What was learned: studies by the consortium for research in ECT (CORE) 1997–2011


  • ‘This paper is commented on by Tom G. Bolwig (Acta Psyciatr Scand 2014;129:415–416.)’



To review the findings of the four-hospital collaborative studies of electroconvulsive therapy (ECT) in unipolar depressed patients known as CORE between 1997 and 2011. Unipolar depressed patients were treated with bilateral ECT, and on remission were randomly assigned to a fixed schedule continuation ECT or to combined lithium and nortriptyline for 6 months. A second study compared three electrode placements in unipolar and bipolar depressed patients.


Nineteen published reports were reviewed. The findings are compared with those of a parallel multi-hospital study of ECT led by a Columbia University Collaboration (CUC) team that studied right unilateral ECT in a similar population with similar inclusion/exclusion and remission criteria. Successful ECT was followed by placebo, nortriptyline alone, or combined lithium, and nortriptyline.


Relapse rates after remission were similar with fixed schedule ECT as with medications. Predictors of outcome (psychosis, suicide risk, polarity, melancholia, atypical depression, age) and technical aspects (electrode placement, seizure threshold, speed of response) are discussed,


The findings offer criteria to optimize the selection of patients, the technique, and outcome of ECT for unipolar and bipolar depressed patients. Continuation ECT is an effective alternative to continuation treatment with lithium and nortriptyline. Bilateral electrode placement is more efficient than alternative placements. ECT relieves both bipolar and unipolar depression.


  • Continuation treatments, either continuation electroconvulsive therapy (ECT) or combined nortriptyline and lithium, sustain 6-month remission rates equally after successful courses of ECT in depressed patients.
  • Unipolar and bipolar depressive mood disordered patients are relieved equally by ECT.
  • ECT with bitemporal electrode placement is more efficient than treatment with right unilateral placement.
  • Seizure threshold determinations to determine electricity dosing are neither necessary nor useful for effective ECT.
  • ECT rapidly relieves active suicide risk.


  • A prescribed fixed treatment schedule for continuation ECT was inefficient. A flexible “as required by symptoms” schedule would have improved outcomes.
  • Determining the seizure threshold in the first treatment unnecessarily reduced the efficiency of the treatment course. Seizure threshold (ST) calibration is not needed for effective dosing.
  • Failure to examine neuroendocrine measures during the study lost an opportunity for greater understanding of the mechanism of action of ECT.


In 1934, the Hungarian neuroscientist Ladislas Meduna induced grand mal seizures by chemical means in hospitalized psychotic patients and reported sufficient relief for some to leave the hospital. His reports in 1935 and 1937 encouraged a world-wide adoption of the treatments [1]. Improvements in technical methods followed rapidly with electrical inductions of seizures introduced by the Italian clinicians Ugo Cerletti and Luigi Bini in 1938 [2]. Electrical inductions of seizures were more efficient and more reliable, rapidly replacing chemical inductions. By the mid-1950s, the risks of patient fear and confusion, bone fractures, and missed seizures were mitigated by barbiturate sedation, supported ventilation with oxygen, and muscle relaxation with succinylcholine.

In the first treatments, electrodes were placed bilaterally, usually one electrode over each temple. Testing electrode placement unilaterally over the non-dominant brain hemisphere minimized the immediate postseizure complaints of confusion and delirium. Systematic studies of these placements by randomly assigning patients referred for electroconvulsive therapy (ECT) to treatments with either unilateral (RUL) or bilateral (BT) placements found that RUL placements were indeed accompanied by lesser immediate effects on cognitive tests and on electrophysiology measures, but at the cost of lower and slower remission rates. RUL treatments were accompanied by greater numbers of missed seizures, required more treatments for a sustained benefit in a treatment course, and were followed by more frequent relapses. One to 3 months after the ECT course the cognitive advantage of RUL over BT treatments was lost with higher relapse rates [3, 4].

Whether a grand-mal seizure was necessary for the treatment's benefits led to UK government-sponsored studies of real and simulated ECT. Seizure inductions were again found essential to the treatments' benefits and that RUL placements were less efficient than BT placements [5].

The advantage in lesser immediate memory complaints for RUL placements was so seductive despite lesser efficacy that treatment guidelines in the 1980s recommended that depressed patients be initially treated with RUL placement. If no measurable benefit was seen with four RUL treatments, the electrode placement was to be switched to bilateral [6]. Treatments with RUL placements alone were effective in up to 40% of depressed patients, but for the majority the number of treatments was greater than for those initially treated with BT placement [7].

One would have thought that such findings in repeated studies would have settled the question, but public outcries against ECT were so strong that the NIMH supported additional cognitive studies. Richard Weiner at Duke University examined EEG and memory test changes finding little persistent difference between the electrode placements and making no new recommendations for optimizing treatment outcomes [7].

In the belief that the intensity of the electric current was a basis for persistent cognitive complaints, Harold Sackeim and his colleagues at Columbia University sought to minimize the electrical energies to induce a seizure by first calibrating a ‘seizure threshold’ (ST). In each patient's first treatment, the instrument's energy was set below the amount anticipated to induce a seizure. Stimulations were repeated in an increasing fixed dosing schedule until a grand mal seizure was elicited; that energy level was labeled the ST. The energies in subsequent treatments were delivered in multiples of this calibrated ST. Seizures induced with energies at 1.5 and 2.5 times ST with RUL placements were grossly inefficient with many seizures of inadequate quality by EEG and duration criteria. [8]. To approximate the clinical efficacy rates of BT placements (with dosing at 1.5 × ST) required energies from 6 to 12 × ST. At these energies, the anticipated advantage in cognitive tests and the justification for this electrode use was lost [9].

Decline and then revival of ECT use

From the 1940s to the 1970s, ECT was commonly an out-patient office procedure. The numbers of treatments were not prescribed in advance, and patients were sustained in the community by treatments given when symptoms suggested the need. Treatments were administered during evenings and on Saturdays to enable patients to work. The success of ECT and its widespread use was so great that the exponential rise in the number of psychiatric hospital beds reached a peak in the mid-1950s as more and more patients were effectively treated without in-patient hospital care [1].

As the reduction of depressive moods by tricyclic antidepressants, psychotic thoughts by neuroleptic drugs, and manic states by lithium were recognized, medications were assumed to also sustain the benefits of ECT. The use of continuation ECT and the numbers of facilities and trained practitioners fell sharply.

Descriptions of medication treatment resistant illnesses led some practitioners to recall ECT and test its use against that of medications. A new impediment to ECT practice then appeared. The demarcation of professional medical departments in hospitals led to demands that muscle relaxation by succinylcholine and sedation by anesthetic agents be done by specialists certified in anesthesiology and not by psychiatrists. ECT was brought into hospital settings, increasing the complexity of staffing, scheduling of treatments, and expense.

Electroconvulsive therapy use was also impeded by the 1978 recommendations of the American Psychiatric Association Task Force on ECT that the patient's consent for ECT acknowledge a maximum number of treatments [10]. Although interest in ECT was revived, the specification of the maximum numbers of treatments, preferential use of continuation medications, and the emphasis on RUL placements led to concerns about relapse rates when medications were the mainstay of continuation treatments.

In a follow-up of delusional depressed patients effectively treated either with ECT or with combined neuroleptic and tricyclic antidepressant, Aronson and his colleagues reported relapse rates greater than 80% within 1 year after hospital discharge regardless of continuation treatments [11]. Such reports challenged the practice of early discharge from hospital with medication continuation and the need for reassessment of ECT continuation treatments was recognized.

Collaborative ECT studies

In the 1990s, two multi-site collaborative programmes were supported by the NIMH to examine ways to sustain the ECT benefit. A multi-site collaborative study led by Columbia University faculty (CUC Study) examined three continuation medication treatments – placebo, the tricyclic antidepressant nortriptyline alone, and the combination of nortriptyline and lithium – after a successful course of ECT in patients with unipolar major depression. At a site visit, the NIMH Review Committee members suggested a fourth arm, that of continuation ECT included. The investigators thought that such addition would make their study unwieldy. The project was funded as requested. The NIMH then encouraged a new study to compare continuation treatments after ECT by the combination of medications or ECT. This study was undertaken by collaborating sites known as CORE, with the study arm, the same CUC nortriptyline-lithium combination treatment schedule and the other, a fixed schedule of ECT. Patient selection was limited to unipolar major depressed patients. The two studies used the same inclusion-exclusion criteria, the same evaluation procedures and rating scales, and the same blood level monitoring for lithium and nortriptyline. A singular difference was made in the electrode placement with bitemporal placement (BT) adopted in the CORE study and right unilateral placement (RUL) in the CUC study.

While the CORE study was ongoing, questions arose as to whether limiting the study to unipolar depressed patients was justified. Many depressed patients referred for ECT met the criteria for bipolar depression and were rejected from the study, yet went on to a course of ECT and responded well.

At the same time, the merits of bifrontal electrode placement (BF) became an interest and its study was incorporated within a new protocol. Patients referred for ECT who met criteria for major depression, both unipolar and bipolar types, who agreed to participate, were randomly assigned to treatments with BT, RUL, or BF electrode placements.

Aims of the review

The Consortium for Research in ECT (CORE) group has published 19 reports between 1997 and 2011 scattered in different journals; this review discusses the main findings and compares the observations with the Columbia University Collaboration CUC studies.

Material and methods

The details of the first CORE study were published in 2006 [12] and those of the second in 2010 [13, 14].

CORE I: Treatments were optimized by specific protocols for anesthesia, seizure monitoring, concurrent medications, and randomization of continuation treatments. A baseline HAMD24 score of 21 or higher was required for inclusion [15]. Remission was defined as a decrease by at least 60% from the baseline and scores ≤10 on the HAMD24 after two consecutive treatments. Continuation treatments for 6 months were either a fixed ECT schedule or a dosing schedule for combined nortriptyline and lithium, monitored and adjusted by periodic blood level measurements, identical to the schedule of the CUC study. The fixed ECT schedule was one ECT weekly for 4 weeks; then one every other week for four treatments, and then once monthly.

CORE II: Patients referred for ECT who met criteria for the diagnosis of major depression, both unipolar (UP) and bipolar (BP) types, were randomized to treatments with either BT, BF, or RUL placements. The seizure threshold was determined at the first treatment and subsequent treatments were set at 1.5 × ST for BF and BT placements and at 6 × ST for RUL.The diagnoses were established using DSM (SCID-1) criteria.


Continuation treatment after a successful course of ECT

Optimized ECT commonly elicits remission rates greater than 80% in severely depressed patients but relapse is frequent when the treatment ends. The CUC study reported a 6-month relapse rate greater than 84% when placebo was offered, 60% with nortriptyline alone, and 40% with combined lithium and nortriptyline [16]. The CORE study confirmed a 6-month relapse rate of 32% for the nortriptyline-lithium combination [12]. In the C-ECT group, 46.1% continued in disease remission, 37.1% experienced illness relapse, and 16.8% dropped out of the study. In the C-Pharm group, 46.3% continued in illness remission, 31.6% experienced relapse, and 22.1% dropped out of the study.

A course of treatments, to be effective with sustained benefit, requires repeated seizures. The expectation that a clinician can prescribe a fixed number of treatments on a predetermined schedule is no longer realistic. A fixed schedule of continuation ECT was tested in the CORE study, and while as effective as the optimized combined medication treatment, it was able to sustain less than 40% of the treated ECT patients in the community for 6 months.

A prescribed treatment schedule is inconsistent with the reality that the benefits are not like those of penicillin in bacterial pneumonia that clears the infectious illness but is more like those of insulin in diabetes where dosing varies with symptom severity and must be continued. This conclusion is consistent with the report of a 1996 Task Force of the Association for Convulsive Therapy reviewing the experience with continuation ECT that concluded that flexible continuation treatments on symptom demand sustained ECT relief best and a planned 6-month continued care served the most patients [17].

Predictors of treatment outcome


Of the first 253 study patients, 77 met criteria for psychosis and 176 did not, using the presence of delusions or hallucinations as the criterion [18]. With an overall remission rate of 87%, the remission rate for psychotic patients was 95% and for the non-psychotic 83%, a significant statistical difference. The presence of psychosis in a severely depressed patient predicted a good response to ECT.

Whether psychotic depression is a distinct syndrome from non-psychotic depression is much debated [19]. In almost all published studies, the more severely ill, those who met criteria for psychotic or delusional illness, responded more rapidly to effective ECT than did the non-psychotic depressed. Considering the presence of psychosis as a marker of severity of illness, we would expect that the psychotic patients would require more treatments for effective response. The evidence finds otherwise, the psychotic depressed are more responsive than the non-psychotic.

That psychotic depression was a distinct mood disorder was recognized by Glassman and co-workers when they reported that such patients failed to respond to well-monitored tricyclic antidepressant (TCA) treatments but did respond to ECT [20]. The recognition of psychosis is difficult, however. In the CUC study, of 52 patients referred for ECT only two had received at least one adequate medication trial by the referring physicians [21]. The definition of psychotic depression using DSM-IV criteria is imprecise, and the role of psychosis in predicting outcome varies with the criteria for psychosis and the behavior rating scale that is used. The CORE criteria and conclusion are consistent with numerous reports [22, 23] but not with others [24-28].

The debate has continued with the recognition that patients with psychotic depression often have abnormal adrenal functions and are unresponsive to the psychotherapies, placebo, and SSRI medications that are often prescribed [29]. A useful image of psychotic depression is as a clinical entity meeting criteria for melancholia [19].

Suicide risk

The benefit of ECT in reducing suicide risk is widely acknowledged by national psychiatric associations in Great Britain and the United States although supporting studies are few [30]. One approach is to examine the changes in Item 3 of the HAMD24 that evaluates the presence and severity of suicidal thoughts/action on a 4-point scale, with zero rated for no report of suicide thoughts during the illness progressing to four as a suicide attempt during the present episode.

For the 444 patients referred for ECT in the first CORE study, 118 (26.6%) received the high score of 3 for recent active suicidal thoughts, actions, or gestures, and 13 (2.9%) received the maximum score of 4 for reporting a suicidal event during the current episode. Of the 131 patients in this high expressed suicidal intent group, the rating dropped to zero in 20 (15.3%) after one ECT session; in 50 (38.2%) after three ECT sessions (1 week); in 80 (61.1%) after six ECT sessions (2 weeks); and in 100 (76.3%) after nine ECT sessions (3 weeks) [31]. In the second CORE study, item 3 of the HAMD24 was similarly examined [14]. In the pooled sample for the two CORE studies (n = 746), 250 (34%) scored high suicide ratings of 3 or 4 at baseline. Nearly half (49.6%) were no longer suicidal (item 3 = 0) after ECT #6. Overall, 63.2% of the high suicide risk patients had complete resolution of suicide thoughts with ECT. Only 10 (4.0%) patients of the two cohorts with high baseline score remained at high risk after the ECT course (Knapp R, pers. comm.).

The parallel CUC study using the same rating scale item of the HAMD in 148 patients reported a similar resolution of suicide intent [32]. The overall average score on item 3 was 1.8 at baseline and was reduced to 0.1 in 72 clinical responders and to 0.9 in 76 non-responders. For the total sample, there was a greater decrease in the suicide item scores than the decrease in overall Hamilton depression scale scores.

Although the benefits of ECT in relieving suicide risk are widely acknowledged, frequent reports of successful suicides, especially among melancholic elderly patients, reflects the persistent position of ECT as a last resort in treatment algorithms. The rapid relief of suicide intent argues for an early referral for ECT in those with suicide risk.

Polarity of mood disorder

Of 220 patients in the CORE II study, 170 (77.3%) were classified as unipolar and 50 (22.7%) as bipolar depression in the intent-to-treat sample [14]. Defining outcome as at least a 50% reduction from baseline in HAMD24, 78.8% of the UP-depressed and 80% of the BP-depressed responded to treatment. After adjustment for baseline value, electrode placement, age, psychosis status, and gender, the odds of responding for BP patients did not differ from that of the UP patients. Further analyses of remission and drop-out rates found no differences.

A common clinical belief in prescribing antidepressant medications to patients with bipolar depression is the risk of precipitation of a manic episode [33]. The CORE studies found no evidence of precipitated mania with ECT. Indeed, the response of bipolar depressed patients to ECT is found to be better than that of unipolar depressed patients, again encouraging a preference for ECT in treating severely ill bipolar depressed patients [34, 35].

The separation of depressed patients ill enough to be referred for ECT into UP- and BP-depressed patients did not affect the outcome of treatment with ECT. The distinction in referral for ECT in UP-depression and non-referral for BP-depression is not valid.


Over the decades of ECT use, elderly patients have been favored for good outcomes, as cited in the clinical adage ‘the older the better’ [3]. In the early years, it was soon learned that younger patients, especially those with thought disorders labeled as ‘schizophrenia’ required more treatments with poorer long-term outcomes [36]. It was among the melancholic, psychotic, and elderly depressed that the benefits were most clearly found. In the CORE study, 253 patients were treated to remission, and age was positively associated with treatment response. The remission rate for those aged ≥ 65 years was 90.2%, for 46–64 years 93.3%, and for those aged ≤45 years 75%. The effects of age were not mitigated by gender, psychosis, race, or baseline severity. The age of onset of illness was a factor in outcome with those who became ill before age 45 had poorer outcomes than those whose illness began after age 65 [37].

Not all studies report age as a predictor to response to ECT [38, 39]. Nevertheless ECT is a favored treatment among the elderly, in part because of its efficacy and their poor response and poor tolerance to known medications.

For many decades ECT has been rejected for use in adolescents and children. The bias resulted from community beliefs that ECT permanently altered brain functions and interfered with learning. In the past two decades, however, as ECT has been tested and found useful in the relief of catatonia in adolescents, and especially in those with autism spectrum disorders, pediatric ECT is better tolerated [40].


The syndrome of severely depressed mood with agitated motor, widespread vegetative, and impaired cognitive symptoms is identified as melancholia [19]. The DSM classification does not recognize melancholia as a distinct syndrome but only as a specifier of two principal mood disorders. It is not identified by a defined numeric code. The label is applied to patients with the loss of pleasure in all, or almost all, activities or lack of reactivity to usually pleasant stimuli and three (or more) of six cited vegetative signs. The specifier is added on the basis of interviews at baseline. Of 489 patients in CORE I, 311 (63.6%) met criteria for melancholic features. The overall remission rate was 68.1%, with higher rates (78.7%) for those who did not meet the melancholia specifier criteria, and 62.1% of those with melancholia specifier [41].

We expected a more rapid and fuller response for those with melancholia, much as we had for those with psychosis, and we were disappointed. Recent studies of melancholia see merit in abnormalities in endocrine measures, both thyroid and adrenal, as essential to the recognition of the syndrome [19, 42]. Cognitive impairment and abnormalities of movement are essential in the recognition by melancholia scholars but are ignored in the DSM label [43]. The absence of these criteria in the DSM definition loses the usefulness of the melancholia specifier as a predictor of treatment choice and outcome.

Atypical depression

Atypical depression is identified by mood reactivity (mood brightening in response to events), hyperphagia or increasing weight, hypersomnia, leaden paralysis (limbs feeling unusually heavy), and sensitivity to interpersonal rejection. The patients commonly have co-morbid anxiety and personality disorders. Of the 489 patients in the CORE I intent-to-treat sample, 36 (7%) met criteria for atypical depression. Their mean age was 10 years younger, the illness onset 10 years earlier, and a higher percentage were women. The number of treatments and percentages that improved were equivalent. The odds of remission were almost three times higher for the atypical compared with the typical group. We concluded that patients with atypical depression do not have poorer outcomes with BT ECT and questioned the cautions against the use of ECT for atypical depression [44].

The conclusions were criticized for the lack of follow-up, the failure to assess the role of accompanying anxiety and personality disorder, and the use of the HAMD24, noting that this scale is a measure of severity of melancholia [45].


In the two CORE studies, 32 patients were black (5%) among the 624 patients examined. Fewer had failed adequate medication trials. They were more likely to have psychotic features, to have greater reductions in HAMD24 scores and higher rates of remission [46]. The data are consistent with the higher ECT response rates of psychotic patients cited earlier and argues against the prejudice by black populations against the use of ECT [47].


Early investigators cited a depressed patient's focus on bodily symptoms as signs of a neurotic depression that predicted poor outcome with ECT. Three items in the HAMD24 are the main signs – 10 (psychic anxiety), 11 (somatic anxiety), and 15 (hypochondriasis). The scores on these items and the combined score of items 11 and 15 were examined for their relation to sustained remission. For the more severely depressed (HAMD24 > 32), no relationship was found; for the less severely depressed (HAMD24 = 21–31), especially those with high scores on items 11 and 15, the sustained remission rates were lower. These findings support the experience that the patients with high degrees of anxiety and somatic symptoms have less successful outcomes with ECT in contrast to the better outcomes in the more severely and psychotic ill [48].

Cognition during continuation treatment

Fears of impaired cognition and memory loss are the principal limitation in the use of ECT. Although the symptoms are transient and are no longer evident in testing 4 weeks after a successful course of ECT, the mantra of ‘memory loss, memory loss’ is repeated daily in the press and in public discourse with any note of ECT.2 We compared the cognitive effects in the C-Pharm and C-ECT treatment cohorts. No differences were found in six cognitive tests at 24 weeks nor in a comparison of the test scores at 12 and 24 weeks, indicating that neither continuation treatment affected the cognitive measures [49].

Technique of treatment

Electrode placement

In the second CORE study, the patients were randomly assigned to one of three electrode placements regardless of their polarity status. Of the 230 patients who completed the outcome criteria, the remission rates at fixed selected dosing were 55% with RUL, 61% with BF, and 64% with BT electrode placements. The rates were not affected by the presence of psychosis, polarity of depressive illness, or severity of HAMD24 scores. Cognitive assessments found no follow-up differences among the patients treated with the three placements [13].

These observations are consistent with the extensive reports of the efficiency of electrode placements. The inferiority of RUL placement would have been much greater if the minimal dosing of the original studies had been followed. Treatments with RUL placements at 1.5 × or 2.5 × ST, as initially recommended to minimize immediate cognitive effects, elicited poor remission rates of 35% and 30% [8]. In the CORE study, the benefits of RUL at 6 × ST sought to approximate the remission rate of BT placement, but even at this high dosing, the benefits were less than that of BT placement at 1.5 × ST.

Comparing the efficiency of the treatments in the CUC and CORE studies, the CUC patients first received a mean of seven RUL treatments followed by bilateral treatments for a total mean of 10.5 treatments to remission. In the CORE study, the average number of BT treatments to remission was 7.3, a savings of at least three anesthesia inductions and seizures [50]. The ease of use of bilateral ECT, in which the energy dosing can be estimated by age-related algorithms and does not need ST determination, its consistent achievement of remission and the only modest cognitive advantage of high-dose unilateral ECT support the continued preferential clinical use of bilateral ECT.

Speed of response

An assessment was made of the speed of response of the first 255 patients treated [51]. The demographic characteristics did not differ from the larger sample later reported [12]. The first response was defined as >50% reduction in baseline HAMD24 score and occurred after one treatment in 12.6% of the sample, after 3 ECT (within 1 week) in 53.8%, and by ECT #6 (2 weeks) in 83.4%. A sustained response was achieved in 34.8% of patients at or before ECT #3, in 64.4% at or before ECT #6, and 75.9% by the third week of treatment. Remission was achieved in 74.7% of those treated, occurring with an average of 7.4 treatments. When the first response was attained at or before ECT#6, 83% exited the study as remitters.

The speed of response of patients with major depressive disorder to ECT within 3 weeks contrasts sharply with the symptomatic outcomes reported in pharmacotherapy trials in which 50% of the patients achieve response and only 35% achieve remission, usually requiring 2–3 months of treatment.

Seizure threshold

In the belief that the intensity of the electrical stimulus drove the severity of the cognitive effects of ECT, a concept of measuring the “seizure threshold” (ST) – the minimum current that elicited a seizure – and treating patients with low doses just above the ST took hold. The first reports cited a 40-fold range in the calibrated ST encouraging the belief that the ST was a salient feature in eliciting effective treatments [52]. With successive treatments, the duration of the seizure shortened, suggesting that the ST rose with repeated seizures.

The ST for BT electrode placement was measured at baseline and again at the end of the course of treatment in the CORE studies [53]. The ST was positively correlated with age, with slightly higher measures in males. Using the electrical characteristics of the Thymatron DGx, a 40% charge elicited seizures in 97.5% of the subjects. To treat at 1.5 × ST, 96% of the patients were successfully treated at the three levels of 15%, 30%, and 60% of the maximum 500 millicoulombs energy delivered by the devices [54].

The ST was measured at the first treatment and again during 1 week after remission. In 80 subjects, the ST was the same in 56 (70%) at the two estimates, increased in 17 (21%), and decreased in 7 (9%). The lack of change in ST negates the hypothesis that the rise in ST was a salient feature of the ECT mechanism of action [55].

We found no justification for the measurement of the ST with either BT or RUL electrode placements. That RUL treatments require high energy levels, from 8× to 12× ST, can be achieved with treatments that begin at 75% of the delivered 500 mC of present devices and assessing the features of the seizure EEG to determine if the treatment is ‘effective’ [56]. Clinicians modify the energies of successive treatments by the quality of the EEG recordings. For BT treatments, the half-age estimate is a reliable dosing strategy, especially with attention to the seizure EEG characteristics [57]. The determination of the individual ST assures an ineffective first treatment, subjecting the patient to an unnecessary, ineffective, and costly anesthesia and treatment session. The ST is a measure of little heuristic value; it is not clinically justified.


After more than 14 years of study in a multi-site national collaboration and the support of hundreds of patients and many staff members, what was learned? The remarkable gamble by a Hungarian neuropsychiatrist to induce seizures in psychiatric patients in 1934 has given us a safe and effective treatment for the severe psychiatric ill. We do not understand why seizures that are damaging when they occur in response to a brain insult should ameliorate disturbances in mood and thought. The relief and safety of the treatments are well defined. Alas, the relief is often transient and like insulin treatment of diabetes and dialysis for renal failure, the treatments must be sustained in continuation treatments, often for weeks and months, and occasionally for years.

In the CORE studies, the patients were treated three times weekly and remission was commonly recorded after 7–10 treatments. In the 6-month continuation phase after initial remission, a fixed schedule of induced seizures was prescribed. The benefit was sustained in less than 40% of the patients. Prescribed scheduled treatments are inefficient. To optimize continuation care, treatments should be available when the patient and family report the recurrence of symptoms, as often and for as long as the treatments offer relief. In the first decades of ECT use, evening and week-end treatment sessions were common; these should be reinstated.

Electroconvulsive therapy relieves both unipolar and the bipolar depressive states. This experience challenges the widely held belief that ECT is not effective in patients with bipolar depression. The treatment guidelines for these patients warrant revision.

The rapid relief of suicide thoughts and actions, commonly within the first week, needs re-assessment of treatment guidelines to give ECT pride of place.

In the belief that the electrical dose affects cognition and memory, much time and energy has been expended seeking the minimum energies needed for effective treatment in depressed patients. Seizures induced with unilateral electrode placement at minimal energies are inefficient and should be discarded. Once the energies are raised to effective treatment levels, the observed advantage of lesser effects on memory is lost. The justification for unilateral electrode placement is questioned, and their use should not be prescribed except under unusual circumstances.

A similar enthusiasm for ultra-brief currents raises the same hazard of inefficient treatments. Such currents are reported to elicit lesser immediate cognitive effects with less effective clinical outcomes [58]. The justification for this enthusiasm is questioned.

The drive to stimulate the brain without inducing seizures using transcranial magnetic pulse stimulation (TMS), vagus nerve stimulation by bodily implanted devices (VNS), and the surgical placing of electrodes in brain nuclei (DBS) are recommended replacements for ECT. These efforts ignore the experience that grand mal brain seizures, not any aspect of electricity or brain stimulation, are the therapeutic mechanism for the relief in patients with mood disorders.

Electroconvulsive therapy devices presently sold in the United States are limited by FDA regulations in the maximum energy delivered [59]. In other countries, the maximum allowed energies are twice that permitted in the US. While many patients can be satisfactorily treated with limited dosing, for some the seizures are inadequate even at maximum energies. This limitation is particularly damaging for the elderly whose seizure thresholds are higher than those of younger adults and adolescents. This mechanical limitation is unjustified and a particular impediment to effective treatments.

In a detailed review of the ECT experience, Ottosson and Odeberg describe many tests of electric currents, energies, and electrode placements that fill the world literature for more than 70 years [60]. Instead of emphasizing treatment efficacy in relieving the severely ill, the research sought to minimize the effects on cognition, thereby distorting the treatments, assuring poor outcomes, and encouraging a conflicted public image. They carefully note that impaired cognition is found in the first 3 days after each treatment. At times longer than 15 days, both self-reports and test results improve. The overemphasis on the ‘memory loss, memory loss’ mantra by researchers has distorted the public image of the treatment and inhibited effective treatment [61].

The American Psychiatric Association diagnostic classification of 1980 increased the number of labeled disorders from 182 to 265. These labels were based on symptom checklists, avoiding any biological tests or markers. The teaching that convulsive therapy is effective for ‘depressive mood disorders’ broadened the subject pool to include many for whom the treatment was not effective. Ottosson and Odeberg cite the principal criteria for patient referral are those with psychotic depression and risk of suicide. More than 90% remit. They also cite efficacy in disorders with a high risk of lethal outcome as delirious mania, catatonic stupor, postpartum psychosis, lethal catatonia, and neuroleptic malignant syndrome. They caution that ‘Widening the indications where the treatment has no obvious therapeutic advantage to other treatments but may give rise to memory complaints, runs the risk of undermining the reputation of the treatment.’

We still do not understand why or how inducing brain seizures can ameliorate psychiatric disorders. That such benefits are well defined justifies the treatment's world-wide use. Seizures are responses to brain insults, whether by brain scarring, electricity, chemicals, trauma, infection, or fever. Focusing on electricity for stimulation has given us an efficient but dangerous instrument and turned our interest away from the seizure to electricity.

In the first decades of convulsive therapy, the inductions of seizures were chemical, using pentylenetetrazol (Metrazol) and flurothyl (Indoklon) [3]. Flurothyl is a congener of diethyl ether that first induces unconsciousness and then after a few additional breaths induces a full seizure. In extensive trials in the 1960s, flurothyl-induced seizures were as effective as electrical inductions with lesser immediate effects on cognition [62]. Treatments then were unmodified, without the routine use of a barbiturate sedative and succinylcholine for muscle relaxation. The authors found flurothyl administration more complex than that of ECT, and discarded flurothyl. Present routine clinical ECT practice with anesthesia and muscle relaxation makes flurothyl inhalations a reasonable, effective, and safe alternative to the complexities of electrical inductions. It properly focuses attention on the seizure and away from the complexities of dosing with electricity.

Acknowledgements: project history

A grant application to the National Institute of Mental Health by Harold Sackeim at Columbia University requested support for a study of continuation placebo, nortriptyline alone, and nortriptyline combined with lithium in unipolar depressed patients treated by right-unilateral ECT. The study title was ‘Continuation Pharmacotherapy Following ECT’. It was funded in 1992.

A Review committee headed by Jonathan O. Cole with Gary Figiel, Max Fink, Ranga R Krishnan, S. Craig Risch, and Charles Welch recommended the investigators consider an additional continuation treatment arm of ECT. The Columbia University investigators demurred. The reviewers recommended the described study be funded and suggested the NIMH call for proposals for a comparison of continuation ECT and continuation combined lithium and nortriptyline in a similar population.

The CORE study was funded by NIMH in 1997 in four facilities in a study programme titled: ‘Continuation ECT vs. Pharmacotherapy: Efficacy and Safety’. The initial CORE collaborators were Max Fink (Stony Brook University, New York), Charles Kellner (Medical University, Charleston SC), Teri Rummans (Mayo Clinic, Rochester MN) and John Rush (University of Texas, Dallas TX).

In the 14 years of this endeavor, many personnel and site changes occurred. Max Fink was replaced by George Petrides, John Rush by Mustafa Husain, and Teri Rummans by Keith Rasmussen and then by Shirlene Sampson as Principal Investigators at their individual sites. The initial site at Stony Brook University was moved to the Long Island Jewish Hillside Hospital, and that at Medical University of South Carolina moved to the New Jersey School of Medicine in Newark NJ.

A second study protocol titled ‘Comparing Three Electrode Placements to Optimize ECT’ was funded by NIMH for the same collaborating investigators in 2002.