Electroconvulsive therapy, depression severity and mortality: Data from the Danish National Patient Registry

Background: The effects of electroconvulsive therapy are usually estimated from changes in depression scales from studies with relatively small patient samples. Larger patient samples can be achieved from epidemiological registers, which provide information on other social and clinical predictors, results and risks. Aims: To examine whether depression severity predicts the use of electroconvulsive therapy, risk of re-hospitalization, suicidal behaviour and mortality following electroconvulsive therapy in patients with major depression. Methods: A cohort of 92,895 patients diagnosed with single or recurrent depression between 2005 and 2016 in the Danish National Patient Registry was followed for electroconvulsive therapy and adverse outcomes. Associations between electroconvulsive therapy and outcomes were analysed using Cox regression. Results: A total of 5004 (5.4%) patients were treated with electroconvulsive therapy. Depression severity was the strongest predictor of electroconvulsive therapy. Electroconvulsive therapy was used more frequently above age 70, in those better educated or married, whereas comorbid alcohol abuse or history of prior stroke at study entry were associated with lower rates. Electroconvulsive therapy was associated with lower mortality. The adjusted hazard ratio for the association between electroconvulsive therapy and suicide in patients with mild depression was 6.99 (3.30–14.43), whereas it was 1.10 (0.55–2.20) in those with severe depression and psychotic symptoms. A similar pattern was seen for emergency contacts and attempted suicide. Conclusions: Electroconvulsive therapy was associated with lower all-cause mortality and the relative risk for re-hospitalization and attempted and committed suicide was lowest in patients with the most severe depression. Electroconvulsive therapy is an important treatment, with significant public health benefits, for patients with severe depression.