Background: Measuring immune function in kidney transplant recipients (KTRs) remains challenging given the precarious balance between over- and underimmunosuppression. This is primarily due to high inter- and intrapatient variability in pharmacokinetics and -dynamics of both tacrolimus (TAC) and mycophenolic acid (MPA) and their narrow therapeutic window. Torque teno virus (TTV), a non-pathogenic commensal virus, has been proposed as marker of immune function: high loads may correspond to over-immunosuppression, and low loads to under-immunosuppression. An inverse relation between TTV load and acute rejection has already been shown. This study aimed to investigate the effect of TAC and MPA exposure on TTV loads in KTRs.

Methods: A cross-sectional cohort study was designed using a unique database of drug exposure data. KTRs transplanted between 2005-2012 and started on TAC/MPA/prednisolone were included. Patients without exposure data, without available samples for TTV load measurement or switches in therapy were excluded, leaving 170 KTR for analysis on Month 3 and 159 on Month 6. Linear regression was used to study the association between TTV loads and TAC through (C0) levels, and TAC and MPA area-under-the-curve (AUC) data, measured on the day of, and 2 weeks before TTV load measurement. Analyses were adjusted for donor and recipient age and gender, number of human leukocyte antigen (HLA) mismatch, days on dialysis, and induction therapy.

Results: Linear regression showed an increase in predicted TTV load at month 3 with TAC C0 and AUC increase in the univariate analysis (C0: β=0.20 per 1 μg/L, p=0.01, R2=0.04); AUC: β=0.18 per 20 mg*h/L, p=0.05, R2=0.04), but not in the analysis adjusted for confounders (C0: β=0.16 per 1 μg/L, p=0.15, R2=0.08); AUC: β=0.01 per 20 mg*h/L, p=0.17, R2=0.08) for exposure samples taken on the same day as the TTV sample. Similar results were found when using samples taken 2 weeks before. For MPA no association was found. Analyses on month 6 showed similar results.

Conclusions: An association between tacrolimus and mycophenolic acid levels and TTV levels was not found in the present study. The low R² values and absence of an association in the adjusted analyses indicate that the high variation in TTV loads is not explained in a linear relationship with drug exposure levels, nor the known covariates. Future study will investigate different timeframes, non-linear relationships and drug exposure and TTV trajectories over time.