Longitudinal study on hippocampal subfields and glucose metabolism in early psychosis

Abstract

Altered hippocampal morphology and metabolic pathology, but also hippocampal circuit dysfunction, are established phenomena seen in psychotic disorders. Thus, we tested whether hippocampal subfield volume deficits link with deviations in glucose metabolism commonly seen in early psychosis, and whether the glucose parameters or subfield volumes change during follow-up period using one-year longitudinal study design of 78 first-episode psychosis patients (FEP), 48 clinical high-risk patients (CHR) and 83 controls (CTR). We also tested whether hippocampal morphology and glucose metabolism relate to clinical outcome. Hippocampus subfields were segmented with Freesurfer from 3T MRI images and parameters of glucose metabolism were determined in fasting plasma samples. Hippocampal subfield volumes were consistently lower in FEPs, and findings were more robust in non-affective psychoses, with strongest decreases in CA1, molecular layer and hippocampal tail, and in hippocampal tail of CHRs, compared to CTRs. These morphometric differences remained stable at one-year follow-up. Both non-diabetic CHRs and FEPs had worse glucose parameters compared to CTRs at baseline. We found that, insulin levels and insulin resistance increased during the follow-up period only in CHR, effect being largest in the CHRs converting to psychosis, independent of exposure to antipsychotics. The worsening of insulin resistance was associated with deterioration of function and symptoms in CHR. The smaller volume of hippocampal tail was associated with higher plasma insulin and insulin resistance in FEPs, at the one-year follow-up. Our longitudinal study supports the view that temporospatial hippocampal subfield volume deficits are stable near the onset of first psychosis, being more robust in non-affective psychoses, but less prominent in the CHR group. Specific subfield defects were related to worsening glucose metabolism during the progression of psychosis, suggesting that hippocampus is part of the circuits regulating aberrant glucose metabolism in early psychosis. Worsening of glucose metabolism in CHR group was associated with worse clinical outcome measures indicating a need for heightened clinical attention to metabolic problems already in CHR.

Fig. 1. Visualization of hippocampal subfields.

Upper: 3D-visualisation of hippocampal subfields generated with FreeSurfer. Lower: Hippocampal subfield segmentation masks shown overlaid on a representative coronal plane of T1-weighted MRI image.

Fig. 2. Associations between insulin or insulin resistance and hippocampal tail volumes.

There is a significant subfield-specific inverse association between fasting plasma insulin (A) or insulin resistance (B) and hippocampal tail volumes in first-episode psychosis patients at the one-year follow-up time point. The model is adjusted for age, sex, TIV and BMI. Adjusting for lifetime antipsychotic exposure did not change these results.

Fig. 3. Volume differences of CA1 and tail subfields.

Illustration of differences in CA1 and tail volumes of FEP, CHR-C, CHR-NC and CTR at baseline (A) and at one-year follow-up (B). The volumes are adjusted for age, sex, TIV and BMI.

Fig. 4. Longitudinal changes in insulin and insulin resistance.

The change of fasting plasma insulin (A) and insulin resistance (B) during one-year follow-up. The increase in fasting plasma insulin and insulin resistance is statistically significant only in CHR-C. The model is adjusted for age, sex, and BMI. Adjusting for lifetime antipsychotic exposure did not change these results.