Studies in many countries have demonstrated a relationship between drinking water mineral content and the risk of death in cardiovascular disease (CVD). Particularly strong relationships have been found for magnesium and it has been suggested that magnesium be added to drinking water. The aim of this article is to evaluate the validity of this suggestion by reviewing information on possible causative agents. Major epidemiological studies on the drinking water content of calcium, magnesium, and hardness were analysed regarding exposure specificity, confounding factors, dose-response relationships and biological plausibility. Intervention experiments were analysed. The risk of death in CVD was related to the content of Ca, Mg and HCO3−. The data demonstrate that Ca and Mg need to be considered together, and that HCO3− could play a role by intervening with the body acid load. There is no evidence to justify the addition of magnesium only to drinking water for preventive purposes. The data suggest that Ca and Mg could be administered together but no data are available regarding the relative proportions for an optimal effect.