Admixture analysis of age at onset in schizophrenia

In schizophrenia, clinical, familial and biological characteristics according to age at onset (AAO) suggest that AAO is a valid candidate symptom for genetic studies. However, none of the various thresholds used to define AAO subgroups in schizophrenia has been validated. We aim to define different AAO subtypes by admixture analysis in a sample of prospectively recruited subjects with schizophrenia. Consecutive inpatients and outpatients (N=141) meeting DSM IV criteria for schizophrenia were included. We used admixture analysis to investigate whether the observed AAO distribution consisted of a mixture of gaussian distributions and then compared clinical features and familial risks in the various groups of subjects. The model that best fitted the observed AAO distribution was a mixture of two gaussian distributions (mean+/-S.D.): (19.91+/-3.56 years) and (33.48+/-8.2 years), with a cutoff point at 28 years. The existence of two subgroups according to AAO was further confirmed by the different clinical and familial profiles of these subgroups. The early-onset group consisted predominantly of male patients, with non-paranoid subtypes and with a higher familial risk of schizophrenia spectrum disorders and affective disorders. The late-onset group of patients presented predominantly paranoid subtype, preponderance of females; they were more likely to be married and to have children. We identified two subgroups of schizophrenic subjects with different clinical and familial profiles. This study provides a mathematical validation of the existence of two subgroups defined by an onset of schizophrenia before or after 28 years. These results may have important implications for the search for schizophrenia susceptibility factors. Working with homogeneous subgroups defined on the basis of AAO may facilitate the identification of genetic vulnerability factors in schizophrenia.