[Water hardness and cardiovascular mortality. Discussion of evidence from geographical pathology (author's transl)]

The existence of a significant negative correlation between hardness of drinking water and mortality rates from cardiovascular disease has been established by numerous authors. The intensity of this relationship varies according to the country concerned, the size of the areas studied, and the confounding factors taken into consideration. Nevertheless, the problem concerning the interpretation of the correlation has not yet been resolved. In particular the methodological problem resulting from the contiguity of the geographic units has been generally neglected. It is shown that these difficulties weaken the conviction that the relationship is a causal one.