Intraoperative versus extended antimicrobial prophylaxis after gastric cancer surgery: a phase 3, open-label, randomised controlled, non-inferiority trial

Background: Although evidence for the efficacy of postoperative antimicrobial prophylaxis is scarce, many patients routinely receive such treatment after major surgeries. We aimed to compare the incidence of surgical-site infections with intraoperative antimicrobial prophylaxis alone versus intraoperative plus postoperative administration.

Methods: We did a prospective, open-label, phase 3, randomised study at seven hospitals in Japan. Patients with gastric cancer that was potentially curable with a distal gastrectomy were randomly assigned (1:1) to receive either intraoperative antimicrobial prophylaxis alone (cefazolin 1 g before the surgical incision and every 3 h as intraoperative supplements) or extended antimicrobial prophylaxis (intraoperative administration plus cefazolin 1 g once after closure and twice daily for 2 postoperative days). Randomisation was stratified using Pocock and Simon's minimisation method for institution and American Society of Anesthesiologists scores, and Mersenne twister was used for random number generation. The primary endpoint was the incidence of surgical-site infections. We assessed non-inferiority of intraoperative therapy with a margin of 5%. Analysis was by intention-to-treat. During hospital stay, infection-control personnel assessed patients for infection, and the principal surgeons were required to check for surgical-site infections at outpatient clinics until 30 days after surgery. This study is registered with UMIN-CTR, UMIN000000631.

Findings: Between June 2, 2005, and Dec 6, 2007, 355 patients were randomly assigned to receive either intraoperative antimicrobial prophylaxis alone (n=176) or extended antimicrobial prophylaxis (n=179). Eight patients (5%, 95% CI 2-9%) had surgical-site infections in the intraoperative group compared with 16 (9%, 5-14) in the extended group. The relative risk of surgical-site infections with intraoperative antimicrobial prophylaxis was 0·51 (0·22-1·16), which revealed statistically significant non-inferiority (p<0·0001).

Interpretation: Elimination of postoperative antimicrobial prophylaxis did not increase the incidence of surgical-site infections after a gastrectomy. Therefore, this treatment is not recommended after gastric cancer surgery.