Combined ECT and neuroleptic therapy in treatment-refractory schizophrenia: prediction of outcome

Psychiatry Research 105 Ž2001. 107᎐115

Combined ECT and neuroleptic therapy in treatment-refractory schizophrenia: prediction of outcome 夽 Worrawat Chanpattana a,U , M.L. Somchai Chakrabhand b a

Department of Psychiatry, Srinakharinwirot Uni¨ ersity, 681 Samsen, Dusit, Bangkok 10300, Thailand b Department of Mental Health, Ministry of Health, Nonthaburi, Thailand Received 5 December 2000; received in revised form 5 June 2001; accepted 21 June 2001

Abstract There have been scanty reports of the clinical features of schizophrenic patients treated with electroconvulsive therapy ŽECT.. This prospective study examined clinical characteristics and predictive factors associated with therapeutic outcome. Two hundred and ninety-three patients with refractory schizophrenia were treated with a combination of ECT and flupenthixol. Outcome assessments included the Brief Psychiatric Rating Scale ŽBPRS., the Global Assessment of Functioning ŽGAF., and the Mini-Mental State Examination ŽMMSE.. One hundred and sixty patients Ž54.6%. met a response criterion. The responders were younger, had shorter durations of illness and current episode, more admissions, and less family history of schizophrenia. The duration of current episode Ž t s 5.0, P- 0.0001., followed by baseline GAF score Ž t s 3.1, P s 0.002., duration of illness Ž t s 3.1, Ps 0.002., baseline MMSE score Ž t s 3.0, P s 0.003., duration of the previously failed neuroleptic trials Ž t s 3.0, P s 0.003., family history of schizophrenia Ž t s 2.1, P s 0.03., and paranoid type Ž t s 2.1, Ps 0.04., could predict the therapeutic outcome. Treatment resulted in marked improvement in positive symptoms but had a minimal effect or led to a worsening of negative symptoms. 䊚 2001 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Electroconvulsive therapy ŽECT.; Schizophrenia; Predictive factors of response; Nature of clinical response; Negative symptoms

夽 Presented at the 154th American Psychiatric Association Annual Meeting, May 8, 2001, Convention Center, New Orleans, Louisiana. U Corresponding author. Tel.: q66-2-292-2192; fax: q66-2-292-2211. E-mail address: [email protected] ŽW. Chanpattana..

0165-1781r01r$ - see front matter 䊚 2001 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 5 - 1 7 8 1 Ž 0 1 . 0 0 3 2 1 - 3

108

W. Chanpattana, M.L. Somchai Chakrabhand r Psychiatry Research 105 (2001) 107᎐115

1. Introduction Electroconvulsive therapy ŽECT. has been used to treat schizophrenia since its inception in 1938 ŽCerletti and Bini, 1938.. Practice surveys indicate that schizophrenia is the second most common diagnostic indication for the use of ECT in the United States ŽThompson and Blaine, 1987; Thompson et al., 1994., and administration of ECT in schizophrenia is substantially greater in Canada, many European and developing countries ŽPippard and Ellam, 1981; Latey and Fahy, 1985; Malla, 1986; Agarwal et al., 1992; Baudis, 1992.. Nonetheless, there is considerable uncertainty about the role of ECT in this disorder, and virtually no information on the patient characteristics associated with clinical outcome ŽKrueger and Sackeim, 1995.. Indeed, the literature that directly assessed the predictive factors of ECT outcome in schizophrenia consists of a single study of a mixed sample of 17 patients with schizophrenia or schizoaffective disorder ŽDodwell and Goldberg, 1989.. We determined the clinical characteristics and predictive factors associated with response in patients with refractory schizophrenia.

2. Methods 2.1. Study design A large sample of refractory schizophrenic patients was prospectively treated with the combination of ECT and a conventional neuroleptic. This study did not compare the efficacy of this combination treatment relative to pharmacotherapy; rather, among this patient group, we examined the clinical characteristics and predictive factors associated with therapeutic outcome. 2.2. Subjects Adapting the criteria used in the original clozapine trial ŽKane et al., 1988., we identified 293 refractory schizophrenic patients. We defined those with refractory schizophrenia as patients for whom at least two distinct periods of neuroleptic

treatment from at least two different classes had failed, with a duration of treatment of at least 6 weeks at neuroleptic dosages of at least 750 mgr day chlorpromazine equivalents ŽCPZE., that did not result in significant symptomatic relief. Duration of illness had to be more than 2 years. Diagnosis was based on DSM-IV ŽAmerican Psychiatric Association, 1994. determined by consensus of three psychiatrists on the diagnosis and subtype of schizophrenia, as well as concurrence for the diagnosis with the patient’s past medical records. Other inclusion criteria required a baseline Brief Psychiatric Rating Scale ŽBPRS. score ) 37 Ž18 items, rated 0᎐6; Overall and Gorham, 1962., patient in an acute psychotic exacerbation, and age between 16 and 50 years. Patients were excluded if they received ECT or depot neuroleptics within 6 months, had alcohol or substance abuse, or had serious medical conditions, as assessed by history, physical examination, and pertinent laboratory tests ŽCBC, electrolytes, ECG.. After a detailed explanation, each subject andror guardian gave written informed consent for ECT and for study participation. Patients were recruited from two university hospitals and two state mental hospitals from 1 October 1995 to 31 March 2000; 253 Ž86.3%. patients were treated on an inpatient basis, and 40 Ž13.7%. patients were outpatients. This study was approved by the Ethics Committee at each participating hospital, and the National Review Board of Research Studies in Humans of Thailand. 2.3. Treatment All patients were free of medicines beginning 5 days prior to the first ECT. Flupenthixol was prescribed without a washout, 12 mgrday Ž; 600 mg CPZE. during the first week and then increased to 24 mgrday Ž; 1200 mg CPZE. depending on tolerability. All patients received benzhexol Ž4᎐15 mgrday. to control extrapyramidal symptoms, and in 48 patients diazepam, 5᎐20 mgrday Žmaximal three doses during study., was used for agitation. No other medications were prescribed. ECT was administered three times per week with either an MECTA SR1 Ž n s 179. or a Thy-

W. Chanpattana, M.L. Somchai Chakrabhand r Psychiatry Research 105 (2001) 107᎐115

matron DGx Ž n s 114. device. Thiopental Ž2᎐4 mgrkg. was used as an anesthetic agent and succinylcholine Ž0.5᎐1 mgrkg. as a muscle relaxant. Bitemporal bilateral electrode placement was used exclusively. Seizure thresholds were estimated at the first and the last treatment sessions, using the empirical titration technique. An adequate seizure was defined as at least 30 s of tonic᎐clonic motor activity with electroencephalographic ŽEEG. evidence of a seizure. Electrical dose was maintained at the threshold level unless short seizures occurred; then, charge was increased by 50%. 2.4. Response criterion A 3-week stabilization period was used as a response criterion, and to limit the number of ECT treatments in the responders ŽChanpattana et al., 1999a,b,c.. Patients who achieved a BPRS score of 25 or less during the regular ECT course entered a 3-week stabilization period. During this period, ECT was given three times per week during the first week, and then once weekly during the second and third weeks. If the BPRS score was greater than 25 during this period and the total number of ECT treatments was less than 20, the patient returned to the ECT schedule of three treatments per week. Patients who received 20 ECT treatments and had final BPRS scores greater than 25 were classified as non-responders. Responders were restricted to patients who completed the 3-week stabilization period, during which the BPRS score assessed prior to each treatment was always 25 or less. 2.5. Measures 2.5.1. BPRS ratings BPRS ratings were assessed just prior to the start of ECT, prior to each treatment, and 1 week after the last treatment Žstudy end.. Only the baseline and final BPRS ratings were examined. BPRS measures included the total BPRS score; scores for each of the five traditional subscales of thought disturbance, anxiety᎐depression, with-

109

drawal᎐retardation Žor a cluster of negative symptoms., activation, and hostility᎐suspiciousness; a positive-symptoms score based on the items of conceptual disorganization, suspiciousness, hallucinations, and unusual thought content; and a negative-symptoms score based on the items of emotional withdrawal, motor retardation, blunted affect, and disorientation. 2.5.2. Other ratings The Global Assessment of Functioning ŽGAF. was assessed at study entry and end, and the Mini-Mental State Examination Ž MMSE, Kongsakon and Vanichtanom, 1994. was assessed at the same time as the BPRS. 2.5.3. Reliability Ratings were provided by five nurses who had undergone extensive training in the use of the three instruments prior to study participation. Inter-rater reliability was assessed every 3 months and intraclass correlation coefficients across the raters at the multiple assessments ranged from 0.83 to 0.97. 2.6. Statistics Analyses were performed using SPSS version 10.0. The significance level was set at P- 0.05, two-tailed. The general improvement Žpre-treatment vs. post-treatment. was tested by paired t-tests. Differences between groups on single, continuous variables were evaluated with t-tests or analysis of variance ŽANOVA.. For discontinuous data, ␹ 2 tests were used to test for significant differences among groups. When the sample size was small, Fisher’s exact test was used. Correlations between the therapeutic outcome and other variables were examined with non-parametric tests using Spearman’s coefficient. The degree to which variables could predict the therapeutic outcome was examined in a stepwise multiple regression analysis. Seizure threshold data were logarithmically transformed to achieve a normal distribution.

110

W. Chanpattana, M.L. Somchai Chakrabhand r Psychiatry Research 105 (2001) 107᎐115

3. Results 3.1. Treatment ¨ ariables As shown in Table 1, 160 of 293 patients Ž54.6%. met our response criteria. t-Tests indicated that responders received markedly fewer treatments, compared to non-responders. There were no gender differences in initial seizure threshold across both the total sample wwomen: 93.6" 32.3 millicoulombs ŽmC. Ž n s 161., men: 94.2" 35.3 Ž n s 132.; t 291 s 0.26, Ps 0.8x and subgroups of responders wwomen: 89.1" 35.8 mC Ž n s 87., men: 92.2" 37.0 mC Ž n s 73.; t 158 s 1.63, Ps 0.11x and non-responders wwomen: 97.6" 27.3 mC Ž n s 74., men: 90.6" 33.0 mC Ž n s 59.; t 131 s 1.67, Ps 0.1x. Initial seizure threshold estimated with the MECTA Ž97.1" 34.4 mC. was higher than with the Thymatron device Ž88.6" 32.2 mC; t 291 s 2.13, Ps 0.034.. 3.2. Relations of demographic and clinical ¨ ariables to treatment response As depicted in Table 2, t-tests indicated that responders were younger, had shorter durations of illness, shorter durations of current episode, and a history of more psychiatric admissions. Re-

sponders were less likely to have a positive family history of schizophrenia in first-degree relatives, and were more likely to be paranoid type, compared to non-responders. In addition, responders had higher baseline scores on the GAF and MMSE, but had lower baseline negative symptoms than non-responders. Two hundred and sixty-three patients Ž89.8%. had a rise in seizure thresholds. The overall increase in seizure threshold was 202.8%" 183.3% Žrange: 0᎐900%; t 1292 s 29.48, P- 0.0001.. The rise in seizure threshold of the responders Ž159.6%" 156.4%. was less than that of the nonresponders Ž254.8%" 199.6%; t 291 s 4.48, P0.0001.. There was no gender difference in seizure threshold increases across both the total sample and subgroups of responders and non-responders. There was no significant difference in the threshold increase between patients treated with the MECTA and Thymatron devices Ž215.3% " 197.1% vs. 183.3%" 158%; t 291 s 1.53, Ps 0.13.. Analyses of covariance ŽANCOVA. were conducted on the threshold increase using number of treatments and initial threshold as the covariates, and therapeutic response as a between-subjects factor. The result indicates that a smaller increase in threshold in the responders appears to be related to fewer ECT treatments Ž F3,288 s

Table 1 Treatment parameters as a function of treatment response Variable

Dosage of flupenthixol Žmgrday. Number of ECT treatments Thiopental Žmg. Succinylcholine Žmg. Seizure threshold, initial ŽmC. Seizure threshold, last ŽmC. Threshold increase Ž%. ECT devices, n Ž%.b MECTA SR1 Thymatron DGx

Responders Ž n s 160.

Non-responders Ž n s 133.

Mean ŽS.D..

Mean ŽS.D..

21.7 Ž3.8. 12.5 Ž4.2. 147.6 Ž30.8. 26.2 Ž9.3. 93.3 Ž36.5. 221.9 Ž129.3. 159.6 Ž156.4.

22.2 Ž3.0. 20.2 Ž0.6. 147.8 Ž21.4. 25.1 Ž6.5. 94.5 Ž30.0. 289.8 Ž129.6. 254.8 Ž199.6.

102 Ž63.8%. 58 Ž36.2%.

77 Ž57.9%. 56 Ž42.1%.

ta

P

1.49 22.84 0.55 1.14 0.85 5.62 4.48

0.14 - 0.0001 0.96 0.26 0.4 - 0.0001 - 0.0001

Abbre¨ iation: ECT, electroconvulsive therapy; mC, millicoulombs. a d.f.s 291. b There was no difference in the percentage of patients who received treatments with either the MECTA SR1 or the Thymatron DGx: ␹ 2 Ž1. s 1.05, Ps 0.31.

W. Chanpattana, M.L. Somchai Chakrabhand r Psychiatry Research 105 (2001) 107᎐115

111

Table 2 Subject characteristics as a function of response status Characteristic

Age Žyears. Education Žyears. Onset of illness Žyears. Duration of illness Žyears. Duration of current episode Žyears. Psychiatric admissions Žnumber. Previously failed neuroleptic trials Duration of Žfailed. neuroleptic trials Žmonths. Average dosage of failed neuroleptic trials Žmg CPZE. BPRS GAF MMSE

Sex Žmale. Family history of schizophrenia Prior failure with flupenthixol History of ECT Diagnostic subtype Žparanoid.

ta

P

37.1 Ž7.7. 8.7 Ž3.0. 20.4 Ž4.4. 16.6 Ž7.9. 4.8 Ž4.3. 6.0 Ž4.5. 3.5 Ž1.2.

5.87 1.25 0.75 6.93 8.60 2.24 0.72

- 0.0001 0.21 0.45 - 0.0001 - 0.0001 0.026 0.47

22.3 Ž21.8.

1.49

0.14

Responders Ž n s 160.

Non-responders Ž n s 133.

Mean ŽS.D..

Mean ŽS.D..

31.9 Ž7.5. 9.2 Ž3.6. 20.8 Ž5.1. 11.0 Ž5.5. 1.4 Ž1.6. 7.3 Ž6.0. 3.4 Ž1.2. 18.8 Ž18.7. 1240 Ž293. 46.8 Ž8.3. 33.1 Ž5.7. 24.7 Ž4.4.

1281 Ž278. 48.2 Ž7.0. 27.6 Ž4.9. 20.8 Ž5.7.

1.25 1.58 8.90 6.46

0.21 0.11 - 0.0001 - 0.0001

N Ž%. 73 Ž45.6. 16 Ž10. 59 Ž36.9. 111 Ž69.4. 113 Ž70.6.

N Ž%. 59 Ž44.4. 26 Ž19.5. 38 Ž28.6. 95 Ž71.4. 68 Ž51.1.

␹2b 0.05 5.39 2.26 0.15 11.69

P 0.83 0.02 0.13 0.70 0.001

Abbre¨ iation: CPZE, chlorpromazine-equivalent dose Žmg.; BPRS, Brief Psychiatric Rating Scale; GAF, Global Assessment of Functioning; MMSE, Mini-Mental State Examination; ECT, electroconvulsive therapy. a d.f.s 291. b d.f.s 1.

23.45, P- 0.0001.. Thus, threshold factor had been deleted from our regression model Žsee below.. 3.3. Predictor ¨ ariables for therapeutic outcome The non-parametric correlation tests revealed that therapeutic outcome was positively related to duration of the current episode Ž r s 0.55, P0.0001; responder s 1, non-responder s 2., duration of illness Ž r s 0.35, P- 0.0001., age Ž r s 0.33, P- 0.0001., negative symptoms Ž r s 0.21, P0.0001., family history of schizophrenia Ž r s 0.14, Ps 0.02., baseline BPRS Ž r s 0.13, Ps 0.03., and duration of failed neuroleptic trials Ž r s 0.12, Ps 0.037.. Therapeutic outcome was negatively related to baseline GAF Ž r s y0.48, P- 0.0001., baseline MMSE Ž r s y0.34, P- 0.0001., and

paranoid type Ž r s y0.20, Ps 0.001.. Table 3 presents a summary of the stepwise multiple regression analysis. Duration of the current episode was the most potent predictor of therapeutic outcome in the total sample. This regression model accounted for 36.1% of the variance that could predict the therapeutic outcome Ž F7,285 s 24.62, P- 0.0001., 63.9% remaining unexplained. 3.4. Profile of symptom change and therapeutic outcome 3.4.1. Baseline symptom profiles and therapeutic outcome BPRS ratings at baseline and study end are shown in Table 4. The non-responders had higher scores of baseline negative symptoms than the responders Ž t 291 s 3.91, P- 0.0001.; the differ-

112

W. Chanpattana, M.L. Somchai Chakrabhand r Psychiatry Research 105 (2001) 107᎐115

Table 3 Stepwise multiple regression analysis: prediction of response status a Variable

t

P

Duration of the current episode Žyears. Baseline GAF score Duration of illness Žyears. Baseline MMSE score Duration of Žfailed. neuroleptic trials Žmonths. Family history of schizophrenia Diagnostic subtype Žparanoid.

5.01 3.13 3.07 3.00 2.97

- 0.0001 0.002 0.002 0.003 0.003

2.14 2.10

0.033 0.037

Abbre¨ iations: GAF, Global Assessment of Functioning; MMSE, Mini-Mental State Examination. a F7,285 s 24.62, P- 0.0001.

ences were most marked in disorientation Ž t 291 s 5.14, P- 0.0001. and emotional withdrawal Ž t 291 s 3.68, P- 0.0001., and were modest in blunted affect Ž t 291 s 1.93, Ps 0.054.. No difference in the baseline positive symptoms between responders and non-responders was found Ž t 291 s 0.001, Ps 1.0.. 3.4.2. Extent and profile of symptom change In the total sample, treatment had no effect on negative symptoms, compared with positive symptoms and other subscales. In the non-responders, although treatment resulted in worsening of negative symptoms Ž10.3" 4.2 vs. 11.6" 3.9; t 1,132 s y3.93, P- 0.0001., the benefits from ECT were still evident, as manifested by marked reductions in positive symptoms Ž15.8" 3.9 vs. 11.5" 3.1;

t 1,132 s 11.96, P - 0.0001. and increased GAF scores Ž27.6" 4.9 vs. 32.6" 5.7; t 1,132 s 12.94, P0.0001. at the end of the study.

4. Discussion In our large sample of patients with treatmentrefractory schizophrenia, the rate of response to combined treatment with ECT and flupenthixol was as impressive Ž54.6%. as that observed in controlled studies of clozapine or other atypical neuroleptics ŽKane et al., 1988; Kane, 1999; Wirshing et al., 1999.. Nonetheless, the short-term response rate observed in our study could have been influenced by its uncontrolled nature and the requirement that all patients have acute exacerbations. Other limitations of this study include that the Structured Clinical Interview for DSM-IV was not used, and that all ratings were made on an open basis, which could have introduced a bias in the assessments. However, all raters were instructed to provide honest, accurate ratings and had no known proclivity to favor either ECT or flupenthixol treatments. Furthermore, all raters were not likely to entertain biases about clinical or demographic variables such as illness duration, episode duration, and family history. 4.1. Clinical characteristics of responders The use of a 3-week stabilization period in this study may have enhanced reliability in identifying

Table 4 BPRS profile at baseline and at study end BPRS profile

Thought disturbance Anxiety-depression Withdrawal-retardationb Activation Hostility᎐suspiciousness Positive symptoms Negative symptoms

At baseline

At study end

Mean ŽS.D..

Mean ŽS.D..

12.6 Ž3.2. 3.9 Ž3.7. 9.3 Ž4.3. 11.4 Ž2.9. 10.1 Ž3.8. 15.8 Ž3.6. 9.3 Ž4.3.

Abbre¨ iation: BPRS, Brief Psychiatric Rating Scale. a d.f.s 1,292. b Also represents a cluster of negative symptoms.

6.4 Ž3.8. 1.3 Ž1.9. 8.7 Ž4.7. 6.7 Ž3.1. 4.1 Ž3.7. 7.8 Ž4.7. 8.7 Ž4.7.

ta

P

24.01 12.20 1.92 20.50 21.69 24.88 1.92

- 0.0001 - 0.0001 0.056 - 0.0001 - 0.0001 - 0.0001 0.056

W. Chanpattana, M.L. Somchai Chakrabhand r Psychiatry Research 105 (2001) 107᎐115

responders ŽChanpattana et al., 1999c.. We found that the responders were younger, had shorter durations of illness and of the current episode, more psychiatric admissions, less family history of schizophrenia, more paranoid type, lower baseline negative symptoms, and higher baseline GAF compared to the non-responders. These findings have potential clinical implications and are similar to our prior studies ŽChanpattana et al., 1999a,b.. They also confirm the results from earlier studies of predictive features of response to ECT in schizophrenia ŽKalinowsky and Worthing, 1943; Herzberg, 1954; May, 1968; WHO, 1979; Taylor and Fleminger, 1980; Dodwell and Goldberg, 1989.. The responders had a higher baseline MMSE score than the non-responders. This finding also replicates our prior studies ŽChanpattana et al., 1999a,b.. Cognitive deficit has long been recognized in the psychopathology of schizophrenia ŽKraepelin, 1919., particularly in attention, memory, and executive functions ŽGold and Harvey, 1993; Tollefson, 1996.. Moreover, the reduced MMSE scores were reported to be associated with more severe negative symptoms ŽJohnstone et al., 1976. but to be independent of positive symptoms ŽLiddle, 1987; Bilder et al., 1992., similar to our findings. 4.2. Seizure threshold and therapeutic response We found no difference in initial seizure threshold between genders. Seizure threshold is known to be higher in men than women ŽSackeim et al., 1991., and also is positively related to age ŽSackeim et al., 1987a.. Our finding appears to be explained by patients’ age, since women were older than men Ž35.2" 8.3 vs. 33.0" 7.5 years; t 291 s 2.37, P s 0.018.. Differences in initial seizure threshold between the two ECT devices might be explained by the gender effect, as there were more men in the MECTA group Ž49%, n s 88. than in the Thymatron group Ž38.6%, n s 44, F s 0.049.. ECT is a remarkably powerful anticonvulsant and raises its own seizure threshold ŽSackeim et al., 1983.. Prior ECT studies in depressed patients found that the magnitude of the threshold in-

113

crease was related to clinical outcome, i.e. patients who failed to manifest a substantial threshold increase were particularly likely to be non-responders ŽSackeim et al., 1987b, 1991.. This was not the case in our study; we found that non-responders had higher threshold increases than responders, which was related to the number of ECT treatments. Our study design might limit probable explanation. 4.3. Predicti¨ e factors of therapeutic response There is a longstanding belief that the duration of illness andror the duration of the current episode are strongly predictive of ECT outcome in schizophrenia ŽKalinowsky and Worthing, 1943; Herzberg, 1954; Taylor and Fleminger, 1980; Dodwell and Goldberg, 1989.. We found that episode duration had a much more powerful predictive value than illness duration. Thus our results suggest that the combination of ECT and neuroleptic treatment is most effective early in an acute exacerbation of chronic schizophrenia. Similar concerns have been raised regarding the potential contribution of prolonged episode duration or ineffective treatment to treatment resistance in major depression ŽFava and Davidson, 1996; Flint and Rifat, 1996.. Medication resistance prior to ECT is known to be associated with a reduced rate of ECT response in depressive patients ŽMedical Research Council, 1965; Hamilton, 1974; Prudic et al., 1990; Devanand et al., 1991.. In our study, the longer durations of trials of the previously failed neuroleptics and the current episode may represent the degree of medication resistance. Thus our finding appears to parallel the result in the depressive patients. Other predictive factors, such as higher baseline GAF and MMSE scores, less family history of schizophrenia, and being of the paranoid type, are the factors that favor responsiveness to ECT ŽChanpattana et al., 1999a,b.. 4.4. Profile of symptom change and therapeutic response We found that greater severity of the baseline negative symptoms also predicted inferior thera-

114

W. Chanpattana, M.L. Somchai Chakrabhand r Psychiatry Research 105 (2001) 107᎐115

peutic response, as is also known to be associated with outcome with neuroleptics ŽLeff and Wing, 1971; WHO, 1979.. ECT had a limited effect on negative symptoms, as indicated by the Royal College of Psychiatrists Ž1995.. Nonetheless, treatment resulted in improvement of negative symptoms in the responders Ž8.4" 4.1 vs. 6.4" 3.8; t 1,159 s 5.22, P- 0.0001. at study end. The following explanations may be relevant to our findings: Ž1. a combination treatment with ECT and neuroleptic is much more effective than either treatment alone as both acute and continuation treatments in patients with treatment-resistant schizophrenia ŽChanpattana et al., 1999b., probably by synergistic effects ŽKrueger and Sackeim, 1995.; thus there is a marked increase in treatment efficacy. Ž2. Conventional neuroleptics are able to reduce negative symptoms to some degree ŽGoldberg, 1985; Breier et al., 1987.; and Ž3. other predictive factors may also have a role in determining the responsiveness to ECT. Another interesting finding is that a subgroup of non-responders also received benefits from ECT treatment; our results may have important clinical implications, e.g. encourage more use of ECT in patients with refractory schizophrenia, in particular in those with short duration of current episode. In summary, in patients with treatment-refractory schizophrenia who received combined treatment with ECT and flupenthixol, the rate of response was encouraging Ž54.6%.. Duration of the current episode, baseline GAF score, duration of illness, baseline MMSE score, duration of the previously failed neuroleptic trials, positive family history of schizophrenia, and paranoid type could predict therapeutic outcome. Treatment resulted in marked improvement in positive symptoms but had a limited effect on or worsened negative symptoms.

Acknowledgements Supported by the Thailand Research Fund, grant BRG 3980009.

References Agarwal, A.K., Andrade, C., Reddy, M.V., 1992. The practice of ECT in India: issues relating to the administration of ECT. Indian Journal of Psychiatry 34, 285᎐297. American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders, 4th ed. APA, Washington, DC. Baudis, P., 1992. Elektrokonvulsivni lecba v Ceske republice v letech 1981᎐1989. Ceskeslovenska Psychiatrie 88, 41᎐47. Bilder, R.M., Lipschutz-Broch, L., Rieder, R.O., 1992. Intellectual deficits in first-episode schizophrenia: evidence for progressive deterioration. Schizophrenia Bulletin 18, 437᎐448. Breier, A., Wolkowitz, O.M., Doran, A.R., 1987. Neuroleptic responsivity of negative and positive symptoms in schizophrenia. American Journal of Psychiatry 144, 1549᎐1555. Cerletti, U., Bini, L., 1938. Un neuvo metodo di shockterapie ‘L’elettro-shock’. Bolletino Accademia Medica Roma 64, 136᎐138. Chanpattana, W., Chakrabhand, S., Kongsakon, R., Techakasem, P., Buppanharun, W., 1999a. Short-term effect of combined ECT and neuroleptic therapy in treatment-resistant schizophrenia. Journal of ECT 15, 129᎐139. Chanpattana, W., Chakrabhand, S., Sackeim, H.A., Kitaroonchai, W., Kongsakon, R., Techakasem, P., Buppanharun, W., 1999b. Continuation ECT in treatment-resistant schizophrenia: a controlled study. Journal of ECT 15, 178᎐192. Chanpattana, W., Chakrabhand, S., Techakasem, P., Prasertsuk, Y., 1999c. The use of stabilization period in ECT research in schizophrenia: II. Implementation. Journal of the Medical Association of Thailand 82, 558᎐568. Devanand, D.P., Sackeim, H.A., Prudic, J., 1991. Electroconvulsive therapy in the treatment-resistant patient. Psychiatric Clinics of North America 14, 905᎐923. Dodwell, D., Goldberg, D., 1989. A study of factors associated with response to electroconvulsive therapy in patients with schizophrenic symptoms. British Journal of Psychiatry 154, 635᎐639. Fava, M., Davidson, R.G., 1996. Definition and epidemiology of treatment-resistant depression. Psychiatric Clinics of North America 19, 179᎐200. Flint, A.J., Rifat, S.L., 1996. The effect of sequential antidepressant treatment on geriatric depression. Journal of Affective Disorders 36, 95᎐105. Gold, J.M., Harvey, P.D., 1993. Cognitive deficits in schizophrenia. Psychiatric Clinics of North America 16, 295᎐312. Goldberg, S.C., 1985. Negative and deficit symptoms in schizophrenia do respond to neuroleptics. Schizophrenia Bulletin 11, 453᎐456. Hamilton, M., 1974. Drug resistant depression: response to ECT. Pharmakopsychiatry and Neuropsychopharmacology 7, 205᎐206.

W. Chanpattana, M.L. Somchai Chakrabhand r Psychiatry Research 105 (2001) 107᎐115 Herzberg, F., 1954. Prognostic variables for electro-shock therapy. Journal of General Psychology 50, 79᎐86. Johnstone, E.C., Crow, T.J., Frith, C.D., 1976. Cerebral ventricular size and cognitive impairment in chronic schizophrenia. Lancet 2, 924᎐926. Kalinowsky, L., Worthing, H.J., 1943. Results with electric convulsive therapy in 200 cases of schizophrenia. Psychiatric Quarterly 17, 144᎐153. Kane, J.M., Honigfeld, G., Singer, J., Meltzer, H.Y., 1988. Clozapine for the treatment-resistant schizophrenic: a double-blind comparison with chlorpromazine. Archives of General Psychiatry 45, 789᎐796. Kane, J.M., 1999. Pharmacologic treatment of schizophrenia. Biological Psychiatry 46, 1396᎐1408. Kongsakon, R., Vanichtanom, R., 1994. Assessment of the differences in MMSE between neurological, psychiatric patients and normal population. Journal of Rajwithi Hospital 5, 99᎐106. Kraepelin, E., 1919. Dementia Praecox and Paraphrenia. E & S Livingstone, Edinburgh. Krueger, R.B., Sackeim, H.A., 1995. Electroconvulsive therapy and schizophrenia. In: Hirsch, S., Weinberger, D. ŽEds.., Schizophrenia. Oxford University Press, Oxford, pp. 503᎐545. Latey, R., Fahy, T., 1985. Electroconvulsive therapy in the Republic of Ireland 1982: a summary of findings. British Journal of Psychiatry 147, 438᎐439. Leff, J.P., Wing, J.K., 1971. Trial of maintenance therapy in schizophrenia. British Medical Journal iii, 599᎐604. Liddle, P.F., 1987. Schizophrenic syndromes, cognitive performance, and neurological dysfunction. Psychological Medicine 17, 49᎐52. Malla, A., 1986. An epidemiological study of electroconvulsive therapy: rate and diagnosis. Canadian Journal of Psychiatry 31, 824᎐830. May, P.R., 1968. Treatment of Schizophrenia: A Comparative Study of Five Treatment Methods. Science House, New York. Medical Research Council, 1965. Clinical trial of the depressive illness. British Medical Journal 5439, 881᎐886. Overall, J.F., Gorham, D.R., 1962. The Brief Psychiatric Rating Scale. Psychological Reports 10, 799᎐812.

115

Pippard, J., Ellam, L., 1981. Electroconvulsive treatment in Great Britain, 1980. Gaskell, London. Prudic, J., Sackeim, H.A., Devanand, D.P., 1990. Medication resistance and clinical response to electroconvulsive therapy. Psychiatry Research 31, 287᎐296. Royal College of Psychiatrists, 1995. The ECT Handbook: The Second Report of the Royal College of Psychiatrists’ Special Committee on ECT. Gaskell, London. Sackeim, H.A., Decina, P., Prohovnik, I., Malitz, S., Resor, S.R., 1983. Anticonvulsant and antidepressant properties of electroconvulsive therapy: a proposed mechanism of action. Biological Psychiatry 18, 1301᎐1310. Sackeim, H.A., Decina, P., Prohovnik, I., Malitz, S., 1987a. Seizure threshold in electroconvulsive therapy: effects of sex, age, electrode placement, and number of treatments. Archives of General Psychiatry 44, 355᎐360. Sackeim, H.A., Decina, P., Portnoy, S., Neeley, P., Malitz, S., 1987b. Studies of dosage, seizure threshold, and seizure duration in ECT. Biological Psychiatry 22, 249᎐268. Sackeim, H.A., Devanand, D.P., Prudic, J., 1991. Stimulus intensity, seizure threshold, and seizure duration: impact on the efficacy and safety of electroconvulsive therapy. Psychiatric Clinics of North America 14, 803᎐843. Taylor, P.J., Fleminger, J.J., 1980. ECT for schizophrenia. Lancet 1, 1380᎐1382. Thompson, J.W., Blaine, J.D., 1987. Use of ECT in the United States in 1975 and 1980. American Journal of Psychiatry 144, 557᎐562. Thompson, J.W., Weiner, R.D., Myers, C.P., 1994. Use of ECT in the United States in 1975, 1980, and 1986. American Journal of Psychiatry 151, 1657᎐1661. Tollefson, G.D., 1996. Cognitive function in schizophrenic patients. Journal of Clinical Psychiatry 57 Žsuppl 11., 31᎐39. Wirshing, D.A., Marshall, B.D., Green, M.F., Mintz, J., Marder, S.R., Wirshing, W.C., 1999. Risperidone in treatment-refractory schizophrenia. American Journal of Psychiatry 156, 1374᎐1379. World Health Organization, 1979. Schizophrenia: An International Follow-Up Study. John Wiley & Sons, New York.