Real‐world incidence of efficacy and safety outcomes in patients on direct oral anticoagulants with left ventricular systolic dysfunction at a tertiary referral center

Abstract

Background

Patients with heart failure (HF) have increased risk for thromboembolic events. Real‐world incidences of efficacy and safety outcomes of direct oral anticoagulants (DOACs) in patients with left ventricular systolic dysfunction (LVSD) are of growing clinical interest.

Hypothesis

Real‐world efficacy and safety outcomes of DOACs in patients with LVSD will be similar to those of LVSD or HF subgroups in the RE‐LY, ROCKET‐AF, and ARISTOTLE trials.

Methods

We performed a retrospective review of adult patients with LVSD (left ventricular ejection fraction ≤40%) on DOAC therapy between 2010 and 2016. Incidences of safety and efficacy outcomes of anticoagulation with DOACs were extracted from primary and secondary hospital discharge diagnoses.

Results

DOACs were prescribed to 287 patients with LVSD over a mean follow‐up of 313.3 ± 52.3 days. Many patients had moderate and severe chronic kidney disease (28.9% and 10.1%, respectively) and indications for anticoagulation therapy other than atrial fibrillation (19.9%). For efficacy outcomes, the calculated incidence rates of ischemic stroke and systemic embolism were 1.2 (95% confidence interval [CI]: 0.25‐3.56) and 0.81 (95% CI: 0.10‐2.94) events per 100 person‐years, respectively. For the safety outcomes, incidence rates of GI bleeding and intracranial hemorrhage were 2.4 (95% CI: 0.8‐5.3) and 0.41 (95% CI: 0.1‐2.2) events per 100 patient‐years, respectively.

Conclusions

Our findings are largely compatible with the results of LVSD or HF subgroups in RE‐LY, ROCKET‐AF, and ARISTOTLE trials and add to increasing confidence that DOACs can be safely used for stroke and systemic embolism prevention in patients with LVSD.

1. INTRODUCTION

Patients with heart failure (HF) are at increased risk for both venous and arterial thromboembolic events through a variety of mechanisms that impact all elements of Virchow's triad: stasis, endothelial dysfunction, and hypercoagulability (Figure 1).1 Patients with left ventricular systolic dysfunction (LVSD) often have other comorbid conditions that lead to an increased risk of stroke, including atrial fibrillation (AF), which itself is an indication for long‐term anticoagulation.2, 3, 4 Large randomized controlled trials (RCTs) and systematic reviews investigating direct oral anticoagulants (DOACs), such as the Randomized Evaluation of Long‐Term Anticoagulation Therapy (RE‐LY), Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE), Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared With Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation (ROCKET‐AF), and Effective Anticoagulation With Factor Xa Next Generation in Atrial Fibrillation–Thrombolysis In Myocardial Infarction 48 (ENGAGE AF‐TIMI 48) trials, have shown that DOAC therapy is not inferior to warfarin in the prevention of stroke and systemic embolism, and that it significantly reduces hemorrhagic complications, especially intracranial hemorrhage (ICH) in patients with nonvalvular AF.5, 6, 7, 8, 9, 10 In subgroup analyses of LVSD patients in the RE‐LY (dabigatran), ROCKET‐AF (rivaroxaban), and ARISTOTLE (apixaban) trials, researchers found similar results, with no significant difference between HF patients with preserved or reduced ejection fractions.11, 12, 13

Figure 1.

All elements of Virchow's triad are present in patients with HF. Abbreviations: HF, heart failure; IL‐1, interleukin 1; TNF a, tumor necrosis factor‐α; vWF, von Willebrand factor

However, these RCTs excluded many patients that we often see in actual medical practice, including patients on anticoagulation therapy for indications other than AF, patients with significant renal dysfunction, and patients with a history of gastrointestinal (GI) bleeding and anemia, valvular heart disease, and indications for anticoagulation besides AF. Because of these differences in patient characteristics, the real‐world safety and efficacy of DOACs in LVSD is not well characterized. To determine whether the real‐world safety and efficacy of DOACs in LVSD patients approximates that observed in RCTs, we retrospectively evaluated these outcomes in patients with LVSD on DOACs at a single center.

2. METHODS

2.1. Data source

The present study is a retrospective cohort study of patients with LVSD at the Mayo Clinic in Rochester, Minnesota, from January 2010 until January 2016. Longitudinal health information for all patients meeting study criteria was extracted from clinical data stored in the Mayo Clinic database. This study was approved by the Mayo Clinic Institutional Review Board.

2.2. Study population

We identified adult patients with LVSD (defined as left ventricular ejection fraction ≤40%) with a confirmatory echocardiographic examination within 12 months of inclusion into the study who were on DOAC (dabigatran, rivaroxaban, or apixaban) therapy. Due to its recent approval by the US Food and Drug Administration, edoxaban was not included in this study, as no patients were taking this medication within the study period.

We also obtained sociodemographic information (including age, sex, and race) and medical comorbidity data.

2.3. Primary endpoints

The primary endpoints are inpatient admissions for (1) clinical efficacy outcomes including ischemic stroke, transient ischemic attacks, and systemic embolism (SE); and (2) clinical safety outcomes including GI bleeding, ICH, and bleeding from other sites. All clinical efficacy and safety outcomes were identified using primary and secondary hospital discharge diagnoses based on International Classification of Diseases, Ninth Revision, Clinical Modification (ICD‐9‐CM) and ICD‐10‐CM codes (Table 1). Patients had to be on anticoagulation therapy within 1 week of the primary outcomes.

Table 1.

ICD‐9‐CM and ICD‐10‐CM codes used for primary endpoints

	ICD‐9‐CM Codes	ICD‐10‐CM Codes
Efficacy outcomes
Ischemic stroke	433.×1, 434.×1, 436	I63.x
TIA	435.x	G45.x
SE	444.x	I74.x
Safety outcomes
GI bleeding	456.0, 456.20, 530.21, 530.7, 530.82, 531.0×, 531.2×, 531.4×, 531.6×, 532.0×, 532.2×, 532.4×, 532.6×, 533.0×, 533.2×, 533.4×, 533.6×, 534.0×, 534.2×, 534.4×, 534.6×, 535.01, 535.11, 535.21, 535.31, 535.41, 535.51, 535.61, 535.71, 537.83, 537.84, 562.02, 562.03, 562.12, 562.13, 568.81, 569.3, 569.85, 578.x	I85.01, I85.11, K22.11, K22.6, K25.0, K25.2, K25.4, K25.6, K26.0, K26.2, K26.4, K26.6, K27.0, K27.2, K27.4, K27.6, K28.0, K28.2, K28.4, K28.6, K29.01, K29.41, K29.61, K29.21, K29.61, K29.71, K29.81, K29.91, K31.811, K31.82, K57.11, K57.13, K57.31, K57.33, K66.1, K62.5, K55.21, K92.x
ICH	430, 431, 432.x, 852.x, 853.x	I60.9, I61.9, I62.x, S06.4×, S06.5×, S06.6×
Bleeding from other sites	423.0, 459.0, 596.7, 599.71, 719.1×, 784.8, 786.3	I31.2, R58, R31.0, M25.0×, R04.x

Abbreviations: GI, gastrointestinal; ICD‐9‐CM, International Classification of Diseases, Ninth Revision, Clinical Modification; ICD‐10‐CM, International Classification of Diseases, Tenth Revision, Clinical Modification; ICH, intracranial hemorrhage; SE, systemic embolism; TIA, transient ischemic attack.

The incidence of these safety and efficacy outcomes in our cohort was compared to that of the subgroup of patients with LVSD (defined as left ventricular ejection fraction ≤40%) and/or clinically diagnosed HF in the RE‐LY, ROCKET‐AF, and ARISTOTLE trials. The data from these trials included in this study were obtained directly from the published manuscripts.11, 12, 13 For the major clinical trials, but not the present study, major bleeding was defined as clinically overt bleeding associated with any of the following: fatal outcome, involvement of a critical anatomic site, drop in hemoglobin concentration > 2 g/dL, transfusion of >2 units of whole blood or packed red blood cells, or permanent disability. Nonmajor clinically relevant bleeding was defined as overt bleeding not meeting criteria for major bleeding but requiring medical intervention, unscheduled contact (visit or telephone) with a physician, temporary interruption of study drug (ie, delayed dosing), pain, or impairment of daily activities.

2.4. Follow‐up

The index date for each patient was determined by the date of initiation of anticoagulation therapy. The patients were followed until (1) their first event related to the efficacy or safety outcomes as determined by hospital discharge summaries, (2) termination of anticoagulation therapy for any reason as determined by clinical notes, or (3) the end of the study period.

2.5. Statistical analysis

Descriptive analyses were performed on the patient demographic information. Continuous variables were reported as mean ± SD and categorical variables as frequencies and percentages. Continuous and categorical variables were compared using 2‐sample t tests and χ² tests, respectively. The incidence rates of the safety and efficacy outcomes were determined in events per 100 person‐years with 95% confidence interval (CI). All statistical analyses were performed using JMP software version 10.0.0 (SAS Institute, Inc., Cary, NC).

3. RESULTS

3.1. Patient demographics and baseline characteristics

We included 287 patients with LVSD on DOAC therapy, specifically dabigatran (n = 189), rivaroxaban (n = 29), or apixaban (n = 116). Because some patients switched between DOACs, they were treated as single patients for analysis. The mean age at the initiation of DOAC therapy was 68.2 ± 12.5 years. Most patients (80.1%) were on DOAC for AF. A significant proportion of patients had chronic kidney disease (CKD) stage 3 (28.9%), stage 4 (7.3%), and stage 5 or end‐stage renal disease (2.8%). The patient demographics and medical comorbidities are summarized in Table 2. Of the 29 patients with severe CKD (≥stage 4), 9 patients (31%) had dose reductions per manufacturing dosing guidelines and 1 patient had an inappropriate dose.

Table 2.

Descriptive statistics on patient demographics and medical comorbidities

	Present Study	RE‐LY	ROCKET‐AF	ARISTOTLE
Patient population	LVSD (LVEF ≤40%)	HF (NYHA Class ≥II)	HF (Clinical HF or LVEF ≤40%)	LVSD (LVEF ≤40%)
Sample size, n	287	1640	4530	2736
Age at index, y	68.2 ± 12.5	68.0 ± 10.5	72 (65–78)	68 (60–74)
Male sex	241 (84.0)	1061 (64.7)	2757 (60.9)	2153 (77)
Caucasian	265 (92.3)	—	—	—
LVEF, %	29.3 ± 15.4	—	—	35 (30–39)
LVEF ≤40%	287 (100)	429 (44.0)	1043 (33.3)	2736 (100)
CAD	111 (38.7)	522 (31.8)	942 (20.8)	1172 (43)
HTN	200 (69.7)	1230 (75.0)	4202 (92.8)	2059 (75)
DM	96 (33.4)	458 (27.9)	1916 (42.3)	736 (27)
CKD stage
3	83 (28.9)	—	—	442 (16)
4	21 (7.3)	—	—	52 (2)
5/ESRD	8 (2.8)	—	—	—
CrCl <50 mL/min	—	362 (22.3)	—	—
CrCl	—	—	68 (37)	—
AF	232 (80.1)	1640 (100)	4530 (100)	2736 (100)
Follow‐up, d	313.3 ± 52.3	—	—	—

Abbreviations: AF, atrial fibrillation; ARISTOTLE, Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation; CAD, coronary artery disease; CKD, chronic kidney disease; CrCl, creatinine clearance; DM, diabetes mellitus; ESRD, end‐stage renal disease; HF, heart failure; HTN, hypertension; IQR, interquartile range; LVEF, left ventricular ejection fraction; LVSD, left ventricular systolic dysfunction; NYHA, New York Heart Association; RE‐LY, Randomized Evaluation of Long‐Term Anticoagulation Therapy; ROCKET‐AF, Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared With Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation; SD, standard deviation.

Data are presented as n (%), mean ± SD, or median (IQR).

3.2. Clinical outcomes

For efficacy outcomes, out of a total of 246 person‐years of follow‐up, there were 3 ischemic strokes and 2 SE events, with a calculated incidence rate of 1.2 (95% CI: 0.25‐3.56) and 0.81 (95% CI: 0.10‐2.94) events per 100 person‐years, respectively. Overall, the incidence rate of combined stroke and SE was 2.0 (95% CI: 0.66‐4.74). For the safety outcomes, there were 6 GI bleeding events and 1 ICH event, with a calculated incidence rate of 2.4 (95% CI: 0.8‐5.3) and 0.41 (95% CI: 0.1‐2.2) events per 100 patient‐years, respectively. Overall, the incidence rate of all hemorrhagic events was 3.5 (95% CI: 1.4‐6.4).

The incidence rates for all clinical outcomes were compared with those of the RE‐LY, ROCKET‐AF, and ARISTOTLE trials (Table 3).

Table 3.

Comparison of efficacy and safety outcomes for DOAC therapy

	Present Study	RE‐LY	ROCKET‐AF	ARISTOTLE
Patient population	LVSD	HF + LVSD	HF	LVSD
Sample size, n	287	429	4530	2736
Incidence rate unit	Events/100 PY	%/year	Events/100 PY	Events/100 PY
Outcome	Incidence Rate (95% CI)a
Ischemic stroke	1.2 (0.25‐3.56)	—	1.84	—
SE	0.81 (0.10‐2.94)	—	0.17	—
Combined SSE	2.0 (0.66‐4.74)	1.23	1.99	1.53
GI hemorrhage	2.4 (0.8‐5.3)	—	—	0.95
ICH	0.41 (0.1‐2.2)	—	0.53	0.45
Combined hemorrhage	3.5 (1.4‐6.4)	—	—	—
Major or NMCR bleeding	—	2.70	14.12	5.53

Abbreviations: ARISTOTLE, Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation; CI, confidence interval; DOAC, direct oral anticoagulant; GI, gastrointestinal; HF, heart failure; ICH, intracranial hemorrhage; LVSD, left ventricular systolic dysfunction; NMCR, nonmajor clinically relevant; PY, person‐years; RE‐LY, Randomized Evaluation of Long‐Term Anticoagulation Therapy; ROCKET‐AF, Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared With Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation; SE, systemic embolism; SSE, stroke and systemic embolism.

^a95% CIs provided for the Mayo Clinic cohort.

4. DISCUSSION

Our study shows the real‐world incidence of efficacy and safety outcomes for patients with LVSD on DOACs at our institution. This represents the clinical experience at a large tertiary referral hospital system and captures higher‐risk patients with comorbid conditions that would generally exclude them from RCTs. This study investigates the use of DOACs in a general cohort of LVSD patients, who are at particular risk for thromboembolic events. Our findings are generally consistent with those in the subgroup analyses of patients with LVSD and/or HF in the RE‐LY, ROCKET‐AF, and ARISTOTLE clinical trials.11, 12, 13

We included patients who likely would have been excluded from the aforementioned RCTs, such as patients with CKD or patients on OACs for indications other than AF, such as venous thromboembolism. We have a greater proportion of patients with moderate and severe CKD as compared with patients enrolled in the major trials. In doing so, we provided a more accurate real‐world evaluation of the safety and efficacy of DOACs in the clinical setting.

Our findings in patients with advanced renal dysfunction should be interpreted with caution. The renal clearance of DOACs requires appropriate dose adjustment in patients with advanced renal dysfunction. Safe dosing for most DOACs has not been defined in CKD stage 5 and end‐stage renal disease. Furthermore, use of DOACs at doses inappropriately high or low considering a patient's renal function have been associated with a higher risk of adverse events.14 In our study, 28 out of 29 patients with severe CKD (≥stage 4) had appropriate dosing adjustments based on manufacturer guidelines. Patients with advanced renal dysfunction who are considering DOACs for anticoagulation should be informed of the alternative of using vitamin K antagonists such as warfarin, which can be monitored. They should be informed of the increased risk of adverse events with DOACs as they make informed decisions regarding their anticoagulation therapy.

Although direct comparisons between our data and the data from the RE‐LY, ROCKET‐AF, and ARISTOTLE clinical trials cannot be made due to differences in study design and patient populations, we were able to identify trends that suggest that the incidence of clinical efficacy and safety outcomes may be similar among patients with LVSD on DOAC therapy. For ischemic stroke and SE, the incidence rate for our cohort (2.0 [95% CI: 0.66‐4.74] events per 100 person‐years) is similar to incidence rates of the RE‐LY, ROCKET‐AF, and ARISTOTLE subgroup analyses (1.89% per year, 1.99 events per 100 person‐years, and 1.53 events per 100 person‐years, respectively). For trials that reported ICH and GI bleed separately, the incidences were also similar, though the incidence of GI bleeding of apixaban in the ARISTOTLE trial appeared to be nominally lower than our mixed DOAC cohort, in which most patients were on dabigatran. Due to differences in how clinically relevant bleeding was defined, no direct comparisons could be made. However, it should be noted that the incidence rates of major bleeding in the RE‐LY and ARISTOTLE trials were similar to those of our combined hemorrhage rates. It is also noteworthy that the incidence of major or nonmajor clinically relevant bleeding in patients with HF on rivaroxaban in the ROCKET‐AF trial appears to be higher than all other DOACs.

4.1. Study limitations

One of the major limitations of this study is our sample size in comparison with the major clinical trials. This is reflected in the small number of events and the wide CIs for the incidence ratios for our cohort. The small sample size also prevents statistically relevant subgroup analyses of the individual DOACs and of the severity of bleeding (major vs nonmajor clinically relevant bleeding). As a retrospective cohort study, we are able to identify associations but unable to establish causality or mechanisms. Furthermore, this study relies on existing medical records and ICD codes, and thus the ability of healthcare providers to accurately document and code for disease states. Another limitation is the inability to properly assess adherence to anticoagulation therapy, which may adversely affect outcomes.15 There may have been other potential confounding factors, such as clinical and socioeconomic factors that play an important role in the decision to pursue DOAC therapy, that were beyond the scope of the database. The demographics of the study population, who were predominantly elderly Caucasian males, may also be difficult to generalize to the wider LVSD population.

5. CONCLUSION

Real‐world clinical experience with DOACs appears to have a similar safety and efficacy as reported in subgroup analyses from RCTs. Our findings add to increasing confidence that DOACs can be safely used for stroke and SE prevention in patients with LVSD.

Tseng AS, William Schleifer J, Shen W‐K, et al. Real‐world incidence of efficacy and safety outcomes in patients on direct oral anticoagulants with left ventricular systolic dysfunction at a tertiary referral center. Clin Cardiol. 2017;40:1328–1332. 10.1002/clc.22833