A placebo-controlled study of interaction between nabumetone and acenocoumarol

Abstract

Aims

The use of nonsteroidal anti-inflammatory drugs (NSAIDs) in patients treated with oral anticoagulants is generally discouraged due to the risk of interactions that could increase the risk of bleeding complications. Available data suggest the NSAID, nabumetone, does not produce such an interaction. We investigated whether nabumetone would interact with acenocoumarol, an oral anticoagulant widely used in some European countries.

Methods

A double-blind, randomized, placebo-controlled study was conducted evaluating nabumetone (1–2 g daily for up to 4 weeks) in osteoarthritis patients with thromboembolic risk previously stabilized on acenocoumarol. The primary efficacy end point was the proportion of patients whose International Normalized Ratio (INR) remained within established margins and whose acenocoumarol dose was not changed. Fifty-six patients were randomized to receive nabumetone (n=27) or placebo (n=29).

Results

Eighteen patients in each group (67% for nabumetone and 62% for placebo) completed the study without showing INR or acenocoumarol dose changes, and were considered as study successes. Nine patients (33%) with nabumetone and 11 (38%) with placebo were considered study failures in the intention-to-treat analysis (one patient on nabumetone and four on placebo did not complete the study due to reasons not related to INR and acenocoumarol dose changes). No significant differences were found between groups with regard to study successes. There were two minor bleeding complications, one in each group. Six patients per group presented with eight adverse experiences in each group.

Conclusions

Treatment with nabumetone did not alter INR levels compared with placebo in patients stabilized on oral acenocoumarol who require NSAID therapy. These results suggest that nabumetone does not produce a clinically relevant interaction with acenocoumarol. In orally anticoagulated patients without other associated risk factors, treatment with nabumetone for up to 4 weeks does not require increased monitoring of INR levels.

Introduction

Maintenance of anticoagulation stability in patients under treatment with oral anticoagulants (OACs) requires close monitoring of coagulation parameters in order to prevent unexpected shifts in anticoagulation and the subsequent risk of thromboembolic or bleeding complications [1]. It is well known that the concomitant use of nonsteroidal anti-inflammatory drugs (NSAIDs) in these patients may modify the coagulation status through various mechanisms: 1) Some NSAIDs enhance the hypoprothrombinaemic effect of OACs through a pharmacokinetic interaction either by displacing OACs from their protein binding sites or by inhibiting their hepatic metabolism [2]; 2) Most NSAIDs inhibit platelet function, thus increasing a patient’s bleeding liability [2]; 3) NSAIDs can induce direct gastrointestinal mucosal damage and bleeding [3]. The global effect of these three mechanisms is an increased risk of bleeding complications when OACs and NSAIDs are taken together.

Several studies suggest that nabumetone, a nonacidic NSAID, does not increase the risk of bleeding through the mentioned mechanisms [2, 4–6]. When nabumetone was given to patients stabilized on chronic warfarin, the International Normalized Ratio (INR) values did not differ significantly from those observed prior to initiating nabumetone use [7]. Similarly, nabumetone did not produce an interaction when given to patients under stable anticoagulation with acenocoumarol an OAC closely related to warfarin [8]: in this study only one of 16 patients showed an increase in the INR value, a proportion similar to that observed in a series of patients treated with acenocoumarol who had not received any concomitant NSAID (A. Fernández-Pavón & A. Pardo, unpublished data).

The objective of this study was to assess whether nabumetone compared with placebo interacts with the OAC acenocoumarol in patients previously stabilized on acenocoumarol.

Methods

This was a prospective, randomized, parallel group, placebo-controlled, double-blind trial carried out at two centres in Spain. The study was approved by the Research Ethics Committees of both participating centres. Male and female patients aged between 18 and 80 years old under stable oral anticoagulation with acenocoumarol were eligible to participate in the study. Stable anticoagulation was defined as no changes in both the INR and acenocoumarol dose in the last three consecutive assessments prior to patient inclusion. Acenocoumarol dose was adjusted for each individual to maintain a level of anticoagulation (INR) within specified margins [9] depending on the particular condition for which it was indicated (Table 1) and/or common practice at the participating centres. In addition, patients were required to have a diagnosis of osteoarthritis requiring treatment with a NSAID. Exclusion criteria included previous treatment with nabumetone, potentially fertile females, allergy to NSAIDs, bleeding episodes or conditions within 3 months prior to inclusion, other uncontrolled serious conditions, concomitant treatment with drugs different from acenocoumarol unless their dose had been stable for at least 3 months prior to inclusion, and concomitant use of other NSAIDs. Prior to beginning their participation in the study, written informed consent was obtained from all patients.

Table 1.

Conditions for which oral anticoagulation was prescribed and recommended INR ranges [9].

Patients were randomly allocated to receive the approved [10] dose of 2–4 tablets of either nabumetone (500 mg tablet⁻¹) or matching placebo once daily for 3–4 weeks. INR values were assessed at baseline and every 7–10 days for 3 follow-up visits. All patients were allowed to use orally administered 575 mg capsules of dipyrone (magnesium metamizole), up to four per day, as analgesic rescue medication. Safety was assessed by routine haematology, biochemistry and urine laboratory tests performed at baseline and end-point visits, and by recording of adverse events at all follow-up visits. Acenocoumarol and nabumetone compliance and concomitant medication use (including dipyrone) were evaluated at all follow-up visits. The primary efficacy endpoint was the proportion of patients whose INR remained within the specified margins and whose acenocoumarol dose remained unchanged during the study and at study end.

Sample size was calculated assuming an expected similar efficacy of 90% in both treatment groups, and a clinically meaningful difference higher than 20% between groups; for an alpha error of 0.05, a power of 80% for a one tail comparison, 28 patients per treatment group were needed [11]. All comparisons were made using the Fisher’s exact test.

Results

Fifty-six Caucasian outpatients (17 male) presenting to the Haematology Departments of two large general hospitals in the Madrid, Spain area, were randomized to receive nabumetone (n=27) or placebo (n=29). No significant differences were found between treatment groups in terms of demographic characteristics and sex distribution (Table 2). The conditions for which acenocoumarol was prescribed are shown in Table 1. Osteoarthitis conditions for which nabumetone was prescribed are shown in Table 2. All randomized patients taking at least one dose of study medication (n=56) were included in the intention-to-treat analysis. One patient in the nabumetone group and four on placebo were withdrawn from the study due to reasons not related to INR and acenocoumarol dose changes.

Table 2.

Demographic characteristics and conditions for which nabumetone was prescribed.

Eighteen patients in each group successfully completed the study without showing INR and acenocoumarol dose changes. Nine patients in the nabumetone group and 11 in the placebo group prematurely terminated their participation and were considered study failures in the analysis (Table 3). Between-groups differences in successes and failures were not statistically significant: 4.6% (95% C.I. −29.7; 20.5).

Table 3.

Patients withdrawn during the study period and successes at study end.

There were no statistically significant differences between the nabumetone and placebo groups, respectively, in the mean number of rescue analgesic medication capsules taken at visit 1 (8.3±8.8 and 9.2±8.3; 95% C.I. −5.55; 3.71), at visit 2 (8.1±10.1 and 10±10.3; 95% C.I. −7.79; 3.87) and at visit 3 (9.3±9.6 and 13.3±12.1; 95% C.I. −10.55; 2.46).

There was not a statistically significant difference either between mean (±s.d.) baseline acenocoumarol dose (mg) in both groups: 1.4 (±2.22) (95% C.I. −5.86; 3.06), or between mean (±s.d.) baseline INR values in both groups: 0.38 (±0.17) (95% C.I. 0.03; 0.073).

Six patients in each group reported eight adverse experiences per group. Only one patient per group was withdrawn due to adverse experiences (a biliary colic and an acute myocardial infarction). There were no clinically significant abnormal laboratory values during the study. There was one minor bleeding complication per group not requiring corrective treatment and not leading to immediate patient’s withdrawal.

Discussion

Nabumetone is a nonacidic NSAID, which unlike other NSAIDs [2], does not inhibit platelet aggregation significantly [12]. In a comparative study between aspirin, nabumetone (1 g day⁻¹), placebo and no treatment, the incidence of gastrointestinal microscopic haemorrhage was not significantly different between the nabumetone, placebo and no treatment groups [5].

Although it has been shown in in vitro studies that, due to its protein binding affinity, 6-MNA, the active metabolite of nabumetone, is able to displace other protein-bound drugs from their binding sites [6], clinical studies have not shown any significant interaction in vivo between warfarin and nabumetone, both in healthy volunteers and in patients [4–7].

The present results support the previous findings concerning the safety of using nabumetone in patients under oral anticoagulation who require the chronic use of a NSAID. However, this conclusion should be taken cautiously due to the relatively short duration of the study (3–4 weeks).

Although previous work on the interaction between nabumetone and oral anticoagulants has generally looked at comparisons between other surrogate endpoints, such as mean INR value changes before and after treatment with nabumetone, or between nabumetone-treated and untreated groups [4, 7, 8], it can be argued that it is clinically more relevant to investigate the actual proportion of individuals that remain stable on their anticoagulation level and that do not require anticoagulant dose adjustments once treated with a NSAID (as compared with placebo). Although the incidence of bleeding episodes is the ultimate outcome measure to look at, ethical reasons preclude such a study design. INR changes, as a surrogate measurement, are good predictors of an increased risk of bleeding. In clinical practice, INR values serve as indicators guiding OAC dose adjustment. The results of this study demonstrate that, compared to patients not treated with a NSAID, treatment with nabumetone for 3–4 weeks in patients previously stabilized with acenocoumarol does not increase the proportion of patients requiring adjustments to acenocoumarol dose. The combination of drugs was well tolerated.

Although the statistical power of this study was lower than planned due to a higher than expected early withdrawal rate, it may be concluded that up to 4 weeks of treatment with nabumetone can be safely given to patients stabilized on oral acenocoumarol. The frequency of monitoring of the anticoagulation level (by INR) does not need to be increased in orally anticoagulated patients treated with nabumetone without other associated risk factors. Further longer term studies are warranted in order to assess whether the concomitant use of nabumetone and acenocoumarol for periods longer than 4 weeks is equally safe.