Clinical trials on hemodiafiltration

In the present chapter, several clinical hemodiafiltration (HDF) studies are discussed, with special emphasis on the reliability of the methodology used. These studies differ widely in design, end points, patient numbers, treatment and comparator groups, and amount of convection volume in the treatment arms. Recently, three large randomized controlled trials (RCT) have been performed comparing survival between online postdilution HDF and hemodialysis (HD). While neither CONTRAST nor the Turkish HDF study showed differences between study arms, in the Spanish ESHOL study the mortality risk was significantly lower in HDF patients (HR 0.70; 95 % CI 0.53-0.92). In all three studies, post-hoc (on-treatment) analysis showed a survival benefit up to 40 % for patients treated with high-volume HDF (convection volume >20-22 L/treatment). Apart from these RCTs, in the last 4 years five meta-analyses on convective therapies have been performed, including four on aggregated study results and one on pooled individual patient data (IPD). Since in the latter approach all individual patient data from trials are ascertained, put together and combined to a new data base, this type of meta-analysis is most reliable for making adjustments and evaluating subgroup results. Notably, in both types of meta-analyses the hazard ratio (HR) for mortality of online HDF (as compared to HD) was about 0.83-0.86, indicating a 15 % lower mortality risk for patients treated with HDF. From the IPD meta-analysis it appeared that the mortality risk was even lower when high convection volumes are applied (HR 0.78; 95 % CI 0.62-0.98). Finally, meta-analysis on modern convective therapies should include only online treatments with a convection volume >20 L/session.