This meta-analysis will be conducted to study the effect on vitamin D supplements on markers for cardiovascular disease and diabetes using individual participant data from 12 RCTs. Given that previous data suggest that vitamin D supplementation might be most effective in those individuals with very low 25-hydroxyvitamin D (25(OH)D) and potentially harmful in those achieving very high levels after vitamin D supplementation, this meta-analyses will be performed in subgroups according 25(OH)D levels.

Large randomized controlled trials (RCTs) are currently on-going to evaluate non-skeletal effects of vitamin D supplementation in the general older population. The results of these studies can be expected in 2017 to 2020 and while they will report important data, they may still leave knowledge gaps on the effects of vitamin D on clinically relevant surrogate parameters in specific groups. Recent data indicate that the association of vitamin D status and outcome is U- or reverse J-shaped and may be modified by the presence of certain risk factors or genetic variations of the vitamin D receptor.

This work is part of the European Union-project 'Food-based solutions for eradication of vitamin D deficiency and health promotion throughout the life cycle' (ODIN). This individual patient meta-analysis of existing high quality vitamin D RCTs aims to evaluate whether there are beneficial or harmful vitamin D effects on surrogate parameters for clinical outcomes (i.e. blood pressure, lipids, parathyroid hormone (PTH), fasting blood glucose and glycated haemoglobin (HbA1c)). With data from almost 3000 randomized subjects from 12 RCTs that are available within this consortium, the investigators will have sufficient power to detect clinically relevant effects of vitamin D supplementation on risk factors for cardiovascular disease.

Subgroup analyses will investigate effect modifiers to evaluate whether certain groups of individuals experience pronounced or attenuated effects from vitamin D supplementation. Given that previous data suggest that vitamin D supplementation might be most effective in those individuals with very low 25(OH)D and potentially harmful in those achieving very high levels after vitamin D supplementation, this individual patient data meta-analyses will be performed in subgroups with serum 25(OH)D levels <30, 40, and 50 nmol/L and >100, 125 and 150 nmol/L. 25(OH)D levels will be re-measured to ensure comparability.

Inclusion Criteria:

• Availability of quality bio-banked samples for uniform 25(OH)D re-measurement
• Prospective data on one or more of the outcomes
• Willingness to collaborate

Exclusion Criteria:

• Trials among pregnant women, children, or performed in patient populations.

The participating RCTs are the Tromso IGT, Tromso OBESITY, Tromso BMD, Tromso CLAMP, Tromso DEPRESSION, Styrian Vitamin D Hypertension Trial, Paravit, Oosterwerff, Chel, Wicherts, UCC1, and UCC2 trials.