eTable 1. Estimated Proportion of US Adults Aged ≥40 Years in Systolic Blood Pressure Categories According to 2014 Evidence-Based Guideline Subgroups
eTable 2. Estimated Proportion of US Adults Aged ≥40 Years in Systolic Blood Pressure Categories According to 2017 ACC/AHA Guideline Subgroups
eTable 3. Events per 1000 Person-Years (95% Confidence Intervals) in Participants Aged ≥40 Years from ARIC, CHS, FHS Offspring, and MESA According to the 2014 Evidence-Based Guideline and 2017 ACC/AHA Guideline Subgroups
eTable 4. Network Meta-Analysis Hazard Ratios (95% Confidence Intervals) Comparing Higher Systolic Blood Pressure Groups to Lower Systolic Blood Pressure Targets
eTable 5. Population Attributable Risks and Annual Reductions in Cardiovascular Disease and All-Cause Mortality With Systolic Blood Pressure Lowering According to the 2014 Evidence-Based Guideline by Subgroup
eTable 6. Population Attributable Risks and Annual Reductions in Cardiovascular Disease and All-Cause Mortality With Systolic Blood Pressure Lowering According to the 2017 ACC/AHA Guideline by Subgroup
eTable 7. Estimated Numbers of Annual Reductions in Cardiovascular Disease and All-cause Mortality (in Thousands) According to the Proportion of Populations Who Achieved Blood Pressure TargeteReferences
Customize your JAMA Network experience by selecting one or more topics from the list below.
Bundy JD, Mills KT, Chen J, Li C, Greenland P, He J. Estimating the Association of the 2017 and 2014 Hypertension Guidelines With Cardiovascular Events and Deaths in US Adults: An Analysis of National Data. JAMA Cardiol. 2018;3(7):572–581. doi:10.1001/jamacardio.2018.1240
What are the associations of the 2017 ACC/AHA and 2014 evidence-based hypertension guidelines with the proportion of US adults defined as being hypertensive or recommended for antihypertensive treatment?
In this analysis of national data, the estimated prevalence of hypertension increased compared with the 2014 guideline, as did the proportions of individuals recommended for antihypertensive treatment. Achieving the 2017 guideline treatment goals may further reduce 340 000 cardiovascular events and 156 000 total deaths annually compared with the 2014 guideline treatment goals.
Compared with the 2014 guideline, the 2017 guideline may increase the proportion of adults recommended for antihypertensive treatment and further reduce major cardiovascular diseases and all-cause mortality.
The 2017 American College of Cardiology/American Heart Association hypertension guideline recommends lower blood pressure (BP) thresholds for initiating antihypertensive medication and treatment goals than the 2014 evidence-based hypertension guideline.
To estimate the potential association of the 2017 and 2014 hypertension guidelines with the proportion of US adults defined as being hypertensive or recommended for antihypertensive treatment and with risk reduction of major cardiovascular disease (CVD) and all-cause mortality.
Design, Setting, and Participants
Using data from the National Health and Nutrition Examination Survey 2013 to 2016, we estimated the proportions of US adults with hypertension or recommended for antihypertensive treatment according to the 2017 and 2014 hypertension guidelines. Using data from the National Health and Nutrition Examination Survey, antihypertensive clinical trials, and population-based cohort studies, we estimated risk reductions of CVD and all-cause mortality assuming the entire US adult population achieved guideline-recommended systolic BP (SBP) treatment goals. Data were analyzed between October 2017 and March 2018.
Main Outcomes and Measures
Proportions and numbers of individuals with hypertension or recommended for antihypertensive treatment and numbers of CVD and all-cause mortality reduction.
According to the 2017 hypertension guideline, the prevalence of hypertension (BP level ≥130/80 mm Hg) was 45.4% (95% CI, 43.9%-46.9%), representing 105.3 (95% CI, 101.9-108.8) million US adults, which was significantly higher than estimates per the 2014 hypertension guideline (BP level ≥140/90 mm Hg): 32.0% (95% CI, 30.3%-33.6%) or 74.1 (95% CI, 70.3-77.9) million individuals, respectively. Additionally, the proportion of individuals recommended for antihypertensive treatment was significantly higher according to the 2017 hypertension guideline (35.9%; 95% CI, 34.2%-37.5%) compared with the 2014 hypertension guideline (31.1%; 95% CI, 29.6%-32.7%). Achieving the 2017 hypertension guideline SBP treatment goals is estimated to reduce 610 000 (95% CI, 496 000-734 000) CVD events and 334 000 (95% CI, 245 000-434 000) total deaths in US adults 40 years and older. Corresponding estimates after achieving the 2014 hypertension guideline SBP treatment goals were 270 000 (95% CI, 202 000-349 000) and 177 000 (95% CI, 123 000-241 000), respectively. Implementing the 2017 hypertension guideline is estimated to increase 62 000 hypotension and 79 000 acute kidney injury or failure events.
Conclusions and Relevance
Compared with the 2014 hypertension guideline, the 2017 hypertension guideline was associated with an increase in the proportion of adults recommended for antihypertensive treatment and a further reduction in major CVD events and all-cause mortality, but a possible increase in the number of adverse events in the United States.
Hypertension is a leading cause of cardiovascular disease (CVD) and premature death in the United States and worldwide.1-3 Observational epidemiology studies have shown a linear association between usual blood pressure (BP) levels and risk of CVD and all-cause mortality, with the lowest risk at a systolic BP (SBP) level of 115 mm Hg.4 The Systolic Blood Pressure Intervention Trial documented that an intensive SBP treatment target of less than 120 mm Hg was associated with significant reductions of 25% for major CVD and 27% for all-cause mortality in individuals at high risk for CVD.5-7 A 2017 meta-analysis of BP-lowering trials8 also demonstrated significant and linear associations between mean achieved SBP and the risk of CVD and all-cause mortality, with the lowest risks in randomization groups with mean achieved SBP of 120 to 124 mm Hg.8
Supported by evidence on intensive BP treatment, the 2017 American College of Cardiology (ACC)/American Heart Association (AHA) Guideline for the Prevention, Detection, Evaluation, and Management of High Blood Pressure in Adults recommended lower BP thresholds for initiation of antihypertensive medication and treatment goals than previous hypertension guidelines.9-11 It is important to estimate the potential associations of these changes with treatment of hypertension and consequent CVD and mortality in the US population. Muntner et al12 reported a hypertension prevalence of 45.6% and a percentage of US adults recommended for antihypertensive treatment of 36.2% in 2011 through 2014 under the new guideline.12 The objectives of our analysis are 2-fold: to compare the estimated proportions of US adults in 2013 through 2016 defined as having hypertension or recommended for antihypertensive treatment per the 2017 and 2014 hypertension guidelines and to estimate the potential associations of treatment goals with risk reduction of major CVD and all-cause mortality and increases in adverse events in the US adult population.
A more detailed description of the methods is available in the eMethods in the Supplement. The National Health and Nutrition Examination Survey (NHANES) uses a complex, stratified, multistage probability cluster sampling design to select representative samples of the US civilian, noninstitutionalized population.13 Our analyses were limited to NHANES 2013 to 2016 participants 20 years and older (n = 11 062). Further exclusions were made for those who did not have 3 BP measurements (n = 565), had missing data on antihypertensive treatment (n = 10), and had missing data needed to define diabetes, chronic kidney disease (CKD), 10-year predicted risk of CVD, or history of CVD (n = 741). After exclusions, 9746 were used for this analysis. This study was approved by the Tulane University institutional review board, and all participants provided written informed consent as part of the individual protocols of the included studies.
In NHANES, 3 BP measurements were taken using a standard protocol, and the mean of them was used for analysis. Diabetes was defined as glycated hemoglobin level greater than 6.5% (to convert to proportion of total hemoglobin, multiply by 0.01), fasting serum glucose level of at least 126 mg/dL (to convert to millimoles per liter, multiply by 0.0555), nonfasting serum glucose level of at least 200 mg/dL, or self-reported history of diabetes and current use of glucose-lowering medication.14 Estimated glomerular filtration rate was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation,15 and CKD was defined as albumin-to-creatinine ratio of at least 30 mg/g or estimated glomerular filtration rate of 60 mL/min/1.73 m2 or less. High CVD risk is defined as self-reported history of CVD (coronary heart disease, stroke, or heart failure) or 10-year predicted CVD risk of at least 10% using the ACC/AHA Pooled Cohort Equation.9,16
We assessed hypertension status and eligibility for pharmacologic treatment according to criteria from the 2014 evidence-based guideline and the 2017 ACC/AHA guideline (Table 1).9,10 All analyses used NHANES sampling weights to obtain US nationally representative estimates. These weights were recalibrated based on the proportion of participants missing data by age, sex, and race/ethnicity within each NHANES cycle. The numbers of CVD or deaths that could be reduced if the entire population achieved SBP treatment goals were computed using SBP distributions in the US population and the population attributable risks (PARs) related to SBP treatment goals.
We estimated the proportion of US adults 40 years and older in each of 8 SBP categories (<130 mm Hg, 130-134 mm Hg, 135-139 mm Hg, 140-144 mm Hg, 145-149 mm Hg, 150-154 mm Hg, 155-159 mm Hg, and ≥160 mm Hg) by subgroups of sex, age, race/ethnicity, and hypertension guideline treatment categories. Data from NHANES 2011 to 2016 were combined to provide sufficient sample size for adults 40 years and older (n = 11 184). Those who did not have 3 BP measurements (n = 896), had missing data on antihypertensive treatment (n = 13), and had missing data needed to define diabetes, CKD, 10-year predicted risk of CVD, or history of CVD (n = 946) were excluded. After exclusions, 9329 participants were included in this analysis.
We pooled individual-level data of 30 285 participants from 4 community-based US cohort studies: the Atherosclerosis Risk in Communities Study,17 the Cardiovascular Health Study,18 the Framingham Offspring Study,19 and the Multi-Ethnic Study of Atherosclerosis.20 We used Poisson regression models to estimate incidence of major CVD (coronary heart disease, stroke, heart failure, and CVD deaths) and all-cause mortality by age, sex, race/ethnicity, and hypertension guideline categories. The estimates were calibrated to reflect an annual 2.6 million major CVD events (795 000 stroke, 935 000 coronary heart disease, and 870 000 heart failure)21 and 2 488 667 all-cause deaths22 in the US population aged 40 years and older.
We conducted a network analysis of 42 antihypertensive clinical trials to estimate hazard ratios for CVD and mortality comparing each of the 8 SBP categories.8 Briefly, clinical trials that randomly allocated patients to an antihypertensive medication, control, or treatment target and reported a difference in mean achieved SBP of 5 mm Hg or more between comparison groups were included. Hazard ratios for each possible SBP comparison were calculated using Markov Chain Monte Carlo simulation. To account for trial heterogeneity in intervention duration and baseline risk of CVD or mortality, we adjusted for trial length and event rate (or mortality) in the reference groups for each trial in the model.23
We calculated age-, sex-, and race/ethnicity–specific PARs of CVD and mortality owing to SBP higher than recommended treatment goals using the following formula:
where pi is the proportion of SBP category i, HRi is the hazard ratio of CVD (or deaths) in SBP category i, and k is the total number of SBP categories. A given PAR represents the proportion of CVD events (or deaths) that could be reduced by lowering the age-, sex-, and race/ethnicity-specific population’s SBP level to treatment goals based on guideline treatment categories. We multiplied the age-, sex-, and race/ethnicity–specific PARs by the total number of annual CVD events (or deaths) in the US population 40 years and older to estimate the absolute reduction in number of events (or deaths). In addition, we estimated CVD event (or death) reductions assuming 25%, 50%, and 75% of individuals achieved guideline-specific SBP targets in a sensitivity analysis.
We accounted for uncertainty in the estimation of the proportion of SBP categories, annual event rates, and hazard ratios of BP treatment effects using Monte Carlo simulations.24 For each age-specific, sex-specific, race/ethnicity-specific, and guideline category-specific calculation for each study outcome, 10 000 simulations were conducted. The 2.5th and 97.5th percentiles were used to form 95% confidence intervals.
We systematically searched published clinical trials targeting SBP less than 130 mm Hg using MEDLINE and EMBASE.8 Data on adverse events were extracted from each study independently and in duplicate by at least 2 investigators. Absolute risk differences between intervention and control groups on adverse events were combined using inverse variance weighting. Risk differences were applied to the number of individuals recommended to lower SBP levels to less than 130 mm Hg per the 2017 guideline to estimate the expected number of adverse events.
Analyses were conducted using SAS, version 9.4 (SAS Institute, Inc) and R, version 3.4.2 (The R Foundation). A 2-sided P value less than 0.05 was considered statistically significant.
According to the 2017 ACC/AHA Hypertension Guideline, 45.4% of US adults (105.3 million) 20 years and older had hypertension, which was significantly higher than estimates per the 2014 hypertension guideline of 32.0% (74.1 million) (Table 2). The difference in reclassified hypertension was 13.5% or 31.3 million between the 2 guidelines. The difference was greatest in men and those aged 40 to 59 years.
According to the 2017 ACC/AHA hypertension guideline, 35.9% of US adults (83.2 million) were estimated to be eligible for antihypertensive treatment, which was significantly higher than the 31.1% (72.2 million) according to the 2014 hypertension guideline (Table 3). Thus, an additional 4.7% or 11.0 million US adults were eligible for treatment per the new guideline. Among those recommended for treatment per the 2017 guideline, 12.1% (27.9 million) were untreated, 12.9% (29.9 million) were treated but have not achieved goal BP, and 10.9% (25.3 million) were treated and have achieved goal BP. Among those recommended for treatment per the 2014 guideline, 7.3% (16.9 million) were untreated, 6.1% (14.2 million) were treated but have not achieved goal BP, and 17.7% (41.0 million) were treated and have achieved goal BP. There are 6.8% (15.7 million) of treated adults with hypertension who met 2014 guideline BP goals but not 2017 guideline goals.
Compared with the 2014 guideline, the 2017 guideline is estimated to recommend slightly more men than women for antihypertensive treatment: 5.4% (6.0 million) increase in men and 4.1% (5.0 million) increase in women; and substantially more individuals 60 years and older (10.5% increase or 6.6 million) than those younger than 60 years (2.6% increase or 4.4 million). Increases in the proportions recommended treatment were similar by race/ethnicity groups. Among guideline categories, the 2017 guideline recommended more individuals 60 years and older without diabetes or CKD (13.4% or 4.7 million increases) and those with high CVD risk, diabetes, CKD, or 65 years and older (14.2% or 9.7 million increases) for antihypertensive treatment (Table 3).
Systolic BP distributions by 2014 and 2017 guideline categories and age, sex, and race/ethnicity among US adults 40 years and older are shown in eTables 1 and 2 in the Supplement, while estimated event rates of CVD and all-cause mortality are shown in eTable 3 in the Supplement. Hazard ratios of CVD and all-cause mortality associated with SBP levels of at least 130 mm Hg, at least 140 mm Hg, and at least 150 mm Hg from the network analysis are presented in eTable 4 in the Supplement.
An estimated 610 000 CVD events and 334 000 total deaths will be reduced annually in US adults 40 years and older if all individuals with hypertension achieve the 2017 hypertension guideline SBP treatment goals (Table 4). The corresponding reductions of CVD and all-cause mortality after achieving the 2014 hypertension guideline SBP treatment goals were 270 000 and 177 000, respectively. Therefore, implementation of the 2017 guideline could potentially further reduce 340 000 CVD events and 156 000 total deaths annually. Estimated risk reductions were observed in all subgroups of sex, age, and race/ethnicity. However, the magnitude of estimated risk reductions associated with implementing the 2017 guideline compared with the 2014 guideline were greater among those 60 years and older and individuals with high CVD risk, diabetes, or CKD (Table 4; eTables 5 and 6 in the Supplement). Additionally, the 2017 guideline resulted in a greater risk reduction than the 2014 guideline even when a comparatively lower percentage of patients with hypertension achieve 2017 guideline target BP levels (eTable 7 in the Supplement). For example, 50% achievement of the 2017 hypertension guideline would reduce an estimated 305 000 CVD events and 167 000 total deaths. On the other hand, 75% achievement of the 2014 hypertension guideline would reduce only 203 000 CVD events and 133 000 total deaths.
The number needed to treat (NNT) to prevent 1 major CVD event was 70 (range, 63-91) in sex, age, and race/ethnicity subgroups according to the 2017 guideline; the NNT was 88 (range, 61-226) according to the 2014 guideline (Table 4). The NNT for CVD was significantly lower in women and those younger than 60 years according to the 2017 guideline compared with the 2014 guideline. The NNT to prevent 1 death was 129 (range, 105-240) in sex, age, and race/ethnicity subgroups according to the 2017 guideline and 134 (range, 90-390) according to the 2014 guideline. The NNT for death was significantly lower among those younger than 60 years according to the 2017 guideline compared with the 2014 guideline.
Three trials achieving SBP levels less than 130 mm Hg in the intensive treatment group and reporting adverse events were included in this analysis.5,25-27 Treated SBP levels less than 130 mm Hg were significantly associated with increased risk of hypotension, syncope, electrolyte abnormality, and acute kidney injury or acute renal failure (Table 5). Implementing a SBP goal of less than 130 mm Hg according to the 2017 ACC/AHA guideline was estimated to increase 62 000 hypotension, 32 000 syncope, 31 000 electrolyte abnormality, and 79 000 acute kidney injury or acute renal failure events.
The 2017 ACC/AHA hypertension guideline differs substantially from the 2014 evidence-based guideline, particularly regarding the definition of hypertension and recommendation for antihypertensive treatment.9,10 The 2014 evidence-based guideline recommended BP treatment goals less than 150/90 mm Hg in the general population 60 years and older without diabetes or CKD, while the 2017 ACC/AHA guideline recommended BP treatment goals of less than 130/80 mm Hg in all adults with hypertension.9 Furthermore, the 2017 ACC/AHA guideline recommends a comprehensive approach to care and treatment of BP, including the use of team-based care, home blood pressure monitoring, and an emphasis on standardized, consistent, and accurate measurement of BP.
The 2017 ACC/AHA hypertension guideline lowers BP thresholds for diagnosis of hypertension from 140/90 mm Hg to 130/80 mm Hg, classifying those with SBP levels between 130 and 139 mm Hg as stage 1 and those with SBP levels of at least 140 mm Hg as stage 2.9 Based on this new criterion, we estimated that 45.4% or 105.3 million US adults 20 years and older have hypertension, which represents a 13.5% or 31.3 million–person increase from estimates based on previous criteria. Given the linear association between BP levels and risk of CVD and all-cause mortality, lower BP thresholds might be justifiable.3,4 The new criteria increase the number of adults who would be newly diagnosed as hypertensive and receive lifestyle interventions and antihypertensive medication treatment. However, the new criteria also result in an increased number of adults diagnosed as hypertensive who might not be expected to develop CVD events. The health and economic consequences of the new diagnostic criteria for hypertension should be evaluated in future studies.
We estimated that with the 2017 ACC/AHA hypertension guideline, 83.2 million US adults (35.9%) would be eligible for antihypertensive pharmacologic treatment. Of them, 27.9 million individuals (12.1%) would require initiation of antihypertensive medications and 29.9 million (12.9%) would require antihypertensive treatment intensification. Compared with the 2014 evidence-based hypertension guideline, an additional 11.0 million US adults (4.7%) were eligible for antihypertensive treatment according to the 2017 ACC/AHA hypertension guideline. The guideline further differentiates recommendations among those with stage 1 hypertension, based on 10-year CVD risk. While comprehensive lifestyle interventions based on 6 major components are recommended among all individuals with hypertension, they are the primary mode of intervention in those with 10-year CVD risk less than 10%.9
The new treatment recommendations have the largest effect in individuals 60 years and older and those with high CVD risk, diabetes, or CKD. Lowering BP treatment goals from less than 150/90 mm Hg to less than 130/80 in adults with hypertension 60 years and older would result in a net increase of 6.6 million US adults (10.5%) requiring initiation of antihypertensive treatment and 9.9 million (15.5%) requiring treatment intensification. Likewise, lowering BP treatment goals from less than 140/90 mm Hg to less than 130/80 mm Hg in individuals with high CVD risk, diabetes, CKD, or 65 years and older would result in an increase of 9.7 million adults (14.2%) requiring initiation of antihypertensive treatment and 8.4 million (12.3%) requiring treatment intensification.
We estimated that 610 000 major CVD events and 334 000 total deaths could be prevented annually in the United States if all individuals 40 years and older achieved the 2017 ACC/AHA hypertension guideline SBP treatment goals. Compared with the 2014 guideline, the 2017 treatment recommendations could reduce an additional 340 000 major CVD events and 156 000 deaths annually among the US adult population aged 40 years and older. Furthermore, we estimated that the reductions of CVD events and total deaths were greater under the 2017 hypertension guideline even when fewer patients achieve the BP treatment target compared with the 2014 hypertension guideline. Additionally, the new treatment recommendations slightly improve NNT,28 especially in individuals younger than 60 years, supporting a 2017 analysis suggesting intensive BP control is cost-effective, despite projected increases in office visits and medication use.29 It has been well documented in clinical trials that BP lowering provided progressively greater absolute risk reductions as baseline risk increased.30 Among individuals with heterogeneous CVD risk, such as those younger than 60 years, use of predicted baseline CVD risk to inform antihypertensive treatment decisions is evidence-based.
We also estimated the likely adverse events of full implementation of the 2017 ACC/AHA hypertension guideline. Achieving SBP treatment goals of less than 130 mm Hg in patients with hypertension might increase risk of hypotension, syncope, electrolyte abnormality, and acute kidney injury or acute renal failure, although these events were uncommon. Furthermore, the numbers needed to harm for adverse events are substantially larger than the NNT for CVD and all-cause mortality risk reduction, suggesting the benefits of the lower SBP goal outweigh the potential risks. Based on potential risks associated with intensive BP lowering, the 2017 ACC/AHA hypertension guideline emphasizes individualized care, especially for elderly patients, using clinical judgment and patient preference for decisions regarding intensity of BP lowering and choice of therapy.
Using data from NHANES 2011 to 2014, Muntner et al12 estimated that the prevalence of hypertension was 45.6% (103.3 million) and antihypertensive medication was recommended for 36.2% of US adults (81.9 million) according to the 2017 ACC/AHA guideline.12 These estimates are similar to our analyses, which used more recent NHANES 2013-2016 data. Furthermore, we estimated the absolute risk reductions of major CVD and all-cause mortality and the potential number of adverse events associated with implementation of the 2017 hypertension guideline. Therefore, our article provides a unique and important contribution to the discussion about the new hypertension guideline and antihypertensive treatment in the US population.
Our study has several limitations. First, we had insufficient sample size from NHANES and US cohort studies to reliably estimate the effect of the new guideline within subgroups. Second, 3 BP measurements were obtained at 1 NHANES visit, while guidelines recommend BP measurements be taken at 2 separate clinical visits. The 2017 ACC/AHA guideline in particular promotes standardization and consistency of BP measurement, which is crucial in accurately identifying and managing those with hypertension. Third, because there were no US national data on CVD incidence, we used 4 US population-based cohort studies to estimate incidence by subgroups, and calibrated event counts to values from AHA/National Heart, Lung, and Blood Institute reports.21 Fourth, we were not able to estimate hazard ratios of CVD or all-cause mortality for SBP levels less than 130 mm Hg by subgroups of CVD absolute risk or comorbidities. However, analyses of antihypertensive trials showed consistent proportional risk reductions across various subgroups.30-32 Additionally, these hazard ratios are based on results from efficacy studies and may represent greater risk reductions than trials of effectiveness or implementation. Fifth, there were insufficient data from BP-lowering clinical trials on CVD subtype or kidney disease. Therefore, these clinical outcomes were not included in the network analysis.8 Finally, most antihypertensive trials did not report adverse events; if reported, they were inconsistent.33 Therefore, our analysis of adverse events was based on 3 trials.
The 2017 hypertension guideline defines a significant increase in the prevalence of hypertension by 31.3 million (13.5%), and the proportion of US adults recommended for antihypertensive treatment by 11.0 million (4.7%). If SBP treatment goals were achieved, the 2017 ACC/AHA hypertension guideline recommendations are estimated to reduce major CVD events by an additional 340 000 and total deaths by an additional 156 000 compared with the 2014 evidence-based hypertension guideline but may increase the number of adverse events.
Corresponding Author: Jiang He, MD, PhD, Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, 1440 Canal St, New Orleans, LA 70112 (firstname.lastname@example.org).
Accepted for Publication: April 2, 2018.
Published Online: May 23, 2018. doi:10.1001/jamacardio.2018.1240
Author Contributions: Dr He had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Drs Bundy and Mills contributed equally to this work.
Concept and design: Bundy, Mills, He.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Bundy, Mills, He.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Bundy, Mills, Chen, Li.
Obtained funding: He.
Administrative, technical, or material support: He.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Funding/Support: Research reported in this publication was supported by grants from the National Heart, Lung, and Blood Institute of the National Institutes of Health (R01HL133790) and by the National Institute of General Medical Sciences of the National Institutes of Health (P20GM109036). Dr Bundy was partially supported by the National Heart, Lung, and Blood Institute (T32HL069771).
Role of the Funder/Sponsor: The National Institutes of Health 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.
Additional Contributions: We thank Hua He, PhD, Department of Epidemiology, Tulane University School of Public Health and Tropical Medicine, New Orleans, Louisiana, for her statistical support of this study; she did not receive compensation.
Disclaimer: The manuscript was prepared using research materials obtained from the National Heart, Lung, and Blood Institute Biologic Specimen and Data Repository Information Coordinating Center and does not necessarily reflect the opinions or views of the Atherosclerosis Risk in Communities Study, the Cardiovascular Health Study, the Framingham Offspring Study, the Multi-Ethnic Study of Atherosclerosis, or the National Heart, Lung, and Blood Institute.
Create a personal account or sign in to: