Predicting Response to vALIC Deep Brain Stimulation for Refractory Obsessive-Compulsive Disorder

Background: Deep brain stimulation (DBS) for treatment-refractory obsessive-compulsive disorder (OCD) is effective in half of patients, but also is invasive and labor-intensive. Objective: Selecting probable responders beforehand would more optimally allocate treatment resources and prevent patients' disappointment. Some centers use clinical and demographic predictors to exclude patients from DBS treatment, but the evidence base remains uncertain. Methods: This observational cohort study examined the association of baseline demographic and disease characteristics with a 1-year prospective course of OCD and depressive symptoms in a cohort of 70 consecutive patients who received DBS of the ventral anterior limb of the internal capsule (vALIC-DBS) for OCD according to DSMIV or DSM-5 criteria between April 2005 and October 2017. Baseline characteristics and symptom decrease were analyzed using Fisher exact tests and binary logistic regression to examine whether they could predict individual response (> 35% reduction in Yale-Brown Obsessive Compulsive Scale score and 50% reduction in Hamilton Depression Rating Scale score, respectively). Results: Insight into illness was the only significant predictor of individual response, with a positive predictive value of 84.4%, while the negative predictive value was 44.0% (b = 0.247, χ21 = 5.259, P = .022). Late-onset OCD was associated with more symptom decrease (β = -0.29; 95% CI, -0.53 to -0.04; P = .023) and comorbid personality disorder with less symptom decrease over time (β = 0.88; 95% CI -0.29 to 1.47; P = .004), but they could not significantly predict vALIC-DBS response. A later age at onset, comorbid personality disorder, and insight into illness were associated with clinical outcomes after vALIC-DBS, but predictive values were not large enough to facilitate clinical patient selection. Conclusions: Clinical and demographic factors cannot yet predict outcome and should not be used to exclude patients from treatment with vALIC-DBS. These first individual prediction analyses for vALIC-DBS response in OCD are important, given that some centers up until now still exclude patients based on clinical characteristics such as comorbid personality disorders.