Predictors of positive angiography in pelvic fractures: a prospective study.

BACKGROUND

Severe pelvic fractures continue to be a major problem for trauma surgeons. Early identification of patients who would benefit from therapeutic angiographic embolization (AE) of pelvic bleeding would be beneficial. We hope to identify simple risk factors that would pinpoint patients who would benefit from therapeutic AE.

STUDY DESIGN

This is a prospective observational study at an academic Level I trauma center. All blunt trauma patients with a pelvic fracture admitted from December 2003 to February 2007 were included. AE was performed for hemodynamic instability (systolic blood pressure < 100 mmHg), fracture pattern (sacroiliac joint [SIJ] disruption, butterfly, open book), or CT demonstrating a large pelvic hematoma. AE was considered therapeutic if contrast extravasation was noted and addressed with embolization. Main outcomes measures were positive angiography and mortality. Stepwise logistic regression was performed to identify predictors of therapeutic AE. A predictive model was built based on these independent risk factors to estimate the probability of a therapeutic AE.

RESULTS

One hundred thirty-seven (23%) of 603 patients with pelvic fractures had angiography. Therapeutic AE was performed in 85 patients (62%). Indications for angiography included hemodynamic instability (58%), fracture pattern (26%), and CT findings (9%). Nineteen patients (22%) with therapeutic AE had no hypotension or tachycardia. Independent predictors for therapeutic AE were SIJ disruption (odds ratio [OR]: 4.5; 95% CI, 1.6 to 12.6; p = 0.005), female gender (OR: 3.9; 95% CI, 1.5 to 10.0; p = 0.005), and duration (in minutes) of systolic blood pressure < 100 mmHg (OR: 1.4; 95% CI, 1.1 to 1.9; p = 0.007).

CONCLUSIONS

Presence of SIJ disruption, female gender, and duration of hypotension can reliably predict patients who would benefit from AE. This predictive model can help early identification of patients who would benefit from pelvic angiography.