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

To the Editor:

Dahm et al. correctly point out that the studies cited by us do not show a statistically significant increase in renal graft survival with tacrolimus as compared with cyclosporine. Yet, there is one randomized study in which renal transplant patients receiving tacrolimus had a significantly higher graft survival than those treated with cyclosporine microemulsion (1).

We disagree with the statement of Dahm et al. that the manner of presentation of our data is misleading. Regarding the graphical presentation of serum creatinine, we do not see the relevance of including the zero point on the scale since serum creatinine never reaches a value of zero (like e.g. body temperature). Furthermore, we do not agree that the use of SEM instead of SD, if clearly indicated, is misleading as it allows the reader to evaluate the statistical robustness of the observed differences, which the SD does not permit.

The way in which the scatter is presented does not influence the outcome of the statistical analysis, nor does it affect the clinical relevance of the findings. The latter depends on the size of the estimated difference between the groups that, we readily admit, was not very large for serum creatinine. However, the fact that the curves of the two groups were diverging suggests that with longer follow-up the use of tacrolimus may prove to be less detrimental to renal function than the use of cyclosporine.

With respect to the statistical analysis, Dahm et al. have a number of concerns to which our response is as follows. Overlapping distributions of data do not preclude statistically significant differences between groups. Normalization of data was undertaken to enable optimal statistical analysis and if omitted could potentially result in misleading conclusions. We excluded patients who switched in their study medication from analysis of metabolic parameters, to examine more closely the effect of each drug (per protocol analysis). For patients developing diabetes mellitus, evaluation of the glucose levels and HbA1c is confounded by the introduction of antidiabetic treatment. However, there was no significant difference between both groups in the number of patients developing diabetes mellitus. The grouping of continuous variables formed no part of the primary analysis on which our conclusions are based.

In conclusion, we think that the statistical methods that we used were appropriate and do not compromise the validity of the data. The clinical relevance of our findings is a separate issue that is mainly related to the magnitude of the differences between the two groups. We feel that the combination of a slightly better preserved renal function, a decrease in Framingham risk score, and a reduction in side effect scores after conversion from cyclosporine to tacrolimus is of considerable relevance for our patients.