New insights into the effect of haemodiafiltration on mortality: the Romanian experience.

BACKGROUND Haemodiafiltration (HDF), by successfully removing the larger solutes and protein-bound compounds, may offer a feasible approach to improve dialysis outcomes. Recently, three large, randomized, controlled trials have tested this hypothesis, but only one showed an improved survival associated with HDF treatment, when compared with haemodialysis (HD). METHODS This is a retrospective analysis of the entire Romanian dialysed population from the European Clinical Database (EUCLID) Fresenius Medical Care Database. We conducted two types of analysis. First, we used an intention-to-treat approach including all patients who were in dialysis (either HDF or HD) at 1 March 2010--'prevalent cohort analysis'. We then considered only the incident patients who started dialysis (either HDF or HD) after 1 March 2010--'incident cohort analysis'. In both analyses, patients were followed until 31 April 2013. RESULTS In the prevalent cohort, we included 1546 patients who were already performing dialysis at the first time point-1322 on HD and 224 on HDF. When compared with HD, HDF treatment was associated with reduced mortality in both univariate and multivariate survival analysis (HR = 0.67, 95% CI 0.46-0.96 and HR = 0.58, 95% CI 0.36-0.93, respectively). In the incident cohort, 2447 patients started dialysis (2181 HD and 266 HDF) during the observation period. Patients in the HDF group maintained a reduced risk for all-cause mortality (HR = 0.20, 95% CI 0.11-0.38 for the univariate and HR = 0.24, 95% CI 0.13-0.46 for the fully adjusted model). CONCLUSIONS This study suggests that HDF treatment could reduce all-cause mortality in incident and prevalent patients even after correction for different confounders. Interestingly, an additional survival benefit could be observed in incident patients. However, as with any observational study, there could have been other unmeasured confounders that could have influenced our final results.