Actually delivered dose of continuous renal replacement therapy is underestimated in hemofiltration.

Dose determination in continuous renal replacement therapy (CRRT) is controversial. Most clinical trials use effluent flow rate as a surrogate of the dose. However, such definition may overestimate actually delivered dose due to declining filter function. The current study aimed to determine the difference between prescribed and delivered clearance and its association with transmembrane pressure. Hemofiltration was done in a mixed pre- and postdilution mode. Creatinine concentrations in serum and effluent fluid were measured simultaneously at 4, 10, 16, 28, 40, 52, and 64 hours for an individual hemofilter. Prescribed clearance (K) was estimated as the effluent flow rate corrected for predilution, and delivered clearance (Kx) was estimated using the ratio of serum and effluent creatinine. A total of 60 patients involving 248 filters were included in our analysis. The mean filter life span was 37.7 hours (standard deviation: 17.6). K overestimated Kx by 9.3% (95% confidence interval: -4.4% to 32.3%). The differences between K and Kx increased progressively over time. Transmembrane pressure was significantly correlated to the reduction with a Spearman's rho of 0.44 (p < 0.001). K significantly overestimates Kx during CRRT, and the difference increases progressively over time. Filters are recommended to be changed at 48-72 hours on a routine basis.