Implementation of an integrated blood-warming pathway during continuous renal replacement therapy: a pre–post quality- improvement study

Abstract Background Inadvertent hypothermia is common during continuous renal replacement therapy (CRRT) and reflects both extracorporeal heat loss and variation in warming and surveillance practices. We evaluated a standardized blood-warming and nursing-surveillance pathway in a mixed adult intensive care service. Methods This single-center, uncontrolled pre–post quality-improvement study compared a retrospective preimplementation service audit (1 February–30 April 2026) with a prospective implementation-period cohort (1 May–31 July 2026) in a 22-bed adult intensive care service in Saudi Arabia. The pathway specified a CRRT platform with integrated blood warming, temperature assessment at least every 4 h, escalation below 36°C, competency-focused nurse education, and recurrent audit-feedback. The primary outcome was the proportion of patients with at least one documented temperature < 36°C during CRRT; temperature < 35°C was a secondary outcome. We calculated exact binomial 95% confidence intervals (CIs), Newcombe–Wilson risk differences, unadjusted risk ratios, and two-sided Fisher exact P values. Results Seventy unique patients were included before implementation and 80 unique patients, representing 80 distinct CRRT-treated admissions, during implementation. Temperature < 36°C occurred in 49/70 patients (70.0%; 95% CI, 57.9%–80.4%) before implementation and 5/80 (6.3%; 95% CI, 2.1%–14.0%) during implementation (risk difference, − 63.8 percentage points; 95% CI, − 73.9 to − 50.0; risk ratio, 0.09; 95% CI, 0.04–0.21; P < 0.001). Temperature < 35°C occurred in 16/70 patients (22.9%; 95% CI, 13.7%–34.4%) and 0/80 (0%; 95% CI, 0%–4.5%), respectively (P < 0.001). All five implementation-period events were mild. No bleeding complication or circuit-clotting event was reported in the project log. Conclusions The pathway was associated with substantially less documented hypothermia and no recorded moderate or severe episode during implementation. Because patient-level case mix, CRRT exposure, temperature-measurement route, and component-level fidelity data were unavailable, the findings should be interpreted as a local quality-improvement association rather than proof of causal efficacy. Trial registration Not applicable. This was a servicelevel qualityimprovement initiative rather than a clinical trial, and patients were not assigned to alternative interventions for research purposes.