Efficacy and Safety of Apixaban Versus Warfarin in Patients With Atrial Fibrillation and Extremes in Body Weight

Introduction

Editorial, see p 2301

Obesity is a comorbid condition of increasing prevalence, particularly in Western societies. In the United States, the prevalence rates of obesity and extreme obesity are 35% and 6.4%, respectively.¹ Obesity is an important risk factor for the development and maintenance of atrial fibrillation (AF).² Current guidelines recommend the use of fixed-dose non–vitamin K antagonist oral anticoagulants (NOACs) for stroke prevention in patients with AF, with some guidance on dose reduction for patients with multiple bleeding risk factors.^3,4 However, the currently available evidence on the use of NOACs in patients with low or high body weight is limited, and data on the efficacy and safety of these compounds in patients with extremes of weight are sparse. A major concern with the use of NOACs in patients with extremes of weight is that fixed dosing schedules in individuals above the average weight of those included in the clinical trials may result in unintentional over- or underdosing, with the consequence of enhanced risk of thromboembolic or bleeding complications. In fact, recent International Society on Thrombosis and Haemostasis (ISTH) guidelines caution against the use of NOACs in patients with extremely high weight (>120 kg) without clear guidance on use NOACs in patients with low (≤60 kg) body weight.⁵ Based on this concern and the lack of respective data on this important topic, we conducted a post hoc analysis of the ARISTOTLE (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) trial, which randomized patients with AF to apixaban or warfarin therapy. The primary aim of our analysis was to determine the efficacy and safety of NOACs for stroke prevention in patients with AF randomized to apixaban versus warfarin according to body weight.

Methods

Study Population

The data, analytic methods, and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedure. The ARISTOTLE trial (ClinicalTrials.gov NCT00412984) design and results have been previously published.^4,6 Briefly, ARISTOTLE was a double-blind, double dummy, randomized clinical trial of 18 201 patients with AF and at least 1 additional risk factor for stroke or systemic embolism. Risk factors included age ≥75 years; prior stroke, transient ischemic attack, or systemic embolism; heart failure or reduced left ventricular ejection fraction; diabetes mellitus; or hypertension requiring pharmacological treatment. Patients were randomized to treatment with 5 mg of apixaban twice daily (n=9120) or dose-adjusted warfarin (n=9081) with a target international normalized ratio of 2.0 to 3.0. Patients received a reduced dose of 2.5 mg of apixaban twice daily if they had ≥2 of the following criteria: age ≥80 years, body weight ≤60 kg, or a serum creatinine level ≥1.5 mg/dL. For the present analyses, we included patients with available baseline measurements of weight.

Institutional review boards at participating centers reviewed and approved the study protocol, and all patients provided written informed consent.

Weight Groups

Weight was assessed at baseline in kilograms. We analyzed weight as a continuous variable, and patients were divided into 3 main groups based on drug dosing criteria (dose reduction ≤60 kg) and ISTH recommendations (no NOAC >120 kg)⁵: low weight (≤60 kg), midrange weight (>60–120 kg), and high weight (>120 kg). For an additional sensitivity analysis, the high-weight group was further subdivided into patients of 121 to 140 kg and >140 kg.

Outcomes

The efficacy outcomes were stroke (ischemic, hemorrhagic, or uncertain type) or systemic embolism, death from any cause, or myocardial infarction at 2 years of follow-up. Safety outcomes included ISTH major bleeding, ISTH clinically relevant nonmajor bleeding, intracranial bleeding, gastrointestinal bleeding, and any bleeding at 2 years of follow-up.⁶ An independent clinical events committee whose members were unaware of treatment assignment adjudicated all outcomes according to prespecified criteria.

Statistical Analysis

Baseline categorical variables are presented as counts, and percentages and continuous variables are presented as medians (25th, 75th percentiles). The characteristics were compared across categories of weight using the χ² test and Kruskal-Wallis test for continuous variables. Cox regression models were used to estimate the HRs and 95% CIs comparing apixaban with warfarin across body weight categories. Analyses of bleeding outcomes included only events occurring during study drug treatment, whereas stroke or systemic embolism and other efficacy outcomes were analyzed using the intention-to-treat principle. Restricted cubic splines were used to allow for nonlinear relationships between body weight and outcomes. Treatment effects were compared in separate models according to weight (both as a categorical variable and continuous), by adding interaction terms to models including body weight and treatment allocation. All analyses were performed using SAS software, version 9.4 (SAS Institute Inc., Cary, NC). A 2-sided P value of <0.05 was considered statistically significant.

Results

Baseline Characteristics

A total of 18 139 patients (99.7% of the entire population) with known baseline body weight at randomization were included in the present analysis and categorized into 3 groups according to body weight: 1985 (10.9%) patients were in the low-weight group (≤60 kg), 15 172 (83.6%) were in the midrange weight group (>60–120 kg), and 982 (5.4%) were in the high-weight group (>120 kg; Figure 1). Throughout study follow-up, there were only small changes in weight in each of the individual weight groups (Table 1). Baseline characteristics across groups of weight are shown in Table 2. Compared with the midrange and low-weight groups, patients in the high-weight group were younger, more likely to be male, and more likely to be randomized in North America. Patients in the low-weight group were more likely to be female and of Asian or Latin American ethnicity; have a history of stroke, transient ischemic attack, or systemic embolism; have abnormal renal function; have higher N-terminal probrain natriuretic peptide levels; and have a CHA₂DS₂-VASc (congestive heart failure, hypertension, age ≥75 yr, diabetes mellitus, stroke or transient ischemic attack or thromboembolism, vascular disease, age 65–74, sex category [female]) score of ≥2. Patients in the low-weight group were also more likely to be on a reduced dose of apixaban (27.2% of cases) when compared with those in the midrange or high-weight groups (1.9% and 0.3%, respectively). There were significant differences in concomitant medications among the 3 groups (Table 2).

• Download figure
• Download PowerPoint

Treatment Effect of Apixaban Versus Warfarin According to Weight

During a median follow-up of 21.8 months (25th, 75th percentiles [16.3, 28.1 months]), there were 475 stroke or systemic embolic events, 1268 all-cause mortality events, and 786 major bleeding events. The distribution of events by body weight group and treatment assignment is presented in Table 3. The relationship between efficacy and safety outcomes and weight as a continuous variable is presented in Figure 2. Overall, patients treated with apixaban had lower rates of both the efficacy and safety outcomes across most of the weight spectrum compared with patients treated with warfarin. The risk of stroke/systemic embolism, all-cause death, and ISTH major bleeding was higher in patients with lower weight at randomization; however, there was no significant interaction between study treatment (apixaban versus warfarin) and weight on efficacy outcomes such as stroke/systemic embolism, all-cause death, or myocardial infarction. There was a significant interaction between weight and study treatment for the outcome of ISTH major bleeding, suggesting that warfarin was associated with a higher risk of major bleeding in the lower weight range when compared with the higher weight categories; this was not seen with apixaban use.

• Download figure
• Download PowerPoint

Efficacy and safety outcomes were further stratified by the specific weight categories (Table 3; Figure I in the online-only Data Supplement) to highlight event rates and interaction testing among the 3 clinically relevant weight groups. The hazards for stroke/systemic embolism in the low-weight group (≤60 kg) were HR, 0.63; 95% CI, 0.41–0.96, those in the midrange weight group (>60–120 kg) were HR, 0.85; 95% CI, 0.70–1.05, and those in the high-weight group (≥120 kg) were HR, 0.39; 95% CI, 0.12–1.22 (interaction P value=0.640). Among all efficacy end points, a significant interaction between weight and randomized treatment was only found for hemorrhagic stroke (interaction P value=0.042), suggesting that only for this efficacy end point, the relative effectiveness of apixaban versus warfarin was modified by differences in weight. Patients in the low-weight group (≤60 kg) who were on apixaban were at a lower risk of hemorrhagic stroke (HR, 0.16; 95% CI, 0.05–0.54) during follow-up compared with patients on warfarin. In patients in the midrange weight group, the risk of hemorrhagic stroke for apixaban compared with warfarin was also lower (HR, 0.64; 95% CI, 0.42–0.97). In the high-weight group (>120 kg), there were no hemorrhagic strokes in the apixaban arm and 4 in the warfarin arm.

Apixaban was associated with lower event rates for all safety end points when compared with warfarin (Table 3). In the case of ISTH major bleeding, the weight groups were associated with a differential risk. Apixaban had a better safety profile than warfarin in patients in the low (≤60 kg; HR, 0.55; 95% CI, 0.36–0.82) and midrange weight (>60–120 kg) groups (HR, 0.71; 95% CI, 0.61–0.83; interaction P value=0.016). However, in patients in the high-weight group (≥120 kg), despite an absolute lower risk of major bleeding, the relative risk between apixaban and warfarin was comparable (HR, 0.47; 95% CI, 0.37–1.50). Patients randomized to receive apixaban were at a lower risk of major or clinically relevant nonmajor bleeding across all 3 weight groups; however, the effect size was largest in the low-weight (≤60 kg; HR, 0.51; 95% CI, 0.37–0.71) and high-weight (≥120 kg) groups (HR, 0.58; 95% CI, 0.35–0.95), when compared with the midrange (>60–120 kg) group (HR, 0.71; 95% CI, 0.63–0.79; interaction P value=0.011). For all other safety outcomes (intracranial, gastrointestinal, or any bleeding), there were no significant interactions between apixaban and warfarin treatment effects and weight.

Treatment Effect of Apixaban Versus Warfarin in the High-Weight Group

In the high-weight category (>120 kg), there was no significant difference in the occurrence of efficacy outcomes associated with either apixaban or warfarin, but there was a lower risk of major and clinically relevant nonmajor bleeding (HR, 0.58; 95% CI, 0.35–0.95) or any bleeding with apixaban (HR, 0.67; 95% CI, 0.53–0.85) as compared with warfarin. The high-weight group was further stratified into 2 groups: 121 to 140 kg (n=724 [4%]) and >140 kg (n=258 [1.4%]). Baseline characteristics for the respective groups can be found in Table I in the online-only Data Supplement. The randomized treatment effect by body weight for the high-weight groups is shown in Table 4. The risk of stroke or systemic embolism associated with apixaban was lower than for warfarin in patients weighing 121 to 140 kg (HR, 0.21; 95% CI, 0.05–0.95). In the >140 kg patient cohort, there were only 2 events in the apixaban group and 1 in the warfarin group, making it impossible to compare the 2 groups. There was no difference between apixaban and warfarin in relation to all-cause death for either of the weight groups (121–140 kg: HR, 1.23; 95% CI, 0.64–2.35; and >140 kg: HR, 1.14; 95% CI, 0.43–3.04). Patients in the 121 to 140 kg weight group randomized to apixaban were at comparable risk for major bleeding as those receiving warfarin (HR, 0.99; 95% CI, 0.46–2.10) but tended to have lower rates of any bleeding (HR, 0.61; 95% CI, 0.46–0.80). In the >140 kg weight group, there were no stroke/systemic embolic events in those receiving apixaban, but there were 5 stroke/systemic embolic events in those receiving warfarin. All-cause bleeding was comparable between both treatment arms (HR, 0.89; 95% CI, 0.57–1.39).

Discussion

The present post hoc analysis of the ARISTOTLE trial represents the largest study in patients with AF evaluating the efficacy and safety of a NOAC in relation to body weight. The main finding is that apixaban is safe and efficacious in patients with AF with extremes of body weight. There was no interaction showing any attenuation of the treatment effect of apixaban compared with warfarin in extreme body weights.

Current guidelines and product labeling provide only limited guidance for NOAC use and dosing in patients with low body weight but none for the high body weight population. In fact, use of NOACs in the high-weight population is cautioned against.⁵ To date there are no large randomized trials that have tested the efficacy and safety of NOACs in the low-weight or high-weight population. Given the absence of randomized trial data, one can gain some insight from the available pharmacodynamic and pharmacokinetic data for NOAC use. In a small study comprising 54 healthy volunteers (18 with body weight ≤50 kg, 18 with body weight 65–85 kg, and 19 body weight ≥120 kg), Upreti et al.⁷ determined both plasma and urine levels of apixaban and measured antifactor Xa activity following administration of a single 10-mg dose of apixaban. Maximum blood levels of the drug and areas under the concentration versus time curve were found to be 31% (90% CI, 18% to 41%) and 23% (90% CI, 9% to 35%) lower in the high-weight patients. On the other hand, the time to maximum concentration of apixaban in the high-weight group was similar to that in the reference group (65–85 kg) and amounted to 3.9 h (1.0–6.0). The authors concluded from their findings that adjustment of the apixaban dose based on body weight is not necessary.

Notably, our post hoc analysis supports previous retrospective subgroup analyses of other smaller NOAC trials, which demonstrated that the use of NOACs (apixaban, dabigatran, edoxaban, and rivaroxaban) in the high-weight population appears to be efficacious and safe when compared with warfarin, irrespective of the underlying condition (AF or venous thromboembolism).^8–12 These findings are also supported by meta-analyses of major NOAC trials in AF and venous thromboembolism in the high-weight groups.^13–15 The meta-analysis by Proietti et al¹⁴ suggests a potential relationship between weight and major adverse outcomes in patients with AF in randomized, controlled trials, as well as observational cohort studies. Investigators reported a lower risk for stroke/systemic embolic events (OR, 0.75; 95% CI, 0.66–0.84 and OR, 0.62; 95% CI, 0.54–0.70, respectively) in obese patients treated with NOACs.¹⁴ For major bleeding, only obese patients were at lower risk compared with normal weight patients (OR, 0.84; 95% CI, 0.72–0.98). Proietti et al¹⁴ concluded that there might be an obesity paradox in patients with AF, particularly for all-cause and cardiovascular death but also for stroke/systemic embolic events because it was also seen in the EORP-AF General Pilot Registry (EURObservational Research Programme Atrial Fibrillation).¹⁶ Interestingly, a treatment benefit favoring NOACs over warfarin for both efficacy and safety was seen only in normal weight but not in overweight/obese patients. Evidence for a potential obesity paradox has also been reported for apixaban utilizing data from the ARISTOTLE trial.¹⁷ Here, higher body mass index was associated with a lower risk of the composite end point of all-cause mortality, stroke, systemic embolism, and myocardial infarction compared with a normal body mass index.¹⁷ There was no significant association between obesity and major bleeding. Although our analysis of the ARISTOTLE trial confirms the preserved treatment effects of apixaban in patients with high body weight, we are the first to present the findings for patients with extremely high body weight (121–140 kg and >140 kg). In both the 121 to 140 kg and >140 kg groups, the benefits of apixaban over warfarin in reducing stroke, death, and bleeding were preserved.

In the ARISTOTLE trial dose-reduction criteria were applied to patients who had at least 2 high risk criteria (age ≥80 years, body weight ≤60 kg, or a serum creatinine level ≥1.5 mg/dL). These criteria were based on pharmacokinetic modeling,^7,18,19 and a reduction of the apixaban dose from 5 mg to 2.5 mg was intended to avoid increased blood drug levels and thus decrease the risk of bleeding. In the ARISTOTLE trial, 831 patients (4.6%) had 2 or 3 dose-reduction criteria and were receiving the reduced dose of apixaban. The safety and effectiveness of the reduced dose of apixaban appeared to be unchanged from the standard dose of apixaban in patients with ≤1 high risk criteria.²⁰ Further, 5 mg of apixaban twice daily appeared to be safe and efficacious in patients with only 1 dose-reduction criterion, despite a higher risk of stroke or systemic embolism and major bleeding in this population. Our analysis confirms and extends the findings by Alexander et al²⁰ because it shows that the effectiveness of apixaban is consistent across weight categories. Importantly, patients treated with apixaban (versus warfarin) with weight in the low to midrange had an even greater reduction in major bleeding when compared with patients with higher weight (>120 kg; interaction P value=0.016). Finally, patients in the predefined low-weight category of ≤60 kg, of whom only 27% were on a reduced dose of apixaban, were found to have better efficacy and safety outcomes with apixaban as compared with warfarin.

Limitations

The present findings are based on results obtained from a post hoc analysis utilizing weight measurements at baseline. Subgroup analyses are subject to many limitations because of their retrospective nature. Furthermore, the relationship between changes in weight over time and outcomes cannot be assessed, although changes in weight throughout follow-up were small. Although residual confounding can never be excluded, this was a randomized clinical trial, and comparisons were only made between randomized groups (warfarin versus apixaban), thus minimizing the risk of confounding. Finally, our results cannot necessarily be generalized to the general population given differences in our trial population from the general population.

Conclusions

The data obtained from the large ARISTOTLE trial provided evidence that apixaban is efficacious and safe across the spectrum of weight, including low- (≤60 kg) and high-weight patients (>120 kg). The superiority on efficacy and safety outcomes of apixaban compared with warfarin persists across weight groups, with even greater reduction in major bleeding in patients with AF with low to normal weight as compared with high weight.

Footnotes