Clinical Outcomes of Continuation and Maintenance Electroconvulsive Therapy

IMPORTANCE: Large-scale evidence for the efficacy of continuation and maintenance electroconvulsive therapy (c/mECT) is lacking. OBJECTIVE: To provide an exhaustive and naturalistic insight into the real-world outcomes and the cost-effectiveness of c/mECT in a large dataset. DESIGN, SETTING, AND PARTICIPANTS: This cohort study included all patients in the Danish National Patient Registry who initiated treatment with ECT from 2003 through 2022. The data were analyzed from October 2023 to February 2024. EXPOSURES: ECT. An algorithm to identify c/mECTs in the dataset was developed: (>3 treatments with ≥7 and <90 days between adjacent treatments, occurring within a time frame of 180 days [cECT] or more [mECT] after an acute [aECT] series). MAIN OUTCOMES AND MEASURES: The association of c/mECT with subsequent 6- to 12-month risk of hospitalization or suicidal behavior using Cox proportional hazard regression with multiple adjustments and aECT only as a reference, propensity score matching, and self-controlled case series analysis using a Poisson regression model. A cost-effectiveness analysis based on hospitalization and ECT expenses was made. RESULTS: A total of 19 944 individuals were treated with ECT (12 157 women [61%], 7787 men [39%]; median [IQR] age, 55 [41-70] years). Of these, 1533 individuals (7.7%) received c/mECT at any time point (1017 [5.1%] cECT only and 516 [2.6%] mECT). Compared with patients receiving aECT only, c/mECT patients more frequently experienced schizophrenia (odds ratio [OR], 2.14; 95% CI, 1.86-2.46) and schizoaffective disorder (OR, 2.42; 95% CI, 1.90-3.09) and less frequently unipolar depression (OR, 0.56; 95% CI, 0.51-0.62). In all models, c/mECT was associated with a lower rate of hospitalization after finishing aECT (eg, 6-month adjusted hazard ratio, 0.68; 95% CI, 0.60-0.78 [Cox regression]; 6-month incidence rate ratio, 0.51; 95% CI, 0.41-0.62 [Poisson regression]). There was no significant difference in the risk of suicidal behavior. Compared with the periods before the end of aECT, c/mECT was associated with a substantial reduction in total treatment costs. CONCLUSIONS AND RELEVANCE: In a nationwide and naturalistic setting, c/mECT after aECT was infrequently used but associated with a lower risk of readmission than aECT alone. The totality of the evidence indicates that c/mECT should be considered more often to prevent relapse after successful aECT in patients whose condition does not respond sufficiently to other interventions.