Blood biomarkers of neuronal injury and astrocytic reactivity in electroconvulsive therapy

Despite electroconvulsive therapy (ECT) being recognized as an effective treatment for major depressive episodes (MDE), its application is subject to controversy due to concerns over cognitive side effects. The pathophysiology of these side effects is not well understood. Here, we examined the effects of ECT on blood-based biomarkers of neuronal injury and astrocytic reactivity. Participants with a major depressive episode (N = 99) underwent acute ECT. Blood was sampled just before (T0) and 30 min after (T1) the first ECT session, as well as just before the sixth session (T2; 48-72 h after the fifth session). Age- and sex-matched controls (N = 99) were recruited from the general population. Serum concentrations of neurofilament light chain (NfL), total tau protein, and glial fibrillary acidic protein (GFAP) were measured with ultrasensitive single-molecule array assays. Utilizing generalized least squares regression, we compared baseline (T0) biomarker concentrations against those of our control group, and calculated the shifts in serum biomarker concentrations from baseline to immediately post-first ECT session (T1), and prior to the sixth session (T2). Baseline analysis revealed that serum levels of NfL (p < 0.001) and tau (p = 0.036) were significantly elevated in ECT recipients compared with controls, whereas GFAP levels showed no significant difference. Relative to T0, serum NfL concentration neither changed at T1 (mean change 3.1%, 95%CI -0.5% to 6.7%, p = 0.088) nor at T2 (mean change -3.2%, 95%CI -7.6% to 1.5%, p = 0.18). Similarly, no change in total tau was observed (mean change 3.7%, 95%CI -11.6% to 21.7%, p = 0.65). GFAP increased from T0 to T1 (mean change 20.3%, 95%CI 14.6 to 26.3%, p < 0.001), but not from T0 to T2 (mean change -0.7%, 95%CI -5.8% to 4.8%, p = 0.82). In conclusion, our findings suggest that ECT induces a temporary increase in serum GFAP, possibly reflecting transient astrocytic activation. Importantly, we observed no indicators of neuronal damage or long-term elevation in any assessed biomarker.