Perioperative antibiotic use in high-risk penetrating hollow viscus injury: a prospective randomized, double-blind, placebo-control trial of 24 hours versus 5 days.

BACKGROUND

The purpose of this study was to compare the safety and therapeutic efficacy of a 24-hour versus 5-day course of ampicillin/sulbactam for the prevention of postoperative infections in high-risk patients sustaining hollow viscus injury from penetrating abdominal trauma.

METHODS

A total of 317 patients from four Level I trauma centers with penetrating abdominal injuries and at least one hollow viscus perforation each received one preoperative and three postoperative doses of ampicillin/sulbactam 3 g intravenously. After receiving 24 hours of unblinded ampicillin/sulbactam, patients were then randomized into one of two groups. Group 1 received 4 additional days of blinded ampicillin/sulbactam (5 days total of antibiotic), and Group 2 received 4 days of placebo (24 hours of antibiotic). Patients were assessed postoperatively for occurrence of deep surgical-site infections (intra-abdominal abscess, fasciitis, and peritonitis) and superficial (wound) surgical-site infections. Development of nonsurgical-site infections (e.g., pneumonia, urinary tract infection, phlebitis, and cellulitis) was also recorded. Continuous variables were analyzed by analysis of variance and discrete variables by the Cochran-Mantel-Haenszel chi2 test. Multivariate logistic regression analyses were also performed to identify independent risk factors for postoperative infection.

RESULTS

A total of 159 patients were randomized into Group 1, and 158 patients were randomized into Group 2. The Injury Severity Score and penetrating abdominal trauma index were 18+/-8 and 21+/-13, respectively, for Group 1 and 18+/-9 and 20+/-15, respectively, for Group 2. A total of 162 (51%) patients sustained one or more colon injuries (82 in Group 1 and 80 in Group 2). There were 16 (10%) surgical-site infections in Group 1 and 13 (8%) surgical-site infections in Group 2 (p = 0.74). Group 1 patients experienced 17 (11%) nonsurgical-site infections, whereas Group 2 had 32 (20%) nonsurgical-site infections. This difference, however, was not statistically significant (p = 0.16). Only the total number of blood units transfused and the presence of a PATI score greater than or equal to 25 were found to be independently associated with the development of a postoperative surgical- and nonsurgical-site infections (p = 0.001 and p = 0.003, respectively). Of note, the presence of a colon injury was not found to be an independent risk factor (p = 0.11) for either surgical or nonsurgical site postoperative infection in our study.

CONCLUSION

High-risk patients with colon or other hollow viscus injuries from penetrating abdominal trauma are at no greater risk for surgical-site or nonsurgical-site infection when treated with only a 24-hour course of a broad-spectrum antibiotic.