Search

Introduction: Heart failure (HF) is a leading cause of death. Despite the high burden of CVD and diabetes affecting American Indian (AI) communities, studies of risk factors for HF and HF subtypes in AIs are lacking and critically needed to optimize prevention strategies. Hypothesis: We hypothesize that diabetes and diabetes-related complications are HF risk factors in AIs. Methods: A total of 3,066 Strong Heart Study (SHS) participants (mean age 56.3±8.1 years, 58% women, all free of CVD) recruited in 1989-91 were followed through 2017. A total of 489 participants developed HF (119 HFpEF, 68 HFmrEF, 98 HFrEF, 204 unclassified). We used Cox proportional hazards models to determine risk factors for HF and HF subtypes progressively adjusting for sociodemographic and traditional CV risk factors (Model 1) and additionally for glycated hemoglobin (HbA1c) and albuminuria (Model 2-main model). Results: HRs (95% CI) are shown in Table 1. Age, smoking status, BMI, hypertension treatment, diabetes status, HbA1c, and albuminuria were independently associated with overall HF (Model 2). Some associations differed across HF subtypes: male sex was associated with higher risk of HFrEF and potentially HFmrEF, whereas female sex was associated with higher risk of HFpEF. BMI was associated with HFpEF but not with HFmrEF and HFrEF. Higher HDL-cholesterol was negatively associated with risk of HFmrEF and HFrEF. Diabetes was consistently associated with all HF subtypes. HbA1c was significantly associated with HFmrEF and albuminuria was significantly associated with HpEF and HFrEF. The association with diabetes status was attenuated after adjustment for HbA1c and albuminuria for all forms of HF. Conclusion: Diabetes and diabetes-related complications are major risk factors for all forms of HF in the SHS. Indicators of HF risk differ across HF subtypes. Including measures of diabetes control (HbA1c) and target organ damage (albuminuria) in HF risk calculation improves risk prediction and could aid in targeting primary prevention interventions.

Background: Dyslipidemia is an established risk factor for CVD. Standard lipid panel lacks description of the full spectrum of blood lipids. A longitudinal profiling of blood lipidome in relation to the risk of CVD is lacking in any racial/ethnic groups. Objective: Identify novel lipid species associated with risk of CVD in American Indians, independent of traditional risk factors (TRFs). Method: Using untargeted LC-MS, we repeatedly measured 1,542 lipid species in 3,916 fasting plasma samples from 1,958 participants who attended two exams (~5.5 years apart) and were followed up to 18 years in the Strong Heart Study. Elastic net was used to select baseline lipids and change in lipids (baseline to 5-year follow-up) associated with 18-year risk of CVD. Multivariate and network analyses were conducted to identify discriminatory lipidomic signatures and differential lipid networks associated with risk of CVD. Results: Of 1,911 participants free of CVD at both baseline (2001-2003) and 5-year follow-up (2006-2009), 222 developed CVD after 18 years of follow-up. High-resolution LC-MS detected 1,542 lipid species (518 known). After adjusting for TRFs (age, sex, center, smoking, BMI, hypertension, diabetes, LDL-c, and albuminuria), elevated baseline levels of 21 known lipids, including cholesterol esters, sphingolipids, glycerophospholipids, and fatty acids, were significantly associated with increased (HR: 1.08 to 1.26) or decreased (HR: 0.81 to 0.91) risk of CVD. Longitudinal change in 13 lipids, including cholesterol esters, sphingolipids, glycerophospholipids, and fatty acids, were either positively (HR: 1.03 to 1.28) or inversely (HR: 0.75 to 0.92) associated with risk of CVD at 18-year follow-up. A lipid score comprising 10 baseline lipids significantly improved risk prediction beyond TRFs (ROC increases from 0.823 to 0.837, P = 0.009). Multivariate and network analyses identified distinct lipidomic signatures and lipid networks associated with risk of CVD. Conclusion: Novel molecular lipids significantly predict risk of CVD beyond TRFs in American Indians. Our findings provide novel lipid molecules for risk stratification and shed light on the mechanism linking lipid dysregulation to CVD in American Indians.

Introduction: Leukocyte telomere length (LTL) is a potential biomarker of aging and associated with several age-related diseases. Current research on an association between LTL and incident stroke has had inconclusive results. We hypothesized that LTL is associated with incident stroke among American Indians (AI) in the Strong Heart Family Study (SHFS). Methods: The SHFS is a population-based cohort study of cardiovascular disease (CVD) and its risk factors. Participants (n=2,769) recruited from regions in Arizona, Oklahoma and the Dakotas were assessed for LTL and CVD risk factors during a clinic visit between 2001 and 2003. Incident stroke events were identified through the end of 2018 (mean follow-up: 16.4 years). We assessed the association between LTL and incident stroke using frailty models based on the proportional hazards, accounting for family relatedness and established stroke risk factors that include sex, geographical location, education, smoking, atrial fibrillation, diabetes mellitus, and hypertension. Results: Among 2,769 participants, the mean age was 40.6±17.2 and 41.4% were male. During follow-up, there were 79 (2.9%) incident stroke cases. In the primary model, which adjusted for demographic variables (sex, location and education), the hazard ratios (HR) for stroke in participants in the first and second LTL quartiles were significantly higher than those in the highest (longest) LTL quartile, with HRs of 3.1 (95%CI: 1.4 - 6.6) and 3.5 (95%CI: 1.7 - 7.5), respectively. After adjusting for smoking, atrial fibrillation, diabetes mellitus, and hypertension, the association between LTL and stroke was attenuated, but remained significant when comparing the second shortest LTL quartile to the longest LTL quartile, HR: 2.3 (95% CI: 1.1 – 5.0). Conclusions: In summary, LTL was associated with incident stroke among SHFS participants. Those with shorter LTL have higher risk of stroke. Longer follow-up time may add more power to data analyses since the SHFS is relatively young, with an average baseline age of 40 years. If results are confirmed in other populations, LTL may serve as a biomarker identifying high risk individuals for the purpose of stroke prevention.

Background: Prior work has shown that fibrinogen, a well-known inflammatory marker, is associated with cardiovascular (CV) events. While it is known that physical activity (PA) imparts CV health benefits, there remains a knowledge gap on the interactions of PA with systemic inflammation on risk of CV death among American Indians. Hypothesis: To investigate the relationship between PA, inflammation, and CV death, we tested the hypotheses that: (i) PA is associated with lower risk of CV death, and (ii) the CV benefit of PA is mediated via an inflammatory pathway. Methods: We examined the association between PA (leisure-time plus occupational as METs[Metabolic Equivalent of Task]-hours per week via a validated questionnaire) and CV death among 3,135 adults (mean age 56 ± 8) without baseline CV disease in Strong Heart Study (1989-1991), a longitudinal study among American Indians. Preacher and Hayes bootstrap method was used for the model assessing mediation by fibrinogen. Results: During 26 years (mean 17.3 ± 8 years) of follow-up, there were 378 CV deaths. Compared to participants with minimal activity, subjects with higher levels of PA had a lower risk of CV death. Odds ratios were 0.91 (95% confidence interval [CI]: 0.65-1.21), 0.69 (95% CI: 0.51-0.92), and 0.56 (95% CI: 0.39-0.71) (p-trend=0.003), for increasing quartile of activity compared to the lowest, after adjustment for age, sex, study site, education, smoking, diabetes, hypertension, BMI, LDL cholesterol, and urine albumin-creatinine ratio. Mediation analysis showed the effect of PA on CV death was fully mediated via fibrinogen level (correlation coefficient: 0.998, 95% CI 0.995-0.999; Sobel test z= -3.11, p=0.03) (Figure). Conclusion: In a population-based study of American Indians, PA compared to inactivity, was associated with lower risk of CV death. This study highlights that PA is associated with improved CV outcomes and this effect is largely driven by inflammation.

Background: In the US, American Indians suffer a disproportionate burden of CHD compared to other racial/ethnic groups. Additional strategies are needed to identify individuals at risk. Objectives: Investigate the association of blood DNA methylation (DNAm) with incident CHD in the Strong Heart Study and the prediction ability of DNAm beyond traditional risk factors. We maximized prediction ability using Bayesian Hierarchical Cox (BHCox) and Survival Random Forest models, which allow large numbers of CpGs in a single model, instead of considering them individually. Methods: Among 2325 men and women 45-74 years old in 1989-1991, 557 CHD events were identified over 20 years of follow-up. DNAm was measured in 790,026 CpGs, pre-processed and corrected for batch effects. We ran adjusted BHCox models for subsets of CpGs selected from similarly adjusted Cox models for individual CpGs. Prediction ability of CpGs in BHCox was further evaluated with Survival Random Forest, which is robust to overfitting. We also conducted a targeted analysis using Cox regression. Results: 26 CpGs associated with CVD in previous studies were nominally associated with incident CHD in a targeted analysis. The cross-validated C index for the model with traditional risk factors was 0.703. In BHCox, further entering 30K CpGs in a single model, resulted in 231 CpGs being significantly associated with incident CHD, with a cross-validation C statistic of 0.855. In Survival Random Forest, further entering the 231 CpGs from the BHCox model resulted in a cross-validation C statistic of 0.771 and 182 CpGs with variable importance (VIMP) greater than zero. The top CpGs in BHCox were also VIMP>0 in random forest models and were located in SLC24A1 (calcium exchanger), SHBG (sex hormone binding globulin), and LINC00346 (non-protein coding RNA 346). Conclusions: We found novel CpG sites prospectively associated with incident CHD and confirmed signals from previous studies. DNAm might help to identify individuals at high risk of developing CHD.

Background: Biological aging assessed by both leukocyte telomere length (LTL) and DNA methylation (DNAm) has been associated with CVD and its risk factors. Moreover, LTL is epigenetically regulated. We hypothesized that LTL-associated epigenetic changes are associated with risk of CVD in the community. Objective: To test whether LTL-associated loci are associated with incident CVD, independent of standard risk factors in multi-ethnic cohorts. Method: We evaluated 3,628 participants with complete LTL and DNAm data in four prospective cohorts, including 1,531 American Indians from the Strong Heart Study (SHS, mean age 56, 60% women), 821 non-Hispanic Whites (NHW) from the Framingham Heart Study (FHS, mean age 60, 51% women), 471 NHW, 150 Hispanics, and 162 African Americans from the Multi-ethnic Study of Atherosclerosis (MESA, mean age 70 , 55% women), and 471 NHW and 342 African Americans from the Women’s Health Initiative (WHI, mean age 65 , all women). LTL was quantified by qPCR (SHS, MESA) or Southern blot (FHS, WHI). DNAm was assayed by Illumina EPIC (SHS) or 450K (FHS, MESA and WHI) arrays. We imputed 450K to 850K using random forest algorithms. CVD events included fatal and nonfatal MI, CHD, heart failure, stroke, peripheral artery diseases, and cardiovascular deaths. Cohort-specific EWAS was conducted to identify CpGs associated with LTL, adjusting for age, sex, race/ethnicity, smoking, alcohol, BMI, site, cell proportion, and batch. Multiple testing was Bonferroni-corrected (genome-wide P < 2.4 x10 -7 ). Results across studies were combined by random-effects meta-analysis. To examine whether LTL-associated epigenetic loci are associated with CVD risk, we used a weighted methylation score to predict incident CVD by Cox regression, adjusting for age, sex, site, smoking, alcohol, BMI, glucose, SBP, LDL-C, and total cholesterol. Results: We ascertained 2,001 CVD events, including 986 in the SHS (average follow-up 15.2 years), 208 in FHS (average follow-up 7.7 years), 74 in MESA (average follow-up 4.9 years), and 733 in WHI (average follow-up 12.2 years). Meta-EWAS identified 22 CpGs (mapped to 17 unique genes) associated with LTL. Of these, 19 loci (15 negatively and 4 positively associated with LTL) had consistent directionality of association across four cohorts. The most significant genes harboring altered CpG sites included TLL2 (cg10549018, P= 2.42 x 10 -12 ) and TPST1 ( cg10691866, P =8.6 x 10 -10 ). A higher composite methylation score, which reflects longer LTL ( P Conclusion: Altered DNA methylation at 19 CpG loci was significantly associated with LTL. Their combined effects may predict a reduced risk of CVD. The observed associations warrant further investigation.

Introduction: About 30% of all U.S. adults and 70-80% of those with Type 2 diabetes mellitus (T2DM) have non-alcoholic fatty liver disease (NAFLD). Atherosclerotic cardiovascular disease (CVD) is more common among NAFLD patients, albeit with limited evidence from large prospective studies. Hypothesis: American Indians (AI) with NAFLD would be more likely to develop carotid atherosclerosis given their high prevalence of obesity and T2DM. Methods: The Strong Heart Family Study (SHFS) is a population-based family study of CVD and its risk factors in AI. Participants (n=2786; 59.6% female, mean age 40.8 ±17.3 y) were recruited from 12 tribes in 3 regions: Arizona, North/South Dakota, and Oklahoma. Carotid ultrasound-assessed plaque and plaque score were obtained at the baseline examination in 2001-03, and again in surviving participants (n=2406) from 2006 to 2009. NAFLD prevalence was estimated using hepatic steatosis index (HSI). NAFLD fibrosis score was used to separate NAFLD patients with or without advanced fibrosis. HSI is calculated using serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), BMI, diabetes status, and sex, while the NAFLD fibrosis score is derived from age, BMI, diabetes status, ALT, AST, platelet count, and serum albumin. Plaque progression was defined as any increase in plaque score (0 to 8 carotid segments with plaque) from baseline to the second examination. Frailty model was used in data analyses to account for the relatedness among family members. Results: Mean BMI of participants was 31.3 ± 7.5 kg/m 2 , 19% had T2DM, 32.6% were hypertensive, and 36.3% were current smokers. NAFLD prevalence (HSI > 36) was 78%, while 6% did not have NAFLD (HSI < 30) and 16% participants were in an intermediate category (HSI 30-36). Among participants with NAFLD (n=2096), three fibrosis classes were defined by NAFLD fibrosis scores: 1 no to moderate fibrosis (61%, fibrosis score < -1.455), 2 indeterminate (31%, -1.455 ≤ fibrosis score ≤ 0.676), and 3 cirrhosis or severe cirrhosis (8%, fibrosis score > 0.676). About 31% of participants had carotid atherosclerosis (plaque score ≥ 1) at baseline. About 14% of the 2252 participants who had two carotid ultrasound evaluations had plaque progression and 19% of those without carotid atherosclerosis at baseline (n=1571) had incident plaque at the second visit. Compared to those with no to moderate fibrosis, those who had cirrhosis or severe cirrhosis had higher risks of plaque progression with a hazard ratio (HR, 95% CI) of 1.8 (1.1-2.9, P=0.026), and of incident plaque with a HR (95% CI) of 2.6 (1.5-4.6, P=0.0009), adjusted for sex, smoking, hypertension, albuminuria, LDL-C, and HDL-C. Similar results were found among participants with intermediate fibrosis scores. Conclusions: The prevalence of HSI-defined NAFLD is high in AI. NAFLD fibrosis score predicts both incident and progressive carotid atherosclerosis.

Introduction: Non-alcoholic fatty liver disease has been associated with an increased risk of cardiovascular disease (CVD). Levels of the circulating hepatic enzymes aspartate amino transferase (AST) and alanine amino transferase (ALT) are used as non-invasive biomarkers of liver fat and fibrosis. A number of studies have investigated the potential for these biomarkers of liver function, as well as their ratio AST/ALT, to predict CVD risk, with equivocal results. We hypothesized that these measurements might be important biomarkers of CVD in American Indians (AI) who experience high rates of both liver disease and type 2 diabetes-related CVD. Methods: The Strong Heart Family Study (SHFS) is a population-based family study of heart disease and its risk factors in AI. Participants (n=2760) were recruited from 12 tribes located within 3 geographical regions in Arizona, North/South Dakota, and Oklahoma. Circulating biomarkers and metabolic measures were assessed during a clinic visit in 2001-2003. Participants were followed for morbidity and mortality events until the end of 2016. Median follow-up time was 14.5 years. Results: In this study, 59.8% were female, mean (± standard deviation) age was 40.8 ±17.2 years, mean BMI was 31.3 ± 7.5 kg/m 2 , 19% had type 2 diabetes (T2D), 32.6% were hypertensive, and 36.1% were current smokers. In univariate analyses, higher ALT and AST levels (IU/L) were associated with all components of the metabolic syndrome (MS), i.e., increased waist circumference, triglycerides, blood pressure, fasting glucose, and decreased HDL-C; all p Conclusions: In American Indians of the SHFS, liver function biomarkers are significantly associated with metabolic syndrome and its components, and predict all-cause and cardiovascular mortality, suggesting that common mechanisms impact MS-related CVD risk factors and liver disease biomarker levels in this high-risk population.

Childhood obesity is a global health concern due to its potential to increase cardiometabolic risk across the life course. In the United States (US) the burden of childhood obesity is highest among Hispanic/Latinos, in particular children or adolescents of Mexican descent. Although the genetic epidemiology of childhood obesity has been studied previously, the potential for novel childhood obesity loci in Hispanic/Latinos and the generalizability of previously reported loci to Hispanic/Latino children and adolescents are still unknown. Thus we aimed to conduct a genome-wide association study of childhood obesity in 1,612 Hispanic/Latino children and adolescents (2-18 years) collected as part of one Mexican (n=794 Mexico City Study) and two US (n=362 Children’s Hospital of Philadelphia; n=456 Viva La Familia Study) studies, and to generalize 11 previously reported childhood obesity loci from European descent samples to our Hispanic/Latino samples. Obesity cases and controls were defined by BMI-for-age percentiles based on the Centers for Disease Control and Prevention smoothed and sex-specific growth curves from 2000, wherein cases had percentiles ≥95 th and controls had percentiles ≤85 th . Each study performed a genome-wide logistic regression analysis of single nucleotide polymorphism (SNPs) after adjusting for sex, population stratification and relatedness, as applicable. We combined study results for SNPs >10 minor allele counts and imputation quality ≥0.5 using fixed-effect inverse-variance weighted meta-analysis. A priori, we estimated that in our sample (n effective =1,498) we would have >80% power to detect common SNPs (>15% minor allele frequency) across the genome (p-8 ) that increase the odds of childhood obesity of 55% per risk allele. Generalizability at 11 known childhood obesity loci was defined as p-6 ), including a SNP that associated with an increased odds of childhood obesity of 54% per risk allele (73% frequent) at ARHGAP21, which is expressed in an enhancer region in brain, muscle and adipose tissues and has been previously implicated with trunk fat mass in Viva la Familia at another SNP (r 2 SEC16B and TMEM18 generalized to Hispanic/Latinos (p≤0.01), corresponding to a 27% and 40% increased odds of obesity per risk allele (22-88% frequency). These preliminary results suggest the presence of novel loci for childhood obesity and the generalizability of genetic loci discovered in samples of European descent to Hispanic/Latinos, albeit with stronger effect sizes. Future work will attempt to identify additional Hispanic/Latino obesity cases and controls to replicate the suggestive associations.

Background: Chronic stress increases the risk of CVD. Allostatic load (AL) is a measure of multisystem physiological dysregulation resulting from long-term response to stress. American Indians (AIs) suffer higher rates of chronic stress and CVD, but no study has examined the association of AL with CVD in this high risk population. Objective: To examine whether baseline AL predicts incident CVD in AIs in the Strong Heart Study. Methods: Among 1400 apparently healthy AIs (free of CVD and diabetes at baseline) followed for an average of 7.8 years, we calculated AL score using 10 markers at baseline, including leukocyte telomere length, fasting glucose, total cholesterol, triglycerides, low- and high- density lipoprotein, estimated glomerular filtration rate, body mass index, and systolic and diastolic blood pressure. Each of the 10 markers was categorized as high or low risk according to criteria listed in Table 1. A composite AL score was calculated by summing the number of high-risk markers. CVD events were defined as fatal and nonfatal MI, congestive heart failure, and stroke. The association of baseline AL with incident CVD was examined by Cox proportional hazard model, adjusting for age, sex, study site, socioeconomic status, smoking and drinking at baseline. Results: The mean age at baseline was 55±8 years. Participants with AL scores of 0, 1, 2, 3 and 4+ comprised of 20%, 30%, 28%, 13%, and 9%, respectively, of the study population. A total of 224 participants developed CVD during the follow-up period. Compared to participants with an AL score of 0, those with a score of 1, (HR=1.62 [1.00 - 2.62]), 2 (HR=2.28 [1.42 - 3.66]), 3 (HR=2.25 [1.33 - 3.80]), or 4+ (HR=3.06 [1.77 - 5.30]) had significantly higher risk for developing CVD. Moreover, there was a dose-response relationship between AL and CVD (trend P Conclusion: A higher AL score at baseline significantly predicts incident CVD among AIs, independent of sociodemographic and behavioral risk factors. The mechanisms underlying this association warrant further investigation.

Background: Atherosclerosis is a complex disease resulting from the joint effects of multiple genetic variants. Cigarette smoking is a strong risk factor for atherosclerotic cardiovascular disease (CVD). Nicotine, the major active component of cigarette smoke, influences CVD by binding to nicotinic acetylcholine receptors (nAChRs). Genetic variants in nAChRs may therefore contribute to atherosclerosis. While each single variant may have only a minor effect, the joint contribution of multiple SNPs to atherosclerosis may be larger. Objective: To investigate the joint impact of 61 SNPs in seven nAChRs genes on subclinical atherosclerosis using a pathway-based approach. Methods: We genotyped 61 tagSNPs in seven nAChRs genes (CHRNA3-A6, CHRNB2-B4) in 3,665 American Indians recruited by the Strong Heart Family Study from Oklahoma (OK), Arizona (AZ) and North/ South Dakota (DK). Subclinical atherosclerosis was assessed by carotid intima-media thickness (cIMT) and carotid plaque score using ultrasound. We performed a pathway-based analysis to evaluate the joint impact of these 61 SNPs on cIMT and plaque score, separately, adjusting for classical cardiovascular risk factors including age, sex, body mass index (BMI), cigarette smoking status (current, former vs. never), alcohol intake (current, former vs. never), high density lipoprotein cholesterol (HDL-C), low density lipoprotein cholesterol (LDL-C), blood pressure, diabetes, levels of physical activity, fibrinogen, and renal function (assessed by estimated glomerular filtration rate). To avoid population substructure, all statistical analyses were stratified by study center (OK, AZ, and DK). Multiple testing was adjusted using Bonferroni correction. Results: Prevalence of smokers (current plus former) was highest in participants from DK (66%), followed by OK (58%) and AZ (51%). Compared to never smokers, smokers had significantly larger cIMT (0.647mm vs. 0.681mm, P Conclusion: Genetic variants in the nAChRs pathway jointly contribute to subclinical atherosclerosis in American Indians. Though the mechanism linking the nAChRs genetic pathway to subclinical atherosclerosis is unclear, our results may provide valuable information for behavioral and/or molecular strategies for reducing cardiovascular risk in American Indians.

Background: Obesity is a strong determinant for cardiovascular disease (CVD). It results from the complex interplay between genetic and environmental factors. Inflammation has been implicated in obesity and its related phenotypes. Leukotrienes are inflammatory mediators generated from arachidonic acid (polyunsaturated n-6 fatty acid) by the enzyme 5-lipoxygenase (5-LO). Genetic variants in genes encoding 5-LO (e.g., ALOX5) are potential contributors to obesity, probably through nutrigenetic interactions. Objectives: To examine whether genetic variants in ALOX5 are associated with obesity measures, and whether the association is modified by dietary intake of n-6 fatty acids. Methods: Thirteen tagging SNPs in ALOX5 were genotyped in 3,665 American Indians recruited by the Strong Heart Family Study from Oklahoma (OK), South/North Dakota (DK) and Arizona (AZ). Dietary intake was assessed with a food frequency questionnaire. Gene-based association and haplotype × diet interactions were investigated using generalized estimating equation, controlling for age, sex, smoking, alcohol intake, lipids, blood pressure, levels of physical activity, fibrinogen, renal function and total caloric intake. Because dietary intakes of n-3 and n-6 fatty acids are highly correlated, the nutrigenetic interactions were additionally adjusted for n-3 fatty acids. All analyses were stratified by geographic sites. Multiple testing was adjusted using the false discovery rate. Results: A 13-SNP haplotype, named HapZ, is significantly associated with obesity. Subjects carrying HapZ have significantly smaller obesity measures, including body mass index (BMI), waist circumference (WC) and body fat, compared to those not carrying this haplotype (all P Conclusion: A novel ALOX5 haplotype is significantly and negatively associated with obesity in American Indians. Dietary intake of n-6 fatty acids modifies the apparent beneficial effect of this haplotype on obesity. If replicated, our findings could potentially lead to individualized lifestyle intervention for reducing the risk of obesity, diabetes and CVD.Funding(This research has received full or partial funding support from the American Heart Association, National Center)