A hypothesis on the etiological role of diet on age of menarche.

Body size and body composition have been suggested as the best explanation for the temporal trend toward early menarche over the last 100 years. There is evidence from human and animal studies that indicates that body size is not the primary factor in influencing the occurrence of menarche. The data actually show that diet may be a primary environmental control mechanism of menarche especially since it alters hormone levels. We see diet as an etiological factor in both the long term and immediate control of menarche. In the long term it influences body size and development leading to menarche. In the short term it acts at a critical state to precipitate the onset of menarche and related physiological changes. This hypothesis does not exclude other less important factors associated with menarche. Our data shows that the present trend toward early menarche can be reversed when a balanced vegetarian diet is selected in place of the ordinary American diet.

Patterns of alcohol and cigarette use in pregnancy.

Maternal smoking and drinking have been implicated as factors in the adverse outcome of pregnancies. This California study of over 12,000 women demonstrates that a large proportion of them smoke or drink during pregnancy. Drinking habits appear to be associated with age, ethnicity, martial status, income, education, church attendance, and prenatal care, but not with gestational age, breast-feeding, or self-reported health. Smoking shows associations with income and education. Drinking and smoking are closely interrelated, showing strikingly parallel associations with such factors as use of caffeine or illicit drugs. Subgroups of pregnant women who exhibit difference drinking habits were found not comparable on certain other characteristics which potentially influence the health of neonates. The importance is reemphasized of utilizing large populations, adjusting for confounding variables, and paying particular attention to smoking habits. Forthcoming multivariate analyses of the data set should disentangle effects of the confounding variables.