RESUMEN
OBJECTIVE: To establish if the correction with estimates of ultraviolet (UV) exposure influences the association between 25-OH-vitamin D (25OHD) levels and metabolic variables. STUDY DESIGN: A cross-sectional study was performed in 575 obese children and adolescents (>6 years of age) in a tertiary referral center. Cardiovascular risk factors were measured. The estimate of UV exposure was evaluated by 3 methods: (1) season; (2) mean of UV radiation (UVR); and (3) mean of UV index (UVI). UVR and UVI were considered at 1 (UVR 1 month prior to testing [UVR1], UVI 1 month prior to testing [UVI1]) or 3 (UVR 3 months prior to testing [UVR3], UVI 3 months prior to testing [UVI3]) months prior to testing. All analyses were corrected for confounders (sex, age, puberty, body mass index, waist circumference, the inclusion and exclusion of estimates of UV exposure). RESULTS: The 25OHD levels were associated with seasons, UVR1, UVR3, UVI1, and UVI3, and best associations with UVR3 and UVI3. In all models, total cholesterol, low-density lipoprotein cholesterol and triglycerides were negatively associated with 25OHD levels. The strength of the association increased with no correction, correction for seasons, UVR, and UVI. UVR3 and UVI3 performed better than UVR1 and UVI3. CONCLUSIONS: Higher lipid concentrations were associated with low 25OHD levels in obese children and adolescents with the power of the association dependent on the estimates of UVR. As the mean values 3 months prior to testing for both UVR and UVI determined the best associations, the interval of the steady state time of 25OHD levels could be preferentially used in the metabolic studies. Controlling for an estimate of UVR is important to decrease the heterogeneity of studies.