The Authors’ Reply

We appreciate the insightful comments by authors Gaynor and Ciancio to our study that showed a significant association between high estimated tacrolimus clearance and biopsy-proven acute rejection (BPAR) the first 90 days after transplantation.^1,2

The authors raise 2 concerns. First, in viewing of the Kaplan-Meier freedom-from-BPAR comparisons (Figures 1–3 in the original article), the unfavorable effect of higher tacrolimus clearance is most pronounced in the first 2 weeks after transplantation. In a new logistic regression analysis (Table 1) (where timing of BPAR is not a component), the odds ratio for BPAR the first 90 days after transplantation showed strikingly similar results as in the Cox regression (Table 2 in the original article). Therefore, timing of BPAR does not seem to affect the overall conclusions of our study.

TABLE 1: Multivariate logistic regression model of risk factors associated with BPAR the first 90 days after transplantation

Secondly, we agree that evaluating tacrolimus clearance as a time-dependent covariate would have strengthened our analyses. Unfortunately, the necessary data to perform such analysis were not available. Apparent clearance of tacrolimus decreases during the first months after transplantation,³ and we understand the authors’ concern that estimating clearance at earlier time points for patients experiencing BPAR may have introduced systematic differences. As discussed in the original article, using the 8-week clearance values for all patients regardless of BPAR led to similar results as the original analysis. We chose to report the pre-BPAR data for those patients, since there is a theoretical risk that the steroid-based antirejection treatment may influence tacrolimus metabolism, also at the 8-week estimates for at least some patients. We also agree that the use of different strengths of maintenance steroids may affect tacrolimus clearance. However, when excluding the high-risk patients who start on 80 mg prednisolone as compared to 20 mg in standard-risk patients, the multivariate analysis (published as Supplemental Digital Content to the original article, Table S1, SDC,https://links.lww.com/TP/B443) showed the same results as the analysis based on the whole population. Actually, several factors may affect the clearance and/or bioavailability of tacrolimus, including steroids, body size and composition, cytochrome P450 enzyme genotypes, concomitant food or drug intake, etc. However, irrespective of the underlying determinants of individual apparent clearance, our analysis indicates that this crude clearance approximation may provide a simple marker to identify patients at higher risk for acute rejection episodes.