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

The management of treatment-resistant depression is challenging. Compared to depression that responds, the severity of the illness is often greater, as are the associated risks of treatment. Nevertheless, decision-making must be logical. The case of a woman with treatment-resistant depression seen at our tertiary outpatient service (CADE Lithium Clinic, Royal North Shore Hospital) highlights a number of concerns where there is a lack of standard guidance to inform the use of repetitive transcranial magnetic stimulation (rTMS) in the treatment of depression.

The patient, a 47-year-old married mother of two, had an extended history of recurrent major depressive disorder compounded by anxiety (see Figure 1). Her early episodes of depression responded to treatment with medications but, as the illness progressed, her episodes became less responsive. Her most recent past episode had responded to electroconvulsive therapy (ECT) and she achieved remission which was sustained for 8 years while she continued to be maintained on several medications. Her current episode of depression began in 2016 after a bout of pneumonia. As symptoms emerged, the dosage of her long-term tricyclic antidepressant treatment was increased and augmented with lithium and adjunctive olanzapine – to no avail. Hence, she was eventually hospitalised in order to receive a course of ECT, totalling 14 sessions, followed by maintenance ECT. Despite responding to ECT, intolerable cognitive side-effects (affecting memory) meant that maintenance ECT was ceased before remission was reached. Instead, aripiprazole was prescribed and supplemented with alprazolam as needed, but with little benefit.

Figure 1.

The patient’s history of depression. (A) She had previously experienced two brief early episodes that responded to SSRI treatment and lifestyle changes.  After 3 years, she entered an extended episode of 3-year duration, followed by a remission of 1 year, and another episode of 2-years duration. During these latter episodes, the patient showed no or partial response to 10 psychopharmaceuticals and combinations thereof. However, the patient did respond to ECT which she received after being hospitalised for her depression, alongside nortriptyline and lithium, and was in remission for the next 8 years, able to work and raise a family. (B) In the current episode, the first strategy trialled was dose increases and reinstatement of the medications that had been on board when she had previously remitted – nortriptyline and lithium, with the addition of olanzapine. However, she was hospitalised shortly after.  At this time, she received ECT and continued with maintenance ECT for the following 8 months. However, the side-effects were not tolerable, and so maintenance ECT was stopped. After 8 months of no change, the patient underwent 40 sessions of rTMS, with little to no benefit. The patient received psychotherapy throughout most of the time pictured.

Having partially responded to ECT which was stopped because of concerns regarding memory loss, the patient was then referred for rTMS, and 8 months after receiving their last ECT session they were administered 30 sessions of rTMS. The cost of these was covered by the public health system and her private health insurance. Subsequently, a further 10 sessions were prescribed, which she paid for herself (out of pocket). On top of this, a further 10 sessions were suggested by the rTMS practice, but the patient refused these. Her original psychiatrist, who had been managing her for the past decade, noted a small improvement during the treatment period, though this was not sustained. However, retrospectively, the patient reported no subjective benefit from the 40 sessions of rTMS.

This case raises several questions: First, was it logical to prescribe rTMS after ECT? Second, was the number of rTMS sessions prescribed appropriate? And third, was the financial cost of rTMS to the health system and the patient justified?

Sequencing of treatment

As to whether referral of the patient to rTMS was logical in this circumstance, it is important to note that she had responded partially to ECT and that this was stopped because of side-effects. While guidelines note that failed ECT should not be followed by a trial of rTMS (e.g. Milev et al., 2016), they do not address the substitution of effective ECT. Two small case series (Cristancho et al., 2013; Noda et al., 2013) have found that rTMS may be successfully substituted (such that it maintains or improves clinical status) for maintenance ECT terminated due to tolerability issues. This suggests that consideration of a trial of rTMS in this case was perhaps appropriate, albeit somewhat experimental. On the other hand, the patient’s depression is characterised by marked anxiety, which is known to be an indicator of poor treatment response in general, and also for rTMS specifically, thereby reducing the likelihood that the patient would respond. Furthermore, it is important to note that the patient managed her anxiety to some extent with alprazolam, which is of relevance because it has been reported that benzodiazepines may also attenuate response to rTMS. Determining the impact of these factors on the likelihood of treatment response and using this information to plan an appropriate sequence of treatment in this case (e.g. washout of the benzodiazepine prior to rTMS treatment) has fallen outside the scope of clinical practice recommendations. Therefore, guidelines must address these gaps by systematically investigating whether rTMS can be substituted for effective ECT, and by clearly informing clinicians about characteristics that moderate a patient’s likelihood of responding to rTMS treatment.

Number of treatment sessions

Second, the patient completed a large number of rTMS sessions without ever showing significant benefit. Studies suggest that rTMS response is best predicted by whether the patient has responded after 26–28 sessions (Milev et al., 2016). Accordingly, the recent Canadian Network for Mood and Anxiety Treatments (CANMAT) guidelines suggest that an adequate trial of rTMS is 20 sessions, and an extension to 25–30 sessions is appropriate only ‘if improvements occur’ (Milev et al., 2016: 563). Indeed, to date, no randomised controlled trial of rTMS has administered more than 30 sessions of treatment (Brunoni et al., 2017). In this case, the patient received 30 full sessions and did not reach remission. She then received another 10, again without reaching remission, after which a further additional 10 sessions were suggested by the rTMS proprietor. This would have resulted in the administration of a total of 50 sessions if the patient had not refused due to a lack of observable benefit up to that point. The clinical rationale for this is unclear and not supported by any randomised controlled trial evidence. What it does highlight is the lack of standardised guidance to aid clinical decision-making as to whether continuation of treatment is appropriate. Given that (1) rTMS is an attractive alternative to established treatments due to its ease of administration and low level of side-effects, (2) it is often the most vulnerable patients who are considering its use and (3) most health practitioners managing depressed patients are not familiar with the rTMS evidence base and so are less able to question the practice of rTMS administrators, it is especially important that a standardised, evidence-based protocol is developed to justify the number of rTMS trials prescribed and to ensure that patients are fully informed of whether they are likely to respond to further treatment sessions.

Cost

Third, the patient completed 40 sessions of rTMS at a substantial cost to the healthcare system and to themselves. In Australia, rTMS is not currently covered by Medicare. The Federal Government has rejected applications for rTMS to be covered by Medicare twice, most recently in 2014, determining that the efficacy and cost-effectiveness of rTMS compared to already available treatments is unclear (Medical Services Advisory Committee, 2014). Hence, patients usually receive rTMS as inpatients. In practice, this means that they either have to deteriorate to the point of needing hospitalisation, or they are unnecessarily admitted to hospital principally for this purpose. Initially, rTMS was touted as a hospitalisation-free alternative to ECT. Requiring hospitalisation for administration of rTMS increases the costs associated with the treatment, both directly, as it is more costly for the healthcare system and, indirectly, as the patients must take more time away from their usual activities.

Conclusion

Ultimately, a lack of research into the specific factors that impact the efficacy of rTMS in the treatment of depression means that clinical practice recommendations on sequencing, counteractive clinical variables and protocols for extending treatment sessions are incomplete and it is thus important for clinicians to consider whether its use (and cost) is justified in each case and to clearly communicate the rationale for treatment decisions and likelihood of response to each patient. It is understandable that in times of treatment non-response any change seems better than no change, but a logical approach informed by evidence must be adopted if effective treatment is to be found. Improved guidance is required; in the interim, greater vigilance is needed to ensure that patients are not administered treatments that are unlikely to be beneficial, but sure to be costly.