1282
BIOL PSYCHIATRY 1990;27:1282-1292
Mechanisms of Action of ECT: Schizophrenia and Schizoaffective Disorder Victor Milstein, Joyce G. Small, Marvin J. Miller, Patricia H. Sti,~rp!ey, and Iver F. Small
A variety of neurophysiological mechanisms !rove been suggested to explain the therapeutic action of electroconvulsive therapy (ECT). Processes of kindling, resolution of hemispheric dysfunctions, anticonvulsant effects, and diencephalic stimulation all have been proposed to account for the beneficial effects of ECT. To investigate these, we analyzed clinical, neuropsychological, and electroencephalographic (EEG) data from 1I0 ECT-treated patients with ¢chizophrenia and schizoaffective disorders, comparing responders with nonresponderf. Fifty.four percent of all the patients were rated as very much or much improved. Mechanisms of kindling or anticonvulsant effects were not supported by the data. Dominant hemispheric dysfunctions in schizophrenics were suggested by the neurop~ychologicai test data. There was tenuous support for the sensitization theory and both the neuropsychological and EEG data contradicted the dominant accentuation theory. Taken together with orcrprevious report on ECT-treated patients with affective disorders, we propose that ECT might act by restoration of equilibrium between the hemispheres.
Introduction Previously we (Small et el. 1988) investigatt~dclinical, neumpsychological, and electroencephalogram (EEG) findings in a series of 150 patients with affective disorders who received ECT. Data from responders and nonresponders we~ compared in light of currently held neurophysiological theories about mechanisms of action. The findings from that series of affectively ill patients revealed essentially no support for theories of kindling (Goddard 1967), sensitization (Antelman and Chiodo 1984), anticonvulsant effects (Sackelm et al. 1987), or diencephalie stimulation (Roth 1951; Abrams 1986). However, patients with affective disorders did show neuropsychological evidence of right hemispheric dysfunctions. Patients who responded to ECT showed a_ combination of more improvements in right hemispheric functions combined with more impairments on the dominant side than did nonresponders. These differences were manifested both by changes in neuropsychological test scores and transient, lateralized EEG slowing. These findings
Based on a paper presented to the Society of Biological Psychiatry, Montreal, Canada, May, 1988. From the Lame D. Carter Memorial Hospital and Indiana University School of Medicine, Indianapolis, IN. Address reprint requests Io Dr. Victor Milstein, Larue D. Carter Memorial Hospital, 1315 West 10lh Slseet, Indianapolis, IN 46202. Received December 30, 1988; revised December 5.19gq. © 1990 Society of B~ological Psychiatry
0006-3223i~$G3.50
Mechanisms of ECT
nlOL PSYCHIATRY i990;27:1282-1292
1283
led us to propose that ECT could act by restoring hemispheric equilibrium, with reciprocal enhancement of right brain functions and dampening of the left.
Methods We further investigated these issues in a series of ECT-treated patients with DSM-III-R diagnoses of paranoid schizophrenia (PS), nonparanoid schizophrenia (NPS), and schizoaffective (SA) disorders to see if comparative data from responde~ and nonresponders wo~,~d be the same or different then in patients with affective disorders. Our setting is a tertiary care, psychiatric facility on the Indiana University Medical School campus which has a statewide catchment area. It is one of only two public hospitals in the state that offers ECT. From 1981 to 1986, all patients referred for ECT who were able to cooperate w ~ evaluated before ECT, "24-32 hours after the fifth or sixth treatment, and 2 or 3 weeks following ECT termination. The decision to refer patients for ECT was made by staff psychiatrists who continued to care for their patients while ECT was administered by a te~n that included anesthesiologist, resident psychiatrist, nurses, and technic:.an. As is generally the case (Small et al. 1986), failure to respond to other therapy or dew~lopment of intolerable side effects were major reasons for patients being referred for ECT. Diagnoses were assigned according to DSM III-R c.riteria by an inder~ndent psychiatrist based on a review of the patient's entire medical record. Data were obtained from a valiety of clinical, neuropsychological, and EEG assessments. Ratings included the 24item Hamilton Depression Scale (HAM-D) (Hamilton 1960), the Brief Psychiatric Rating Scale (BPRS) (Overall and Gorham 1962), and the Clinical Global Impression (CGI) (Guy 1976). Experienced clinicians who completed these ratings were blind to the side or sides of treatment and other ECT procedures. Decisions about when to begin and end ECT and whether to switch from unilateral to bilateral treatment or vice versa were made by attending psychiatrists who were unaware of the ratings. Prior to ECT, a complete waking and sleep EEG and the Halstead-Reitan Neuropsychological Test Battery (Reitan and Wolfso;l 1985) were accomplished. In addition, the Wechsler Adult Intelligence Scale (WAIS) (Wechsler 1981) was administered. All clinical ratings were repeated 24 hours after the fifth or sixth seizure along with a subset of neuropsychological tests we judged as least sensitive to practice effects. All assessments were repeated 2-3 weeks ai'ter ECT termination. Multichannel EEGs were recorded immediately before, during, and after the first and fifth or sixth ECT (48 hours after the previous seizure) and 2-3 weeks after ECT termination. ECT was given three times weekly with routine unilateral induction unless bilateral treatment specifically was prescribed by the attending psychiatrist. Bilateral treatment was ordered if the patient was at great medic~,l or psychiatric risk, had a diagnosis of bipolar manic disorder (Small et al. 1985; Milstein et al. 1987), or failed to respond to unilateral ECT. Glycopyrrolate, methohexital, and succinylcholine were used along with preoxygenation and support of respiration during and after the seizure. (For additional details concerning specifics of the ECT procedure, see Milstein et al. 1986.) For almost all of ~he treatments, the instrument used was the monitored EL~I' apparatus (MECTA Model D). We considered factors such as gender, elecu~ode placement, and age in determining the stimulus for the first treatment. Afterwards, stimulus parameters were adjusted dependin~ u ~ n clinical response, seizuce duraticn, and the development of side effects. Electrodes were applied with the Lancaster unilateral positi~on (Lancaster et al. 1958) and with bitemporal placement for bilateral ECT.
12 84
BIOLPSYCHIATRY 1990;27:1282-1292
V. Milstein et al.
Table 1. Identifying Data Gender (%) n
M
F
Schizoaffective Responder Nonresponder
25 14 I1
44 36 54
56 64 45
30 30 29
i 1.9 11.5 12.4
20 28 9
16 7 27
64 64 64
Paranoid schizophrenia Responder Nonresponder
39 26 !3
46 42 54
54 58 46
27 26 28
13.2 14.2 11.4
13 12 15
18 15 23
69 73 62
46 20 2_66
54 70 42
46 30 58
24 25 23
11.2 11.2 11.2
I1 20 4
2 0 4
89 80 92
DSM-III Diagnosis
Not paranoid schizophrenia qesponder Nonresponder
Education (mean years)
Maritial status (%)
~ge ° (mean years)
Ever M a r r i e d married
Single
Iis't l|u
"Two-wayANOVA:Diagnosissignificant;response and interaction not significant.
The data were exan~ined using multivariate analyses of variance (ANOVA) for measures (the pre- and post-ECT termination scores) repeated over time. The neuropsycho. logical test scores were converted to z-scores and compared with age- and gender-matched normal subjects. We combined a number of these to represent right, left, anterior, and posterior functions (Reitan 1985). Scores from left-hand Tapping and Tactual Performance Test (TPT) and Trails A and Performance IQ made up the right hemisphere z-score; those from right-hand Tapping and TPT and Trails B and Performance IQ comprised the left hemisphere z-score. The anterior z-score consisted of right- plus left-hand Tapping, Category Test Total score, Trails B, w~d Picture Arrangement subtest of the WAIS. TPT Total score and WAIS Block Design :~lbtest made up the posterior z-score.
Results Table I presents identifying data ~'orthe 110 ECT-treated patients with schizophrenia and schizoaffective disorders, grouped according to DSM-III-R diagnoses. Treatment response was based on CGI ratings 2-3 weeks after ECT termination. A rating of "very much improved" or "much improved" classified the patient as a responder. All other ratings including "minimally improved" defined nonresponders. The NPS group included disorganized, undifferentiated, catatonic, and residual classifications. The groups were similar except that the NPS patients were significantly younger with more single male patients. Concomitant medications (Table 2) that the patients received dm'ing their course of ECT were recorded and a senior clinician (J.G.S.) established dosage categories of low, medium, and high for each class. No patient i~ceived lithium or any stimulant medication and only a very few patients received o~,e. dose of antianxiety, antidepressant, anticonvulsant, or antihistaminic medications during their co,:rse of ECT. Seven patients (6.4%) were completely drug free during their ECT course. As Table 2 shows, concomitant antipsychotic medicatiov, was given to most of the patients with only 2 in the PS group receiving none. Differences in concomitant antipsychotic drug therapy during the course of ECT among ~¢ "~ree diagnostic groups approached statistical significance. Anticho-
Mechanisms of ECT
BIOL PSYCHIATRY 199o;27:!282-1292
~285
Table 2. Concomitam Medications (Percent) Total ~,; = I|0)
Schizoaffective (n = 25)
Paranoid schizophrenia (n -- 39)
Not paranoid schizophrenia (n = 46)
X2
p (df = 2)
Antipsychotics Any dose High dose
87 26
76 21
95 27
87 33
4.876 2.347
>0.05, 0.25
Anticholinergic Any dose High dose
61 I
64 4
61 0
80 0
4. ! --
>0.10
19 7
24 ~2
20 8
13 4
!.591 0. I ! 8°
>0.25 >0.975
Hypnotic/Sedative
Any dose High dose
°High versus low + mediumdose.
linergic medication was given to 61% of the sample with greatest use in 80% of the NPS patients. The three groups of patients did not differ significantly. Only 1 patient from the SA group received high doses of antiparkinson medication. The only other class of drags that was prescribed for more than I or 2 patients was hypnotics/sedatives, with 19% of the patients receiving ~uch drugs one or more times during the course of their ECT, but not on the night preceding treatment. This was the oniy class of drugs in which high doses were given more frequently than low or medium doses. When patients receiving high doses of concomitant antipsychotic medication were examined separately, the groups did not differ. As Table 3 shows, there were no significant associations between response to ECT and dose of concomitant antipsychotic medications. This held true for the SA, PS, and NPS groups separately and when they were considered together. There also were no significant associat_~e-~g ~,h~.n only those responders and nonresponders who had received high-dose medications were compared with those who were drug free during their course of ECT. Table 3. ECT Response and Concomitant Antipsychotic ~!edications Dt)se (%) P
Group
n
None
Low
lVledium
High
Schizoaffective Responder Nonresponder
25 14 1i
12 12
16 4
20 20
8 8
Paranoid schizophrenic Rcsponder Nonresponder
39 26 13
2 2
20 5
26 18
!8 8
Not paranoid schizophrenic Responder Nonresponder
46 20 26
6 6
6 6
13 28
17 15
110 60 50
6 6
14 5
19 23
15 !!
All patients Responder Nomesponder
X2
(df = 3)
!.464
>0.50
!.630
>0.50
1.960
>0.50
4.142
0.25
1286
V. Milstein et al.
BIOL PSYCHIATRY
1990;27:1282-1292
Table 4. Pre-ECT Clinical Ratings (mean scores)
BPRS factors I. Thinking disturbance I!. Withdrawal-Retardation I11. H~le-Suspiciousness IV. Anxious depression V. Schizophrenia VI. Total score Hamilton-D factors 1. Somatization IL Diurnal variation !!I. Sleep disturbance IV. Weight loss V. Reality disturbance VI. Mood depression VII. Agitation-Anxiety Vegetative depression Cos~'~,tive depression Total score
Schizoaffective
Paranoid schizophrenia
Not paranoid schizophrenia
ANOVA
(n = 25)
(n = 39)
(n = 46)
p
9.9 8.2 8.2 7.8 18.8 49.2
10.9 9.0 8.0 5. I 20.2 45.2
i0.9 9.0 4.4 6. I 18.3 41.0
ns ns ns 0.0022 ns 0.01
2.2 4.2 3. I 4.2 2.0 8.9 4.8
2.0 4.6 2.5 4.3 1.6 6.4 3.6
i.g 2.7 2.7 3.7 1.5 7.6 3.7
ns
BIOL PSYCHIATRY 1990;27:1282-1292
Mechanisms of Action of ECT: Schizophrenia and Schizoaffective Disorder Victor Milstein, Joyce G. Small, Marvin J. Miller, Patricia H. Sti,~rp!ey, and Iver F. Small
A variety of neurophysiological mechanisms !rove been suggested to explain the therapeutic action of electroconvulsive therapy (ECT). Processes of kindling, resolution of hemispheric dysfunctions, anticonvulsant effects, and diencephalic stimulation all have been proposed to account for the beneficial effects of ECT. To investigate these, we analyzed clinical, neuropsychological, and electroencephalographic (EEG) data from 1I0 ECT-treated patients with ¢chizophrenia and schizoaffective disorders, comparing responders with nonresponderf. Fifty.four percent of all the patients were rated as very much or much improved. Mechanisms of kindling or anticonvulsant effects were not supported by the data. Dominant hemispheric dysfunctions in schizophrenics were suggested by the neurop~ychologicai test data. There was tenuous support for the sensitization theory and both the neuropsychological and EEG data contradicted the dominant accentuation theory. Taken together with orcrprevious report on ECT-treated patients with affective disorders, we propose that ECT might act by restoration of equilibrium between the hemispheres.
Introduction Previously we (Small et el. 1988) investigatt~dclinical, neumpsychological, and electroencephalogram (EEG) findings in a series of 150 patients with affective disorders who received ECT. Data from responders and nonresponders we~ compared in light of currently held neurophysiological theories about mechanisms of action. The findings from that series of affectively ill patients revealed essentially no support for theories of kindling (Goddard 1967), sensitization (Antelman and Chiodo 1984), anticonvulsant effects (Sackelm et al. 1987), or diencephalie stimulation (Roth 1951; Abrams 1986). However, patients with affective disorders did show neuropsychological evidence of right hemispheric dysfunctions. Patients who responded to ECT showed a_ combination of more improvements in right hemispheric functions combined with more impairments on the dominant side than did nonresponders. These differences were manifested both by changes in neuropsychological test scores and transient, lateralized EEG slowing. These findings
Based on a paper presented to the Society of Biological Psychiatry, Montreal, Canada, May, 1988. From the Lame D. Carter Memorial Hospital and Indiana University School of Medicine, Indianapolis, IN. Address reprint requests Io Dr. Victor Milstein, Larue D. Carter Memorial Hospital, 1315 West 10lh Slseet, Indianapolis, IN 46202. Received December 30, 1988; revised December 5.19gq. © 1990 Society of B~ological Psychiatry
0006-3223i~$G3.50
Mechanisms of ECT
nlOL PSYCHIATRY i990;27:1282-1292
1283
led us to propose that ECT could act by restoring hemispheric equilibrium, with reciprocal enhancement of right brain functions and dampening of the left.
Methods We further investigated these issues in a series of ECT-treated patients with DSM-III-R diagnoses of paranoid schizophrenia (PS), nonparanoid schizophrenia (NPS), and schizoaffective (SA) disorders to see if comparative data from responde~ and nonresponders wo~,~d be the same or different then in patients with affective disorders. Our setting is a tertiary care, psychiatric facility on the Indiana University Medical School campus which has a statewide catchment area. It is one of only two public hospitals in the state that offers ECT. From 1981 to 1986, all patients referred for ECT who were able to cooperate w ~ evaluated before ECT, "24-32 hours after the fifth or sixth treatment, and 2 or 3 weeks following ECT termination. The decision to refer patients for ECT was made by staff psychiatrists who continued to care for their patients while ECT was administered by a te~n that included anesthesiologist, resident psychiatrist, nurses, and technic:.an. As is generally the case (Small et al. 1986), failure to respond to other therapy or dew~lopment of intolerable side effects were major reasons for patients being referred for ECT. Diagnoses were assigned according to DSM III-R c.riteria by an inder~ndent psychiatrist based on a review of the patient's entire medical record. Data were obtained from a valiety of clinical, neuropsychological, and EEG assessments. Ratings included the 24item Hamilton Depression Scale (HAM-D) (Hamilton 1960), the Brief Psychiatric Rating Scale (BPRS) (Overall and Gorham 1962), and the Clinical Global Impression (CGI) (Guy 1976). Experienced clinicians who completed these ratings were blind to the side or sides of treatment and other ECT procedures. Decisions about when to begin and end ECT and whether to switch from unilateral to bilateral treatment or vice versa were made by attending psychiatrists who were unaware of the ratings. Prior to ECT, a complete waking and sleep EEG and the Halstead-Reitan Neuropsychological Test Battery (Reitan and Wolfso;l 1985) were accomplished. In addition, the Wechsler Adult Intelligence Scale (WAIS) (Wechsler 1981) was administered. All clinical ratings were repeated 24 hours after the fifth or sixth seizure along with a subset of neuropsychological tests we judged as least sensitive to practice effects. All assessments were repeated 2-3 weeks ai'ter ECT termination. Multichannel EEGs were recorded immediately before, during, and after the first and fifth or sixth ECT (48 hours after the previous seizure) and 2-3 weeks after ECT termination. ECT was given three times weekly with routine unilateral induction unless bilateral treatment specifically was prescribed by the attending psychiatrist. Bilateral treatment was ordered if the patient was at great medic~,l or psychiatric risk, had a diagnosis of bipolar manic disorder (Small et al. 1985; Milstein et al. 1987), or failed to respond to unilateral ECT. Glycopyrrolate, methohexital, and succinylcholine were used along with preoxygenation and support of respiration during and after the seizure. (For additional details concerning specifics of the ECT procedure, see Milstein et al. 1986.) For almost all of ~he treatments, the instrument used was the monitored EL~I' apparatus (MECTA Model D). We considered factors such as gender, elecu~ode placement, and age in determining the stimulus for the first treatment. Afterwards, stimulus parameters were adjusted dependin~ u ~ n clinical response, seizuce duraticn, and the development of side effects. Electrodes were applied with the Lancaster unilateral positi~on (Lancaster et al. 1958) and with bitemporal placement for bilateral ECT.
12 84
BIOLPSYCHIATRY 1990;27:1282-1292
V. Milstein et al.
Table 1. Identifying Data Gender (%) n
M
F
Schizoaffective Responder Nonresponder
25 14 I1
44 36 54
56 64 45
30 30 29
i 1.9 11.5 12.4
20 28 9
16 7 27
64 64 64
Paranoid schizophrenia Responder Nonresponder
39 26 !3
46 42 54
54 58 46
27 26 28
13.2 14.2 11.4
13 12 15
18 15 23
69 73 62
46 20 2_66
54 70 42
46 30 58
24 25 23
11.2 11.2 11.2
I1 20 4
2 0 4
89 80 92
DSM-III Diagnosis
Not paranoid schizophrenia qesponder Nonresponder
Education (mean years)
Maritial status (%)
~ge ° (mean years)
Ever M a r r i e d married
Single
Iis't l|u
"Two-wayANOVA:Diagnosissignificant;response and interaction not significant.
The data were exan~ined using multivariate analyses of variance (ANOVA) for measures (the pre- and post-ECT termination scores) repeated over time. The neuropsycho. logical test scores were converted to z-scores and compared with age- and gender-matched normal subjects. We combined a number of these to represent right, left, anterior, and posterior functions (Reitan 1985). Scores from left-hand Tapping and Tactual Performance Test (TPT) and Trails A and Performance IQ made up the right hemisphere z-score; those from right-hand Tapping and TPT and Trails B and Performance IQ comprised the left hemisphere z-score. The anterior z-score consisted of right- plus left-hand Tapping, Category Test Total score, Trails B, w~d Picture Arrangement subtest of the WAIS. TPT Total score and WAIS Block Design :~lbtest made up the posterior z-score.
Results Table I presents identifying data ~'orthe 110 ECT-treated patients with schizophrenia and schizoaffective disorders, grouped according to DSM-III-R diagnoses. Treatment response was based on CGI ratings 2-3 weeks after ECT termination. A rating of "very much improved" or "much improved" classified the patient as a responder. All other ratings including "minimally improved" defined nonresponders. The NPS group included disorganized, undifferentiated, catatonic, and residual classifications. The groups were similar except that the NPS patients were significantly younger with more single male patients. Concomitant medications (Table 2) that the patients received dm'ing their course of ECT were recorded and a senior clinician (J.G.S.) established dosage categories of low, medium, and high for each class. No patient i~ceived lithium or any stimulant medication and only a very few patients received o~,e. dose of antianxiety, antidepressant, anticonvulsant, or antihistaminic medications during their co,:rse of ECT. Seven patients (6.4%) were completely drug free during their ECT course. As Table 2 shows, concomitant antipsychotic medicatiov, was given to most of the patients with only 2 in the PS group receiving none. Differences in concomitant antipsychotic drug therapy during the course of ECT among ~¢ "~ree diagnostic groups approached statistical significance. Anticho-
Mechanisms of ECT
BIOL PSYCHIATRY 199o;27:!282-1292
~285
Table 2. Concomitam Medications (Percent) Total ~,; = I|0)
Schizoaffective (n = 25)
Paranoid schizophrenia (n -- 39)
Not paranoid schizophrenia (n = 46)
X2
p (df = 2)
Antipsychotics Any dose High dose
87 26
76 21
95 27
87 33
4.876 2.347
>0.05, 0.25
Anticholinergic Any dose High dose
61 I
64 4
61 0
80 0
4. ! --
>0.10
19 7
24 ~2
20 8
13 4
!.591 0. I ! 8°
>0.25 >0.975
Hypnotic/Sedative
Any dose High dose
°High versus low + mediumdose.
linergic medication was given to 61% of the sample with greatest use in 80% of the NPS patients. The three groups of patients did not differ significantly. Only 1 patient from the SA group received high doses of antiparkinson medication. The only other class of drags that was prescribed for more than I or 2 patients was hypnotics/sedatives, with 19% of the patients receiving ~uch drugs one or more times during the course of their ECT, but not on the night preceding treatment. This was the oniy class of drugs in which high doses were given more frequently than low or medium doses. When patients receiving high doses of concomitant antipsychotic medication were examined separately, the groups did not differ. As Table 3 shows, there were no significant associations between response to ECT and dose of concomitant antipsychotic medications. This held true for the SA, PS, and NPS groups separately and when they were considered together. There also were no significant associat_~e-~g ~,h~.n only those responders and nonresponders who had received high-dose medications were compared with those who were drug free during their course of ECT. Table 3. ECT Response and Concomitant Antipsychotic ~!edications Dt)se (%) P
Group
n
None
Low
lVledium
High
Schizoaffective Responder Nonresponder
25 14 1i
12 12
16 4
20 20
8 8
Paranoid schizophrenic Rcsponder Nonresponder
39 26 13
2 2
20 5
26 18
!8 8
Not paranoid schizophrenic Responder Nonresponder
46 20 26
6 6
6 6
13 28
17 15
110 60 50
6 6
14 5
19 23
15 !!
All patients Responder Nomesponder
X2
(df = 3)
!.464
>0.50
!.630
>0.50
1.960
>0.50
4.142
0.25
1286
V. Milstein et al.
BIOL PSYCHIATRY
1990;27:1282-1292
Table 4. Pre-ECT Clinical Ratings (mean scores)
BPRS factors I. Thinking disturbance I!. Withdrawal-Retardation I11. H~le-Suspiciousness IV. Anxious depression V. Schizophrenia VI. Total score Hamilton-D factors 1. Somatization IL Diurnal variation !!I. Sleep disturbance IV. Weight loss V. Reality disturbance VI. Mood depression VII. Agitation-Anxiety Vegetative depression Cos~'~,tive depression Total score
Schizoaffective
Paranoid schizophrenia
Not paranoid schizophrenia
ANOVA
(n = 25)
(n = 39)
(n = 46)
p
9.9 8.2 8.2 7.8 18.8 49.2
10.9 9.0 8.0 5. I 20.2 45.2
i0.9 9.0 4.4 6. I 18.3 41.0
ns ns ns 0.0022 ns 0.01
2.2 4.2 3. I 4.2 2.0 8.9 4.8
2.0 4.6 2.5 4.3 1.6 6.4 3.6
i.g 2.7 2.7 3.7 1.5 7.6 3.7
ns
