We attempted to assess compliance using both a pharmacologic indicator (low-dose phenobarbital) and a return tablet count in 225 patients who were taking part in three separate studies. There were 216 patients (96%) who kept a follow-up appointment after 28 days; 161 patients appeared to have good compliance (90% to 109%) by return tablet count. Of these 161 patients, 51 (32%) had plasma phenobarbital concentrations (corrected for dose and weight) that were less than 90% of the lowest value previously found in normal volunteers, which suggested poorer compliance. When compared with the age-related volunteer values, 77 (48%) had values that were less than 90% of the lowest volunteer value. There were 6 of 10 patients with apparently excessive (greater than or equal to 110%) compliance by return tablet count and 4 of 12 who failed to return their container who also had phenobarbital concentrations that were less than 90% of the lowest volunteer value. We concluded that return tablet count grossly overestimates compliance.

SummaryControl of oral anticoagulant therapy in outpatients is often unsatisfactory. The contribution of poor compliance with prescribed warfarin to unstable anticoagulant control was investigated prospectively using low-dose phenobarbitone as an indicator of compliance in 30 out-patients, 15 with stable and 15 with unstable control. Following entry to the study, there was no significant change in anticoagulation (p = 0.36) in the group with stable control. In the group who previously had unstable control, there was a significant change in INR (p = 0.0045) and anticoagulant control greatly improved. It appears that the considerable fluctuation in INR seen in many of the latter patients before the study was due to poor compliance and that entering them into the study modified their behaviour. Two patients in this group who continued to have unstable anticoagulant control were shown to be poorly compliant using the phenobarbitone indicator. The results suggest that, in outpatients, poor compliance is the major cause of unstable anticoagulation with warfarin.

Summary There is a discrepancy in the results of reported studies of levels of vitamin K dependent coagulation factors in patients on warfarin therapy. This may have arisen partly because of the problem of assuring compliance with therapy in outpatients. The plasma concentrations of the vitamin K dependent clotting factors II, VII, IX and X were studied in 23 outpatients whose adherence to prescribed warfarin therapy was determined using a pharmacological indicator of complicance. In these patients, who were shown to have consistently good compliance and stable anticoagulant control over a period of 3–6 months, the activities in plasma of the four coagulation factors were not equally suppressed. Factor IX levels were significantly greater than those of factor VII (P<0.0001) which in turn were significantly greater than the levels of factor II (P<0.0001) or factor X (P<0.0001). There was no significant difference between the levels of factors II and X which were depressed to a similar extent. The proportion of variability of the International Normalized Ratio (INR) explained by linear regression was 51–77% and a model was derived to predict the INR from the mean of the levels of the four clotting factors. The concentrations of the coagulation factors II, VII, IX and X are likely to be highly dependent on the degree of compliance with warfarin therapy which should be taken into account when investigating the behaviour of these factors.

No abstract

There is a discrepancy in the results of reported studies of levels of vitamin K dependent coagulation factors in patients on warfarin therapy. This may have arisen partly because of the problem of assuring compliance with therapy in outpatients. The plasma concentrations of the vitamin K dependent clotting factors II, VII, IX and X were studied in 23 outpatients whose adherence to prescribed warfarin therapy was determined using a pharmacological indicator of compliance. In these patients, who were shown to have consistently good compliance and stable anticoagulant control over a period of 3-6 months, the activities in plasma of the four coagulation factors were not equally suppressed. Factor IX levels were significantly greater than those of factor VII (P less than 0.0001) which in turn were significantly greater than the levels of factor II (P less than 0.0001) or factor X (P less than 0.0001). There was no significant difference between the levels of factors II and X which were depressed to a similar extent. The proportion of variability of the International Normalized Ratio (INR) explained by linear regression was 51-77% and a model was derived to predict the INR from the mean of the levels of the four clotting factors. The concentrations of the coagulation factors II, VII, IX and X are likely to be highly dependent on the degree of compliance with warfarin therapy which should be taken into account when investigating the behaviour of these factors.