To elucidate the association between the duration of untreated psychosis and long-term outcome in schizophrenic patients, we followed up a cohort of first-episode schizophrenic patients in Nagasaki. The present study was conducted in the framework of the World Health Organization Collaborative Study on Determinants of Severe Mental Disorders, which our research group participated in as The Nagasaki World Health Organization Collaborating Center for Research and Training in Mental Health. The cohort was established during the period from 1979 to 1980 and consisted of 107 patients initially diagnosed as schizophrenia according to ICD-9. The subjects of the present study were 97 members of the cohort (54 males and 43 females) in whom we could measure the duration of untreated psychosis. The 97 study subjects were followed up for 15 years since they were enrolled in the cohort and were checked their mental conditions at 1-, 2-, 5-, 10- and 15-year follow-up points. The triplet of the 1st, 2nd and 3rd quartiles of the duration of untreated psychosis in them was (1, 4, 12) months (the mean was 9.9 months). During the whole period of 15-year follow-up, 5 patients died, 40 were lost for follow-up and 52 remained contacted at 15-year follow-up point; out of these 52 patients, 22 were contacted through the whole period. During the first 10 years of the follow-up, the patients diagnosed to have completely been remitted at follow-up point showed a significantly or marginally shorter duration of untreated psychosis as compared to those diagnosed not to have completely remitted; the triplet of the 1st, 2nd and 3rd quartiles of the duration of untreated psychosis in those diagnosed to have completely remitted at 1-year follow-up point was (1, 3, 4) months, while that in those diagnosed not to have completely remitted at the same follow-up point was (2, 6, 12) months and the difference was significant (p=0.036, Wilcoxon rank-sum test). Similarly, the triplets of the 1st, 2nd and 3rd quartiles of the duration of untreated psychosis in the two groups diagnosed with and without complete remission at 2-, 5- and 10-year follow-up points were (1, 3, 4) and (2, 6, 17) months (p=0.021), (1, 3, 8) and (1, 6, 17) months (p=0.149), and (1, 2, 3) and (3, 6, 12) months (p=0.008), respectively. However, no difference was observed between the two groups at 15-year follow-up point; (1, 4, 12) and (1, 4, 9) months (p=0.828). The results of the present study indicate that the duration of untreated psychosis will probably influence on the outcome of schizophrenia at least 10 years.