Bleeding in patients with mechanical heart valves on warfarin: impact of target INR and of comorbidity

Abstract

Dear Sir, Filippatos et al. suggest that our relatively high target for warfarin therapy in our patients with mechanical heart valves, Internalized Normalized Ratio (INR) 2.5–4.2, might explain the high incidence of bleeding [1, 2]. They argue that when INR is out of this target range, bleedings could be more severe, and ask for the last INR values and the time intervals between last control and the bleeding episodes [1]. At the last visit prior to the bleeding, median INR was 3.5 and the two highest INR values were 4.0 and 6.0 in our patients. The time interval between this visit and the admittance with bleeding was 1–2 weeks in four patients, 4–5 weeks in two patients, and excessive (about 8 weeks) in one patient. This patient had the highest INR, 7.5, at admittance. The median INR at admittance was 4.8, and above the therapeutic range in four patients. The mean interval between INR controls was 21 days in the total material, but 24 days in patients with bleeding (the difference was not statistically significant). For patients with venous thromboembolism or atrial fibrillation in our study, the therapeutic range was INR 2.0–3.0; few bleeding episodes occurred and the net gain was probably considerable [2]. These data thus confirm the suggestions of Filippatos et al. that both a relatively high therapeutic INR range, and long intervals between controls, appear to increase the risk of bleeding. Our study concerned patients monitored up to 1996 [2], and we have since then employed a lower therapeutic range in patients with mechanical heart valves (INR 2.5–3.5). Further, we recommend intervals of no more than 3–4 weeks between INR controls. We would like to add a short comment about comorbidity which is known to increase the risk of bleeding [3]. This was probably the case in five of the six mechanical valve patients with bleeding. In three patients with gastrointestinal bleeding, two bled from a malignant tumour and one patient was admitted twice with bleeding from angiodysplasia. One patient had infectious endocarditis with probable Disseminated Intravascular Coagulation (DIC), and one patient was admitted with a traumatic bleeding. As limitations to our study, we had mentioned that the subgroup of patients with mechanical heart valves was small [2]. As several larger studies had reported a much lower incidence of bleeding, we concluded that our results underestimated the value of warfarin therapy in these patients. We appreciate this opportunity of presenting additional data, confirming that both intense warfarin therapy and comorbidity may contribute to bleeding. No conflict of interest was declared.