Response to: Do we need to know more about repetitive transcranial magnetic stimulation in the treatment of depression?

Professor Malhi and colleagues (2019) in their recent debate article have raised a number of interesting and important issues with regard to the use of repetitive transcranial magnetic stimulation (rTMS) treatment of patients with depression that are worthy of comment.

Sequencing of treatment

First, the authors discuss the sequencing of treatment and in particular question the use of rTMS as a substitution treatment under the circumstances of ‘effective electroconvulsive therapy (ECT)’. However, it is debatable that the course of ECT treatment provided could truly be considered effective. The patient responded initially to ECT but intolerable cognitive side effects determined that maintenance ECT could not be continued and remission was not reached. Clearly, the definition of effective treatment must take into account whether this can be continued to the point at which a patient has made a meaningful and persistent therapeutic response. Under the circumstances, the authors question whether rTMS is an appropriate therapy and whether there is an evidence base to support its use. However, they do not provide any discussion of the broader therapeutic context.

In fact, there is no substantive evidence base to support any therapy at all under the circumstances of medication and ECT failure (including the other treatments used in this patient including trials of aripiprazole and alprazolam). Does this mean that we should send the patient away and offer no treatment at all (as I have seen some psychiatrists do as they are ‘no longer too able to offer anything’)? This seems completely unreasonable. Patients should be able to consider the use of any therapy to which they have not previously failed to respond. Trialling rTMS is as justifiable as any other option – perhaps more so given that the evidence for its value in treatment-resistant depression is considerably greater than further courses of medication. Of importance, we have previously found that a failure to respond to ECT in a previous treatment episode was not associated with a poorer response to rTMS in a substantive cohort of heterogeneous patients receiving rTMS in relatively real-world circumstances (Fitzgerald et al., 2013).

Patient selection

The second issue raised by Malhi et al. was to whether there were clinical characteristics suggesting that rTMS therapy should not be chosen for this patient. Specifically, whether her anxiety and use of alprazolam were contraindications or meaningful negative predictors of likelihood of response. In the largest study exploring potential predictors of antidepressant response, response rates in patients with comorbid generalised anxiety disorder were lower than in patients with no comorbidity at all, but the overall response rate in patients with this comorbidity was still 47.3% (Fitzgerald et al., 2016). Unfortunately higher, or in this case somewhat lower, response rates based on a single clinical characteristic have almost no predictive power in individual patients. Minimal research has explored whether benzodiazepine use predicts clinical response. One recent naturalistic study did suggest that this might be the case, but as acknowledged by the authors this result requires replication before it should be used to alter clinical practice (Hunter et al., 2019). This is certainly the case where weaning a medication like alprazolam may cause a worsening of anxiety symptoms and in the short term raises TMS-related seizure risk. Of interest, concurrent treatment with mood stabilisers, some of which also affect the GABAergic system, was associated with significantly higher response rates than non-use in our previous analysis (52.7% vs 43.8%; Fitzgerald et al., 2016).

Treatment course

As identified by the authors, there is a relatively sparse literature exploring what is a sufficiently long course of TMS therapy prior to a determination of non-response. However, studies are starting to emerge suggesting that longer courses of treatment may result in therapeutic responses not seen with briefer therapy (Tendler et al., 2018), and within trials, more days of treatment have generally been associated with greater effects. The referenced Canadian Network for Mood and Anxiety Treatments (CANMAT) guidelines do suggest that the extension of treatment beyond 20 sessions should occur when some improvement has been seen to occur. This appears to have been the case here where the authors document that the patient’s independent psychiatrist had noticed some degree of improvement during the course of treatment, although this was not noticeable to the patient. Adequate consideration of this does need to take into account the paucity of other therapeutic options available to the patient: clearly, a recommendation to continue for a longer period of time is more likely to be appropriate under circumstances where the patient has few options. We do, however, require further research as to ‘how many treatments are enough’. Decisions around this in the meantime are hopefully being made in collaboration with fully informed patients.

Treatment costs

The final issue raised in the debate piece refers to the cost of rTMS therapy. Malhi et al. stated that Medicare coverage of the treatment was rejected based upon a finding that the ‘efficacy and cost effectiveness of rTMS compared to already available treatments is unclear’ referencing the 2014 Medicare assessment. This assessment is now out of date. In 2018, the Medicare Services Advisory Committee (MSAC) reassessed an application for funding of rTMS and deferred its advice to the Minister pending a further submission addressing questions raised by the committee, which is currently progressing through the Medicare assessment process. The context here is also worth noting. Due to the rather unique circumstances around the provision of rTMS treatment (especially, the lack of a single commercial entity with a strong financial incentive), the submissions to MSAC in the past have not been supported by the same resources that typically come to bear on applications where there are strong financial interests – applications authored primarily by members of the Royal Austra-lian and New Zealand College of Psychiatrists with an interest in this area, rather than professional agencies funded by large device or pharmaceutical corporations. This, in my view, has certainly contributed to the failure of rTMS therapy to successfully progress through this system to date. Of note, several industry-supported applications for rTMS approvals and funding in other countries have been successful.

I would strongly agree that there are major clinical and financial disadvantages to our current state of rTMS treatment provision in Australia (where most rTMS is currently provided to inpatients, in contrast to the rest of the world where it is safely used as an outpatient treatment). Administration of rTMS through often unnecessary and prolonged admissions has substantial clinical disadvantages, exorbitant and unnecessary direct and indirect costs. Patients are likely to do better when not dislocated from family and supports. Treatment tapering may also play a role in reducing the likelihood of early relapse (this is not usually done during inpatient treatment). Hopefully a successful resolution of the Medicare assessment process will provide patients more meaningful options – this will have significant clinical advantages and save our system many tens of millions of dollars annually.

Conclusion

Yes, we need to know more about how best to use rTMS in depression. But we also lack evidence-based guidance for most treatment decisions in psychiatry after the initial choices of therapy. All clinical practice balances an awareness of what we know and don’t know with hopefully judiciously applied clinical judgement and collaborative decision making with informed patients.