Repetitive Transcranial Magnetic Stimulation for Depression: Managing Poor or Late Response

Tue, 11/09/2021 - 10:03

By Timothy Buckley a,c, and  Kevin Reeves a,b,

aClinical TMS Society, Clinical Standards Committee, Fresno, California

bOhio State University, Wexner Medical Center, Department of Psychiatry, Columbus, Ohio

cHines Veterans Affairs Medical Center, Mental Health Service Line, Hines, Illinois

Abstract:

Repetitive Transcranial Magnetic Stimulation (rTMS) is now widely used for depression, with several devices FDA approved for this indication. While the field is evolving rapidly, the question of what to do with patients who have little or no response early in treatment (e.g., after 10-15 treatments) to the standard protocol remains difficult. A similar dilemma can occur at the end of a series when a patient has had only a partial response. There is intriguing work outlining new protocols and augmentation strategies, but at this time, there is still insufficient randomized controlled (RCT) data to guide management in these non-responders. There is, however, a body of open-label continuation studies regarding extended courses of treatment. There are also a few retrospective studies examining the trajectories of response over a treatment course. While RCTs are the gold standard for clinical data, it is worth noting that it may be unlikely that any new RCTs extending the current FDA protocol for depression emerge, given that the field is now focusing more of its resources on newer protocols and various indications. As such, the clinician seeking guidance on this question may need to rely more on open-label studies in conjunction with many individual patient factors. Clinicians may decide to abort a course of treatment when significant benefit seems unlikely, thereby sparing the patient the burden of several weeks of futile effort and expense. An alternative argument may be that completing a course of treatment despite minimal response provides a baseline for future reference. As treatment protocols evolve, an improved response may be a possibility.  Furthermore, many clinicians are already using alternative protocols to augment or switch treatments when the risk/benefit ratio seems favorable.

 

Background

The standard protocol for rTMS is 30 treatments over six weeks, which can be a significant commitment and burden for patients. While metanalyses show encouraging response rates, most study subjects still fall in the non-response or partial-response categories12. It is common for patients to show little or no improvement at the 2- or 3-week assessment points3. The dilemma for clinicians is recommending continuing treatment for another 3 to 4 weeks when there are minimal gains. Just as with medication trials, along the lines of sound clinical methodology, clinicians generally want to ensure the current treatment has been given adequate time to take effect before proceeding with the next treatment option.  Neither physicians nor their patients want to give up hope prematurely. When both parties support continuing treatment, the dilemma may not be as acute. However, when patients make significant sacrifices to   undergo the treatment and seek guidance from the physician, or when other factors such as insurance reimbursement or limited clinical resources are  in play, the question of whether to continue or not becomes more pressing. Complicating this question is the fact that rTMS is evolving rapidly and hopefully will become more effective in the future.

Moreover, as we learn more about the mechanism of action, we may find that there are benefits that are only seen when augmentation with other treatments are paired with rTMS. For example, are there delayed or occult benefits such as priming the patient for psychotherapy even when rTMS all by itself has no apparent benefit4. In short, there is no easy answer to this dilemma, and it will inevitably be a decision made on an individual basis depending on numerous variables. Below we will try to summarize and rank 5 some of  the research evidence that may assist clinicians in making these difficult treatment decisions. Other practices in rTMS, such as redoing the motor threshold, double-checking the location of the treatment target, revisiting the initial evaluation and diagnosis, reviewing medications, and identifying other psychosocial factors at play, will not be addressed as these should be part of standard TMS practice.   

 

Defining “non-response” in the clinic

Before looking at the evidence, it may be helpful to examine the definition of non-response and its relevance to clinical care. Translating the research literature into real-world patient care can be tricky. Non-response has a precise meaning in clinical trials, and numerical cut-offs of measures determine it. The standards themselves are chosen with many considerations for research, including the logistics of administration, the unique population being studied, or the need to compare outcomes to other studies within the field. Tools such as the HAM-D are a standard by which studies can compare themselves 6. The downside is that the standards tend to change slowly.

Therefore, a measure that suits a research study may not ideally suit a particular clinic patient or may provide only a limited data source 7. As Nierenberg 6 points out, the research categories of remission, response, and non-response can be misleading when applied to clinical care. Clinicians can consider several different metrics for depression and may be surprised at the variability in results. What’s more, subjective reports by patients can conflict with their metric results. Their accounts can also be rich in nuance that is not captured by the metrics

being used. At Baseline, a broader array of metrics may help identify what the key markers of improvement are likely to be for a given patient7. We must remind ourselves that for treatment-resistant (TRD) patients, even a small improvement can represent a glimmer of hope after years of disappointment.

 

Trajectories of response and prediction

Trajectories of response or dose-response patterns can help guide clinical care. Most experienced psychiatrists have spent countless hours advising their patients that it may take 4 to 6 weeks to see a response with SSRIs. rTMS providers may find themselves wondering if a patient’s flat or sluggish response at two weeks might suddenly “kick-in” (be non-linear) at week three or beyond. Conversely, they may wonder if most of the benefit occurs in the early stages of treatment. What does the data say about trajectories? A couple of secondary analyses of level I evidence trials have examined trajectories specifically, along with one retrospective case series of 101 patients. The findings are mixed: Kaster et al. 8. conducted a secondary analysis of the Three-D study (N-388), the non-inferiority study comparing the standard 10hz protocol to intermittent theta-burst (iTBS). Using a different type of modeling than the original study (group-based vs. linear mixed), they identified four distinct trajectories of response, two of which were roughly linear, and one, a fast-responding group, which was non-linear.  The fourth was a non-responding group.

Additionally, their modeling found that these distinct trajectories or subgroups of patients were associated with specific clinical characteristics. For example, higher severity is a negative predictor, and older age is a positive predictor. Notably, the linear response groups did not reach a plateau at six weeks, which is a feature of early large rTMS trials. (The relevance of this to extended treatment will be discussed below.)  Another secondary analysis by Kaster et al. 9. examined data from two trials (N=189) which compared accelerated treatment to standard (one using accelerated iTBS and the other accelerated standard 10hz). Group-based modeling of MADRAS scores again found four distinct trajectories, although in a different pattern.  However, in this study, all trajectories were roughly linear. These findings could prove useful as it gives us a way to predict the likelihood, and roughly the amount, of response that will be achieved at the end of a course of treatment based on a patient’s early response, say at ten treatments. There is also some interesting and perhaps more generalizable level IV evidence. Following an approach used with both ECT and pharmacotherapy, Feffer et al.10, in a retrospective case series of 101 patients receiving bilateral dorsomedial prefrontal cortex stimulation with either 10hz or iTBS, analyzed the results to determine the negative predictive value (NPV) of responses at the end of 2 weeks (total treatment was four weeks). Using a cut-off of <20% improvement they achieved a NPV of 88.2% using conventional non-response criteria and 80.4% using more stringent criteria (<35% vs. <50% response). Interestingly, examining the distribution of outcomes identifies three distinct (trimodal distribution) groups, echoing the trajectories analysis above.

Finally, they created from their data a tool for predicting outcome based on a patient’s early response, broken down to every % point from 0-50% and providing NPV for both conventional and strict non-response criteria.

A later study11, replicating the methods of Feffer et al., but using standard 10hz to LDLPC extended to 6 weeks and emphasis on different metrics (IDS-SR and PHQ9 vs. BDI-II in the earlier study), found that the NPV for similar cut-offs fell to 72% when attempting to predict out to 6 weeks. Their 4-week outcomes yielded an NPV of 93%. In their conclusions, they supported further validation of such an approach but cautioned about missing late responders. If validated by future studies and perhaps even refined by adding in subgroups of clinical characteristics, tools like this could be extremely useful for clinicians, saving individual patients time and effort when non-response is highly likely. It would also help with resource utilization.

Last but certainly not least, there is important level II evidence supporting the findings above regarding subgroups or patterns of response in the form of a pooled analysis (N= 1,132) of 11 trials (several, but not all, of which were sham-controlled) by Fitzgerald et al. 12. Their 2016 analysis identified a bimodal distribution of outcomes with 46% responding and 32 % not responding. The 32% of non-responders were defined as 0-25% reduction in symptoms. Another 20.5% were partial responders (25-50%). Separating partial responders is significant here, as this group fell roughly between the other groups at the nadir. While the pattern is not identical to those described in the secondary analyses above, it does seem to support the finding that there may indeed be distinct subgroups of patients. They also found associated clinical characteristics such as older age and recurrent episodes (positive

predictors of response) and treatment resistance (negative predictor of response) which partly confirm findings in the studies mentioned above. This would seem to lend weight to the idea of using NPV tools for making decisions about aborting treatment. However, they also caution that the signals of associated clinical characteristics were not strong enough to warrant use in patient selection for treatment.

 

Extended treatment

Closely related to the question of trajectories is that of extending treatments. As Fitzgerald noted in his recent update13, most of the major clinical trials did not identify a plateau of response at the end of the studies (ranging from 3 to 6 weeks). There is, therefore, a possibility that maximal response, or a plateau, occurs on average somewhere beyond six weeks of the standard FDA-approved protocol. There is still no level I evidence for additional benefit beyond six weeks, and there is only one study to date that would meet Level II evidence criteria on this question for the standard protocol for the left dorsolateral prefrontal cortex14. This is the only clinical trial for this protocol that has extended randomization beyond six weeks, and notably, it found a plateauing of response at five weeks. Notably, a recent consensus statement addresses the question and evidence briefly.15

Another extended trial that was randomized into an extended phase was the Brainsway deep TMS pivotal trial16. However, this was a post-hoc analysis, blinded but active only (no comparator group) extension and therefore still falls within level II evidence (see below). The former study by Zhang et al. 14. was a large, randomized study with an N of 221 extended randomization to an optional 12 weeks but had a couple of significant limitations. It compared four arms examining 5Hz and 10 Hz and both left and bilateral treatment. However, the number of pulses per session was lower (1,200) than most other major trials. Secondly, it did not include a placebo arm, so while patients were blinded to the specific treatment parameters, they were essentially analogous to open-label subjects. This would simultaneously seem to cast doubt on the effect size (larger than average for randomized trials) but also strengthen the finding of a plateau at five weeks in contrast to the aforementioned open-label extension trials, which would be presumed to be skewed by a placebo effect. In short, more randomized controlled trials beyond six weeks are needed.

Regarding level III evidence, some open-label (OL) extensions of major clinical trials have shown additional responses beyond six weeks ranging roughly between 20 and 60%. In  2007 O’Reardon et al.17. published the findings of one of the early large multicenter RCTs for the safety and efficacy of TMS for depression. As an add-on study to this, Janicak et al. 18. conducted a durability follow-up study, part of which included a 6-week OL extension for non-responders (0-25% response). While a group of non-responders did achieve partial response in the OL phase, the study focused more on the question of relapse occurrence across various levels of response. McDonald et al. 19, in 2011, conducted an OL extension of the OPT-TMS RCT, which addresses the question of optimizing the number of treatments more directly. Phase I was the randomized, 3–6- week study which started with an N of 190 and utilized standard 10hz at 120% motor threshold over the dorsolateral prefrontal cortex (DLPFC), 3,000 pulses per session. These subjects had between 3-6 weeks of treatment and up to 90,000 pulses. Remission in this study was defined as a HAM-D score of <3 or two consecutive weeks under 10. Subjects who failed to achieve remission in phase I had the option to join this OL extension study (phases 2 and 3). Phase 2a was an additional 3-6 weeks of fast left treatment. Those who failed to remit 2a could opt to enter phase 2b, which was up to 4 weeks of slow right DLPFC (1hz for 30 min, 120% motor threshold). 16% (22/141) of subjects in the OL fast left (phase 2a) remitted. 26% (21/81) of the patients in the OL slow right (phase 2b) remitted for a total of 30.5% (43/141) of patients remitting who initially joined the OL phases. The two main drawbacks of the study are that one, it was open-labeled, and two, subjects were receiving treatment for as many as 16 weeks which begins to include a significant temporal confound, i.e., some of these subjects may have remitted in this time frame even without treatment. However, since this was a treatment-resistant group initially, this is likely not as great a factor. Interestingly, as the authors note, the 2b group (slow right) did not show a relative correlation with medication resistance as the 2a group had, suggesting that 1hz to the RDLPFC may represent a distinct subgroup as such, support switching protocols for non-remitters.

With regard to deep TMS, the Brainsway study 14 mentioned previously extended their 4-week double-blinded initial phase with up to another 12 weeks of twice-weekly double-blinded but all active continuation treatment. For this continuation study, only subjects with active treatment in the acute and continuation phases were studied. 61% (20/33) of non-responders to the acute phase (<50% reduction in HDRS-21 score) went on the achieve responder status in the continuation phase within four weeks of twice-weekly treatments. Lastly, but importantly, one large registry trial (N=307) 2021 showed a  subset of patients (22%) attained response (<3 on the CGI-S) beyond six weeks, echoing the results on the above open-label studies in real-world settings.

 

Conclusion

Choosing an endpoint for treatment in patients who respond minimally or late in the course of rTMS, for now, is a difficult dilemma and one that may not receive much more guidance from randomized controlled trials concerning the current FDA-approved protocol. The good news is that the field is evolving rapidly. New data on biomarkers and predictors of response may eventually help us with decisions about patient selection.222324 New protocols may offer  options for switching from the standard protocol when it is ineffective.25 Accelerated protocols now being studied may relieve the burden of time for patients, making decisions about extending treatment less troublesome.

Perhaps more importantly, the field of psychiatry is seeing a burgeoning of new treatments, including novel neuromodulation and pharmacological therapies and new paradigms of treatment such as psychedelic-assisted psychotherapy. As the toolbox grows, our algorithms for selecting treatments and combining treatments (in series or in parallel) become exponentially more complex. This is excellent news for the field at large but a daunting task for the busy clinician who will find it increasingly taxing to stay abreast of the developments in all of these areas. It is our hope that the summary of the literature provided above may provide some grounding to the very difficult clinical decision at hand. While falling well short of being a set of guidelines, it is hopefully a context and reference point to assist with these decisions.

 

References

  1. Berlim MT, van den Eynde F, Tovar-Perdomo S, Daskalakis ZJ. Response, remission and drop-out rates following high-frequency repetitive transcranial magnetic stimulation (rTMS) for treating major depression: a systematic review and meta-analysis of randomized, double-blind and sham-controlled trials. Psychol Med. 2014;44(2):225-239. doi:10.1017/S0033291713000512
  2. Fitzgerald PB, Hoy K, Gunewardene R, et al. A randomized trial of unilateral and bilateral prefrontal cortex transcranial magnetic stimulation in treatment-resistant major depression. Psychol Med. 2011;41(6):1187-1196. doi:10.1017/S0033291710001923
  3. Feffer K, Lee HH, Mansouri F, et al. Early symptom improvement at 10 sessions as a predictor of rTMS treatment outcome in major depression. Brain Stimul. 2018;11(1). doi:10.1016/j.brs.2017.10.010
  4. Dalhuisen, Iris, Eveline Ackermans, Liede Martens, Peter Mulders, Joey Bartholomeus, alex de bruijn, jan spiker, philip van Eijndhoven and IT. Longitudinal Effects of RTMS on Neuroplasitcity in Chronic Treatment-Resistant Depression. Eur Arch Psychiatry Clin Neurosci. 2021;271(1):39-47.
  5. Evans D. Hierarchy of evidence: a framework for ranking evidence evaluating healthcare interventions. J Clin Nurs. 2003;12(1):77-84. doi:10.1046/j.1365-2702.2003.00662.x
  6. Nierenberg, Andrew A. and DeCecco LM. Definitions of Antidepressant Treatment Response, Remission, Non-response, Partial Response, and Other Relevand Coutcomes: A focus on Treatment-Resistant Depression. J Clin Psychiatry. 2001;62:5-9.
  7. McClintock SM, Haley C, Bernstein IH. Psychometric considerations of depression symptom rating scales. Neuropsychiatry (London). 2011;1(6):611-623. doi:10.2217/npy.11.59
  8. Kaster TS, Downar J, Vila-Rodriguez F, et al. Trajectories of Response to Dorsolateral Prefrontal rTMS in Major Depression: A THREE-D Study. Am J Psychiatry. 2019;176(5):367-375. doi:10.1176/appi.ajp.2018.18091096
  9. Kaster TS, Chen L, Daskalakis ZJ, Hoy KE, Blumberger DM, Fitzgerald PB. Depressive symptom trajectories associated with standard and accelerated rTMS. Brain Stimul. 2020;13(3):850-857. doi:10.1016/j.brs.2020.02.021
  10. Feffer K, Lee HH, Mansouri F, et al. Early symptom improvement at 10 sessions as a predictor of rTMS treatment outcome in major depression. Brain Stimul. 2018;11(1):181-189. doi:10.1016/j.brs.2017.10.010
  11. Trevizol AP, Downar J, Vila-Rodriguez F, Thorpe KE, Daskalakis ZJ, Blumberger DM. Predictors of remission after repetitive transcranial magnetic stimulation for the treatment of major depressive disorder: An analysis from the randomised non-inferiority THREE-D trial. EClinicalMedicine. 2020;22:100349. doi:10.1016/j.eclinm.2020.100349
  12. Fitzgerald PB, Hoy KE, Anderson RJ, Daskalakis ZJ. A STUDY OF THE PATTERN OF RESPONSE TO rTMS TREATMENT IN DEPRESSION. Depress Anxiety. 2016;33(8):746-753. doi:10.1002/da.22503
  13. Fitzgerald PB, Hoy KE, Anderson RJ, Daskalakis ZJ. A STUDY OF THE PATTERN OF RESPONSE TO rTMS TREATMENT IN DEPRESSION. Depress Anxiety. 2016;33(8). doi:10.1002/da.22503
  14. Zhang T, Huang Y, Jin Y, Ma X, Liu Z. Treatment for Major Depressive Disorder by Repetitive Transcranial Magnetic Stimulation in Different Parameters: A Randomized Double-Blinded Controlled Trial. Front Psychiatry. 2021;12. doi:10.3389/fpsyt.2021.623765
  15. McClintock SM, Reti IM, Carpenter LL, et al. Consensus recommendations for the clinical application of repetitive transcranial magnetic stimulation (rTMS) in the treatment of depression. J Clin Psychiatry. 2018;79(1):35-48. doi:10.4088/JCP.16cs10905
  16. Yip AG, George MS, Tendler A, Roth Y, Zangen A, Carpenter LL. 61% of unmedicated treatment resistant depression patients who did not respond to acute TMS treatment responded after four weeks of twice weekly deep TMS in the Brainsway pivotal trial. Brain Stimul. 2017;10(4). doi:10.1016/j.brs.2017.02.013
  17. Ballenger JC. Efficacy and Safety of Transcranial Magnetic Stimulation in the Acute Treatment of Major Depression: A Multisite Randomized Controlled Trial. Yearb Psychiatry Appl Ment Heal. 2009;2009:219-220. doi:10.1016/s0084-3970(08)79304-6
  18. Janicak PG, Nahas Z, Lisanby SH, et al. Durability of clinical benefit with transcranial magnetic stimulation (TMS) in the treatment of pharmacoresistant major depression: Assessment of relapse during a 6-month, multisite, open-label study. Brain Stimul. 2010. doi:10.1016/j.brs.2010.07.003
  19. McDonald WM, Durkalski V, Ball ER, et al. Improving the antidepressant efficacy of transcranial magnetic stimulation: maximizing the number of stimulations and treatment location in treatment-resistant depression. Depress Anxiety. 2011;28(11):973-980. doi:10.1002/da.20885
  20. Carpenter LL, Janicak PG, Aaronson ST, et al. TRANSCRANIAL MAGNETIC STIMULATION (TMS) FOR MAJOR DEPRESSION: A MULTISITE, NATURALISTIC, OBSERVATIONAL STUDY OF ACUTE TREATMENT OUTCOMES IN CLINICAL PRACTICE. Depress Anxiety. 2012;29(7):587-596. doi:10.1002/da.21969
  21. Sackeim HA, Aaronson ST, Carpenter LL, et al. Clinical outcomes in a large registry of patients with major depressive disorder treated with Transcranial Magnetic Stimulation. J Affect Disord. 2020;277:65-74. doi:10.1016/j.jad.2020.08.005
  22. Bailey N, Hoy K, Rogasch N, et al. Differentiating responders and non-responders to rTMS treatment for depression after one week using resting EEG connectivity measures. J Affect Disord. 2019;242:68-79. doi:10.1016/j.jad.2018.08.058
  23. Downar J, Geraci J, Salomons T V., et al. Anhedonia and reward-circuit connectivity distinguish nonresponders from responders to dorsomedial prefrontal repetitive transcranial magnetic stimulation in major depression. Biol Psychiatry. 2014;76(3):176-185. doi:10.1016/j.biopsych.2013.10.026
  24. Shalbaf R, Brenner C, Pang C, et al. Non-linear Entropy Analysis in EEG to Predict Treatment Response to Repetitive Transcranial Magnetic Stimulation in Depression. Front Pharmacol. 2018;9. doi:10.3389/fphar.2018.01188
  25. Baeken C, Brem A-K, Arns M, et al. Repetitive transcranial magnetic stimulation treatment for depressive disorders. Curr Opin Psychiatry. 2019;32(5):409-415. doi:10.1097/YCO.0000000000000533