The present study determined the rates of concordance between the Diagnostic and Statistical Manual of Mental Disorders Version III-R (DSM-III-R) and the next version (DSM-IV) lifetime diagnoses for Substance Abuse and Dependence in a population (N=900) homogeneous for chronic and severe substance use disorder. The substance use disorder sections of the Structured Clinical Interview for DSM-III-R and DSM-IV were combined into a single interview and administered by trained clinical research interviewers. Analysis for each drug class was restricted to patients who reported prior use of the substance. Kappa values indicated excellent agreement between the 2 classification systems for Dependence diagnoses and fair to excellent concordance for Abuse diagnoses. For some drug classes, cannabis in particular, DSM-IV nosology resulted in lower rates of Dependence and higher rates of Abuse diagnoses. Overall, in this population with severe and multiple drug use, the effect of differences between DSM-III-R and DSM-IV was of modest magnitude with the exception of cannabis use disorders.