Is it time to refine? An exploration and simulation of optimal antibiotic timing in general surgery

Background Postoperative infections increase morbidity, resource use, and costs. Our objective was to examine whether within guideline recommendations an optimal time exists for an initial dose of antibiotic to reduce postoperative infections in general surgery, and to simulate the magnitude of a reduction in infections should an optimal time be implemented. Study Design The population consisted of 6,731 patients who underwent 7,095 general surgery procedures between January 5, 2006 and June 25, 2012. Patients with pre-existing infections, such as pneumonia and sepsis, and patients with no recorded use of antibiotics were excluded, as were patients on vancomycin and surgical procedures longer than 4 hours in duration. The final analysis dataset included 4,453 patients. The National Surgical Quality Improvement Program was used for perioperative variables and outcomes. The end point was a composite of wound disruption; superficial, deep, organ space, surgical site infections; and sepsis. Semi-parametric logistic regression was used to study the association between antibiotic timing and infection. Results There were 444 (10%) patients with a primary end point of infectious complication. A nonlinear "bowl-shaped" relationship between duration of interval from antibiotic administration and surgical incision and infection was observed; lowest risk corresponding to administration time close to incision was 4 minutes before incision (95% one-sided CI, 0-18 minutes). The model suggested optimal timing would result in an 11.3% reduction in the primary infection end point. Conclusions Risk of infectious complications decreased as antibiotic administration moved closer to incision time. These data suggest an opportunity to reduce infections by 11.3% by targeting initial antibiotic administration closer to incision.