Successful Selective Nonoperative Management of Abdominal Gunshot Wounds Despite Low Penetrating Trauma Volumes

PAPER Successful Selective Nonoperative Management of Abdominal Gunshot Wounds Despite Low Penetrating Trauma Volumes

Karim Fikry, MD; George C. Velmahos, MD, PhD; Athanasios Bramos, MD; Sumbal Janjua, MD; Marc de Moya, MD; David R. King, MD; Hasan B. Alam, MD

Objective: To determine whether selective nonoper- tially managed by selective nonoperative management (25 ative management of abdominal gunshot wounds (AGSW) of 99 anterior and 13 of 26 posterior AGSW patients). is safe in trauma centers with a low volume of penetrat- Seven selective nonoperative management patients re- ing trauma. ceived delayed laparotomy as late as 11 hours after admission. At the end, 30 of the 125 patients (24%) were Design: Retrospective study. successfully managed without an operation (20 of 99 anterior and 10 of 26 posterior AGSW patients). There were Setting: Academic level 1 trauma center with approxi- no predictors of delayed laparotomy and no complica- mately 10% penetrating trauma. tions or mortality attributed to it. Ten patients (8%) had a nontherapeutic laparotomy, and 3 of them developed Patients: All patients with anterior and posterior AGSW complications. (January 1, 1999, through December 31, 2009), exclud- ing tangential injuries, transfers, and deaths in the emergency department. Patients with hemodynamic instabil- Conclusions: Selective nonoperative management of ity or peritonitis received an urgent laparotomy. The AGSW is feasible and safe in trauma centers with low pen- remaining patients had selective nonoperative manage- etrating trauma volumes. Nearly 1 in 4 AGSW patients ment. A delayed laparotomy was offered for worsening does not need a laparotomy, and nontherapeutic lapa- symptoms or worrisome computed tomography find- rotomies are associated with complications. The vol- ings. ume of AGSW per se should not be an excuse for routine laparotomies. These data become particularly Main Outcome Measures: Hospital stay, complica- important because penetrating trauma volumes are de- tions, and mortality. creasing around the country.

Results: Of 125 AGSW patients, 38 (30%) were ini- Arch Surg. 2011;146(5):528-532

FTER DECADES OF OPERAT- volumes are dwindling around the coun- ing on abdominal gunshot try,14-16 and experience with the manage- wounds (AGSW) rou- ment of AGSW is concentrated in only a tinely, Shaftan and few urban trauma centers. Most other McAlvanah1,2 proposed in trauma centers admit limited volumes of theA late 1960s “selective conservatism” as penetrating trauma. This lack of exposure has been repeatedly mentioned as a rea- See Invited Critique son to avoid SNOM and to prefer routine laparotomy.4,13,17,18 at end of article In a pattern of admissions similar to that of most other trauma centers, our pen- an alternative. In the 1990s, studies from etrating trauma admissions comprise less South Africa and Los Angeles reported on than 15% of our entire volume. The ratio the effectiveness and safety of selective non- of stab wounds to gunshot wounds is 2:1, operative management (SNOM) of AGSW which allows a rather limited annual ex- Author Affiliations: Division of in a large series of patients evaluated ret- posure of the staff and residents to AGSW. Trauma, Emergency Surgery, 3-8 and Surgical Critical Care, rospectively and prospectively. Despite We hypothesized that by following a Massachusetts General Hospital the encouraging results and follow-up stud- simple algorithm for the management of and Harvard Medical School, ies in the 2000s,9-13 this practice is still not AGSW, as previously described in the lit- Boston, Massachusetts. universally accepted. Penetrating trauma erature, SNOM of AGSW is feasible and

ARCH SURG/ VOL 146 (NO. 5), MAY 2011 WWW.ARCHSURG.COM 528

©2011 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 safe despite the lack of large penetrating injury volumes. Initial period Subsequent period Hemodynamic instability Hemodynamic instability Peritonitis Peritonitis METHODS Nonevaluable patient

This is an 11-year retrospective study (January 1, 1999, through Immediate laparotomy Immediate laparotomy December 31, 2009) of patients with anterior or posterior AGSW admitted to our academic level 1 trauma center. Patients with All others All others obvious tangential injuries and deaths in the emergency department (ED) were excluded. Similarly, transfers from other SNOM CT scan hospitals were excluded because we did not have precise admission data for analysis. The abdomen was defined as the area between the nipple line Negative Equivocal Positive and pubic symphysis anteriorly and between the tip of the scapu- lae and the gluteal folds posteriorly. The midaxillary lines served Individualization to divide the anterior from the posterior part. The patients were managed by a trauma team that, inevitably, was subjected to Figure. Algorithm of management of abdominal gunshot wound patients multiple changes in personnel and protocols during the study during the entire study period ( January 1, 1999, through December 31, period. After 2004, the trauma center was managed by an un- 2009). There was no specific transition year between the initial and changing dedicated trauma and acute care surgery team. The subsequent periods. The transition was gradual because computed basic infrastructure and focus in trauma care was present tomography (CT) was increasingly used in the 2000s. SNOM indicates selective nonoperative management. throughout the study. The management of AGSW patients fol- lowed a relatively simple protocol. Patients who were hemo- dynamically unstable, had peritonitis, or were clinically un- center’s infrastructure, which allows immediate access to re- evaluable (eg, with intubation or head injury) received an sources such as CT and the operating room. It was assumed immediate exploratory laparotomy (IMMLAP). Selective non- that if a patient were not taken to the operating room within 2 operative management was offered to all other patients. The hours, a deliberate decision was made against IMMLAP. The patients were initially observed either in the ED or regular wards, SNOM patients who eventually had a laparotomy were in- but after 2004 a dedicated and highly monitored observa- cluded in the DELLAP group. The IMMLAP, DELLAP, and tional unit was built to house such patients during the first 12 SNOM patients were compared. The t test was used for con- to 24 hours after injury. The increasing use of computed to- tinuous variables and the ␹2 or Fisher exact test for categorical mography (CT) for penetrating trauma patients in the past 10 variables. A P value of .05 was considered statistically signifi- years interfered with the initial simple algorithm because more cant. The study was approved by the institutional review board. patients underwent CT before the final decision was made. In the last 5 years of the study, all patients managed by SNOM had a CT scan. In short, AGSW patients with hemodynamic RESULTS instability or obvious peritonitis were still taken directly to the operating room. The remaining patients underwent CT in a fre- Of 150 AGSW patients admitted during the study pe- quency that increased over the years. There was no specific point riod, 12 died in the ED, 11 were transferred from other of transition from no CT to CT but rather a gradual increase in use because the technology evolved and the confidence of phy- hospitals, and 2 had tangential wounds and were admit- sicians in its findings grew. Patients with clear evidence of clini- ted for other reasons. The remaining 125 patients con- cally significant injuries on CT were taken to the operating room, stitute the study population. Of them, 87 patients (70%) even if diffuse peritonitis or hemodynamic instability was not received an IMMLAP because of hemodynamic instabil- yet developed. Patients with bullet trajectories away from vital ity (n=30), peritonitis (n=38), inability to evaluate clini- structures were offered SNOM. Patients with equivocal signs cally (n=5), or CT scan findings suggestive of clinically on CT were managed operatively or nonoperatively at the dis- significant organ injury (n=14). Of the 38 SNOM pa- cretion of the trauma surgeon (Figure). If, during observa- tients, 30 (79% of SNOM patients and 24% of all pa- tion, patients developed clinical symptoms consistent with sig- tients) were successfully discharged without the need for nificant organ injury, a delayed laparotomy (DELLAP) was an exploratory laparotomy. Seven patients received a offered. An operation was considered nontherapeutic if it failed to find any clinically significant organ injuries, ie, injuries re- DELLAP (18% of SNOM patients and 6% of all patients) quiring surgical intervention. For example, a nonbleeding su- because of worsening abdominal signs and symptoms, perficial liver injury was not considered clinically significant often in the presence of suspicious CT scan findings. No and, if this were the only finding, the laparotomy was defined patients had a DELLAP because of hemodynamic insta- as nontherapeutic. bility. Five of the DELLAP patients had a CT scan that The following information was collected: age, sex, site of showed either a solid parenchymal injury or free fluid gunshot wound (anterior or posterior), vital signs and hemat- without an apparent injury. The patients did not satisfy ocrit level on admission, Abbreviated Injury Score for the ab- criteria for immediate exploration and were observed un- domen, Injury Severity Score, time from arrival at the ED to til the abdominal tenderness increased. There was no dif- transfer to the operating room, and hospital course, including ference in the distribution of DELLAP throughout the morbidity, duration of hospital stay, and mortality. Patients who were taken to the operating room within 2 hours of arrival in study period. The mean interval from arrival at the ED the ED were included in the IMMLAP group. Patients who had to DELLAP was 5 hours (range, 3-11 hours). Table 1 a clear note in the medical record about SNOM or who were describes the 7 DELLAP patients. One of them died on not operated on within 2 hours were included in the SNOM the day after admission because of an associated gun- group. The 2-hour limit was set arbitrarily on the basis of our shot wound to the head. None of the other DELLAP pa-

ARCH SURG/ VOL 146 (NO. 5), MAY 2011 WWW.ARCHSURG.COM 529

Patient Site of Time From No. Age, y ISS AGSW ED to OR, h Laparotomy Findings Operation LOS, d 1 35 59 A 3.5 Kidney grade V Nephrectomy 1a 2 36 10 A 3.2 Spleen grade II, diaphragm Repair of diaphragm 4 3 23 25 P 4.5 Small bowel Enterectomy 8 4 23 25 A 3.0 Spleen grade III, distal pancreas Splenectomy, distal pancreatectomy, gastric 7 grade II, stomach repair 5 27 34 P 5.0 Stomach, diaphragm Gastric and diaphragmatic repair 6 6 20 9 A 4.8 Stomach, liver, kidney Gastric repair 10 7 28 20 A 11.2 Right colon Primary repair 4

Abbreviations: A, anterior; AGSW, abdominal gunshot wound; ED, emergency department; ISS, Injury Severity Score; LOS, length of stay; OR, operating room; P, posterior. a Patient died of an associated gunshot wound to the head.

Table 2. Comparison of Patients With AGSW Who Received Table 3. Comparison of Patients Who Received an an IMMLAP With Those Initially SNOMa Operation During Their Hospital Stay With Those Who Were Successfully Managed Nonoperativelya IMMLAP SNOM P Characteristic (n = 87) (n = 38) Value Successfully Managed Age, y 25 (7) 25 (8) .85 Operated On Nonoperatively P Male sex, No. (%) 84 (97) 37 (97) .81 Characteristic (n = 95) (n = 30) Value Anterior AGSW, No. (% of 74 (85) 25 (66) .02 category) (n = 99) Age, y 25 (7) 25 (8) .66 Posterior AGSW , No. (% of 13 (15) 13 (34) Male sex, No. (%) 92 (97) 29 (97) .96 category) (n = 26) Anterior AGSW, No. (% 79 (83) 20 (67) .05 Abbreviated Injury Score abdomen 3 (1) 2 (2) Ͻ.01 of category) (n = 99) Injury Severity Score 17 (10) 14 (12) .03 Posterior AGSW, No. (% of 16 (17) 10 (33) Blood pressure on admission, 132 (28) 143 (24) .03 category) (n = 26) mm Hg Abbreviated Injury Score 3 (1) 1 (1) Ͻ.01 Heart rate on admission, beats/min 97 (22) 103 (21) .25 abdomen Hematocrit 38 (5) 38 (6) .72 Injury Severity Score 28 (11) 11 (10) Ͻ.01 Length of stay, d 11 (12) 6 (6) Ͻ.01 Blood pressure on admission, 132 (27) 143 (27) .05 Complications, No. (%) 28 (32) 4 (11) .01 mm Hg Mortality, No. (%) 3 (3) 1 (3) .81 Heart rate on admission, 97 (22) 104 (23) .18 beats/min Abbreviations: AGSW, abdominal gunshot wound; IMMLAP, immediate Hematocrit 38 (5) 38 (7) .32 exploratory laparotomy; SNOM, selected for nonoperative management. Length of stay, d 11 (11) 6 (7) Ͻ.01 aData are given as mean (SD) unless otherwise indicated. Complications, No. (%) 29 (31) 3 (10) .03 Mortality, No. (%) 4 (4) 0 .25

tients developed a complication. Of the 31 patients who aAbbreviation: AGSW, abdominal gunshot wound. Data are given as mean were successfully managed by SNOM, 24 had an abdomi- (SD) unless otherwise indicated. nal CT scan. In 15 patients, there were no injuries found; the remaining 9 had the following findings: nonbleed- tients, 3 developed complications. One had pneumonia, ing liver lacerations (n=5), renal lacerations (n=1), com- another had a protracted urinary tract infection, and the bined liver and renal lacerations (n=2), and bowel wall third had deep venous thrombosis, pneumonia, and mul- edema (n=1). tiple infections. By comparison of SNOM patients with IMMLAP patients, we found that SNOM patients had fewer complications and shorter hospital stays (Table 2). By com- COMMENT parison of patients with successful SNOM with all patients who received a laparotomy (IMMLAP and DELLAP), we Since the introduction of selective conservatism for pen- again found fewer complications and shorter hospital stays etrating abdominal wound by Shaftan,1 McAlvanah and among successful SNOM patients (Table 3). Shaftan,2 NanceandCohn,19 DemetriadesandRabinowitz,20,21 There were 10 patients (8%) with a nontherapeutic and Demetriades et al,22 the concept has found wide- laparotomy, all of them included in the IMMLAP group. spread application for stab wounds but not for gunshot Two patients had nonbleeding liver injuries, 1 had a non- wounds. Born out of necessity more than ingenuity in bleeding splenic injury, 1 had a small serosal tear in the the high-volume trauma centers of South Africa3,7,20-22 and small bowel, 1 had a nonexpanding renal hematoma, and the county hospitals of the United States,1,2,19 where some 1 showed intraperitoneal blood without evidence of ac- patients were unexpectedly found to improve while wait- tive bleeding. In addition, there were 4 laparatomies that ing for an operation, the selective approach emerged revealed no intra-abdominal injuries. Of these 10 pa- as a valid alternative to routine laparotomy. For reasons

ARCH SURG/ VOL 146 (NO. 5), MAY 2011 WWW.ARCHSURG.COM 530

©2011 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 that are poorly understood, stab wounds were quickly imity, free air, or free fluid.12 We, like others, have been considered appropriate for SNOM but AGSW were not. using CT scan with increasing frequency during the past The arguments for routine laparotomy after AGSW—the few years and have incorporated it in our therapeutic al- unreliability of a clinical examination, the high likelihood gorithm. We have been reluctant to discharge patients of organ injury, and the “benign” nature of a nonthera- from the ED on the basis of negative CT scan results un- peutic laparotomy—have been debated for years and re- less the trajectory is obviously tangential. Such patients futed in a number of studies.4,8,18,23 In short, we and others with tangential wounds were excluded from our analy- have argued that (1) if a clinical examination is reliable sis. In general, the precise role of CT scan for AGSW is for stab wounds, then it should be as reliable for AGSW6,8; not supported by high-level evidence. It has become a (2) the likelihood for intraperitoneal penetration and or- widespread practice and there is evidence supporting its gan injury is high, but the likelihood for clinically signifi- use, but to our knowledge, there have been no compara- cant organ injury is lower than initially thought4; and (3) tive studies proving that patient outcomes have im- nontherapeutic laparotomies are anything but “benign” proved compared with patient outcomes relying primar- and are associated with a significant rate of postoperative ily on clinical examination. 24,25 complications. Seven SNOM patients developed symptoms during the For the most part, SNOM for AGSW has been prac- period of observation and received a DELLAP that proved ticed in large trauma centers with experienced trauma therapeutic on all of them. Five of them had anterior teams that manage high volumes of penetrating trauma. AGSW in contrast to the common belief that posterior A legitimate concern remains that this concept may not AGSW are more likely to produce delayed symptoms from be applicable to centers with lower numbers and inevi- retroperitoneal organ injuries. Selective nonoperative man- tably limited experience in managing AGSW. However, agement for patients with gunshot wounds to the back this argument would prevent the majority of trauma cen- and buttocks is similarly safe.6,22,29 The mean delay to op- ters around the country from practicing SNOM because eration was only 5 hours, with the longest delay being penetrating trauma volumes are on the decrease and most 11 hours. Within these short periods, no adverse events trauma centers—even level 1—treat only a few AGSW were caused by the delays. This study confirms the find- patients per year. Such volumes, distributed across the ings of previous research, indicating that an observation entire trauma faculty of a specific center, would allow only period of 24 hours is adequate for most patients and that single-digit numbers of AGSW to be managed by any given the operative risk does not increase in patients initially trauma surgeon, offering theoretically a rather inad- managed by SNOM and eventually requiring an opera- equate experience. The problem may be even worse in tion.4,6-8,10 smaller trauma centers.14-16,26 The benefits of SNOM are obvious. Patients who even- To evaluate this argument, we set out to examine the management of AGSW in our trauma center. The Mas- tually did not receive an operation had fewer complica- sachusetts General Hospital is the largest trauma center tions and shorter hospital stays than those operated on. in New England but, nevertheless, admits penetrating The financial superiority of SNOM over routine lapa- trauma volumes that are relatively low compared with rotomy has been proved in previous studies unequivo- cally.4 The rate of nontherapeutic laparotomy in our study those of centers that have published on SNOM previ- 2-4,8,13 ously. On average, every year we admit 278 patients with was 8% and similar to that of other studies. It seems penetrating injuries. Approximately 120 of these inju- that this rate represents an irreducible statistic and a point ries are caused by a gunshot wound, with an even smaller of reference for successful SNOM protocols. number (approximately 15) located in the abdomen. Dur- Our study makes the case that a large volume of AGSW ing those years, 5 to 8 surgeons shared the trauma calls, is not a prerequisite for practicing SNOM. Obviously, the indicating that each was managing approximately 2 or 3 infrastructure of a mature level 1 trauma center, the col- AGSW per year. Thirty-eight of these patients (30%) were lective experience of the trauma surgeons group, and the offered a trial of SNOM (approximately one-third of an- functioning protocols assuring around-the-clock moni- terior [66%] and two-thirds of posterior [34%] AGSW). toring were crucial elements of the uneventful applica- Finally, 30 patients (24%) (25 [66%] with anterior and tion of SNOM in our AGSW population. Our hospital may 13 [34%] with posterior AGSW) were discharged from serve as an example to other trauma centers with pre- the hospital without an operation. These numbers are re- dominantly blunt trauma populations on the feasibility markably similar to those reported in the largest analy- and safety of SNOM. The modest delays may be a result sis of AGSW to date.4 In that study, 42% of the patients of our strict definition of IMMLAP (within 2 hours of ar- were initially managed by SNOM and 38% were dis- rival) or our focus on repeat evaluation of SNOM pa- charged without an operation. The proportions of ante- tients, which allowed early identification of evolving symp- rior and posterior AGSW were similar to those in our ex- toms. We accept that the concept still needs to be tested perience. in smaller centers. We agree that the success of SNOM Clinical examination was and still is the main method relies on a multitude of factors, including the appropri- for repeat evaluation of patients managed by SNOM. The ate resources and group commitment. However, we find addition of CT scan has expedited the diagnosis of asymp- little rationale in the vague arguments about lack of ex- tomatic organ injury and allowed early discharge in the perience or expertise as reasons for avoiding SNOM. The absence of a suspicious bullet trajectory.12,27,28 On occa- assessment of abdominal pain, hemodynamic stability, sion, it has contributed to less desired outcomes, such and CT findings is a skill familiar to every surgeon and as a nontherapeutic laparotomy based on trajectory prox- should not be considered the privilege of only a few. If a

ARCH SURG/ VOL 146 (NO. 5), MAY 2011 WWW.ARCHSURG.COM 531

©2011 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 surgeon can assess a patient with a stab wound to the ab- 8. Demetriades D, Velmahos G, Cornwell E III, et al. Selective nonoperative man- domen, the surgeon can assess a patient with AGSW. This agement of gunshot wounds of the anterior abdomen. Arch Surg. 1997;132 (2):178-183. claim does not argue against regionalization of trauma 9. Navsaria PH, Nicol AJ. Selective nonoperative management of kidney gunshot care. Complex trauma patients should be managed in spe- injuries. World J Surg. 2009;33(3):553-557. cialized centers by specialized teams. By the same to- 10. Navsaria PH, Nicol AJ, Krige JE, Edu S. Selective nonoperative management of ken, not every AGSW patient needs an urban trauma liver gunshot injuries. Ann Surg. 2009;249(4):653-656. megacenter. Given the commitment of human and other 11. Demetriades D, Hadjizacharia P, Constantinou C, et al. Selective nonoperative management of penetrating abdominal solid organ injuries. Ann Surg. 2006; resources implicit in the definition of a trauma center, 244(4):620-628. we believe that SNOM for AGSW is safe and effective in 12. Velmahos GC, Constantinou C, Tillou A, Brown CV, Salim A, Demetriades D. Abdomi- most of them, and as such it should be practiced. nal computed tomographic scan for patients with gunshot wounds to the abdomen selected for nonoperative management. J Trauma. 2005;59(5):1155-1161. Accepted for Publication: January 31, 2011. 13. Pryor JP, Reilly PM, Dabrowski GP, Grossman MD, Schwab CW. Nonoperative Correspondence: Karim Fikry, MD, Division of Trauma, management of abdominal gunshot wounds. Ann Emerg Med. 2004;43(3): 344-353. Emergency Surgery, and Surgical Critical Care, Massa- 14. Centers for Disease Control and Prevention. Surveillance for fatal and nonfatal chusetts General Hospital, 165 Cambridge St, Ste 810, firearm-related injuries—United States 1993-1998. http://www.cdc.gov/mmwr Boston, MA 02114 ([email protected]). /preview/mmwrhtml/ss5002a1.htm#top. Published April 13, 2001; reviewed May Previous Presentation: This study was presented at the 2, 2001; accessed September 16, 2010. 91st Annual Meeting of the New England Surgical Soci- 15. Centers for Disease Control and Prevention. Methods of suicide among persons aged 10-19 years—United States, 1992-2001. MMWR Morb Mortal Wkly Rep. ety; October 31, 2010; Saratoga Springs, New York. 2004;53(22):471-474. Author Contributions: Study concept and design: Fikry, 16. Cheng TL, Wright JL, Fields CB, et al. Violent injuries among adolescents: de- Velmahos, Bramos, Janjua, de Moya, King, and Alam. Ac- clining morbidity and mortality in an urban population. Ann Emerg Med. 2001; quisition of data: Fikry and Bramos. Analysis and inter- 37(3):292-300. pretation of data: Fikry and Velmahos. Drafting of the manu- 17. Como JJ, Bokhari F, Chiu WC, et al. Practice management guidelines for selec- script: Fikry, Velmahos, Bramos, and Janjua. Critical tive nonoperative management of penetrating abdominal trauma. J Trauma. 2010; 68(3):721-733. revision of the manuscript for important intellectual con- 18. Salim A, Velmahos GC. When to operate on abdominal gunshot wounds. Scand tent: Fikry, de Moya, King, and Alam. Statistical analy- J Surg. 2002;91(1):62-66. sis: Fikry and Bramos. Administrative, technical, and ma- 19. Nance FC, Cohn I Jr. Surgical judgment in the management of stab wounds of terial support: Fikry. Study supervision: Velmahos, de Moya, the abdomen: a retrospective and prospective analysis based on a study of 600 King, and Alam. stabbed patients. Ann Surg. 1969;170(4):569-580. 20. Demetriades D, Rabinowitz B. Selective conservative management of penetrat- Financial Disclosure: None reported. ing abdominal wounds: a prospective study. Br J Surg. 1984;71(2):92-94. 21. Demetriades D, Rabinowitz B. Indications for operation in abdominal stab wounds: REFERENCES a prospective study of 651 patients. Ann Surg. 1987;205(2):129-132. 22. Demetriades D, Rabinowitz B, Sofianos C, et al. The management of penetrating injuries of the back: a prospective study of 230 patients. Ann Surg. 1988;207 1. Shaftan GW. Selective conservatism in penetrating abdominal trauma. J Trauma. (1):72-74. 1969;9(12):1026-1028. 23. Saadia R, Degiannis E. Non-operative treatment of abdominal gunshot injuries. 2. McAlvanah MJ, Shaftan GW. Selective conservatism in penetrating abdominal wounds: a continuing reappraisal. J Trauma. 1978;18(3):206-212. Br J Surg. 2000;87(4):393-397. 3. Demetriades D, Charalambides D, Lakhoo M, Pantanowitz D. Gunshot wound 24. Ross SE, Dragon GM, O’Malley KF, Rehm CG. Morbidity of negative coeliotomy of the abdomen: role of selective conservative management. Br J Surg. 1991; in trauma. Injury. 1995;26(6):393-394. 78(2):220-222. 25. Renz BM, Feliciano DV. Unnecessary laparotomies for trauma: a prospective study 4. Velmahos GC, Demetriades D, Toutouzas KG, et al. Selective nonoperative man- of morbidity. J Trauma. 1995;38(3):350-356. agement in 1,856 patients with abdominal gunshot wounds: should routine lapa- 26. Patton JH Jr, Woodward AM. Urban trauma centers: not quite dead yet. Am Surg. rotomy still be the standard of care? Ann Surg. 2001;234(3):395-403. 2002;68(4):319-323. 5. Velmahos GC, Demetriades D, Cornwell EE III. Transpelvic gunshot wounds: rou- 27. Schwab CW. Selection of nonoperative management candidates. World J Surg. tine laparotomy or selective management? World J Surg. 1998;22(10):1034- 2001;25(11):1389-1392. 1038. 28. Ginzburg E, Carrillo EH, Kopelman T, et al. The role of computed tomography in 6. Velmahos GC, Demetriades D, Foianini E, et al. A selective approach to the man- selective management of gunshot wounds to the abdomen and flank. J Trauma. agement of gunshot wounds to the back. Am J Surg. 1997;174(3):342-346. 1998;45(6):1005-1009. 7. Muckart DJ, Abdool-Carrim ATO, King B. Selective conservative management of 29. Velmahos GC, Demetriades D, Cornwell EE, Asensio J, Belzberg H, Berne TV. abdominal gunshot wounds: a prospective study. Br J Surg. 1990;77(6):652- Gunshot wounds to the buttocks: predicting the need for operation. Dis Colon 655. Rectum. 1997;40(3):307-311.

ARCH SURG/ VOL 146 (NO. 5), MAY 2011 WWW.ARCHSURG.COM 532