79
The individual patient data meta-analysis comparing haemodiafiltration and haemodialysis for kidney failure by Robin W M Vernooij and colleagues1 provides an important contribution to the evidence. However, as the accompanying Comment by Bruno Ranchin and Rukshana Shroff notes, such an analysis “can only be as good as the individual trial data input into it”2 and several issues exist with this meta-analysis. First, 1688 (40%) of 4153 patients in the meta-analysis came from two trials that removed patients at higher cardiovascular risk from the haemodiafiltration group after randomisation.
