Funnel plot, Egger's test, and Begg's test were used to examine publication bias. Either of the two tests was significant for schizophrenia/any psychotic disorder, OCPD and narcissistic personality disorder. Funnel plots showed asymmetry for dysthymia, agoraphobia, PTSD, any eating disorder, and schizoid personality disorder. The pooled prevalence rates for these comorbidities may therefore have been affected by publication bias. Nearly 50% of the 91 included studies in the meta-analysis reported rates for MDD, however, personality disorders, and psychotic disorders were reported in <20% of the studies, suggesting that these comorbidities are left out in the evaluation of comorbidities in clinical studies, perhaps due to their comparatively lower prevalence.
