A Bayesian decision support tool for efficient dose individualization of warfarin in adults and children

Abstract

Background: Warfarin is the most widely prescribed anticoagulant for the prevention and treatment of thromboembolic events. Although highly effective, the use of warfarin is limited by a narrow therapeutic range combined with a more than ten-fold difference in the dose required for adequate anticoagulation in adults. An optimal dose that leads to a favourable balance between the wanted antithrombotic effect and the risk of bleeding as measured by the prothrombin time International Normalised Ratio (INR) must be found for each patient. A model describing the time-course of the INR response can be used to aid dose selection before starting therapy (a priori dose prediction) and after therapy has been initiated (a posteriori dose revision).

Results: In this paper we describe a warfarin decision support tool. It was transferred from a population PKPD-model for warfarin developed in NONMEM to a platform independent tool written in Java. The tool proved capable of solving a system of differential equations that represent the pharmacokinetics and pharmacodynamics of warfarin with a performance comparable to NONMEM. To estimate an a priori dose the user enters information on body weight, age, baseline and target INR, and optionally CYP2C9 and VKORC1 genotype. By adding information about previous doses and INR observations, the tool will suggest a new dose a posteriori through Bayesian forecasting. Results are displayed as the predicted dose per day and per week, and graphically as the predicted INR curve. The tool can also be used to predict INR following any given dose regimen, e.g. a fixed or an individualized loading-dose regimen.

Conclusions: We believe that this type of mechanism-based decision support tool could be useful for initiating and maintaining warfarin therapy in the clinic. It will ensure more consistent dose adjustment practices between prescribers, and provide efficient and truly individualized warfarin dosing in both children and adults.

Figure 1

Schematic picture of PKPD-based warfarin model. This is a schematic picture of the basic structure of the published PKPD-model for warfarin. The predictors necessary for individual dose predictions (e.g. CYP2C9 and VKORC1 genotype, age and bodyweight, baseline and target INR) are not included in the picture.

Figure 2

Example of the a priori dose estimation function. This shows an example of an a priori dose estimation for a 5 year old child, with bodyweight 20 kg, genotypes CYP2C9 *2/*2 and VKORC1 A/A, with a target INR of 2.0-3.0 and a baseline INR of 1.2. The estimated maintenance dose is 0.7 mg/24h, or 4.9 mg/week. The graph indicate that with this dose regimen, time to reach a target INR is ~6 days, and time to steady state is ~12 days.

Figure 3

Example of the a priori INR prediction function. This shows an example of an a priori INR prediction for a 20 year old, with bodyweight 75 kg, genotypes CYP2C9 *3/*3 and VKORC1 A/G, with a target INR of 2.0-3.0 and a baseline INR of 1. The predicted INR after a total of 15 doses, including a 3-day loading dose regimen of 7.5 mg, 5 mg and 5 mg (not seen here but defined in the Set doses option) and followed by daily doses of 1.5 mg, is an INR of 2.57. The graph indicate that a target INR is reached after ~3 days with this dose regimen.

Figure 4

Example of the estimation of individual parameters. This provides an example of the output from the estimation of individual model parameters, showing both typical and individual parameter estimates, and the population predicted (black) and the individually predicted (red) INR curves for a given dose history. The individually predicted INR is presented with an optional 90% confidence interval.

Figure 5

Example of the a posteriori dose estimation function. This shows an example of an a posteriori dose estimation for a 1.53 year old child, with bodyweight 20 kg, genotypes CYP2C9 *1/*1 and VKORC1 A/A, and target INR 2.0-3.0 and a baseline INR of 1 using the individual model parameters estimated in Figure 4. The estimated a posteriori dose is 1.08 mg/24 h, or 7.56 mg/week. The graph shows the predicted INR curve after administration of the estimated daily dose (1.08 mg).

Figure 6

Comparison of a priori dose predictions. This figure provides results from a comparison of a priori dose predictions from NONMEM and the Java-based tool. The validation was performed using treatment data from 49 external children, and the results indicated no differences in computational performance between the two methods.

Figure 7

Comparison of individual parameter estimates and a posteriori dose predictions. This figure provides results from comparisons of individual parameter estimates (K₁₀ and EC₅₀) and a posteriori dose predictions from NONMEM and the Java-based tool. The validation was performed using treatment data from 49 external children, and the results indicated minor differences in computational performance between the two methods.