The extent to which dependence occurs with or without abuse is important because of the potential for underestimation and biased estimates of drug dependence in surveys that rely on abuse as a screening method for dependence. The purpose of this paper was to present the prevalence of DSM-IV drug dependence with and without drug abuse in a nationally representative sample, as well as in subgroups defined by sex, age and race/ethnicity. Among all respondents with current drug dependence, 22.0% did not additionally meet criteria for abuse (19.5% among males and 27.8% among females). Current drug dependence without abuse was especially common among females age 45-64 (52.6% of all cases). Among those with lifetime diagnoses of drug dependence, a small proportion overall, 5.0% had no symptoms of abuse, with the highest proportion again found among females aged 45-64 (19.5% of all cases). The use of drug abuse as a screening method for drug dependence in large epidemiologic studies will differentially underestimate the prevalence of dependence by subgroup, affecting many types of research. Dependence with and without abuse may represent heterogeneous phenotypes for genetic and gene-environment research, which should be explored.