Duration of untreated illness (DUI) is a predictor of outcome in psychotic and affective disorders. Data available on DUI and its relationship with outcome in obsessive-compulsive disorder (OCD) suggest an association between longer DUI and poorer treatment response. The present study investigated socio-demographic and clinical predictors of DUI and its association with long-term outcome in OCD patients. Eighty-three OCD outpatients were treated with serotonin reuptake inhibitors (SRIs) and prospectively followed-up for 3 years. Baseline information was collected on demographic and clinical characteristics using standard assessments. Each patient was assessed through the structured clinical interview for DSM-IV axis I disorders (SCID-I), the structured clinical interview for DSM-IV axis II personality disorders (SCID-II), the Yale-Brown obsessive–compulsive scale (Y-BOCS) and the 17-item Hamilton rating scale for depression (HDRS). The DUI was explored by interviewing patients, family caregivers and clinicians. OCD subjects had a mean DUI of 7.3 (5.8) years. A younger age at onset and a greater severity of OCD symptoms at baseline were associated with a longer DUI. The DUI of patients with a “good outcome” was shorter than that of patients with a “poor outcome”. Logistic regression analysis revealed indeed a possible association between longer DUI and “poor outcome”. In the logistic multivariable model, the association of DUI with treatment outcome held true whilst controlling for socio-demographic and clinical variables.