The American Journal of Emergency Medicine Contents VOLUME 25 • NUMBER 4 • MAY 2007

ORIGINAL CONTRIBUTIONS Expanding the Use of Topical Anesthesia in Wound Management: Sequential Layered Application 379 of Topical Lidocaine With Epinephrine Slava V. Gaufberg, Michael J. Walta, and Tom P. Workman Management of Severe Acute Pain in Emergency Settings: Ketamine Reduces Morphine 385 Consumption Michel Galinski, François Dolveck, Xavier Combes, Véronique Limoges, Nadia Smaïl, Veronique Pommier, François Templier, Jean Catineau, Frédéric Lapostolle, and Frédéric Adnet The Value of Serum Tau Protein for the Diagnosis of Intracranial Injury in Minor Head Trauma 391 Cemil Kavalci, Murat Pekdemir, Polat Durukan, Necip Ilhan, Mustafa Yildiz, Selami Serhatlioglu, and Dilara Seckin Frequency of Radiology Self-Referral in Abdominal Computed Tomographic Scans and the 396 Implied Cost Michael Blaivas and Matthew Lyon Prehospital Use of Analgesics at Home or En Route to the Hospital in Children With Extremity 400 Injuries Alex L. Rogovik and Ran D. Goldman A Risk Score to Predict Silent Myocardial Ischemia in Patients With Coronary Artery Disease 406 Under Aspirin Therapy Presenting With Upper Gastrointestinal Hemorrhage Chien-Chih Chen, Chee-Fah Chong, Cheng-Deng Kuo, and Tzong-Luen Wang A Clinical Score Predicting the Need for Hospitalization in Scorpion Envenomation Semir Nouira, 414 Riadh Boukef, Noureddine Nciri, Habib Haguiga, Souheil Elatrous, Lamia Besbes, Mondher Letaief, and Fekri Abroug External Cardiac Defibrillation During Wet-Surface Cooling in Pigs Alexandra Schratter, 420 Wolfgang Weihs, Michael Holzer, Andreas Janata, Wilhelm Behringer, Udo M. Losert, William J. Ohley, Robert B. Schock, and Fritz Sterz BRIEF REPORTS

Clinical Measures Associated With FEV1 in Persons With Asthma Requiring Hospital Admission 425 Donald H. Arnold, Tebeb Gebretsadik, Patricia A. Minton, Stanley Higgins, and Tina V. Hartert Splenic Artery Aneurysms Encountered in the ED: 10 Years’ Experience Chu-Feng Liu, 430 Chia-Te Kung, Ber-Ming Liu, Shu-Hang Ng, Chung-Cheng Huang, and Sheung-Fat Ko Utility of Impedance Cardiography for Dyspneic Patients in the ED Hsiang-Yun Lo, 437 Shu-Chen Liao, Chip-Jin Ng, Jen-Tse Kuan, Jih-Chang Chen, and Te-Fa Chiu Outcomes After Environmental Hyperthermia Frank LoVecchio, Anthony F. Pizon, 442 Christopher Berrett, and Adam Balls Effectiveness of Nonnarcotic Protocol for the Treatment of Acute Exacerbations of Chronic 445 Nonmalignant Pain James E. Svenson and Thomas D. Meyer CLINICAL NOTES Sixty-Four–Slice Multidetector Computed Tomography: The Future of ED Cardiac Care 450 Alexander T. Limkakeng, Ethan Halpern, and Kevin M. Takakuwa

(Continued on next page) Contents continued DIAGNOSTICS Broad Complex Atrial Fibrillation Huck Chin Chew and Swee Han Lim 459 The VIDAS D-dimer Test for Venous Thromboembolism: A Prospective Surveillance 464 Study Shows Maintenance of Sensitivity and Specificity When Used in Normal Clinical Practice David Mountain, Ian Jacobs, and Andrew Haig THERAPEUTICS Ultrasound-Guided Supraclavicular Block for the Treatment of Upper Extremity Fractures, 472 Dislocations, and Abscesses in the ED Michael B. Stone, Daniel D. Price, and Ralph Wang CONTROVERSIES The Use of Long-Term Controller Medications in Asthmatic Patients Being Discharged 476 From the ED—Why the Controversy? A. Dosanjh EDITORIAL Therapeutic Hypothermia and the Need for Defibrillation: Wet or Dry? Joseph Varon 479 CORRESPONDENCE Obtundation in a Toddler: Naloxone is Fundamental Kevin C. Osterhoudt and Diane P. Calello 481 Infantile Case of Seizure Induced by Intoxication After Accidental Consumption of Eperisone 481 Hydrochloride, an Antispastic Agent Katsutoshi Tanno, Eichi Narimatsu, Yoshihiro Takeyama, and Yasufumi Asai Hydronephrosis During Pregnancy Todd McArthur, Chad S. Crystal, and Michael A. Miller 482 Recurrent Myelin Basic Protein Elevation in Cerebrospinal Fluid as a Predictive Marker of 483 Delayed Encephalopathy After Carbon Monoxide Poisoning Yoshito Kamijo, Toshimitsu Ide, and Kazui Soma Reply to Multiple Dose Activated Charcoal in Carbamazepine Poisoning Nozha Brahmi 485 Reply to Proper Observation of Patient-Related Factors is an Important Determinant in the 485 Utility of the D-Dimer Test for Exclusion of Venous Thromboembolism in the ED Patrick Ray Emergency Short-Stay Unit as an Effective Alternative to in-Hospital Admission for Acute Chronic 485 Obstructive Pulmonary Disease Exacerbation Albert Salazar, Antoni Juan, Ricard Ballbe, and Xavier Corbella Differences Between Various Glomerular Filtration Rate Calculation Methods in Predicting 487 Patients at Risk for Contrast-Induced Nephropathy Cenker Eken and I˙sa Kilicaslan

AMERICAN JOURNAL OF EMERGENCY MEDICINE ELECTRONIC-EXTRA PAGES (Full text articles available online at http://journals.elsevierhealth.com/periodicals/yajem)

CASE REPORTS Traumatic Inferior Gluteal Artery Pseudoaneurysm: Case Report and Review of Literature 488 Ani Aydin, Christopher C. Lee, Eric Schultz, and Jeremy Ackerman Acute Emphysematous Cholecystitis With Initial Normal Radiological Evaluation: A Fatal 488 Diagnostic Pitfall in the ED Vei-Ken Seow, Chiu-Mei Lin, Tzong-Luen Wang, Chee-Fah Chong, and I-Yin Lin The American Journal of Emergency Medicine

THE AMERICAN JOURNAL OF EMERGENCY MEDICINE (ISSN any means now or hereafter known, electronic or mechanical, 0735-6757) is published nine times a year by Elsevier Inc., 360 including photocopy, recording, or any information storage and Park Avenue South, New York, NY 10010-1710. Months of retrieval system, without permission in writing from the Pub- issue are: January, February, March, May, June, July, September, lisher. Printed in the United States of America. October and November. Business and Editorial Offices: PERMISSIONS: Permissions may be sought directly from 1600 John F. Kennedy Blvd., Ste. 1800, Philadelphia, PA Elsevier’s Rights Department in Philadelphia, PA, USA: phone 19103-2899. Customer Service Office: 6277 Sea Harbor 215-239-3600, e-mail [email protected]. Requests Drive, Orlando, FL 32887-4800. Periodicals postage paid at may also be completed on-line via the Elsevier homepage (www. New York, NY and additional mailing offices. elsevier.com/locate/permissions). POSTMASTER: Send address changes to THE AMERICAN JOUR- REPRINTS: For 100 or more copies of an article in this NAL OF EMERGENCY MEDICINE, Elsevier Customer Service, 6277 Sea publication, please contact the Commercial Reprints Depart- Harbor Drive, Orlando, FL 32887-4800. ment, Elsevier Science Inc., 360 Park Avenue South, New Manuscripts, correspondence, and editorial material should York, New York 10010-1710. Tel. (212) 633-3813 Fax: be sent to The Editor, THE AMERICAN JOURNAL OF EMERGENCY (212) 633-3820 e-mail: [email protected] MEDICINE, P.O. Box 1494, West Bethesda, MD 20827-1494. ADVERTISING: Display Advertising and related corres- For enquiries relating to the submission of articles (in- pondence should be addressed to Inez Herrero, 360 Park Avenue, cluding electronic submission where available) please visit South, New York, NY 10010. Tel: 212-633-3122 / Fax: 212-633- Elsevier’s Author Gateway at http://authors.elsevier.com. The 3820; E-mail: [email protected]. Classified & Recruitment Author Gateway also provides the facility to track accepted Advertising and related correspondence should be addressed to articles and set up e-mail alerts to inform you of when Simone Imbert, 360 Park Avenue, South, New York, NY 10010. Tel: an article’s status has changed, as well as detailed artwork 212-462-1908 / Fax: 212-633-3820; E-mail: [email protected] guidelines, copyright information, frequently asked questions The ideas and opinions expressed in THE AMERICAN JOURNAL OF and more. EMERGENCY MEDICINE do not necessarily reflect those of the Editor Contact details for questions arising after acceptance of an or the Publisher. Publication of an advertisement or other product article, especially those relating to proofs, are provided after mention in THE AMERICAN JOURNAL OF EMERGENCY MEDICINE should registration of an article for publication. not be construed as an endorsement of the product or the manufac- YEARLY SUBSCRIPTION RATES: United States and pos- turer’s claims. Readers are encouraged to contact the manufacturer sessions: individual, $262.00; institution, $384.00; student and resi- with any questions about the features or limitations of the products dent, $123.00. All other countries: individual, $378.00; institution, mentioned. The Publisher does not assume any responsibility for $501.00; student and resident, $189.00. For all areas outside the any injury and/or damage to persons or property arising out of or United States and possessions, there is no additional charge for related to any use of the material contained in this periodical. The surface delivery. To receive student/ resident rate, orders must reader is advised to check the appropriate medical literature and the be accompanied by name of affiliated institution, date of term, product information currently provided by the manufacturer of and the signature of program/residency coordinator on institution each drug to be administered to verify the dosage, the method and letterhead. Orders will be billed at individual rate until proof of duration of administration, or contraindications. It is the responsi- status is received. bility of the treating physician or other health care professional, Prices are subject to change without notice. Current prices relying on independent experience and knowledge of the patient, to are in effect for back volumes and back issues. Single issues, determine drug dosages and the best treatment for the patient. both current and back, exist in limited quantities and are offered The appearance of the code at the bottom of the first page for sale subject to availability. Back issues sold in conjunction of an article in this journal indicates the copyright owner’s with a subscription are on a prorated basis. consent that copies of the article may be made for personal or Checks should be made payable to Elsevier and sent to internal use, or for the personal or internal use of specific THE AMERICAN JOURNAL OF EMERGENCY MEDICINE, W.B. Saun- clients, for those registered with the Copyright Clearance Cen- ders, Periodicals Department, PO Box 628239, Orlando, FL ter, Inc (222 Rosewood Drive, Danvers, MA 01923; (978) 32862-8239. 750-8400; www.copyright.com). This consent is given on Correspondence regarding subscriptions or change of address the condition that the copier pay the stated per-copy fee for should be directed to THE AMERICAN JOURNAL OF EMERGENCY that article through the Copyright Clearance Center, Inc for MEDICINE, Elsevier, Periodicals Department, 6277 Sea Harbor Dr, copying beyond that permitted by Sections 107 or 108 of the Orlando, FL 32887-4800. Telephone number, (800) 654-2452; out- US Copyright Law. This consent does not extend to other side the United States and Canada, (407) 345-4000. Changes of kinds of copying, such as copying for general distribution, for address should be sent preferably 60 days before the new address advertising or promotional purposes, for creating new collec- becomes effective. Missing issues will be replaced free of charge tive works, or for resale. Absence of the code indicates that the if the Publisher is notified at the above address within 2 months of material may not be processed through the Copyright publication of the issue for US and Canadian subscribers and within Clearance Center, Inc. 4 months for subscribers from all other countries. THE AMERICAN JOURNAL OF EMERGENCY MEDICINE is indexed © 2007 Elsevier Inc. All rights reserved. No part of this and abstracted in Index Medicus, EMBASE/Excerpta Medica, publication may be reproduced or transmitted in any form or by 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, Richard M. Nowak, MD, MBA, Henry Ford Hospital, Charlottesville Detroit Neal E. Flomenbaum, MD, Cornell Medical Center, New York Jonathan Olshaker, MD, Boston University, Boston Glenn C. Hamilton, MD, Wright State University, Joseph P. Ornato, MD, Medical College of Virginia, Dayton Richmond Gabor D. Kelen, MD, Johns Hopkins University, Baltimore Norman A. Paradis, MD, University of Colorado, Denver Toby L. Litovitz, MD, Georgetown University, Howard A. Werman, MD, Ohio State University, Columbus Washington, DC Charles J. McCabe, MD, Massachusetts General Loren Yamamoto, MD, MPH, MBA, University of Hawaii, Hospital, Boston Honolulu

EDITORIAL CONSULTANTS 2006 Neal Abarbanell, MD Michael Heller, MD Patrick Ray, MD David Amponsah, MD Johan Herlitz, MD, PhD Philip Rice, MD Joel Bartfield, MD C. James Holliman, MD Alfred Sacchetti, MD Steven L. Bernstein, MD James F. Holmes, MD, MPH Philip Salen, MD Paul Biddinger, MD Stephen Huff, MD David M. Schreck, MD, MS Polly Bijur, MD Fredric Husty, MD William Scruggs, MD Michael Blaivas, MD Ken Iserson, MD Donna Seeger, MD Judith Brillman, MD Jeanne Jacoby, MD Philip Shayne, MD Sean Bush, MD Gary Josephsen, MD Ronald Sing, MD Christopher R. Carpenter, MD Marshall Kapp, MD Adam Singer, MD Jeffrey Caterino, MD Lawrence Edward Kass, MD David E. Slattery, MD William Chiang, MD Ijaz Khan, MD Corey M. Slovis, MD Richard Christensen, MD, MA Bruce Klein, MD Richard Sobel, MD William Cordell, MD Wendy Klein-Schwartz, PharmD Matthew Spencer, MD Frank Counselman, MD Michael Kontos, MD Tom Stair, MD Cameron Crandall, MD John G. Laffey, MD LG Stead, MD Sandra J. Cunningham, MD Jerroid Leikin, MD Milton Tenenbein, MD Daniel Davis, MD Philip D. Levy, MD Kevin M. Terrell, DO, MS Robert Derlet, MD Siu Fai Li, MD Stephen Thomas, MD Deborah Dierecks, MD Joseph Losek, MD Joseph Varon, MD Charles Emerman, MD Frank Lovecchio, MD Arvind Venkat, MD Amy Ernst, MD Michael Lyons, MD Gary Michael Vilke, MD Lorrie Garces, MD Scott Melanson, MD Rade Vukmir, MD, JD Leslie A. Geddes, ME, PhD James R. Miner, MD Terry Walman, MD Nina Gentile, MD Antonio Muniz, MD Daniel Walsh, MD Louis Graff, MD Kristen E. Nordenholz, MD Richard Wersman, MD Colin Graham, MD David Overton, MD Howard Werman, MD Steven Green, MD Manish Patel, MD Michael D. Witting, MD Eric A. Gross, MD W. Frank Peacock, MD Tim Wolfe, MD Blaine Hannafin, MD Andrew Perron, MD Allan B. Wolfson, MD Raymond G. Hart, MD, MPH Michael Phelan, MD Keith Wrenn, MD Mark Hauswald, MD Jesse M. Pines, MD, MBA Huiyun Xiang, MD Kennon Heard, MD Rumen D. Powers, MD Leslie Zun, MD Manuscript Submission and Editorial Review Policy

The scope of The American Journal of Emergency Medicine is as Cover Letter broad as the definition of emergency medicine itself, encompassing all The cover letter accompanying all submitted manuscripts must (1) activities concerned with acute medical care. AJEM invites the submission be signed by all authors, and (2) contain the following language: ‘‘The of original research, reports, correspondence, and opinion relating to manuscript, as submitted or its essence in another version, is not under acute adult and pediatric medicine and surgery and the related fields consideration for publication elsewhere, and will not be published of trauma, toxicology, critical care, resuscitation, emergency medical elsewhere while under consideration by AJEM. The authors have no services, behavioral emergencies, and environmental medicine. commercial associations or sources of support that might pose a Original contributions will be accepted on the basis of significance, conflict of interest. All authors have made substantive contributions to validity, and clarity. Authors will be expected to justify conclusions by the study, and all authors endorse the data and conclusions.’’ Authors the data presented, maintain a lucid prose style, and describe method- with a potential conflict of interest should cite it in the cover letter. ology in sufficient detail for readers to evaluate results accurately. AJEM, in turn, is committed to a confidential, expeditious, and professional Author Responsibility editorial process. Reviews will be objective, rigorous, and responsible. Submissions are reviewed for possible publication with the under- Articles published in AJEM are indexed and abstracted in Index standing that they are original and not simultaneously under consid- Medicus, Excerpta Medica, Current Contents/Clinical Medicine, eration by another journal. Accepted manuscripts become the property ISI/BIOMED, and BIOSIS. of AJEM and may not be published elsewhere without the written per- For the convenience of prospective authors, AJEM is a participating mission of AJEM. Any material previously published elsewhere must journal in the International Committee of Medical Journal Editors’ be accompanied by written consent of its author and publisher when ‘‘Uniform Requirements for Manuscripts Submitted to Biomedical submitted to AJEM. Photographs of an identifiable subject should be Journals’’ (N Engl J Med 1997;336:309-315). This agreement provides accompanied by a release signed by the subject or responsible party for a standardized manuscript format, allowing authors to submit authorizing publication. If required, institutional clearance to publish articles to any one of over 500 scholarly medical publications without should be submitted with the manuscript. AJEM is not responsible for revision simply to accommodate the vagaries of any individual jour- statements made by any contributor. Authors should keep copies of nal’s technical and stylistic requirements. all submitted materials. Photographs and figures are not returned, REVIEW POLICY even if a manuscript is not accepted. All original contributions, investigations, and reports will be subjected Repetitive Publication to multiple-peer review. To protect the integrity and anonymity of the review process, all reviews will be conducted in double-blinded fashion. Authors submitting papers to this journal must confirm that the paper, To promote quality composition and investigation, legible comments as submitted or its essence in another version, has not been published from all referees will accompany returned manuscripts. To encourage elsewhere, is not under consideration for publication elsewhere, and will criticism, correspondence, and open discussion of controversial issues, not be published elsewhere while under consideration by AJEM. Prior letters to the editor will be printed promptly. As a courtesy to contributors publication of some content of the paper may not preclude the paper’s and to ensure the timeliness of AJEM’s content, authors will routinely publication in AJEM. Authors must provide full information in the be notified of the action taken upon their manuscripts within 10 weeks cover letter sent with the submitted manuscript on any possibly of submission. Case reports will receive expedited in-house review and repetitive publication of content, including: (1) reworked data already will be accepted or rejected without specific comments. reported; (2) cases or subjects in a study cohort already described in a published report; (3) previously reported single or multiple cases; (4) GUIDE FOR AUTHORS content already published or to be published in another format such as the We invite submissions on clinical and laboratory research and topics proceedings of a meeting or symposium, a chapter in a book, or a letter to pertinent to adult and pediatric emergency medicine including emergency the editor; (5) content published in a language other than English. medical and health services, trauma, toxicology, resuscitation, behavioral emergencies, critical care, and environmental medicine. In general Conflict of Interest AJEM does not publish surveys, papers that focus on patient satis- Authors are expected to disclose any commercial associations or faction, quality assurance, or didactics. The following are journal sources of support that might pose a conflict of interest in connection with features for which we invite submissions: the submitted article. All funding sources supporting the work must be ORIGINAL CONTRIBUTIONS: Reports of new clinical and laboratory acknowledged in a footnote on the title page. All affiliations with or investigations and research. financial involvement in any organization on entity with a direct financial BRIEF REPORTS: Short papers, series of cases, and preliminary reports interest in the subject matter or materials of the research discussed (eg, of work in progress; studies with small numbers pointing to the need for employment, consultancies, stock ownership or other equity interest, further investigation. Brief reports should be limited to 2,000 words of text patent-licensing arrangements) should be cited in the cover letter. (exclusive of tables, references, and figure legends). Human Research and Informed Consent RESEARCH SEMINARS: Discussions of the history, methodology, and future of a particular area or subject in emergency medicine research. When appropriate, manuscripts reporting the results of experimental REVIEWS: Definitive, in-depth, state-of-the-art reviews of clinical and investigations on human subjects should include a statement indicating research subjects. Unsolicited reviews are not generally published in approval by the institution’s Human Research Committee. AJEM. Before submitting any unsolicited reviews, please forward an Author Approval outline to the Editor for consideration. All accepted manuscripts are subject to copyediting. Authors will THERAPEUTICS: Detailed reviews of important devices and drugs receive page proof of their article before publication. used in the practice of emergency medicine. DIAGNOSTICS: Concise articles guiding clinical practice, with refer- Manuscript Submission ences to additional, authoritative sources. All manuscripts (including figures) must be submitted to AJEM CONTROVERSIES: Editorial viewpoints on current controversies. through our Web site (http://ees.elsevier.com/ajem/). Submission CLINICAL NOTES: Descriptions of new techniques and procedures in items should include separate files for a cover letter, title page, emergency medicine practice and investigation. abstract, manuscript text, references, legends for table/figure, tables, CORRESPONDENCE: Letters to the editor are limited to 800 words and figures. Revised manuscripts should also be accompanied by a of text (exclusive of references, tables, and figure legends). These unique file (separate from the covering letter) with responses to submissions should not contain an abstract. reviewers’ comments. CASE REPORTS: Case reports should describe a case unique to the The preferred order of files for electronics submission is as follows: emergency medicine literature, and are limited to 800 words of text cover letter, response to reviews (revised manuscripts only), title page, and an abstract < 250 words. Accepted case reports in their entirety manuscript file(s), table(s), figure(s). Files should be labelled with will be published digitally at our web site (www.ajemjournal.com), appropriate and descriptive file names (e.g., SmithText.doc, Fig1.eps, while the abstract will be published in each printed issue. Table3.doc). Upload text, tables and graphics as separate files. Do not import figures or tables into the text document; submit them as separate Acknowledgments files. Complete instructions for electronic artwork submission can be Acknowledge only people who have made substantive contributions found on the journal home page. to the study, and specify the contributions. Authors are responsible for All manuscripts must be submitted double-spaced in English. Please visit obtaining written permission from everyone acknowledged by name http://ees.elsevier.com/ajem to submit your manuscript electronically. because readers may infer their endorsement of the data and conclusions. The website guides authors stepwise through the creation and upload- ing of the various files. Note that original source files, not PDF files, are Abbreviations, Symbols, and Nomenclature required. Once the submission files are uploaded the system automati- Usage should conform to that recommended in Council of Biology cally generates electronic (PDF) proof, which is then used for review- Editors Style Manual (5th ed., 1983) available from the American ing. All correspondence, including the Editor’s decision and request for Institute of Biological Sciences, 1950 Rockville Pike, Bethesda, MD revisions, will be by e-mail. 20814. Avoid abbreviations. Do not abbreviate names of organizations, institutions, symptoms, diseases, or anatomic characteristics. A list of Copyright acceptable abbreviations is included in ‘‘Uniform Requirements for A copyright transfer agreement will be sent to corresponding authors Manuscripts Submitted to Biomedical Journals’’ (see below). Generic of each manuscript accepted for publication. Authors are responsible names of drugs are preferred; a brand name may be given for applying for permission for both print and electronic rights for all only with the first use of generic name. When the brand name of a borrowed materials and are responsible for paying any fees related to product or pharmaceutical is used, supply the manufacturer’s the application of these provisions. name and location (city and state). Title Page Units of Measurement On the title page include (1) the title; (2) a short running head of Use SI units for linear dimensions, weight, clinical chemistry, and fewer than 50 characters/spaces placed at the foot of the page; (3) hematology. Use the Celsius scale for all temperatures. The use of author(s) names, highest degrees, department(s) and institution(s); (4) other SI units is encouraged. name and address of author to whom reprint requests should be sent; References (5) source(s) of support in the form of equipment, drugs, or grants Cite references consecutively in the text. Do not cite review (including grant numbers); (6) the name of organization and date of articles. Use the same number each time the reference appears in assembly if the article has been presented; and (7) ‘‘Key Words,’’ a list the text. At the conclusion of the article, list references in numerical of three to ten important words or phrases for indexing. Whenever pos- order, typed double spaced. Abbreviate journal titles according to sible, use terms from the medical subject heading of Index Medius. Index Medicus style. Please provide inclusive pagination Punctuation To ensure blinded, impartial review, do not indicate the authors of the is shown below. article on any other page. Journal articles: List all authors when three or fewer; when four or Abstract more, list first three and add et al. On the second page include a structured abstract of fewer than Abraham E, Baraff LJ: Oral versus parenteral therapy of pye- 150 words stating the objective, methods, results, and conclusion. Be lonephritis. Curr Ther Res 1982;31:536-542 concise yet detailed. Books: Capitalize all important words in title. Ludwig S, Fleisher GR, Henretig FM, et al (eds): Pediatric Text Emergency Medicine. Baltimore, MD, Williams & Wilkins, When appropriate, divide the text into Introduction, Methods, Re- 1983, pp 203-209. sults, and Discussion. Chapter in a book: List editors of book. Introduction: Clearly state the purpose of the article, summarize the Eliastam M: Cardiac emergencies. In Eliastam M, Sternbach rationale for the study or observation, give only strictly pertinent GL, Bresler MJ (eds): Manual of Emergency Medicine. Chi- references, and do not review the subject extensively. cago, IL, Yearbook, 1983, pp 1-28 Methods: Identify the methods, apparatus, and procedures in sufficient detail to allow other workers to reproduce the results. Give references to References to unpublished information should be included parenthet- established methods, including statistical method; provide references and cally in the text. Do not cite review articles. brief descriptions of methods that have been published but may not be well known; describe new or substantially modified methods, giving Tables reasons for using them and evaluating their limitations. Type tables double-spaced on separate sheets with number and title. Results: Present your results in logical sequence in the text, tables, and Do not submit tables as photographs. Omit internal horizontal and illustrations. Do not repeat in the text all of the data in the tables and/or vertical rules. Cite each table in the text in consecutive order. illustrations; emphasize or summarize only important observations. Figures Discussion: Emphasize the new and important aspects of the study and Submit figures electronically as separate files. Complete instructions conclusions that follow from them. Do not repeat in detail data given in for electronic artwork submission can be found on the Author Gateway, the Results sections. Include in the Discussion the implications of the accessible through the journal home page, http://ees.elsevier.com/ajem/. findings and their limitations and relate the observations to other relevant studies. Link the conclusions with the goals of the study, but avoid Editorial Inquiries unqualified statements and conclusions not completely supported by your Authors are strongly encouraged to use the Journal gateway for all data. Avoid claiming priority and alluding to work that has not been inquiries regarding submitted manuscripts for the fastest response and most completed. State new hypotheses when warranted, but clearly label them current information/status. Other inquiries (i.e., unrelated to a submitted as such. Recommendations, when appropriate, may be included. manuscript) can be directed to [email protected]. American Journal of Emergency Medicine (2007) 25, 379–384

www.elsevier.com/locate/ajem

Original Contribution Expanding the use of topical anesthesia in wound management: sequential layered application of topical lidocaine with epinephrine

Slava V. Gaufberg MDa,*, Michael J. Walta MDb, Tom P. Workman MDa aDepartment of Emergency Medicine, The Cambridge Hospital, Cambridge Health Alliance, Division of Emergency Medicine, Harvard Medical School, Boston, MA 02139, USA bDepartment of Internal Medicine, Cambridge Health Alliance, Harvard Medical School, Boston, MA 02139, USA

Received 4 April 2006; revised 1 November 2006; accepted 2 November 2006

Abstract Topical anesthesia eliminates the need for injection of anesthetic. Most studies on the use of topical anesthesia were done on children, using 3 active ingredients (lidocaine, epinephrine, tetracaine, or tetracaine, adrenaline, cocaine) for relatively small wounds of the face and scalp. Objectives: To demonstrate that topical anesthesia is effective and safe in adults of all ages and for larger wounds, using a preparation with 2 active ingredients, topical lidocaine and epinephrine (TLE). Methods: One hundred patients were enrolled in a randomized controlled trial, with 50 in each group. The study group received TLE using a unique method of bsequential layered application.Q The control group received 2% lidocaine infiltration anesthesia. Patients rated the pain from the application of anesthesia and from suturing, using a 0 to 10 visual analog pain scale. Follow-up interviews were conducted to assess for complications and to rate patients’ wound repair experience. Results: Patients in the study group reported significantly less pain from TLE application, with 66% reporting no pain vs 0% reporting no pain from the infiltration in the control group ( P b .001). There was no difference in pain during wound repair between the 2 groups ( P ~ .59). On follow-up, 95% of patients contacted in the TLE group rated their experience in regard to pain as bexcellent,Q compared to 5% of patients in the control group ( P b .001). Conclusion: Topical lidocaine and epinephrine bsequential layered applicationQ is an effective, safe, and less painful method of anesthesia for a wide variety of lacerations. Patients recall their experience with this technique very favorably. D 2007 Elsevier Inc. All rights reserved.

1. Introduction primary goal of the emergency physician. There are a number of methods available for wound anesthesia, includ- More than 10 million lacerations are treated annually in ing topical anesthesia, infiltration of the wound, field or EDs in the United States. Pain relief during wound repair is a nerve block, hematoma block, and conscious sedation [1]. The main advantage of topical anesthesia is the * Corresponding author. Tel.: +1 617 665 1712. elimination of injection of anesthetic. It effectively elimi- E-mail address: [email protected] (S.V. Gaufberg). nates pain from injection, distortion of infiltrated wound

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2006.11.013 380 S.V. Gaufberg et al. tissues, and decreases the risk of needle-stick injury to the 2.3. Study protocol operator [2]. However, the use of topical anesthesia has been limited. Subjects in both groups were given anesthesia before Most previous studies have used preparations that contain 3 wound irrigation and cleaning. active ingredients, either lidocaine, epinephrine, tetracaine In the study group, all lacerations were anesthetized with (LET), or tetracaine, adrenaline, cocaine (TAC). LET has TLE solution, containing 5% lidocaine and 0.025% epi- b been studied extensively in children [1,3], but its use has nephrine. A novel technique of sequential layered appli- Q been largely confined to small wounds (b2 cm) on the face cation, developed by Slava V. Gaufberg, MD, was used for and scalp [2,4,5] or superficial wounds [6]. Anesthetic the application of TLE. A piece of cotton just large enough to failure has been reported to exceed 5% on extremities [6,7]. cover the laceration and 2 mm of surrounding skin was The use of TAC has declined because of concerns about the soaked with TLE. The soaked cotton was placed on the potential for severe toxicity from systemic absorption of wound for 10 to 15 minutes, then removed. A second, cocaine [1,8]. similarly sized piece of cotton soaked in TLE was packed We therefore undertook a study to investigate the use of deeper inside the wound for 10 to 15 minutes, then removed. topical anesthetics in adults for a large variety of wounds For deeper wounds, a third layer of cotton soaked with TLE and lacerations. Because we planned to use topical may be packed deeper into the wound in the same fashion anesthesia for large wounds that require a significant (see Fig. 1). The wound was presumed to be anesthetized and amount of anesthetic, we used a preparation with only ready for suturing when 3-mm-wide or greater area of 2 active ingredients, topical lidocaine and epinephrine vasoconstrictive pallor formed along all wound edges. The (TLE), to limit the potential for toxicity [9]. Because the wound was then probed with the teeth of surgical forceps to traditional method of applying topical anesthetic to the assure anesthesia along the entire length and depth of the wound surface does not provide sufficient anesthesia to wound. If an area of the wound was inadequately anesthe- deep wounds, we developed a novel method of bsequential tized, more TLE was applied to the wound until adequate layered applicationQ of TLE. anesthesia was achieved. We hypothesized that sequential layered application In the control group, infiltration anesthesia with buffered of TLE would allow us to achieve anesthesia equal to 2% lidocaine solution with 8.4% sodium bicarbonate that of infiltration with lidocaine, minus the pain caused solution 9:1 was performed using the standard technique. by infiltration. The wound was then probed with the teeth of surgical forceps in the same fashion as in the study group. If an area of the wound was inadequately anesthetized, more lidocaine 2. Methods 2.1. Study design We performed a prospective, randomized controlled trial of the effectiveness of TLE anesthesia using the sequential layered application technique. Patients presenting with lacerations to the ED were enrolled in either the study group (using topical anesthetic) or the control group (injection anesthetic). The primary outcome measured was pain reported by the patient during application of anesthetic and during wound repair using a 0-to-10 visual analog pain scale. Informed consent was obtained from all subjects. The study was approved by the institutional review board. 2.2. Study setting and population The setting of the study was a community teaching hospital ED with an annual patient volume of 30000 serving an urban population. The study population consisted of 100 adults greater than 18 years old, having lacerations in need of suturing. The following exclusion criteria were applied: allergy to Fig. 1 A small piece of TLE-soaked cotton is packed on the lidocaine or epinephrine, altered mental status, inability to wound surface. Vasoconstrictive pallor develops, indicating anes- verbalize pain according to the visual analog pain scale, thesia. Another piece of cotton is packed deeper and remains for 10 pregnancy, age younger than 18 years. to 15 minutes. This is repeated until the wound is fully anesthetized. Expanding the use of topical anesthesia in wound management 381

Table 1 Characteristics of control and study group Control Study P group, n group, n Total 50 50 Age (y) .75 18-30 20 19 31-50 21 22 51-70 6 8 N70 3 1 Laceration length (cm) .75 b110 1-2 5 3 2-3 14 16 3-5 19 22 N5119 Laceration depth (cm) .95 b0.5 13 14 0.5-2.0 31 29 Fig. 2 Percent reporting pain during application of anesthetic. 2.0-4.0 6 7 Laceration location .10 Scalp 7 8 Follow-up assessment 3 to 6 months later was made by Face 5 10 ED volunteers (premed students) who asked patients Lower extremity 9 4 whether they had encountered bany problem with wound Upper extremity 11 4 healing, including infection,Q and then asked patients to rate Hands 18 24 their overall wound repair experience as bexcellent,Qbgood,Q bsatisfactory,Q or bunsatisfactory.Q 2.5. Data analysis was infiltrated into the wound until adequate anesthesia was achieved. Data for all 100 patients were analyzed without any exclusion. Baseline characteristics and follow-up data 2.4. Measurements were compared using v2 tests. v2 values were converted to In each group, baseline characteristics including age of P values using standard conversion tables. Pain scores were the subject, length and depth of the wound, location of compared using Wilcoxon rank sum tests. The difference wound, and time to achieve anesthesia were recorded. between means was used to construct 95% confidence Information about the number of cotton layers and amount intervals (CIs) around the difference in the mean pain scores of TLE used in the study group was recorded. between the study and control groups. P values of less than The effectiveness of anesthesia was assessed by the .05 were considered significant. patient immediately after the procedure using a 0-to-10 visual analog pain scale administered by third-party (a 3. Results resident, nurse, or multilingual interpreter). The subject was instructed to assess the pain from the application of We enrolled 100 consecutive adult patients presenting for anesthesia (either from TLE application or lidocaine laceration repair. Fifty subjects were randomly enrolled in infiltration) and the pain from suturing the wound. These each group. pain assessments were recorded immediately. Baseline characteristics for the 50 patients in each group were similar (Table 1). In the control group, time necessary for adequate Table 2 Application of anesthesia infiltration anesthesia varied from 1 to 12 minutes, with a Control group Study group P mean time of 5 minutes. The amount of lidocaine needed to Time to achieve achieve adequate anesthesia varied between 10 and 340 mg, anesthesia (min) with a median amount of 124 mg. For the study group, the b Range 1-12 20-40 .001 time for the application of TLE necessary for adequate Mean 5 29 wound anesthesia varied from 20 to 40 minutes, with a Amount of lidocaine mean time of 29 minutes. The time taken for TLE to achieve used (mg) Range 10-340 25-400 .90 adequate anesthesia was significantly longer than for the b Mean 124 135 control group ( P .001). The number of application layers needed to achieve adequate anesthesia varied between 1 and 382 S.V. Gaufberg et al.

Table 3 Pain during application of anesthetic Table 4 Effectiveness of anesthesia during wound repair Control Study P Control group, n Study group, n P group, n group, n Visual analog .59 Visual analog b.001 pain scale pain scale 0424 0033 164 1416 222 261 300 390 400 4100 z500 590 Mean 0.20 0.16 650 Difference 0.04 720 between means 820 95% CI (range of (0.23 to À0.15) 910 difference) 10 2 0 Mean 4.34 0.36 Difference 3.98 pain from the application of TLE. The remaining 17 patients between means reported mild pain, with no patients reporting a score greater 95% CI (range of (4.60-3.36) difference) than 2 (Fig. 2). The mean pain score for the control group is 4.34 vs a pain score of 0.36 in the study group ( P b .001), with a difference of 3.98 in the mean pain scores between 4, with a mean of 2.7. Twenty-six patients (52%) required groups. We are 95% confident that the difference between 3 layers, 17 patients (34%) required 2 layers, and 5 patients pain scores is between 4.60 and 3.36 (Table 3). (10%) required 4 layers of TLE. The amount of TLE needed During wound repair, effectiveness of anesthesia was to achieve adequate anesthesia varied from 25 to 400 mg of similar in both groups. In the control group, 42 patients lidocaine, with a median amount of 135 mg (Table 2). The (84%) reported no pain from suturing, and no one reported a amount of lidocaine used in TLE application was compa- pain level above 2. In the study group, 44 patients (88%) rable to that (mean 124 mg) in the control group ( P ~ .90). reported no pain from suturing, and no patients reported pain scores greater than 2 (Fig. 3). In the control group, the 3.1. Pain during application of anesthetic and mean pain score was 0.20 vs a pain score of 0.16 in the wound repair study group ( P ~ .59), with a difference of 0.04 in the mean pain scores between groups. We are more than 95% During the application of anesthesia, patients in the confident that there is no difference in the pain scores control group reported significant pain from infiltration of between the control and study groups during wound repair lidocaine, with 21 patients (42%) reporting pain of 5 or (Table 4). greater. In the study group, 33 patients (66%) reported no 3.2. Follow-up interview We obtained follow-up interviews with 37 patients in the control group and 42 patients in the study group (Table 5). Patients in the TLE application group reported a better

Table 5 Follow-up interview after wound repair Control Study P group, n group, n Patients contacted 37 (74%) 42 (84%) (% of total) Recall of experience b.001 (% of patients contacted) Excellent 2 (5%) 40 (95%) Good 23 (62%) 2 (5%) Satisfactory 10 (27%) 0 Unsatisfactory 2 (5%) 0 Difficulty with wound 00 healing or infection Fig. 3 Percent reporting pain during wound repair. Expanding the use of topical anesthesia in wound management 383 overall experience than patients in the control group. A infiltrated wound tissue. It also shows the high rate of majority (95%) of patients contacted in the TLE application patient satisfaction. We feel this justifies the additional time group rated their experience as bexcellent,Q as compared to spent on applying topical anesthesia. only 5% of patients contacted in the control group. The difference in satisfaction between the 2 groups was 4.3. Potential for toxicity b significant ( P .001). No patients in either group reported The generally accepted limit for the use of injectable difficulty with wound healing or infection. lidocaine with epinephrine is 7 mg/kg [12,13]. We propose the same limit be used for TLE, although substantially less lidocaine is absorbed when used topically. 4. Discussion As such the limit for a 70-kg person would be: 70 Â 7= 490 mg, or 10 mL of the 5% solution that was used in Topical anesthesia for laceration repair has been prac- this study. Therefore, the limit for 5% lidocaine solution ticed for many years. Factors responsible for limiting its would be 1 mL for every 7 kg of body weight. This imposes use include concern for toxicity, lack of optimal technique a practical limit on the size of lacerations that can be for application in large wounds, and increased time to treated with this technique. The largest laceration treated achieve anesthesia. in this study was 12 cm long in a 79-kg patient who needed a total of 8 mL of TLE (400 mg of lidocaine) to achieve 4.1. Sequential layer application technique good anesthesia. In our study, adults of all ages with a wide variety of Two percent lidocaine solution in control group was lacerations on various areas of the body achieved effective chosen over 1% solution to decrease the volume of injected anesthesia using sequential layered application of TLE. solution and subsequent tissue distortion. Lidocaine without Application of TLE was significantly less painful than the epinephrine was used for injection anesthesia in the control infiltration technique used in the control group. Overall, group to avoid excluding patients with lacerations on patients who received TLE reported a significantly better acral areas. wound repair experience than patients receiving injectable lidocaine. No complications in wound healing or infection were reported during follow-up interviews in either group. 5. Limitations Most of the topical anesthesia studies showed poor success on trunk and extremity wounds in comparison with 5.1. Data collection and statistical analysis scalp and face wounds [9-11]. However, in previous studies, Our study may be limited by the way in which primary application of anesthetic was limited to the surface of the data were collected. Background information on age, wound [2-12]. We postulate that, as the circulation on the laceration depth, and laceration length was collected with trunk and extremities is relatively poor compared to the face respect to somewhat arbitrary ranges and placed into tables and scalp, local tissue absorption of single-layered topical for statistical analysis. The effect was to break otherwise anesthetic in these areas is limited. Using sequential layer continuous variable into discrete variables. t-Test analysis application of topical anesthetic may overcome relatively would have been the preferred statistical method had these poor absorption in these areas. data been recorded on a continuous scale. It is impossible to 4.2. Time to achieve anesthesia know whether this had more than a nominal effect on our results and analysis. The time taken to achieve adequate anesthesia for TLE is the major pitfall of this technique. Time ranged between 5.2. Follow-up interview 20 and 40 minutes, and 1 to 4 layers of TLE were required. When we obtained follow-up interview, we were unable Deeper lacerations requiring several layers are the most time- to contact 26% of the patients from control group vs 16% consuming. However, waiting 10 to 15 minutes for anes- from study group. This may weaken the follow-up thesia to take effect for each layer was arbitrarily used in this comparisons between the groups, although it is difficult to study. It is possible that this time could be reduced without know whether the subjects lost to follow-up were more compromising the effectiveness of anesthesia, although we prone to good or bad wound repair experiences. did not evaluate this in the present study. Training ancillary staff, including triage nurses, in the application of TLE 5.3. Use on acral areas might reduce the problems associated with longer application times [10,11]. Fully anesthetized patients would then be It is generally accepted that epinephrine should not be ready for wound repair by the ED physician. injected in acral areas such as the fingers, toes, nose, penis, bIs it worth the time?Q In our opinion, it is. The technique and ears. Such injection can theoretically cause vasospasm leads to the elimination of the following factors: painful and threaten the viability of the part. However, recent injection, risk of hollow-bore needle injury, and distortion of studies suggest low risk of ischemic vasospasm when local 384 S.V. Gaufberg et al. epinephrine is injected into acral areas [14,15]. Our study [6] Adler AJ, Dubinisky I, Eisen J. Does the use of topical lidocaine, does not provide any additional information regarding this epinephrine and tetracaine solution provide sufficient anesthesia for issue. We therefore cannot recommend that the technique laceration repair? Acad Emerg Med 1998;5(2):108-12. [7] Ernst AA, Marvez E, Nick TG, et al. Lidocaine adrenaline tetracaine described here be used on acral areas, although topical gel versus tetracaine adrenaline cocaine gel for topical anesthesia in lidocaine-adrenaline-tetracaine preparations have been used linear scalp and facial lacerations in children aged 5 to 17 years. safely on children with finger lacerations [16]. Pediatrics 1995;95(2):255-8. [8] Grant SA, Hoffman RS. Use of tetracaine, epinephrine, and cocaine as a topical anesthetic in the emergency department. Ann Emerg Med 6. Conclusions 1992;21(8):987-97. [9] Blackburn PA, Butler KH, Hughes MJ, et al. Comparison of Topical anesthesia with only 2 active ingredients tetracaine-adrenaline-cocaine (TAC) with topical lidocaine-epinephrine (TLE): efficacy and cost. Am J Emerg Med 1995;13(3):315-7. (lidocaine and epinephrine) using our sequential layered [10] Priestly S, Kelly AM, Chow L, et al. Application of topical local application technique is a highly effective and safe method anesthetic at triage reduces treatment time for children with of pain control for suturing a wide variety of lacerations. lacerations: a randomized controlled trial. Ann Emerg Med 2003; Patients graded their experience with this technique very 42(1):34-40. favorably. [11] Priestly S, Kelly AM, Chow L, et al. Topical local anesthetic application to wounds in children: does application at the time of triage decrease the overall treatment time in the emergency department? Acad Emerg Med 2002;9:449-50. References [12] Carpenter RL, Mackey DC. Local anesthetics. In: Barash PG, Cullen BF, Stoelting RK, editors. Clinical anesthesia. Philadelphia (Pa)7 J.B. [1] Singer AJ, Hollander JE, Quinn JV. Evaluation and management of Lippencott, 1992, 1992. p. 509-41. traumatic lacerations. N Engl J Med 1997;337(16):1142-8. [13] Strichartz GR, Berde CB. Local anesthetics. In: Miller RD, editor. [2] Hollander JE, Singer AJ. Laceration management. Ann Emerg Med Anesthesia. Philadelphia (PA)7 Churchill Livingstone; 1994. 1999;34(3):356-67. p. 489-521. [3] Anderson AB, Collechi C, Baronski R. Local anesthesia in pediatric [14] Krunic A, Wang L, Soltani K, et al. Digital anesthesia with patients: topical TAC versus lidocaine. Ann Emerg Med 1990; epinephrine: an old myth revisited. J Am Acad Dermatol 2004; 19(5):519-22. 51(5):755-9. [4] Pryor GJ, Kilpatrick WR, Opp DR. Local anesthesia in minor [15] Radovic P, Smith RG, Shumway D. Revisiting epinephrine in foot lacerations: topical TAC vs lidocaine infiltration. Ann Emerg Med surgery. J Am Podiatr Med Assoc 2003;93(2):157-60. 1980;9(11):568-71. [16] White NJ, Kim MK, Brousseau DC, et al. The anesthetic effectiveness [5] Bonadio WA, Wagner V. Efficacy of TAC topical anesthetic for repair of lidocaine-adrenaline-tetracaine gel on finger lacerations. Pediatr of pediatric lacerations. Am J Dis Child 1988;142(2):203-5. Emerg Care 2004;20(12):812-5. American Journal of Emergency Medicine (2007) 25, 385–390

www.elsevier.com/locate/ajem

Original Contribution Management of severe acute pain in emergency settings: ketamine reduces morphine consumptionB

Michel Galinski MDa,*, Franc¸ois Dolveck MDb, Xavier Combes MDe, Ve´ronique Limoges MDc, Nadia Smail!! MD, PhDd,Ve´ronique Pommier MDc, Franc¸ois Templier MDb, Jean Catineau MDa, Fre´de´ric Lapostolle MDa, Fre´de´ric Adnet MD, PhDa aSamu 93-EA 3409 Avicenne Hospital, 93009 Bobigny Cedex, France bSamu 92-Raymond Poincare´ Hospital, 92100 Garches, France cEmergency Department-Smur 77, Meaux’s Hospital, 77100 Meaux, France dSamu 31-Purpan Hospital, 31000 Toulouses, France eSamu 94-Henri Mondor Hospital, 94000 Cre´teil, France

Received 1 July 2006; revised 1 November 2006; accepted 2 November 2006

Abstract Objective: The aim of the study was to compare in emergency settings 2 analgesic regimens, morphine with ketamine (K group) or morphine with placebo (P group), for severe acute pain in trauma patients. Methods: This was a prospective, multicenter, randomized, double-blind, clinical trial. Seventy-three trauma patients with a severe acute pain defined as a visual analog scale (VAS) score of at least 60/100 were enrolled. Patients in the K group received 0.2 mg d kgÀ1 of intravenous ketamine over 10 minutes, and patients in the P group received isotonic sodium chloride solution. In both groups, patients were given an initial intravenous morphine injection of 0.1 mg d kgÀ1, followed by 3 mg every 5 minutes. Efficient analgesia was defined as a VAS score not exceeding 30/100. The primary end points were morphine consumption and VAS at 30 minutes (T30). Results: At T30, morphine consumption was significantly lower in the K group vs the P group, with 0.149 mg d kgÀ1 (0.132-0.165) and 0.202 mg d kgÀ1 (0.181-0.223), respectively ( P b .001). The VAS score at T30 did not differ significantly between the 2 groups, with 34.1 (25.6-42.6) in the K group and 39.5 (32.4-46.6) in the P group ( P = not significant). Conclusion: Ketamine was able to provide a morphine-sparing effect. D 2007 Elsevier Inc. All rights reserved.

1. Introduction Presented at the European Society of Anaesthesiology Congress in Madrid, Spain, June 3-6, 2006. Morphine titration infusion usually provides rapid and B Support was provided solely by institutional and/or departmental sources. effective analgesia in severe acute pain [1]. However, * Corresponding author. Tel.: +33 1 48 96 44 55; fax: +33 1 48 96 44 45. adverse effects sometimes occur and may require discon- E-mail address: [email protected] (M. Galinski). tinuation of morphine titration before sufficient pain relief is

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2006.11.016 386 M. Galinski et al. obtained [2]. The combination of nonopioid analgesics with randomly allocated to receive either morphine and ket- morphine provides a morphine-sparing effect and should amine (K group) or morphine and placebo (P group). decrease toxicity. This concept is the basis of multimodal Patients were eligible for inclusion if they presented a analgesia [3]. trauma with a severe acute pain defined as a visual analog Small doses of ketamine possess N-methyl-d-aspartate scale (VAS) score of at least 60/100; were aged between (NMDA) receptor noncompetitive antagonist properties 18 and 70 years; and were without acute respiratory, through a magnesium-dependent channel blockade [4]. hemodynamic, or neurologic compromise (respiratory Several studies demonstrated that it improves opioid distress signs, systolic blood pressure V90 mm Hg, analgesia in postoperative settings [5,6]. Glasgow Coma Score b15). Exclusion criteria included Potentiation between opioids and ketamine was demon- the presence of a psychiatric history; chronic respiratory, strated in animal studies [7] and was suggested in a renal, or hepatic failure; known ketamine sensitivity; volunteer study [8], whereas another report favored only known opioid allergies; treatment of chronic pain or an additive association [9]. Furthermore, ketamine attenu- treatment with opioids; incapacity to understand the VAS; ates the development of acute analgesic tolerance to opioids pregnancy; or indication for local or regional analgesia. in rats [10] and suppresses the rebound hyperalgesia Patients who had already received an opioid analgesic observed after opioid exposure in volunteers [11]. A recent (either by self-administration or by another attending study demonstrated that the combined administration of physician) were also excluded. small-dose ketamine and morphine promptly and satis- Intravenous morphine was given and titrated according to factorily resolved pain that was unresponsive to intra- the VAS pain score. Ketamine and placebo were adminis- venous (IV) morphine alone [12]. In emergency settings, a tered from syringes of similar appearances prepared by a 0.1-mg d kgÀ1 dose of IV morphine was not effective for nurse anesthetist who was otherwise not involved in the controlling severe acute pain in most patients [13]. There is study. The dilution of ketamine was 1 mg d mLÀ1.An evidence to suggest that the lack of effectiveness of independent physician-observer blinded to the analgesic morphine is due to the activation of NMDA receptors; and treatment group did all assessments of patients. The first if these are not effectively inhibited in time, the process may volume administered was 0.2 mL d kgÀ1 (ie, 0.2 mg d kgÀ1) evolve into a complex change in neural plasticity, resulting of ketamine (Ketamine; Panpharma, France) in the K group in central sensitization [14]. To our knowledge, there is no or 0.2 mL d kgÀ1 of placebo in the P group given over study assessing the interest of the morphine and ketamine 10 minutes with 0.1 mg d kgÀ1 of morphine, followed by association for trauma patients with severe acute pain in additional doses of 3 mg every 5 minutes until pain relief emergency settings. was obtained as defined by a VAS score not exceeding We tested the hypothesis that the combination of small- 30/100. The study protocol is shown in Fig. 1. dose ketamine and morphine would promptly reduce pain The end points of the study were the VAS and the perception and morphine consumption compared with morphine consumption at 30 minutes (T30). morphine alone in trauma patients with severe acute pain The VAS used was a 100-mm ruler and a marker that in emergency settings. the patient moves to the point indicating his or her pain intensity. The VAS was presented as a horizontal line on which the patient’s pain intensity was represented by a 2. Materials and methods point between the extremes of bno pain at allQ and bworst pain imaginable.Q We performed a prospective, multicenter, randomized, Patients were asked to assess the intensity of their pain double-blind, controlled study. The trial was coordinated by by providing a VAS score on inclusion (VAS [T0]) and then the Avicenne University Hospital (Bobigny, France). Five every 5 minutes until arrival at hospital. emergency departments using mobile intensive care units All VAS scores were recorded at T0, T15, and T30. previously described [15] were involved in this study. Thirty minutes after the first injection, overall patients’ and Mobile intensive care units are staffed by an attending investigators’ satisfaction regarding analgesia (pain relief emergency physician, a nurse anesthetist, and an emergency classified as excellent, good, mild, or weak) was recorded. medical technician. The safety evaluation included monitoring of the blood The Human Subjects Committee of the Robert Ballanger pressure, heart rate, respiratory rate, and oxygen saturation Hospital (Aulnay, France) approved this study, and all by pulse oximetry, as well as sedation level using the patients provided written informed consent. Patients were Ramsay score ranging from 1 to 6 [16] (1, anxious and recruited between January 01, 2003, and January 31, 2005. agitated or restless; 2, cooperative, oriented, and tranquil; A table of random numbers determined the randomiza- 3, responds to command only; 4, brisk response to light tion sequence, using a restricted randomization scheme to glabellar tap; 5, sluggish response to light glabellar tap; 6, no ensure roughly equal numbers in each group. Group response to glabellar tap). Sedation was defined as a Ramsay assignments were sealed in opaque envelopes and opened score of greater than 2. The presence of hallucination, sequentially by the investigators. Eligible patients were dysphoria, weakness sensation, diplopia, nausea, vomiting, Management of severe acute pain in emergency settings 387

Fig. 1 The study protocol involving administration of ketamine vs placebo. dizziness, itching, and bradypnea was likewise recorded. score not exceeding 30/100, and standard deviation was These data were recorded at T0 and T30. estimated to be approximately 15/100 [17]. To reach a VAS difference of more than 13/100 [18] in favor of the 2.1. Statistical analysis ketamine group, the appropriate sample size using an a error of .05 and a b error of .10 was calculated with The Wilcoxon test was used for quantitative nonpara- the formula of Casagrande et al [19]. A minimum of metric variables such as VAS, the Student t test for 29 patients for each group should be included to see a parametric quantitative variables, and the v2 test for difference of 13 mm between groups. We chose to include qualitative variables. The aim of pain relief was a VAS 30 patients in each group to increase the power of this 388 M. Galinski et al.

Table 1 Baseline characteristics of patients from groups K and P Characteristics K group P group P (n = 33) (n = 32) Age (y, mean F SD) 35 F 13 40 F 14 .3 Sex ratio (male-female) 25:8 23:9 .6 BMI (kg d mÀ2, mean F SD) 25 F 425F 4.8 Etiology of trauma (n [%]) .14 Suspicion of bone fracture 19 (58) 24 (75) Burns 2 (6) 2 (6) Others 12 (36) 6 (19) BMI indicates body mass index. Fig. 2 The effect of morphine with and without ketamine on the variation of the VAS score at T0, T15, and T30: DVAS (VAS [T0] À VAS [Tx]) (mean F 95% confidence interval). There was no study. Statistical analysis was performed using Statview difference between the 2 groups. Software (StatView version 5; Abacus Concepts, SAS Institute, Berkeley, Calif). A P value of less than .05 was considered statistically significant. significantly fewer morphine boluses in the K group than those in the P group (Table 2). The VAS score was not significantly different between 3. Results groups K and P at T0 and T30 (Table 2). Evolution of 3.1. Patient characteristics the VAS score variation (DVAS [Tx], defined as VAS [T0] À AS [Tx]) was not significantly different between Between January 01, 2004, and June 30, 2005, seventy- the groups (Fig. 2). At T30, 20 patients (61%) of the three patients were enrolled in the study (Fig. 1). Seven K group had a VAS score not exceeding 30/100 vs patients (5 in the K group and 2 in the P group) were 13 patients (41%) of the P group, representing a nonsignif- withdrawn from the analysis because of incomplete data or icant difference ( P = .2). disrespect of study protocol. One patient was excluded because of an anaphylactoid reaction after antibiotic 3.3. Secondary outcome injection in the P group. Thus, data from 65 patients were Satisfaction of patients was not significantly different completed and analyzed, 33 in the K group and 32 in between the 2 groups. There were no differences between the P group. groups with regard to blood pressure, heart rate, respiratory Baseline characteristics were similar between the rate, or oxygen saturation at T0 and T30 (Table 3). The 2 groups (Table 1). incidence of neuropsychological adverse effects was signif- 3.2. Primary outcome icantly greater in the K group, with hallucinations (n = 4), dizziness (n = 6), diplopia (n = 2), and dysphoria (n = 6), At T30, morphine consumption was significantly lower whereas there was only 1 case of dysphoria in the P group. in the K group than that in the P group, corresponding to Some patients had 2 or more symptoms.

Table 2 Comparison of morphine requirements, number of morphine boluses, VAS, and patient satisfaction between groups K and P K group P group P (n = 33) (n = 32) Morphine requirements (mg d kgÀ1, mean [95% CI]) T0 0.099 (0.097-0.102) 0.096 (0.0917-0.100) .2 T30 0.149 (0.132-0.165) 0.202 (0.181-0.223) b.001 No. of morphine boluses (median [interquartile range]) 1.0 (0.0-2.0) 2.3 (1.8-3.8) b.0001 VAS (mean [95% CI]) T0 80.4 (75.6-85.2) 80.7 (75.8-85.6) NS T30 34.1 (25.6-42.6) 39.5 (32.4-46.6) NS Patient satisfaction T30 (n [%]), excellent or good 18 (56) 22 (69) 0.3 NS indicates not significant; CI, confidence interval. Management of severe acute pain in emergency settings 389

results were similar whatever their designs, namely, that Table 3 Comparison between groups K and P for clinical parameters and adverse effects small-dose ketamine reduced morphine requirements and pain intensity [5,6,12]. Weinbroum [12] evaluated the Parameters K group P group P effects of postoperative coadministration of small doses of (n = 33) (n = 32) ketamine (0.250 mg d kgÀ1) and morphine on pain intensity Heart rate in surgical patients who complained of pain of at least 6/10 F (beats/min, mean SD) on a VAS despite more than 0.1 mg d kgÀ1 of IV morphine T0 87 F 15 81 F 18 .3 administration within 30 minutes. This study demonstrated T30 82 F 14 78 F 15 .5 d À1 Systolic blood pressure that a single dose of 0.250 mg kg of ketamine plus d À1 (mm Hg, mean F SD) 0.015 mg kg of morphine provided analgesia in 68% of T0 139 F 17 139 F 21 .9 patients compared with only 3.5% of patients who received 1 T30 136 F 20 133 F 20 .5 0.030 mg d kgÀ of morphine with saline. Pain intensity Respiratory rate was still lower in the former group 2 hours after. (breaths/min, mean F SD) The idea of combining opiates and NMDA receptor T0 20 19 .6 antagonists is based on numerous experimental data T30 18 17 .5 suggesting different and complementary mechanisms of Pulse oximetry action for these 2 classes of analgesic agents [21,22]. F (%, mean SD) Morphine acts at both the spinal and supraspinal levels. The T0 99 F 299F 1.8 spinal action is well documented [23]. It is generally T30 99 F 198F 2.2 No. of patients with 7 (21) 2 (6) .2 accepted that opioids reduce the spinal transmission of Ramsay score z3 (n [%]) nociceptive signals predominantly at presynaptic sites Adverse effects (n [%]) (ie, reduction in transmitter release from afferent C fibers), Nausea and vomiting 8 (6) 4 (6) NS although a postsynaptic inhibition of spinal dorsal horn Neuropsychological 12 (36) 1 (3) .002 nociceptive neurons has also been demonstrated. In contrast, Itching 1 (3) 1 (3) NS NMDA receptor antagonists preferentially act postsynapti- Bradypnea 0 1 (3) NS cally by reducing the hyperexcitability (ie, central sensitization) of spinal nociceptive neurons [21,24]. Thus, the reduction in nociceptive inputs associated with the reduction The level of sedation, nausea, vomiting, and itching did in postsynaptic neuronal hyperexcitability may explain the not differ between the 2 groups (Table 3). synergistic interaction of the 2 drugs [8]. Bossard et al [8] analyzed the effects of morphine, ketamine, and their combination on electrophysiological recordings of the 4. Discussion nociceptive flexion RIII reflex in 12 healthy volunteers. The stimulus response curve of the nociceptive RIII reflex In trauma patients with severe acute pain, we found that was significantly reduced after injection of a combination of the association of low-dose ketamine with morphine ketamine and morphine, and they concluded there was a reduced morphine requirements by approximately 26% synergistic interaction between the 2 drugs. However, we within 30 minutes. However, pain intensity measured on a found that there were significantly more adverse effects VAS at T30 was not different between the 2 groups. To our with ketamine. Analgesic doses of 0.150 to 0.500 mg d kgÀ1 knowledge, this is the first study that associated morphine of ketamine have been reported to produce dose- or plasma with low-dose ketamine for trauma patients with severe concentration–dependent cognitive, perceptual, and mood acute pain in emergency settings. disturbances, as well as psychotomimetic adverse effects Gurnani et al [20] studied low-dose ketamine in patients [25,26]. This could be a limitation to the use of ketamine in with acute pain after musculoskeletal trauma in emergency such settings, especially if the benefit regarding pain is poor. settings. However, in this study, patients initially received In our study, all adverse effects were weak and brief because 0.250 mg d kgÀ1 of IV ketamine, followed by a constant-rate no one needed treatment and patients’ satisfaction was not infusion of subcutaneous ketamine (0.100 mg d kgÀ1 d hÀ1) different between the 2 groups. In a review article about without morphine, whereas the control group received IV perioperative ketamine for acute postoperative pain, Bell injections of morphine (0.1 mg d kgÀ1 every 4 hours). They et al [27] found that the occurrence of adverse effects was found that the onset of analgesia was slower in the morphine similar in ketamine- and placebo-treated groups. Especially, group and that analgesia was more intense in the ketamine 21 of 37 trials specifically stated that there were no group [20]. Furthermore, the rate of patients requiring psychotomimetic adverse effects. However, in an experi- supplementary analgesia was lower in the ketamine group. mental study on 12 healthy volunteers, Bossard et al found However, these patients did not have severe pain. many psychosensory effects (8-10/12 volunteers) after the Most studies about small doses of ketamine in painful administration of ketamine, morphine, or their combination, patients were performed in postoperative settings. Their but the incidence was similar between groups. 390 M. Galinski et al.

The aim of a coanalgesic association is to avoid opioid [9] Sethna NF, Liu M, Gracely R, et al. Analgesic and cognitive effects of overdose and adverse effects with equal or better analgesia. intravenous ketamine-alfentanil combinations versus either drug alone Weinbroum [12] obtained significantly fewer adverse effects after intradermal capsaicin in normal subjects. Anesth Analg 1998; 86:1250-6. such as nausea and vomiting with ketamine and morphine [10] Kissin I, Bright CA, Bradley Jr EL. The effect of ketamine on opioid- compared with morphine and placebo (12% and 38%, induced acute tolerance: can it explain reduction of opioid consump- respectively). We did not reach this conclusion because it tion with ketamine-opioid analgesic combinations? Anesth Analg was not the first end point of our study and the patient 2000;91:1483-8. number was too small for that. [11] Angst MS, Koppert W, Pahl I, et al. Short-term infusion of the A- opioid agonist remifentanil in humans causes hyperalgesia during The observation period also has to be longer than withdrawal. Pain 2003;106:49-57. 30 minutes because there is probably a ketamine activity [12] Weinbroum AA. A single small dose of postoperative ketamine that lasts longer than its simple pharmacokinetic action provides rapid and sustained improvement in morphine analgesia particularly on hyperalgesia. This point was demonstrated in in the presence of morphine-resistant pain. Anesth Analg 2003; postoperative studies [12]. Furthermore, the ketamine we 96:789-95. used was a racemic mixture consisting of equal shares of [13] Bijur PE, Kenny MK, Gallagher EJ. Intravenous morphine at 0.1 mg/kg is not effective for controlling severe acute pain in the 2 optical enantiomers. Pharmacologic investigations show majority of patients. Ann Emerg Med 2005;46:362-5. differences between those enantiomers, with a clinical [14] Mao J, Price DD, Mayer DJ. Mechanisms of hyperalgesia and superiority of S(+)-ketamine particularly with fewer adverse morphine tolerance: a current view of their possible interaction. Pain effects [28]. Unfortunately, this enantiomer is not yet 1995;62:259-74. available in our country. [15] Adnet F, Jouriles NJ, Le Toumelin P, Hennequin B, Taillandier C, Rayeh F, et al. Survey of out-of-hospital emergency intubations in the The observation period of our trial is a limitation because French prehospital medical system: a multicenter study. Ann Emerg 30 minutes was probably too short. Indeed, Weinbroum [12] Med 1998;32:454-60. demonstrated that the analgesic effects of the ketamine [16] Ramsay MA, Savege TM, Simpson BR, Gooodwin R. Controlled group were clearly evident throughout the 120-minute sedation with alphaxalone-alphadolone. Br J Med 1974;2:256-9. observation period. [17] Ricard-Hibon A, Chollet C, Saada S, et al. A quality control program for acute pain management in out-of-hospital critical care medicine. Ann Emerg Med 1999;34:738-44. 5. Conclusion [18] Todd KH, Funk KG, Funk JP, Bonacci R. Clinical significance of reported changes in pain severity. Ann Emerg Med 1996;27: This study showed that low-dose ketamine in trauma 485-9. [19] Casagrande JT, Pike MC, Smith PG. An improved approximate patients with severe acute pain reduced morphine require- formula for calculating sample sizes for comparing two binomial ments. We could not demonstrate any reduction in pain distributions. Biometrics 1978;34:483-6. intensity with ketamine in this study. [20] Gurnani A, Sharma PK, Rautela RS, Bhattacharya A. Analgesia for acute musculoskeletal trauma: low-dose subcutaneous infusion of ketamine. Anaesth Intensive Care 1996;24:32-6. References [21] Dickenson AH. NMDA receptor antagonists: interactions with opioids. Acta Anaesthesiol Scand 1997;41:112-5. [1] Belpomme V, Ricard-Hibon A, Cholet C, et al. De´lai d’obtention de [22] Wiesenfeld-Hallin Z. Combined opioid-NMDA antagonist therapies. l’analge´sie en pre´hospitalier. Ann Fr Anesth Reanim 2002;21:324S What advantages do they offer the control of pain syndromes? Drugs [abstract]. 1998;56:1-5. [2] Paqueron X, Lumbroso A, Mergoni P, et al. Is morphine-induced [23] Yaksh TL. Pharmacology and mechanisms of opioid analgesic sedation synonymous with analgesia during intravenous morphine activity. Acta Anaesthesiol Scand 1997;41:94-111. titration? Br J Anaesth 2002;89:697-701. [24] Chapman V, Dickenson AH. The combination of NMDA antagonism [3] Kehlet H, Dahl JB. The value of bmultimodalQ or bbalanced analgesiaQ and morphine produces profound antinociception in the rat dorsal in postoperative pain treatment. Anesth-Analg 1993;77:1048-56. horn. Brain Res 1992;573:321-3. [4] Richebe´ P, Rivat C, Rivalan B, Maurette P, Simonnet G. Low doses [25] Oye I, Paulsen O, Maurset A. Effects of ketamine on sensory ketamine: antihyperalgesic drug, non-analgesic. Ann Fr Anesth perception: evidence for a role of N-methyl-d-aspartate receptors. Reanim 2005;24:11349-59. J Pharmacol Exp Ther 1992;260:1209-13. [5] Kapfer B, Alfonsi P, Guignard B, Sessler DI, Chauvin M. Nefopam [26] Krystal JH, Karper LP, Seibyl JP, et al. Subanesthetic effects of and ketamine comparably enhance postoperative analgesia. Anesth noncompetitive NMDA antagonist, ketamine, in humans: psychomi- Analg 2005;100:169-74. metic, perceptual, cognitive and neuroendocrine responses. Arch Gen [6] Suzuki M, Kentaro T, Lansing PS, Tolan MM, Fuhrman TM, Ignacio Psychiatry 1994;14:144-53. CI, et al. Small-dose ketamine enhances morphine-induced analgesia [27] Bell RF, Dahl JB, Moore RA, Kalso E. Perioperative ketamine after outpatient surgery. Anesth Analg 1999;89:98-103. for acute post-operative pain: a quantitative and qualitative systema- [7] Alvarez P, Saavedra G, Hernandez A, et al. Synergistic antinociceptive tic review (Cochrane review). Acta Anaesthesiol Scand 2005;49: effects of ketamine and morphine in the orofacial capsaicin test in the 1405-28. rat. Anesthesiology 2003;99:969-75. [28] Pfenninger EG, Durieux ME, Himmelseher S. Cognitive impairment [8] Bossard AE, Guirimand F, Fletcher D, et al. Interaction of a after small-dose ketamine isomer in comparison to equianalgesic combination of morphine and ketamine on the nociceptive flexion racemic ketamine in human volunteers. Anesthesiology 2002; reflex in human volunteers. Pain 2002;98:47-57. 96:357-66. American Journal of Emergency Medicine (2007) 25, 391–395

www.elsevier.com/locate/ajem

Original Contribution The value of serum tau protein for the diagnosis of intracranial injury in minor head trauma

Cemil Kavalci MDa, Murat Pekdemir MDa,*, Polat Durukan MDa, Necip Ilhan PhDb, Mustafa Yildiz MDa, Selami Serhatlioglu MDc, Dilara Seckin MDb aDepartment of Emergency Medicine, Firat University Faculty of Medicine, ElazVg˘/Turkey bDepartment of Biochemistry and Clinical Biochemistry, Firat University Faculty of Medicine, ElazVg˘/Turkey cDepartment of Radiodiagnostic, Firat University Faculty of Medicine, ElazVg˘/Turkey

Received 16 June 2006; revised 25 July 2006; accepted 14 October 2006

Abstract Objective: Tau protein localizes in the axons of neuron cells, and it is released secondarily from the central nervous system because of hypoxia and trauma. In the present study, it was aimed to investigate the value of serum tau protein levels in diagnosing intracranial pathologies in minor head trauma. Methods: Patients were categorized into 2 groups: those without intracranial lesions in head CTs (group 1) and those with lesions in head CTs (group 2). Serum tau protein levels were determined. Results: Group 1 (n = 55) median serum tau protein level was 16.29 pg/mL (2.12-215.97 pg/mL) and group 2 (n = 33) median serum tau protein level was 18.39 pg/mL (2.19-714.47 pg/mL). Statistical analysis revealed no significant difference between the 2 groups for tau protein values, sex, age, mechanism of trauma, and Glasgow Coma Scale score. Conclusion: It is suggested that serum tau protein has limited value in minor head trauma. D 2007 Elsevier Inc. All rights reserved.

1. Introduction Minor or mild traumatic brain injuries (defined as loss of consciousness b30 minutes, amnesia b24 hours, or peri- Head injury is one of the major health problems in injury confusion/disorientation in a patient with a Glasgow emergency medicine. There are approximately 1 million Coma Scale [GCS] score of 13-15) constitute nearly 70% to emergency department (ED) visits annually for traumatic 90% of traumatic brain injuries that occur worldwide [3-5]. brain injury in the United States [1], or an incidence of 56.4/ Ruling out pathology is essential for patients with closed 100000 of US population [2]. head injury because it may require immediate intervention or cause rapid deterioration. At present, the ED workup of intracranial hemorrhage, brain edema, and impending herni- ation is predominantly limited to clinical evaluation and This study was presented at Tqrkiye Acil TVp Sempozyumu & 3. Acil computed tomography (CT) [6]. However, the routine use of Y Hem irelig˘i ve Paramedik Sempozyumu, 24-27 KasVm 2004, Gaziantep. CT in screening patients with minor head trauma (MHT) for * Corresponding author. Department of Emergency Medicine, Faculty of Medicine, Kocaeli University, Kocaeli, Turkey. Tel.: +90 262 3038547; intracranial lesions is expensive. It is estimated that even a fax: +90 262 3038003. 10% reduction in the number of CT scans in patients with E-mail address: [email protected] (M. Pekdemir). MHT would save more than $20 million per year [7].

Emergency physicians have sought for the development of a Table 2 Pathologic findings in cranial CT clinical decision rule, which is essential for more standardized and efficient use of cranial CT in MHT [8]. Moreover, to Findings Head CT (À) Head CT (+) standardize and improve the emergency management of group (n = 55) group (n = 33) patients with MHT, diagnostic and prognostic use of serum Brain edema – 26 markers such as neuron-specific enolase, creatine kinase–BB, Linear fracture 6 6 myelin basic protein, and S-100b for traumatic brain injury Depressed fracture – 4 have also been intensely investigated. Tau protein is also one Pneumocephalus – 5 Cerebral contusion – 3 of these serum markers; however, to date, it has not been Intraparenchymal –2 extensively studied in patients with MHT. hemorrhage Tau protein has a microtubule structure and is localized in SAH – 1 the axons of central nervous system (CNS) neurons. [9-11]. Epidural hematoma – 1 It is expressed from a single gene that undergoes alternate SAH indicates subarachnoid hemorrhage. splicing resulting in 6 tau isoforms with apparent molecular masses between 48 and 68 kd [12]. It binds to axonal microtubules and forms axonal microtubule bundles. These trauma, GCS scores, physical examination findings, cranial bundles form important structural elements in the axonal CT results, and ED outcomes were also gathered. cytoskeleton, and they are also critical elements in the Venous blood samples taken from each patient were axoplasmic flow of proteins between the nerve terminal and centrifuged for 15 minutes at 4000g. After the centrifuga- neuronal cell body [13]. Tau also stabilizes the microtubules tion, each sample and serum was stored at À808C while and arranges the internal cell vesicular transport and interacts awaiting assay for tau protein. Human tau immunoassay kit with actins in neuronal formation and cell skeleton. Hypoxic (BioSource International, Camarillo, CA) with sandwich or traumatic axonal injury causes tau protein to be released enzyme-linked immunosorbent assay method was used in extracellularly by CNS neurons [14]. the analysis. Sensitivity of enzyme-linked immunosorbent The aim of our study was to evaluate the role of serum assay assessment was less than 12 pg/mL [14]. tau protein concentrations in patients with MHT. We also All of the cranial CT studies done in the study were investigated the concordance between cranial CT findings evaluated by the same radiologist. and serum tau levels. The relationship between other The patients were divided into 2 groups. Group 1 demographic and diagnostic clinical factors and serum tau consisted of patients with normal cranial CT findings and level was also investigated. linear fracture, whereas the patients in group 2 had intracranial lesions such as brain edema, epidural hemato- 2. Methods ma, subdural hematoma, subarachnoid bleeding, cerebral contusion, intraparenchymal bleeding, and depressed frac- This prospective cross-sectional study was approved by ture on cranial CT. the local ethics committee of Firat University. Informed SPSS 11.0 software (SPSS Inc, Chicago, Ill) was used for consent was obtained from every patient. the statistical analysis. Demographic and clinical features of The patients with diagnosis of MHT admitted to the ED the patients were examined according to mean F SD, of the university hospital between January 8 and June 26, median, range, and percentage. The normal distributions 2003, were prospectively analyzed. Patients with blunt head were tested with Shapiro-Wilk test. Kruskal-Wallis test was trauma in the first 24 hours requiring screening with cranial used for multiple continuous group comparisons, whereas CT were included in the study. The inclusion and exclusion Student t test and Mann-Whitney U test were used for criteria are shown in Table 1. 2 continuous group comparisons. v2 Test was used for the Demographic features of the patients, mechanisms of the categorical variables. Linear regression was used to examine trauma, presence accompanying injuries, time passed after the influence of independent variables (GCS score, pathology in cranial CT, sex, age, injury mechanism, presence Table 1 Inclusion and exclusion criteria of study in patients with MHT accompanying injuries, and time from the trauma to blood sampling) on serum tau protein levels. P b .05 was Inclusion criteria Exclusion criteria considered to be statistically significant. Power of the study Blunt head trauma Penetrating head trauma was determined as 0.23. Patients with cranial CT Coagulation disorder or use of anticoagulants Presentation at first History of brain operation 24 hours of trauma 3. Results Old neurologic deficit Presentation after 24 hours Eighty-eight patients fitting the inclusion criteria were after trauma included in the study. There were 55 patients (63%) in group 1 and 33 patients (37%) in group 2. The value of serum tau protein in minor head trauma 393

examinations were scalp laceration (56.8%), scalp hemato- Table 3 Demographic and clinical features of patients ma (20.5%), depressed fracture (4.5%), periorbital hemato- Features Head CT (À) Head CT (+) P ma (3.4%), focal neurologic findings (2.3%), otorrhea group (n = 55) group (n = 33) (1.1%), and rhinorrhea (1.1%). Age (mean F SD) 24.36 F 20.4 24.91 F 16.8 .892 Admission time, min 160 (30-1080) 180 (60-840) .474 3.2. Tau protein (median [range]) Sex (male/female) 40/15 22/11 .717 The median serum tau protein level was 16.29 pg/mL Children/adult 24/31 10/23 .309 (2.12-215.97 pg/mL) in group 1, whereas it was 18.39 pg/ Mechanism of mL (2.19-714.47 pg/mL) in group 2, with no significant trauma difference between the groups ( P = .515). Motor vehicle 17 12 The median serum tau protein levels were 16.62 pg/mL accidents (2.12-215.97 pg/mL) and 17.60 pg/mL (3.42-714.47 pg/ Pedestrian accidents 10 3 .892 mL) for the male and female patients, respectively, also with Falls from height 25 17 no statistically significant difference ( P = .714). Violent assaults 3 1 Concomitant injury 16 (29%) 17 (51%) .061 There was also no significant difference in serum tau SBP (mean F SD) 109.82 F 21.5 120.91 F 26.3 .045 protein levels between children and adults (18.61 vs 16.62 DBP (mean F SD) 67.45 F 13.8 72.42 F 14.8 .123 pg/mL, P = .377). Pulse rate 95.44 F 18.1 93.64 F 16.1 .630 Median serum tau protein levels also showed no (mean F SD) significant difference according to the mechanism of injury RR (mean F SD) 20.07 F 2 21.12 F 2.9 .077 (Table 4). SBP indicates systolic blood pressure; DBP, diastolic blood pressure; Serum levels of tau protein according to GCS scores are RR, respiration rate. also shown in Table 5. There was no statistically significant difference in serum tau protein levels according to GCS scores ( P = .408). Fifty-five patients (62.5%) had isolated head injuries, and A multiple regression model was constructed to inves- 33 patients (37.5%) had accompanying injuries. These tigate the influence of the variables on tau protein levels. injuries were extremity fracture (13 patients), pneumothorax Glasgow Coma Scale score, pathology in cranial CT, sex, (3 patients), vertebrae fracture (3 patients), splenic rupture age, injury mechanism, presence of accompanying injuries, (2 patients), liver laceration (1 patient), and pelvic fracture and the time from the trauma to blood sampling were (1 patient). Pathologic findings in cranial CT are shown examined by multiple regression analysis. Glasgow Coma in Table 2. Scale score was the only influential variable on tau protein 3.1. Demographic and clinical features levels (b = À0.301, P = .004). Sixty-two patients (70.5%) were male, and 54 (61.4%) were adults. The mean age was 24.57 F 19.0 years 4. Discussion (3 months-80 years). The demographic and clinical features of the patients are presented in Table 3. No significant Serum tau protein levels were increased in patients with difference was found between the groups in terms of MHT who had intracranial lesions demonstrated in cranial demographic and clinical features except for mean systolic CT; however, the increase was not statistically significant. blood pressure. In addition, there was no difference in serum tau protein Falling from height was the most frequent trauma levels between subgroups (age, sex, mechanism of injury, mechanism in the study (47.7%). The most frequently seen and GCS scores). symptoms were drowsiness (47.7%), headache (40.9%), Traditionally, GCS score has been used to determine the amnesia (23.9%), unconsciousness (10.2%), and seizure category of brain injury. The term MHT is usually used for (4.5%). The most frequently observed findings in physical patients with a GCS score of 13 or higher. However, some

Table 4 Tau protein levels according to trauma mechanisms Mechanism Head CT (À) group* (median) Patients (n) Head CT (+) group**median) Patients (n) P MVA 19.76 (3.57-215.97) 17 15.82 (2.19-37.79) 12 .479 Pedestrian 24.55 (5.91-201.09) 10 24.35 (9.59-50.29) 3 .866 Falls 15.07 (2.12-175.82) 25 23.72 (3.42-714.47) 17 .163 Violent assault 14.76 (6.13-17.99) 3 19.94 1 – MVA indicates motor vehicle accident. * P = .694, v2 = 1.447; Kruskal-Wallis test. ** P = .564, v2 = 2.039; Kruskal-Wallis test. 394 C. Kavalci et al.

Table 5 The patients’ tau protein levels according to GCS scores GCS score Head CT (À) group* (median) Patients (n) Head CT (+) group**median) Patients (n) P 13 – – 25.06 (6.23-714.47) 7 – 14 14.47 (3.17-175.82) 9 66.57 (10.96-197.21) 6 .099 15 16.67 (2.12-215.97) 46 16.98 (2.19-125.21) 20 .586 * P = .838, v2 = .042; Kruskal-Wallis test. ** P = .101, v2 = 4.579; Kruskal-Wallis test. authors have suggested that patients with a GCS score of 13 To date, there is no report in the literature showing a could be excluded from the mild category and could be relationship between serum tau protein level and trauma placed in the moderate-risk group because of their high mechanism, presence of other injuries, age, and sex. incidence of lesions requiring neurosurgical intervention Similarly, there was also no significant correlation between [15]. As a limitation in the study, patients with a GCS score serum tau protein levels and these variables in our study. of 13 were also included; however, the results show that Some researchers have investigated S-100b protein as an there was no significant difference between the serum tau indicator of CNS injury. Woertgen and colleagues [27] have protein levels of patients with a GCS score of 13 and those demonstrated that serum S-100b levels might be useful in of patients with GCS score of 14 and 15. reflecting clinical conditions of patients with severe head The tau protein level that can be measured in serum is injuries. High serum levels of S-100b related to intracranial released from CNS neurons secondarily because of trauma pathologies in magnetic resonance imaging have also been and hypoxia. Previous studies suggest that measuring the reported [28]. Yamazaki and colleagues [29] have reported levels of the microtubule-associated protein tau in cerebro- that the sensitivity and specificity of neuron-specific spinal fluid (CSF) may provide an alternative method to enolase levels higher than 20 ng/mL are 87% and 82%, assessing axonal injury in traumatic brain injury. Tau is a respectively, in intracranial pathologies demonstrated in protein localized primarily in neurons and demonstrates cranial CT. However, in our study, no significant correlation selective axonal stabilization as well, resulting in tau was found between tau protein levels and pathologic sequestration in the axonal compartment [9,16]. findings in cranial CT. Schunk and colleagues [17] have shown that there is a 4.1. Limitations of the study positive correlation between high serum and CSF tau protein levels and intracranial pressure, but a negative The statistical power of the study is poor because of the correlation with poor clinical prognosis. Few postmortem relatively small number of patients. Secondly, a small studies have reported that tau protein levels were high in number of patients (7 patients, 8%) with a GCS score of oligodendrogliocytes of people who died after sustaining 13 were also included in the study. However, although all head injuries [18]. It has also been suggested that it is these patients had intracranial pathologies demonstrated in clinically useful to measure CSF tau protein levels in axonal cranial CT, the results showed that there was no significant injuries by head traumas [19]. Furthermore, there is a difference between the serum tau protein levels of patients relationship between serum tau protein levels and intracra- with a GCS score of 13 and those of patients with GCS nial injury, disability, and death only in the closed head score of 14 and 15. injuries [20]. The discrepancy between findings in this study In conclusion, our study showed that there was a and those in aforementioned studies may be attributed to relationship between serum tau protein concentration and larger sample sizes and the inclusion of patients with only severity of trauma, although it did not achieve statistical MHT [21-25]. significance. However, serum tau protein concentration Bulut and colleagues [26] have reported that serum tau does not seem to be useful in indicating the intracranial protein levels of patients with mild traumatic brain injury lesions in patients with MHT. There was also no were relatively higher than those of healthy volunteers; significant correlation between serum tau protein levels however, the difference was not statistically significant. Our and cranial CT, age, sex, trauma mechanism, presence of study also revealed that serum tau protein levels in patients other injuries, GCS score, symptoms, and physical with intracranial lesions increased in correspondence to their examination findings. Although the results of this study cranial CT, although there was no statistically significant should be viewed with caution because of the relatively difference. The lack of statistical significance may be due to small number of patients, they still provide an evidence for the poor power of the study. the significance of serum tau protein concentrations in The GCS score negatively relates with serum tau protein patients with MHT. Tau can assist physicians in deciding level. The correlation between severity of head injury and whether cranial CT is necessary for patients with MHT. serum tau protein level appears to be consistent with the Further studies with higher number of patients can provide literature. The GCS score has also a negative relationship better information about the relationship between tau with both CNS degeneration and serum tau protein level. protein and head trauma. The value of serum tau protein in minor head trauma 395

References [15] Dietrich AM, Bowman MJ, Ginn-Pease ME, et al. Pediatric head injuries: can clinical factors reliably predict an abnormality on [1] Jager TE, Weiss HB, Coben JH, et al. Traumatic brain injuries computed tomography? Ann Emerg Med 1993;22:1535-40. evaluated in U.S. emergency departments, 1992-1994. Acad Emerg [16] Gruskin KD, Schutzman SA. Head trauma in children younger than Med 2000;7:134-40. 2 years: are there predictors for complications? Arch Pediatr Adolesc [2] Bazarian JJ, McClung J, Cheng YT, et al. Emergency department Med 1999;153:15-20. management of mild traumatic brain injury in the USA. Emerg Med J [17] Schunk JE, Rodgerson JD, Woodward GA. The utility of head 2005;22:473-7. computed tomographic scanning in pediatric patients with normal [3] Centers for Disease Control and Prevention, National Center for neurologic examination in the emergency department. Pediatr Emerg Injury Prevention and Control. Report to Congress. Mild traumatic Care 1996;12:160-5. brain injury in the United States: steps to prevent a serious public [18] Quayle KS, Jaffe DM, Kuppermann N, et al. Diagnostic testing for health problem. Atlanta (Ga)7 Centers for Disease Control and acute head injury in children: when are head computed tomography Prevention; 2003. and skull radiographs indicated? Pediatrics 1997;99:11-23. [4] Cassidy JD, Carroll LJ, Peloso PM, et al. Incidence, risk factors and [19] Ingebrigtsen T, Romner B, Kock-Jensen C. Scandinavian guidelines prevention of mild traumatic brain injury: results of the WHO for initial management of minimal, mild and moderate head injuries. Collaborating Centre Task Force on mild traumatic brain injury. J Trauma 2000;18:760-6. J Rehabil Med 2004;(Suppl 43):28-60. [20] Jagoda AS, Cantrill SV, Wears RL, et al. Clinical policy: neuro- [5] von Holst H, Cassidy JD. Mandate of the WHO Collaborating Centre imaging and decision-making in adult mild traumatic brain injury in Task Force on mild traumatic brain injury. J Rehabil Med 2004; the acute setting. Ann Emerg Med 2002;40:231-49. (Suppl 43):8-10. [21] Kanai Y, Hirokawa N. Sorting mechanisms of tau and MAP2 in [6] Neumar RW. Rule out TBI? Serum markers for traumatic brain injury. neurons: suppressed axonal transit of MAP2 and locally regulated Ann Emerg Med 2002;39:342-3. microtubule binding. Neuron 1995;14:421-32. [7] Reinus WR, Wippold II FJ, Erickson KK. Practical selection criteria [22] Zemlan FP, Jauch EC, Mulchahey JJ, et al. C-tau biomarker of for noncontrast cranial computed tomography in patients with head neuronal damage in severe brain injured patients: association with trauma. Ann Emerg Med 1993;22:1148-55. elevated intracranial pressure and clinical outcome. Brain Res [8] Stiell IG, Wells GA, Vandemheen K, et al. Variation in ED use of 2002;947:131-9. computed tomography for patient with minfr head injury. Ann Emerg [23] Irving EA, Nicoll J, Graham DI, et al. Increased tau immunoreactivity Med 1997;30:14-22. in oligodendrocytes following human stroke and head injury. Neurosci [9] Binder LI, Frankfurter A, Rebhun LI. The distribution of tau in the Lett 1996;213:189-92. mammalian central nervous system. J Cell Biol 1985;101:1371-8. [24] Raabe A, Grolms C, Keller M, et al. Correlation of computed [10] Kosik KS, Finch EA. MAP2 and tau segregate into dendritic and tomography findings and serum brain damage markers following axonal domains after the elaboration of morphologically distinct severe head injury. Acta Neurochir 1998;140:787-92. neurites. J Neurosci 1987;7:3142-53. [25] Shaw GJ, Jauch EC, Zemlan FP. Serum cleaved tau protein levels and [11] Litman P, Barg J, Rindzoonski L, et al. Subcellular localization of tau clinical outcome in adult patients with closed head injury. Ann Emerg mRNA in differentiating neuronal culture: implications for neuronal Med 2002;39:254-7. polarity. Neuron 1993;10:627-38. [26] Bulut M, Koksal O, Dogan S, et al. Tau protein as a serum marker of [12] Goedert M, Spillantini MG, Jakes R, et al. Multiple isoforms of human brain damage in mild traumatic brain injury: preliminary results. Adv microtubule-associated protein tau: sequence and localization in neuro- Ther 2006;23:12-22. fibrillary tangles and Alzheimer’s disease. Neuron 1989;3:519-26. [27] Woertgen C, Rothoerl RD, Metz C, et al. Comparison of clinical, [13] Zemlan FP, Rosenberg WS, Luebbe PA, et al. Quantification of axonal radiologic and serum marker as prognostic factors after severe head damage in traumatic brain injury: affinity purification and charac- injury. J Trauma 1999;47:1126-30. terization of cerebrospinal fluid tau proteins. J Neurochem 1999;72: [28] Ingebrigtsen T, Romner B. Serial S-100 protein measurements related 741-50. to early magnetic resonance imaging after minor head injury. [14] Human Tau (total) Biosource Immunoassay Kit Catalog #KHB0042/ J Neurosurg 1996;85:945-8. KHB0041. Biosite website available at: http://www.biosource.com/ [29] Yamazaki Y, Yada K, Morii S, et al. Diagnostic significance of serum content/catalogContent/tds3/moreinfo/KHO0631%20pr328%20revA1% neuron specific enolase and myelin basic protein assay in patients with 20mar2706%20(Hu%20Tau%20[pT181]).pdf [Accessed: 1.06.2006]. acute head injury. Surg Neurol 1995;43:267-71. American Journal of Emergency Medicine (2007) 25, 396–399

www.elsevier.com/locate/ajem

Original Contribution Frequency of radiology self-referral in abdominal computed tomographic scans and the implied cost

Michael Blaivas MD*, Matthew Lyon MD

Section of Emergency Ultrasound, Department of Emergency Medicine, Medical College of Georgia, AF-2056, Augusta, GA 30912-4007, USA

Received 2 June 2006; accepted 9 September 2006

Abstract Concerns over rising imaging costs have led to the consideration by Medicare to limit the ability of clinicians to bill for image interpretation. This move has often been justified as a method to limit self- referral. However, clinicians may not be the only ones capable of referring imaging business to themselves. Objective: This study was conducted to evaluate the frequency and potential cost of self-referral by radiologists found in dictated reports of abdominal computed tomographic (CT) scans for emergency patients. Methods: A retrospective chart review of all abdominal CTs performed at a level 1 academic urban emergency department (ED) with an annual census of 80000 patients for a 12-month period was performed. Two investigators reviewed the medical record dictation on each abdominal CT performed on ED patients older than 18 years, for specific recommendations for additional radiologic testing. To check for agreement, both investigators reviewed approximately 20% of the charts. Recommended additional radiologic tests were recorded, and their costs were estimated by the lowest regional Medicare reimbursements for each test; professional and facilities fees were combined. Statistical methodology included descriptive statistics and interrater agreement. Results: A total of 785 reports of abdominal CTs were reviewed. Of these reports, 246 (31%) specifically recommended an additional imaging study be obtained for a specific finding. In 38 (5%) cases, 2 separate imaging studies were suggested. The total lowest cost for additional imaging among all of the patients studied was $58157. The mean suggested charge per patient with additional imaging self- referral was $242.32. The additional suggested imaging averaged to $74.09 (95% confidence interval, 63.67-84.50) for each patient receiving an abdominal CT scan in the ED. The largest suggested cost was $1045. Extrapolation to a national level means that more than $226 million of additional costs are seen annually from such CTs. No attempt was made to evaluate the appropriateness of the suggested imaging. Conclusions: A type of radiology self-referral is possible and can add considerable cost to patient care. In our study, an average of $74 of extra imaging was suggested for each patient who received an abdominal CT. If this holds up nationwide, Medicare can expect at least $226 million worth of radiology self-referrals per year on patients getting abdominal CT. D 2007 Elsevier Inc. All rights reserved.

* Corresponding author. Tel.: +1 706 721 2613. E-mail address: [email protected] (M. Blaivas).

1. Introduction Table 2 Most frequently suggested follow-up imaging studies among abdominal CTs with additional imaging Recent cost overruns in the Medicare system have suggested prompted authorities to seek cost-cutting measures [1].As medical imaging has been responsible for annual cost Study suggested in report No. of suggestions (%) increases of 9% or more, diagnostic medical imaging has Pelvic ultrasound 89 (36) been targeted for utilization review by the Medicare Abdominal CT 39 (16) Payment Advisory Commission (MedPac) [2]. A concern Right upper quadrant ultrasound 28 (11) of MedPac is the potential for self-referral of diagnostic Chest CT 25 (10) Renal ultrasound 23 (9) imaging studies, particularly clinician-performed ultrasound. Because the ordering physician is the same as the performing physician, there is a potential that clinicians will examinations were performed by the department of radiol- order tests that are motivated by financial concerns rather ogy and read by radiology attending physicians. than clinical indications. Although this has been viewed as All emergency medicine patients, 18 years and older, who legal, some organizations continue to argue for additional received abdominal CTs during a 12-month period were regulations to limit payment for clinician-performed studies, eligible for the study. The medical records department was as it is a significant conflict of interest that adds to increasing asked to generate a list containing the names and medical imaging costs [2,3]. In contrast, radiologists have tradition- record numbers as well as dates of procedure for all patients ally been thought to not be capable of self-referral and receiving one of the International Classification of Diseases, therefore better able to make decisions regarding imaging Ninth Revision codes for abdominal CT. All reports were needs and cost containment. However, as most radiologic reviewed in entirety by 1 of 2 board-certified emergency departments are single group entities, self-referral is physicians. To evaluate for concordance between report possible as one radiologist is able to refer additional studies analyses and identify any differences in recording or that will be read by them or their partners. interpretation, both physicians reviewed approximately 20% We sought to determine the rate of self-referral by radio- of the patient charts. For each abdominal CT report, the logists in abdominal computed tomography (CT) reports researchers noted any recommended imaging studies and the performed in a busy urban emergency department (ED). As finding that caused them. Multiple recommended studies and a secondary objective, we tabulated the costs associated findings were recorded when present. Regional and local with the suggested additional imaging studies. Medicare reimbursement allowances for a variety of imaging tests were identified, and the lowest rate was noted for each 2. Methods test. Facilities charges and professional charges were combined for each examination. When more than 1 possible This was a retrospective review of all abdominal CT test was suggested to evaluate a finding, the less expensive reports generated for ED patients for a 12-month period. test was selected and recorded. When multiple tests were The study was approved by our institutional review board. suggested for multiple findings, each test finding and test Written informed consent was waived, as the authors were recorded, and costs were added. Although, in many reviewed reports in a blinded fashion, and no identifying cases, incidental findings of possible abdominal pathology or personal information was collected on any patient. This would result in ordering abdominal and pelvic CT in study took place at an urban, level I trauma center ED with combination, we recorded abdominal CT only unless both an annual census of 80000 patients. The ED is a regional were specifically suggested to assume the lowest cost. The referral center and has an emergency medicine residency same was true for abdominal ultrasound. Only 1 quadrant program. Patients can receive abdominal CTs 24 hours a was recorded instead of comprehensive examinations, which day. Computed tomography with or without contrast can be are performed most commonly in radiology laboratories. performed, and the facility uses 1 of several multislice CT Data were entered into a Microsoft Access database scanners (GE, Milwaukee, WI). Computed tomographic (Redmond, Wash) that was then used for queries. Data were then exported to a Microsoft Excel spreadsheet used for statistical analysis. All data were analyzed using commer- Table 1 Most common findings noted on abdominal CT cially available statistical software, StatsDirect (Cheshire, report, leading to additional imaging UK). Descriptive statistics were used to evaluate the data, Finding No. of occurrences and interrater reliability was checked. Ovarian or adnexal cyst 55 Renal cyst 28 3. Results Lung nodule 23 Gallbladder pathology 15 The 2 study physicians reviewed a total of 785 reports Liver cyst 13 of abdominal CTs on ED patients. There were no disagree- ments between the reviewing physicians. Of these reports, 398 M. Blaivas, M. Lyon

246 (31%) specifically suggested an additional imaging $74 of paid charges per abdominal CT ordered. This resulted study be obtained for a specific finding. The most common in a total of $58157 of additional imaging reimbursement findings are listed in Table 1. In 38 (5%) cases, 2 separate during the study period. We used the most conservative imaging studies were suggested, whereas in 5 (1%) cases, estimates for regional Medicare reimbursement for this 3 separate imaging studies were suggested. Table 2 contains calculation. Using data from 2003 when Medicare paid out the most commonly suggested imaging tests. on at least 3064000 (rounded off) claims to radiologists for The mean Medicare payment for each suggested test was abdominal CTs, we found that an additional expense of $242.32. On average, $74.09 (95% confidence interval, $74 per patient would total to more than $226 million per 63.67-84.50) of additional Medicare imaging cost was year to Medicare alone. The total number of CT scans suggested per patient receiving an abdominal CT. The total performed by radiologists per year is actually closer to cost for all of the patients studied was $58157. The largest 41 million and will yield an additional expense of nearly suggested cost for a single patient was $1045. $3 billion if the same trend is maintained for all CTs performed [5,6]. Actual bills sent would be much higher. 4. Discussion Although these numbers are impressive, this in no way suggests that the additional diagnostic studies were inten- The controversy over imaging in the emergency setting is tionally designed to generate extra fees or that they were all nothing new and has centered on the inability of emergency or in part inappropriate. However, we were surprised to note physicians to obtain appropriate studies 24 hours a day [4]. that referral patterns were inconsistent, suggesting that self- In addition, lack of contemporaneous reading of imaging policing by imaging specialists may not be as reliable as studies by radiologists often occurs despite Medicare suggested by imaging specialty advocates [7].When b Q regulations mandating contemporaneous interpretation by possible renal cysts were seen on stone hunt CTs, some an attending physician [4]. Many clinicians are now filling reports recommended ultrasound for follow-up, yet others the void of available diagnostic imaging themselves by suggested MRI, and others suggested simply clinical performing and interpreting their own diagnostic imaging correlation or nothing at all. Similarly, for simple pelvic tests, whether those are ultrasound, CT, or magnetic reso- cysts, recommendations ranged from clinical correlation to nance imaging (MRI). Although clinician-performed imag- pelvic ultrasound and pelvic CT to MRI. Furthermore, ing studies only make up a small percentage of all imaging follow-up was sometimes suggested for findings clearly studies performed each year, there has been a concern that diagnosed on the primary CT. An example of this included clinicians may perform (and bill insurers for) substandard functional follicles on ovaries, which often lead to a and inappropriate imaging tests [3]. recommendation for a pelvic ultrasound examination. Despite the potential that clinician-performed and clini- We did not analyze individual cases for appropriateness of cian-interpreted studies may lead to improvement in patient suggested additional imaging tests. However, it is clear that care, there is a possibility that a clinician may perform an clinicians are not the only ones potentially driving the imaging test for less honorable reasons. Incentives for increased use of imaging tests and thus imaging costs clinician-performed imaging studies include easier access to for Medicare. Although guidelines for imaging appropriate- the imaging test, contemporaneous interpretation and there- ness, as suggested by the American College of Radiology, fore faster diagnosis, and income from billing for the may help decrease costs, such guidelines cannot be made examination. With a potential financial incentive, there is a by imaging specialists alone because of the tendency to possibility for self-referral of diagnostic studies by the self-refer at least in some situations [7]. Furthermore, not only clinician. In contrast, there has been a justifiable assumption should the development of such guidelines include clinician that self-referral by radiologists is not possible because radio- input, additional guidelines should be in place describing logists are not clinicians and do not order tests. However, under what circumstances radiologists should suggest further because of the current medicolegal environment in the United imaging and which tests are appropriate for specific findings. States, many clinicians feel compelled to order follow-up This study had several limitations. We did not analyze studies if recommended by radiology consultants. Further- individual cases for appropriateness of self-referral. The more, most radiologists practice in multiphysician groups that study made no effort to control for variability among have exclusive contracts with a hospital. Hence, it is possible radiologists. Although a sample of patients was checked by for a radiologist from a multiperson group to indirectly refer both emergency physicians and no differences in analyzing additional studies to partners in their group by including in CT reports were noted, both physicians did not evaluate all their report a recommended follow-up radiologic study. Many reports. We did not track how often suggested imaging was such studies are likely to be ordered during the same ED visit obtained. This was outside the scope of study allowed by our unless long-term follow-up is specified. Even later studies are institutional review board. In addition, because this study still likely to be performed at the same hospital or imaging was conducted at 1 academic medical center, these results center where the patient was initially sent. may not be applicable to all radiologic practices. Our data showed that radiologists frequently recommen- In conclusion, a type of radiology self-referral is possible ded further diagnostic studies that averaged an additional and can add considerable cost to patient care. In our study, Radiology self-referral 399 an average of $74 of extra imaging was recommended for [2] Hackbarth GM, Reischauer RD, Miller ME. Assessing payment each patient who received an abdominal CT. If this holds up adequacy and updating payments for physician services. Medicare nationwide, Medicare can expect at least $226 million worth Payment Advisory Commission report to Congress, March 2003. [3] Maitino AJ, Levin DC, Parker L, Rao VM, Sunshine JH. Practice of radiology self-referrals per year on patients getting patterns of radiologists and nonradiologists in utilization of noninvasive abdominal CT. diagnostic imaging among the Medicare population 1993-1999. Radiology 2003;228:795-801. [4] Carroll TJ. Trends in on-call workload in an academic medical center References radiology department 1998-2002. Acad Radiol 2003;10:1312-20. [5] Maureen Glabman MANAGED CARE January 2005. [1] Levin DC, Rao VM. Turf wars in radiology: the overutilization of [6] Miller JC. Radiology rounds Feb 2004, 1-3. imaging resulting from self-referral. J Am Coll Radiol 2004;1:169-72. [7] Gardner J. Is MedPAC stacked? Mod Healthc 2000;30:64. American Journal of Emergency Medicine (2007) 25, 400–405

www.elsevier.com/locate/ajem

Original Contribution Prehospital use of analgesics at home or en route to the hospital in children with extremity injuries

Alex L. Rogovik MD, PhD, Ran D. Goldman MD*

Pediatric Research in Emergency Therapeutics (PRETx) Program, Division of Pediatric Emergency Medicine, Department of Pediatrics, The Hospital for Sick Children, University of Toronto, Toronto, ON, Canada M5G 1X8

Received 2 August 2006; revised 23 November 2006; accepted 27 November 2006

Abstract Purpose: The purpose of the study was to document prehospital analgesia (PA) for children with extremity injuries at home or en route to the hospital, as assessed by research personnel at the pediatric emergency department. Methods: Two parallel groups of patients with fractures or soft-tissue injuries (STIs) were chosen for this prospective observational study. Patients 3 to 18 years of age with a limb or clavicle injury were enrolled. Parents or children were interviewed, pain assessed, and data from the emergency department charts collected. Results: A total of 310 patients were recruited; their mean age was 10.2 years, and 62% had fractures. The median pain score was 4.0, with no significant difference between fractures and STI. Of the patients, 78% had PA, 73% received first aid (icing, immobilization), and 37% had medication, mostly acetaminophen and ibuprofen. Children with fractures and STI received PA at a similar rate; however, the time to first aid was shorter in those with fractures. Conclusion: Most patients with moderate or severe pain did not receive prehospital pain medication. Parental education and moderate over-the-counter analgesics are needed for better pain relief. D 2007 Elsevier Inc. All rights reserved.

1. Introduction Previous reports documented inadequate prehospital use of analgesia by emergency medical services (EMS) personnel Pain is one of the most common symptoms in patients [6,7] in children with injuries, especially among the younger presenting to the emergency department (ED) [1] and is age group [8,9]. perceived as a difficult issue for the pediatric ED care Underestimation of patients’ needs [1], uncertainty as to providers [2,3] despite evidence that a child’s experience of routes of administration, pediatric drug dosing, and per- pain is as severe as that of an adult [4,5]. ceived discomfort with the care of an injured child are all Current knowledge on prehospital analgesia (PA) for barriers to administration of analgesia in children [6]. In its children at home or en route to the hospital is limited. recent emergency practice guidelines, the American Acad- emy of Pediatrics recommended eliminating any barriers preventing appropriate and timely administration of analge- * Corresponding author. Tel.: +1 416 813 5418; fax: +1 416 813 5043. sia to the child who requires emergency treatment and E-mail address: [email protected] (R.D. Goldman). taking every opportunity to use available methods of pain

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2006.11.021 Analgesia for pediatric extremity injuries 401 control [10]. The National Association of EMS Physicians To minimize bias between physicians, researchers, and [11] recommends including both nonpharmacologic and parents [15], the children were asked to report and score pharmacologic interventions in prehospital pain manage- their pain directly. The 1-page pain assessment form ment protocols, with first aid measures such as application contained the Wong-Baker pain scale [16] for children of ice packs, immobilization of fractures, or elevation of aged 3 to 6 years and a visual analogue scale [17] for extremities to be provided regardless of whether or not children aged 7 years and older. Mild pain was defined as a medications are used. score of 1 to 3, moderate pain as 4 to 6, and severe pain as We are unaware of any previous studies providing insight 7to10[18]. into parental analgesia except for one survey of parents/ Nonpharmacologic first aid pain relief measures such as guardians of children with head injuries or limb problems, application of ice pack or cold water, immobilization, and including burns, conducted at an accident and emergency dressing [19] were also documented. The data collected (A&E) department in the UK by Spedding et al [12]. They retrospectively from the ED charts included triage time, area found that 74% of children did not receive pain medication of injury, and outcome (fracture or STI). before attendance at the A&E department. First aid non- The primary end point was difference in PA (both pharmacologic pain relief by parents was not documented. pharmacologic and nonpharmacologic [first aid]) rate Studies conducted in EMS adult patients have also between patients with and without fractures. Secondary shown that patients with suspected fractures rarely receive measures included time from injury to PA, characteristics of PA [13,14]. analgesia, and age differences. The objective of this study was to document and assess PA, both pharmacologic and nonpharmacologic, at home or 2.3. Data analysis en route to the hospital for limb and clavicle injuries in patients admitted to the pediatric ED. We hypothesized that A sample size of 96 patients per group would allow 80% children with fractures receive analgesia more often and power to detect a 25% difference [20] in the rate of sooner than patients with soft-tissue injuries (STI). analgesia administration between the groups (a = .05). The data were collected on the Microsoft Excel program (Microsoft Corp, Redmond, Wash) and analyzed using SPSS release 13.0 (SPSS Inc, Chicago, Ill). Results are 2. Methods expressed as mean (95% confidence intervals) for normal 2.1. Design, setting, and population distributions, median (interquartile range) for non-Gaussian distributions and ordinal data, and absolute numbers (%) for We performed a prospective observational study with 2 categorical variables. Student t test and nonparametric parallel groups of injured pediatric patients: with and Mann-Whitney statistical tests were used for comparison without fractures. Children aged 3 years and older seen in of continuous variables, and Pearson v2 test for categorical our ED at a tertiary pediatric hospital in Toronto, Canada, variables. Results were significant at P b .05. with a limb or clavicle-area injury were recruited as part of another study from August 1, 2004, to February 28, 2005, consecutively between 9:00 am and 11:00 pm, 7 days a week. The study was approved by the Hospital for Sick 3. Results Children’s research ethics board. Written informed consent and a patient’s assent form (for children aged z7 years) 3.1. Patient characteristics were obtained from all study participants. We excluded A total of 310 patients were recruited for this study; 192 patients with underlying chronic conditions, those predis- (62%) had fractures (group 1), and 118 (38%) had STI posed to limb pain (sickle cell disease, osteogenesis (group 2). Patient characteristics are presented in Table 1. imperfecta), those with a condition influencing their pain Mean age was 10.2 years (95% confidence interval, 9.8-10.7 perception or ability to express it (developmental delay, years; range, 3.3-18.0 years). Children presenting with autism), those taking analgesic or anti-inflammatory med- fractures were younger (9.8 vs 11.0 years, P = .005), and ications on a regular basis, those with triage score most of them were males (69.3% vs 56.8%, P = .03). b Q resuscitation, and those with multiorgan trauma. The median time between injury and arrival at the ED was 2.2. Measures and outcomes 9.8 hours (1.8-27.9 hours). Although there was a trend toward earlier presentation in patients with fractures, the The data from parents and children were collected by an difference was not statistically significant (5.9 [1.5-26] vs experienced research assistant and included patient charac- 16 [2-32], P = .1). The areas of injury were upper limb (201, teristics, pain severity while waiting to be seen in the ED, 64.9%), lower limb (101, 32.6%), and clavicle (8, 2.6%). information on the time of injury, medications used before Fractures occurred more frequently than STI in upper limbs arrival at the ED, and when medications were used after ( P b .001) and clavicles ( P = .04) and less frequently in the injury. lower limbs ( P b .001). 402 A.L. Rogovik, R.D. Goldman

Table 1 Patient characteristics, pain severity, and elapsed time Characteristic All patients Group 1: fracture Group 2: no fracture P between groups (N = 310), n (%) (n = 192), n (%) (n = 118), n (%) Age (y)a 10.2 (9.8-10.7) 9.8 (9.3-10.3) 11.0 (10.3-11.7) .005 Maleb 200 (64.5) 133 (69.3) 67 (56.8) .03 Area of injuryb b.001 Upper limb 201 (64.9) 143 (74.5) 58 (49.2) b.001 Lower limb 101 (32.6) 41 (21.4) 60 (50.8) b.001 Clavicle 8 (2.6) 8 (4.2) 0 .04 Time between injury and ED arrival, hc (range) 9.75 (1.77-27.91) 5.94 (1.53-26.0) 16.4 (1.98-31.81) .10 Pain score in EDc (range) 4.0 (2.0-6.5) 4.5 (2.0-7.0) 4.0 (2.0-6.0) .55 Pain severityb .84 No pain (V1) 30 (9.7) 19 (9.9) 11 (9.3) .87 Mild (1-3) 94 (30.3) 55 (28.6) 39 (33.1) .41 Moderate (4-6) 111 (35.8) 69 (35.9) 42 (35.6) .95 Severe (z7) 75 (24.2) 49 (25.5) 26 (22.0) .49 PAb 242 (78.1) 150 (78.1) 92 (78.0) .97 Time between injury and analgesia, minc (range) 5.0 (1.0-30.0) 5.0 (1.0-21.3) 15.0 (2.0-60.0) .001 a Mean (95% confidence interval); P value calculated using t test. b n (%); Pearson v2 test. c Median (interquartile range); Mann-Whitney test.

3.2. Pain severity and those with STI received analgesia at a similar rate ( P = .97); however, it was administered significantly sooner for The median pain score in the study cohort was 4.0 (2.0- patients with suspected fractures (5.0 [1.0-21.3] vs 15.0 6.5) out of 10 (Table 1). Pain scores were 4.5 (2.0-7.0) for [2.0-60.0] minutes, P = .001) (Table 1). group 1 and 4.0 (2.0-6.0) for group 2 ( P = .55). Of the One hundred fourteen (36.8%) received prehospital patients, 30 (9.7%) reported no pain, 94 (30.3%) mild pain, pharmacologic analgesia 105 minutes (30-213 minutes) 111 (35.8%) moderate pain, and 75 (24.2%) severe pain. after the injury (Table 2). Fifty-eight (18.7%) patients were 3.3. Analgesia given acetaminophen, 50 (16.1%) received ibuprofen, and 10 (3.2%) were given opioids (morphine or codeine). Most Prehospital analgesia was administered to 78.1% of the (103, 90%) children received only one dose of an analgesic patients with a limb or clavicle injury. Patients with fractures medication. No significant difference between the groups

Table 2 Characteristics of PA Characteristic All patients Group 1: fracture Group 2: no fracture P between groups (N = 310), n (%) (n = 192), n (%) (n = 118), n (%) Pharmacologic analgesiaa 114 (36.8) 70 (36.5) 44 (37.3) .88 Time to medication, minb (range) 105 (30-213) 120 (22.5-209) 90 (30-300) .75 Medicationsa .48 Acetaminophen 58 (18.7) 34 (17.7) 24 (20.3) .56 Ibuprofen 50 (16.1) 32 (16.7) 18 (15.3) .74 Opioids 10 (3.2) 9 (4.7) 1 (0.8) .06 Other 6 (1.9) 2 (1.0) 4 (3.4) .15 Nonpharmacologic analgesiaa 226 (72.9) 145 (75.5) 81 (68.6) .19 Time to first aid, minb (range) 5.0 (1.0-30.0) 5.0 (1.0-15.0) 10.0 (1.0-60.0) .001 First aid measuresa .21 Cold 184 (59.4) 115 (59.9) 69 (58.5) .81 Immobilization 88 (28.4) 65 (33.9) 23 (19.5) .006 Other 27 (8.7) 15 (7.8) 12 (10.2) .48 First aid by parents/familya 101 (45.7) 58 (40.8) 43 (54.4) .05 First aid by school or camp staff or coacha 68 (30.8) 49 (34.5) 19 (24.1) .11 First aid by paramedicsa 12 (5.4) 11 (7.7) 1 (1.3) .04 a n (%); Pearson v2 test. b Median (interquartile range); Mann-Whitney test. Analgesia for pediatric extremity injuries 403

Table 3 Age differences between children 3 to 6 years and older than 6 years Characteristic Children b6 y old (n = 48), n (%) Children z6 y old (n = 262), n (%) P between groups Age, ya (range) 4.7 (4.4-4.9) 11.3 (10.9-11.6) b.001 PAb 30 (62.5) 212 (80.9) .005 Time to analgesia, minc (range) 10 (1-45) 5 (1-30) .98 Pharmacologic analgesiab 22 (45.8) 92 (35.1) .16 Medicationsb .41 Acetaminophen 13 (27.1) 45 (17.2) .11 Ibuprofen 9 (18.8) 41 (15.6) .59 Other 3 (6.3) 13 (5.0) .71 Time to medication, minc (range) 45 (8.8-150) 120 (31-270) .049 First aidb 28 (58.3) 198 (75.6) .01 First aid measuresb .07 Cold 20 (41.7) 164 (62.6) .007 Immobilization 14 (29.2) 74 (28.2) .90 Other 3 (6.3) 24 (9.2) .51 First aid by parents/familyb 19 (70.4) 82 (42.3) .006 Time to first aid, minc (range) 8.0 (1.0-37.5) 5.0 (1.0-30.0) .66 Pain score in the EDc (range) 4.0 (2.0-6.0) 4.75 (2.0-6.5) .61 a Mean (95% confidence interval); P value calculated using t test. b n (%); Pearson v2 test. c Median (interquartile range), Mann-Whitney test. was recorded in medications given and time to medication children (45 [8.8-150] vs 120 [31-270] minutes, P = .049). (Table 2). Most patients (65%) with moderate or severe pain Pain score in the ED was not different between the groups did not receive prehospital pain medication. (4.0 [2.0-6.0) vs 4.75 [2.0-6.5], P = .61) (Table 3). Two hundred twenty-six (72.9%) patients received first aid, and most (45.7% of the cohort) was administered by parents or family members (Table 2). Patients with fractures 4. Discussion received first aid significantly sooner (5 [1-15] vs 10 [1-60] minutes, P = .001), more by paramedics (11 [7.7%] vs 1 Limb and clavicle injuries are among the most painful [1.3%], P = .04) and less by parents (58 [40.8%] vs 43 pediatric emergencies [21], and late or inadequate pain [54.4%], P = .05). First aid was administered by school or management may lead to unnecessary suffering. camp staff or a coach in 31% of patients, without significant This is the first study to evaluate PA at home or en route differences between the groups. to the hospital for extremity injuries in pediatric patients Whereas 184 (59.4%) of the 310 children had ice or cold admitted to the ED. We found that most children received water applied to their skin, 88 (28.4%) injuries were analgesia; however, it was usually a nonpharmacologic immobilized with a splint, sling, or tensor-type bandage. treatment. Patients with an injury that proved later to be a Limbs with a fracture were immobilized more often (65, fracture received first aid sooner in the prehospital phase, 33.9%) than limbs that suffered only an STI (23, 19.5%, P = but the rate of analgesia for these children was the same as .006) (Table 2). Other first aid measures (dressing, in children who were later found to have an STI. School- elevation, superficial massage of the soft tissues) were used aged children received first aid more often, with a more in only 27 (8.7%) of the patients. frequent use of ice compared with preschoolers. There were no significant differences in the ED pain score between children receiving and not receiving preho- 4.1. Pharmacologic analgesia spital pain medication (4.0 [2.0-6.0] vs 4.25 [2.0-7.0], P = Studies assessing prehospital use of analgesics in .42) or first aid (4.5 [2.0-7.0] vs 4.0 [2.0-6.0], P = .49). pediatric patients have demonstrated suboptimal use of 3.4. Patients’ age pharmacologic analgesia by paramedics and by parents [6,8,9,12]. First aid was administered significantly more often (198 In our study, 90% of the patients reported some degree of [75.6%] vs 28 [58.3%], P = .01), and ice was used more pain, and 60% had moderate to severe pain. Although 78% frequently (62.6% vs 41.7%, P = .007) in children older than received some type of analgesia, only 37% received pain 6 years (Table 3). We found no significant differences in the medication. The number of children given pharmacoanal- age of children receiving pharmacologic analgesia or the gesia was higher than reported in a study from the UK [12], medications used by the different age groups. However, time where only 26% received pain medication before visiting to pain medication was significantly shorter for younger the A&E department. Nevertheless, we found that 65% of 404 A.L. Rogovik, R.D. Goldman children with moderate to severe pain did not receive extremity fractures delivered to the ED by EMS [13], only pharmacoanalgesia. According to protocols suggested for 17% received ice packs. This difference could be explained pediatric emergency analgesia [18,22], acetaminophen or by the younger age of patients in our study, paramedics’ ibuprofen is recommended for mild pain, codeine or codeine concern about keeping the injured extremity immobilized, with acetaminophen is suggested for moderate pain, and and shorter time between injury and arrival at the ED for morphine is recommended for severe pain. Thus, appropri- patients delivered by EMS. In the same study, 16% had ate analgesia for moderate and severe pain cannot be bandages or dressings [13], twice the rate compared with reached with over-the-counter medications available to our cohort, probably because of higher injury severity in parents who do not have access to strong analgesics such EMS-transported patients, almost all of whom had sus- as opioids. pected fractures. The rate of splinting (25%) [13] was Safe and efficacious analgesic medications are needed for similar to the rate in our study (34%). appropriate prehospital pain management. Ease of adminis- Although patients with fractures and STI reported similar tration and rapid onset are their desirable properties; levels of pain in the ED, patients with fractures received first caregivers’ recommendations include improved taste, stron- aid significantly sooner, possibly because the parents had ger pain relief, fewer side effects, and less frequent dosing perceived the potential fracture based on the mechanism of [23]. Over-the-counter availability also plays an important the injury, their previous experience, or a visualized role for a pain medication to be readily accessed by parents deformity of the limb. in emergency situations. The oral nonopioid analgesic More patients with fractures than STI received first aid metamizol (dipyrone) is effectively used for treatment of from paramedics in our study because patients transported mild and moderate pain [24] in many countries, including by EMS usually have more severe injuries, and the level of Germany, France, Switzerland, Netherlands, Spain, South training of paramedics is higher. Interestingly, children with Africa, Latin America, Russia, Israel, and India. However, it fractures had significantly less first aid from parents and has been banned in the United States and the UK because of family compared with those with STI, possibly because its association with potentially life-threatening blood dys- fractures were encountered more frequently outside the crasias such as agranulocytosis. home, for example, during sports activities, and because of A combination of acetaminophen and ibuprofen could parental concern with moving the injured extremity. also promote better pain relief [18]. However, combining analgesics may cause parental confusion, especially 4.3. Age differences with young children, and potential harm that may result In a study by Watkins [8], no child aged less than from dosing errors makes the practice of combining anal- 5 years arriving by EMS received pain medication, gesics unfavorable. compared with 51% between 5 and 15 years of age. Thus, a great need exists for efficacious and safe over- Conversely, we found no significant age differences in the-counter analgesic medications that could be readily pharmacologic analgesia and medications used; moreover, available in an emergency situation for moderate injury time to medication was significantly shorter for younger pain. However, even available mild analgesics, acetamino- children. The difference could stem from increased phen or ibuprofen, were not given to these children by parental pain recognition and perception for 3- to 6-year- parents in our study. Factors associated with omitting old children and awareness to pharmacology measures in pharmacoanalgesia could be a lack of medications at home our community. No difference in pharmacologic analgesia or parental concern that analgesics may harm the child, as between children and adults (21% vs 26%) delivered by suggested previously [12]. EMS was observed in another study [6]. The median time to analgesic drug administration in our Because parents spend more time with younger children, study (105 [30-213] minutes) is much longer than the 22 parents administered first aid more frequently to preschool minutes described in the United States because we included children compared with school-aged children. patients coming independently to the ED, whereas Swor et al [6] included only patients arriving by EMS. Pain score in 4.4. Limitations our ED was similar in patients receiving and not receiving prehospital pain medication, probably because of the higher This study was limited because patients transported to the initial pain in those children to whom the medication ED by EMS received analgesia at a higher rate than others. was given. However, the number of patients delivered by paramedics 4.2. Nonpharmacologic analgesia was small and therefore could not significantly confound the results of the study. First aid was much more common than administration of We did not recruit patients between 11:00 pm and 9:00 medications because of the accessibility of ice or cold water am because of a low number of children with injuries (59.4%) and abridged safety concerns. We found no arriving at the ED at night and because this was unlikely to previous reports on pediatric prehospital nonpharmacologic cause a selection bias. In addition, we did not measure pain analgesia in the literature. Among adults with suspected scores before arrival at the hospital. Besides, we presented Analgesia for pediatric extremity injuries 405 data from the Faces pain scale (3-6 years) and visual [5] Wille C, Bocquet N, Cojocaru B, Leis A, Cheron G. Oral morphine analogue scale (z7 years) together, which have not been administration for children’s traumatic pain. Arch Pediatr 2005; previously validated; however, there was no difference in 12:248-53. [6] Swor R, McEachin CM, Seguin D, Grall KH. Prehospital pain the median pain score between preschool and school-aged management in children suffering traumatic injury. Prehosp Emerg children (Table 3), which justifies the presentation of a Care 2005;9:40-3. combined median score. [7] Hennes H, Kim MK, Pirrallo RG. Prehospital pain management: a The number of patients in the study was relatively small, comparison of providers’ perceptions and practices. Prehosp Emerg and larger studies of PA may be needed to determine the Care 2005;9:32-9. [8] Watkins N. Paediatric prehospital analgesia in Auckland. Emerg Med most efficacious ways to enhance it. In addition, patients Australas 2006;18:51-6. with and without fractures differed in terms of age, sex, and [9] Alexander J, Manno M. Underuse of analgesia in very young area of injury, which could bias comparisons between these pediatric patients with isolated painful injuries. Ann Emerg Med groups. However, we found no age differences in pharma- 2003;41:617-22. cologic analgesia and medications used. Our study was [10] Zempsky WT, Cravero JP. American Academy of Pediatrics Com- mittee on Pediatric Emergency Medicine and Section on Anesthesi- conducted in a single tertiary pediatric ED, and the results ology and Pain Medicine. Relief of pain and anxiety in pediatric may not be extrapolated to other institutions. patients in emergency medical systems. Pediatrics 2004;114:1348-56. [11] Alonso-Serra HM, Wesley K. National Association of EMS Physi- 4.5. Conclusion cians Standards and Clinical Practices Committee. Prehospital pain management. Prehosp Emerg Care 2003;7:482-8. Pain is often undertreated in pediatric limb and clavicle [12] Spedding RL, Harley D, Dunn FJ, McKinney LA. Who gives pain injuries. Children with fractures and STI receive PA at a relief to children? J Accid Emerg Med 1999;16:261-4. [13] White LJ, Cooper JD, Chambers RM, Gradisek RE. Prehospital use of similar rate; however, time to first aid is shorter in those analgesia for suspected extremity fractures. Prehosp Emerg Care with fractures. Most patients (65%) with moderate or severe 2000;4:205-8. pain do not receive prehospital pain medication. Educational [14] McEachin CC, McDermott JT, Swor R. Few emergency medical guidance should be provided to parents to increase their services patients with lower-extremity fractures receive prehospital awareness of appropriate analgesia for children. Efficacious analgesia. Prehosp Emerg Care 2002;6:406-10. [15] Kelly AM, Powell CV, Williams A. Parent visual analogue scale and safe over-the-counter analgesics are needed for moder- ratings of children’s pain do not reliably reflect pain reported by child. ate pain relief. Pediatr Emerg Care 2002;18:159-62. [16] Wong DL, Hockenberry-Eaton M, Wilson D, et al, editors. Wong’s Essentials of Pediatric Nursing. 6th ed. St. Louis (Mo)7 Mosby; 2001. p. 1301. Acknowledgments [17] Abu-Saad H. Assessing children’s responses to pain. Pain 1984;19: 163-71. We thank all ED staff of the Hospital for Sick Children [18] Kowalczyk A, editor. The 2005-2006 Drug Handbook and Formulary. 7 for their cooperation during the study. 24th ed. Toronto (ON) The Hospital for Sick Children; 2005. p. 276, 282. [19] O’Donnell JJ, Maurice SC, Beattie TF. Emergency analgesia in the paediatric population. Part III Non-pharmacological measures of pain relief and anxiolysis. Emerg Med J 2002;19:195-7. References [20] Nelson BP, Cohen D, Lander O, Crawford N, Viccellio AW, Singer AJ. Mandated pain scales improve frequency of ED analgesic [1] Cordell WH, Keene KK, Giles BK, Jones JB, Jones JH, Brizendine administration. Am J Emerg Med 2004;22:582-5. EJ. The high prevalence of pain in emergency medical care. Am J [21] Kennedy RM, Luhmann JD, Luhmann SJ. Emergency department Emerg Med 2002;20:165-9. management of pain and anxiety related to orthopedic fracture care: a [2] Friedland LR, Pancioli AM, Duncan KM. Pediatric emergency guide to analgesic techniques and procedural sedation in children. department analgesic practice. Pediatr Emerg Care 1997;13:103-6. Paediatr Drugs 2004;6:11-31. [3] O’Donnell J, Ferguson LP, Beattie TF. Use of analgesia in a paediatric [22] Maurice SC, O’Donnell JJ, Beattie TF. Emergency analgesia in the accident and emergency department following limb trauma. Eur J paediatric population. Part II Pharmacological methods of pain relief. Emerg Med 2002;9:5-8. Emerg Med J 2002;19:101-5. [4] O’Rourke KP, Mun S, Browne M, Sheehan J, Cusack S, Molloy M. [23] Drendel AL, Lyon R, Bergholte J, Kim MK. Outpatient pediatric pain A retrospective study of the demographics of sport and exercise management practices for fractures. Pediatr Emerg Care 2006;22:94-9. injuries in 1143 children presenting to an Irish emergency department [24] Flemming A, Adams HA. Analgesia, sedation and anaesthesia in over a 6-month period. Eur J Pediatr 2005;164:421-6. emergency service. Anaesthesiol Reanim 2004;29:40-8. American Journal of Emergency Medicine (2007) 25, 406–413

www.elsevier.com/locate/ajem

Original Contribution A risk score to predict silent myocardial ischemia in patients with coronary artery disease under aspirin therapy presenting with upper gastrointestinal hemorrhageB

Chien-Chih Chen MD, MSa,b, Chee-Fah Chong MD, MSa,b, Cheng-Deng Kuo MD, PhDc,d, Tzong-Luen Wang MD, PhDa,b,e,* aDepartment of Emergency Medicine, Shin-Kong Wu Ho-Su Memorial Hospital, Taipei 111, Taiwan bSchool of Medicine, Fu Jen Catholic University, Taipei 510, Taiwan cInstitute of Emergency and Critical Care Medicine, School of Medicine, National Yang-Ming University, Taipei 112, Taiwan dDepartment of Medical Research and Education, Veterans General Hospital, Taipei 112, Taiwan eDepartment of Medicine, Taipei Medical University, Taipei 110, Taiwan

Received 7 August 2006; revised 11 September 2006; accepted 25 September 2006

Abstract Background: Silent myocardial ischemia (SMI) is a relatively common complication in patients with coronary artery disease (CAD) under aspirin therapy presenting with upper gastrointestinal hemorrhage (UGIH). Aim: This study was conducted to develop and prospectively validate a risk prediction score to identify SMI in patients undergoing aspirin therapy for CAD presenting with UGIH in the emergency department (ED). Methods: This was a 2-phase noninterventional study. In the derivation phase, adults with CAD under aspirin therapy (100 mg once daily) presenting to the ED with UGIH were retrospectively recruited. By multiple logistic regression analysis, we derived a risk score from 224 patients that predicts the patients’ risk of SMI. In the validation phase, we prospectively validated this score using receiver operating characteristic curves with data from 110 patients. We also developed a fast-track screening procedure from this score. Results: There were 56 patients (25.0%) and 29 patients (26.4%) with SMI in the derivation and validation sets, respectively. Independent multivariate predictors of SMI were age of older than 75 years, severity of CAD, systolic blood pressure of less than 110 mm Hg, diastolic blood pressure of less than 85 mm Hg, hematocrit of less than 30%, and blood urea nitrogen–creatinine ratio of more than 30. The

B The authors contributed in the study as follows: Tzong-Luen Wang and Chien-Chih Chen conceived the study. Tzong-Luen Wang and Chien-Chih Chen supervised the conduct of the study and data collection. Chee-Fah Chong and Cheng-Deng Kuo provided statistical advice on study design and analyzed the data. Chien-Chih Chen drafted the manuscript. Tzong-Luen Wang takes responsibility for the article as a whole. * Corresponding author. Emergency Department, Shin-Kong Wu Ho-Su Memorial Hospital, No.95 Wen Chang Road, Shih Lin District, Taipei City 111, Taiwan, ROC. Tel.: +886 2 28389425; fax: +886 2 28353547. E-mail address: [email protected] (T.-L. Wang).

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2006.09.009 Risk score to myocardial ischemia GI hemorrhage 407 area under receiver operating characteristic curve for the rule was 0.93 in the derivation set and 0.96 in the validation set. At the cutoff value of 5 points or higher, the sensitivity and specificity of the fast-track screening procedure for SMI were 96.6% and 86.4%, respectively. The positive and negative predictive values were 71.8% and 98.6%, respectively. Conclusions: This simple risk prediction score is easily calculated and is based on rapidly available clinical and laboratory data in the ED. It can be used to stratify patients undergoing aspirin therapy for CAD presenting with UGIH by risk of SMI. D 2007 Elsevier Inc. All rights reserved.

1. Introduction January 1, 2002, and October 30, 2004, with UGIH were retrospectively enrolled into the study. In the validation Acute upper gastrointestinal hemorrhage (UGIH) is a phase, all adult patients with CAD under aspirin therapy common reason for emergency hospital admission, most (100 mg once daily for 1 month) with UGIH who presented after evaluation in the emergency department (ED). Upper to ED between November 1, 2004, and March 31, 2006, gastrointestinal hemorrhage is a common problem in were prospectively enrolled into the study. Coronary artery patients with coronary artery disease (CAD) presenting to disease, defined as 1 or more vessels with greater than 50% the ED. Myocardial ischemia or infarction is relatively diameter stenosis, was proven by previous angiography. All common in patients with UGIH. Two studies have shown patients were known to have CAD proven by coronary that myocardial ischemia or infarction occurs in 10% to angiography before the presentation with UGIH. All 12-lead 25% of patients admitted to the intensive care unit with electrocardiograms (ECGs) were compared with the latest significant UGIH, with a mortality rate of 15% to 20% [1,2]. 12-lead ECGs of the previous year. The following were the Millions of individuals use aspirin for the secondary criteria for exclusion from the study: chronic renal failure; prevention of cardiovascular events. Because of inhibition liver cirrhosis; absence of ECG in the ED; refusal to of prostaglandin synthesis, long-term aspirin therapy may be undergo upper endoscopy; chest pain or tightness at or after associated with peptic ulcer and its complications. Upper arrival to the ED; and use of anticoagulants, corticosteroids, gastrointestinal hemorrhage may be accompanied by con- nonsteroidal anti-inflammatory drugs, and proton pump siderable cardiopulmonary stress, so myocardial ischemia or inhibitors in the previous month. The study protocol was infarction might be expected in patients with CAD present- approved by our institutional review board. ing with UGIH [3]. Low-dose aspirin use (V325 mg/d) is associated with a 2.5-fold increase in risk of UGIH, but it 2.2. Data collection and definitions of covariates reduced the number of myocardial infarctions by 30% [4]. Aspirin use for the secondary prevention of cardiovascular In this study, SMI was defined as ischemia on the 12-lead events has a favorable benefit-to-risk profile and should be ECG without typical chest pain indicating CAD. The criteria encouraged in those at high risk. In our previous study [5], for myocardial ischemia on the 12-lead ECG were horizon- we retrospectively examined the frequency and identified tal or downsloping ST-segment depressions of more than independent predictors of silent myocardial ischemia (SMI) 1.0 mm at 40 milliseconds after the J point or T-wave in a cohort population with CAD under aspirin therapy inversion in 2 or more contiguous electrocardiographic leads presenting with UGIH in the ED. The incidence of in the absence of left ventricular hypertrophy, left bundle- simultaneous SMI and UGIH in patients with CAD under branch block, or digitalis effect. All patients who satisfied aspirin therapy is 20.4%. In this study, we developed and the criteria for myocardial ischemia on the initial 12-lead prospectively validated a risk prediction score to identify ECG were continuously monitored using 3-lead ECG and SMI in patients undergoing aspirin therapy for CAD serial 12-lead ECG recordings for 12 hours. The serial presenting with UGIH in the ED. 12-lead ECG recordings were obtained at the time of presentation, 2 to 4 hours, 6 to 8 hours, and 10 to 12 hours after presentation. The lead positions were marked on the anterior chest wall for these serial recordings. Blood 2. Materials and methods samples for cardiac enzymes were drawn from all patients 2.1. Study population and study design who satisfied the criteria for myocardial ischemia on the initial 12-lead ECG at the time of presentation, 2 to 4 hours, This was a 2-phase noninterventional study conducted in 6 to 8 hours, and 10 to 12 hours after presentation. The a 921-bed urban medical center with an adult intensive myocardial ischemia group included myocardial infarction care unit bed capacity of 60 and approximately 75000 ED and non–myocardial infarction. Myocardial infarction was visits annually. In the derivation phase, all adult patients defined as the satisfaction of enzymatic criteria for (N18 years) with CAD under aspirin therapy (100 mg once myocardial infarction or ST-segment elevation on the initial daily for 1 month) who presented to the ED between 12-lead ECG or subsequent serial 12-lead ECG. Enzymatic 408 C.-C. Chen et al.

troponin I level; results of upper endoscopy; length of hospital stay; and inhospital mortality. All continuous variables were categorized by using commonly defined ranges [6-9]. Tachycardia was defined as an HR of more than 100 beats/min, tachypnea as an RR of more than 20 breaths/min, anemia as an Hct of less than 30%, high BUN/Cr ratio as that higher than 30, and old age as that older than 75 years. The SBP and DBP cutoff values are 110 and 85 mm Hg, respectively. 2.3. Data analysis and statistics The 12-lead ECG in the ED was presented to a senior cardiologist who interpreted it in a blinded fashion. We stratified patients into 1 of 2 groups with the serial 12-lead ECG and cardiac enzyme. The ischemia group, classified as patients who had myocardial ischemia, included myocardial infarction and non–myocardial infarction. The nonischemia group was classified as patients who had non–myocardial ischemia. All analyses were performed on SPSS 10.0 for Windows (SPSS Inc, Chicago, Ill). Continuous data were presented as mean F SD. Univariate relationships between the categorical covariates and SMI were evaluated by using

Fig. 1 Patient enrollment and group allocation. MI indicates Table 1 Patient characteristics in the derivation and myocardial infarction. validation sets Characteristic Derivation Validation evidence of myocardial infarction required 1 of the (n = 224) (n = 110) following factors: (1) cardiac troponin I level of more than Male (%) 71.4 69.1 0.5 lg/L and (2) increased total creatine phosphokinase Age (y) 70.6 F 10.2 69.1 F 11.2 (CK, z150 U/L) and myocardial band (MB) isoenzyme Diabetes (%) 43.3 44.5 fraction (at least 5% of total CK). It was considered as non– Hypertension (%) 50.9 51.8 myocardial infarction if patients did not satisfy enzymatic Severity of CAD (%) criteria for myocardial infarction but satisfied the criteria for SVD 44.6 43.6 myocardial ischemia on the initial 12-lead ECG and DVD 28.6 20.9 resolution of ST-segment deviation occurred on the subse- TVD 26.8 25.5 quently serial 12-lead ECG within 12 hours. It was Previous PCI (%) 95.1 94.6 S/P CABG (%) 12.1 12.9 considered as non–myocardial ischemia if they did not Finding of upper endoscopy (%)a satisfy the criteria for myocardial ischemia or infarction on GU 73.7 76.4 the initial 12-lead ECG, or if they satisfied the criteria for DU 52.2 51.2 myocardial ischemia on the initial 12-lead ECG but HR (beats/min) 93.1 F 14.6 92.8 F 15.4 resolution of ST-segment deviation did not occur on the RR (breaths/min) 19.3 F 2.2 18.2 F 2.4 subsequent serial 12-ECG within 12 hours. Detailed SBP (mm Hg) 125.5 F 18.3 124.0 F 23.3 information of patients in the derivation and validation sets DBP (mm Hg) 72.3 F 15.8 71.8 F 13.7 was obtained after the patients were discharge through Hct (%) 30.9 F 7.1 31.4 F 9.6 F F medical record abstraction by 1 of 2 physicians with exten- BUN (mg/dL) 34.7 18.6 34.0 17.9 F F sive experience in chart review procedures. The following Cr (mg/dL) 1.3 0.5 1.3 0.6 BUN/Cr ratio 27.3 F 13.0 28.1 F 13.5 data were recorded for each patient: age; sex; comorbidities; Length of stay (d) 6.3 F 4.0 6.3 F 5.2 how many vessels were narrowed greater than 50% in Myocardial ischemia (%) 25.0 26.4 patients with CAD; whether primary coronary intervention Myocardial infarction (%) 4.0 4.5 and coronary artery bypass graft would be performed; Mortality (%) 4.9 4.5 findings of ECG; hematocrit (Hct) in the ED; lowest systolic SVD indicates single-vessel disease; DVD, double-vessel disease; and diastolic blood pressure (SBP and DBP) in the ED; TVD, triple-vessel disease; PCI, primary coronary intervention; CABG, highest heart rate (HR) in the ED; highest respiratory rate coronary artery bypass graft; GU, gastric ulcer; DU, duodenal ulcer; S/P (RR) in the ED; blood urea nitrogen (BUN) level; creatinine CABG, status of post coronary artery bypass graft status. a Some patients had both gastric and duodenal ulcers. (Cr) level; ratio of BUN/Cr; CK and MB isoenzyme levels; Risk score to myocardial ischemia GI hemorrhage 409

Table 2 Univariate predictors of myocardial ischemia Variable Yes, n No, n P (% ischemia) (% ischemia) Age N75 y 90 (32.2) 134 (20.1) .04 Male 160 (21.3) 64 (34.4) .04 Diabetes 97 (25.8) 127 (24.4) .82 Hypertension 114 (28.1) 110 (21.8) .28 Severity of CAD (%) b.001 SVD 100 (6.0) 124 (40.3) DVD 64 (28.1) 160 (23.8) TVD 60 (53.3) 164 (14.6) Previous PCI (%) 212(22.2) 12 (16.7) 1.00 S/P CABG (%) 14 (35.7) 210 (21.0) .34 GU 165 (23.6) 59 (28.8) .44 DU 117 (24.8) 107 (25.2) .94 HR N100 beats/min 65 (41.5) 159 (18.2) b.001 Fig. 2 Silent myocardial ischemia by risk score in derivation set. N RR 20 breaths/min 37 (32.4) 187 (23.5) .26 ROC curve of 1.0 indicating perfect discriminations. We b b SBP 110 mm Hg 36 (61.1) 188 (18.1) .001 assessed the fit of the model with the Hosmer-Lemeshow DBP b85 mm Hg 183 (30.1) 41 (2.4) b.001 goodness-of-fit test. Finally, we derived a fast-track screen- Hct b30% 117 (43.6) 107 (4.7) b.001 BUN/Cr ratio N30 78 (47.4) 146 (13.0) b.001 ing procedure from the full risk score that placed a patient who scores 5 or higher in a high-risk category. We also calculated the sensitivity, specificity, positive predictive 2 v test and Fisher exact test as appropriate. Univariate value, and negative predictive value of this screening tool. predictors of SMI that were significant at a level of P b .05 were identified. Factors that were significant at a level of P b .1 were then eligible for inclusion in a forward selection 3. Results multiple logistic regression model, which identified factors in the ED that were independent predictors of SMI at a level In the derivation phase, a total of 303 patients with CAD of P b .05 (2-tailed). under aspirin therapy presented to the ED with UGIH. The results of multivariate analysis were then used to Seventy-nine patients were excluded from the study because develop a clinical prediction rule. Each b coefficient was of chronic renal failure (n = 10); liver cirrhosis (n = 6); divided by 2 and rounded off to the nearest integer. Points refusal to undergo upper endoscopy (n = 8); simultaneous were assigned to each predictor, and a score of SMI in CAD chest pain in the ED (n = 22); absence of ECG in the ED with UGIH was calculated for each patient. (n = 16); and use of anticoagulants (n = 6), corticosteroids The model and prediction rule were then validated on the (n = 3), nonsteroidal anti-inflammatory drugs (n = 5), and population of patients in the validation set. The discrimina- proton pump inhibitors (n = 3) in the previous month. The tory capacity of the risk score was assessed by using the area derivation set consisted of 224 patients with 11 deaths (4.9% under the receiver operating characteristic (ROC) curves as mortality rate). There were 56 patients (25%) with SMI and an index of model performance in both derivation and 9 patients (4.0%) with myocardial infarction in the validation sets. The ROC curve reflects the concordance of predictions with actual outcomes in rank order, with an ROC curve of 0.5 indicating no better than change and an

Table 3 Independent predictors identified by multivariate analysis Variable b OR (95% CI) P Points Intercept À6.35 Age N75 y 1.33 3.78 (1.21-11.82) .022 1 SBP b110 mm Hg 2.75 15.62 (3.08-79.27) .001 1 DBP b85 mm Hg 3.34 28.31 (2.91-275.84) .004 2 Hct b30% 2.57 13.11 (3.51-48.91) b.001 1 BUN/Cr N30 1.87 6.47 (2.08-20.10) .001 1 Severity of CAD DVD 2.01 7.43 (1.99-27.76) .003 1 TVD 4.33 75.78 (14.01-409.84) b.001 2 Fig. 3 Receiver operating characteristic curves for derivation OR indicates odds ratio; CI, confidence interval. and validation sets. 410 C.-C. Chen et al.

4. Discussion

Upper gastrointestinal hemorrhage is a common medical emergency with an incidence of about 103 per 100000 adults per year in the United Kingdom, and reported mortality rates (5%-14%) attributable to UGIH have remained unchanged for many decades [10]. Few studies have investigated the risk factors for myocardial ischemia complicating gastrointestinal hemorrhage. In the study by Cappell [11], patients who had concomitant myocardial infarction and UGIH had lower mean arterial pressures and Hct than did patients who had UGIH without myocardial infarction or with previous myocardial infarction. This study also suggested that myocardial infarction, before or after UGIH, was associated with a higher total mortality rate than Fig. 4 Silent myocardial ischemia by risk score in validation set. the sum of the individual mortality rates in patients with UGIH alone and myocardial infarction alone. Most patients derivation set. In the validation phase, the validation set with UGIH were initially presented and evaluated in the ED. consisted of 110 patients with 5 deaths (4.5% mortality At the time of initial presentation, most patients with rate). There were 29 patients (26.4%) with SMI and 5 patients myocardial infarction tended to nonspecific symptoms of (4.5%) with myocardial infarction in the validation set. A dizziness or confusion, not chest pain or tightness. In the flow diagram illustrating patient enrollment and group study by Bhatti et al [1], patients with myocardial infarction allocation is shown in Fig. 1. The samples of patients in complicating admission for gastrointestinal hemorrhage the derivation and validation sets were similar (Table 1). may not present with typical symptoms of myocardial Each of the clinical characteristics in the derivation set was ischemia or infarction. Patients with myocardial infarction evaluated by univariate analysis (Table 2). had a higher acuity of illness, a larger number of risk factors A multiple logistic regression model was constructed for CAD, and lower Hct than those without myocardial using the statistically significant univariate variables ( P b.1) infarction. In the study by Emmanuel et al [2], patients with and was used to assign point values for a clinical decision myocardial infarction are likely to be older, have more risk rule (Table 3). The 6 independent correlates of SMI were age factors for CAD, and have a more acutely ill condition than of older than 75 years, severity of CAD, SBP of less than 110 those without myocardial infarction. In a study on 51 pa- mm Hg, DBP of less than 85 mm Hg, Hct of less than 30%, tients with UGIH related to aspirin therapy, 9 patients had a and BUN/Cr ratio of more than 30. We developed a clinical simultaneous myocardial infarction [12]. Risk factors for a risk prediction score using the b coefficients from the model. myocardial infarction included a prior history of ischemic Patients’ risk scores were calculated by adding the score heart disease and severe bleeding, as evidenced by severe components associated with each clinical risk predictors at anemia. In our previous study [5], the incidence of the initial presentation. The range of risk score is from 0 to simultaneous SMI and UGIH in patients with CAD under 8 points. Fig. 2 shows the degree of association between the aspirin therapy is 20.4%, and severity of CAD, higher BUN risk predictors and the observed SMI in each category of the level, lower DBP, and lower Hct were the best independent risk score. The rates of SMI increase as the risk score predictors for SMI. increases. The area under the ROC curve for the risk score Rockall et al [13] found that advanced age, shock, other was 0.93 (Fig. 3). The Hosmer-Lemeshow goodness-of-fit comorbidities, diagnosis, major stigmata of recent hemor- statistic was 0.35 for the model. rhage, and rebleeding are independent predictors of mortal- The rule maintained its predictive ability when applied to ity in patients with UGIH. They developed a simple the validation set. It also showed that the rates of SMI numerical score that can be used to categorize patients increase as the risk score increases (Fig. 4). The area under presenting with UGIH by risk of death. Blatchford et al [6,7] the ROC curve for the risk prediction score in the validation devised and prospectively evaluated a risk stratification set was 0.96 (Fig. 3). The Hosmer-Lemeshow goodness-of- score that included patients’ admission hemoglobin level, fit statistic was 0.68 for the model in the validation cohort. BUN level, pulse rate, SBP, presentation with syncope or From the full risk score, we derived a rapid risk- melena, and evidence of hepatic disease or cardiac failure to screening procedure that classified patients as being at high identify the patient’s need for treatment. In this study, we risk of SMI if, at presentation, the risk score is 5 or higher. found that SMI (approximately 25%) occurs commonly in The sensitivity and specificity of the rapid risk-screening patients with CAD presenting to the ED with UGIH. Our procedure for SMI were 96.6% and 86.4%, respectively. score discriminated well between patients who had SMI and The positive and negative predictive values were 71.8% and those who did not. Age older than 75 years, severity of 98.6%, respectively. CAD, SBP of less than 110 mm Hg, DBP of less than Risk score to myocardial ischemia GI hemorrhage 411

85 mm Hg, Hct of lower than 30%, and BUN/Cr ratio of tool to identify patients affected with perioperative myocar- higher than 30 were independent predictors of SMI in dial ischemia and subsequent myocardial cell damage. In patients with CAD undergoing aspirin therapy presenting to our study, myocardial ischemia, suspected when initial the ED with UGIH. 12-lead ECG was compared with the last 12-lead ECG, was Besides the definite linkage of severity of CAD, higher confirmed with subsequent serial 12-lead ECG according to BUN/Cr ratio, which has been shown to be linked to UGIH whether resolution of ST-segment deviation had occurred instead of lower gastrointestinal bleeding, may also indicate after supportive treatments. In the derivation set, 20 patients the severity of UGIH [8,14]. In addition to reflecting the (42.5%), 21 patients (44.7%), and 6 patients (12.8%) who severity of UGIH and kidney function, BUN levels may rise did not have myocardial infarction and were classified in the independent of changes in glomerular filtration rate or ischemia group had resolution of ST-segment deviation serum creatinine owing to enhanced proximal tubular within 4 hours, between 4 and 8 hours, and between 8 and reabsorption under the activation of the sympathetic nervous 12 hours after presentation, respectively. In the validation and renin-angiotensin-aldosterone systems [15]. Blood urea set, 9 patients (37.5%), 11 patients (45.8%), and 4 patients nitrogen has been associated with adverse outcomes in (16.7%) who did not have myocardial infarction and were patients with acute coronary syndromes with or without classified in the ischemia group had resolution of myocardial infarction, and it has been previously incorpo- ST-segment deviation within 4 hours, between 4 and 8 hours, rated into myocardial infarction risk prediction model and between 8 and 12 hours after presentation, respectively. [16-18]. Patients with gastrointestinal hemorrhage compli- Seven patients and 4 patients who satisfied the criteria for cated by myocardial infarction are likely to be older than myocardial ischemia on the initial 12-lead ECG but whose those without myocardial ischemia or infarction [2]. resolution of ST-segment deviation did not occur on the Lower lowest SBP and lower Hct may therein be the subsequent serial 12-lead ECG within 12 hours were most possible pathophysiologic mechanisms for the delete- classified in the nonischemia group in the derivation and rious effects of UGIH on those who are taking aspirin for validation sets, respectively. These patients still did not have CAD. The massive bleeding produces hypovolemia and resolution of ST-segment deviation on the 12-lead ECG at myocardial hypoperfusion, which contributes to the devel- the time of discharge. opment of SMI. In the study by Emmanuel et al [2], patients Without a high index of suspicion, myocardial ischemia with gastrointestinal hemorrhage complicated by myo- or infarction simultaneous with UGIH will be neglected in cardial infarction had lower lowest SBP than those most patients because of the lack of typical symptoms of without myocardial ischemia or infarction. On the other myocardial ischemia. Because myocardial ischemia or hand, lower Hct always indicates the severity of anemia, infarction could contribute to increased morbidity and which, in turn, has been proven to be associated with the mortality in patients with UGIH, early detection is grave outcome for those with CAD, whereas maintenance of necessary. The fast-track risk screen procedure is simple adequate Hct may provide protective benefits [19-21]. It has enough to be used by clinicians to predict SMI. It has also been shown that lower lowest DBP is strongly sensitivity adequate to provide a substantial margin of associated with myocardial ischemia because of reduced clinical safety. We advise routine 12-lead ECG screening of coronary perfusion [9,22]. Patients with combined CAD and all patients with CAD presenting to the ED with UGIH and left ventricular hypertrophy experience a 25% fall in completion of the risk score as soon as possible. When the coronary flow, accompanied by enhanced oxygen extraction risk score is 5 points or higher, early aggressive treatments and an increase in CAD death rate when DBP falls below and close monitoring are suggested, including monitoring 85 mm Hg [23,24]. levels of troponin I and CK-MB. There are several diagnostic tools such as serial or Upper endoscopy plays an important role in the diagnosis continuous 12-lead ECG monitors [25,26], radionuclide and the therapy for UGIH and may also provide benefits for imaging techniques [27], or stress echocardiography [28] to those with both CAD and UGIH. Diagnosis of the specific provide evidence of SMI in patients classified in the cause of UGIH may be particularly important in a patient ischemia group. Although the radionuclide imaging techni- with simultaneous myocardial ischemia or infarction to ques and echocardiography are more sensitive and specific apply appropriate specific therapy and to prevent further diagnostic procedures [29], they are both more labor blood loss and myocardial hypoperfusion. Several recent intensive and costly than serial 12-lead ECG. Moreover, studies have suggested that upper endoscopy can be patients with myocardial ischemia due to UGIH were performed safely in medically stable patients with UGIH potentially unstable; we could not let them leave the ED and simultaneous myocardial infarction, but it had a high or intensive care unit to perform the radionuclide imaging complication rate in highly unstable patients [30,31]. In this study, and the echocardiography could not be available for study, no complications occurred in these patients undergo- 24 hours at our ED. Serial 12-lead ECG monitoring ing the initial upper endoscopy. However, we cannot provides virtually continuous information about ST seg- eliminate the possibility that more myocytes could have ments. In the study by Bottiger et al [25], the intermittent died from the stress caused by the upper endoscopy in 12-lead ECG recordings can be used as an early warning this study. 412 C.-C. Chen et al.

This study has some limitations. Firstly, the quality of [8] Ernst AA, Haynes ML, Nick TG, et al. Usefulness of the blood urea data obtained in a retrospective phase is only as accurate as nitrogen/creatinine ratio in gastrointestinal bleeding. Am J Emerg Med 1999;17:70-2. that which was recorded and stored. Secondly, patients who [9] Cruickshank JM. Coronary flow reserve and the J curve relation did not satisfy the criteria for myocardial ischemia on the between diastolic blood pressure and myocardial infarction. BMJ initial 12-lead ECG were not continuously monitored using 1988;197:1227-30. 3-lead ECG, serial 12-lead ECG recordings, and serial [10] Rockall TA, Logan RF, Devlin HB, et al. Incidence of and mortality cardiac enzymes for 12 hours. The patients who did not from acute upper gastrointestinal haemorrhage in the United Kingdom. BMJ 1995;311:222-6. satisfy the criteria for myocardial ischemia on the initial [11] Cappell MS. A study of the syndrome of simultaneous acute upper 12-lead ECG cannot be completely ruled out for the gastrointestinal bleeding and myocardial infarction in 36 patents. possibilities of SIM. Thirdly, the medical treatments at the Am J Gastroenterol 1995;90:1444-9. ED were not recorded and analyzed. This may raise [12] Bar-Dayan Y, Amital H, Levy Y, et al. Low dose aspirin in patients questions on whether the treatment would also affect the with ischemic heart disease may precipitate secondary myocardial infarction. Isr J Med Sci 1996;32:288-91. patients’ outcome. Fourthly, this was a single-center study, [13] Rockall TA, Logan RF, Devlin HB, et al. Risk assessment after acute and the representation of our database may not be upper gastrointestinal haemorrhage. Gut 1996;38:316-21. representative to the general patient cohort in Taiwan. [14] Urashima M, Toyoda S, Nakano T, et al. BUN/Cr ratio as an index of gastrointestinal bleeding mass in children. J Pediatr Gastroenterol Nutr 1992;15:89-92. [15] Usberti M, Federico S, Di Minno G, et al. Effects of angiotensin II on 5. Conclusions plasma ADH, prostaglandin synthesis, and water excretion in normal humans. Am J Physiol 1985;248:F254-9. Our data suggest that SMI is a relatively common [16] Normand ST, Glickman ME, Sharma RG, McNeil BJ. Using complication in patients with CAD under aspirin therapy admission characteristics to predict short-term mortality from myo- presenting with UGIH. This new risk prediction score is cardial infarction in elderly patients. Results from the Cooperative easily calculated and is based on rapidly available clinical Cardiovascular Project. JAMA 1996;275:1322-8. [17] Luria MH, Knoke JD, Margolis RM, et al. Acute myocardial and laboratory data in the ED. It can be used to stratify infarction: prognosis after recovery. Ann Intern Med 1976;85:561-5. patients undergoing aspirin therapy for CAD presenting [18] Kirtane AJ, Leder DM, Waikar SS, et al. Serum blood urea nitrogen as with UGIH by risk of SMI. an independent marker of subsequent mortality among patients with acute coronary syndromes and normal to mildly reduced glomerular filtration rates. J Am Coll Cardiol 2005;45:1781-6. [19] Lee PC, Kini AS, Ahsan C, et al. Anemia is an independent predictor Acknowledgments of mortality after percutaneous coronary intervention. J Am Coll Cardiol 2004;44:541-6. We express our gratitude to Prof Shi-Chuan Chang, [20] Bar-Dayan Y, Levy Y, Amital H, et al. Aspirin for prevention of Assoc Prof David Hung-Tsang Yang, and Mei-Jy Jeng for myocardial infarction. A double-edged sword. Ann Med Interne (Paris) 1997;148:430-3. the critical evaluations and suggestions on the study design. [21] Wu WC, Rathore SS, Wang Y, et al. Blood transfusion in elderly patients with acute myocardial infarction. N Engl J Med 2001;345: 1230-6. References [22] Zaslavskaya RM, Lilitsa GV, Dilmagambetova GS, et al. Melatonin, refractory hypertension, myocardial ischemia and other challenges in [1] Bhatti N, Amoateng-Adjepong Y, Qamar A, et al. Myocardial nightly blood pressure lowering. Biomed Pharmacother 2004; infarction in critically ill patients presenting with gastrointestinal 58(Suppl 1):S129-34. hemorrhage: retrospective analysis of risks and outcomes. Chest 1998; [23] Polese A, De Cesare N, Montorsi P, et al. Upward shift of the lower 114:1137-42. range of coronary flow autoregulation in hypertensive patients with [2] Emmanuel E, Sangeeta S, Yaw AA, et al. Myocardial infarction hypertrophy of the left ventricle. Circulation 1991;83:845-53. complicating gastrointestinal hemorrhage. Mayo Clin Proc 1999;74: [24] Cruickshank JM, Polese A. Left ventricular hypertrophy and the 235-41. possible harmful effect of the excessive lowering of diastolic blood [3] Gostout CJ, Wang KK, Ahlqquist DA, et al. Acute gastrointestinal pressure. In: Cruickshank JM, Messerli FH, editors. Left ventricular bleeding: experience of a specialized management team. J Clin hypertrophy and its regression. London7 Science Press; 1992. p. 61-9 Gastroenterol 1992;14:260-7. [with permission of Kannel WV]. [4] Weisman SM, Graham DY. Evaluation of the benefits and risks of [25] Bottiger BW, Motsch J, Teschendorf P, et al. Postoperative 12-lead low-dose aspirin in the secondary prevention of cardiovascular and ECG predicts peri-operative myocardial ischaemia associated with cerebrovascular events. Arch Intern Med 2002;162:2197-202. myocardial cell damage. Anaesthesia 2004;59:1083-90. [5] Chen CC, Chong CF, Kuo CD, et al. Silent myocardial ischemia in [26] Fesmire FM. A rapid protocol to identify and exclude acute coronary artery disease patients under aspirin therapy presenting with myocardial infarction: continuous 12-lead ECG monitoring with upper gastrointestinal hemorrhage. J Gastroenterol Hepatol 2006 2-hour delta CK-MB. Am J Emerg Med 2000;18:698-702. [Online publication date: 29 Mar]. [27] Caglar M, Mahmoudian B, Aytemir K, et al. Value of 99mTc- [6] Blatchford O, Davidson LA, Murray WR, et al. Acute upper methoxyiso-butylisonitrile (99mTc-MIBI) gated SPECT for the gastrointestinal haemorrhage in west of Scotland: case ascertainment detection of silent myocardial ischemia in hemodialysis patients: study. BMJ 1997;315:510-4. clinical variables associated with abnormal test results. Nucl Med [7] Blatchford O, Murray WR, Blatchford M. A risk score to predict the Commun 2006;27:61-9. need for treatment for upper-gastrointestinal haemorrhage. Lancet [28] Segar DS, Brown SE, Sawada SG, et al. Dobutamine stress 2000;356:1318-21. echocardiography: correlation with coronary lesion severity as Risk score to myocardial ischemia GI hemorrhage 413

determined by quantitative angiography. J Am Coll Cardiol 1992;19: tion: a six-year study of 42 endoscopies in 34 consecutive patients 1197-202. at two university teaching hospitals. Am J Gastroenterol 1993;88: [29] Schinkel AF, Bax JJ, Geleijnse ML, et al. Noninvasive evaluation of 344-50. ischaemic heart disease: myocardial perfusion imaging or stress [31] Lee JG, Krucoff MW, Brazer SR. Periprocedural myocardial ischemia echocardiography? Eur Heart J 2003;24:789-800. in patients with severe symptomatic coronary artery disease under- [30] Cappell MS. The safety and clinical utility of esophagogastroduode- going endoscopy: prevalence and risk factors. Am J Med 1995;99: noscopy for acute gastrointestinal bleeding after myocardial infarc- 270-5. American Journal of Emergency Medicine (2007) 25, 414–419

www.elsevier.com/locate/ajem

Original Contribution A clinical score predicting the need for hospitalization in scorpion envenomation

Semir Nouira MDa,*, Riadh Boukef MDa, Noureddine Nciri MDb, Habib Haguiga MDd, Souheil Elatrous MDc, Lamia Besbes MDb, Mondher Letaief MDe, Fekri Abroug MDb aEmergency Department, Fattouma Bourguiba University Hospital, Monastir 5000, Tunisia bMedical Intensive Care Unit, Fattouma Bourguiba University Hospital, Monastir 5000, Tunisia cMedical Intensive Care Unit, Tahar Sfar University Hospital, Mahdia 5100, Tunisia dEmergency Department and Intensive Care Unit, District Hospital, Tozeur 2200, Tunisia eDepartment of Epidemiology and Statistics, Faculty of Medicine, Monastir 5000, Tunisia

Received 29 July 2006; accepted 21 August 2006

Abstract Objective: Predicting complications is a clinical challenge in the assessment of victims of scorpion envenomation (SE). We sought to develop a clinical score to predict need for hospitalization after scorpion sting. Methods: We prospectively collected data in patients attending the emergency department after SE in derivation (n = 868) and validation groups (n = 435). A score was derived from a multiple regression analyses using clinical variables as dependent variables and hospitalization as independent variable. Results: Discrimination power of the constructed score was good (area under the receiver operating characteristic curve, 0.85 and 0.83 in derivation and validation group, respectively). Goodness-of-fit tests indicated that the score performed well in the derivation and the validation groups ( P = .88 and P = .67 respectively). The score has a good sensitivity and negative predictive value at cutoff value of 2. Conclusion: Our clinical score could be used for efficient hospital admission decision in patient’s victims of SE. D 2007 Elsevier Inc. All rights reserved.

1. Introduction deaths include pulmonary edema and congestive heart failure. Most of these complications and deaths are Of the 40000 scorpion-envenomed patients that are preventable. The main preventable factors include inability reported each year in Tunisia, 90% occur in rural areas, to early recognize the severity and delay in referring victims affecting predominantly young victims. The epidemiologic to specific health care facilities. Hospital-related preventable picture is quite similar in many other developing countries factors in endemic areas include inadequate skills and lack where mortality ratio remains unacceptably high (250-300/ of intensive care units. Indeed, delay in seeking emergent 100000) [1-6]. The leading causes of complications and treatment contributes to most of rural deaths. In endemic Tunisian areas, most patients live more than 50 km from the * Corresponding author. Tel.: +216 73460678; fax: +216 73460678. nearest facility offering intensive care. Approximately more E-mail address: [email protected] (S. Nouira). than 2 hours are needed to reach the district hospital.

Tunisian heath authorities began to recognize the impor- (systolic blood pressure b90 mm Hg or a sustained decrease tance of the problem and the need for emergent intervention. in systolic blood pressure N40 mm Hg) or clinical signs of They understood that primary health care system at the peripheral choc; (2) respiratory failure defined by the district and subdistrict levels need strengthening to provide presence of tachypnea (respiratory rate N30 breaths per adequate medical services. During the past few years, minute), accessory muscle use, and requirement of oxygen projects were designed to upgrade some district hospitals, therapy or mechanical ventilation; (3) neurologic distress and several intensive care units (ICUs) were created. defined as a Glasgow Coma Scale score of 13 or lower. We Although there seems to have been a slight drop in scorpion also excluded patients with serious comorbidity. Elderly envenomation (SE) mortality, the number of needless death patients were not excluded. remains high. Failure to early identify high-risk patients at the first-line health care was considered as one of the 2.3. Data collection potential causes of this failure. Although there is no All the data were prospectively collected by treating consensus on the criteria for hospital admission, several physicians. Patients received information concerning the recommendations suggest that all patients should be study and were asked to give informed consent for observed during up to 4 hours in the first-line care facility participation. Baseline data were recorded on standard form, before transfer to the nearest community-based hospital including the following: age, sex, comorbidity, scorpion [7,8]. However, for low-risk patients, the time required for color, time between scorpion sting and arrival to the hospital, observation in the emergency department (ED) is often standard vital signs, and treatment received before reaching useless. Likewise, for high-risk patients, this observation the ED. Any case without complete data sheet was excluded. period will be a waste of time and could adversely affect Patients were provided symptomatic treatment and re- outcome because their transfer from rural setting to the evaluated at hourly intervals for up to 4 hours. At the end nearest hospital will likely be unsafe and too late. Clinicians of the ED observation period, patients were discharged home dealing with SE need more objective guidelines for the or hospitalized according to standard accepted criteria: assessment of their patients to guide their decision making patients totally asymptomatic or with only local symptoms in a way that minimizes both complications and unnecessary were discharged home. According to the Tunisian Ministry of health care costs. Prediction rules using scoring systems Health recommendations, those patients with systemic designed to estimate the probability of complication and the symptoms or who showed serious worsening of their clinical need for hospital admission may provide the key to condition were admitted in the ward or the ICU. achieving this goal [9]. In this report, we developed and validated a clinical score based on initial ED presentation 2.4. Construction of the score that would allow one to determine if hospital admission and therapeutic intervention is needed. The score was constructed using data of patients enrolled during the first study period between June 1995 and December 1997 (derivation group). All data analyses were 2. Patients and methods performed with the Statistical Package for the Social Sciences (SPSS 10.0, SPSS, Chicago, Ill). Association of The study was prospectively conducted in the ED of categorical and dependent variables with the hospitalization Tozeur Hospital between June 1995 and December 1997 decision was assessed with v2 test, and the significance of (first period) and between January 1998 and September 1999 continuous variables was assessed with Student t and (second period). This community-based hospital is the only Wilcoxon rank sum tests. Variables were eligible for entry one that covers all southwest Tunisia, which is a rural and endemic area for SE. Yearly, almost 20000 hospital visits due into a multiple logistic regression model if they were significantly associated with hospitalization decision at a to SE are reported (Androctonus australis Hector). We P value of less than .1. Variables associated with P b .05 prospectively included all patients attending the ED. were retained in the final model. To develop a practical 2.1. Inclusion criteria scoring system, all continuous variables were dichotomized. Cutoff points were chosen to make optimal use of the Consecutive patients older than 10 years who came to the information, with the condition that the cutoff values ED were included in the study if they have a documented demarcate a clearly abnormal state and, if possible agree, history of scorpion sting, with the scorpion being seen or with cutoff values used in the literature. No dichotomized captured by the patient or bystander. covariates were entered into the model unless the continu- 2.2. Exclusion criteria ous analogue had a significant independent predictive effect. This strategy was used to ensure that the selection of Patients were excluded from the study if there was a predictive factors for the model would be independent of the clinical evidence of a serious life-threatening symptoms choice of the various cutoff points. The coefficients requiring prompt admission to ICU. Life-threatening symp- obtained from the logistic regression were multiplied by a toms are defined by the presence of (1) hypotension scaling factor to produce a scoring system that required the 416 S. Nouira et al.

September 1999. The value of the score for every patient Table 1 Patients’ characteristics in the derivation and validation groups in this group was calculated using the same equation of the score described in the development group. Predictive Derivation Validation P performance of the score was analyzed using the same group (n= 868) group (n= 435) indices (calibration and discrimination). Mean age (y) (SD) 34.8 (17.9) 35.9 (17.9) .28 The study was approved by our hospital’s institutional Men (%) 530 (61.1) 277 (63.7) .36 review board. It was designed, conducted, and analyzed Recurrence of scorpion 97 (11.8) 49 (11.3) .72 independently of any sponsor. sting (n [%]) Time delay to ED 40.3 (15-112) 42.1 (21-135) .45 arrival (min) (median [range]) 3. Results Localization of sting (n [%]) Arms 407 (47) 217 (50) .43 Overall, 1399 patients presented to the ED after a scorpion Legs 377 (43) 187 (43) .95 sting during the 2 periods of the study. Of these, 96 patients Others 84 (10) 31 (7) (59 in the derivation group and 37 in the derivation group) Vital signs (mean [SD]) failed to satisfy entry criteria and were excluded. The reason Blood pressure (mm Hg) Systolic 124 (25) 123 (25) .57 Diastolic 65 (16) 64 (16) .49 Table 2 Univariate analysis comparing hospitalized and non Heart rate 85 (13) 86 (14) .21 hospitalized patients (beats per min) Respiratory rate 21 (4) 22 (4) .48 Hospitalized Non hospitalized P value (breaths per min) n = 121 n = 747 Temperature N388C 113 (13.0) 61 (14.0) .72 Mean age (y) (SD) 35.6 F 19 35.0 F 18 NS (n [%]) Men (%) 77 (63.6) 453 (60.6) NS Priapism (n [%]) 15 (1.7) 8 (1.8) .88 Recurrence 9 (7.4) 88 (11.8) NS Agitation (n [%]) 16 (1.8) 10 (2.3) .58 of scorpion Vomiting (n [%]) 32 (3.7) 20 (4.6) .32 sting (n [%]) Abdominal distension 17 (1.9) 4 (0.9) .24 Time delay to ED 81.3 F 87 36.2 F 61 b.001 (n [%]) arrival (min) Diarrhea (n [%]) 8 (0.9) 4 (0.9) 1.00 (median [range]) Local pain (n [%]) 798 (91.9) 404 (92.6) 0.73 Localization of sting (n [%]) Treatment received before arrival to ED Legs 70 (57.8) 307 (41.1) b.001 Scorpion antivenom 469 (53.6) 213 (48.6) .10 Arms 14 (36.4) 363 (48.6) b.001 (n [%]) Head and Trunk 7 (5.8) 77 (10.3) b.001 Corticosteroids 125 (14.4) 54 (12.4) .10 Vital signs (mean [SD]) (n [%]) Blood pressure (mm Hg) Length of ED stay 4.7 (1.5) 4.9 (1.5) .63 Systolic 144 F 32 122 F 23 b.001 (hours) (SD) Diastolic 74 F 20 68 F 15 b.001 Hospital admission 121 (13.9) 61 (14.0) .73 Heart rate 93 F 27 84 F 12 b.001 (n [%]) (beats per minute) Length of hospital 1.1 (1.4) 1.5 (1.3) .09 Respiratory rate 24 F 621F 3NS stay (d) (SD) (breaths per min) Temperature N388C 24 (20.5) 89 (12) b.001 (n [%]) addition of integer values. The ability of the score to classify Priapism (n [%]) 12 (10.5) 3 (0.4) .000 b patients (discrimination) was assessed by the area under the Vomiting (n [%]) 22 (18.1) 10 (1.3) .001 Abdominal 9 (7.4) 8 (1.1) b.001 receiver operating characteristic curve (ROC) for dichoto- distension mous outcomes [10]. To evaluate model calibration, we (n [%]) performed Hosmer-Lemeshow goodness-of-fit test compar- Diarrhea (n [%]) 4 (3.3) 4 (0.5) NS 2 ing observed with expected hospitalization rate [11]. Low v Local pain (n [%]) 114 (94.2) 684 (91.6) NS values and high corresponding P values for the Hosmer- Treatment received before arrival to ED Lemeshow statistic indicate that the data can be adequately Scorpion 96 (79.3) 373 (49.9) b0.001 fit to a logistic function. antivenom (n [%]) 2.5. Validation of the score Corticosteroids 67 (55.3) 58 (7.8) b0.001 (n [%]) The validation group included patients who met the same NS, nonsignificant. entry criteria and enrolled between January 1998 and Scorpion envenomation hospitalization score 417

Table 3 Multivariable analysis of variables associated to 3.3. Assessment of the score performance hospital admission and related coefficients for score calculation The mean score in the derivation group was 4.1 F 1.5 in Variable OR 95% CI Coefficient hospitalized patients, as compared to 0.9 F 0.4 in Priapism 150.59 54.20-257.12 +3 nonhospitalized patients ( P b .001). The area under the Vomiting 15.82 5.50-25.49 +2 ROC curve of our score was 0.85 (95% confidence interval Systolic blood pressure 13.38 5.75-21.12 +2 [CI], 0.78-0.92) (Fig. 1A). The goodness-of-fit test showed N160 mm Hg a good agreement between observed and expected hospi- Corticosteroids 10.06 3.80-19.16 +2 talization rate with a Hosmer Lemeshow statistic of 0.64 administration ( P = .88). Using a cutoff value of 2, our score identified before arrival to ED 108 patients among 121 (sensitivity, 89.2%) of those who Time delay to ED 3.71 1.53-9.01 +1 arrival N30 min Temperature N388C 3.66 1.75-7.63 +1 Heart rate N100 beats 3.35 1.38-8.13 +1 per minute OR, odds ratio. of the exclusion included immediate clinical severity (pulmonary edema [n = 12], acute circulatory failure [n = 2], altered consciousness [n = 1]), age less than 10 years (n = 40), serious comorbidity (n = 9) and missing covariate data (n = 32). Thus, 1303 victims of scorpion sting satisfied all entry criteria and comprised the development (n = 868) and the validation population (n = 435). 3.1. Description of the derivation and validation groups Approximately two thirds of the overall population were women, and most of them had no comorbid illness, whereas more than 10% were stung by scorpion for the second time. One hundred eighty-two patients (13.9%) were admitted to the hospital after a 4-hour evaluation period in the ED. Of these, 35 patients (19.2%) were admitted to ICU. The derivation group and the validation group were similar with respect to age and sex distribution, previous health status, and clinical features (Table 1). 3.2. Correlates of hospitalization decision in the Derivation group One hundred twenty-one patients (13.9%) were hospitalized in the derivation group. Results of univariate analysis comparing baseline characteristics of hospitalized and nonhospitalized patients are shown in Table 2. Multivariable analysis revealed the following 7 independent predictors of hospitalization with their correspondent coefficients (Table 3): treatment with corticosteroids received before reaching ED, time delay between scorpion sting and the first presentation in the ED of 30 minutes or longer, vomiting, priapism, temperature above 388C, heart rate above 100 beats per minute, and systolic arterial pressure above 160 mm Hg. For each patient, risk score was calculated by adding the score coefficients associated to each significant Fig. 1 A, Receiver operating characteristic curve in the clinical variable derived from multivariate analysis. The derivation group. B, Receiver operating characteristic curve in score ranged from 0 to 12. the validation group. 418 S. Nouira et al.

discharge induced by scorpion venom [14-17]. Previous Table 4 Observed and predicted hospitalization rate using a cutoff value 2 studies have clearly demonstrated the importance of this hormonal activation in SE and correlation of hypertension to Hospitalized (n) Nonhospitalized (n) Total (n) cardiovascular dysfunction. Corticosteroids were commonly a Derivation group administered in SE despite the lack of evidence of benefit. z Score 2 108 309 417 They may merely be a marker for more serious envenoma- b Score 2 13 438 451 tion. However, the single randomized controlled trial Total 121 747 868 performed by our group addressing this issue provided Validation group Score z2 51 153 204 negative results [18]. That corticosteroids administration is Score b2 10 221 231 independently associated with hospitalization decision Total 61 374 435 suggests that these drugs are not only ineffective in SE a Sensitivity, 89.2%; specificity, 41.4%; positive predictive value, but are likely harmful. Our study found that the time delay 25.9%; negative predictive value, 97.1%. between scorpion sting and ED arrival is an independent predictor of hospitalization. The explanation is not obvious but delay may relate to location and specific scorpion characteristics [19]. In agreement with previous studies, needed hospitalization and 438 patients among 451(negative fever was frequently observed in SE. In animal models, it predictive value, 97.1%) of those did not require hospital- was demonstrated that fever could interfere with toxicoki- ization (Table 4). netic properties of scorpion venom by decreasing its half life elimination and its volume distribution and, hence, increas- 3.4. Validation of the score ing its blood levels [20]. In addition, hyperthermia has been The mean score was 3.8 F 1.5 in the hospitalized group, associated to reduction in survival time and electrocardio- as compared to 1.0 F 0.3 in the nonhospitalized group ( P b graphic disorders in animals [20]. .001). The area under the ROC curve was 0,83 (95% CI, Our model was designed to be independent of biologic 0,72-0,90) (Fig. 1B), and Hosmer Lemeshow statistic was markers because these data may not be available in rural 1.52 ( P = .67). health care facilities. This design allows for broader use because many victims of scorpion sting do not undergo biologic assessment during their initial assessment. 4. Discussion Our study has several limitations. First, we modeled the score on the process rather on the outcome of SE; thus, it Our results indicate that standard clinical characteristics identifies which patients may need hospitalization, rather routinely obtained during the initial medical evaluation of than identifying those at risk for complication and death. patients who are victims of SE can be used to construct a The assumption is that every patient admitted needed simple classification system aiming to aid early decision admission and that every patient discharged did not need about hospitalization. The developed score includes varia- admission. Because this outcome is the basis of examining bles that can be easily obtained when a patient with scorpion how well our score performed, we checked that no patient sting presents to the ED. Moreover, the score can be who was discharged from the ED actually was admitted. In calculated without the aid of a computer and could be addition, most of patients who were admitted needed active assessed by either medical or paramedical staff. Our score observation and treatment. Second, because the data used in discriminate well between patients who needed hospitaliza- our modeling were obtained from one center only, the tion and those who did not. A cutoff value of 2 provided a results may not be applicable to other hospital settings. Of good sensitivity and negative predictive value, indicating note, the toxicity, variation, and duration of symptoms may that it is safe to discharge patients at the first clinical depend on the scorpion, the features of the sting, the features evaluation when the score is less than 2. of the victim, and treatment. Despite these limitations, we The most common important predictor factor for hospital should highlight that our study represents one of the largest admission found in our study was priapism. Nineteen databases on SE and provide the unique opportunity to patients (82.6%) presenting with priapism at the initial study the natural history of this injury during the first hours physical examination have been eventually hospitalized. In in human. Moreover, unlike some previous risk classifica- those for whom hospitalization was not decided (4 patients), tion systems [21], the present score is based on objective priapism was the only symptom associated with local signs. clinical criteria. Vomiting was an important predictor of hospitalization, a In conclusion, one of the most important challenges for finding that is in agreement with previous reports associat- clinicians dealing with SE is to assess accurately and rapidly ing vomiting with the severity of SE [12]. Vomiting was the need for hospital referral. They need objective and usually related to acute pancreatitis, an explanation that is readily available risk scheme to assess accurately and early not widely accepted [12,13]. Other significant variables the need for hospital transfer or discharge. Our score is a included in our score were those related to catecholamine simple screening tool including 7 important clinical risk Scorpion envenomation hospitalization score 419 factors of hospitalization with excellent calibration and [10] Hanley JA, McNeil BJ. The meaning and use of the area under a discrimination. It can serve as a useful aid for early triage of receiver operating characteristic (ROC) curve. Radiology 1982;143: patients with SE. If validated in a large independent cohort, 29-36. [11] Hosmer DW. Applied logistic regression. New York (NY)7 Johon this score could be used to optimize patient management. Wiley & Sons Inc; 1989. [12] Sofer S, Shalev H, Weizman Z, et al. Acute pancreatitis in children following envenomation by the yellow scorpion Leiurus quinques- Acknowledgments triatus. Toxicon 1991;29:125-8. [13] D’Suze G, Sevcik C, Ramos M. Presence of curarizing polypeptides and a pancreatitis-inducing fraction without muscarinic effects in the The authors thank Dr Khadija Beloulied, MD, for her venom of the Venezuelan scorpion Tityus discrepans (Karsch). support in reviewing the data. Toxicon 1995;33:333-45. [14] Gueron M, Adolph RJ, Grupp IL, et al. Hemodynamic and myocardial consequences of scorpion venom. Am J Cardiol 1980;45:979-86. [15] Zeghal K, Sahnoun Z, Guinot M, et al. Characterization and References mechanisms of the cardiovascular and haemodynamic alterations induced by scorpion venom in rats. Fundam Clin Pharmacol 2000;14: [1] Sofer S, Shahak E, Gueron M. Scorpion envenomation and antivenom 351-61. therapy. J Pediatr 1994;124:973-8. [16] Abroug F, Nouira S, El Atrous S, et al. A canine study of immu- [2] Ismail M. The scorpion envenoming syndrome. Toxicon 1995;33: notherapy in scorpion envenomation. Intensive Care Med 2003;29: 825-58. 2266-76. [3] Newlands G. Review of Southern African scorpions and scorpionism. [17] Gwee MC, Nirthanan S, Khoo HE, et al. Autonomic effects of some S Afr Med J 1978;54:613-5. scorpion venoms and toxins. Clin Exp Pharmacol Physiol 2002;29: [4] Vachon M. Studies on scorpions; identification of the scorpions of 795-801. northwest Africa. Arch Inst Pasteur Alger 1951;29:46-104. [18] Abroug F, Nouira S, Haguiga H, et al. High-dose hydrocortisone [5] Santhanakrishnan BR, Balagopal Raju V. Management of scorpion hemisuccinate in scorpion envenomation. Ann Emerg Med 1997;30: sting in children. J Trop Med Hyg 1974;77:133-5. 23-7. [6] Dehesa-Davila M. Epidemiological characteristics of scorpion sting in [19] Njah M, Ben Abdelaziz A, Abdouli M, et al. Health program and use Leon. Guanajuato, Mexico. Toxicon 1989;27:281-6. of community health workers: the example of scorpion envenomation [7] Ministe`re de la Sante´ Publique de Tunisie. Envenimation scorpioni- in Tunisia. Sante´ 2001;11:57-62. que. Tunis7 Circulaire N85798 du 27 Mai; 1998. [20] Ismail M, Abd-Elsalam MA, Morad AM. Do changes in body [8] Ben-Abraham R, Eschel G, Winkler E, et al. Triage for Leiurus temperature following envenomation by the scorpion Leiurus quinquestriatus scorpion envenomation in children: is routine ICU quinquestriatus influence the course of toxicity? Toxicon 1990;28: hospitalization necessary? Hum Exp Toxicol 2000;19:663-6. 1265-84. [9] Ruttimann UE. Statistical approaches to development and validation [21] Abroug F, Elatrous S, Nouira S, et al. Serotherapy in scorpion of predictive instruments. Crit Care Clin 1994;10:19-35. envenomation: a randomised controlled trial. Lancet 1999;354:906-9. American Journal of Emergency Medicine (2007) 25, 420–424

www.elsevier.com/locate/ajem

Original Contribution External cardiac defibrillation during wet-surface cooling in pigsB

Alexandra Schratter MDa, Wolfgang Weihs Mag Veta, Michael Holzer MDa, Andreas Janata MDa, Wilhelm Behringer MDa, Udo M. Losert DVM, PhDb, William J. Ohley PhDc, Robert B. Schock PhDd, Fritz Sterz MDa,* aDepartment of Emergency Medicine, Medical University of Vienna, 1090 Vienna, Austria bCore Unit for Biomedical Research, Medical University of Vienna, 1090 Vienna, Austria cDepartment of Electrical and Computer Engineering, University of Rhode Island, Kingston, RI 02881, USA dLife Recovery Systems, HD, LLC, Kinnelon, NJ 07405, USA

Received 25 February 2007; accepted 28 February 2007

Abstract Objective: During surface cooling with ice-cold water, safety and effectiveness of transthoracic defibrillation was assessed. Methods: In a pig ventricular fibrillation cardiac arrest model, once (n = 6), defibrillation was done first in a dry and then in a wet condition using the ThermoSuit System (Life Recovery Systems, HD, LLC, Kinnelon, NJ), which circulates a thin layer of ice-cold water (648C) over the skin surface. Another time (n = 6), defibrillation was done first in a wet and then in a dry condition. Success of defibrillation was defined as restoration of spontaneous circulation, and the current and voltage of the defibrillation signal was measured. Results: There was a tendency toward less number of shocks needed for achieving restoration of spontaneous circulation in the wet condition as compared with the number of shocks needed in the dry condition. The energy delivered in both dry and wet conditions was 144 F 3J. Discussion: Transthoracic defibrillation is safe and effective in a wet condition after cooling with ice- cold water. D 2007 Elsevier Inc. All rights reserved.

1. Introduction body temperature) can reduce the level of damage to vital organs, including the brain, after cardiac arrest [1-5]. Prior studies have shown that resuscitative mild Conventional noninvasive means that are currently hypothermia (approximately 38Cto58C below normal available for cooling include water-filled cooling blankets and cool air–emitting coverings [5,6]. Some investigators

B have sought to increase the rate of cooling and have This work was supported by Life Recovery Systems, HD, LLC developed invasive methods, such as endovascular cool- (Kinnelon, NJ). * Corresponding author. Tel.: +43 1 40400 1964; fax: +43 1 40400 1965. ing catheters [7-9] and cold saline [8,10] or iced saline E-mail address: [email protected] (F. Sterz). slurry injected into the circulatory system [11]. However,

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2007.02.044 Cold wet trans-thoracic defibrillation 421 these would be challenging, if not impossible, to (20 mg/kg/h) and boluses of piritramide (15 mg IV) were implement in the field. given via a peripheral IV cannula (ear vein, 18 G). Life Recovery Systems, HD, LLC (Kinnelon, NJ) has Rocuronium (20 mg) was given for muscle relaxation. developed an advanced surface cooling system, the Thermo- Saline (5 mL d kg À1 d hÀ1) was administered to maintain Suit System (TSS), for rapid cooling of the patient. The central venous pressure N3 mm Hg. fundamental approach of using ice water to cool the human body is already known to be rapid [12], and it has been shown 2.2. Monitoring to be safe when used in certain surgical applications requiring Electrocardiogram (ECG) electrodes were attached to the extended periods of cardiac arrest [12]. The TSS is being extremities and connected to the ECG monitor. A gastric developed so that cooling could be efficiently delivered even tube, featured with an esophageal temperature probe (Mon- while the patient is in a state of cardiac arrest or if a converted a-therm General Purpose, 9 F, Mallinckrodt Medical, St patient requires defibrillation again while being cooled. The Louis, MO) was inserted. The esophageal temperature probe ability to use defibrillation during the use of the TSS would was connected to a data acquisition system (DATAQ avoid interruption in the cooling of the patient and eliminate Instruments, Akron, Ohio). A Foley catheter with a the time delay in returning the patient to normal sinus rhythm temperature probe (Ruesch Sensor Ch 12; Ruesch, Kernen, because of the need to drain the water from the TSS [13]. Germany) (Tbl) was inserted for collecting urine. Therefore, our objective was to prove the safety of A catheter was placed by Seldinger technique in the left performing transthoracic defibrillation in a wet condition brachial artery to monitor mean arterial pressure (MAP) and during the use of the TSS. We compared the current leakage for blood sampling. A pulmonary artery catheter (Thermo- and defibrillation success in dry and wet conditions after dilution Paceport Catheter, 7.5 F, Edwards Lifesciences 1 minute of ventricular fibrillation (VF) cardiac arrest. LLC, Irvine, CA) was inserted via the right jugular vein by Seldinger technique for monitoring of pulmonary artery blood temperature and administration of medications and 2. Methods infusions as well as for inserting a pacing wire to induce VF. Animals in which a sequence of induced VF was After insertion of all catheters, heparin (50 IU/kg) was followed by defibrillation were prospectively studied. This applied to avoid clotting. study was conducted in compliance with the good labora- The following parameters were monitored continuously: tory practice regulations set forth in part 58 of title 21 of the ECG, MAP, central venous pressure, pulmonary artery United States Code of Federal Regulation. The study was pressure, tidal volume, respiratory rate, ventilation pressures approved by our institutional animal investigation commit- (peak, mean, and end-expiratory), and all temperatures tee. Animal care and use was performed by qualified described previously. Blood gases, electrolytes, hematocrit, personnel and supervised by veterinarians. All animal lactate, and glucose were measured at selected time points. facilities and transportations complied with the current legal Baseline blood temperature was taken as a reference requirements and guidelines. temperature and maintained within the reference range for pigs (38.58C F 0.28C) using heating blankets (Warm Air 2.1. Animal preparation Hyperthermiasystem 134, CSZ Cincinnati Subzero Products, Cincinnati, OH) or fans. Female pigs (Large White breed) weighing 28 to 35 kg To record voltage and current of the defibrillation signal, were obtained from the licensed farm of the University 2 Pearson current probes (Pearson Electronics Inc, Palo for Veterinary Medicine of Vienna, Vienna, Austria Alto, Calif) were interposed between defibrillator and (Hochschulgut Medau) and brought to the testing facility automatic external defibrillator (AED) pads (for details, 14 days before the experiment. They were given see Supplementary Material). cefquinom sulfate (2 mL) intramuscularly for infection prophylaxis and fasted with free access to water 12 hours 2.3. Test item before the experiment. After premedication with intramuscularly administered The TSS provides highly efficient cooling of the skin by 6 atropine sulfate (0.5 mg), ketamine hydrochloride (20 mg/ circulating a thin layer of ice-cold water ( 48C) over most kg), acepromazine maleate (1.75 mg/kg), and piritramide of the skin surface by means of a suit component that (22.5 mg), sedation was induced with an IV propofol bolus conforms to the body and an external water-circulating (40 mg). Then, the pigs were intubated and mechanically pump. The TSS was set up before the experiment. The ventilated (Servo 300, Maquet Critical Care, Solna, Swe- equipment was specially designed for the 30-kg swine den) with tidal volumes of 10 mL/kg, positive end- model used in this study (Fig. 1). io expiratory pressure of 5 cm H2O, F 2 of 0.3, and a ratio 2.4. Experimental protocol of inspiration to expiration of 1:2. The respiratory rate was adjusted to achieve a Paco2 of 35 to 40 mm Hg (4.7 to 5.3 After cardiopulmonary parameters and temperatures kPa). To maintain anesthesia during preparation, propofol were stabilized, 2 baseline measurements were made, and 422 A. Schratter et al.

delivered via commercially available self-adhesive monitor-defibrillator electrode pads (model M3713A; Philips Medical Systems, Andover, Mass), placed on the shaved chest of the pig. The 2 pads were placed on either side of the chest, just below the upper extremities. After restoration of spontaneous circulation (ROSC), the procedure was repeated after 20 minutes. In a crossover design, 2 different conditions were chosen to represent conditions that may exist during the use of the TSS system. These were (1) subject dry (n = 6) and (2) subject wet (n = 6) with water contaminated with saline solution. Once, animals received VF in the dry condition first followed by the wet condition. Another time, the animals first had the wet condition and then were Fig. 1 Life Recovery Systems (Life Recovery Systems, HD, dried completely for the dry condition. During TSS LLC, Kinnelon, NJ) TSS, an advanced surface cooling system, for application with cooling water, VF was initiated after the the rapid cooling used in the experiments (external water- core temperature had fallen to 338C F 18C. The sequence circulating pump not shown). of treatment was randomly selected by sealed envelopes after data recording had begun. Randomization of animals a bolus of 40 mg of propofol was administered; VF was to a treatment sequence was performed using a computer- initiated electrically by the passage of a 60-Hz alternating printed randomization schedule. The defibrillation for current through a catheter to the right ventricular apex. achieving ROSC in the wet setting was performed while Abrupt reduction of blood pressure and ECG readings the suit was pumping. The water was drained from the suit confirmed the arrest. Heating devices, IV fluids, and as soon as ROSC was achieved. Consequently, 6 experi- anesthesia were discontinued, and the endotracheal tube ments were performed in a wet condition and 6 in a dry was disconnected from the ventilator. Ventricular fibrilla- condition. Pigs that underwent wet defibrillation first were tion was allowed to persist for 1 minute. Defibrillating carefully dried before the procedure was repeated in the shocks were delivered from an AED (Heartstart 4000; dry condition. Afterwards, pigs were killed with a Laerdal, Stavanger, Norway), which was determined to potassium chloride solution. deliver biphasic truncated exponential waveform shocks of 2.5. Outcome equal pulse duration and a selected energy of 150 J. The biphasic waveform shocks composed of equal-duration Defibrillation success was defined as ROSC with a MAP pulses included a negative pulse (4.5 ms/4 ms) and a of 50 mm Hg for at least 5 minutes. In addition, the number positive pulse (4.5 ms/4 ms). Defibrillation shocks were of defibrillation shocks needed to achieve ROSC was

Fig 2 Example of voltage and current curves of the defibrillation signal in dry and wet conditions (arbitrary values on the y-axis, voltage in volts, and current in amperes). Cold wet trans-thoracic defibrillation 423 recorded. Voltage and current of the defibrillation signal Table 2 Physiologic variables at baseline and 5 minutes after were measured. Energy over time was calculated by ROSC in dry (n = 6) and wet (n = 6) conditions multiplying the voltage by the current and then, integrated, Baseline 5 min after ROSC dividing them by the delivery time of the shock. This resulted in total joules delivered. Leakage current, defined Heart rate Wet 110 (97-120) 126 (111-135) as current that could flow from the surface of the human (beats/min) Dry 116 (113-149) 150 (126-162)* body to the ground and therefore reflects a shock hazard to Mean arterial Wet 70 (66-72) 65 (48-86) pressure Dry 70 (62-86) 63 (54-67) the rescuer if there is a failure, was estimated by calculation (mm Hg) of the area under the curve (AUC) of the energy-time curves pH Wet 7.50 (7.46-7.53) 7.43 (7.41-7.45) during the first shock in both dry and wet conditions. At the Dry 7.47 (7.45-7.55) 7.38 (7.34-7.42) end of the experiment, the skin areas under the defibrillation Potassium Wet 4.0 (3.8-4.2) 3.6 (3.2-4.0) pads were carefully examined for skin lesions and burns. (mmol/L) Dry 3.7 (3.4-4.3) 4.1 (3.3-4.4)* After the animals were killed, a necropsy was performed to Base excess Wet 3.6 (2.6-4.1) 1.6 (0.6-3.5) exclude severe cardiac or lung diseases that could interfere (mEq/L) Dry 2.5 (1.2-4.4) À0.3 (À2.0 to 1.3)** with defibrillation success. Lactate Wet 1.5 (1.1-2.3) 2.0 (1.8-2.3) (mmol/L) Dry 1.7 (0.9-2.3) 3.5 (2.3-4.4)** 2.6. Statistical analysis Variables are given as median and the range from the first to the third quartile. Temperature and AUC values were normally distributed * P = .03. and thus reported as mean and SD. Continuous variables ** P = .04. not normally distributed are given as median and the range from the first to the third quartile. Heart rate and MAP as well as physiologic parameters at baseline and 5 minutes Heart rate was lower in the wet vs dry condition at 5 after ROSC were compared between dry and wet minutes after ROSC (126 [111-135] vs 150 [126-162] beats conditions with the Wilcoxon signed rank test using the per minute, P = .03]. There was no significant difference in exact version of the test. The energy AUC between dry the MAP between dry and wet conditions. Potassium and and wet conditions was compared with the t test. All lactate were higher during the dry as compared with the wet calculations were performed with SPSS for Windows, 10.0 condition at 5 minutes after ROSC (4.1 [3.3-4.4] vs 3.6 [3.2- (SPSS Inc, Chicago, Ill). P b .05 was considered 4.0] mmol/L, P = .03, and 3.5 [2.3-4.4] vs 2.0 [1.8-2.3] mmol/ statistically significant. L, P = .04). Base excess was À0.3 mEq/L (À2.0 to 1.3 mEq/ L) during the dry vs 1.6 mEq/L (0.6-3.5 mEq/L) during the wet condition at 5 minutes after ROSC ( P = .04) (Table 2). 3. Results Furthermore, no skin lesions or burns were found on Pigs weighed 30.7 F 1.0 kg, and ROSC was achieved in the skin areas located underneath the defibrillation pads in all animals. There was a tendency toward less number of particular or elsewhere, and no frostbite was observed. shocks needed for achieving ROSC in the wet condition (1, Within the 20-minute postdefibrillation observation peri- 1, 1, 1, 1, 2) as compared with the number of shocks needed od, ventricular tachycardia occurred in 3 pigs at hypo- in a the dry condition (1, 1, 2, 2, 2, 2). The energy AUC in thermia (53, 349, and 30 seconds) and in 2 pigs at both dry and wet conditions was 144 F 3J(P = .96). This normothermia (3 seconds each). However, these tachycar- concordance indicates that no appreciable leakage current dia periods did not affect the hemodynamic stability and existed in the wet condition (Fig. 2). The impedance was did not have to be treated either pharmacologically or similar in both dry and wet conditions (Table 1). electrically (Table 2).

4. Discussion Table 1 Transthoracic impedance Wet condition Dry condition External cardiac defibrillation from VF with an AED is (n = 6) (n = 6) safe and effective in a wet-surface cooling setting. Defibril- Leading-edge 53.2 (46.8-58.8) 48.6 (47.0-54.0) lation success was defined as ROSC, which could be impedance (V) achieved in all included animals. In the dry condition, 2 pigs Trailing-edge 53.7 (47.5-104.4) 51.4 (47.2-61.3) achieved ROSC after 1 shock and 4 pigs after 2 shocks. In impedance (V) the wet condition, 5 pigs achieved ROSC after 1 shock and Variables are given as median and the range from the first to the third 1 pig after 2 shocks. The approach performed seems to be quartile. Leading-edge impedance is the value that corresponds to the safe for personnel using this device. left edge of the impedance plateau. Trailing-edge impedance is the The safety of external cardiac defibrillation with an value that corresponds to the right edge of the impedance plateau. AED in a wet condition has already been assessed [14]. 424 A. Schratter et al.

However, the study of Lyster et al [14] focused only on Acknowledgments the safety for the bystanders and did not use large animals in conditions after cardiac arrest. Similarly, we did The authors gratefully acknowledge the help of all the not observe any appreciable leakage current in the wet laboratory technicians and nurses of the Core Center of condition because the energy AUC did not differ between Biomedical Research, Vienna, Austria. dry and wet settings. There appears to be a tendency that with a drop in core temperature, the defibrillation success increases. In fact, this References assumption has already been investigated [15,16]. In these studies, the beneficial effect of hypothermia on defibrillation [1] Leonov Y, Sterz F, Safar P, et al. Mild cerebral hypothermia during success could be shown. and after cardiac arrest improves neurologic outcome in dogs. J Cereb Since the beneficial effects of therapeutic hypothermia Blood Flow Metab 1990;10:57-70. [2] Sterz F, Safar P, Tisherman S, et al. Mild hypothermic cardiopulmo- after cardiac arrest have been shown in multiple studies nary resuscitation improves outcome after prolonged cardiac arrest in [17], various cooling methods, including surface cooling dogs. Crit Care Med 1991;19:379-89. devices, have become an important issue to emergency and [3] Weinrauch V, Safar P, Tisherman S, et al. Beneficial effect of mild intensive care units. Especially for the handling of surface hypothermia and detrimental effect of deep hypothermia after cardiac cooling devices using freely circulating water, such as the arrest in dogs. Stroke 1992;23:1454-62. [4] Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose TSS, it is fundamental to guarantee the safety for the survivors of out-of-hospital cardiac arrest with induced hypothermia. attending personnel and patient. Furthermore, the possibility N Engl J Med 2002;346:557-63. of defibrillating effectively in a wet condition is very [5] Hypothermia after Cardiac Arrest Study Group. Mild therapeutic important in case a successfully resuscitated patient requires hypothermia to improve the neurologic outcome after cardiac arrest. defibrillation again while being cooled. This study suggests N Engl J Med 2002;346:549-56. [6] Felberg RA, Krieger DW, Chuang R, et al. Hypothermia after cardiac that no risks result from the use of an AED during surface arrest feasibility and safety of an external cooling protocol. Circulation cooling with contaminated ice water. 2001;104:1799-804. The more frequent and prolonged durations of ventricular [7] Dae MW, Gao DW, Sessler DI, et al. Effect of endovascular cooling tachycardia, higher heart rates, lower potassium levels, and on myocardial temperature, infarct size, and cardiac output in human- less lactate acidosis within the 20-minute postdefibrillation sized pigs. Am J Physiol Heart Circ Physiol 2002;282:H1584-91. [8] Behringer W, Safar P, Wu X, et al. Veno-venous extracorporeal blood observation period observed after wet conditions could be shunt cooling to induce mild hypothermia in dog experiments and explained by the physiologic effects of cooling. However, review of cooling methods. Resuscitation 2002;54:89-98. these tachycardia periods seemed not to have significantly [9] Holzer M, Behringer W, Janata A, et al. Extracorporeal venovenous affected the experiments during the observation period, such cooling for induction of mild hypothermia in human-sized swine. Crit as in the alterations of the laboratory results, which were Care Med 2005;33:1346-50. [10] Polderman KH, Rijnsburger ER, Peerdeman SM, Girbes AR. within clinical tolerable ranges (Table 2). Induction of hypothermia in patients with various types of neurologic A limitation of our study could be the small sample size injury with use of large volumes of ice-cold intravenous fluid. Crit and its crossover design because of the possible interference Care Med 2005;33:2744-51. of the first 1-minute cardiac and respiratory arrest event on [11] Vanden Hoek TL, Kasza KE, Beiser DG, et al. Induced hypothermia outcome of the second arrest event in the same subject. by central venous infusion: saline ice slurry versus chilled saline. Crit Care Med 2004;32:S425-31. Even so, we think that it is justified to draw the conclusions, [12] Miyazaki M, Yoda K, Tanaka Y, et al. A study of profound especially with regard to the safety and efficacy of hypothermia by surface cooling. Can Anaesth Soc J 1980;27: defibrillation in a wet condition, because this design equally 370-80. allows both conditions to follow each other 6 times. [13] Kuboyama K, Safar P, Radovsky A, et al. Delay in cooling negates the Transthoracic defibrillation via AED pads is safe and beneficial effect of mild resuscitative cerebral hypothermia after cardiac arrest in dogs: a prospective, randomized study. Crit Care Med effective in a wet condition after cooling with ice-cold 1993;21:1348-58. water in a pig VF cardiac arrest model because ROSC [14] Lyster T, Jorgenson D, Morgan C. The safe use of automated could be achieved in all animals. Thus, this new cooling external defibrillators in a wet environment. Prehosp Emerg Care device needs further exploration in cases of cardiac arrest 2003;7:307-11. in humans. [15] Boddicker KA, Zhang Y, Zimmerman MB, et al. Hypothermia improves defibrillation success and resuscitation outcomes from ventricular fibrillation. Circulation 2005;111:3195-201. [16] Rhee BJ, Zhang Y, Boddicker KA, et al. Effect of hypothermia on Appendix A. Supplementary material transthoracic defibrillation in a swine model. Resuscitation 2005;65:79-85. [17] Nolan JP, Morley PT, Vanden Hoek TL, et al. Therapeutic Supplementary data associated with this article can be hypothermia after cardiac arrest: an advisory statement by the found, in the online version, at doi:10.1016/j.ajem.2007. Advanced Life Support Task Force of the International Liaison 02.044. Committee on Resuscitation. Circulation 2003;108:118-21. American Journal of Emergency Medicine (2007) 25, 425–429

www.elsevier.com/locate/ajem

Brief Report

Clinical measures associated with FEV1 in persons with asthma requiring hospital admissionB Donald H. Arnold MD, MPHa,b,*, Tebeb Gebretsadik MPHc, Patricia A. Minton RNd,e, Stanley Higgins PhDd, Tina V. Hartert MD, MPHd,e,f aDepartment of Emergency Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee, USA bDepartment of Pediatrics, Vanderbilt University School of Medicine, Nashville, Tennessee, USA cDepartment of Biostatistics, Vanderbilt University School of Medicine, Nashville, Tennessee, USA dDivision of Allergy, Pulmonary and Critical Care Medicine, Vanderbilt University School of Medicine, Nashville, TN 37232-4700, Tennessee, USA eDepartment of Medicine, Vanderbilt University School of Medicine, Nashville, Tennessee, USA fCenter for Health Services Research, Vanderbilt University School of Medicine, Nashville, Tennessee, USA

Received 19 July 2006; revised 30 August 2006; accepted 5 September 2006

Abstract Objective: We sought to determine the association of select clinical measures of asthma severity with percent predicted forced expiratory volume in one-second (%FEV1). Methods: We studied a prospective cohort of adult subjects (N = 129) with asthma exacerbations requiring hospital admission. Clinical data was acquired, including medical and social history, symptoms, vital signs, physical assessment, and spirometry. Predictor variables for this study included manually determined pulsus paradoxus (PP), percent predicted peak expiratory flow rate (%PEFR) and accessory muscle use. The outcome measure was %FEV1. Multiple linear regression analyses were performed to determine the independent associations between predictor variables and %FEV1. Results: In univariate analysis, %PEFR correlated with %FEV1 (rho = 0.77, P b .001) and PP correlated negatively with %FEV1 (rho = À0.384, P b .001). %FEV1 was significantly lower in participants with accessory muscle use (Median %FEV1 = 37.5%, IQR: 27.0-49.0) than in those without accessory muscle use (Median %FEV1= 55.0%, IQR: 39.0-69.0), ( P = .004). In multivariable analysis including the covariates %PEFR, accessory muscle use, PP, age, sex, heart rate and respiratory rate, %PEFR ( P b .0001) and accessory muscle use ( P = .003) remained significantly associated with %FEV1, whereas PP did not ( P = .52). D 2007 Elsevier Inc. All rights reserved.

Presented at the International Conference of the American Thoracic Society, San Diego, Calif, May, 2006. 1. Introduction B This study was supported by National Institutes of Health grant KO8 AI001582, Bethesda, MD (to TVH), American Lung Association Clinical 1.1. Background and importance Research grant New York, NY (to TVH), and also supported in part by grant M01 RR-00095 from the National Center for Research Resources, NIH. Asthma is the most common chronic disease in children * Corresponding author. Tel.: +1 615 936 0095; fax: +1 615 936 1316. and one of the most common chronic diseases in adults E-mail address: [email protected] (D.H. Arnold). [1-4]. Approximately 2 million emergency department (ED)

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2006.09.006 426 D.H. Arnold et al. visits in the United States each year are for acute asthma congestive heart failure that could account for the acute exacerbations, and 7% to 15% of these patients relapse after illness or if they were previously enrolled in the study. The ED discharge [1,5,6]. These high rates of disease exacerba- study was approved by the institutional review board, and tion and relapse are due, at least in part, to the difficulty in the subjects provided written informed consent. assessing asthma severity and predicting subsequent course and outcome of the exacerbation at the bedside. 2.2. Data collection and processing National Heart Lung and Blood Institute guidelines for The study nurses completed a standardized form, management of acute asthma exacerbations recommend including medical and social history and clinical symptoms. assessment of signs, symptoms, and functional tests of lung We measured chronic asthma disease severity with the Johns function. These assessments include measurement of pulsus Hopkins Asthma and Allergy Composite score, in which paradoxus and peak expiratory flow rate (PEFR) and lower values indicate milder disease [19]. All participants appraisal of accessory muscle use [7]. Adherence to these had spirometry performed, had vital signs measured, and guidelines has been limited [7-10]. Further, physicians’ underwent a physical assessment. We followed up patients ability to assess severity has been found to be variable and daily during hospitalization and collected clinical data at limited in accuracy, and patient perceptions of severity have admission 24 hours after admission, at discharge, and at not correlated well with measured severity of airflow 3 months after discharge. This study pertains to data obstruction [5,11-15]. obtained at the time of hospital admission. Percent predicted FEV1 (%FEV1) is a criterion standard for assessing the severity of airway obstruction in asthma 2.3. Predictor and outcome measures and the response to therapy [16,17]. However, spirometry is often not available in the ED environment, and patients Predictor variables used for this study included PP, experiencing clinically significant airflow obstruction are %PEFR, and accessory muscle use. The outcome measure frequently unable to perform the necessary forced airway was %FEV1 at the time of admission to the hospital, maneuvers. Accessory muscle use has been shown to be performed in accordance with American Thoracic Society standards [17,20]. associated with significant decline in %FEV1 [18]. Further understanding of the relationship between these clinically 2.4. Data Analysis available measures of asthma severity and %FEV1 will better inform decision making for patients with acute asthma We generated descriptive statistics for participants’ exacerbations. demographic and clinical characteristics. Categorical variables are presented as number and proportions, and continuous 1.2. Goals of this investigation variables, as medians and interquartile range (IQR). Corre- We sought to determine whether clinical measures of lations between %FEV1 and %PEFR and PP were performed asthma severity, including percent predicted peak expiratory using Spearman rank correlation coefficients (q). The flow rate (%PEFR), manually determined pulsus paradoxus Wilcoxon rank sum test was performed to assess whether the use of accessory muscles (yes/no) was statistically (PP), vital signs and accessory muscle use, predict %FEV1 in acutely ill patients evaluated and determined to require associated with %FEV1. Adjusted multiple linear regression admission to the hospital. models were used to evaluate the independent associations between %PEFR, PP, and accessory muscle use and %FEV1. Residual analyses of the multiple linear regression were performed graphically by plotting residuals against predicted 2. Methods values and plotting normal Q-Q plot to assess normality. 2.1. Setting and selection of participants Regression covariates were age, sex, respiratory rate, and heart rate. The maximum number of independent variables The Bronchopulmonary Response during Episodes of were chosen a priori and based on the 10 events per variable Asthma and the Treatment and History of Exacerbations rule to prevent overfitting [21]. Nonlinearity for continuous (BREATHE) cohort is an ongoing, prospective, convenience independent variables was assessed, and %PEFR was sample of subjects 18 years and older, recruited from included as a nonlinear term using restricted cubic splines eligible patients with asthma exacerbations admitted to a [22]. All analyses used a 5% 2-sided significance level and single tertiary university teaching hospital. The BREATHE were performed with STATA 8.0 and R 2.1.0 (http://www. database includes demographic, historical, physical exami- r-project.org). nation, and laboratory variables. During a full 5-year period, 5 days per week, all adult patients hospitalized with a diagnosis of acute asthma were approached for study 3. Results inclusion. Charts on these subjects were reviewed to assure that the hospital admission was for asthma. Patients were During recruitment hours, we approached 136 of 235 excluded if they had other chronic pulmonary disease or unique eligible patients admitted during the entire study Association of clinical measures with %FEV1 427

Table 1 Demographic characteristics of 129 adults hospitalized for an acute asthma exacerbation Characteristic Value Age (y) (mean F SD) 41.6 F 12.5 Female (%) 80% Race (n [%]) White 68 (53%) African American 58 (45%) Other 3 (2%) Insurance (n [%]) TennCare (Medicaid) 55 (43%) Other Medicaid/Medicare 12 (9%) Commercial 50 (39%) None 12 (9%) Type of outpatient physician (n [%]) Primary care physician 76 (59%) Pulmonologist 40 (31%) Allergist 3 (2%) Primary care physician with 6 (5%) pulmonologist or allergist Other 4 (3%) Fig. 1 Top, Distribution of %FEV1 by presence or absence of accessory muscle use (Wilcoxon rank sum test, P = .004). Each th th period (December 1999 through March 2006) with an box extends from 25 (lower hinge) to 75 percentile (upper asthma exacerbation and recruited 129 to participate. This hinge). The median is shown as a line across the box. An outside value is defined as a value that is smaller than the lower quartile resulted in a participation rate of 95% of those approached minus 1.5 times the IQR, or larger than the upper quartile plus 1.5 and 55% of eligible patients overall. Demographic features times the IQR (inner fences or the lines outside boxes). A far-out of the participants are included in Table 1. The majority of value is defined as a value that is smaller than the lower quartile participants were female (80%) and white (53%). There was minus 3 times the IQR, or larger than the upper quartile plus 3 representation of both Tennessee Medicaid (43%) and times the IQR. Bottom, Points represent the scatterplot of %FEV1 commercial (39%) insurance coverage, and most subjects by %PEFR. The solid line indicates %FEV1 by %PEFR using (59%) received their outpatient care from a primary care multiple linear regression (model adjusted for age, sex, pulsus physician alone. paradoxus, heart rate, respiratory rate and accessory muscle use). The mean age of asthma onset in this cohort was early Dashed lines indicate 95% CIs for the regression line. adulthood (Table 2), and most participants were either current or past smokers. Inhaled corticosteroids had been prescribed intensive care unit (ICU) for asthma (38%) or had undergone to most (80%) participants. A substantial proportion had been endotracheal intubation (20%) for respiratory failure. admitted to the hospital within 5 years (54%) or to an In the multiple linear regression adjusting for age, sex, PP, respiratory rate, heart rate, and accessory muscle use, the association of %PEFR and %FEV1 remained significant ( P b Table 2 Asthma characteristics of 129 adults hospitalized for .0001 for main effect and P = .08 for nonlinearity) (Fig. 1), an acute asthma exacerbation whereas the association of PP and %FEV1 did not (b = À.13; Baseline asthma characteristics Value 95% confidence interval [CI], À0.53 to 0.27, P = .52). The Known earliest age of asthma (y) (mean F SD) 21 F 17.8 distribution of %FEV1 was significantly lower in participants Participant monitors PEFR at home 53 (42%) with accessory muscle use (median %FEV1 = 37.5%; IQR, Hospitalized for asthma within the past 5 y 70 (55%) 27.0-49.0) than in those without accessory muscle use Admitted to ICU during lifetime (%) 48 (37%) (median %FEV1 = 55.0%; IQR, 39.0-69.0) ( P = .004) Intubated for asthma during lifetime (%) 25 (20%) (Fig. 1) and remained significant in multiple linear regression Asthma severity scorea (median [IQR]) 57 (40-80) model (b = À12.30; 95% CI, À20.38 to À 4.38, P = .003).

Acute asthma characteristics Value %FEV1 at hospital admission (median [IQR]) 52 (36-67) %PEFR at hospital admission (median [IQR]) 54 (35-72) 4. Discussion Respiratory rate at admission (breaths per min) 22 (20-24) Clinicians who care for patients with acute asthma Accessory muscle use 19 (15%) Pulsus paradoxus (mm Hg) (median [IQR]) 17 (12-22) exacerbations must have metrics with which to measure severity to make treatment and disposition decisions. FEV a Johns Hopkins Asthma and Allergy Composite score. 1 is one such metric and serves as a criterion standard for 428 D.H. Arnold et al. measurement of airways obstruction. Our findings of the quantify asthma severity, we found that those participants association of easily measured clinical signs and symptoms using accessory muscles to breathe had a clinically important with spirometry may be of use to clinicians evaluating difference in %FEV1 (median, 37.5%) as compared with acutely ill asthmatics and who do not have immediate access those without accessory muscle use (median, 55.0%). It to spirometry. The findings are informative to clinicians by follows that the patient with accessory muscle use, a furthering the understanding of how %PEFR, manually relatively simple and easily observed physical sign, warrants determined pulsus paradoxus, and accessory muscle use are a higher level of therapeutic intervention and monitoring. associated with or quantify airways obstruction. However, A strength of this study is the use of trained clinical study accurate, objective, quantifiable, and practical measures of nurses with expertise in performing spirometry and in acute asthma severity are needed and should be evaluated as manually determining pulsus paradoxus. In addition, we single and composite measures. were able to achieve a participation rate of 95% of eligible A substantial proportion of our study cohort had been patients approached for enrollment. These features enhance admitted to the hospital within 5 years or to an ICU for the internal validity of our findings. asthma or had undergone endotracheal intubation for In addition, our trained study nurses focused on complete respiratory failure. The prevalence of prior respiratory and accurate acquisition of clinical signs and data. Our failure (20%) in this cohort is consistent with other studies findings may have limited external validity to other ED conducted in inner-city teaching hospitals [27,28]. These environments, where less optimal conditions of data adverse events, together with the baseline asthma severity acquisition may prevail. In addition, the data pertain to a scores and baseline %FEV1 recorded on an outpatient basis, single, prospective convenience sample of adults presenting indicate less than optimal overall asthma control for this with acute asthma exacerbations and admitted to the cohort. Moreover, %FEV1 and clinical signs, including hospital. Patients evaluated, treated, and discharged from tachypnea, accessory muscle use, and pulsus paradoxus, the ED for an asthma exacerbation were not included in this represent clinically severe levels of airway obstruction at the cohort. These study features have selected a cohort with time of hospital admission. high rates of severe and life-threatening asthma. However, Given this degree of asthma severity, we are surprised at this cohort is representative of hospitalized patients with the failure of pulsus paradoxus to be associated with %FEV1 asthma and those at highest risk for dying from their asthma. (adjusted coefficient P value = .52). This finding is more This may have introduced spectrum bias and may limit the understandable given that pulsus paradoxus is influenced by ability to generalize our findings to a general ED population. systemic and pulmonary venous return as well as intra- Lastly, the precision of PP measurement could not be pleural pressure change during the respiratory cycle [29]. evaluated in this study, and manual blood pressure Although airway obstruction directly influences %FEV1 and measurement alone has been shown to be correlate modestly indirectly influences pulsus paradoxus as a result of with intra-arterial measurement [33]. increasing magnitude of intrapleural pressure change, failure In patients with acute asthma exacerbations assessed by a of these 2 measures to correlate may not be unexpected. trained study nurse and requiring admission to hospital, This finding is contrary to previous studies and to the %PEFR is modestly associated with %FEV1, whereas ma- honored position this physical sign has enjoyed in the nual pulsus paradoxus is not. Accessory muscle use, a simple traditions and history of medicine, as well as in National and easily observed physical examination sign, is associated Heart Lung and Blood Institute guidelines for asthma care with, but does not precisely quantify, clinically important [7,18,30,31]. Our finding is nonetheless consistent with the decreases in %FEV1. These commonly used bedside clinical recognized challenges in manually measuring pulsus para- measures are limited as correlates of disease severity. doxus in a noisy clinical environment, particularly among those with rapid respiratory rates where the systolic difference between inspiration and expiration could be very References difficult to determine [32]. We found that %PEFR was associated with %FEV1 [1] Mannino DM, Homa DM, Akinbami LJ, et al. Surveillance for (adjusted coefficient P = .001, see Fig. 1). That %PEFR is asthma—United States, 1980-1999. MMWR Surveill Summ 2002;51: 1-13. associated with %FEV1 is perhaps not surprising because both measure primarily large and medium airway function, [2] Asthma prevalence and control characteristics by race/ethnicity— United States, 2002. MMWR Morb Mortal Wkly Rep 2004;53:145-8. and persons adept at performing one measure should be able [3] Akinbami LJ, Schoendorf KC. Trends in childhood asthma: prevalence, to similarly perform the other. In contrast, other investi- health care utilization, and mortality. Pediatrics 2002;110:315-22. gators have found that %PEFR does not adequately predict [4] Yawn BP, Wollan P, Kurland M, et al. A longitudinal study of the prevalence of asthma in a community population of school-age %FEV1 in children and adults with variable degrees of airway obstruction [23-26]. children. J Pediatr 2002;140:576-81. [5] McFadden Jr ER. Acute severe asthma. Am J Respir Crit Care Med Our finding of the association of accessory muscle use 2003;168:740-59. with the degree of %FEV1 impairment is of practical use to [6] McCaig LF, Burt CW. National Hospital Ambulatory Medical Care clinicians. While accessory muscle use did not precisely Survey: 2002 emergency department summary. Adv Data 20041-34. Association of clinical measures with %FEV1 429

[7] National Heart, Lung, and Blood Institute, National Asthma Education [21] Peduzzi P, Concato J, Kemper E, et al. A simulation study of the and Prevention Program. Expert panel report 2: guidelines for the number of events per variable in logistic regression analysis. J Clin diagnosis and management of asthma. National Asthma Education Epidemiol 1996;49:1373-9. and Prevention Program (Publication No. 97-4051). Bethesda, MD7 [22] Harrell FE. Regression modeling strategies. New York7 Springer; US Department of Health and Human Services, National Institutes of 2001. Health; 1997. [23] Giannini D, Paggiaro PL, Moscato G, et al. Comparison between peak [8] Finkelstein JA, Lozano P, Shulruff R, et al. Self-reported physician expiratory flow and forced expiratory volume in one second (FEV1) practices for children with asthma: are national guidelines followed? during bronchoconstriction induced by different stimuli. J Asthma Pediatrics 2000;106:886-96. 1997;34:105-11. [9] Milks CJ, Oppenheimer JJ, Bielory L. Comparison of emergency room [24] Llewellin P, Sawyer G, Lewis S, et al. The relationship between FEV1 asthma care to national guidelines. Ann Allergy Asthma Immunol and PEF in the assessment of the severity of airways obstruction. 1999;83:208-11. Respirology 2002;7:333-7. [10] Canny GJ, Reisman J, Healy R, et al. Acute asthma: observations [25] Klein RB, Fritz GK, Yeung A, et al. Spirometric patterns in childhood regarding the management of a pediatric emergency room. Pediatrics asthma: peak flow compared with other indices. Pediatr Pulmonol 1989;83:507-12. 1995;20:372-9. [11] Teeter JG, Bleecker ER. Relationship between airway obstruction and [26] Emerman CL, Cydulka RK. Use of peak expiratory flow rate in respiratory symptoms in adult asthmatics. Chest 1998;113:272-7. emergency department evaluation of acute exacerbation of chronic [12] Burdon JG, Juniper EF, Killian KJ, et al. The perception of obstructive pulmonary disease. Ann Emerg Med 1996;27:159-63. breathlessness in asthma. Am Rev Respir Dis 1982;126:825-8. [27] Dhuper S, Maggiore D, Chung V, et al. Profile of near-fatal asthma in [13] Rubinfeld AR, Pain MC. Perception of asthma. Lancet 1976;1:882-4. an inner-city hospital. Chest 2003;124:1880-4. [14] Spiteri MA, Cook DG, Clarke SW. Reliability of eliciting physical [28] Wisnivesky JP, Leventhal H, Halm EA. Predictors of asthma-related signs in examination of the chest. Lancet 1988;1:873-5. health care utilization and quality of life among inner-city patients [15] Holleman Jr DR, Simel DL. Does the clinical examination predict with asthma. J Allergy Clin Immunol 2005;116:636-42. airflow limitation? JAMA 1995;273:313-9. [29] Jardin F, Farcot JC, Boisante L, et al. Mechanism of paradoxic pulse [16] Miller A. Lung function testing: selection of reference values and in bronchial asthma. Circulation 1982;66:887-94. interpretative strategies. Am Rev Respir Dis 1992;146:1368-9. [30] Wright RO, Steele DW, Santucci KA, et al. Continuous, noninvasive [17] American Thoracic Society. Standardization of spirometry, 1994 measurement of pulsus paradoxus in patients with acute asthma. Arch update. Am J Respir Crit Care Med 1995;152:1107-36. Pediatr Adolesc Med 1996;150:914-8. [18] Kerem E, Canny G, Tibshirani R, et al. Clinical-physiologic [31] Rebuck AS, Pengelly LD. Development of pulsus paradoxus in the correlations in acute asthma of childhood. Pediatrics 1991;87:481-6. presence of airways obstruction. N Engl J Med 1973;288:66-9. [19] Togias A, Horowitz E, Joyner D, et al. Evaluating the factors that relate [32] Jay GD, Onuma K, Davis R, et al. Analysis of physician ability in the to asthma severity in adolescents. Int Arch Allergy Immunol 1997;113: measurement of pulsus paradoxus by sphygmomanometry. Chest 87-95. 2000;118:348-52. [20] American Thoracic Society. Lung function testing: selection of refe- [33] Gravlee GP, Brockschmidt JK. Accuracy of four indirect methods of rence values and interpretative strategies. Am Rev Respir Dis 1991; blood pressure measurement, with hemodynamic correlations. J Clin 144:1202-18. Monit 1990;6:284-98. American Journal of Emergency Medicine (2007) 25, 430–436

www.elsevier.com/locate/ajem

Brief Report Splenic artery aneurysms encountered in the ED: 10 years’ experience

Chu-Feng Liu MDa, Chia-Te Kung MDa, Ber-Ming Liu MDa, Shu-Hang Ng MDb, Chung-Cheng Huang MDb, Sheung-Fat Ko MDb,* aDepartment of Emergency Medicine, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University, College of Medicine, Kaohsiung 833, Taiwan bDepartment of Radiology, Chang Gung Memorial Hospital-Kaohsiung Medical Center, Chang Gung University, College of Medicine, Kaohsiung 833, Taiwan

Received 26 July 2006; revised 9 August 2006; accepted 13 August 2006

Abstract Objective: Our objective was to report 7 cases of splenic artery aneurysm (SAA) encountered in the emergency department (ED). Methods: A retrospective survey of our ED database revealed 7 cases of SAA (6 men, 1 woman; mean age, 56 years) of 651347 ED visits over the last decade. Their clinical and imaging features, management, and outcomes were evaluated. Results: Splenic artery aneurysm in the ED was rare (prevalence, 0.011%). Common presentations included acute abdomen (n = 5) and shock (n = 2). Five cases had liver cirrhosis and portal hypertension. Abdominal radiographs (n = 7) revealed 2 atherosclerotic patients with SAA. Abdominal computed tomography (n = 7) depicted all SAAs (size, 1.5-8 cm; mean, 3.8 cm). Four ruptured SAAs were successfully managed with coils embolization. Among them, 1 patient with ruptured mycotic SAA also received surgery, but the patient died of Klebsiella sepsis 3 months later. Conclusions: In the ED, ruptured SAA should be included as a rare differential consideration of acute abdomen, especially in middle-aged men with liver cirrhosis and portal hypertension. Although SAA may be an unexpected computed tomographic finding, once diagnosed, endovascular treatment is recommended. D 2007 Elsevier Inc. All rights reserved.

1. Introduction autopsy [1]. Rupture of SAA may lead to catastrophic hemorrhage. Approximately 10% of SAAs are ruptured at Splenic artery aneurysm (SAA) is uncommon, although the time of diagnosis, with a 75% mortality [2,3]. Modern it is the third most common intra-abdominal aneurysm, imaging techniques allow early detection of asymptomatic following abdominal aorta aneurysm and iliac artery SAA [1,4,5]. In addition to conventional surgery, alternative aneurysm. In one study, the prevalence of SAA was 0.8% therapeutic options include endovascular treatment and in unselective visceral angiograms and 0.04% to 0.1% at laparoscopic repair [1,5-10]. However, SAA encountered in the emergency department (ED) has not been specifically * Corresponding author. addressed. The objective of this study was to present our

Table 1 Summary of 7 patients with SAA Case no. Age Sex Initial presentations Associated SAA Location Size Diagnosis Management Outcome Follow-up conditions rupture (cm) 1 77 M Acute abdomen, Gallstone, DM, + Middle 2.5 AR, CT, AG TAE, cholecystectomy, Died of RP, fever, mycotic aneurysm splenectomy, Klebsiella shock after 4 h aneurysmectomy pneumonia after 3 mo 2 55 M Acute abdomen, RP HTN (180/ + Middle 5.8 AR, CT, AG TAE Alive 18 mo 100mm Hg) 3 54 M Acute abdomen, RP, LC with PH + Distal 1.5 AR, CT, AG TAE Alive 4 y shock after 3 h 4 41 M Acute abdomen, RP LC with PH, + Distal 8 AR, US, CT, AG TAE Alive 6 y alcoholism, prior pancreatitis 5 39 M Acute abdomen, RP LC with PH À Distal 3.6 AR, US, CT Conservative Alive 13 m 6 72 F Abdominal fullness LC with PH, HTN À Proximal 2.2 AR, CT, AG Conservative Alive 9 y (190/98 mm Hg) ASA 7 54 M Upper gastrointestinal LC with PH, DU À Distal 3 AR, CT, Conservative Died of bleeding endoscopy DU perforation after 1 y AG indicates angiography; AR, abdominal radiographs; ASA, aberrant splenic artery; DM, diabetes mellitus; DU, duodenal ulcer; HTN, essential hypertension; LC, liver cirrhosis; PH, portal hypertension; RP, rebound pain; US, ultrasonography. 432 C.-F. Liu et al.

CT. Transarterial embolization was done using coils with or without gelfoam cubes. Technical success was defined as complete exclusion of the aneurysm on controlled arterio- gram after the procedure. One patient underwent surgery immediately after TAE. A follow-up CT was performed 5 days, 3 months, and 9 months after TAE.

3. Results

The clinical and imaging findings of the patients are summarized in Table 1. 3.1. Clinical features There were 6 men and 1 woman in our study, ranging in age from 39 to 77 years (mean, 56 years). Of the 7 patients, 5 presented with an acute abdomen with sudden-onset diffuse abdominal or epigastric tenderness with rebound pain; patient 1 also had fever and subsequently developed shock 4 hours after arrival in the ED, and patient 3 was hemodynamically stable at admission but suddenly developed shock 3 hours later. One patient presented with abdominal fullness and another with gastrointestinal bleeding. Medical histories revealed liver cirrhosis with portal hypertension in 5 patients, essential hypertension in 2, diabetes mellitus and gallstones in 1, duodenal ulcer in 1, and alcoholism with a prior history of pancreatitis in 1. Laboratory studies revealed elevated white blood cell counts in 4 patients, mildly elevated levels of serum aspartate and alanine aminotransferase in 5 patients, and mildly decreased hemoglobin levels in 5 patients. 3.2. Imaging features Abdominal radiographs (n = 7) showed splenomegaly with a calcified-rim nodule in the splenic hilum in patient 7 Fig. 1 Patient 6. A, Abdominal radiograph shows a nodule with and a calcified-rim nodule in the left L1 through L2 a calcified rim (arrows) over the left L1 through L2 paraspinal region. B, Digital subtraction angiogram shows an SAA at an paraspinal region in patient 6, suggesting possible calcified aberrant splenic artery (arrows) originating from the anomalous SAAs (Fig. 1). Two patients underwent transabdominal splenomesenteric trunk. experience with 7 cases of SAA encountered in the ED, with an emphasis on patient characteristics, clinical presentations, and diagnostic and therapeutic considerations.

2. Materials and methods

From March 1995 to March 2006, a total of 7 cases of SAA were found after a retrospective investigation of the ED database. The medical records of the patients were reviewed for clinical manifestations, known prior diseases, pertinent laboratory data, and outcomes. All patients had plain abdominal radiographs and abdominal computed tomography (CT). Two of them also received sonographic evaluation. Fig. 2 Patient 4. Transabdominal sonogram reveals a cystic Emergent transarterial embolization (TAE) was per- lesion (arrows) in the splenic hilar region. Doppler study revealed formed in 4 patients with ruptured SAAs diagnosed by artery pulsations of the lesion, compatible with SAA. Splenic artery aneurysms encountered in the ED 433

aneurysmectomy was performed immediately after TAE. However, the patient died 3 months later of Klebsiella pneumonia and sepsis. After TAE, patients 2, 3, and 4 showed mildly elevated levels of serum amylase (186- 214 U/L; mean, 200 U/L; reference range, 27-137 U/L) and serum lipase (196-257 U/L; mean, 227 U/L; reference range, b190 U/L), which normalized after 1 to 2 weeks. They were alive and well at 18 months, 4 years, and 6 years after the procedure, respectively. Of our 7 patients, 3 received conservation treatment. Patient 5 had a 3.6-cm SAA in the distal portion of the splenic artery; he refused surgery and remained alive and well at the 13-month follow- up. In patients 6 and 7, the SAAs had calcified walls and were probably atherosclerotic. Abdominal angiographic evaluation of patient 6 showed an atherosclerotic SAA in the proximal part of an aberrant splenic artery originating from the splenomesenteric trunk (Fig. 1), and this aberrant Fig. 3 Patient 1. Abdominal CT reveals a mycotic SAA (arrows) SAA remained stable at the 9-year follow-up. Patient 7 died with mottled gas and a disrupted calcified rim in the middle part of of duodenal ulcer perforation and multiple-organ failure the splenic artery, surrounded by inflammatory strands. 1 year after treatment. sonography, which revealed SAA in patient 4 (Fig. 2) but missed the diagnosis in patient 5. All 7 patients underwent 4. Discussion abdominal CT, which consistently revealed SAA. The lesion was located in the distal part of the splenic artery in Splenic artery aneurysms are more common in women 4 patients, the middle part in 2, and the proximal part in 1. than men, with a 4:1 female-to-male ratio, and commonly The aneurysm sizes ranged from 1.5 to 8 cm, with a mean of affect multiparous women during pregnancy [2-4,11]. 3.8 cm. Two SAAs had calcified rims. Four SAAs were Approximately 80% to 95% of SAAs are asymptomatic ruptured focally with irregular arterial outpouching from the and are incidentally found during evaluation of unrelated anerusymal sac, contrast-medium leakage, perisplenic he- symptoms [1,3,6-8,11]. However, SAAs may rupture, matoma, and regional fluid collection and/or ascites. The resulting in severe abdominal pain or even hypovolemic sizes varied (1.5, 2.5, 5.8, and 8 cm; mean, 4.45 cm). In shock. In these circumstances, the patient may be brought to patient 1, who had diabetes, perianeursymal gas with the ED for management. Nevertheless, as shown by our adjacent inflammatory strands was demonstrated, compati- experience, an SAA is indeed rare in the ED, with a ble with mycotic aneurysm (Fig. 3). In patient 3, the SAA prevalence of approximately 0.011%. Among 100 cases of was small and was initially overlooked; the images were reviewed, and SAA was identified retrospectively (Fig. 4). Liver cirrhosis and portal hypertension (engorged portal vein, presence of varices and/or splenomegaly) were demonstrated in 5 patients. Gallstones were demonstrated in 1 patient. 3.3. Treatment outcome and follow-up All 4 ruptured SAAs were managed with emergent TAE. In patients 1 and 2, the ruptured SAAs were located in the middle portion of the splenic artery and were embolized with the sandwich method by applying coils 1 to 2 cm upstream and downstream of the aneurysm; gelfoam cubes were also added in patient 2 for increased packing of the SAA (Fig. 5). In patients 3 and 4, the SAA s were located in the distal portion of the splenic artery and were successfully occluded with coil packing at the distal supplying artery; gelfoam cubes were also applied in patient 3. Immediate follow-up splenic arteriograms revealed technical success in Fig. 4 Patient 3. Abdominal CT shows a small ruptured SAA all 4 patients. Because patient 1 had a mycotic aneurysm, (arrow), which was initially overlooked, at the distal part of the surgical treatment with a cholecstectomy, splenectomy, and splenic artery. 434 C.-F. Liu et al.

Fig. 5 Patient 2. A, Abdominal CT reveals a ruptured SAA (arrows) at the middle part of the splenic artery. B, Digital subtraction angiogram confirms the ruptured SAA (arrows). C, Follow-up angiogram after coil embolization shows complete obliteration of the SAA (arrows). D, Follow-up CT angiogram 9 months after embolization shows persistent obliteration of the SAA (arrows) and collateral reconstitution of the distal part of the splenic artery (open arrows).

SAA reported by Trastek et al [3], 80 were women, but only The pathogenesis of SAA is not well understood. 3 had acute rupture. In contrast, 6 of 7 ED-encountered Atherosclerosis, essential hypertension, trauma, septic em- SAAs in our series affected middle-aged men who bolism, pancreatitis, liver disease, and portal hypertension commonly presented with acute abdomen, and there was a have been reported as risk factors for SAA [4-8,11,12]. surprisingly high rupture rate of 59% (4/7 cases). Notewor- Essential hypertension seems to play a role in atherosclerotic thily, as in 2 of our cases, there is a so-called double rupture weakening of the splenic artery wall and formation of an phenomenon, which may initially occur as bleeding aneurysm. Atherosclerotic changes are observed in up to 99% confined to the lesser sac, with the patient in a transient of the SAAs examined histologically but are most likely hemodynamically stable status and decrease clinical alert- secondary to medial degeneration [3]. In our series, essential ness. This is followed by an unpredictable onset of shock hypertension associated with an atherosclerotic SAA could due to subsequent intraperitoneal SAA rupture, which can only be identified in 1 case. In pregnant women, etiologic be fatal [1,5]. factors including aorta compression by the uterus with portal Splenic artery aneurysms encountered in the ED 435 congestion, hormonal changes with vascular intimal hyper- SAA rarely exceeds 3 cm in diameter, although giant SAAs plasia, and fragmentation of the internal elastic membrane up to 30 cm have been documented [1-12,16]. The risk factors have been described [1-3]. However, in our hospital, we had for SAA rupture are difficult to assess, and various factors no pregnant women with SAA among more than 650000 ED including the presence of calcification, patient’s age, and admissions and approximately 2000 deliveries each year aneurysmal size have been evaluated, but no definitive throughout the last 20 years. Dave et al [5] also reported that relation to aneurismal rupture can be established no SAA occurrence associated with pregnancy was noted in [1,4,5,7,8]. Lee et al [12] reported that ruptured SAAs in the 1990s, despite 5000 deliveries performed in the Mount patients with portal hypertension were larger than in those Sinai–New York University Hospital each year. Although without portal hypertension (5.5 vs 5.1 cm). In our series, the liver disease and portal hypertension are reported to be average size of the ruptured SAAs was 4.45 cm. Although a associated with only 20% of all SAA ruptures [3,11], patients ruptured SAAwith a diameter of less than 2.5 cm is rare [3,6], with aneurysmal rupture who have portal hypertension have a the smallest ruptured SAA in this series was only 1.5 cm. higher mortality rate than those without this condition (56% Regardless of how SAA is discovered, surgical or vs 17%) [12]. Among Asia Pacific countries, our country is endovascular treatment is advocated in patients with SAA- one of the regions with high prevalence rates of hepatitis and related symptoms, patients with expanding SAA, women liver cirrhosis, and this may be the reason that most SAA who are pregnant or expecting to be pregnant, in liver cases in this series were associated with liver cirrhosis with transplant candidates, and patients with SAA measuring portal hypertension. 2.5 cm or greater [1-7]. Classically, SAA can be surgically Various imaging modalities including plain radiographs, repaired or directly ligated with or without splenectomy or a sonography, CT, and angiography have been used to partial pancreatectomy [6,7,12]. Elective repair of SAA is identify SAA [1,4,5,12-14]. Radiographically, as in our safe, with a 0% to 1.3% mortality. However, emergent repair study, only 2 SAAs were identified as calcified-rim nodule of ruptured SAA is associated with mortality of up to 40% in the left upper abdomen [13]. On sonograms, SAA [1,3,5,7,12]. Laparoscopic SAA repair has also been typically manifests as a left hypochrondrial pulsatile cystic described as a minimally invasive method that may be lesion if it is not obliterated by overlying stomach gas preferable for selected patients in whom open surgery [14]. Magnetic resonance angiography plays a role in the entails prohibitively high risks [9,10]. However, the clinical diagnosis of SAA in liver transplantation but may not feasibility of laparoscopic surgery of a ruptured SAA in the always be available for ED patients [4]. In our experience, ED setting has not been validated. CT is an excellent tool for diagnosing SAA. Besides The application of TAE for SAA as a treatment alternative assessing the location and size of the SAA, it also reveals has been substantiated by low morbidity rates ranging from ruptured aneurysms, intra-abdominal hemorrhage, and 14% to 25% and high success rates ranging from 75% to associated underlying diseases. As in patient 1, CT was 100%. Some studies have even advocated TAE as the capable of revealing a clinically occult mycotic SAA. treatment of choice for all visceral aneurysms [1,8,17,18]. However, SAAs may be small and may be overlooked if Some TAE-related complications of SAA have been de- radiologists and ED physicians do not keep this condition scribed [1,5-8], including splenic infarction, inadvertent in mind, as occurred in patient 4. On the other hand, in embolism to other visceral arteries, abscess formation, the ED, SAA rupture with intra-abdominal hemorrhage arterial disruption, and access site hematoma. Saltzberg et al may be an incidental or unexpected finding on CT, as a [7] reported that major complications developed in 4 of part of the workup of abdominal pain. When CT reveals 11 patients with distal SAAs treated with TAE, but those intra-abdominal hemorrhage, the bleeding site or origin of TAEs were performed with coil plus adjunctive n-butyl-2- active extravasation of the contrast material should cyanoacrylate (NBCA) adhesive. Because NBCA is a liquid cautiously be identified. The differential diagnoses of adhesive, embolization using NBCA will lead to permanent intra-abdominal hemorrhage include traumatic solid vis- occlusion of the distal-end arteries, and thus, severe splenic ceral or mesenteric injuries, intra-abdominal tumor bleed- infarcts and pancreatitis are probable. In our study, the TAEs ing (most commonly hepatocellular carcinoma rupture in in all 4 patients with ruptured SAAs (coil embolization in 2, our country), abdominal aortic aneurysm rupture, and, coil and minimal use of gelfoam cubes in 2) were technically uncommonly, visceral aneurysm rupture. Although rare, successful. Except for patient 1 who had a mycotic SAA and awareness of the possibility of SAA rupture in patients in whom subsequent surgery was needed to remove a septic with cirrhosis with an acute abdomen and detailed scrutiny focus and for complete treatment, the other 3 patients had of CT findings in the splenic artery are important for good recoveries with only minimal transient elevation of establishing the correct diagnosis. pancreatic enzymes and no recurrence in long-term follow- Consistent with prior reports [1-12], in 4 of 7 cases up. In our experience, TAE is an effective method for treating encountered in the ED, the SAA originated from the distal most of the SAAs, even if the patient is hemodynamically splenic artery, which is the most frequently affected site. Only unstable. However, as in patient 1, who had a mycotic SAA, 1 SAA was found in the proximal part of an aberrant splenic urgent referral to a vascular surgeon after TAE is mandatory. artery, originating from the splenomesenteric trunk [15].An In addition, in SAA not amendable by TAE due to complex 436 C.-F. Liu et al. anatomy or technical failure even after repeated TAE, referral [7] Saltzberg SS, Maldonado TS, Lamparello PJ, et al. Is endovascular for surgical repair is warranted [1,7]. therapy the preferred treatment for all visceral artery aneurysms? In summary, this case series highlights that SAA should Ann Vasc Surg 2005;19:507-15. [8] Salam TA, Lumsden AB, Martin LG, et al. Nonoperative management be included as a rare differential consideration of acute of visceral aneurysms and pseudoaneurysms. Am J Surg 1992;164: abdomen, especially in middle-aged men with liver cirrhosis 215-9. and portal hypertension. Although SAA rupture may be an [9] Matsumoto K, Ohgami M, Shirasugi N, et al. A first case report of the unexpected CT finding as part of the workup of abdominal successful laparoscopic repair of a splenic artery aneurysm. Surgery pain in the ED, once diagnosed, timely TAE is usually 1997;121:462-4. [10] Arca MJ, Gagner M, Heniford BT, et al. Splenic artery aneurysms: effective. If SAA is not amendable by endovascular methods of laparoscopic repair. J Vasc Surg 1999;30:184-8. treatment, surgical repair is necessitated. [11] Stanley JC, Fry WJ. Pathogenesis and clinical significance of splenic artery aneurysms. Surgery 1974;76:898-909. [12] Lee PC, Rhee RY, Gordon RY, et al. Management of splenic artery References aneurysms: the significance of portal and essential hypertension. J Am Coll Surg 1999;189:483-90. [1] Guillon R, Garcier JM, Abergel A, et al. Management of splenic artery [13] Chou YH, Tiu CM, Pan HB, et al. Diagnosis of splenic artery aneurysm aneurysms and false aneurysms with endovascular treatment in by Doppler and real-time sonography. J Formos Med Assoc 1985; 12 patients. Cardiovasc Intervent Radiol 2003;26:256-60. 84:747-50. [2] Vassalotti SB, Scahaller JA. Spontaneous rupture of splenic artery [14] Mattar SG, Lumsden AB. The management of splenic artery aneurysm in pregnancy. Obstet Gynec 1967;30:264-8. aneurysms: experience with 23 cases. Am J Surg 1995;169:580-4. [3] Trastek VF, Pairolero PC, Joyce JW, et al. Splenic artery aneurysm. [15] Settembrini PG, Jausseran JM, Roveri S, et al. Aneurysms of Surgery 1982;91:694-9. anomalous splenomesenteric trunk: clinical features and surgical [4] Heestand G, Sher L, Lightfoote J, et al. Characteristics and treatment in two cases. J Vas Surg 1996;24:687-92. management of splenic artery aneurysm in liver transplant candidates [16] Bornet P, Medjoubi SA, Tissot A, et al. Giant aneurysm of the splenic and recipients. Am Surg 2003;69:933-40. artery—a case report. Angiology 2000;51:343-7. [5] Dave SP, Reis ED, Hossain A, et al. Splenic artery aneurysm in the [17] Gabelmann A, Gorich J, Merkle EM. Endovascular treatment of 1990s. Ann Vasc Surg 2000;14:223-9. visceral artery aneurysms. J Endovasc Ther 2002;9:38-47. [6] Sessa C, Tinelli G, Procu P, et al. Treatment of visceral artery [18] DiMuzio P, Mandel E, Sullivan K. Transcatheter embolization: an aneurysms: description of a retrospective series of 42 aneurysms in alternative treatment for splenic artery aneurysms. Contemp Surg 34 patients. Ann Vasc Surg 2004;18:695-703. 2002;58:617-20. American Journal of Emergency Medicine (2007) 25, 437–441

www.elsevier.com/locate/ajem

Brief Report Utility of impedance cardiography for dyspneic patients in the ED

Hsiang-Yun Lo MD, Shu-Chen Liao MD, Chip-Jin Ng MD, Jen-Tse Kuan MD, Jih-Chang Chen MD, Te-Fa Chiu MD*

Department of Emergent Medicine, Chang Gung Memorial Hospital, Linko Medical Center

Received 11 August 2006; revised 11 September 2006; accepted 1 October 2006

Abstract Background: Dyspnea is one of the most common emergency department (ED) symptoms, but early diagnosis and treatment are challenging because of multiple potential causes. Hemodynamic parameters may aid in the evaluation of dyspnea, but are difficult to assess. Impedance cardiography is a noninvasive hemodynamic measurement method that may assist in early ED decision making. Methods: This study is intended to determine the accuracy in differentiating cardiac from noncardiac causes of dyspnea using impedance cardiography–derived hemodynamic parameters compared to ED physician opinion in light of initial history, and physical and laboratory tests. The final diagnosis, made after patient hospital record review, was compared with ED physician and impedance cardiography diagnoses. Results: A total of 52 patients were included: 14 women and 38 men, aged 68.5 F 14.2 years. There were significant differences in values of stroke index (25.7 vs 32.9, P b .05), cardiac index (2.3 vs 3.1, P b .0001), velocity index (35.1 vs 53.2, P b .01), and systolic time ratio (0.55 vs 0.44, P b .05) between the cardiac and noncardiac groups, respectively. Impedance cardiography measurements demonstrated better sensitivity (75% vs 60%), specificity (88% vs 66%), and positive and negative predictive values (79% vs 52% and 85% vs 72%, respectively) compared with those of the ED physician in distinguishing cardiac from noncardiac causes of dyspnea. Conclusion: Impedance cardiography data result in improvement in ED physician differentiation of cardiac from noncardiac causes of dyspnea. D 2007 Elsevier Inc. All rights reserved.

1. Introduction causes, including heart disease, obstructive disease of the airways, diffuse parenchymal lung disease, and pulmonary Dyspnea is one of the most common complaints seen in embolism may occur alone or in combination [1]. Evalua- the emergency department (ED). Multiple confounding tion is often challenging, especially when such patients have a history of both cardiac and pulmonary disorders and early diagnosis is required in the ED. Accurate history, physical * Corresponding author. Department of Emergent Medicine, Linko Chang Gung Memorial Hospital, Taoyuan, Taiwan, ROC. Tel.: +886 3 examination, oxygen saturation by oximeter, arterial blood 3281 200x2140; fax: +886 3 328 7715. gases, and chest x-rays are all helpful for diagnosis, but E-mail address: [email protected] (T.-F. Chiu). none is useful in identifying patients with depressed

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2006.10.009 438 H.-Y. Lo et al. ventricular performance usually present in patients with 2.4. Study protocol cardiac dyspnea. Studies have shown that clinician estima- tion of cardiac output (CO) and systemic vascular resistance All patients received a history and physical examination (SVR) shows poor correlation to measured CO and SVR from an ED physician, as well as the following standard values [2]. The most commonly used hemodynamic tests: electrolyte, kidney and liver function, complete blood measure is obtained by the indicator-dilution technique of count, electrocardiogram, chest radiography, and arterial thermodilution, which uses pulmonary artery catheterization blood gas. When deemed medically necessary, an echocardiogram was performed. (PAC) [3]. However, PAC is not a standard procedure for The treating physicians were blind to noninvasive the management of such patients in the ED because of risks hemodynamic monitoring by ICG, and all ICG data were and costs [4-6]. Therefore, rapid and objective hemody- collected before pharmacologic agents were administered by namic measurements in the ED may be valuable in clinical a member of the research staff not involved in patient care. decision making. After hospital discharge, each patient’s medical record Impedance cardiography (ICG) is a valid, noninvasive was reviewed by a board-certified ED physician blind to method that measures beat-to-beat changes of thoracic ICG results and not involved in the treatment of any study bioimpedance via 4 dual sensors applied on the neck and patients. Using all medical record data, this physician thorax to calculate hemodynamic parameters [7-10]. Previ- determined a final hospital diagnosis, including whether ous ICG studies have demonstrated its accuracy and utility the dyspnea was cardiac or noncardiac in origin. in differential diagnosis of dyspnea [11-13], in prediction of hospital charges and length of stay [14], and its ability to 2.5. Data analysis change real-time diagnosis and treatment decisions in dyspneic patients [15]. Impedance cardiography data were evaluated retrospectively, and criteria for ICG-derived cardiac causes of dyspnea were defined as either a cardiac index of 2.4 or lower or a systolic time ratio (STR) of 0.55 or higher, 2. Methods concurrent with a cardiac index of less than 3.0. Continuous variables were expressed as mean F SD. t Tests were 2.1. Experiment design performed to determine the statistical difference between A prospective blind study was conducted to compare ICG results in differentiating cardiac from noncardiac causes of dyspnea to ED physician diagnoses. 2.2. Impedance cardiography monitor (BioZ; CardioDynamics, San Diego, Calif) Disposable sensors transmit a small electrical signal through thorax impedance (resistance). The electrical signal is measured and displayed as an ICG waveform. As the volume and velocity of blood in the aorta change with each heartbeat, DISQ (Digital Impedance Signal Quantifier) technology processes the impedance changes. The impedance changes are inputted to the innovative Z MARC (Modulating Aortic Compliance) algorithm to provide hemodynamic parameters such as CO, stroke volume, SVR, contractility, and fluid status. These parameters have an average of 30 regular ICG beats. 2.3. Setting and population A convenience sample of patients at the Linko Chang Gung Memorial Hospital ED was enrolled. Emergency department patients were included if they met 1 or more of Fig. 1 Trial population profile. Asterisk indicates lung cancer, the following criteria: complaints of shortness of breath, hepatic cell carcinoma with massive pleural effusion, neuromus- respiratory rate of more than 20 breaths/min, and hypoxemia cular disease, and anxiety state; AAD, against advised discharge; (arterial oxygen concentration of b90% on room air). Dx, diagnosis; CHF, congestive heart failure; ACS, acute coronary Patients were excluded if they were younger than 18 years, syndrome; Pn, pneumonia; UARS, upper airway resistance identified as trauma patients, or pregnant. syndrome; COPD, chronic obstructive pulmonary disease. Utility of impedance cardiography for dyspneic patients in the ED 439

Table 1 Summary of hemodynamic statistics (N = 52) Characteristic Cardiac final Dx (n = 20) Noncardiac final Dx (n = 32) P SI (mL/m2) 25.7 (21.5-30.0) 32.9 (29.1-36.7) b.05 CI (L/[min d m2]) 2.31 (2.06-2.56) 3.1 (2.8-3.3) b.001 VI (/1000/s) 35.1 (28.3-41.9) 53.2 (44.8-61.6) b.01 STR 0.55 (0.46-0.63) 0.44 (0.38-0.50) b.05 Values are expressed as mean (95% confidence interval). Dx indicates diagnosis; SI, stroke index; CI, cardiac index; VI, velocity index. cardiac and noncardiac causes of dyspnea ( P b .05). For disease can exceed 90% [16]. It can also be used to identify each patient, the final diagnosis was compared with ICG- cardiac causes of dyspnea. However, its cost and complexity derived and ED physician diagnoses. Sensitivity, specificity, reduces its ED availability. Pulmonary artery catheterization positive predictive value, negative predictive value, and is the most commonly used method to obtain hemodynamic diagnostic accuracy were calculated. data [3], but the risks are well documented, including infection, sepsis, and arrhythmia, as well as increased morbidity and mortality [4-6]. Therefore, most hospitals 3. Results reserve this technique for only the most critically ill patients in the intensive care unit. Clearly, there is a need for a low- Sixty patients were enrolled in the study. Eight were cost, accurate, and noninvasive alternative. In this study, excluded because of a lack of definite diagnosis and advice ICG is the alternative. against discharge. There were 14 women and 38 men. The The latest ICG technology for determining cardiac F average age was 68.5 14.2 years. Fourteen patients had a performance is as proven as PAC in continuous monitor- history of congestive heart failure, 16 had chronic lung ing [7,8,10,17-20]. Our goal was to ascertain whether disease, and 6 had both. Twenty patients had a final ICG-derived hemodynamic data could be used to distin- diagnosis of cardiac-caused dyspnea, and what the other guish between cardiac and noncardiac dyspnea in an adult 32 had were noncardiac (Fig. 1). population in the ED. In this study, ICG was able to Those with cardiac-caused dyspnea had significantly accurately differentiate between cardiac- and non–cardiac- b ( P .05) lower stroke, cardiac, and velocity indices, with related causes of dyspnea with better sensitivity, specific- higher STRs and SVR. Hemodynamic values for cardiac ity, and positive and negative predictive values than ED and noncardiac groups are shown in Table 1. physicians using conventional methods. However, our Compared with the final diagnoses, the overall diagnostic study showed lower sensitivity than some previous studies accuracy for ED physicians was 69% (36/52) vs 83% (43/ [11,13]. Of the 5 patients misdiagnosed with noncardiac 52) for ICG. Emergency department physicians diagnosed dyspnea, 3 patients had acute myocardial infarction, and 13 of 20 patients correctly with a final diagnosis of cardiac- the ejection fraction (EF) for 2 of them were 55% and caused dyspnea, and 23 of 32 for noncardiac-caused 71% by echocardiography. Another had pericarditis (EF, dyspnea. If the ED physician diagnosed both cardiac and 62%), and the last had acute pulmonary edema (EF, noncardiac causes in the same patient, we judged the 18%). For the first 4 patients, we considered that the favored diagnosis by ED physicians according to the dyspnea symptoms were induced by pain rather than by treatment at the ED. Impedance cardiography correctly cardiac causes directly. Furthermore, ICG was able to diagnosed 15 of 20 patients with cardiac cause, and 28 of 32 with noncardiac cause. Impedance cardiography demon- differentiate between cardiac- and non–cardiac-related strated superior sensitivity, specificity, positive predictive dyspnea in all patients with histories of both cardiac value, and negative predictive value over ED physicians in and pulmonary disease in this study. the final diagnosis of cardiac vs noncardiac causes of McCullough et al [21] reported that ED physician dyspnea (Table 2). accuracy in diagnosing cardiac vs noncardiac causes of dyspnea was 74% in a study on 1586 patients. Although our patient number was significantly smaller, ED physician 4. Discussion Table 2 Summary diagnosis statistics (N = 52) Although today’s ED physicians have a variety of Method Sensitivity Specificity PPV NPV laboratory and diagnostic tools available, an accurate initial (%) (%) (%) (%) determination of the underlying cause of dyspnea remains ICG 75 88 79 85 challenging. Echocardiography is the most common diag- ED physician 60 66 52 72 nostic test to determine left ventricular dysfunction, and PPV indicates positive predictive value; NPV, negative predictive value. under optimal conditions its accuracy for detecting heart 440 H.-Y. Lo et al. accuracy was similar (69%). B-natriuretic peptide (BNP) rationale for considering ICG as an important and emerging and N-terminal pro-BNP are cardiac neurohormones secret- tool in dyspnea diagnosis and treatment. ed from the ventricles in response to left ventricular volume expansion and pressure overload. Levels of BNP are known to be elevated in patients with left ventricular dysfunction and be correlated with echocardiography findings [22,23]. References McCullough et al [21] also sought to determine if overall [1] Scano G, Ambrosion N. Pathophysiology of dyspnea. Lung 2002; diagnostic accuracy could be improved by adding a 180:131. screening test for BNP. The addition of BNP testing [2] Eisenberg PR, Jaffee AS, Schuster DP. Clinical evaluation compared increased overall diagnostic accuracy to 81.5%, but it was to pulmonary artery catheterization in the hemodynamic assessment of also concluded that BNP testing had little impact on medical critically ill patients. Crit Care Med 1984;12:549-53. decision making for patients already given a primary [3] Swan HJC, Ganz W, Forrester J, et al. Catheterization of the heart in man with the use of a flow-directed balloon-tipped catheter. N Engl J diagnosis of heart failure. In our study, ICG testing had Med 1970;283:447-51. 83% diagnostic accuracy. [4] Connors Jr AF, Speroff T, Dawson NV, et al. The effectiveness of right Impedance cardiography information can be obtained in heart catheterization in the initial care of critically ill patients. JAMA 3 to 5 minutes and does not require blood draw or waiting 1996;276:889-97. [5] Mermel LA, McCormick RD, Springman SR, et al. The pathogenesis time for laboratory analysis. Furthermore, ICG measures and epidemiology of catheter-related infection with pulmonary artery have been shown to be available in real time and be reliable Swan-Ganz catheters: a prospective study utilizing molecular subtyp- in reflecting various aspects of cardiac function, including ing. Am J Med 1991;91:197S-205S. blood flow (stroke volume and CO), contractility (velocity [6] Guyatt G. A randomized controlled trail of right heart catheterization index and STR), and changes in fluid status (thoracic fluid on critically ill patients: Ontario Intensive Care Study Group. J Intensive Care Med 1991;6:91-5. content) [24]. Results from the Emergency Department [7] Sageman WS, Riffenburgh RH, Spiess BD. Equivalence of bioimpe- IMPedance Cardiography–aided Assessment Changes Ther- dance and thermodilution in measuring cardiac index after cardiac apy trial demonstrated that rapid provision of ICG hemo- surgery. J Cardiothorac Vasc Anesth 2002;16:8-14. dynamic information resulted in a treatment plan change in [8] Drazner M, Thompson B, Rosenberg P, et al. Comparison of impedance cardiography with invasive hemodynamic measurements 24% of dyspneic patients [15]. In patients with severe sepsis in patients with heart failure secondary to ischemic or nonischemic and septic shock, for example, the early hemodynamic cardiomyopathy. Am J Cardiol 2002;89:993-5. assessment on the basis of physical findings, vital signs, [9] Van De Water JM, Miller TW. Impedance cardiography: the next vital central venous pressure, and urinary outputs fail to detect sign technology. Chest 2003;123:2028-33. persistent global tissue hypoxia. A more definite resuscita- [10] Albert N, Hail M, Li J, et al. Equivalence of bioimpedance and TD in measuring CO/CI in patients with advanced, decompensated chronic tion strategy involves goal-oriented manipulation of cardiac heart failure hospitalized in critical care. J Am Coll Cardiol 2003; preload, afterload, and contractility to achieve a balance 41(6 Suppl):211A. between systemic oxygen delivery and oxygen demand [11] Marrocco A, Eskin B, Nashed A, et al. Noninvasive bioimpedance [25]. Impedance cardiography data may reflect this infor- monitoring differentiates cardiogenic from pulmonary causes of acute dyspnea in the emergency department. Acad Emerg Med 1998;5: mation in real time and lead to suitable treatments. 476-7. Our study collected only 1 ICG data measurement, not [12] Han J, Lindsell C, Tsurov B, et al. The clinical utility of impedance serial ICG data measurement, for measuring intramethod cardiography in diagnosing congestive heart failure in dyspneic variability. However, we tried to expand the ICG to a single emergency department patients. Acad Emerg Med 2002;9:439-40. laboratory data in dyspneic patients, not just as a continuous [13] Springfield CF, Sebat D, Johnson S, Lengle C. Utility of impedance cardiography to determine cardiac vs. noncardiac cause of dyspnea monitoring method but in the way that previous studies have in the emergent department. Congest Heart Fail 2004;10(Suppl 2): revealed this to be highly reproducible [9,26,27]. 14-6. This study’s limitations include a small sample size and [14] Milzman D, Morrisey J, Pugh C, et al. Occult perfusion deficits in retrospective criteria for ICG diagnosis. We did not exclude heart failure patients: identification through noninvasive central a group of clinical conditions such as late stage cirrhosis, hemodynamic monitoring. Crit Care Med 1999;27:A88. [15] Peacock F, Summers R, Emerman C. Emergent dyspnea impedance hyperdynamic state, tachycardia, and cardiac dysrhythmia in cardiography–aided assessment changes therapy: the ED IMPACT which CO measurements by ICG underestimate the ther- trial. Ann Emerg Med 2003;42:S82. modilution values [28]. This study did not compare ICG to [16] Dao Q, Krishnaswamy P, Kazanegra R, Harrison A, Amirnovin R, BNP in improving physician diagnostic accuracy or Lenert L, et al. Utility of B-type natriuretic (BNP) in the diagnosis of CHF in an urgent care setting. J Am Coll Cardiol 2001;37:379-85. evaluate the ability of ICG to change medical decision. A [17] Yung GL, Fedullo PF, Kinninger K, Johnson W, Channick RN. prospective trial with a larger set of subjects is needed to Comparison of impedance cardiography to direct fick and thermodi- provide greater confidence in which hemodynamic param- lution cardiac output determination in pulmonary arterial hyperten- eters of ICG have the greatest value in assessing patients sion. Congest Heart Fail 2004;10(2 Suppl 2):7-10. with dyspnea in the ED. [18] Yung GL, Fletcher CC, Fedullo PF, Johnson FW, Kinninger K, Knowlton KU, et al. Noninvasive cardiac index using bioimpedance We conclude that ICG provided more accurate differen- in comparison to direct fick and thermodilution methods in patients tial diagnosis of dyspnea in the ED. This offers a sufficient with pulmonary hypertension. Chest 1999;116(4):281S. Utility of impedance cardiography for dyspneic patients in the ED 441

[19] Ziegler D, Grotti L, Krucke G. Comparison of cardiac output without symptoms of congestive heart failure at a veterans hospital. measurements by TEB vs. intermittent bolus thermodilution in Am J Med 2001;111:274-9. mechanical ventilated patients. Chest 1999;116(4):281. [24] Summers R, Schoemaker W, Peacock WF, et al. Bench to bedside [20] Demaria AN, Raisinghani A. Comparative overview of cardiac output series: impedance cardiography (ICG). Acad Emerg Med 2003; measurement methods: has impedance cardiography come of age? 10(6):669-80. Congest Heart Fail 2000;6(2):7-18. [25] Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, et [21] McCullough PJ, Nowak RM, McCord J, et al. B-type natriuretic al. Early goal-directed therapy in the treatment of severe sepsis and peptide and clinical judgement in emergency diagnosis of heart failure septic shock. N Engl J Med 2001;345(19):1368-77. analysis from breathing not properly (BNP) multinational study. [26] Paul EV, Cynthia AC, Segalit T. Reproducibility of noninvasive Circulation 2002;106:416-22. bioimpedance measurements of cardiac function. J Card Fail 1998; [22] Maisel AS, Koon J, Krishnaswamy P, Kazenegra R, Clopton P, 4(3 Suppl):53. Gardetto N, et al. Utility of B-natriuretic peptide as a rapid, point-if- [27] Greenberg BH, Hermann DD, Pranulis MF, Lazio L, Cloutier D. care test for screening patients undergoing echocardiography to Reproducibility of impedance cardiography hemodynamic measures determine left ventricular dysfunction. Am Heart J 2001;141:367-74. in clinically stable heart failure patients. Congest Heart Fail 2000; [23] Krishnaswamy P, Lubein E, Clopton P, Koon J, Kazanegra R, Wanner 6(2):19-26. E, et al. Utility of B-type natriuretic peptide (BNP) in elucidating left [28] Summer RL. Noninvasive hemodynamics monitoring using imped- ventricular dysfunction (systolic and diastolic) in patients with and ance cardiography. Crit Care Int 1999;9:9-12. American Journal of Emergency Medicine (2007) 25, 442–444

www.elsevier.com/locate/ajem

Brief Report Outcomes after environmental hyperthermia

Frank LoVecchioa,b,c,*, Anthony F. Pizon MDa, Christopher Berrett DOc, Adam Balls MDb aBanner Good Samaritan Regional Poison, Phoenix, AZ 85006, USA bMaricopa Medical Center, Phoenix, AZ 85308, USA cArizona College of Osteopathic Medicine, Phoenix, AZ 85308, USA

Received 21 August 2006; revised 19 November 2006; accepted 21 November 2006

Abstract Objectives: This study was conducted to describe the characteristics and outcomes of patients who presented to the emergency department (ED) with presumed environmental hyperthermia. Methods: A retrospective chart review was performed in 2 institutions with patients who were seen in the ED and had a discharge diagnosis of hyperthermia, heat stroke, heat exhaustion, or heat cramps. Exclusion criteria were an alternative diagnosis potentially explaining the hyperthermia (pneumonia, etc). Research assistants, who were blinded to the purpose of the study, performed a systematic chart review after a structured training session. If necessary, a third reviewer acted as a tiebreaker. Data regarding patient demographics, comorbidities, vital signs, laboratory results, and short-term outcome were collected. Data were analyzed with Excel and STATA software. Results: We enrolled 52 patients with a mean age of 42.6 years (range, 0.4-81 years) from August 1, 2003 to August 31, 2005. The mean high daily temperature was 103.68F (range, 88-1188F). At presentation, the mean body temperature was 105.18F (range, 100.2-111.28F) and the Glasgow Coma Scale score was less than 14 in 36 (69.2%) patients. Laboratory results demonstrated that 21 (40.4%) patients had a creatinine level of more than 1.5 mg/dL, 35 (67.3%) patients had a creatine kinase (CK) of more than 200 U/L, 30 patients (57.7%) had a prothrombin time of more than 13 seconds, 29 (55.8%) patients had an aspartate aminotransferase (AST) of more than 45 U/L, and only 3 patients (5.7%) had a glucose of less than 60 mg/dL. Ethanol or illicit drugs were involved in 18 (34.6%) cases. The mean hospital stay was 4.7 days (range, 1-30 days), and there were 15 deaths (28.8%). A kappa score for interreviewer reliability was 0.69. Major limitations were the retrospective nature and lack of homogeneity in patient evaluation and test ordering. Conclusions: Hyperthermic patients with higher initial temperatures, hypotension, or low Glasgow Coma Scale score were more likely to die. D 2007 Elsevier Inc. All rights reserved.

1. Introduction

This work was previously presented as an oral abstract at SAEM Environmental hyperthermia encompasses a broad spec- Western Conference 2005, Los Angeles, CA. trum of heat-related illness ranging from minor conditions * Corresponding author. Banner Good Samaritan and Maricopa Medical Center, Emergency Medicine, Phoenix, AZ 85006, USA. Tel.: +1 such as heat cramps or heat syncope to life-threatening heat 602 239 2358; fax +1 602 239 4138. stroke [1]. Exposure to elevated temperatures can have E-mail address: [email protected] (F. LoVecchio). deleterious effects on the human body. The body’s internal

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2006.11.026 Outcomes after environmental hyperthermia 443 mechanisms for temperature regulation may be quickly Table 1 Characteristics of environmental hyperthermia in compromised under extreme temperature environments. A 52 patients paucity of literature exists in the emergency medicine Characteristic No. (%) literature on characteristics of patients presenting with heat-related illness. Mean age (y) 42.6 We hypothesized that clinical parameters exist upon Range (y) 0.3-76 initial presentation, and that these variables can be used to Mean LOS (d) 4.7 LOS range (d) 1-30 predict mortality in emergency department (ED) patients Age, N60 y 16 (31) with environmental hyperthermia. We performed a retro- Altered mental status 38 (73) spective chart review on patients at 2 large academic GCS score b3 11 (21) institution in Phoenix, Ariz, diagnosed with heat-related Intubated 28 (53) illness to determine the presence of any useful clinical Deaths 15 (29) characteristics. SBP, b90 mm Hg 13 (25) HR, N100 bpm 40 (77) CK, N500 15 (29) Creatinine, N1.5 mg/dL 21 (40) 2. Methods Ethanol/illicit drugs present 18 (35) A retrospective chart review was performed at Good Samaritan Medical Center and Maricopa Medical Center in as age and sex were included. Other measurements Phoenix, Ariz. Both of these hospitals serve as tertiary care including electrocardiogram changes, altered mental status, facilities and level I trauma centers for the Phoenix or Glasgow Coma Scale (GCS) were identified. Short-term metropolitan area. Each facility supports a large number outcome measures such as need for endotracheal intubation, of residency programs and the ED see a combined average cardiopulmonary resuscitation, or death were extracted. volume of 96,000 patients per year. Electronic medical records were reviewed by research assistants from August 2003 to August 2005. Charts of 3. Results patients who are older than 18 years evaluated in the ED and given a discharge diagnosis of hyperthermia, heat cramps, We enrolled 52 patients with a mean age of 42.6 years heat exhaustion, and heat stroke or the related ICD-9 code (range, 0.4-81 years) over 2 years (Table 1). The mean high for these illnesses were reviewed. Exclusion criteria were daily temperature was 103.68F (range, 88-1188F). At alternative diagnoses potentially explaining the presence of presentation, the mean body temperature was 105.18F an elevated temperature from nonenvironmental causes such (range, 100.2-111.28F) and the GCS score was less than as pneumonia and urosepsis, among others. Research 14 in 36 (69.2%) patients. Laboratory results demonstrated assistants, blinded to the purpose of the study, took part in that 21 patients (40.4%) had a creatinine level of more than a brief 1-hour structured training session on systematic chart 1.5 mg/dL, 35 (67.3%) patients had a CK of more than 200 review. After the training session, 2 separate research U/L, 30 (57.7%) patients had a prothrombin time of more assistants performed a systematic chart review of charts than 13 seconds, 29 (55.8%) patients had an AST of more with a heat-related diagnosis. They extracted patient than 45 U/L, and only 3 (5.7%) patients had a glucose of demographics, comorbidities, vital signs on presentation, less than 60 mg/dL. Ethanol or illicit drugs were involved in laboratory results, and short-term outcome measures. If any 18 (34.6%) cases. The mean hospital stay was 4.7 days discrepancy existed between the data extracted by the 2 (range, 1-30 days), and there were 15 deaths (28.8%). A j research assistants, a third reviewer acted as a tiebreaker. score for interreviewer reliability was 0.69. Data were analyzed with Excel and STATA software. Odds The OR for death if the initial presenting systolic blood ratios (OR), multivariate analysis, and Student t tests were pressure was less than 90 mm Hg was 17.0 (95% CI, 3.8- used where appropriate. 75.8) ( P b .0001). The OR for death if the GCS score was Vital signs and laboratory results obtained as part of the less than 12 was 3.39 (95% CI, 0.91-12.6) ( P b .001). The ED evaluation were abstracted in an attempt to predict OR for death if initial temperature was higher than 42.08C outcomes from presenting patient characteristics. Vital sign was 3.28 (95% CI, 0.91-11.8) ( P b .001). The other measurements included systolic and diastolic blood pres- parameters were not clinically significant. sure, heart rate, body temperature, respiratory rate, and location of temperature measurement (rectal/oral). The maximum daily temperature on the day of presentation 4. Discussion was also recorded. Laboratory results collected included blood urea nitrogen (BUN), creatinine, potassium, AST, Hyperthermia has a significant impact on public health. creatinine phosphokinase, troponin, bicarbonate, serum In particular, patients with underlying disease, extremes of glucose, and urine drug screen. Patient demographics such age, and poor social equity are most at risk [2-4]. Estimated 444 F. LoVecchio et al. acute mortality from heat stroke ranges widely depending strates what is likely intuitive to most, namely, in addition to on the source [2-5]. Yet, the current study demonstrates high body temperature, low blood pressure or low GCS nearly 30% mortality rate from all heat-related illness score can suggest which patients are at most risk for death. presenting to the ED. Without coroner records, the true Physicians know that patients’ survival depends on prompt mortality rate is unknowable. Nevertheless, early recogni- diagnosis, cooling, and supportive therapies. This information and rapid initiation of treatment can reduce the high tion is invaluable in providing physicians the necessary morbidity and mortality from this disease. To prevent heat- clues to recognize the patient’s condition and then initiate related illness and death, public health agencies have prompt treatment of a disease with a high mortality. identified populations and risk behaviors. Extremes of age The major limitations of our study were the retrospective and persons without access to air conditioning are at nature and lack of homogeneity in patient evaluation, increased risk for heat-related illness and death. In addition, treatment, and test ordering. The sensitivity of our search persons with chronic mental illness or cardiopulmonary criterion in identifying all cases of heat stroke is unknown. disease, and those receiving medications such as diuretics, We completed a structured chart review as outlined by anticholenergic agents, and tranquilizers that decrease Gibert et al [8], in an attempt to overcome these limitations. perspiration, are at greater risk for heat-related illness and In addition, the presenting signs predicting the highest death. Drinking ethanol, abusing illicit drugs (eg, cocaine or mortality require validation in a prospective study. Howev- amphetamines), and decreased fluid intake while participat- er, to date our work is the largest study to analyze the ing in physical activities are risk behaviors associated with presence of clinical variables in hyperthermic patients heat-related illness [6-8]. presenting to the ED. Unfortunately, no study to date has analyzed the In summary, a presenting temperature higher than 428C, characteristics of patients with heat-related illness present- a systolic blood pressure of less than 90 mm Hg, or a GCS ing to the ED. Few studies have described the clinical score lower 12 resulted in the greatest risk of death in our features of near-lethal or lethal heat stroke or addressed patients. treatment in a controlled trial [5-8]. Most practitioners agree that aggressive cooling measures as rapidly as possible to minimize end-organ damage, to a core temperature should References be about 388C to avoid overshoot. [1] Lugo-Amador NM, Rothenhaus T, Moyer P. Heat-related illness. Controversy remains over which method is faster, and no Emerg Med Clin North Am 2004;22(2):315-27. controlled head-to-head studies comparing times or out- [2] McGeehin MA, Mirabelli M. The potential impacts of climate comes between the various techniques are available. variability and change on temperature-related morbidity and mortality Evaporative cooling is safe, effective, easily accom- in the United States. Environ Health Perspect 2001;109(Suppl 2): plished, and well tolerated. In general, we undress the 185-9. [3] Impact of heat waves on mortality—Rome, Italy, June-August 2003. patient, spray with tepid (not cold) water, and cool by large MMWR Morb Mortal Wkly Rep 2004;53(17):369-71. fans to maximize evaporative heat loss. Many experts [4] Semenza JC, Rubin CH, Falter KH, et al. Heat-related deaths during the suggest ice packs applied to the patient’s neck, axillae, and July 1995 heat wave in Chicago. N Engl J Med 1996;335(2):84-90. groin as well as cooling blankets. Other modalities with [5] Dematte JE, O’Mara K, Bueschr J, Whitney CG, Forsythe S, McNamee anecdotal success include ice water peritoneal and thoracic T, et al. Near-fatal heat stroke during the 1995 heat wave in Chicago. Ann Intern Med 1998;129(3):173-81. lavage, and cardiopulmonary bypass. Gastric or bladder [6] Varghese GM, John G, Thomas K, Abraham OC, Mathai D. Predictors lavage likely adds very little to the effect of evaporative of multi-organ dysfunction in heatstroke. Emerg J Med 2005;22(3): cooling when performed properly, and the former addition- 185-7. ally carries the risk of aspiration. Alcohol sponge baths are [7] LoVecchio F, Stapczynski JS, Hill J, et al. Heat-related mortality— dangerous and should never be used. Despite this lack of Arizona, 1993-2002, and United States, 1979-2002. MMWR 2005; 54(25):628-30. hard science, the emergency physician needs to know what [8] Gilbert EH, Lowenstein SR, Koziol-McLain J, Barta DC, Steiner J. predicts a poor outcome in patients presenting with any Chart reviews in emergency medicine research: where are the methods? heat-related illness, not just heat stroke. Our study demon- Ann Emerg Med 1996;27(3):305-8. American Journal of Emergency Medicine (2007) 25, 445–449

www.elsevier.com/locate/ajem

Brief Report Effectiveness of nonnarcotic protocol for the treatment of acute exacerbations of chronic nonmalignant pain James E. Svenson MD, MS*, Thomas D. Meyer MD

Section of Emergency Medicine, University of Wisconsin, Madison, WI 53792, USA

Received 7 September 2006; revised 28 September 2006; accepted 29 September 2006

Abstract Introduction: Emergency department (ED) overcrowding is a growing problem. Frequent visits for chronic pain are a significant subset of patients. The use of narcotics in these patients is controversial. The purpose of this study was to test a strict nonnarcotic protocol in reducing need for and number of ED visits for chronic pain while at the same time addressing their pain. Methods: This was a prospective observational study. We identified patients with more than 10 ED visits for exacerbations of chronic nonmalignant pain in the last 12 calendar months. Each patient and their physician were sent letters informing them of the concern of frequent ED use and the use of opioids for rescue therapy. Furthermore, the patient would receive medications other than narcotics in subsequent ED visits, and follow-up with the primary physician for alternatives was encouraged. Use of the ED for pain-related visits was then monitored for the subsequent 12-month period. Clinic use and outpatient medication uses were also monitored. Results: Fifteen patients were identified for the initial study. These patients averaged 19 ED visits per 12 months for pain-related complaints. All of them had a regular physician. After notification of the new protocol, ED visits decreased to an average of 2 visits per year. Visits with primary care physicians also dropped from an average of 19 visits per year to 7 visits. There were 7 patients who had been weaned off narcotic medications, 4 who had been converted to methadone maintenance, and 1 who had been switched to a fentanyl patch. Conclusions: Initiation of a strict nonnarcotic protocol for treatment of patients with frequent ED visits for chronic nonmalignant pain results in a significant drop in the number of pain-related visits to the ED. These visits were not offset by a significant elevation in the number of clinic visits for pain complaints, and many were weaned off narcotics. Nonnarcotic protocols for acute exacerbations of chronic nonmalignant pain may be a viable alternative for reducing frequent pain-related ED visits in a select population. D 2007 Elsevier Inc. All rights reserved.

1. Introduction account for a disproportionate number of ED visits [1-3]. Prior studies have reported that 3% to 4% of patients may Emergency department (ED) overcrowding is a national account for 12% to 20% of total ED visits [1,2]. One problem. One contributor is a small group of patients who subgroup of these consists of patients with chronic, frequent migraines and chronic nonmalignant pain [4,5].These Presented at the 2006 SAEM Annual Meeting, San Francisco, Calif. patients account for about 7% of those frequent users of * Corresponding author. Tel.: +1 608 265 5808; fax: +1 608 262 2641. the ED. This subgroup of frequent ED users represents a E-mail address: [email protected] (J.E. Svenson). challenging area of patient management.

Programs have been instituted to improve the care and inclusion in this study. Patients were excluded if they did management of frequent ED users in the hopes of not have more than 10 pain-related visits in the 12-month decreasing their use. These include, for example, intensive period, if their pain was not treated with narcotics, if their case management [6], keeping lists of narcotics abusers [7], pain was malignancy-related, or if they had significant and limiting the number of providers and pharmacies the underlying medical comorbidities. Comorbidities included patient may use [8]. These approaches may have limited renal failure, sickle cell, and so on. Both study coordinators usefulness in the subset of patients with chronic migraines agreed on all exclusions. A letter was sent to each patient by or chronic malignant pain. Limiting the number of ED visits certified mail, which discussed the use of narcotics as rescue for treatment of acute exacerbations of pain has been one therapy for such syndromes. The observation was made that strategy, but again, has not been completely effective [9]. the patient’s physician could supply rescue therapies The optimal management of acute exacerbations of provided in the ED in alternative outpatient forms and the chronic pain is controversial, both to avoid the use of problems associated with frequent use of narcotics. We then potentially inappropriate opioids and at the same time avoid informed the patient that we would no longer treat them inadequately addressing the patient’s pain. Often, patients with parenteral narcotics in the ED and encouraged to be with chronic pain treated with chronic opioids can produce a seen by their personal physician in the near future to discuss condition of enhanced pain sensitivity [10,11]. As medica- alternative therapeutic options. A similar letter was sent to tion wears off, they feel worse. In these cases, short-term their personal physician (Appendices A and B). We parenteral narcotics may be worsening the problem. The continued to offer nonnarcotic alternatives to treat their American Academy of Pain Management contains several exacerbations if they required ED evaluation. We allowed essential recommendations on chronic pain management: patients a 1-month period to make contact with their (1) treatment should be provided solely by a single physicians to discuss and initiate alternative treatment practitioner or clinic; (2) narcotic use should be limited to regimens. When presenting to the ED after this period, circumstances where it enhances function at work and alternative nonnarcotic pain management regimens were home; (3) escalation of use or seeking narcotics from offered to the patient. Alternative regimens offered were at multiple providers should be considered an indication that the discretion of the treating ED physician. narcotic treatment needs to be reviewed and possibly Patient use of the ED over the next 12 months (excluding discontinued. Thus, for the chronic pain patient with a the 1-month period after the letter was sent out in which the personal physician, the current pattern of providing addi- patient was to meet with their own physician) was then tional narcotics in the ED is contrary to the recommenda- monitored by review of clinic and ED notes in the patient’s tions of the American Academy of Pain Management. In electronic record. addition, most opioids used in the ED can be given in alternative forms as an outpatient medication [12]. Some of these may be more appropriate as they have slower onset 3. Results compared with the rapid onset action of parenteral narcotics [13]. Although narcotics may be suboptimal therapy for There were 32 patients who had greater than 10 visits per acute exacerbations of chronic pain, other alternative year for acute exacerbations of chronic pain. Of these, 17 therapies must be offered to each patient to help alleviate were excluded: 8 did not receive narcotics at each visit, 4 their pain. had sickle cell disease, and 4 had complex medical The purpose of this study was to test a strict nonnarcotic problems or malignancy-related pain. Thus, there were 15 protocol in reducing frequent ED visits for patients with patients included in this preliminary study. One patient was chronic nonmalignant pain while at the same time address- lost during the follow-up periods. Of the remaining, during ing their pain. the 12-month period before entry, these patients had an average of 19 visits per 12-month period for pain-related complaints at which they were treated with narcotic pain 2. Methods medications (range, 14-28). In the 12-month period after institution of this protocol, This was a prospective observational study. We identified ED use fell for each patient. The average number of visits for all patients with greater than 10 ED visits to the XXXX ED pain-related complaints fell to 2 per patient per 12-month in the last 12 calendar months. The XXXX ED is a period (range, 0-8). No violations of the new protocol were community teaching hospital in a small city environment. recorded. Patients reported pain relief with the new protocols. Because this was a new protocol, we tried to make the Every patient reported a pain level of 10/10 on arrival to the criteria for entry as strict as possible. As a feasibility study, ED. After institution of our new protocol, all reported pain no formal power calculation was made. The records of these relief, but at varying levels (average, 5/10; range, 2-8/10). patients were reviewed, and if it was determined that they Corresponding clinic visits also significantly dropped were treated at these visits for an exacerbation of chronic from an average of 19 visits to 7 visits per 12-month period. nonmalignant pain or headache, then they were eligible for During the follow-up period, 7 patients had been weaned off Effectiveness of nonnarcotic protocol 447 of narcotic medications, 4 had been switched to methadone benzodiazepines [10] among others. Thus, in all of our maintenance therapy, and 1 converted to a fentanyl patch. patients with chronic nonmalignant pain, treatment with the The total cost before the institution of this policy, use of alternative nonnarcotic regimens were open for the ED including physician, pharmacy and hospital bills, averaged physician to use during repeat visits in the time after the $800 per patient. The average length of stay calculated from institution of our policy. triage to discharge was slightly more than 3 hours. Thus, in If narcotic treatment is withheld in the ED setting, patients the first year after initiation of this protocol, there was a may seek this treatment in other urgent care or clinic settings. reduction of approximately 255 visits in the ED for chronic The institution of our protocol did not result in a pain, with a calculated cost saving of more than $200,000 corresponding increase in outpatient clinic use, but we have and the reduction of approximately 765 patient hours. There no data on the use of urgent care centers or other EDs in the were no patient complaints or board action brought in area. However, we noted that clinic use actually declined relation to the new policy. during the follow-up period, and many patients were successfully weaned from their chronic narcotic medications. The Joint Commission on Accreditation of Healthcare 4. Discussion Organizations (JACHO) recently published standards that called for the evaluation, treatment, and assessment of In this preliminary study, we have shown that the use of a effectiveness of pain improvement [31]. In addition, there strict nonnarcotic protocol can significantly reduce ED use have been reports of inadequately treated pain in the ED in those patients presenting with chronic nonmalignant pain. [32]. Given these guidelines and findings, the question is Treatment of bfrequent migraineursQ or those with chronic whether the ED physician is compelled both ethically and nonmalignant pain is controversial [14-16]. Although legally to provide narcotics for pain relief to those with opioids have been used frequently in such patients, these chronic nonmalignant pain. The standards state, bpatients drugs are associated with the potential for abuse, addiction, have the right to appropriate assessment and management of and tolerance. There is evidence that opioids may be pain.Q There are no specific references to opioids, and the ineffective in neuropathic and idiopathic pain [17].In necessary degree of relief is not specified [33]. The intent of addition, treatment with opiates frequently contributes to the standards is that a patient’s pain should be treated in the the psychological aspects of the disease. Many patients use best way possible. This may exclude narcotics for many opioids on a daily basis [18]. Chronic use and the use of patients with chronic pain. opioids for acute exacerbations of pain can be associated Many patients with chronic pain actually have improve- with a rebound phenomenon, limiting their effectiveness ment in their pain when weaned off opioids [34,35]. Some both in the short and long term [19]. patients’ pain resolves completely with stopping narcotics Regardless of the appropriateness of narcotics for rescue [36]. In this study, we noted that many patients had actually therapy for patients with acute exacerbation of chronic been weaned from their chronic narcotics after we instituted nonmalignant pain, many narcotics are available in alternative our new policy apparently without increasing their pain. formulations and can be used effectively at home for rescue If narcotics are used for relief of chronic nonmalignant therapy. For example, effective narcotics are available for pain, the medication should be slow-onset. Rapid-onset rectal, oral (transmucosal), or nasal or even home intramus- medication should be avoided because it may cause problem cular administration [20-22]. Oral protocols have been used with reinforcing pain behavior and gives a short-lived successfully for treatment of acute exacerbations of chronic psychological relief [13]. We noted to each of our patients migraines even if there is associated nausea and vomiting [12]. that alternative, more appropriate rescue therapies than Limits on the number of acute visits for rescue therapy parenteral narcotics in the ED could be worked out with have been tried but still result in frequent ED visits [9]. their treating physician and gave them ample time to seek out Proposals have been made to try to shift treatment of such regimen. We also notified the patient’s physician of our exacerbations of chronic nonmalignant pain to home or concerns and change in policy, thus again, encouraging the family physician’s offices, but these have not been really formulation of a better rescue regimen. We noted that to avoid met with great success [23]. If the outpatient treatment these reinforcement issues, the patient’s physician has to protocol is effective, then these strategies could lead to more appropriately address rescue medications and regimens. decreased use [12]. The effective rescue therapeutic regimen Most pain specialists agree that to maintain control and could be based on the medications used in repeated ED optimize the patient’s drug usage, only one physician should visits given in an alternative outpatient form. prescribe opioids [37]. We felt that our protocol encouraged Many alternative nonnarcotic regimens for the treatment the reinforcement of this behavior by referring the patient of acute exacerbations of headache or nonmalignant pain back to their own physician for discussion of more have been studied, with variable results [24-26]. However, appropriate rescue regimens. there are alternatives that have been shown to be effective, We felt that our protocol, although seemingly draconian, including ketamine [27], droperidol [28] and other neuro- was ethically and medically in the patient’s best interest in leptics [29], continuous dihydroergotamine (DHE) [30], and treating their chronic nonmalignant pain in the optimal 448 J.E. Svenson, T.D. Meyer manner. With this change in policy, our patients experienced the XXXX ED for medical care. If you need help finding a improvement in their clinic and ED use and their use of primary care physician within the XXXX Health system, narcotics in controlling their pain. please contact our Patient Relations Department at XXXXX or you can discuss this with a social worker in the ED. As always, you are welcome to come to the ED for 5. Conclusions evaluation of any condition, as well as to receive nonnarcotic Initiation of a strict nonnarcotic protocol for treatment of treatment for intermittent worsening of your chronic pain. patients with frequent ED visits for chronic nonmalignant pain results in a significant drop in the number of pain-related Sincerely, visits to the ED. These visits were not offset by a significant Cc: Patient’s PCP & Pain Specialist elevation in the number of clinic visits for pain complaints. Nonnarcotic protocols for acute exacerbations of chronic nonmalignant pain may be a viable alternative for reducing Appendix B frequent pain-related ED visits in a select population. Dear Doctor:

Appendix A We are writing about our mutual patient ______. As you know he/she frequently uses the ED for acute Dear Patient: exacerbations of chronic pain. He/she has been in the ED over 10 times in the last 12 months. During these visits he/she has As a patient who has come to the XXXX ED many times requested and received rescue therapy including IM/IV during the last year to receive treatment for pain, we are narcotics. We would note that this medication is available writing to inform you of a new guideline being used at the in equivalent doses in oral, intranasal, or rectal forms. XXXX ED for those who suffer from intermittent worsening Current literature would suggest that treatment of acute of chronic pain. exacerbations of chronic pain conditions, for example, While an important role for physicians is to relieve pain headache, with narcotics is not optimal. Pain relief from IV and suffering, we also want to do no harm. Pain relief from or IM narcotics lasts only a short period of time, while the injectable narcotics lasts only a short time, while the frequency of rebound pain increases. The literature also frequency of rebound pain increases. The medical literature reports that short-term narcotic treatment can actually suggests that treatment with short-term narcotics can actually increase the stress and disability associated with chronic pain. increase the stress and disability associated with chronic pain. Because of this, we do not feel that we can continue to In addition, long waits in the ED are common and the noise offer narcotics for rescue therapy to patients with acute and chaos of the ED may actually worsen the pain. exacerbations of chronic non-malignant pain. If you feel that We have spoken with primary care and pain physicians in narcotics are absolutely necessary for the treatment of this our area about developing a more consistent, coordinated patient’s pain, we would suggest that you work out a plan approach to managing those people with chronic pain. From for rescue therapy using the alternative form mentioned these conversations, our observations, and the medical above, in an alternative venue, or that you primarily provide literature, we have concluded that frequent, short-term this therapy in the ED. narcotic pain treatment is not a healthy approach to pain Since it takes time to arrange for alternative strategies, management. The best results are seen when a patient’s pain we will continue to treat acute exacerbations of your is managed through a primary care physician (PCP). The patient’s chronic pain with narcotic rescue therapy for up management may include prevention as well as specific pain to 30 days from the time of this letter, though we now feel treatments to be used at home. The PCP may also elect to that therapy is suboptimal. We would expect that within that consult with an area pain specialist who can assist in time frame, you could see your patient and formulate an developing a pain management plan. alternative treatment protocol. Given the above considerations, XXXX emergency If you have any questions regarding our new policy physicians can no longer support the use of injectable please do not hesitate to contact us. narcotics for the treatment of intermittent worsening of chronic pain. We realize that it will take some time for you Sincerely, to set up an alternative treatment plan with your PCP. To help you begin this process, we will be contacting your primary care physician with this information. In the interim, References we will continue to provide you with your usual care for up to thirty (30) days from the date of this letter. [1] Sun BC, Burstin HR, Brennan TA. Predictors and outcomes of We hope that this information has been informative and frequent emergency department users. Acad Emerg Med 2003;10: that it will help you know what to expect when you come to 320-8. Effectiveness of nonnarcotic protocol 449

[2] Mandelberg JH, Kuhn RE, Kohn MA. Epidemiologic analysis of an [21] Dale O, Hjortkjaer R, Kharasch ED. Nasal administration of opioids for urban, public emergency department’s frequent users. Acad Emerg pain management in adults. Acta Anaesthesiol Scand 2002;46:759-70. Med 2000;7:637-46. [22] Burton AW, Driver LC, Mendoza TR, Syed G. Oral transmucosal [3] Owens HJ, Chan BT. Heavy users of emergency services: a fentanyl citrate in the outpatient management of severe cancer pain population-based review. Can Med Assoc J 2001;165:1049-50. crises: a retrospective case series. Clin J Pain 2004;20:195-7. [4] Chan BTB, Owens HJ. Chronic migraineurs: an important subgroup [23] Fishman SM, Brandman TB, Edwards A, Borsook D. The opioid of patients who visit emergency departments frequently. Ann Emerg contract in the management of chronic pain. J Pain Symptom Manage Med 2004;43:238-42. 1999;18:27-37. [5] Malone RE. Heavy users of emergency services: social construction of [24] Mehl-Madrona LE. Comparison of ketorolac-chlorpromazine with a policy problem. Soc Sci Med 1995;40:469-77. meperideine-promethazine for treatment of exacerbations of chronic [6] Pope D, Fernandes CM, Bouthillette F, Etherington J. Frequent users pain. J Am Board Fam Prac 1999;12:188-94. of the emergency department: a program to improve care and reduce [25] Larkin GL, Prescott JE. A randomized, double-blind, comparison visits. Can Med Assoc J 2000;162:1017-20. study of the efficacy of ketorolac tomethamne versus meperidine in [7] Zechnich AD, Hedges JR. Community-wide emergency department the treatment of severe migraine. Ann Emerg Med 1992;21:919-24. visits by patients suspected of drug-seeking behavior. Acad Emerg [26] Kudrow L, Kudrow DB, Sandweiss JH. Rapid and sustained relief of Med 1996;3:312-7. migraine attacks with intranasal lidocaine: preliminary findings. [8] Stewart J. Manitoba targets patient overuse as it tackles abuse of Headache 1995;35:79-82. health care system. Can Med Assoc J 1995;152:1483-4. [27] Carr DB, Goudas LC, Denman WT, Brookoff D, Staats PS, Brennen [9] Fishman SM, Brandman TB, Edwards A, Borsook D. The opioid L, et al. Safety and efficacy of intranasal ketamine for the treatment of contract in the management of chronic pain. J Pain Symptom Manage breakthrough pain in patients with chronic pain: a randomized double- 1999;18:27-37. blind, placebo-controlled, crossover study. Pain 2003;108:17-27. [10] Banllantyne JC, Mao J. Opioid therapy for chronic pain. N Engl J [28] Richman PB, Allegra J, Eskin B, Doran J, Reischel U, Kaiafas C, et Med 2003;349:1943-53. al. A randomized clinical trial to assess the efficacy of intramuscular [11] Streltzer J. Pain management in the opioid-dependent patient. Curr droperidol for the treatment of acute migraine headache. Am J Emerg Psychiatry Rep 2001;3:489-96. Med 2002;20:39-42. [12] Von R, Seggern L, Adelman JU. Oral narcotic protocol to reduce [29] Siow HC, Young WB, Silberstein SD. Neuroleptics in headache. narcotic injections in refractory migraine patients. Headache 1997;37: Headache 2005;45:358-71. 341-5. [30] Ford RG, Ford KT. Continuous intravenous dihydroergotamine in the [13] Bushnell TG, Justins DM. Choosing the right analgesic: a guide to treatment of intractable headache. Headache 1997;37:129-36. selection. Drugs 1993;46:394-408. [31] www.jcaho.org accessed on 11 November 2005. [14] Mitka M. Experts debate widening use of opioid drugs for chronic [32] Rupp T, Delaney KA. Inadequate analgesia in emergency medicine. nonmalignant pain. JAMA 2003;289:2347-8. Ann Emerg Med 2004;43:494-503. [15] Wilsey B, Fishman S, Rose JS, Papzian J. Pain management in the [33] Hansen GR. Management of chronic pain in the acute care setting. ED. Am J Emerg Med 2004;22:51-7. Emerg Med Clin North Am 2005;12:307-38. [16] Bodley SC. Narcotics for chronic nonmalignant pain. Can Med Assoc [34] Taylor CB, Zlutnick SI, Corley MJ, Flora J. The effects of J 2001;164:461. detoxification, relaxation, and brief supportive therapy on chronic [17] Norton LL. The use of opioids in the treatment of nonmalignant pain. Pain 1980;8:319-29. pain: clinical issues and guidelines. Am J Pain Manage 1997;l7: [35] Schofferman J. Long-term use of opioid analgesics for the treatment of 42-52. chronic pain of nonmalignant origin. J Pain Symptom Manage 1993;8: [18] Maizels M. Health resource utilization of the emergency department 279-88. headache brepeaterQ. Headache 2002;42:747-53. [36] Finlayson RE, Maruta T, Morse RM, Martin MA. Substance [19] Savage SR. Long-term opioid therapy: assessment of consequences dependence and chronic pain: profile of 50 patients treated in an and risks. J Pain Symptom Manage 1996;11:274-86. alcohol and drug dependence unit. Pain 1986;26:167-74. [20] Pasero C, McCaffery M. Opioids by the rectal route. Am J Nurs 1999; [37] Savage SR. Opioid use in the management of chronic pain. Med Clin 99:20. North Am 1999;83:761-86. American Journal of Emergency Medicine (2007) 25, 450–458

www.elsevier.com/locate/ajem

Clinical Notes Sixty-four–slice multidetector computed tomography: the future of ED cardiac care

Alexander T. Limkakeng MDa,*, Ethan Halpern MDb, Kevin M. Takakuwa MDa aDepartment of Emergency Medicine, Chest Pain Center, Thomas Jefferson University, Philadelphia, PA 19107-5004, USA bDepartment of Radiology, Thomas Jefferson University, Philadelphia, PA 19107-5004, USA

Received 20 September 2006; accepted 26 October 2006

Abstract Multidetector computed tomography (MDCT) imaging, a technological advance over traditional CT, is a promising possible alternative to cardiac catheterization for evaluating patients with chest pain in the emergency department (ED). In comparison with traditional CT, MDCT offers increased spatial and temporal resolution that allows reliable visualization of the coronary arteries. In addition, a btriple scan,Q which includes evaluation for pulmonary embolism and thoracic aortic dissection, can be incorporated into a single study. This test will enable emergency physicians to rapidly evaluate patients for life-threatening illnesses and may allow safer and earlier discharges of many patients with chest pain in comparison with a traditional rule-out protocol. In this article, we will highlight the technological advances of MDCT imaging, review the literature on coronary angiography via MDCT, and discuss the future of this technology as it relates to the ED. D 2007 Elsevier Inc. All rights reserved.

1. Introduction million of these procedures were performed in the inpatient setting alone [8,11]. Approximately one fourth of patients Evaluation for acute coronary syndrome accounts for undergoing the test will have normal coronary vessels and more than 5 million emergency department (ED) visits [1]. will not require percutaneous intervention [12,13]. Although Symptom presentation varies widely, and initial workup is relatively safe, the procedure carries the risk of complications often nondiagnostic [2-7]. However, cardiac disease remains such as bleeding, infection, stroke, and coronary vessel a leading cause of morbidity, mortality, and medicolegal damage or clotting. According to recent American Heart burden [8]. Thus, emergency physicians are required to have Association statistics [8], cardiac catheterization carried an a low threshold to admit patients to an inpatient setting or inhospital mortality rate of 1.0%, with an average length of observation unit for additional testing to brule outQ coronary stay of 3.7 days. In addition, the test requires an interventional artery disease [9,10], at great cost to the health care system cardiologist and specialized space and equipment. The mean and inconvenience to the patient. charge for patients hospitalized for diagnostic cardiac Cardiac catheterization is currently the gold standard test catheterization was $24,893 [8]. Because of limitations of for identifying coronary artery disease. In 2003, at least 1.4 time, cost, risk, and availability, most ED patients with chest pain do not currently receive cardiac catheterization [1,8]. Multidetector computed tomography (MDCT) has * Corresponding author. Tel.: +1 215 955 6844; fax: +1 215 923 6225. emerged as a promising test for diagnosing coronary artery E-mail address: [email protected] (A.T. Limkakeng). disease. This new technique may prove to be the new

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2006.10.012 Sixty-four–slice multidetector computed tomography 451 diagnostic gold standard. In addition to coronary artery the time required to rotate the detector array 1808 and is disease, MDCT can diagnose other life-threatening diseases termed temporal resolution [15,16]. in an undifferentiated ED chest pain population. In this The single detector row CT scanner created 2-dimen- article, we seek to provide the emergency physician an sional planar images, or bslices,Q one at a time as the patient overview of this technology and discuss future applications. was advanced through the gantry. Single row CT scanners require approximately 1.0 to 2.0 seconds per slice. The first helical scanner was introduced in 1988 with slip-ring 2. Methods technology that allowed continuous rotation of the x-ray tube along with continuous patient motion. However, single We reviewed all articles published from 1995 to present row helical detector CT scanner still required 0.5 to 1.0 regarding MDCT for coronary angiography. A Pubmed second for acquisition of each slice. Inasmuch as the b search was performed using the terms multi-detector craniocaudal length of the heart approaches 10 cm, a single Qb Qb Q CT, noninvasive coronary angiography, multi-slice, detector row scanner requires 50 to 100 seconds to scan b Qb Q b Q angiography, coronary, and MDCT. We focused our through a heart. Motion artifact, from continuous beating of review to articles reporting sensitivity and specificity values the heart, prevents adequate visualization of coronary in reference to cardiac catheterization. vessels using single row CT [15,16]. In the 1990s, CT scanners with multiple detector rings were introduced, allowing simultaneous acquisition of 3. Background multiple slices. The first double detector row scanner was 3.1. Technical Aspects of CT/MDCT introduced in 1992. With MDCT, the patient is moved continuously through the gantry as the x-ray source and Sir Godfrey Hounsfield introduced medical imaging by detectors rotate within the gantry. The rate at which the table CT scanning in the 1970s [14]. A CT scanner consists of an is advanced for each rotation of the detector array is termed x-ray source that emits radiation, a detector unit that detects the pitch. For cardiac imaging, spatial resolution is typically the emitted radiation and converts it into an electric signal, 0.5 mm in the xy-imaging plane. Image resolution along the and a computer that transforms the raw x-ray data into an z-axis, also known as slice thickness, is related to detector image (Fig. 1). The x-ray source and the detectors are thickness and pitch. Thinner slices are obtained at the positioned opposite of each other within a ring-shaped expense of increased radiation exposure by reducing the housing called the gantry. X-ray radiation is tightly pitch. A typical pitch for imaging of the coronary arteries is collimated along a thin plane within the gantry termed the 0.2 with a slice thickness of 0.5 to 0.67 mm. The data from xy-imaging plane. The patient, centered within the bore of these scans can be reconstructed in any plane because the the gantry, is moved in a direction perpendicular to the xy- resolution is isotropic; that is, individual point resolution is imaging plane, along what is termed the z-axis. Image detail equal in the x-, y-, and z-axes. within the xy-imaging plane is quantified by spatial Early-generation 2-, 4-, and 8-slice CT scans created resolution and contrast resolution and is determined to a images that were not of sufficient quality to reliably large extent by the detector array. The time to acquire a identify or exclude coronary artery disease [17].The single image in the xy-imaging plane is generally equal to current decade has witnessed a rapid increase in the number of detector rows used in MDCT. Sixteen-slice CT scanners provided a marked improvement in both spatial and temporal resolution needed to visualize the coronary artery lumen. Initial studies reported sensitivity and specificity rates of 59% to 95% and 79% to 98%, respectively, primarily in patients with stable angina pectoris referred for cardiac catheterization (Table 1). The rate of adequate visualization of coronary vessels in these studies ranged from 68% to 96% [18-28]. These results have compared favorably with other modalities such as intravascular ultrasound and cardiac magnetic resonance imaging [29,30] but are still inadequate. In the largest multicenter study to date, Garcia and colleagues have confirmed some of the problems with 16-slice scans. In 238 outpatients, almost 30% of the scans had images of insufficient quality to determine coronary vessel disease, and sensitivity and specificity were poor [31]. Fig. 1 An example of an MDCT scanner. Image courtesy of Sixty-four–slice MDCT increases spatial and temporal Philips Medical Systems. resolution further and reduces overall scan time to as short 452 A.T. Limkakeng et al.

Table 1 Sixteen-slice MDCT scan studies Authors Year N Patient-based Sensitivity Specificity PPV NPV Calcium Premedication Inadequate Adequate unit of scores used images visualization analysis reported excluded rate Achenbach et al 2001 64 Y 91 84 59 98 N N Y 68 Kopp et al 2002 102 N 93 97 81 99 N N N NR Nieman et al 2002 59 Y 95 86 80 97 Y Y N 93 Ropers et al 2003 77 N 92 93 79 97 N Y N 88 Achenbach et al 2004 22 N 82 88 91 76 Y Y N NR Hoffman et al 2004 33 Y 90 75 86 81 N Y N 83 Kuettner et al 2004 60 Y 72 97 72 97 Y Y N 79 Achenbach et al 2005 50 Y 94 96 69 99 N Y Y 96 Burgstahler et al 2005 117 N 83 97 88 95 N Y Y NR Heuschmid et al 2005 37 Y 59 87 61 87 Y Y N 88 Hoffman et al 2005 103 Y 95 98 87 99 N Y Y 94 Kefer et al 2005 52 Y 82 79 46 95 Y N N NR Kuettner et al 2005 124 N 85 98 91 96 Y Y N 94 Kuettner et al 2005 72 Y 82 98 87 97 Y Y N 93 JACC Results patient-based where available. PPV indicates positive predictive value; NPV, negative predictive value; NR, not reported.

as 8 seconds. Newer scanners in development use 256 reported their experience on 67 consecutive patients who detector rows or flat-panel technology that actually elimi- were referred for suspected coronary artery disease or before nates the need for individual rows of detectors. coronary artery bypass graft [32]. Patients with prior stents Temporal resolution has been a major obstacle to cardiac or coronary artery bypass grafts were excluded. Vessels CT imaging. Most cardiac imaging must be performed N1.5 mm were evaluated, and all vessels imaged were of within the quiescent period of the heartbeat during diastole. adequate image quality. Based on individual vessels, At a heart rate of 60 beats per minute, the quiescent period is sensitivity and specificity for N50% stenosis were 94% no more than 200 milliseconds. Multidetector CT provided and 97%, respectively; and on a patient level, no false- increased z-axis coverage in a shorter time interval, but it negative or false-positive results were reported. did not address the issue of cardiac motion until the Mollet et al studied 70 patients who were scheduled for introduction of electrocardiogram gating. Electrocardiogram cardiac catheterization for atypical chest pain, stable or gating is the coordination of image acquisition with the unstable angina, or non-ST elevation myocardial infarction cardiac cycle to address cyclical changes in the position of [33]. They excluded those with previous stents or coronary the heart with each heartbeat. Because the length of diastole artery bypass grafts. In addition, 18 patients were excluded is closely related to the heart rate, b-blockers are given to from analysis because of logistical issues such as scheduling lower the heart rate for coronary CT. of the study, arrhythmias, renal insufficiency, or contrast Temporal resolution of MDCT in this decade has been allergy. All vessel sizes were imaged, with 97% of vessels further improved by faster gantry rotation speeds. More adequately imaged to allow interpretation. They reported recently, one manufacturer has introduced a scanner with 2 patient-based sensitivity and specificity of 100% and 92%, x-ray sources that may double the temporal resolution and respectively, for N50% vessel stenosis, misidentifying only eliminate the need for the use of b-blockers. Improvements one patient with normal coronaries by cardiac catheteriza- in temporal resolution reduce motion artifact by allowing tion. They used 2 observers for MDCT, with an interob- scans to acquire data at the same quiescent phase of the server j of 0.73. cardiac cycle over multiple beats of the heart [18]. Ropers et al reported use of 64-slice MDCT in 84 stable patients referred for cardiac catheterization for suspected coronary artery disease, excluding patients with contraindications to contrast and radiation or with previ- 4. The present ously documented coronary artery disease [34]. Vessels N 4.1. Sixty-four–slice MDCT in outpatients 1.5 mm were analyzed, with 96% of the vessels being adequately imaged. They found patient-based sensitivity Five studies comparing 64-slice CT with cardiac cathe- and specificity of 96% and 91%, respectively, compared terization in outpatients suspected of having coronary artery with cardiac catheterization, with only 4% of vessels disease have been published (Table 2) [32-36]. Leschka et al inadequately visualized. Sixty-four–slice multidetector computed tomography 453

Table 2 Sixty-four–slice MDCT scan studies Authors Year N Patient- Sensitivity Specificity PPV NPV CAD Calcium Premedication Inadequate Adequate based prevalence scores images visualization unit of excluded rate analysis Mollet et al 2005 52 Y 100 92 97 100 73% Y Y N 97 Leschka 2005 67 Y 100 100 100 100 70% Y N N 100 et al Ropers et al 2006 84 Y 96 91 83 98 31% N Y Y 96 Leber et al 2005 59 Y 88 85 88 85 42% N Y Y 93 Raff et al 2005 70 Y 95 90 93 93 57% Y Y Y 88 Results patient-based where available. CAD indicates coronary artery disease.

Raff et al studied 84 consecutive patients being referred 3 years [37]. Although this may increase specificity of the for cardiac catheterization for suspected coronary artery test, it comes at a cost of decreased sensitivity. Of the disease, excluding 14 patients with arrhythmias or contra- aforementioned studies, only two examined MDCT’s ability indications to contrast and b-adrenergic blocking medica- to quantify lesion severity, finding a correlation coefficient tion [35]. They reported sensitivity and specificity of 95% of 0.54 and 0.76 compared with cardiac catheterization and 90%, respectively, for N50% stenosis. They imaged all [34,36]. It remains to be seen whether MDCT can accurately vessel sizes and found that only 88% of vessels were imaged quantify the percentage of coronary vessel stenosis, with high quality, but no patients needed to be excluded information that is valuable in guiding treatment. because of image quality. A final limitation of some MDCT studies has been the Furthermore, Leber et al imaged 59 patients scheduled for exclusion of some segments or scans from analysis cardiac catheterization, excluding those with atrial fibrilla- [38,39]. Although somewhat limited in research settings, tion, contraindications to contrast, previous coronary artery the number of bnondiagnosticQ or technically inadequate bypass graft, or more than one coronary stent [36].In studies may be substantial in real-life settings. Nonetheless, addition, 4 patients were excluded because of inadequate CT the existing studies demonstrate the promise of this imaging. They included patients with only one stent and technology and indicate that this is fertile ground for those whose heart rates were N60 and found sensitivity of emergency medicine research. only 88% for vessels with N50% stenosis. Their limited accuracy was largely due to inclusion of patients with stents: 4.2. Emergency medicine literature in such patients, 6 of 13 vessels were incorrectly identified. These studies demonstrate the technical capability of 64- The literature on MDCT in ED patients is rapidly slice MDCT to accurately identify patients with N50% emerging. Gallagher and colleagues prospectively com- coronary stenosis. Although follow-up for adverse effects pared the accuracy of 64-slice MDCT with that of was not routinely or systematically performed, in the 332 traditional stress tests in 92 low-risk ED observation patients studied, there were no reported adverse effects from patients with symptoms suggestive of coronary artery the test, indicating its safety in a selectively screened disease [40]. Patients with ischemic electrocardiographic population. The high negative predictive values reported findings; positive cardiac markers; existing cardiomyopa- suggest a role for MDCT as a noninvasive screening test for thy, coronary artery disease, or heart failure; irregular heart coronary artery disease. rhythm; renal insufficiency; or a contraindication to However, several limitations apply to these results as a contrast or b-blockade medication were excluded. Seven group. First, the study populations were stable outpatients patients’ MDCT images were of insufficient quality to who were being referred for cardiac catheterization. There evaluate. All patients received both nuclear imaging stress was a high prevalence of coronary artery disease in these tests and MDCT. The results of all studies were available populations. Therefore, it is uncertain whether these results to the treating physicians, who determined which patients will translate to an ED population that is acutely symptom- subsequently underwent cardiac catheterization. Outcomes atic and has a different prevalence of coronary artery disease. were defined as 30-day adverse cardiac events or cardiac A second limitation is the use of 50% stenosis as the catheterization showing N70% stenosis. A total of 7 criterion for detecting a coronary lesion. This is a lower patients in the study group were found to have coronary threshold than what is considered clinically significant artery disease, all by cardiac catheterization. Multidetector based on angiography. It has been shown that lesions CT identified 6 of these patients. Eleven patients’ MDCT b50% portend a slight risk for myocardial infarction at results showed N50% stenosis; 6 were ultimately found to 454 A.T. Limkakeng et al. have N70% stenosis on catheterization. They failed to find (9%), depending on which tests the patient ultimately significant differences between MDCT and nuclear stress received. Multidetector CT diagnosed 4% of patients with testing in terms of sensitivity (86% vs 71%, respectively), noncardiac diseases for which CT is the standard reference specificity (92% vs 90%), or negative (99% vs 97%) and technique of diagnosis. They noted high sensitivity (83%) positive (50% vs 38%) predictive values. Among patients and specificity (96%) for N50% stenosis using a 16-slice CT with discordant results between MDCT and nuclear stress scanner, with only 2 false-positive and 2 false-negative tests, there was 1 patient with a negative MDCT but results in the 69 patients. positive stress test and 2 patients with a positive MDCT Another study evaluated 66 consecutive patients admit- but negative stress test who ultimately were determined to ted to the hospital for bacute chest pain syndromeQ with a have acute coronary syndrome. 16-slice CT scanner [44]. Vessels N2 mm were analyzed, The Departments of Radiology, Emergency Medicine, with 3.1% of vessels inadequately imaged. After excluding and Cardiology of Massachusetts General Hospital reported 7 patients with inadequate studies, they found a vessel- their experience using 64-slice MDCT in ED patients with based sensitivity and specificity of 80% and 89% for N50% chest pain who were being admitted despite an initially stenosis compared with cardiac catheterization, respectively. normal or nondiagnostic workup [41,42]. Patients were The rate of coronary artery disease in this group was high excluded if they had pain for N24 hours, evidence of (88.8% had acute myocardial infarction or unstable angina), hemodynamic instability or definite acute coronary syn- and the average time between cardiac catheterization and drome, an arrhythmia, or a contraindication to iodinated MDCT was 4 days. contrast. All eligible patients received an MDCT, then Chase et al [45] reported their experience with 64-slice standard clinical care throughout hospital admission. Two MDCT in 41 patients at low risk for ischemia with a reviewers, blinded to the MDCT results, assigned the Thrombolysis In Myocardial Infarction (TIMI) risk score diagnosis of acute coronary syndrome to study patients of 0 to 2 in abstract form only. Patients with negative based on evidence for myocardial infarction or unstable MDCT results (b50% stenosis) were immediately dis- angina after the patient’s index hospitalization workup. In an charged home instead of being placed in an observation initial trial of 40 patients using this protocol, MDCT unit. Thirty-three patients from this group were discharged, identified all 5 patients with acute coronary syndrome, and none had any adverse events within 30 days. Although without any false-negative results [41]. At the same time, analysis of this study is limited in the absence of more the MDCT results indicated or could not exclude stenosis on detailed information on their methods and results, they are 9 patients who were determined not to be having acute suggestive of the possibilities of MDCT in ED patients. coronary syndrome. In a second, more recent trial of 103 Savino and colleagues reported their initial experience on patients, they identified all 14 patients with acute coronary 23 ED patients with chest pain with nondiagnostic electro- syndrome without any false-negative results in the 72 cardiograms and cardiac markers, finding moderate to patients without MDCT findings. However, a significant severe coronary artery disease in 8 patients, all of which stenosis was detected or could not be excluded in 30 were later confirmed angiographically. Two patients with patients, providing an overall positive predictive value of pulmonary embolism were diagnosed on the basis of the CT 47%. Fifteen of the 17 inconclusive studies were attributed scan and treated with fibrinolytics. They note that 9 patients to either previous stent placement or severe vessel calcifi- with normal scans were discharged from the ED, but no cation. Follow-up was completed in 81 of the 89 patients follow-up was reported. Whether this experience can be without acute coronary syndrome on an average of 5 months replicated in a population with lower prevalence of disease after the index hospitalization, and there were no subsequent remains to be seen [46]. adverse outcomes. Logistic regression analysis indicated Moloo et al [47] compared MDCT with myocardial that information on plaque severity enhanced risk stratifi- perfusion imaging via single photon emission CT and cation using traditional risk factors or a clinical gestalt found a slightly lower rate of diagnostic abnormality using estimate [42]. the MDCT. However, the rate of agreement was high White et al studied 78 ED patients with acute chest pain, (93%) when both studies were diagnostic; and only 1 of excluding those who were clinically unstable, had definite 63 patients had single photon emission CT perfusion myocardial infarction, or deemed unlikely to have a defects with a negative MDCT, whereas there were 2 significant cause of chest pain [43]. Patients consented to patients with abnormal MDCT examinations and normal the 16-slice MDCT as an additional study. Nine patients single photon emission CT studies. Currently, there is not were excluded because of loss of data or patient being enough data to determine whether MDCT is better at unavailable for the MDCT after consenting. Approximately identifying high-risk ED patients than proven modalities 1 month after each patient’s ED visit, the final diagnosis was such as stress testing [48]; but if so, it would represent an determined by consensus of a group consisting of an advance in diagnosis, as MDCT can be more readily emergency physician, a cardiologist, and a radiologist. available than stress testing. Diagnoses were based on clinical data (49%), radionuclide Although it was not the primary focus of some studies of testing (22%), cardiac catheterization (16%), and stress echo ED patients, the ability of MDCT to detect noncardiac Sixty-four–slice multidetector computed tomography 455 disease processes in patients with chest pain seems 4.3.2. Heart rate promising. White et al diagnosed 3 patients with non- The first studies of MDCT recognized that heart rate had coronary diseases: one patient each with pulmonary significant impact on the quality of vessel imaging. Faster embolism, pericardial effusion, and pneumonia [43]. One heart rates were associated with more motion artifact, which study of 151 low-risk patients with chest pain found that was consistently the leading cause of inadequate images. In 11.9% of patients had significant noncardiac findings, studies aimed specifically at evaluating this effect, a including hiatal hernias, esophageal inflammation, pulmo- statistically significant inverse correlation between heart nary infiltrate, and pericardial effusion. An additional 6.6% rate and vessel visibility was found [23,53,56]. For this had findings requiring follow-up such as enlarged lymph reason, it has become standard protocol to administer b- nodes or noncalcified masses [49]. blocker medication to patients whose heart rates are N60 Two investigators have estimated the financial impact beats per minute. This premedication significantly decreases of using MDCT in ED patients with chest pain. Nagurney heart rates and improves images [34]. et al [50] performed an analysis that showed that MDCT In all of the 64-slice scan studies except that of Leschka significantly changed the posttest probability of low-risk and colleagues, patients were premedicated with b-blockers. ED patients with chest pain with and without coronary Leschka et al noted that 2 significant lesions were missed artery disease. This would have resulted in 13 of 35 fewer because of motion artifact associated with increased heart admissions using MDCT. Khare [51] found that an rate [32]. Mollet et al premedicated patients with a heart rate N MDCT-only strategy was less costly yet more effective 70, but motion artifacts still accounted for 60% of the than 3 other strategies (electrocardiogram stress observa- images rated as poor [33]. Raff et al noted that sensitivity tion unit, echocardiogram stress observation unit, or and specificity dropped to 88% and 71%, respectively, for N enzyme testing without observation unit) based on patients with a heart rate 70 [35]. marginal cost-effectiveness. However, although motion artifact and increased heart rate accounted for a high proportion of unevaluable images, these represent a minority of the total number of studies. In 4.3. MDCT technical limitations one study, 6 of 9 patients with heart rates N70 could still be Use of MDCT requires knowledge of its technical imaged without deleterious effect [36]. Thus, although it considerations to optimize results. As with other contrast- seems prudent to premedicate patients without contra- enhanced CT scans, the timing of dye injection is critical to indications before 64-slice MDCT, increased heart rate is highlight the areas of interest. The study is contraindicated not an absolute contraindication to MDCT. in patients with contrast dye allergy or renal insufficiency. 4.3.3. Obesity Patients need to be able to lie still and hold their breath for Because of greater tissue radiation interference, obese the duration of the study. An additional limitation to the patients are more likely to have lower-quality CT scan technology is that the patient must have a regular cardiac images. One study found that although patients in the rhythm. Patients with irregular rhythms have been excluded lowest body mass index (BMI) category had significantly from studies on MDCT [32-36]. 4.3.1. Radiation exposure One of the majors concerns regarding MDCT is Table 3 Reporting criteria for future MDCT scan studies radiation exposure. Because of the low pitch and thin Population slice thickness, the amount of radiation to which the % CAD patient is exposed is greater than that with conventional BMI/rates of obesity CT scanning. Doses of radiation exposure range from Heart rate Calcium score 3.9 to 10.1 mSv, which is approximately equal to the Inpatient vs outpatient vs ED radiation of a sestamibi scan (approximately 8-10 mSv) Traditional cardiac risk factors [16,25,52,53]. One method for limiting dose is using an Inclusion criteria electrocardiogram-dependent tube current modulation in Stents/CABG excluded? which the tube pulses radiation triggered by R waves Size of vessels evaluated detected by electrocardiogram [54]. This method was used Premedication regimen by Ropers and colleagues and resulted in radiation doses Vessel-based vs patient-based unit of analysis averaging 7.45 mSv for men and 10.24 mSv for women What is comparison reference standard? [55]. However, this technique does not allow retrospective When was reference standard performed relative to MDCT? reconstruction of images in additional phases of the cardiac Was MDCT reader blinded to reference standard? cycle and will result in suboptimal image quality whenever Interobserver j reported? How are unevaluable segments handled? there is a slight arrhythmia. In our experience, retrospec- What constitutes a positive rating for CAD on MDCT? tive reconstruction in systolic phases is useful in approx- CABG indicates coronary artery bypass grafts. imately 20% of patients. 456 A.T. Limkakeng et al. higher-quality images, there were no differences in likely become a part of the standard workup of ED patients sensitivity and specificity among groups [20].Fifty with acute chest pain, possibly allowing immediate percent of patients in the study of Raff et al had a BMI discharge of patients found to be free of stenosis. In z30 kg/m2. For these patients, sensitivity and specificity addition, important prognostic information such as the dropped to 90% and 86%, respectively. These patients level of plaque calcification, the degree of stenosis, and the accounted for all but one of their inaccurate results [35]. ejection fraction can be determined [11,15,16]. Other studies failed to systematically report the effect of Unlike cardiac catheterization, other intrathoracic dis- obesity on image quality. eases causing chest pain in the undifferentiated ED patient can be identified [40,46]. The cost of the study (approx- 4.3.4. Coronary calcification imately $2000) is considerably less than cardiac catheter- Heavily calcified coronary vessels can present a problem ization [8,11]. Furthermore, MDCT adapts familiar forCTscanimagesbycreatingbloomingandbeam technology that is more widely available in EDs than hardening artifact. Blooming refers to the appearance that modalities such as stress testing. It is possible that MDCT the calcified plaque is larger than its true size. Circumfer- may prove to be more sensitive for detecting coronary ential calcified plaque, for example, may appear to occlude artery disease lesions than cardiac catheterization and may the lumen of a patent vessel. Beam hardening refers to become the new gold standard for detection of coronary artifacts in the CT image adjacent to dense materials. There artery disease. is often a dark area adjacent to calcified plaque that is Before this occurs, more ED-based MDCT scan studies related to beam hardening, and that mimics the appearance are needed to determine the feasibility and utility of this of soft plaque. The amount of calcification a patient technology (Table 3). As more EDs upgrade their CT demonstrates on CT scan can be quantified using a system technology and more radiologists become trained to read called the Agatston score [57,58]. Calcifications accounted the studies, more data will emerge. Many studies on the use for 32% of unevaluable images in one series. This study of MDCT in ED populations are currently under way. Ideal found that by limiting analysis to patients with Agatston studies would include a representative ED population of score equivalents b1000, the sensitivity increased from 72% symptomatic patients and follow them for a period of time to 98% [21]. Another study increased sensitivity from 59% to correlate MDCT results with patient outcomes. They to 93% in similar fashion [51]. would also report the rates of inaccurate results due to The presence of calcium within the coronary circulation increased heart rates, obesity, and coronary calcifications. remains problematic despite advances in 64-slice CT Cost analysis is an additional, important area for future scanners. Ropers et al found that 64% of unevaluable investigation. If these studies continue to show such segments were due to calcifications [34]. Leschka and promising results, then MDCT will truly become the future colleagues reported that half of the vessels segments they of ED cardiac care. imaged had calcifications, 18% severe enough to cause artifacts. This resulted in 8 false-negative and all 24 of their false-positive readings [32]. Mollet et al found a 5.8% false- References positive reading rate in vessels of patients with a calcium score N400 Agatston units compared with 3.2% in other [1] McCaig LF, Burt CW. National Hospital Ambulatory Medical Care patients [33]. Raff et al noted that specificity dropped to 67% Survey: 2003 emergency department summary. Advance data from N vital and health statistics. no. 358. Hyattsville (MD)7 National Center for patients with a calcium score 400 Agatston units [35]. for Health Statistics; 2005. New techniques are under development that use dual-energy [2] Goldman L, Cook EF, Brand DA, et al. A computer protocol to predict CT to differentiate calcium from true vessel lumen and to myocardial infarction in emergency department patients with chest limit the artifact resulting from coronary calcium. However, pain. N Engl J Med 1988;318(13):797-803. until these new technologies are implemented, CT images of [3] Goldman L, Weinberg M, Weisberg M, et al. A computer-derived protocol to aid in the diagnosis of emergency room patients with acute patients with a high level of calcifications should be chest pain. N Engl J Med 1982;307(10):588-96. interpreted with caution despite the increased visualization [4] Pope JH, Ruthazer R, Beshansky JR, et al. Clinical features of accomplished with 64-slice CT scans. emergency department patients presenting with symptoms suggestive of acute cardiac ischemia: a multicenter study. J Thromb Thrombol- ysis 1998;6(1):63-74. [5] Jayes Jr RL, Beshansky JR, D’Agostino RB, et al. Do patients’ 5. Conclusions coronary risk factor reports predict acute cardiac ischemia in the emergency department? A multicenter study. J Clin Epidemiol We believe MDCT will be a vital tool to the emergency 1992;45(6):621-6. physician. The current generation of MDCT scanners has [6] Swap CJ, Nagurney JT. Value and limitations of chest pain history in been able to visualize coronary vessel luminal stenosis at a the evaluation of patients with suspected acute coronary syndromes. JAMA 2005;294(20):2623-9. level comparable with cardiac catheterization. Given its [7] Limkakeng Jr A, Gibler WB, Pollack C, et al. Combination of Goldman high negative predictive value for coronary artery disease risk and initial cardiac troponin I for emergency department chest pain and the rapidity with which it can be performed, it will patient risk stratification. Acad Emerg Med 2001;8(7):696-702. Sixty-four–slice multidetector computed tomography 457

[8] American Heart Association C. Heart disease and stroke statistics— [27] Nieman K, Cademartiri F, Lemos PA, et al. Reliable noninvasive 2006 update. Dallas (TX)7 American Heart Association; 2006. coronary angiography with fast submillimeter multislice spiral [9] Pope JH, Selker HP. Acute coronary syndromes in the emergency computed tomography. Circulation 2002;106(16):2051-4. department: diagnostic characteristics, tests, and challenges. Cardiol [28] Ropers D, Baum U, Pohle K, et al. Detection of coronary artery Clin 2005;23(4):423-51. stenoses with thin-slice multi-detector row spiral computed tomo- [10] Lee TH, Rouan GW, Weisberg MC, et al. Clinical characteristics and graphy and multiplanar reconstruction. Circulation 2003;107(5): natural history of patients with acute myocardial infarction sent home 664-6. from the emergency room. Am J Cardiol 1987;60(4):219-24. [29] Kefer J, Coche E, Legros G, et al. Head-to-head comparison of three- [11] Schussler JM, Dockery WD, Moore TR, et al. Computed tomographic dimensional navigator-gated magnetic resonance imaging and 16-slice coronary angiography: experience at Baylor University Medical computed tomography to detect coronary artery stenosis in patients. Center/Baylor Jack and Jane Hamilton Heart and Vascular Hospital. J Am Coll Cardiol 2005;46(1):92-100. Proc (Bayl Univ Med Cent) 2005;18(3):228-33. [30] Achenbach S, Moselewski F, Ropers D, et al. Detection of calcified [12] Bashore TM, Bates ER, Berger PB, et al. American College of and noncalcified coronary atherosclerotic plaque by contrast-en- Cardiology/Society for Cardiac Angiography and Interventions hanced, submillimeter multidetector spiral computed tomography: a clinical expert consensus document on cardiac catheterization segment-based comparison with intravascular ultrasound. Circulation laboratory standards. A report of the American College of Cardiology 2004;109(1):14-7. task force on clinical expert consensus documents. J Am Coll Cardiol [31] Garcia MJ, Lessick J, Hoffmann MH, CATSCAN Study Inves- 2001;37(8):2170-214. tigators. Accuracy of 16-row multidetector computed tomography [13] Scanlon PJ, Faxon DP, Audet AM, et al. American College of for the assessment of coronary artery stenosis. JAMA 2006;296(4): Cardiology/American Heart Association guidelines for coronary 403-11. angiography: executive summary and recommendations. A report of [32] Leschka S, Alkadhi H, Plass A, et al. Accuracy of MSCT coronary the American College of Cardiology/American Heart Association task angiography with 64-slice technology: first experience. Eur Heart J force on practice guidelines. Circulation 1999;99:2345-57. 2005;26(15):1482-7. [14] Beckmann EC. CT scanning the early days. Br J Radiol [33] Mollet NR, Cademartiri F, van Mieghem CA, et al. High-resolution 2006;79(937):5-8. spiral computed tomography coronary angiography in patients [15] Schoenhagen P, Stillman AE, Halliburton SS, et al. CT of the heart: referred for diagnostic conventional coronary angiography. Circula- principles, advances, clinical uses. Cleve Clin J Med 2005;72(2): tion 2005;112(15):2318-23. 127-38. [34] Ropers D, Rixe J, Anders K, et al. Usefulness of multidetector row [16] de Roos A, Kroft LJ, Bax JJ, et al. Cardiac applications of multislice spiral computed tomography with 64- Â 0.6-mm collimation and computed tomography. Br J Radiol 2006;79(937):9-16. 330-ms rotation for the noninvasive detection of significant coronary [17] Ritman EL. Cardiac computed tomography imaging: a history and artery stenoses. Am J Cardiol 2006;97(3):343-8. some future possibilities. Cardiol Clin 2003;21(4):491-513, vii. [35] Raff GL, Gallagher MJ, O’Neill WW, et al. Diagnostic accuracy of [18] Achenbach S, Giesler T, Ropers D, et al. Detection of coronary artery noninvasive coronary angiography using 64-slice spiral computed stenoses by contrast-enhanced, retrospectively electrocardiographical- tomography. J Am Coll Cardiol 2005;46(3):552-7. ly-gated, multislice spiral computed tomography. Circulation [36] Leber AW, Knez A, von Ziegler F, et al. Quantification of 2001;103(21):2535-8. obstructive and nonobstructive coronary lesions by 64-slice com- [19] Achenbach S, Ropers D, Pohle FK, et al. Detection of coronary artery puted tomography: a comparative study with quantitative coronary stenoses using multi-detector CT with 16 x 0.75 collimation and angiography and intravascular ultrasound. J Am Coll Cardiol 375 ms rotation. Eur Heart J 2005;26(19):1978-86. 2005;46(1):147-54. [20] Burgstahler C, Beck T, Kuettner A, et al. Image quality and diagnostic [37] Ellis S, Alderman E, Cain K, et al. Prediction of risk of anterior accuracy of 16-slice multidetector computed tomography for the myocardial infarction by lesion severity and measurement method of detection of coronary artery disease in obese patients. Int J Obes stenoses in the left anterior descending coronary distribution: a CASS (Lond) 2006. Registry Study. J Am Coll Cardiol 1988;11(5):908-16. [21] Heuschmid M, Kuettner A, Schroeder S, et al. Electrocardiogram- [38] Garcia M. Noninvasive coronary angiography: hype or new para- gated 16-MDCT of the coronary arteries: assessment of image quality digm? JAMA 2005;293(20):2531-3. and accuracy in detecting stenoses. AJR Am J Roentgenol [39] Sechtem U, Vohringer M. The clinical role of bnon-invasiveQ coronary 2005;184(5):1413-9. angiography by multidetector spiral computed tomography: yet to be [22] Hoffmann U, Moselewski F, Cury RC, et al. Predictive value of 16- defined. Eur Heart J 2005;26(26):1942-4. slice multidetector spiral computed tomography to detect significant [40] Gallagher MJ, Ross MA, Raff GL, Goldstein JA, O’Neill WW, obstructive coronary artery disease in patients at high risk for coronary O’Neil B. The diagnostic accuracy of 64-slice computed tomography artery disease: patient-versus segment-based analysis. Circulation coronary angiography compared with stress nuclear imaging in 2004;110(17):2638-43. emergency department low-risk chest pain patients. Ann Emerg [23] Hoffmann MH, Shi H, Schmitz BL, et al. Noninvasive coronary Med 2006;49(2):125-36. angiography with multislice computed tomography. JAMA [41] Hoffmann U, Pena AJ, Moselewski F, Ferencik M, Abbara S, 2005;293(20):2471-8. Cury RC, et al. MDCT in early triage of patients with acute chest [24] Kopp AF, Kuttner A, Heuschmid M, et al. Multidetector-row CT pain. AJR Am J Roentgenol 2006;187(5):1240-7. cardiac imaging with 4 and 16 slices for coronary CTA and imaging of [42] Hoffmann U, Nagurney JT, Moselewski F, Pena AJ, Ferencik M, atherosclerotic plaques. Eur Radiol 2002;12(Suppl 2):S17-S24. Chae CU, et al. Coronary multidetector computer tomography in [25] Kuettner A, Beck T, Drosch T, et al. Diagnostic accuracy of assessment of patients with acute chest pain. Circulation 2006 noninvasive coronary imaging using 16-detector slice spiral computed [114 article in press available via http://circ.ahajournals.org/rapidac- tomography with 188 ms temporal resolution. J Am Coll Cardiol cess.shtml accessed 11/14/06]. 2005;45(1):123-7. [43] White CS, Kuo D, Kelemen M, et al. Chest pain evaluation in the [26] Kuettner A, Trabold T, Schroeder S, et al. Noninvasive detection of emergency department: can MDCT provide a comprehensive evalu- coronary lesions using 16-detector multislice spiral computed ation? AJR Am J Roentgenol 2005;185(2):533-40. tomography technology: initial clinical results. J Am Coll Cardiol [44] Ghersin E, Litmanovich D, Dragu R, et al. 16-MDCT coronary 2004;44(6):1230-7. angiography versus invasive coronary angiography in acute chest pain 458 A.T. Limkakeng et al.

syndrome: a blinded prospective study. AJR Am J Roentgenol [52] Achenbach S, Ropers D, Regenfus M, et al. Noninvasive coronary 2006;186(1):177-84. angiography by magnetic resonance imaging, electron-beam comput- [45] Chase M, Brown AM, Robey JL, et al. Clinical implementation of ed tomography, and multislice computed tomography. Am J Cardiol computed tomography in emergency department patients with low 2001;88(2A):70E-3E. risk chest pain. Acad Emerg Med 2006;13(5):s103-4 [abstract]. [53] Schroeder S, Kopp AF, Kuettner A, et al. Influence of heart rate on [46] Savino G, Herzog C, Costello P, et al. 64 slice cardiovascular CT in the vessel visibility in noninvasive coronary angiography using new emergency department: concepts and first experiences. La Radiologia multislice computed tomography: experience in 94 patients. Clin Medica 2006;111(4):481-96. Imaging 2002;26(2):106-11. [47] Moloo J, Pena AJ, Nichols JH, et al. Multidetector computed [54] Kuettner A, Kopp AF, Schroeder S, et al. Diagnostic accuracy of tomography (MDCT) coronary angiography vs. myocardial perfusion multidetector computed tomography coronary angiography in patients imaging for early triage of patients with suspected acute coronary with angiographically proven coronary artery disease. J Am Coll syndrome. Acad Emerg Med 2006;13(5):s104 [abstract]. Cardiol 2004;43(5):831-9. [48] Amsterdam EA, Kirk JD, Diercks DB, et al. Immediate [55] Ropers D, Ulzheimer S, Wenkel E, et al. Investigation of aortocoro- exercise testing to evaluate low-risk patients presenting to the nary artery bypass grafts by multislice spiral computed tomography emergency department with chest pain. J Am Coll Cardiol 2002; with electrocardiographic-gated image reconstruction. Am J Cardiol 40:251-6. 2001;88(7):792-5. [49] Romey A, Ross MA, Gallagher M, et al. Noncardiac findings by [56] Hoffmann MH, Shi H, Manzke R, et al. Noninvasive coronary multislice computed tomographic coronary angiography aid patient angiography with 16-detector row CT: effect of heart rate. Radiology diagnosis. Acad Emerg Med 2006;13(5):s189 [abstract]. 2005;234(1):86-97. [50] Nagurney JT, Moselewski F, Pena AJ, et al. Multidetector computed [57] Agatston AS, Janowitz WR, Hildner FJ, et al. Quantification of tomography of the coronary arteries improves path probabilities in coronary artery calcium using ultrafast computed tomography. J Am emergency department patients being admitted with chest pain. Acad Coll Cardiol 1990;15(4):827-32. Emerg Med 2006;13(5):s104. [58] Detrano R, Hsiai T, Wang S, et al. Prognostic value of coro- [51] Khare R. Cost-effectiveness decision analysis model comparing nary calcification and angiographic stenoses in patients under- 64-slice computed tomography with other means of evaluating chest going coronary angiography. J Am Coll Cardiol 1996;27(2): pain. Acad Emerg Med 2006;13(5):S105. 285-90. American Journal of Emergency Medicine (2007) 25, 459–463

www.elsevier.com/locate/ajem

Diagnostics Broad complex atrial fibrillation

Huck Chin Chew*, Swee Han Lim

Department of Emergency Medicine, Singapore General Hospital, Singapore 169608, Singapore

Received 3 June 2006; revised 18 June 2006; accepted 14 October 2006

Abstract The management of broad complex atrial fibrillation is complex and may be a source of morbidity and mortality if not correctly recognized and treated appropriately. We present a case series of 3 patients who were managed in our emergency department after complaints of palpitations. They presented with varying forms of rapid atrial fibrillation that had broad complexes on the 12-lead electrocardiogram. The first 2 patients were treated with calcium channel blockers for rate control, and treatment was complicated by rapid arrhythmia that required cardioversion. The final patient was correctly treated with intravenous procainamide. The diagnosis of Wolff-Parkinson-White syndrome was eventually made in all these patients. Broad complex atrial fibrillation must be treated with respect. Cases with rapid ventricular rate can decompensate from mismanagement due to poor ability to recognize the possibility of Wolff-Parkinson-White syndrome in such patients. Procainamide forms the cornerstone of treatment in hemodynamically stable rapid broad complex atrial fibrillation of unknown origin. D 2007 Elsevier Inc. All rights reserved.

1. Clinical presentation 2. Discussion

This 30-year-old man with a history of recurrent Wolff-Parkinson-White (WPW) syndrome was first palpitations presented to the emergency department with described in 1930 by Louis-Wolff, Sir John Parkinson, palpitations and shortness of breath. On examination, his and Paul Dudley White [1]. In these patients, instead of vital signs were stable; but he was noted to be tachycardic, normal conduction from the atria to the ventricles via the with a pulse rate of 150/min. The following 12-lead atrioventricular (AV) node–His-Purkinje system, an acces- electrocardiogram (ECG) was done Figs. 1-3. sory pathway directly connects the atria and ventricles, This 50-year-old patient presented with shortness of bypassing the AV node. Earlier activation (pre-excitation) of breath with palpitations. He denied any history of note. the ventricles occurs in those who have AV conduction via Clinical examination showed that he was alert and the bypass tract (the bundle of Kent). This accessory hemodynamically stable, with a pulse rate of 140/min that pathway may conduct rapidly and recovers its excitability was noted to be irregular. There was no evidence of heart rapidly because of its short refractory period, resulting in failure. The following ECG was done Figs. 4 and 5. rapid arrhythmias (Table 1). Different types of arrhythmias occur with WPW syndrome, although the incidence is rare (about 1%-2% of * Corresponding author. Tel.: +65 98560027; fax: +65 64624178. patients) [2,3]. Approximately 80% of arrhythmias are AV E-mail address: [email protected] (H.C. Chew). reentrant tachycardia; 15% to 30%, atrial fibrillation; and

Fig. 1 This ECG shows an irregular broad complex tachycardia at 170/min with varying R-R intervals and absence of P waves.

5%, atrial flutter [4]. Rapid atrial fibrillation may degenerate in the penetration of the accessory pathway by into ventricular fibrillation because of rapid conduction. normally conducted beats. Ventricular fibrillation in WPW syndrome results from rapid ventricular response during atrial fibrillation. This is All these result in fractionation of the ventricular wave especially so if the accessory pathway has a very short front, and ventricular fibrillation occurs. As such, verap- antegrade refractory period (b250 sec) [5]. amil is contraindicated in the immediate termination of Drug therapy may contribute to the risk for ventricular atrial response in patients with WPW. Other drugs that fibrillation in patients with pre-excitation syndromes. should be avoided include h-blockers, adenosine [8], and Intravenous verapamil can increase the ventricular response digoxin. Digoxin is associated with degeneration of atrial to atrial fibrillation and has resulted in ventricular fibrilla- fibrillation into ventricular fibrillation in patients with pre- tion in some patients. This is due to 2 mechanisms [6,7]. excitation syndromes [9]. The shortening of the atrial and accessory pathway refractory period plus the increasing AV 1. The increased sympathetic discharge after hypoten- nodal block causes a decrease in concealed retrograde sion, produced by verapamil-induced vasodilation, penetration of the accessory pathway by normal conducted enhances accessory pathway conduction. beats, hence preventing its inactivation [10]. Lignocaine 2. Calcium channel blocker directly slows AV nodal has also been associated with degeneration into ventricular conduction and increases its refractoriness, resulting fibrillation [11].

Fig. 2 The patient was given intravenous verapamil for rate control, and this subsequent ECG shows the worsening of the arrhythmia with increasing tachycardia at 200/min. The R-R intervals are still irregular, and no P waves are seen. Broad complex atrial fibrillation 461

Fig. 3 The patient had to undergo cardioversion after he became hemodynamically unstable. This ECG after cardioversion shows a return to normal sinus rhythm. In addition, features suggestive of WPW syndrome are seen, with shortened PR interval (0.08 second) and slurring upslope of the delta wave.

Fig. 4 This ECG shows an irregular broad complex tachycardia with rate at 180/min. The QRS complex measures 0.2 second, P waves are absent, and the R-R intervals are irregular. The managing physician suspected underlying possible WPW syndrome. The patient was given intravenous procainamide. The following ECG was done after conversion.

Fig. 5 This is the ECG done after conversion with procainamide. It shows a return to normal sinus rhythm and features suggestive of WPW, with shortened PR interval at 0.06 second and delta waves noted best in leads V4 through V6. 462 H.C. Chew, S.H. Lim

Table 1 Electrocardiographic features of WPW syndrome ECG component Features Etiology PR interval Short (b0.12 s) Rapid AV conduction through accessory pathway and bypass of AV node QRS interval Prolonged QRS (N0.12 s); slurring of Slow muscle fibre–to–muscle fibre conduction upstroke of QRS complex (delta wave) with early ventricular activation due to pre-excitation QRS interval Prolonged QRS (N0.12 s); subsequent Transmission of conduction through the normal upstroke of QRS complex AV node and the infranodal conduction system to the ventricle

Procainamide slows conduction and prolongs refractori- 3. Conclusion ness in atrial and ventricular myocardium, accessory path- ways, and the His-Purkinje system, although it has no effect Broad complex atrial fibrillation must be treated with on or causes slight shortening of AV nodal refractory period. respect. Cases with rapid ventricular rate can decompensate It is the drug of choice in a wide QRS complex tachycardia from mismanagement due to poor ability to recognize the due to functional or preexisting chronic bundle-branch possibility of WPW in such patients. Procainamide forms block. Intravenous procainamide is the safest drug to the cornerstone of treatment in hemodynamically stable administer for the immediate treatment of a wide QRS rapid broad complex atrial fibrillation of unknown origin. complex tachycardia of unknown etiology [12]. Amiodarone is a second-line choice in the immediate treatment of hemodynamically stable broad complex atrial References fibrillation. It slows the ventricular rate as a result of its effect on accessory pathway refractoriness and conduction [13]. [1] Wolff L, Parkinson J, White PD. Bundle-branch block with short P-R interval in healthy young people prone to paroxysmal tachycardia. Am Table 2 summarizes the distinguishing features of WPW Heart J 1930;5:685. syndrome associated atrial fibrillation (AFib) and non– [2] Fitzsimmons PJ, McWhirter PD, Peterson DW, Kruyer WB. WPW AFib. The natural history of Wolff-Parkinson-White syndrome in 228

Table 2 Distinction between WPW AFib and non–WPW AFib WPW AFib Non–WPW AFib 12-lead ECG features QRS complexes are broad (N0.12 sec). QRS complexes may be broad or narrow. Heart rate potentially higher, with rates Heart rate is usually slower because the often in the 250/min range or higher individual is protected from exceptionally because of shorter anterograde refractory high ventricular rates by relatively long period in the accessory pathway allowing refractory period of the AV node. for much faster transmission of impulses and correspondingly higher rates. The shortened refractory period and No risk for of ventricular increase in the ventricular rate may result fibrillation in ventricular fibrillation. Medical therapy Treatment principle in WPW AFib is to Goal of treatment of non–WPW AFib is to prolong anterograde refractory period of slow the refractory period of the AV node. accessory pathway relative to the AV node. This slows the rate of impulse transmission through the accessory pathway and, thus, the ventricular rate. Drugs that prolong the refractory period of Treated in the conventional manner by the AV node (eg, calcium channel blockers, drugs that prolong the refractory period of h-blockers, digoxin) increases the rate of the AV node (eg, calcium channel blockers, transmission through the accessory pathway, h-blockers, digoxin) with a corresponding increase in ventricular rate. This can possibly cause the arrhythmia to deteriorate into ventricular fibrillation. Procainamide, which slows down conduction via the accessory pathway, forms the cornerstone of treatment. Broad complex atrial fibrillation 463

military aviators: a long-term follow-up of 22 years. Am Heart J studies correlated with plasma verapamil concentrations. Ann Intern 2001;142:530. Med 1980;93:682. [3] Leitch JW, Klein GJ, Yee R. Prognostic value of electrophysiologic [8] Belardinelli L, Berne RM. The cardiac effects of adenosine. Prog testing in asymptomatic patients with Wolff-Parkinson-White pattern. Cardiovasc Dis 1989;32:73. Circulation 1990;82:1718. [9] Sellers Jr TD, Bashore TM, Gallagher JJ. Digitalis in the pre-excitation [4] Pappone C, Santinelli V, Rosanio S, et al. Usefulness of invasive syndrome. Analysis during atrial fibrillation. Circulation 1977;56:260. electrophysiologic testing to stratify the risk of arrhythmic events in [10] Smith TW. Digitalis: mechanisms of action and clinical use. N Engl J asymptomatic patients with Wolff-Parkinson-White pattern. Results Med 1988;318:358. from a large prospective long-term follow-up study. J Am Coll Cardiol [11] Akhtar M, Gilbert CJ, Shenasa M. Effect of lidocaine on atrioven- 2003;41:239. tricular response via the accessory pathway in patients with Wolff- [5] Klein GJ, Bashore TM, Sellers TD, et al. Ventricular fibrillation in the Parkinson-White syndrome. Circulation 1981;63:435. Wolff-Parkinson-White syndrome. N Engl J Med 1979;301:1080. [12] Mandel WJ, Laks MM, Obayashi K, et al. The Wolff-Parkinson-White [6] Rinkenberger RL, Prystowsky EN, Heger JJ, et al. Effects of syndrome: pharmacologic effects of procainamide. Am Heart J 1975; intravenous and chronic oral verapamil administration in patients 90:744. with supraventricular tachyarrhythmias. Circulation 1980;62:996. [13] Gomes JA, Kanf PS, Hariman RJ, et al. Electrophysiologic effects and [7] Sung RJ, Elser B, McAllister Jr RG. Intravenous verapamil for mechanisms of termination of supraventricular tachycardia by termination of reentrant supraventricular tachycardias. Intracardiac intravenous amiodarone. Am Heart J 1984;107:214. American Journal of Emergency Medicine (2007) 25, 464–471

www.elsevier.com/locate/ajem

Diagnostics The VIDAS D-dimer test for venous thromboembolism: a prospective surveillance study shows maintenance of sensitivity and specificity when used in normal clinical practice

David Mountain FACEM*, Ian Jacobs PhD, Andrew Haig FACEM

Department of Emergency Medicine, Sir Charles Gairdner Hospital, Perth, Australia Department of Academic Emergency Medicine, University of Western Australia, Perth, Australia Department of Emergency Medicine, Royal Perth Hospital, Perth, Australia

Received 26 June 2006; received in revised form 18 September 2006; accepted 25 September 2006

Abstract Background: As a result of a number of clinical management studies, D-dimer (DD) tests such as VIDAS (BioMe´rieux Australia P/L-Sydney, NSW) have been recommended to reduce venous thromboembolism (VTE) investigations. Surveillance studies for new tests are recommended. We prospectively assessed VIDAS DD in normal practice. Methods: Consecutive emergency patients and inpatients (IPs) with DD or VTE investigations were prospectively identified. Investigation results and early chart review including predefined factors reducing specificity were documented. A latex DD was also performed. Patients were followed for at least 3 months for recurrent VTE. Results: Four hundred three patients (emergency, 64%; VTE-positive, 12%; 95% followed up) were analyzed. VIDAS sensitivity was 96% (95% confidence interval 86%-99%), specificity 38% (confidence interval, 34%-44%; negative likelihood ratio, 0.11), and emergency specificity 51%. Latex sensitivity was 76%. Cancer, trauma, recent operations, IP status, and advanced age were associated with markedly reduced specificity. Specificity in older emergency patients (N70 years old) and younger IPs (b70) without comorbidities was 20% to 30%, but sensitivity was maintained at 100%. Conclusions: VIDAS DD probably maintains adequate sensitivity in normal clinical practice for low- or even intermediate-risk patients. Latex agglutination had poor sensitivity. Specificity is best in younger low- morbidity emergency patients. These findings need validation in larger multicenter surveillance studies. D 2007 Elsevier Inc. All rights reserved.

Data in this study have been presented as: Mountain D, Haig A. Presentation. A prospective observational study in an unselected population of the VIDAS D-dimer (VIDAS DD) test for excluding acute venous thrombo-embolic (VTE) disease shows retention of sensitivity and specificity in a real environment. ACEM/ASEM joint scientific meeting, November 2000, Canberra, Australia. * Corresponding author. Department of Emergency Medicine, Sir Charles Gairdner Hospital, Hospital Ave, Nedlands, Perth WA, Australia 6009. Tel.: +61 8 93463333; fax: +61 8 93463516. E-mail address: [email protected] (D. Mountain).

1. Introduction have been the most promising of these products. Standard enzyme-linked immunosorbent assays (ELISAs) have high Providing a diagnosis of pulmonary embolism (PE) and sensitivity but are logistically unsuited to clinical practice. deep venous thrombosis (DVT) clinically is difficult [1,2]. Newer rapid ELISA tests have been introduced and show Definitive diagnosis normally relies on radiologic inves- similar results to standard ELISA testing in study con- tigations. At the time of this study, recommended testing ditions [5]. often required pulmonary angiography (PA) or serial ultra- Few DD tests have been adequately tested against gold- sounds (USs). Clinical audit confirmed that compliance to standard diagnostic testing or in management studies. recommended diagnostic regimens was poor at our institu- A study of 1000 outpatients used the VIDAS DD test tion. Studies have confirmed low rates of appropriate testing (a modified nonbatched ELISA) to manage patients [6].A for PE in normal clinical practice [3,4]. Many clinicians are negative test result (confirmed by negative US finding if averse to PA because of its perceived morbidity, mortality, DVT was clinically suspected) meant the patient had no and limited availability. In current practice, multislice further investigation or therapy for VTE. Using 3-month computed tomographic pulmonary angiography (CTPA) is follow-up as part of the gold standard, VIDAS DD was 99% often used as the default investigation, but radiation dosing, sensitive. Based on this (and other studies demonstrating problems with readings (particularly by junior staff), good sensitivity against gold-standard tests), our institution dye issues, and suboptimal views in up to 15% of patients changed from using a latex agglutination (LA) DD (Agen) make universal use of this investigation problematic, to VIDAS [5,7]. particularly in low-prevalence emergency department (ED) A prospective audit was performed to demonstrate populations [5]. adequate clinical test performance and allay clinician Sensitive blood tests for excluding venous thromboem- concerns. Surveillance and clinical management studies bolism (VTE) would be useful. Most target markers of (prospective audit) have been suggested for monitoring the thrombus generation and breakdown. D-dimer (DD) tests performance and utility of diagnostic tests in normal clinical

Fig. 1 Eligible patients—outcomes using STARD criteria [14]. 466 D. Mountain et al.

were not managed at this institution; follow-up was Table 1 Excluded patients—reasons for exclusion and numbers of samples excluded from analysis impossible; or no suitable blood samples were available. Reasons for exclusions No. of samples/patient episodes excluded 3.3. Investigations/follow-up No samples 16 All patients had a VIDAS DD test performed either Repeated samples—similar results 30 clinically or on stored samples taken (within 48 hours of the Tests (DD/VQ) done for non-VTE 11 index symptoms) for other tests. VIDAS is normally diagnosis performed on citrated samples, giving quantitative results. Patient not managed at primary 15 EDTA samples were used if citrate was unavailable. Samples institution were stored as per standard laboratory protocols. Latex Patients not available to follow-up 16 Total excluded 88 agglutination (Agen) DD was performed for comparison on almost all samples (6 patients were missed because of lost samples), but the results were unavailable to clinicians. All use, similar to type IV surveillance studies for therapeutic patients were followed up until they reached defined end drugs [8,9]. points. Positive VTE was confirmed by either a high- probability VQ scan, positive CTPA, PA, or echocardiogram; deep venous thrombosis was confirmed by positive US, 2. Aims venogram, computed tomography, or any VTE at postmortem (PM). Pulmonary embolism was excluded by a negative PA The primary hypothesis for this study was that the result PM or a normal VQ scan. Deep venous thrombosis was VIDAS DD would maintain sensitivity for proven PE/DVT excluded by negative venogram or US finding (for low- or of more than 95% with adequate specificity when compared intermediate-risk DVT patients). Patients with a definitive with a gold standard. Secondary aims were to examine alternative diagnosis clearly explaining all symptoms (eg, factors affecting specificity and to compare performance chest x-ray (CXR) showing pneumothorax, US showing with the Agen LA test. ruptured Baker cyst) were considered negative for VTE. Any negative CTPA required clinical follow-up because CTPA was not considered to be adequately validated at the time. 3. Methods Patients with an indeterminate workup but clinically treated as VTE were included as positive for VTE even if DD was 3.1. Background negative to avoid overestimating sensitivity. All other patients were followed up for a minimum of The setting was a 600-bed adult teaching hospital serving 3 months by a combination of chart review (performed on a population of approximately 400000 people. Emergency all patients) and contacting general practitioners (GPs) or department attendances were 39000 yearly, with 45% the patients/relatives for telephone interview. Patients who admitted. Approximately 400 ventilation perfusion lung died had their notes reviewed (by DM) and were classified scans (VQ) were performed, with 150 VTEs diagnosed according to whether PE was likely or unlikely to have annually. At the time, a single-slice helical CTPA was only contributed to death. Comorbidities expected to affect being used occasionally and was of marginal significance to specificity, such as recent surgery or major trauma, active PE diagnosis at this center before the study ended. cancer, or severe concurrent disease (eg, sepsis or recent Investigations ordered were purely at the discretion of intensive care unit [ICU] admission) were documented for treating clinicians. The VIDAS DD test was introduced in all patients [10-12]. The ED population was compared to July 1999. Any clinician could order DD, and a negative inpatients (IPs) for demographics, outcomes, test perfor- result was considered to exclude significant VTE. Staff mance, and factors affecting specificity. members were aware of the prospective audit but did not collect data. 3.4. Data retrieval 3.2. Population All definitions (including end points) were confirmed between the investigators before data retrieval. Data were All patients who had either DD or a VQ scan ordered from July 1999 to December 1999 were included. This Table 2 Type of VTE suspected and prevalence of disease included those patients with DD but only investigated for Type of VTE suspected No. of VTE-positive/No. DVT. Coagulation laboratory and nuclear medicine depart- investigated (%) ments provided lists of patients on a daily basis (except PE 30/353 (8%) weekends). Patients were excluded for the following DVT 12/33 (36%) reasons: DD was done for a diagnosis other than VTE, for PE + DVT 7/17 (41%) example, disseminated intravascular coagulation; patients VIDAS D-dimer is sensitive for DVT/PE in usual practice 467

Table 3 Population characteristics—total and for selected subgroups Total VTE- VTE- ED IP VIDAS- VIDAS- Latex- Latex- positive negative positive negative positive negative n (% of total population) 403 (100) 49 (12) 354 (88) 257 (64) 143 (36) 265 (66) 138 (34) 151 (38) 246 (62) Male (% of subgroup) 190 (47) 25 (51) 166 (47) 119 (46) 69 (48) 135 (51) 55 (40) 81 (54) 105 (43) Female (%) 213 (53) 24 (49 ) 188 (53) 138 (54) 74 (52) 130 (49) 83 (60) 70 (46) 141 (57) Age, mean (median) 62 (64) 61 (63) 62 (64) 60 (61) 66 (68) 67 (69) 52 (55) 66 (69) 60 (62) retrieved by chart review directly into a labeled Excel The population studied had a VTE prevalence of 12%. spreadsheet by 2 authors (DM, AH). Data included demo- Prevalence was higher in the IP group (17% vs 9%) and in graphics; tests ordered and their results; if symptoms of patients investigated for DVT (Tables 2-4). Patients with VTE were present in ED; risk factors for VTE and coexistent VTE were not significantly older than the whole population, active conditions or previous VTE; presenting features; and there was no sex bias (Table 3). Characteristics for total whether pretest probability was documented; and results of population and population subgroups are outlined in Table 3. CXR, arterial blood gas, and pulse oximetry. The authors Among the subjects, 47% were male, and mean age was performed follow-up without blinding to clinical details. 62 years. Inpatients were significantly older than ED patients (66 vs 60 years; P b .001; 95% CI for difference 3.5. Statistics in means, À10.42 to À3.25 years). It was also noted that Data were analyzed using SPSS (version 11.5; SPSS, patients with a positive VIDAS DD were much older than Chicago, Ill). Frequency counts and proportions were those with a negative result (67 vs 52 years). The VIDAS calculated for all categorical data and primary end points. test had a sensitivity of 96% and a specificity of 38%, giving Because of the skewed nature of the distribution, medians a negative LR of 0.11 (Table 5). Agen LA test showed a were calculated for age. Sensitivity, specificity, and positive sensitivity of 79% and a specificity of 66% in the 397 (of and negative likelihood ratios (LRs) were all calculated with 403) patients who had latex performed as well (2 Â 2 data 95% confidence intervals (CIs) derived for these values. withheld for brevity—available on request from corresponding author). When death or VTE was used as 3.6. Ethics the end point of interest, sensitivity of VIDAS was maintained at 97% (95% CI, 95%-99%). This was true The study was presented to the chair of the ethics even if patients with other causes for death or minimal risk committee. It was exempted from ethics review because it of VTE death were excluded (see Table 6). was performed as a prospective clinical audit and had no The overall specificity of the VIDAS test was 38% interventions apart from patient contact for follow-up. (Table 5), 51% in ED and 15% for IP (Table 4). Prevalence of factors suspected to affect specificity and the likelihood 4. Results of a positive VIDAS test are detailed in Table 7. All of these factors, including age of older than 70 years, cancer, recent There were 491 consecutive episodes (see Fig. 1 for operations or trauma, or recent ICU admission, were Standards for Reporting of Diagnostic Accuracy [STARD] associated with poor specificity of the VIDAS test. In outcome details) reviewed, with 88 excluded for the reasons patients older than 70 years, 16/84 (19%) of ED patients had shown in Table 1 [13]. Of the 16 lost to follow-up, 8 were a negative VIDAS compared with 1/65 (2%) of IPs. In older subsequently found at a later review (N2 years) to have been ED patients, 16/74 (22%) had negative VIDAS if other free of VTE. However, these additional negative follow-ups factors associated with raised DD were removed (eg, were not included in the statistical analysis. 10 were postoperative or had cancer). In IPs younger than

Table 4 Emergency vs IP VIDAS test performance ED (prevalence—9%) IP (prevalence—17%) Condition present (24) Absent (235) Condition present (25) Absent (119) Test-positive 22 116 25 102 Test-negative 2 119 0 17 Sensitivity (95% CI) 92% (74%-98%) 100% (84%-100%) Specificity (95% CI) 51% (45%-57%) 15% (9%-22%) Negative LR (95% CI) 0.16 (0.04-0.62) not applicable Positive LR (95% CI) 1.86 (1.56-2.23) not applicable 468 D. Mountain et al.

prospective investigation of pulmonary embolism diagnosis Table 5 VIDAS test performance—2 Â 2 table [PIOPED] criteria) and was treated as PE. The treating VTE-positive VTE-negative respiratory consultant was not certain of the diagnosis of (49) (354) acute PE when contacted for follow-up. She probably had Test-positive (265) TP 47 FP 218 intermediate risk of PE before VQ and had a posttest Test-negative (138) FN 2 TN 136 probability (by PIOPED criteria) of 30% to 50% [1]. Sensitivity (95% CI) 96% (86%-99%) The second patient had had a previous PE/DVT and was a Specificity (95% CI) 38% (34%-44%) smoker on the pill with 7 days of leg pain and 1 day of Negative LR (95% CI) 0.11 (0.03-0.42) Positive LR (95% CI) 1.56 (1.4-1.7) pleuritic chest pain, and dyspnea with oxygen saturation of 99% on room air. She was assigned a high pretest probability TP indicates true positives; TN, true negatives; FP, false positives; and FN, false negatives. of PE. D-dimer was negative and VQ intermediate. No other tests were performed, and she was anticoagulated. By PIOPED criteria, she had a 55% chance of PE [1]. 70 years and without raised specificity associated comorbidities, 16/54 (30%) had negative test results. 5. Discussion 4.1. Follow-up and VTE testing after DD This study looks at some key issues for real-life Of 354 patients without confirmed VTE, 47 died implementation of a new DD test (eg, in unselected (Table 6). Of the 47 deaths, 45 had a positive VIDAS. Of patients for whom any clinician, of any experience, can the 2 patients with a negative DD, one patient was order the test). Firstly, does that test retain sensitivity when definitively without PE at PM (died of acute myocardial performed outside of trial situations? Secondly, for which infarction secondary to gastrointestinal bleeding). The populations does the test have adequate specificity (utility)? second patient had known pulmonary hypertension due to Finally, how does LA perform against ELISA (VIDAS) in atrial septal defect (ASD) shunt and died during admission. these conditions? Pulmonary embolism could not be definitively excluded in We have found that the sensitivity was maintained at this case (clinical risk low/VQ intermediate). 96%, although with a lower CI of 86%. Negative LR (0.11) There were 76 patients who had a positive DD test but no suggests a moderate-to-large reduction in pretest probability further investigations for PE. All of these patients were for a negative VIDAS. Confidence intervals are wide followed up clinically without a VTE event. Of 138 patients because of the low numbers of confirmed VTE. This is a with a negative DD, 28 patients had other investigations for feature of populations weighted to outpatients [14,15].In PE. Ten had definitive testing excluding PE (8 normal VQ, addition, this was not a study protocol, and, therefore, 1 alternative diagnosis on CTPA, and 1 negative PA), 2 with ordering was allowed for any patient with any concern for indeterminate investigations were deemed positive for VTE VTE. In many studies, there is a selection bias to high- (clinical details of cases follow), and the rest had indeter- prevalence populations, probably related to strict inclusion minate testing with negative follow-up. Patients positive for criteria and the reluctance of physicians to enter patients PE had the following positive confirmatory tests: 19 high- with low risk of VTE into trials with invasive diagnostic probability VQ, 7 CTPA, 3 PA, 14 US, and 4 VTE picked tests [16]. on other imaging (echocardiogram, computed tomography The sensitivity was not affected if patients who died of the thorax/abdomen). during follow-up (1/47 negative VIDAS without PM Both cases assigned as false-negative diagnosis of VTE by sensitivity of 97%; 95% CI, 95%-99%) were included as VIDAS did not have a definitive diagnosis of PE. The first VTE. Even if only those deemed possible or probable VTE- patient presented dyspneic with chest tightness and had a associated deaths (1/26 negative VIDAS) are included, there history of previous PE but normal arterial blood gases (Po2, is no change to the initial sensitivities for VIDAS. The same 95; Pco2, 35) and CXR. A DD was negative, and the patient was true if patients treated clinically as PE without an was discharged from ED. She continued to have symptoms adequate diagnostic workup were included because we had and was sent for a VQ scan by her GP. This showed a included the 2 of 7 with negative VIDAS and therapy for PE segmental perfusion-ventilation mismatch (intermediate by as VTE-positive in the original calculations.

Table 6 Deaths in patients with an incomplete or negative definitive initial workup Risk group for Very low PE Definitive—PE-negative Palliative care—expected Death—PE possible Unexpected death—PE PE at death risk (b10%) workup death (10%-30%) likely (N30%) n (47) 6 8 9 19 7 DD-positive (45) 6 7 9 18 7 VIDAS D-dimer is sensitive for DVT/PE in usual practice 469

Table 7 Effect of comorbidities on DD specificity Risk factors for reduced DD specificity No. with risk factor VIDAS-positive (%) Active cancer 58 (12 also postoperative, 2 in ICU) 54 (93) Postoperative/major trauma or delivery within 4/52 48 (12 also had ca, 6 in ICU) 47 (98) ICU admission within 2/52 9 (6 postoperative [2 with Ca]) 9 (100) No. of IP with at least 1 risk factor from above (total IP no.) 54 (145) 54 (100) Age N70 yr 149 131 (88) Ca, cancer.

Because of the wide CIs for sensitivity and negative LR, expected in most trials). Therefore, we would expect that the data from this study cannot by itself justify using this test this unselected population would have decreased sensitivity in normal clinical practice. However, because the sensitivity for DD tests. This has been demonstrated by the SimpliRed (and negative LR) in this unselected population is similar to test where initial sensitivities were reported as more than that achieved in previous management studies, it suggests 95% in gold-standard trials, dropping to approximately 85% that this test can safely be used in low- and probably in management studies, and have been reported as low as intermediate-risk outpatients to exclude VTE [5,6,17]. 65% in observational studies[5,7,20]. Finally, it is probable Specificity for VIDAS overall was 38%, with outpatient that some patients with VTE were missed because the (ED) specificity at 51% and IPs at 15%. This difference in minority of this population had a gold-standard diagnostic specificity is similar to findings for this and other similar workup. However, it is the ability to safely discharge DD tests [6,17]. We looked at some common conditions and patients and know that they will remain clinically disease- patient characteristics (see Table 6) that have been associ- free that clinicians want from a test [21]. In this study, a ated with high false-positive rates. It seems that these negative DD had a 99% negative predictive value for conditions (recent operations or major trauma, active cancer, excluding VTE based on follow-up or standard diagnosis. and possibly age N70 years) make specificity too low for Because not all patients had DD performed as part of their DD to be a useful screening test. However, there are some diagnostic workup, some tests were performed on stored caveats to this. We found that in outpatients older than samples or EDTA tubes. However, if there were any effect, 70 years, 16/84 (19%) had negative VIDAS. This percent- we would expect this to be reduced specificity (more false age was not changed if other risk factors were used to stop positives) from activated samples. testing (14/73 [19%]). If age of older than 70 years and The small numbers of positive patients and wide CI for prothrombotic risk factors were used to stop VIDAS testing sensitivity were inevitable because of the large numbers of in the 144 IP, 16/54 (30%) would have negative test results. patients being tested and relatively low prevalence rate. This suggests that a significant number of younger IP Logistically, we were unable to collect data for longer (b70 years) without obvious prothrombotic conditions could than 6 months, and this was the available population. usefully have DD testing. Because this is subgroup analysis Clearly, other similar observational studies or larger studies with only modest numbers, these findings should be would clarify further the real sensitivity of this test in confirmed in larger prospective cohort studies. normal practice. We compared VIDAS to the LA Agen that was Was our follow-up adequate? Follow-up was achieved in previously used and had been freely available to clinicians. a high number of cases (N96%). The follow-up was via The LA Agen had poor sensitivity (78%) but better verbal contact with GP/patient or by review of further IP specificity (67%). This is similar to previous findings for admissions after the 3-month period. Although this is not as commercial LA tests and confirms their unsuitability in real good as direct patient contact (eg, clinic review), we felt that practice unless a very low-prevalence VTE population can patients (or GPs) were unlikely to forget being diagnosed be accurately identified [2,5,7]. The latex test (Agen) had with DVT or PE and being started on anticoagulants. Chart not been used very frequently for VTE diagnosis in our review was used to extract data on readmission, symptoms, institution before changing to VIDAS probably because of and coexistent disease. Only 2 investigators reviewed the justifiable concerns about the test’s lack of sensitivity charts and used preagreed criteria for the major findings (Mr Nick Micholopoulos, coagulation laboratory senior looked for. A formal interrater reliability was not calculated scientist, personal communication). or compared to an independent third party. The investigators were not blinded to results when extracting data. Formal 5.1. Methodological concerns for this study interrater testing, blinding to results, and follow-up by an This population was unselected, and the whole range of independent reviewer would strengthen further studies in patients with VTE would be expected (excepting massive this area [22]. PE) with a smaller proportion of the larger VTEs seen in Finally, we did not control ordering practices, so that clinical trials[18,19]. There were no exclusions on the basis variations in practice and test ordering were almost certainly of delay from symptom onset in this study (as would be present between various units and seniority of staff. 470 D. Mountain et al.

However, this is how tests are ordered in reality, and it is the very unlikely to be negative because of comorbidities [15]. ability of a test to maintain performance in variable They would assist in clarifying guidelines for reducing conditions that is important for demonstration of real inappropriate testing particularly relating to underlying clinical safety and utility. illness, age, and IP status. Studies are also needed to further investigate how DD testing is integrated into clinical 5.2. Implications for current practice practice and how its use affects investigation patterns and diagnosis rates for VTE. This study suggests (but cannot, in isolation, confirm) that VIDAS maintains sensitivity when used in standard clinical conditions. Given that VIDAS (and other tests without management study tests) is already being used in 6. Conclusion clinical practice, these results improve our knowledge of We have found that in normal practice, the VIDAS test real-life performance. We can certainly suggest use in low- probably maintains adequate sensitivity for VTE with a b risk patients ( 10% risk of VTE) where a posttest negative LR of 0.11. Specificity in outpatients was probability of disease would be less than 2%. Its use in reasonable but is low in IPs. A latex DD did not perform intermediate-risk patients could not be recommended on the adequately for routine clinical usage. Other prothrombotic basis of this study, but in association with evidence from conditions (cancer, postoperative state and need for ICU, previous management studies, its use is probably reason- advanced age, and IPs) have been identified as probably able. We have found that a number of conditions or patient reducing specificity to the point of clinical futility. In characteristics reduce specificity to the point where testing combination with evidence from other studies, we suggest is unlikely to be useful. In particular, patients who are within that negative VIDAS DD is a suitable exclusionary test for 2 weeks of surgery, have major trauma, or have active ED patients with a low (and possibly intermediate) cancer have very low rates of negative VIDAS. In addition, probability of VTE. IP as a group has poor specificity (15%). Education or ordering control would need to be rigorous to select the group of younger IPs without obvious prothrombotic confounding conditions who could benefit from VIDAS Acknowledgments testing. Older patients (older than 70 years) also show poor The authors are indebted to Dr Nick Michalopoulos, specificity (particularly IP). Older patients in ED popula- senior scientist in the coagulation laboratory, and his staff, tions may have marginal utility for testing (20% negative for performance of the DD tests and the very generous test results). These findings would need to be validated donation of time and information provided to the authors. prospectively in other populations. Finally, there seems little place for older latex tests in excluding VTE unless used in well-selected, very low-risk populations. Our study suggests that this test may be overused. In References addition to the low prevalence of disease (particularly in [1] Value of the ventilation/perfusion scan in acute pulmonary embolism. outpatients ([9%]), there was evidence of indiscriminate Results of the prospective investigation of pulmonary embolism testing where a positive DD did not lead to other testing diagnosis (PIOPED). The PIOPED Investigators. JAMA 1990;263: (78 patients). This probably suggests that the clinical risk of 2753-9. PE was very low before the test was ordered or another [2] Manganelli D, Palla A, Donnamaria V, Giuntini C. Clinical features condition was diagnosed soon after. If ordering is indis- of pulmonary embolism. Doubts and certainties. Chest 1995;107: 25S-32S. criminate, there would be concerns with the cost benefit of [3] Egermayer P, Town GI. Compliance with guidelines for the the test (laboratory charges are approximately AUS $30) investigation and management of patients with suspected pulmonary and the potential for overuse of more expensive radiologic embolism at Christchurch Hospital. N Z Med J 1998;111:70-3. investigations for VTE. This has been noted in other studies, [4] Hagen PJ, van Strijen MJ, Kieft GJ, et al. The application of a Dutch although the overall effect was to increase VTE diagnosis consensus diagnostic strategy for pulmonary embolism in clinical practice. Neth J Med 2001;59:161-9. and numbers of investigations [15,23]. [5] Stein PD, Hull RD, Patel KC, et al. D-dimer for the exclusion of acute venous thrombosis and pulmonary embolism: a systematic review. 5.3. Further studies Ann Intern Med 2004;140:589-602. [6] Perrier A, Desmarais S, Miron MJ, et al. Non-invasive diagnosis of Prospective observational studies in multiple centers or venous thromboembolism in outpatients. Lancet 1999;353:190-5. over longer time frames would improve our knowledge of [7] Legnani C, Pancani C, Palareti G, et al. Contribution of a new, rapid, these tests’ safety in real life. Follow-up could be improved quantitative and automated method for D-dimer measurement to by formal clinic review at 3 months and better controls for exclude deep vein thrombosis in symptomatic outpatients. Blood Coagul Fibrinolysis 1999;10:69-74. interrater reliability. In addition, larger prospective multi- [8] Buller HR, Lensing AW, Hirsh J, ten Cate JW. Deep vein throm- center trials are needed to clarify subgroups where VTE can bosis: new non-invasive diagnostic tests. Thromb Haemost 1991;66: be clinically ruled out without a DD test or where testing is 133-7. VIDAS D-dimer is sensitive for DVT/PE in usual practice 471

[9] Sackett DL, Haynes RB. The architecture of diagnostic research. ing the Localization of Pulmonary Embolism. Thromb Haemost 2001; Br Med J 2002;324:539-41. 85:604-8. [10] Bounameaux H, Miron MJ, Blanchard J, et al. Measurement of [17] Miron MJ, Perrier A, Bounameaux H, et al. Contribution of plasma D-dimer is not useful in the prediction or diagnosis of noninvasive evaluation to the diagnosis of pulmonary embolism in postoperative deep vein thrombosis in patients undergoing total knee hospitalized patients. Eur Respir J 1999;13:1365-70. arthroplasty. Blood Coagul Fibrinolysis 1998;9:749-52. [18] Heit JA, Minor TA, Andrews JC, et al. Determinants of plasma fibrin [11] Owings JT, Gosselin RC, Anderson JT, et al. Practical utility of the D-dimer sensitivity for acute pulmonary embolism as defined by D-dimer assay for excluding thromboembolism in severely injured pulmonary angiography. Arch Pathol Lab Med 1999;123:235-40. trauma patients. J Trauma 2001;51:425-9 [discussion 29-30]. [19] De Monye W, Sanson BJ, MacGillavry MR, et al. Embolus location [12] Lee AY, Julian JA, Levine MN, et al. Clinical utility of a rapid whole- affects the sensitivity of a rapid quantitative D-dimer assay in the blood D-dimer assay in patients with cancer who present with diagnosis of pulmonary embolism. Am J Respir Crit Care Med 2002; suspected acute deep venous thrombosis. Ann Intern Med 1999;131: 165:345-8. 417-23. [20] Farrell S, Hayes T, Shaw M. A negative SimpliRED D-dimer assay [13] Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and result does not exclude the diagnosis of deep vein thrombosis or accurate reporting of studies of diagnostic accuracy: the STARD pulmonary embolus in emergency department patients. Ann Emerg initiative. BMJ 2003;326:41-4. Med 2000;35:121-5. [14] Bounameaux H, Perrier A, Wells PS. Diagnostic strategies for [21] Perrier A, Bounameaux H. Validation of helical computed tomogra- suspected pulmonary embolism among outpatients. Semin Vasc phy for suspected pulmonary embolism: a near miss? J Thromb Med 2001;1:189-94. Haemost 2005;3:14-6. [15] Kline JA, Mitchell AM, Kabrhel C, et al. Clinical criteria to prevent [22] Gilbert EH, Lowenstein SR, Koziol-McLain J, et al. Chart reviews in unnecessary diagnostic testing in emergency department patients with emergency medicine research: where are the methods? Ann Emerg suspected pulmonary embolism. J Thromb Haemost 2004;2:1247-55. Med 1996;27:305-8. [16] Hartmann IJ, Prins MH, Buller HR, Banga JD. Acute pulmonary [23] Goldstein NM, Kollef MH, Ward S, Gage BF. The impact of the embolism: impact of selection bias in prospective diagnostic studies. introduction of a rapid D-dimer assay on the diagnostic evaluation of ANTELOPE Study Group. Advances in New Technologies Evaluat- suspected pulmonary embolism. Arch Intern Med 2001;161:567-71. American Journal of Emergency Medicine (2007) 25, 472–475

www.elsevier.com/locate/ajem

Therapeutics Ultrasound-guided supraclavicular block for the treatment of upper extremity fractures, dislocations, and abscesses in the ED

Michael B. Stone MD, RDMSa,*, Daniel D. Price MDa, Ralph Wang MDb aDepartment of Emergency Medicine, Alameda County Medical Center - Highland Campus, Oakland, CA 94602, USA bDepartment of Internal Medicine, Division of Emergency Medicine, University of California San Francisco, San Francisco, CA 94143, USA

Received 27 July 2006; revised 19 August 2006; accepted 21 August 2006

Abstract Emergency department (ED) patients with fractures, dislocations, or abscesses of the upper extremities often require closed reduction or incision and drainage to treat these conditions. Procedural sedation is often necessary when infiltration of local anesthetic provides insufficient analgesia. Anesthesiologists commonly perform supraclavicular brachial plexus nerve blocks to achieve analgesia for upper extremity surgery. We report a series of 5 ED patients in whom supraclavicular brachial plexus nerve blocks using real-time ultrasound guidance provided excellent analgesia and obviated the need for procedural sedation. D 2007 Elsevier Inc. All rights reserved.

1. Introduction time ultrasound-guided supraclavicular brachial plexus nerve blocks has been reported extensively in the anesthe- Emergency physicians often encounter patients who siology literature [2-6]. These studies used both real-time require anesthesia for the treatment of acute traumatic or ultrasound guidance and nerve stimulation to confirm infectious processes. Direct infiltration of a local anesthetic needle position. agent may be insufficient to obtain adequate anesthesia for We hypothesize that real-time ultrasound-guided brachial upper extremity fractures, dislocations, and abscesses. plexus nerve blocks can be performed without nerve Although procedural sedation can facilitate the treatment stimulation and can provide an excellent alternative to of these patients, it requires patients to have fasted for 6 or procedural sedation for the management of upper extremity more hours and still involves the risk of apnea, hypotension, fractures, dislocations, or abscesses in the emergency and other adverse effects. department (ED). We report a series of 5 ED patients in Peripheral nerve blocks (of the median, ulnar, and radial whom supraclavicular brachial plexus nerve blocks were nerves) in the upper extremity are effective but require performed using ultrasound guidance. multiple injections and are unable to provide effective anesthesia proximal to the forearm [1]. The success of real- 2. Methods

* Corresponding author. Tel.: +1 510 437 4564. The procedure for ultrasound-guided supraclavicular E-mail address: [email protected] (M.B. Stone). brachial plexus nerve block was modified from the

3. Case reports

The following patients presented to our ED during a 3-month period, from October to December 2005. Each case was reviewed by the authors and met the following case definitions: (1) An upper extremity fracture, dislocation, or abscess was present; (2) the patient would typically require procedural sedation for treatment of the upper extremity condition; and (3) procedural sedation would be delayed for several hours due to preprocedural fasting or lack of available monitored beds. Sensory and motor examinations were conducted every 5 minutes; beginning at 20 minutes Fig. 1 Still image of brachial plexus and surrounding anatomy. after the injection was complete. BP indicates brachial plexus; SA, subclavian artery. 3.1. Case 1 technique originally described by Chan [4]. After written A 36-year-old man with a history of active injection drug informed consent was obtained, the supraclavicular fossa use presented with a 6 Â 5 cm right forearm abscess and a was prepared and draped in sterile fashion. A sonographic low-grade fever. Despite 10 mg of morphine IV, the patient view of the brachial plexus was obtained with a 10 to reported 10/10 pain and refused all attempts at local 5.0 MHz linear transducer oriented transversely in the infiltration of anesthetic. Procedural sedation could not be supraclavicular fossa, just above the clavicle. In this view, performed because there were no available monitored beds the brachial plexus is superficial and lateral to the in the ED. After written informed consent was obtained, a subclavian artery and is visualized as a group of hypoechoic supraclavicular brachial plexus nerve block was performed nodules (Fig. 1). Arterial flow was confirmed by pulsed using ultrasound guidance with a 27-gauge needle and wave Doppler flow to ensure the correct identification of 30 mL of lidocaine 1.0% with epinephrine. The patient the subclavian artery. A 27-gauge or a 22-gauge noncutting demonstrated full sensory and motor blockade 25 minutes spinal needle was inserted from the lateral aspect of the after injection, and incision and drainage were performed linear transducer and directed in parallel with the transducer without the need for any additional intravenous or local to allow visualization of the full length of the needle analgesia. Total time from injection to completion of throughout the procedure (Fig. 2). When the needle tip was procedure was 35 minutes. visualized adjacent to the hypoechoic nodules representing 3.2. Case 2 the brachial plexus, 30 mL of lidocaine 1% with epinephrine was instilled with frequent aspiration to avoid A 40-year-old right hand–dominant man presented to the intravascular injection. The spread of local anesthetic within ED with 1 week of right hand pain after a fight 1 week the brachial plexus was visualized as an expanding earlier. Radiographs revealed a 100% displaced, bbayonetQ hypoechoic collection within the brachial plexus. This fracture of his right third metacarpal. Despite 2 mg of technique is similar to the one described by Chan et al midazolam IV and 10 mg of morphine IV, the patient [4], yet does not involve the use of a nerve stimulator reported severe pain and would not allow attempts at closed needle given the lack of availability and familiarity with this reduction. He had eaten solid food on arrival to the ED and device in the ED. was thus not a candidate for procedural sedation. After

Fig. 2 A, Insertion of needle parallel to long axis of transducer during real-time ultrasound. B, The needle is seen (small arrows) approaching the BP. 474 M.B. Stone et al. written informed consent was obtained, a supraclavicular thesiologists (ASA) class. After written informed consent brachial plexus nerve block was performed using ultrasound was obtained, a supraclavicular brachial plexus nerve block guidance with a 27-gauge needle and 30 mL of lidocaine was performed using ultrasound guidance with a 27-gauge 1.0% with epinephrine. The patient demonstrated full needle and 30 mL of lidocaine 1.0% with epinephrine. The sensory blockade and partial motor blockade 25 minutes patient reported complete pain relief 30 minutes after after injection, and a closed reduction and cast application injection, and a coaptation splint and sling was applied after was performed without the need for any additional closed reduction was performed. Total time from injection to analgesia. Total time from injection to completion of completion of procedure was 50 minutes. procedure was 40 minutes. 3.3. Case 3 4. Discussion A 30-year-old woman status post motor vehicle collision presented with a closed midshaft humerus Procedural sedation is commonly used for patients with fracture. Despite 10 mg of morphine IV, the patient upper extremity fractures, dislocations, or abscesses that are reported unrelenting 10/10 pain. The patient had a con- not amenable to local infiltration of anesthetic agents. comitant closed head injury and was thus a poor candidate Procedural sedation involves prolonged fasting, multiple for procedural sedation. After written informed consent providers, a monitored bed in the ED, time for preparation, was obtained, a supraclavicular brachial plexus nerve block sedation and recovery, and risks of deep sedation. In was performed using ultrasound guidance with a 27-gauge addition, certain conditions such as head injury, hypoten- needle and 30 mL of lidocaine 1.0% with epinephrine. The sion, or underlying cardiopulmonary disease may make the patient reported complete pain relief 35 minutes after use of procedural sedation unacceptable for some patients, injection, and physical examination revealed full motor given the risk of hypotension and the inability to closely and sensory blockade. A coaptation splint and sling were monitor neurologic status during sedation. Brachial plexus applied after closed reduction was performed, and the nerve block facilitates motor and sensory blockade of the patient was admitted for observation given a closed head upper extremity; requires no additional staff, monitoring, or injury. The patient remained in the ED for 6 hours before other ED resources; and can be performed without the risks admission to the wards, and during this time required no of procedural sedation. additional analgesia. Total time from injection to comple- Complications of regional nerve blocks of the brachial tion of procedure was 45 minutes. plexus include arterial puncture, pneumothorax, paresthe- sias, recurrent laryngeal or phrenic nerve paralysis, and 3.4. Case 4 rarely permanent neurologic dysfunction. Using real-time ultrasound guidance, the risks of pneumothorax and arterial A 39-year-old woman with morbid obesity and well- puncture are dramatically reduced because the operator is controlled schizoaffective disorder presented to the ED able to visualize and avoid the pleura and subclavian artery, after a fall onto her outstretched hand. Radiographs respectively. Prior anesthesiology research has demonstrated revealed a posterior elbow dislocation. There were no reduced rates of pneumothorax when using ultrasound monitored beds available in the ED, and procedural guidance for the supraclavicular approach [4,7]. In addition, sedation was not pursued because of concerns regarding the rate of block success is increased from approximately the patient’s morbid obesity and ability to maintain an 85% using the landmark technique to 95% using ultrasound airway. After written informed consent was obtained, a guidance [7]. supraclavicular brachial plexus nerve block was performed Blaivas and Lyon [8] have described the use of the using ultrasound guidance with a 22-gauge noncutting interscalene approach to ultrasound-guided brachial plexus spinal needle and 30 mL of lidocaine 1.0% with blocks in the ED management of glenohumeral disloca- epinephrine. The patient reported complete pain relief tions. Although this application appears useful, we feel that 30 minutes after injection, and physical examination the supraclavicular approach presented here is preferable to revealed full sensory and partial motor blockade. Closed the interscalene approach. The interscalene approach is reduction was performed without the need for any associated with high rates of symptomatic recurrent additional analgesia. Total time from injection to comple- laryngeal nerve paralysis and a nearly 100% incidence of tion of procedure was 45 minutes. phrenic nerve paralysis [9], which may have a significant 3.5. Case 5 effect on pulmonary function [10]. These effects are undesirable in the relatively uncontrolled setting of the A 90-year-old woman presented with a closed midshaft ED. In contrast, the supraclavicular brachial plexus block is humerus fracture after a mechanical fall. She had congestive associated with a much lower incidence of recurrent heart failure, diabetes, and severe chronic obstructive laryngeal nerve paralysis and an approximately 50% pulmonary disease, and procedural sedation was relatively incidence of phrenic nerve paralysis [11], which is mild contraindicated because of her American Society of Anes- and asymptomatic in healthy subjects. Ultrasound-guided supraclavicular block in the ED 475

Over the 3-month period over which these cases were References collected, we performed a total of 12 ultrasound-guided supraclavicular brachial plexus blocks. None of our [1] Gray AT, Schafhalter-Zoppoth I. Ultrasound guidance for ulnar nerve patients experienced pneumothoraces, symptomatic phrenic block in the forearm. Reg Anesth Pain Med 2003;28(4):335-9. or recurrent laryngeal nerve palsies, or permanent neuro- [2] Brull R, McCartney CJ, Chan VW. A novel approach to infracla- vicular brachial plexus block: the ultrasound experience. Anesth logic dysfunction. Eleven patients had adequate anesthesia Analg 2004;99(3):950 [author reply 950-1]. after a single injection of 30 mL of 1% lidocaine with [3] Chan VW. Applying ultrasound imaging to interscalene brachial epinephrine. One patient required an additional injection of plexus block. Reg Anesth Pain Med 2003;28(4):340-3. 15 mL of 0.5% bupivicaine to achieve a satisfactory level [4] Chan VW, et al. Ultrasound-guided supraclavicular brachial plexus of anesthesia. block. Anesth Analg 2003;97(5):1514-7. [5] De Andres J, Sala-Blanch X. Ultrasound in the practice of brachial The 5 cases described in this report suggest the ability of plexus anesthesia. Reg Anesth Pain Med 2002;27(1):77-89. this technique to save time in a busy ED with limited [6] Kapral S, et al. Ultrasound-guided supraclavicular approach for resources while achieving excellent analgesia for a wide regional anesthesia of the brachial plexus. Anesth Analg 1994; range of painful upper extremity procedures. 78(3):507-13. [7] Williams SR, et al. Ultrasound guidance speeds execution and improves the quality of supraclavicular block. Anesth Analg 2003; 97(5):1518-23. 5. Conclusion [8] Blaivas M, Lyon M. Ultrasound-guided interscalene block for shoulder dislocation reduction in the ED. Am J Emerg Med 2006; Real-time ultrasound-guided supraclavicular brachial 24(3):293-6. [9] Urmey WF, Talts KH, Sharrock NE. One hundred percent incidence of plexus nerve block in the ED provided excellent analgesia hemidiaphragmatic paresis associated with interscalene brachial for 5 patients requiring treatment of upper extremity plexus anesthesia as diagnosed by ultrasonography. Anesth Analg fractures, dislocations, and abscesses. This may represent 1991;72(4):498-503. a useful alternative to procedural sedation. A randomized [10] Urmey WF, McDonald M. Hemidiaphragmatic paresis during controlled trial comparing ED length of stay in patients interscalene brachial plexus block: effects on pulmonary function and chest wall mechanics. Anesth Analg 1992;74(3):352-7. receiving procedural sedation versus ultrasound-guided [11] Neal JM, et al. Quantitative analysis of respiratory, motor, and supraclavicular brachial plexus nerve block is currently sensory function after supraclavicular block. Anesth Analg 1998; underway at our institution. 86(6):1239-44. American Journal of Emergency Medicine (2007) 25, 476–478

www.elsevier.com/locate/ajem

Controversies The use of long-term controller medications in asthmatic patients being discharged from the ED— why the controversy?

A. Dosanjh MD

Department of Pediatrics, UCSD School of Medicine, La Jolla, CA, USA

Received 17 August 2006; revised 6 September 2006; accepted 6 September 2006

1. Introduction cheap and safe. The role of inhaled corticosteroids to some authors then becomes superfluous in the ED management The use of long-term controller medication (LTCM) in of asthma. the management of persistent asthma is part of standard This perception may be related to 2 key factors: (i) the practice guidelines in the National Asthma Education and role of the ED physician in managing the asthmatic patient; Prevention Program [1]. The LTCMs include numerous (ii) the lack of distinction between long- and short-term therapeutic options, all aimed at controlling the inflamma- controller therapy. tion that underlies persistent asthma. Controller medications Inhaled steroids offer some distinct advantages to the can include inhaled corticosteroids, leukotriene modifiers, patient and are likely to be used to prevent further asthma mast cell stabilizers, long-acting inhaled b2 adrenergic attacks by the primary physician. Local inhaled agents are medications, or combinations of these medications [1]. more rapid in onset and reduce local edema and vasocon- Although the use of the controller medications is widely striction [3,4]. accepted practice, they are rarely prescribed when discharg- ing a patient from the emergency department (ED). The discordance between the traditional ED practices of dispensing short-term medications to patients with chronic 3. The use of inhaled versus oral controller asthma is leading to reanalysis of the issue [2]. medications

Several trials have compared the uses of oral steroids 2. The use of LTCM in the ED patient versus inhaled corticosteroids in the ED setting. A meta- analysis of 7 trials was conducted, involving over 1200 The use of short-term controller medications, such as patients, 772 of which were adults, and the other patients oral steroids, upon discharge from the ED is the traditional ranged in age from 6 months to 16 years. The authors approach to the asthmatic patient. Some authors advocate summarized some of the advantages of high-dose inhaled continuing this practice, pointing out that oral steroids are corticosteroids in acute asthma therapy [5]. These included greater efficacy in reducing local edema and local vasocon- E-mail address: [email protected]. striction. The side effects of oral corticosteroids (CS) can

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2006.09.013 The use of LTCMs in asthmatic patients being discharged from the ED 477 also be a compliance barrier in many patients. The study 3.1. The use of both inhaled and oral compared 7 studies using high-dose inhaled corticosteroids corticosteroids (ICS) (N2 mg/d) to standard doses of oral CS. There was no significant difference at either 7 to 10 days or 16 to 21 days When studies analyze the addition of inhaled CS to oral in asthma relapse rates between the 2 routes of delivery. The CS, there appears to be added benefit and more agreement trials analyzed only included mild asthmatics, and the among authors who prefer the traditional use of oral CS. In possibility of type II error could not be ruled out. Type II one study of 44 children, 6 months to 18 years of age, with error is defined as failure to reject a given hypothesis when an acute asthma exacerbation, inhaled BUD was compared an alternative hypothesis may be true. with placebo as an adjunct to oral prednisone in the acute In one study of pediatric patients, the impact of inhaled setting. The authors found that patients receiving BUD after anti-inflammatory therapy on hospitalization and ED visits a loading dose of oral prednisone had greater improvement for children with asthma was evaluated. The authors included in symptoms and fewer hospitalizations and found that 11,195 children up to 15 years of age in major metropolitan patients discharged from the ED after treatment for acute areas and analyzed the use of any anti-inflammatory inhaled asthma benefit from added treatment with high-dose inhaled BUD for 21 days as compared to using only a short course agent on the risk of hospitalization and ED visits in children of oral prednisone [12]. with asthma. They concluded that both inhaled cromolyn or Marik and Varon [13] commented that although they see inhaled CS were significantly protective in reducing risk of no particular reason to abandon the use of oral CS in the ED visits and hospitalization [6]. discharged ED patient, they do see some benefit in the Some studies have examined the rate of action of inhaled addition of ICS in these patients. CS compared with oral CS in the setting of acute asthma [7] b and concluded that with reference to the early administration 4. The changing role of the physician of CSs in the ED treatment of adult patients with acute asthma, this evidence-based evaluation suggests the follow- Although oral CS are often prescribed for patients with ing: parenteral administration probably requires N6to an acute asthma attack and shown to be effective in studies, 24 hours to improve pulmonary function in the patient (on there is as problem with adherence in actual practice [14].In the contrary, high doses of inhaled CSs increase pulmonary a study of 161 patients, patient caregivers reported function 1 to 3 hours after therapy) [1]; comprehensible adherence only 64% of the time. conclusions about admission rates in the ED setting are To determine the frequency with which ED physicians difficult to make (overall, there is evidence about a reduction prescribe LTCMs for children with asthma and to assess in hospital admissions, but a separate analysis, including only their awareness of national guidelines for LTCM use, a high-quality studies, showed no benefit) [2]; IV and oral CSs survey was conducted. The findings among the 391 appear to have equivalent effects on pulmonary function [3]; surveyed indicated that although 99% believe that children and there is a tendency toward improvement in pulmonary with persistent asthma should be treated with LTCMs, less function with medium and high doses, although the evidence than 20% provided LTCMs for children at discharge from does not support the use of very high doses [4]Q. the ED [15]. In a Cochrane review of 352 patients enrolled in a total of The most often cited reason for this discordance was that 6 trials (4 adults and 2 pediatrics), comparing inhaled to oral 74% of those surveyed do not believe it to be their role to CS, the patients receiving inhaled steroids had reduced provide LTCMs, and that this duty was to be performed by admission rates. The results were inconclusive with respect the primary caregiver (PCP). to the addition of oral CS to inhaled CS [8]. In other studies, only 23% of children reported to a PCP The rapid onset of action of inhaled CS is an advantage after an ED visit for asthma, and 22% of adults received all in the ED when decisions regarding admission to the their asthma care in the ED. However, other studies hospital need to be made in a short period. conducted on the effectiveness of follow-up in the ED Other studies have concluded that inhaled CS are just as itself, rather than at the PCP, enhanced accessibility and effective as using oral prednisone (2 mg/kg) in reducing asthma management [16]. Results of an ED-based random- symptoms of acute asthma in children treated in the ED [9]. ized clinical trial showed improved asthma outcomes in a One such study of 22 children ranging in age from 6 to high-morbidity pediatric population. Other authors suggest 16 years received terbutaline and either budesonide (BUD) additional innovative approaches such as providing a1-800 or oral prednisone (2 mg/kg) in a double-blinded fashion. # for ongoing counseling and outreach advice [17]. Efforts A short course of BUD over 1 week was shown to be as at asthma teaching have yielded mixed results, and one of effective as oral prednisone and further was not associated the obstacles to providing maintenance therapy has been the with a small transient decrease in serum cortisol levels. factor of time management in the ED and reluctance to The group on oral prednisone had a small transient provided time-consuming services. decrease in serum cortisol levels. Some studies have also Emergency department physicians commonly encounter shown better improvement in FEV1 when using a local children with persistent asthma who are not on LTCM. The inhaled steroid solution [10,11]. lack of controller medication among these children may be 478 A. Dosanjh one reason for frequent use of the ED for acute asthma. The [8] Edmounds ML, Camargo CA, Pollack CV, Rowe BH. Early use of use of LTCM in patients upon discharge has been shown to inhaled corticosteroids in the emergency department treatment of be effective in preventing hospitalization and promoting acute asthma. Cochrane Database Syst Rev 2001. [9] Scarfone RJ, Loiselle JM, Wiley II JF, et al. Nebulized dexamethasone more rapid resolution of symptoms [18]. The opportunity to versus oral prednisone in the emergency treatment of asthmatic intervene and prescribe LTCM in addition to short-term children. Ann Emerg Med 1995;26:480-6. traditional therapy such as prednisone could undoubtedly [10] Chipps B, Chipps D. A review of the role of inhaled corticosteroids in impact the quality of clinical management and long-term the treatment of acute asthma. Clin Pediatr 2001;40:185-9. care of the patient. Emergency department physicians are [11] Volovitz B, Bentur L, Finkelstein Y, et al. Effectiveness and safety of inhaled corticosteroids in controlling acute asthma attacks in children thus in a unique position to enhance the care of the who were treated in the emergency department: a controlled persistent asthmatic. comparative study with oral prednisolone. J Allergy Clin Immunol 1998;102:605-9. [12] Sung L, Osmond MH, Klassen TP. Randomized controlled trial of inhaled budesonide as an adjunct to oral prednisone in acute asthma. References Acad Emerg Med 1998;5:209-13. [13] Marik P, Varon J. Oral vs inhaled corticosteroids following emergency department discharge of patients with acute asthma. Chest 2002;121: [1] Strek M. Difficult asthma. Proc Am Thorac Soc 2006;3:116-23. 1735-6. [2] Singer A. A call for expanding the role of the emergency physician in [14] Butler K, Cooper WO. Adherence of pediatric asthma patients with the care of patients with asthma. Ann Emerg Med 2005;45:295-8. oral corticosteroid prescriptions following pediatric emergency [3] Rodrigo G, Rodrigo C. Inhaled Flunisolide for acute severe asthma. department visit or hospitalization. Pediatr Emerg Care 2004;20(11): Am J Respir Crit Care Med 1998;157:698-703. 730-5. [4] Gibbs MA, Camargo Jr CA, Rowe BH, et al. Therapeutic controversies [15] Scarfone R, Zorc J, Angsuco C. Emergency physicians prescribing of in severe acute asthma. Acad Emerg Med 2000;7:800-15. asthma controller medications. Pediatrics 2006;117(3):821-7. [5] Edmounds M, Carmargo C, Brenner B, Rowe B. Replacement of oral [16] Teach SJ, Guagliardo MF, Crain E, McCarter R, et al. Spatial corticosteroids with inhaled corticosteroids in the treatment of acute accessibility of primary care pediatric services in an urban: asthma following emergency department discharge. Chest 2002;121: association with asthma management and outcome. Pediatrics 2006; 1798-805. 117(4):S78-S85. [6] Adams RJ, Fuhlbrigge A, Finkelstein JA, Lozano P, et al. Impact of [17] Rowe B, Majumdar S. Improving quality of asthma care after inhaled antiinflammtory therapy on hospitalization and emergency emergency department discharge: evidence before action. Ann Emerg department visits for children with asthma. Pediatrics 2001;107(4): Med 2004;45:299-301. 706-11. [18] Fernandez J, Onis-Gonzalez E, Pitti J, Zache S, et al. Factors [7] Rodrigo G, Rodrigo C. Corticosteroids in the emergency department associated with the short-term clinical outcomes after acute treatment therapy of acute adult asthma: an evidence-based evaluation. Chest of asthma in a pediatric emergency department. Pediatr Pulmonol 1999;116:285-95. 2004;38:123-8. American Journal of Emergency Medicine (2007) 25, 479–480

www.elsevier.com/locate/ajem

Editorial Therapeutic hypothermia and the need for defibrillation: wet or dry?

Joseph Varon MD*

The University of Texas Health Science Center Houston, St Luke’s Episcopal Hospital, Houston, TX 77030, USA The University of Texas Medical Branch at Galveston, St Luke’s Episcopal Hospital, Galveston, TX 77030, USA

Received 19 March 2007; accepted 20 March 2007

Therapeutic hypothermia (TH) has been used for demonstrated to improve the outcome in patients with millennia for a variety of reasons. However, the use of cold coma after resuscitation from out-of-hospital cardiac arrest, temperatures in modern clinical medicine is probably only this intervention is now endorsed by the International 200 years old. The Russian method of resuscitation, Liaison Committee on Resuscitation, the European Resus- described in 1803, consisted in covering a patient with citation Council, and the American Heart Association [6- snow, hoping for return of spontaneous circulation [1]. 8]. Indeed, these organizations now recommend the use of Therapeutic hypothermia was used during Napoleon’s induced mild hypothermia in comatose survivors of out-of- Russian campaign in 1812, by the surgeon Baron de Larrey, hospital cardiac arrest caused by ventricular fibrillation. in an attempt to preserve injured limbs as well as for its Evidence-based medicine compels that a moderate degree anesthetic effects during amputation [2]. However, the first of selective brain cooling, if instituted promptly, may actual recorded clinical use of TH was in 1937, when Dr protect against ischemic neuronal injury of the type that Temple Fay bcooledQ a lady to 328C for 24 hours, in a occurs commonly after a cardiac arrest (class IIa recom- desperate attempt to prevent cancer cells from further mendation) [9]. multiplying and progressing at lower temperatures [3,4]. During TH, some patients may have recurrent ven- The use of TH after cardiac arrest in humans in modern tricular fibrillation requiring cardiac defibrillation. If medicine was first reported in 1957 by Benson and the patient is being cooled using surface cooling, con- colleagues [5]. They cooled 4 patients (to 308C-338C) after cerns arise regarding safety for the operator and the patient in-hospital cardiac arrest for periods of 24 to 72 hours. All 4 as well as efficacy. Theoretically, TH could impair patients recovered. Interest in the use of mild to moderate defibrillation by reducing sodium channel conductance induced TH (328C-348C) after cardiac arrest was spurred by and increasing electrical heterogeneity via dispersion in the pioneering work of the late Dr Peter Safar. the action potential duration [10]. In this issue of The After 2 landmark randomized controlled studies pub- American Journal of Emergency Medicine, Schratter and lished in 2002, in which moderate hypothermia was associates [11] report the results of an animal model of cardiac arrest in which they assessed the safety and effectiveness of transthoracic defibrillation during surface * The University of Texas Health Science Center Houston, St Luke’s Episcopal Hospital, Houston, TX, USA. Tel.: +1 713 669 1670; fax: +1 713 cooling with cold water as compared with dry conditions. 669 1671. These investigators demonstrate that transthoracic defibril- E-mail address: [email protected]. lation is feasible and safe during TH in wet conditions and

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.ajem.2007.03.022 480 Editorial that the energy delivered to the victims is the same as in Harbans L, Harbans L, editors. Emerging strategies in neuroprotec- dry conditions. tion. Boston7 Birkhauser; 1992. p. 289-306. Therapeutic hypothermia is now seen as a new neuro- [3] Fay T. Observations on prolonged human refrigeration. N Y State J Med 1940;40:1351-4. protective strategy and is gaining ground as an emergency [4] Alzaga AG, Salazar G, Varon J. Breaking the thermal barrier: Dr therapy. The best cooling technique is one that is easy to use Temple Fay. Resuscitation 2006;69:359-64. and at the same time effective. In the future, TH trials to [5] Benson DW, Williams GR, Spencer FC. The use of hypothermia after lower temperatures may produce an unstable electrophysi- cardiac arrest. Anesth Analg 1958;38:423-8. ologic state. However, data such as those presented by [6] Hypothermia after Cardiac Arrest Study Group. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Schratter and colleagues [11] suggest that TH may be used Engl J Med 2002;346:549-56. to slow metabolic processes without concerns over the [7] Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose ability to successfully and safely defibrillate and treat survivors of out-of-hospital cardiac arrest with induced hypothermia. hypothermia-induced dysrhythmias. It is clear that, based N Engl J Med 2002;346:557-63. on this study, in surface cooling with ice conditions, [8] Nolan JP, Morley PT, Vanden Hoek TL, et al. Therapeutic hypothermia after cardiac arrest: an advisory statement by the electrical defibrillation is safe and efficacious. advanced life support task force of the International Liaison Committee on Resuscitation. Circulation 2003;108:118-21. [9] Alzaga A, Cerdan M, Varon J. Therapeutic hypothermia. Resuscita- References tion 2006;70:369-80. [10] Sprung J, Laszlo A, Turner LA, et al. Effects of hypothermia, [1] Varon J, Sternbach GL. Cardiopulmonary resuscitation: lessons from potassium and verapamil on the action potential characteristics of the past. J Emerg Med 1991;9:503-7. canine cardiac Purkinje fibers. Anesthesiology 1995;82:713-22. [2] Mordecai Y, Globus RB, Dietrich WD, Sternau L, Morikawa E, [11] Schratter A, Weihs W, Holzer M, et al. External cardiac defibrillation Ginsberg MD. Temperature modulation of neuronal injury. In: during wet surface cooling in pigs. Am J Emerg Med 2007;25:420-4. American Journal of Emergency Medicine (2007) 25, 481–487

www.elsevier.com/locate/ajem

Correspondence Obtundation in a toddler: naloxone is fundamental high index of suspicion of opioid poisoning when treating patients with coma, even children. To the Editor, Kevin C. Osterhoudt MD, MSCE A 2-year-old girl was heard by her parents to be making Diane P. Calello MD, MSCE bgurglingQ breathing noises and was subsequently found to Section of Clinical Toxicology be unresponsive. Emergency medical service personnel Poison Control Center noted her to be cyanotic and provided bag-mask ventilatory Division of Emergency Medicine support. Her pupils were small but equal. She had no history The Children’s Hospital of Philadelphia of preceding illness or trauma. Philadelphia, PA 19104, USA The girl was afebrile upon arrival at the ED, and no E-mail address: [email protected] etiology for her neurological and respiratory depression was readily apparent to treating physicians. An arterial blood gas doi:10.1016/j.ajem.2005.09.011 demonstrated significant respiratory acidosis. Endotracheal intubation was performed with lidocaine, midazolam, and Reference vecuronium, but successful tube placement was described as bdifficultQ and took considerable time and several attempts. [1] Osterhoudt KC. No sympathy for a boy with obtundation. Pediatr A commercial urine drug immunoassay (including amphet- Emerg Care 2004;20:403-6. amines, benzodiazepines, barbiturates, cocaine, cannabis, and opiates) was negative. Infantile case of seizure induced by intoxication after An exhaustive laboratory and radiographic evaluation accidental consumption of eperisone hydrochloride, was performed to determine a cause of illness, and the girl an antispastic agent was transferred to a tertiary-care pediatric facility. Comput- ed tomography of the brain demonstrated bwatershedQ hypodensities consistent with ischemic injury. Over the To the Editor, course of the next 36 hours, she regained alertness. More elaborate immunoassay of the urine detected methadone, the An 18-month-old (9 kg in body weight) female, whose presence of which was confirmed via gas chromatography/ family history was unremarkable, accidentally took 100 mg mass spectroscopy. It is speculated that the girl took an (2 pills) of eperisone hydrochloride. Approximately 30 min- exploratory ingestion of methadone while visiting her utes after swallowing the tablets, her voice became pe- neighbor, a patient with cancer. culiar, and she then lost consciousness. When emergency In the emergency resuscitation of children, an bABCDQ medical technicians arrived, she was already in a coma, approach has been advocated [1]. Those children with accompanied by recurrent grand mal. Her cyanosis (Spo2, neurological disability merit careful and rapid consideration 93%) was immediately resolved by administration of of oxygenation and blood glucose concentration. Adminis- oxygen (Spo2, 100%). tration of an empiric trial of naloxone may also be At about 50 minutes after the accidental dosing, the considered, even among children, and especially among patient arrived at our emergency department. The coma pediatric patients whose neurological and respiratory (Glasgow Coma Scale: 9 [E4M4V1]) and recurrent grand depression are accompanied by miotic pupils. In this case mal accompanying spastic eye opening were still continu- of a child with obtundation and miotic pupils, it is possible ing. Both pupil sizes were 6 mm without light reflex. Elbow that a naloxone trial, fundamental to emergency resuscita- flexion, weak rigidity, and tremor were observed in limbs tion, may have led to early diagnosis and have been more during the interval of seizures. Electrocardiographs showed useful than the commercial urine drug screen, prevented ventricular tachycardia (120-180 beat/min), but blood potential morbidity associated with the bdifficultQ airway, pressure was maintained (130/78 mm Hg). Respiration and saved further emotional, physical, and monetary costs. was unstable; shallow breathing (36/min) and apnea Emergency medical providers are encouraged to maintain a repeated alternately. Body temperature was 37.58C, and

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. 482 Correspondence stiffness of the neck was not obvious. Intravenous injection Katsutoshi Tanno MD of 2 mg midazolam terminated the seizure, and a tracheal Eichi Narimatsu PhD, MD tube was immediately inserted to assist ventilation. Subse- Yoshihiro Takeyama MD quently, brain computed tomography showed no neurolog- Yasufumi Asai PhD, MD ical abnormalities. Department of Trauma and Critical Care Medicine About 3 hours after the dosing, she suddenly regained Sapporo Medical University consciousness. Thereafter, seizure and other neurological Sapporo, 060-8543 Japan symptoms did not recur, and ventricular tachycardia resolved itself spontaneously. About 4 hours after admis- doi:10.1016/j.ajem.2006.09.002 sion, the tracheal tube was removed following confirma- tion of respiratory stabilization. On the second hospital References day, electroencephalography showed no abnormalities, and on the third day, she was discharged with no sequela. The [1] Noma S, Sasa M, Ohno Y, et al. Effects of eperisone applied by serum concentration of eperisone hydrochloride, belatedly microiontophoresis on neurons in the medial and lateral vestibular reported 14 days after admission, was 162.93 ng/mL. nuclei. Jpn J Pharmacol 1986;40:283-90. Eperisone hydrochloride (4-ethyl-2-methyl-3-piperidino- [2] Fujioka M, Kuriyama H. Eperisone, an antispastic agent, processes vasodilating actions on the guinea-pig basilar artery. J Pharmacol Exp prophenone hydrochloride), an antispasmodic, acts to Ther 1985;235:757-63. reduce tonus of the skeletal muscles by depressing spinal [3] Asano M, Ohkubo C. Effects of eperisone on microcirculation as cord reflexes through both mono- and polysynaptic path- observed by intravital-microscopy in the conscious rabbit. Bull Inst ways. The detailed mechanisms of eperisone attenuating the Public Health 1989;38:17-22 [Japanese]. spinal cord reflexes have not been revealed satisfactorily [1]. [4] Shiraishi H, Ishibashi K, Urao N, et al. Two cases of polymorphic ventricular tachycardia induced by the administration of verapamil Eperisone also acts to dilate vessels by blocking voltage- ++ against paroxysmal supraventricular tachycardia. Intern Med dependent Ca channels [2] and increases peripheral blood 2002;41:445-8. flow [3]. The normal oral dose of eperisone is 50 mg 3 times (150 mg) a day for adults. Considering the small body size Hydronephrosis during pregnancy (9 kg) and infantile immature drug metabolism of the patient, the effect of 100 mg of eperisone on the patient was To the Editor, suspected to correspond to that of at least 1000 to 2000 mg eperisone on adults. Actually, the elevated plasma concen- Recently, a 30-year-old woman at 22 weeks’ gestation tration of eperisone (162.93 ng/mL, toxic level) was more presented to our emergency department (ED) with unre- than 20 times higher compared to the peak therapeutic mitting right flank pain. Her history was suspicious for plasma concentration (around 7 ng/mL). urolithiasis, and a urinalysis showing 67 red blood cells Severe symptoms, such as conscious loss, seizure, apnea, was obtained. Because of the patient’s extreme pain after and ventricular tachycardia, were observed in the present medication, a renal ultrasound was obtained to evaluate case, all of which have, to the best of our knowledge, not been both kidneys for hydronephrosis secondary to presumed reported as the side effects of eperisone. Because intravenous ureteral obstruction. ++ verapamil, a voltage-dependent Ca channel blocker, has the The examination showed a unilateral right hydroneph- potential to elicit ventricular tachycardia [4], it is suspected rosis. Although many emergency medicine and urology that ventricular tachycardia observed in the present case was texts discuss hydronephrosis as a common finding in induced by an excessive effect of eperisone to block voltage- pregnancy, the exact incidence is not clearly known. There ++ dependent Ca channels [2]. Loss of consciousness and is also a lack of knowledge regarding the significance of apnea can be explained by the after-effect of seizure or the unilateral hydronephrosis in pregnant patients with flank probable action of eperisone depressing the excitability of the pain. In cases such as ours, it may be unclear if the central nervous system, which is suspected from its attenu- hydronephrosis is secondary to an impacted stone or due to ating effect on spinal cord reflexes and its side effect of physiological changes during pregnancy. To better delineate inducing a somnolent state. However, it is difficult to this issue, we reviewed the literature regarding the theoretically explain the cause of seizure from the given characteristics of hydronephrosis in pregnancy, current actions of eperisone. It is suspected that some unrevealed evaluation techniques, and recommendations for disposition neurological mechanisms of eperisone may exist, which are when patients present to the ED in a similar fashion. manifested only when its plasma concentration is extraordi- Physiological hydronephrosis is a common occurrence narily elevated above its therapeutic one. in pregnancy. Peake et al [1] used ultrasound to evaluate This case report reveals that an overdose of transoral 154 pregnant patients and found that the incidence of any eperisone has the potential to elicit life-threatening states of seizure and arrhythmia, although eperisone is considered to These views are those of the authors and do not necessarily reflect the be a neurodepressant. views of the U.S. Army or the Department of Defense. Correspondence 483 hydronephrosis was 90%. Several other studies report a References 60% to 100% incidence rate. Most studies describe small amounts of hydronephrosis generally occurring after the [1] Peake SL, Roxburgh HB, Langlois SL. Ultrasonic assessment of 10th week of pregnancy. It was also found that right-sided hydronephrosis of pregnancy. Radiology 1983;146:167. hydronephrosis was greater than left-sided hydronephrosis [2] Fowler KA. US for detecting renal calculi with nonenhanced CT as a reference standard. Radiology 2002;222:109. in 62% of patients, 32% of the patients had a right equal to [3] Wachsberg R. Unilateral absence of ureteral jets in the third trimester of left hydronephrosis, and only 6% of the patients had a left pregnancy: pitfall in color Doppler US diagnosis of urinary obstruction. greater than right presentation. A true unilateral hydro- Radiology 1988;209:279. nephrosis on either side was rare, but on average, the right kidney was 4 to 7 mm larger than the left [1]. Further evaluation in these cases becomes complicated Recurrent myelin basic protein elevation in because some standard techniques are not used secondary to cerebrospinal fluid as a predictive marker of delayed radiation exposure. Renal ultrasound is standard in evalu- encephalopathy after carbon monoxide poisoning ating signs of hydronephrosis, but studies have attempted to expand its use by locating and measuring stones in the ureter. Although promising for ureteral stone detection, To the Editor, sensitivity ranges from 63% to 85%, and specificity ranges from 79% to 100%. It can be used as a diagnostic tool in Although neuron-specific enolase (NSE) has been shown pregnant patients with suspected urolithiasis, but variation to reflect neuronal damage, myelin basic protein (MBP), a in operator experience and the extent of the disease major myelin component in the central nervous system, has contribute to the imbalance in these numbers. An inability been shown to reflect the degree of demyelination in to visualize stones smaller than 5 mm and to differentiate cerebral white matter [1]. However, we know of no report between obstructed and nonobstructed stones does not make dealing with NSE or MBP in cerebrospinal fluid (CSF) after this technique useful in the disposition of ED patients [2]. carbon monoxide (CO) poisoning. Attempts using color Doppler to detect obstruction by A 36-year-old woman found semicomatose in a car measuring ureteral jets into the bladder have also been with traces of burnt charcoal was recently transferred to our unreliable secondary to operator dependence, positional emergency center. On admission the patient was lethargic. variables, and lack of ureteral jets in healthy subjects [3]. The carboxyhemoglobin (Hbco) content of the blood was No well done prospective study has been performed to 0.3%. She was treated with hyperbaric oxygen 3 times differentiate obstructive from physiologic hydronephrosis in during the first 3 days of the hospital stay. When she became pregnant patients. Although most cases can be handled with alert, she showed only anterograde amnesia in neurologic outpatient pain management and routine referral to an assessments and admitted to have attempted suicide using obstetrician, diagnostic surety on occasions is a must. burning charcoal 2 days before admission. T2-weighted, Instances of hydronephrosis complicated by the presence diffusion-weighted, and fluid-attenuated inversion-recovery of an infection or uncontrollable pain, as was the case for magnetic resonance imaging (MRI) performed on hospital our patient, should be admitted. In our patient, the stone day 5 demonstrated bilateral symmetric signal intensity passed spontaneously after inpatient pain management, and abnormalities in the globus pallidus (Fig. 1). Cerebrospinal she continued with an uncomplicated pregnancy. Future fluid examination on the same day showed no abnormalities studies should endeavor to further clarify the incidence, in routine items, but NSE and MBP were 22 ng/mL (normal degree, and significance of pregnancy-related hydroneph- value, b10 ng/mL) and 1010 pg/mL (normal value, rosis in the presence and absence of ureteral obstruction. b102 pg/mL), respectively. After obtaining consent from the patient as well as her families, we assayed these proteins serially in the following weeks. Although NSE in CSF Todd McArthur MD gradually decreased, MBP in CSF rapidly decreased to Chad S. Crystal MD below the detectability limit (b40 pg/mL) on hospital day 13 Department of Emergency Medicine but rose again to 324 and 641 pg/mL on hospital days 20 C.R. Darnall Army Medical Center and 27, respectively (Fig. 1). No notable change in neuro- Ft Hood, TX 76544 logic symptoms occurred until hospital day 30 when hypo- Michael A. Miller MD tonia of the extremities and lack of spontaneity developed, Department of Emergency Medicine rapidly progressing to an apallic state for a few days. Darnall Army Community Hospital T2-weighted, diffusion-weighted, and fluid-attenuated in- Ft Hood, Texas version-recovery MRI on hospital day 44 revealed bilateral, Central Texas Poison Control Center confluent increased signal intensity in cerebral white matter Temple, TX 76508 (Fig. 1). On hospital day 60, neurologic abnormalities began to gradually abate. On hospital day 105, the patient was doi:10.1016/j.ajem.2006.08.018 discharged from the hospital, showing mild cognitive 484 Correspondence

Fig. 1 Changes of MBP (filled circles) and NSE (open triangles) in CSF measured after 3 days of hyperbaric oxygen (HBO) sessions. Diffusion-weighted magnetic resonance images on hospital days 5 (inset A) and 44 (inset B) also are shown, respectively, demonstrating bilateral increased in signal intensity in the globus pallidus and in cerebral white matter. Asterisk indicates MBP in CSF below the detectability limit for the assay (b40 pg/mL). deficit. Three months later, she had fully recovered, and subjecting myelin to autoimmune attack causing delayed abnormal MRI findings disappeared. demyelination and dysfunction [3]. In the present report, the Her neuroimaging finding of bilateral symmetric lesions second elevation of MBP in CSF may have involved a in the globus pallidus—the most common site of abnor- demyelinating autoimmune response. If the autoimmune mality in the acute phase of CO poisoning—suggested that hypothesis is correct, timely antiautoimmune therapy such she had a very severe CO exposure 2 days before she was as steroid administration as well as prolonged hospitaliza- found, with a decline in Hbco and neurologic severity tion might be considered upon detection of a recurrent MBP during the interval preceding admission. Later the patient elevation in CSF. manifested clinical findings of delayed encephalopathy, with MRI demonstrating typical lesions suggesting demy- Yoshito Kamijo MD elination of cerebral white matter [2]. Kazui Soma MD Myelin basic protein and NSE were found abnormally Department of Emergency and Critical Care Medicine elevated in the first CSF specimen in addition to examina- School of Medicine, Kitasato University tions to exclude inflammatory central nervous system Sagamihara Kanagawa 228-8555, Japan diseases. Neuron-specific enolase gradually decreased, E-mail address: [email protected] whereas MBP rapidly decreased only to increase again Toshimitsu Ide MD preceding clinical onset of delayed encephalopathy (Fig. 1). Department of Neurology, School of Medicine Our findings suggest that early elevation of both NSE and Kitasato University MBP in CSF may reflect cerebral lesions including those in Sagamihara Kanagawa 228-8555, Japan the globus pallidus, whereas recurrent elevation of MBP in CSF may serve as a predictive marker of delayed sequelae allowing prolonged hospitalization. doi:10.1016/j.ajem.2006.06.019 As Hbco values correlate poorly with clinical severity of delayed sequelae, mechanisms other than CO-induced References hypoxic stress are considered responsible for delayed sequelae. Animal experiments suggest that MBP reacts [1] Liuzzi GM, Mastroianni CM, Vullo V, et al. Cerebrospinal fluid myelin chemically with aldehydes produced during lipid peroxida- basic protein as a predictive marker of demyelination in AIDS dementia tion. This alteration could render MBP immunogenic, complex. J Neuroimmunol 1992;36:251-4. Correspondence 485

[2] Zagami AS, Lethlean AK, Mellick R. Delayed neurological deteriora- Reply to Proper Observation of Patient Related Factors tion following carbon monoxide poisoning: MRI findings. J Neurol is an Important Determinant in the Utility of the 1993;240:113-6. D-Dimer Test for Exclusion of Venous Thromboembolism [3] Thom SR, Bhopale VM, Fisher D, et al. Delayed neuropathology after carbon monoxide poisoning is immune-mediated. Proc Natl Acad Sci in the ED USA 2004;101:13660-5. To the Editor,

Reply to Multiple-Dose Activated Charcoal We appreciate the thoughtful letter from Dr. Houdijk [1] in Carbamazepine Poisoning that summarizes important issues regarding the accuracy of D-dimer enzyme-linked immunosorbent assay in suspected venous thromboembolic disease (VTED). We do agree that To the Editor, previous studies demonstrated that a negative BioMe´rieux VIDAS BioMe´rieux, Marcy l’Etoile, France could rule out In response to the letter from Drs Beecroft, Lu, & VTED in patients with low or intermediate pretest probability Mycyk in the March issue [1], in our study [2] we based in a large proportion of suspected outpatients [2,3]. With this results on the serum carbamazepine (CBZ) concentration recent report, we only highlighted the risks taken by at admission and the peak value, which were similar in emergency physicians in performing D-dimer referent the 2 groups. Moreover, their supposed ingestion time was enzyme-linked immunosorbent assay without proper obser- similar (6 F 3 hours [G1] vs 5 F 2 hours [G2]); in vation of patients’ clinical conditions (prior anticoagulation, addition, the increase of the CBZ concentration in the location of the clot, duration of symptoms, etc) [4]. non-multiple dose activated charwal (MDAC) group was Indeed, exact knowledge of the location of the clot and not statistically significant. duration of symptoms is very difficult in emergency What we deduced from these results is that the MDAC medicine. However, it should be pointed out that other group was more seriously poisoned than the non-MDAC studies have also reported lower sensitivity of referent D- group with a higher SAPSII and APACHE II. dimer, such as our study [5,6]. Moreover, in the prospective On the other hand (and as everyone knows), neither the studies evaluating the capability of D-dimers to rule out ingested dose nor the ingested time could be used as a referent VTED, the studied population was selected (inclusion and prognostic factor in toxicological studies because of the lack exclusion criteria), and thus their results might not be simply of reliability of the anamnesis. Ingestion time could be extrapolated to the whole population seen in the emergency incriminated in the non-MDAC group—as that it was department [3]. reported by Winnicka—but we could not confirm this in One of the aims of our practical study was to determine our study. the causes of false-negative D-dimers in patients with Our study’s aim was to evaluate 2 modalities of confirmed VTED. Unfortunately, the small sample size of charcoal administration (MDAC vs single dose), and we the different subgroups limited our conclusion. Thus, we think that this aim was achieved. As for the second confirm that larger studies are required to conclude on these question, it is true that the half-life was calculated important points because too much few studies focused on between 2 CBZ measurements. However, the first point negative referent D-dimer in confirmed VTED. considered was the CBZ blood peak; so one cannot say that the half-lives calculated after MDAC were not Patrick Ray MD calculated at a much lower concentration. As a conse- Pitie-Salpetriere Hospital quence, the resulting difference would not be due to the Paris, France simple serendipity. doi:10.1016/j.ajem.2006.12.004

Nozha Brahmi MD References Centre d’Assistance Medicale Urgente Tunis, Tunisia [1] Houdijk WPM. Proper Observation of patient-related factors is an important determinant in the use of D-Dimer Test for Exclusion of doi:10.1016/j.ajem.2006.12.003 venous thromboembolism in the ED. Am J Emerg Med 2007;25(2):255. [2] Stein PD, Hull RD, Patel KC, et al. D-dimer for the exclusion of acute venous thrombosis and pulmonary embolism: a systematic review. Ann Intern Med 2004;140:589-602. References [3] Perrier A, Desmarais S, Miron MJ, et al. Non-invasive diagnosis of venous thrombolembolism in out patients. Lancet 1999;353:190-5. [1] Beecroft MJ, Lu TJ, Mycyk MB. Multiple-Dose activated charcoal in [4] Ray P, Belick B, Birolleau S, et al. Referent D-dimer ELISA testing is of carbamazepine poisoning. Am J Emerg Med 2007;25(3):373. limited value in the exclusion of thrombo-embolic disease: a practical [2] Brahmi N, Kouraichi N, Thabet H, et al. Influence of activated charcoal study in an emergency department. Am J Emerg Med 2006;24:313-8. on the pharmacokinetics and the clinical features of carbamazepine [5] De Monye W, Sanson BJ, Mac Gillavry MR, et al. Embolus location poisoning. Am J Emerg Med 2006;24(4):440-3. affect the sensitivity of a rapid quantitative D-dimer assay in the 486 Correspondence

diagnosis of pulmonary embolism. Am J Respir Crit Care Med 2002; 60 mm Hg or less (HU: 59.6% vs EDSSU: 48.3%; P b.001), 165:345-8. long-term oxygen therapy (HU: 13.4% vs EDSSU: 8.8%; P = [6] Sijens PE, Oudkerk M, Berghout A, et al. Comparison of a quantitative .004), mean length of stay (HU: 12.0 days vs EDSSU: latex and a quantitative ELISA plasma d-dimer assay in the exclusion b of segmental and subsegmental pulmonary embolism. Thromb Hae- 3.4 days; P .001), mortality (HU: 8.1% vs EDSSU: 1.7%; most 2001;86:1580-2. P b .001), and readmission rate of 10 days or less (HU: 7.0% vs EDSSU: 9.9%; P = .02). There were no statistically Emergency short-stay unit as an effective alternative to significant differences regarding sex, chronic cor pulmonale, in-hospital admission for acute chronic obstructive and number of associated conditions. The mean hospital stay pulmonary disease exacerbation in HU was not significantly modified over the period of the study (1996 = 11.9 days; 1997 = 12.1 days; 1998 = 11.0 days). To the Editor, The strengths of our study include a large sample size and Observation and short-stay units are becoming common follow-up data on repeat ED visits after discharge. We in hospitals and are an increasingly important component of demonstrated that the introduction of an EDSSU at a tertiary the modern emergency department (ED) because they are an university hospital was associated with a decrease in the alternative to admission or discharge [1]. Acute exacerba- length of stay of patients with COPD exacerbation, along tion of chronic obstructive pulmonary disease (COPD) with a corresponding increase in the rate of repeat visits to the causes frequent hospitalizations in winter when there is ED within 10 days of discharge. It is likely that part of the often a coexisting in-hospital bed crisis and ED over- increased length of stay for the HU group is a feature of the crowding [2]. However, there are few data to establish the nature and function of inpatient services. The mean hospital duration of hospitalization in individual patients to achieve length of stay in the HU was not significantly modified over maximal benefit and to identify those patients with COPD the period of the study, which rules out a possible bias effect suitable for early discharge from the hospital [3]. of less seriously ill patients with COPD admitted to the new The objectives of the study were to determine whether a EDSSU. The increasing occurrence of relapse during the 10- new emergency department short-stay unit (EDSSU) was an day period is remarkable and suggests the need for intensified effective alternative to conventional hospital units (HUs) for ambulatory care or home care in the weeks after a short patients with exacerbation of COPD and to identify the admission in an EDSSU. clinical factors predictive of short stay at the time of However, several important limitations must be addressed. presentation to the ED in those patients. First, pulmonary function and the presence of hypercarbia co A comparative analysis (v2 or t test) was used to identify (Pa 2 of 50 mm Hg or more) were not recorded to compare differences between patients admitted to the HU (n = 1961) disease severity between the 2 groups. Other authors have and those admitted to the EDSSU (n = 545) during the described poor outcomes after hospitalization for exacerba- winter months (November 1, 1997, to March 31, 1998, and tion of COPD associated with hypercarbia [5]. Second, we November 1, 1998, to March 31, 1999). The study was cannot exclude the possibility that some patients had repeat performed at Bellvitge Hospital, a 1000-bed teaching visits to their physicians’ offices or other institutions in the 10 tertiary care referral center in Barcelona, Spain. The ED days after hospital discharge. Last, as we did not conduct a attends to approximately 110000 emergency visits per year, formal economic analysis, further work is needed to quantify excluding pediatrics and obstetrics. We retrospectively the economic impact of the introduction of an EDSSU in the studied the characteristics of patients hospitalized with an management of COPD. acute exacerbation of COPD between November 1, 1996, We conclude that selected emergency patients with acute exacerbation of COPD can be effectively and safely treated and March 31, 1999 (n = 2506). We chose charts of patients in the EDSSU. The identification of clinical factors from the hospital discharge database and selected according predictive of short stay at the time of presentation to the to the ninth revision of the International Classification of ED in patients with acute exacerbation of COPD, such as Diseases codes [4]. We used the computerized database to age of less than 65 years, Pao of greater than 60, and obtain outcome data on all patients. Clinical and demo- 2 absence of long-term oxygen therapy proved to be an graphic factors were available through chart review. Patients effective and safe measure in emergency care and a helpful were excluded from the study if they had pulmonary intervention that alleviated winter in-hospital bed crises. diagnoses at the time of hospital admission other than or in addition to COPD or if they were intubated and ventilated Albert Salazar MD on the day of admission. Discharge medications for both Antoni Juan MD groups included prednisone and adrenergic and steroid Department of Emergency Medicine metered-dose inhalers with a spacer device. Hospital de Bellvitge Statistically significant differences were found for mean University of Barcelona age (HU: 69.5 years vs EDSSU: 63.7 years; P b.05), Pao2 of Feixa Llarga s/n 08907 L’Hospitalel This study has been presented at the American College of Emergency Barcelona, Spain Physicians 2001 Research Forum, October 15 to 16, 2001, Chicago, Ill. E-mail address: [email protected] Correspondence 487

Ricard Ballbe MD to the body surface area [2]. A more current GFR calculation Department of Clinical Documentation formula is the Modification of Diet in Renal Disease Hospital de Bellvitge (MDRD) equation [3]. For serum creatinine in milligram University of Barcelona per deciliter, the estimating equation is GFR = 175 Â Feixa Llarga s/n standardized ScrÀ1.154 Â ageÀ0.203 Â 1.212 (if the subject is 08907 L’Hospitalel black) Â 0.742 (if the subject is female) [3]. Body surface Barcelona, Spain area and also ethnicity were taken into account in the MDRD equation because black people on average have higher Xavier Corbella MD creatinine levels than white people because of having Medical Direction increased muscles mass. The correlation between estimated Hospital de Bellvitge GFR and measured GFR is closer with the MDRD (R2 = 0.88) University of Barcelona than with the Cockcroft-Gault equation (R2 = 0.83). Fifty- Feixa Llarga s/n nine percent of the population studied by Brand et al was 08907 L’Hospitalel African American, so different GFR results may be calculated Barcelona, Spain by using these 2 equations. We can also see these differences by citing an example from the present study. For instance, a 43-year-old (the mean age of the study population) 70-kg doi:10.1016/j.ajem.2007.03.010 male patient with a creatinine level of 1.5 mg/dL (the critical creatinine level), as Dr Band pointed out, has a GFR level of 62.87 mL/min if it is calculated by the Cockcroft-Gault References equation. However, the same patient will have a GFR level of 65.68 mL/min if he is an African American and 54.28 mL/min [1] Graff L. Overcrowding in the emergency department: an international (which means impaired renal function if he is not an African symptom of health care system failure. Am J Emerg Med 1999;17:208-9. American) if it is calculated by the MDRD equation. [2] Hanratty B, Robinson M. Coping with winter bed crises. New Finally, some patients with impaired renal function may surveillance systems might help. Br Med J 1999;319:1511-2. be overlooked by using only the Cockcroft-Gault equation. [3] Kong GK, Belman MJ, Weingarten S. Reducing length of stay for The MDRD equation should also be used with the patients hospitalized with exacerbation of COPD by using a practice Cockcroft-Gault equation to prevent missing patients with guideline. Chest 1997;111:89-94. [4] International Classification of Diseases, 9th revision, Clinical Mod- impaired renal function and critical creatinine levels. ifications. 4th ed. Salt Lake City7 Med-Index Publications, Utah; 1993. [5] Connors Jr AF, Dawson NW, Thomas C, et al. Outcomes following acute exacerbation of severe chronic obstructive lung disease. The Cenker Eken MD SUPPORT investigators (Study to Understand Prognoses and Prefer- Department of Emergency Medicine ences for Outcomes and Risks of Treatments). Am J Respir Crit Care Akdeniz University Medical Faculty Med 1996;154:959-67. 07059 Antalya, Turkey E-mail address: [email protected] ˙ Differences between various glomerular filtration rate Isa Kilicaslan MD calculation methods in predicting patients at risk for Department of Emergency Medicine contrast-induced nephropathy Guven Hospital Ankara, Turkey E-mail address: [email protected] To the Editor, doi:10.1016/j.ajem.2007.03.023 Band et al [1] make an important point when they alert readers to the hazards of predicting renal function by a single creatinine measurement, which can be affected by References multiple factors. We write to emphasize several other points that are also worthy of note in measuring glomerular [1] Band RA, Gaieski DF, Mills AM, et al. Discordance between serum creatinine and creatinine clearance for identification of ED patients with filtration rate (GFR) in patients undergoing contrast-related abdominal pain at risk for contrast-induced nephropathy. Am J Emerg diagnostic procedures in the emergency setting. Med 2007;25:268-72. Band et al used the Cockcroft-Gault equation to calculate [2] Stevens LA, Coresh J, Greene T, et al. Assessing kidney function— the GFR levels of the study patients. The Cockcroft-Gault measured and estimated glomerular filtration rate. N Engl J Med equation uses age, weight, and serum-measured creatinine 2006;354:2473-83. [3] Levey AS, Coresh J, Greene T, et al. Expressing the modification of levels. But it may overestimate GFR because of the free diet in renal disease study equation for estimating glomerular filtration secretion of creatinine from the proximal tubules. The rate with standardized serum creatinine levels. Clin Chem 2007 [Epub Cockcroft-Gault equation has not been adjusted according ahead of print]. American Journal of Emergency Medicine (2007) 25, 488.e1–488.e5

www.elsevier.com/locate/ajem

Case Reports Traumatic inferior gluteal artery pseudoaneurysm: vomiting, any change in bowel habits, chest pain, or case report and review of literature shortness of breath. At the time of admission, vital signs were normal: blood Inferior gluteal artery pseudoaneurysms are rare, with pressure, 122/68 mm Hg; pulse rate, 69 beats per minute; few cases reported in the recent literature. Such cases often respiratory rate, 20 breaths per minute; temperature, 35.68C present with a variable time course, mode of injury, and (96.18F); SaO2, 99% to 100% on room air; and complaint of associated symptoms, leading to their misdiagnoses and 10/10 pain. The physical examination was remarkable for a 4 improper treatment. We report a case of a posttraumatic Â 5-cm, warm, indurated, well-circumscribed, fluctuant mass inferior artery pseudoaneurysm that presented 3 months on the right buttock, which was very tender to palpation. after the initial injury with a warm, tender, enlarging mass A bruit was not appreciated on auscultation, and no thrill and numbness in the ipsilateral lower extremity. This article or pulsations were felt over the mass. Erythema highlights the importance of placing a pseudoaneurysm in was not appreciated secondary to the patient’s dark skin the differential diagnosis of an indurated, fluctuant, warm, color. An ultrasonographic examination of the mass revealed erythematous posttraumatic mass, despite the absence of a well-demarcated, fluid-filled structure with loculations. No thrills, bruits, and pulsations. photographs were printed of this ultrasound examination, and Gluteal artery pseudoaneurysms are rare, and most of a color-flow Doppler ultrasonography was not performed. these cases are due to blunt or penetrating trauma [1]. Most A field block of the region was done with 10 mL of 2% such cases involve the superior gluteal arteries [2]. A recent lidocaine, with good anesthesia. The region was prepared review of the literature found only 10 cases of inferior and draped sterilely. A 1.5-cm horizontal incision was made gluteal artery involvement in the past 30 years [2].We along the mass, yielding a small amount of pus. Immedi- report a case of an inferior gluteal artery pseudoaneurysm ately after the pus drainage, a high-pressure arterial bleed 3 months after sustaining a penetrating trauma to the was visualized, which was controlled with direct pressure. right buttock to highlight the problems in their diagnosis Repeated vital signs revealed the following: blood pressure, and management. 83/65 mm Hg; pulse rate, 69 beats per minute; respiratory A 34-year-old African American man presented with an rate, 16 breaths per minute; and temperature, 36.48C binfectionQ to his right buttock after sustaining a penetrating (97.58F). The patient denied any nausea, vomiting, or trauma with an apple corer 3 months prior. The stab wound lightheadedness at that time. Surgery was called for a had been closed with sutures 3 months ago at a local consultation and evaluation of the lesion. hospital, with no apparent complications at that time. The wound was further explored by the surgical team in However, the extent of the wound exploration before closure the ED, resulting in the removal of several large clots. The was unknown. As per the patient, he was started on a course patient lost approximately 1.5 L of blood during this of penicillin with poor compliance. procedure, resulting in subjective feelings of dizziness and Four days before his presentation at our emergency nausea, with 2 episodes of vomiting. Repeated vital signs at department (ED), the patient was again evaluated at his local this time were as follows: blood pressure, 70/41 mm Hg; hospital for increasing pain and pressure to the right gluteal pulse rate, 70 beats per minute; respiratory rate, 18 breaths per region, with pain and tingling radiating down his right minute; and SaO2, 98% on room air. Two large-bore leg. The lesion was evaluated, and the patient was given intravenous lines were placed in his extremities, and the pain medication and sent to our ED for drainage of a right patient was given 2 L of isotonic sodium chloride, with an gluteal babscess.Q improvement in his vital signs and symptoms. Chemistries, Upon presentation, the patient complained of increasing complete blood count, prothrombin time/international nor- pain and pressure in the right gluteal area, with increased malized ratio/partial thromboplastin time, and a type and swelling, firmness, and warmth for 1 week. Four days cross were sent; and 2 U of packed red blood cells was prior, the patient developed pain and tingling radiating ordered. The patient’s initial hemoglobin and hematocrit down his right leg, 10/10, with no alleviating factors and levels were 11.3 g/dL and 33.4%, respectively. A repeated made worse with straightening of the leg. The patient complete blood count sent 2 hours later noted a drop in his denied a fever, chills, lightheadedness, dizziness, nausea, hemoglobin and hematocrit to 8.6 g/dL and 24.7%, respec-

0735-6757/$ – see front matter D 2007 Elsevier Inc. All rights reserved. 488.e2 Case Reports tively. The patient was transfused with 1 U of packed red identified in the literature include after bone grafts, hip blood cells immediately before repair of the defect. prostheses, pelvic fractures, intramuscular injections, muscle The patient underwent an inferior gluteal artery catheter- strain, and penetrating traumas such as stabbings or gunshot ization and pelvic angiogram in the interventional radiology wounds [5,6]. suite (Fig. 1). An unsuccessful attempt was made to embolize The symptoms associated with gluteal artery pseudoa- the pseudoaneurysm, and the patient was immediately trans- neurysms are also variable, including localized swelling, ferred to the operating room for ligation of the defect. Post- pain, and pressure [4]. The bclassic signQ is a pulsation over operatively, the patient remained hemodynamically stable the mass, which may be absent secondary to an incomplete and was discharged home within a few days of admission. or small defect in the arterial wall [7]. Sciatic nerve A review of the literature demonstrates that inferior compression secondary to increased pressure on the nerve gluteal artery pseudoaneurysms are a rare phenomenon. The bundle is also a common presentation associated with aim in presenting this case is to highlight the importance of gluteal artery pseudoaneurysms. Our patient reported considering a pseudoaneurysm in the differential diagnosis numbness and tingling radiating down the back of his right when confronted with an indurated, fluctuant, warm, leg, which was immediately relieved upon decompression of erythematous posttraumatic mass, despite the absence of the mass. However, it is important to bear in mind that the thrills, bruits, and pulsations. compressive force can also result in neurological impair- A firm posttraumatic mass should heighten a clinician’s ments in the superior or inferior gluteal, pudendal, and index of suspicion for a possible pseudoaneurysm. Howev- posterior cutaneous nerves [8]. er, these lesions are often misdiagnosed, especially when Given the variable time course, mode of injury, and pseudoaneurysms are not considered in the differential symptoms associated with pseudoaneurysms, they are often diagnosis. Other possible diagnoses include abscesses, misdiagnosed and improperly treated. Gilroy et al provide an lipomas or sarcomas, sciatic hernias, or cysts [1]. algorithm for the management of suspected cases of arterial Misdiagnoses of these lesions may be attributed to their involvement in posttraumatic masses [4]. The management variable presentations. The time course from the traumatic of arterial lesions varies depending on the time course since incident to the development of a gluteal pseudoaneurysm presentation. In late presentations, the importance of early can vary. The literature cites presentations ranging from angiography cannot be overemphasized, as the consequence weeks to several years after the initial insult [2-4]. Our of an erroneous diagnosis, and the incision and drainage of patient presented with symptoms 3 months after the initial such a lesion, can be disastrous [4]. penetrating trauma, with no prior warning signs to indicate a Diagnostic techniques for such lesions include angiog- pseudoaneurysm. raphy, aspiration, computed tomography, magnetic reso- Formation of a gluteal artery pseudoaneurysm can occur by several different mechanisms. Modes of injury nance imaging, and color-flow Doppler ultrasonography [9]. Angiography has several benefits, including the ability to identify the distal and proximal tributaries of the affected artery [4]. According to Agarwal et al, a routine aortic flush angiogram may reveal a normal anatomy [2]. Therefore, a selective angiography of the internal iliac artery or the feeding vessel would be best suited to diagnose an inferior gluteal artery pseudoaneurysm [2]. Aspiration with a needle can identify arterial involvement before an incision for drainage; however, the lack of a sanguinous aspirate should not be used to rule out the diagnosis [4]. Bennett et al discussed the use of computed tomography to correctly diagnose the lesion and suggested that dynamic imaging techniques such as magnetic resonance imaging and Doppler ultrasound can also be beneficial [9]. As with the diagnostic techniques, several options exist for the treatment of inferior artery pseudoaneurysms. Such lesions may be repaired using either surgical or less invasive techniques, such as embolization during angiography. The advantages of angiography with embolization include a decreased risk of infection, the avoidance of the release of the tamponade effect of the hematoma, and the evasion of the retroperitoneal space [2]. However, when visualization Fig. 1 Angiogram and embolization of gluteal artery of associated structures, such as the sciatic nerve, are pseudoaneurysm. important, an open surgical technique is preferable [4]. Case Reports 488.e3

The illustrated case demonstrates the importance of predominantly affects elderly diabetic men. The apparently placing arterial pseudoaneurysms on the differential diag- high mortality and morbidity associated with acute emphy- noses of posttraumatic masses. The variability of the time sematous cholecystitis have previously emphasized the course, mode of injury, and symptoms necessitates a high importance of prompt diagnosis and emergent surgical clinical suspicion, as the incision and drainage of such intervention. Radiological evaluation including plain ab- lesions can result in life-threatening consequences. Algo- dominal radiograph (KUB), abdominal sonography, and rithms of the diagnoses and management of such lesions computed tomography of abdomen is the cornerstone of indicate angiographic examination followed by repair; diagnosis of acute emphysematous cholecystitis. We de- however, prompt recognition remains paramount to reduce scribe a nondiabetic woman who developed sepsis due to morbidity and mortality. acute emphysematous cholecystitis with rapid deterioration within 24 hours. She was misdiagnosed as having peptic Ani Aydin ulcer disease initially due to normal KUB, abdominal SUNY Stony Brook School of Medicine sonography, and computed tomography of abdomen. Stony Brook, NY 11794, USA

Christopher C. Lee MD Case report Eric Schultz MD Jeremy Ackerman MD A 69-year-old nondiabetic woman was brought to the Department of Emergency Medicine emergency department (ED) because of epigastric pain for Stony Brook University Hospital 2 hours. She had history of peptic ulcer disease. The pain Stony Brook, NY 11794, USA persisted without vomiting and diarrhea. Notably, she was E-mail address: [email protected] afebrile (35.68C). The following initial vital signs were stable: blood pressure, 140/58 mm Hg; pulse rate, doi:10.1016/j.ajem.2006.11.015 71 beats/min; and respiratory rate, 18 breaths/min. Physical examination showed mild tenderness at epigastric area References without obvious Murphy sign. Blood test revealed slight leukocytosis (white blood cell count, 10,000/mm3) and [1] Herber SC, Ajalat GM, Smith DC, Hinshaw Jr DB, Killeen Jr JD. normal hepatobiliary values (glucose, 112 mg/dL; aspartate Transcatheter embolization facilitating surgical management of a aminotransferase (AST), 23 U/L; total bilirubin, 0.4 mg/ giant inferior gluteal artery pseudoaneurysm. J Vasc Surg 1998;8(6): 716-20. dL; lipase, 33 U/L). Both KUB and bedside sonography [2] Agarwal M, Giannoudis PV, Syed AA, Hinsche AF, Matthews SJ, showed no significant finding. Abdominal computed Smith RM. Pseudoaneurysm of the inferior gluteal artery following tomography was arranged to determine the nature of the polytrauma: diverse presentation of a dangerous complication: a report acute abdomen but was negative as well (Fig. 1). After of two cases. J Orthop Trauma 2003;17(1):70-4. observation for several hours with improved symptoms, [3] Williams Jr W, Jackson Jr GF, Greene C. Superior gluteal artery aneurysm. J Trauma 1997;17(6):477-9. [4] Gilroy D, Saadia R, Hide G, Demetriades D. Penetrating injury to the gluteal region. J Trauma 1992;32(3):294-7. [5] Barker SG, Anthony AA, Pillay SS, Porter AJ, Davies RP, Jury P. Sporting ’groin strains’: not always muscular!. Aust N Z J Surg 1995; 65(6):451-3. [6] Kaplan JL, Challenor Y. Posttraumatic osseous tunnel formation causing sciatic nerve entrapment. Arch Phys Med Rehabil 1993;74(5):552-4. [7] Mikulin T, Walker EW. False aneurysm following blunt trauma. Injury 1984;15(5):309-10. [8] Papadopoulos SM, McGillicuddy JE, Messina LM. Pseudoaneurysm of the inferior gluteal artery presenting as sciatic nerve compression. Neurosurgery 1989;24(6):926-8. [9] Bennett JD, Brown TC, Coates CF, MacKenzie D, Sweeney J. Pseudoaneurysm of the inferior gluteal artery. Can Assoc Radiol J 1992;43(4):296-8. Case Report Acute emphysematous cholecystitis with initial normal radiological evaluation: a fatal diagnostic pitfall in the ED

Acute emphysematous cholecystitis is a relatively rare Fig. 1 Initial abdominal computed tomography showed no disease, a severe variant of acute cholecystitis, that obvious finding. 488.e4 Case Reports she was discharged with the possible diagnosis of peptic ulcer disease. Unfortunately, she returned to the ED with sepsis 8 hours after discharge due to recurrent epigastric pain. She was found to have fever (38.58C), tachycardia (115 beats/min), and normal blood pressure (129/62 mm Hg). Physical examination showed severe tenderness at epigastric area. Left decubitus abdominal x-ray was taken approximately 24 hours after the previous abdominal computed tomography for ruled out peptic ulcer perforation. It incidentally showed characteristic gallbladder distention, a circumferential gallbladder wall gas lucency, and an intraluminal air fluid level (Fig. 2). Acute emphysematous cholecystitis was highly suspected then. Abdominal sonography showed gas- forming lesion near gallbladder with nonvisible gallbladder. Repeated abdominal computed tomography confirmed our suspicion (Fig. 3), and she underwent emergent open cholecystectomy. Acalculous, ischemic, and gangrenous gallbladder was found. The patient recovered well after surgical intervention. Fig. 3 Abdominal computed tomography showed emphysema- Acute emphysematous cholecystitis is a relatively rare tous cholecystitis. disease, a severe variant of acute cholecystitis, which exists with several differences compared with simple acute the importance of prompt diagnosis and emergent surgical cholecystitis. It predominantly affects elderly men (men intervention [1-3]. forming 71% of patients with emphysematous cholecystitis Radiological evaluation including plain KUB, abdominal but only 27% in acute cholecystitis). A high incidence of sonography, and computed tomography of the abdomen is diabetes mellitus (up to 50%) and high frequency of the cornerstone of diagnosis of acute emphysematous acalculous cholecystitis were noted too. The mortality rate cholecystitis; indeed, there is no report of diagnosis being for uncomplicated acute cholecystitis is approximately made through high clinical suspicion without a radiological 1.4%. The mortality rate for acute emphysematous workup. Laboratory tests are nonspecific and might only cholecystitis, however, is 15% to 20%, owing to the show indirect evidence of acute systemic infection, such as increased incidence of gallbladder wall gangrene and leukocytosis [4,5]. This makes the early diagnosis of acute perforation in these patients. This variance emphasizes emphysematous cholecystitis challenging for any emergency physician. The fatal diagnostic pitfall in this patient was that she presented to the ED with no fever, relatively normal blood test, normal KUB, normal abdominal sonography, and even normal abdominal computed tomography initially. Furthermore, she is a nondiabetic woman, which considerably lowered the index of suspicion for acute emphysematous cholecystitis. She returned within 24 hours with severe gangrenous emphysematous cholecystitis, an insidious and rapidly progressive disease with probable fatal outcome [6]. In conclusion, acute emphysematous cholecystitis should be kept in mind as a cause of epigastric pain even in an afebrile, nondiabetic patient with a negative thorough preliminary imaging (plain KUB, abdominal sonography, and abdominal computed tomography) to avoid missing this curable but rapidly progressing fatal disease.

Vei-Ken Seow MD Chiu-Mei Lin PhD, MD Emergency Department Fig. 2 Plain abdominal radiograph showed circumferential Shin-Kong Wu Ho-Su Memorial Hospital gallbladder wall gas lucency with intraluminal air-fluid level. Taipei, Taiwan, ROC Case Reports 488.e5

Tzong-Luen Wang PhD, MD [2] Garcia-Sancho Tellez L, Rodriguez-Montes JA, Fernandez De Lis S, Chee-Fah Chong MS, MD et al. Acute emphysematous cholecystitis. Report of twenty cases. I-Yin Lin MD Hepatogastroenterology 1999;46:2144-8. Emergency Department [3] Mentzer RM, Golden GT, Chandler JG, et al. A comparative appraisal of emphysematous cholecystitis. Am J Surg 1975;129:10-5. Shin-Kong Wu Ho-Su Memorial Hospital [4] de Araujo DB, Renck DV, de Britto MAP, et al. Emphysematous Taipei, Taiwan, ROC cholecystitis: an unusual presentation of a rare disease. MJM Fu Jen Catholic University 2004;8:28-30. E-mail address: [email protected] [5] Gill KS, Chapman AH, Weston MJ, et al. The changing face of emphysematous cholecystitis. Br J Radiol 1997;70:986-91. [6] Yeatman TJ. Emphysematous cholecystitis: an insidious variant of doi:10.1016/j.ajem.2006.11.023 acute cholecystitis. Am J Emerg Med 1988;4:163-6.

References

[1] Bedirli A, Sakrak O, Sozuer EM, et al. Factors effecting the complications in the natural history of acute cholecystitis. Hepatogas- troenterology 2001;48:1275-8.