Background: Angiographic embolization has emerged as the treatment modality of choice for bleeding pelvic fractures. The purpose of this study is to identify potential indicators for ongoing pelvic hemorrhage despite initial therapeutic or non-diagnostic angiography.
Methods: The trauma registry of a Level I trauma center was used to identify patients with pelvic fractures between January 2000 and June 2002. Records were reviewed for demographics, severity of injury, hemodynamic status, initial and subsequent base deficit, blood and fluid requirements, length of stay, and mortality. Statistical analysis was performed using Student's t test, and univariate and multivariate analysis, significance was assigned to p < or = 0.05.
Results: During the study period, 678 patients had pelvic fractures. Angiography was performed in 31 (4.6%) of these patients. Arterial hemorrhage was diagnosed initially on 16 (51.6%) patients requiring embolization. Three (18.8%) of these embolized patients required repeat angiography and embolization due to ongoing pelvic hemorrhage. Of the initial 15 patients with negative angiograms, five (33.3%) had repeat angiograms due to continued hypotension and acidosis. Four (80.0%) of these five patients were found to have arterial hemorrhage requiring embolization. Of the seven (22.6%) patients requiring repeat angiography for control of ongoing pelvic hemorrhage, three independent factors were predictive: continued or recurrent hypotension (SBP < 90), absence of intra-abdominal injury, and persistent base deficit of 10 for greater than 6 hours. The presence of all three independent predictors was associated with a 97% probability of pelvic bleeding (p = 0.001).
Conclusion: Angiographic embolization is highly effective in controlling arterial bleeding associated with pelvic fractures. However, repeat angiography should be performed in patients with pelvic fractures with ongoing evidence of hemorrhage demonstrated by persistent base deficit and hypotension once other potential sources of bleeding have been excluded.