|
|
| |||||||||||||||||||||||||||||||
Background The efficacy of electroconvulsive therapy in major depression is established, but the importance of the electrical dosage and electrode placement in relation to efficacy and side effects is uncertain.
Methods In a double-blind study, we randomly assigned 96 depressed patients to receive right unilateral or bilateral electroconvulsive therapy at either a low electrical dose (just above the seizure threshold) or a high dose (2.5 times the threshold). Symptoms of depression and cognitive functioning were assessed before, during, immediately after, and two months after therapy. Patients who responded to treatment were followed for one year to assess the rate of relapse.
Results The response rate for low-dose unilateral electroconvulsive therapy was 17 percent, as compared with 43 percent for high-dose unilateral therapy (P = 0.054), 65 percent for low-dose bilateral therapy (P = 0.001), and 63 percent for high-dose bilateral therapy (P = 0.001). Regardless of electrode placement, high dosage resulted in more rapid improvement (P<0.05). Compared with the low-dose unilateral group, the high-dose unilateral group took 83 percent longer (P<0.001) to recover orientation after seizure induction, whereas the combined bilateral groups took 252 percent longer (P<0.001). During the week after treatment, there was three times more retrograde amnesia about personal information with bilateral therapy (P<0.001). There were no differences between treatment groups in cognitive effects two months after treatment. Forty-one of the 70 patients who responded to therapy (59 percent) relapsed, and there were no differences between treatment groups.
Conclusions Increasing the electrical dosage increases the efficacy of right unilateral electroconvulsive therapy, although not to the level of bilateral therapy. High electrical dosage is associated with a more rapid response, and unilateral treatment is associated with less severe cognitive side effects after treatment. .
Positioning electrodes over the right hemisphere causes less severe cognitive side effects than bilateral placement7,8. Despite many trials, however, the relative efficacy of right unilateral and bilateral electroconvulsive therapy is controversial2,9,10. Some studies have found superior efficacy with bilateral therapy,11,12,13 whereas others have reported equivalent efficacy14,15,16. Given this uncertainty, the American Psychiatric Association Task Force on Electroconvulsive Therapy recently recommended that electrode placement be determined on a case-by-base basis2.
Until recently, the standard practice has been to administer the same electrical dose to virtually all patients. Among patients, however, the seizure threshold, defined as the minimal electrical intensity needed for a generalized seizure of a specified minimal duration, appears to vary by approximately 40-fold6,17. When a fixed dose is used, many patients receive stimulation grossly above the threshold, and cognitive side effects are probably increased2,18,19.
In a study using an electrical intensity just above each patient's seizure threshold, we found that titrated, low-dose bilateral electroconvulsive therapy had a powerful antidepressant effect but that low-dose right unilateral therapy was ineffective13. These results led to the hypothesis that the efficacy of right unilateral electroconvulsive therapy depends on the electrical dose20,21. Other research has raised the possibility that for both unilateral and bilateral electroconvulsive therapy, a higher electrical dose leads to a more rapid clinical response3,16,22. We report a prospective, randomized, double-blind trial of the effects of both stimulus intensity and electrode placement on the efficacy and safety of electroconvulsive therapy in 96 patients.
Methods
Patients
The study patients met the Research Diagnostic Criteria23 for major depressive disorder, had pretreatment scores of 18 or higher on the 24-item Hamilton Rating Scale for Depression (HRSD),24 and gave written informed consent. All had clinical indications for electroconvulsive therapy that included a lack of response to previous treatment with antidepressant medication, intolerance to such medication, a particular need for a rapid clinical response due to conditions such as inanition, psychosis, or suicidality, or a history of clinical response to electroconvulsive therapy2. Patients were excluded if they had a history of schizophrenia, schizoaffective disorder, other functional psychosis, rapid-cycling bipolar illness, organic mental syndrome, neurologic illness, alcohol or other drug abuse within the previous year, electroconvulsive therapy within the previous six months, or severe medical illness. All the patients had normal results of routine hematologic and biochemical tests, including thyroid-function tests, chest roentgenography, and electroencephalography. The 31 patients who had computed tomography or magnetic resonance imaging of the brain all had normal findings. All the patients were inpatients at the New York State Psychiatric Institute, where the institutional review board approved the study.
Of 100 consecutive patients admitted to the study, 4 were considered to have dropped out because they received fewer than five treatments; the reasons were withdrawal of consent (1 patient), the need to institute concomitant psychotropic treatment (1 patient), and intercurrent illness (2 patients). With the exception of lorazepam (up to 3 mg of which was allowed per day), all psychotropic medications were discontinued at least 5 days (mean [±SD], 17 ±8; maximum, 30) before electroconvulsive therapy and until 1 week after the therapy had ended. The mean dose of lorazepam during electroconvulsive therapy was 1.0 ±1.0 mg per day and was similar in the four treatment groups.
Electroconvulsive Therapy
The patients were randomly assigned to treatment groups in blocks of 20. Atropine (0.4 mg), methohexital (0.75 mg per kilogram of body weight), and succinylcholine (0.5 mg per kilogram), all given intravenously, were used as the anesthetic medications. The standard bifrontotemporal (bilateral) placement and the d'Elia (right unilateral) placement5 of electrodes were used. Electroconvulsive therapy was administered three times per week with a square-wave, brief-pulse, constant-current device (MECTA, Lake Oswego, Oreg.). The seizure threshold was quantified at the time of the first and last treatments with the empirical titration procedure2,6. Electrical stimulations were administered at subconvulsive levels of increasingly higher intensity until a generalized tonic-clonic seizure of adequate duration was induced. For low-dose treatment, the electrical intensity that resulted in a generalized seizure in the first session was the one administered at the next treatment. If this intensity proved adequate, a lower intensity was used for the next treatment. In the low-dose groups, this procedure was followed throughout the treatment course. For high-dose treatment, at the second and subsequent treatments each patient received stimulation of an electrical intensity that in units of charge was 2.5 times the threshold identified in the first treatment, except that at the time of the last treatment the seizure threshold was determined again. The frequency of brief pulses (range, 20 to 140 Hz) was the primary electrical variable manipulated, because the device used for electroconvulsive therapy had greater range and sensitivity in the frequency domain than pulse width, current intensity, or the duration of the stimulus. The tourniquet method25 and two channels of prefrontal electroencephalography were used to assess the duration of seizures. In the first five treatments, the criterion for an adequate generalized seizure was at least 25 seconds of tonic-clonic movement or 30 seconds of electroencephalographic seizure activity. After the fifth treatment, the cutoff values were reduced to 20 and 25 seconds, respectively. This change was made to account for the spontaneous decrease in seizure duration that occurs during electroconvulsive therapy17.
Clinical Evaluations
The patients and the clinical evaluation team (a research psychiatrist and a social worker) were unaware of the treatment-group assignments. The team rated each patient on the HRSD two days before the first treatment, at least twice weekly during treatment, within two days after the end of treatment, and one week after the end of treatment. The interrater reliability coefficients for the HRSD scores exceeded 0.98 at all times. Analyses of these scores were based on the mean scores obtained by the raters.
The patients considered to have responded initially to therapy had a decrease of at least 60 percent in their HRSD scores from before treatment to immediately after the final treatment, and they had a maximal score of 16 after treatment. To be classified in the final analysis as having responded to therapy, patients had to maintain this level of improvement for at least one week after electroconvulsive therapy. No minimal or maximal number of treatments was imposed on patients who improved progressively. These patients continued to receive treatment until they were asymptomatic or had reached a plateau in clinical improvement after 2 consecutive treatments; they had a range of 4 to 17 treatments overall. At least 10 treatments were required before a patient was classified as not having responded, with 8 treatments required of patients who had reductions of 20 percent or less in HRSD scores during therapy. These patients then received a course of high-dose bilateral electroconvulsive therapy. The evaluation of efficacy during this crossover phase was made with the same procedures that were used in the randomized phase. Patients who responded to therapy during either phase were followed until relapse, as defined elsewhere,26 or for one year. Interviews to determine HRSD scores were conducted every two weeks for the first three months after electroconvulsive therapy and monthly thereafter. During the follow-up period, virtually all the patients received maintenance pharmacotherapy with antidepressant medications.
Cognitive Evaluations
Short-term neuropsychological effects were assessed at the time of each treatment during the randomized phase. Before each treatment, a 10-item interview evaluating the patient's orientation was administered. The patients then memorized sets of words, geometric shapes, nonsense shapes, and faces that were either neutral or emotionally expressive. After the presentation of each set, immediate learning was assessed by testing the patient's recall and recognition for the word task and by testing the patient's recognition for the shape and face tasks. After the seizure, the recovery of orientation was assessed continuously for 90 minutes. Patients who did not meet the criteria for the recovery of orientation7 were assigned scores of 100 minutes. Once the criteria were met, each patient's retrograde memory was tested for the material learned previously. Twelve equivalent sets of stimuli were available for each task, with the assignment randomized at the time of each treatment7.
Neuropsychological tests focusing on anterograde and retrograde memory were administered before the start of treatment, the day after some treatments, during the week after the end of the randomized and crossover phases, and two months later. The tests included modified versions of the Mini-Mental State Examination,27 recognition memory of paired words and paired faces,28 the Selective Reminding Test,29 and an expanded Autobiographical Memory Interview8. The paired-word and paired-face tasks assessed both immediate learning and memory after a four-hour delay. Recall after a two-hour delay was included in the Selective Reminding Test. Patients evaluated their own memory function by completing the Squire Subjective Memory Questionnaire30.
Statistical Analysis
The results are expressed as means ±SD. All statistical tests were two-tailed. Analyses of electrical variables and measures of seizure duration were performed on logarithmically transformed values, and arc-sine transformations were applied to scores reflecting percentages of change. The treatment groups were compared with respect to base-line measures with two-way analyses of variance for continuous variables and with log-linear analyses for dichotomous variables. As an omnibus test of short-term differences in efficacy, a multivariate analysis of covariance was conducted of the HRSD scores, with electrode placement and dosage condition as between-subject factors, time (before electroconvulsive therapy, after six treatments, immediately after treatment, and one week after treatment) as a repeated-measures factor, and the patient's age as a covariate. This was followed by analyses of covariance of the HRSD scores obtained at the most recent three time points, with age and HRSD scores before treatment used as covariates. Post hoc Newman-Keuls comparisons of covariate-adjusted means were used to identify pairwise group differences. Chi-square tests were used to compare rates of clinical response. To examine the speed of clinical response, the number of treatments required for reductions from base line of 40, 50, 60, and 70 percent in HRSD scores to be achieved was analyzed by analysis of variance. Multiple regression analyses were used to determine whether either the dosage relative to threshold or the absolute electrical dose administered contributed to efficacy. In these analyses, dosage (low or high) and mean absolute dose per treatment served as predictors of the percentage change in HRSD scores. Survival analyses, with both regression models and life-table methods, provided tests of differences between treatment groups with respect to the likelihood and speed of relapse. These analyses were conducted both with patients who responded to crossover treatment considered as a separate group and with these patients included in the group that had responded to high-dose bilateral electroconvulsive therapy. In the regression models, age and HRSD scores one week after electroconvulsive therapy were covariates.
Similar methods were used to examine how the conditions of treatment affected cognitive performance. For short-term measures of orientation recovery and retrograde memory, average scores were derived for each patient, with the first and last treatments excluded because they involved low dosages in all groups. Scores for the percentage of amnesia were computed by subtracting the scores for delayed learning from the scores for immediate learning and dividing the difference by the scores for immediate learning.
Results
Characteristics of Patients and Treatment Measures
The demographic and clinical characteristics of the treatment groups were similar (P>0.05 for each comparison) (Table 1). The doses of methohexital and succinylcholine and all the measures of seizure duration were also similar among all groups (Table 2). In a replication of previous findings,6 the initial seizure threshold was higher in the patients treated with bilateral therapy than in those treated with right unilateral therapy. Consequently, the bilateral-therapy groups exceeded the unilateral-therapy groups, and the high-dose groups exceeded the low-dose groups, on all measures of stimulus intensity. The increase in seizure threshold during electroconvulsive therapy was greater with low-dose bilateral therapy than with low-dose unilateral therapy, resulting in a smaller proportional difference in charge between low-dose and high-dose groups when therapy was given bilaterally rather than unilaterally.
|
|
Both electrode placement (P = 0.001) and electrical dosage (P = 0.005) strongly influenced clinical outcome (Figure 1). After the sixth treatment, the antidepressant response in the low-dose unilateral group was significantly worse than in the high-dose unilateral group (P = 0.01), the low-dose bilateral group (P<0.05), and the high-dose bilateral group (P<0.001). At this time, the HRSD score was higher in the low-dose bilateral group than in the high-dose bilateral group (P<0.05). Immediately after the end of treatment, the low-dose unilateral group also had a higher HRSD score than the other groups (P<0.01). One week after electroconvulsive therapy, the low-dose unilateral group continued to have a higher HRSD score than the high-dose unilateral group (P<0.05) and the two bilateral groups (P<0.001). At this time, the high-dose unilateral group had higher HRSD scores than the high-dose bilateral group (P<0.05).
|
0.001). The final response rate in the low-dose unilateral group (17 percent) was also lower than those in the high-dose unilateral group (43 percent, P = 0.054), the low-dose bilateral group (65 percent, P = 0.001), and the high-dose bilateral group (63 percent, P = 0.001). Early relapse (within one week) was most common in the high-dose unilateral group. The final response rate tended to be lower in this group than in the combined groups receiving bilateral electroconvulsive therapy (P = 0.099).
|
The low-dose unilateral group was excluded from the analyses of speed of response, since too few patients in this group had improvement. Except for the 70 percent improvement cutoff, the remaining three treatment groups differed in each analysis (P<0.05). Approximately 1.5 more treatments were administered in the low-dose bilateral group than in either of the high-dose groups in order to achieve reductions in HRSD scores of 40, 50, and 60 percent (P<0.05 for each comparison).
Absolute Electrical Dose or Dose Relative to Threshold
Among patients treated with right unilateral electroconvulsive therapy, improvement in symptoms was associated with dosage group after the sixth treatment (P = 0.008), immediately after treatment (P = 0.005), and one week later (P = 0.08). Clinical improvement was not related to the absolute electrical dose at any point. In the case of bilateral therapy, dosage group was associated with improvement after the sixth treatment (P = 0.007), but not after that point. Again, the absolute electrical dose had no relation with clinical outcome. Therefore, efficacy was influenced by whether the electrical dose exceeded the patient's seizure threshold, rather than by the absolute dose administered.
Crossover Phase
Of the 40 patients who did not respond to therapy and were not randomly assigned to high-dose bilateral therapy, 34 (85 percent) completed the crossover phase. They received 8.5 ±3.2 high-dose treatments bilaterally. Their HRSD scores were 35 ±9 before any treatment, 27 ±10 after the randomized phase, and 9 ±6 one week after the crossover phase; 79 percent were classified as having responded. Improvement in symptoms was equivalent regardless of the randomized assignment, and 12 of the 15 patients who had not responded to low-dose unilateral therapy (80 percent) responded after crossover treatment.
Relapse
Of the 73 patients who responded to randomized or crossover treatment (46 and 27 patients, respectively), 70 were followed for up to one year. Three patients declined to participate in monitoring for relapse and were lost to follow-up. The conditions in which electroconvulsive therapy was provided were unrelated to the likelihood or the speed of relapse (Figure 2). Forty-one of the 70 patients (59 percent) met the criteria for relapse; 78 percent of the relapses occurred within the first six months.
|
Eight patients were excluded from short-term neuropsychological testing because they had visual or auditory deficits, difficulties with language, or impaired orientation at base line. At the postictal assessments, cognition was generally more impaired in the groups receiving bilateral or high-dose therapy (Table 4). The rare instances of prolonged disorientation occurred exclusively with bilateral electroconvulsive therapy (P = 0.008). Time to the recovery of orientation was a function of electrode placement (P<0.001), electrical dosage (P = 0.001), and the interaction of the two (P = 0.03). The recovery of orientation took 83 percent longer (P<0.001) with high-dose unilateral electroconvulsive therapy than with low-dose unilateral therapy, whereas the combined bilateral-therapy groups took 252 percent longer (P<0.001). Bilateral therapy resulted in more profound amnesia than unilateral therapy with regard to the recall and recognition of words (P<0.001). Bilateral therapy and high dosage both resulted in more impaired performance on the shape tests and the test of emotionally expressive faces (P<0.05). For these tasks, the amnesia produced by high-dose unilateral electroconvulsive therapy exceeded that produced by low-dose unilateral therapy (P<0.05). For the cognitive measures that showed dosage effects, regression analyses indicated that dosage was the determining factor, rather than the absolute dose administered (data not shown).
|
|
Long-Term Cognitive Side Effects
Sixty-six patients were evaluated at the two-month follow-up, of whom 54 (82 percent) had responded to electroconvulsive therapy during either the randomized or the crossover phase. There were no effects related to treatment conditions in which electroconvulsive therapy was administered (data not shown). For these 66 patients, scores improved significantly from the pretreatment values on the Mini-Mental State Examination (P<0.001) and for paired-word learning (P = 0.02), paired-face learning (P = 0.03), and learning on the Selective Reminding Test (P = 0.03). Forgetting scores on these anterograde tests were unchanged relative to the pretreatment scores. Scores on the retrograde Autobiographical Memory Interview indicated no change in the consistency of recall relative to the post-treatment value. Subjective evaluations of memory were markedly improved at follow-up, in relation to the pretreatment scores (P<0.001), and were improved further in relation to those after treatment (P = 0.04).
Discussion
This study demonstrates that electrical dose affects the efficacy of electroconvulsive therapy. Despite the advantage of diminished cognitive side effects,7 the extremely poor efficacy of low-dose right unilateral therapy indicates that it should be abandoned. At all times examined, the high-dose right unilateral group had an antidepressant response superior to that of the low-dose right unilateral group. Dosage effects may account for the inconsistent results of previous research comparing right unilateral and bilateral electroconvulsive therapy10,11,12,13,14,15,16. Dosage also influenced the speed of clinical response with both right unilateral and bilateral therapy. Both high-dose groups had more rapid improvement than the low-dose groups. This is of particular importance, since electroconvulsive therapy is frequently used when rapid improvement is needed.
One week after electroconvulsive therapy, there were indications that high-dose right unilateral therapy was less efficacious than either of the bilateral-therapy approaches. The choice for the high-intensity conditions of a dosage 2.5 times the initial seizure threshold was arbitrary. The efficacy of right unilateral electroconvulsive therapy might be enhanced by a further increase in electrical dose, yet retain the advantages of unilateral therapy with respect to cognitive side effects, because electrode placement had more pronounced and persistent effects on cognition than did dosage. For virtually all objective cognitive measures during the week after treatment, bilateral therapy resulted in greater deficits than did right unilateral therapy, whereas adverse effects of the higher-stimulus dose were not evident. Few studies have demonstrated differences among types of electroconvulsive therapy in cognitive side effects two or more weeks after therapy8,14,19. Likewise, at the two-month follow-up in this study, cognitive function was similar in all four treatment groups. As compared with the pretreatment values, the follow-up results revealed enhanced performance on measures of global cognitive function and learning and in the subjective evaluation of memory, without change in the capacity to retain information. This pattern reflected improvement in the objective and subjective cognitive deficits associated with severe depression. There is substantial evidence that major depression is characterized by deficits in the acquisition of new information (learning), but not by disturbances in the retention of information (forgetting)19,28.
The effects of stimulus intensity on efficacy, speed of response, and short-term cognitive side effects were not associated with the absolute electrical dose administered, but with whether or not the dosage substantially exceeded the seizure threshold. These findings concur with basic research suggesting that the seizure threshold acts as a filter for many of the neurobiologic and behavioral effects of electrical stimulation32,33. Our findings indicate that a patient with an initial threshold of 60 millicoulombs (mC) who is treated with a 150-mC stimulus is likely to respond more quickly -- or to respond at all, if treated with right unilateral electroconvulsive therapy -- than a patient with a threshold of 300 mC who is treated with an electrical dose just above that threshold. Similarly, the first patient is likely to have more short-term cognitive side effects. This underscores the desirability of adjusting the stimulus intensity in relation to each patient's seizure threshold2. Case reports have raised the possibility that stimulation at subconvulsive levels that is used to identify the seizure threshold increases the risk of cardiovascular complications34,35,36. In a recent study of depressed patients with and without preexisting cardiac disease, however, cardiovascular complications were slightly more common after treatment sessions involving only stimulation at convulsive levels, as compared with those in which stimuli were also administered at subconvulsive levels37.
As in other recent follow-up studies of patients who responded to electroconvulsive therapy,38,39 the relapse rate was high despite conventional maintenance pharmacotherapy. Relapse is a critical problem. In other work, we found that failure to respond to antidepressant medications during the acute phase of illness was a potent predictor of relapse after therapy26. It is common for patients to receive the same class of antidepressant medication after electroconvulsive therapy that proved ineffective during treatment of the acute episode, but this practice should be reconsidered.
Low-dose right unilateral electroconvulsive therapy resulted in generalized seizures of adequate duration, but with little benefit. This contradicts the view that the duration of seizures provides a valid index of maximally effective treatment1,3,4 and indicates the need to identify more useful markers. It also contradicts the longstanding principle that the generalized seizure provides both the necessary and the sufficient conditions for the efficacy of electroconvulsive therapy. It now appears that the generalized seizure may be necessary, but not sufficient, to ensure a clinical response. The finding that there are electrical dose-response functions for electroconvulsive therapy offers new avenues to investigate the mechanisms of action of this treatment.
Supported in part by grants (R37 MH35636 and R01 MH47739) from the National Institute of Mental Health.
We are indebted to Dr. R.P. Brown, Dr. P. Decina, Dr. D. Kahn, Ms. B. Kerr, Dr. D. Malaspina, Dr. S. Malitz, Dr. L. Rosnick, Ms. F. Ross, Ms. J. Silverman, and the staff of the General Clinical Research Service, without whose help this study could not have been carried out.
Source Information
From the Department of Biological Psychiatry, New York State Psychiatric Institute, and the Department of Psychiatry, College of Physicians and Surgeons, Columbia University -- both in New York.
Address reprint requests to Dr. Sackeim at the Department of Biological Psychiatry, New York State Psychiatric Institute, 722 W. 168th St., New York, NY 10032.
References
-adrenergic blockade. Am J Psychiatry 1984;141:298-300.
| |||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
