Low Certainty of Evidence and Heterogeneity Dominate in Systematic Review of Antimicrobial Drug Use and Antimicrobial Resistance in Livestock-The Example of Cattle and Salmonella.

Abstract

Introduction: Predicting the public health impact of policies limiting antimicrobial use (AMU) in livestock requires quantifying the link between AMU and antimicrobial resistance (AMR) in foodborne pathogens. Using cattle and Salmonella as an example, we conducted a systematic review and meta-analysis (SRMA) (PROSPERO #CRD42023399764) to elucidate AMU's impact on AMR in bacteria from animals raised both conventionally (CONV) and without AMU (RWA).

Methods: Using a predefined population (cattle), intervention (AMU), comparison (CONV vs. RWA), and outcome (AMR in Salmonella or commensal Escherichia coli) framework, 36 studies met the inclusion criteria. We estimated pooled odds ratios (POR) describing the association between AMU and Salmonella prevalence, tetracycline, ciprofloxacin, and third-generation cephalosporin (3GC) resistances and evaluated evidence certainty using a GRADE approach. Predictive intervals (PIs) incorporating heterogeneity (τ²) were calculated along with POR to illustrate the effect of between-study differences on association estimates.

Results: Poor evidence certainty was driven by a high risk of bias, imprecise odds ratio estimates, and inconsistency among the included studies. Substantial heterogeneity was observed, and PIs reflected non-significant associations for all AMR outcomes.

Conclusions: Given the poor certainty of evidence and between-studies differences, pooled estimates should not be trusted, necessitating a suitable alternative to estimate the effects of AMU reduction on human health.