A CASE-CONTROL STUDY OF RISK-FACTORS THAT PREDICT FEMORAL ARTERIAL INJURY IN PENETRATING THIGH TRAUMA

Study hypothesis: Femoral artery injuries can be predicted by the mech anism of injury, wound location and tract, and physical findings follo wing penetrating thigh trauma. Design: Retrospective case-control stud y. Participants: All 808 consecutive patients undergoing femoral arter iography for penetrating thigh trauma from September 1986 through Dece mber 1990 were eligible for inclusion in the study. All 50 patients in the eligible population with proven femoral artery injuries diagnosed by angiogram were the study subjects. Fifty patients with penetrating thigh trauma who had angiograms negative for injury and were systemat ically chosen from the eligible population served as controls. Interve ntions: Data included mechanism of injury; location of wound entrance, tract, exit, and retained missile; physical findings (including ankle -brachial index); and the presence of femur fractures. Physical findin gs were divided into hard findings (pulse abnormality, expanding hemat oma, or pulsatile bleeding) or soft findings (neurologic deficit, hypo tension without another source, or bruit/thrill). Results: Of the 808 eligible patients, 50 (6.2%) had a femoral arterial injury on angiogra phy, 20 (40%) of which were clinically occult injuries. A medial thigh tract made an arterial injury 58 times more likely (odds ratio [OR], 57.5; P<.001) and was present in 100% of cases and 64% of controls. An anteromedial thigh tract made an arterial injury 12 times more likely (OR, 11.5; P<.001) and was present in 92% of cases and 50% of control s. A wound with hard physical findings was 118 times more likely to ha ve an arterial injury (OR, 118; P<.00001) and was found in 54% of case s and none of the controls. The presence of any physical finding made an arterial injury 36 times more likely (OR, 36; P<.00001) and was fou nd in 60% of cases and 4% of controls. The presence of a femur fractur e or a gunshot mechanism was not predictive of injury. Conclusion: Onl y patients with medial thigh wounds need to undergo angiography for th e detection of femoral artery injuries. This approach would have reduc ed the angiography rate by 36% in this series. Had angiography been pe rformed only on patients with any physical findings, a 70% reduction i n the rate of angiography would have been achieved, although five occu lt arterial injuries per year would have been missed. Angiography shou ld not be peformed solely because of a gunshot mechanism or the presen ce of a femur fracture.