Caveats of Metronidazole Addition for Surgical Site Infection Prophylaxis in Gynecologic Surgeries

Emerging evidence supports the addition of metronidazole to the standard regimen with cefazolin for surgical site infection (SSI) prophylaxis in gynecologic surgeries [1-3]. In specific, enhanced coverage against anaerobic bacteria with metronidazole was reportedly associated with a lower risk of SSI compared to cefazolin alone [1-3]. Whereas these findings may hold important clinical implications, particularly in optimizing perioperative antibiotic regimens for high-risk procedures, there remain unacknowledged limitations within these studies that could bias clinical practices. We would like to comment on important aspects and suggest approaches to avoid these issues in future investigations. For the design of these studies, the most (or only) performed procedure was hysterectomy [1-3], which cannot reflect all microbial characteristics in gynecologic surgeries. The retrospective approach is not appropriate for detecting all SSI cases, especially in patients without medical insurance for data tracking. SSI-related out-of-hospital mortality was also likely to be missed in case of rapidly or severely progressing infection (when SSI diagnosis could not be given timely). Doses of cefazolin were inconsistent or poorly reported, which could complicate the interpretation, generalisability, and synthesis of the findings. In particular, the variations in cefazolin dosage can lead to a violation of the stable unit treatment value assumption of causal inference. Most noteworthily, a before–after design (as in the study of Knisely et al. [1]) could introduce bias and uncontrollable confounding due to time-varying factors. We suggest future investigations focusing on various types of gynecologic surgeries, widely recommended dosage of cefazolin, and 30-day SSI or infection-related mortality as the primary outcome. For the statistical analysis, none of the findings have been adjusted for social determinants of health (SDoH), which were reportedly associated with SSI risk [4-6]. SDoH represent important factors for post-discharge care and wound healing. Future observational studies should consider these factors to avoid obtaining biased estimates. Another issue is that some observational findings may not be suitable for causal inference due to unresolved lack of overlap in SSI risk factors. This phenomenon occurs when patients with a higher risk of SSI are purposely and preferentially given cefazolin and metronidazole combination. Separating the effects of antibiotics from SSI risk factors can be challenging with multivariable regression models in these cases. We suggest measuring all potentially confounding covariates (to the best of the authors' knowledge and capability) and applying propensity score methods to mitigate this issue [7]. If propensity score matching is to be used, the balances of the covariates should be checked using standardized mean differences, not by p-values from statistical tests as in the study of Till et al. [3] p-values are for statistical significance and are usually driven by sample size, which may not demonstrate the magnitude of covariate imbalance. For the interpretation of data and findings, without evidence from randomized controlled trials or adequate causal inferences, the use of causal language, as in these studies [1-3], should be avoided [8]. In this case, potential errors arising from the study design and multiple comparisons are likely to preclude the changes in clinical guidelines. Therefore, to avoid misleading implications and unconfirmed practices, the interpretation of the findings should only focus on the association between cefazolin/metronidazole and SSI. Moreover, as these studies did not investigate the safety of adding metronidazole, there is a lack of evidence to weigh the risks and benefits of this regimen. Further investigations should explore the safety profile of metronidazole for gynecologic surgeries to fill in the knowledge gap. For the applicability of the findings, most studies have not looked into the risk of antimicrobial resistance against metronidazole [1-3], a key antibiotic for anaerobic infections. Many may think that a single dose of metronidazole in gynecologic surgeries will not increase the resistance rate considerably. However, considering the bigger picture of many gynecologic surgeries, either regionally or globally, the accumulation of resistance genes can pose a threat to the healthcare systems worldwide [9], especially for one of the most commonly used antibiotics for anaerobic organisms. Thus, this should be carefully evaluated before updating the clinical guidelines. While it is indeed a challenge to conventionally culture and monitor the resistance profile of anaerobes, investigators can obtain more molecular information using sequencing or DNA amplification on fecal/vaginal samples [10]. In resource-limited settings, issues with advanced equipment may be addressed by commercial kits implementing PCR, loop-mediated isothermal amplification, or recombinase polymerase amplification with lateral flow strips [11, 12]. The goal of this approach is to assess the presence, accumulation, and spread of resistance genes against metronidazole within a population, which can provide more evidence to improve the guidelines. Hence, screening studies are needed to cover this piece of the bigger picture. M.-H.T. and K.-H.T.-N. reviewed the literature and this paper, drafted and revised the manuscript, read, and agreed to the final manuscript. The authors have nothing to report. The authors have nothing to report. The authors have nothing to report. M.-H.T. reported receiving travel reimbursement from Pfizer Vietnam Ltd. and Viatris Vietnam Ltd. speaking fees and grants from Servier Vietnam Ltd. and speaking fees from Aguettant Asia Pacific Pte Ltd. outside the submitted work. K.-H.T.-N. reported receiving travel reimbursement from Pierre Fabre Vietnam Ltd. outside the submitted work. Data sharing not applicable to this article as no datasets were generated or analysed during the current study.