An exploratory study of glycemic biomarkers and rTMS treatment outcomes in treatment-resistant depression

Repetitive transcranial magnetic stimulation (rTMS) is an efficacious therapy for treatment-resistant depression (TRD) and is thought to drive neuroplastic changes in affective networks. However, up to half of patients with TRD do not achieve response or remission with rTMS. Metabolic dysfunction is more prevalent in major depressive disorder (MDD) and may impair neuroplasticity, potentially reducing rTMS efficacy. This study examined associations between baseline serum systemic glucose regulation measures and rTMS outcomes in TRD. 46 adults (aged 20–65) with moderate-to-severe MDD and ≥ 2 failed antidepressant trials completed an open-label trial of connectivity-guided, accelerated intermittent theta burst stimulation (iTBS). Participants were classified according to the American Diabetic Association (ADA) (ADA-normal vs. ADA-elevated). Glycated hemoglobin (HbA1c), fasting glucose (FG), and fasting insulin (FI) were collected at baseline, and homeostatic model assessment of insulin resistance (HOMA-IR) was calculated from FG and FI. Baseline associations were tested using MLR, and Week-4 outcomes were examined using ANCOVA models adjusting for baseline symptom severity, age, and sex. Response (≥ 50% Montgomery–Åsberg Depression Rating Scale (MADRS) reduction) and remission (MADRS ≤ 10) at week 4 were the primary outcomes and were compared across metabolic groups with Fisher’s exact tests. Sensitivity analyses excluded glucose-lowering medication users, repeated in the full cohort, and included BMI as an additional covariate. Across both glycemic groups, none of the baseline metabolic biomarkers (HbA1c, FG, FI, and HOMA-IR) predicted baseline depression severity (all p > 0.05) or Week-4 MADRS outcomes after adjustment (all p > 0.10). FI and HOMA-IR were not significantly associated with Week-4 MADRS outcomes in the ANCOVA models (both p = 0.099). Group differences in response and remission rates were not statistically significant for both response (p = 0.314) and remission (p = 0.695). Sensitivity analyses were also non-significant. Given the low power (5–41%), these findings should be interpreted cautiously. Glycemic biomarkers and metabolic status were not associated with clinical outcomes following accelerated iTBS. Post-hoc power analyses indicated that larger samples are required to detect effects. Future studies should investigate insulin-related mechanisms and dynamic glucose markers as potential moderators of depression severity and treatment response. Prospectively registered on ClinicalTrials.gov (NCT05813093). Registration date: 13th of March 2023.