Deep Brain Stimulation for Highly Refractory Depression

Pharmacotherapies

The early antidepressants iproniazid and imipramine were first developed for tuberculosis and psychosis, respectively. Their antidepressant effects were discovered serendipitously; patients treated for those other illnesses had reduced depressive symptoms. The insight that these and related agents affected monoamine neurotransmission allowed the field to “improve on serendipity.” Thus drugs such as selective serotonin reuptake inhibitors were developed, and eventually became first-line antidepressants due to their better tolerability and reduced lethality in overdose. However, the earlier classes of antidepressants remain in use as second- or third-line medications in refractory cases. More than 20 antidepressants are commonly used. The drugs are usually grouped by their chemical classes or pharmacological actions, such as tricyclics and tetracyclics; serotonin reuptake inhibitors, which include the more selective medications; monoamine oxidase inhibitors; and those affecting other or combinations of biogenic amine systems. Medications from different classes are frequently combined, particularly in refractory cases.

Treatment of Refractory Patients

While efficacy of antidepressants is well demonstrated, they benefit many but not all patients. A key point to emphasize at the outset is that there are a number of different degrees of refractoriness or “treatment resistance.” It is instructive to review classifications of levels of poor responses to treatment (e.g., ). The important methodological point here is that entry criteria for studies of “resistant patients” may vary substantially. This applies to trials of any potential antidepressant treatment, including neurosurgical therapies. Thus differences in the degree of refractoriness, along with other characteristics of study patients, may be expected to affect efficacy rates of any given trial. This can complicate attempts to compare outcomes from different studies. It appears beyond dispute that affected individuals who have an inadequate response to all the treatments discussed below—medications, psychotherapies, and electroconvulsive therapy (ECT)—currently have little prospect of sustained recovery. But how specific differences in refractory or resistance criteria might affect outcomes remains poorly understood.

By any measure, the limits of conventional treatments remain a serious problem. The STAR∗D trial, a stepwise treatment protocol involving thousands of patients with major depressive disorder, highlights this issue. Actually a relatively small proportion of patients experience remission with their first antidepressant trial ( ). Overall, antidepressant monotherapy may bring about and maintain remission in about half of patients ( ). The most affected group remains refractory to all standard medication treatments for depression ( ). A proportion of this group might improve after more aggressive “augmentation” trials where other classes of psychotropic medications are added to antidepressants. Augmenting agents include mood stabilizers (lithium or anticonvulsants), neuroleptics, thyroid hormone, and other medications. Use of certain dietary supplements or “nutraceuticals,” including omega-3 fatty acids and S-adenosyl methionine, appears to be increasing (though systematic data is scant). The few agents approved for augmentation in refractory depression in the United States include the second-generation antipsychotic aripiprazole and the prescription “medical food” methylfolate.

Psychotherapies for depression, while considered first line in their own right, are very often used together with medications, especially in depressions of moderate or greater severity. Various forms of psychotherapy have been studied to different degrees. There is strong evidence for the efficacy of cognitive behavior therapy (and variants), interpersonal therapy, and family therapy for depression (e.g., ). But, as noted above, many patients remain severely affected despite aggressive use of conventional treatments such as these.

Brain Stimulation Techniques

ECT remains a therapeutic gold standard after 75 years (e.g., see ). In ECT electrical current is delivered to the brain across the large electrical resistance of the scalp and skull. ECT can be associated with significant adverse effects, however, particularly memory loss, which can limit its acceptance. Moreover, ECT’s therapeutic effects are transient in a large proportion of patients, and so continuation or “maintenance” treatment may be needed ( ). On the other hand, the recent development of an ECT technique using much briefer electrical pulses to induce convulsions reported a much lower rate of adverse effects on cognition, and is seeing expanded clinical use ( ).

Various stimulation methods have been studied as potential treatments for depression. These device-based stimulation modalities can alter brain electrical activity directly or indirectly. Transcranial magnetic stimulation (TMS) magnetically induces electrical currents in brain tissue using an electromagnetic coil placed on the scalp, and was approved for the treatment of depression by the United States Food and Drug Administration (FDA) in 2008. Both a manufacturer-sponsored trial and an independent trial sponsored by the National Institutes of Health demonstrated efficacy ( ). As for ECT, variations in how TMS is delivered are beginning to be explored. These include magnetic seizure therapy, different magnetic pulse sequences, and markedly different designs of the electromagnetic coils themselves. Variations in coil design can result in advantages such as markedly lower requirements for electric current. Intriguing new TMS devices create magnetic field geometries that should allow effective stimulation deeper in the brain, and one such TMS approach was approved for the treatment of depression by the FDA in 2013 ( ). Transcranial direct-current stimulation and transcranial alternating-current stimulation, both of which involve inducing low-amplitude electrical current directly to the scalp, are currently under investigation for clinical use ( ). Lastly, vagus nerve stimulation, approved by the FDA for the treatment of treatment-resistant depression in 2006, uses electrodes surgically wrapped around the left vagus nerve in the neck to activate its afferent projections to target nuclei and related neural circuits (see Chapter 42 in the present work).

In addition to the brain stimulation methods described above, neurosurgery remains an option for patients with otherwise untreatable and severe psychiatric illnesses, primarily depression and obsessive–compulsive disorder (OCD). Stereotactic ablative procedures like anterior cingulotomy and anterior capsulotomy continue in small-scale and/or research use in North America, Europe, and elsewhere. For MDD, symptom improvement has been reported in up to two-thirds of otherwise intractably ill patients after lesion procedures ( ). Rates of persistent serious adverse effects have been generally modest at the most experienced expert centers. But this is not true when the volume of tissue lesioned has been large, particularly for some procedures such as thermocapsulotomy or high-dose multiple-target gamma knife procedures (e.g., ). An advantage of deep brain stimulation (DBS) compared to ablative neurosurgery is that the effects of stimulation itself are reversible, though long-term or even irreversible side-effects of brain lead implantation have occurred. Another key issue in assessing the risks and burdens of DBS versus lesion procedures is the need for patients to have access to highly specialized expert treatment centers, essentially in perpetuity. This model of care, with all its advantages, can impose important logistical and financial burdens on patients, who by virtue of long-term disability and psychosocial dysfunction may have few resources.

History

DBS for psychiatric illness, and specifically for depression, is not a new idea. But the devices are new, and there are now empirical findings from stereotactic lesion procedures and neuroimaging that have allowed theoretical models of depression neurocircuitry to advance dramatically since earlier attempts in the 20th century. In 1948 used a silver electrode implanted in the caudate nucleus to try to treat a woman with depression and anorexia. Over subsequent decades, Heath, Sem-Jacobsen, and Delgado exemplified an earlier era of intracranial stimulation (see below).

Over the past 20 years the introduction and refinement of DBS for movement disorders have resulted in a renaissance in this branch of functional neurosurgery, and in the field more generally. In the United States DBS is approved for tremor and Parkinson’s disease and, under a Humanitarian Device Exemption, for dystonia. Worldwide, DBS is or is becoming a standard of care for such patients.

These developments spurred renewed interest in the use of such procedures for the treatment of other refractory neurologic conditions. As of this writing, DBS remains investigational for primary psychiatric disorders. Investigational uses of DBS for neurologic illness include epilepsy, pain, cluster headaches, tardive dyskinesia, Gilles de la Tourette syndrome, brain injury, and persistent vegetative states. For OCD, a Humanitarian Device Exemption enables DBS use in medication/psychotherapy-refractory patients, with the understanding that clinical benefit is not proven.

DBS was conceived as a treatment for psychopathology in the 1940s, when caudate nucleus stimulation was tried for treatment of depression and anorexia. In work that began soon afterwards and was contemporary with Sem-Jacobsen’s, Heath et al. stimulated the “septal region,” an area including the ventral anterior capsule (VC) and ventral striatum (VS) that is just posterior to our current target. Heath chose it in part because tumors there and nearby in the forebrain had been related to psychiatric symptoms. Heath et al. selected 20 patients with heterogeneous symptoms including delusions, hallucinations, poverty of speech or near-mutism, depression, and compulsions, though all had a formal diagnosis of schizophrenia ( ). Stimulation was limited to 1–3 days after electrode implantation, at an amplitude of 2–15 mA. Three of the 20 patients had “no objective signs” and a further two “could not be evaluated” during stimulation. The others had acute effects: “patients became more alert [13 of 15]; … had increased motor activity and spontaneous [speech] production; …[in] previously almost inaudible or expressionless [subjects], speech became louder and enunciation clearer and inflection more appropriate [in five who had been the least verbal].” One of these, “who had been almost mute, became talkative and later almost hypomanic.” Three patients appeared acutely more tense, two less so ( ).

Accompanying behavioral changes included improved social interaction and enhanced emotional expression. As observed by Monroe and Heath, DBS subjects demonstrated “ability to relate to other people, increased responsiveness to pleasure, gradual appearance of a sense of humor, and more overt expression of anxiety and ambivalence,” as well as improved functioning, e.g., “Less negativism… everyday problems were approached more realistically and more interest was shown in ward activities.” Eleven patients, described as generally “idle, seclusive, and withdrawn before operation, afterward participated actively in some or all of the ward activities.” Improved emotional responsiveness in social settings was “even more dramatic.” “Twelve patients showed significant improvement in their ability to relate to other people,” one of the “outstanding aspects” of which was the “emergence of pleasurable feelings.” Nine patients showed the “development of humor.” Some of these effects apparently persisted after stimulation ceased, though for how long is not fully clear. Monroe and Heath believed that “patients who respond particularly well… [were those] whose main abnormalities seem to consist of flattened affect or disturbed motor behavior.” The time course and persistence of therapeutic benefit after stimulation ceased is not entirely clear in this work, although effects apparently could be transient. Some lasting or emerging benefit might have been due to concerted multidisciplinary therapies also used in these patients, described as a “total push” approach—which had, however, also been tried before stimulation without improvement.

In our own experience to date, and that of others, ongoing DBS is required for persistent behavioral and emotional change. A potential exception to this, however, is the sustained benefit seen in two OCD patients after chronic stimulation. In these individuals, stimulation facilitated completion of courses of behavioral therapy (exposure and ritual prevention), which had been impossible for these patients before DBS treatment ( ). In this sense, lasting effects after DBS ceases might be possible. This intriguing possibility will require systematic study.

It is important to note that the early work, from the 1950s and later, predated modern research methods. Diagnostic and severity measures used did not meet current standards for reliability or construct validity, limiting interpretation. However, recorded observations of acute and subacute DBS effects (in patients diagnosed with schizophrenia) have high face validity as manifestations of affective state. These include enhanced production, volume, and prosody of speech, greater affective range, social relatedness, sense of humor, and functioning, and increased level of activation or hypomania.

Affect and Mood Effects Observed During Depth Electrode Stimulation

Understanding where brain stimulation effects may converge at a systems level is now a reasonable goal. Observations of how DBS for movement disorders changes affect and mood continue to accumulate. They point to neural networks that might represent potential therapeutic targets for primary psychiatric illness. Taken together with early attempts with focal brain stimulation, they suggest that multiple stimulation sites may be useful for depression. In this context, considering efforts of an earlier era is worthwhile, with a view to integrating them with evolving anatomical models of pathophysiology.

In the early 1950s Sem-Jacobsen began recording effects of acute and chronic (several days) stimulation in 220 movement-disordered patients over more than two decades ( ). Most patients subsequently underwent lesion procedures for Parkinson’s disease, but some were studied before ablative surgery. Stimulation of sites throughout the frontal lobes induced affective/mood changes, with apparent selectivity noted for stimulation of ventromedial brain areas. Positive effects ranging from mild relaxation and feelings of tranquility (most common) to marked euphoria were observed twice as often as negative mood effects. The latter ranged from mild tension and/or sadness (most common) to more pronounced sadness and overt sobbing necessitating stimulation cessation.

The same responses were elicited by unilateral stimulation on the right (at 327 sites) or left (316 sites), with no significant laterality differences ( ), suggesting stimulation of many different brain loci could induce positive and negative mood states. Further, effects of opposite affective valence (e.g., mild tension and sadness versus mild euphoria) were sometimes seen with stimulation of sites 5–10 mm apart in the same individual.

Modern DBS for movement-disordered patients has at times had dramatic effects on the affective state of patients. Case reports have described effects ranging from induction of depressive dysphoria, anhedonia, apathy, and blunted affect to hypomania, merriment, and involuntary laughter. These findings are extremely intriguing, especially given the possibility of mood effects when the subthalamic nucleus (STN) is stimulated to treat OCD. Case reports of DBS of the STN in two patients with severe Parkinson’s disease who also had moderately severe OCD produced improvement in OCD symptoms by 2 weeks after the start of therapy. In one of the two patients, OCD improvement was seen despite little change in Parkinson symptoms. A controlled trial of STN stimulation for OCD itself by a collaborative group in France was published in 2008. Yale–Brown OCD Scale (YBOCS) scores decreased from 30 to 19 after 3 months of active stimulation in eight patients who received active DBS first. In contrast, in the sham group YBOCS severity declined from 31 at baseline to 26 after 3 months of sham stimulation. The YBOCS score was 24 at the end of 3 months of the subsequent active DBS period. There were 15 serious adverse effects, including hemorrhage and infection ( ).

DBS for Obsessive–Compulsive Disorder

Work using DBS for OCD, the first contemporary report of DBS for psychiatric illness, is described more fully in Chapter XX of the present work. The rationale for development of DBS for OCD in large part paralleled that for tremor, Parkinson’s disease, and dystonia, where DBS was applied to structures where lesions had therapeutic effects. Case studies of severely ill, highly treatment-refractory OCD patients treated with DBS of the anterior limb of the internal capsule and/or the adjacent striatum were published beginning in 1999 ( ). These reports support the therapeutic potential of DBS in this population, and suggest that DBS is generally well tolerated ( ).

For any surgery for psychiatric illness, a key issue is long-term outcome, as is true in established uses of DBS in movement disorders. Treatment decisions need to be based on the probability that therapeutic effects will be durable, while taking into account burdens imposed by potential adverse effects. A related issue is the need to determine the likely rate at which therapeutic effects will develop in multiple domains. This is in part necessary to give patients and family members a realistic idea of the potential unfolding of benefits when they occur. Based on our own experience and that of others with lesion procedures for OCD ( ), even cases with ultimately positive outcomes take time to improve. Beneficial changes in symptom severity, functioning, and quality of life may develop gradually (and at different rates) in individuals who have had chronic and severely impairing illnesses that have disrupted not only the patients’ functional capacities but also their family and social relationships. A related point is that a description of therapeutic outcomes that will be most meaningful to patients and families needs to go beyond symptom severity reductions and take into account functioning and quality of life.

In 2006 our research group reported on 10 OCD patients meeting stringent criteria for severity and treatment resistance who underwent DBS of a ventral internal capsule/ventral striatum (VC/VS) target ( ). This work followed and was based upon pioneering work by Nuttin et al. which began in 1998, which was itself influenced by earlier results of anterior capsulotomy for OCD. The OCD patients, who met rigorous criteria for diagnosis and failure to respond to multiple adequate conventional treatments, had quadripolar stimulating leads implanted bilaterally in the VC/VS. DBS was activated openly 3 weeks later. Mean YBOCS scores decreased significantly from baseline to 36 months ( P < .001). Four patients had at least 35% threshold decrease in YBOCS severity at 36 months, and scores declined between 25% and 35% for two others, consistent with the categorical response definition commonly used in modern treatment trials for OCD. Mood and nonOCD anxiety symptoms improved in these patients, and there were improvements in self-care, independent living, and work, school, and social functioning. Surgical adverse effects included asymptomatic hemorrhage (n = 1), intraoperative seizure (n = 1), and superficial infection (n = 1). Psychiatric adverse effects included transient mood elevation, which met diagnostic criteria for a hypomanic episode, in 1 of the 10 patients.

Long-term effects observed by our research group during open-label VC/VS DBS include worsened depression followed by a more gradual exacerbation of OCD symptoms at the point when DBS is interrupted by stimulator battery depletion. These observations are in accord with a hypothesis of overlapping neurocircuitry mediating at least some dimensions of depression and OCD. Another interesting observation from this OCD patient series is that two patients had sufficient improvement with VC/VS DBS to be able to engage in adjunct cognitive behavioral therapy. Later we found a similar overall picture of benefits and adverse effect burden in an expanded series including these 10 individuals and 16 others ( ). In this combined series it was easier to discern a “learning curve,” in which patients in the second or third cohorts implanted did better over the long term than those enrolled when experience was more limited.

DBS of a target closely related to the VC/VS, the nucleus accumbens (NAcc), also appears promising ( ). Sixteen patients had open-label DBS first. Exposure-based behavior therapy and pharmacotherapy were both ongoing (the case for almost all studies). Then there was a double-blind randomized cross-over to 2-week periods of active or sham DBS, followed by open DBS. In the initial open phase, 56% of patients responded. In the sham-controlled crossover (n = 14), symptoms were less intense during active versus sham DBS. The most prominent stimulation-related adverse effect was elevated mood/hypomania in half the patients, always judged nonserious. Other side-effects were a surgical wound infection (n = 1), forgetfulness (n = 5), and word-finding difficulties (n = 3). Objective testing (in other DBS studies in psychiatric illness) found no cognitive deterioration or improvement.

A controlled cross-over study tested DBS at a different target, the STN ( ), which is also within the general cortical–basal ganglia–thalamic circuitry implicated in OCD. Of 16 patients, half were randomized to active DBS and half to sham stimulation, after which each group crossed over to the other condition. There was a difference in favor of active treatment. Adverse events included intracerebral hemorrhage on device insertion (n = 1) leading to a permanent hand motor deficit, and infection leading to pulse generator removal (n = 2). One patient had hypomania and one had mania related to DBS, both reversible. Other putative DBS targets for OCD include the inferior thalamic peduncle and ventral caudate nucleus, each in small studies.