Comparison of INR stability between self-monitoring and standard laboratory method: Preliminary results of a prospective study in 67 mechanical heart valve patients

Summary Introduction Thromboembolic accidents and haemorrhage are the main complications observed during long-term follow-up of mechanical heart valve patients. Several suggestions for improving anticoagulation quality have been made, including international normalised ratio (INR) self-monitoring. Objectives We report the preliminary results of a single-centre, open, randomised study (scheduled population of 200 patients), which compares monthly laboratory monitoring (group A) versus weekly self-monitoring of INR (group B). The primary aim is INR stability improvement within the target range, and the secondary aim is adverse events reduction. Patients and methods Between May 2004 and June 2005, 67 patients with an average age of 56.6 years (± 9.6), were enrolled in the study (group A: 34 patients, group B: 33 patients). The mean follow-up was 47 weeks (± 11.5). The two groups differed only in the sex ratio (44.1 and 21.2% of women in groups A and B respectively, p = 0.0459). Mechanical heart valves were aortic in 73% of patients, mitral in 13.5%, and multiple in 13.5%. Sixty-five patients (97%) were treated with fluindione, the others with acenocoumarol. The intraclass correlation coefficient between the self- and laboratory-monitored INR was 0.75. Results The time spent in the INR target range (group A: 53 ± 19%, group B: 57 ± −19%, p = 0.45) and the time spent in the INR therapeutic range, between 2 and 4.5, (group A: 86 ± 14%, group B: 91 ± 7%, p = 0.07) are longer in group B, but not significantly so. For patients outside the range, the absolute mean deviation of INR from the target or therapeutic range (range standardized between 0 and 100) is lower for the self-monitoring group (41.1 ± 39.3 and 11.27 ± 11.2) than for the control group (62.4 ± 72.6 and 39.2 ± 52.8). This difference is significant ( p = 0.0004 and p = 0.0005). Eighteen adverse events were reported: 17 haemorrhages, 13 in group A (9 mild, 4 serious) and four in group B (all mild), and one sudden death in group B, two days after the patient's discharge. No thromboembolic events were reported. Six patients (8.8 %), 3 in each group, dropped out of the study. Conclusion This first study evaluating INR self-monitoring in France shows that this method leads to better stability of the INR within the target range. On the basis of these preliminary data, this appears to be related to a decrease in serious haemorrhages (11.8% serious haemorrhage cases in group A versus 0% in group B, p = 0.06, NS