Secondary Stroke Prevention in A-Fib: Do DOACs PROSPER in High-Risk Patients?

Author(s)
Blaire White, PharmD
Amber Cizmic, PharmD, BCACP
Tisha Smith, PharmD, BCACP, CACP

Reviewed By
Robert DiDomenico, PharmD, BCPS, AQ Cardiology
Elisa Greene, PharmD, BCACP
Kevin Phan, PharmD

Citation
Xian Y, Xu H, O’Brien EC, et al. Clinical effectiveness of direct oral anticoagulants vs warfarin in older patients with atrial fibrillation and ischemic stroke: findings from the patient-centered research into outcomes stroke patients prefer and effectiveness research (PROSPER) study. JAMA Neurol. 2019 76 (10): 1192-1202

The Problem

Several guidelines now recommend direct oral anticoagulants (DOACs) as the preferred anticoagulants for patients with non-valvular atrial fibrillation (a-fib).^1,2 However, the landmark clinical trials focused largely on the primary prevention of stroke in patients with a-fib.  Moreover, real-world data using DOACs for secondary prevention, especially in older adults, is lacking. Many have argued that warfarin might be a better choice in these high-risk patients because it requires routine monitoring and increases the patient’s contact with the healthcare system.  Leaving many to wonder – does the choice of anticoagulant make a difference in preventing recurrent stroke?

 

What’s Known

The current guideline recommendations are based on the results of four key trials: ARISTOTLE, RE-LY, ROCKET AF, and ENGAGE AF-TIMI 48.^3-6 The results of these studies are summarized in Table 1. Meta-analysis from clinical trials suggests that DOACs can reduce the risk of stroke, intracranial hemorrhage, and mortality but have similar risk of major bleeding and an increased risk of gastrointestinal bleeding compared with warfarin.⁷

 

Table 1: Summary of Trials for Anticoagulation in Atrial Fibrillation^3-6

Trials (DOAC)	Design	Primary Efficacy Outcome (DOAC vs. Warfarin)	Relative Risk (DOAC vs. Warfarin) HR (95% CI)	Major Bleeding Rate (DOAC vs. Warfarin)
ARISTOTLE (Apixaban)	Apixaban 5 or 2.5 mg BID vs. warfarin	Stroke or systemic embolism: 1.27% vs. 1.6%	0.79 (0.66-0.95); p<0.001 for noninferiority, p=0.01 for superiority	2.13% vs. 3.09%; p<0.001
RE-LY (Dabigatran)	Dabigatran 110 or 150 mg BID vs. warfarin	Stroke or systemic embolism: 110 mg: 1.53% 150 mg: 1.11% Warfarin: 1.69%	110 mg: 0.91 (0.74-1.11); p<0.001 150 mg: 0.66 (0.53-0.82); p<0.001	110 mg: 2.71%; p=0.003 150 mg: 3.11%; p=0.31 Warfarin: 3.36%
ROCKET AF (Rivaroxaban)	Rivaroxaban 20 or 15 mg daily vs. warfarin	Stroke or systemic embolism: 1.7% vs. 2.2%	0.79 (0.66-0.96); p<0.001	3.6% vs. 3.4%; p=0.58
ENGAGE AF-TIMI 48 (Edoxaban)	Edoxaban 30 or 60 mg daily vs. warfarin	Stroke or systemic embolism: 30 mg: 1.61% 60 mg: 1.18% Warfarin: 1.5%	30 mg: 1.07 (97.5% CI, 0.87-1.31); p=0.005 60 mg: 0.79 (97.5% CI, 0.63-0.99); p<0.001	30 mg: 1.61%; p<0.001 60 mg: 2.75%; P<0.001 Warfarin: 3.43%

HR: hazard ratio       CI: confidence interval

 

What’s New

The PROSPER study was conducted to evaluate the real-world clinical effectiveness of DOACs versus warfarin for secondary prevention in older adults with atrial fibrillation who experienced an ischemic stroke.⁸ This retrospective analysis used data from the American Heart Association/American Stroke Association (AHA/ASA) Get with the Guidelines-Stroke (GWTG-Stroke) clinical registry and Centers for Medicare & Medicaid Services (CMS) claims between October 2011 and December 2014. Patients included were 65 years or older with atrial fibrillation or atrial flutter and newly prescribed oral anticoagulants at discharge following an acute ischemic stroke. Only anticoagulant-naïve patients were included in an attempt to avoid selection bias. Patients were categorized into those prescribed a DOAC (dabigatran, rivaroxaban, or apixaban) or warfarin. Patients who were discharged to hospice, transferred to another hospital, received comfort measures only, had contraindications for anticoagulation treatment, or with missing National Institutes of Health Stroke Scale (NIHSS) scores were excluded from the analysis. Those with renal insufficiency, dialysis, creatinine clearance less than 15 mL/min, or with missing data were also excluded.

 

The co-primary outcomes were “home time” and major adverse cardiovascular events (MACE). Home time is a patient-centered outcome and defined as the total number of days alive and out of the hospital or a skilled nursing facility during the first year after the index hospital discharge. The MACE end point is a composite of all-cause mortality, cardiovascular, or cerebrovascular readmission. Secondary outcomes included all-cause mortality, fatal bleeding, all-cause readmission, cardiovascular readmission, ischemic stroke readmission, systemic embolism readmission, hemorrhagic stroke readmission, gastrointestinal bleeding, and any bleeding requiring hospitalization. A negative binomial model was used for the home time outcome, and a Cox proportional hazards model was used for MACE. A propensity score overlap weighting approach was used to control for selection bias.

 

In the primary analysis, 11,662 patients were included (4041 [34.7%] prescribed DOACs and 7621 [65.3%] prescribed warfarin). The median age of patients was 80 years old, 56% were female, and 87% were white. The median pre-stroke CHA₂DS₂-VASc score was 4, and 98% of patients had a score of 2 or greater, indicating that most should have received anticoagulant therapy prior to the stroke. Patient characteristics were similar between groups with the exception of NIHSS scores and zip code-level unemployment rates. Compared to warfarin, patients prescribed DOACs had significantly more days at home and were significantly less likely to experience major adverse cardiovascular events (see Table 2). Additionally, patients receiving DOACs had significantly fewer deaths, all-cause readmissions, cardiovascular readmissions, hemorrhagic stroke readmissions, and hospitalizations due to any bleeding, but there was a significantly higher incidence of gastrointestinal bleeding (see Table 2). There were no detectable differences in fatal bleeding, ischemic stroke readmissions, and systemic embolism readmissions.  

 

Table 2: Major Results from PROSPER

	DOACs vs. Warfarin (Events/ 100 Patient Years)	Adjusted HR (CI)a
Primary Outcomes
Home time (mean)	287.2 vs. 263 days	15.6 (9.0-22.1); p <0.001
MACE	1930 vs. 4476 (34 vs. 40.4)	0.89 (0.83-0.96); p<0.001
Secondary Outcomes
Mortality
All-cause	1183 vs. 3028 (15.8 vs. 19.6)	0.88 (0.82-0.95); p<0.001
Fatal bleeding	59 vs. 164 (0.8 vs. 1.1)	0.84 (0.63-1.12); p=0.23
Readmission
All-cause	2353 vs. 5052 (53.1 vs. 62.4)	0.93 (0.88-0.97); p=0.003
Cardiovascular	1267 vs. 2770 (22.3 vs. 25)	0.92 (0.86-0.99); p=0.02
Ischemic stroke	380 vs. 770 (5.6 vs. 5.6)	1.01 (0.89-1.14); p=0.91
Systemic embolism	42 vs. 100 (0.6 vs. 0.7)	0.95 (0.66-1.38); p=0.8
Hemorrhagic stroke	54 vs. 151 (0.8 vs. 1)	0.69 (0.5-0.95); p=0.02
GI bleeding	367 vs. 714 (5.4 vs. 5.1)	1.14 (1.01-1.3); p=0.03
Any bleeding	728 vs. 1717 (11.4 vs. 13.4)	0.89 (0.81-0.97); p=0.009

HR: hazard ratio       CI: confidence interval

^a 99% CI for primary outcomes, 95% CI for secondary outcomes

 

Our Critical Appraisal

The PROSPER study has several strengths. First, PROSPER provides real-world data regarding the secondary prevention of stroke – information that was previously lacking. Previous trials (ARISTOTLE, RE-LY, ROCKET AF, and ENGAGE AF-TIMI 48) focus largely on the primary prevention of stroke.^3-6 A meta-analysis of DOACs versus warfarin looked at patients similar in age (75 years and older) to PROSPER and showed comparable results but did not include a patient-centered outcome (such as home time) or secondary prevention.⁹ While other real-world studies have been published, only a minority (≤10%) of the patients in these analyses had a prior stroke.^10,11 Additionally, patients in the PROSPER study draw from a nationally representative sample. As a result, findings from this study can be applied more confidently to older adults. Compared to other landmark trials, the participants in this study were significantly older, more likely to be women, had multiple comorbidities, and had several risk factors for recurrent stroke. These factors likely explain the higher rate of MACE and recurrent ischemic strokes seen in this cohort. Importantly, these data provide valuable insights about the risks and benefits of DOACs in a much broader population.

 

Like any study, PROSPER has some limitations. This was a retrospective study and confounding may exist — those prescribed a DOAC may be different in important (but unknown) ways from those prescribed warfarin.  The investigators attempted to mitigate confounding by using propensity score overlap weighting but these statistical manipulations aren’t perfect. The study does not provide any information about specific DOAC use – this would be important particularly as it relates to gastrointestinal bleeding. Nearly 75% of DOAC patients were on dabigatran or rivaroxaban, both of which have higher rates of gastrointestinal bleeding compared to warfarin. Apixaban represented a minority of DOAC patients in this analysis because the FDA approval of this agent occurred after October 2011 whereas the other DOACs were approved prior to the analysis period. It is also uncertain whether DOACs were dosed appropriately for renal dysfunction. Thus, stroke risk would potentially be higher in patients who were under-dosed and bleed risk higher in patients who were over-dosed. In the ARISTOTLE, RE-LY, ROCKET-AF, and ENGAGE AF-TIMI 48 trials, median time in therapeutic range for those patients treated with warfarin was 58-68%.^5,6,11,12 Information for time in therapeutic range and DOAC adherence is not available for the PROSPER study. One may argue that PROSPER better reflects real-world clinical practice compared to the highly monitored patients who enrolled in the Phase 3 studies. Lastly, one of the co-primary outcomes in the PROSPER study was “home time” – an outcome that has not been employed in other anticoagulation stroke-prevention trials. However, “home time” is an intuitively attractive patient-centered outcome and correlates with a patient’s modified Rankin scale score which is the commonly used gold standard to assess functional status in stroke trials.

 

The Bottom Line

Overall, the results of PROSPER suggest that prescribing a DOAC at discharge is associated with better long-term outcomes compared with warfarin for older patients with atrial fibrillation following acute ischemic stroke. The results from PROSPER provide much needed data regarding the secondary stroke prevention in older adults, and the results support guideline recommendations to preferentially use DOACs instead of warfarin for patients with atrial fibrillation.¹

 

Key Points

• The PROSPER study provides important new insights into the benefits and risks of DOACs in older adults (median age = 80 years!) following a cardioembolic stroke due to a-fib.
• In a real-world population sample, DOACs significantly reduced the risk of major cardiovascular events and all-cause mortality as well as significantly increased home time.
• Home time represents an intuitively attractive patient-centered measure of effectiveness that could be considered in future stroke prevention studies.