To the Editor: We read with great interest the article by Carole Di Maggio, M.D., and her colleagues R1592CIHIEJFF. Although the study presented some interesting data about decreasing age at onset of schizophrenia, it also raised some questions about the interpretation and analyses of the data. The authors presented retrospective data for medical records (N=877) from 1976, 1986, and 1996. Individuals in the study were divided into three birth cohorts: 1905–1944, 1945–1964, and 1965–1984. However, the significant decrease in age at onset through time could be due to bias caused by the different age structures of the cohorts. Individual ages of the subjects in the 1905–1944 cohort were from 32 to 91 years, the subjects in the 1945–1964 cohort were aged 12 to 51 years, and the subjects in the 1965–1984 cohort were aged 1 to 31 years. Thus, all persons who died before the age of 32 could not be part of the 1905–1944 cohort. This might have been an important source of bias in the study because mortality is greater for patients with schizophrenia, but at the population level, the standardized mortality ratio for schizophrenia decreases exponentially with age R1592CIHCDAEH. Only patients with illness onset before age 32 could be present in the youngest cohort.

We calculated mean ages at onset on the basis of our data R1592CIHEGJFB for cohorts from 1950–1955 (N=4,023) and 1956–1961 (N=2,941), with follow-up assessments made, respectively, at 1969–1978 and 1975–1984. The mean ages at onset were 21.8 and 21.9 years. We suggest that the data presented by Dr. Di Maggio and colleagues should be reanalyzed with consideration of the different age structures.