Correlates of high density lipoprotein cholesterol in women studied by the method of co-twin control

This study examines the relation between each of the following risk variables and high density lipoprotein (HDL) cholesterol: cigarette smoking, alcohol use, leisure exercise, postmenopausal estrogen use, and body mass, before and after adjustment for genetic and nongenetic variables shared by monozygous and dizygous co-twins. Subjects were 179 dizygous and 255 monozygous twin pairs from the Kaiser Permanente Twin Registry who participated in a special examination in 1978-1979 in Oakland, California. A multivariate co-twin analysis made it possible to adjust for measured covariables and also to adjust for unmeasured familial variables shared by co-twins. After adjustment for measured covariables, further adjustment for the unmeasured familial variables shared by monozygous co-twins reduced a positive association between alcohol and HDL cholesterol and eliminated a positive association between exercise and HDL cholesterol. On the other hand, adjustment for unmeasured familial variables shared by monozygous co-twins had little effect on significant associations between smoking and HDL cholesterol, postmenopausal estrogen use and HDL cholesterol, and body mass and HDL cholesterol. Although an important role for random error was not ruled out in the sample size available, findings are consistent with the following interpretations: 1) alcohol consumption and leisure exercise may have smaller effects on HDL cholesterol than predicted by studies unadjusted for familial factors because effects of these variables are confounded by familial factors; and 2) smoking, postmenopausal estrogen, and body mass have effects predicted by studies unadjusted for familial factors and may therefore be good targets for interventions to raise HDL cholesterol.