Results of several studies suggest that either a reduction in the seru
m level of total cholesterol level or a persistently low cholesterol l
evel may be associated with an increase in violent deaths. Although th
ere are several possible explanations for these observations, it has b
een suggested that the cholesterol level could influence various behav
iors. We therefore examined the cross-sectional relation of several ps
ychologic characteristics, assessed by the Diagnostic Interview Schedu
le and the Minnesota Multiphasic Personality Inventory, to levels of t
otal cholesterol, high-density lipoprotein cholesterol, and triglyceri
des among 3,490 men aged 31-45 years who were examined in 1985-1986. (
All men had served in the US Army between 1965 and 1971). Compared wit
h that of other men, the mean total cholesterol level was 5 mg/dl high
er among 697 men diagnosed with generalized anxiety disorder (possibly
because of increased catecholamine levels) and 7 mg/dl lower among 32
5 men with antisocial personality disorder (p < 0.01 for each associat
ion). These differences could not be attributed to education, relative
weight, cigarette smoking, use of various medications, or other poten
tial confounders. In contrast, cholesterol levels were not significant
ly associated with major depression or hostility; levels of high-densi
ty lipoprotein cholesterol and triglycerides were not related to any d
iagnosis. If the serum level of total cholesterol is found to be predi
ctive of antisocial personality disorder in longitudinal analyses, thi
s association may have implications for cholesterol-lowering recommend
ations.