Objective: To evaluate efficacy, effect on mood, and safety of deep brain stimulation (DBS) for obsessive-compulsive disorder (OCD) at different target sites. Data Sources: Electronic records from databases MEDLINE, EMBASE, and CENTRAL up to November 2019 were searched. Search terms included OCD, depression, and DBS. Study Selection: Eight randomized controlled trials (RCTs) (n = 85) and 38 observational studies (case reports and case series) (n = 225) were included. Data Extraction: In RCTs, the differences in outcomes between sham and active stimulation for OCD and depression were evaluated and the proportion of responders was determined. In all included studies, at last follow-up, the improvement from baseline in OCD (Yale-Brown Obsessive Compulsive Scale [Y-BOCS score]) and a scale of weighted depression scores (WDS) were determined. Predictors of response (age, illness duration and severity, frequency parameters, and response in depression) were evaluated. The proportions of adverse events and dropouts were calculated. Results: In RCTs, mean differences between sham and active stimulation in Y-BOCS and Hamilton Depression Rating Scale (HDRS) scores were-7.8 (95% CI =-11.2 to-4.3, I2 = 40%, P =-0001) and-7.3 (95% CI =-11.5 to-3.0, I2 = 0%, P =-0009), respectively. No differences between limbic and non-limbic targets were identified (÷2 = 0.21, I2 = 0%, P = .0006). At last follow-up, improvements in Y-BOCS and WDS were-15.0 (95% CI =-18.3 to-11.7, I2 = 90%, P < .001) and-13.7 (95% CI =-20.1 to-7.3, I2 = 76%, P < .001), respectively. No consistent predictors of response were found. There were 0.68 adverse events (95% CI = 0.59 to 0.78, I2 = 88%), 0.32 serious adverse events (95% CI = 0.12 to 0.62, I2 = 96%), and 0.13 dropouts (95% CI = 0.07 to 0.16, I2 = 16%) per treated patient. Conclusions: DBS can significantly decrease Y-BOCS score and depressive symptoms in refractory OCD.