TY - JOUR
T1 - Performance of the American Heart Association/American College of Cardiology Pooled Cohort Equations to Estimate Atherosclerotic Cardiovascular Disease Risk by Self-reported Physical Activity Levels
AU - Pandey, Ambarish
AU - Mehta, Anurag
AU - Paluch, Amanda
AU - Ning, Hongyan
AU - Carnethon, Mercedes R.
AU - Allen, Norrina B.
AU - Michos, Erin D.
AU - Berry, Jarett D.
AU - Lloyd-Jones, Donald M.
AU - Wilkins, John T.
N1 - Funding Information:
Pooling Project was supported in its inception by the NIH National Heart, Lung, and Blood Institute grant R21HL085375 and is currently supported by funds from the Northwestern University Feinberg School of Medicine. Dr Pandey received research funding from Texas Health Resources Clinical Scholarship, Gilead Sciences Research Scholar Program, and National Institute of Aging GEMSSTAR grant (1R03AG067960-01), and an investigator-initiated grant from Applied Therapeutics.
Funding Information:
reported serving in an unpaid position on the advisory board of Roche Diagnostics. Dr Berry reported receiving grants from the National Institutes of Health (NIH), grants from Abbott, and personal fees from Cooper Institute during the conduct of the study; and personal fees from Roche and Astra Zeneca outside the submitted work. Dr Lloyd-Jones reported receiving grants from the NIH and American Hospital Association during the conduct of the study. No other disclosures were reported.
Publisher Copyright:
© 2021 American Medical Association. All rights reserved.
PY - 2021/6
Y1 - 2021/6
N2 - Importance: The American Heart Association/American College of Cardiology pooled cohort equations (PCEs) are used for predicting 10-year atherosclerotic cardiovascular disease (ASCVD) risk. Pooled cohort equation risk prediction capabilities across self-reported leisure-time physical activity (LTPA) levels and the change in model performance with addition of LTPA to the PCE are unclear. Objective: To evaluate PCE risk prediction performance across self-reported LTPA levels and the change in model performance by adding LTPA to the existing PCE model. Design, Setting, and Participants: Individual-level pooling of data from 3 longitudinal cohort studies - Atherosclerosis Risk in Communities, Multi-Ethnic Study of Atherosclerosis, and Cardiovascular Health Study - was performed. A total of 18824 participants were stratified into 4 groups based on self-reported LTPA levels: inactive (0 metabolic equivalent of task [MET]-min/wk), less than guideline-recommended (<500 MET-min/wk), guideline-recommended (500-1000 MET-min/week), and greater than guideline-recommended (>1000 MET-min/wk). Pooled cohort equation risk discrimination was studied using the C statistic and reclassification capabilities were studied using the Greenwood Nam-D'Agostino χ2goodness-of-fit test. Change in risk discrimination and reclassification on adding LTPA to PCEs was evaluated using change in C statistic, integrated discrimination index, and categorical net reclassification index. Main Outcomes and Measures: Adjudicated ASCVD events during 10-year follow-up. Results: Among 18824 participants studied, 10 302 were women (54.7%); mean (SD) age was 57.6 (8.2) years. A total of 5868 participants (31.2%) were inactive, 3849 (20.4%) had less than guideline-recommended LTPA, 3372 (17.9%) had guideline-recommended LTPA, and 5735 (30.5%) had greater than guideline-recommended LTPA level. Higher LTPA levels were associated with a lower risk of ASCVD after adjustment for risk factors (hazard ratio [HR] per 1-SD higher LTPA, 0.91; 95% CI, 0.86-0.96). Across LTPA groups, PCE risk discrimination (C statistic, 0.76-0.78) and risk calibration (all χ2P >.10) was similar. Addition of LTPA to the PCE model resulted in no significant change in the C statistic (0.0005; 95% CI, -0.0004 to 0.0015; P =.28) and categorical net reclassification index (-0.003; 95% CI, -0.010 to 0.010; P =.95), but a minimal improvement in the integrated discrimination index (0.0008; 95% CI, 0.0002-0.0013; P =.005) was observed. Similar results were noted when cohort-specific coefficients were used for creating the baseline model. Conclusions and Relevance: Higher self-reported LTPA levels appear to be associated with lower ASCVD risk and increasing LTPA promotes cardiovascular wellness. These findings suggest the American Heart Association/American College of Cardiology PCEs are accurate at estimating the probability of 10-year ASCVD risk regardless of LTPA level. The addition of self-reported LTPA to PCEs does not appear to be associated with improvement in risk prediction model performance.
AB - Importance: The American Heart Association/American College of Cardiology pooled cohort equations (PCEs) are used for predicting 10-year atherosclerotic cardiovascular disease (ASCVD) risk. Pooled cohort equation risk prediction capabilities across self-reported leisure-time physical activity (LTPA) levels and the change in model performance with addition of LTPA to the PCE are unclear. Objective: To evaluate PCE risk prediction performance across self-reported LTPA levels and the change in model performance by adding LTPA to the existing PCE model. Design, Setting, and Participants: Individual-level pooling of data from 3 longitudinal cohort studies - Atherosclerosis Risk in Communities, Multi-Ethnic Study of Atherosclerosis, and Cardiovascular Health Study - was performed. A total of 18824 participants were stratified into 4 groups based on self-reported LTPA levels: inactive (0 metabolic equivalent of task [MET]-min/wk), less than guideline-recommended (<500 MET-min/wk), guideline-recommended (500-1000 MET-min/week), and greater than guideline-recommended (>1000 MET-min/wk). Pooled cohort equation risk discrimination was studied using the C statistic and reclassification capabilities were studied using the Greenwood Nam-D'Agostino χ2goodness-of-fit test. Change in risk discrimination and reclassification on adding LTPA to PCEs was evaluated using change in C statistic, integrated discrimination index, and categorical net reclassification index. Main Outcomes and Measures: Adjudicated ASCVD events during 10-year follow-up. Results: Among 18824 participants studied, 10 302 were women (54.7%); mean (SD) age was 57.6 (8.2) years. A total of 5868 participants (31.2%) were inactive, 3849 (20.4%) had less than guideline-recommended LTPA, 3372 (17.9%) had guideline-recommended LTPA, and 5735 (30.5%) had greater than guideline-recommended LTPA level. Higher LTPA levels were associated with a lower risk of ASCVD after adjustment for risk factors (hazard ratio [HR] per 1-SD higher LTPA, 0.91; 95% CI, 0.86-0.96). Across LTPA groups, PCE risk discrimination (C statistic, 0.76-0.78) and risk calibration (all χ2P >.10) was similar. Addition of LTPA to the PCE model resulted in no significant change in the C statistic (0.0005; 95% CI, -0.0004 to 0.0015; P =.28) and categorical net reclassification index (-0.003; 95% CI, -0.010 to 0.010; P =.95), but a minimal improvement in the integrated discrimination index (0.0008; 95% CI, 0.0002-0.0013; P =.005) was observed. Similar results were noted when cohort-specific coefficients were used for creating the baseline model. Conclusions and Relevance: Higher self-reported LTPA levels appear to be associated with lower ASCVD risk and increasing LTPA promotes cardiovascular wellness. These findings suggest the American Heart Association/American College of Cardiology PCEs are accurate at estimating the probability of 10-year ASCVD risk regardless of LTPA level. The addition of self-reported LTPA to PCEs does not appear to be associated with improvement in risk prediction model performance.
UR - http://www.scopus.com/inward/record.url?scp=85105210751&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85105210751&partnerID=8YFLogxK
U2 - 10.1001/jamacardio.2021.0948
DO - 10.1001/jamacardio.2021.0948
M3 - Article
C2 - 33909016
AN - SCOPUS:85105210751
SN - 2380-6583
VL - 6
SP - 690
EP - 696
JO - JAMA cardiology
JF - JAMA cardiology
IS - 6
ER -
