To the Editor:
The recently published PROTECT trial  evaluated the renal sparing effect of using everolimus and tapering calcineurin inhibitors (CNI) in liver transplantation (LT). After 1 month, 54% patients who met the randomization criteria (among these rejection free for at least 2 weeks and glomerular filtration rate >50 mL/min) were randomized as follows: (a) an experimental arm in which everolimus was started, and CNIs dose was progressively reduced for 8 weeks and then stopped; (b) a control arm in which CNI dosage was maintained. The primary endpoint was renal function at 12 months from LT. However the benefit of using everolimus and discontinuing CNIs was very limited in terms of glomerular filtration rate. The authors concluded that renal sparing strategies using everolimus deserve further investigation in LT patients.
We believe, as do the authors, that minimizing renal impairment is a prime goal of management after LT, since 18% of patients develop chronic renal dysfunction at 5 years post-LT, and this is independently related to death . In a recent systematic review of randomized controlled trials using tacrolimus in LT, we described higher rates of renal dysfunction at 1 year post-LT in most recent trials, clearly influenced by the use of high trough concentrations (TC) of tacrolimus within the first month after LT (>10 ng/mL) . Indeed among six randomized controlled trials including 1594 patients, tacrolimus TC within the first month after LT were positively correlated with renal impairment rates at 1 year (r = 0.73; p = 0.003). Lower TC of tacrolimus (6–10 ng/mL) within this period halved the rate of renal impairment (RR = 0.51 [0.38–0.69]), while acute rejection rates were identical. In the PROTECT trial the target TC of CNIs within the first month after LT (prior to randomization) varied between centers, as no thresholds were set by protocol, and this could have influenced the randomization process. Furthermore the exposure to CNIs was described only by CNI dose which, given the high pharmacokinetic variability, is poorly correlated with TC, and is not related to renal impairment . Thus the lack of both documentation and defined target TC of CNIs, is a crucial limitation, and a potential source of bias in the PROTECT trial. In addition, it may well be that the absence of reduced CNI levels in the first month after LT, and a theoretical imbalance between groups regarding early CNI exposure, could explain why no advantage was seen in terms of renal function in the everolimus arm.
Future randomized trials using CNIs in LT, especially those aiming to lessen renal dysfunction, should fully describe the TC of CNIs used within the first month after LT, either with mean TC or area under curve over time, controlling this potential confounding factor in the analysis. Lower target TC of tacrolimus early after LT (6–10 ng/mL within the first month) are safe in terms of acute rejection, and reduce renal impairment, so this should be included in future trial design .