To the Editor.—
The association between light drinking and low birth weight reported by Mills et al1 in "Maternal Alcohol Consumption and Birth Weight" must be cautiously considered. In a cohort of 31,604 pregnancies, the 3% who drank the most included women reporting consumption of one to two drinks per day (2.4%), those reporting three to four drinks per day (0.4%), and those reporting six or more drinks per day (0.1%). Decreased birth weights were observed in offspring of these 3% of the mothers. In light of many other surveys of drinking practices among pregnant women, which have identified 2% to 14% of pregnant women as heavy drinkers,2 we suggest that the top 3% in this study may have been underreporting their alcohol use. Thus, while there is clearly a higher incidence of low birth weight among offspring of the women in this top 3%, an association between light drinking and reduced birth weight is not demonstrated.