Regional Citrate Anticoagulation Versus No Anticoagulation During Continuous Renal Replacement Therapy in Critically Ill Patients at High Risk of Bleeding: A Meta-analysis of Randomized Controlled Trials
摘要
Continuous renal replacement therapy (CRRT) in critically ill patients often requires anticoagulation to maintain circuit patency, but systemic anticoagulation may increase bleeding risk. Regional citrate anticoagulation (RCA) has been proposed as an alternative. This meta-analysis evaluated the efficacy and safety of RCA versus no anticoagulation (NA) during CRRT in critically ill patients at high risk of bleeding.
MethodsPubMed, Embase, Web of Science, Cochrane Library, Wanfang, and China National Knowledge Infrastructure were searched for randomized controlled trials comparing RCA with NA during CRRT in adult critically ill patients with high bleeding risk. Mean differences (MDs) or risk ratios (RRs) with 95% confidence intervals (CIs) were pooled using random-effects models accounting for potential heterogeneity.
ResultsSeventeen RCTs involving 893 patients were included. Compared with NA, RCA significantly prolonged filter lifespan (MD: 13.82 h; 95% CI 10.83–16.81) and reduced the incidence of circuit clotting (RR: 0.20; 95% CI 0.12–0.35). RCA was associated with higher risks of citrate accumulation (RR: 4.67; 95% CI 1.30–16.77) and hypocalcemia (RR: 1.78; 95% CI 1.12–2.83), while acid–base disturbances were not significantly different between groups (RR: 1.34; 95% CI 0.60–3.03). In addition, RCA significantly reduced major bleeding (RR: 0.37; 95% CI 0.16–0.87), but did not significantly affect the in-hospital mortality (RR: 0.82; 95% CI 0.65–1.04).
ConclusionsRCA improves circuit patency during CRRT in critically ill patients with a high risk of bleeding, although it increases the risk of certain metabolic complications but does not affect mortality. While RCA was associated with fewer reported major bleeding events than no anticoagulation, the certainty of evidence for this outcome was low and the finding should be interpreted with caution.