Calf circumference is more effective than body mass index in predicting emerging care-need of older adults – Results of a national cohort study

• Published in: Clinical nutrition (Edinburgh, Scotland) Vol. 35; no. 3; pp. 735 - 740
• Main Authors: Hsu, Wei-Chung, Tsai, Alan C, Wang, Jiun-Yi
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01.06.2016
• Subjects: Aged; Aged, 80 and over; Aging; Body Mass Index; Body Size; Calf circumference; Care-need; Cohort Studies; Early Diagnosis; Female; Frail Elderly; Frailty - diagnosis; Frailty - epidemiology; Frailty - therapy; Gastroenterology and Hepatology; Geriatric Assessment; Health Services for the Aged; Health Surveys; Humans; Leg; Longitudinal Studies; Lost to Follow-Up; Male; Needs Assessment; Older adults; Predictive Value of Tests; Risk; Sex Factors; Taiwan - epidemiology; Body mass index; Care-need; Calf circumference; Older adults
• ISSN: 0261-5614; 1532-1983
• DOI: 10.1016/j.clnu.2015.05.017

Summary Background & aims The usefulness of body mass index (BMI) and calf circumference (CC) for predicting emerging care-need of older adults has not been carefully evaluated. We attempted to compare the abilities of these two anthropometrics in predicting emerging care-need in older adults. Methods We analyzed the 1999 (baseline) and 2003 (end-point) datasets of the Taiwan Longitudinal Survey on Aging. Participants were 2521 ≥ 65-year old adults without care at baseline. To derive proper cut-offs of BMI and CC, we first drew a plot to show the relationship between the risk of care-need and the cumulative distribution of BMI or CC. We then divided the risk into three levels and calculated the corresponding percentiles of BMI and CC. Multivariable logistic regression was used to build up predictive models. Akaike Information Criterion (AIC) and net reclassification improvement (NRI) were performed to compare the models as well as the predictive abilities of BMI and CC. Results After controlled other independent variables, CC was significantly associated with emerging care-need in the subsequent 4 years, but BMI was not. The AIC value of the model with CC (1956.3) was remarkably lower than BMI (1968.7). The NRI was 4.8% (p = 0.007) which quantified the improvement of the model with inclusion of CC instead of BMI to predict emerging care-need. Conclusions All indices we performed suggest that CC has better ability to predict emerging care-need in older Taiwanese compared to BMI. The potential application of anthropometric indicators, especially CC, for predicting emerging care-need deserves further investigation.