Real-world studies have evaluated the use of anticoagulants in obese patients with nonvalvular atrial fibrillation (NVAF), but they have been limited by sample size or the use of diagnosis codes on claims to define obesity. This retrospective study used body weight data of ≥100 kg or a body mass index of ≥30 kg/m 2 to identify elderly (aged ≥65 years) NVAF patients with obesity in dually enrolled Veterans Affairs and fee-for-service Medicare patients. It evaluated the risk of stroke/systemic embolism (SE) and major bleeding (MB) in patients that initiated apixaban versus warfarin. Stabilized inverse probability treatment weighting was used to balance the baseline characteristics between patients prescribed apixaban and warfarin in obese patients. Cox models were used to evaluate the relative risk of stroke/SE and MB. Overall, 35.9% (n = 26,522) of the NVAF population were obese, of which 13,604 apixaban and 12,918 warfarin patients were included. After inverse probability treatment weighting, patient characteristics were balanced. The mean age was 75 years, the mean CHA 2 DS 2 -VASc score (Congestive Heart Failure, Hypertension, Age ≥75 [Doubled], Diabetes Mellitus, Prior Stroke or Transient Ischemic Attack [Doubled], Vascular Disease, Age 65–74, Female) was 3.8, the mean HAS-BLED (Hypertension, Abnormal Renal/Liver Function, Stroke, Bleeding History or Predisposition, Labile INR, Elderly, Drugs/Alcohol Concomitantly) Score was ∼2.6, and >98% of patients were males. Obese apixaban patients were associated with a similar risk of stroke/SE (hazard ratio: 0.82; 95% confidence interval: 0.66 to 1.03) and a significantly lower risk of MB (hazard ratio: 0.62; 95% confidence interval: 0.54 to 0.70) versus warfarin. No significant interaction was observed between treatment and obesity status (nonobese, obese/nonmorbid, obese/morbid) for stroke/SE (interaction p = 0.602) or MB (interaction p = 0.385). In obese patients with NVAF, apixaban was associated with a similar risk of stroke/SE and a significantly lower risk of MB versus warfarin.
Atrial fibrillation (AF) is the most common cardiac arrhythmia in the United States, currently affecting ∼6 million, primarily patients aged >65 years. AF is associated with a considerable risk of stroke and resulting mortality. In addition to advanced age, , NVAF is associated with obesity, with NVAF risk being positively correlated with body mass index (BMI). Clinical guidelines recommend direct oral anticoagulants (apixaban, dabigatran, edoxaban, and Rivaroxaban) as first-line thromboprophylaxis for patients with NVAF, including obese patients (BMI 30 to 39 kg/m 2 ). However, obesity complicates NVAF anticoagulation, and data regarding clinical outcomes in this subpopulation is mixed. All anticoagulation increases bleeding risk, and obesity has been independently associated with increased bleeding risk in anticoagulated patients with NVAF. , , Moreover, the obese, and especially the severely obese populations, have been underrepresented in clinical trials of direct oral anticoagulants, and clinicians may be reluctant to prescribe fixed-dosed agents to patients with extremely high BMIs (>40 kg/m 2 ). , Thus, this study evaluated the risk of stroke/SE and MB in obese anticoagulated patients in a linked Veterans Affairs-Medicare database.
METHODS
This retrospective cohort study used linked national patient-level data from fee-for-service Medicare data from the United States Centers for Medicare and Medicaid Services database and from the Veterans Affairs database from January 1, 2013, to December 31, 2017. The Veterans Affairs data included vital measures (BMI, body weight, height) and laboratory results (international normalized ratio and time in therapeutic range) that are not available in Medicare claims data. Patients were first identified in the Veterans Affairs dataset and then linked to Medicare data using a unique patient identifier; linked patient records were deidentified before analysis. There were no duplicate records, and all outcomes were counted only once in the linked database.
Patients aged ≥18 years prescribed apixaban or warfarin were identified between July 1, 2013, and December 31, 2017. Patients were required to have ≥1 AF diagnosis claim in any position on or before the index date, which was defined as the first prescription date. Patients were required to have continuous health plan enrollment (Parts A and B) for ≥6 months before the index date. Exclusion criteria are listed in Figure 1 . After applying the selection criteria, patients were assigned to 1 of 3 subgroups based on body weight and BMI measures during the study period:
- •
Nonobese Patients: Patients with a BMI <30 kg/m 2 or bodyweight <100 kg were classified as nonobese patients.
- •
Obese Patients: Patients with a BMI ≥30 kg/m 2 or bodyweight ≥100 kg were classified as obese patients. For patients with multiple body weights during the study period, the body weight closest to the index date was used. BMI was given a preference over body weight in assigning patients to the previously mentioned cohorts. Obese patients were further classified as morbidly obese and obese/nonmorbid patients.
- o
obese/morbid: patients with BMI >40 kg/m 2 or weight >120 kg were classified as morbidly obese or obese/morbid patients.
- o
obese/nonmorbid: patients with BMI 30 to 40 kg/m 2 or body weight 100 to 120 kg were classified as obese/nonmorbid patients.
- o
Patients were followed from the day after the index date to the earliest of 30 days after the date of discontinuation, switch, death, end of continuous medical enrollment (Parts A and B), or the study end.
Patient demographic characteristics were summarized, including age, gender, race, and the geographic region at the index date. In addition, baseline patient clinical characteristics were summarized using available data in the 6-month preindex period, including Deyo-Charlson comorbidity index score, CHA 2 DS 2 -VASc score, HAS-BLED score, and comorbidities of interest. BMI and weight baseline characteristics were evaluated for 6 months preindex and postindex date.
The primary effectiveness outcome was stroke/SE—including ischemic stroke, hemorrhagic stroke, and SE—and the primary safety outcome was MB, stratified by gastrointestinal bleeding, intracranial hemorrhage, and bleeding at other key sites (e.g., the genitourinary tract, respiratory tract, or ocular area).
Stabilized inverse probability treatment weighting (IPTW) was used to balance patient characteristics between the treatment cohorts. Propensity scores were used to obtain estimates of the average treatment effect using a logistic model for the treatment cohorts. Baseline demographics and clinical characteristics were included as covariates in the propensity score model. Covariates included in the propensity score model were based on clinical rationale (such as gender, race, chronic coronary insufficiency [CCI], and baseline comorbidities) and those with a standardized difference (STD) greater than 10.0 (such as age, geographic region). After the propensity score calculation, each patient was weighted by the inverse of the probability of their treatment option (weight = 1/propensity score). The weights were stabilized by multiplying the original weights with a constant, which was equal to the expected value of being in the treatment or comparison cohorts, respectively. After IPTW, the baseline characteristics were well balanced between the cohorts.
The risk of stroke/SE and MB in each weighted cohort was evaluated using Cox proportional hazard models. Kaplan-Meier curves were visually inspected to ensure the proportional hazard assumption was not violated. Interaction analysis was conducted to evaluate the treatment effect across nonobese, obese/nonmorbid, and obese/morbid patients. The statistical significance (p <0.10) of the interactions between treatment and obesity status in the effect of effectiveness and safety outcomes were evaluated. Cox proportional hazard models were used to evaluate the risk of clinical outcomes. No covariates were included in the models, as they were balanced after IPTW.
To check the robustness of the analysis, 1:1 propensity score matching was conducted to evaluate the risk of stroke/SE and MB in obese patients who initiated apixaban versus warfarin. Additionally, a dose subgroup analysis was also conducted. In obese patients that initiated apixaban standard dose (5 mg twice a day) versus warfarin, IPTW was conducted, and the risk of stroke/SE and MB was evaluated using Cox proportional hazard models.
Results
After applying the selection, 519,114 unique dual beneficiaries were identified in the Veterans Affairs and Medicare databases with ≥1 pharmacy claim for apixaban or warfarin during the identification period (July 1, 2013, through December 31, 2017). After applying the selection criteria, a total of 73,807 patients with NVAF who initiated apixaban or warfarin were identified ( Figure 1 ). Of these, 26,522 were classified as obese patients of which 51.3% initiated apixaban and 48.7% initiated warfarin.
Overall, the study population was predominately male, White, from the Southern United States, and of older age. Before IPTW, patients who initiated warfarin had a significantly higher baseline Charlson comorbidity index score and higher rates of congestive heart failure and diabetes mellitus in the baseline as compared with apixaban ( Supplementary Table 1 ). After applying IPTW, the apixaban and warfarin cohorts were evenly balanced for all baseline variables. During follow-up, apixaban patients had a significantly lower risk of stroke/SE and MB compared with warfarin patients ( Figure 2 ). When stratified by obesity status (nonobese, obese/morbid, and obese/nonmorbid), no significant interaction was observed between treatment and obesity status for stroke/SE and MB ( Figure 2 ).
Table 1 lists the pre- and post-IPTW baseline characteristics in NVAF patients with obesity. Before IPTW, the average age of patients was similar across apixaban versus warfarin. After IPTW, patient characteristics were balanced. The mean age was 75 years, the mean CHA 2 DS 2 -VASc score was 3.8, the mean HAS-BLED score was ∼2.6, and most patients were males ( Table 1 ). During follow-up, apixaban patients with obesity were associated with a similar risk of stroke/SE but a significantly lower risk of MB compared with warfarin patients ( Figure 3 ).
| Pre-IPTW | Post-IPTW * | |||||
|---|---|---|---|---|---|---|
| Warfarin Cohort (Reference) | Apixaban Cohort | STD † | Warfarin Cohort (Reference) | Apixaban Cohort | STD † | |
| Sample size | 12,918 | 13,604 | 12,918 | 13,604 | ||
| Age (years), mean (SD) | 74.4 (7.9) | 75.4 (7.6) | 12.6 | 74.9 (7.9) | 74.9 (7.6) | 0.0 |
| Age category (years) | ||||||
| 18–54 | 96 (1%) | 62 (0%) | 3.7 | 80 (1%) | 77 (1%) | 0.6 |
| 55–64 | 780 (6%) | 482 (4%) | 11.7 | 688 (5%) | 548 (4%) | 6.2 |
| 65–74 | 6,434 (50%) | 6,722 (49%) | 0.8 | 6,196 (48%) | 6,960 (51%) | 6.4 |
| 75–79 | 2,231 (17%) | 2,420 (18%) | 1.4 | 2,286 (18%) | 2,375 (18%) | 0.6 |
| ≥80 | 3,377 (26%) | 3,918 (29%) | 6.0 | 3,669 (28%) | 3,644 (27%) | 3.6 |
| Gender | ||||||
| Male | 12,739 (99%) | 13,430 (99%) | 0.9 | 12,745 (99%) | 13,423 (99%) | 0.1 |
| Female | 179 (1%) | 174 (1%) | 0.9 | 173 (1%) | 181 (1%) | 0.1 |
| Race | ||||||
| White | 11,820 (92%) | 12,473 (92%) | 0.7 | 11,835 (92%) | 12,462 (92%) | 0.0 |
| Black | 801 (6%) | 791 (6%) | 1.6 | 771 (6%) | 814 (6%) | 0.0 |
| Other | 191 (1%) | 229 (2%) | 1.6 | 205 (2%) | 215 (2%) | 0.1 |
| Unknown | 106 (1%) | 111 (1%) | 0.1 | 107 (1%) | 113 (1%) | 0.1 |
| US. Geographic region | ||||||
| Northeast | 2,462 (19%) | 2,413 (18%) | 3.4 | 2,364 (18%) | 2,487 (18%) | 0.1 |
| Midwest | 4,118 (32%) | 3,748 (28%) | 9.5 | 3,837 (30%) | 4,039 (30%) | 0.0 |
| South | 4,270 (33%) | 5,289 (39%) | 12.2 | 4,654 (36%) | 4,900 (36%) | 0.0 |
| West | 2,037 (16%) | 2,103 (16%) | 0.9 | 2,021 (16%) | 2,135 (16%) | 0.1 |
| Other/unknown | 31 (0%) | 51 (0%) | 2.4 | 41 (0%) | 43 (0%) | 0.0 |
| Body Weight (kg) , ‡ mean (SD) | 111.5 (19.4) | 109.5 (16.8) | 11.2 | 111.0 (19.1) | 109.9 (17.0) | 6.2 |
| BMI (kg/m 2 ) , ‡ mean (SD) | 37.1 (63.5) | 35.4 (36.3) | 3.2 | 37.2 (69.5) | 35.5 (33.3) | 3.2 |
| Baseline Bleed Scores, mean (SD) | ||||||
| Deyo-Charlson comorbidity index | 3.1 (2.6) | 2.6 (2.4) | 20.3 | 2.8 (2.5) | 2.9 (2.6) | 0.5 |
| CHA 2 DS 2 -VASc score | 3.8 (1.6) | 3.8 (1.5) | 5.3 | 3.8 (1.5) | 3.8 (1.5) | 1.5 |
| HAS-BLED score § | 2.7 (1.1) | 2.6 (1.1) | 5.5 | 2.7 (1.1) | 2.6 (1.1) | 5.5 |
| Baseline comorbid conditions, n (%) | ||||||
| Bleeding history | 1,951 (15%) | 1,678 (12%) | 8.1 | 1,767 (14%) | 1,866 (14%) | 0.1 |
| Congestive heart failure (CHF) | 4,824 (37%) | 4,397 (32%) | 10.6 | 4,463 (35%) | 4,690 (34%) | 0.2 |
| Diabetes mellitus | 7,248 (56%) | 6,958 (51%) | 10.0 | 6,899 (53%) | 7,270 (53%) | 0.1 |
| Hypertension | 11,193 (87%) | 11,812 (87%) | 0.5 | 11,206 (87%) | 11,798 (87%) | 0.1 |
| Renal disease | 1,759 (14%) | 2,858 (21%) | 19.6 | 2,287 (18%) | 2,390 (18%) | 0.4 |
| Liver disease | 524 (4%) | 547 (4%) | 0.2 | 519 (4%) | 542 (4%) | 0.2 |
| Myocardial infarction | 1,845 (14%) | 1,862 (14%) | 1.7 | 1,802 (14%) | 1,888 (14%) | 0.2 |
| Dyspepsia or stomach discomfort | 1,421 (11%) | 1,319 (10%) | 4.3 | 1,317 (10%) | 1,395 (10%) | 0.2 |
| Non-stroke/ SE peripheral vascular disease | 2,903 (22%) | 2,950 (22%) | 1.9 | 2,838 (22%) | 2,984 (22%) | 0.1 |
| Stroke/SE | 1,240 (10%) | 1,236 (9%) | 1.8 | 1,184 (9%) | 1,248 (9%) | 0.0 |
| Transient ischemic attack (TIA) | 786 (6%) | 1,258 (9%) | 11.9 | 1,009 (8%) | 1,051 (8%) | 0.3 |
| Anemia and coagulation defects | 3,157 (24%) | 2,875 (21%) | 7.9 | 2,918 (23%) | 3,070 (23%) | 0.0 |
| Alcoholism | 4,532 (35%) | 5,439 (40%) | 10.1 | 4,865 (38%) | 5,107 (38%) | 0.3 |
| Peripheral artery disease | 2,821 (22%) | 2,743 (20%) | 4.1 | 2,741 (21%) | 2,793 (21%) | 1.7 |
| Coronary artery disease | 6,814 (53%) | 7,288 (54%) | 1.7 | 6,729 (52%) | 7,365 (54%) | 4.1 |
| Dose of the index OAC prescription, n (%) | ||||||
| Standard Dose (5 mg apixaban) | 11,813 (87%) | 11,890 (87%) | ||||
| Lower Dose (2.5 mg apixaban) | 1,791 (13%) | 1,714 (13%) | ||||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
