Utility of a Body Shape Index Parameter in Predicting Cardiovascular Disease Risks

Background: Anthropometric indices are used as predictors of cardiovascular disease (CVD). The most used indices are body mass index (BMI) and waist circumference (WC); however, there are limitations regarding their validity to address different body shapes, fat and lean mass distribution. A body shape index (ABSI) has been proposed as an alternative parameter to reflect differences in body shape and potentially be more useful for predicting CVD. ABSI is calculated by ABSI = WC / (BMI(2/3 )• Height(1/2)). The purpose of this cross-sectional study was to determine the utility of ABSI as a predictor or modifiable risk factor of CVD compared to other commonly used measures in clinical practice. Methods: The sample population was from the baseline interview and health examination included in the National Health and Nutrition Examination Survey (NHANES) 2013-2014. Patients (n=5,924, 52% female) were aged 18-80 years (47.4 ± 18.4 years) who completed a series of questionnaires on a spectrum of health-related risks. After the interview, health technicians performed a standardized examination of the participants to collect data on weight, height, BMI, WC, and sagittal abdominal diameter (SAD). Statistical analysis was done using R Studio, version 0.99.903 (RStudio, Inc. Boston, MA). Using logistic regression, the correlation between each predictor (ABSI, BMI, WC, SAD) as a continuous variable, and CVD outcomes was evaluated with two models: a univariable model and a multivariable model. In a secondary analysis, ABSI was reclassified into categorical values based on quartiles of the NHANES dataset. Logistic regressions were again run for overall CVD and all CVD sub-categories, followed by chi-square tests for significance. For comparison, BMI categories of normal, overweight, obese, and severely obese were tested with overall CVD and all CVD subcategories as outcome measures, followed by chi-square tests for significance. Results: Approximately 10% of the sample population had at least one prior manifestation of CVD, the most common being myocardial infarction (MI) (4.0%). ABSI showed little correlation with weight, BMI, WC, and SAD (r<0.3), while BMI had a strong correlation with weight, BMI, WC, and SAD (r ≈ 0.9). In univariable logistic regression, ABSI showed the most robust associations of all predictors with overall CVD and all CVD subcategories. ABSI demonstrated stronger correlations than BMI for all CVD outcomes (except CHF in the multivariable model). This study attempted to create classifications of ABSI and compare them to the normative classifications of BMI. In this categorical analysis, ABSI was also stronger than BMI in all logistic regression analyses for CVD outcomes, except for CHF in the multivariable model. Severe obesity (BMI ≥40 kg/m(2)) almost doubled the odds of having CVD, while being categorized in Q2, Q3, and Q4 for ABSI increased odds by double, triple, and eight-fold, respectively. Conclusion: An ABSI parameter in the upper three quartiles increases the risk of CVD manifestations more significantly than an elevated BMI per category of overweight, obese, and severely obese, respectively. Since the categories for ABSI were created based on quartiles of a large sample size reflecting the US population, this suggests that the increased risk from an elevated ABSI is more widespread than previously understood. Thus, ABSI should be monitored more closely and managed in preventative medical care than BMI alone.