Among 2,408 participants — 1,456 MRI-confirmed lacunar stroke cases and 952 hospital controls — Mendelian randomization (MR) analysis distilled two genuinely causal metabolite associations from 211 observational signals. Higher genetically instrumented glycine levels were inversely associated with isolated lacunar infarcts (OR 0.704, 95% CI 0.585–0.848), while a greater proportion of cholesterol to total lipids within medium LDL (%C/Total in M-LDL) was inversely associated with the more severe multiple-infarct/leukoaraiosis subtype (OR 0.541, 95% CI 0.401–0.730). Omega-3 fatty acids also showed MR-supported links to diffusion tensor imaging markers of small vessel disease.
The use of bidirectional MR to filter 211 observational associations down to two robust causal candidates is methodologically sound and substantially reduces the confounding problem endemic to metabolomics cohort studies. Glycine's protective signal is biologically plausible — the amino acid modulates vascular inflammation, endothelial nitric oxide signalling, and platelet aggregation, all implicated in cerebral small vessel disease. The LDL subfraction finding adds nuance beyond standard LDL-cholesterol metrics and suggests lipoprotein particle composition matters more than total burden for this stroke subtype. Practically, glycine is inexpensive and available as a supplement, making it an actionable candidate for intervention trials. However, MR establishes directionality, not a therapeutic dose-response curve, and the cohort is UK-based and predominantly older — limiting global generalisability. This is an incremental but genuinely clarifying advance for SVD pathogenesis research.