Deep Brain Stimulation for Intractable Obsessive-Compulsive Disorder: Progress and Opportunities

In this issue of the Journal, Denys et al. (1) describe the response of 70 patients with severe and intractable obsessive-compulsive disorder (OCD) treated with bilateral ventral anterior limb of the internal capsule (vALIC) deep brain stimulation (DBS). This is the largest cohort study of DBS for OCD ever reported and includes detailed clinical outcomes and safety data that led the authors to conclude that vALIC is generally effective and safe for patients with severe and chronic OCD whose symptoms were nonresponsive to a wide range of medications as well as exposure and response prevention therapy.

At the 12-month follow-up, 52% of patients were categorized as “responders,” and 17% were categorized as “partial responders,” as determined by ≥35% and 25%−34% decreases, respectively, in scores on the Yale-Brown Obsessive Compulsive Scale (Y-BOCS) (2). The 35% metric is used as a more stringent criterion for DBS, whose invasiveness warrants a higher bar, but a 25% decrease is also considered to be an acceptable measure of response in clinical trials (3). Combining these two categories, 69% of patients showed meaningful clinical improvement. On the other hand, that leaves 31% who experienced no relief from their obsessive-compulsive symptoms.

This study adds to a growing body of evidence that DBS is generally safe and well tolerated for patients with intractable OCD (4, 5). Adverse events were categorized as either related to surgery, the hardware, or programming. Among the 11 serious adverse events, two patients developed postsurgical infections involving components of the device that required explantation and reimplantation several months later. Six patients required revision surgery to correct malposition of the electrodes. Importantly, no intracerebral hemorrhages or seizures were reported. There were three suicide attempts without sequelae, and only one was classified as related to stimulation changes in a patient who was reported to be disappointed with her response to DBS.

Transient hypomania occurred in 39% of patients, along with agitation (30%) and impulsivity (19%). It is unclear whether the authors’ use of the term hypomania, which connotes a clinically significant mood disorder lasting days, is synonymous with a mirth response, the immediate induction of a smile/laughter and euphoria during DBS programming. Some studies maintain that a mirth response during initial programming may be a positive predictor of eventual OCD outcome (6, 7). In the Denys et al. study, implantation of DBS leads was performed under general anesthesia, rendering behavioral testing impossible. In our hands, we use intraoperative behavioral effects (i.e., mirth response and absence of anxiety) to confirm electrode placement, and, if necessary, we will adjust the position of the lead until a mirth response is obtained. We have used this approach in our last seven patients and have elicited a mirth response from at least one hemisphere (one lead) in each case. To date, six of these seven patients are responders (unpublished data). Together, these published (and unpublished) data suggest that the presence of a mirth response—whether intraoperatively or in the first programming session—may be a necessary (but not sufficient) condition tied to OCD response. Further research is needed to confirm this observation and to determine whether the mirth response is truly a signature of target engagement or simply a convenient clinical marker to guide programming.

Currently, programming adjustments for DBS for OCD are made largely on the basis of acute beneficial effects on “mood,” “energy,” and “anxiety” as described by the patient and evaluated by the clinician. In contrast to DBS for tremor, in which immediate symptomatic benefits are observable, direct effects on the core obsessive-compulsive symptoms are not discernable during a programming session. Instead, parameters are adjusted, in a largely trial-and-error fashion, based on changes in OCD symptom severity since the last visit. We are exploring the use of Automated Facial Affect Recognition (AFAR) (8), a computer-vision machine-learning based approach that objectively measures real-time changes in anatomically based facial actions and positive (and negative) valence of emotional state to aid in programming.

As stated in this article and elsewhere, sustained positive effects on mood and anxiety invariably precede improvement in OCD during DBS targeting the vALIC (or related neighboring regions such as the ventral striatum, nucleus accumbens, or bed nucleus of the stria terminalis) (9). On the other hand, improvement in mood does not guarantee a successful outcome for OCD. Patient-rated measures of increased energy and motivation accompany the positive mood effects of ventral striatum DBS for OCD (9). It is noteworthy that no patient in the Denys et al. cohort requested explantation, including those who were OCD nonresponders. The authors attribute this preference to the beneficial effects of DBS on mood and anxiety.

The prevailing neurocircuit-based framework for the pathophysiology of OCD points to dysfunction in specific cortico-striato-thalamo-cortical loops (10). The first study of DBS for OCD (11) targeted the ALIC based on prior positive experience with stereotactic ablation in this region and supported by the rationale that fibers coursing through the ALIC, forming part of the relevant cortico-striato-thalamo-cortical circuit, would be interrupted, along with the symptoms of OCD. The original assumption that high-frequency DBS (e.g., 130 Hz) would act as “functional ablation” has been challenged by emerging basic neuroscience research showing that the therapeutic mechanisms of DBS are far more complex (12). Several research groups have proposed that DBS exerts neuromodulatory effects, both locally and distally throughout the cortico-striato-thalamo-cortical circuit (10, 13), that normalize hyperconnected network activity in OCD (14, 15).

An alternate but complementary hypothesis is that ventral striatum DBS for OCD affects the balance between positive and negative valence systems in the brain such that circuit bias is changed in the direction of reward/approach over harm avoidance (9). Figee et al. (16) highlighted the importance of disordered reward processing in OCD based on several lines of evidence, including a functional MRI study showing diminished nucleus accumbens activity during a reward anticipation task in patients with OCD compared with healthy control subjects (16). A study using single-photon emission computed tomography found that nucleus accumbens DBS induced striatal dopamine release in patients with OCD (17). Using AFAR, we have been able to demonstrate that ventral striatum DBS can turn up the gain on positive valence affect (8). For the most part, patients with OCD are driven by harm avoidance (18, 19), and their compulsions are aimed at preventing negative outcomes and are not inherently pleasurable, thus differentiating compulsions from behavioral addictions. During ventral striatum DBS, patients with OCD become more engaged in rewarding activities, but without close monitoring and careful adjustments, there can be overshoot into excessive reward-seeking behaviors as manifestations of a hypomanic state (20). Better management of this behavioral side effect of ventral striatum DBS is one of the aims of a current project funded by the National Institutes of Health (NIH) to develop adaptive DBS for OCD (clinical trials identifier, NCT03457675) (21). In this NIH study, local field potentials are being recorded chronically from the ventral striatum together with other neural and behavioral data in an effort to identify classifiers of hypomania and exacerbations of OCD.

The main limitation of the Denys et al. study is that the majority of patients received DBS in an open-label fashion. The first 16 patients were enrolled in a double-blind sham-controlled crossover study that showed active DBS as superior to sham DBS (22); the remaining 54 patients were treated openly. Concerns have emerged about higher than expected sham response rates in double-blind trials of DBS for treatment-resistant depression (23), including for the ventral striatum target (24). In a recent meta-analysis (N=24), Schruers et al. (25) found evidence for nonstimulation effects of DBS on symptoms of OCD. However, Denys et al. point out that the magnitude of the sham effects identified by Schruers el al. (25) is smaller than the mean Y-BOCS reduction shown in their study. To address sham response—due to either insertion effects or expectation bias—our preferred study design for DBS in psychiatric disorders is to conduct a blinded, discontinuation phase at the conclusion of open-label DBS. This approach allows for individualized programming to optimize outcome, establish a stable continuation period, and then identify sham responders by gradually withdrawing stimulation over a period of several weeks.

An important takeaway from the Denys et al. study is the clinical value of providing exposure and response prevention therapy during DBS (26). Importantly, all patients had previous failed trials of exposure and response prevention before becoming eligible for DBS. One of the confounds of the study is that it is hard to disentangle the effects of DBS alone from that of combined DBS and exposure and response prevention therapy. The majority of patients (N=57) received individualized exposure and response prevention sessions from a skilled therapist starting at different time points after the activation of DBS. The first 16 patients received weekly sessions for 24 weeks during open-label DBS (26). Exposure and response prevention therapy appeared to augment the effects of DBS on OCD beyond benefits already achieved through DBS; patients attained an average point reduction of 7.3 (SD=11.3) on the Y-BOCS after exposure and response prevention (in addition to the 8.3 [SD=7.8] Y-BOCS point reduction post-DBS). However, the absence of a control group for exposure and response prevention does not allow for definitive conclusions (26). During a subsequent double-blind discontinuation phase of this cohort (N=16), all responders rapidly and completely relapsed once DBS was turned off, despite having been treated with exposure and response prevention (27). The authors suggest that the gains seen with exposure and response prevention therapy were dependent on the presence of active DBS (27). Preclinical studies show that nucleus accumbens DBS enhances fear extinction (28), which may help explain how it augments exposure and response prevention in humans (26).

Overall, the response rates reported by Denys et al. are in line with recent meta-analytic findings (4) showing that DBS is a highly effective intervention for more than 50% of patients with severe, chronic, and treatment-resistant OCD. The current findings from the Netherlands are germane to clinicians and patients in the United States, where an overlapping brain region, the ventral striatum, is approved by the Food and Drug Administration (under a Humanitarian Device Exemption) as a DBS target in the treatment of refractory OCD. While ventral striatum DBS is an important option for intractable OCD, there is room for improvement in outcome rates, magnitude of response, and mitigation of DBS-induced side effects, particularly hypomania (6).

There are several opportunities for improving outcomes of ventral striatum DBS for OCD. One such approach is individualization of DBS lead placement using measures of structural connectivity within the brain, usually with diffusion tensor imaging (DTI) (10). Proponents argue that because the connectivity profile of each patient’s brain is unique, the target should be defined in a personalized manner, taking into account individual variability. Although Denys et al. did not use DTI tractography in this series, they have separately reported their retrospective experience, using it to help identify optimal lead location (29). There is still controversy in this arena, however, as other investigators have arrived at sometimes conflicting conclusions (30). Studies that test these methods using prospective targeting are needed.

Unfortunately, there are no established clinical predictors informing us which patients are most likely to respond to DBS (5). We rely mostly on illness severity, chronicity, and well-documented evidence of treatment resistance to determine eligibility for DBS in adults with a primary diagnosis of OCD. This assessment is coupled with a robust informed consent process that covers risks, potential benefits, and alternatives. The creation of a worldwide database for OCD, like the one developed for DBS in Tourette’s syndrome, would be an important step toward defining predictors of response. Finally, we need to leverage state-of-the-art neurotechnologies (like chronic sensing of local field potentials) (21) to learn more about the neurocircuitry of OCD to discover biomarkers of response and to identify nodes (besides those already studied) that may be targeted to directly modulate the symptoms of OCD.