Abstract
Repetitive transcranial magnetic stimulation (rTMS) is a rapidly developing technique for the investigation of brain function and the treatment of neuropsychiatric disorders. Over recent years, considerable interest has developed in the use of rTMS in the treatment of depression and a considerable number of applied and basic research studies have been conducted. The change in status of rTMS, from a method under research to an accepted therapeutic modality, appears to be currently underway. This article will address whether this is timely or whether the more widespread application of rTMS should await further research data.
History and brief overview
Although the use of magnetic fields has a long history in medicine, the modern application of rTMS is relatively new and dates from the mid 1980s when the first stimulators were developed in Sheffield, UK [1]. Initially stimulators were used in neurology and neurosciences research but psychiatric attention was drawn to the technique in the mid 1990s following the publication of a number of promising treatment studies. Since this time, research has prospered with the publication of numerous studies into the mechanisms of action of rTMS and its therapeutic potential.
Repetitive transcranial magnetic stimulation involves the application of pulsed magnetic fields to the cortical surface of the brain with the aim of stimulating or disrupting ongoing brain activity. The magnetic field generated during rTMS is produced through rapidly switching on and off an electrical current in a coil held over the scalp. In commercially available stimulators, two types of coils are used: circular and figure-of-eight shaped. Figure-of-eight shaped coils, which consist of two round coils placed side by side, produce more focused magnetic fields. The orientation and intensity of the current that passes through the coil determines the type of tissue stimulated as well as strength of that stimulation. Using small figure-of-eight shape coils, neurones are activated in a cortical area of approximately 2 cm2 and to a depth of approximately 2 cm [2].
The use of rTMS in therapeutic paradigms is based on the application of repetitively applied pulses, usually at a frequency of 1Hz or greater. The repetitive stimulation of the underlying cortical region is hypothesized to result in a change in the activity of the corresponding neural tissue. There is evidence that low frequency stimulation (∼1Hz) can result in a down regulation of local activity and that higher frequency stimulation results in increased activity [3]. These changes have been demonstrated with the use of electrophysiological techniques as well as through the demonstration of stimulation related alterations in activity levels with functional imaging (for review see [4]).
rTMS and depression
Initial rTMS treatment studies in patients with depression provided stimulation at the vertex and showed no significant antidepressant effects [5, 6]. Specific targeting of the left dorsolateral prefrontal cortex (DLPFC) was suggested by studies identifying abnormal activity levels in this region in depressed subjects on functional imaging scans [7]. Subsequently, initial trials of DLPFC stimulation produced more promising effects using cross over designs [8, 9]. Since then, a number of randomized sham-controlled parallel design studies have been published. These studies have varied in patient selection but improvement greater than sham has been shown with both in- [8] and out-patients [9, 10]; with patients considered medication non-responsive [8, 9, 11] and those that were not [12]; and with patients concurrently treated with medication and those medication-free [10, 13]. The considerable majority of published studies have shown a greater effect of active over sham treatment and a recent meta-analysis of five trials identified a clear benefit of active over sham rTMS treatment [14]. The exception in published studies to date is the study of Loo et al. which showed a similar improvement in depression in both active and sham treated groups [15]. Difficulties with interpretation of the results of this study have been raised as there are questions as to the degree of intracortical stimulation provided to the sham/control group [16].
Doubts about the magnitude or clinical relevance of the effects seen in some have been raised. For example, in the study of Berman et al. only one patient experienced a greater than 50% reduction in Hamilton Depression Rating Scale (HDRS) score [13]. The reduction of HDRS in the study of Padberg et al. was also significant but only marginally clinical relevant [11]. Substantially greater effects were demonstrated in the studies of George et al. (nine out of 20 patients with active stimulation experienced a greater than 50% reduction HDRS score) [10] and Klein et al. (17 out of 35 ‘responders’ using the same criteria) [12]. Promising results were also produced by Avery et al. and Stikhina et al. [17, 18]. Of note, the studies of George et al. [10] and Klein et al. [12] used a higher stimulation intensity (100 and 110% of the resting motor threshold (RMT)) compared to several of the studies in which limited clinical affects have been achieved with intensities between 80 and 90% of RMT. The study of George et al. [10] also used a considerably higher total number of pulses in each subject than several of the studies with marginal results (16 000 vs 1250 for Padberg et al. [11] and 4000 for Berman et al. [13]). This may suggest that there is a dose related effect both in number and intensity of stimulation pulses. This is supported by more basic investigative studies, which have shown differences in the magnitude of biological effects with increasing pulse number and intensity of stimulation (e.g. [19–21]).
Several recent studies have reported promising findings with an alternative method involving the use of slow rTMS administered to the right prefrontal cortex (PFC). Studies of this sort include the randomized study of Klein et al. [12] involving a medication ‘responsive’ group, an additional open study [22] and a randomized study in medication non-responsive patients by our group. In contrast, high frequency rTMS to the right PFC appears ineffective in depression but potentially effective in mania [23]. It is possible that this lateralization of effects is due to a relative imbalance in prefrontal cortical activity [24]: the left PFC being differentially under-active in depression or the right PFC over-active and that the situation is reversed in mania. As rapid rTMS increases underlying activity levels this would increase prefrontal activity on the left in depression or the right in mania. Slow rTMS reduces activity and could correct a relative over-activity on the right in depression. This hypothesis is supported by studies that have looked at the physiological response to slow and rapid stimulation in the motor cortex and by functional imaging data (see review in [4]).
A number of promising studies have also compared rTMS with electroconvulsive therapy (ECT). Grunhaus et al. randomized 40 patients to either rTMS or ECT [25]. Treatment efficacy was equivalent for patients with non-psychotic depression although patients with psychotic symptoms responded better to ECT. Similar results were achieved in a separate study where severely ill patients were randomly assigned to receive ECT or up to 20 sessions of rTMS at 10 Hz and 110% of the RMT [26]. In a novel, single-blind design, Pridmore et al. randomized 22 patients to either ECT alone (standard six-treatment course over 2 weeks) or a combination of two ECT treatments and eight TMS sessions over 2 weeks [27]. There was similar clinical response between the two groups and fewer side-effects in the combination group.
The established rTMS research base has been criticized due to the lack of evidence of persistence of antidepressant effects beyond the time of treatment trials [28]. Although these studies are by necessity difficult to perform, some evidence is now emerging that rTMS effects may persist. A study by Dannon et al. found that 6 month relapse rates were identical for 20 patients having received ECT and 21 patients having received rTMS, all initially treated in a randomized treatment protocol [29].
Limitations in study design
Several issues have been raised in the interpretation of these key intervention studies. For example, concerns exist about the interpretation of the cross over studies, especially related to concerns about the unblinding of subjects [30]. The application of the type of sham can also be problematic with some coil orientations used for sham ‘stimulation’ found to produce intracortical stimulation [16, 30]. Ideally, a balance needs to be achieved between the production of some type of sensation and the minimization of intracortical effects. Finally, there appears to be a relationship between age and antidepressant response. Older age appears be associated with a poorer response to treatment [31], which may be associated with a greater coil to cortex distance [32].
Several other issues confound the interpretation of these studies and may contribute to the limited clinical effects seen in some. First, most studies to date have included fairly limited numbers of patients, especially compared to trials of pharmaceutical agents. It is therefore not surprising, given the well-recognized high degree of placebo response to depression, that some studies will have negative results. The inclusion of treatment resistant subjects in studies could also limit the magnitude of therapeutic responses, especially if we consider the likelihood of comorbid psychopathology that can be very difficult to exclude. Finally, no consistent method for precisely localizing the DLPFC site of stimulation has been applied in these studies and it has been reported that the method of targeting of the DLPFC in the published studies may be incorrect in a substantial proportion of patients [33]. This heterogeneity in the site of stimulation is likely to limit the degree of response across subjects.
Safety
The greatest concern raised with rTMS is the possibility of seizure induction, which is related to the parameters of rTMS applied and patient selection. There have been no reported seizures in therapeutic trials in depression with the application of safety guidelines developed following initial reports of seizures with high frequency stimulation [34–36]. In fact, low frequency stimulation, like that applied in right PFC treatment in depression [37] appears to reduce seizure activity [38, 39] and it has been suggested that the risk of seizures with carefully assessed patients may not be greater than that for antidepressant medication [24]. Studies that have tried to use rTMS to induce seizures as a form of nonelectrical convulsive therapy indicate that stimulation parameters considerably in excess of those used in standard rTMS protocols are required [40].
Another concern with rTMS has been its potential to induce cognitive or memory impairment. This has been assessed by a number of studies and while rTMS can clearly be shown to produce transient cognitive change (including both improvement [41, 42] and disruption [43]), rTMS treatment courses in depression are not associated with cognitive impairment (see review in [4]). It is also possible that rTMS has other deleterious effects on the brain. However, no problematic changes have been found with studies of the blood brain barrier [44], gross brain structure (with magnetic resonance imaging), electroencephalogram, electrocardiogram and neurohormonal levels [45]. One human pathological study revealed no adverse changes in the brains of two patients with epilepsy who underwent rTMS prior to operation [46]. One animal study reported microvacuolar changes with stimulation intensities equivalent to three times motor threshold (approximately three times the intensity applied in human studies) but this has not been replicated in at least four other studies that have shown no adverse changes (as reviewed in [47]). Finally, there has been one report of the development of delusions during rTMS treatment of non-delusional depression with the consideration that this may have been related to enhanced dopaminergic activity [48]. While this is a single case report and it is quite possible that the emergence of delusions in this patient was unrelated to rTMS, it an important reminder that the field is relatively young and vigilance is required in monitoring for emergent problems.
The actual reported side-effects of rTMS are relatively minor. Repetitive transcranial magnetic stimulation can induce a head or neck ache and it is possible that stimulation related noise could cause changes in auditory thresholds (this could be associated with hearing impairment) if ear-plugs are not used [45]. Although rTMS is well tolerated and positively judged by patients [49], it can be associated with site related pain or discomfort during the procedure, especially with stimulation at high intensity. Perhaps supporting the antidepressant effects of rTMS, there are reports of rTMS related switch to mania in patients receiving treatment for the depressive phase of bipolar disorder [50].
Discussion and comments
Clinical studies with rTMS show effects on mood and these appear to be clinically relevant. The findings to date are more impressive if we consider that many of the limitations of the studies (e.g. small sample sizes and poor accuracy in DLPFC targeting) would be likely to reduce the chance of positive study outcomes. The crucial question at the current time is whether the findings of these studies are sufficient to justify more widespread use of rTMS as treatment. This process has commenced with the approval by the Therapeutic Products Directorate, Medical Devices Bureau of the Canadian Health Ministry in Canada of one stimulator for use in clinical practice with certain clinical populations. This approval is for patients who meet the following criteria:
Are resistant to an adequate trial of antidepressant drugs. Adults aged 18–64. Meet the American Psychiatric Association Guidelines for ECT.
Applications for similar approval have been made to the Food and Drug Administration in the US although approval has not be granted to date [51]. The mechanism through which rTMS will be regulated will differ in Australasia than in North American countries. Repetitive transcranial magnetic stimulation devices are ‘listed’ by the Therapeutics Goods Administration in Australia but this listing refers more to electrical safety than therapeutic benefit. As has been recognized by the Royal Australian and New Zealand College of Psychiatrists, there is no clear alternative approval mechanism and as such the promulgation of guidelines by the college has great significance. The original guidelines produced by the college Psychotropic Drugs and other Physical Therapies committee in 1998 identified the experimental nature of the technique and indicated that rTMS should only be applied in the context of Human Research and Ethics committee approved clinical trial. However, this lack of ‘approval’ of rTMS has not stopped unregulated applications of its clinical use. As was discussed and made clear at a recent meeting of the International Society for Transcranial Magnetic Stimulation (ISTS) (Philadelphia, May 2002), rTMS is offered in clinical practice in a number of Western countries including Australia and there is no process of public regulation and audit of this use.
The decision on the regulatory approval of a method such as rTMS must take into account several considerations. These include the strength of the evidence of therapeutic efficacy as well as the accumulated data on safety and tolerability. In addition are considerations as to the therapeutic alternatives available to the patient population. This principle is accepted in pharmaceutical development and regulation where treatments may be ‘fast tracked’ or introduced with a more limited evidence base, when there is considerable clinical need. Although treatment-resistant depression (TRD) is poorly defined [52], it is clearly associated with considerable suffering and social morbidity [53]. Patients with TRD are often offered ECT as the most appropriate therapeutic option. However, there are considerable difficulties with the use of ECT despite its therapeutic efficacy. These include the stigma associated with its use, concerns associated with the application of multiple anaesthetic agents and the cognitive side-effects. Many patients will refuse ECT due to these concerns and others remain unable to access the procedure due to risks associated with medical comorbidity. Therefore, this clinical need may be said to argue for the ‘early’ introduction of a modality such as rTMS.
However, the possible benefits of the widespread introduction of rTMS must be balanced by consideration of the disadvantages of this action. First, if we introduce a therapeutic method before it is thoroughly evaluated, there is a danger of raising false expectations that will not bear out as the technique in applied in clinical practice. It may, however, prove very difficult to reverse the ‘approval’ decision once the technique is widely available in clinical settings. In addition, a premature release of a therapeutic technique may prematurely limit ongoing research and refinement of the method. This is of particular importance with rTMS where many questions remain to be answered. For example, we cannot be sure of the best site for treatment, the best method for the localization of that site and the optimal stimulation parameters. It remains unclear whether concomitant medications act synergistically with rTMS or even if they may lessen its efficacy as has been suggested with the use of anticonvulsants when rTMS is applied in the treatment of auditory hallucinations [54].
Addressing these questions is likely to require larger trials than those that have been performed to date, especially to look for subtle between group differences. As rTMS trials thus far have attracted limited industry support, it seems unlikely that the generous finances available for pharmaceutical trials are likely to be provided in the future. Therefore, these trials can only be conducted in a limited number of academic centres with the research infrastructure and funding available. Widespread availability of rTMS in community locations may divert patients away from these programmes, ultimately limiting the development of the field.
How is it best to balance these competing concerns? This involves balancing the needs and choices of individual patients who may choose to have rTMS, even when given a balanced appraisal of the research evidence, and the interests of future patients in supporting an environment best able to ensure that the evidence develops rapidly and appropriately. I would suggest that a balance of these needs could be best met under the following circumstances:
1. That rTMS be provided to patients only with a confirmed diagnosis of a depressive disorder that has failed to respond to other treatments or in a patient who is unable to tolerate other treatment strategies. At this stage, although studies are being conducted in a number of other disorders (for example [23, 54–56]), there is limited evidence for the efficacy of rTMS except in depression. Treatment for other disorders should remain limited to clinical research trials.
2. That rTMS is applied in adults between the ages of 18 and 70. There is a suggestion of reduced efficacy in the elderly from the limited trials in this population to date and more research is required before rTMS is more widely applied in this age group. Similarly, there is an absence of research in adolescent depression and rTMS should be applied with caution to the developing brain.
3. Repetitive transcranial magnetic stimulation should only be applied in centres that are active in ongoing rTMS evaluation and research. It could be argued that rTMS should still only be applied in subjects who are entered in Human Research Ethics Committee approved clinical trials. However, clinical trials are by necessity exclusive and this provision will considerably limit the application of rTMS in subjects traditionally excluded from trials and in indications that are a natural progression from the conduct of trials, for example repeat treatments for relapsed treatment responders. On the other hand, where considerable questions remain of the clinical utility of rTMS, its limitation to research ‘settings’ will have a two-fold effect. First, it will help ensure that all rTMS is conducted in an environment of continual evaluation and research, if not always actually in a clinical trial. Second, it will actively support the ongoing conduct of the sort of trials that are required to answer the outstanding questions in the current literature. The reporting of serious new adverse events is also more likely to be facilitated from this environment compared, for example, with private practice where the report of adverse events may be counter to economic interests. This should not inhibit the use of rTMS in new settings but argues that new groups establishing rTMS facilities actively engage in the evaluation of its efficacy.
4. As rTMS is associated with both known and potentially unknown risks, it should only be applied in medical settings, especially with the capacity to manage the development of seizure activity. This use should be consistent with the ISTS consensus statement on managing risks with rTMS [57]. In addition, procedures should be in place for the continuous monitoring of potential or new emergent side-effects. In the absence of a system, such as the adverse drug reaction reporting system, there is an obligation on rTMS practitioners to promptly report any significant side-effects in the professional literature.
Conclusions
That rTMS has antidepressant properties is supported by a number of lines of research including the body of clinical trial research. However, the high number of parameters that may be varied in a treatment protocol complicates its use and questions remain as to the best way to apply rTMS in clinical practice. As the main indication for rTMS is likely to be treatment for nonresponsive subjects for whom other options are limited, a good argument can be made that rTMS should now be more available. I support this but believe that widespread dissemination of rTMS in community settings remains premature. Greater availability, however, in academic and tertiary referral settings is appropriate and should be encouraged.
Footnotes
Acknowledgements
This paper was supported by a National Health & Medical Research Council grant (143651). Thanks to Tim Brown and Natasha Marston.