OTA 1997 Posters - Pelvic & Acetabular Fractures

Angiography and Embolization in Pelvic Ring Fractures

Mark D. Mellinger, MD, Jeffrey M. Wallentine, MD, John T. Ruth, MD, Haiyan Cui, PhD

Tucson, Arizonia, USA

Purpose: We wished to determine if fracture classification was predictive of pelvic arterial injury, resuscitative requirements, and mortality. This information would be extremely useful in refining selection criteria for angiography and in determining more effective triage of pelvic fractures.

Methods: This is a retrospective study of 40 trauma patients (29 male, 11 female, mean age 41) with pelvic ring fractures who underwent angiography for suspected pelvic arterial injury. The selection criteria for angiography was based solely on clinical suspicion of pelvic arterial injury. There was no hemodynamic threshold. Study time period was from 7/88 to 4/96. Hemodynamic and resuscitation data were obtained. Radiology studies were reviewed and AO fracture classifications assigned. Angiographic studies were reviewed and arterial injury and embolizations recorded. Associated injuries were recorded. Trauma scores were calculated. Statistical analyses were performed using ANOVA for continuous variables, and the Chi Square test for categorical variables. Data with small sample size were analyzed using the Wilcox Rank test for continuous variables and the Fisher Exact test for categorical variables.

Results: The most frequent mechanisms of injury were MVA (n=20), pedestrian struck (n=10), and motorcycle accident (n=4). Other mechanisms of injury include bicycle accident (n=2), fall from height (n=1), straddle injury from bucking horse (n=1), hang glider accident (n=1), and pedestrian versus train (n=1). The mean injury severity score (ISS) for all patients was 31.6 (8-75). The mean ISS for patients with and without angiographic confirmation of pelvic arterial injuries were 31.3 (10-50) and 32.2 (8-75) respectively (p>0.05). There was correlation between ISS and mortality rate (p=0.008), survival probability (SP) and mortality rate (p=0.035), and Glasgow Coma Score (GCS) and shock (p=0.017). AO Fracture classification and pelvic arterial injury did not have statistically significant correlation with mean ISS, SP, GCS, or mortality rate.

Angiography detected 25 (62.5%) patients with pelvic arterial injuries (n=52). Multiple pelvic arterial injuries were found in 12 (30%) patients, 9 (22.5%) with bilateral injuries. The most common pelvic arteries injured were the internal iliac artery (n=29), pudendal artery (n=6), and the superior gluteal artery (n=5), obturator (n=3), inferior vesicular (n=2), inferior epigastric (n=2), unspecified internal iliac branch (n=2), inferior gluteal (n=1), external iliac (n=1), and superficial femoral (n=1). A total of 46 pelvic arteries were embolized on 27 (68%) patients. Bilateral internal iliac artery embolizations were performed on 12 (30%) patients.

The mortality rate was 32.5 percent (n=13). Seven deaths (58%) were attributed to blood loss. Five patients died within 24 hours of arriving at the emergency department. All of these early deaths were secondary to blood loss. There was no correlation between GCS and mortality. The mortality rate for patients undergoing embolization was 25 percent (n=10). These 10 patients averaged an additional 4 units of PRBCs prior to embolization, and 5.5 units after embolization. One patient died secondary to blood loss in the angiography suite prior to embolization.

Pelvic ring fractures included AO Type A (n=2), Type B (n=23), and Type C (n=13). There were 2 isolated acetabular fractures and 8 concomitant acetabular fractures. Incidence of bleeding pelvic vessel by fracture classification was as follows: AO Type A (1/2, 50%), Type B (14/23, 61%), Type C (9/13, 69%), and isolated acetabular fracture (1/2, 50%). There were no statistically significant differences between AO fracture classes and pelvic arterial injuries (number or site). External fixation was used in 11 (27.5%) patients. Ten (25%) patients underwent ORIF of the pelvic ring fracture.

The average 24 hour blood requirement for all patients was 15.8 units (0 - 41). The average blood requirements prior to and after embolization were 12.8 units (0 - 40) and 10.9 units (0 - 32) respectively. There was a statistically significant difference in 24 hour blood requirements for patients with and without pelvic arterial injuries (19.0 units vs. 11.2 units, p = 0.017), patients with and without coagulopathy (18.4 units vs. 6.9 units, p = 0.017), patients with and without shock (18.4 units vs. 10.8 units, p = 0.017), and patients surviving and not surviving (13.0 units vs. 21.5 units, p = 0.022). Shock on admission did not correlate with pelvic arterial injury, AO fracture classification, or mortality.

There were numerous concomitant and associated injuries: CHI (n=18), visceral bleed / liver laceration / splenic injury (n=13), pneumothorax (n= 3), bowel disruption (n=4), closed skeletal fracture (n=29), open skeletal fracture (n=12), spine injury (n=10). Controlling for other sources of blood loss did not change the statistical outcome for hemodynamic requirements.

Discussion: In this study population, there is no statistically significant angiographic, hemodynamic, or mortality differences between the unstable AO fracture types. However, the relative risk for pelvic arterial injuries in stable and unstable fractures were 1.55 and 2.25 for AO fracture classifications Type B and Type C. This obvious trend could not be statistically verified due to the small sample size of Type A fractures (n=2). This small sample probably resulted from the selection bias for unstable fractures. However, the 24 hour blood requirements averaged an additional 11.5 units with coagulopathy, 8.5 units with mortality, 7.8 with bleeding pelvic artery, and 7.6 units with hypotension. This retrospective study reconfirmed that ISS, AISmax, and SP correlate with mortality. These trauma-related "scores" did not correlate with pelvic ring fracture classifications. Due to the lack of consistent use and relatively small number of patients, no comparisons were made regarding the use of external fixation.

Conclusion: We were unable to identify any true predictors of arterial injuries requiring angiography and embolization in patients with pelvic ring fractures. Other studies have suggested that angiography be performed when PRBC requirements exceed 4 units in the first 24 hours. Our data does not support this as the average blood requirement in the first 24 hours was 10.5 units in patients without pelvic arterial injury, and 19.0 units in patients with pelvic arterial injury by angiography. We would recommend angiography in patients with clinically unstable fracture patterns (AO Types B and C), who remain hemodynamically unstable despite adequate resuscitation efforts in whom no other obvious source of blood loss can be identified. Given the high rate of concomitant injuries involving substantial blood loss, clinicians should maintain a high degree of suspicion for multiple sources of hemorrhage, including pelvic arterial injury.