Hippocampal and amygdala volumes according to psychosis stage and diagnosis: a magnetic resonance imaging study of chronic schizophrenia, first-episode psychosis, and ultra-high-risk individuals.

Context: Magnetic resonance imaging studies have identified hippocampal volume reductions in schizophrenia and amygdala volume enlargements in bipolar disorder, suggesting different medial temporal lobe abnormalities in these conditions. These studies have been limited by small samples and the absence of patients early in the course of illness.

Objective: To investigate hippocampal and amygdala volumes in a large sample of patients with chronic schizophrenia, patients with first-episode psychosis, and patients at ultra-high risk for psychosis compared with control subjects.

Design: Cross-sectional comparison between patient groups and controls.

Setting: Individuals with chronic schizophrenia were recruited from a mental health rehabilitation service, and individuals with first-episode psychosis and ultra-high risk were recruited from the ORYGEN Youth Health Service. Control subjects were recruited from the community.

Participants: The study population of 473 individuals included 89 with chronic schizophrenia, 162 with first-episode psychosis, 135 at ultra-high risk for psychosis (of whom 39 subsequently developed a psychotic illness), and 87 controls.

Main Outcome Measures: Hippocampal, amygdala, whole-brain, and intracranial volumes were estimated on high-resolution magnetic resonance images and compared across groups, including first-episode subgroups. We used 1- and 2-way analysis of variance designs to compare hippocampal and amygdala volumes across groups, correcting for intracranial volume and covarying for age and sex. We investigated the effects of medication and illness duration on structural volumes.

Results: Patients with chronic schizophrenia displayed bilateral hippocampal volume reduction. Patients with first-episode schizophrenia but not schizophreniform psychosis displayed left hippocampal volume reduction. The remaining first-episode subgroups had normal hippocampal volumes compared with controls. Amygdala volume enlargement was identified only in first-episode patients with nonschizophrenic psychoses. Patients at ultra-high risk for psychosis had normal baseline hippocampal and amygdala volumes whether or not they subsequently developed a psychotic illness. Structural volumes did not differ between patients taking atypical vs typical antipsychotic medications, and they remained unchanged when patients treated with lithium were excluded from the analysis.

Conclusions: Medial temporal structural changes are not seen until after the onset of a psychotic illness, and the pattern of structural change differs according to the type of psychosis. These findings have important implications for future neurobiological studies of psychotic disorders and emphasize the importance of longitudinal studies examining patients before and after the onset of a psychotic illness.