Validation of Body Mass Index-for-age percentile curves in older adults using data from the Canadian Longitudinal Study on Aging

Abstract

Background: The use of a standard body mass index (BMI) cut-point (e.g., BMI ≥ 30 kg/m2) to define obesity across the lifespan does not reflect our current understanding of changes in adiposity that occur throughout the aging process. The objective of the present study was to examine the validity of BMI-for-age percentiles as a measure of obesity in older adults, compared to an objective gold-standard measure of obesity status, total body fat percent (BF%) measured by Dual Energy X-ray Absorptiometry scan.

Methods: Using data from the Canadian Longitudinal Study on Aging (n = 28,764), sensitivity (Se), specificity (Sp), positive predictive value (PPV), and negative predictive value (NPV) were calculated for BMI ≥ 85th percentile and other thresholds against BF% ≥ 35%, 38%, and 40%, respectively.

Results: When comparing obesity defined by BMI percentile values (≥ 85th percentile) to BF% (≥ 40% BF) as a measure of true obesity status, Se, Sp, PPV, and NPV values were 0.32 (0.31, 0.33), 0.98 (0.98, 0.98), 0.94 (0.93, 0.95) and 0.61 (0.60, 0.62) in women, and 0.85 (0.81, 0.89), 0.86 (0.85, 0.86), 0.14 (0.13, 0.16), and 1.00 (0.99, 1.00) in men, respectively.

Conclusion: Incorporating age and sex information, BMI-for-age percentile is a highly specific and moderately sensitive measure of obesity in older adults, with a potential as a valuable screening tool. Further research on BMI-for-age percentiles could help develop more accurate obesity screening tools than the current standard BMI cut-points.

Keywords: Aging; BMI; CLSA; Geriatrics; Obesity; Validation study.