A predictive model of vitamin D insufficiency in older community people: From the Sao Paulo Aging & Health Study (SPAH)

Abstract

Objective: To evaluate the prevalence of 25-hydroxyvitamin D insufficiency (25OHD < 20 ng/mL) and to develop a predictive model for this status. Methods: This is a cross-sectional study including 908 community-dwelling older subjects, 18% (158) of which were randomly selected to be a ""test"" sample, with the remaining (750) composing a ""development"" sample. A radioimmunoassay technique was used to measure 25OHD levels. Anthropometrical data, information about lifestyle habits and co-morbidities were obtained. Multiple logistic regression models were created. An Index Risk of Vitamin D Insufficiency (IRVDI) was designed and subsequently validated. The performance of this tool was assessed through ROC analysis. Results: The prevalence of 25OHD <20 ng/mL was of 58.0% (CI 95% 51.6-64.6). The clinical independent factors for 25OHD <20 ng/mL were female gender (OR=2.16; 95%CI 1.13-4.13; p=0.020), diabetes (OR=1.84; 95%CI 1.23-2.74; p=0.003) and season (winter/spring) (OR=3.63, 95%CI 2.62-4.88; p < 0.001). After statistical adjustments, the IRVDI was able to identify older people at risk for vitamin D insufficiency with a sensitivity of 55.9%, specificity 72.3% and ROC area of 0.685 (p < 0.001). Conclusions: Our results suggest that vitamin D insufficiency is common among Brazilian community-dwelling elderly. Female gender, diabetes and the season (winter/spring) were the important parameters that predicted this status. The clinical use of these parameters can be help to design and target appropriate public health interventions. The IRVDI is a convenient tool for the selection of older people at risk for vitamin D insufficiency.