Causal associations between congenital adrenal hyperplasia and neuropsychiatric conditions- a Mendelian Randomization Study.

BACKGROUND

Congenital adrenal hyperplasia (CAH), predominantly caused by 21-hydroxylase deficiency (21-OHD) due to CYP21A2 mutations, disrupts cortisol synthesis and adrenal androgen homeostasis. Observational studies suggest CAH patients exhibit elevated risks of neuropsychiatric disorders, but causal mechanisms remain unestablished. We hypothesized that reduced CYP21A2 expression, reflecting CAH, differentially influences psychiatric outcomes via tissue-specific pathways.

METHODS

Using two-sample Mendelian randomization (MR), we analyzed tissue-specific CYP21A2 expression quantitative trait loci (eQTLs) from adrenal (GTEx v8) and whole blood (GTEx v8 and eQTLGen meta-analysis). Genetic instruments were validated via positive control MR with classical CAH biomarkers. Associations with ten neuropsychiatric disorders were assessed using inverse-variance-weighted MR, supplemented by sensitivity analyses (LCV, SMR) and LD score regression.

RESULTS

Adrenal-derived CYP21A2 downregulation reduced Alzheimer's disease (AD) risk (discovery: OR = 1.245, replication: OR = 1.100) but increased autism spectrum disorder (ASD) susceptibility (discovery: OR = 0.766, replication: OR = 0.659). Conversely, blood-derived eQTLs showed opposing effects that decreased ASD risk (discovery: OR = 1.072, replication: OR = 1.071) and elevated AD risk (OR = 0.968 for both discovery and replication). Both tissues linked reduced CYP21A2 expression to elevated bioavailable testosterone (adrenal: OR = 0.972, p = 0.04; blood: OR = 0.983, p = 0.01), consistent with CAH pathophysiology.

CONCLUSION

Our research indicates that adrenal-driven pathways of CYP21A2 deficiency may reduce the risk of AD while increasing the ASD risk. These findings underscore the pivotal role of endocrine mechanisms in the pathogenesis of neuropsychiatric disorders and advocate for personalized CAH management integrating mental health monitoring.