Objectives - To present a comprehensive review of the studies assessing the frequency of PTSD in the general population, in clinical populations and among subjects at risk. Design - An extensive search was made using the Excerpta Medica Psychiatry CD-ROM 1980-1993 (October). Using post-traumatic stress disorder as a key word, a total of 1,057 papers published in this timespan were retrieved. A further search was made using the same key word with MEDLINE CD-ROM 1988-1993. A manual search was also performed for all issues of the Journal of Traumatic Stress. Results - A total of 135 studies meeting inclusion criteria were selected for this review. Almost two thirds (n = 86, 64%) of these studies were carried out in the United States; only 8 (6%) studies have been carried out in developed countries. The sample size ranges from a low of 11 subjects up to a high of 22,463 subjects. The mean sample size is 500 and the median is 108. In terms of assessment methods, in one third (n = 45, 33%) of the studies the investigators used a PTSD symptom checklist (either self- administered or administered by a clinician), based on DSM criteria, to generate a PTSD diagnosis. In more than in one third of the studies (n = 44, 33%) of the studies, a structured interview was administered (e.g., the DIS, the SCID, the SADS), while in the remaining surveys the diagnostic evaluation was based either on an un- structured clinical assessment or on the administration of other specific assessment instruments (e.g., M-PTSD, IES, SCL-90-R or others). In 77 studies (57%) the investigators based their assessment on DSM-III diagnostic criteria for PTSD, whereas in 55 studies (41 %) DSM-III-R criteria were employed. Prevalence rates are discussed separately for the different population groups. Conclusions - There has been a remarkable increase in the number and quality of epidemiological studies on PTSD over the past decade. A substantial amount of information is now available about the frequency of PTSD among different population groups and among at risk subjects, and about risk factors, common symptom patterns, and natural history of this disorder. However, a number of important limitations of these studies are to be noted and should be addressed by future research efforts.