Objective: The study aims to evaluate the association of methylenetetrahydrofolate reductase (MTHFR) C677T and A1298C polymorphisms with cerebral small vessel disease (CSVD), cerebral large vessel disease (CLVD), vascular cognitive impairment (VCI) and global and regional brain volumes. Methods: Clinical, laboratory, neuroimaging, volumetric, cognitive and MTHFR genotyping data were retrospectively collected from 62 patients with CSVD (n = 40) or CLVD (n = 22), all carriers of at least one MTHFR variant. Brain volumes were compared between patients with VCI carrying C677T (n = 30) and patients without VCI (n = 32). Analyses were repeated in the open-source platform jamovi and p-values were corrected for multiple comparisons with the Holm procedure. Results: The cohort comprised 41 women and 21 men (mean age 56.9 years). The CT and TT genotypes of C677T, the AC and CC genotypes of A1298C, blood pressure, total cholesterol and triglycerides differed between CSVD and CLVD at the unadjusted level (p < 0.05), but none of these differences remained significant after Holm correction. VCI carriers of C677T showed lower white matter (332.9 ± 25.5 vs. 437.9 ± 35.7 cm³) and gray matter volumes (626.0 ± 45.7 vs. 699.7 ± 51.5 cm³), higher cerebrospinal fluid volume (238.3 ± 42.2 vs. 197.1 ± 24.8 cm³) and reduced hippocampal, amygdalar, thalamic, superior frontal, middle temporal, entorhinal, parahippocampal and middle occipital volumes (all Holm-adjusted p < 0.05). Conclusions: Differences in MTHFR genotype distribution between CSVD and CLVD were not robust to correction for multiple testing, whereas VCI carriers of C677T showed global and selective regional volume loss that withstood correction. MTHFR variants are susceptibility markers rather than direct causes of vascular injury; these associations, which could not be adjusted for comorbid vascular risk factors or for concomitant neurodegenerative pathology, require confirmation in prospective, controlled studies with homocysteine measurements.

Background: Elevated plasma homocysteine, arising from impaired one-carbon metabolism, has been implicated in cerebral microvascular injury and in neurodegeneration. The methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism is the commonest genetic determinant of homocysteine concentration, but whether it leaves a measurable structural footprint on the brain remains unresolved. We synthesised genotype-stratified structural MRI findings in Alzheimer’s disease, mild cognitive impairment, and vascular cognitive disorders. Methods: Four sources (PubMed, Europe PMC, OpenAlex, Semantic Scholar) were searched up to 20 September 2026. Eligible reports gave C677T-stratified structural MRI outcomes in adults with clinically defined cognitive disorders. Quality was appraised with an adapted Newcastle–Ottawa Scale, and exploratory random-effects meta-analyses supplemented a narrative synthesis. Results: Thirteen studies (3880 participants across partly overlapping cohorts) were included. Moderate-to-severe white matter lesions were modestly more frequent in TT homozygotes, consistently across contributing studies, with no consistent excess in heterozygotes. Regional volume findings were inconsistent and largely confined to one cohort, whereas whole-brain atrophy rate, hippocampal and medial temporal measures, lacunes, microbleeds, and diffusion metrics showed no genotype-only association. Homocysteine was higher in T-allele carriers wherever measured. Conclusions: The evidence is more compatible with a modest association with vascular white matter injury than with a reproducible neurodegenerative imaging phenotype and does not support genotype-guided clinical decision-making. Longitudinal, biomarker-characterised cohorts with harmonised imaging, adequately powered genotype groups, and measured nutritional status are the main priorities.