Prevalence, Clinical Correlates, and Prognostic Impact of Tricuspid Regurgitation in Older Adults: The ARIC Study

Abstract

Background

Tricuspid regurgitation (TR) is associated with older age and a heightened mortality rate. Limited data exist on TR prevalence and prognostic significance in community‐based older adults.

Methods

Among 3046 participants in the ARIC (Atherosclerosis Risk in Communities) study who underwent echocardiography at the seventh study visit, TR severity was assessed as none/trace, mild, moderate, or severe by a board‐certified cardiologist. Multivariable linear and logistic regression models assessed associations of TR severity with clinical characteristics, echocardiographic measures, and self‐reported dyspnea. Multivariable Cox proportional hazard models evaluated associations with incident adjudicated heart failure (HF) and death.

Results

Mean age was 81±4 years, 58% were women, and 25% reported Black race. TR prevalence was 29% mild, 7% moderate, and 1% severe. Greater TR severity was associated with older age, female sex, prevalent HF, and worse cardiac structure and function. Associations with dyspnea were not statistically different in fully adjusted models. Over a median follow‐up of 3.7 (interquartile range, 2.6–4.3) years, there were 154 incident HF events and 412 deaths. Higher TR severity was associated with heightened risk of incident HF after adjusting for demographics, comorbidities, and measures of left ventricular structure and function (hazard ratio, 1.28 [95% CI, 1.01–1.64] per increment in severity category; P=0.04). Greater TR severity was associated with all‐cause death after adjusting for demographics (P=0.003) but not after further adjustment (P=0.3).

Conclusions

TR is common in older adults and is associated with worse cardiac structure and function. Greater TR severity is independently associated with a greater risk of developing HF.

Nonstandard Abbreviations and Acronyms

ARIC

Atherosclerosis Risk in Communities

MR

mitral regurgitation

PASP

pulmonary artery systolic pressure

TR

tricuspid regurgitation

TRILUMINATE Pivotal

Clinical Trial to Evaluate Cardiovascular Outcomes in Patients Treated With the Tricuspid Valve Repair System Pivotal

TRISCEND II

Transcatheter Valve Replacement I Severe Tricuspid Regurgitation

TRJA

tricuspid regurgitation jet area

Clinical Perspective.

What Is New?

• Tricuspid regurgitation is prevalent in late life (37%) and is associated with higher comorbidity burden, worse cardiac function, and higher risk of incident heart failure beyond traditional risk factors and left ventricular function.

What Are the Clinical Implications?

• Detection of tricuspid regurgitation should prompt comprehensive evaluation of left ventricular systolic and diastolic function.

• Community‐dwelling older adults with greater tricuspid regurgitation severity may warrant enhanced heart failure surveillance and risk stratification.

Tricuspid regurgitation (TR) is associated with significant morbidity and death. ¹ , ² , ³ The estimated prevalence of moderate to severe TR in the general population is 3%, increases with age, and is ≈7% in those aged >75 years. ⁴ Milder TR is even more common, detectable in 82% of men and 86% of women. ⁵ The burden of TR therefore is expected to grow substantially as the population ages, with individuals aged >65 years anticipated to account for 20% of the US population by 2030. ⁶ The pathogenesis of TR is physiologically complex and involves altered leaflet structure, tricuspid annular size, and right‐sided hemodynamics. ⁷ Among patients clinically referred for echocardiography, greater TR is associated with female sex, atrial fibrillation, pulmonary hypertension, lower ejection fraction, and left‐sided valvular heart disease. Severe TR, especially in association with left‐sided valvular heart disease and heart failure (HF), is associated with a higher mortality rate, ⁸ , ⁹ , ¹⁰ but the clinical importance of mild and moderate TR is unclear. ⁹ , ¹¹

To our knowledge, contemporary data quantifying the burden of TR in older adults in the general population not referred for clinically indicated echocardiography is lacking. We therefore evaluated clinical and echocardiographic data from a diverse community‐based cohort of older adults aiming to (1) estimate the prevalence of TR; (2) identify clinical and echocardiographic correlates of TR; and (3) assess the association of TR severity with functional impairment, incident cardiovascular disease, and death.

Methods

Data Availability

Preexisting data access policies specify that research data requests can be submitted to the ARIC (Atherosclerosis Risk In Communities) steering committee; these will be promptly reviewed for confidentiality or intellectual property restrictions and will not be refused unreasonably. Individual‐level patient data may further be restricted by consent, confidentiality, or privacy laws/considerations.

Study Population

The ARIC study is a prospective epidemiologic cohort study. Its design and methods have been previously described. ¹² , ¹³ Between 1987 and 1989, 15 792 subjects aged 45 to 64 years were enrolled in 4 communities in the United States: Forsyth County, North Carolina; Jackson, Mississippi; suburban Minneapolis, Minnesota; and Washington County, Maryland. Participants returned for a seventh study exam (visit 7) in 2018 to 2019, at which time they underwent protocol echocardiogram and phlebotomy for laboratory testing. This analysis includes 3046 participants who attended ARIC visit 7 and underwent echocardiography (Figure 1). The study protocol was approved by institutional review boards at all participating institutions, and all participants provided written informed consent at each study visit.

Figure 1. Flowchart of study population in the ARIC study at visit 7 (2018–2019).

ARIC indicates Atherosclerosis Risk in Communities.

Echocardiography

Procedures and design for echocardiography in ARIC at visit 5, including reproducibility metrics, have been previously described and were similar to those used at visit 7. ¹⁴ Studies were prospectively acquired by certified sonographers using uniform imaging machines (Philips iE33, Koninklijke Philips, Amsterdam, the Netherlands), probes (Philips XMatrix), and acquisition protocols. Quantitative measures were conducted at a dedicated echocardiography reading center by trained analysts blinded to clinical data and in accordance with American Society of Echocardiography recommendations ¹⁵ , ¹⁶ , ¹⁷ and overread by reading center staff cardiologists with Core Cardiology Training Symposium level III advanced training in echocardiography or American Society of Echocardiography board certification in comprehensive adult echocardiography.

TR was qualitatively assessed by overreading echocardiographers on the basis of color Doppler signal, right atrial and ventricular size, and tricuspid valve structure. TR color Doppler was assessed in the parasternal right ventricular (RV) inflow view, parasternal short‐axis view at the aortic valve level, and apical 4‐chamber view focused on the RV and right atrium (RA). Its severity was classified as trace/none, mild, moderate, and severe by the overreading echocardiographers. Tricuspid regurgitation jet area (TRJA) was also measured by tracing the maximal TR color Doppler regurgitation signal in the apical 4‐chamber view.

Clinical Covariates, Functional Measures, and Biomarkers

Race and sex were self‐reported. Hypertension was defined on the basis of antihypertensive medication use or recording of blood pressure ≥140/90 mm Hg at any ARIC visit. ¹⁸ , ¹⁹ , ²⁰ Diabetes was defined on the basis of a nonfasting blood glucose level ≥ 200 mg/dL, a fasting blood glucose ≥126 mg/dL, or hemoglobin A_1C ≥6.5% at any ARIC visit. ¹⁹ , ²¹ Body mass index was calculated from weight and height assessed at visit 7. Estimated glomerular filtration rate was calculated using the 2023 Chronic Kidney Disease Epidemiology Collaboration method using plasma creatinine and cystatin‐C. ²² Chronic kidney disease was defined as estimated glomerular filtration rate <60 mL/min per 1.73m². Prevalent coronary artery disease was defined as self‐reported diagnosis before ARIC visit 1 or adjudicated myocardial infarction or revascularization between visit 1 and visit 7 as previously described. ²³ Smoking status was self‐reported. Prevalent atrial fibrillation (AF) was determined by electrographic evidence of AF and International Classification of Diseases (ICD) codes associated with AF during prior hospitalizations or death between visits 1 and 7. ²⁴ Prevalent HF was assessed on the basis of prior hospitalizations with HF‐associated ICD codes between visits 1 and 7 as previously described. ²⁵ HF stages were classified according to the 2021 Universal Definition and Classification of Heart Failure. ²⁶

NT‐proBNP (N‐terminal pro‐B‐type natriuretic peptide) was measured using electrochemiluminescent immunoassay (Roche Diagnostics, Indianapolis, IN), with a lower detection limit of ≤5 ng/mL. ²⁷ High‐sensitivity troponin T was measured using a highly sensitive assay (Elecsys Troponin T; Roche Diagnostics), and the limit of the blank was 3 ng/L. ²⁷ , ²⁸

Dyspnea was assessed at visit 7 using the modified Medical Research Council scale as described previously and graded on a scale of 0 to 4. ²⁹ , ³⁰ Participants were categorized as having no or mild dyspnea (modified Medical Research Council scale score, 0 or 1), or moderate or severe (modified Medical Research Council scale score, 2, 3, or 4) dyspnea. ³⁰ , ³¹

Incident HF, AF, and Death

As previously detailed, ARIC participants undergo surveillance through biannual telephone questionnaires, review of hospitalization discharge codes, state death records, and linkage to the National Death Index. ¹² HF events were ascertained through chart abstraction and physician adjudication by 2 independent reviewers of hospitalizations with HF‐related ICD discharge codes as previously described. ³² AF events were ascertained on the basis of hospitalizations with an AF‐related ICD code or on visit ECGs. ³³ The follow‐up for these analyses started at visit 7 and continued through December 2022, excluding prevalent events.

Statistical Analysis

TR severity prevalence was stratified by age (<75, 75–79, 80–84, 85–90, >90 years), sex, self‐reported race, mitral regurgitation (MR) severity, and HF stage. ²⁶ Cross‐sectional associations of clinical characteristics, echocardiographic measures and cardiac biomarkers with TR severity were assessed using linear regression models for continuous variables and logistic regression models for categorical variables. For age, sex, race, and field center, P values were obtained using univariable regression models (no adjustment). For all other clinical variables (Table 1), P values were obtained from regression models adjusted for age, sex, race, and field center (model 1). Models for echocardiographic measures (Table 2) were adjusted for age, sex, race, visit center, estimated glomerular filtration rate, body mass index, prevalent diabetes, prevalent hypertension, prevalent coronary artery disease, prevalent AF, and prevalent HF (model 2). NT‐proBNP and high‐sensitivity troponin T levels were log‐transformed. TR severity categories were assigned ordinal values (0–3) to test for a linear trend. This approach assumes that the categories are equally spaced in terms of effect, that is, that the increase in risk is proportional with each step in severity.

Table 1.

Characteristics of the Study Population TR Severity in the ARIC Study at Visit 7 (2018–2019)

	None/Trivial TR (n=1881)	Mild TR (n=891)	Moderate TR (n=226)	Severe TR (n=40)	P value for trend
Age, y, mean±SD	80.2±4.3	81.2±4.6	82.0±4.5	83.3±5.0	<0.001
Male sex, n (%)	887 (47)	314 (35)	68 (30)	18 (45)	<0.001
Black race, n (%)	486 (26)	201 (22)	53 (24)	9 (23)	0.09
Forsyth County, NC	467 (25)	238 (27)	67 (30)	11 (28)	0.1
Jackson, MS	445 (24)	173 (19)	48 (21)	8 (20)	0.04
Minneapolis, MN	528 (28)	320 (36)	78 (34)	14 (35)	<0.001
Washington County, MD	441 (23)	168 (19)	33 (15)	7 (18)	<0.001
Ever smoker, n (%)	1184 (63)	529 (59)	126 (56)	23 (58)	0.6
Hypertension, n (%)	1684 (90)	789 (88)	204 (90)	36 (90)	0.4
Diabetes, n (%)	775 (41)	291 (32)	73 (32)	14 (35)	0.002
Chronic kidney disease (eGFR <60), n (%)	745 (41)	379 (44)	100 (46)	19 (51)	0.7
Coronary artery disease, n (%)	350 (19)	158 (18)	43 (19)	12 (31)	0.4
Myocardial infarction, n (%)	189 (10)	71 (8)	22 (10)	6 (15)	0.8
HF, n (%)	211 (12)	122 (14)	45 (21)	12 (32)	<0.001
AF, n (%)	164 (9)	145 (16)	59 (27)	15 (39)	<0.001
Loop diuretic use, n (%)	164 (9)	96 (11)	32 (14)	15 (38)	<0.001
Body mass index, kg/m2, mean±SD	28.7±5.4	27.2±5.3	26.3±4.9	26.8±4.7	<0.001
Systolic blood pressure, mm Hg	135±19	135±19	132±19	127±19	<0.001
eGFR, mL/min−1 per −1.73 m−2, mean±SD	63.5±18.0	62.3±18.2	61.9±19.2	55.6±18.2	0.4
High‐sensitivity troponin T, ng/L, median (IQR)	13 (8.0–19.0)	13 (9.0–18.0)	13 (8.0–19.0)	16 (9.0–26.0)	0.02*
NT‐proBNP, μg/L, median (IQR)	145 (79–286)	216 (111–449)	333 (171–742)	1172 (466–1722)	0.001*

For age, sex, race, and field center, P values were obtained from univariable regression models (no adjustment). For all other variables, P values were obtained from regression models adjusted for age, sex, race, and field center (model 1). Binary variables were analyzed with logistic regression and continuous variables with linear regression. AF indicates atrial fibrillation, ARIC, Atherosclerosis Risk in Communities; eGFR, estimated glomerular filtration rate, HF, heart failure; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; and TR, tricuspid regurgitation.

^* P value used is for log‐transformed variables.

Table 2.

Echocardiographic Measures by TR Severity Category in the ARIC Study Population at Visit 7 (2018–2019)

	Overall (n=3046)	None/Trivial TR (n=1881)	Mild TR (n=891)	Moderate TR (n=226)	Severe TR (n=40)	P value
Left‐sided valvular heart disease
Moderate to severe aortic stenosis, n (%)	47 (2.5)	22 (1.2)	19 (2.1)	6 (2.7)	0	0.2
Moderate to severe aortic regurgitation, n (%)	38 (1.2)	17 (0.9)	19 (2.1)	1 (0.4)	1 (2.5)	0.3
Moderate to severe MR, n (%)	170 (5.6)	52 (2.9)	75 (8.7)	30 (13.3)	13 (32.5)	<0.001
Moderate to severe mitral stenosis, n (%)	0	0	0	0	0	NA
LV structure
LV end‐diastolic volume, mL/m2	46.0±11.0	46.2±10.9	45.8±10.8	44.9±10.2	47.6±20.0	0.4
LV mass index, g, mean±SD	81.8±21.0	81.8±20.0	81.3±21.4	81.7±22.1	91.7±33.5	0.9
LV systolic function
LV ejection fraction, mean±SD	63.4±7.7	63.7±7.2	63.3±8.0	62.7±9.1	58.2±12.9	0.002
Global longitudinal strain, %, mean±SD	−18.0±2.5	−18.0±2.4	−18.0±2.4	−18.0±2.7	−16.6±3.2	0.1
LV diastolic function
Left atrial volume index, mean±SD	27.9±9.5	26.4±8.1	29.2±9.7	33.1±12.8	42.4±15.1	<0.001
E wave, m/s, mean±SD	75.6±21.7	72.8±19.5	78.4±23.2	84.7±24.8	94.1±34.4	<0.001
Tissue Doppler imaging e′, m/s, mean±SD	6.0±1.5	5.8±1.4	6.0±1.5	6.5±1.9	7.5±2.4	<0.001
E/e′ mean ratio, mean±SD	13.3±4.7	13.1±4.5	13.6±4.7	14.0±5.6	13.3±6.3	0.1
Doppler E/A ratio, mean±SD	0.87±0.31	0.83±0.26	0.91±0.35	0.96±0.37	1.31±0.63	<0.001
RA and RV structure/function
RVEDA, cm2, mean±SD	20.0±5.3	19.8±5.0	19.7±5.2	21.3±5.9	24.7±8.1	<0.001
RA volume index, mL/m2, mean±SD	19.7±9.0	17.9±6.5	20.4±8.6	26.9±13.6	37.7±19.1	<0.001
RV fractional area change, mean±SD	0.49±0.08	0.49±0.08	0.50±0.08	0.49±0.09	0.47±0.10	0.3
PASP, mm Hg, mean±SD	32.3±8.2	29.3±6.1	33.6±7.4	40.2±10.1	45.2±13.5	<0.001
TAPSE, cm, mean±SD	20.4±5.0	20.4±4.8	20.3±5.0	20.7±5.7	20.1±6.2	0.1
TAPSE/PASP, cm/mm Hg, mean±SD	0.67±0.22	0.73±0.22	0.63±0.19	0.55±0.20	0.49±0.21	<0.001

P values were obtained from regression models adjusted for age, sex, self‐reported race, field center, diabetes, hypertension, estimated glomerular filtration rate, atrial fibrillation, body mass index, coronary artery disease, and heart failure. Binary variables were analyzed with logistic regression and continuous variables with linear regression. ARIC indicates Atherosclerosis Risk in Communities; LV, left ventricular; LVEDV, left ventricular end‐diastolic volume; MR, mitral regurgitation; PASP, pulmonary artery systolic pressure; RA, right atrial; RV, right ventricular; RVEDA, right ventricular end‐diastolic area; TAPSE, tricuspid annular plane systolic excursion; and TR, tricuspid regurgitation.

Cross‐sectional associations of TR severity with functional outcomes were assessed using multivariable logistic regression for dyspnea scores. Associations of TR severity with incident AF, HF, and death were assessed using multivariable Cox proportional hazards models, where TR severity categories were assigned ordinal values to test for a linear trend in hazard across severity levels. These analyses were adjusted for model 1 and model 2 covariates described above, in addition to model 3, which further adjusted for left ventricular (LV) ejection fraction, LV end‐diastolic volume, and E/e′ mean. The proportional hazards assumption was assessed using Schoenfeld residuals (both global and variable‐specific tests). There was no meaningful violation of the proportional hazards assumption. Participants with prevalent HF and AF were excluded from analyses of incident HF and AF, respectively. Kaplan–Meier curves were used for descriptive visualization of unadjusted cumulative incidence by TR severity. To account for potential systematic differences between participants who attended visit 7 and did not undergo an echocardiogram versus those who did, we performed inverse probability of attrition weighting analyses with model 2 covariates as predictors of echocardiogram performance. Missing data were not imputed for any analysis. Sensitivity analyses with TRJA as exposure were performed. TRJA was standardized to 0±1.

All analyses were performed using STATA 18.0 (StataCorp, College Station, TX). Two‐sided P values of <0.05 were considered significant.

Results

Prevalence and Clinical Correlates of TR

Among the 3046 included participants, mean age was 80.6±4.4 years; 58% were women, and 25% reported Black race (Table 1). Cardiovascular risk factors and comorbidities, such as hypertension, diabetes, chronic kidney disease, and smoking history were common. Based on an integrated qualitative assessment of study echocardiograms, 62% had no/trivial TR, 29% had mild TR, 7% had moderate TR, and 1% had severe TR. Individuals who attended visit 7 but did not undergo an echocardiogram were more frequently from the Washington County, Maryland, field center and were older, more frequently women, and had a higher prevalence of cardiovascular disease (including coronary artery disease, HF, and AF) and chronic kidney disease compared with included participants (Table S1), but TR prevalence with inverse probability of attrition weighting was similar.

The prevalence and severity of TR was greater among older participants, even in this late‐life cohort (Figure 2A). TR prevalence (mild, moderate, or severe) was greater among White participants compared with Black participants. White men demonstrated higher TR severity compared with Black men, but there was no statistical difference by race among women (Figure 2B). Greater TR severity was also associated with a more advanced HF stage, although no difference in the prevalence of TR severity was observed between stage C participants with reduced and preserved ejection fraction HF (Figure 2D). After adjustment for demographic factors (model 1), greater TR severity was associated with a greater prevalence of HF and AF, greater loop diuretic use, and higher plasma levels of NT‐proBNP and high‐sensitivity troponin T (Table 1). Notably, lower body mass index, lower systolic blood pressure, and lower diabetes prevalence were each associated with higher TR severity.

Figure 2. Prevalence of TR severity in the ARIC study at visit 7 (2018–2019).

Prevalence of TR severity in ARIC at visit 7 (n=3046) by age category (A) and by race and sex (B), by stage of MR (C), and by heart failure stage (D). ***P<0.001 for pairwise χ² test. ARIC indicates Atherosclerosis Risk in Communities; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; MR, mitral regurgitation; NS, nonsignificant; and TR, tricuspid regurgitation.

Echocardiographic Correlates of TR

Mean LV ejection fraction was 63±8%, LV end‐diastolic volume index was 46±11 mL/m², pulmonary artery systolic pressure (PASP) was 32±8 mm Hg (Table 2), and 6% had moderate or greater MR. Greater TR severity was associated with lower LV ejection fraction, worse LV diastolic function (higher left atrial volume index, higher E wave velocity, and higher E/A ratio; Table 2), greater PASP, and with greater MR severity (Figure 2C) but not with other left‐sided valvular heart disease. Greater TR severity was also associated with larger RV size (greater RV end‐diastolic area), RA size (higher RA volume index), and worse RV function accounting for RV afterload (lower tricuspid annular plane systolic excursion/PASP ratio; Table 2).

TR Severity and Participant‐Reported Dyspnea

Of the 3046 participants, 2926 had dyspnea scores reported at visit 7. Among these participants, 2243 (76%) reported no or mild dyspnea, and 726 (24%) reported moderate or severe dyspnea. In models adjusted for demographics (model 1), severe TR, but not mild or moderate TR, was associated with moderate/severe dyspnea, an association that was attenuated and no longer statistically significant after additional adjustment for comorbidities (model 2; Table 3).

Table 3.

Association of TR Severity With Participant‐Reported Dyspnea in the ARIC Study at Visit 7 (2018–2019)

	TR severity	No.	Moderate or severe dyspnea, n (%)	Odds ratio (95% CI), model 1	P value, model 1	Odds ratio (95% CI), model 2	P value, model 2
Moderate/severe dyspnea score	None/Trivial	1828	436 (24)	Reference	NA	Reference	NA
	Mild	883	211 (24)	1.0 (0.8–1.2)	0.7	1.1 (0.8–1.3)	0.6
	Moderate	219	62 (28)	1.2 (0.8–1.6)	0.4	1.2 (0.8–1.7)	0.3
	Severe	39	17 (44)	2.3 (1.2–4.5)	0.01	2.1 (1.0–4.4)	0.06

Logistic regression was used for analysis of TR severity and dyspnea scores. Model 1: adjusted for age, sex, race, and field center; model 2: adjusted for age, sex, race, field center, hypertension, diabetes, coronary artery disease, heart failure, body mass index, and estimated glomerular filtration rate. ARIC indicates Atherosclerosis Risk in Communities; NA, not applicable; and TR, tricuspid regurgitation.

TR Severity and Incident Cardiovascular Events

Over a median follow‐up of 3.7 (interquartile range, 2.6–4.3) years, there were 154 incident HF events, and 412 participants died. Higher TR severity was associated with a higher risk of incident HF in all models, including models adjusting for demographic covariates, cardiovascular risk factors, and echocardiographic measures of LV structure, function, and filling pressure (model 3; hazard ratio, 1.28 [95% CI, 1.01–1.64], P=0.04; Table 4). Higher TR severity was associated with a higher mortality rate after adjustment for demographic covariates (model 1), with severe TR associated with an ≈2.5‐fold increase in risk of death (Table 4; Figure 3). Associations of TR severity with all‐cause death were attenuated and no longer significant after further adjustment for clinical cardiovascular risk factors (model 2) and LV measures (model 3). Over a median follow‐up of 3.6 (interquartile range, 2.4–4.6) years, there were 195 AF events. There was no statistically significant difference observed between TR severity and incident AF. Analyses with inverse probability of attrition weighting showed similar results (Table S2). In a sensitivity analysis, higher TRJA was significantly associated with higher risk of HF in models 1 and 2, but there was no statistical difference in mortality outcomes (Table S3).

Table 4.

Association of TR Severity With Incident HF, AF, and Death in the ARIC Study (2019–2020)

Outcome	n	Events, n (%)	HR (95% CI), model 1	P value, model 1	HR (95% CI), model 2	P value, model 2	HR (95% CI), model 3	P value, model 3
HF	2557	154 (6)	1.33 (1.08–1.64)	0.008	1.28 (1.03–1.61)	0.02	1.28 (1.01–1.64)	0.04
AF	2619	195 (7)	1.09 (0.88–1.35)	0.4	1.12 (0.90–1.40)	0.3	1.05 (0.83–1.33)	0.7
Death	3046	412 (14)	1.21 (1.07–1.38)	0.003	1.07 (0.92–1.23)	0.3	1.06 (0.91–1.24)	0.5

Linear trend tests across categories of severity qualitative assessment. Model 1: adjusted for age, sex, race, and field center; model 2: adjusted for age, sex, race, field center, hypertension, diabetes, coronary artery disease, heart failure, body mass index, and estimated glomerular filtration rate; model 3: adjusted for age, sex, race, field center, hypertension, diabetes, coronary artery disease, heart failure, body mass index, estimated glomerular filtration rate, left ventricular ejection fraction, left ventricular end‐diastolic volume, E/e′ mean. HRs for multivariable Cox proportional hazards models are shown here. AF indicates atrial fibrillation; ARIC, Atherosclerosis Risk in Communities; HF, heart failure; HR, hazard ratio; and TR, tricuspid regurgitation.

Figure 3. Kaplan–Meier curves for heart failure incidence and mortality rates per TR severity.

Cumulative incidence of heart failure (A) and death (B). P value for multivariable Cox proportional hazards models and HRs are adjusted for age, sex, race, and study center. P value for trend shows TR severity categories assigned ordinal values to test for a linear trend in hazard across severity levels. Median follow‐up time was 3.7 (interquartile range, 2.6–4.3) years. Total number of heart failure events was 154, and total number of deaths was 412. Event rate is expressed per 100 person‐years. HR, hazard ratio; py, person‐years; and TR, tricuspid regurgitation.

Discussion

Among 3046 older adults in a community‐based cohort, we quantified the prevalence of TR; identified clinical, structural, and functional cardiac correlates of TR; and assessed the association of TR with functional impairment, incident cardiovascular diseases, and death. We report 3 major findings. First, TR is common among older adults, with a prevalence of nearly 40%. Higher TR severity associated with older age, female sex, prevalent cardiovascular disease (HF and AF), MR, and concurrent LV dysfunction. Second, greater TR severity is associated with RV enlargement, RV dysfunction, and RV–pulmonary artery uncoupling. Finally, greater TR severity is associated with a higher risk of developing HF and death, with incident HF remaining significant even after adjusting for clinical comorbidities and LV structure and function. These findings characterize the burden of TR in late‐life adults and clarify the prognostic value of TR severity in this population.

Our findings of a 9% prevalence of at least moderate TR and 38% prevalence of at least mild TR in this cohort with a mean age of 81 years is concordant with prior community‐based studies. The Framingham Heart Study reported a 30% prevalence of mild or greater TR in adults aged >70 years, ⁵ while the OxValve Study noted at least moderate TR in 3% of those aged >65 years and in 7% of those aged >75 years. ⁴ In contrast, a recent study among asymptomatic individuals with a mean age of 69 years and free of known cardiac disease reported a much lower TR prevalence of 14%, with 13% being mild. ³⁴ This discrepancy is likely due to younger age and exclusion of known cardiac disease. With an older population, our prevalence estimates are expectedly higher than those from previous younger community‐based cohorts. ⁴ , ⁵ Notably, our prevalence estimates are also similar to those from previous studies done on hospitalized or clinically referred patients who were generally younger. ⁸ , ³⁵ These studies reported a 30% to 55% prevalence of mild or greater TR, highlighting the bias inherent in estimating population prevalence from clinically referred samples.

Consistent with previous studies in patients with clinically indicated echocardiography, ¹⁰ , ¹¹ , ³⁶ we identified older age, female sex, higher prevalence of HF and AF, and higher pulmonary pressure as correlates of greater TR severity. In contrast, among this older cohort, other cardiovascular comorbidities including hypertension, coronary heart disease, and prior myocardial infarction were not associated with TR severity. We observed an association between higher body mass index and diabetes prevalence and lower TR severity, which was unexpected given their association with LV diastolic dysfunction and higher pulmonary pressure. However, given the older age of our study cohort and the association of TR severity with HF and significant MR, we speculate that cardiac cachexia may contribute to this observation and paradoxical relationship. ³⁷ , ³⁸ , ³⁹

Among patients with HF with either reduced or preserved ejection fraction, TR severity is correlated with markers of right‐ and left‐sided heart disease, including elevated PASP, lower ejection fraction, left atrial area, MR, and RV dysfunction. ⁴⁰ , ⁴¹ We now extend these findings to a predominantly HF‐free community cohort and demonstrate that TR is associated with LV systolic dysfunction (lower LV ejection fraction), LV diastolic dysfunction (higher E wave and E/A ratio, lower tissue Doppler imaging e′, and higher PASP), left atrial enlargement (left atrial volume index), RA enlargement (RA volume index), and RV enlargement and dysfunction (larger RV end‐diastolic area, lower tricuspid annular plane systolic excursion/PASP ratio). These broad associations of TR with LV and RV structure and function highlight the value of TR as a marker for heart disease.

Trace and mild TR are often found incidentally on echocardiography and have no physiologic consequence. ⁸ , ¹¹ , ⁴² In patients with HF, however, even moderate TR severity has been associated with increased symptoms, reduced exercise capacity, and worse clinical outcomes. ⁴³ Among this community‐based cohort of older adults, only severe TR was associated with a higher prevalence of dyspnea, but more modest degrees of TR were not. These results expand on the prior literature to suggest that mild and even moderate TR is largely subclinical in older community‐dwelling adults.

Numerous studies in ambulatory and hospitalized patient populations have consistently demonstrated that moderate and severe isolated TR is associated with a heightened risk of HF hospitalization and all‐cause death. ⁸ , ⁹ , ³⁵ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ Consistent with these findings, we observed that greater TR severity was associated with a higher risk of incident HF over a median follow‐up of 3.7 years, even after adjustment for cardiovascular comorbidities and echocardiographic markers of left‐sided heart disease. We also observed a graded association between TR severity and death, although this association was attenuated after adjusting for comorbid conditions. Recent randomized clinical trials of percutaneous interventions for severe symptomatic TR in older patients, such as TRILUMINATE Pivotal (Clinical Trial to Evaluate Cardiovascular Outcomes in Patients Treated With the Tricuspid Valve Repair System Pivotal) and TRISCEND II (Transcatheter Valve Replacement I Severe Tricuspid Regurgitation), demonstrate symptomatic improvement and reduced HF risk, ⁴⁸ , ⁴⁹ , ⁵⁰ highlighting the evolving understanding of TR as an active contributor to cardiac symptoms and events in late life. In this context, our findings suggest that TR is a relevant prognostic marker also among community‐dwelling older adults.

This study has limitations. Nonattendance of surviving ARIC participants at visit 7 may introduce healthy selection bias and limit generalizability. Likewise, among individuals who attended visit 7, those who underwent echocardiograms demonstrated some difference from those who did not. However, sensitivity analyses incorporating inverse probability of attrition weighting demonstrated similar results to our primary analysis, suggesting a modest impact of these differences on the reported TR prevalence and associations with outcomes. Additionally, TR classification was primarily based on an integrated qualitative echocardiographic assessment, although all studies were measured in an imaging core center by Core Cardiology Training Symposium level III trained cardiologists using American Society of Echocardiography criteria. Furthermore, analyses using TRJA, a quantitative measure of TR, demonstrated overall consistent results with our primary findings. While the prognostic implications of TR may differ by mechanism of regurgitation, data on the cause of TR were not available, and analysis by TR phenotype was not possible. The small number of participants with severe TR resulted in few events and consequently wide CIs, limiting the precision of effect estimates for this subgroup. Finally, residual confounding of the associations between TR severity and study outcomes cannot be excluded despite the use of multivariable models. Notably, pulmonary hypertension (estimated by PASP) is a potential confounder of the association of TR severity with clinical outcomes, but we were unable to adjust for PASP due to substantial and differential missingness with resulting loss of power and potential for selection bias.

Conclusions

TR is common in older adults, particularly among women and those with structural heart disease. Greater TR severity is associated with greater risk of developing HF over a 4‐year follow‐up after accounting for clinical comorbidities and LV structure and function. The extent to which treating TR in this community‐based population with less advanced disease can alter HF or mortality risk remains an important area for future investigation.

Sources of Funding

The ARIC study was funded in whole or in part with federal funds from the National Heart, Lung, and Blood Institute, National Institutes of Health, Department of Health and Human Services, under contract numbers 75N92022D00001, 75N92022D00002, 75N92022D00003, 75N92022D00004, and 75N92022D00005. Dr Shah was supported by National Institutes of Health/National Heart, Lung, and Blood Institute grants R01HL135008, R01HL143224, R01HL150342, R01HL148218, R01HL160025, and K24HL152008.

Funding for laboratory testing and biospecimen collection at ARIC visit 7 was supported by grant R01DK089174 from the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health. Grant R01HL134320 also funded high‐sensitivity cardiac troponin I, high‐sensitivity cardiac troponin T, and NT‐proBNP. Reagents for the high‐sensitivity cardiac troponin and NT‐proBNP assays were donated by the Roche Diagnostics Corporation. Reagents for high‐sensitivity cardiac troponin I were donated by Abbott Diagnostics.

Disclosures

None.

Supporting information

Tables S1–S3