Deep brain stimulation for treatment-refractory obsessive-compulsive disorder should be an accepted therapy in Australia

Mosley and colleagues (2021) recently published an influential report arguing for deep brain stimulation (DBS) to be an accepted therapy for treatment refractory-obsessive-compulsive disorder (TR-OCD). We support this argument based on additional, complimentary evidence within our systematic review (Acevedo et al., 2021) and more recent publications.

Mosley et al. propose that four randomized controlled trials (RCTs) and four open-label (OL) trials, from differing sites, have demonstrated sufficient evidence in 300 cases worldwide for DBS to be an accepted therapeutic option for TR-OCD. Worldwide, 369 cases are identified (outcomes of 33 are unpublished), across 6 RCTs, 2 randomized non-controlled trials, 15 OL follow-up reports (up to 9 years) and several case studies (see Figure 1). In addition to the four RCTs discussed by Mosley with optimized programming, another RCT demonstrated efficacious outcomes (Barcia et al., 2019), with no programming optimization.

Figure 1.

The number of reported OCD DBS cases per year.

Acevedo et al. systematically identified that the ventral capsule/ventral striatum (VC/VS) had the greatest supporting evidence for an optimized DBS target for OCD, with response rates of 62–83% from chronic therapy. Yet, more recent reports have identified the adjacent white matter region, the ventral anterior limb of the internal capsule (ALIC) as a more appropriate classification (i.e., Denys et al., 2020; Guzick et al., 2020). Thus, it is critical to consider anatomical precision in DBS targeting, programming and reporting. Particularly for striatal regions in which stimulation of one region may modulate others, different regions may be concurrently targeted within an individual (through the same electrode trajectory) or a cohort (through shifts in targeting). Although an optimal striatal location has been identified, a one size fits all approach cannot be assumed, and the precise refinement of the location will depend on individual pathological circuitry. Hence, for the purpose of assessing efficacy, we argue that grouping of striatal regions is necessary.

The World Society for Stereotactic and Functional Neurosurgery (WSSFN) defines an established therapy as

at least two blinded (if feasible) randomized controlled trials from two different groups of researchers need to be published, both showing an acceptable risk-benefit ratio, at least comparable with other existing therapies, the clinical trials should be on the same brain area for the same psychiatric indication. (Nuttin et al., 2014)

This criterion has been met by the RCTs of Denys et al. (2010) and Luyten et al. (2016) that demonstrated ⩾25% difference between active and sham DBS outcomes of the ventral ALIC and bed nucleus of the stria terminalis (BNST).

However, we argue that sham conditions are not feasible or ethical in such scenarios involving extremely vulnerable patients that have been offered DBS as a ‘last resort’, also long-term outcomes are more ecologically relevant. Although double-blinded phases have been achieved and were necessary in early investigations, sham conditions are often ended early due to worsening of symptoms (Denys et al., 2020; Luyten et al., 2016), or the anxiety associated with stimulation being turned off is too drastic to ethically discontinue therapy (unpublished findings from St Vincent’s Hospital, Melbourne). Also, it is established that optimization of programming generally takes 6–12 months in OCD (Guzick et al., 2020). The lack of closed-label investigations in recent years is supportive of this stance. In consideration of this ethical debate, there is substantial evidence to support the WSSFN criteria has been met; with chronic DBS therapy of striatal regions (excluding the inferior region, the nucleus accumbens), achieving response in 44–83% (median 67%), supported by 10 centers.

These benefits are remarkable considering TR-OCD DBS patients are resistant to all available therapies, and response to pharmacology is 40–60%. Furthermore, our systematic review (of 181 patients) did not identify severe adverse events directly associated with DBS, although seven individuals attempted suicide, and there was one suicide, the relationship to DBS is not clear. Adverse events of DBS include transient psychiatric symptoms (hypomania, anxiety, deterioration of mood and suicidal thoughts) which were resolved through programming adjustments. Indeed, one could argue for a greater symptom profile without DBS.

Nevertheless, treatment protocols need to be refined and standardized, particularly for the validation of pre-operative predictors and scenarios of non-responders. Development of an international registry would allow machine learning analysis of clinical features and demographics to develop pre-operative predictors of response. For example, psychiatric comorbidities may resolve from DBS of motor regions (for the treatment of Tourette’s syndrome), yet DBS of affective regions (for the treatment of OCD) appear to lead to continued symptoms including comorbidities, undesirable side effects or non-response. Such registries have achieved impact in DBS management of Tourette’s syndrome (Deeb et al., 2016). Currently, the most validated predictor of clinical response is white matter connectivity. Connectomics should be implemented pre-operatively to target optimal fibers (rather than regions), and post-operatively to alter the stimulation, or shift the electrode in non-responders based on individual connectivity pathways. We have demonstrated sufficient evidence for DBS to be an accepted therapy for TR-OCD; further optimization is achievable and would clearly be a significant bonus for TR-OCD, a population with substantial unmet need.