Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Young Adult Exposure to Cardiovascular Risk Factors and Risk of Events Later in Life: The Framingham Offspring Study

  • Mark J. Pletcher ,

    mpletcher@epi.ucsf.edu

    Affiliation Departments of Epidemiology and Biostatistics, & Medicine, University of California, San Francisco, San Francisco, California, United States of America

  • Eric Vittinghoff,

    Affiliation Departments of Epidemiology and Biostatistics, & Medicine, University of California, San Francisco, San Francisco, California, United States of America

  • Anusorn Thanataveerat,

    Affiliation Division of General Medicine, Columbia University Medical Center, New York, New York, United States of America

  • Kirsten Bibbins-Domingo,

    Affiliation Departments of Epidemiology and Biostatistics, & Medicine, University of California, San Francisco, San Francisco, California, United States of America

  • Andrew E. Moran

    Affiliation Division of General Medicine, Columbia University Medical Center, New York, New York, United States of America

Abstract

Background

It is unclear whether coronary heart disease (CHD) risk factor exposure during early adulthood contributes to CHD risk later in life. Our objective was to analyze whether extent of early adult exposures to systolic and diastolic blood pressure (SBP, DBP) and low-and high-density lipoprotein cholesterol (LDL, HDL) are independent predictors of CHD events later in life.

Methods and Findings

We used all available measurements of SBP, DBP, LDL, and HDL collected over 40 years in the Framingham Offspring Study to estimate risk factor trajectories, starting at age 20 years, for all participants. Average early adult (age 20–39) exposure to each risk factor was then estimated, and used to predict CHD events (myocardial infarction or CHD death) after age 40, with adjustment for risk factor exposures later in life (age 40+). 4860 participants contributed an average of 6.3 risk factor measurements from in-person examinations and 24.5 years of follow-up after age 40, and 510 had a first CHD event. Early adult exposures to high SBP, DBP, LDL or low HDL were associated with 8- to 30-fold increases in later life CHD event rates, but were also strongly correlated with risk factor levels later in life. After adjustment for later life levels and other risk factors, early adult DBP and LDL remained strongly associated with later life risk. Compared with DBP≤70 mmHg, adjusted hazard ratios (HRs) were 2.1 (95% confidence interval: 0.8–5.7) for DBP = 71–80, 2.6 (0.9–7.2) for DBP = 81–90, and 3.6 (1.2–11) for DBP>90 (p-trend = 0.019). Compared with LDL≤100 mg/dl, adjusted HRs were 1.5 (0.9–2.6) for LDL = 101–130, 2.2 (1.2–4.0) for LDL = 131–160, and 2.4 (1.2–4.7) for LDL>160 (p-trend = 0.009). While current levels of SBP and HDL were also associated with CHD events, we did not detect an independent association with early adult exposure to either of these risk factors.

Conclusions

Using a mixed modeling approach to estimation of young adult exposures with trajectory analysis, we detected independent associations between estimated early adult exposures to non-optimal DBP and LDL and CHD events later in life.

Introduction

High blood pressure and cholesterol are risk factors for coronary heart disease (CHD)[1, 2], and many randomized controlled trials have demonstrated reduced CHD event rates when these risk factors are treated with medications in middle-aged and older adults[35]. It remains unclear, however, whether cardiovascular risk factor exposure during young adulthood is important for future CHD risk, and whether treatment at this age might modify risk later in life. While blood pressure guidelines recommend medication to treat high blood pressure throughout life including young adulthood[6, 7], cholesterol guidelines depend primarily on predicted 10-year cardiovascular disease risk to target treatment and thus tend not to recommend cholesterol-lowering medication in young adults with high cholesterol unless cholesterol is extremely high[8].

Several lines of evidence suggest that both high blood pressure and high cholesterol cause lasting cardiovascular damage during young adulthood. Early life risk factors are associated with atherosclerosis[912] that persists into middle age[1315]. Cohorts with very long term follow up have shown that blood pressure and cholesterol measured once during young adulthood predict CHD events much later in life[1621], but these associations could be wholly attributable to later-life risk factor levels, which usually track with early life levels. Prior analyses of the Framingham Heart Study demonstrated persistent effects of remote exposure to cholesterol and blood pressure on atherosclerosis[22] and CHD events[23], but these studies focused on exposure during middle-age and not young adulthood. A recent analysis of the Framingham Offspring Cohort, which included many young adults at its inception, found that duration of exposure to hyperlipidemia at age 35–55 was associated with CHD events[24], but did not adjust for any other young adult risk factor exposures (e.g., blood pressure) that correlate with hyperlipidemia. No prior studies have teased apart the independent effects of risk factor levels during young adulthood from risk factor levels later in life on the occurrence of CHD events.

We used all available measurements of cardiovascular risk factors (blood pressure and cholesterol) measured repeatedly in the Framingham Offspring Study over 40 years to model complete risk factor trajectories starting at age 20 for all study participants, accounting for the discontinuous effects of time-varying medication use. We used those trajectories to estimate average exposure from age 20–39 (young adulthood), and analyzed associations between average risk factor exposure during young adulthood and CHD events occurring later in life.

Methods

Approach, study design and participants

Our approach, which relies on mixed effects modeling, is designed to take advantage of all available risk factor measurements to model age-based risk factor trajectories for each study participant. Mixed models can be used to estimate the latent trajectories underlying the observed values for each participant, and, “borrowing strength” across participants, to extrapolate those trajectories a moderate distance beyond the range of the observed data if trajectories are relatively smooth. Using these methods, we imputed risk factor trajectories for each participant to obtain complete estimated risk factor trajectories for each participant from age 20 years forward. We used these trajectories to estimate average risk factor exposure during young adulthood (age 20–39 years) and then assess the associations of these early exposures with cardiovascular events.

We used measurements and follow up from the Framingham Offspring Study, an institutional review board-approved community-based cohort study. The study was comprised primarily of children of participants in the Framingham Heart Study Original Cohort and their spouses, began examinations and follow up in 1971[25], and has conducted in-person examinations every 4–8 years with continuous follow up for cardiovascular events. We included participants with at least one measurement each, at baseline or during follow-up, of body mass index, blood pressure, lipids and medication use, and without a prior myocardial infarction or stroke who were followed beyond the age of 40 for cardiovascular events. Our analysis of the de-identified dataset, which we accessed through the National Heart, Lung and Blood Institute’s BioLINCC repository[26], was approved by Columbia University’s Institutional Review Board (IRB-AAAI5961).

Cardiovascular risk factor measurements

Cardiovascular risk factors were measured at seven in-person examinations using standard protocols[27]. Systolic and diastolic blood pressures (SBP and DBP) were measured by a study physician two times while seated. Fasting total cholesterol, high-density lipoprotein cholesterol (HDL) and triglycerides were measured using standard enzymatic methods and low-density lipoprotein cholesterol (LDL) was calculated according to the Friedewald equation[28]; we used only LDL and HDL for this analysis. Height and weight measurements were used to calculate body mass index (BMI) in kg/m2. Diabetes status was defined by fasting hyperglycemia (>126 mg/dl) or self-reported diabetes treatment at each examination[26], and assumed always to be present after meeting this definition once. Smoking status (current, past, or never) and cigarette packs/day smoked were assessed by self-report. Current medication lists (either with specific names or more general categories, depending on the exam) were collected at each examination, and used to categorize participants as either using or not using any blood pressure and any cholesterol medication at each examination.

Cardiovascular events

Incident cardiovascular events were ascertained by passive reporting between visits, active surveillance of local newspaper obituaries, hospital, and mortuary records, and by survey during examination interviews, and adjudicated according to standard protocols. For this analysis, we defined our primary outcome as the occurrence of a CHD event including myocardial infarction (recognized, unrecognized or silent) or CHD death. We censored observation time after the occurrence of a cerebrovascular event including fatal and non-fatal stroke (embolic, hemorrhagic or other) or transient ischemic attack.

Statistical analysis

We used a series of mixed models to estimate age-based risk factor trajectories and obtain fitted values for each risk factor from age 20 years until the end of observed follow up time for each participant, including for participants whose first Framingham examination was after age 40. Gender-specific age trajectories were flexibly modeled using restricted cubic splines with knots at the 10th, 50th, and 90th percentiles of the age distribution. We started by modeling BMI trajectories as a function of age and gender, with random effects for participant (intercept plus both age spline components), and used the model to estimate annual best linear unbiased predictions (BLUPs) of BMI for each participant. Age at diabetes onset and first use of blood pressure and cholesterol medication, which could be right, left, or interval-censored, were estimated based on log-normal survival models, with gender and BMI as covariates. We estimated age at subsequent cessation and/or re-initiation of medication use by assuming these events occurred at the midpoint between sequential visits where use differed. Finally, we modeled SBP, DBP, LDL and HDL (log-transforming SBP, LDL and HDL), which are medication-responsive, using fixed effects for age, gender, diabetes, BMI, and medication use (both current and history of), and random effects for participant intercept, age (both spline components), and current medication use (allowing the effect of medication to vary by participant). Smooth trajectories were obtained from the BLUPs using a standard back-transformation for log-transformed outcomes. Cholesterol medication was used in the models for both LDL and HDL, and blood pressure medication in the model for SBP and DBP. Among smokers, packs/day was linearly interpolated between visits, while age at cessation among current smokers at the final visit was estimated using a pooled logistic model.

We then used these individual trajectories to estimate time-weighted average exposures. Specifically, young adult exposure was calculated as the time-weighted average for ages 20–39. We also estimated the time-weighted average from age 40 on (also using estimated trajectories), as well as the “current” measurement (most recent measurement with the last directly measured value carried forward, not using estimated trajectories), and used these as time-varying covariates to adjust for “later life exposure” to risk factors. The inclusion of these multiple ways of measuring later life exposure in our statistical models should allow relatively complete accounting for their effects, and more fully isolate the putative effects of early life exposure, which is the primary goal of this analysis.

We used survival analysis methods to analyze the association between young adulthood risk factors exposure (age 20–39) and CHD events later in life (after age 40). With the origin for time to event set at age 40 years and censoring for fatal and non-fatal stroke and non-CHD deaths, we estimated unadjusted CHD event rates, and then used a series of Cox proportional hazards models (with age as the time scale) to adjust for covariates. All multivariable models were adjusted for sex, calendar year (via spline), BMI, diabetes, years with diabetes, smoking status (current/past/never), pack-years of tobacco exposure (via spline), current use of blood pressure and lipid medications, and other cardiovascular risk factors (SBP, DBP, LDL and HDL). For example, our initial model for early life exposure to LDL cholesterol was adjusted for current levels of SBP, DBP and HDL as well as average exposures to SBP, DBP and HDL from age 20–39 and from age 40 on, and for time-varying exposure to lipid-lowering medications along with the other variables. We then adjusted for later life exposure to the predictor of interest; in the example above, we additionally adjusted for current LDL levels and average exposure from age 40 on.

To examine the robustness and consistency of our findings, we reran statistical analyses restricted to persons with at least one direct risk factor measurement before the age of 40, tested for interactions by gender, excluded persons using blood pressure or lipid medication, used only SBP without DBP and vice-versa, used continuous versions of risk factor exposure variables using splines to account for non-linearity of associations, and explored different versions of the continuous models guided by analyses of collinearity.

We used SAS version 9.3 or Stata version 13 for all analyses.

Results

Of 5124 total Framingham Offspring Study participants in the BIOLINCC dataset, we excluded 186 with completely missing information on BMI, blood pressure, lipids or medication use, 36 for pre-existing cardiovascular disease, and an additional 42 who did not reach the age of 40 before the end of their observed follow up. The remaining 4860 participants included in the study sample were approximately half male (48%) and ranged from 20–70 years of age at the time of their first examination; 2969 (61%) contributed direct measurements of cardiovascular risk factors at an in-person Framingham examination before the age of 40. Most participants (97%) contributed more than one direct measurement over time (mean 6.3/person, range 1–7), and the average length of uncensored observation time after age 40 was over 20 years (Table 1).

thumbnail
Table 1. Framingham Offspring Study Sample Examinations and Observation Period.

https://doi.org/10.1371/journal.pone.0154288.t001

All available direct measurements were used to estimate trajectories of SBP, DBP, LDL, and HDL from age 20 through the end of follow up for each participant. The Figure in S1 Appendix illustrates fitted trajectories for 20 individual participants in comparison with measured values. By the end of follow up, 1719 participants (35%) were using blood pressure medications and 699 (14%) were using lipid medications. Time-weighted average measurements of SBP, DBP, LDL and HDL from age 20–39 were strongly correlated with measurements later in life (Table A in S1 Appendix). Early adult exposure to SBP, DBP, LDL and HDL were all very strongly associated with each other and with other cardiovascular risk factors (see Table 2, below, for LDL, and Tables B-D for SBP, DBP and HDL in S1 Appendix).

thumbnail
Table 2. Characteristics of participants at the beginning of follow up, stratified by early life exposure to LDL.

https://doi.org/10.1371/journal.pone.0154288.t002

510 participants suffered a first CHD event during follow-up. Unadjusted event rates were 8-fold to 30-fold higher in persons with adverse levels of risk factor exposure during young adulthood (age 20–39 years) compared with persons with optimal levels of exposure (Table 3). These associations were attenuated after adjustment for other CHD risk factors and later life exposure, but remained highly significant for DBP and LDL (Table 3). Compared with DBP≤70 mmHg, adjusted hazard ratios (HRs) for non-optimal DBP levels were 2.1 (95% confidence interval (CI): 0.8–5.7) for DBP = 71–80, 2.6 (0.9–7.2) for DBP = 81–90, and 3.6 (1.2–11) for DBP>90 (p-trend = 0.019). Compared with LDL≤100 mg/dl, adjusted HRs for non-optimal LDL levels were 1.5 (0.9–2.6) for LDL = 101–130, 2.2 (1.2–4.0) for LDL = 131–160, and 2.4 (1.2–4.7) for LDL>160 (p-trend = 0.009). Early adult exposures to SBP and HDL were not independently associated with future CHD events (Table 3).

thumbnail
Table 3. Coronary Heart Disease Events in Framingham Participants with Differing Exposure to Risk Factors During Young Adulthood.

https://doi.org/10.1371/journal.pone.0154288.t003

In the fully adjusted model, current SBP was also strongly associated with CHD events; early adult SBP was not (Fig 1). While none of the three HDL variables were independently associated with CHD events in the full model (p-trend = 0.28–0.47), a test of the overall contribution of all HDL variables to the model was highly significant (p = 0.0013, Fig 1). When only the young adult HDL variable was included in the model, it was highly significant (Table 3); the same is true for current HDL (HRs 1.3–2.1, p-trend < .001). In the full model, the two later life LDL variables (time-weighted average after age 40, and current level) were not independently associated with the outcome (Fig 1), but these variables were strongly correlated (Table A in S1 Appendix), and a joint test of significance for both of these later life LDL variables was borderline significant (p = .053). When we dropped the time-weighted average after age 40 variable, the current level variable became significantly associated with the outcome (p for trend = 0.017), and the early life LDL association became even stronger (p for trend = .0005).

thumbnail
Fig 1. Adjusted associations between SBP, DBP, LDL and HDL levels at different ages and CHD events.

Hazard ratios (95% confidence intervals) are adjusted for age (via Cox model), sex, calendar year (via spline), body mass index, diabetes, years with diabetes, smoking status (current/past/never), pack-years of tobacco exposure (via spline), and use of blood pressure and lipid medications. The first column of results (for age 20–39) corresponds to the right-hand column of Table 2. Categories for systolic blood pressure (SBP) are ≤120 (reference), 121–140, 141–160 and >160 mmHg; for diastolic blood pressure (DBP) are ≤80, 81–90, 91–100, and >100; for low-density lipoprotein cholesterol (LDL) are ≤100 (reference), 101–130, 131–160 and >160 mg/dl; and for high-density lipoprotein cholesterol (HDL) are >65 (reference), 51–65, 36–50, and ≤35 mg/dl. “P Overall” refers to a test of the overall contribution of the risk factor (including early, later, and current exposures) to the model. No participants had an average SBP>160 mmHg from age 20–39. The * indicates a truncated confidence interval.

https://doi.org/10.1371/journal.pone.0154288.g001

Sensitivity analyses demonstrated consistent patterns of association. When we limited our analysis to only the 2969 participants with at least one direct risk factor measurements before age 40, point estimates were similar or even larger for early adult DBP (HRs 1.5–4.4, p-trend = 0.03) and LDL (HRs 2.1–3.0, p-trend = 0.05) and remained null for early adult SBP and HDL (see Table E in S1 Appendix for a full tabulation of results). When early adult DBP was omitted from the model, early adult SBP remained only marginally associated with CHD events (HRs 1.3–1.4, p-trend = 0.39). Among the 2667 participants who never reported using blood pressure or lipid medications, early adult LDL remained significantly associated with CHD events (HRs 1.3–3.4, p-trend = 0.03) but DBP did not, although the HRs were similar in magnitude (HRs 1.2–2.6, p-trend = 0.52); only 12 participants in this subgroup had early adult DBP>90. Among early adult risk factors, only DBP interacted significantly with gender (p < .001); but only 55 men had early adult DBP< = 70, and none of these men had CHD events. When this variable was recoded collapsing categories < = 70 with 71–80, no interaction was noted (p = 0.52). In a model containing continuous instead of categorical risk factors, we found no evidence of non-linearity in any of the predictors, and the significant predictors were the same: current SBP HR = 1.3 (1.1–1.6) per 20 mmHg, early adult DBP HR = 1.6 (1.1–2.3) per 10 mmHg, and early adult LDL HR = 1.5 (1.1–2.0) per 30 mg/dl. Using continuous predictors in the 2969 participants with at least one direct risk factor measurement before age 40 yielded similar results after modifying the analysis to account for collinearity (Table F in S1 Appendix).

Discussion

In this analysis of the Framingham Offspring Study, we estimated trajectories of SBP, DBP, LDL and HDL during early adulthood and later life, and found early adult exposures to DBP and LDL to be associated with subsequent CHD events. These associations were independent of later life exposures and current risk factor levels. SBP and HDL were also associated with CHD events, but in contrast to DBP and LDL, we could not confirm that exposure to SBP or HDL early in life contributes independently to CHD events.

Our finding that DBP and LDL exposure early in life is associated with CHD events extends a line of inquiry on blood pressure, cholesterol and CHD that began in the 1950’s with early analyses of the Framingham Heart Study’s Original Cohort[29, 30]. Findings in adults led to studies of children and young adults showing associations between these risk factors and atherosclerosis measured by autopsy[9, 12, 31, 32] and non-invasive testing[10, 11, 15, 3336]. Prior longitudinal analyses of the Framingham Heart Study[22], Atherosclerotic Risk in Community (ARIC) Study[37], and the Coronary Artery Risk Development in Young Adults (CARDIA) Study[13, 14, 38] suggest that atherosclerotic damage may accumulate and persist over many years into middle and late adulthood, and that early life risk factor exposure may be associated with later life atherosclerosis independent from later life exposure levels[13, 14]. Long-term follow-up studies have demonstrated associations between blood pressure and cholesterol measured once during young adulthood with CHD events later in life[1621], but whether this association is independent of later life risk factor levels has remained unclear. A recent analysis of the Framingham Offspring Cohort, which included many young adults at its inception, found that duration of exposure to hyperlipidemia at age 35–55 was associated with CHD events[24], but did not adjust for any other young adult risk factor exposures (e.g., blood pressure) that strongly correlate with hyperlipidemia (Table 2). Because of strong correlations between risk factors and within risk factor across the life course (“tracking”), teasing apart the effects of early versus later life exposure to each individual risk factor is difficult. Our analysis, which used direct risk factor measurements from over 30,000 study visits and nearly 120,000 person-years of follow up for first CHD events accrued by the Framingham Offspring Study, allowed us to tease apart these factors and detect independent associations between DBP and LDL exposure during early adulthood and CHD events later in life.

Our finding that DBP, not SBP, was a predictor of later life CHD risk is consistent with a previous Framingham analysis showing that DBP is a stronger predictor of events early in life than SBP[39]. Those investigators proposed that different pulse-wave reflection effects on SBP in normotensive and hypertensive young adults obscured the association of SBP with CHD risk in adults <50 years old, a mechanism that changes with stiffening of the arteries later in life. Later in life, both our and prior analyses[39] show SBP to be a strong predictor of events. Because few participants experienced raised SBP during young adulthood, we may also have lacked statistical power to make reliable inference about the association of SBP exposure during ages 20–39 years with later life risk.

Our results are also consistent with recent genetic analyses that bolster the case for a lasting, causal effect of early life exposure. These studies demonstrate that lifelong reductions in blood pressure and cholesterol starting at birth due to genetic variation lead to larger reductions in CHD risk than would be predicted from the degree of differences in risk factor levels later in life[4043]. These effects are consistent both for single genes (e.g., lower LDL from the R46L allele in PCSK9[40, 41]) and for multivariate genetic risk factor scores for both blood pressure[42] and cholesterol[40, 41, 43]. In contrast, and consistent with our results, a multi-polymorphism score for HDL was not independently associated with CHD[43].

Our analytic methods allow us to estimate risk factor trajectories using all available data starting at age 20 and through the end of follow up, but they are imperfect. Extrapolating backwards to age 20 years is prone to measurement error and may be biased by failure to adequately specify cohort and secular effects (though calendar year was included in all adjusted models). If we assume that our measurement error is mostly non-differential and trajectory estimates for persons with relatively fewer or no measurements before the age of 40 are subject to extra “shrinkage” towards the sample means, our point estimates of effect are likely biased towards the null[44], and perhaps variably so (participants with more direct measurements earlier in life should have less measurement error in the young adult average exposure estimate). However, using all participants despite variable measurement error and borrowing strength across participants using our mixed modeling approach appears to have provided adequate statistical power to detect the presence of an independent association between early adult risk factor exposure and later life CHD risk factors that would have been hard to detect without these methods. We believe our results provide some evidence in support of the presence of an early adult exposure effect on later life CHD events, but are likely biased from regression dilution and should not be taken as an accurate estimate of the size of such an effect. Additional methodologic work is needed to validate the approach we have taken and account for variable measurement error and regression dilution. There is also the real possibility of residual confounding. We attempted to adjust systematically for potential confounders, but residual confounding from imperfect measurement or unmeasured factors remains a possible explanation for our results. For example, exposure to blood pressure and lipids even before age 20 could confound our results if childhood (or even prenatal) exposure to these risk factors is the true driver of later life atherosclerosis[45, 46]; we did not have measurements before age 20 and could not therefore adjust for these factors. Our analysis was focused on detecting associations with early life exposure to risk factors that were independent of later life risk factor exposure, and not on quantifying associations with later life risk factors. By including two versions of later life risk factor exposure in our models, we helped ensure isolation of early life exposure, but simultaneously made it difficult to quantify the independent associations with later life exposure because of collinearity. Thus, the lack of an association between later life LDL and outcomes in our primary outcome model (Fig 1) should not be taken as evidence of an absence of association, especially given very strong prior evidence to the contrary[3, 4] and the results of our sensitivity analysis dropping one of the two later life LDL variables and the joint association test (as described in the Results), which both support the expected association with later life LDL. As with all Framingham-based analyses, the lack of racial and ethnic diversity in the cohort limits our ability to make inferences about non-White populations.

If our findings truly indicate causal and modifiable effects of early life risk factor exposure, they could have important implications. Though current hypertension guidelines recommend treating elevated DBP in young adults, current guidelines for treatment of high cholesterol recommend first estimating medium-term (10-year) atherosclerotic cardiovascular disease risk and recommend statin treatment only when risk is greater than 7.5% in 10 years[8]. These recommendations effectively exclude treatment of young adults unless LDL is extremely high (e.g., >190 mg/dl). If our findings are true, however, young adulthood may be an important window during which LDL-lowering could reduce CHD events later in life. Benefits from treatment, however, would be extremely difficult to confirm. Randomized controlled trials of statin treatment in young adults would require large sample sizes and very long follow-up to detect any delayed effects of statins, and would need to “compete” with later life statin initiation, which is clearly beneficial in many persons[4, 8]. It is unclear, also, whether potential benefits of pharmacotherapy accruing far in the future would be valued by young patients concerned about immediate downsides including cost, side effects, convenience, and other factors that might reduce quality of life. Modeling can help clarify these theoretical tradeoffs, and additional observational studies including diverse participants should be analyzed to confirm our results. Meanwhile, current guidelines strongly recommend optimizing risk factors by maintaining a heart-healthy lifestyle throughout life, including young adulthood[8, 47]; our findings suggest that doing so may indeed provide benefits later in life.

Supporting Information

Acknowledgments

This study was supported by U.S. National Heart, Lung, and Blood Institute grant (R01 HL107475-01) to Dr. Moran and Dr. Pletcher, and a Columbia University Irving Scholarship to Dr. Moran. We used the Framingham Cohort and Framingham Offspring Research Materials obtained from the U.S. National Heart, Lung and Blood Institute (NHLBI) Biologic Specimen and Data Repository Information Coordinating Center. The manuscript does not necessarily reflect the opinions or views of the Framingham Cohort, Framingham Offspring, or the NHLBI. Kirsten Bibbins-Domingo is a member of the United States Preventive Services Task Force (USPSTF) and current co-Vice Chair. This work does not necessarily represent the views and policies of the USPSTF.

Author Contributions

Conceived and designed the experiments: MJP EV KBD AEM. Analyzed the data: MJP AT EV. Wrote the paper: MJP EV AT KBD AEM.

References

  1. 1. Vasan RS, Larson MG, Leip EP, Evans JC, O'Donnell CJ, Kannel WB, et al. Impact of high-normal blood pressure on the risk of cardiovascular disease. The New England journal of medicine. 2001;345(18):1291–7. pmid:11794147
  2. 2. Lewington S, Clarke R, Qizilbash N, Peto R, Collins R. Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies. Lancet. 2002;360(9349):1903–13. pmid:12493255
  3. 3. Efficacy and safety of cholesterol-lowering treatment: prospective meta-analysis of data from 90,056 participants in 14 randomised trials of statins. Lancet. 2005;366(9493):1267–78. pmid:16214597
  4. 4. Mihaylova B, Emberson J, Blackwell L, Keech A, Simes J, Barnes EH, et al. The effects of lowering LDL cholesterol with statin therapy in people at low risk of vascular disease: meta-analysis of individual data from 27 randomised trials. Lancet. 2012;380(9841):581–90. pmid:22607822
  5. 5. Ninomiya T, Perkovic V, Turnbull F, Neal B, Barzi F, Cass A, et al. Blood pressure lowering and major cardiovascular events in people with and without chronic kidney disease: meta-analysis of randomised controlled trials. BMJ. 2013;347:f5680. pmid:24092942
  6. 6. James PA, Oparil S, Carter BL, Cushman WC, Dennison-Himmelfarb C, Handler J, et al. 2014 Evidence-Based Guideline for the Management of High Blood Pressure in Adults: Report From the Panel Members Appointed to the Eighth Joint National Committee (JNC 8). JAMA. 2014.
  7. 7. Go AS, Bauman MA, Coleman King SM, Fonarow GC, Lawrence W, Williams KA, et al. An effective approach to high blood pressure control: a science advisory from the American Heart Association, the American College of Cardiology, and the Centers for Disease Control and Prevention. Hypertension. 2014;63(4):878–85. pmid:24243703
  8. 8. Stone NJ, Robinson J, Lichtenstein AH, Merz CN, Blum CB, Eckel RH, et al. 2013 ACC/AHA Guideline on the Treatment of Blood Cholesterol to Reduce Atherosclerotic Cardiovascular Risk in Adults: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2013.
  9. 9. Berenson GS, Srinivasan SR, Bao W, Newman WP 3rd, Tracy RE, Wattigney WA. Association between multiple cardiovascular risk factors and atherosclerosis in children and young adults. The Bogalusa Heart Study. N Engl J Med. 1998;338(23):1650–6. pmid:9614255
  10. 10. Mahoney LT, Burns TL, Stanford W, Thompson BH, Witt JD, Rost CA, et al. Coronary risk factors measured in childhood and young adult life are associated with coronary artery calcificaction in young adults: the Muscatine Study. J Am Coll Cardiol. 1996;27(2):277–84. pmid:8557894
  11. 11. Raitakari OT, Juonala M, Kahonen M, Taittonen L, Laitinen T, Maki-Torkko N, et al. Cardiovascular risk factors in childhood and carotid artery intima-media thickness in adulthood: the Cardiovascular Risk in Young Finns Study. JAMA. 2003;290(17):2277–83. pmid:14600186
  12. 12. Relationship of atherosclerosis in young men to serum lipoprotein cholesterol concentrations and smoking. A preliminary report from the Pathobiological Determinants of Atherosclerosis in Youth (PDAY) Research Group. JAMA. 1990;264(23):3018–24. pmid:2243430
  13. 13. Pletcher MJ, Bibbins-Domingo K, Lewis CE, Wei GS, Sidney S, Carr JJ, et al. Prehypertension during young adulthood and coronary calcium later in life. Ann Intern Med. 2008;149(2):91–9. pmid:18626048
  14. 14. Pletcher MJ, Bibbins-Domingo K, Liu K, Sidney S, Lin F, Vittinghoff E, et al. Non-optimal lipids commonly present in young adults and coronary calcium later in life: the CARDIA (Coronary Artery Risk Development in Young Adults) Study. Annals of Internal Medicine. 2010;153(3):137–46. pmid:20679558
  15. 15. Davis PH, Dawson JD, Riley WA, Lauer RM. Carotid intimal-medial thickness is related to cardiovascular risk factors measured from childhood through middle age: The Muscatine Study. Circulation. 2001;104(23):2815–9. pmid:11733400
  16. 16. Stamler J, Daviglus ML, Garside DB, Dyer AR, Greenland P, Neaton JD. Relationship of baseline serum cholesterol levels in 3 large cohorts of younger men to long-term coronary, cardiovascular, and all-cause mortality and to longevity. JAMA. 2000;284(3):311–8. pmid:10891962
  17. 17. Gray L, Lee IM, Sesso HD, Batty GD. Blood pressure in early adulthood, hypertension in middle age, and future cardiovascular disease mortality: HAHS (Harvard Alumni Health Study). Journal of the American College of Cardiology. 2011;58(23):2396–403. pmid:22115646
  18. 18. Anderson KM, Castelli WP, Levy D. Cholesterol and mortality. 30 years of follow-up from the Framingham study. JAMA. 1987;257(16):2176–80. pmid:3560398
  19. 19. Klag MJ, Ford DE, Mead LA, He J, Whelton PK, Liang KY, et al. Serum cholesterol in young men and subsequent cardiovascular disease. N Engl J Med. 1993;328(5):313–8. pmid:8419817
  20. 20. Seshadri S, Wolf PA, Beiser A, Vasan RS, Wilson PW, Kase CS, et al. Elevated midlife blood pressure increases stroke risk in elderly persons: the Framingham Study. Arch Intern Med. 2001;161(19):2343–50. pmid:11606150
  21. 21. Miura K, Daviglus ML, Dyer AR, Liu K, Garside DB, Stamler J, et al. Relationship of blood pressure to 25-year mortality due to coronary heart disease, cardiovascular diseases, and all causes in young adult men: the Chicago Heart Association Detection Project in Industry. Arch Intern Med. 2001;161(12):1501–8. pmid:11427097
  22. 22. Wilson PW, Hoeg JM, D'Agostino RB, Silbershatz H, Belanger AM, Poehlmann H, et al. Cumulative effects of high cholesterol levels, high blood pressure, and cigarette smoking on carotid stenosis. N Engl J Med. 1997;337(8):516–22. pmid:9262494
  23. 23. Vasan RS, Massaro JM, Wilson PW, Seshadri S, Wolf PA, Levy D, et al. Antecedent blood pressure and risk of cardiovascular disease: the Framingham Heart Study. Circulation. 2002;105(1):48–53. pmid:11772875
  24. 24. Navar-Boggan AM, Peterson ED, D'Agostino RB, Sr., Neely B, Sniderman AD, Pencina MJ. Hyperlipidemia in early adulthood increases long-term risk of coronary heart disease. Circulation. 2015;131(5):451–8. pmid:25623155
  25. 25. Kannel WB, Feinleib M, McNamara PM, Garrison RJ, Castelli WP. An investigation of coronary heart disease in families. The Framingham offspring study. American journal of epidemiology. 1979;110(3):281–90. pmid:474565
  26. 26. Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC): National Heart, Lung, and Blood Institute; [cited 2014 Sept 1]. Available: https://biolincc.nhlbi.nih.gov/home/.
  27. 27. Friedman GD, Cutter GR, Donahue RP, Hughes GH, Hulley SB, Jacobs DR Jr., et al. CARDIA: study design, recruitment, and some characteristics of the examined subjects. J Clin Epidemiol. 1988;41(11):1105–16. pmid:3204420
  28. 28. Bild DE, Jacobs DR, Liu K, Williams OD, Hilner JE, Perkins LL, et al. Seven-year trends in plasma low-density-lipoprotein-cholesterol in young adults: the CARDIA Study. Ann Epidemiol. 1996;6(3):235–45. pmid:8827159
  29. 29. Dawber TR, Moore FE, Mann GV. Coronary heart disease in the Framingham study. Am J Public Health Nations Health. 1957;47(4 Pt 2):4–24. pmid:13411327
  30. 30. Kannel WB, Dawber TR, Kagan A, Revotskie N, Stokes J 3rd. Factors of risk in the development of coronary heart disease—six year follow-up experience. The Framingham Study. Ann Intern Med. 1961;55:33–50. pmid:13751193
  31. 31. Newman WP 3rd, Freedman DS, Voors AW, Gard PD, Srinivasan SR, Cresanta JL, et al. Relation of serum lipoprotein levels and systolic blood pressure to early atherosclerosis. The Bogalusa Heart Study. N Engl J Med. 1986;314(3):138–44. pmid:3455748
  32. 32. McGill HC Jr., McMahan CA, Malcom GT, Oalmann MC, Strong JP. Effects of serum lipoproteins and smoking on atherosclerosis in young men and women. The PDAY Research Group. Pathobiological Determinants of Atherosclerosis in Youth. Arterioscler Thromb Vasc Biol. 1997;17(1):95–106. pmid:9012643
  33. 33. Tonstad S, Joakimsen O, Stensland-Bugge E, Leren TP, Ose L, Russell D, et al. Risk factors related to carotid intima-media thickness and plaque in children with familial hypercholesterolemia and control subjects. Arterioscler Thromb Vasc Biol. 1996;16(8):984–91. pmid:8696963
  34. 34. Sanchez A, Barth JD, Zhang L. The carotid artery wall thickness in teenagers is related to their diet and the typical risk factors of heart disease among adults. Atherosclerosis. 2000;152(1):265–6. pmid:11203160
  35. 35. Knoflach M, Kiechl S, Kind M, Said M, Sief R, Gisinger M, et al. Cardiovascular risk factors and atherosclerosis in young males: ARMY study (Atherosclerosis Risk-Factors in Male Youngsters). Circulation. 2003;108(9):1064–9. pmid:12952827
  36. 36. Li S, Chen W, Srinivasan SR, Bond MG, Tang R, Urbina EM, et al. Childhood cardiovascular risk factors and carotid vascular changes in adulthood: the Bogalusa Heart Study. JAMA. 2003;290(17):2271–6. pmid:14600185
  37. 37. Nieto FJ, Diez-Roux A, Szklo M, Comstock GW, Sharrett AR. Short- and long-term prediction of clinical and subclinical atherosclerosis by traditional risk factors. J Clin Epidemiol. 1999;52(6):559–67. pmid:10408996
  38. 38. Loria CM, Liu K, Lewis CE, Hulley SB, Sidney S, Schreiner PJ, et al. Early adult risk factor levels and subsequent coronary artery calcification: the CARDIA Study. J Am Coll Cardiol. 2007;49(20):2013–20. pmid:17512357
  39. 39. Franklin SS, Larson MG, Khan SA, Wong ND, Leip EP, Kannel WB, et al. Does the relation of blood pressure to coronary heart disease risk change with aging? The Framingham Heart Study. Circulation. 2001;103(9):1245–9. pmid:11238268
  40. 40. Cohen JC, Boerwinkle E, Mosley TH Jr., Hobbs HH. Sequence variations in PCSK9, low LDL, and protection against coronary heart disease. N Engl J Med. 2006;354(12):1264–72. pmid:16554528
  41. 41. Benn M, Nordestgaard BG, Grande P, Schnohr P, Tybjaerg-Hansen A. PCSK9 R46L, low-density lipoprotein cholesterol levels, and risk of ischemic heart disease: 3 independent studies and meta-analyses. Journal of the American College of Cardiology. 2010;55(25):2833–42. pmid:20579540
  42. 42. Ference BA, Julius S, Mahajan N, Levy PD, Williams KA, Sr., Flack JM. Clinical effect of naturally random allocation to lower systolic blood pressure beginning before the development of hypertension. Hypertension. 2014;63(6):1182–8. pmid:24591335
  43. 43. Voight BF, Peloso GM, Orho-Melander M, Frikke-Schmidt R, Barbalic M, Jensen MK, et al. Plasma HDL cholesterol and risk of myocardial infarction: a mendelian randomisation study. Lancet. 2012;380(9841):572–80. pmid:22607825
  44. 44. Clarke R, Shipley M, Lewington S, Youngman L, Collins R, Marmot M, et al. Underestimation of risk associations due to regression dilution in long-term follow-up of prospective studies. Am J Epidemiol. 1999;150(4):341–53. pmid:10453810
  45. 45. Strong JP, Mc GH Jr. The natural history of coronary atherosclerosis. Am J Pathol. 1962;40:37–49. pmid:13917861
  46. 46. Webber LS, Osganian V, Luepker RV, Feldman HA, Stone EJ, Elder JP, et al. Cardiovascular risk factors among third grade children in four regions of the United States. The CATCH Study. Child and Adolescent Trial for Cardiovascular Health. Am J Epidemiol. 1995;141(5):428–39. pmid:7879787
  47. 47. Eckel RH, Jakicic JM, Ard JD, de Jesus JM, Houston Miller N, Hubbard VS, et al. 2013 AHA/ACC guideline on lifestyle management to reduce cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2014;129(25 Suppl 2):S76–99. pmid:24222015