Multicenter evaluation of six methods for predicting warfarin maintenance-dose requirements from initial response

Abstract

Warfarin maintenance-dose requirements predicted by six mathematical methods based on initial response to therapy were compared with patients' actual dose requirements in a multicenter trial. Data were collected for patients who had received an initial regimen of warfarin sodium 10 mg orally every 24 hours for three days and for whom a prothrombin time (PT) had been determined 16-20 hours after the third dose. Patients' individual dose requirements and PT values were recorded during one to three follow-up visits after discharge from one of seven medical centers. Prothrombin ratios (patient PT divided by control PT) calculated on day 4 were used for maintenance-dose prediction by five methods; a sixth method was based on the cumulative warfarin dose-PT response curve up to a PT value of 20 seconds. For 54 men and 30 women who qualified for the study, 197 maintenance-dose-PT response measurements were recorded; 95 in the first four weeks of therapy, 76 during weeks 5-12, and 26 at 6-12 months after initial treatment. Prothrombin ratios were within the therapeutic range (PT 1.5-2.5 times the control value) in 154 observations, and the mean actual warfarin sodium maintenance dose associated with therapeutic response was 7.6 mg/day. For patients with therapeutic prothrombin ratios, dose predictions by the five methods using prothrombin ratios (PRs) correlated significantly with actual dose requirements. The formula that predicted doses numerically closest to the actual dose is as follows: Dose = 11.17 - 21.08 (log PR). Only 45 observations were obtained for the sixth prediction method, and the correlation between actual and predicted doses was not significant. Initial warfarin maintenance-dose requirements can be predicted effectively based on one PT determination after administration of three daily 10-mg doses.(ABSTRACT TRUNCATED AT 250 WORDS)