Predictors of prolonged seizures in electroconvulsive therapy: a 10-year retrospective study of 48,571 treatments

Background

Electroconvulsive therapy (ECT) is a highly effective treatment for severe and treatment-resistant psychiatric disorders. Its therapeutic efficacy depends on the induction of a generalized seizure, with seizure duration and quality routinely monitored using motor and electroencephalographic (EEG) measures. While an adequate seizure is necessary for clinical benefit, excessively prolonged seizures may increase the risk of post-ECT confusion, delirium, cognitive impairment, hypoxia, and other medical complications. Existing literature has suggested potential risk factors for prolonged seizures, including younger age, specific medications, anesthetic agents, and treatment parameters; however, findings remain inconsistent. Improved identification of predictors may support individualized risk stratification and safer ECT delivery.

Aims

This study aimed to determine the incidence of prolonged seizures during ECT and identify demographic and clinical predictors associated with their occurrence in a large clinical cohort from a major academic ECT center.

Methods

We conducted a retrospective cohort study of patients who received ECT at the Centre for Addiction and Mental Health, Toronto, Canada, between June 2010 and September 2021. The study included 48,571 ECT treatments administered to 1,551 unique patients. Prolonged seizure was defined as an EEG-recorded seizure duration greater than 120 seconds during any ECT treatment. Patients receiving standard unilateral ECT were excluded. Demographic and clinical variables were extracted from medical records, including age, gender, ethnicity, diagnosis, comorbid psychiatric and medical conditions, substance use disorder, concurrent psychotropic medications, and ECT electrode placement. Descriptive statistics were used to summarize cohort characteristics. Group comparisons were performed using t-tests, chi-square tests, or Fisher’s exact tests, as appropriate. Univariate and multivariable logistic regression models were used to identify predictors of prolonged seizures. Missing data were handled using multiple imputation with 50 imputed datasets, and pooled estimates were calculated using Rubin’s rules. Stepwise logistic regression was used to select variables for the final multivariable model.

Results

Among 1,551 patients, 287 individuals (18.5%) experienced at least one prolonged seizure during their ECT course. Across all treatments, 625 prolonged seizures were identified, corresponding to an incidence of 1.29% per ECT treatment. Patients who experienced prolonged seizures were significantly younger than those who did not, with a mean age of 36.3 years compared with 48.8 years (p<0.001). In univariate analysis, older age was associated with lower odds of prolonged seizures, and tricyclic antidepressant use was also associated with reduced odds. In the final multivariable model, younger age remained a robust independent predictor of prolonged seizures (OR=0.947 per year increase, 95% CI 0.938–0.956, p<0.001). Diagnosis was also significantly associated with prolonged seizure risk. Compared with schizophrenia, unipolar mood disorder was associated with higher odds of prolonged seizures (OR=1.865, 95% CI 1.247–2.788, p=0.002). Antiepileptic medication use showed a trend toward increased risk, while tricyclic antidepressant use showed a trend toward reduced risk, though neither reached statistical significance.

Conclusions

Prolonged seizures during ECT were uncommon on a per-treatment basis but occurred in a clinically meaningful proportion of patients over a treatment course. Younger age and diagnosis of unipolar mood disorder were independent predictors. These findings may inform clinical monitoring and risk stratification during ECT.

References

no references