The difficulties in achieving stable international normalised ratios (INRs) are a common concern in clinical practice and have been related to thromboembolic and bleeding risks in a number of clinical trials in patients with atrial fibrillation.1 The paper by Grzymala-Lubanski and colleagues fills an important gap since it provides an in-depth analysis of the clinical impact of warfarin treatment quality in the specific population of patients with mechanical heart valve prostheses, who are at particular risk of thromboembolism and bleeding.2

The time in therapeutic range (TTR) of 72% in the paper by Grzymala-Lubanski and colleagues is consistent with trials performed in atrial fibrillation. However, adherence to treatment is often better in clinical trials than in real life. In addition, this analysis was based on INR management by the Auricula network which uses a computerised dosing system allowing for good INR stability. In Sweden, TTR rates were among the highest in recent international clinical trials on atrial fibrillation. Therefore, it is likely that warfarin treatment quality is less satisfactory in other countries than in Sweden.

The relationship between warfarin treatment quality and outcome has seldom been studied in the specific subgroup of patients with mechanical heart valve prostheses. A study published in 2002 found a strong relationship between poor INR stability and all-cause mortality in 1272 patients and this remained significant in multivariable analysis.3 The strength of the paper by Grzymala-Lubanski and colleagues is to confirm and expand these findings in a much larger patient sample and in a contemporary setting with regard to the type of prosthesis and target INRs. The quality of the SWEDEHEART registry is widely acknowledged and the prospective collection of clinical events through the Auricula network ensures …