a b s t r a c t

Introduction: Massive hemorrhage is often associated with unstable pelvic fractures with posterior ring injury. Initial pelvic radiography alone may not detect these posterior lesions. We examined whether the presence of an anterior pelvic fracture on initial pelvic radiography alone may identify patients who are at a high risk of major hemorrhage.

Materials and methods: A total of 288 patients with pelvic fractures were admitted to the Fukui Prefectural Hos- pital during an 11-year period. After excluding 33 patients who were in cardiopulmonary arrest on arrival and nine with concomitant abdominal Organ injuries requiring emergency laparotomy, 246 eligible patients were ret- rospectively reviewed. Anterior pelvic fractures were defined as displacement of the obturator ring, obturator ring with laterality, or displacement of the pubic symphysis on pelvic radiography.

Results: Massive hemorrhage was identified in 106 of 246 patients. Patients with massive hemorrhage had a higher frequency of anterior pelvic fractures on pelvic radiography and higher frequency of posterior pelvic frac- tures on computed tomography than those without massive hemorrhage. Logistic regression analysis identified displacement of the obturator ring by >=5 mm, obturator ring with laterality of >=5 mm, and displacement of the pubic symphysis by >=4 mm on pelvic radiography as predictors of massive pelvic hemorrhage.

Conclusion: The results of the present study suggested that the presence of displaced anterior lesions of the pelvic ring on pelvic radiography alone, without the use of computed tomography during the initial treatment stage, may promptly identify patients at high risk of massive pelvic hemorrhage who require intervention for hemor- rhage control.

(C) 2018

Introduction

The management of hemodynamically unstable patients with pelvic fractures is a multidisciplinary challenge with high mortality [1-4]. Hemorrhage due to disruption of the surrounding venous and arterial vessels is the leading cause of death in hemodynamically unstable pa- tients and requires prompt therapy for hemostasis [1,2]. Numerous pre- vious investigations have reported the predictors of pelvic massive hemorrhage associated with pelvic fractures. Some clinical prediction factors for identifying subjects at high risk of pelvic massive hemorrhage are based on the classification of Young-Burgess or the presence of arte- rial extravasation or massive hematoma on contrast-enhanced pelvic computed tomography (CT) [5-7]. Blackmore et al. reported that initial

? There are no potential conflicts of interest.

?? This paper has not been presented anywhere.

* Corresponding author at: The Department of Emergency Medicine, Fukui Prefectural Hospital, 2-8-1, Yotsui, Fukui 910-8526, Japan.

E-mail address: [email protected] (S. Tanizaki).

pelvic radiography findings, pulse rate, and Hematocrit levels may be used for predicting the risk of pelvic Arterial hemorrhage associated with pelvic fractures [8]. Cryer et al. performed a retrospective study of 245 patients with pelvic fractures using their modification of the Pennel and Sutherland classification and reported that a displacement or gap of >=0.5 cm at any fracture site within pelvic ring (anterior, poste- rior, or acetabulum) was correlated with increased requirement for blood transfusion [9]. However, no rapid and reliable noninvasive method for determining the predictors of pelvic massive hemorrhage on initial pelvic radiography alone has been reported.

Massive hemorrhage is often associated with unstable pelvic frac- tures with posterior ring injury. However, 68%-96.8% of patients with anterior pelvic fractures on pelvic radiography showed a posterior ring injury on CT. Previous studies have concluded that radiography alone cannot detect posterior ring injury and pelvic instability [10,11].

The purpose of this retrospective study was to analyze the impact of anterior pelvic fracture on initial pelvic radiography in the management of massive hemorrhage due to pelvic trauma. Our hypothesis was that the presence of anterior lesions of the pelvic ring on pelvic radiography

https://doi.org/10.1016/j.ajem.2018.03.058

0735-6757/(C) 2018

alone would identify patients at high risk of massive pelvic hemorrhage, without the use of CT, at the initial treatment stage.

Methods

The institutional review board of Fukui Prefectural Hospital ap- proved this retrospective study, and the need for patient consent was waived because of the retrospective study design. We conducted a re- view of a consecutive series of 288 trauma patients with pelvic ring frac- tures who were admitted to our institution between January 1, 2006 and December 31, 2016.

The patient data were obtained from the hospital records. Patients with any type of pelvic fracture were included. After excluding 33 pa- tients with cardiopulmonary arrest on arrival and nine with concomi- tant abdominal organ injuries requiring emergency laparotomy, 246 eligible patients were retrospectively reviewed. Patient data included age, gender, systolic blood pressure (SBP) and heart rate (HR) on arrival, injury severity score (ISS), hemodynamic instability, mortality, require- ment for the Packed red blood cells (PRBCs) in the initial 24 h, length of stay in the intensive care unit (LOS-ICU), anterior pelvic fracture on ini- tial pelvic radiograph, posterior pelvic ring fracture on pelvic CT, arterial extravasation or on contrast-enhanced pelvic CT, and positive angiogra- phy findings. Hemodynamic instability was defined as SBP of b90 mm Hg after initiation of PRBCs transfusion. Massive hemorrhage was defined as hemodynamic instability, arterial extravasation on con- trast-enhanced pelvic CT, requirement of N10 units of PRBCs in the ini- tial 24 h, or positive angiography findings.

Anteroposterior pelvic radiography was used to diagnose pelvic frac- tures. If hemodynamic stability permitted radiographic workup, whole body CT was performed. Patients with arterial extravasation on con- trast-enhanced pelvis CT underwent angiography and embolization. If patients had ongoing hemodynamic instability after the initiation of transfusion of PRBCs, resuscitative Balloon occlusion of the aorta was performed. If patients were hemodynamically unstable with only pelvic hemorrhage, angiography and embolization were immediately per- formed. Laparotomy was indicated for concomitant hemodynamic in- stability and progressive Abdominal effusion or existence of a pneumoperitoneum. Preperitoneal packing was not performed during the current study period. Noninvasive external fixation devices were used because of their rapid placement ability compared with that of in- vasive external fixation devices.

Positive angiography was defined as signs of Vascular injury, such as contrast extravasation, vasospasm, pseudoaneurysm, and arteriovenous fistula, which required embolization. Anterior pelvic fractures on pelvic radiography were defined as follows: Displacement of the obturator ring was defined as the maximal distance along the axis of the displace- ment of superior or inferior ramus fractures (Fig. 1). Laterality of the ob- turator ring was defined as the maximal difference between the left and right perpendicular distance from the inferior line of the superior ramus to the superior line of the inferior ramus (Fig. 2). Displacement of the pubic symphysis was defined as the transverse distance between the left and right superior rami of the pubis (Fig. 3).

Categorical variables were reported as percentages and continuous variables as mean +- SD unless otherwise specified. Univariate analysis of variables was performed using the Chi-square test for categorical var- iables and Student’s t-test for continuous variables. Significance was de- fined as p b 0.05. Multiple regression analysis was performed to predict massive hemorrhage based on the displacement of the obturator ring, obturator ring with laterality, or displacement of the pubic symphysis on pelvic radiography. The association between massive pelvic hemor- rhage and anterior pelvic fractures was examined using multivariate lo- gistic regression analysis. Some variables with p b 0.05 by univariate analysis, such as SBP, HR, ISS, the displacement of the obturator ring, ob- turator ring with laterality, or displacement of the pubic symphysis on pelvic radiography, and posterior ring fractures on CT, were included in the regression model. Other variables, such as hemodynamic

Fig. 1. The solid double arrow indicates the maximal perpendicular distance to the axis of the displaced superior or inferior ramus fractures. The superior or inferior ramus fracture is considered presented if the distance is gapped 5 mm or more.

instability, the requirement for PRBCs, arterial extravasation on CT, and positive angiography, were excluded because massive hemorrhage was defined as their variables. Statistical analyses were performed with R version 3.3.0 (R Foundation for Statistical Computing, Vienna, Austria).

Results

The mean patient age was 62.0 +- 27.6 years. Of 246 eligible patients, 119 were males and 127 were females. Mean SBP and HR on arrival were 112 +- 33.2 mm Hg and 91.2 +- 20.8 beats per min, respectively. The mean ISS was 23.1 +- 14.2. The mean LOS-ICU was 2.29 +-

4.26 days. All injuries were the result of blunt trauma; 93 (37%) patients

Fig. 2. The solid double arrow indicates the maximal difference between the left and right perpendicular distance from the inferior line of the superior ramus to the superior line of the inferior ramus. The dashed double arrow indicates the left and right perpendicular distance from the inferior line of the superior ramus to the superior line of the inferior ramus. The ischium or ramus fracture, or rotational displacement of coxa are considered presented if there was the difference of 5 mm or more.

Fig. 3. The solid double arrow indicated the transverse distance between the left and right superior rami of the pubis.

were pedestrians struck by motor vehicles, 72 (29%) were individuals who fell from a height, and 68 (27%) were individuals who experienced motor vehicle crashes. A total of 29 (11%) patients died, of which eight patients died due to hemorrhage, one due to acute respiratory distress syndrome because of hemorrhagic shock, and three due to Multiple organ dysfunction; 18 succumbed to Severe head injuries.

In total, 106 patients (43%) had massive pelvic hemorrhage. There were significant differences in SBP on arrival, HR on arrival, hemody- namic instability, mortality, ISS, PRBCs requirement in the initial 24 h, and the LOS-ICU between patients with and without massive hemor- rhage (Table 1). Patients with massive hemorrhage had greater dis- placement of the obturator ring or pubic symphysis and greater laterality of the obturator ring than those without massive hemorrhage (Table 1). Furthermore, patients with massive hemorrhage had a higher frequency of posterior pelvic fractures on CT than those without mas- sive hemorrhage. Of the 106 patients with massive hemorrhage, 65 (61%) had arterial extravasation on CT. Angiography was positive in 77 patients (72%) with massive hemorrhage.

receiver operating characteristic curve analysis showed that the op- timal cut-off values for obturator ring with laterality, displacement of the obturator ring, and displacement of the pubic symphysis were 5 mm [area under the curve (AUC), 0.759; 95% confidence interval (CI): 0.697-0.821; sensitivity, 54.7%; specificity, 92.1%], 5 mm (AUC,

0.784; 95% CI, 0.725-0.843; sensitivity, 64.2%; specificity, 86.4%), and

4 mm (AUC, 0.622; 95% CI, 0.551-0.693; sensitivity, 59.0%; specificity, 61.7%), respectively. Fractures were defined as significant when the dis- placement of the obturator ring was >=5 mm, laterality of the obturator ring was >=5 mm, or displacement of the pubic symphysis was >=4 mm on pelvic radiography.

Patients with massive hemorrhage had higher frequencies of obtura- tor ring with laterality of >=5 mm, displacement of the obturator ring by

>=5 mm, and displacement of the pubic symphysis by >=4 mm on pelvic radiography than those in patients without massive hemorrhage (Table 2).

Of the 106 patients with massive hemorrhage, 90 (84%) had poste- rior pelvic fractures on CT, of which 80 (75%) had anterior pelvic injury on pelvic radiography. Contrarily, of the 140 patients without massive hemorrhage, 52 (37%) had posterior pelvic fractures on CT, of which only 15 (10%) had anterior pelvic injury on pelvic radiography.

Multiple regression analysis found that standardized partial regres- sion coefficients for laterality and displacement of the obturator ring were 0.024 (95% CI, 0.012-0.037; p b 0.001) and 0.044 (95% CI, 0.031-

0.056; p b 0.001), respectively. Laterality and displacement of the obtu- rator ring were significantly correlated with massive hemorrhage. How- ever, standardized partial regression coefficient for displacement of the pubic symphysis was 0.008 (95% CI, -0.004-0.020; p = 0.190), with no correlation with massive hemorrhage.

A significant regression equation was derived with a coefficient of multiple determination (R*²) of 0.3027 and variance inflation factor of 1.170:

Massive pelvic hemorrhage = 0.024

• (laterality of the obturator ring)

+ 0.044

• (displacement of the obturator ring)

+ 0.104

Multiple regression analysis showed that the presence of a displaced obturator ring or obturator ring with laterality on the pelvic radiography could predict massive pelvic hemorrhage in 30% of cases.

multiple logistic regression analysis identified displacement of the obturator ring >=5 mm, obturator ring with laterality >=5 mm, and dis- placement of the pubic symphysis >=4 mm on pelvic radiography as pre- dictors of massive pelvic hemorrhage (Table 3).

Discussion

Numerous previous investigations have reported predictors of pelvic massive hemorrhage related to pelvic fractures [5-8]. However, rapid and easy Diagnostic procedures, such as echography or pelvic

Table 1

Comparison of patients with pelvic fractures with and without massive hemorrhage.

Characteristics	Massive hemorrhage (n = 106)	No massive hemorrhage (n = 140)	p value
Age (yr)	65.8 +- 21.1	59.2 +- 31.4	0.063
Male gender (%)	45 (42.4)	74 (52.8)	0.105
SBP on arrival (mm Hg)	98.1 +- 34.4	122.0 +- 28.3	b0.001
HR on arrival (beats/min)	96.5 +- 21.9	82.2 +- 18.9	b0.001
Hemodynamic instability (%)	50 (47.1)	0 (0)	b0.001
Mortality (%)	25 (23.5)	4 (2.8)	b0.001
ISS	30.0 +- 13.5	17.8 +- 12.5	b0.001
Requirement for PRBCs in the initial 24 h (units)	10.4 +- 11.0	0.93 +- 2.33	b0.001
LOS-ICU (days)	4.12 +- 4.94	0.90 +- 2.97	b0.001
Displacement of obturator ring on x-p (mm)	6.6 +- 5.0	2.1 +- 2.8	b0.001
Laterality of obturator ring on x-p (mm)	6.4 +- 5.0	2.7 +- 3.4	b0.001
Displacement of pubic symphysis on x-p (mm)	5.4 +- 5.9	4.1 +- 2.1	b0.001
Posterior pelvic ring fractures on CT (%)	88 (83.0)	52 (37.1)	b0.001
Arterial extravasation of contrast-enhanced pelvic CT (%)	65 (61.3)	0 (0)	b0.001
Positive angiography (%)	77 (72.6)	0 (0)	b0.001

SBP; systolic blood pressure, HR; heart rate, ISS; injury severity score, PRBC; packed red blood cells, LOS-ICU; length of stay in the intensive care unit, CT; computed tomography.

Table 2

Radiographic characteristics of patients with anterior pelvic injury.

ring disruptions, hemipelvic cephalad displacement exceeding 0.5 cm was characterized as unstable, and pelvic instability was accurately di-

Obturator ring with laterality

>=5 mm (%)

Displacement of obturator ring >=5 mm (%)

Displacement of pubic symphysis

>=4 mm (%)

More than 2 types of injury (%)

agnosed in 88% of cases [15]. In our study, obturator ring with laterality was an indirect marker for coxa displacement; these findings were con- sistent with those of Edeiken-Monroe et al. who retrospectively

Massive hemorrhage (n = 106)

No massive hemorrhage (n = 95)

58 (54.7) 66 (62.2) 61 (57.5) 64 (60.7)

11 (7.8) 19 (13.5) 51 (36.4) 17 (12.1)

reviewed 42 patients with anteroposterior compression (APC) II inju- ries with >=25 mm of diastasis of the pubis symphysis who underwent pelvic radiography and found that patients with posterior iliac offset had greater displacement of anterior sacroiliac widening, pubic sym- physis diastasis, and static vertical ramus offset [16]. However, the stud-

p value b0.001 b0.001 b0.001 b0.001

radiography, have not been reported to identify hemorrhage related to pelvic fractures. And also, no definitive method for determining predic- tors of pelvic massive hemorrhage on initial treatment has been reported.

The present study found that patients with massive hemorrhage re- lated to pelvic fractures had a higher frequency of anterior pelvic frac- tures on pelvic radiography and of posterior pelvic fractures on CT, than that in patients without massive hemorrhage. Multiple regression analysis showed that the presence of anterior pelvic injuries with large displacement on pelvic radiography could predict the presence of mas- sive hemorrhage in 30% of cases of with pelvic fractures. Logistic regres- sion analysis identified anterior pelvic injury with large displacement on pelvic radiography as a predictor of massive pelvic hemorrhage.

Pennal et al. first reported that anterior injuries of the pelvic ring were associated with posterior ones [12]. They also suggested that these posterior lesions were often difficult to identify because of sponta- neous reposition of fractured fragments. Scheyerer et al. conducted a study involving 233 patients diagnosed with Pubic rami fractures and found an association between undetected posterior ring injury and prolonged pain or increased morbidity; posterior pelvic ring injury was found on CT in 96.8% of patients without other obvious injuries on pelvic radiography [11]. They concluded that posterior pelvic ring in- jury could easily be missed on initial pelvic radiography. They also re- ported that anterior displacement of the pelvis was more distinctive in patients with posterior ring injury requiring surgical stabilization than in those treated nonsurgically (6.5 mm vs. 4.5 mm, respectively). Courtney et al. found that 68% of patients with inferior ramus fractures had a posterior ring injury and concurrent superior ramus fractures. Displaced unilateral inferior ramus fracture and parasymphyseal inju- ries were associated with higher incidence of posterior ring injury and were more likely to be markers of pelvic instability [10]. Weaver et al. performed a retrospective review of 318 patients with lateral compres- sion (LC) fractures and reported that patients with anterior displace- ment were thrice more likely to have posterior displacement than those without anterior displacement [13]. Lefaivre et al. retrospectively reviewed 100 patients with LC type I pelvic ring disruptions and found an association between the presence of a comminuted rami fracture and complete sacrum injury (p = 0.003) [14]. They concluded that se- verity of anterior ring disruption was a predictive factor for complete disruption of the sacrum. In a previous report by Edeiken-Monroe et al. on anteroposterior pelvic radiography of 154 patients with pelvic

Table 3

Predictors of major hemorrhage related pelvic fractures.

Predictor	Odds ratio	95% CI	p value
SBP on arrival	0.98	0.97-1.00	0.060
HR on arrival	1.01	0.99-1.03	0.212
ISS	1.04	1.01-1.07	0.004
Posterior ring fractures on CT	2.45	1.14-5.27	0.022
Displacement of obturator ring >=5 mm	1.23	1.11-1.36	b0.001
Obturator ring with laterality >=5 mm	1.16	1.04-1.29	0.009
Displacement of pubic symphysis >=4 mm	1.44	1.06-1.94	0.018

SBP: systolic blood pressure, HR; heart rate, ISS; injury severity score, CT; computed to- mography, CI; confidence interval.

ies of Scheyerer et al., Courtney et al., Weaver et al, Lefaivre et al.,

Edeiken-Monroe et al, and Tonne et al. did not examine the relationship between severity of pelvic injury and pelvic hemorrhage. The definition of unstable displacement proposed by Cryer et al. was similar to that of the current study; displacement of the obturator ring by >=5 mm, obtura- tor ring with laterality of >=5 mm, and displacement of the pubic sym- physis by >=4 mm on pelvic radiography are predictors of massive pelvic hemorrhage [9].

The increased pelvic volume in pelvic fractures involving injury of a ligament that maintains the pelvic stability, such as the sacroiliac liga- mentous complex, is thought to reduce the tamponade effect of the ret- roperitoneal space, leading to uncontrollable massive hemorrhage into the pelvic space. This hemorrhage can originate from injury to the arter- ies or venous plexus in the pelvis or fracture bleeding. Huittinen et al. re- ported that approximately 85% of bleeding in case of pelvic fractures is venous in nature, and few cases have arterial bleeding [17]. Although there are no tools for the detection of venous bleeding and evaluation of the lesion and the amount of venous hemorrhage, Pelvic angiography is useful for detecting arterial bleeding associated with pelvic fractures. Hamill et al. evaluated the association between the pattern of pelvic frac- tures and need for pelvic arterial embolization and found that emboliza- tion was required more frequently in patients with Fracture patterns indicative of major pelvic ligamentous disruption, such as APC II, APC III, LC III, vertical shear (VS), and combined mechanical injuries, than in patients without these fracture patterns [18]. In a retrospective review of 231 hypotensive patients with pelvic fractures, hemorrhage in pa- tients with unstable fracture patterns characterized as APCII, APC III, LC III, or VS was predominantly from a pelvic source [19]. Unstable pelvic fracture with posterior ring injury was associated with higher risk of con- comitant massive hemorrhage. Our study indicated that displacement of the obturator ring by >=5 mm, obturator ring with laterality of >=5 mm, and displacement of the pubic symphysis by >=4 mm on pelvic radiography were predictors of posterior ring injury and massive hemorrhage.

This study had several limitations. First, the inferences drawn from

these observations were limited by the small sample size and uncon- trolled bias inherent to a retrospective study. Second, the comparison be- tween the exposed and control groups was limited by the heterogeneity of the sample because of lack of patients without massive hemorrhage who underwent angiography. Finally, the bias toward preperitoneal pack- ing being performed to treat massive pelvic hemorrhage was unclear, be- cause preperitoneal packing was not performed during the study period. In summary, the present study demonstrated a simple tool to predict possible severe hemorrhages by plain pelvic radiography before com- puted tomography or angiography is available. And this study suggested that the presence of displaced anterior lesions of the pelvic ring on pel- vic radiography alone may promptly identify patients at high risk of massive pelvic hemorrhage requiring intervention for hemorrhage con- trol. Further prospective and multi-institutional study with large num-

ber of participants is warranted to support the present findings.

References

1. Dalal SA, Burgess AR, Siegel JH, Young JW, Brumback RJ, Poka A, et al. Pelvic fracture in multiple trauma: classification by mechanism is key to pattern of organ injury, re- suscitative requirements and outcome. J Trauma 1989;29:981-1000.
2. Moreno C, Moore EE, Rosenberger A, Cleaveland HC. Haemorrhage associated with major pelvic fracture. J Trauma 1986;26:987-94.
3. Rotheberger DA, Fischer RP, Strate RG, Velasco R, Perry Jr JF. The mortality associated with pelvic fractures. Surgery 1978;84:356-61.
4. Rommens PM, Hofmann A, Hessmann MH. Management of acute hemorrhage in pelvic trauma: an overview. Eur J Trauma Emerg Surg 2010;36:91-9.
5. Young JWR, Burgess AR, Brumback RJ, Poka A. Pelvic fractures: value of plain radiog- raphy in early assessment and management. Radiology 1986;160:445-51.
6. Hauschild O, Aghayev E, von Heyden J, Strohm PC, Culemann U, Pohlemann T, et al. Angioembolization for pelvic hemorrhage control: results from the German pelvic injury resister. J Trauma Acute Care Surg 2012;73:679-84.
7. Brown CV, Kasotakis G, Wilcox A, Rhee P, Salim A, Demetriades D. Does pelvic hema- toma on admission computed tomography predict active bleeding at angiography for pelvic fracture? Am Surg 2005;71:759-62.
8. Blackmore CC, Cummings P, Jurkovich GJ, Linnau KF, Hoffer EK, Rivara FP. Predicting major hemorrhage in patients with pelvic fracture. J Trauma 2006;61:346-52.
9. Cryer HM, Miller FB, Evers BM, Rouben LR, Seligson DL. Pelvic fracture classification: correlation with hemorrhage. J Trauma 1988;28:973-80.
10. Courtney PM, Taylor R, Scolaro J, Donegan D, Mehta S. Displaced inferior ramus frac- tures as a marker of posterior pelvic injury. Arch Orthop Trauma Surg 2014;134: 935-9.
11. Scheyerer MJ, Osterhoff G, Wehrle S, Wanner GA, Simmen HP, Werner CM. Detec- tion of posterior pelvic injuries in fractures of the pubic rami. Injury 2012;43: 1326-9.
12. Pennal GF, Tile M, Waddell JP, Garside H. Pelvic disruption: assessment and classifi- cation. Clin Orthop Relat Res 1980(151):12-21.
13. Weaver MJ, Bruinsma W, Toney E, Dafford E, Vrahas MS. What are the patterns of injury and displacement seen in lateral compression pelvic fractures? Clin Orthop Relat Res 2012;470:2104-10.
14. Lefaivre KA, Padalecki JR, Starr AJ. What constitutes a Young and Burgess lateral compression- I (OTA 61-B2) pelvic ring disruption? A description of computed to- mography-based fracture anatomy and Associated injuries. J Orthop Trauma 2009; 23:16-21.
15. Edeiken-Monroe BS, Browner BD, Jackson H. The role of standard roentgenograms in the evaluation of instability of pelvic ring disruption. Clin Orthop Relat Res 1989 (240):63-76.
16. Tonne BM, Kempton LB, Lack WD, Karunakar MA. Posterior iliac offset: description of a new radiological measurement of sacroiliac Joint instability. Bone Joint J 2014; 96-B:1535-9.
17. Huittinen VM, Slatis P. Postmortem angiography and dissection of the hypogastric artery in pelvic fractures. Surgery 1973;73:454-62.
18. Hamill J, Holden A, Paice R, Civil I. Pelvic fracture pattern predicts pelvic arterial haemorrhage. Aust N Z J Surg 2000;70:338-43.
19. Eastridge BJ, Starr A, Minei JP, O’Keefe GE, Scalea TM. The importance of fracture pat- tern in guiding therapeutic decision-making in patients with hemorrhagic shock and pelvic ring disruption. J Trauma 2002;53:446-50.