Births of schizophrenics tend to occur more often than expected during the first months of the year. This phenomenon has been repeatedly demonstrated in large samples of patients, but data from France are not yet available. METHOD. Medical records of 2,215 inpatients born in France, have been studied with DSM III-R. Admissions were unduplicated. Inpatient population is described table I. The 230 schizophrenic patients were also classified into three groups according with family history for psychotic diseases (Positive, Negative and Unknown). The year was divided in quarters, but also in two climatic half-years (Cold and Warm). Two different populations were used as controls: first, the live births data in France collected by the INSEE; and second, the 1726 non psychotic inpatients hospitalised in the same unit as the psychotics, during the same period. RESULTS. The season of birth effect can be demonstrated from our sample. When the whole schizophrenic group is compared to the INSEE data, a significant excess of births is found for the winter quarter (p less than .04) and for the cold half-year (p less than .01). In the DSM III-R disorganized patients the seasonal variation becomes significant at a level of p less than .001. The other subtypes of schizophrenia do not show significant seasonal variations (table II). The pattern of birth of the other DSM III-R diagnoses, does not significantly differ from INSEE control data (table III). Schizophrenic patients without psychiatrically ill relatives show a slight, non significant excess of births in winter which become significant for the cold half-year (p less than .03). Subjects with possible family history do not show any trend suggesting a seasonal effect (table IV). DISCUSSION. Few reports include an entire inpatient population as we have done. Such data in our study offers an element of useful comparison with the schizophrenics. Clinic data: Most of seasonal studies have used ICD-8, ICD-9 or Tsuang and Winokur criteria, to compare only paranoid and non paranoid subjects, and some of them found a very small excess of paranoid births in winter. The discrepancies between these studies and ours, can reflect both variability among environmental factors and variability in the diagnostic criteria.
Control data: Most of the subjects were born between 1920 and 1960. All live births in France for the same years could be used as controls, but unfortunately INSEE data were not available by months or quarters before 1949. But since in western and southern European general population, the season of birth appears to be relatively uniform and stable over time, it seemed possible to use the posterior INSEE data to compare with our anterior inpatients data. Familial data: We found a birth excess in the cold season, only for schizophrenics without any family history. This finding is consistent with various other reports. CONCLUSION. In conclusion, our findings are consistent with the suggestion that the season of birth is important in the aetiology of the disorganized subgroup of schizophrenic aetiology of the disorganized subgroups of schizophrenic patients. Moreover, the excess of winter births is only observed in the "negative family history" group.