Since the diagnosis of delayed-onset posttraumatic stress disorder (PTSD) was introduced in DSM-III, there has been controversy over its prevalence and even its existence. The authors sought to resolve discrepant findings concerning the prevalence of delayed-onset PTSD by conducting a systematic review of the evidence. A literature search was conducted for case reports and group studies with adequate measurement of delayed-onset PTSD according to DSM criteria. Studies that met inclusion criteria were examined for the defined length of delay for delayed-onset PTSD, presence of symptoms before full diagnostic criteria were met, length of follow-up, prevalence estimates, and other variables. Studies were also examined for differences between immediate-onset PTSD, delayed-onset PTSD, and no-PTSD cases. Ten case studies and 19 group studies met criteria for inclusion in the review. Studies consistently showed that delayed-onset PTSD in the absence of any prior symptoms was rare, whereas delayed onsets that represented exacerbations or reactivations of prior symptoms accounted on average for 38.2% and 15.3%, respectively, of military and civilian cases of PTSD. The discrepant findings in the literature concerning prevalence can be largely, but not completely, explained as being due to definitional issues. Little is known about what distinguishes the delayed-onset and immediate-onset forms of the disorder. Continuing scientific study of delayed-onset PTSD would benefit if future editions of DSM were to adopt a definition that explicitly accepts the likelihood of at least some prior symptoms.