoms. It is very rare manifestation for spontaneous pneumothorax. In English literature, we found only one report on tension pneumothorax presenting with epigastralgia and epigastric tenderness described by Hollins et al [5]. Very recently, Lien et al [6] reported pneumothorax presenting with right upper quadrant pain and a positive Murphy sign, mimicking acute cholecystitis. The mechanism by which spontaneous pneu- mothorax causes epigastric pain is still uncertain. In the Fig. 1 Chest radiogram reveals around 80% pneumothorax on patients reported by Hollins et al [5], epigastric pain might the left side without pleural effusion. be elicited by tension pneumothorax that depressed the right hemidiaphragm of the affected side. In the patient reported by Lien et al [6], no tension pneumothorax was found, and mimicking peptic ulcer disease or acute pancreatitis. The therefore, they speculated that right upper quadrant pain pain subsided after chest tube thoracostomy. might be caused by a small pleural effusion. In our patient, A 41-year-old man, smoking for 20 years, had gradual however, there was neither tension pneumothorax nor onset of epigastric pain 2 weeks before visit to our hospital. pleural effusion. We suspected that other mechanism such He had no history of heavy alcohol intake. Reviewing his as influence on diaphragm of the affected side by highly history, the respiratory symptoms, such as cough, dyspnea, collapsed lung toward pulmonary ligament would be chest or shoulder pain, did not occur after the onset. No concerned with the epigastric pain. It would be possible in trauma history was noted. He had no medical history of our case rather than tension pneumothorax and pleural peptic ulcer disease and pancreatobiliary disease. On arrival, effusion. In addition, Lien et al [6] suggested that young age he was clear and ambulatory. The vital signs were stable: and high pain tolerance were the reason why the patient had blood pressure was 130/80 mm Hg, pulse rate 84 beats per no respiratory symptoms. With this phenomenon, we minute, respiratory rate 16 breaths per minute, and body suppose that very slow progression of pneumothorax would temperature 37.08C. Blood gas analysis at room air was Po2 be concerned in our patient. 71.4 mm Hg and Pco2 40.1 mm Hg, and his peripheral In a patient with epigastric pain, an intra-abdominal oxygen saturation was 94%. Physically, auscultation of the pathology, especially gastrointestinal and pancreatobiliary lung revealed clear breath sounds except over the left lung diseases, should be sought initially. However, consider a field where breath sounds were diminished. The abdomen spontaneous pneumothorax if there are negative findings on was soft but with epigastric tenderness. The remainder of abdominal sonography and gastric endoscopy. the physical examination was unremarkable. The differential diagnoses were peptic ulcer disease and acute pancreatitis. A peripheral white cell count showed 7100 cells per References microliter without a left shift. Serum amylase and lipase were within normal limits. Abdominal ultrasound revealed [1] Wissberg D, Refaely Y. Pneumothorax: experience with 1,199 patients. normal a pancreas without pancreatobiliary tract dilatation. Chest 2000;117:1279-85. In gastric endoscopy, neither gastric nor duodenal ulcer was [2] Sahn SA, Heffner JE. Spontaneous pneumothorax. N Engl J Med found. Posterior-anterior chest radiograph disclosed a left 2000;342:868-74. pneumothorax (around 80%) without a mediastinal shift [3] Baumann MH, Strange C, Heffner JE, et al. Management of spontaneous pneumothorax. An American College of Chest Physicians (Fig. 1). Pleural effusion was not observed on the Delphi consensus statement. Chest 2001;119:590-602. radiograph. Soon after chest tube thoracostomy, his epigas- [4] Noppen M, Baumann MH. Pathogenesis and treatment of primary tric pain subsided. Full reexpansion of the left lung was spontaneous pneumothorax: an overview. Respiration 2003;70:431-8. achieved, and he was discharged uneventfully 4 days later. [5] Hollins GW, Beattie T, Happer I, et al. Tension pneumothorax: report of No more epigastric pain was reported after discharge. two cases presenting with acute abdominal symptoms. J Accid Emerg Med 1994;11:43-4. Primary spontaneous pneumothorax is a common disease [6] Lien WC, Yuan A, Tsai KC, et al. Primary spontaneous pneumothorax encountered in hospital practice. It is defined as pneumo- with clinical manifestation mimicking acute cholecystitis. J Emerg Med thorax without underlying lung disease and typically occurs 2004;26:354-6. 574 Correspondence
Ryoko Ogawa, MD Table 1 Demographics and clinical variables by group Yusuke Yamamoto, MD Norihiro Haraguchi, MD Dehydrated Febrile Significance Hiroaki Satoh, MD Male sex [n (%)] 7 (70) 12 (71) RR = 0.99 Kiyohisa Sekizawa, MD (0.59 to 1.64) Division of Respiratory Medicine Breastfed [n (%)] 4 (40) 13 (76) RR = 0.53 Institute of Clinical Medicine (0.24 to 1.18) University of Tsukuba Median age 64 [38-84] 52 [32-81] P = .69 (d [IQR]) Tsukuba City Mean weight 52 F 36 61 F 30 P = .55 Ibaraki, 305-8575, Japan (percentile) (À22 to 39)a E-mail address: [email protected] Mean temperature 38.2 F 1 38.4 F 1 P = .66 (8C) (À1to1)a doi:10.1016/j.ajem.2004.12.007 Mean pulse 175 F 24 164 F 16 P = .23 (beats/min) (À30 to 8)a A cross-sectional ED survey of infantile subclinical Mean respirations 41 F 841F 10 P = .98 methemoglobinemiaB (beats/min) (À7to7)a Mean oxygen 99 F 299F 1 P = .33 a To the Editor, saturation (À1to2) (arterial) (%) Median dehydration 5 [5-10] 0 [0-0] P b .001 The prevalence of clinically significant methemoglobi- level (% [IQR]) nemia (MHb) is unknown among previously healthy infants Total 10 17 presenting to the emergency department (ED) with diarrhea a The 95% CI of the difference in the means. and dehydration. However, it is well described in these infants and, when present, can be associated with significant morbidity [1-4]. Nevertheless, it may not be clinically apparent until levels reach almost 7 times normal levels [5]. tion scale [6]. Well-hydrated comparison patients were Based on this, we hypothesized that dehydrated infants selected from a convenience sample of well-hydrated younger than 3 months with diarrhea would commonly have infants in the same age range undergoing a fever MHb levels above normal but below that causing cyanosis evaluation. Patients were excluded if they had a history and obvious distress. of MHb, an acidemia-inducing illness (eg, metabolic We conducted a cross-sectional study of MHb in disorder), or received oxidizing medication via breast acyanotic dehydrated infants with diarrhea and a compar- milk or direct ingestion. ison group of well-hydrated febrile infants, all 3 months old The research was approved by the hospital’s institutional or younger, presenting to an urban pediatric ED. The age review board. Written informed consent was obtained from group and severity of illness of the dehydrated patients were the parents of all infants. The study was conducted at chosen for the highest likelihood of MHb [3]. A sample size the Children’s National Medical Center General Clinical of 10 was needed in each group, with a one-tailed a level of Research Center. .05 and power of 80% to detect a mean group MHb level of Thirty patients were approached for enrollment and 5% F 5%. (An MHb level of 5% was chosen as a cutoff 3 parents refused (1 in the dehydration group and 2 in the level for which symptoms are likely to occur but clinical fever group). We enrolled a total of 27 subjects: 10 with signs will not be evident.) dehydration and 17 febrile patients. The mean age of the A convenience sample was enrolled during 1 year entire study population was 59 F 29 days; 70% were boys. from among age-appropriate patients with diarrhea and The 2 groups were demographically similar (Table 1). The diagnosis of dehydration as determined by the on-duty prevalence of subclinical MHb (sMHb) (ie, MHb N1.5%) attending pediatric emergency physician. Diarrhea was was 2 of 10 (20%; 95% CI, 2.5-55.6) dehydrated subjects defined as 3 or more loose or watery stools in the prior and 1 of 17 (5.9%; 95% CI, 0.2-28.7) well-hydrated 24 hours. Degree of dehydration was established by the comparison patients ( P = NS). It is noteworthy that of the use of a generally accepted pediatric emergency dehydra- patients with sMHb, the 2 dehydrated patients had levels of 5.9% and 6%, whereas the single well-hydrated patient had a level of 1.9%. Presented in part at the Clinical Research 2003 National Meeting None of the potential risk factors for sMHb were (March 23, 2003), the National Meeting of the Pediatric Academic significantly associated with group designation. Among all Societies (May 3, 2003), and the National Meeting of the Society for 27 patients enrolled, v2 analysis and the Student t test Academic Emergency Medicine (May 29, 2003). B This research was supported by GCRC grant M01-RR13297 from the revealed the following factors significantly associated with National Center for Research Resources, National Institutes of Health, sMHb: lower mean body weight percentile ( P = .027) and Bethesda, MD. higher mean chloride level ( P = .033) (Table 2). Correspondence 575
applicability to ED patients, in whom sMHb is probably Table 2 Clinical and laboratory measurements by presence more likely to go unsuspected and therefore undetected of sMHb before disposition. Our data provide a first estimation of the N V MHb 1.5% MHb 1.5% Significance prevalence of elevated MHb levels in dehydrated infants Breastfed 0 (0) 8 (33) OR = 0.07 with diarrhea presenting to the ED. [n (%)] (0.00 to 1.54) Median age 55 [42-59] 62 [36-86] P = .64 (d [IQR]) References Mean weight 20 F 963F 30 P = .027 (percentile) (5 to 79)a [1] Babbitt CJ, Garrett JS. Diarrhea and methemoglobinemia in an infant. Mean 38.7 F 0.5 38.2 F 1 P = .22 Pediatr Emerg Care 2000;16:416-7. temperature (1.3 to À0.4)a [2] Gebara BM, Goetting MG. Life-threatening methemoglobinemia in (8C) infants with diarrhea and acidosis. Clin Pediatr 1994;33:370-3. Median 10 [0-10] 0 [0-5] P = .10 [3] Lebby T, Roco JJ, Arcinue EL. Infantile methemoglobinemia in infants with diarrhea and acidosis. Am J Emerg Med 1993;11:471-2. dehydration [4] Oski JA. A 5-week-old infant with cyanosis, diarrhea, and dehydration. level Case report. Curr Opin Pediatr 1995;7:73-6. (% [IQR]) [5] Lanir A. Transient methemoglobinemia with acidosis and hyper- Mean 11 F 312F 5 P = .68 chloremia in infants. Am J Pediatr Hematol Oncol 1986;8:353-5. a anion gap (6 to À8) [6] Shaw KN. Dehydration. In: Fleisher GR, Ludwig S, Editors. Textbook Mean pH 7.25 F 0.09 7.34 F 0.07 P = .25 of pediatric emergency medicine, 4th ed. Philadelphia7 Lippincott, (0.13 to À0.29)a Williams & Wilkens; 2000. p. 197-202. Mean 10.3 F 1.2 11.8 F 2.2 P = .13 [7] Bricker T, Jefferson LS, Mintz AA. Methemoglobinemia in infants with hemoglobin (0.7 to À3.7)a enteritis [Letter]. J Pediatr 1983;102:161. (g/dL) [8] Danish EH. Methemoglobinemia in infants with enteritis [Letter]. J Pediatr 1983;102:162-3. Mean 114 F 11 106 F 5 P = .033 a [9] Hanugloku A, Fried D, Dodner D. Methemoglobinemia in infants with chloride (15 to 1) enteritis [Letter]. J Pediatr 1983;102:161-2. (mEq/dL) [10] Lebby T, Roco JJ, Acinue EL. Infantile methemoglobinemia associated Total 3 24 with acute diarrheal illness. Am J Emerg Med 1993;11:471-2. a The 95% CI of the difference in the means. [11] Seeler RA. Methemoglobinemia in infants with enteritis [Letter]. J Pediatr 1983;102:162. [12] Pollack ES, Pollack CV. Incidence of subclinical methemoglobinemia in infants with diarrhea. Ann Emerg Med 1994;24:652-6. Despite the often described MHb etiology of oxidant [13] Yano SS, Danish EH, Hsia YE. Transient methemoglobinemia with stress from enteritis, it remains unclear why an otherwise acidosis in infants. J Pediatr 1982;100:415-8. healthy infant with diarrhea develops MHb when another [14] Wright RO, Lewander WJ, Woolf AD. Methemoglobinemia: etiology, does not. Traditionally, acidemia from dehydration has pharmacology, and clinical management. Ann Emerg Med 1999; been thought to be the underlying factor in these children 34:646-56. [7-13]. Despite this, a single inciting factor for methe- moglobinemia in these patients has not been identified Robert J. Freishtat, MD, MPH and several factors are thought to play a contributory Division of Emergency Medicine role. First, by virtue of enzymatic immaturity, infants Children’s National Medical Center have only slightly more than half of the functioning MHb The George Washington University reductase levels of healthy adults [14]. In addition, fetal School of Medicine and Health Sciences hemoglobin is more easily oxidized than adult hemoglo- Washington, DC 20010-2970, USA bin. Finally, the high pH level in an infant’s gut promotes Center for Genetic Medicine Research the growth of nitrite-producing flora that can lead to Children’s National Medical Center added oxidative stress. The George Washington University In this investigation, we identified a potentially clinically School of Medicine and Health Sciences important, but not statistically significant, difference in the Washington, DC 20010-2970, USA prevalence of sMHb and degree of elevation between an ED E-mail address: [email protected] group of dehydrated infants and a group of well-hydrated infants. The lack of statistical significance for this finding is James M. Chamberlain, MD prone to type II error given the small sample size. Of note, it Christina M.S. Johns, MD is not clear whether the elevated MHb levels in either group Stephen J. Teach, MD, MPH are causal for morbidity or merely an indicator of disease Division of Emergency Medicine status. The lack of a true control group (ie, healthy infants) Children’s National Medical Center is an additional limitation. The George Washington University Previous studies of infantile MHb had included obvious- School of Medicine and Health Sciences ly ill or hospitalized patients [12]. These had only limited Washington, DC 20010-2970, USA 576 Correspondence
Cynthia Ronzio, PhD Center for Health Services Research Children’s National Medical Center The George Washington University School of Medicine and Health Sciences Washington, DC 20010-2970, USA
Melissa M. Murphy-Smith, MPH Neelam Gor, BS The George Washington University School of Medicine and Health Sciences Washington, DC 20010-2970, USA doi:10.1016/j.ajem.2004.12.008
Common causes of hemiballism
To the Editor,
Hemiballism is a rare hyperkinetic movement disorder, characterized by irregular, wide amplitude and vigorous involuntary movements of the unilateral limbs. The term ballismus is derived from the Greek word meaning bto throwQ because the abnormal movements resemble the motions of throwing. The movements are usually continu- ous but may be intermittent and can be voluntarily restrained by the patient, although only for a few minutes. They are most prominent during periods of rest but are absent during sleep. It may occur with other types of involuntary movements, such as dystonia, myoclonus, or orofacial gestures [1]. Hemiballism is often used interchangeably but distin- Fig. 1 A, The axial diffusion-weighted MR image shows small guished from hemichorea by the fact that hemichoreic high-signal ovoid lesion in the left basal ganglia suggesting lacunar movements are slower, more randomly distributed, less infarction. B, The coronal T2-weighted MR image shows the violent, and primarily involve distal musculature [2]. Some lesion located in the left STN. patients with hemiballism also have choreiform movements; therefore, the bhemiballism-hemichoreaQ is often used to hospital day with improved condition. The involuntary describe this clinical spectrum [2-5]. movements completely disappeared after 4 months. We present 2 cases of hemiballism, 1 subthalamic A 76 year-old woman presented with left-sided involun- nucleus (STN) infarction and 1 nonketotic hyperglycemia, tary movement for 7 days. She has known hypertension on with a literature review. intermittent medication for several years, but not known A 69 year-old woman was brought to the ED because of diabetes. She had not taken neuroleptic drugs and there was right-sided involuntary movements. The medical history no family history of movement disorder. The irregular but included diabetes and hypertension on medications for persistent movements were started on her lips and extended 10 years. The unintentional, intermittent flinging move- to arm and leg. It resembled rolling and kicking motion, and ments of her arm and leg started several days before aggravated when excited, but resolved when sleeping. She admission. The movements were most prominent during was alert and her vital signs included blood pressure of periods of rest but were absent during sleep. The arm was 150/80 mm Hg and heart rate of 84 beats per minute. On more severely involved than the leg. Vital signs on neurologic examination there were no other unusual admission included blood pressure of 120/80 mm Hg and findings. The initial fasting glucose was 177 mg/dL and pulse rate of 64 beats per minute. She was alert and oriented. hemoglobin A1c level was 10.1%. The urinalysis showed There was no other remarkable neurologic abnormality glucose 3+ and no ketone. The MRI showed high signal except the involuntary movements. The magnetic resonance intensity in the right basal ganglia on the T1-weighted imaging (MRI) showed lacunar infarction of the left STN image (Fig. 2). She had taken haloperidol orally and insulin (Fig. 1). The movements were decreased with the admin- infusion. She was referred to the neurology department istration of oral haloperidol. She was discharged on the sixth under the impression of hemiballism associated with Correspondence 577
[9]. The pathogenesis and the unilaterality of the lesion despite of its metabolic origin are uncertain. The MRI is the imaging study of choice for hemiballism because the STN is a small structure and the sensitivity of computed tomography scan for lacunar infarct is low. It is also useful in the hyperglycemic hemiballism. They have characteristic hyperintense lesion in the contralateral stria- tum on T1-weighted MR image (Fig. 2). All reported cases show putaminal lesion and similar changes are found variably in the globus pallidus and head of caudate [2]. The nature of high signal intensity in the striatum on the T1- weighted images has been the subject of considerable controversy, including petechial hemorrhage, demyelin- ation, or swollen astrocytosis [10]. The T2-weighted images in the corresponding area are more variable [11,12]. The computed tomography scan is reported to show high densities in some cases [10]. In case 2, the patient has a not so high glucose level of 177 mg/dL, but had confirmed hyperglycemic hemiballism Fig. 2 The axial T1-weighted MR image shows hyperintensities because of the typical striatal hyperintensities on T1- in the right caudate nucleus, the putamen, and the globus pallidus weighted images. The initial glucose levels have been without edema or mass effect. reported variously and appear not to be proportional to the severity of hyperglycemic hemiballism [11-13], whereas the glycosylated hemoglobin A1c, which reflects recent glucose nonketotic hyperglycemia. The involuntary movement was status, is always increased. Even a diabetic patient with improved within the fifth hospital day. hemiballism who have low glucose level has typical lesions Hemiballism is reported to be most commonly due to a on MRI [11]. Therefore, the MRI is useful to differentiate lesion in the STN on the side contralateral to the involved whether the cause of hemiballism is hyperglycemia or not. limbs as seen in case 1. The STN is a lens-shaped structure The treatment and prognosis of the hemiballism depends located along the medial and cephalad margin of the on their etiologies. As hemiballism can be life threatening, peduncular portion of the internal capsule [3]. The STN early recognition of its cause is important. Patients who has a tonic inhibitory influence on the thalamus by means of have severe ballism may injure the affected limbs by excitatory projections onto inhibitory neurons in the medial striking walls and bedrails. So, padding of the limb may globus pallidus. Therefore, a destructive lesion in the STN be important to prevent injury in severe cases. The long- could result in disinhibition of these excitatory pathways, term prognosis of vascular hemiballism is similar to that of resulting in production of hemiballistic movements [2].Itis other stroke patients [5]. In case of hyperglycemic hemi- usually caused by lacunar infarction but may also result ballism, the clinical and radiological abnormalities im- from various conditions such as neoplastic, infectious, proved on control of blood glucose levels. Resolution of traumatic, iatrogenic, or metabolic origin [1,2,6]. the lesions on the imaging studies may be slower than the The lesion outside the STN is also reported to develop clinical improvement [13]. However, atypical cases have hemiballism, at points along the afferent or efferent path- been reported with delayed onset after resolution of the ways connecting the STN to its projection areas, which hyperglycemia, unremitting severe movements, and late include the globus pallidus, putamen, thalamus, and recurrence [14]. Most patients respond well to the dopamine brainstem [1,2,7]. Surprisingly, recent review of modern antagonists, such as haloperidol or phenothiazines, because case series has reported that the lesion outside STN of their antidopaminergic activity resulting in inhibition of accounted for more than half the hemiballism [2]. The STN neuronal firing [1,2]. However, pharmacologic therapy striatum (caudate nucleus and putamen) is reported to be the is not always necessary because some patients have mild most common non-STN site [8] and usually associated with and self-limited course. hyperglycemia as seen in case 2. When an emergency physician encounters a patient with The hyperglycemic hemiballism is now regarded as the hemiballism, he should keep in mind common causes of second most commonest cause of hemiballism [2]. It tends hemiballism such as STN infarction and nonketotic hyper- to be more prevalent in Asian elderly women with poorly glycemia. The MRI should be considered to look for lesions controlled type 2 diabetes mellitus. The hemiballism may be in the contralateral STN or striatal hyperintensity on T1- the only manifestation of hyperglycemia [4]. Most patients weighted images. Correction of underlying hyperglycemia have nonketotic hyperglycemia, although a patient with with temporary use of haloperidol, if necessary, will lead to hemiballism associated with ketoacidosis has been reported clinical improvement. 578 Correspondence
References Group A streptococcal necrotizing fasciitis: reducing the risk of unwarranted litigation [1] Dewey Jr RB, Jankovic J. Hemiballism-hemichorea. Clinical and pharmacologic findings in 21 patients. Arch Neurol 1989;46:862-7. To the Editor, [2] Postuma RB, Lang AE. Hemiballism: revisiting a classic disorder. Lancet Neurol 2003;2:661-8. [3] Provenzale JM, Schwarzschild MA. Hemiballismus. AJNR Am J Invasive group A streptococcal (GAS) disease is a Neuroradiol 1994;15:1377-82. disease of public health importance in the United States. [4] Lietz TE, Huff JS. Hemiballismus as a presenting sign of hypergly- The annual incidence of invasive GAS disease is 4 times cemia. Am J Emerg Med 1995;13:647-8. that of meningococcal disease [1]. The case-fatality rates of [5] Ristic A, Marinkovic J, Dragasevic N, et al. Long-term prognosis of GAS necrotizing fasciitis (NF) and streptococcal toxic vascular hemiballismus. Stroke 2002;33:2109-11. [6] Vidakovic A, Dragasevic N, Kostic VS. Hemiballism: report of 25 shock syndrome, 2 severe manifestations of invasive GAS cases. J Neurol Neurosurg Psychiatry 1994;57:945-9. disease, are high (20% and 45%, respectively). [7] Provenzale JM, Glass JP. Hemiballismus: CT and MR findings. Clinicians involved with the treatment of GAS NF are at J Comput Assist Tomogr 1995;19:537-40. risk for malpractice claims [2]. The rapidity with which GAS [8] Lang AE. Persistent hemiballismus with lesions outside the subthala- NF can progress from a minor condition to a life-threatening mic nucleus. Can J Neurol Sci 1985;12:125-8. [9] Romero Blanco M, Joao G, Monteiro E. Hemichorea induced by situation may trigger a lawsuit even though the patient has diabetic ketoacidosis and striatal hyperdensity on computerized axial received appropriate care [2]. The rapid diagnosis of GAS tomography. Rev Neurol 2002;34:256-8. NF can be difficult. Unrelenting pain out of proportion to the [10] Shan DE, Ho DM, Chang C, et al. Hemichorea-hemiballism: an physical findings is an important symptom of GAS NF [3]. explanation for MR signal changes. AJNR Am J Neuroradiol 1998; Competent physicians may initially miss the diagnosis of 19:863-70. [11] Lee BC, Hwang SH, Chang GY. Hemiballismus-hemichorea in older GAS NF because several of the signs and symptoms of GAS diabetic women: a clinical syndrome with MRI correlation. Neurology NF may mimic those of a muscle strain, influenza [3],or 1999;52:646-8. gastroenteritis due to a foodborne pathogen. Another reason [12] Lee EJ, Choi JY, Lee SH, et al. Hemichorea-hemiballism in primary why competent physicians may initially miss the diagnosis diabetic patients: MR correlation. J Comput Assist Tomogr of GAS NF is self-medication by the patient with 2002;26:905-11. [13] Lai PH, Tien RD, Chang MH, et al. Chorea-ballismus with nonketotic nonsteroidal antiinflammatory drugs (NSAIDs). Nonsteroi- hyperglycemia in primary diabetes mellitus. AJNR Am J Neuroradiol dal antiinflammatory drug may delay presentation to the 1996;17:1057-64. emergency department. [14] Ahlskog JE, Nishino H, Evidente VG, et al. Persistent chorea The goal of this analysis is to provide emergency medicine triggered by hyperglycemic crisis in diabetics. Mov Disord 2001; specialists who have treated patients with GAS NF with data 16:890-8. from a population-based study of invasive GAS disease that may reduce the risk of an unwarranted malpractice lawsuit. Hahn-Shick Lee, MD Data from a retrospective study of invasive GAS Department of Emergency Medicine infections conducted in Florida were analyzed. Details of Yongdong Severance Hospital the original study are found elsewhere [4]. These cases were Yonsei University Medical College 195 patients who were hospitalized throughout Florida Seoul, 135-270, South Korea between 1996 and 2000 for invasive GAS disease and reported to the Florida Department of Health, Tallahassee, Fla. Invasive GAS disease is a reportable condition in Seung-Whan Kim, MD Florida. For the current analysis, data on 42 patients with Department of Emergency Medicine GAS NF were extracted from the main database. Chungnam National University Hospital Table 1 shows the frequency distribution of the admit- Daejeon, 301-721, South Korea ting diagnosis listed on the Florida Department of Health
In-Sool Yoo, MD Department of Emergency Medicine Table 1 Admitting diagnosesa of 42 patients hospitalized Chungnam National University Hospital throughout Florida for GAS NF Daejeon, 301-721, South Korea Admitting diagnosis No. (%) Cellulitis 13 (31.0) Sung-Pil Chung, MD Cellulitis with sepsis 2 (4.8) Department of Emergency Medicine Invasive GAS disease or toxic shock syndrome 2 (4.8) Chungnam National University Hospital NF 4 (9.5) Daejeon, 301-721, South Korea Sepsis/rule out sepsis/septic shock/septicemia 10 (23.8) E-mail address: [email protected] Other diagnoses 11 (26.2) 42 a As noted on a health department case report form. doi:10.1016/j.ajem.2004.12.014 Correspondence 579 surveillance case report form of 42 cases of GAS NF. Few the individual drugs categorized in this study and their patients (9.5%) had NF listed as an admitting diagnosis. A effects on gastric motility. We agree with the authors’ total of 13 patients (31%) had an admitting diagnosis of conclusion that poisoning from drug overdose (in general) cellulitis noted on the health department case report form. results in hypomotility and delayed gastric emptying. Twenty-five percent of these patients (8/32) had diarrhea The power of the study would be improved if individual at the time of admission or developed it within 48 hours of drug categories were eliminated and each patient was used as admission. (Ten of these patients with GAS NF had missing their own control. In addition, as gastric instrumentation has values for this variable). been shown to delay emptying, the authors should consider Limited data were available on the recent use of NSAIDs by publishing the data that compare patients treated with and these GAS NF cases. Twenty percent of these patients (5/25) without gastric lavage [2]. We hope that this work will help had used NSAIDs after the onset of invasive GAS disease. advance the discussion of gastrointestinal decontamination Investigators have recently discovered a host genetic and cast doubt on the dogmatic approach in patients factor (a certain human leukocyte antigen class II haplotype) presenting more than 1 hour after ingestion [3,4]. that strongly predisposes certain individuals to develop GAS NF or streptococcal toxic shock syndrome after infection References with GAS [5]. In this situation, it is conceivable that rapid and appropriate treatment may do little to reduce the chance [1] Adams BK, Mann MD, Aboo A, Isaacs S, Evans A. Prolonged gastric of mortality. emptying half-time and gastric hypomotility after drug overdose. Am J If GAS NF is suspected, then rapid consultations with Emerg Med 2004;22:548-54. surgeons and infectious disease clinicians are imperative. [2] Read NW, Al Janabi MN, Bates TE, Barber DC. Effect of Even though penicillin and clindamycin are the antibiotics gastrointestinal intubation on the passage of a solid meal through the stomach and small intestine in humans. Gastroenterology 1983;84: of choice in this setting [4], surgery remains the foundation 1568-72. of treatment for GAS NF [2]. [3] Vale JA, Kulig K. American academy of clinical toxicology; European association of poisons centres and clinical toxicologists. Position paper: gastric lavage. J Toxicol Clin Toxicol 2004;42:933-43. References [4] Chyka PA, Seger D. Position statement: single dose activated charcoal. American Academy of Clinical Toxicology (AACT); European Associ- [1] Centers for Disease Control. Summary of notifiable diseases—United ation of Poison Centers and Clinical Toxicologists (EAPCCT). J Toxicol States, 2000. MMWR Morb Mortal Wkly Rep 2002;49:xix, 3. Clin Toxicol 1997;35(7):721-41. [2] Fink S, Chaudhuri TK, Davis HH. Necrotizing fasciitis and malpractice claims. South Med J 1999;92:770-4. [3] Bisno AL, Cockerill III FR, Bermudez CT. The initial outpatient- Klementyna Breyer, MD physician encounter in group A streptococcal necrotizing fasciitis. Clin Department of Emergency Medicine Infect Dis 2000;31:607-8. Christiana Care Health System [4] Mulla ZD, Leaverton PE, Wiersma ST. Invasive group A streptococcal Newark, Del, USA infections in Florida. South Med J 2003;96:968-73. [5] Kotb M, Norrby-Teglund A, McGeer A, et al. An immunogenetic and molecular basis for differences in outcomes of invasive group A S. Eliza Halcomb, MD streptococcal infections. Nat Med 2002;8:1398-404. Nicole C. Bouchard, MD Zuber D. Mulla, MSPH, PhD Robert S. Hoffman, MD Division of Epidemiology New York University School of Medicine School of Public Health New York, NY, USA University of Texas Health Science Center at Houston New York City Poison Control Center El Paso, TX 79902, USA New York, NY 10016, USA doi:10.1016/j.ajem.2004.12.016 Paul Dohrenwend, MD Department of Emergency Medicine Finally, a paper documenting the delay Morristown Memorial Hospital Morristown, NJ, USA To the Editor, Mary Ann Howland, PharmD We compliment Adams et al [1] for their elegant use of College of Pharmacy scintography to verify the commonly held belief that St. John’s University significant delays in gastric emptying occur after intentional New York City Poison Control Center overdose. Given that histories obtained from suicidal New York, NY 10016, USA patients are inherently unreliable and that toxicology screening was not obtained in a standardized manner, we DOI of original article: 10.1016/j.ajem.2004.08.017 caution readers not to overinterpret the association between doi:10.1016/j.ajem.2005.02.047 580 Correspondence
Oligoanalgesia in ED patients with isolated extremity aggressive pain management than those without fractures, injury without documented fracture despite similar pain scores. This suggests that physicians perceive soft-tissue injuries as less painful than fractures. In To the Editor, fact, in this group of patients, those with soft-tissue injuries reported a slightly higher degree of subjective discomfort. In the Effective management of pain is a fundamental aspect to search for objective evidence of painful injury, physicians may the humane provision of emergency care. Multiple studies look to laboratory and x-ray results. The point is that the have identified oligoanalgesia as a significant problem in subjective feeling of pain reported is highly variable upon the emergency medicine [1,2]. Pain is a subjective entity that physiological and psychological make-up of the host, and manifests differently based upon specific injury, disease patients in pain may have no objective sign of injury or disease. process, and the patients’ physiological and psychological The data also show that patients are being discharged make-up [3]. Therefore, a similar anatomic injury or disease from the ED in a significant pain. Wilson and Pendleton [1] process may produce a very different sensation of pain found that of 198 patients presenting with painful con- depending on the host. ditions, only 44% of patients received analgesia while in the In the process of diagnosing injury, physicians seek ED. Our data collected 14 years after this study show an objective indicators of pain such as the presence of an acute improvement in care, but still 25% of patients with fractures fracture. We hypothesized that in patients with acute and more than 40% of patients with soft-tissue injury did not extremity injuries, those with radiographically diagnosed receive analgesia while in the ED. fractures would receive more aggressive pain management Lessons that can be learned from this study include the compared with those with only soft-tissue injuries. reinforcement of the principle of taking patients’ report of Patients with isolated acute extremity injury were asked pain at face value and to ensure that patients have achieved by research assistants to rate their pain on a visual analog an adequate level of pain relief before ED discharge. In scale at ED admission and discharge. Demographic data and addition, physicians must be more aware even in the pain medicines administered in the ED and at discharge absence of documented injury on x-ray that patients still were concurrently recorded from the medical record. report significant pain. Also, a determination must be made Patients who received an x-ray of an extremity and who about the need for ongoing pain control after discharge in could consent were included. the form of a limited narcotic prescription with arrangement One hundred three patients with acute extremity injuries for follow-up or at least the recommendation for or were enrolled. Thirty-one had fractures. Initial 10-point prescription of nonsteroidal anti-inflammatory medication. visual analogue scale pain scores were not significantly different ( P = .68) for the fracture and nonfracture groups at ED admission (6.0 vs 6.2, P = .68) or ED discharge (4.2 vs References 4.5, P = .60). The differential between admission pain and discharge was also not significantly different ( P = .80). [1] Wilson JE, Pendleton J. Oligoanalgesia in the emergency department. A higher percentage of patients with fractures received Am J Emerg Med 1989;7:620-3. [2] Ngai B, Ducharme J. Documented use of analgesics in the emergency narcotic analgesia in the ED compared with those without department and upon release of patients with extremity fractures. Acad fracture (45% vs 36%); however, this did not reach Emerg Med 1997;4:1176-8. predetermined statistical significance ( P = .51). Patients [3] Ader R. Conditioning factors in pharmacological pain research. with fractures also received nonnarcotic analgesia more Proceedings Annual Meeting of the American Pain Society, San Diego; frequently than those without (32% vs 22%, P = .32). 1998. [4] Todd K, Lee T, Hoffman J, et al. The effect of ethnicity on physician Patients with fractures were more likely to receive a narcotic estimates of pain severity in patients with isolated extremity trauma. prescription (52% vs 39%, P = .20) at ED discharge. JAMA 1994;271:925-8. These data corroborate published reports on pain [5] Jones JS, Johnson K, McNich M. Age as a risk factor for inadequate management in the ED [1,2]. Certain populations are at emergency department analgesia. Am J Emerg Med 1996;14:157. greater risk for oligoanalgesia such as pediatric patients, critically ill patients, and patients who have an impaired Jesse M. Pines, MD, MBA ability to communicate. Ethnicity, age, and sex have also Department of Emergency Medicine been shown to play a role in the provision of analgesia [4,5]. University of Pennsylvania These data suggest that those with traumatic extremity Philadelphia, PA 19104, USA injury without documented fracture may also be at risk. Although not a statistically significant result, our study Andrew D. Perron, MD showed that patients with fractures tend to receive more Department of Emergency Medicine Maine Medical Center Portland, ME 04102-3175, USA This paper was presented under the same title as part of ACEP Research Forum, October 12 to 13, 2003, Boston, Mass. doi:10.1016/j.ajem.2004.11.006 Correspondence 581
Slow supraventricular tachycardia in a patient on drawal, bronchitis, and pneumonia. With a 2-week history b-blocker therapy of cough, worse at night with no fever or chills, this patient was first worked up for a respiratory illness. After a negative To the Editor, chest x-ray, he was worked up for a pulmonary embolism, which was also eventually ruled out. Tests for myocardial A 77-year-old man presented to the emergency room with ischemia were negative, and his electrocardiogram failed to a heart rate (HR) of 130 beats per minute and a 2-week history give any clues to an etiology as his PR interval and QRS of cough that was worse at night. His current medications interval were both within the reference range and of normal included metoprolol for hypertension and, although not a morphology. The patient also lacked common presenting current smoker, he did have a 30-year smoking history. symptoms of SVT. These include palpitations, dizziness, He described no exacerbation with exercise, and his physi- shortness of breath, syncope, chest pain, fatigue, diaphore- cal examination was negative for jugulovenous distension, sis, and nausea [1,2]. However, the main factor in excluding calf edema, fevers, wheeze, or neck mass. His chest x-ray was SVT as a diagnosis initially was the relatively slow HR of negative. The patient’s pulse oximetry was steady at 95%, 130 beats per minute. and his electrocardiogram demonstrated possible left ven- SVT is most often associated with a more rapid HR tricular hypertrophy, a regular rhythm, a PR interval of 148 (N140 beats per minute) [1,3]. We attribute the slow SVT in milliseconds, a QRS interval of 92 milliseconds, and a HR of this patient to his chronic b-blocker therapy. b-Blockers 130 beats per minute. He was given 500 mL of normal saline inhibit sympathetic action at the heart by competing at without a change in his pulse rate. Given the pulse oximetry the b1-adrenergic receptor and have a negative inotropic reading, the tachycardia, and the cough, the possibility of and a negative chronotropic effect. The slowing of the pulmonary embolism was entertained. A d-dimer test and a reentry mechanism that our patient experienced was enough computed tomography pulmonary angiogram were done, and to disguise itself as a tachycardia caused by some other both were normal. Troponin I was negative, and his chemistry mechanism. It was not until the patient bcuredQ himself by panel was normal except for a slightly elevated blood glucose crawling into his bed and creating a vagal response to an (117 mg/dL) and blood urea nitrogen (23 mg/dL). After increase in pressure that the SVT revealed itself. Thus, a routine trip to the bathroom, the patient was swing- bslowQ SVT should be at least considered in patients ing his legs up onto the gurney, and his HR dropped to 90 with sustained tachycardias of unclear cause who are on beats per minute (see Fig. 1). By exerting an effort to lift his b-blocker therapy. legs and flexing his abdominal muscles, the patient had performed a Valsalva maneuver, which terminated the tachycardia. The diagnosis of supraventricular tachycardia References (SVT) was subsequently made. [1] Baerman JM, Morady F, DiCarlo Jr LA, de Buitleir M. Differentiation Sustained tachycardia in the elderly can have many of ventricular tachycardia from supraventricular tachycardia with etiologies. Common causes include congestive heart failure, aberration: value of the clinical history. Ann Emerg Med 1987;16(1): pulmonary embolism, myocardial infarction, alcohol with- 40-3.
Fig. 1 Rhythm strip. 582 Correspondence
[2] Robinson AW. Common varieties of supraventricular tachycardia: 5 mEq/L, arterial pH 7.06, and LA of 70 mg/dL. Again, she differentiation and dangers. Heart Lung 1996;25(5):373-81. remained in stable condition and was feeling well after [3] Herbert ME, Votey SR, Morgan MT, Cameron P, Dziukas L. Failure to intravenous fluids and bicarbonate therapy, but her LA agree on the electrocardiographic diagnosis of ventricular tachycardia. persisted above 60 mg/dL for 1 week. We then administered Ann Emerg Med 1996;27(1):35-8. riboflavin (50 mg) and thiamine 1 mg intravenously, and her LA subsequently normalized over 2 days, and she was Steve Allison, BS discharged to home. At a follow-up visit 2 weeks later, she Sean O. Henderson, MD remained well. Department of Emergency Medicine The NRTIs are nucleotide analogues lacking the 3V-OH LAC+USC Medical Center group, to which the 5V-PO4 group of the next nucleotide to Los Angeles, CA 90033, USA be added is joined, so that when they are incorporated into E-mail address: [email protected], [email protected] newly transcribed viral DNA, they act as chain terminators, inhibiting viral RNA-directed DNA polymerase (reverse doi:10.1016/j.ajem.2004.11.008 transcriptase). Mammalian DNA polymerase-a is not affect- ed, but DNA polymerase-c, involved in mitochondrial DNA replication, is inhibited by NRTIs. When the mitochondrial Lactic acidosis caused by nucleoside analogues DNA is poisoned, 2 important catabolic pathways become impaired: (1) oxidative phosphorylation, causing a shift To the Editor, in pyruvate metabolism toward lactate, (2) fatty acid b-oxidation, causing triglyceride accumulation in hepato- Type B lactic acidosis, lactate accumulation without cytes, which are highly dependent on oxidative metabolism any clinical evidence of tissue hypoxia, occurs because of [1-4]. The symptoms of this syndrome—anorexia, nausea, impairment in lactate use. Type B is further divided into vomiting, and abdominal pain—are typically seen after 3 to subcategories based on underlying etiology. Type B1 is 15 months of NRTI therapy, but merely discontinuing the associated with systemic disease such as renal and hepatic drug does not reverse the disorder [3,5]. In addition to the failure, diabetes, and malignancy. Type B3 is associated metabolic acidosis that results from elevated LA, abnormal with inherited metabolic disorders. Type B2 is known to prothrombin time and liver enzymes may be found, but they be caused by several classes of drugs including bigua- may be only mildly elevated despite severe hepatic steatosis. nides, alcohols, iron, isoniazid, and salicylates. Since the Severe pancreatitis and diffuse weakness, caused by early 1990s, the nucleoside reverse transcriptase inhibitors myopathy from mitochondrial toxicity in these tissues, are (NRTIs) have also been implicated in causing this serious also common features in this syndrome [1,3,5]. Skeletal syndrome at a rate of 1.3 per 1000 person-years of drug muscle samples will have pathognomonic features including therapy [1,2]. The mortality rate is high and may be up red-ragged fibers on microscopy, mitochondrial inclusions, to 80% in those patients with lactic acid (LA) levels and reduced amounts of mitochondrial DNA [6,7]. greater than 10 mmol/L [2,3]. The following case Underlying nutritional deficiencies in the cofactors presentation and review of the literature will discuss the required for cellular respiration, other defects in oxidative pathophysiology, clinical features, and potential therapy enzymes, or genetic factors may predispose a subset of for this disorder. patients for this disorder [4]. Therapy thus far has been A 56-year-old woman HIV-positive for 7 years, with a aimed at attempting to rescue mithochondrial metabolism CD4 cell count of 268 cells/microliter and undetectable viral by supplying necessary cofactors. Riboflavin (B ), the load, presented complaining of 1 week of vomiting and 2 precursor of flavin mononucleotide and flavin adenine diffuse abdominal pain; her medications included the dinucleotide, required in the Krebs cycle, electron transport NRTIs, stavudine (d4T) and lamivudine (3TC). She was chain, and oxidation of fatty acids, has been reported found to have an anion gap of 27 mEq/L, serum bicarbonate effective in treating several other cases of NRTI-induced of 8 mEq/L, arterial pH 7.20, and LA level of 63 mg/dL lactic acidosis and has been studied more extensively in (reference range 5-15 mg/dL). Remaining hemodynamically pediatric patients with inherited mitochondrial disorders stable, she clinically improved with intravenous fluids and and in animal models [8-12]. Thiamine (B ), a cofactor for bicarbonate infusions despite her LA continuing to be 1 pyruvate dehydrogenase, which converts pyruvate to elevated (50-60 mg/dL); workup was only significant for acetyl-CoA (instead of lactate), has also been used in fatty infiltration of the liver by computed tomography scan. treating congenital mitochondrial disorders and could When tolerating a regular diet on day 6, she was discharged, theoretically be useful in this syndrome [3,11]. Lastly, a but she returned 4 days later with the same symptoms and deficiency of l-carnitine, an important cofactor for the was found to have an anion gap of 30 mEq/L, bicarbonate of intramitochondrial transport of fatty acids, is thought to result from NRTIs and may also contribute to the acidosis Presented as a poster at the 2001 Research Day for the American and steatosis. It has been used previously in treating College of Physicians, Maryland Chapter. inherited oxidative disorders, and there was one case report Correspondence 583 of successfully treating NRTI-induced lactic acidosis with with uncommon manifestations, such as behavioral change, l-carnitine supplementation [3,13]. sensorimotor deficit, psychosis, or hallucinosis [1], which make it difficult to establish an accurate diagnosis abruptly. References We add an unusual flaccid bladder caused by detrusor hyporeflexia as an early symptom for encephalitis. [1] Brinkman K, Hofstede HJ, Burger DM, Smeeitink JA, Koopmans PP. A 24-year-old woman caught common cold and had Adverse effects of reverse transcriptase inhibitors: mitochondrial an acute onset of urinary retention 4 days later. Urinary toxicity as common pathway. AIDS 1998;12:1735-44. tract infection was diagnosed based on a presence of [2] Falco V, Crespo M, Ribera E. Lactic acidosis related to nucleoside 8 leukocytes per high-power field in urinalysis. While she therapy in HIV-infected patients. Expert Opin Pharmacol 2003;4(8): was hesitating for further procedure, intermittent fever, 1321-9. [3] Carr A, Cooper DA. Adverse effects of antiretroviral therapy. Lancet neck stiffness, vomiting, and mental change subsequently 2000;356(9239):1423-30. developed on the following 2 days. On presentation, she [4] Miller KD, Cameron M, Wood LV, Dalakas MC, Kovacs JA. Lactic was drowsy. Glasgow Coma Scale was E2V2M4. Blood acidosis and hepatic steatosis associated with use of stavudine: report pressure, pulse rate, and respiratory rate were normal, but of four cases. Ann Intern Med 2000;133:192-6. body temperature was 37.88C. Neck was stiff. There was [5] Sundar K, Suarez M, Banogon PE, Shapiro JM. Zidovudine-induced fatal lactic acidosis and hepatic failure in patients with AIDS: report of no proptosis or conjunctival congestion. Abnormal neu- two patients and review of the literature. Crit Care Med 1997;25(8): rological findings included unreactive pupils, papilledema, 1425-30. generalized brisk tendon reflex, bilateral Babinski sign, [6] Araudo E, Dalakas M, Shanske S, et al. Depletion of muscle DNA in and snout reflex. Several episodes of tonic eyes deviation AIDS patients with zidovudine-induced myopathy. Lancet 1991; to left side were observed. Encephalitis was interpreted. 337(8740):508-10. [7] Asuncion JG, Olmo ML, Sastre J, et al. AZT treatment induces Cranial computer tomography revealed brain edema. molecular and ultrastructural oxidative damage to muscle mitochon- Mastoid and paranasal sinuses were clear. Electroenceph- dria: prevention by antioxidant vitamins. J Clin Invest 1998; alogram showed excessive slowing activity in diffuse. 102(1):4-9. Biochemistry, immunology, hematology, serology, VDRL [8] Fouty B, Frerman F, Feves R. Riboflavin to treat analogue-induced test, Treponema pallidum hemagglutination test, HIV test, lactic acidosis [Letter]. Lancet 1998;352:291-2. [9] Luzzati R, Del Bravo P, Di Perri G, Luzzani A, Concia E. Riboflavin and lupus study were normal except for leukocytosis 3 and severe lactic acidosis [Letter]. Lancet 1999;353:901-2. (12400 cells/mm ) and neutrophilia (78%). Cerebrospinal [10] Mokryycki MH, Harris C, May H, Laut J, Palmisano J. Lactic acidosis fluid analysis disclosed an opening pressure to be 250 associated with stavudine administration: a report of five cases. Clin mmCSF, cytology 141 cells/mm3, neutrophils 4%, glucose Infect Dis 2000;30:198-200. 44 mg%, total proteins 190 mg%, and immunoglobulin [11] Tanaka J, Nagai T, Arai H, Inui K, Yamanouchi H, Goto Y, et al. (Ig) G–link index 0.23. Culture for bacterium, fungus, or Treatment of mitochondrial encephalomyopathy with a combination tubercle bacillus was sterile. Virology showed an initial of cytochrome C and vitamins B1 and B2. Brain Dev 1997;19:262-7. [12] Vasseur BG, Kawanishi H, Shah N, et al. Type B lactic acidosis: a rare increase of IgM antibody but over 4-fold IgG antibody 2 complication of antiretroviral therapy after cardiac surgery. Ann weeks later to respiratory syncytial virus in serum and Thorac Surg 2002;74(4):1251-2. cerebrospinal fluid. Other viral antibody titer and poly- [13] Claessens YE, Cariou A, Chiche JD, Dauriat G, Dhainaut JF. l- merase chain reaction for Mycobacterium tuberculosis and Carnitine as a treatment of life-threatening lactic acidosis induced by herpes simplex virus were negative. Viral encephalitis nucleoside analogues [Correspondence]. AIDS 2000;14:472-3. was established. Osmotherapy was initiated. Her fever Joseph R. Shiber, MD subsided, and consciousness recovered rapidly, but dys- Department of Emergency Medicine uria still persisted. Urine was sterile, and residual urine East Carolina University was 350 mL. Cystometry revealed a severe decrease of Greenville, NC 27834, USA detrusor pressure upon bladder distention and micturition, Department of Internal Medicine suggesting detrusor hyporeflexia. No detrusor hyperre- East Carolina University flexia was seen. Urethral pressure and detrusor-sphincter Greenville, NC 27834, USA synergy were normal. Micturition was relied only on E-mail address: [email protected] abdominal strain (Fig. 1). Nerve conduction velocity study, F wave, H reflex, and somatosensory evolved doi:10.1016/j.ajem.2004.10.002 potential were normal. Her dysuria spontaneously recov- ered 1 month later. Detrusor hyporeflexia presents as an early In contrast to the previous concept that micturition is a manifestation in encephalitis spinal reflex, the pontine micturition center (PMC) is recently found to control micturitional process also [2,3]. To the Editor, An increase in bladder pressure activates the ascending afferent to PMC, from where it sends 2 distinct descending Encephalitis is an emergent condition urgent for rapid pathways to the lower urinary tract. The first one is and correct diagnosis. However, some patients may present parasympathetic pathway that innervates the detrusor 584 Correspondence
Fig. 1 Residual urine was 350 mL. Cystometry showed detrusor areflexia but normal urethral pressure upon bladder distention. No detrusor hyperreflexia or detrusor-sphincter dyssynergy was seen. Micturition was committed by abdominal strain. Pabd indicates abdominal pressure; Pdet, detrusor pressure; Pves, pelvic pressure; Pura, urethral pressure.
muscle, contracts, and empties the bladder to excrete urine bladder dysfunction is obscure. Early recognition is the rule [3]. The second one is the c-aminobutyric acid pathway, to reduce the mortality in acute brain disease. which relaxes the urethra and external sphincter and facilitates the urine outflow [3]. A high frequency of neurogenic bladder in hemispheric References stroke [4] supports a cortical regulation of the PMC [5]. However, bladder dysfunction has not been mentioned as [1] Swartz MN. Intracranial infections. In: Rosenberg RN, editor. 7 an early manifestation in case of neuroinfection as in our Neurology. New York Grune & Strattin, Inc.; 1980. p. 1-40. [2] Blaivas JG. The neurophysiology of micturition: a clinical study of patient. Tonic ocular deviation to left side and presence of 550 patients. J Urol 1982;127:958-63. snout reflex reflect an involvement of her right frontal lobe. [3] de Groat WC. Central neural control of the lower urinary tract. Ciba In fact, the right frontal cortex (anterior and medial Found Symp 1990;151:27-44. surface), left sensorimotor cortex, and bilateral supplemen- [4] Sakakibara R, Hattori T, Yasuda K, et al. Micturitional disturbance after tary motor area [6] are activated during micturition, acute hemispheric stroke: analysis of the lesion site by CT and MRI. J Neurol Sci 1996;137:47-56. suggesting a functional link between PMC and these areas. [5] Bradley WE. Cerebro-cortical innervation of the urinary bladder. Descending fibers may travel through the anterior edge of Tohoku J Exp Med 1980;131:7-13. the paraventricular white matter or genu of the internal [6] Fukuyama H, Matsuzaki S, Ouchi Y, et al. Neural control of micturition capsule [4] and are integrated by putamen, thalamus [4],or in man examined with single photon emission computed tomography cerebellum [7]. However, these anatomic sites cannot using 99mTc-HMPAO. Neuroreport 1996;7:3009-12. [7] Burney TL, Senapati M, Desai S, et al. Acute cerebrovascular accident predict the type of bladder dysfunction, except that detrusor and lower urinary tract dysfunction: a prospective correlation of the site hyporeflexia is more frequent in cerebellar lesion [7]. This of brain injury with urodynamic findings. J Urol 1996;156:1748-50. confliction is similar to pontine lesion that can provoke [8] Sakakibara R, Hattori T, Yasuda K, et al. Micturitional disturbance and detrusor areflexia or hyperreflexia [8], which may eventu- the pontine tegmental lesion: urodynamic and MRI analyses of vascular ally be determined by the severity of damage of excitatory cases. J Neurol Sci 1996;141:105-10. [9] Kalita J, Shah S, Kapoor R, et al. Bladder dysfunction in acute or inhibitory pathway. transverse myelitis: magnetic resonance imaging and neurophysiolog- An involvement of spinal cord may occasionally cause ical and urodynamic correlations. J Neurol Neurosurg Psychiatry detrusor hyporeflexia, but a negative electrophysiology 2002;73:154-9. study and an absence of other urodynamic finding and associated neurological deficit at lower limbs [9] may Wei Hsi Chen, MD, MSc be argued for her bladder dysfunction by lower spi- Yi Fen Kao, MD, BS nal involvement. Department of Neurology Preceding bladder disorder may be neglected or under- Chang Gung Memorial Hospital estimated in case of acute brain disease. Our patient reminds Kaohsiung 833, Taiwan us to consider central lesion promptly when the cause for Email-address: [email protected] Correspondence 585
Ming Chin Hsu, MD common formulation of aripiprazole ingested was the 15-mg Department of Neurology tablet commonly prescribed as initial therapy. There were Tze Che Buddhist Hospital 4 accidental ingestion and 4 intentional ingestion. The Chiayi, Taiwan average amount of aripiprazole ingested was 81.66 mg (data from 6 cases; in 2 cases, the dosage was unknown). In the doi:10.1016/j.ajem.2004.10.007 accidental ingestion group, 3 of the 4 patients were observed at home with poison center follow-up conducted at 24 hours. One-year experience with aripiprazole exposures All of these patients had favorable outcomes. The remaining patient of this group was observed in the emergency To the Editor, department (ED) and eventually discharged home also with a favorable outcome. In the group of intentional overdoses, Aripiprazole is a second generation, atypical antipsychot- all 4 were observed in the ED. Two of these 4 patients were ic medication used in the treatment of schizophrenia, schizo- admitted to the hospital for excessive somnolence. All of affective disorder, and acute bipolar mania. The mechanism these patients in this group demonstrated favorable out- of action of this drug is not entirely known; however, it has comes (Table 1). been shown to have high affinity agonistic activity at D2 In premarketing studies of the efficacy and safety of dopamine and serotonin 5-HT1A receptors, and antagonistic aripiprazole, there have been minimal reports of toxic activity at serotonin 5-HT2A receptors. Other affinities noted exposures. Among those that did have aripiprazole toxicity, in premarketing studies [1] demonstrated moderate affinities no one succumbed to the adverse effects. Areas of concern for dopamine D4, serotonin 5-HT2C, histamine H1,anda1- associated with aripiprazole toxicity as well as most adrenergic receptors. There have been no appreciable antipsychotic medications include cardiovascular derange- affinities for peripheral muscarinic cholinergic receptors. ments, central nervous system depression, and autonomic Aripiprazole toxicity has been reported in a minority of instability. The most concerning of these would be cases. In a premarketing study, of 5500 patients receiving neuroleptic malignant syndrome. In clinical trials with aripiprazole therapy, there were 7 reported overdoses. Two aripiprazole compared to placebo or other antipsychotics, of these patients reportedly ingested 180 mg of aripiprazole. there have not been any cases of neuroleptic malignant These patients were observed in the hospital and were syndrome. The most common adverse events associated provided supportive care only. One of the 2 patients who with aripiprazole overdose included orthostatic hypotension, ingested this dose reportedly suffered somnolence as well as likely secondary to the drugs action on a1-adrenergic nausea and vomiting [1]. No deaths have been reported from receptors. The cardiovascular effects observed were primar- aripiprazole overdose in the medical literature. ily rhythm disturbances, namely, QTc prolongation. These 2 This was a retrospective chart review study of patients situations required continuous cardiac monitoring and with reported aripiprazole toxicities over 1 year (2003). supportive care. Given the drug’s high protein binding and Data were compiled and reviewed for patient demographics, large volume of distribution (4.9 L/kg), hemodialysis would adverse outcomes, and clinical course associated with either not be efficacious in treating acute or chronic toxicity [1-3]. accidental or intentional overdoses. Our experience with aripiprazole is similar to that The inclusion criteria for this study consisted of either observed with larger clinical trials. Most of our patients accidental or intentional overdose of aripiprazole above the had little to no serious consequences associated with the recommended dosage (15 mg/d) and no known congestion. overdose. A common treatment of these patients simply In addition to these criteria, poison control center follow- involved supportive care with most patients remaining home up finally occurred until a final outcome was documented. as opposed to being admitted. Those patients who were Specifically, all patients were followed by the poison center observed in the ED had an uneventful clinical course. until the cessation of symptoms or for 24 hours if the A limitation to this study is the sample size. To make patients were asymptomatic. significant conclusions, the sample size should be larger. Reviewers received training in systematic chart review. A However, in review of the premarketing trial [1], the sample third reviewer acted as a tiebreaker. Ages, outcomes, signs, size of those with aripiprazole toxicity involved 7 patients and symptoms were recorded. (n = 5500). Regardless, a larger sample of those with The study population consisted of 8 patients with an aripiprazole toxicity would yield substantial data to draw average age of 23.75 years (range, 3-43 years). The most significant conclusions. This was a study based solely on data
Table 1 Patient group characteristics Patient group Average age (y) Average dose ingested (mg) Clinical course Outcome Accidental (n = 4) 23.75 81.66 3/4 Observed at home; 1/4 observed in the ED All favorable Intentional (n = 4) 23.75 81.66 All observed in ED; 2/4 admitted for All favorableM excessive somnolence 586 Correspondence collected from one poison control center (Good Samaritan Three-year experience with ziprasidone exposures Regional Poison Control Center, Phoenix, Ariz). The patient population was limited to those residing in this region. When To the Editor, reproducing this study, it would be advantageous to include multiple poison control centers to build the patient population Ziprasidone (ZIP) is an atypical antipsychotic medica- more as well as to sample various populations. tion approved for treatment of schizophrenia, acute Our outcomes are consistent with the literature on psychomotor agitation, and the treatment of bipolar aripiprazole adverse effects. The new second and third disorder. Although the side effects of ZIP during standard generation antipsychotic drugs have proven to be less clinical use are well described, there is a relative dearth of detrimental to patient outcomes in the setting of overdoses. information regarding toxicity in either accidental ingestion These newer drugs are not more effective than previous or intentional overdose. antipsychotics. Rather, they appear to have significantly less We conducted a retrospective chart review of isolated adverse effects. Aripiprazole is a promising drug in the ZIP ingestions reported to our Poison Center during 2001 treatment of schizophrenia, schizo-affective disorder, and to 2003. After a brief training, reviewers blinded to the acute bipolar mania. However, in the setting of toxic purpose of the study completed a standardized data exposures, this drug does exhibit potentially significant collection sheet. Age, outcomes, and amount ingested consequences and each overdose should be evaluated were recorded. individually with appropriate treatment rendered on the basis Approximately 150000 human exposures were reviewed. of these adverse effects. Thirty patients met the criteria. The average amount of ZIP ingested was 205.62 mg (data from 29 cases), and the average patient age was 22.7 years (range, 1-41 years). The References most common formulation ingested was the 20-mg tablet. Eight patients accidentally ingested ZIP, and 22 admitted [1] Marder SR, et al. Aripiprazole in the treatment of schizophrenia: safety intentional ZIP ingestion. Of the 8 accidental ZIP inges- and tolerability in short term, placebo-controlled trials. Schizophr Res tions, 7 were observed at home with Poison Control Center 2003;61(2-3):123-36. follow-up in 24 hours, all with good outcomes. The other [2] Nakai S. Diminished catalepsy and dopamine metabolism distinguish aripiprazolefrom haloperidol or risperidone. Eur J Pharmacol patient was observed in the ED and discharged home 2003;472(1-2):89-97. without incident. Nineteen of the intentional ingestions were [3] Sugiyama A. In-vivo canine model comparison of cardiohemodynamic observed in the ED with good outcomes, 1 patient was lost and electrophysiological effects of a new antipsychotic drug: aripipra- to follow-up, and 1 patient was evaluated by paramedics and zole (OPC-14597) to haloperidol. Toxicol Appl Pharmacol 2001;173(2): watched at home. One patient ingested an unknown quantity 120-8. of ZIP and presented to the ED comatose. Endotracheal intubation was performed and the patient recovered after Frank LoVecchio, DO, MPH 8 hours. Banner Good Samaritan Regional Poison Center As an atypical antipsychotic, the efficacy in treating Phoenix, AZ 85006, USA schizophrenia with ZIP is purportedly due to the Maricopa Medical Center antagonism of specific serotonin and dopamine receptor Department of Emergency Medicine subclasses and serotonin and norepinephrine reuptake Phoenix, AZ 85008, USA blockade. Ziprasidone is an a-1 adrenergic as well as AZ College of Osteopathic Medicine histamine (H1) antagonist, resulting in the side effects of Glendale, AZ, USA orthostatic hypotension and somnolence, although these E-mail address: [email protected] effects are less pronounced with ZIP use than with other atypical antipsychotics. Even in routine use, there is David Watts, MD unique concern for corrected QT interval prolongation, Banner Good Samaritan Regional Poison Center which appears dose-related and has been shown to be as Phoenix, AZ 85006, USA much as 10 to 20 milliseconds with the higher standard Maricopa Medical Center dosages [1] Although no case has yet been reported, this Department of Emergency Medicine produces concern for torsades de pointes, especially in an Phoenix, AZ 85008, USA overdose situation. Other complications noted to date include, but are not limited to, neuroleptic malignant Joe Winchell, MD syndrome, tardive dyskinesia, mania, and rhabdomyolysis. Maricopa Medical Center Often, these cases are difficult to assess given the Department of Emergency Medicine possibility of mixed ingestion, recent discontinuation of Phoenix, AZ 85008, USA other medication to initiate ZIP therapy, and a history of similar symptoms in patients initially treated with other doi:10.1016/j.ajem.2004.10.008 medications [2-5]. Correspondence 587
Accidental ingestions were typically a sub- or therapeutic [5] Yang S, McNeely M. Rhabdomyolysis, pancreatitis, and hyperglycemia dose, and with the relatively high safety profile of ZIP at with ziprasidone. Am J Psychiatry 2002;159:1435 normal doses, were without complication. Given the more substantial dosage of ZIP in intentional overdoses, most of Frank LoVecchio, DO, MPH these patients were monitored in the ED for adverse effects, Banner Good Samaritan Regional Poison Center as well as for further psychiatric intervention. We report Phoenix, AZ 85006, USA only 1 case requiring significant medical intervention, that Department of Emergency Medicine being secondary to the sedative properties of ZIP. In our Maricopa Medical Center experience, it appears that low-dose accidental ingestions AZ 85006, USA may be monitored at home with close follow-up, whereas E-mail address: [email protected] overdose situations may continue to benefit from observa- tion in the ED. However, given the relatively recent David Watts, MD introduction of ZIP, more information on exposures and Banner Good Samaritan Regional Poison Center management can be expected in the future. Phoenix, AZ 85006, USA Department of Emergency Medicine Maricopa Medical Center References Phoenix, AZ 85006, USA
[1] Carnahan R, Lund B, Perry P. Ziprasidone, a new atypical antipsychotic P. Eckholdt drug. Pharmacotherapy 2001;21:717-30. [2] Burton S, Heslop K, Harrison K, Barnes M. Ziprasidone overdose. Am Department of Emergency Medicine J Psychiatry 2000;157:835. Maricopa Medical Center [3] House M. Overdose of ziprasidone. Am J Psychiatry 2002;159:1061-2. Phoenix, AZ 85006, USA [4] Bryant SM, Zilberstein J, Cumpston KL, Magdziarz DD, Costerisan DD. A case series of ziprasidone overdoses. Vet Hum Toxicol 2003;45(2): 81-2. doi:10.1016/j.ajem.2004.09.003 American Journal of Emergency Medicine (2005) 23, 588
www.elsevier.com/locate/ajem
Erratum
Regarding the article bThe value of protocol-driven CT scanning in stab wounds to the head,Q which appeared in Am J Emerg Med 2002 Jul;20(4):295–7, by Exadaktylos AK, Stettbacher A, Bautz PC.
Dr Jane Terries, of St. Thomas’s Hospital, London, UK, was inadvertently left off the manuscript as an author when it was submitted to AJEM. The co-authors wish to acknowledge Dr. Terries as a co-author at this time to correct this oversight.
DOI of the original article: 10.1053/ajem.2002.33784
0735-6757/$ – see front matter D 2005 Published by Elsevier Inc. doi:10.1016/S0735-6757(05)00220-2