Bruce NJ, Thomson AH, †Undre N, †Stadler P, Elliott HL.
University Department of Medicine and Therapeutics, Western Infirmary, Glasgow, Scotland. †Fujisawa GmbH, Munich, Germany.
The population pharmacokinetics of the immunosuppressant drug tacrolimus were investigated in 172 patients, aged 18-66 years, following liver transplantation. Tacrolimus was initially administered by intravenous infusion at doses of 0.2-7mg/day then given orally at doses of 1-20mg twice daily. Trough concentrations were measured in whole blood from 1 to 382 days after transplantation (median 62 days). Before analysis, patients were randomly allocated to either the population data set (115 patients) or the îtest” data set (57 patients).
The population data set comprised 905 concentration measurements and was analysed using NONMEM. Absorption was assumed to be first-order and the absorption rate constant was fixed at 0.5h-1. Elimination was satisfactorily described using a monoexponential decline. The following clinical factors were tested for their influence on clearance (CL), volume of distribution (V), and bioavailability (F): sex; age; height; weight; body mass index; body surface area; day; indices of hepatic function (concentrations of protein, albumin, bilirubin, alkaline phosphatase (ALKP), alanine amino transferase (ALT), aspartate amino transferase (AST) and glutamyl amino transferase (GGT)); indices of renal function (creatinine concentration, estimated creatinine clearance); and haematocrit. The influence of these covariates was investigated using linear, non-linear and step models. No graphical evidence of covariate relationships could be seen, either with the population data set or by analysing the concentrations as separate individuals. Model comparisons were therefore based on the change in objective function value (chi-squared test, p<0.01 significant), estimates of variability and standard errors of the parameter estimates.
Clearance was reduced in patients with evidence of liver impairment, as indicated by high concentrations of AST, ALT and bilirubin. The best model was an exponential function of clearance including AST and bilirubin. No relationships were identified between any of the covariates and volume of distribution, or bioavailability. The range of parameter estimates determined using the full model and the simplest model are summarised in Table 1.
Table 1
| Model | CL (cv) | V (cv) | F (cv) |
|
| Lh-1 | L | % |
| Simplest model | 2.83 (61%) | 114 (76%) | 7.2 (65%) |
| Full model | 1.0-4.1 (54%) | 104 (63%) | 9.61 (57%) |
These population derived results were evaluated using the îtest” data set which comprised 440 concentrations. The predicted distribution of each concentration (mean ±(SD) was determined using the parameter estimates obtained in the population analysis. For each individual the percentage of measured concentrations that lay within this distribution was calculated. Eighty-six percent of the measured concentrations for the group as a whole were within their predicted distributions but there was a small bias in the mean predicted concentration (2.8ng/ml (range -13-75.6 ng/ml)).
This analysis has identified significant relationships in liver transplant patients between the pharmacokinetics of tacrolimus and indices of hepatic function.
Reference: PAGE 5 () Abstr 555 [www.page-meeting.org/?abstract=555]
Poster: poster