Predicting ambient ultraviolet from routine meteorological data; its potential use as an instrumental variable for vitamin D status in pregnancy in a longitudinal birth cohort in the UK

Methods Using linear regression we described the relationship between estimated ambient-erythemal ultraviolet (eUV) exposure in Oxford (1990-95) and total hours of sunshine and month in order to forecast eUV in nearby regions, whilst adjusting for regional variations in weather. The forecast was validated with empirical data collected from Cornwall and then predicted for the Avon region. Total 98-day prenatal ambient-eUV was then predicted in 355 expectant mothers in the Avon Longitudinal Study of Parents and Children (ALSPAC) cohort and its relationship with maternal vitamin D status was determined. Results Estimated ambient-eUV was strongly associated with measured ambient-eUV (r(2) = 0.989) with a near 1:1 prediction for the validation data set [beta = 0.99, 95% confidence interval (CI) 0.913, 1.067 r(2) = 0.980]; strong seasonal associations were observed between eUV in the last trimester of pregnancy and maternal serum 25-(OH)D concentrations (r(2) = 0.40). Conclusion This technique of prediction could be applied to existing cohorts allowing the relationship between maternal vitamin D status and the health of the offspring to be studied via instrumental variable analysis.