Discordance between Body-Mass Index and Body Adiposity Index in the Classification of Weight Status of Elderly Patients with Stable Coronary Artery Disease

• Published in: Journal of clinical medicine Vol. 10; no. 5; p. 943
• Main Authors: Hudzik, Bartosz, Nowak, Justyna, Szkodzinski, Janusz, Danikiewicz, Aleksander, Korzonek-Szlacheta, Ilona, Zubelewicz-Szkodzińska, Barbara
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 01.03.2021; MDPI
• Subjects: Body fat; Body mass index; Cardiovascular disease; Clinical medicine; Obesity; Older people; Womens health; coronary artery disease; discordance; weight status classification; elderly; body-mass index; body adiposity index
• ISSN: 2077-0383; 2077-0383
• DOI: 10.3390/jcm10050943

Body-mass index (BMI) is a popular method implemented to define weight status. However, describing obesity by BMI may result in inaccurate assessment of adiposity. The Body Adiposity Index (BAI) is intended to be a directly validated method of estimating body fat percentage. We set out to compare body weight status assessment by BMI and BAI in a cohort of elderly patients with stable coronary artery disease (CAD). A total of 169 patients with stable CAD were enrolled in an out-patient cardiology clinic. The National Research Council (US) Committee on Diet and Health classification was used for individuals older than 65 years as underweight BMI < 24 kg/m , normal weight BMI 24-29 kg/m , overweight BMI 29-35 kg/m , and obesity BMI > 35 kg/m . In case of BAI, we used sex- and age-specific classification of weight status. In addition, body fat was estimated by bioelectrical impedance analysis (BImpA). Only 72 out of 169 patients (42.6%) had concordant classification of weight status by both BMI and BAI. The majority of the patients had their weight status either underestimated or overestimated. There were strong positive correlations between BMI and BImpA (FAT%) (R = 0.78 < 0.001); BAI and BImpA (FAT%) (R = 0.79 < 0.001); and BMI and BAI (R = 0.67 < 0.001). BMI tended to overestimate the rate of underweight, normal weight or overweight, meanwhile underestimating the rate of obesity. Third, BMI exhibited an average positive bias of 14.4% compared to the reference method (BImpA), whereas BAI exhibited an average negative bias of -8.3% compared to the reference method (BImpA). Multivariate logistic regression identified independent predictors of discordance in assessing weight status by BMI and BAI: BImpA (FAT%) odds ratio (OR) 1.29, total body water (%) OR 1.61, fat mass index OR 2.62, and Controlling Nutritional Status (CONUT) score OR 1.25. There is substantial rate of misclassification of weight status between BMI and BAI. These findings have significant implications for clinical practice as the boundary between health and disease in malnutrition is crucial to accurately define criteria for intervention. Perhaps BMI cut-offs for classifying weight status in the elderly should be revisited.