Objective:  To review the extant literature on the long‐term outcome of child/adolescent‐onset obsessive–compulsive disorder (OCD).

Method:  Medline and Psychlit databases were systematically searched for articles regarding long‐term outcomes of child/adolescent‐onset OCD. Meta‐analysis regression was applied to evaluate predictors and persistence of OCD.

Results:  Sixteen study samples (n = 6–132; total = 521 participants) in 22 studies had follow‐up periods ranging between 1 and 15.6 years. Pooled mean persistence rates were 41% for full OCD and 60% for full or subthreshold OCD. Earlier age of OCD onset (z = −3.26, P = 0.001), increased OCD duration (z = 2.22, P = 0.027) and in‐patient vs. out‐patient status (z = 2.94, P = 0.003) predicted greater persistence. Comorbid psychiatric illness and poor initial treatment response were poor prognostic factors. Although psychosocial function was frequently compromised, most studies lacked comprehensive outcome measures.

Conclusion:  Long‐term persistence of pediatric OCD may be lower than believed. Future studies should include broader measures of outcome including symptomatic persistence and functional impairment in multiple domains.