PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Int J Surg. Author manuscript; available in PMC 2017 April 1.
Published in final edited form as:
Published online 2016 February 13. doi:  10.1016/j.ijsu.2016.02.042
PMCID: PMC4826292
NIHMSID: NIHMS766264

EXPERIENCE OF DAMAGE CONTROL TRAUMA LAPAROTOMY IN A LIMITED RESOURCE HEALTHCARE SETTING

Abstract

Introduction

Damage control surgery (DCS) is an established option for managing severely injured trauma patients. However, its role in the management of similar patients in the developing world is debatable. The purpose of this study is to describe characteristics and outcomes of patients undergoing DCS.

Methods

All trauma patients requiring laparotomies from 1996–2011 at a tertiary care hospital in South Asia were reviewed. DCS was defined in a patient who underwent a truncated laparotomy where the fascia was primarily left open, with the intention of physiological optimization in the Intensive Care Unit, followed by definitive surgery. The primary outcome was in-hospital mortality. Multivariate logistic regression was used to determine the independent predictors of mortality after adjustment for potential confounders.

Results

Of 258 patients, 47 underwent DCS. 40% patients were transferred from other hospitals. The time between injury and operation was 152 minutes (IQR: 90–330). Intra-operative laboratory parameters revealed a median pH of 7.16 (IQR: 7.10–7.27), median temperature of 34.7 (IQR: 34.0–35.4) and median PT of 15.9 (IQR: 12.4–21.2). 55% of the patients survived to discharge from hospital. Of those who died, 86% died before the first take back operation. Packed red blood cell transfusion and vascular injury were independently associated with mortality.

Discussion

Damage control surgery is feasible in developing countries, with more than 50% survival reported at one hospital. Future research should focus on critical care management.

Conclusion

Damage Control trauma laparotomy is feasible in tertiary care hospitals with multidisciplinary trauma teams in lesser-developed countries.

Keywords: Damage Control Surgery, trauma, mortality

1. Introduction

Damage control surgery (DCS) is a widely accepted surgical approach to manage trauma patients in extremis, as a temporising measure to salvage otherwise non-survivable patients [1]. Credited with improving survival in up to 50% of the patients [1, 2], DCS involves pausing an operation to achieve physiologic optimisation after control of surgical bleeding in an exsanguinating patient when ongoing blood loss is from hypothermia, acidosis and coagulopathy. An attempt to reverse the lethal triad of death in the critically ill patient is made in the intensive care unit, and the patient is then taken back for definitive and reconstructive operation [3].

DCS relies on optimal resources for pre-hospital in-transit care, intra-operative treatment, interval stabilization and post-operative care. Most studies in literature describe outcomes of DCS in resource-rich healthcare settings [2, 4]. However, there is paucity of literature describing the application and outcomes of damage control surgery in resource poor countries. The spectrum of injury and outcomes of trauma care in resource-poor settings are affected by several barriers including limited en-route resuscitation, lack of efficient transport, limited resuscitation resources in the trauma bay, increased time to assemble operating teams and limited resources for intensive care before the take-back operation [57]. Hence, the presentation and operative care of patients treated with DCS in these settings are expected to be different from experiences reported in resource-rich environments. The primary objective of this study is to describe demographic, injury characteristics and in-hospital outcomes of patients undergoing damage control surgery and their outcomes at a tertiary care hospital in Pakistan.

2. Materials and Methods

2.1 Study Setting

The Aga Khan University Hospital (AKUH) is a tertiary care hospital located in the metropolitan city of Karachi, Pakistan. It serves an estimated population of 15.5 million people in Karachi as well as northern Pakistan and Eastern Afghanistan [810]. Patients managed at AKUH are considered representative of the patient population throughout the region [810]. The Aga Khan University Hospital serves as a primary and referral trauma center, equipped with trauma resuscitation room, diagnostic radiology, dedicated emergency operating room, and a multidisciplinary trauma team [11]. This hospital has a total capacity of 577 beds distributed over 15 inpatient units. Approximately 10% (55/577) of the beds are critical care. AKUH is the largest privately owned hospital in Pakistan with a pay-for-service model of health care delivery [12].

Currently, there are 63 full-time and 23 part-time faculty members in the department of surgery. Surgical expertise of five surgeons was employed in this study. All surgeons have completed certified post-graduate training in general surgery. In 2002, trauma quality improvement (TQI) initiatives were implemented. Important TQI initiatives in this context included: 1) Establishing a multidisciplinary trauma team, 2) Training of faculty and residents using adapted American College of Surgeons (ACS) Advanced Trauma Life Support protocols, 3) Establishing an Emergency Room trauma rush call generation protocol, 24-hour availability of dedicated trauma resuscitation, operating rooms and computed tomography scanning, 4) Collaborating with Interventional Radiology to perform angioembolization, 5) Availability of modern intraoperative surgical devices, 6) Collaborating with Hematology to setup a massive transfusion protocol, 7) Improving post-operative care with dedicated trauma care nurses [13, 14].

2.2 Data collection and statistical analysis

A retrospective review of medical records of all trauma patients admitted to AKUH who underwent a laparotomy from January 1, 1996 to July 31, 2011 was conducted. The inclusion criteria were all patients who underwent DCS at AKU. In this study, DCS was defined as a patient who was in extremis and undergoing a truncated laparotomy (DCS part 1) followed by a plan for physiological optimization in the ICU (DCS part 2) and further definitive management at a later, second operation (DCS part 3). Different techniques (towel clips, Op-Site and home-made vacuum dressing) were used for closure at the time of the initial operation for temporary abdominal wall closure. The medical records of these selected patients were reviewed by a trauma fellow or senior surgical resident.

Information was recorded on a standardized questionnaire. Data was collected on: demographic details (age, gender), transfer status, injury characteristics (type of injury, Injury Severity Score (ISS), Revised Trauma Score (RTS), vitals on arrival to the ED, Glasgow coma scale score, associated injuries), laboratory parameters at admission (hemoglobin, PT, pH, core temperature), transfusion details, DCS and definitive surgery characteristics (vitals, hemoglobin, PT, pH, core temperature, blood loss, transfusions, bowel, solid viscera and vascular injury and procedures), ICU management (fluids, transfusions, inotropic support, ventilator parameters), short-term post-operative complications and in-hospital mortality (Figure 1). The Injury Severity Score is an anatomical scoring system that provides an overall score for patients with multiple injuries. Each injury is assigned an Abbreviated Injury Scale (AIS) score. AIS divides the body into six regions (head and neck, face, chest, abdomen, pelvis, and extremities and general) and classifies the severity of injuries in each region based on clinical experience (1 = minor; 2 = moderate; 3 = severe, not life-threatening; 4 = severe, life threatening, survival probable; 5 = critical, survival uncertain; 6 = fatal). The ISS score is defined as the sum of the squares of the single highest AIS scores in each of the three most severely injured body regions (score 0–75) [15]. Injury-ED time and ED-OR times were also calculated based on time of injury, time of arrival to ED and OR.

Figure 1
Cumulative distribution of patients undergoing DCS from 1996–2011. Solid line represents moving average.

The primary outcome of this study was mortality. Student’s t test was used to compare continuous variables. If the assumptions of this test were not met (Shapiro-Wilk normality test), the Mann-Whitney U test was used. Fisher's exact analysis was used to compare categorical variables (Expected value in cell less than five). Multivariate logistic regression was used to determine the independent predictors of mortality and their effect sizes after adjustment for potential confounders. Only variables which were found to be significant on univariate analysis (p<0.05) were included in the final model. The final mortality model included the following variables: hypotension on arrival to ED, Injury Severity Score, number of pRBC transfusions and presence of vascular injury. Temperature on arrival to ED was not included in the final model due to the high missing rate (24/47 patients). All analysis was done in Stata/MP version 11 (Stata, College Station, TX), and statistical significance was defined as a p-value of less than 0.05.

This study was exempt from the Aga Khan University ethical committee review.

3. Results

There were 258 patients who underwent a trauma laparotomy from 1996–2011 at AKU. Of these, 47 patients met the criteria for DCS. Figure 1 describes the cumulative distribution of patients undergoing DCS over 14 years.

3.1 Overall group

Demographics

The mean age of the patients was 33 years (standard deviation (SD): 15 years) (Table 1). Approximately 91% of the patients were male. 40% of the patients were transferred, mostly from other tertiary care hospitals. Majority of the patients were brought by their relatives in an ambulance. The median injury-ED time was 60 minutes (interquartile range (IQR): 31–120). The median injury-OR time was 152 minutes (IQR: 90–330)

Table 1
Socio demographic and Injury characteristics

Physiologic and Laboratory characteristics

Approximately 32% of the patients had tachycardia (heart rate > 100 beats per minute) and 83% had hypotension (systolic blood pressure < 90 mmHg) (Table 2). The mean ISS was 24.7 (SD: 6.3). The mean RTS was 7.21 (SD: 0.85). Majority of the patients had a GCS score of 13–15. 68% of the patients had a penetrating injury (30% blunt and 2% missing), with 34% having multiple torso penetrations. 11% had injuries severe enough to cause exsanguinations from more than one anatomic site. The most common associated injury was thoracic injury (45%). For patients with data recorded on pH, 79% had acidosis (11/13), with a median pH of 7.31 (IQR: 7.25–7.34). For patients with data recorded on temperature, 14% (3/23) had hypothermia (temperature < 35 ‘C, with a median temperature of 36.6 (IQR: 35.7–37.0). 19% of the patients had a raised PT (≥ 15.00).

Table 2
Pre-operative and Intra-operative parameters

Operative characteristics

The median ER to OR was 53 minutes (IQR: 25–120) (Table 2). The lowest recorded value for pH and temperature during the operating room was analyzed. The median pH, temperature and PT were 7.16 (IQR: 7.10–7.27), 34.7 (IQR: 34.0–35.4) and 15.9 (IQR: 12.4–21.2) respectively. The median duration of surgery was 105 minutes (IQR: 75–147), with a median number of 6 packed red blood cells (pRBCs) transfusions during surgery.

Injury and Type of Procedure characteristics

The major injuries encountered are summarized in Table 2. The most common bowel injury was small bowel and colon (28% each). Liver laceration (43%) was the most common solid viscera injury and portal vein injury (13%) was the most common major vascular injury. Techniques for temporary abdominal wall closure at the time of initial operation included: towel clips (49%), Op-site (26%), vacuum dressing (11%). Data on closure was missing for 15% of the patients.

Mortality and Length of stay

The overall mortality rate was 45% in this patient population (Table 3). Of these, 86% of the patients died before or in the ICU before the first take back operation (29% in the OR and 57% in the ICU). Table 4 demonstrates the results of the associations of patient, injury and operative characteristics and mortality. Hypotension and hypothermia on arrival to ED, higher injury severity score, increasing requirement of pRBC’s transfusions and presence of major vascular injury were significantly associated with mortality (p<0.05). On multivariable analysis, only increasing number of pRBC’s transfusions (OR, 1.76; CI, 1.14–2.71), and vascular injury (OR, 18.19; CI, 1.94–170.6) were independent predictors of mortality (Table 5). The median length of stay in patients who survived to discharge was 16 days (IQR: 12–29).

Table 3
Outcomes
Table 4
Univariate analysis: Associations of mortality
Table 5
Adjusted Odds ratios for predictors of mortality (n=44)

3.2 Transferred group

The median injury-ER time was significantly higher in patients that were transferred from other hospitals as compared to non-transferred patients (112 (IQR: 52–169) versus 34 (IQR: 24–110) minutes). In addition, transferred patients had lower pH (7.29 (IQR: 7.10–7.36) versus 7.31 (IQR: 7.27–7.36) and higher PT (14.3 (IQR: 12.9–17.5) versus 13.5 (IQR: 12.2–20.6) on presentation to the ED and in the OR.

4. Discussion

This study reviewed all patients undergoing damage control surgery at a tertiary care hospital in Pakistan. The overall survival rate of these patients was 55%. This was despite the fact that 40% of the patients were transferred from other facilities. In patients who incurred in-hospital death, majority of the patients died in the second resuscitating phase of damage control surgery. We identified five factors associated with mortality in patients undergoing damage control surgery on univariate analysis: hypotension and hypothermia on arrival to ED, higher injury severity score, increasing number of pRBC’s transfusions and the presence of major vascular injury. On multivariate analysis, packed red blood cell transfusion and vascular injury were independent predictors of mortality.

Damage control surgery (DCS) is an increasingly used tool in trauma surgery since its inception in 1993 by Rotondo [3]. A successful outcome in earlier and subsequent studies is attributed to the interruption of the lethal cascade of physiological derangements [1618]. A comparison of our patients with previous studies of DCS from Rotondo et al and Johnson et al is shown in Table 6 [3, 19]. Our patients were similar with regards to age, gender and Injury Severity Score. However, the physiologic insult to our patients was much greater manifesting as markedly lower systolic blood pressures and acidosis. The poor physiology in the operating room is reflective of delays in definitive treatment. For optimal results in resource-rich environments, patients are moved to trauma care centers rapidly via ambulance or air transport where resuscitative therapy is initiated and or continued. However, in lesser-developed countries, this aspect of care is most neglected. Minimal or no medical aid is available to patients at the injury scene. Patients are mostly transported by relatives or by-standers using public transport or personal vehicles, and in many cases are transferred from other hospitals [20]. Because pre-hospital care does not exist, patients arrive without established IV access or interventional airways [7]. Further, hospitals are usually not alerted as to when trauma patients may arrive, how many patients are involved, or about the severity of the injuries. Our patients on average reach the Operating Room 2.5 hours after their injury, and this is much higher for patients with multiple hospital transfers. Thus, most patients do not receive definitive care in the “golden hour”.

Table 6
Comparison of demographic, physiologic characteristics and outcomes after DCS

In this study, majority of the patients (57%) died in the Intensive Care Unit before the first take back operation (DCS part 3). The intensive care unit is the optimal location for resuscitation for the critically ill patient [21, 22]. Damage Control Part II is resource intensive, and requires continued bedside surgical presence during the first critical hours and a minimum ratio of 2 nurses to 1 patient [23]. There are several roadblocks for ventilation, resuscitation, rewarming, coagulopathy and acidosis correction in our ICU. There is no standardized ventilator strategy; rotokinetic therapy and oscillating ventilators are not available. There is a shortage of pRBCs, FFPs, fibrinogen, cryoprecipitate. We also do not have the financial liberty to repeat laboratory markers such as pH, lactate, base deficit (endpoints of resuscitation) as frequently as indicated [24]. There is shortage/unavailability of dedicated fluid warming devices. Continuous arteriovenous rewarming [25], a technique used in persistently hypothermic patients was not employed in any of our patients. Thus, either these patients continued to bleed despite undergoing a truncated laparotomy or were not sufficiently optimized in the ICU for the take back operation. As a result, majority of the patients died soon after the first surgery in the OR or in the ICU.

These early deaths suggest that in most patients we were unable to reverse the lethal triad of acidosis, hypothermia and coagulopathy. There are a number of factors, which may be responsible for this. Firstly, the opportunity to perform damage control surgery may be too late. Secondly, we may be lacking collective experience, as there are no evidence-based guidelines on the practice of damage control surgery in Pakistan. Thirdly, early use of blood products in the resuscitative phase could not be implemented due to delayed presentation.

Fourthly, there is lack of availability of blood products. Furthermore, we did not have a massive transfusion protocol for the majority of the study period. A protocol for massive blood transfusion, a component of damage control resuscitation has been recently introduced at our institute as part of the trauma quality improvement initiatives [26]. Finally, costly pharmacological adjuncts such as recombinant factor VIIa [27, 28] were not available. The role of angio-embolisation in arresting ongoing hemorrhage in trauma patients is relatively new [29]. None of our patients underwent angio-embolisation as part of initial trauma resuscitation/damage contol surgery. There are all areas of improvement. In 2002, trauma quality improvement (TQI) initiatives were implemented at AKUH. The objective of this program was to reduce in-hospital morbidity and mortality in trauma patients [13, 14]. Despite a comparatively high mortality rate in this study, damage control surgery is an acceptable technique in a subset of severely injured trauma patients in developing countries who may not survive otherwise.

This study has several limitations. This study represents a small group of patients from a single institution, and thus may not be generalized to rest of the country or region. As data was retrospectively collected, we were concerned that we could potentially include other patients who had undergone a laparatomy whereby primary closure had been delayed. These patients do meet the criterion for DCS part I. For example, a surgeon may have delayed closure in a patient with bowel edema at the time of initial operation. To eliminate this error, data collection was done by our trauma fellow or senior surgical resident. And finally, there was a high missing rate on variables capturing pre-hospital and ED characteristics and we could not include these factors in our final multivariate analysis.

In conclusion, damage control surgery is an effective way of managing severely injured patients in lesser-developed countries. Future research should focus on critical care management.

Highlights

  • Damage control surgery is an effective way of managing severely injured patients in lesser developed countries
  • 55% of the patients survived to discharge
  • Majority of the patients died in the second resuscitating phase
  • Increasing number of transfusions and vascular injury are independent predictors of mortality

Acknowledgments

Financial support for this work was provided by: National Institutes of Health/NIGMS K23GM093112-01 and American College of Surgeons C. James Carrico Fellowship for the study of Trauma and Critical Care (Dr. Haider and Dr. Kisat)

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Authors’ contribution Statement:

Mehreen Kisat: literature search, study design, data analysis, data interpretation, figures, manuscript writing

Syed Nabeel Zafar: Data analysis, data interpretation, manuscript review

Zain G. Hashmi: Data analysis, manuscript writing and review

Amyn Pardhan: Data collection, manuscript writing and review

Tahreem Mir: Data Entry, manuscript writing and review

Adil Shah: Data Entry, manuscript writing and review

Adil H. Haider: Study design, data interpretation, manuscript revision and review

Hasnain Zafar : Study design, data interpretation, manuscript revision and review

References

1. Lee JC, Peitzman AB. Damage-control laparotomy. Current opinion in critical care. 2006;12(4):346–350. [PubMed]
2. Rotondo MF, Zonies DH. The damage control sequence and underlying logic. The Surgical clinics of North America. 1997;77(4):761–777. [PubMed]
3. Rotondo MF, Schwab CW, McGonigal MD, Phillips GR, 3rd, Fruchterman TM, Kauder DR, Latenser BA, Angood PA. 'Damage control': an approach for improved survival in exsanguinating penetrating abdominal injury. The Journal of trauma. 1993;35(3):375–382. discussion 382-373. [PubMed]
4. Shapiro MB, Jenkins DH, Schwab CW, Rotondo MF. Damage control: collective review. The Journal of trauma. 2000;49(5):969–978. [PubMed]
5. Jamali AR. Trauma care in Pakistan. JPMA The Journal of the Pakistan Medical Association. 2008;58(3):102–103. [PubMed]
6. Jat AA, Khan MR, Zafar H, Raja AJ, Hoda Q, Rehmani R, Lakdawala RH, Bashir S. Peer review audit of trauma deaths in a developing country. Asian journal of surgery / Asian Surgical Association. 2004;27(1):58–64. [PubMed]
7. Kobusingye OC, Hyder AA, Bishai D, Hicks ER, Mock C, Joshipura M. Emergency medical systems in low- and middle-income countries: recommendations for action. Bulletin of the World Health Organization. 2005;83(8):626–631. [PubMed]
8. Shah AA, Rehman A, Haider AH, Sayani R, Sayyed RH, Ali K, Zafar SN, Rehman ZU, Zafar H. Angiographic embolization for major trauma in a low-middle income healthcare setting--A retrospective review. International journal of surgery. 2015;18:34–40. [PubMed]
9. Shah AA, Rehman A, Sayyed RH, Haider AH, Bawa A, Zafar SN, Zia Ur R, Ali K, Zafar H. Impact of a predefined hospital mass casualty response plan in a limited resource setting with no pre-hospital care system. Injury. 2015;46(1):156–161. [PubMed]
10. Shah AA, Rehman A, Shah SJ, Haider AH, Zogg CK, Zafar SN, Hadi YB, Zia Ur R, Zafar H. Abdominal gunshot wounds-a comparative assessment of severity measures. The Journal of surgical research. 2015;198(2):334–339. [PubMed]
11. Zafar H, Rehmani R, Raja AJ, Ali A, Ahmed M. Registry based trauma outcome: perspective of a developing country. Emergency medicine journal : EMJ. 2002;19(5):391–394. [PMC free article] [PubMed]
12. Shah AA, Latif A, Zogg CK, Zafar SN, Riviello R, Halim MS, Rehman Z, Haider AH, Zafar H. Emergency general surgery in a low-middle income health care setting: Determinants of outcomes. Surgery. 2016;159(2):641–649. [PubMed]
13. Noordin S, Allana S, Ahmad T, Bhatti AU, Zafar H, Wajid MA. Evolution of trauma management at a tertiary care hospital: a cohort study. International journal of surgery. 2011;9(1):75–78. [PubMed]
14. Hashmi ZG, Haider AH, Zafar SN, Kisat M, Moosa A, Siddiqui F, Pardhan A, Latif A, Zafar H. Hospital-based trauma quality improvement initiatives: first step toward improving trauma outcomes in the developing world. The journal of trauma and acute care surgery. 2013;75(1):60–68. discussion 68. [PubMed]
15. Baker SP, O'Neill B, Haddon W, Jr, Long WB. The injury severity score: a method for describing patients with multiple injuries and evaluating emergency care. The Journal of trauma. 1974;14(3):187–196. [PubMed]
16. Feliciano DV, Mattox KL, Jordan GL., Jr Intra-abdominal packing for control of hepatic hemorrhage: a reappraisal. The Journal of trauma. 1981;21(4):285–290. [PubMed]
17. Lucas CE, Ledgerwood AM. Prospective evaluation of hemostatic techniques for liver injuries. The Journal of trauma. 1976;16(6):442–451. [PubMed]
18. Walt AJ. Founder's lecture: The mythology of hepatic trauma--or Babel revisited. American journal of surgery. 1978;135(1):12–18. [PubMed]
19. Johnson JW, Gracias VH, Schwab CW, Reilly PM, Kauder DR, Shapiro MB, Dabrowski GP, Rotondo MF. Evolution in damage control for exsanguinating penetrating abdominal injury. The Journal of trauma. 2001;51(2):261–269. discussion 269–271. [PubMed]
20. Mock CN, Adzotor KE, Conklin E, Denno DM, Jurkovich GJ. Trauma outcomes in the rural developing world: comparison with an urban level I trauma center. The Journal of trauma. 1993;35(4):518–523. [PubMed]
21. Parr MJ, Alabdi T. Damage control surgery and intensive care. Injury. 2004;35(7):713–722. [PubMed]
22. Moore EE, Burch JM, Franciose RJ, Offner PJ, Biffl WL. Staged physiologic restoration and damage control surgery. World journal of surgery. 1998;22(12):1184–1190. discussion 1190-1181. [PubMed]
23. Sagraves SG, Toschlog EA, Rotondo MF. Damage control surgery--the intensivist's role. Journal of intensive care medicine. 2006;21(1):5–16. [PubMed]
24. Abramson D, Scalea TM, Hitchcock R, Trooskin SZ, Henry SM, Greenspan J. Lactate clearance and survival following injury. The Journal of trauma. 1993;35(4):584–588. discussion 588–589. [PubMed]
25. Gentilello LM, Rifley WJ. Continuous arteriovenous rewarming: report of a new technique for treating hypothermia. The Journal of trauma. 1991;31(8):1151–1154. [PubMed]
26. Rehman A, Shah AA, Sayyed RH, Tareen A, Rehman ZU, Moiz B, Shah SJ, Zafar H. Massive Transfusion for Trauma in a Lower Middle Income Country. The American surgeon. 2015;81(11):395–397. [PubMed]
27. Bickell WH, Wall MJ, Jr, Pepe PE, Martin RR, Ginger VF, Allen MK, Mattox KL. Immediate versus delayed fluid resuscitation for hypotensive patients with penetrating torso injuries. N Engl J Med. 1994;331(17):1105–1109. [PubMed]
28. Jansen JO, Thomas R, Loudon MA, Brooks A. Damage control resuscitation for patients with major trauma. BMJ. 2009;338:b1778. [PubMed]
29. Teo LT, Punamiya S, Chai CY, Go KT, Yeo YT, Wong D, Appasamy V, Chiu MT. Emergency angio-embolisation in the operating theatre for trauma patients using the C-Arm digital subtraction angiography. Injury. 2011 [PubMed]