TMS is a psychiatric treatment that uses a rapidly alternating magnetic field to induce electric currents in the brain. These currents stimulate neurons, causing them to “fire.” When used repetitively, TMS is said to alter the excitability of the brain area that has been stimulated. In the psychiatric field, TMS is being used increasingly as a treatment for depression, particularly with so-called treatment-resistant clients. I Googled the string “TMS + depression” and got 1.35 million hits. So the idea is attracting attention.
One of the abstracts presented at the May 2013 APA annual conference in San Francisco was called A Multisite, Longitudinal, Naturalistic Observational Study of Transcranial Magnetic Stimulation (TMS) For Major Depression in Clinical Practice, by Mark Demitrack, MD, et al.
The abstract was a study of the efficacy of TMS involving 307 depressed participants who had failed to benefit from prior antidepressant treatment. The authors’ conclusions were:
“These data support the view that TMS demonstrates a statistically and clinically meaningful durability of acute response over 52 weeks of follow up. Maintenance of benefit was observed under a pragmatic regimen of continuation antidepressant medication and access to TMS reintroduction for symptom recurrence.”
The abstract also indicated that the study had been “supported by a grant from Neuronetics, Inc.”
The presentation generated some enthusiasm. Medscape ran an article by Caroline Cassels on May 24, 2013, under the headline TMS for Resistant Depression: Long-Term Results Are In. In this article Mark George, MD, a long-term advocate of TMS, is quoted:
“This is great news for our field and for the millions of patients who suffer from depression and do not respond well to medications…”
With regards to formal publication, the study is divided into three parts.
1. L. Carpenter et al, 2012, focused on the effect of TMS on depressive symptoms.
2. P. Janicak et al, 2013, focused on the effect of TMS on quality of life measures.
3. Analysis of the 52-week follow-up, submitted for publication but not yet published. Some of the data from the follow-up, however, was presented in the APA conference abstract.
The authors of the 2012 and 2013 articles are identical except for one person – Dafna Bonneh-Barkay, PhD – whose name appears on the 2013 article only. Dr. Demitrack, listed as lead author on the APA conference abstract, is shown as an author on both the 2012 and 2013 papers.
The study, which started with 339 participants, was conducted on 42 sites which used TMS. There was considerable participant attrition during the study. By baseline the number was down to 307, and only 265 actually completed the acute treatment phase. Reasons for drop-out were:
|REASON FOR DROP-OUT||NUMBER|
|No post-baseline evaluation||17|
|Non-standard treatment protocol used||7|
|Did not meet diagnostic criteria||4|
|Prior TMS treatment||4|
|Failed to return||18|
|Unsatisfactory response efficacy||7|
|Satisfactory response efficiency||2|
|Adverse event (seizure)||1|
Of the 265 individuals who completed the acute treatment phase, 257 entered the 12-month follow-up phase, and of those only 204 participants provided data across the entire study period. Presumably the reasons for these follow-up phase attritions will be listed in the formal write-up, but at present this information is not available.
The 2012 paper begins by stating that
“…20-40% of patients do not benefit from, or cannot tolerate, adequate trials of antidepressant medications even after repeated attempts.”
The authors go on to describe TMS, and state that
“When used as an antidepressant, TMS produces clinical benefit without the systemic adverse effects associated with medications, and has no adverse cognitive effects.”
The authors state that although the efficacy of TMS has been established, there have been no multisite studies of its utility and effectiveness in routine clinical practice. Their goal in the present study was
“…to summarize outcomes experienced by a large population of depressed patients treated with TMS therapy in various clinical settings.”
Clients were eligible for participation if
1. Their primary “diagnosis” was a major depressive episode without psychotic features, using DSM-IV criteria
2. They had no medical condition that precluded the use of TMS
3. They had not received TMS previously
4. They were pharmacoresistant as determined by the Antidepressant Treatment Record (ATR). The ATR is a screening instrument developed by Neuronetics
5. Their attending psychiatrists had determined that TMS was the most appropriate clinical option.
All treatments used the NeuroStar TMS system – a Neuronetics product. The intensity of the magnetic field was set at 120% of motor threshold, which in turn was determined on an individual basis with the help of a computerized mathematical algorithm called MT Assist, which is also a Neuronetics product. Each treatment session involved about 3,000 pulses (average 3,216), usually over the left dorsolateral prefrontal cortex. The motor threshold, incidentally, is the minimum amount of power output that will cause the client’s thumb to twitch. The actual treatment is administered at a point 5cm in front of the motor threshold determination point.
Participants who were taking antidepressant drugs at the time of enrollment were permitted to continue taking these at the discretion of the treating psychiatrist. At study baseline, 300 of the 307 participants were taking one or more psychotropic drugs.
The average number of TMS treatments administered was 28.3 (range 2-94) across an average duration of 42 days (range 2-130). The normal protocol was about five sessions per week.
Outcome measures of depressive “symptoms” were reported in the 2012 paper, (Carpenter et al, 2012). The primary outcome measures were the changes from baseline to the end of the acute phase of treatment on:
- the Clinical Global Impression – Severity Scale (CGI-S)
- the Inventory of Depressive Symptoms, Self-Report version (IDS-SR)
- the 9-Item Patient Health Questionnaire (PHR-9).
The CGI-S is completed by the treating clinician; the other two scales are completed by the client.
The mean scores on all three scales declined (i.e., improved) during the treatment period. Remission rates (as defined by a score of 2 or 1 on CGI-S, a score of less than 5 on the PHQ-9, or a 50% or greater drop in score on the IDS-SR) were 37.10%, 28.70%, and 26.50% respectively.
The actual duration of acute-phase treatment was determined by the treating psychiatrist’s estimate of the “point of maximum benefit.” The authors report that 22% of the participants were treated beyond the 6 weeks duration. The maximum duration was 130 days, or a little over 4 months.
The 2013 study focused on quality of life outcome measures. These were assessed by:
- Medical Outcomes Study-36-Item Short Form Health Survey (SF-36)
- EuroQol 5 – Dimensions (EQ-5D)
Each participant was assessed on these self-administered measures at baseline and again at the end of acute treatment. As with the depressive “symptoms,” mean scores on these quality of life measures improved during the acute treatment period.
As mentioned earlier, the formal write-up of this phase of the study is not yet available, but the APA abstract provided the following information:
CGI-S, PHQ-9, and IDS-SR scores were obtained at the 3, 6, 9, and 12-month marks. The reduction in mean scores, noted at the end of acute treatment, was sustained throughout the 52-week follow-up
|End of Acute Treatment||52-weeks|
Remission rates were also maintained.
|End of Acute Treatment||52-weeks|
[Note there is an erroneous transposition of PHQ-9 and IDS-SR scores in either the 2012 article or the APA Conference abstract. I believe that the way I’ve presented the figures above is correct, but, as the numbers are fairly close anyway, it makes little difference to the overall result.]
The APA Conference abstract provides no information on quality-of-life scores during the follow-up period.
On the face of it, the two papers and the APA abstract suggest that TMS is an effective treatment for depression, and that its beneficial results are long-lasting. But such an interpretation is misleading.
TMS is routinely presented as a safer alternative for the treatment of depression than ECT (e.g. TMS Center of Princeton). It is well known that ECT provides, for some individuals at least, a temporary feeling of well-being, but that these effects are short-lived, and the treatment often needs to be repeated more or less indefinitely. When Demitrack et al state (in the APA abstract) that TMS demonstrates
“…clinically meaningful durability of acute response over 52 weeks of follow up”
most people would, I suggest, interpret this to mean that the beneficial effects of the acute phase treatment lasted for 52 weeks. In fact, both the TMS and the antidepressants were continued as needed to maintain therapeutic benefit for many of the participants for most of the year. This is mentioned in the various documents. For instance, in the APA Conference abstract it says:
“Maintenance of benefit was observed under a pragmatic regimen of continuation antidepressant medication and access to TMS reintroduction for symptom recurrence.” [Emphasis added]
In fact, 30.2% of participants required subsequent TMS after the 3 month assessment.
Earlier in the abstract, however, they had said that the participants who benefitted from “…acute TMS treatment were tapered from their TMS regimen, consented to long-term follow up over 52 weeks, and were evaluated for statistical analysis.” [Emphasis added]
This, I suggest, implies that the TMS was discontinued after the acute treatment phase. In my view, the fact that the TMS was reintroduced on an as-needed basis to combat “symptoms” might have been reported more prominently.
Further to this point, in Caroline Cassel’s Medscape article quoted earlier, there’s another quote from Mark George, MD:
“This is a very important and exciting study. Several prior studies have shown that prefrontal rTMS works to treat depression acutely. Until this study, we have had only limited information about how well these patients do a year after completing a course of TMS. Long-term data following remission produced by medications or electroconvulsive therapy [ECT] in these treatment-resistant patients have been disappointing, with only about 13% being still remitted a year later.”
“For example, over half of patients who remit with ECT are ill again 6 months later. Thus, having 45% in remission a full year later is very, very encouraging that rTMS is perhaps changing the underlying pathological circuitry associated with depression and producing a more stable remission than the other treatments.”
In this quote, Dr. George laments the fact that more than half the people who receive ECT are “ill” again 6 months later.
The transient nature of the ECT effect is a well-established fact, and these individuals are usually returned to the hospitals for further “treatment.” By contrasting these individuals with the 45% still “in remission a full year…” after TMS , it appears that Dr. George is under the impression that these individuals did not receive further sessions. Or at any rate, that is the impression he is giving, particularly with the phrase: “…a more stable remission…”
Incidentally, I’m not sure where Dr. George got the 45% remission figure. In the APA Conference abstract, the highest remission rate was 40.4% (on CGI-S). The others were 33.6% on PHQ-9 and 26.10% on IDS-SR. Perhaps the 45% refers to individuals who remitted by any of the three measures – but there’s no mention of that in the paper.
Also incidentally, Dr. George is the director of the Brain Stimulation Lab in Charleston, South Carolina, which has received grant funding from Neuronetics. (George et al, 2013, p 17).
After the conference PsychCongress ran an article highlighting the abstract. Their headline read: APA News: Transcranial Magnetic Stimulation Shows Long-Term Benefit In Treating Depression, suggesting, here again, that the author of the article had been misled.
Conflicts of Interest
It is also of note that all of the authors of the 2012 and 2013 papers have significant conflicts of interest.
Scott Aaronson, MD: speakers fee from Neuronetics.
Dafna Bonneh-Barkay, PhD: Neuronetics, employee, with salary and stock options.
Terrence Boyadjis, MD: Neuronetics research support contract.
David Brock, MD: Neuronetics employee, with salary and stock options.
Ian Cook, MD: Neuronetics, speaker’s bureau, honoraria, research support, grant.
Linda Carpenter, MD: Neuronetics research support.
David Dunner, MD: Neuronetics, grant support, research support, consultant fees.
Mark Demitrack, MD: Neuronetics, employee (Chief Medical Officer) with salary and stock options.
Hugh Solvason, PhD, MD: Neuronetics research support.
Philip Janicak, MD: Neuronetics, research support, advisor, consultant.
Karl Lanocha, MD: Neuronetics, research support, speaker’s bureau.
On June 4, 2012, Shiv Gaglani posted an article on Medgadget.com. The title was The Promise of TMS: Interview with Neuronetics Chief Medical Officer Mark Demitrack, M.D. Here’s a quote:
“The makers of the NeuroStar TMS System, Neuronetics, contacted us to let us know about their new study released at the American Psychiatric Association meeting in May.” [Emphasis added]
This, incidentally, was not the same APA 2013 abstract mentioned earlier. Dr. Demitrack had presented an earlier abstract – a preview of the Carpenter et al 2012 study – at the 2012 APA conference. The Medgadget interview was based on the earlier abstract, and gave a brief summary of the study’s main findings, with quotes from an interview with Dr. Demitrack, but there is no mention of the concomitant use of the antidepressants, or of the continued use of TMS during the follow-up period.
An additional point of concern in this regard is Dr. Demitrack’s statement concerning safety. Here’s the quote
“Medgadget: Were any side effects observed?”
“Dr. Demitrack: The most common side effect directly related to the device is scalp pain or discomfort at the site of stimulation. Most patients find this to be mild to moderate in intensity, and it usually becomes unnoticeable after the first week.”
In fact, as the 2012 paper makes perfectly clear, one of the participants experienced a generalized tonic-clonic seizure (formerly known as grand mal seizure) during a TMS treatment session.
“There was one medical event considered probably or definitely related to the device and that was filed with the FDA as a Medical Device Report. This event was a generalized tonic-clonic seizure that occurred in a female patient during her 10th TMS treatment session. The patient had no prior history of seizure, however she had several clinical factors that may have contributed to altering her seizure threshold. Specifically, the evening before her treatment she had completed a night shift of work, and was therefore sleep-deprived at the time of the TMS session. In addition, she was also taking bupropion, sertraline, and dextroamphetamine/ levoamphetamine at the time of her TMS acute-phase participation. The patient recovered fully from the event without neurological sequelae.
Seizure is a known, but rare, medical risk associated with TMS. In the entire postmarketing experience with this system, there have been six reports of seizure filed as MDRs to the FDA. Based on this experience, the estimated risk of seizure is approximately 0.003% per treatment exposure, and <0.1% per acute treatment course.”
It seems obvious to me that this should have been mentioned in the interview and included in Medgadget’s article. This is particularly important in that Medgadget is a credible and sophisticated website operated by a team of MDs and biomed engineers. They have 11,700 Twitter followers, they have their own YouTube channel, and they have a large following on their Facebook site. In addition, the article was picked up and hyperlinked by at least three other sites: Regator, AllVoices, and Organized Wisdom.
At the end of the Medgadget article there’s a link to a 20-minute Neuronetics promotional video. At about 18:28 minutes, the spokesperson says:
“Over 10,000 active treatments were performed across all NeuroStar clinical trials, demonstrating its safety. During trials, no seizures or systemic side effects were seen.”
However, as mentioned earlier, the Carpenter et al study states that six seizures had occurred in the post-marketing experience, and one during the Carpenter et al study itself. On this basis, the video statement seems deceptive, or at least misleading.
The authors of the 2012 paper estimate the seizure risk at 0.003% per treatment session. This is about 1 in 33,000, and on the face of it seems like a very small risk. But a treatment course typically consists of about 30 sessions (5 per week for 6 weeks), which brings the risk per treatment course up to 0.09%, or 1 in 1,100. But, further, some individuals receive considerably more than 30 sessions. The maximum in the present study was 94. The risk for a person receiving 94 sessions is 0.28%, or 1 in 357. And when we remember that many people receive repeat treatments, it is clear that the risk for these individuals becomes increasingly significant. It is also clear that these risk estimates are based on very limited actuarial experience, and may have to be revised (upwards or downwards) as TMS use increases.
TMS – Non-Invasive?
As mentioned earlier, TMS is getting a good deal of air-time in recent years. In particular, it is being promoted as a “non-invasive” treatment. I Googled the string “TMS + non-invasive” and got 317,000 hits, and the present authors use the term to describe the treatment (2012 paper, p 588).
This is misleading. TMS involves sending multiple strong bursts of magnetic energy into the brain. The pulses are increased in strength until they are actually causing neurons (brain cells) to fire, and are repeated at that intensity about 3,000 times in each session. In the present study, the average number of sessions received by participants was 28. Strictly speaking, a medical procedure is considered non-invasive if it doesn’t break the skin, but it seems a misnomer in the present context. X-rays don’t break the skin, but would hardly be considered non-invasive.
Method of Action
George et al also note that “Mood regulating centers in the brain overlap significantly with the neural pathways involved in pain regulation…” They also report that TMS, applied to the prefrontal cortex (the same general area that is used to treat depression), has an analgesic effect, from which they tentatively conclude that opiate receptors play a part in the effects of TMS.
The precise mechanism of action of TMS is not fully understood. However, George et al suggest that depression results from neural “dysregulation” in the prefrontal and limbic regions. They hypothesize that:
“…chronic, frequent, sub-convulsive stimulation of the prefrontal cortex might initiate a therapeutic cascade of events that rebalances and normalizes the dysregulated prefrontal and limbic circuitry.” [Emphasis added]
This sounds remarkably similar to the now discredited notion of antidepressants therapeutically correcting chemical imbalances.
Similarity to ECT
TMS is being routinely presented as a safer alternative to ECT for “treatment-resistant” clients. In some instances (e.g. the quote from Dr. George, mentioned earlier), it is being promoted as more effective than ECT. And based on present information, TMS does not seem to be as immediately traumatic to the brain as ECT. However, there is a fundamental similarity between these two interventions: they both interfere with the normal electrical activity of the brain. We know that in the case of ECT, the long-term results are catastrophic, and that it is only a matter of time before this procedure will be banned. Though meanwhile psychiatry, to its shame, continues to tout it as safe and effective.
With TMS, the picture is less clear. But as this technology becomes more widely used, and used more frequently for given individuals, we may see increasing evidence of cumulative neural damage, including even more pronounced anhedonia than the technology purports to address. Certainly the already established risk of seizure suggests that TMS has more in common with ECT than its proponents might be willing to acknowledge.
There are already indications that TMS protocols may call for higher doses and more frequent treatments in the future, at least for some individuals. Mark George, MD, et al wrote in January 2013:
“Another issue is determining the optimal dose over the optimal time period for alleviating depression. Most studies have stimulated patients at or above motor threshold. This is particularly important in elderly patients, in whom prefrontal atrophy may outpace motor cortex atrophy… There have never been dose-finding studies with rTMS. Thus, some groups are studying whether higher doses of TMS might produce more rapid or more effective results …Also, there are a few case series suggesting that weekly or monthly rTMS can serve as maintenance therapy for acute responders…” [Emphases added]
A great deal of psychiatric research, including the present Neuronetics study, uses questionnaires and rating scale scores as outcome measures. This is an inherent deficit in all research of this kind, in that there is inevitably some uncertainty as to whether positive responses on a questionnaire reflect positive changes in a person’s life.
An additional problem in this regard is that often the questionnaire items are ambiguous, or at the very least, difficult to unravel.
It is clear from material presented above that there are already close links between the TMS industry and psychiatric researchers. As has become glaringly obvious in the pharma sphere, these kinds of links do not generally foster research objectivity and integrity.
The concept of “treatment-resistant depression” is an integral component of TMS promotion, and has become a hot topic in recent years (5.9 million hits on Google). The idea is that there are some individuals whose feelings of depression are not alleviated by antidepressants to any appreciable degree despite multiple trials and high dosages.
Many of us on this side of the debate are of the opinion that this kind of chronic anhedonia is actually caused by long-term use of antidepressants, and this notion has been presented persuasively by Rif El-Mallakh et al, 2011. Not surprisingly, this concept has garnered little support in psychiatric circles, where depression is assumed (groundlessly) to be an illness; antidepressants are promoted as the “cure;” and people for whom the cure is ineffective are labeled treatment-resistant.
Instead of recognizing and acknowledging the destructive effects of the drugs, psychiatry is salving its conscience by pretending that these individuals have some undefined (and hence unrefutable) quirk that prevents them from benefiting from psychiatric drugs, and is now busily developing and promoting TMS as a treatment for a problem that psychiatry very likely created in the first place.
The idea that people can find happiness or contentment or peace or mind by having their brains tweaked by chemicals or electromagnetism is a fundamentally dehumanizing and disempowering concept. Feelings of loss and despondency are an integral part of the human condition, best resolved by dealing with the source of the problem, with the assistance and support of friends and loved ones. The notion that we can banish these feelings with an effortless brain-altering intervention is precisely the same philosophy that underpins the marketing of street drugs. It provides an increasingly short-lived sense of relief for some people, but always entails more problems than it solves. Ask any member of Narcotics Anonymous.
Mad in America hosts blogs by a diverse group of writers. These posts are designed to serve as a public forum for a discussion—broadly speaking—of psychiatry and its treatments. The opinions expressed are the writers’ own.