A cross-sectional multicentre study aimed at evaluating the prevalence of DSM III-R Axis I morbidity among drug addicts seeking treatment. Consecutive outpatients of the National Health Service's treatment units (aged 18-60 years, fulfilling criteria for drug dependence and without primary diagnosis of alcohol dependence) were enrolled at intake. In the first phase of the study, professional staff members evaluated current psychiatric morbidity of 317 clients clinically. In the second phase, trained interviewers administered the Composite International Diagnostic Interview and a European adaptation of the Addiction Severity Index to 65 probands. Some current psychiatric morbidity was found in 26.2% of subjects recruited during the first phase and in 22.2% during the second; lifetime comorbidity was ascertained in 32.3% of probands recruited in the second phase. Unfortunately, remarkable rates of undetermined diagnostic assessment during the second phase are due to refusers. The distribution of categorical diagnoses is somewhat different between the two samples, but the most frequent are Anxiety and Mood disorders in both. Primary versus secondary psychiatric diagnoses were also distinguished when evaluated over a lifetime. The rates of psychiatric comorbidity are low in this study as compared with previous investigations, but suffer from a poor compliance with structured diagnostic procedures by unselected probands. Other methodological factors may also affect this kind of research. Prospective studies are needed to evaluate the validity of comorbid diagnostic constructs and their prognostic significance. Also a concurrent assessment of Axis II disorders may be helpful to explain symptomatic polymorphism in case of multiple Axis I comorbidity.