Conversion from Cyclosporine to Tacrolimus Improves Quality-of-Life Indices, Renal Graft Function and Cardiovascular Risk Profile

Abstract

To the Editor: We have read with great interest the recent article by Artz et al. (1). We congratulate the authors for performing a prospective randomized trial on the important and controversial topic of optimal choice of calcineurin inhibitor in renal transplantation. Yet, we would like to voice some concerns on a few issues. First, the opening statement ‘… several studies report an increased graft survival in patients using tacrolimus as initial immunosuppressive treatment’, is inaccurate. In fact, neither the cited studies (2, 3) nor others have shown superior graft or patient survival with tacrolimus-based immunosuppression, although acute rejection and crossover rates were lower (4). Secondly, the manner of data presentation is somewhat misleading: differences between groups are inflated by a y-axis not reaching 0 and scatter seems reduced by giving SEM. A graph showing the values on an absolute scale with the correct measure of scatter reveals that there are no clinically relevant differences between groups (see Figure 1). Serum creatinine at randomization and after two years of follow-up. Thirdly, and most importantly, the analysis is questionable: significant differences in endpoints serum creatinine, blood pressure and lipids are surprising, as visual inspection of distributions reveals wide overlap. Significance was only achieved by additional statistical methods of normalization. Dropouts were apparently excluded from the analysis of metabolic effects. Additionally, the groping of continuous variables into categories (‘50% had a 10% rise in serum creatinine’, etc.), and then testing for significance, is a questionable practice of subgroup analysis. Although statistical significance was achieved in several endpoints by the unusual methods applied, we conclude that this kind of data analysis may not imply clinical relevance and might therefore be misleading. Clinical practice should not be changed solely based on this data.