Objective: To longitudinally evaluate the role of depression in the development of impulse control disorders (ICDs) in Parkinson disease (PD) patients. Methods: Using data from the Parkinson's Progression Markers Initiative, we included PD patients without ICDs at baseline according to the Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease (QUIP). Patients were prospectively evaluated first quarterly and then biannually. Development of an ICD was defined as an increase in QUIP scores during follow-up. Using survival proportional hazard models, we studied the effect of baseline depression on ICD risk. We also evaluated this effect controlling for dopamine agonist use as a time-dependent variable and for other potential confounders. Results: Among 354 patients, 68 were depressed at baseline. The median follow-up was 4.08 years. Depression at baseline was associated with higher ICD risk (hazard ratio [HR] = 1.96, 95% confidence interval [CI] = 1.32-2.9, p < 0.001). This risk remained significant after controlling for dopamine agonist use (HR = 1.97, 95% CI = 1.33-2.9, p < 0.001), which was also independently linked to ICD development (HR = 1.87, 95% CI = 1.3-2.7, p < 0.001). Therefore, depressed patients faced an even higher ICD risk when receiving dopamine agonists. Controlling for multiple potential confounders did not alter these results. Interpretation: Depression predisposes to the development of ICDs in PD. This risk is magnified by dopamine agonists. Dopamine agonists should thus be used cautiously in depressed PD patients. ANN NEUROL 2019;86:762-769 View this article online at wileyonlinelibrary.com.