Background Non-steroidal anti-inflammatory drugs (NSAIDs) are widely used as ancillary therapy for bovine respiratory disease (BRD) alongside antimicrobials, and interest in NSAID monotherapy has grown amid antimicrobial stewardship and welfare concerns. We evaluated the clinical effectiveness of NSAIDs for BRD as adjunctive or sole therapy. Methods We conducted a systematic review and meta-analysis of randomized controlled trials (RCTs) in cattle with naturally occurring BRD. Databases searched were MEDLINE (Ovid), Embase (Ovid), CAB Abstracts (Ovid), Biological Abstracts (Ovid), Web of Science Core Collection, and Scopus (search: 22 Apr 2024; update: 12 Sep 2025). Primary outcomes were short-term treatment failure (need for re-treatment); secondary outcomes included relapse, mortality, adverse events, and performance where available. Random-effects pairwise meta-analyses estimated risk ratios (RRs) with 95% CIs. Risk of bias was assessed with a modified RoB 2.0 tool; certainty of evidence was appraised using GRADE. Registration/protocol: this review extends a prior ENOVAT BRD protocol; no separate registration was created. Funding: COST Action CA18217. Results Seventeen RCTs (22 comparisons; 4,909 animals) compared NSAID+antimicrobial versus antimicrobial alone. Adding an NSAID did not reduce re-treatment (RR 0.94, 95% CI 0.84–1.05; I 2 = 0%; moderate certainty). Subgroup (same vs different antimicrobial) and sensitivity analyses (handling of multi-arm trials; follow-up restricted to ≤14 or ≤ 10 days; risk-of-bias restrictions) did not change conclusions. Two RCTs in UK dairy calves compared NSAID monotherapy with antimicrobial monotherapy (RR 1.19, 95% CI 0.72–1.97; very low certainty). Re-treatment is an indirect outcome and may not capture analgesia, fever reduction, or growth effects. Conclusions Across RCTs, NSAID use did not confer a clinically relevant reduction in re-treatment when added to antimicrobials, and evidence is very uncertain for NSAID monotherapy versus antimicrobials. Future trials should include validated pain/welfare measures, clinical and ultrasonographic outcomes, and performance metrics, with pathogen identification to explore effect modification.
Scahill et al. (2026) studied this question.