[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
Purchase Options:
[Skip to Content Landing]
Figure.
Population Attributable Fractions of Coronary Heart Disease Variables in the Early Era (1983-1990) vs Late Era (1996-2002)
Population Attributable Fractions of Coronary Heart Disease Variables in the Early Era (1983-1990) vs Late Era (1996-2002)

Abbreviations: CHD, coronary heart disease; SBP, systolic blood pressure; total:HDL-C, total to high-density lipoprotein cholesterol ratio. Error bars indicate 95% CIs.

Table.  
Characteristics and Risk Factor Burden in Adults With vs Without CHD in 2 Different Periodsa
Characteristics and Risk Factor Burden in Adults With vs Without CHD in 2 Different Periodsa
1.
Moser  M.  From JNC I to JNC 7—what have we learned?  Prog Cardiovasc Dis. 2006;48(5):303-315.PubMedGoogle ScholarCrossref
2.
Talwalkar  PG, Sreenivas  CG, Gulati  A, Baxi  H.  Journey in guidelines for lipid management.  Indian J Endocrinol Metab. 2013;17(4):628-635.PubMedGoogle ScholarCrossref
3.
Giffen  CA, Carroll  LE, Adams  JT, Brennan  SP, Coady  SA, Wagner  EL.  Providing contemporary access to historical biospecimen collections.  Biopreserv Biobank. 2015;13(4):271-279.PubMedGoogle ScholarCrossref
4.
Laaksonen  MA, Knekt  P, Härkänen  T, Virtala  E, Oja  H.  Estimation of the population attributable fraction for mortality in a cohort study using a piecewise constant hazards model.  Am J Epidemiol. 2010;171(7):837-847.PubMedGoogle ScholarCrossref
5.
Tobin  MD, Sheehan  NA, Scurrah  KJ, Burton  PR.  Adjusting for treatment effects in studies of quantitative traits.  Stat Med. 2005;24(19):2911-2935.PubMedGoogle ScholarCrossref
6.
Laaksonen  MA, Virtala  E, Knekt  P, Oja  H, Härkänen  T.  SAS macros for calculation of population attributable fraction in a cohort study design.  J Stat Softw. 2011;43(7):1-25. https://www.jstatsoft.org/article/view/v043i07. Accessed October 18, 2016.PubMedGoogle ScholarCrossref
Research Letter
November 15, 2016

Temporal Changes in the Association Between Modifiable Risk Factors and Coronary Heart Disease Incidence

Author Affiliations
  • 1Duke Clinical Research Institute, Duke University Medical Center, Durham, North Carolina
  • 2Regeneron Pharmaceuticals, Tarrytown, New York
  • 3Mike Rosenbloom Laboratory for Cardiovascular Research, McGill University Health Centre, Montreal, Canada
  • 4Department of Mathematics and Statistics, Boston University, Boston, Massachusetts
JAMA. 2016;316(19):2041-2043. doi:10.1001/jama.2016.13614

Diagnosis and control of coronary heart disease (CHD) risk factors have received particular emphasis in guidelines issued since 1977 (blood pressure) and 1985 (lipids).1,2 Yet on a population level, little is known about how these efforts have altered CHD incidence and its association with modifiable risk factors. This study explored (1) how the associations between modifiable risk factors and CHD events changed from 1983 through 1995 and from 1996 through 2011, and (2) during this timeframe, whether the population attributable fractions (PAFs) of CHD due to modifiable risk factors were altered.

Methods

Individual patient-level data from 5 observational cohort studies (Table) available in the National Heart, Lung, and Blood Institute Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC) were pooled.3 Two analytic data sets were created: 1 set with baseline data collected from 1983 through 1990 (early era) with follow-up from 1996 through 2001 and 1 set with baseline data collected from 1996 through 2002 (late era) with follow-up from 2007 through 2011. Participants aged 40 to 79 years who were free of cardiovascular disease were selected from each era and matched on age (within 2 years), race (black vs nonblack), and sex. Each cohort was followed for up to 12 years for new-onset CHD (ie, myocardial infarction, coronary death, angina, coronary insufficiency) using outcomes available in BioLINCC. A piecewise constant hazards model adjusted for age, sex, and race was used to estimate the hazard ratios (HRs) of CHD due to systolic blood pressure (SBP), diabetes, smoking (current or within past year), and total to high-density lipoprotein cholesterol (total:HDL-C) ratio.4 Blood pressure and total cholesterol were adjusted for treatment using a nonparametric approach.5 HRs were compared by testing the interaction between risk factor and era. PAFs were computed and compared using 10-year survival probabilities from the model with CIs derived using the delta method and log transformation.6 Analysis was approved by the Duke University institutional review board and conducted using SAS (SAS Institute), version 9.4. Statistical significance used 2-sided α of .05.

Results

The Table shows characteristics of 14 009 pairs of participants in 2 cohorts. In the late era, the proportions of adults who smoked, had SBP of 140 mm Hg or greater, or had a total:HDL-C of 4.0 or greater were lower than in the early era, whereas the rates of blood pressure treatment and lipid-lowering therapy were higher. The incidence of new-onset CHD declined from 1.18 (95% CI, 1.12-1.25; 1428 events) to 0.98 (95% CI, 0.93-1.04; 1317 events) per 100 person-years (P < .001).

For the outcome of CHD, the HRs of smoking, SBP, and total:HDL-C were not significantly different between eras (Figure), but diabetes HRs were lower in the late era (2.01 [95% CI, 1.76-2.28] for the early era vs 1.49 [95% CI, 1.29-1.71] for the late era; P = .002). Numerical decreases in the PAFs of CHD due to smoking, systolic hypertension (SBP ≥140 mm Hg), and dyslipidemia (total:HDL-C ≥4.0) over time were not statistically significant (Figure). However, the PAF of diabetes declined significantly from 9.8% (95% CI, 7.7%-11.9%) in the earlier era to 5.4% (95% CI, 3.3%-7.6%) in the later era (difference, 4.4% [95% CI, 1.4%-7.4%], P = .004).

Discussion

Examination of adults from 5 large observational cohort studies led to several findings. First, the incidence of CHD declined almost 20% over time. Second, although the prevalence of diabetes increased, the fraction of CHD attributable to diabetes decreased over time, due to attenuation of the association between diabetes and CHD. This may have resulted from changing definitions and awareness of diabetes, improvements in diabetes treatment and control, and/or better primary prevention. Third, there was no evidence that the strength of the association between smoking, SBP, or dyslipidemia and CHD changed between eras, nor was there evidence that the proportion of CHD due to these factors changed. This underscores the importance of continued prevention efforts targeting these risk factors. Study limitations include the unknown effect of increasing treatment rates on follow-up, the variability of PAFs with classification thresholds, and 1 of the cohorts contributing only to the late era.

Section Editor: Jody W. Zylke, MD, Deputy Editor.
Back to top
Article Information

Correction: This article was corrected for an error in the Figure on December 13, 2016.

Corresponding Author: Michael J. Pencina, PhD, Duke Clinical Research Institute, 2400 Pratt St, Durham, NC 27705 (michael.pencina@duke.edu).

Author Contributions: Mr Wojdyla had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Navar, Peterson, Sanchez, D'Agostino, Pencina.

Acquisition, analysis, or interpretation of data: Navar, Wojdyla, Sniderman, D'Agostino, Pencina.

Drafting of the manuscript: Navar, Peterson, Sanchez, Pencina.

Critical revision of the manuscript for important intellectual content: Navar, Wojdyla, Sanchez, Sniderman, D'Agostino, Pencina.

Statistical analysis: Peterson, Wojdyla, D'Agostino.

Obtained funding: Navar, Pencina.

Administrative, technical, or material support: Navar, Peterson, Sanchez.

Study supervision: Navar, Pencina.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Navar reports receiving grant funding from Sanofi and Regeneron and consulting fees from Sanofi. Dr Peterson reports receiving grant funding from American College of Cardiology, American Heart Association, and Janssen; and consulting fees from Bayer, Boehringer Ingelheim, Merck, Valeant, Sanofi, AstraZeneca, Janssen, Regeneron, and Genentech. Dr Sanchez reports holding stock in Regeneron. Dr Pencina reports receiving grant funding from Regeneron and Sanofi to his institution. No other disclosures were reported.

Funding/Support: This study was supported by Regeneron and Sanofi Pharmaceuticals.

Role of the Funder/Sponsor: The funders 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 beyond the contributions of its employees as outlined in the Acknowledgments and in the description of Dr Sanchez’s role.

Disclaimer: Dr Peterson, an associate editor for JAMA, was not involved in the editorial review of or the decision to publish this article.

Additional Contributions: We thank Tony Schibler, MPA, and Brian Tinga, MSAE, for their data programming assistance, and Erin Hanley, MS, for her editorial assistance with this article (all from the Duke Clinical Research Institute). We also recognize the intellectual contributions of Irfan Khan, PhD (Sanofi), and Joseph Elassal, MD (Regeneron), in data interpretation and study design. These contributors did not receive compensation beyond their regular salary provided by their employers.

References
1.
Moser  M.  From JNC I to JNC 7—what have we learned?  Prog Cardiovasc Dis. 2006;48(5):303-315.PubMedGoogle ScholarCrossref
2.
Talwalkar  PG, Sreenivas  CG, Gulati  A, Baxi  H.  Journey in guidelines for lipid management.  Indian J Endocrinol Metab. 2013;17(4):628-635.PubMedGoogle ScholarCrossref
3.
Giffen  CA, Carroll  LE, Adams  JT, Brennan  SP, Coady  SA, Wagner  EL.  Providing contemporary access to historical biospecimen collections.  Biopreserv Biobank. 2015;13(4):271-279.PubMedGoogle ScholarCrossref
4.
Laaksonen  MA, Knekt  P, Härkänen  T, Virtala  E, Oja  H.  Estimation of the population attributable fraction for mortality in a cohort study using a piecewise constant hazards model.  Am J Epidemiol. 2010;171(7):837-847.PubMedGoogle ScholarCrossref
5.
Tobin  MD, Sheehan  NA, Scurrah  KJ, Burton  PR.  Adjusting for treatment effects in studies of quantitative traits.  Stat Med. 2005;24(19):2911-2935.PubMedGoogle ScholarCrossref
6.
Laaksonen  MA, Virtala  E, Knekt  P, Oja  H, Härkänen  T.  SAS macros for calculation of population attributable fraction in a cohort study design.  J Stat Softw. 2011;43(7):1-25. https://www.jstatsoft.org/article/view/v043i07. Accessed October 18, 2016.PubMedGoogle ScholarCrossref
×