Antibiotic prophylaxis in head and neck cancer surgery: Systematic review and Bayesian network meta-analysis

Postoperative surgical site infections (SSI) are worrisome complications after head and neck cancer surgery. Due to the lack of direct comparisons, it is difficult to indicate the best antibiotic in the setting of SSI prophylaxis for patients with head and neck cancer. For this reason, we decided to conduct a Bayesian network meta-analysis of clinical studies evaluating various antibiotic classes for SSI prophylaxis; thus, we directly and indirectly compared all the available antibiotics in the setting of head and neck oncological surgery. We performed a systematic review and a network meta-analysis according to the PRISMA-NMA checklist. Regarding inclusion criteria, we included articles with retrospective or prospective design recruiting adult participants with head and neck neoplasm of any subsite that compared different antibiotics or that compared different antibiotic treatment duration with sample sizes of more than 10 patients. Overall, 310 citations were identified by the search on all databases. Of these, 250 were excluded based on title and abstract, leaving 60 publications for full text examination. Finally, 20 papers were included for data extraction and analysis. In total, 8627 patients were evaluated across the included studies. Results of the Bayesian network meta-analysis showed that when compared to clindamycin short course, the antibiotics effective in preventing SSI were ampicillin/sulbactam or other penicillin short course (OR: 0.37, 95%CrI: 0.19-0.72), cefazolin/metronidazole short course (OR: 0.26, 95%CrI: 0.06-0.93), cefazolin (OR: 0.36, 95%CrI: 0.17-0.79), ampicillin/sulbactam long course (OR: 0.20, 95%CrI: 0.04-0.91), cefazolin/metronidazole long course (OR: 0.27, 95%CrI: 0.09-0.64), cefoperazone (OR: 0.05, 95%CrI: 0.002-0.89), cefotaxime (OR: 0.04, 95%CrI: 0.002-0.85). There was no significant difference between clindamycin and no antibiotic (OR: 2.3, 95%CrI: 0.59-9.9). Clindamycin plus aminoglycoside seemed to give a slight protection from SSI compared to clindamycin alone (OR: 0.30, 95%CrI: 0.09-0.99) or no antibiotic (OR: 0.13, 95%CrI: 0.02-067). Antibiotic prophylaxis is important in preventing SSI in head and neck cancer setting. Current evidence suggests that penicillins and cephalosporins are the best choice. Moreover, long duration course does not give any advantage compared to short course prophylaxis. Finally, it is important to implement appropriate antibiotic prophylaxis in patients that are labeled as penicillin allergic, in this regard clindamycin seems to be ineffective and, for this reason, further research is needed to provide a better care for this subset of patients.