The initial evidence linking vitamin D deficiency with increased risk of cardiovascular disease (CVD) extends back to 1970s when case control studies showed lower circulating concentrations of 25-hydroxyvitamin D [25(OH)D] in myocardial infarction cases compared with controls. These findings were supported by the identification of vitamin D receptors in both cardiac and smooth muscle tissue in 1980s. The results from the case control studies were confirmed by cohort studies published in the 2000s, which also extended the outcome from myocardial infarction to CVD more broadly. The latter studies provided stronger evidence because of measurement of 25(OH)D prior to the onset of CVD. However, concerns remained about possible residual confounding as the reason for the inverse association between circulating 25(OH)D levels and CVD risk. This prompted the initiation of large scale randomised controlled trials (RCTs) of vitamin D supplementation with CVD as a pre-specified outcome. Results from these studies, which include (in order of publication date) ViDA in New Zealand (n=5108), VITAL in the United States (n=25,871) and D-Health in Australia (n=21,315), have been published in the last few years. Their findings will be presented, along with those from recent meta-analyses that combine them with findings from earlier RCTs. Because RCTs are known to have limitations, such as relatively short follow-up of long-term chronic diseases, poor adherence, likely contamination in the control group from self-supplementation, and recruitment of participants who are not vitamin D deficient (the so-called ‘worried-well’), the RCT results will also be compared for consistency with those from mendelian randomisation studies of genetically predicted 25(OH)D which are not prone to the same biases. This evidence will be reviewed and conclusions made about the causal inferences regarding vitamin D and prevention of CVD.