Exploring Vascular Endothelial Growth Factor and Other Blood-Brain Barrier Biomarkers in Cognition of First-Episode Psychosis: An Observational Study.

Background

Cognitive deficits are a core feature of early stages of schizophrenia. However, according to neurodevelopmental models, the extent to which chemokines and growth factors are involved in cognitive function remains debatable. We aimed to investigate whether homeostatic/inflammatory chemokines and growth factors are associated with cognitive impairment in patients with first-episode psychosis (FEP) in remission.

Methods

Fifty patients, 21 healthy siblings, and 24 controls participated in the study. The primary outcomes were associations between cognition and growth factors (brain-derived neurotrophic factor [BDNF] and vascular endothelial growth factor [VEGF]), homeostatic markers (CXCL12), and inflammatory chemokines (CCL2, CCL3, CX3CL1, and CCL11) using a whole-blood immunoassay procedure. Differences between the FEP group, siblings, and controls were also examined to understand distinct group profiles.

Results

The VEGF levels were significantly higher in the FEP group than in the control group. High VEGF levels are significantly associated with lower social cognition scores. Moreover, a post hoc hierarchical regression model explained 34.5% of the variance in social cognition (F11, 32 =1.533, P=.168), with inflammatory variables explaining 13.5% and VEGF showing statistical significance (β=-1.936, P=.022). No additional significant results were found for the other inflammatory biomarkers.

Conclusions

Our preliminary results suggest that an increase in VEGF might help preserve social cognition after first- episode psychosis. These findings might suggest that a compensatory mechanism could outweigh other VEGF- related hypotheses, such as blood-brain barrier opening and chronic neuroinflammation. However, this hypothesis requires further investigation to address the methodological challenges of determining chemokine levels and controlling for confounding variables.