Composite guideline-adherent antibiotic prophylaxis and 30-day surgical site infection after caesarean delivery: a single-centre retrospective cohort study

Aims: Surgical antibiotic prophylaxis reduces post-caesarean surgical site infection (SSI), yet adherence to the combined principles of agent, weight-adjusted dose, pre-incisional timing and intraoperative redosing varies widely. We audited composite compliance in a consecutive four-year cohort and asked how much outcome discrimination a binary four-pillar indicator delivers where compliance with each component is already high. Methods: We identified 1100 consecutive women delivered by caesarean between 1 January 2022 and 31 December 2025 in a single-centre retrospective cohort. The exposure was a binary composite of guideline-adherent prophylaxis, an undocumented component counting as not met; the outcome was any 30-day SSI by CDC-NHSN criteria. The primary model was Firth penalised logistic regression on a graph-derived adjustment set, with a co-primary propensity-score analysis and Benjamini Hochberg adjustment of secondary outcomes. Results: Composite compliance reached 63.2%, driven by dosing (75.9%) and timing (85.6%); agent appropriateness was 98.7% and a redosing indication arose in 3 women, so the composite functioned as a dose-and-timing indicator. Among 1053 women with verifiable follow-up the 30-day SSI rate was 3.61%. On 37 events, compliance was not associated with SSI (adjusted odds ratio 0.79; 95% confidence interval (CI) 0.40-1.56; p=0.49); adjusted odds ratios were 0.54 (0.27-1.09) for endometritis and 0.72 (0.37-1.41) for readmission, and no secondary outcome approached significance after adjustment (lowest q=0.332). Conclusion: Composite compliance was not associated with 30-day SSI, and on 37 events the interval remained compatible with either substantial benefit or moderate harm. Because component compliance was already high the contrast between strata was narrow, so these data describe the composite indicator rather than any individual component. Weight-based dosing and pre incisional timing were the weakest components and the most direct local targets.