Across different study groups stratified by left ventricular hypertrophy (LVH) and subclinical myocardial injury status. Subclinical myocardial injury refers to elevated troponin-I levels (≥4 ng/L in women and ≥6 ng/L in men). P interaction (LV mass times high-sensitivity cardiac troponin-I) < .001.
The model is adjusted for age, sex, hypertension status, systolic blood pressure, diabetes, hemoglobin A1C, renal function, body mass index (calculated as weight in kilograms divided by height in meters squared), smoking status and physical activity levels. Subclinical myocardial injury refers to high-sensitivity cardiac troponin-I levels of 4 ng/L or more in women and 6 ng/L or more in men.
eTable 1. Baseline characteristics of JHS participants among participants with vs. without LV hypertrophy of left ventricular mass
eTable 2. Baseline characteristics of JHS participants stratified by presence of LVH and subclinical myocardial injury
eTable 3. Multivariable adjusted association between baseline levels of hs-cTnI, and LVH status and risk of heart failure including only adjudicated HF events from January 1st, 2005 onwards
eTable 4. Prevalence of LVH, subclinical myocardial injury and heart failure incidence across different study groups
eTable 5. Multivariable adjusted association between study groups defined by baseline LVH status and subclinical myocardial injury levels of hs-cTnI, and LVH status and risk of incident HF across different subgroups by hypertension status, age, and interval history of acute myocardial infarction
eTable 6. Multivariable adjusted association between study groups defined by baseline LVH remodeling patterns (no LVH vs. concentric LVH vs. eccentric LVH) and subclinical myocardial injury levels of hs-cTnI, risk of incident HF
eFigure 1. Restricted cubic splines showing the adjusted association between continuous measures of hs-TnI levels and risk of heart failure in men and women.
eFigure 2. Levels of hs-cTnI (left) and LV mass (right) among men and women participants with LVH and subclinical myocardial injury and without LVH or chronic myocardial injury
eFigure 3. Proportion of participants with incident HF on follow up across different study group stratified by LVH remodeling patterns (no LVH vs. concentric LVH vs, eccentric LVH) and subclinical myocardial injury status
Customize your JAMA Network experience by selecting one or more topics from the list below.
Pandey A, Keshvani N, Ayers C, et al. Association of Cardiac Injury and Malignant Left Ventricular Hypertrophy With Risk of Heart Failure in African Americans: The Jackson Heart Study. JAMA Cardiol. 2019;4(1):51–58. doi:10.1001/jamacardio.2018.4300
What is the independent and joint contribution of left ventricular hypertrophy and subclinical myocardial injury toward the risk of heart failure in African Americans?
This longitudinal cohort study from the Jackson Heart Study cohort suggests that the combination of left ventricular hypertrophy and subclinical myocardial injury, as measured by high-sensitivity cardiac troponin assay, is associated with very high absolute (35% incidence) and relative risk of heart failure (5-fold higher risk), particularly in African American men (15-fold higher risk).
Left ventricular hypertrophy and subclinical myocardial injury may identify a malignant, subclinical heart failure phenotype and have implications in developing screening strategies to prevent heart failure in African Americans.
African Americans have a higher burden of heart failure (HF) risk factors and clinical HF than other racial/ethnic groups. However, the factors underlying the transition from at-risk to clinical HF in African Americans are not well understood.
To evaluate the contributions of left ventricular hypertrophy (LVH) and subclinical myocardial injury as determined by abnormal high-sensitivity cardiac troponin-I (hs-cTnI) measurements toward HF risk among African Americans.
Design, Setting, and Participants
This prospective, community-based cohort study was conducted between July 2016 and September 2018 and included African American participants from Jackson, Mississippi enrolled in the Jackson Heart Study without prevalent HF who had hs-cTnI measurements and an echocardiographic examination at baseline. Participants were stratified into categories based on the presence or absence of LVH and subclinical myocardial injury (category 1: hs-cTnI <4 ng/L in women and <6 ng/L in men; category 2: 4-10 ng/L in women and 6-12 ng/L in men; category 3: >10 ng/L in women and >12 ng/L in men).
Main Outcomes and Measures
Adjusted associations between LVH, subclinical myocardial injury, and the risk of incident HF hospitalization were assessed using Cox proportional hazards models.
The study included 3987 participants (2552 women [64%]; 240 (6.0%) with LVH; 1003 (25.1%) with myocardial injury) with 285 incident HF events over a median follow-up of 9.8 years (interquartile range, 8.9-10.6 years). In adjusted analyses, higher LV mass and subclinical myocardial injury were independently associated with the risk of HF with a significant interaction between the 2 (Pint < 0.001). The highest risk of HF was noted among individuals with both LVH and myocardial injury (absolute incidence, 35%; adjusted hazard ratio [aHR; vs no LVH and no myocardial injury], 5.35; 95% CI, 3.66-7.83). A significant interaction by sex was also observed. Men with LVH and subclinical myocardial injury had an almost 15-fold higher risk of HF (aHR, 14.62; 95% CI, 7.61-28.10) vs those with neither LVH nor injuries. By contrast, women with this phenotype had a nearly 4-fold higher risk of HF (aHR, 3.81; 95% CI, 2.40-6.85).
Conclusions and Relevance
The combination of LVH and subclinical myocardial injury identifies a malignant, preclinical HF phenotype in African Americans with a very high risk of HF, particularly among men. This finding could have implications for future screening strategies that are designed to prevent HF in the population.
Heart failure (HF) disproportionately affects African Americans, with substantially higher rates compared with other racial/ethnic groups.1,2 The racial/ethnic disparities in HF incidence and outcomes are particularly notable among adults younger than 50 years.3,4 This excess burden of HF among African Americans has been partially attributed to a higher prevalence and worse control of traditional risk factors, such as hypertension and diabetes, in African Americans.5-7
Prior studies from multiethnic cohorts have demonstrated that the progression from at-risk to symptomatic HF occurs through a series of intermediate subclinical cardiac phenotypes characterized by left ventricular (LV) hypertrophy (LVH) or LV dysfunction and adverse remodeling.8-14 Recent population-based studies have also demonstrated that subclinical myocardial injury, now detectable using high-sensitivity assays for cardiac troponin (hs-cTn), is strongly associated with prevalent structural heart disease and with future HF risk.15-17 Preliminary studies suggest that a subclinical phenotype defined by concomitant LVH and myocardial injury is particularly hazardous and appears to be more prevalent among African Americans.18 However, these prior studies were limited by insufficient numbers of clinical events among African Americans, and thus the natural history of this subclinical phenotype and its role in downstream development of HF in African Americans is not well understood. This represents an important knowledge gap in the development of novel preventive strategies in this high-risk population.17,19 Therefore, we sought to determine the independent and joint associations of LVH and markers of subclinical myocardial injury with the risk of HF in a cohort of community-dwelling African American adults. We hypothesized that subclinical myocardial injury and LVH are independently associated with a higher risk of HF among African Americans. We further hypothesized that subclinical myocardial injury modifies the risk of HF that is associated with LVH and identifies a subset of African Americans who are at a particularly high risk of HF.
The Jackson Heart Study (JHS) is a community-based, prospective, observational cohort study of African American participants aged 21 to 84 years from the tricounty Jackson, Mississippi area (Hinds, Madison, and Rankin).20 The study enrolled 5306 participants and was designed to evaluate risk factors for cardiovascular disease (CVD) in African Americans. The details of the study design, recruitment strategy, and visit protocols have been described previously.20-22 Participants were examined at a baseline (2000-2004), visit 2 (2005-2008), and visit 3 (2009-2013).22 Written informed consent was obtained from all study participants and the study protocol was approved by local institutional review boards. The institutional review board at University of Texas Southwestern Medical Center, Dallas, Texas, approved the use of JHS data for this analysis. Our study included participants from the baseline visit that were free of cardiovascular diseases and heart failure (n = 4734) at baseline and had measured hs-cTnI levels (n = 4518). The final study population included 3987 participants with nonmissing data on relevant clinical covariates and follow-up.
Study participants underwent an in-person interview and a baseline clinical examination using standardized protocols as described previously.21,23 Demographic details, medical history, and lifestyle factors (physical activity levels, tobacco use, and alcohol use) were self-reported. Anthropometric parameters (height and weight) and blood pressure were measured using standard protocols. Body mass index was calculated as weight in kilograms divided by height in meters squared. Fasting blood samples were drawn and analyzed to measure hemoglobin A1C, high-density lipoprotein, low-density lipoprotein, and serum creatinine levels using standardized methods. The estimated glomerular filtration rate was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation.24 Participants were identified as hypertensive based on the measure of blood pressure levels (≥140/90 mm Hg) and/or the use of blood pressure–lowering medications at baseline. Diabetes status was identified based on laboratory measurements (hemoglobin A1C >6.5%; fasting blood glucose ≥126 mg/dL [to convert to millimoles per liter, multiply by 0.0555]) and/or the use of diabetes medications.
Plasma samples were collected at the baseline visit (2000-2004) and stored at −70°. High-sensitivity cardiac troponin-I was measured from thawed samples in a core laboratory (Vermont) by personnel who were masked to participant information and outcomes. Measurements were performed with the ARCHITECT hs-cTnI assay platform (Abbott Diagnostics), a 2-step, double-monoclonal immunoassay that uses antibody-coated paramagnetic microparticles.25 The lower limit of detection for this assay is 1.2 ng/L (to convert to micrograms per liter, multiply by .001) with a coefficient of variation of 10% at a concentration of 3.0 ng/L.26 The 99% cutoff in the general population for this assay is 15.6 ng/L in women and 34.2 ng/L in men.
A detailed echocardiographic examination was performed at the baseline visit by trained sonography technicians using previously reported standardized protocols (Sonos-4500; Philips Medical Systems).27,28 The examination included a 30-minute 2-dimensional and M-mode echocardiography examination with parasternal, apical, and subcostal views. Left ventricular mass was estimated in grams as previously described using the formula 0.8 × (1.04 × [([interventricular septal thickness in the end diastole + LV end diastolic dimension + posterior wall thickness in the end diastole]3−LV end diastolic dimension)3]) + 0.6. Left ventricular mass was indexed to body surface area, and LVH was defined as an LV mass index of 96g/m2 or more in women and 116 g/m2 or more in men.29 Relative wall thickness was defined using the formula right wall thickness = 2 × posterior wall thickness/LV end diastolic diameter.
The primary outcome of interest for this analysis was HF hospitalization incidence. The HF adjudication protocol used in JHS has been described previously.30-32 Briefly, incident HF hospitalizations were identified through annual follow-up telephone interviews and hospital discharge lists and confirmed with reviews and abstractions of HF hospitalization records. The HF hospitalization adjudication was performed by trained medical personnel using abstracted information from hospital records. The formal adjudication of HF events in the JHS started in January 2005 and events before that were self-reported. Consistent with recent reports from the JHS, we have included all HF events (from baseline visit to the end of follow-up in 2012) in the primary analysis and formally adjudicated HF events only in the sensitivity analysis.31
For this analysis, we prospectively defined 3 categories for hs-cTnI (category 1, hs-cTnI levels <4 ng/L in women and <6 ng/L in men [no significant subclinical myocardial injury]; category 2, hs-cTnI levels 4-10 ng/L in women and 6-12 ng/L in men [mild subclinical myocardial injury]; and category 3, hs-cTnI levels >10 ng/L in women and >12 ng/L in men [moderate to severe subclinical myocardial injury]). These categories are similar to those used in previous studies to define the burden of subclinical myocardial injury.33 The baseline characteristics of the study participants were reported as medians (for continuous variables) or percentage (for categorical variables) across the previously mentioned categories of hs-cTnI. The Jonckheere trend test or Wilcoxon rank-sum test were used to compare the continuous and categorical baseline characteristics, respectively, across categories. Multivariable-adjusted Cox proportional hazards models were constructed to evaluate the association of categorical and continuous hs-cTnI values and risk of HF. The models were adjusted for the following a priori selected potential confounders: demographic characteristics (age and sex), traditional CVD risk factors (history of hypertension, systolic blood pressure, history of diabetes, hemoglobin A1C, estimated glomerular filtration rate [eGFR], current smoking status), lifestyle risk factors (body mass index and physical activity levels), and LV mass. For continuous analyses, hs-cTnI was log transformed to account for its skewed distribution. Restricted cubic spline plots were used to assess for any threshold effects in the association between h-TnI levels and risk of HF in men and women. The association between the presence of LVH and risk of HF was also assessed using similar Cox models with adjustment for demographic characteristics, CVD risk factors, lifestyle risk factors, and hs-cTnI levels. The contribution of LV mass and hs-cTnI toward the prediction of incident HF risk independent of traditional HF risk factors (age, sex, body mass index, history of hypertension, diabetes, hemoglobin A1C, eGFR, current smoking status, and physical activity levels) was assessed by calculating the Harrell C statistic for discrimination and the net reclassification index. The calibration of the model was assessed using the Gronnesby and Borgan test.
Multiplicative interaction terms were included in the adjusted models to determine if the association between LV mass and risk of HF hospitalizations was modified by hs-TnI levels. Interaction testing was also performed between hs-TnI levels, LV mass, and eGFR (hs-TnI × eGFR and LV mass × eGFR) in the adjusted Cox models to determine if the associations of LV mass and hs-TnI levels with HF risk are modified by the baseline renal function. To further characterize the interaction between subclinical myocardial injury and LVH, we evaluated the unadjusted and adjusted risk of HF across the following 4 groups of study participants: group 1, no LVH/no subclinical myocardial injury (hs-cTnI <4 ng/L in women and <6 ng/L in men); group 2, no LVH/subclinical myocardial injury (hs-cTnI ≥4 ng/L in women and ≥6 ng/L in men); group 3, LVH/no subclinical myocardial injury; and group 4, LVH and subclinical myocardial injury. The Cox proportional hazards assumption was tested and found to be appropriate for this analysis.
In exploratory analyses, LVH was stratified into concentric and eccentric LVH. Concentric LVH was identified by presence of LVH as defined previously and a relative wall thickness of 0.42 or more. Eccentric LVH was identified by presence of LVH and a relative wall thickness of less than 0.42. Sensitivity analyses were performed, restricting HF events to those formally adjudicated HF events after January 1, 2005, and after excluding individuals that developed incident myocardial infarction (MI) on follow-up. Analyses were also performed across relevant subgroups, including sex (men vs women), age (above vs below median), and history of hypertension (yes vs no). Statistical analyses were performed using SAS (version 9.1; SAS Institute) and P < .05 was considered statistically significant for all statistical tests.
In the study cohort of 3987 African American adults without known coronary heart disease or HF at baseline, the prevalence of LVH was 6.0% and subclinical myocardial injury (hs-cTnI ≥6 ng/L in men and ≥4 ng/L in women) was 25.1%. The phenotype characterized by both LVH and subclinical myocardial injury was present at baseline in 145 study participants (3.7%). Table 1 compares the baseline characteristics of the study participants, stratified by the hs-cTnI categories. Individuals with subclinical myocardial injury were older, more commonly men, and had a higher burden of traditional cardiovascular risk factors, such as hypertension and diabetes. The use of antihypertensive medications, including β-blockers and diuretics, was higher among participants with a higher burden of subclinical myocardial injury. There were no clinically meaningful differences in body mass index and lipid levels across the hs-cTnI categories. In contrast, LV mass and the prevalence of LVH increased across hs-cTnI categories.
Compared with participants without LVH, those with LVH were older with higher hs-cTnI levels and had a greater burden of traditional CVD risk factors (eTable 1 in the Supplement). When stratified by both LVH and hs-cTnI, similar trends were observed in baseline characteristics across hs-cTnI groups (no subclinical myocardial injury [<4 ng/L in women and <6 ng/L in men] vs subclinical myocardial injury present (≥4 in men and ≥6 in women]), with the highest burden of CVD risk factors among those with LVH and subclinical myocardial injury at baseline (eTable 2 in the Supplement).
There were 285 incident HF events over a median follow-up of 9.8 years (interquartile range, 8.9-10.6 years). In adjusted analyses, the presence of LVH was independently associated with a higher risk of HF after adjusting for potential confounders, including hs-cTnI levels (adjusted hazard ratio [aHR] LVH vs no LVH, 2.05; 95% CI, 1.50-2.80; P < .001; Table 2). Subclinical myocardial injury, as assessed by hs-cTnI levels, was also significantly associated with a higher risk of HF in a dose-dependent manner (Table 2). This association was consistent across categorical and continuous measures of hs-cTnI (aHR per 1 SD higher log Hs-cTnI, 1.51; 95% CI, 1.38-1.65; P < .001). The addition of LV mass and hs-cTnI to the traditional risk factor model (age, sex, body mass index, history of hypertension, diabetes, hemoglobin A1C, GFR, current smoking status, and physical activity levels) was associated with significant improvement in C statistic (from 0.81 [95% CI, 0.79-0.83] to 0.84 [95% CI, 0.82-0.86]; P < .001]) and the net reclassification index (0.34; 95% CI, 0.13-0.54; P = .002). The Gronnesby and Borgan test was used to evaluate the calibration of the model. The difference in −2 Loglikelihoods yielded a statistic of 2.58 (P = .98), suggesting that the models were well calibrated. The results were also consistent in landmark analysis using HF outcome events after January 1, 2005 (n = 218 events) (eTable 3 in the Supplement). No significant interaction was noted between eGFR and hs-TnI levels (P interaction = .48) and GFR and LV mass (P interaction = .84) for risk of HF.
In a restricted cubic spline analysis that evaluated the adjusted associations between continuous measures of hs-TnI and HF risk, the risk of HF increased linearly with higher levels of hs-TnI levels up to the 99th percentile of the population distribution of hs-TnI (eFigure 1 in the Supplement). Furthermore, the risk of HF became statistically significant at a lower absolute threshold of hs-TnI in men than in women (more than 2.2 ng/L in men and more than 4.8 ng/L in women).
A statistically significant interaction was noted between LV mass and hs-cTnI levels in the multivariable adjusted model for the risk of HF (P interaction < .001). The risk of HF varied markedly across groups stratified by the baseline LVH and subclinical myocardial injury status, with a 10-fold higher unadjusted incidence of HF among individuals with LVH and subclinical myocardial injury at baseline vs those without LVH or elevated hs-TnI levels (10-year incidence, 35.0% vs 3.56%; Figure 1). Furthermore, among participants with LVH, those with subclinical myocardial injury had up to a 3.5-fold higher incidence of HF as compared with those without myocardial injury (35% vs 10.6%). Fifty-one HF events (18% of all HF events) were observed in the 146 (3.7%) of the study population with LVH and subclinical myocardial injury (eTable 4 in the Supplement). In multivariable analyses, a more than 5-fold higher adjusted risk of HF was observed among those with LVH and subclinical myocardial injury (aHR, 5.35; 95% CI, 3.66-7.83; P < .001 [reference group: no LVH/no myocardial injury]). In contrast, when LVH was present without myocardial injury, HF risk was attenuated and not statistically significant (aHR, 1.81; 95% CI, 0.94-3.50; P = .08; Figure 2).
A statistically significant interaction by sex between the presence of LVH and subclinical myocardial injury and risk of HF was noted (P interaction < .001). In sex-stratified analyses, men with LVH and subclinical myocardial injury had an approximately 15-fold higher risk of HF (10-year absolute incidence, 48%; HR; 14.62; 95% CI, 7.61-28.10; P < .001; reference group: no LVH no subclinical myocardial injury; Table 3). In contrast, women with this phenotype had an approximately 4-fold higher risk of HF (10-year absolute incidence, 30%; HR, 3.81; 95% CI, 2.40-6.85; P < .001). The median hs-cTnI levels and LV mass were 6-fold and 2-fold higher in men with the malignant LV phenotype as compared with those without LVH or subclinical myocardial injury. In contrast, among women, the median hs-cTnI levels and LV mass were 3-fold and 1.5-fold higher among those with the malignant LV phenotype vs those without these cardiac abnormalities (eFigure 2 in the Supplement). In other subgroup analyses stratified by age and hypertension status, the associations between LVH, hs-cTnI, and the risk of adverse clinical outcomes were consistent across subgroups (eTable 5 in the Supplement). Interval MI before HF hospitalization was noted in 37 participants (13%) who developed HF on follow-up. In a sensitivity analysis excluding HF events occurring after MI, results were unchanged with that observed in the primary analysis (eTable 5 in the Supplement).
Among different LVH phenotypes, the incidence of HF was significantly higher with concentric and eccentric patterns of LVH (vs no LVH) (eFigure 3 in the Supplement). Furthermore, when stratified by the presence of subclinical myocardial injury, the risk of HF was increased similarly among individuals with concentric and eccentric LVH in the presence of subclinical myocardial injury in unadjusted and adjusted analyses (referent group: no LVH or subclinical myocardial injury) (eFigure 3 and eTable 6 in the Supplement).
In this study, we observed several important findings. First, both LVH and hs-cTnI levels were independently associated with an increased risk of HF among African Americans. Second, there was a highly significant and clinically relevant synergistic interaction between LV mass and hs-cTnI levels for the risk of HF, such that the joint presence of both LVH and subclinical myocardial injury was associated with a risk of HF that was markedly increased above the expected level from the joint effect. An interaction was also observed by sex, such that the risk of HF associated with the presence of LVH and myocardial injury was higher among men ( ~ 15-fold higher risk of HF) than women ( ~ 4-fold higher risk). Our study findings confirm that the value of hs-cTnI for HF risk prediction extends to African Americans who are at high risk for HF. Moreover, the findings underscore the heterogeneity in the risk of HF associated with LVH and confirm that subclinical myocardial injury among those with LVH identifies a malignant preclinical phenotype that, although low in prevalence, is associated with a very high risk of HF among African Americans, particularly among men.
Our findings confirm and extend prior studies that link small elevations in cardiac troponins, detected with high-sensitivity assays, with future HF risk. To our knowledge, this is the first study to evaluate the association of hs-cTn with HF risk in an exclusively African American cohort, which is important as prior multiethnic studies had too few HF events among African American participants, precluding rigorous evaluation of this question. The association of hs-cTnI with HF risk was dose-dependent and independent of potential confounders, including LV mass and major CVD risk factors.
Left ventricular hypertrophy has been identified as an intermediate subclinical phenotype in the development of clinical HF.8,10,11 However, the natural history of the transition from LVH to clinical HF remains poorly understood. Prior studies have implicated interval MI in the transition from LVH to HF, particularly HF with reduced ejection fraction.34 The findings from this study suggest that subclinical myocardial injury as indicated by elevated hs-cTnI levels may also play an important role in mediating the progression from LVH to clinical HF among African Americans, as evidenced by the substantially higher absolute and relative risk of HF among those with vs without subclinical myocardial injury. The risk of HF that was associated with myocardial injury among those with vs without LVH was consistent across age-based and hypertension-based subgroups and among those without interval MI over the follow-up period.
Prior studies have observed divergent findings regarding the interaction between LVH and measures of subclinical myocardial injury for the risk of HF in multiethnic cohorts.18,35 In the Dallas Heart Study and Multiethnic Study of Atherosclerosis, significant interactions between hs-cTnT and LVH were observed for the risk of HF or cardiovascular death.18,36In these studies, participants with LVH and biomarker evidence of subclinical myocardial injury had a substantially augmented risk of developing HF or cardiovascular death as compared with those with LVH in the absence of subclinical myocardial injury. In contrast, the Cardiovascular Health Study also observed a significantly higher risk of HF among participants with elevated hs-cTnT levels and LVH but failed to observe a statistically significant interaction.35 These prior studies included multiethnic study participants with 50% or fewer African American participants and an insufficient power to detect an interaction between hs-cTn and LVH among African Americans.
While prior studies from these multiethnic cohorts have reported a similar risk of HF associated with the presence of LVH and subclinical myocardial injury in men and women,35,36 we observed a significant interaction by sex for the risk of HF that was associated with this malignant phenotype in African Americans. African American men with this malignant phenotype had a particularly high risk of HF as compared with those without LVH and subclinical myocardial injury (14-fold higher risk in men vs 3-fold higher risk in women). This could be associated with the sex differences in the severity of myocardial injury (6-fold higher hs-cTnI levels in men vs 4-fold higher in women) among African Americans with vs without the malignant LV phenotype.
Our study findings have important implications for HF prevention in African Americans. Several studies have demonstrated that African Americans have a higher burden of HF as compared with other racial/ethnic groups.1-3 Consistent with prior observations, the risk of HF among African Americans in our study cohort was substantial (7% at 10-year follow-up). We observed that subclinical myocardial injury in African Americans may identify individuals who are at an increased risk of HF, even in absence of LVH, highlighting its importance as a potential risk-stratification tool. It is notable that more than 60% of the HF events occurred among the 25% of JHS participants with evidence of subclinical myocardial injury, an observation that supports the potential value of hs-cTnI measurement as an initial screening test for HF risk among African Americans.
This study also provides additional support for the hypothesis that a malignant form of LVH may represent a particularly hazardous subphenotype. Among African American individuals, the presence of both subclinical myocardial injury and LVH identified a small subgroup (146 [3.7%]) with a 35% risk of HF within 10 years of follow-up. This high absolute risk of HF is remarkable considering the relatively young age of the study population (median age, 54 years) and lack of known CVD at baseline. Simple screening strategies to identify this malignant phenotype may allow targeting of HF prevention efforts to the highest-risk African Americans who may benefit from aggressive preventive interventions. It is noteworthy that prior studies have demonstrated that electrocardiogram-based LVH can substitute for imaging-based LVH, which would make such a screening strategy more cost-effective and readily available.18 Future studies are needed to determine if a “high-risk approach” to targeted prevention, focusing on hs-cTn and LVH, may lower the incidence of HF and reduce disparities among African Americans. Prospective studies are also needed to determine if the high risk of HF among individuals with the malignant LVH phenotype is modifiable with an aggressive management of traditional risk factors, such as hypertension, diabetes, obesity, and physical inactivity.
Our study is not without limitations. First, owing to the observational nature of the study, there is a possibility of residual measured or unmeasured confounding. Second, we do not have data on HF subtypes in the outcome events, and thus we could not characterize associations of our phenotypes with HF with preserved ejection fraction vs HF with reduced ejection fraction. Third, the formal adjudication of HF events began in 2005, and some HF events before 2005 could have been missed. However, the observed associations and interactions between LVH, hs-cTnI, and the risk of HF were consistent in the sensitivity analysis that used only adjudicated HF hospitalizations from January 1, 2005, through December 31, 2012. Finally, our study findings may not be generalizable to other racial/ethnic groups.
Among community-dwelling African Americans without clinical CVD, the combination of LVH and subclinical myocardial injury identifies a malignant, subclinical HF phenotype with a remarkably high-risk absolute risk of HF. Future studies are needed to delineate the underlying mechanisms that are responsible for this phenotype and to determine if targeted preventive strategies aimed at this high-risk subgroup are warranted to address the disproportionately high risk of HF in African Americans.
Corresponding Author: Jarett D. Berry, MD, MS, Division of Cardiology, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-9047 (firstname.lastname@example.org).
Accepted for Publication: November 2, 2018.
Published Online: December 19, 2018. doi:10.1001/jamacardio.2018.4300
Author Contributions: Dr Berry had full access to all data in the study and had final responsibility for the decision to submit for publication.
Concept and design: Pandey, Correa, Hall, Seliger, Neeland, de Lemos.
Study concept and design: Berry.
Acquisition, analysis, or interpretation of data: Pandey, Keshvani, Ayers, Correa, Drazner, Lewis, Rodriguez, Fox, Mentz, deFilippi, Ballantyne, de Lemos, Berry.
Drafting of the manuscript: Pandey, Keshvani, Fox, Berry.
Critical revision of the manuscript for important intellectual content: Ayers, Correa, Drazner, Lewis, Rodriguez, Hall, Fox, Mentz, deFilippi, Seliger, Ballantyne, Neeland, de Lemos, Berry.
Statistical analysis: Pandey, Ayers, Correa.
Obtained funding: Correa, de Lemos, Berry.
Administrative, technical, or material support: de Lemos, Berry.
Study supervision: Berry.
Supervision: Hall, Mentz, de Lemos.
Conflict of Interest Disclosures: Dr Mentz reported grants and personal fees from Novartis, Amgen, and AstraZeneca; receives research support from the National Institutes of Health (grants U01HL125511-01A1, U10HL110312, and R01AG045551-01A1), Akros, Amgen, AstraZeneca, Bayer, GlaxoSmithKline, Gilead, Luitpold, Medtronic, Merck, Novartis, Otsuka, and ResMed; receives honoraria from Abbott, Amgen, AstraZeneca, Bayer, Janssen, Luitpold Pharmaceuticals, Medtronic, Merck, Novartis, and ResMed; and has served on an advisory board for Amgen, Luitpold, Merck and Boehringer Ingelheim. Dr deFilippi reported grants from Roche Diagnostics, Abbott Diagnostics, FujiRebio, and Siemens Healthcare Diagnostics and personal fees from Alere, Radiometer, Ortho Diagnostics, UpToDate, WebMD, Siemens Healthcare, Roche Diagnostics, and Metabolomics. In addition, Dr deFilippi had a patent to US20170234888 issued. Dr. Seliger reported grants from Roche Diagnostics and personal fees from Abbvie Inc. In addition, Dr Seliger had a patent to “Methods for Assessing Differential Risk of Developing Heart” pending. Dr Ballantyne reported grants and personal fees from Roche and Abbott. In addition, Dr Ballantyne had a patent to patent 61721475 ("Biomarkers to Improve Prediction of Heart Failure Risk") filed by Baylor College of Medicine and Roche pending. Dr Neeland reported personal fees from Boehringer Ingelheim/Lilly Alliance and AMRA Medical and grants from Novo Nordisk. Dr de Lemos reported grants from Abbott Diagnostics; personal fees from Abbott Diagnostics, Ortho Clinical Diagnostics, Novo Nordisc, Amgen, and Regeneron; and grants and personal fees from Roche Diagnostics. Dr Berry reported grants from Abbott and the American Heart Association. No other disclosures were reported.
Funding/Support: This project was funded by the Strategically Focused Research Network Grant for Prevention from the American Heart Association to University of Texas Southwestern Medical Center, Dallas, and Northwestern University School of Medicine, Chicago. Dr. Berry received funding from grant 14SFRN20740000 from the American Heart Association prevention network and salary support from Abbott Diagnostics. The funding for biomarker assays was provided by Abbott Diagnostics.
Role of the Funder/Sponsor: The funding organizations had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Create a personal account or sign in to: