Anxiety and autistic traits in adults: a systematic review and meta-analysis

Abstract

Background: Autistic traits are distributed along a continuum, and some individuals exhibit subclinical characteristics without meeting diagnostic criteria for autism spectrum disorder (ASD). This population, referred to as the broader autism phenotype (BAP), has been associated with increased prevalence of anxiety symptoms. Given that these individuals often do not access clinical services or receive interventions, understanding how anxiety manifests within this group is essential for improving psychological well-being and quality of life. Although research on autism and psychopathology has expanded in recent years, few studies have explored this relationship in adults with BAP from a dimensional and transdiagnostic perspective. Objective: This study aimed to systematically review and synthesize recent empirical evidence on the relationship between autistic traits and anxiety symptoms in adults, and to assess whether this association is statistically significant. Methods: A systematic search was conducted across four databases (PubMed, Web of Science, Scopus, Dialnet) for peer-reviewed articles published between 2013 and 2023. Studies were included if they used validated instruments to assess autistic traits (e.g., AQ, ADOS-2) and anxiety (e.g., HADS, STAI, GAD-7, BAI). A total of 18 independent samples from 13 studies were included. Effect sizes (Hedges’ g) were calculated and synthesized using a random-effects model. Heterogeneity and publication bias were also examined. Results: Findings were mixed: 55% of the included studies reported positive effect sizes and 45% negative. However, the overall effect size was not statistically significant (g = 0.0234, SE = 0.235, 95% CI: −0.438 to 0.483, p = 0.921), with substantial heterogeneity across studies (I2 = 99.83%). Larger studies tended to report positive associations, while smaller studies yielded negative or inconsistent effects. Inconsistencies in measurement tools, particularly across AQ versions, contributed to this variability. Conclusion: Although a significant association was not confirmed, the high heterogeneity highlights the need for more standardized approaches to evaluating autistic traits in non-clinical adult populations. These findings underscore the complexity of subclinical autism and support the relevance of transdiagnostic research frameworks to better understand its relationship with anxiety.