The purpose of the study was to determine whether the total number of days spent in hospital during a 5-year follow-up period could be predicted by determining IgA, IgG and IgM concentrations in 62 schizophrenics on initial admission. A high IgA concentration on initial admission indicated less frequent need for hospital care during follow-up. A scattering diagram of this negative correlation showed that the correlation was based on a group of seven patients whose IgA concentrations markedly exceeded normal values. This group with exceedingly high IgA differed from the control group in several areas, which would indicate that these patients suffer from a "different" schizophrenia than the patients in the control group. This finding supports the idea of the heterogeneity of schizophrenia, and raises the possibility of developing biochemical methods to delineate the diagnostic category of schizophrenia.