ORIGINAL STUDY
Depression Severity in Electroconvulsive Therapy (ECT) Versus Pharmacotherapy Trials Charles H. Kellner, MD,* David C. Kaicher, MD,† Hiya Banerjee, PhD,‡ Rebecca G. Knapp, PhD,‡ Rachael J. Shapiro, BA,§ Mimi C. Briggs, BA,∥ Rosa M. Pasculli, BA,¶ Dennis M. Popeo, MD# Gabriella M. Ahle, BA,* and Lauren S. Liebman, BA* Objective: We sought to compare the level of severity of depressive symptoms on entry into electroconvulsive therapy (ECT) clinical trials versus pharmacotherapy clinical trials. Data Sources: English-language MEDLINE/PubMed publication databases were searched for ECT literature (search terms: ECT, electroconvulsive therapy, depression, and Hamilton) for clinical trials in which depressed patients had baseline Hamilton Rating Scale for Depression (HRSD) scores. For comparison, we used a convenience sample of 7 large pharmacotherapy trials in major depression (N = 3677). The search included articles from 1960 to 2011. Study Selection: We included 100 studies that met the following criteria: ECT trial for depression, patients adequately characterized by diagnosis at baseline, and patients rated at baseline by 15-item HRSD (HRSD15), HRSD17, HRSD21, HRSD24, or HRSD28, with mean (SD) and sample size (n) reported. For the comparator pharmacotherapy trials, we chose to use a subset of the studies (excluding one study of minor depression) in the widely publicized meta-analysis of Fournier et al, as well as the STAR*D study and one additional study by Shelton et al. This provided 7 studies of major depression using HRSD17 (total N = 3677). Data Extraction: Data extracted included number of subjects and baseline and final HRSD scores, with mean (SD) values. Results: Of 100 ECT studies, 56 studies (N = 2243) used the HRSD17 version. The mean baseline HRSD17 score in the ECT trials was 27.6, the mean in the pharmacotherapy trials was 21.94, a statistically, and clinically, significant difference. In a subanalysis of the 16 ECT studies that used the HRSD24 version, the mean baseline score was 32.2. Conclusions: This selective literature review confirms that patients who entered ECT clinical trials were more severely ill than those who entered the selected comparator pharmacotherapy trials. Such data highlight the critical role of ECT in the treatment of severe and treatmentresistant mood disorders.
From the *Icahn School of Medicine at Mount Sinai, New York, NY; †CrozerKeystone Health Network Chester, PA; ‡Medical University of South Carolina, Charleston, SC; §New York Medical College Valhalla, NY; ∥State University of New York Upstate Medical University, Syracuse, NY; ¶New York University School of Medicine, New York, NY; and #New York University, New York, NY. Received for publication December 3, 2012; accepted March 17, 2014. Reprints: Charles H. Kellner, MD, Mount Sinai Hospital, Department of Psychiatry One Gustave L Levy Place, New York, NY 10029 (e‐mail: [email protected]). At the time this work was done, David C. Kaicher, Mimi C. Briggs, and Rosa M. Pasculli were affiliated with the Icahn School of Medicine at Mount Sinai. All authors contributed equally to this manuscript, with the exception of Gabriella Ahle and Lauren Liebman, whose contribution is limited to the revision of this manuscript. Drs. Kellner, Banerjee, Knapp, and Popeo report NIMH grant support (09-0429) and no potential conflicts of interest. Dr. Kellner reports the following financial interests: honoraria from the North Shore-LIJ Health System, UpToDate, and Psychiatric Times, and royalties from Cambridge University Press. Dr. Kaicher, Ms. Shapiro, Ms. Briggs, and Ms. Pasculli report no financial interests and no potential conflicts of interest. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. DOI: 10.1097/YCT.0000000000000135
Key Words: electroconvulsive therapy, ECT, pharmacotherapy, major depression, Hamilton Rating Scale for Depression (J ECT 2015;31: 31–33)
E
lectroconvulsive therapy (ECT) is generally considered the most effective treatment for severe and treatment-resistant depression.1 In fact, severity of illness is the most important factor compelling the recommendation to proceed with ECT. In addition to being severely ill, many ECT patients have not responded to multiple trials of antidepressant medications, thus qualifying them as “treatment resistant”.2 To our knowledge, there are no reports in the literature of overall mean depression severity ratings across ECT clinical trials. Such data are important, as they provide a snapshot of how ill patients referred for ECT trials are. We hypothesized that patients entered into ECT trials would be much more severely ill at baseline, compared with patients entered into antidepressant medication trials. Given the heterogeneity of trial design and data reporting, it was expected that the methodology of collecting meaningfully comparable data would be complex, and that criteria for eligible studies would need to be determined, at times arbitrarily.
MATERIALS AND METHODS Selection of Papers We searched the English-language ECT literature from 1960 to 2011 for clinical trials in which depressed patients had baseline Hamilton Rating Scale for Depression (HRSD) rating scores using either the 15-item HRSD (HRSD15), the 17-item HRSD (HRSD17), 21-item HRSD (HRSD21), 24-item HRSD (HRSD24), or the 28-item HRSD (HRSD28) version.3 Electroconvulsive therapy trials for patients with diagnoses other than depression (eg, mania and schizophrenia) were excluded. We included studies referenced in standard textbooks of ECT1,4 as well as the recently compiled FDA executive summary.5 We searched PubMed using the search terms electroconvulsive therapy, ECT, depression, and Hamilton. We then read each paper and culled data from each (Table 1). In total, 100 studies met the following criteria: ECT trial for depression, patients adequately characterized by diagnosis at baseline, and patients rated at baseline by HRSD15, HRSD17, HRSD21, HRSD24, or HRSD28, with mean (SD) and sample size (n) reported. Some studies did not report the exact baseline score in the article text; in such cases, it was read off a figure, creating some degree of inaccuracy. Of the 100 studies containing sufficient information to be included (mean [SD] and sample size [n]), 5 studies reported data for HRSD15, 56 studies reported data for HRSD17, 17 studies reported data for HRSD21, 16 studies reported data for the RSD24, and 6 studies reported data for HRSD28. For the comparator pharmacotherapy trials, we chose to use a subset of the studies (excluding one study of minor depression) in the widely publicized meta-analysis of Fournier et al6 and the
Journal of ECT • Volume 31, Number 1, March 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.ectjournal.com
31
Journal of ECT • Volume 31, Number 1, March 2015
Kellner et al
TABLE 1. Pooled Hamilton Rating Scale for Depression (HRSD) Baseline Means and 95% CI for ECT and Pharmacotherapy Studies ECT Studies HRSD Scale 15-item 17-item 21-item 24-item 28-item
Pharmacotherapy Studies
Pooled Mean
95% CI
No. Studies/ Combined Sample Size
26.8 27.6 27.4 32.2 25.8
(22.4, 31.2) (26.4, 28.8) (24.7, 30.0) (28.9, 35.6) (20.0, 31.6)
5/116 56/2243 17/494 16/1652 6/112
Pooled Mean
95% CI
No. Studies/ Combined Sample Size
P*
21.94
(21.4, 22.5)
7/3677
Depression Severity in Electroconvulsive Therapy (ECT) Versus Pharmacotherapy Trials Charles H. Kellner, MD,* David C. Kaicher, MD,† Hiya Banerjee, PhD,‡ Rebecca G. Knapp, PhD,‡ Rachael J. Shapiro, BA,§ Mimi C. Briggs, BA,∥ Rosa M. Pasculli, BA,¶ Dennis M. Popeo, MD# Gabriella M. Ahle, BA,* and Lauren S. Liebman, BA* Objective: We sought to compare the level of severity of depressive symptoms on entry into electroconvulsive therapy (ECT) clinical trials versus pharmacotherapy clinical trials. Data Sources: English-language MEDLINE/PubMed publication databases were searched for ECT literature (search terms: ECT, electroconvulsive therapy, depression, and Hamilton) for clinical trials in which depressed patients had baseline Hamilton Rating Scale for Depression (HRSD) scores. For comparison, we used a convenience sample of 7 large pharmacotherapy trials in major depression (N = 3677). The search included articles from 1960 to 2011. Study Selection: We included 100 studies that met the following criteria: ECT trial for depression, patients adequately characterized by diagnosis at baseline, and patients rated at baseline by 15-item HRSD (HRSD15), HRSD17, HRSD21, HRSD24, or HRSD28, with mean (SD) and sample size (n) reported. For the comparator pharmacotherapy trials, we chose to use a subset of the studies (excluding one study of minor depression) in the widely publicized meta-analysis of Fournier et al, as well as the STAR*D study and one additional study by Shelton et al. This provided 7 studies of major depression using HRSD17 (total N = 3677). Data Extraction: Data extracted included number of subjects and baseline and final HRSD scores, with mean (SD) values. Results: Of 100 ECT studies, 56 studies (N = 2243) used the HRSD17 version. The mean baseline HRSD17 score in the ECT trials was 27.6, the mean in the pharmacotherapy trials was 21.94, a statistically, and clinically, significant difference. In a subanalysis of the 16 ECT studies that used the HRSD24 version, the mean baseline score was 32.2. Conclusions: This selective literature review confirms that patients who entered ECT clinical trials were more severely ill than those who entered the selected comparator pharmacotherapy trials. Such data highlight the critical role of ECT in the treatment of severe and treatmentresistant mood disorders.
From the *Icahn School of Medicine at Mount Sinai, New York, NY; †CrozerKeystone Health Network Chester, PA; ‡Medical University of South Carolina, Charleston, SC; §New York Medical College Valhalla, NY; ∥State University of New York Upstate Medical University, Syracuse, NY; ¶New York University School of Medicine, New York, NY; and #New York University, New York, NY. Received for publication December 3, 2012; accepted March 17, 2014. Reprints: Charles H. Kellner, MD, Mount Sinai Hospital, Department of Psychiatry One Gustave L Levy Place, New York, NY 10029 (e‐mail: [email protected]). At the time this work was done, David C. Kaicher, Mimi C. Briggs, and Rosa M. Pasculli were affiliated with the Icahn School of Medicine at Mount Sinai. All authors contributed equally to this manuscript, with the exception of Gabriella Ahle and Lauren Liebman, whose contribution is limited to the revision of this manuscript. Drs. Kellner, Banerjee, Knapp, and Popeo report NIMH grant support (09-0429) and no potential conflicts of interest. Dr. Kellner reports the following financial interests: honoraria from the North Shore-LIJ Health System, UpToDate, and Psychiatric Times, and royalties from Cambridge University Press. Dr. Kaicher, Ms. Shapiro, Ms. Briggs, and Ms. Pasculli report no financial interests and no potential conflicts of interest. Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. DOI: 10.1097/YCT.0000000000000135
Key Words: electroconvulsive therapy, ECT, pharmacotherapy, major depression, Hamilton Rating Scale for Depression (J ECT 2015;31: 31–33)
E
lectroconvulsive therapy (ECT) is generally considered the most effective treatment for severe and treatment-resistant depression.1 In fact, severity of illness is the most important factor compelling the recommendation to proceed with ECT. In addition to being severely ill, many ECT patients have not responded to multiple trials of antidepressant medications, thus qualifying them as “treatment resistant”.2 To our knowledge, there are no reports in the literature of overall mean depression severity ratings across ECT clinical trials. Such data are important, as they provide a snapshot of how ill patients referred for ECT trials are. We hypothesized that patients entered into ECT trials would be much more severely ill at baseline, compared with patients entered into antidepressant medication trials. Given the heterogeneity of trial design and data reporting, it was expected that the methodology of collecting meaningfully comparable data would be complex, and that criteria for eligible studies would need to be determined, at times arbitrarily.
MATERIALS AND METHODS Selection of Papers We searched the English-language ECT literature from 1960 to 2011 for clinical trials in which depressed patients had baseline Hamilton Rating Scale for Depression (HRSD) rating scores using either the 15-item HRSD (HRSD15), the 17-item HRSD (HRSD17), 21-item HRSD (HRSD21), 24-item HRSD (HRSD24), or the 28-item HRSD (HRSD28) version.3 Electroconvulsive therapy trials for patients with diagnoses other than depression (eg, mania and schizophrenia) were excluded. We included studies referenced in standard textbooks of ECT1,4 as well as the recently compiled FDA executive summary.5 We searched PubMed using the search terms electroconvulsive therapy, ECT, depression, and Hamilton. We then read each paper and culled data from each (Table 1). In total, 100 studies met the following criteria: ECT trial for depression, patients adequately characterized by diagnosis at baseline, and patients rated at baseline by HRSD15, HRSD17, HRSD21, HRSD24, or HRSD28, with mean (SD) and sample size (n) reported. Some studies did not report the exact baseline score in the article text; in such cases, it was read off a figure, creating some degree of inaccuracy. Of the 100 studies containing sufficient information to be included (mean [SD] and sample size [n]), 5 studies reported data for HRSD15, 56 studies reported data for HRSD17, 17 studies reported data for HRSD21, 16 studies reported data for the RSD24, and 6 studies reported data for HRSD28. For the comparator pharmacotherapy trials, we chose to use a subset of the studies (excluding one study of minor depression) in the widely publicized meta-analysis of Fournier et al6 and the
Journal of ECT • Volume 31, Number 1, March 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.ectjournal.com
31
Journal of ECT • Volume 31, Number 1, March 2015
Kellner et al
TABLE 1. Pooled Hamilton Rating Scale for Depression (HRSD) Baseline Means and 95% CI for ECT and Pharmacotherapy Studies ECT Studies HRSD Scale 15-item 17-item 21-item 24-item 28-item
Pharmacotherapy Studies
Pooled Mean
95% CI
No. Studies/ Combined Sample Size
26.8 27.6 27.4 32.2 25.8
(22.4, 31.2) (26.4, 28.8) (24.7, 30.0) (28.9, 35.6) (20.0, 31.6)
5/116 56/2243 17/494 16/1652 6/112
Pooled Mean
95% CI
No. Studies/ Combined Sample Size
P*
21.94
(21.4, 22.5)
7/3677
