Use of Direct‐Acting Oral Anticoagulants in Patients With Atrial Fibrillation and Significant Tricuspid Regurgitation

Abstract

Background

There are limited data on the efficacy and safety of direct oral anticoagulants (DOACs) in patients with atrial fibrillation with significant tricuspid regurgitation (TR), which can lead to hepatic dysfunction and intestinal malabsorption. We aimed to compare the efficacy and safety of DOACs and warfarin for patients with atrial fibrillation with significant (moderate to severe) TR.

Methods and Results

A total of 1215 patients with significant TR and atrial fibrillation who were treated with warfarin (N=491) or DOACs (N=724) were retrospectively analyzed. The primary outcomes were ischemic stroke, systemic embolic events, and hospitalization for major bleeding. The secondary outcomes were intracranial hemorrhage, hospitalization for gastrointestinal bleeding, all‐cause mortality, and a composite outcome. The median follow‐up duration was 2.4 years. In the inverse probability treatment weighting–adjusted cohort, DOACs and warfarin had a similar risk for ischemic stroke and systemic embolic events (adjusted hazard ratio [aHR], 0.95 [95% CI, 0.67–1.36]; P=0.79) and major bleeding (aHR, 0.78 [95% CI, 0.57–1.06]; P=0.11). For the secondary outcomes, relative to warfarin, DOACs had a lower risk of intracranial hemorrhage and the composite outcome, and a comparable risk for gastrointestinal bleeding and all‐cause mortality. In the subgroup analysis, the effects of DOACs on ischemic stroke and systemic embolic events were comparable to the effects of warfarin, even in patients with inferior vena cava plethora (increased right atrial pressure) or severe TR.

Conclusions

In this study, relative to warfarin, DOACs demonstrated comparable efficacy for ischemic stroke and systemic embolic events and major bleeding, with a lower intracranial hemorrhage risk in patients with significant TR and atrial fibrillation, indicating their effectiveness and safety.

Nonstandard Abbreviations and Acronyms

DOAC

direct oral anticoagulant

ICH

intracranial hemorrhage

IPTW

inverse probability treatment weighting

IS/SE

ischemic stroke and systemic embolic event

IVC

inferior vena cava

TR

tricuspid regurgitation

Clinical Perspective

What Is New?

• Direct oral anticoagulants (DOACs) are the treatment of choice for anticoagulation in nonvalvular atrial fibrillation, showing noninferior efficacy and superior safety compare with warfarin. However, for patients with significant tricuspid regurgitation, which can lead to hepatic dysfunction and intestinal malabsorption, there are limited data comparing DOACs and warfarin.

• In this retrospective observational study, DOACs and warfarin had similar risks for ischemic stroke/embolic events and major bleeding, whereas for intracranial hemorrhage, DOACs had a lower risk than warfarin.

• Between on‐label dose DOACs and off‐label underdose DOACs, there were no significant differences in the risks for ischemic stroke/embolic events or major bleeding.

What Are the Clinical Implications?

• The prevalence of significant tricuspid regurgitation is increasing as the age of the population increases, and for patients with significant tricuspid regurgitation and nonvalvular atrial fibrillation, DOACs and warfarin showed comparable efficacy and safety.

• In subgroup analyses, DOACs were effective and safe anticoagulants, even in patients with inferior vena cava plethora (increased right atrial pressure) or severe tricuspid regurgitation, which can lead to hepatic dysfunction and splanchnic congestion.

In a retrospective study including all adults who underwent echocardiography between 1990 and 2000 at a single US center, the prevalence of significant (moderate or greater) tricuspid regurgitation (TR) was 0.55%, and it increased with age; its prevalence was estimated at 5% in the population >75 years old. ¹ In a UK cohort study, among 2500 subjects >65 years old, 2.7% of them had moderate to severe TR. ² The majority of cases of significant TR arise secondary to causes dilating the right atrium and right ventricle, including left‐sided valvular disease, left ventricular dysfunction, and chronic atrial fibrillation (AF). The late stages of severe TR manifest as right heart failure, with fatigue, peripheral edema, and hepatomegaly. ³ Additionally, significant TR can contribute to renal and hepatic dysfunction, by elevation of the central venous pressure and splanchnic congestion. This hepatic dysfunction can increase the chance of coagulopathy and cause a bleeding tendency. ⁴ , ⁵ Also, an elevation of central venous pressure can induce intestinal congestion, resulting in increased gut permeability and bacterial translocation, which can lead to malabsorption in the gastrointestinal tract. ⁶ , ⁷

The safety and efficacy of direct oral anticoagulants (DOACs) in patients with AF were well established by 4 landmark randomized controlled trials (RCTs), ⁸ , ⁹ , ¹⁰ , ¹¹ and DOACs subsequently became the main anticoagulant therapy in patients with nonvalvular AF. In patients with nonvalvular AF and significant valvular heart disease, DOACs also showed consistent efficacy and safety in the subgroup analyses of these 4 RCTs. ¹² , ¹³ , ¹⁴ , ¹⁵ However, there are limited data on DOAC use in patients with significant TR with AF, especially in cases with hepatic or renal dysfunction due to elevated central venous pressure caused by significant TR. Therefore, in the present study, we focused on significant TR with AF and compared the efficacy and safety of DOACs and warfarin in these patients.

METHODS

Data Sources

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Study Design and Population

This retrospective study was conducted at Asan Medical Center. We included patients with significant (moderate or severe) TR with nonvalvular AF seen from January 2010 to December 2020. The exclusion criteria were patients who underwent (1) percutaneous mitral balloon valvuloplasty or valve surgery, (2) percutaneous left atrial appendage closure device insertion, or patients with (3) rheumatic mitral stenosis, (4) congenital heart disease, (5) moderate to severe aortic or mitral valve disease, (6) constrictive pericarditis, (7) cardiac amyloidosis, (8) idiopathic pulmonary hypertension, (9) other diseases requiring anticoagulation such as pulmonary thromboembolism, (10) prescription of oral anticoagulants for <1 month, (11) end‐stage renal disease requiring dialysis, or (12) chronic obstructive pulmonary disease. The study population flow chart is shown in Figure 1. The institutional review board of Asan Medical Center approved this study protocol (number: 2020‐1872), and the need for informed consent was waived, because the study was retrospective. Our investigations were performed in accordance with the Declaration of Helsinki.

Figure 1. Flow diagram of patient selection and grouping.

The number of patients enrolled in the present study are shown. AF indicates atrial fibrillation; AR, aortic regurgitation; AS, aortic stenosis; ASD, atrial septal defect; AVR, aortic valve replacement; COPD, chronic obstructive pulmonary disease; CTEPH, chronic thromboembolic pulmonary hypertension; DOACs, direct oral anticoagulants; ESRD, end‐stage renal disease; HTN, hypertension; LA, left atrial; MR, mitral regurgitation; MS, mitral stenosis; MV, mitral valve; MVR, mitral valve replacement; PDA, patent ductus arteriosus; PMV, percutaneous mitral valvotomy; PR, pulmonary regurgitation; TAP, tricuspid annuloplasty; TE, thromboembolism; TGA, transposition of great arteries; TOF, tetralogy of Fallot; TR, tricuspid regurgitation; TV, tricuspid valve; and VSD, ventricular septal defect.

Clinical and Laboratory Data

The baseline clinical covariates were age, sex, weight, body mass index, and comorbidities, including hypertension, diabetes, chronic kidney disease, liver disease, heart failure, cancer; and any history of stroke, coronary intervention, percutaneous transluminal angioplasty, intracranial hemorrhage (ICH), gastrointestinal bleeding, major bleeding; or the presence of a permanent pacemaker. CHA₂DS₂‐VASc scores were also calculated: congestive heart failure, hypertension, diabetes, vascular disease, age of 65 to 74 years, and female sex were each assigned 1 point, and an age of ≥75 years and a prior stroke or transient ischemic attack were each assigned 2 points. ¹⁶ Antiplatelet medication (aspirin, clopidogrel, prasugrel, or ticagrelor) during the follow‐up period was evaluated. Laboratory data, including hemoglobin levels, platelet counts, creatinine clearance (Cockcroft‐Gault), total cholesterol, liver function test results (aspartate aminotransferase, alanine transaminase, total bilirubin, alkaline phosphatase, and γ‐glutamyl transferase), and B‐type natriuretic peptide were also included.

Echocardiographic Data

Baseline echocardiographic parameters at the time of TR diagnosis were included: left ventricular (LV) dimension in diastole/systole, LV posterior wall thickness, interventricular septum thickness in diastole, LV mass index, left atrium, ejection fraction, TR grade, tricuspid valve peak velocity, and the presence of inferior vena cava (IVC) plethora. The LV dimension in diastole, LV posterior wall thickness, interventricular septum thickness in diastole, and the left atrium were measured in the parasternal long‐axis view. Using these values, the LV mass was calculated using the linear method cube formula and was divided by the body surface area to calculate the LV mass index. ¹⁷ The LV ejection fraction was measured using the biplane Simpson volumetric method, combining apical 4‐ and 2‐chamber views. ¹⁷ Moderate to severe TR was graded using the echocardiographic criteria of the 2017 European Society of Cardiology. ¹⁸ A group with a peak TR velocity of ≥3.4 m/s is considered to have a high probability of pulmonary hypertension according to a 2015 European Society of Cardiology review article. ¹⁹ IVC plethora is defined as a <50% decrease in the IVC diameter after a deep inspiration, and it suggests increased right atrial pressure and a chance of hepatic congestion. ²⁰ , ²¹ The right atrial area was measured in the apical 4‐chamber view at end‐systole by a 2‐dimensional method. Also, right ventricular dysfunction was defined as a fractional area change <35%. ¹⁷

Clinical Outcomes and Follow‐Up

The primary efficacy and safety outcomes were clinical outcomes to compare the effectiveness and safety of DOACs versus warfarin: ischemic stroke (IS) and systemic embolic events (SEs), and hospitalization for major bleeding. A major bleeding event was defined as fatal bleeding, symptomatic bleeding in a critical organ, or bleeding either causing a fall in hemoglobin levels of >2 g/dL or leading to the transfusion of ≥2 units of whole blood or red cells, as defined by the International Society on Thrombosis and Hemostasis. ²² Secondary efficacy outcomes were all‐cause mortality and a composite outcome (IS/SEs+hospitalization for major bleeding+all‐cause mortality). Secondary safety outcomes were ICH and gastrointestinal bleeding. ICH was defined as any bleeding within the intracranial vault, including bleeding into the brain parenchyma and surrounding meningeal spaces. ²³ The index date was the date of initial warfarin or DOAC prescription. Patients were censored at the time of the outcome event, valve surgery, or last follow‐up period (September 2021). In cases with an anticoagulant change, if we ensured the separation of the data for each anticoagulant period, the patients were included twice, once for each treatment period.

Statistical Analysis

The propensity score method was used for comparisons between the warfarin and DOAC treatment groups. The propensity score of being in the warfarin or the DOAC group was assessed by a logistic regression model, which included the following baseline characteristics: age, weight, diastolic blood pressure, CHA₂DS₂‐VASc score, creatinine clearance, hemoglobin, total cholesterol, low‐density lipoprotein, total bilirubin, albumin, left atrial diameter, LV mass index, ejection fraction, peak TR velocity, and right atrial area as continuous variables, and sex, antiplatelet medication, heart failure, hypertension, diabetes, chronic kidney disease, liver disease, history of stroke, history of coronary intervention, history of percutaneous transluminal angioplasty, history of major bleeding, cancer, presence of pacemaker, TR grade (moderate or severe), presence of IVC plethora, and right ventricular dysfunction as categorical variables. Variables related to IS/SEs and major bleeding with clinical relevance or a P value <0.1 in univariable analysis were included. Based on the calculated propensity score, inverse probability treatment weighting (IPTW) was used to balance covariates between the 2 treatment groups in the total study population and the severe TR group. ²⁴ The C statistics for the propensity score models of the total population and patients with severe TR were 0.69 and 0.75, respectively. The balance of covariates between the 2 groups was assessed by standardized mean differences. A standardized mean difference ≤0.1 is considered a good balance between 2 groups. ²⁵ The balance of the baseline covariates before and after weighting in the total population is presented (Table S1). A weighted Cox proportional hazard regression model was used for clinical outcome analysis, and the proportional hazards assumption was confirmed by examination of the log (−log [survival]) curves and by testing of partial (Schoenfeld) residuals, and no significant violations were found. The hazard ratios of the DOACs group for clinical outcomes were calculated using the warfarin group as a reference.

In a subgroup of patients with severe TR, the warfarin and DOAC groups were also evaluated using the propensity score method (Table S2).

Statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC) and R software version 4.0.5 (R Foundation for Statistical Computing, Vienna, Austria). P values <0.05 indicated significance.

Subgroup Analysis

To compare efficacy and safety within the DOAC group, we divided patients by dose regimen, as on‐label dose, and off‐label underdose groups. Individuals on full‐dose DOAC therapy (5 mg apixaban, 20 mg rivaroxaban, 60 mg edoxaban, or 150 mg dabigatran) or apixaban 2.5 mg if patients fulfilled 2 of 3 criteria (age ≥80 years, weight ≤60 kg, serum creatinine ≥1.5 mg/dL); rivaroxaban15 mg if creatinine clearance was ≤50 mL/min; edoxaban 30 mg if weight was ≤60 kg or creatinine clearance was ≤50 mL/min; and dabigatran 110 mg if age was ≥80 years and creatinine clearance ≥30 mL/min, were classified as on‐label dose users. ²⁶ Individuals taking a reduced dose without fulfilling the above criteria were classified as off‐label underdose users. Each treatment group was also rebalanced using IPTW by the calculated propensity score (Table S3). Analysis according to multiple groups, including warfarin and DOAC types (dabigatran, rivaroxaban, apixaban, and edoxaban) could not be conducted because of imbalances after the use of the IPTW method. In the total study population, subgroup analysis was performed according to the presence of IVC plethora, TR grade, age strata (<65, 65–74, and ≥75 years), sex, body weight (<60 and ≥60 kg), liver function (total bilirubin <2 and ≥2 mg/dL), and renal function (creatinine clearance <50 and ≥50 mL/min). Subgroup analysis was performed using weighted Cox proportional hazard models in a well‐balanced total population cohort, and P for interaction was also calculated.

RESULTS

Baseline Characteristics

A total of 1215 patients with moderate to severe TR and nonvalvular AF, of whom warfarin (n=491) or DOACs (n=724) were prescribed, were included. Among the patients on DOACs, dabigatran was prescribed to 9.7% (n=70), rivaroxaban to 30.8% (n=223), apixaban to 33.4% (n=242), and edoxaban to 26.1% (n=189). Among the patients on DOACs, off‐label underdoses were prescribed to 34.4% (n=249). Time in the therapeutic range by the traditional method in the warfarin group was 57.5%.

Before propensity score weighting, DOAC users were older and had less frequent history of heart failure. For the echocardiographic data, DOAC users tended to have a less severe TR grade, lower LV mass index, smaller left atrial diameter, higher ejection fraction, and less IVC plethora. For laboratory data, DOAC users had a lower total bilirubin (Table 1). After propensity score weighting, the 2 treatment groups were well‐balanced in terms of the baseline covariates (Table S1).

Table 1.

Baseline Characteristics of Patients on Warfarin or DOACs

Characteristics (total N=1215)	Unadjusted
	Warfarin (N=491)	DOACs (N=724)	SMD
Men	245 (49.9)	334 (46.1)	0.075
Age, y	71.5±10.0	74.1±9.4	0.269
Antiplatelet drugs	80 (16.3)	104 (14.4)	0.054
Severe TR	136 (27.7)	120 (16.6)	0.289
CHA2DS2‐VASc score	3.0±1.5	3.2±1.5	0.163
Heart failure	108 (22.0)	102 (14.1)	0.207
Hypertension	304 (61.9)	482 (66.6)	0.097
Diabetes	105 (21.4)	175 (24.2)	0.066
Liver disease	72 (14.7)	80 (11.0)	0.108
Chronic kidney disease	275 (56.0)	344 (47.5)	0.171
History of stroke	121 (24.6)	154 (21.3)	0.08
History of PCI/CABG	51 (10.4)	94 (13.0)	0.08
History of PTA	11 (2.2)	11 (1.5)	0.053
History of ICH	7 (1.4)	17 (2.3)	0.068
History of gastrointestinal bleeding	10 (2.0)	15 (2.1)	0.002
History of other bleeding	2 (0.4)	3 (0.4)	0.001
History of major bleeding	18 (3.7)	35 (4.8)	0.058
Cancer			0.052
Active	15 (3.1)	29 (4.0)
In remission	48 (9.8)	70 (9.7)
Permanent pacemaker	25 (5.1)	26 (3.6)	0.074
SBP, mm Hg	128±55	126±19	0.035
DBP, mm Hg	73±11	75±13	0.153
Weight, kg	62.4±10.8	63.4±11.4	0.089
Body mass index, kg/m2	24.3±3.6	24.7±3.5	0.103
Echocardiographic data
LVID, diastolic, mm	49.6±6.06	49.0±5.7	0.107
LVID, systolic, mm	33.8±7.2	32.8±6.3	0.146
LVMI, g/m2	102.9±22.8	97.6±23.4	0.231
IVS, diastolic, mm	9.6±1.8	9.4±1.6	0.131
LVPW, diastolic, mm	9.5±1.3	9.3±2.2	0.121
Left atrium, mm	49.6±6.7	48.6±6.7	0.154
EF, %	54.5±11.1	56.6±9.8	0.200
EF <45%	84 (17.1)	74 (10.2)	0.202
TR Vmax, m/s	2.9±0.4	2.9±0.4	0.07
TR Vmax ≥3.4	57 (11.6)	75 (10.4)	0.04
IVC plethora	146 (29.8)	149 (20.6)	0.212
Right atrium, cm2	31.9±14.7	30.9±17.9	0.059
Right ventricle dysfunction	184 (37.5)	245 (33.8)	0.076
Fractional area change	35.6±7.1	36.2±6.8	0.083
Laboratory data
Hemoglobin, g/L	12.8±2.1	12.8±2.1	0.007
Platelets, 103/μg	197±68	196±67	0.018
Creatinine clearance, mg/dL	62±24	63±24	0.037
Total cholesterol, mg/dL	150.6±35.8	147.6±36.2	0.082
LDL, mg/dL	96±29	97±31	0.019
Albumin, g/dL	3.7±0.6	3.7±0.5	0.014
AST, IU/L	32±25	39±183	0.057
ALT, IU/L	24±25	30±112	0.072
ALP, IU/L	81±33	80±53	0.023
γ‐GTP, IU/L	59±71	56±87	0.034
Total bilirubin, mg/dL	1.1±0.8	0.9±0.6	0.224
Log‐transformed BNP,  pg/mL	5.5±1.0	5.5±1.0	0.022

Data are presented as mean±SD or number (percent). γ‐GTP indicates γ‐glutamyl transferase; ALP, alkaline phosphatase; ALT, alanine transaminase; AST, aspartate aminotransferase; BNP, B‐type natriuretic peptide; CABG, coronary artery bypass grafting; DBP, diastolic blood pressure; DOAC, direct oral anticoagulant; EF, ejection fraction; ICH, intracranial hemorrhage; IPTW, inverse propensity treatment weighting; IVC, inferior vena cava; IVS, interventricular septum thickness; LDL, low‐density lipoprotein; LVID, left ventricular internal dimension; LVMI, left ventricular mass index; LVPW, left ventricular posterior wall; PCI, percutaneous coronary intervention; PTA, percutaneous transluminal angioplasty; SBP, systolic blood pressure; SMD, standard mean difference; TR, tricuspid regurgitation; and TR V_max, peak velocity of tricuspid regurgitation.

Clinical Outcomes in Patients With Moderate to Severe TR

The median duration of follow‐up was 2.4 years (interquartile range, 1.0–4.1 years). The incidence rate of all the clinical outcomes is shown in Table 2. Nine cases of major bleeding events, other than ICH and gastrointestinal bleeding, were reported (4 cases of muscle hematoma, 1 each of hemarthrosis, hemoptysis, hematuria, vaginal bleeding, and chest tube bleeding). Six cases of systemic embolic events were reported (2 renal infarctions, 1 superior mesenteric infarction, 1 splenic infarction, 3 acute limb events). Compared with warfarin (reference), DOACs had a comparable risk of IS/SEs (adjusted hazard ratio [aHR], 0.94 [95% CI, 0.66–1.34]; P=0.72) and major bleeding (aHR, 0.77 [95% CI, 0.57–1.05]; P=0.09). For the secondary outcomes, DOACs had a lower risk of ICH (aHR, 0.27 [95% CI, 0.14–0.53]; P<0.001) and the composite outcome (aHR, 0.80 [95% CI, 0.66–0.97]; P=0.02) compared with that of warfarin. DOACs showed similar risks of gastrointestinal bleeding (aHR, 1.14 [95% CI, 0.77–1.68]; P=0.52) and all‐cause mortality (aHR, 0.87 [95% CI, 0.64–1.19]; P=0.38) (Table 2). The weighted incidence curves of the primary and secondary outcomes in the moderate to severe TR group are shown in Figure 2 and Figure S1.

Table 2.

Incidence Rates and HRs for Clinical Outcomes With Warfarin Versus DOACs in the Total Study Population

Clinical outcomes (N=1215)	Warfarin	DOACs	Unadjusted			IPTW‐adjusted
	(N=491)	(N=724)	HR	95% CI	P value	HR*	95% CI	P value
Stroke and systemic embolism	31 (2.10%)	31 (1.64%)	0.76	0.46–1.25	0.28	0.94	0.66–1.34	0.72
Major bleeding	39 (2.66%)	43 (2.31%)	0.87	0.56–1.35	0.54	0.77	0.57–1.05	0.09
Intracranial hemorrhage	16 (1.06%)	7 (0.37%)	0.33	0.13–0.81	0.02	0.27	0.14–0.53	<0.001
Gastrointestinal bleeding	19 (1.28%)	32(1.70%)	1.33	0.74–2.38	0.34	1.14	0.77–1.68	0.52
All‐cause mortality	41 (2.70%)	42 (2.18%)	0.81	0.52–1.26	0.35	0.87	0.64–1.19	0.38
Composite outcome†	99 (6.92%)	102 (5.55%)	0.77	0.58–1.02	0.06	0.80	0.66–0.97	0.02

The incidence rate is presented as the number of total events per 100 patient‐years. HR was adjusted for age, weight, diastolic blood pressure, CHA₂DS₂‐VASc score, creatinine clearance, hemoglobin, total cholesterol, low‐density lipoprotein, total bilirubin, albumin, left atrial diameter, left ventricular mass index, ejection fraction, and peak tricuspid regurgitation velocity as continuous variables, and sex, antiplatelet medication, heart failure, hypertension, diabetes, history of stroke, history of coronary intervention, history of percutaneous transluminal angioplasty, history of major bleeding, cancer, presence of pacemaker, tricuspid regurgitation grade, and presence of inferior vena cava plethora as categorical variables. DOAC indicates direct oral anticoagulant; HR, hazard ratio; and IPTW, inverse probability treatment weighting.

^*Adjusted HR.

^† Composite=all‐cause mortality+stroke and systemic embolism+major bleeding.

Figure 2. Weighted cumulative incidence curves of the primary outcomes for patients with moderate to severe TR treated with DOACs or warfarin.

A, Stroke and systemic embolism. B, Major bleeding. DOAC indicates direct oral anticoagulant; and TR, tricuspid regurgitation.

Clinical Outcomes in Patients With Severe TR

In the total study population, 21% of patients (n=256) were classified as having severe TR; of these, 53% (n=136) were on warfarin and 47% (n=120) on DOACs. Before IPTW, the DOAC users tended to be older, have higher CHA₂DS₂‐VASc scores, less heart failure, more diabetes, more hypertension, more likely to have a history of ICH, less IVC plethora, and a lower total bilirubin level (Table S2).

The distribution of patients with severe TR before and after IPTW is presented in Table S2. After IPTW analysis, the key baseline covariates were well balanced. There was 1 systemic embolic event (a renal infarction) and 3 bleeding events (2 muscle hematomas and 1 hemarthrosis) in the severe TR group. The incidence rates and hazard ratios of the clinical outcomes are shown in Table S4. In the severe TR group, DOAC users had a comparable risk with warfarin users for primary and secondary outcomes: IS/SEs (aHR, 1.06 [95% CI, 0.53–2.13]; P=0.86), major bleeding (aHR, 1.01 [95% CI, 0.55–1.86]; P=0.98), ICH (aHR, 0.71 [95% CI, 0.18–2.76]; P=0.62), gastrointestinal bleeding (aHR, 1.10 [95% CI, 0.50–2.44]; P=0.82), all‐cause mortality (aHR, 0.77 [95% CI, 0.37–1.60]; P=0.48), and the composite outcome (aHR, 0.88 [95% CI, 0.59–1.32]; P=0.53). The weighted incidence curves of the primary outcomes in the severe TR group are shown in Figure S2.

Subgroup Analysis: DOAC Doses

Among DOAC users (n=724), 34% were prescribed off‐label underdoses (n=252) and 66% on‐label doses (n=478). Baseline characteristics and echocardiographic data according to DOAC dose, off‐label underdose, and on‐label dose are presented in Table S3. After weighting, key baseline covariates were relatively well balanced with standardized mean difference, although some variables were slightly higher than 0.1 (Table S3). The incidence rate of the clinical outcomes according to DOAC dose, off‐label underdose, and on‐label dose are shown in Table S5. Both the on‐label dose DOAC and off‐label underdose DOAC showed a lower risk for ICH compared with that for warfarin (aHR, 0.15 [95% CI, 0.06–0.38]; P<0.001; and aHR, 0.44 [95% CI, 0.23–0.85]; P=0.02; respectively). On‐label dose DOAC users had a better composite outcome than warfarin (aHR, 0.75 [95% CI, 0.60–0.93]; P=0.01; Table S6). On‐label dose DOAC and off‐label underdose DOAC users showed similar risks for the primary and secondary outcomes, except for ICH. For intracerebral hemorrhage, patients with on‐label dose DOACs tended to be at lower risk than those with off‐label underdose DOACs (aHR, 0.33 [95% CI, 0.11–0.95]; P=0.04). The weighted incidence curves of the primary outcomes in the total group, according to DOAC dose and warfarin, are shown in Figure S3.

Subgroup Analyses: IVC Plethora, TR Grade, Age, Sex, Body Weight, Total Bilirubin, and Creatinine Clearance

The crude incidences of the clinical outcomes, according to treatment with DOACs or warfarin in various subgroups, are presented in Table S7. The aHR for the primary outcomes, according to several groups, are shown in Table 3. Interaction with treatment was significant for IVC plethora (major bleeding), TR grade (major bleeding), and body weight (stroke and embolic events, major bleeding). DOACs tended to be more effective in patients who weighed <60 kg (aHR, 0.53 [95% CI, 0.31–0.91]; P for interaction=0.01) and safer in groups with weight <60 kg (aHR, 0.53 [95% CI, 0.33–0.86]; P for interaction=0.03), groups with no IVC plethora (aHR, 0.61 [95% CI, 0.43–0.89]; P for interaction=0.03), and groups with moderate TR grades (aHR, 0.66 [95% CI, 0.47–0.94]; P for interaction=0.04) (Figures S4 and S5). The aHR for secondary outcomes, according to several groups, are shown in Table S8.

Table 3.

Hazard Ratios for Primary Outcomes With Direct Oral Anticoagulants Versus Warfarin by Subgroup

Subgroups	IS/SE		Major bleeding
	aHR (95% CI)	P value*	aHR (95% CI)	P value*
IVC plethora		0.61		0.03
No	1.00 (0.64–1.57)		0.61 (0.43–0.89)
Yes	0.82 (0.46–1.45)		1.38 (0.76–2.50)
TR grade		0.13		0.04
Moderate	0.80 (0.52–1.23)		0.66 (0.47–0.94)
Severe	1.50 (0.79–2.87)		1.34 (0.70–2.56)
Age, y		0.50		0.12
<65	1.31 (0.46–3.72)		1.48 (0.51–4.28)
65–74	1.01 (0.54–1.90)		0.87 (0.51–1.49)
≥75	0.77 (0.48–1.23)		0.59 (0.39–0.88)
Sex		0.70		0.43
Men	1.05 (0.52–2.13)		0.73 (0.46–1.16)
Women	0.83 (0.55–1.25)		0.72 (0.48–1.09)
Body weight		0.01		0.03
<60 kg	0.53 (0.31–0.91)		0.53 (0.33–0.86)
≥60 kg	1.43 (0.88–2.31)		0.95 (0.64–1.43)
Total bilirubin		0.15		0.24
<2 mg/dL	0.83 (0.57–1.21)		0.76 (0.55–1.03)
≥2 mg/dL	2.82 (0.78–10.16)		2.40 (0.32–18.14)
CrCL		0.89		0.69
<50 mg/dL	0.94 (0.55–1.60)		0.93 (0.59–1.47)
≥50 mg/dL	0.95 (0.59–1.52)		0.70 (0.46–1.06)

Hazard ratios were adjusted for age, weight, diastolic blood pressure, CHA₂DS₂‐VASc score, creatinine clearance, hemoglobin, total cholesterol, low‐density lipoprotein, total bilirubin, albumin, left atrial diameter, left ventricular mass index, ejection fraction, and peak TR velocity as continuous variables, and sex, antiplatelet medication, heart failure, hypertension, diabetes, history of stroke, history of coronary intervention, history of percutaneous transluminal angioplasty, history of major bleeding, cancer, presence of pacemaker, TR grade, and presence of IVC plethora as categorical variables. aHR indicates adjusted hazard ratio; CrCL, creatinine clearance; IS/SE, ischemic stroke and systemic embolic event; IVC, inferior vena cava; and TR, tricuspid regurgitation.

^* P for interaction.

DISCUSSION

We compared the efficacy and safety of DOACs and warfarin in patients with moderate to severe TR and nonvalvular AF on anticoagulants. The main findings were as follows: (1) DOACs and warfarin had similar risks for IS/SEs and major bleeding. (2) For secondary outcomes, DOACs had a lower risk for ICH and the composite outcome compared with warfarin. (3) On‐label dose DOACs and off‐label underdose DOACs both showed reduced risk for ICH compared with warfarin. On‐label dose DOACs and off‐label dose DOACs had a similar risk for the primary and secondary outcomes. (4) In subgroup analyses, DOACs tended to be safer in the group with weight <60 kg. Also, even in the patients who had IVC plethora and severe TR resulting in hepatic dysfunction and intestinal malabsorption, DOACs can be a comparable treatment choice to warfarin, whereas DOACs had a lower risk for major bleeding in patients without severe TR and IVC plethora.

There are several meta‐analyses and subgroup studies that evaluated the efficacy and safety of DOACs in patients with valvular heart disease. ¹² , ¹³ , ¹⁴ , ¹⁵ , ²⁷ , ²⁸ Previous studies showed that there was no significant difference between DOACs and warfarin in terms of stroke and systemic embolism prevention in patients with AF with valvular heart disease. Additionally, for major bleeding, DOACs, except rivaroxaban, did not show a significant difference compared with warfarin in patients with AF with valvular heart disease. Therefore, these studies suggested that DOACs are a reasonable alternative to warfarin in patients with AF with valvular heart disease. ¹² , ¹³ , ¹⁴ , ¹⁵

However, these results should be interpreted with caution because the inclusion criteria of these studies had marked differences in the definition of valvular heart disease. The subgroup studies of RCTs included patients with at least moderate valvular heart disease, most of which was mitral regurgitation. Also, only the ARISTOTLE (Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation) and RE‐LY (Randomized Evaluation of Long‐Term Anticoagulation Therapy) subgroup studies included patients with TR, accounting for 44.2% and 29.8%, respectively. ¹³ , ¹⁴ The other 2 trials excluded patients with TR because it is less likely to be associated with arterial thromboembolic risk and only included mitral regurgitation, aortic stenosis, and aortic regurgitation. Therefore, the application of these results to patients with significant TR and AF has limitations.

The present study was limited to patients with moderate to severe TR with nonvalvular AF and excluded other significant valvular heart disease. The incidence rates of the clinical outcomes were comparable with those of 4 previous landmark RCTs. ⁸ , ⁹ , ¹⁰ , ¹¹ For treatment outcomes, DOACs had comparable IS/SE and major bleeding risks, but a significantly lower risk for ICH. Other than the lower risk for IS/SE with apixaban and dabigatran 150 mg in 2 RCT substudies, ¹³ , ¹⁴ this study result is similar to that of previous substudies. ¹² , ¹³ , ¹⁴ , ¹⁵ In the severe TR group, DOACs had comparable IS/SEs, major bleeding, and secondary outcomes, including ICH. It is uncertain why the favorable risk profile of DOACs was not observed in the severe TR group. The comparable risk profile of DOACs and warfarin can be attributed to renal and hepatic dysfunction due to prolonged splanchnic congestion in severe TR. Further studies with large sample sizes are needed for severe TR.

In subgroup analysis by DOAC dose, both on‐label and off‐label dose DOACs were at a lower risk for intracranial hemorrhage compared with warfarin. Between on‐label dose and off‐label underdose DOACs, on‐label dose DOACs showed a similar risk for the primary and secondary outcomes, except for ICH. For ICH, on‐label dose DOACs tended to be at lower risk than off‐label dose DOACs; however, because of the small number of events (3 and 4), this needs to be interpreted cautiously.

In subgroup analysis, in the low‐weight group (<60 kg), DOACs had a reduced risk for IS/SEs and major bleeding compared with warfarin. The effects of DOACs are intricately linked to their plasma concentration, which is affected by body distribution volume. Consequently, low body weight may affect the efficacy and safety of DOACs. In a previous study of patients with AF without valvular heart disease, the use of DOACs in low body weight (<60 kg) was associated with lower risk of IS/SEs, ICH, hospitalization for gastrointestinal bleeding, and major bleeding compared with the use of warfarin. ²⁹ Our study also showed these findings were consistent in patients with AF with significant TR.

Furthermore, the findings of our current study demonstrated that the effect of DOACs and warfarin on IS/SEs did not differ between patients with and without IVC plethora or severe TR. In patients with IVC plethora, which suggests high right atrial pressure, or severe TR, splanchnic and hepatic congestion can lead to hepatic dysfunction (increase the chance of coagulopathy) and intestinal malabsorption (decrease in drug concentration). This can affect hepatic drug clearance and drug metabolism, ³⁰ , ³¹ potentially influencing the efficacy and safety of DOACs. Although our data suggest that DOACs tend to be safer in groups without IVC plethora or moderate TR in terms of major bleeding, there was no significant interaction between the anticoagulation regimen and IS/SEs, ICH, gastrointestinal bleeding, death, or the composite outcome based on the TR grade and the presence of IVC plethora. This result suggests that DOACs can be considered as a comparable treatment choice to warfarin even in patients with severe TR and the presence of IVC plethora.

The prevalence of moderate to severe TR is estimated at 5% in the population >75 years old. ¹ Considering that most cases of TR are secondary to AF, the incidence of TR will increase. There are limited data on DOAC efficacy and safety in these patients. This study suggests that DOACs are also effective and safe in patients with moderate to severe TR.

Time in the therapeutic range by the conventional method of warfarin in this study was 57.5%. In the 4 landmark randomized trials of DOACs in patients with nonvalvular AF, the mean time in the international normalized ratio therapeutic range was 64% in the RE‐LY trial, 55% in the 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) trial, 62.2% in the ARISTOTLE trial, and 64.9% in the ENGAGE‐AF (Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation) trial. ⁸ , ⁹ , ¹⁰ , ¹¹ In a multicenter retrospective study in Korea, the mean time in the international normalized ratio therapeutic range was 49.1% for patients taking warfarin for nonvalvular AF and ischemic stroke. ³² Time in the therapeutic range of warfarin in this study is comparable with previous studies. The low quality of warfarin therapy, which is <50% of the time in the therapeutic range, can be associated with a higher thromboembolic risk. ³³ This difficulty in maintaining a constant drug therapeutic range is considered warfarin's main intrinsic limitation, which can lead to at least similar or subinferior efficacy and an inferior safety profile.

This study has several limitations. First, it is a retrospective study, and therefore latent confounding variables may have influenced the findings, and they should be considered as hypothesis generating only. Second, the study population was relatively small, so propensity‐score matching could not be applied among the DOACs. Previous studies have shown different risks among the DOACs, such as higher bleeding risk in valvular heart disease with rivaroxaban. We could not compare the efficacy and safety among the DOACs. Further large analyses of patients with significant TR would help identify the different profiles of the DOACs. Third, this study did not include a control group of patients with AF without significant TR on anticoagulation, so it was difficult to compare the effect of significant TR on patients with AF directly. However, many previous studies with patients with AF and without significant TR can be referenced. Fourth, this study population is composed of 69% moderate TR and 21% severe TR. Because of the patients with relatively less severe TR in this study population, the results need to be cautiously interpreted for patients with severe TR. Fifth, the subgroup analysis in this study is a post hoc analysis; therefore, it is considered exploratory or hypothesis generating. The result of subgroup analysis should not be generalized, especially in groups with a small number of events, such as ICH cases in on‐label dose DOACs and off label underdose DOACs. Further prospective confirmatory studies are needed to clarify. Nevertheless, our study is the first to evaluate the efficacy and safety of DOACs in patients with significant TR. This study could provide evidence that DOACs have comparable efficacy and safety in patients with significant TR.

CONCLUSIONS

In this retrospective study, in patients with significant TR and nonvalvular AF, DOACs had comparable risks for IS/SEs and major bleeding compared with warfarin. In patients with severe TR, DOACs had a similar risk for IS/SEs and major bleeding. Between on‐label dose DOACs and off‐label underdose DOACs, there was no significant difference in risk for clinical outcomes. In subgroup analysis, DOACs tended to be safer in patients who weighed <60 kg. In addition, even in patients with IVC plethora and severe TR, which can lead to hepatic dysfunction and intestinal malabsorption, DOACs can be considered comparable to warfarin as a treatment choice.

Sources of Funding

This research was supported by a grant from the Korea Health Technology Research and Development Project through the Korean Health Industry Development Institute, funded by the Ministry of Health and Welfare, Republic of Korea (HI22C0154).

Disclosures

None.

Supporting information

Tables S1–S8.

Figures S1–S5.