Optimal Electrical Stimulus During Electroconvulsive Therapy

Purpose and aims The purpose of the study is to determine the electrical current stimulus for electroconvulsive therapy (ECT) that produces the optimal balance between antidepressant effect and memory disturbance. Specifically, this study aims to compare 0.5 ms and 1.0 ms pulse width stimuli.

Primary objective: To test the hypothesis that a 1.0 ms pulse width produces a higher remission rate (\< 11 on the Montgomery-Åsberg Depression Rating Scale, self assessed version (MADRS-S)) than a 0.5 ms pulse width within 1 week after the treatment series.

The secondary objectives are to test for differences in:

1. self-rated global health measured with the EuroQual-group 5 Dimensions Scale Visual analogue scale (EQ5D-VAS)
2. subjective memory worsening (increase of 2 on the memory item of the Comprehensive Psychopathological Rating Scale (CPRS))
3. antidepressive response (decrease of 50% on the MADRS-S)
4. number of ECTs in the treatment series
5. readmission and suicide rate within 6 months
6. self-rated global health measured with the EQ5D-VAS at 6 month follow-up
7. subjective memory worsening (increase of 2 on the memory item of the CPRS) at 6 month follow-up
8. remission rate (\< 11 on MADRS-S) at 6 month follow-up

Hypotheses:

1. The longer pulse width produces a higher remission rate (MADRS-S \< 11) than the shorter pulse width within 1 week after treatment.
2. The longer pulse width produces a higher self-rated health status (EQ5D VAS) than the shorter pulse width within 1 week after the treatment.
3. The longer pulse width produces higher rates of subjective memory worsening than the shorter pulse width.
4. The longer pulse width produces a more rapid antidepressive effect than the shorter pulse width (fewer ECTs per treatment series).
5. The longer pulse width produces fewer relapses (rehospitalizations and suicides) than the shorter pulse width within 6 months after the treatment.
6. There are differences in the optimal pulse width between subgroups of patients based on sex, age-group, or initial psychosis status.

3.1 Procedure for randomization: After documented consent, the patient's personal number and hospital are registered in the web-based Q-ECT. This computer-based system presents and records the randomized treatment allocation stratified by hospital and age group.

3.2 Evaluations during the study Patients can be either hospitalized or ambulant. Clinical and adverse effects will be evaluated and at least weekly evaluations are recommended during the index period (14).

3.3 ECT The ECT will be performed according to clinical routine, usually three times per week during the index series (14). The anesthetic doses, electrode placement, and stimulus parameters will be adjusted throughout the treatment course, based on seizure quality, clinical improvement, and adverse effects.

3.4 Depressive symptoms Depressive symptoms will be examined by the MADRS-S (37), before ECT, within 1 week after termination of treatment, and at 6 month follow-up. Remission is defined as a score of 10 or less on the MADRS-S. Patients not able to complete the MADRS-S will be rated by the clinician-completed MADRS (38). All patients will also be rated on the Clinical Global Impression Scale (CGI)(39). Antidepressive response (50% reduction on the MADRS-S) will also be reported.

3.5 Quality of life The EQ5D (40) is used to document the self-rated health status before ECT, within 1 week after treatment termination, and at 6 month follow-up.

3.6 Subjective memory Patients will be asked to self-rate their level of subjective memory impairment before ECT, within 1 week after treatment termination, and at 6-month follow-up, using the memory item from the CPRS.

3.7 Hospital readmission After completion of the study, information about previous hospital admissions and hospital admissions during the follow-up period will be collected from the Patient register for statistical analyses.

3.8 Medication After completion of the study, information about previous psychotropic medication and psychotropic medication administered during the follow-up period will be collected from the Medication register for statistical analyses. Information about medication during ECT will be extracted from the Q-ECT.

3.9 Social factors After completion of the study, information about education level, employment status, and cohabitation status will be collected from Statistics Sweden for statistical analyses.

3.10 Co-morbidities After the completion of the study, co-morbidities data will be collected from the Patient register.

4.1 Patients ending their participation in the study

• Patients can choose to end their participation in study activities at any time, including completion of self-assessments; available register data will be used if the patients agree.

5.1 Collection of safety data The safety of the different ECT dosages will be assessed according to clinical routine. Any adverse events during ECT or within 1 week after ECT will be recorded. Adverse events will also be recorded 6 months after the treatment. Moreover, hospital admissions and diagnosis in ambulant care, as well as causes of deaths, will be collected from the Patient register and Causes of Death register.

7.1 Statistical analyses The primary outcome (\< 11 on the MADRS-S within 1 week after termination of treatment) will be analyzed using the sample that was randomized and received the intended treatment in the first treatment session (modified intention-to-treat sample). Randomized patients that not receive the intended allocation will be presented, but not included in the primary analysis. If data on MADRS-S and MADRS are missing, remission status will be estimated and imputed according to CGI. The primary outcome will be analyzed using logistic regression in a model including treatment allocation, site, psychosis prior to ECT, age group, and number of antidepressant medications during the last year as independent factors (34). The results of a univariate model including only treatment allocation will also be presented. Stratified results will be presented according to sex, age group, and initial psychosis status. The same model without imputed data will be presented in a sensitivity analysis.

A power analysis indicated that 324 patients are required in each group assuming a 44% remission rate in one group and a 55% remission rate in the other group, a power of 80%, and a double-sided test with alpha of 0.05. Assuming that some patients will not receive the intended allocation, or will be lost to follow-up, we aim to include 400 patients in each group.

The secondary outcome of memory worsening from before ECT to within 1 week after ECT and at 6 month follow-up will be analyzed using repeated measures logistic regression in a model including treatment allocation, site, age group, and sex as independent factors. Stratified results will be presented according to sex and age group.

Self-rated overall health measured with the EQ5D-VAS within 1 week after the treatment series and at 6 month follow-up will be analyzed using repeated measures linear regression in a model with treatment allocation, site and age-group as independent factors. Stratified results will be presented according to sex, age group, and initial psychosis status.

The secondary outcome antidepressive response (reduction of 50% on the MADRS-S) from before ECT to within 1 week after ECT and at 6 month follow-up will be analyzed using repeated measures logistic regression in a model including treatment allocation, site, psychosis prior to ECT, age group, and number of antidepressant medications during the last year as independent factors. Stratified results will be presented according to sex and age-group.

The number of ECTs in the treatment series will be analyzed using repeated measures linear regression in a model including treatment allocation and site as independent factors, and stratified by remission status. Stratified results will be presented according to sex, age-group and initial psychosis status.

Hospital readmission and suicide rates will be presented using the Kaplan-Meier technique, and differences between treatment arms will be analyzed using Cox regression in a model including treatment allocation, site, number of previous hospitalizations, number of antidepressant medications during the last year, and age group as independent factors. Stratified results will be presented according to sex, age group and initial psychosis status.