Proceedings of the 2013 annual meeting of the Netherlands Epidemiology Society
Volume 1 Issue S1 Abstract 41
E. Sohl, VU Medical Center, Amsterdam, the Netherlands
M.W. Heymans, VU Medical Center, Amsterdam, the Netherlands
R.T. de Jongh, VU Medical Center, Amsterdam, the Netherlands
P. Lips, VU Medical Center, Amsterdam, the Netherlands
N.M. van Schoor, VU Medical Center, Amsterdam, the Netherlands
Vitamin D deficiency is currently diagnosed by measuring serum 25-hydroxyvitamin D (25(OH)D).This study aimed to develop a risk profile that can be used to easily identify older individuals at risk for vitamin D deficiency.
This study was performed within the Longitudinal Aging Study Amsterdam, which is an ongoing cohort study of a representative sample of the Dutch older population. Vitamin D deficiency was defined as serum 25(OH)D < 50 nmol/L. A backwards logistic regression procedure was used to select predictors using Akaike’s Information Criterion (p < 0.157). A total risk score was calculated by, first, dividing the individual regression coefficients by the lowest regression coefficient in order to create simple scores and, second, by adding up these simple scores.
Vitamin D deficiency was present in 46.2%. The following predictors for vitamin D deficiency were identified: gender, age, bicycling, gardening, sporting, BMI, smoking, alcohol use, season, multivitamin use, the presence of appetite, and having a partner. The resulting AUC was 0.78. With a total risk score cut-off point of 46 (range 6-90), the model predicted vitamin D deficiency with a sensitivity of 73% and specificity of 70%. With a cut-off point of 61, the sensitivity was 35% and specificity 93%.
A total risk score, including twelve predictors that can easily be assessed, was developed and able to predict serum 25(OH)D below 50 nmol/L accurately. This risk score may be useful in clinical practice to identify persons at risk for vitamin D deficiency.
Published: 06 Jun, 2013