The findings in a recent article in the Archives regarding the relation of low cholesterol to violence in the Honolulu Heart Program1 should be viewed in context. Iribarren et al1 conclude from 7309 subjects and only 75 traumatic deaths and 24 suicides that the hypothesis of an inverse relation between serum cholesterol and suicide is contradicted. To put this in perspective, the two largest prior cohort studies and one pooled analysis of smaller studies, all vastly larger than the present study (incorporating around 350 000,2 27 000,3 and 160 000 men4 and hundreds of violent deaths), have each found a significant link between low cholesterol and violent death in men. Studies that have failed to show this association, including the study of Iribarren and colleagues, have involved much smaller samples.
In addition, it should be noted that "controlling" for alcohol consumption in evaluating the relation