Browse the corpus

Fragile X syndrome (FXS) arises from an expansion exceeding 200 CGG trinucleotide repeats in the Fragile X Messenger Ribonucleoprotein 1 ( FMR1 ) gene located on the X chromosome. This expansion results in epigenetic silencing of FMR1 , thereby preventing the production of Fragile X Messenger Ribonucleoprotein (FMRP) – a protein essential for brain development and synaptic plasticity (–). Clinically, FXS affects approximately 1 in 2,500 to 7,000 males and 1 in 2,500 to 11,000 females (, ). Its neurobehavioral phenotype is characterized by intellectual disability (ID) (), co-occurring autism (), and autism-like traits (). Psychiatric difficulties are also prevalent, including anxiety (), attention-deficit/hyperactivity disorder (ADHD) (), and self-injurious and aggressive behaviors (, ).

Despite the seemingly well-defined clinical picture, much of the available evidence has been derived predominantly from male samples. This has produced a male-centric understanding of FXS and corresponding clinical guidance that may not fully reflect the female phenotype. Consequently, the more variable and less predictable presentation in females remains underexplored and underserved. To address this critical gap, the current systematic review and meta-analysis focus on psychiatric difficulties in females with FXS, examining their associations with intellectual ability and autism. It also provides the first systematic evaluation of sex representation across research on psychiatric difficulties in FXS over recent decades.

The limited understanding of the female FXS phenotype largely stems from the substantial variability observed among affected females. The presence of a second, typically unaffected X chromosome provides partial protection, as some cells continue to express FMR1 and produce FMRP. The degree of this protective effect depends on several interacting biological factors, including CGG repeat size mosaicism (inherent to all females), methylation of the expanded allele, and the activation ratio – the proportion of cells in which the affected X chromosome is active (–). These mechanisms contribute to a phenotype in females that is markedly more heterogeneous and less predictable than in males (). As a result, females with FXS exhibit a wide range of intellectual abilities – from moderate or severe ID to above-average intelligence (). This variability has perpetuated the misconception that FXS is uniformly “milder” in females () and has even fostered the notion of a “fragile X spectrum”, with males at one end and “unaffected” females at the other (). Consequently, research and clinical efforts have historically focused on males, presumed to have greater support needs, while neglecting the complex and variable presentation in females.

Methodological challenges have further perpetuated this bias. Few psychiatric assessment tools are validated across the full range of intellectual abilities seen in FXS, limiting accurate characterization of female presentations. Instruments such as the Anxiety, Depression, and Mood Scale (ADAMS; ) and the Aberrant Behavior Checklist (ABC; ) were developed for individuals with ID, whereas more mainstream measures like the Child Behavior Checklist (CBCL; ) may not retain validity in this population (). Consequently, comparisons between females and males with FXS are often confounded, as group-level analyses may inadvertently contrast individuals with ID against those without. Although overlap exists – some females present with severe ID and some males fall within the borderline range () – concern remains that apparent sex differences are often artifacts of sampling or measurement bias.

Moreover, the greater variability observed among females necessitates larger sample sizes to achieve adequate statistical power. However, given the rarity of FXS, achieving such samples remains a major challenge. Although rarity is an inherent limitation in genetic syndrome research (), it presents particular difficulties when studying females with FXS, who exhibit a lower clinical prevalence than males (, ). This apparent disparity likely reflects diagnostic bias, as females without intellectual disability often go undiagnosed — perpetuating under recognition and their limited representation in research. Consequently, methodological and sampling constraints have produced a research base disproportionately focused on males — valuable, yet incomplete in representing the full phenotypic spectrum.

Importantly, these challenges extend beyond individual studies to highlight a broader systemic issue within research funding. Studies focused on females are likely perceived as less feasible due to methodological constraints and are therefore less likely to secure grant support. Consistent with this pattern, data from the National Institutes of Health RePORTER database (searches for “fragile X” or “ FMR1 ”, accessed December 2025) indicate that research attention to females with FXS has been far more limited than to males. This creates a self-perpetuating cycle in which funding priorities reinforce existing research gaps, continuing to hinder progress in understanding and addressing the distinct needs of females with FXS.

The resulting knowledge gap carries tangible consequences. Females with FXS are frequently underdiagnosed, often identified only after a male relative receives a diagnosis (). Even when diagnosed, genetic counseling has historically centered on male presentations, providing limited information about the distinct challenges faced by females and often relying on the outdated notion that their symptoms are “milder” (). Although more recent clinical guidelines acknowledge psychiatric and emotional difficulties as key concerns for females (), these recommendations remain general and lack specificity regarding symptom onset, prevalence, or developmental trajectory.

This is particularly concerning given that, despite often higher intellectual abilities, many females with FXS do not attain greater functional independence than males. National survey data () indicate that a large proportion of females require assistance in daily living (63%) and continue residing with parents well into adulthood (51%), with affect problems (i.e., anxiety/depression) and social challenges emerging as the strongest predictors of reduced independence. In contrast, for males, limitations in adult independence are more closely linked to deficits in adaptive functioning, reflecting the central role of co-occurring intellectual disability in the male phenotype. This divergence underscores a critical issue: intellectual ability continues to serve as the default proxy for clinical need in FXS. To more accurately represent and support the experiences of females, research must move beyond intellectual functioning as the primary indicator of vulnerability and instead prioritize psychiatric difficulties – particularly anxiety and depression – which appear to be key determinants of independence, well-being, and long-term quality of life.

To date, only one systematic review and meta-analysis has focused exclusively on females with FXS. Marlborough et al. () reported an elevated prevalence of autism among females with FXS (14%) compared to the general population (2%; ), aligning with the recognition of FXS as the most common monogenic cause of autism. However, it is important to emphasize that neurodivergence itself is not inherently pathological. Evidence from broader neurodivergent female populations shows that the most debilitating challenges often stem from co-occurring psychiatric difficulties such as anxiety, depression, and suicidality (–);. These difficulties likely arise from navigating predominantly neurotypical environments and receiving inadequate or poorly tailored support (, ). Although this framework has not yet been explicitly applied to FXS, existing data () suggest similar vulnerabilities among females with FXS. Emerging evidence indicates growing research interest in recent years. However, existing findings remain fragmented and unsynthesized, limiting translation into meaningful clinical guidance for families and practitioners. Together, these gaps underscore the urgency and relevance of the present review.

This review represents the first systematic synthesis of the literature specifically examining psychiatric difficulties in females with FXS. Where data permit, meta-analyses estimate the prevalence of key difficulties, including anxiety, depression, ADHD, aggression, and self-injury. The review further explores associations with intellectual ability and co-occurring autism to identify factors that may increase vulnerability to psychiatric symptoms. Additionally, it evaluates patterns of sex representation within the FXS psychiatric literature over time to highlight the degree of female underrepresentation. By integrating findings across decades of research, this review aims to clarify the developmental course and prevalence of psychiatric symptoms in females with FXS and to strengthen the clinical understanding of their unique psychiatric vulnerabilities – ultimately informing more equitable and effective support.

This review examines the prevalence of both psychiatric conditions and psychiatric symptoms in females with FXS. A psychiatric condition – also referred to as a mental health condition – encompasses patterns of difficulties in cognitive, emotional, behavioral, and/or social functioning that adversely affect well-being and daily life (). Such conditions are frequently underdiagnosed in the general population (), with even higher rates of underdiagnosis reported among individuals with genetic syndromes, including FXS (, ). Symptoms associated with psychiatric conditions may be both experienced by the individual and observed by others familiar with them, even in the absence of a formal diagnosis (). Assessing psychiatric symptoms, therefore, provides valuable insight into challenges experienced during the prodromal or subclinical stages of psychiatric conditions. Although Attention-Deficit/Hyperactivity Disorder (ADHD) is formally classified as a neurodevelopmental disorder, it is included in this review under psychiatric difficulties due to its clinical management within psychiatry and its frequent co-occurrence with other psychiatric conditions. For this reason, the term psychiatric difficulties – rather than mental health difficulties – is used throughout this review to synthesize findings across studies.

Literature searches were conducted in PsycINFO, MEDLINE, Embase, and Web of Science. The following inclusion filters were applied: (a) English language, (b) peer-reviewed publication, (c) publication date between 1980 and 2025, and (d) studies involving human participants. Search strategies incorporated terms relating to psychiatric conditions recognized within standard diagnostic frameworks, including the Diagnostic and Statistical Manual of Mental Disorders (DSM; APA) and the International Classification of Diseases (ICD; WHO). Both subject headings (e.g., Medical Subject Headings, MeSH) and free-text keywords were adapted from a previous review examining psychiatric conditions and symptoms in genetic syndromes (). A deliberately broad search approach was adopted, rather than applying a priori restrictions on which psychiatric difficulties might be relevant to the FXS phenotype, recognizing that much of the existing synthesis has been based primarily on male populations. That said, a key modification from the strategy used by Glasson et al. () was the inclusion of search terms related to ADHD, which were not part of their review. ADHD was explicitly incorporated in the present study to address the absence of prior systematic analyses of its presentation in females with FXS.

Terms related to autism were also included to explore potential associations between psychiatric difficulties and co-occurring autism in females with FXS. However, autism prevalence data were not synthesized independently, as this has been comprehensively reviewed elsewhere (). Instead, autism prevalence was reported descriptively within the sample characteristics to provide contextual information. Similarly, ID-related terms were incorporated to investigate possible associations between intellectual ability and psychiatric outcomes. Due to limited available data, analyses of ID prevalence were beyond the scope of this review; thus, IQ information is also presented descriptively. To ensure comprehensive coverage and allow assessment of sex representation across studies, no search terms were used to restrict results by sex or gender. The complete list of search terms is available in the ().

The initial search yielded 3,533 studies (see ). After the removal of duplicates ( n = 1,808), 1,725 articles remained and were independently screened by two authors (L.J. and C.K.) based on titles and abstracts, resulting in the exclusion of 1,558 studies. The remaining 167 articles underwent independent full-text review by both authors to determine eligibility. Inter-rater reliability demonstrated substantial agreement during the title and abstract screening (κ = 0.70) and almost perfect agreement during the full-text review (κ = 0.87). Any discrepancies between reviewers were discussed and resolved through consensus.

To be included in this review, studies were required to report either the prevalence of psychiatric diagnoses or the presence of psychiatric symptoms in females with FXS. The review encompasses research across childhood, adolescence, and adulthood. While it is ideal to examine these developmental stages separately, the rarity of FXS and the limited number of studies focused specifically on females necessitated the inclusion of research spanning the full lifespan. Studies were excluded if participants carried the FMR1 premutation rather than FXS ( n = 2) or if findings were collapsed across these groups ( n = 1), given that the premutation is associated with a distinct phenotype (). Reports were also excluded if they focused exclusively on males ( n = 58) or included females but did not analyze data by sex ( n = 42) – a likely consequence of the limited resources available to support research that sufficiently captures variability among females.

A risk-of-bias and quality assessment framework was adapted from tools used in previous meta-analyses on autism in genetic syndromes () and on females with FXS (). While the criteria pertaining to sample identification and confirmation of genetic diagnosis were not changed, the assessment tool was modified to better reflect the evaluation of psychiatric difficulties specific to the aims of this review (see ). Each study was rated on these three domains – sample identification, genetic syndrome confirmation, and assessment tool used – with scores ranging from 0 to 3, where higher scores indicate higher study quality and lower risk of bias. An average score was then calculated for each study to provide an overall quality index.

Two authors (C.K. and either R.H. or L.J.) independently evaluated all studies for quality and risk of bias. Inter-rater agreement was excellent across all domains, with weighted kappa values of 0.77 (95% CI: 0.51–1.00) for sample identification, 0.73 (95% CI: 0.49–1.00) for genetic syndrome confirmation, and 0.81 (95% CI: 0.60–1.00) for the assessment tool used. For overall quality scores, a two-way random-effects model using consistency and average-measures intraclass correlation () indicated a high level of agreement between raters (ICC = 0.85, 95% CI: 0.65–0.98). In instances of disagreement, the first author (L.J.) made the final determination of the ratings reported in this paper. Consistent with Marlborough et al. () and prior methodological recommendations (), no studies were excluded on the basis of quality or risk-of-bias scores, given the limited research available on this population. However, study quality was incorporated into the meta-regression analyses to assess whether it accounted for variability in prevalence estimates, as described below.

Extracted data included study details (authors and country) and participant characteristics such as the sample size of females with FXS, age range (including mean and standard deviation where available), race/ethnicity, and socioeconomic status. When reported, information on IQ and co-occurring autism was also recorded. Details regarding exclusion criteria were collected to further characterize study samples and evaluate their representativeness.

Data on the prevalence of psychiatric difficulties were extracted along with information on the methods used to assess psychiatric symptoms and/or diagnoses. Prevalence estimates were derived from multiple sources, including diagnostic classifications based on established frameworks such as the International Classification of Diseases, 11th Revision (ICD-11; ) and the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5; ), caregiver-reported clinical diagnoses, medical records, and standardized symptom screening tools with validated clinical cut-offs (e.g., CBCL; ). When the proportion of participants scoring above clinical thresholds on symptom screening measures was not provided ( n = 12), study authors were contacted, resulting in additional data for two studies (, ). A narrative synthesis was then conducted to summarize key findings on psychiatric difficulties, encompassing symptom profiles, comparative prevalence rates, developmental trajectories, and associations with intellectual ability and autism.

Meta-analyses were conducted in R using the metafor package (v6.9) () to estimate the prevalence of clinically significant psychiatric difficulties in females with FXS, where sufficient data were available. For each study, the total number of females with FXS and the number of cases (i.e., females diagnosed with a psychiatric condition) were extracted to calculate proportions. Because some studies reported data across multiple psychiatric domains, an overall pooled prevalence and between-domain comparisons were not conducted. Instead, separate meta-analyses were performed for each psychiatric domain. The primary analytic approach employed a random-effects inverse-variance meta-analysis of logit-transformed proportions, yielding pooled logit estimates with corresponding confidence intervals. Results are presented alongside study-level data, including sample size, number of cases, assessment tools, and quality ratings (see ). Pooled logit estimates were back-transformed to percentages with 95% confidence intervals (CIs) for ease of interpretation. Between-study variance (τ²) was estimated using the DerSimonian–Laird method, and heterogeneity was quantified using I², with values below 75% interpreted as indicating negligible to moderate heterogeneity ().

To assess potential sources of variability in prevalence estimates, meta-regression analyses were conducted, examining whether study quality contributed to heterogeneity. To account for differences in participant age ranges, studies were coded based on the breadth of their age distribution: those restricted to a single developmental stage (e.g., preschool, school-age, or adult-only) were classified as narrow, while those spanning multiple developmental stages (e.g., preschool through adulthood) were classified as wide. This categorical approach was adopted because several studies reported age in ranges or categorical form, precluding the use of age as a continuous moderator. The age-based moderator was then used in subgroup analyses to evaluate whether more homogeneous versus heterogeneous age samples influenced prevalence estimates.

This section begins with an overview of sex representation in research on psychiatric difficulties in FXS to date, followed by a summary of the final pool of included studies ( n = 37) that described females. presents the sample characteristics, psychiatric outcomes, and quality ratings for these studies. The section then provides a narrative synthesis and meta-analytic findings for the most commonly reported psychiatric difficulties - anxiety, depression, ADHD, aggression, and self-injury – and concludes with a description of whether these difficulties relate to intellectual ability and autism.

Among the studies included in this review, 23 conducted sex-specific analyses (i.e., analyzed males and females separately) within mixed-sex samples, while 14 focused exclusively on females. In contrast, 58 studies were excluded because they included only male participants. Overall, this pattern indicates that research on males has historically outnumbered that on females by approximately 4:1 – meaning that for every study investigating psychiatric difficulties in females with FXS, about four have focused on males. As shown in , however, recent years reveal a positive shift toward greater balance, now approaching a ratio of 2:1, underscoring the relevance and timeliness of the present review.

Across the 37 studies reviewed, participants’ mean ages ranged from 3.77 years () to 30.27 years (). The majority of studies (89%; n = 33) were conducted in the United States, with two based in Australia (, ), one in Chile (Baker et al., 2019), and one in the United Kingdom (). Reporting on participants’ race and ethnicity was generally limited; of the studies that included such data (24%; n = 9), most samples were primarily White/Caucasian and non-Hispanic (, ; Baker et al., 2019; , , , , , ). Similarly, socioeconomic indicators such as annual household income were rarely reported – among the few studies that did include such data (11%; n = 4), substantial variability prevented meaningful interpretation (, –, , , ). Regarding exclusion criteria, four studies (11%) explicitly excluded participants with psychosis or bipolar disorder (, , , ), while one excluded individuals with significant language or cognitive impairments ().

Among the 13 studies that provided full-scale or verbal IQ scores, mean values ranged from 61.33 () to 97.40 (). Autism prevalence was also reported in 12 studies, with rates spanning from 10% () to 48% (). An earlier study () reported a substantially lower prevalence of 3%, which reflected the narrower DSM-III diagnostic criteria for “autistic disorder” ().