Efficacy and safety of treating patients with refractory schizophrenia with antipsychotic medication and adjunctive electroconvulsive therapy: a systematic review and meta-analysis.

BACKGROUND: The efficacy and safety of the combined treatment of refractory schizophrenia with antipsychotic medications and electroconvulsive therapy (ECT) remain uncertain. AIMS: Conduct systematic review and meta-analysis of available literature in English and Chinese about ECT in the treatment of refractory schizophrenia.
METHODS: English and Chinese databases were searched for studies published prior to May 20, 2015 regarding the efficacy and safety of the combined treatment of refractory schizophrenia with antipsychotic medications and ECT. Two researchers selected and evaluated studies independently using pre-defined criteria. Review Manager 5.3 software was used for data analysis.
RESULTS: A total of 22 randomized control studies, 18 of which were conducted in mainland China, were included in the analysis. Meta-analysis of data from 18 of the 22 studies with a pooled sample of 1394 individuals found that compared to treatment with antipsychotic medications alone, combined treatment with antipsychotic medications and ECT had significantly higher rates of achieving study-specific criteria of 'clinical improvement' (RR=1.25, 95%CI=1.14-1.37). Based on the Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) criteria, the quality of evidence for this assessment of efficacy was 'moderate'. However, the proportion of participants who experienced headache during the treatment was significantly higher in the combined treatment group (RR=9.10, 95%CI=3.97-20.86, based on a pooled sample of 517 from 8 studies) and the proportion who experienced memory impairment was also higher in the combined treatment group (RR=6.48, 95%CI=3.54-11.87, based on a pooled sample of 577 from 7 studies). The quality of evidence about these adverse events was rated as 'very low'.
CONCLUSIONS: There are very few high quality randomized controlled clinical trials about the combination of antipsychotic medications and ECT in the treatment of refractory schizophrenia. This meta-analysis found that the combination of antipsychotic medications and ECT could improve psychiatric symptoms in patients with refractory schizophrenia, but the incomplete methodological information provided for most of the studies, publication bias (favoring studies with better outcomes in the combined treatment group), and the low quality of evidence about adverse outcomes, cognitive impairment, and overall functioning raise questions about the validity of the results.

Introduction

About 20-30% of patients with schizophrenia are classified as ‘refractory schizophrenia’.[1] The original diagnostic criteria for refractory schizophrenia proposed by Kane in 1995[2,3] were as follows: 1) partially nonresponsive over the past 5 years when treated with three kinds of antipsychotic medications (at least two of which were of different chemical structures) which were administered at appropriate dosages for a sufficient duration; 2) intolerance of side effects of the antipsychotic medications; and 3) relapse or symptomatic deterioration even when taking sufficient doses of appropriate medication. Other widely accepted criteria of refractory schizophrenia include a duration of illness of more than five years; psychiatric symptoms that show no improvement after two-years of regular, full dose and full course treatment with two kinds of antipsychotics; and no response to clozapine.[2]

The difficulty of treating patients with refractory schizophrenia can lead to a poor quality of life for affected individuals.[3] Clozapine is considered an effective medicine for most patients with refractory schizophrenia, but another therapeutic option considered in several studies is the combined use of antipsychotic medication and electroconvulsive therapy (ECT).[1] However, the findings from these studies have been inconsistent: compared to continued use of standard antipsychotic medications,some studies find the combined use of antipsychotic medication and ECT beneficial,some find it no different,and some find it inferior due to the increased occurrence of memory loss and headache.[4] This paper reports on the first known systematic review and meta-analysis on this topic, combining studies reported both in English and in Chinese.

Methods

Search strategy

We searched the following databases for studies published before May 1, 2015: Pubmed, Excerpta Medica dataBASE (EMBASE), The Cochrane Library, EBSCO, Chinese National Knowledge Infrastructure (CNKI), Chongqing VIP database for Chinese Technical Periodicals, WANFANG DATA, Chinese Biological Medical Literature Database, Taiwan Electronic Periodical Services, and ClinicalTrials.gov. We used the keywords ‘refractory’, ‘schizophrenia’, ‘psychosis’, ‘electric shock’, ‘electroconvulsive’, ‘clinical control study’, ‘randomly, placebo’,and ‘randomly, trial’ (and the Chinese equivalents) in the searches. Various combinations of these keywords were used to search for articles, reference lists of included articles were hand-checked for further relevant studies, and experts in the field were asked about ongoing studies.

Inclusion and exclusion criteria

All reports of randomized controlled trails (RCTs) about the combined treatment of refractory schizophrenia with antipsychotic medications and ECT were screened using the following inclusion criteria: a) a diagnosis of refractory (or ‘treatment-resistant’) schizophrenia made by psychiatrists; b) the control group was treated with antipsychotic medications; and c) the intervention group was treated with antipsychotic medication and ECT. Studies published in either English or Chinese were considered. Observational studies, anthropologic studies, review articles, research protocols, case reports, and duplicated reports were excluded.

Screening of articles

All search results were imported into Endnote X5 software. Two authors (WWZ and PCC) independently screened titles and abstracts after eliminating duplicates. The full texts of the remaining articles were screened according to the above inclusion and exclusion criteria. When the two authors disagreed about the inclusion of an article and were unable to agree after discussing the article, a third author (LCB) made the final determination. As shown in Figure 1, 22 studies were included in the final analysis.

Evaluation of risk of bias

Two authors (WWZ and PCC) assessed the risk of bias independently for all included articles using the Cochrane Risk of Bias (RoB)[5] tool which considers seven specific items: random sequence generation; allocation concealment; blinding of participants and treating clinicians about group assignment; blinding of evaluators of outcomes about group assignment; incomplete data (attrition and exclusions); selective outcome reporting; and other biases (including study-specific biases or concerns about fraudulent results). Each aspect was rated as ‘low risk of bias’, ‘high risk of bias’, or ‘unclear’ (if insufficient information was provided in the article to make a determination). A third author’s (LCB) opinion was sought when the two raters disagreed.

We also evaluated the quality and level of evidence of each outcome variable for each of the 22 included studies using the Cochrane collaboration’s Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) software. This software assesses limitations of the design, consistency of results, indirect evidence, precision of results, publication bias, and effect size for each outcome.[5,6] The overall level of evidence is rated as ‘high’, ‘medium’, ‘low’, or ‘very low’.

Outcome measures

This meta-analysis was conducted to assess the efficacy and safety of the combination of antipsychotic medication and ECT in the treatment of patients with refractory schizophrenia. The primary outcome measure of effectiveness was the reduction in the total score of the main scale used to assess psychiatric symptoms during treatment. The primary outcome measure of adverse events was based on the score of the Treatment Emergent Symptom Scale (TESS). Secondary outcomes were the changes of cognitive and overall functioning after treatment.

Data extraction

For each included study, two authors (WWZ and PCC) independently extracted data using a pre-designed data extraction form including the names of authors, publication year, sample size, number of outcome events, age of participants, and types of antipsychotics used. Discrepancies between the two coders were checked by a third author (LCB).

Analysis

Based on the results of a previous study about risk of bias,[25] the overall risk of bias for each of the 22 studies was classified as ‘low’ if the ratings were ‘low’ for all seven items on the ROB tool, ‘unclear’ if any item is rated as ‘unclear’ and all other items are rated as ‘low’, and ‘high’ if any of the items are rated as ‘high’. The kappa statistic was used to measure the inter-rater agreement between the two independent raters for the ratings of each item and for the overall rating.[7] Review Manager (RevMan 5.3) was used to estimate pooled the mean difference (MD) when the same continuous measure was used as the outcome measure in all included studies, the standard mean difference (SMD) when different continuous measures were used as outcome measures in included studies,and risk ratios (RR) for outcomes that were categorical measures. Heterogeneity was measured using I2.[8] When I2 is less than 50% and p>0.10, the results were considered homogeneous and the fixed-effect model was used; when I2 is greater than 50% but less than 75%, results were considered heterogeneous and the random-effect model was used. If I2 is 75% or greater, we conducted sensitivity analysis to identify potential contributors to heterogeneity; if I2 remained 75% or greater after removing outliers, we only provided descriptive results without pooling estimates. A funnel plot was used to evaluate potential publication bias.[5]

Results

Characteristics of included studies

As shown in Figure 1, we identified a total of 1065 articles in the selected databases. After removing 166 duplicated articles using Endnote X5 software, there were 899 unduplicated reports. Based on the inclusion and exclusion criteria, 773 articles were excluded by reading the title and abstract and a further 104 were excluded by reading the full text. The remaining 22 articles[9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,26,27,28,29,30,31,32] were included in the subsequent analyses: 18 (81.8%) were from China [10,12,13,14,15,16,18,19,20,21,22,26,27,28,29,30,31,32] and 4 (18.2%) from other countries; [9,11,17,23] all were published between 1999 and 2015.

The characteristics of these 22 included studies are shown in Tables 1 and 2.

a) Definition of refractory schizophrenia: The inclusion and exclusion criteria varied across studies, but most studies used either the diagnostic criteria for schizophrenia established by the Chinese Society of Psychiatry (CCMD)[33] or the diagnostic criteria for schizophrenia developed by the American Psychiatric Association (DSM-IV),[34] required a duration of illness of at least 2 years (with one exception[9]), and required unsatisfactory clinical results when previously using at least two types of antipsychotic medications.

b) Gender of participants: One study[32] only included females, but all other studies[9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,26,27,28,29,30,31] included both males and females.

c) Age of participants: Three studies[9,17,27] did not describe the age of participants; the remaining 19 studies[10,11,12,13,14,15,16,18,19,20,21,22,23,26,28,29,30,31,32] were conducted with adults ranging from 18 to 74 years of age.

d) Duration of illness: Among the 17 studies that provided the mean duration of illness among participants,[9,11,12,13,15,17,18,19,22,23,26,27,28,29,30,31,32] the range in the mean duration of illness was from 6 to 21 years.

e) Type of antipsychotic medication used in study: All participants received antipsychotic medications either alone or in combination with ECT. The medications employed included clozapine, olanzapine, risperidone,quetiapine, ziprasidone, chlorpromazine,and flupenthixol.

f) ECT sessions: Two studies[9,17] did not provide information on the number of ECT sessions; the remaining studies[10,11,12,13,14,15,16,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32] reported using 6 to 24 ECT sessions. Among the 18 studies[10,11,12,13,14,15,16,18,20,21,22,23,26,27,28,29,30,31,32] that provided information on ECT frequency, the range in frequency was from three sessions per week to two sessions per month. Only eight studies[11,12,13,17,22,23,28,31] reported on the type of ECT stimulus,which was either bilateral or bitemporal.

g) Outcome measures: Standard measures of psychiatric symptoms,primarily the Positive and Negative Syndrome Scale (PANSS)[35] and the Brief Psychiatric Rating Scale (BPRS),[36] were used to assess effectiveness,though the criteria used to determine clinical improvement varied somewhat across studies. Overall adverse events were assessed using TESS[37] in six studies.[10,13,18,20,28,30] Five studies reported results of cognitive functioning: two studies[11,23] used the Mini-Mental State Examination (MMSE),[41] one study[18] used the Wisconsin Card Sorting Test (WCST),[42] and two studies[19,31] used the Wechsler Memory Scale (WMS).[43] Overall functioning was assessed using the Global Assessment of Functioning (GAF)[38] in three studies[11,18,28] and the Clinical Global Impression score (CGI) [44] in one study.[17]

h) Duration of follow-up: One study[26] did not provide information about the duration of follow-up; one study[14] had a 3 to 5 week follow-up; four studies[17,21,27,31] followed patients for 4 weeks; nine studies[9,10,15,19,20,22,23,30,32] followed patients for 8 weeks; six studies[12,13,16,18,28,29] followed patients for 12 weeks; and one study[11] followed patients for 24 weeks.

Risk of bias and publication bias

The results of the assessment of risk of bias in the 22 studies are shown in Table 3. Most of the studies, particularly those from China,did not provide sufficient details about the methods used in the study,so the majority of the assessments for the seven items included in the Risk of Bias (ROB) tool[5] were rated as ‘uncertain’. One study[17] used sham ECT and, thus, was double-blind; two other studies[11,23] used blinded outcome evaluators; and one study[23] reported that the participants and treating clinicians were not blinded; but none of the other studies provided information about blinding of the participant, the treating clinician, or the outcome evaluator. Only two studies[17,18] described the method of randomization. Overall, one study[23] was rated as ‘high’ risk of bias and the other studies were all rated as ‘uncertain’ risk of bias; none of the studies were classified as being at low-risk of bias. The inter-rater reliability of the two independent coders’ assessment of overall risk of bias in the studies was acceptable (kappa=0.75).

The funnel plot of the primary outcome of efficacy provided by 18 of the studies[9,10,12,13,14,15,16,18,19,20,22,23,26,27,29,30,31,32] is shown in Figure 2. There is substantial publication bias; the three smaller studies[9,19,23] were more likely to find a significant advantage for combined treatment with ECT plus antipsychotic medications versus the sole use of antipsychotic medications in the treatment of refractory schizophrenia.

As shown in Table 4, based on the GRADE criteria the overall quality of the evidence about effectiveness (defined as percent of patients who showed ‘improvement’ by the end of the trial) was rated as ‘moderate’, but the quality of evidence about general side effects (assessed using TESS), cognitive functioning, and overall functioning was rated as ‘very low’. The six studies that provided data about general side effects based on the TESS score[10,13,18,20,28,30] had very heterogeneous results and, thus, could not be pooled in a meta-analysis.

Efficacy

Eighteen studies with a pooled sample of 1394 participants provided information on efficacy (i.e., percent of cases that achieved the study-specific criteria for ‘improvement’) at the end of the trial.[9,10,12,13,14,15,16,18,19,20,22,23,26,27,29,30,31,32] The I2 was satisfactory but the corresponding p-value was too small to justify considering the results of the various studies homogeneous (I2=38%, p=0.05), so a random-effect model was conducted. The result, shown in Figure 3, indicate that patients with refractory schizophrenia treated with the combination of ECT and antipsychotic medications are more likely to experience symptomatic improvement than patients treated with antipsychotic medications alone (RR=1.25, 95% CI =1.14-1.37).

Adverse effects

As shown in Table 5, six studies with a pooled sample of 534 participants reported TESS scores.[10,13,18,20,28,30] The TESS results from these studies were quite heterogeneous, so it is only possible to provide a descriptive assessment of the results. Using the overall TESS score as the measure of the severity of adverse events, three of the studies[18,28,30] showed no significant differences between the two groups, one study[13] found significantly more severe adverse events in the ECT plus antipsychotic medication group, and two studies[10,20] found significantly more severe adverse events in the group that only received antipsychotic medication.

Thirteen studies[13,14,15,18,22,23,26,27,28,29,30,31,32] provided information about specific adverse events. As shown in Table 6, a total of 26 kinds of adverse events were reported. Meta-analysis of the results identified two adverse effects that were reported at a significantly higher frequency by patients receiving combined treatment with ECT plus antipsychotic medication than by patients treated with antipsychotic medications alone: headache (based on a pooled sample of 517 individuals from eight studies, OR=9.10, 95%CI=3.97-20.86) and memory impairment (based on a pooled sample of 577 individuals from seven studies, OR=6.48, 95%CI=3.54-11.87).

Cognitive function

Five studies reported cognitive function results,[11,18,19,23,31] but only four of them had data that could be pooled in a meta-analysis.[11,18,23,31] The meta-analysis (Figure 4) comparing the standardized mean difference (SMD) in the final scale scores between groups used the total MMSE score[41] for the two studies that used MMSE,[11,23] the perseverative errors score from the WCST[42] for the study that used the WCST,[18] and the picture recognition score from the WMS[43] for the study that use the WMS.[31] The results for the study that use the WCST and for the study that used the WMS indicate significantly greater cognitive impairment in the group treated with ECT and antipsychotic medications, but when pooled with results from the two studies that use MMSE to assess cognitive functioning, the overall results were not significantly different between the two groups (SMD=-0.28, 95%CI=-0.77~0.20). However, as stated previously (Table 4), the quality of evidence from these studies was rated as ‘very low’.

Overall function

The meta-analysis of the three studies[9,16,26] that used the GAF to assess overall functioning at the end of the trial is shown in Figure 5. (The study that employed CGI to assess overall functioning[17] could not be included because the report for the study did not include a standard deviation for the mean CGI score.) For each of the three studies and for the pooled result for the studies, the overall functioning at the end of the study was significantly better in the ECT plus antipsychotic medication group than in the group treated with antipsychotic medications alone (MD=10.25, 95%CI=7.48-13.01). However,the quality of evidence for these studies was rated as ‘very low’ based on the GRADE criteria.

Discussion

Main findings

Since the operational definition of ‘refractory schizophrenia’ was first proposed by Kane in 1995,[3] there has been increasing research about the identification and management of these hard-totreat patients. Previous meta-analyses conclude that clozapine can improve psychiatric symptoms and decrease the occurence of extrapyramidal side effects for about half of these refractory patients.[24] ECT is an option for increasing the proportion of refractory patients who achieve satisfactory clinical outcomes, but there are few studies addressing this issue, possibly because of the reluctance of patients and clinicians to use ECT.

Extensive screening of English-language and Chinese-language databases identified 22 RCTs on this topic, among which 4 were in English and 18 were in Chinese, suggesting that the use of ECT as an adjunctive treatment in refractory schizophrenia is less controversial in China than abroad. The pooled result about efficacy (defined as the proportion of participants that achieved the study-specific criteria for ‘improvement’) included 1394 individuals that participated in 18 studies. It clearly indicated a superior clinical result for refractory patients who received combined treatment with ECT and antipsychotic medications (RR=1.25, 95% CI=1.14~1.37). Moreover, the quality of the evidence for this outcome using the GRADE criteria was rated as ‘moderate’, which means that the result is reasonably robust. We also found that patients who received combined treatment were much more likely to report headaches and memory impairment during the treatment and to have higher overall functioning at the end of treatment, but the quality of evidence for these findings was rated as ‘very low’.

Limitations

The low quality of the RCTs identified for inclusion in this review seriously limits the validity of the results, particularly the results related to the occurrence of adverse effects, cognitive impairment, and overall functioning. As shown in our assessment of the risk of bias, the main problem was that most of the reports on the included studies did not provide details about the methods employed in the study, resulting in an overall rating of the assessment of risk of bias of ‘uncertain’ in 21 of the 22 included studies.

The most troubling problem was lack of blinding. Double blinding is difficult to achieve in studies involving ECT because of the practical and ethical problems of administering ‘sham ECT’, so it is not surprising that sham ECT was only used in 1 of the 22 included studies. On the other hand,it is relatively easy to blind raters who assess the outcome measures,but this was only done in 3 of the 22 studies. We also identified a clear publication bias, favoring publication of studies that find superior outcomes among patients receiving both ECT and antipsychotic medication.

Implication

The current study systematically reviewed and evaluated all available RCTs about the efficacy and safety of the combined use of antipsychotics and ECT in the treatment of patients with refractory schizophrenia. Meta-analysis of the 22 identified studies found that compared to treatment with antipsychotic medications alone,combined treatment with ECT and antipsychotic medication resulted in improved clinical outcomes, higher rates of headaches and memory impairment during treatment, and better overall functioning at the end of treatment. However, given the low quality of the evidence in the studies and the existence of publication bias,these results must be considered suggestive,not definitive. Better designed studies that include more detailed description of the methods employed are needed to confirm (or disprove) these results.

Studies about the use of ECT in schizophrenia are relatively common in China, so this is one area in which China could make an important contribution to the global literature. However, Chinese mental health researchers will not be able to make useful international contributions until they resolve common problems in the design and implementation of their studies, problems that could be easily addressed by strict adherence to the recommendations for RCTs in the CONSORT statement.[5]

Table 1.

Characteristics of 22 randomized controlled trials that compare treatment of refractory schizophrenia with combined ECT and antipsychotic medication versus treatment with antipsychotic medication alone

study	treatmentECT group/control group	ageECT group/control group	sample sizeECT group/control group	trial duration(weeks)	number of ECT sessions	blind evaluation	outcome measures(change required for ‘improvement’)
Chanpatana 1999[11]	ECT+flupenthixol/flupenthixol	20-49	17/18	24	14	yes	BPRS; GAF; MMSE
Goswami 2003[17]	ECT+chlorpromazine/sham ECT+chlorpromazine	NA	15/10	4	NA	yes	BPRS(20%) and CGI≤3 or BPRS≤35
Yang 2005[27]	ECT+clozapine/clozapine	NA	30/30	4	6-12	NA	PANSS(20%)
Ding 2007[13]	ECT+risperidone/risperidone	36.5/38.7	30/30	12	≥12	NA	PANSS(20%); TESS
Cai 2008[10]	ECT+clozapine/clozapine	18-60	50/50	8	6-12	NA	BPRS(25%); TESS
Braga 2009[9]	ECT+clozapine/clozapine	NA	21/17	8	NA	NA	BPRS(20%)
Jiang 2009[18]	ECT+risperidone/risperidone	38.3/39.7	34/35	12	8-12	NA	PANSS(50%); TESS; WCST; GAF
Zhou 2009[31]	ECT+olanzepine/olanzepine	43.1/42.2	31 /32	4	8-12	NA	PANSS(25%); WMS
liu 2010a[21]	ETC+various antipsychotics/various antipsychotics	38.4/39.4	37/33	4	12	NA	SANS(50%); SAPS(50%)
Liu 2010b[22]	ECT+risperidone/clozapine	28.6/29.6	30/30	8	≥12	NA	PANSS(25%)
Ding 2011[14]	ETC+various antipsychotics/various antipsychotics	29.8/31.5	100/100	3-5	9-15	NA	PANSS(50%)
Du 2011[15]	ECT+clozapine/clozapine	38.6	30/30	8	10	NA	BPRS (25%)
Duo 2003[16]	ECT+olanzepine/olanzepine	43.8/42.7	30/30	12	8-12	NA	PANSS (25%)
Jiang 2011[19]	ECT+various antipsychotics/various antipsychotics	43.4	23/23	8	17-24	NA	PANSS (50%); WMS
Yang 2011[28]	ECT+risperidone/risperidone	18-65	35/36	12	12	NA	PANSS (25%); GAF; TESS
Chen 2012[29]	ECT+clozapine/clozapine	31.9/33.6	36/35	12	7-12	NA	PANSS (25%)
Wang 2012[32]	ECT+quetapine/quetapine	28.9/29.3	31/31	8	10	NA	PANSS (25%)
Zhang 2012[30]	ECT+olanzepine/olanzepine	38.4	42/42	8	16	NA	PANSS (25%); TESS
Chen 2013[12]	ECT+various antipsychotics/various antipsychotics	18-60	50/40	12	8-12	NA	PANSS (25%)
Jiang 2013[20]	ECT+ziprasidone/ziprasidone+clozapine	21-74	81/81	8	12	NA	BPRS (25%); TESS
Wang 2013[26]	ECT+olanzepine/olanzepine	45.5	36/36	NA	10-12	NA	PANSS (25%)
Petrides 2015[23]	ECT+clozapine/clozapine	18-60	20/19	8	20	yes	BPRS (40%); MMSE

ECT, electroconvulsive therapy

NA, data not available

BPRS, Brief Psychiatric Rating Scale[36]

GAF, Global Assessment of Functioning[38]

MMSE, Mini-Mental Status Exam[41]

PANSS,Positive and Negative Syndrome Scale[35]

CGI, Clinical Global Impression[44]

TESS, Treatment Emergent Symptoms Scale[37]

WCST,Wisconsin Card Sort Test[42]

WMS, Weschler Memory Scale[43]

SANS, Scale for Assessment of Negative Symptoms[40]

SAPS, Scale for Assessment of Positive Symptoms[39]

a only had female participants

Table 2.

Supplemental information for the 22 studies included in the systematic review

study	definition of ‘treatment refractory schizophrenia’					mean (sd) years of illness ECT group /control group	type of ECT stimulus	ECT frequency
	diagnostic frequency criteria	number of failed drug trials	chlorpromazine equivalent dose (mg/d)	duration of illness	severity of illness
Chanpatana1999[11]	NA	≥2	>750	≥2 years	BPRS>35	13.7 (5.5)/14.2 (6.4)	bilateral	2-4/month
Goswami 2003[17]	DSM-IV	≥3	>1000	≥5 years	NA	7.6/6.9	bitemporal	NA
Yang 2005[27]	CCMD-3	≥3	NA	≥5 years	PANSS>60	6.3 (4.3)/6.0 (4.9)	NA	3/week
Ding 2007[13]	CCMD-3	≥3	>450	≥5 years	PANSS>65	9.7 (11.0)/9.7 (10.4)	bitemporal	1-2/week
Cai 2008[10]	CCMD-3	≥3	>600	NA	BPRS>50	1-10a	NA	3/week
Braga 2009[9]	NA	NA	>250	≥8 weeks	NA	NA	NA	NA
Jiang 2009[18]	CCMD-3	≥3	NA	≥5 years	PANSS≥60	12.6 (5.1)/12.4 (5.0)	NA	2-3/week
Zhou 2009[31]	CCMD-3	≥3	>600	≥5 years	PANSS≥60	21.0 (7.8)/19.4 (9.9)	bitemporal	2-3/week
Liu 2010a[21]	CCMD-3	NA	NA	NA	NA	NA	NA	3/week
Liu 2010b[22]	CCMD-3	≥3	NA	≥5 years	PANSS≥60	8.3 (4.2)/8.0 (3.2)	bitemporal	1-3/week
Ding 2011[14]	CCMD-3	≥3	NA	NA	NA	NA	NA	3/week
Du 2011[15]	CCMD-3	≥3	>1000	≥5 years	BPRS≥45	10.0 (4.0)	NA	NA
Duo 2011[16]	NA	NA	NA	NA	NA	17.5 (5.6)/18.6 (4.2)	NA	2-3/week
Jiang 2011[19]	CCMD-3	≥2	NA	≥2 years	PANSS≥60	7.5 (7.6)	NA	NA
Yang 2011[28]	CCMD-3	≥3	NA	≥5 years	PANSS≥60,	11.2 (5.2)/10.3 (4.9)	bitemporal	2-3/week
Chen 2012[29]	CCMD-3	≥2	>600	≥5 years	PANSS>60	16.1 (11.6)/15.8 (11.2)	NA	2-3/week
Wang 2012[32]	CCMD-3	≥2	>600	≥5 years	PANSS≥60	8.7	NA	3/week
Zhang 2012[30]	CCMD-3	≥3	NA	≥5 years	PANSS>60	NA	NA	1-3/week
Chen 2013[12]	CCMD-3	≥3	NA	≥5 years	PANSS>60	9.0 (1.7)/8.5 (0.9)	bitemporal	2-3/week
Jiang 2013[20]	CCMD-2-R	≥3	NA	NA	NA	7.9	NA	1-3/week
Wang 2013[26]	CCMD-3	≥3	NA	≥5 years	NA	13.2 (5.2)	NA	3/week
Petrides 2015[23]	DSM-IV	≥2	>600	≥2 years	BPRS; CGIb	NA	bilateral	2-3/week

ECT, electroconvulsive therapy

NA, data not available

BPRS, Brief Psychiatric Rating Scale[36]

PANSS, Positive and Negative Syndrome Scale[35]

CGI, Clinical Global Impression[44]

DSM-IV, 4th edition of Diagnostic and Statistical Manual for mental disorders[34]

CCMD-3, 3rd edition of Chinese Classification of Mental Disorders[33]

CCMD-2-R, Revised 2nd edition of Chinese Classification of Mental Disorders[33]

a range (not mean) in duration of illness

b one of the four psychotic items of BPRS≥4 or total score≥12; and CGI≥4

Table 3.

Evaluation of risk of bias in the 22 included studies based on the seven items in the Cochrane Risk of Bias (RoB) tool

study	random sequence generation	allocation concealment	blinding of participants and providers	blinding of outcome assessment	incomplete outcome data	selective reporting	other biasesa	overall risk of biasb
Chanpatana 1999[11]	Unclear	Low	Unclear	Low	Low	Unclear	Unclear	Unclear
Goswami 2003[17]	Low	Low	Low	Low	Low	Unclear	Unclear	Unclear
Yang 2005[27]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Ding 2007[13]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Cai 2008[10]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Braga 2009[9]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Jiang 2009[18]	Low	Unclear	Unclear	Unclear	Low	Unclear	Unclear	Unclear
Zhou 2009[31]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Liu 2010a[21]	Unclear	Unclear	Unclear	Unclear	Low	Unclear	Unclear	Unclear
Liu 2010b[22]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Ding 2011[14]	Unclear	Unclear	Unclear	Unclear	Low	Unclear	Unclear	Unclear
Du 2011[15]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Duo 2011[16]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Jiang 2011[19]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Yang 2011[28]	Unclear	Unclear	Unclear	Unclear	Low	Unclear	Unclear	Unclear
Chen 2012[29]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Wang 2012[32]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Zhang 2012[30]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Chen 2013[12	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Jiang 2013[20]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Wang 2013[26]	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear	Unclear
Petrides 2015[23]	Unclear	Unclear	High	Low	Low	Unclear	Unclear	High
kappac	1.00	1.00	0.49	0.82	1.00	1.00	1.00	0.75

a Other biases considered include including study-specific biases or concerns about fraudulent results

b If any of seven items are coded high-risk of bias the overall study is classified as high-risk, if all seven items are coded as low-risk the overall study is classified as low-risk; all other studies (i.e., those with some items coded ‘unclear’ and no items coded as high-risk) are classified as ‘unclear’

c Weighted kappa values for inter-rater reliability of the two independent coders who assessed each item for the 23 studies

Table 4.

Summary of meta-analysis and GRADE assessments of quality of data about different outcome measures of randomized compared trials comparing ECT and antipsychotic medication versus medication alone in the treatment of treatment refractory schizophrenia

outcomes	number of studies (pooled sample)	test for heterogeneity		analytic model	test for overall effect		estimate	95% confidence interval of estimate	GRADE
		I2	p		Z	p
effectiveness	18 (1394)	38%	0.05	fixed	7.51	<0.001	1.31 (RR)	1.22-1.41	moderate
side effect (TESS)	6 (534)	81%	< 0.01	-	-	-	-	-	very low
cognitive function	4 (199)	65%	0.04	random	1.14	0.25	-0.28 (SMD)	-0.77-0.20	very low
overall function	3 (157)	48%	0.14	fixed	7.26	<0.01	10.25 (MD)	7.48-13.01	very low

GRADE, Grades of Recommendation, Assessment, Development, and Evaluation

ECT, electroconvulsive therapy

TESS, Treatment Emergent Symptom Scale[37]

RR, risk ratio

SMD, standardized mean difference

MD, mean difference

Table 5.

Results for six studies that used the Treatment Emergent Symptom Scale (TESS) to assess adverse events in patients with refractory schizophrenia treated with combined electroconvulsive therapy (ECT) and antipsychotic medication versus those treated with antipsychotic medication alone

study	ECT + medication group			medication only group		mean difference [95% CI]
	n	mean (sd)		n	mean (sd)
Ding, 2007[13]	30	2.7 (2.4)		30	1.3 (1.2)	1.44 [0.47,2.41]
Cai 2008[10]	50	2.9 (1.6)		50	3.7 (2.0)	-0.76 [-1.47,-0.05]
Jiang 2009[18]	32	7.7 (5.9)		35	6.8 (5.4)	0.95 [-1.78,3.68]
Yang 2011[28]	30	8.2 (6.1)		31	7.9 (5.9)	0.34 [-2.69,3.37]
Zhang 2012[30]	42	5.9 (2.7)		42	6.9 (2.7)	-0.98 [-2.13,0.17]
Jiang 2013[20]	81	3.4 (0.8)		81	4.5 (1.1)	-1.10 [-1.40,-0.80]

Table 6.

Comparison of incidence of different adverse events in patients with refractory schizophrenia treated with electroconvulsive therapy and antipsychotic medications versus those treated only with antipsychotic medications

events	number of studies	participants	I2	p	statistical method (risk ratio)	effect estimate [95% C]
abnormal electrocardiogram	6	383	32%	0.20	fixed	0.83 [0.58,1.19]
abnormal electroencephalogram	2	123	85%	0.01	random	1.97 [0.31,12.65]
abnormal liver enzymes	7	448	0%	0.80	fixed	0.67 [0.34,1.34]
akathisia	4	263	0%	0.80	fixed	0.69 [0.37,1.26]
anorexia/decreased appetite	1	60	not applicable		fixed	2.00 [0.19,20.90]
blurred vision	3	203	0%	0.95	fixed	0.92 [0.37,2.26]
constipation	6	388	42%	0.13	fixed	0.52 [0.37,0.73]
decreased motor activity	2	140	71%	0.06	random	0.22 [0.02,3.10]
dizziness	5	323	51%	0.08	random	1.23 [0.56,2.70]
drowsiness	7	457	61%	0.02	random	0.71 [0.35,1.42]
dry mouth	3	203	0%	0.85	fixed	1.03 [0.49,2.18]
Extrapyramidal symptoms	3	182	0%	0.40	fixed	1.74 [0.84,3.58]
headache	8	517	7%	0.37	fixed	9.10 [3.97,20.86]
hypotension	1	60	not applicable		fixed	0.33 [0.01,7.87]
increased salivation	5	328	54%	0.07	random	0.41[0.18,0.93]
insomnia	5	322	59%	0.05	random	0.46 [0.07,3.22]
leukopenia	5	336	0%	0.49	fixed	0.31 [0.12,0.82]
memory impairment	7	577	36%	0.16	fixed	6.48 [3.54,11.87]
menstrual disorder	1	60	not applicable		fixed	1.50 [0.27,8.34]
nasal congestion	1	63	not applicable		fixed	3.09 [0.13,73.17]
nausea/vomiting	5	334	0%	0.94	fixed	2.33 [0.99,5.49]
rigidity	2	140	0%	0.65	fixed	0.69 [0.40,1.18]
tachycardia	4	268	0%	0.58	fixed	0.77 [0.50,1.21]
tremors	2	140	0%	0.82	fixed	0.77 [0.43,1.38]
weakness	3	185	0%	0.43	fixed	0.25 [0.11,0.59]
weight gain	7	468	4%	0.40	fixed	0.61 [0.43,0.87]