Regional citrate anticoagulation versus no-anticoagulation for continuous venovenous hemofiltration in patients with liver failure and increased bleeding risk: A retrospective case-control study.

Objective There are controversial opinions on anticoagulation for continuous venovenous hemofiltration (CVVH) in patients with liver failure (LF) and increased bleeding risk. Therefore, we conducted a retrospective study to evaluate the efficacy and safety of regional citrate anticoagulation (RCA) versus no-anticoagulation for CVVH in these patients. Methods The included patients were divided into RCA and no-anticoagulation group according to the CVVH anticoagulation strategy they accepted for CVVH. Filter lifespan, bleeding, citrate accumulation, catheter occlusion, and totCa/ionCa ratio were evaluated as outcomes. Results In the original cohort, the filter lifespan of the RCA group (41 patients, 79 filters) was significantly longer than the no-anticoagulation group (62 patients, 162 filters) (> 72 hours vs 39.5 hours (IQR 31.2–47.8), P = 0.002). The adjusted results demonstrated that RCA could significantly reduce the risk of filter failure (HR = 0.459, 95%CI 0.26–0.82, P = 0.008). Four episodes of totCa/ionCa > 2.5 were observed in the RCA group and continuously accepted RCA-CVVH after the reduction of citrate dose and blood flow. No obvious citrate accumulation was observed in these patients. In the matched cohort, the filter lifespan of the RCA group was significantly longer than the no-anticoagulation group (P = 0.013) as well. No significant difference in the episodes of totCa/ionCa > 2.5 was observed between the two matched groups (P = 0.074). Both in the original cohort and the matched cohort, the bleeding, acidosis, alkalosis, and catheter occlusion incidences were not significantly different between the two groups. Conclusions In LF patients with increased bleeding risk who underwent CVVH, RCA could prolong the filter lifespan and be safely used with careful blood gas monitoring and citrate dose adjusting. Further prospective, randomized, control studies are warranted to obtain robust evidences.