Bidirectional Mendelian Randomization Analysis of 338 Cerebrospinal-Fluid Metabolites and Cluster-Headache Risk.

BACKGROUND

Cluster headache (CH) is a rare but highly disabling primary headache disorder characterized by severe unilateral attacks and autonomic symptoms. The metabolic mechanisms underlying CH remain poorly understood.

OBJECTIVE

To investigate the potential causal effects of cerebrospinal fluid (CSF) metabolite levels on CH risk, and to explore possible reverse causal effects of CH on CSF metabolites, using a bidirectional Mendelian randomization (MR) approach.

METHODS

We performed a bidirectional two-sample Mendelian randomization (MR) analysis integrating genome-wide association study (GWAS) data for 338 cerebrospinal fluid (CSF) metabolites and CH (1,833 cases and 498,515 controls from FinnGen release 12). Genetic instruments were selected at P < 1×10- 5 (LD r2 < 0.01). The primary causal estimates were derived using the inverse-variance weighted (IVW) method under a random-effects model, complemented by MR-Egger, weighted median, and MR-PRESSO sensitivity tests. Multiple testing correction was performed using both Bonferroni and false discovery rate (FDR) approaches.

RESULTS

In the forward MR analysis, 11 CSF metabolites were significantly associated with CH risk (P<0.05). The strongest associations were observed for orotate (β = 0.53, 95% CI: 0.23-0.82, P = 0.0006), betaine (β = 0.47, 95% CI: 0.16-0.79, P = 0.0035), and 5-oxoproline (β = 0.57, 95% CI: 0.17-0.97, P = 0.0053). In the reverse MR analysis, eight metabolites, including lysine (β = 0.015, P = 0.029) and kynurenine (β = 0.025, P = 0.020), were nominally associated with genetic liability to CH. Sensitivity analyses showed no evidence of directional pleiotropy or heterogeneity (all P > 0.05).

CONCLUSION

This bidirectional MR study provides the first genetic evidence linking central metabolic alterations to CH susceptibility. While these results highlight potential metabolic biomarkers and mechanistic pathways, the findings remain preliminary due to modest statistical power and should be replicated in larger and ethnically diverse cohorts.