Tacrolimus dose requirements in paediatric renal allograft recipients are characterized by a biphasic course determined by age and bone maturation

Abstract

Aims: Despite longstanding recognition of significant age-dependent differences in drug disposition during childhood, the exact course and the underlying mechanisms are not known. Our aim was to determine the course and determinants of individual relative dose requirements, during long-term follow-up in children on tacrolimus.

Methods: This was a cohort study in a tertiary hospital with standardized annual pharmacokinetic (PK) follow-up (AUC_0-12hr ) in recipients of a renal allograft (≤19 years), between 1998 and 2015. In addition, the presence of relevant pharmacogenetic variants was determined. The evolution of dose-corrected exposure was evaluated using mixed models.

Results: A total of 184 PK visits by 43 children were included in the study (median age: 14.6). AUC_0-12h corrected for dose per kg demonstrated a biphasic course: annual increase 4.4% (CI: 0.3-8.7%) until ±14 years of age, followed by 13.4% increase (CI 8.7-18.3%). Moreover, exposure corrected for dose per m² proved stable until 14 years (+0.8% annually; CI: -3.0 to +4.8%), followed by a steep increase ≥14 years (+11%; CI: 7.0-16.0%). Analysis according to bone maturation instead of age demonstrated a similar course with a distinct divergence at TW2: 800 (P = 0.01). Genetic variation in CYP3A4, CYP3A5, and CYP3A7 was associated with altered dose requirements, independent of age.

Conclusions: Children exhibit a biphasic course in tacrolimus disposition characterized by a high and stable drug clearance until a specific phase in pubertal development (TW2: 800 at age: ±14 years), followed by an important decline in relative dose requirements thereafter. Pharmacogenetic variation demonstrated an age/puberty independent effect. We suggest a critical reappraisal of current paediatric dosing algorithms for tacrolimus and drugs with a similar disposition.

Keywords: children; drug disposition; ontogeny; pharmacokinetics; tacrolimus.

Figure 1

Subject‐specific profiles of absolute tacrolimus exposure (AUC_0–12h in ng h ml⁻¹) for ‘time after transplantation’ (A), and ‘age’ (B) in years. Each dot represents a visit with an AUC_0–12h assessment. Consecutive visits from the same individual are connected with lines

Figure 2

The relation of age (years) with tacrolimus exposure corrected for dose per kg bodyweight (AUC_0−12h/doseBW (ng h ml⁻¹/mg kg⁻¹)) depicted in the upper panels (A, B, C) versus exposure corrected for dose per m² body surface area (AUC_0−12h/doseBSA (ng h ml⁻¹/mg m⁻²)) in the lower panels (D, E, F). Panel A and D represent subject‐specific profiles of relative tacrolimus exposure corrected for dose. Each dot represents a visit with an AUC_0−12h assessment. Consecutive visits from the same individual are connected with lines. Panel B and E represent a plot based upon the mixed model for the relation of relative tacrolimus exposure corrected for dose with age, allowing non‐linearity in the relation. The dashed straight green line refers to the model assuming linearity. Panel C and F represent a plot based upon the segmented regression model with two linear phases for the predicted relation of relative tacrolimus exposure corrected for dose with age, and a cut off point at 14 years (dashed lines referring to the pointwise 95 percent confidence interval)

Figure 3

Segmented regression model with two linear phases for the relation of age (years) and genotype with tacrolimus exposure corrected for dose per kg bodyweight (ng h ml⁻¹/mg kg⁻¹) depicted in the upper panels (A, B, C) versus exposure corrected for dose per m² body surface area (ng h ml⁻¹/mg m⁻²) in the lower panels (D, E, F). Panel A and D illustrate the genotypic variation for CYP3A4*1B, panel B and E for CYP3A5 and panel C and F for CYP3A7. A line with a cut off point at 14 years was fitted for each genotypic variant (dashed lines referring to the pointwise 95 percent confidence interval). Note that for better visual resolution the data on the Y‐axis were log‐transformed

Figure 4

The relation of bone maturation (TW2 score) with tacrolimus exposure corrected for dose per kg bodyweight (ng h ml⁻¹/mg kg⁻¹) depicted in the upper panels (A, B, C) versus exposure corrected for dose per m² body surface area (ng h ml⁻¹/mg m⁻²) in the lower panels (D, E, F). Panel A and D represent subject‐specific profiles of relative tacrolimus exposure corrected for dose. Each dot represents a visit with an AUC_0−12h assessment. Consecutive visits from the same individual are connected with lines. Panel B and E represent a plot based upon the mixed model for the relation of relative tacrolimus exposure corrected for dose with TW2 score, allowing non‐linearity in the relation. The dashed straight green line refers to the model assuming linearity. Panel C and F represent a plot based upon the segmented regression model with two linear phases for the predicted relation of relative tacrolimus exposure corrected for dose with TW2 score and a cut off point at 800 (dashed lines referring to the pointwise 95 percent confidence interval)