Abstract P241: 2013 AHA/ACC/TOS and 1998 NHLBI Clinical Guidelines for the Management of Overweight and Obesity Predict Higher Cardiovascular Disease Risk in Adults Recommended versus Not Recommended for Weight Loss Treatment: the Atherosclerosis Risk in Communities Study
Introduction: In 2013, AHA/ACC/TOS released an algorithm to assist primary care providers in the identification of patients who should be considered for weight loss treatment to promote the prevention and treatment of cardiovascular disease (CVD). This guidance updated the recommendations released in 1998 by NHLBI. We know of no studies that have compared risk of incident CVD in adults recommended or not recommended for treatment by these two guidelines.
Hypothesis: We hypothesized that adults recommended for weight loss treatment by the 1998 and 2013 algorithms have higher CVD risk than those not recommended for treatment. Additional analyses explored sources of differences in the algorithms as predictors of CVD risk.
Methods: The ARIC study included 13,020 African American and White adults aged 45-64 years. Baseline data collected in 1987-1989 on BMI, waist circumference and other CVD risk factors in participants were used to form treatment groups that differed between the algorithms in regard to the choice, definition and number of risk factors. We calculated hazard ratios (HR) for first CVD event defined as ischemic stroke, fatal and non-fatal myocardial infarction (MI), silent MI and coronary revascularization procedures. Covariates included age, gender, ethnicity and study center.
Results: At baseline, mean BMI was 28 kg/m2 (SD 5.3). During a median follow-up of 19.8 years, 2,698 incident CVD outcomes were recorded. The 1998 and 2013 algorithms recommended weight loss for 58% and 63% of participants, respectively, with 13% discrepant. The HR for CVD in adults recommended vs not recommended for treatment were 1.85 (95% CI: 1.69, 2.00) for the 1998 algorithm and 1.59 (95% CI: 1.47, 1.72) for the 2013 algorithm. The higher HR for the 1998 algorithm was driven in part by the inclusion of three risk factors (age, smoking and family history of premature CHD) not included in the 2013 algorithm (1998 HR reduced from 1.85 to 1.78 (95% CI: 1.65, 1.93) when those risk factors omitted). In the 2013, but not the 1998 algorithm, all overweight adults with a large waist circumference or pre-diabetes as their only risk factor were included in the treated group, but the HR’s for those subgroups were not elevated. The 2013 HR changed from 1.59 to 1.77 (95% CI: 1.63, 1.93) when those criteria were omitted.
Conclusions: Both algorithms identified patients for weight loss treatment who were at elevated CVD risk, and the difference in the HR’s associated with the algorithms can be explained by differences in the specification of risk factors.
Author Disclosures: J. Stevens: None. E. Erber-Oakkar: A. Employment; Significant; Quintiles. Z. Cui: None. J. Cai: None. D. Wormser: A. Employment; Significant; F. Hoffmann-La Roche. H. Other; Significant; Stock in F. Hoffmann-La Roche. S.S. Virani: None.
- © 2015 by American Heart Association, Inc.