Defining a threshold for tacrolimus intra-patient variability associated with late acute cellular rejection in paediatric kidney transplant recipients

Background Late acute cellular rejection (LACR) is associated with poorer graft outcomes and non-adherence. Non-adherence to tacrolimus can be indirectly assessed by the intra-patient variability (IPV) of tacrolimus trough levels. The threshold of IPV associated with rejection is not known. Methods We conducted a case-control study comparing 25 patients with biopsy-proven LACR against 25 stable controls matched for age group, primary diagnosis and time post-transplant. IPV was calculated using coefficient of variance (CV) and mean absolute deviation (MAD) using tacrolimus levels in the preceding 12 months. We also assessed the percentage time for tacrolimus levels < 4 mu g/L (Tac < 4) and the concentration/weight-adjusted dose (C/D) ratio as a proxy marker of tacrolimus metaboliser status. Results LACR patients had higher CV (median, IQR 44%, 36-61% v. 24%, 19-35%, p < 0.0001) and higher MAD (33%, 25-48% v. 19%, 15-26%, p < 0.0001). TheMAD was less affected by outlying tacrolimus results. Receiver operating curve analysis of the MAD resulted in a sensitivity of 76% and specificity of 76% at a threshold of 26% (AUC 0.85, p < 0.05). LACR patients had more Tac < 4 (50% v. 26%, p < 0.05). There was no difference in C/D suggesting that good IPV can be maintained in fast metabolisers. Patients with LACR had significantly increased creatinine at 12-month follow-up despite treatment (108 v. 5 umol/L increase from baseline) and four patients lost their allograft. Conclusions Monitoring of tacrolimus IPV using the MAD may be a clinical marker for LACR. A threshold IPVof 26% can potentially be used as a therapeutic target pending further validation studies.