Dosing methods in electroconvulsive therapy (ECT): towards the modal ECT technique

Bergsholm and Bjølseth [1] (this issue of the Nordic Journal of Psychiatry) have written an important paper about the state-of-the-art of ECT technique. They present a scholarly, insider’s view of the history of the five major ECT stimulus dosing techniques. It is a superb historical account of dosing methods in ECT. The good news is that ECT is such an effective treatment that most techniques work well; the not-so-good news is that there is no consensus in the field about the optimal approach for a particular patient. Optimized ECT should preserve the remarkable antidepressant/antipsychotic/anti-catatonic efficacy of the treatment while capping adverse cognitive effects to a low and acceptable level. In this editorial, we will comment on the technique issues reviewed by Bergsholm and Bjølseth from the point of view of ECT clinicians and researchers. In addition, we will make some recommendations about moving towards more standardized ECT techniques.

Electrical stimulus dosing and electrode placement cannot be considered separately; they come as an interactive package. As much as one would have thought that the electrode placement disagreements in our field would be settled by now, alas it is not so; the relative merits and applications of bilateral vs. right unilateral remain disputed, with practitioners having their ‘favorite.’ What is needed, in our opinion, is a reasonable algorithm that recognizes the pros and cons of each and guides the practitioner in selecting the appropriate technique for a particular patient’s clinical situation. The choice should be patient-centered, not based on the idiosyncratic experience of the practitioner. Perhaps when we have accurate methods to predict response to a particular technique, the issue will be settled; until then, we recommend an inclusive approach based on clinical urgency (see below).

Now to some specific issues raised by Bergsholm and Bjølseth about the five dosing methods. Since it has been conclusively proven that excess charge results in greater adverse cognitive effects, we can agree that the fixed high-dose technique is suitable only for selected urgent cases. Age- and formula-based methods, with considerable evidence to back them up, are in widespread use, and likely suitable for the majority of patients. The benchmark method, while very interesting, requires much more study.

As part of their plea to bring back the Scandinavian time-titration method, Bergsholm and Bjølseth consider whether individualized stimulus dose titration is necessary and helpful. They discuss some recent literature on this topic and quote several skeptical authors, one of us (CHK), included. A central question relevant to this is whether or not there is great variability in seizure threshold (ST). We would suggest, that while a few outliers exist, the vast majority of patients have highly predictable STs. This view is supported by several large datasets [2,3] and the fact that age- and formula-based dosing methods work for the vast majority of patients. If a patient has an unusually high ST, this will be discovered at the first treatment session, and appropriate steps can be taken to ensure effective treatments at subsequent sessions. If ST outliers comprise 1-10% of ECT patients, the technique should be geared to the 90%, not the 10%. In addition, the concept that there exists an optimal charge dose that is a multiple of ST has never been adequately replicated. In summary, while the promise of ‘personalized’ or ‘individualized’ medicine is tempting, there is no solid evidence that dose titration leads to better ECT outcomes.

Bergsholm and Bjølseth tout the Scandinavian time-titration method as a form of individualized dosing, dose titration at every treatment, as it were. Let us look first at the theory behind this technique, and then the practical logistics of implementing it. We are not aware of studies that can tell us exactly when the seizure starts during stimulus delivery; the MECTA and Thymatron devices start recording EEG only after the stimulus has ended. The idea that stimulating until there is robust evidence of a sustained seizure makes good sense, and the history reviewed by Bergsholm and Bjølseth proves that, in the right hands, it is highly effective. Whether it minimizes cognitive effects is less certain. A well-trained time-titration practitioner will likely err on the side of ample, rather than stingy, stimulus applications, particularly when the desired endpoint is not crystal clear. Bergsholm and Bjølseth describe two types of abortive seizures (‘dissociate convulsion’ and ‘clonic convulsion’) that may result from the premature release of the stimulus button, but the description of them is confusing. Long duration, low pulse frequency stimulus packages are the centerpiece of the time-titration method; since there is robust evidence that such parameters are very efficient for seizure induction, we agree that this part of the method is fine.

A related point is that the ability to set each of the four stimulus parameters (pulse width, frequency, current, and duration) in nearly infinite combinations, while theoretically advantageous in the hands of a very knowledgeable practitioner, may, in fact, be counterproductive in most situations, leading to the choice of suboptimal stimulus packages. Better to have fewer choices and have them all biased towards low frequency and long duration, as recommended by Bergsholm and Bjølseth.

But how reproducible and consistent is the time-titration method in practice? What does it require to adequately train a provider? Can the method be taught virtually, or does it require apprenticeship? Could a video do the trick? And is there a component of uncertainty introduced by varying levels of muscle relaxation? We are skeptical that the method could be reliably taught on a large scale. The desired endpoint (‘interruption of the stimulus when the initial muscle contractions shift to the generalized tonic phase, ultimately observed by maximal tonic extension of the big toes’) may not be quite as simple to determine as they suggest.

Bergsholm and Bjølseth’s excellent and extensively referenced review of the relevant literature leads one to reflect on the matter of absolute stimulus charge and the curious situation of America making due with 500 mC, while the rest of the world gets 1000 mC. Is the ‘200%’ really necessary and helpful? If it were really necessary, should not ECT outcomes in America be inferior? Are cognitive outcomes worse in Europe? If, as we believe, ST outliers on the high side are rare, why do we need this extra capability? If it is available, practitioners will certainly use it, perhaps sometimes indiscriminately, with what results? Literature meta-analyses of stimulus charge and cognitive outcomes in Europe vs. America could provide some answers.

Bergsholm and Bjølseth rely heavily on the postictal reorientation time as an important metric related to seizure adequacy. They state, ‘The optimal time of disorientation, i.e., the postictal reorientation time (PRT), is not known. However, it may be about half an hour…A sufficiently long PRT may be the ultimate sign of a therapeutic seizure.’ In our opinion, this is dangerously close to saying that efficacy is tied to at least a moderate amount of cognitive impairment. Does a ‘good’ seizure necessarily result in long recovery-of-orientation times and possibly other cognitive (adverse) effects? Certainly, many patients get completely well with ECT and exhibit negligible cognitive effects. Noted ECT researchers have long suggested that efficacy and cognitive effects in ECT are not tightly linked: e.g., Sackeim et al. in 2008 wrote, ‘… the therapeutic and cognitive effects of ECT are dissociable. In general, correlational studies have not found associations between the extent of amnesia and the therapeutic effects of ECT’ [4].

Again, optimal ECT should have high efficacy and low cognitive impairment.

Studies to answer all of the above questions are easily doable with the right resources. Perhaps the recently formed international ECT research consortia (e.g., GEMRIC—The Global ECT-MRI Research Collaboration) can turn their attention to issues of technique, as well as mechanisms of action. So here is our suggestion: develop an ECT technique algorithm to serve as a template for more standardized treatment around the world. This, of course, is not a novel idea, but no such authoritative, detailed guideline has achieved widespread acceptance. ECT technique should be geared to the severity/urgency of the clinical situation and take into account patient preferences, if feasible and reasonable. The most urgently ill patients should be offered bilateral (BL) electrode placement with moderately high initial stimulus dose packages, optimally with long duration and relatively low pulse frequency. Less urgently ill patients may be offered right unilateral (RUL) electrode placement with brief, or ultra-brief stimuli, with similar parameters, but higher overall dosage. Dose titration at the first treatment, primarily for RUL electrode placement, should be optional.

A technique algorithm also needs to guide practitioners in what to do in cases of slow- or non-response: in other words, how to transition from less to more powerful forms of ECT. After several (and a range for ‘several’ will need to be specified) treatments with ultra-brief pulse RUL ECT at an adequate dose, is the next step brief-pulse RUL or BL electrode placement? This part of the algorithm is sure to elicit the most debate.

And what to conclude about the time-titration method? Bergsholm and Bjølseth make several well-reasoned and compelling arguments, but we suspect the method is hard to implement with good reliability. Scandinavians have led the way for innovation and advancement in ECT technique; this time, these authors’ suggestion to revert to an old-fashioned method seems retrogressive, not progressive. Certainly, since the time titration method involves stimulation until a toe extension is observed and not ended when a predetermined dose is reached, it involves the risk of excessively high dosing with subsequent otherwise avoidable cognitive side effects. But it is possible that more research on the time-titration method might prove these concerns wrong.

As noted at the top, ECT is so remarkably effective (and safe), that discussions of optimizing technique, while important, take a back seat to the overriding problem that ECT is grossly underutilized; greater, and earlier, adoption of ECT would certainly be a public health benefit [5]. Perhaps the two issues overlap, and greater consistency of technique might actually contribute to wider acceptance of the treatment. In any case, Bergsholm and Bjølseth are to be applauded for a major contribution to the literature on the ECT technique that is sure to ‘stimulate’ and inform much discussion.

Disclosure statement

CHK receives fees from UpToDate and Northwell Health. He receives royalties from Cambridge University Press.

Additional information

Notes on contributors

Charles H. Kellner

Charles H. Kellner, M.D. is a geriatric psychiatrist who has specialized in ECT and ECT research. He is a former Editor-in-Chief of the Journal of ECT and has published over 300 articles, mostly related to ECT. He is the author of Handbook of ECT from Cambridge University Press.

Martin B. Jørgensen

Martin Balslev Jørgensen is a chief physician at the Psychiatric Center, University Hospital Copenhagen, and a professor at the Department of Clinical Medicine, UCPH.