Endocrine Abstracts

Introduction: The prevalence of obesity in elderly population increased. Obesity leads to increased morbidity and mortality as well as worsens quality of life in older individuals. Body-mass index (BMI) is the most widely used adiposity index. BMI correlates with body fatness but this index does not account for fat mass versus fat-free mass. Recently a new index of body adiposity the body adiposity index (BAI) was introduced. This novel index correlates well (in comparison to BMI) with body fat percentage.

Aim: We investigated the discordance between BMI and BAI in measuring adiposity status in patients hospitalized in the geriatric department.

Material and methods: The study group comprised 391 patients above 60 years, hospitalized in the geriatric department (age 76±7 years, women 66%). In each patients we calculated BMI and BAI. Based on the BMI and WHO classification the patients were divided into three categories: normal weight, overweight and obesity. The value of BAI was classified too as normal weight, overweight and obesity. We calculated rates of misclassification to adiposity categories according BMI and BAI.

Results: Median BMI was 27.7 (24.8–31.3) kg/m² and median BAI was 33.9 (29.6–39.3) % We observed normal weight among 26.3% patients by BMI and 25.3% using BAI. Overweight was diagnosed among 40.7% using BMI and only 27.9% by BAI. However obesity was identified among 32.3% by BMI and as 46.8% using BAI index. There was a strong positive correlation between BMI and BAI index (Spearman R=0.70; P<0.0001). We observed a 54.7% rate of concordant assessment of adiposity status between BAI and BMI. In 28.6% of patients BMI underclassified patients and in 16.7% overclassified patients in comparicon with BAI. We observed a high rate of misclassification of adiposity status according to BMI and BAI index, especially in the normal weight and overweight subgroups. BMI tended to overclassify patients as normal weight and underclassify patients as overweight and obese compared to BAI. Multivariate logistic regression identified independent predictors of underestimation od adiposity status by BMI: age (per 1 year increment) OR 0.96 95%CI: 0.94–0.99 P=0.05; male sex (female reference) OR 0.32 95%CI: 0.20–0.50 P<0.0001.

Conclusion: The accuracy of BMI in predicting adiposity status in elderly patients is insufficient as compared to BAI classification. BMI tended to overestimate the rate of normal weight and underestimate the rate of overweight and obesity.