Long-term outcome of obsessive-compulsive disorder in adults: a meta-analysis

To study the long-term rate and predictors of remission in adults with obsessive-compulsive disorder (OCD), using meta-analysis.

The MEDLINE database was searched to May 2013 using the search terms obsessive-compulsive disorder, prospective, outcome study, clinical course, remission, prognosis, follow-up, and long-term and limits for language (English), species (humans), and age (adults). This was supplemented by manual bibliographic cross-referencing.

English-language studies from peer-reviewed journals on adults with DSM-III-R, DSM-IV, DSM-IV-TR, ICD-9, or ICD-10 diagnosis of OCD followed up for ≥ 1 year and treated with serotonin reuptake inhibitors and/or cognitive-behavioral therapy that reported rate of remission (Yale-Brown Obsessive Compulsive Scale [YBOCS] score < 16 at longest follow-up) were included.

Data were gathered as numbers/means/percentages/categories on sample size, study design, follow-up duration, age at assessment, illness duration, age at illness onset, gender, marital status, inpatient/outpatient status, family history, baseline YBOCS score, comorbidities, and remission.

Seventeen studies (pooled N = 1,265) fit the selection criteria and were used for the meta-analysis. The pooled sample had a mean follow-up duration 4.91 years and was predominantly male and outpatient and had onset of illness in the second decade, illness duration more than 10 years, and moderate-to-severe OCD. Pooled remission rate was 53% (95% CI, 42%-65%). Prospective studies showed higher pooled remission rate than retrospective studies (55% [95% CI, 45%-65%] vs 50% [95% CI, 27%-73%], P < .001). Indian studies showed higher pooled remission rate than others (71% [95% CI, 59%-83%] vs 48% [95% CI, 37%-59%], P < .001). Age at onset (t = -7.08, P = .019), illness duration (t = -8.13, P = .015), baseline YBOCS score (t = -6.81,P = .021), and male gender (t = -5.92, P = .027) had significant negative association with remission on meta-regression.

A high long-term remission rate found in this meta-analysis is contrary to generally held beliefs about poor outcome of individuals with OCD. Multicenter, prospective, long-term studies should systematically examine course and outcome in larger samples, emphasizing symptomatic and functional recovery.