Association of high-frequency repetitive transcranial magnetic stimulation with 6-month cognitive outcomes and white matter lesion progression in patients with cerebral small vessel disease and cognitive impairment

To examine the association of high-frequency repetitive transcranial magnetic stimulation (rTMS) over the left dorsolateral prefrontal cortex (DLPFC) with 6-month cognitive outcomes and white matter lesion progression in patients with cerebral small vessel disease (CSVD) and cognitive impairment. A retrospective cohort study design was used. Data from 427 patients with CSVD and cognitive impairment who attended the Department of Neurology of our hospital from January 2022 to December 2025 were retrospectively analyzed. Based on treatment documented in the medical records, patients were classified into an rTMS group (n = 228), who had received 10 Hz high-frequency rTMS over the left DLPFC as part of routine clinical care, and a control group (n = 199), who had received conventional pharmacotherapy alone. Both groups received standardized basic CSVD treatment; the rTMS group additionally received rTMS once daily, 5 days per week, for 2 consecutive weeks (10 sessions). Complete baseline and 6-month neuropsychological and MRI records were required for inclusion; therefore, all 427 patients had complete data for the prespecified outcomes and no imputation was required. The primary outcome was the between-group difference in change from baseline in Montreal Cognitive Assessment (MoCA) score at 6 months (ΔMoCA). Secondary outcomes included changes in MMSE, TMT-A, TMT-B, Rey-Osterrieth Complex Figure delayed recall, MoCA responder rates, imaging outcomes, safety, and exploratory subgroup analyses. A propensity score matching (PSM) sensitivity analysis was performed for ΔMoCA, MoCA responder rates, and WMHV change. The rTMS group showed significantly greater delta-MoCA than the control group (mean difference [MD] = 1.11, 95% confidence interval [CI]: 0.56–1.66, p < .001, Cohen d = 0.38). The rTMS group also showed significantly greater improvement in MMSE (MD = 0.87, 95% CI: 0.42–1.32, p < .001, d = 0.37), TMT-B (MD = 14.35 s, 95% CI: 0.86–27.83, p = .037, d = 0.20), and Rey delayed recall (MD = 1.73, 95% CI: 0.55–2.90, p = .004, d = 0.28) compared with the control group. The between-group difference in TMT-A did not reach statistical significance (MD = 4.35 s, 95% CI: −0.37 to 9.08, p = .071), but became significant after ANCOVA adjustment for age, education, and baseline score (p = .001). The rTMS group had significantly higher responder rates than the control group at all three thresholds: delta-MoCA ≥ 1 point (85.1% vs. 70.9%, odds ratio [OR] = 2.35, p < .001), ≥2 points (74.1% vs. 60.3%, OR = 1.89, p = .002), and ≥3 points (61.4% vs. 46.2%, OR = 1.85, p = .002). No statistically significant between-group differences were observed for any imaging parameter (WMHV, FA, ADC, Fazekas grade progression; all p > .05). A PSM sensitivity analysis (195 matched pairs) yielded consistent results for the primary outcome (MD = 1.23, 95% CI: 0.64–1.81, p < .001, d = 0.29). In exploratory subgroup analyses (not prespecified), cognitive severity (interaction p < .001) and diabetes status (interaction p = .014) suggested potential treatment effect modification, with numerically greater benefit in patients with moderate cognitive impairment and those without diabetes. The overall incidence of adverse events did not differ significantly between groups (15.4% vs. 13.6%, p = .602); all events were mild to moderate and transient, and no serious adverse events occurred. In this retrospective cohort study, high-frequency rTMS over the left DLPFC was associated with greater improvement in long-term cognitive function in patients with CSVD and cognitive impairment, particularly in global cognition, executive function, and memory, with a favorable safety profile. Exploratory subgroup analyses suggested that cognitive severity and diabetes status may influence treatment outcomes, warranting prospective validation. rTMS did not significantly delay white matter lesion progression; its cognitive benefits may be mediated primarily through enhanced neuroplasticity and functional network reorganization rather than structural lesion modification.