Is rTMS Ready for Primetime?

Increasingly, repetitive transcranial magnetic stimulation (rTMS) is vying for a front row seat in the management of major depression, even though the majority of studies testing its efficacy have been conducted in depressed patients with varying degrees of “treatment-resistance” —loosely referred to as treatment-resistant depression (TRD). This is problematic because TRD is a descriptor that is variably defined and used in a variety of contexts to describe the presentation of depressed patients where all manner of treatments, in particular pharmacotherapy, have been trialled (to varying extents) and have failed. ¹ In studies examining treatments for TRD, a specific definition for TRD may have been utilized, but because the definition varies considerably between studies, meaningful cross-comparisons of treatment efficacy are difficult, and the strength of evidence for any one particular definition and the positioning of rTMS remains weak. Thus, we argue that the logical positioning of rTMS in the treatment of depression is among therapies that are trialled when patients have failed to respond to tried and tested interventions such as lifestyle, psychological, and pharmacological strategies, and when rTMS is one of the options among remaining alternatives, either because of patient preference or because electroconvulsive therapy (ECT) is unsuitable.

As a potential antidepressant treatment, rTMS is appealing because of its tolerability and interesting putative mechanism of action. However, it is yet to demonstrate a robust therapeutic effect of its own, and even as an add-on to antidepressant treatment, rTMS has failed to show a significant augmentative effect in a large, randomized, double-blind, sham-controlled multicentre trial. ² Hence, rTMS does not warrant positioning alongside other standard treatments such as psychological interventions and antidepressant medications (on the basis of comparative efficacy).

Therefore, in this brief perspective piece we review how rTMS gained Food and Drugs Administration (FDA) approval via medical device pathways and argue that the threshold of evidence should have been set much higher and at least on par with that used for medications.

Questioning the Indication for rTMS

In 2008, the FDA approved the first TMS device for “the treatment of a Major Depressive Episode in adult patients who have failed to achieve satisfactory improvement from one prior antidepressant medication at, or above the minimal effective dose and duration in the current episode” (NeuroStar®). The positioning of rTMS after only one failure of an antidepressant medication is notable because research from STAR*D had already clearly shown that effective responses could be achieved with pharmacotherapy, provided clinicians persisted with treatment; and that in real-world practice, this sometimes required 2 to 3 courses of antidepressants. ³ In other words, even though the initial antidepressant may not lead to recovery in all cases, by persevering with medication, the majority of depressed patients will eventually recover. And given the widespread availability and ease of access to antidepressants, the appropriate positioning of rTMS should have been after several (at least 3) antidepressants have been trialled. Thus, the approval of this indication by the FDA remains puzzling. Since then, additional TMS devices have been approved, and rather than specifying and restricting the indication to an antidepressant treatment failure “at or above the minimal effective dose and duration” for one treatment failure, it was broadened further to the “treatment of major depressive disorder (MDD) in adult patients who have failed to achieve satisfactory improvement from prior antidepressant medication in the current episode.” In other words, anyone who has trialled an antidepressant is eligible.

Similarly, and perhaps equally puzzling is the approval given by Health Canada in 2002 for the use of rTMS for depression, and the licensing of the Brainsway deep TMS system in 2013 for the “treatment of depressive episodes in patients suffering from major depressive disorder who have failed to benefit from or are intolerant to antidepressant drugs.” ⁴ Notably, this indication lacks specificity with regard to the positioning of rTMS, as it does not indicate the number or type of medications to be trialled before proceeding with rTMS, nor does it specify what exactly defines a failure to benefit from an antidepressant. It is important here to note that clinical indications for other treatments are also notably vague. However, these indications and regulatory approvals are based on evidence, and unlike the robust evidence base supporting the indications for various pharmacological treatments, the evidence supporting rTMS, and the way in which it has been evaluated, means that indications and the way in which these are used to inform treatment algorithms lack specificity and are open to misapplication.

Why “Satisfactory” Is Unsatisfactory!

The difficulty of defining treatment failure, difficult-to-treat depression and TRD, is widely acknowledged. ^5,6 By definition, TRD is a mixed population that comprises patients with diverse aetiologies that present with many complexities and comorbidities. This is because TRD is usually defined primarily in terms of pharmacotherapeutic failures and therefore the threshold for designation is somewhat arbitrary, and little consideration is given to psychosocial factors. ⁷ Furthermore, because patients are “treatment-resistant,” clinical response within these populations is usually modest at best. Furthermore, because the definition of TRD lacks sufficient specificity, any response that does occur in one TRD population is unlikely to generalize to another. Therefore, extrapolating from response in TRD to other presentations of depression is highly problematic and unjustified. Another concern is that remarkably little is known about both patient and illness factors that make an individual a good candidate for rTMS. ⁶

As written, the positioning of rTMS by the FDA fails to take into account many baseline clinical complexities such as use alongside other interventions but also broader challenges posed by the neurobiological complexity of MDD and clinical response to therapeutic interventions, which we are only beginning to understand more thoroughly.

Clinical outcomes for example, such as response, partial response, remission, relapse prevention, and failure to achieve any of these have increasingly well-defined parameters and that should be included and clearly specified in any rTMS indication. In this regard, the non-specific description “failed to achieve satisfactory improvement” does not take into account important factors such as the magnitude of response to prior medication, the different types of antidepressants and the different dosages and combination strategies that can be used to try to achieve a satisfactory response, and other common reasons for declaring failure, such as side effects and patient preferences.

Finally, the emphasis on “current episode” ignores the recurrent nature of depressive disorders and the importance of knowing about previous patterns of response. Thus, both the initial indication of rTMS and its subsequent broadening are extremely problematic.

The FDA Approval Process Leading to a Broadening of the Indication for rTMS

It is important to note that the FDA approval process for medical devices is different to that for pharmaceuticals. ^8,9 And as outlined below, its weaknesses have led to ever broadening utilization of rTMS, especially in commercial settings, with significant clinical and health economic consequences.

Strategy 1: Justified Failure

In 2007, an initial application for rTMS use was classified in the Class III category through the 510(k) pathway (see Table 1). This pathway requires that a new device demonstrates substantial equivalence (SE) to an existing medical device that is already in the market. This is referred to as the “predicate,” and any new device only needs to prove that its performance is comparable to an existing therapeutic device. The 510(k) pathway is much less stringent, faster, and less expensive than other pathways and is designed for devices deemed to present only moderate risk to patients.

Table 1.

FDA Classification and Process Through Which rTMS Was Approved (Modified From ^5,6 ).

Classification	Risk status	Mechanism	Process
Class I	Low	Not applicable to rTMS.	Not applicable to rTMS.
Class II	Moderate	Reclassification from Class III requires demonstration of substantial equivalence (SE) in safety and effectiveness to another device on the market ‘predicate’ or where no predicate exists through the ‘de novo’ process with substantial support/evidence that Class III status is not necessary.	rTMS developers petitioned for the reclassification through the de novo process. They presented post hoc evidence from previously declined application. This post hoc evidence was acceptable and approved by FDA, thus allowing rTMS to achieve ‘predicate’ status on which all subsequent applications have been based on.
Class III	High and novel, need to demonstrate safety and effectiveness; stringent controls	Typically, Premarket Amendment Approval (PMA), similar to New Drug Application (NDA), required unless device is substantially equivalent to another device on the market e.g. ECT as ‘predicate’ through the fast-track 510(k) pathway.	For rTMS, initially submitted under 510(k) pathway and when SE with ECT could not be demonstrated, submission was declined.

Developers of rTMS initially claimed that their new treatment was SE to the predicate, which in this case was ECT. This is because ECT is the recognized standard for managing TRD, and rTMS had been trialled in TRD populations. However, ECT has a very different mechanism of action to that of rTMS and it is often used as a targeted, first-line strategy for treating specific depressive subtypes (e.g., patients with psychosis), where ECT is preferable because it provides a rapid and robust response. These patients are not treatment-resistant per se. Instead, the populations for which ECT is particularly effective are well-defined, and on this basis, it cannot be regarded as an appropriate comparator—especially as it is far superior in efficacy in these populations.

In any event, this initial application was unsuccessful as there were no clear benefits of rTMS compared to sham treatment. Interestingly, the developers of rTMS had presented evidence from one multi-site randomized clinical trial and subsequent uncontrolled, or open label follow-up, of these same patients if they had not responded adequately to treatment (<25% reduction of baseline HAMD17). ¹⁰ Furthermore, although all patients were medication-free at the time of investigation, they had failed an unnecessarily broad range (1 to 4) of adequate antidepressant treatments in the current or previous episode. Naturally, because of a lack of efficacy, rTMS was not deemed to be SE to ECT; an appropriate outcome that seems self-evident in retrospect.

Strategy 2: Unjustified Success

A subsequent application was lodged through the de novo process, which was set up for devices where the safety profile and technology are thought to be reasonably well-understood, but the device itself has never been marketed in the United States. Approval through this pathway also requires a public presentation and review by the FDA Advisory Committee, and it is a route that is generally reserved for low-risk devices that have been unsuccessful in gaining approval through the 510(k) pathway, and where the indication has been further modified. Significantly, this process is also utilized for submissions where there is no predicate (see Fargen et al. ⁸ ), and therefore, in this de novo submission, there was no need to demonstrate equivalence to ECT (see Hines et al. ⁹ ). However, in addition to the vagaries of the pathway, the data presented for consideration were also unusual.

In this second submission, the proponents of rTMS presented a post-hoc analysis of the same data set that they had previously submitted in their first submission. ¹⁰ Interestingly, this post-hoc analysis now revealed superior effectiveness of active compared to sham rTMS, but only in those patients who had failed one adequate antidepressant treatment in their current episode of depression, when compared to those who had failed 2 to 4 adequate antidepressant treatments. ¹¹ Astoundingly, on the basis of this post-hoc analysis of previously considered data, the FDA approved and reclassified the treatment under Class II, that is, a moderate risk category (see Table 1). This only required modification of its indication to a “new type of therapeutic device” for the treatment of MDD that induces neural action potentials in the prefrontal cortex (PFC) “without inducing seizures in patients who have failed at least 1 antidepressant medication and are currently not on any antidepressant therapy.” In other words, rTMS could now be used second line after a single trial of an antidepressant. Furthermore, and even more remarkably, under the de novo process, devices that have been approved may be marketed and used as predicates for future 510(k) submissions. This approach, involving post-hoc reanalysis of existing data and the substitution of the standard used for comparison, has been rightly criticized (see Hines et al. ⁹ ).

In practice, clinicians are highly unlikely to view patients as candidates for physical interventions such as ECT after they have failed just one antidepressant trial, regardless of the severity of their episode, unless as discussed, they have a distinct subtype of depression for which ECT is indicated. Furthermore, they are also very unlikely to consider rTMS as an equivalent alternative to ECT. By allowing an alternative physical treatment that lacks equivalent efficacy to serve as a predicate, the threshold for gaining an indication has been compromised significantly, and especially since this 510(k) data submission and indication in no way reflects clinical practice or thinking. Indeed, to appropriately test rTMS and demonstrate that it has a place as a physical therapy, or has SE to ECT, comparisons between treatments need to involve patients who would otherwise have been selected to receive ECT.

Subsequent FDA Approvals

Since the initial FDA approval of NeuroStar®, 6 subsequent TMS device applications have been approved along with further modifications of the initial NeuroStar® device. These applications have all used the Class II Neurostar® approval under the generic name of Repetitive Transcranial Magnetic Stimulation (rTMS) System as the predicate. It is important to note, and emphasize once again, that for an indication through this alternative and much easier pathway, all that is needed is equivalence with this system—that is, equivalence with an earlier version of itself. Subsequent approvals have been based on modifications, but what is perhaps most striking is the increasing lack of specificity introduced in terms of the clinical populations being targeted. For example, instead of aiming for one treatment failure in the current episode (already a rather dubious predicate), applications have been vaguely directed at the “treatment of MDD in adult patients who have failed to achieve satisfactory improvement from previous antidepressant medication in the current episode.” This indication seems to be unnecessarily broad and vulnerable to very variable interpretation, especially as there are no additional exclusionary factors or cautions.

The Clinical Parameters of rTMS Treatment Indications

Satisfactory Improvement

In clinical trials, therapeutic benefit is often indicated by either full treatment response (≥ 50% decrease in symptoms from baseline) or remission parameters that can be reliably quantified. Cut-offs on rating scales can be used in conjunction with subjective ratings to attain a reliable and valid measure of treatment induced change (see Riedel et al. ¹² ). This is necessary to detect differences between groups and also to determine change over time. But in many rTMS studies, the cut-offs and/or differences in ratings are seldom stipulated in relation to defining what is “satisfactory or adequate” improvement (e.g., Health Quality Ontario ⁴ ).

At the same time, it is likely that a lack of consensus regarding how to define treatment response and remission has enabled imprecision, and the use of subjective measures in rTMS studies, such as “satisfactory improvement.” As a consequence, even a minimum level of improvement (e.g., a 25% improvement) is considered to be satisfactory in open label rTMS studies (see Hines et al. ⁹ ). Thus, overall, because the expected outcome is vague, poor standards of assessing response and remission are being utilized in rTMS research and regarded as acceptable for clinical use as compared to standards for pharmacological randomized controlled trials. ^9,13

Variability in Treatment Failure

Similarly, failure from prior antidepressant medication is also formulated somewhat ambiguously, as it does not take into account many factors, such as the different types and classes of antidepressants or dosage changes in a particular episode, which inevitably vary among patients and clinicians. For example, it is unlikely that failure from one antidepressant trial without a change in dosage or antidepressant type, or the use of augmentation can be considered equivalent to failure from multiple antidepressant types.

Therefore, it is important to emphasize that the view that one trial of antidepressants is a failure of all potential antidepressants is factually incorrect and has no foundation in trials or clinical experience. Due to a lower standard than that applied to novel pharmacotherapies, the evidence for the acute efficacy of rTMS as first line treatment for patients who have failed at least 1 antidepressant, is somewhat skewed and of limited quality. ^9,13,14 Furthermore, this problem cannot be overcome by meta-analyses. ¹³

Explicit Focus on Current Episode

Finally, focusing on a failure of achieving satisfactory improvement in the current episode is also extremely limiting and does not reflect clinical practice. ¹⁵ It undermines the importance of assessing responses to previous treatments which are valuable for clinical decision making. ^15,16 In practice, to treat a current episode effectively, a strategy that considers past-history and allows for rational decision-making is necessary so as to take full advantage of the diverse armamentarium available.

Summary

An indication for a treatment that is approved by a regulatory body is supposed to be informed by the available evidence. However, in many instances, the necessary evidence may not exist, or if it does it may not be decisive, and thus indications can be an extrapolation based on existing evidence. A similar and subsequent leap is then made from both the evidence and regulatory indication to clinical practice guidelines, which generally try to synthesize the evidence and clinical experience and inform clinical practice. But ultimately, real-world treatment still differs because of the many contextual factors that these processes and original studies have not been able to take into consideration. This means that indications for treatments undergo a natural transmogrification that leads to successive changes and modifications being made at every step depending on whether the extant evidence is catering to regulatory needs, the needs of guidelines or the needs of clinical practice. Given the inherent flux within these perspectives on treatment indications, it is all the more important to ensure that the ultimate use of an intervention remains true to its evidence. In this context, the unusual pathway that rTMS has taken to achieve its indication for the treatment of MDD—involving a significant compromise of standards—has meant that rTMS has gained indications that are not keeping with clinical practice. This poses serious challenges for clinicians as the clinical validity of rTMS as an effective treatment alternative for depression remains unknown. Therefore, rTMS needs to be regarded as an experimental tool and further research needs to be conducted to determine its clinical potential. The FDA urgently needs to reform its de novo classification processes and regulations so as to provide structure, clarity, and transparency for medical device submissions.