Data from 27,616 current and former drinkers interviewed in the 1992 National Longitudinal Alcohol Epidemiologic Survey were used to examine the relationship between age at first use of alcohol and the prevalence of lifetime alcohol abuse and alcohol dependence, among all U.S. adults 18 years of age and over and within subgroups defined by sex and race. The rates of lifetime dependence declined from more than 40% among individuals who started drinking at ages 14 or younger to roughly 10% among those who started drinking at ages 20 and older. The rates of lifetime abuse declined from just over 11% among those who initiated use of alcohol at ages 16 or younger to approximately 4% among those whose onset of use was at ages 20 or older. After using multivariate logistic regression models to adjust for potential confounders, the odds of dependence decreased by 14% with each increasing year of age at onset of use, and the odds of abuse decreased by 8%. These findings are discussed with respect to their implications for prevention policies and the need to integrate epidemiological and intervention research.