Postictal suppression correlates with therapeutic efficacy for depression in bilateral sine and pulse wave electroconvulsive therapy

Authors

  • HIDEKI AZUMA md , phd,

    1. Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya, Japan
    Search for more papers by this author
  • AKIKO FUJITA md ,

    1. Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya, Japan
    Search for more papers by this author
  • KIYOE SATO md , phd,

    1. Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya, Japan
    Search for more papers by this author
  • KEIKO ARAHATA md ,

    1. Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya, Japan
    Search for more papers by this author
  • KAZUYUKI OTSUKI md ,

    1. Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya, Japan
    Search for more papers by this author
  • MIKI HORI md ,

    1. Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya, Japan
    Search for more papers by this author
  • YOSHIHITO MOCHIDA md ,

    1. Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya, Japan
    Search for more papers by this author
  • MEGUMI UCHIDA md ,

    1. Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya, Japan
    Search for more papers by this author
  • TOMOKO YAMADA md ,

    1. Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya, Japan
    Search for more papers by this author
  • TATSUO AKECHI md , phd,

    1. Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya, Japan
    Search for more papers by this author
  • TOSHI A. FURUKAWA md , phd

    1. Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya, Japan
    Search for more papers by this author

Hideki Azuma, MD, PhD, Department of Psychiatry and Cognitive-Behavioral Medicine, Nagoya City University Graduate School of Medical Sciences, Mizuho-cho, Mizuho-ku, Nagoya 467–8601, Japan. Email: azma@med.nagoya-cu.ac.jp

Abstract

Abstract  Ictal electroencephalography (EEG) parameters in electroconvulsive therapy (ECT) for depression reportedly correlate with therapeutic response and stimulus dosage, particularly in right unilateral (RUL) ECT. The authors examined ictal EEG parameters as predictors of therapeutic effectiveness in bilateral (BL) sine and pulse wave ECT. A total of 30 consecutive depressed inpatients who had not responded to standard pharmacotherapy were treated using BL ECT given in either sine or pulse wave mode. Ictal EEG parameters (e.g. regularity, postictal suppression) were manually rated by three trained psychiatrists. Polyspike phase duration was significantly longer in sine wave ECT than in pulse wave ECT. Postictal suppression emerged as the only significant predictor of therapeutic outcome when baseline Hamilton Rating Scale for Depression and mode of stimulation were controlled for. Postictal suppression appears to offer a useful predictor of clinical outcome of depression in BL ECT. No EEG parameters were found to be differentially predictive between sine and pulse wave ECT.

INTRODUCTION

Electroconvulsive therapy (ECT) is widely used to treat depression and is considered to be a rapidly acting and effective therapy.1 This approach is especially indicated for patients who are unresponsive to conventional antidepressant pharmacotherapies.

For decades, generalized seizure was thought to provide necessary and sufficient conditions for the antidepressant effects of ECT.2 Various aspects of ictal and immediately postictal electroencephalography (EEG) parameters have been studied as possible predictors of antidepressant effectiveness. In the first place, seizure duration has been shown to have limited relevance to efficacy. Conversely, several studies have found that patients with superior clinical outcome had greater ictal amplitude in the delta frequency band and greater postictal suppression in right unilateral (RUL) or bilateral (BL) pulse wave ECT.3–6

In Japan, the authors used the sine wave ECT apparatus until 2003, when use of the pulse wave ECT device was approved by the regulatory agency. Sine wave ECT required approximately threefold greater stimulus energy compared to pulse wave ECT producing equivalent supra-threshold seizures.7 However, no difference exists between the effectiveness of sine and pulse wave stimuli on depressive symptoms, and degree of stimulus dosage correlated with degree of cognitive impairment.1 Pulse wave ECT, therefore, appears superior to sine wave ECT because of fewer cognitive problems.

Ictal EEG parameters have been shown to differentiate low dosage RUL pulse wave ECT from moderately supra-threshold RUL pulse wave ECT,6 but differences in ictal EEG expression due to wave mode or associated relationships to therapeutic effects have not yet been fully elucidated. The present study, therefore, compared ictal and peri-ictal EEG parameters between BL sine and pulse wave ECT, and investigated ictal EEG parameters predicting antidepressant therapeutic outcomes in BL sine and pulse wave ECT.

METHODS

Subjects

The subjects comprised 33 consecutive depressed inpatients who were referred for ECT in the Department of Psychiatry at Nagoya City University Hospital, Nagoya, Japan, following lack of response to previous treatment with at least one full-dose antidepressant medication for ≥4 weeks or intolerance to such medication. All subjects fulfilled Diagnostic and Statistical Manual of Mental Disorders, fourth edition criteria for unipolar major depression or bipolar disorder, with depression as the most recent episode.8 Diagnosis was determined by psychiatrists through a clinical interview and review of psychiatric records. Patients with the following conditions were excluded: past or present history of schizophrenia, schizoaffective disorder, significant neurological illness, substance abuse or substance dependence; any other significant medical illness; grave abnormality on chest radiography, brain computed tomography or EEG; or administration of ECT within the previous 6 months. Within 10 days before and after a course of ECT, depression severity was assessed using the Structured Interview Guide for the Hamilton Rating Scale for Depression (HRSD).9 Ratings were completed by an independent psychiatrist not directly involved in the clinical management. This study protocol was approved by the Ethics Committee of Nagoya City University Medical School. All subjects were informed about purposes and procedures of ECT and provided written consent to participate in the study.

Electroconvulsive therapy and medication

ECT was given through electrodes positioned at the standard bifrontotemporal location. In that position, each electrode was placed on a perpendicular line 3 cm above the midpoint of the line joining the external auditory meatus and outer canthus of the eye. For sine wave stimuli, a CS-1 sine wave device (Sakai Medical, Tokyo, Japan) was used. Stimulation was delivered at 100–120 V for 5–6 s to produce a tonic clonic seizure. Two channels of EEG (C3-A1, C4-A2, international 10–20 system) were monitored and recorded onto paper from the end of stimuli using a Lifescope 12 monitor (Nihon Kohden, Tokyo, Japan). For pulse wave stimuli, a Thymatron System IV ECT apparatus (THYMATRON System IV, Somatics, Inc., Lake Bluff, IL, USA) containing an inbuilt EEG system (Fp1-A1, Fp2-A2, international 10–20 system) was used. Stimulation dose was calculated using the ‘half age’ method.10 A LOW 0.5 preset program using 0.5 ms pulse width was selected, adjusting frequency to maximize duration.11 The criterion for an adequate seizure was an electroencephalographic seizure lasting ≥20 s. If no electroencephalographic seizure had occurred after 20 s, re-stimulation at a higher stimulus intensity was immediately performed by increasing voltage by 5 V up to 120 V for sine wave stimuli, or by 10% for pulse wave stimuli, to a maximum of three stimulations/session. Motor seizures were further monitored in a cuffed leg. For patients undergoing sine wave ECT, a mixture of 1.0% sevoflurane and oxygen was inhaled, and 3 mg vecronium bromide was used intravenously for muscle relaxation. Intravenous injection of 2.5 mg/kg propofol and 3 mg vecronium bromide was performed for patients given pulse wave stimuli. When the authors started using the pulse wave device in January 2004, electroencephalographic seizures were harder to produce. The authors, therefore, consulted with an anesthesiologist and changed anesthetic agent from sevoflurane to propofol.

Antidepressants remained unchanged at a minimal dose throughout the course of ECT. Lithium carbonate and sodium valproate were withdrawn before first ECT. Due to clinical considerations, use of benzodiazepine was permitted during the study to alleviate insomnia and anxiety (sine wave group, n = 13; pulse wave group, n = 6). Benzodiazepine dosage at the time of the first ECT session was ≤4 mg/day of lorazepam or equivalent. In the sine wave group, four patients were given major tranquilizers, for example, chlorpromazine, levomepromazine or sulpiride, and one patient received the antiparkinson drug promethazine at 75 mg/day. In the pulse wave group, six patients were given major tranquilizers, for example, chlorpromazine, levomepromazine, risperidone or perphenazine, and one patient received biperiden 2 mg/day.

Electroencephalography analysis

A total of 73.8% of EEG in the sine wave group and 97.3% of EEG in the pulse wave group were examined. In the sine wave group, some ictal EEG were missing or were unsuitable for analysis due to artifacts during recordings. Ictal and peri-ictal EEG parameters, including polyspike phase maximal amplitude (mV), polyspike phase duration (s), slow wave phase maximum amplitude (mV), slow wave phase duration (s), regularity (global seizure strength, 7-point scale ranging from 0 to 6), stereotypy (global seizure patterning, 4-point scale ranging from 0 to 3), and postictal suppression (degree of postictal bioelectric suppression, 4-point scale ranging from 0 to 3) were manually rated by three experienced psychiatrists using standard methods from the literature.12–14

The polyspike phase was defined as starting with the offset of stimulation and terminating point at which visually discernible slow wave activity fully replaced early chaotic polyspike activity. The slow wave phase was defined as the period from this time point until seizure termination. Maximal amplitudes during the polyspike and slow wave phases were defined by the largest peak to peak deflections in the relevant phase, then mean maximal amplitude was determined for each patient. Seizures were rated as more stereotypic if a clear progression from low amplitude chaotic polyspike activity to high amplitude slow wave activity was seen without reappearance of chaotic polyspike activity or marked variability in amplitude during phases. Seizures were rated as having greater regularity if slow wave activity of high amplitude predominated regularly during the slow wave phase. Higher numbers indicate increasing stereotypy and regularity of the ictal EEG recording. Postictal suppression was rated as follows: 0, cannot tell where the seizure ends; 1, seizure termination is clear, but suppression is poor (not flat); 2, good seizure suppression (very flat), but transition to flat is gradual; and 3, good seizure suppression (very flat), and transition is abrupt. Interrater reliability among the three raters was examined by calculating anova intraclass correlation coefficient (anova ICC).

Statistical analysis

All results are expressed as mean ± standard deviation (SD). The level of statistical significance was set at α = 0.05. To investigate possible differences in ictal EEG parameters between sine and pulse wave stimuli, t-tests and χ2 tests were used (two-tailed). Predictors of HRSD improvement among ictal EEG parameters were investigated for each parameter separately through multiple regression, always controlling for mode of stimulation and baseline HRSD. SPSS version 11.5 software (SPSS Inc., Chicago, IL, USA) was used for all statistical analyses.

RESULTS

Among consecutive 33 patients (sine wave group, n = 18; pulse wave group, n = 15), two patients developed angina (sine wave group, n = 1; pulse wave group, n = 1) and one patient in the sine wave group refused ECT due to delirium during the study period. As a result, 30 patients participated in the study (sine wave group, n = 16; pulse wave group, n = 14). The pulse wave group contained a significantly greater proportion of women than the sine wave group (Table 1). All patients displayed normal results on routine hematological, biochemical and thyroid function testing, chest radiography and EEG.

Table 1.  Clinicodemographic characteristics of the patient sample
 Sine wave ECT (n = 16)Pulse wave ECT (n = 14)P
MeanSDMeanSD
  1. P <0.05; t-test (two-tailed), χ2 test.

  2. ECT, electroconvulsive therapy; HRSD, Hamilton Rating Scale for Depression; SD, standard deviation.

Age (years)53.911.049.413.30.32
Gender (% female)25.0 60.0 0.03
Age of onset (years)44.015.438.514.50.33
Number of previous episodes1.81.62.72.90.27
Duration of current episode (months)18.421.617.116.90.85
Polarity (% bipolar)12.5 13.3 0.89
Number of ECT administrations10.01.810.72.30.34
Baseline HRSD18.86.617.77.40.67
Post-treatment HRSD7.55.110.06.80.26

Prior to ECT, each subject in the sine wave group had received 1.8 ± 0.7 antidepressant medication trials and each subject in the pulse wave group had received 2.2 ± 0.7 antidepressant medication trials. Maximal antidepressant dose in imipramine equivalents prior to ECT was 223.7 ± 116.9 mg/day in the sine wave group and 273.9 ± 92.1 mg/day in the pulse wave group. During the course of ECT, patients in the sine and pulse wave groups received 165.9 ± 83.4 mg/day and 246.1 ± 147.4 mg/day of imipramine equivalents and 40.9 ± 86.7 mg/day and 58.4 ± 133.4 mg/day of chlorpromazine equivalents, respectively. No significant differences in antidepressant medication trials, maximal antidepressant dose, antidepressant dosage during ECT or major tranquilizer during ECT were noted between sine and pulse wave groups.

The actual amount of electricity delivered with the pulse wave apparatus was 224.2 ± 91.7 mC, while electrical dose in the sine wave machine could not be measured. Mean HRSD scores decreased from 18.8 ± 6.6 to 7.5 ± 5.1 with sine wave ECT (P < 0.001, paired t-test, two-tailed) and from 17.7 ± 7.4 to 10.0 ± 6.8 with pulse wave ECT (P = 0.002, paired t-test, two tailed). The proportions of patients showing >50% reduction or scoring ≤8 on the HRSD were 63% and 43%, respectively (P = 0.28, χ2 test, two-tailed). Neither pretreatment HRSD nor post-treatment HRSD differed significantly between the two groups.

Mean number of analyzable ictal EEG was significantly fewer in the sine wave group than in the pulse wave group (sine wave group, 7.4 ± 3.0; range, 3–12; pulse wave group, 10.4 ± 2.3; range, 6–14; P = 0.004, t-test, two-tailed). anova ICC were 0.61, 0.71, 0.92, 0.89, 0.84, 0.67 and 0.82 for polyspike phase amplitude, polyspike duration, slow wave phase amplitude, slow wave phase duration, regularity, stereotypy and postictal suppression, respectively. Less-than-satisfactory reliability was observed for polyspike phase amplitude and stereotypy.

Table 2 shows a comparison of ictal EEG parameters between sine and pulse wave groups. Only polyspike phase duration differed between the two groups, being significantly longer in the sine wave group.

Table 2.  Comparison of ictal electroencephalography characteristics between sine wave and pulse wave electroconvulsive therapy
 Sine wave ECT (n = 16)Pulse wave ECT (n = 14)P
MeanSDMeanSD
  1. P <0.05; t-test (two-tailed).

  2. ECT, electroconvulsive therapy; SD, standard deviation.

Polyspike phase duration (s)8.45.44.72.30.02
Polyspike phase amplitude (µV)132.561.799.532.30.08
Slow wave phase duration (s)46.59.460.829.40.10
Slow wave phase amplitude (µV)528.1131.1466.1138.60.22
Regularity (0–6)4.50.864.40.830.66
Stereotypy (0–3)1.670.421.860.440.25
Postictal suppression (0–3)1.950.561.70.520.26

Each ictal EEG characteristic was individually examined through multiple regression predicting post-treatment HRSD, always controlling for mode of stimulation (sine or pulse wave) and baseline HRSD. Only postictal suppression emerged as a significant predictor (Table 3).

Table 3.  Multiple regression predicting post-treatment Hamilton Rating Scale for Depression, controlling for mode of stimulation (sine or pulse wave) and baseline Hamilton Rating Scale for Depression
 βSEP
  1. P <0.05; Adjusted R2 = 0.13, F = 13.9, P <0.01.

  2. None of the other six electroencephalography parameters similarly entered in multiple regression was significant.

  3. HRSD, Hamilton Rating Scale for Depression; SE, standard error of the mean.

Baseline HRSD0.300.05<0.01
Sine wave or pulse wave0.180.670.03
Postictal suppression−0.150.340.01

DISCUSSION

When ictal EEG parameters were compared between BL sine and pulse wave ECT, only polyspike phase duration was significantly longer in the former mode. This finding is inconsistent with previous results, which have shown that higher stimulus dosage results in better therapeutic efficacy and shorter polyspike duration.1,7,12 The significantly longer polyspike phase duration for BL sine wave ECT compared to pulse wave ECT may be associated with cognitive impairments and needs be examined in a future study.

The authors also found that postictal suppression was significantly associated with post-treatment HRSD when baseline HRSD and mode of stimulation were controlled for. This finding is in accordance with those of previous studies, which have shown a relationship between postictal suppression and therapeutic outcomes.4–6 Evidence also indicates that a higher amplitude of slow wave activity is associated with therapeutic effects,15–16 but the authors could not show this in the current study. Regarding the effectiveness of ECT and EEG features, for example, postictal suppression, Sackeim et al. hypothesized that postictal inhibitory response to the seizure, rather than seizure itself, is therapeutic.17 In effective ECT, postictal suppression, prefrontal slowing following the ECT course, diminished cerebral blood flow (CBF) and metabolism suggest that relative enhancement in prefrontal inhibition is involved in the mechanism of action for ECT.18 The authors have also reported decreased regional CBF in the left medial frontal area and left limbic region after successful ECT.19 As effective ictal EEG could be differentiated from noneffective ictal EEG in RUL ECT, the suggestion has been made that ictal EEG parameters might be utilized to guide adequate stimulus dosage and, therefore, optimize therapeutic effectiveness.3,20,21 However, even in RUL ECT, EEG features reportedly began to saturate when ECT voltage dose was only modestly above stimulus threshold. The window for electrical dosage effect on EEG seizure expression was, therefore, considerably narrower than that for efficacy.6 Conversely, ictal EEG seizure in BL pulse wave ECT could be differentiated between low and high stimulus doses using computed fractal geometry methods.22 Although in this study the authors could not examine the electrical dosage effect in ictal EEG parameters between BL sine and pulse wave ECT, the correlation between ictal EEG parameters and electrical dose in BL pulse wave ECT is thought to be a matter of clinical importance.

Several caveats should be noted, including: (i) a greater proportion of inadequate EEG in the sine wave group; (ii) lower agreement of anova ICC for polyspike phase duration, polyspike phase amplitude and stereotypy, indicating that these parameters were ambiguous and difficult to define; (iii) change of anesthetic agents for sine and pulse wave ECT by the anesthesiologists worked with (however, in a randomized controlled trial, both sevoflurane and propofol groups exhibited equally good seizures and the sevoflurane group displayed slightly better morphology, which the authors failed to differentiate from)23; (iv) some patients in both groups were given antidepressants, minor tranquilizers such as benzodiazepines, antiparkinson drugs and major tranquilizers, and the psychotropics they received might have influenced seizure threshold, although the degree of effects on ictal seizure expression is unclear; and (v) the current study was not a randomized controlled trial and may, therefore, be subject to unknown biases.

This study adopted manual ratings for ictal EEG expression, as manual ratings correlate with computerized analysis of ictal EEG parameters.6 If ictal EEG parameters in BL ECT had some potential as a useful marker for guiding stimulus dosage, psychiatrists would have to manually rate ictal EEG parameters in the clinical scene and then adjust stimuli dosage accordingly. Explicit criteria for rating ictal EEG parameters, including stereotypy, regularity and postictal suppression, will, therefore, be required to improve interrater reliability of ictal EEG,24,25 and further ictal EEG parameters or other physiological markers need to be explored as clinically useful tools to adjust electrical dose.

In conclusion, when BL sine and pulse wave ECT were compared in terms of ictal EEG parameters, polyspike phase duration was significantly longer in sine wave mode. When baseline HRSD score and mode of ECT were controlled for, postictal suppression predicted around 13% of the variance in treatment response of cases. Further accumulation of clinical experience with BL pulse wave ECT is required to detect ictal EEG parameters or other physiological markers that could usefully predict and guide effective administration of ECT for depression.

Ancillary