Inter-Individual Differences in Baseline Coagulation Activities and Their Implications for International Normalized Ratio Control During Warfarin Initiation Therapy

Genetic polymorphisms of cytochrome P450 (CYP) 2C9 (CYP2C9) and vitamin K epoxide reductase complex subunit 1 (VKORC1) and patient demographic characteristics are responsible for inter-individual differences in warfarin maintenance dosage requirements. At present, however, the factors associated with over-anticoagulation responses, especially before achieving the maintenance phase, have not been completely clarified. In this study, we investigated the effects of baseline coagulation activity assessed in terms of the level of fully carboxylated plasma normal prothrombin (NPT) on international normalized ratio (INR) control during the induction phase of warfarin therapy. Our objectives were to (1) identify factors associated with inter-patient variability in baseline NPT (NPT₀); (2) estimate the therapeutic NPT (NPT_tx) levels that can achieve an INR of 2–3; and (3) investigate the influence of NPT₀ on the INR response to warfarin by employing modelling and simulation techniques.

We measured NPT before (NPT₀) and during the introduction of warfarin therapy for up to 3 months and analysed functional single nucleotide polymorphisms (SNPs) of VKORC1 and CYP4F2 in 179 Chinese patients. The patients were classified into tertile groups according to NPT₀ values (i.e. high, intermediate and low groups), and in each group the NPT_tx achieving therapeutic INR, the absolute reduction of NPT from NPT₀ to NPT_tx, and the percentage inhibition of NPT₀ [{(NPT₀ − NPT_tx)/NPT₀} × 100] were obtained. The nonlinear relationship between NPT and INR was modelled on the basis of the INR value before warfarin treatment (INR₀) added by the nonlinear increase in INR after warfarin initiation, which was predicted using the percentage inhibition of NPT₀ and a nonlinear coefficient (λ). The population parameter λ and its inter-individual variability and intra-individual variability in INR in the NPT–INR model were estimated by nonlinear mixed-effect modelling software NONMEM^®.

Multivariate analysis identified age and liver disease as covariates of NPT₀, but none of the SNPs had a significant influence. Although the mean absolute NPT reduction necessary to achieve NPT_tx was dependent on NPT₀ (i.e. the higher the NPT₀, the larger the reduction in NPT), the percentage inhibition was within the narrow range of 67–72 % of NPT₀, irrespective of NPT₀. However, a significantly higher percentage inhibition (80 % on average) was observed in patients with INR values exceeding 4.0. As the nonlinear coefficient λ in the developed model was dependent on NPT₀ (i.e. the higher the NPT₀, the larger the nonlinear λ value), the simulated nonlinear NPT–INR curves were superimposable in the three respective NPT₀ groups, and the only difference was the starting median NPT₀ level. As a result, a steeper increase in the slope of the nonlinear NPT–INR curve might be expected in patients with a lower NPT₀ after initiation of warfarin.

The present study suggests that INR may be prolonged by warfarin nonlinearly as a function of the percentage inhibition of NPT₀. Furthermore, these results indicate that NPT₀ may contribute to inter-individual variability in the INR response, and that patients with low NPT₀ may have the potential to show a sharp increase in INR during initiation therapy with warfarin.