To compare the efficacy and safety of apixaban and rivaroxaban for the prevention of stroke in patients with nonvalvular atrial fibrillation (NVAF) by way of a meta-analysis informed by real-world evidence. Systematic review and meta-analysis of observational studies including patients with NVAF on apixaban and rivaroxaban, which reported stroke/systemic embolism and/or major bleeding. Prospero registration number: CRD42021251719. Estimates of relative treatment effect (based on hazard ratios[HRs]) were pooled using the inverse variance method. Fixed-effects and random effect analyses were conducted. Exploratory meta-regression analyses that included study-level covariates were conducted using the metareg (meta-regression) command of Stata Statistical Software: Release 15.1 (College Station, Texas. StataCorp LLC.). Study level covariates explored in the meta-regression analyses were CHA 2 DS 2 -VASc and HAS-BLED scores. A total of 10 unique retrospective real-world evidence studies reported comparative estimates for apixaban versus rivaroxaban in patients with NVAF and were included in the meta-analysis. Adjusted HR was 0.88 (95% [confidence interval] CI 0.81 to 0.95), indicating a significantly lower hazard of stroke/systemic embolism associated with apixaban versus rivaroxaban. Pairwise meta-analysis for a major bleeding episode was significantly lower with apixaban compared with rivaroxaban (HR 0.62; 95% CI 0.56 to 0.69), whereas apixaban was associated with a lower risk of gastrointestinal bleeding compared with rivaroxaban (HR 0.57; 95% CI 0.50 to 0.64). In conclusion, this study suggests that patient CHA 2 DS 2 -VASc and HAS-BLED scores might be an important factor when selecting which direct oral anticoagulants to use, given the relation these scores have on treatment outcomes. Apixaban is associated with lower rates of both major and gastrointestinal bleeding than rivaroxaban, with no loss of efficacy.
Atrial fibrillation (AF) is a major risk factor for stroke, and strokes associated with AF are particularly severe, highly recurrent, often fatal, or with a permanent disability. The management of non-valvular AF (NVAF) includes the use of anticoagulants to reduce the risk of stroke, with a concomitant risk of bleeding, particularly gastrointestinal and intracranial hemorrhage. Apixaban and rivaroxaban are direct-acting oral anticoagulants (DOACs) that work by directly inhibiting Factor Xa in the coagulation cascade, and they are 2 of the most commonly prescribed DOACs by clinicians in treating patients with NVAF. Several systematic literature reviews (SLR) and meta-analyses of both randomized controlled trial (RCT) and real-world evidence (RWE) demonstrated that, whereas there was no apparent difference in the efficacy/effectiveness profiles between apixaban and rivaroxaban in the treatment of NVAF, apixaban was associated with fewer major bleeding events than rivaroxaban. However, those studies did not consider the potential impact of patients treated with lower doses of DOACs on the outcomes studied or consider the influence of CHA 2 DS 2 -VASc and HAS-BLED on efficacy and safety end points. The objective of this study was to compare the effectiveness and safety of apixaban and rivaroxaban for the prevention of stroke in patients with NVAF using an RWE meta-analysis.
Methods
The SLR protocol is registered with Prospero, unique identifier CRD42021251719. The study was conducted following PRISMA (preferred reporting items for systematic reviews and meta-analyses) guidelines. Searches of PubMed and Embase were conducted between January 1, 2012, and February 2018 3 and updated to May 5, 2021, to capture the earliest prescription of DOACs for treatment of NVAF in the prevention of stroke. The date of censoring for publications included in this meta-analysis was December 31, 2020. The search strategy is provided in the Supplementary Appendix (Supplementary materials, Search strategy). Search results were screened for RWE studies investigating rivaroxaban and apixaban in adult patients with NVAF.
All English language abstracts of published and peer-reviewed studies extracted from the search were reviewed to assess whether they were consistent with the following inclusion criteria: (1) studies to compare apixaban with rivaroxaban, (2) studies to report stroke/systemic embolism and/or major bleeding, (3) studies must include data about a lower dose (either adjusted or separated) or have sensitivity analysis, (4) studies to report an appropriate method used to adjust for confounding and matching groups such as propensity score matching or inverse probability treatment weighting.
Citations of interest were identified by a reviewer (JAM) and verified by an independent reviewer (KP), based on the title and abstract. Full publications were obtained for all citations of interest and were assessed by 1 reviewer and verified by a second reviewer. Any uncertainties were resolved through discussion. Data were extracted to tables in Microsoft Word by 1 reviewer and checked against the original publication by a second reviewer. RWE studies were assessed for risk of bias using the Agency for Healthcare Research and Quality (AHRQ) assessment tool.
The statistical analysis was conducted in Review Manager (RevMan) Version 5.4.1 (The Cochrane Collaboration, 2020) and Stata Statistical Software: Release 15.1 (College Station, Texas. StataCorp LLC.). Study-level estimates of relative treatment effect (based on hazard ratios [HRs]) were pooled using the inverse variance method. Fixed-effects and random effect analyses were conducted and the results of the random effect models were considered the base case. Exploratory meta-regression analyses that included study-level covariates were conducted using the metareg (meta-regression) command of Stata Statistical Software: Release 15.1 Study level covariates explored in the meta-regression analyses were CHA 2 DS 2 -VASc score and HAS-BLED score.
Results
The PRISMA flow diagram and results are depicted in Figure 1 . The initial 2018 search yielded 5,340 results. After the exclusion of duplicates, a total of 4,457 results were screened based on the title and abstract. Finally, after screening 1,370 full-text articles, 5 articles were included. The updated search identified a further 12 articles (4 of 5 articles identified in the original search were superseded by publications identified by the updated search). In summary, a total of 13 publications reporting on 10 independent studies were included with 592,772 participants included (281,013 patients in the apixaban group and 311,759 patients in the rivaroxaban group) (Supplementary Table 1). These 10 independent retrospective RWE studies were included in the meta-analysis.
All the identified RWE studies evaluated retrospective patient cohorts; 6 studies used patient data from administrative/claims databases and 4 studies used patient registry data. Study periods ranged from 2012 to 2019, and primary outcomes included all-cause hospitalization (2 studies reporting stroke and major bleeding), stroke and systemic embolism (5 studies) , , and major bleeding (3 studies). , , A total of 6 studies were conducted in the United States, , , and single studies were conducted in Denmark, France, Norway , and Sweden. Patient cohort size varied considerably across studies and ranged from 3,676 11 to 100,163 15 per treatment cohort. Study designs of the 10 RWE studies included in this analysis are summarized in Supplementary Table 1.
The mean age of patients ranged from 68.0 years to 78.3 years per treatment cohort. One study only evaluated patients with NVAF aged 65 years and over; however, the mean age of the patients in the apixaban and rivaroxaban cohorts were comparable. There was notable variability across studies in the proportions of patients according to gender but the proportion of male patients within studies were comparable between the apixaban and rivaroxaban cohorts; the proportion of male patients ranged from 47.2% to 63.2%. All but 2 studies reported CHA 2 DS 2 -VASc scores, with a mean score ranging from 2.4 10 to 4.6 8 per treatment cohort. By contrast, the CHADS 2 score was inconsistently reported across the studies; if reported (n = 4), these scores had notable variability across studies. Similarly, there was also notable variability across studies reporting HAS-BLED scores (n = 7), and the mean HAS-BLED score ranged from 1.8 14 to 3.2. Baseline characteristics of the patients evaluated in the 10 included RWE studies are summarized in Table 1 .
| Author, Year | Treatment Groups | N (Matched) | Age, Years | Male | CHADS 2 | CHADS 2 -VASc Score | HAS-BLED | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Mean | SD | Median | IQR | n | % | Mean | SD | Mean | SD | Median | IQR | Mean | SD | Median | IQR | |||
| Amin et al | apixaban | 8,440 | 72.84 | 11.06 | NR | NR | 4,734 | 56.09 | 2.3 | 2.29 | 3.79 | 1.83 | NR | NR | 2.75 | 1.3 | NR | NR |
| Rivaroxaban | 8,440 | 72.53 | 10.79 | NR | NR | 4,740 | 56.16 | 2.25 | 2.27 | 3.77 | 1.83 | NR | NR | 2.74 | 1.29 | NR | NR | |
| Amin et al | apixaban | 38,820 | 78.3 | 7.4 | NR | NR | 18,414 | 47.4 | 2.7 | 1.4 | 4.5 | 1.7 | NR | NR | 3.2 | 1.2 | NR | NR |
| Rivaroxaban | 38,820 | 78.3 | 7.2 | NR | NR | 18,334 | 47.2 | 2.4 | 1.4 | 4.6 | 1.7 | NR | NR | 3.2 | 1.2 | NR | NR | |
| Noseworthy et al | apixaban | 6,565 | NR | NR | 73 | 65–81 | NR | 54 | NR | NR | NR | NR | 4 | 3–5 | NR | NR | 2 | 3–5 |
| Rivaroxaban | 6,565 | NR | NR | 73 | 65–81 | NR | 54.4 | NR | NR | NR | NR | 4 | 3–5 | NR | NR | 2 | 3–5 | |
| Tepper et al | apixaban | 8,785 | 70 | 12 | 70 | 61–80 | NR | 62.7 | 1.7 | 1.1 | 2.5 | 1.5 | NR | NR | 1.9 | 1.2 | NR | NR |
| Rivaroxaban | 30,529 | 68 | 12 | 68 | 60–78 | NR | 63.2 | 1.6 | 1.1 | 2.4 | 1.5 | NR | NR | 1.8 | 1.2 | NR | NR | |
| Andersson et al | apixaban | 3,676 | 71.9 | 9.1 | NR | NR | 2057 | 56 | NR | NR | NR | NR | NR | NR | NR | NR | NR | NR |
| Rivaroxaban | 3,676 | 72 | 9.8 | NR | NR | 2041 | 56 | NR | NR | NR | NR | NR | NR | NR | NR | NR | NR | |
| Jansson et al | apixaban | 11,493 | 73.5 | 10.3 | NR | NR | NR | 57 | NR | NR | NR | NR | 3 | 2–4 | NR | NR | NR | NR |
| Rivaroxaban | 7,897 | 73.5 | 10.3 | NR | NR | NR | 57 | NR | NR | NR | NR | 3 | 2–4 | NR | NR | NR | NR | |
| Rutherford et al | apixaban | 13,699 | 72.7 | 11.66 | 73 | NR | 7,946 | 58 | NR | NR | 3.23 | 1.74 | NR | NR | 2.43 | 1.15 | NR | NR |
| Rivaroxaban | 13,699 | 72.7 | 11.08 | 73 | NR | 7,943 | 58 | NR | NR | 3.22 | 1.71 | NR | NR | 2.43 | 1.12 | NR | NR | |
| Fralick et al | apixaban | 39,351 | 69.4 | 10.5 | NR | NR | NR | 60.3 | NR | NR | NR | NR | NR | NR | NR | NR | NR | NR |
| Rivaroxaban | 39,351 | 69.3 | 10.6 | NR | NR | NR | 60.4 | NR | NR | NR | NR | NR | NR | NR | NR | NR | NR | |
| Van Ganse et al | apixaban | 87,565 | 74.7 | 11.5 | 76 | NR | NR | 51.2 | 1.6 | 1.3 | 3.1 | 1.7 | 3 | NR | 2.2 | 1 | 2 | NR |
| Rivaroxaban | 100,163 | 72 | 12 | 73 | NR | NR | 54.4 | 1.3 | 1.2 | 2.7 | 1.7 | 3 | NR | 2 | 1 | 2 | NR | |
| Lip et al | apixaban | 62,619 | 73.1 | NR | NR | NR | NR | 55.2 | NR | NR | 3.5 | NR | NR | NR | 2.9 | NR | NR | NR |
| Rivaroxaban | 62,619 | 72.9 | NR | NR | NR | NR | 54.8 | NR | NR | 3.5 | NR | NR | NR | 2.8 | NR | NR | NR | |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
