Antibiotic prophylaxis in colorectal surgery: are oral, intravenous or both best and is mechanical bowel preparation necessary?

Background The benefit of adding oral antibiotic prophylaxis (OA) to intravenous prophylaxis (IV) in elective colorectal surgery to prevent surgical site infection (SSI) and whether the benefit of OA requires a mechanical bowel cleansing (MBP) are assessed in a systematic review. Meta-analyses compare randomized trials of IV versus IV plus OA, both with MBP; OA versus IV plus OA, both again with MBP; OA plus IV in studies randomizing patients to MBP or no MBP; and IV versus IV plus OA in patients with no MBP. Methods MEDLINE, EMBASE, and the Cochrane Library were searched for eligible studies from 1965 to April 1, 2020. The outcome assessed was SSI, superficial and deep, but not organ space. For each included study, risk of bias was assessed using the Cochrane Risk of Bias tool version 1. For each comparison, meta-analysis was performed from data from eligible studies to obtain a summary effect and heterogeneity using RevMan. Sensitivity analyses were performed excluding studies of poor quality. Certainty of evidence was assessed using GRADE for each comparison. Results Sixty-one studies published in 1971-2020 from 55 publications reporting 12,297 patients were eligible for inclusion. A total of 36 studies compared IV to OA plus IV with MBP. The risk ratio (RR) and 95% confidence interval (CI) for SSI with oral and IV vs. IV alone are 0.47, 0.40-0.56. The RR in 19 studies for IV plus OA versus OA alone is 0.48, 0.38-0.62. The RR for OA plus IV with MBP versus without MBP in 5 studies is 1.17, 0.84-1.64. The RR for OA plus IV versus IV alone when no bowel prep was used in two studies is 0.36, 0.18-0.72. RRs were similar in sensitivity analyses. The GRADE is high for the first two comparisons, moderate for the 3rd, and low for the 4th due to imprecision and heterogeneity. Conclusions Combined OA and IV is superior to either alone in preventing SSI. The certainty of evidence is such that further research is unlikely to alter this relationship when MBP is used. In randomized trials of MBP, OA plus IV shows no benefit from MBP versus no MBP. The last comparison shows in just two studies that as in the first meta-analysis, but in the absence of MBP, combined OA plus IV is also superior to IV alone.