Objective: The effectiveness of continuous renal replacement therapy (CRRT) increases when unplanned circuit failure is prevented. Adequate anticoagulation is an important component. Although heparin is the predominating anticoagulant, calcium chelation with citrate is an alternative, but systemic calcium monitoring and supplementation increase the complexity of CRRT. We assessed efficacy and safety of citrate delivery via integrated software algorithms against an established regional heparin protocol.
Design: Prospective computer randomisation allocated eligible patients to regional citrate or heparin between April and December 2012. Citrate fluids were Prismocitrate 18 mmol/L predilution and Prism0cal B22 dialysate. Hemosol B0 was the default fluid for heparin. The primary outcome was filter running time. Electively terminated circuits were censored. Intention-totreat (ITT) and per-protocol analyses were performed. Filter survival was compared by log-rank tests and hazard ratios were explored with a mixed-effects Cox model.
Results: 221 filters were analysed from 30 patients (of whom 19 were randomly allocated to citrate filters and 11 to heparin filters). Patients randomly allocated to citrate were older, sicker, with a higher male:female ratio, but of similar weight. Mortality was 37% in the citrate arm and 27% in the heparin arm. All deaths were attributed to underlying disease. Significant crossover occurred from the citrate arm to use of heparin. Median filter survival, by ITT, was not significantly different (citrate, 34 hours; heparin, 30.7 hours; P=0.58). Per-protocol survival favoured citrate (citrate, 42.1 hours; heparin, 24 hours; χ(2)=8.1; P=0.004). Considerable variation in filter life existed between patients, and between vascular access sites within patients. Safety end points were reached in one heparin and three citrate patients.
Conclusion: Although the per-protocol results favoured citrate when it was actually delivered, the significant crossover between treatment arms hampered more definitive conclusions. Until further studies support improved patient outcomes, increased complexity and complications suggest that anticoagulation choice be made using patient-specific indications.