• More than half of adults with ADHD met criteria for at least one personality disorder • Borderline, avoidant, depressive, and antisocial PDs showed the highest prevalence • Prevalence estimates varied substantially by diagnostic instrument and study setting • Structured interviews and self-report tools yielded higher PD rates than clinical diagnosis • Personality disorders were overrepresented in ADHD compared with community samples This meta-analysis investigated the prevalence and moderators of personality disorders (PDs) in adults with Attention-Deficit/Hyperactivity Disorder (ADHD), synthesizing data from 11 studies encompassing 2,120 participants across clinical and nonclinical settings. The results indicate a substantial but highly variable burden of personality pathology, with a pooled prevalence for the presence of at least one comorbid PD estimated at 57% (95% CI 42–71). However, this figure was largely driven by specialized outpatient populations, and statistical heterogeneity was exceptionally high (I² > 90%), suggesting that prevalence is deeply context-dependent. Among specific diagnoses, the highest co-occurrence rates were observed for Passive-Aggressive (25.3%), Avoidant (23.1%), and Borderline (21.9%) personality disorders. Moderator analyses revealed that these estimates were significantly influenced by diagnostic methodology; structured clinical interviews (such as the SCID-II) and self-report inventories (like the MCMI) consistently yielded higher rates than standard clinical assessments. Furthermore, setting played a critical role, with ADHD-specific clinics showing significantly higher coexistence than prison or student samples. While tests for publication bias were non-significant, they were limited by low statistical power. Ultimately, these findings underscore the high frequency of personality-related impairment in adult ADHD while highlighting the challenges of "diagnostic noise" and phenotypic overlap. The extreme fluctuations in reported rates suggest that clinicians must look beyond categorical labels, acknowledging that the identified prevalence is often a byproduct of the specific instruments and clinical frameworks employed.
Adamis et al. (Sun,) studied this question.