Pharmacological management of fragile X syndrome: a systematic review and narrative summary of the current evidence

ABSTRACT

Introduction

Fragile X syndrome (FXS) is the most common inherited cause of Intellectual Disability. There is a broad phenotype that includes deficits in cognition and behavioral changes, alongside physical characteristics. Phenotype depends upon the level of mutation in the FMR1 (fragile X messenger ribonucleoprotein 1) gene. The molecular understanding of the impact of the FMR1 gene mutation provides an opportunity to target treatment not only at symptoms but also on a molecular level.

Methods

We conducted a systematic review to provide an up-to-date narrative summary of the current evidence for pharmacological treatment in FXS. The review was restricted to randomized, blinded, placebo-controlled trials.

Results

The outcomes from these studies are discussed and the level of evidence assessed against validated criteria. The initial search identified 2377 articles, of which 16 were included in the final analysis.

Conclusion

Based on this review to date there is limited data to support any specific pharmacological treatments, although the data for cannabinoids are encouraging in those with FXS and in future developments in gene therapy may provide the answer to the search for precision medicine. Treatment must be person-centered and consider the combination of medical, genetic, cognitive, and emotional challenges.

KEYWORDS:

• Autism spectrum disorder
• CGG
• co-morbidities
• holistic
• FMR1
• FMRP

1. Introduction

The FMR1 (fragile X messenger ribonucleoprotein 1) gene (full mutation > 200 CGG repeats) is located on chromosome X (q27.3) [1] Mutations in the FMR1 gene cause large expansion of the CGG trinucleotide repeat. This leads to decreased or absent levels of FMRP (fragile X messenger ribonucleoprotein) [2]. Fragile X syndrome (FXS) is considered the most common inherited cause of Intellectual Disability (ID). A systematic review and meta-analysis found the prevalence of full mutation to be 1.4 (95% CI: 0.1–3.1) per 10,000 males and 0.9 (95% CI: 0.0–2.9) per 10,000 females [3]. FXS has a broad phenotypic spectrum and is associated with cognitive deficits, behavioral challenges, and physical characteristics [4]. The penetrance of the physical and behavioral phenotype is dependent on the level of mutation (full mutation/premutation/premutation carrier/carrier/methylation mosaicism) and the subsequent impact upon FMRP levels.

Males with FXS generally have a behavioral phenotype including a level of ID, anxiety, and traits consistent with other neurodevelopmental conditions, including autism (shyness/repetitive stereotyped behaviors/hyper-arousal/sensory needs) and attention deficit hyperactivity disorder (ADHD) (attention deficit/impulsivity). For females with FXS, the phenotype is generally milder (25% ID), with less behavioral challenges [5]. Almost one in three males with FXS may meet the formal criteria for autism, and almost all will have notable traits [6]. The etiology for autism is multifactorial, with numerous candidate genes and genetic variants that may contribute. However, given the level of association of autism with FXS, it has been considered as a surrogate to target autistic features.

FXS has been the target for investigations into pharmacological treatments for a number of reasons [6]. Firstly, there was focus on supplementing identified deficits (folic acid) through dietary intake. Secondly, to target the associated neurodevelopmental/neuropsychiatric disorder (FXAND). This also extends to behaviors that challenge (BtC), with a systematic review identifying prevalence estimates of self-injurious behavior (SIB) at 10–81%, and aggression at 12.5–60.9%, higher than rates estimated in the general ID population [7]. Furthermore, a study of children and adults with FXS (58 males, 39 females) found that 86% of males and 77% of females met the DSM-IV (Diagnostic and Statistical Manual of Mental Disorders IV) criteria for any anxiety disorder [8]. Thirdly, there has been the development of molecules targeted at normalizing/reducing the impact of FMR1 on synaptic connectivity based upon molecular understanding. The impact of these molecules is often measured against more fundamental cognitive functions [6].

The aim of this review was to systematically consider the current evidence base for pharmacological treatment in FXS and provide a narrative summary. This review focused strictly on FXS populations only and aimed to identify high-quality interventional studies. The focus of this review was on efficacy of the pharmacological intervention for clinically relevant outcomes.

2. Methods

We conducted a systematic search strategy of electronic databases Embase, Medline, CINAHL, and PsychINFO (Appendix 1. Example search strategy). The search dates were 1 January 1980 to 2March 2023. The search did not include any other registers or gray literature. The ancestry method was utilized, to identify any potentially eligible articles within the references of articles selected for inclusion. The search was limited to English language journal articles.

2.1. Inclusion criteria

Study design: Randomized, blinded, placebo controlled clinical trials. (RCTs)

Population: Fragile X (only) Excluding FXTAS (Fragile-X Associated Tremor/Ataxia Syndrome)

Intervention: Pharmacological only

Outcomes: Validated objective psychological or behavioral assessment tools with clinical relevance, excluding basic physiological parameters

Focus on efficacy over safety and tolerability.

Published in a peer-reviewed journal (excluding conference abstracts/posters)

We did not include studies with mixed population (unless data could be separated), reviews (unless systematic and meeting inclusion criteria), studies with multiple simultaneous interventions, post-hoc analyses of already considered data, or basic biological parameters.

2.2. Study selection

The search strategy was conducted by an information specialist and clinical academic. The initial abstract results were screened by two reviewers against the inclusion criteria. Full-text reviews were conducted blind (each reviewer blind to results of the other reviewer) by the same two reviewers. Any conflicts were then discussed against the inclusion criteria with the team. The references of included studies were screened for any missed studies that met the inclusion criteria. Final decisions on inclusion were made by the information specialist and clinical academic.

The results were presented in a descriptive narrative. The level of evidence assessed against the Oxford Centre for Evidence Based Medicine (OCEBM) Levels of evidence [9]. For the context of this review, Level 1 evidence consists of good-quality RCTs. A systematic review of RCTs with homogeneity being the highest level of evidence (1a). Level 2 evidence includes cohort studies or lower powered individual RCTs (2b).

3. Results

The literature search identified 2377 articles for screening. Duplicates were removed (364 articles), leaving 1733 abstracts. Following screening against the including criteria, 42 articles were identified for full-text screening (1731 removed). Full text screening identified 22 articles for potential inclusion, of which 16 studies were finally included. Articles were excluded at full-text review for not being a peer reviewed journal article (7); editorials (5); no clinical application of results (5); post-hoc data already considered (3); wrong population (3); no randomization (1); dual intervention (1); pharmaceutical company communication (1) (Figure 1).

Figure 1. PRISMA flow diagram.

3.1. Included studies

The 16 studies included in this review are summarized in Table 1. The focus of this review was to consider the efficacy of the pharmacological interventions on meaningful clinical parameters using validated measures. The results are reported alongside appraisal of the evidence.

Table 1. Summary of studies included in final narrative review (n = 16).

Author	Year	Intervention	Population	Study Design	Outcome measures	Results
Hagerman et al. [10], United States	1986	Folic acid	N = 25 Sex: All male; Age range: 1–31 years (y)	Double-blind, crossover study. Computer randomization, switch at 6 months.	Peabody Picture Vocabulary Test (PPVT) [11], Test of Language Development (TOLD) and an apraxia battery [12]. Yale Revised Developmental Schedules, Stanford-Binet [13], or the Leiter International Scale [14]. Autism Behavior Checklist (ABC) [15], Autism Screening Instrument for Educational Planning (ASIEP) [16], and the Childhood Autism Rating Scale (CARS) [17].	Standardized testing did not show statistically significant changes. Psychological testing demonstrated statistically significant and clinically relevant improvement in IQ for the 8 pre-pubertal males who completed testing.
Hagerman et al. [18]	1988	Methylphenidate and Dextroamphetamine	N = 15 Sex: 13 male, 2 female. Further details: Cytogenic and clinical diagnoses, plus observed attention problems.	Double-blind, Randomized, crossover design, 3 weeks duration	Behavioral Checklists: the Comers Abbreviated Parent-Teacher Questionnaire [19], ADD-H: Comprehensive Teacher Rating Scale (AcTeRs) [20].	Improvement was seen with methylphenidate in socialization skills aid attention span according to teacher checklists. Ten children were clinically considered responders and treatment was continued after the study.
Strom et al. [21]	1992	Folinic acid	N = 21 Sex: All male; Age range: 2–22 y (mean 8.3 y); Further details: Caucasian participants, including one sibling pair	Randomized, double-blind, placebo-controlled crossover study. Two treatment intervals of 12 week each	Vineland Adaptive Behavioral (VAB) Scales [22], PPVT-Revised; the Conners Parent and Teacher Rating Scales; ACTeRS; and a questionnaire designed by the investigators.	No statistically significant differences between the treatment and placebo phases with any instrument. No significant side effects have been noted from long-term folic acid therapy.
Berry-Kravis et al. [23]	2006	CX516	N = 49 Sex: 38 male, 11 female; Age range: 18-50y; Further details: FXS diagnosis by DNA analysis; cognitive impairment with measured IQ 20 to 85; normal hearing and normal or corrected vision; stable medication regimen, negative pregnancy test	Phase II, 4-week randomized, double-blind, placebo-controlled clinical trial was conducted to evaluate the safety and efficacy.	Clinical Global Impression Scale-Improvement (CGI-I) [24], Visual Analogue Scale (VAS), Swanson, Nolan, and Pelham Questionnaire IV (SNAP-IV) [25].	No significant difference between the placebo and C×516groups in safety parameters, or any of the cognitive and behavioral outcome measures.
Torrioli et al. [26]	2008	L-acetylcarnitine (LAC)	N = 63 Sex: All males; Age range: 6–13 y; Further details: Participants had co-occurring ADHD	Phase II, randomized, double- blind, placebo-controlled, parallel, multicenter study.	CGI-P; CGI-T; VAB Composite, and Vineland Socialization domain.	A stronger reduction of hyperactivity and improvement of social behavior in patients treated with LAC, compared with the placebo group.
Wirojanan et al. [27]	2009	Melatonin	N = 18 Sex: 16 male, 2 female; Age range: 2–15 y (mean 6 y, standard deviation [SD] 3.5 y); Further details: Participants had co-occurring reported sleep problems.	Randomized, double blind, placebo-controlled, 4-week crossover design.	Sleep variables: sleep-onset time, total night sleep duration, sleep-onset latency time, and the number of night awakenings.	Melatonin was associated with modest but significant improvement in mean night sleep duration, mean sleep-onset latency was shorter, and mean sleep-onset time was earlier.
Sahu et al. [28]	2012	Donepezil	N = 20 Sex: All males; Age range: 6–15 y; Further details: Cytogenic diagnosis of FXS, stable treatment for 8 weeks.	12-week, randomized, double-blind, placebo-controlled, parallel-group study.	Stanford-Binet Intelligence Scale, Conners 3 Parent Rating Scale (Short) and CARS.	No significant difference with placebo.
Leigh et al. [29]	2013	Minocycline	N = 66^2 Sex: male 47, female 8; age range: 3.5–16y (mean 9.2 y, SD 3.6 y); Further details: FXS confirmed by FMR1 DNA testing; stable regimen of pharmacological treatment.	Randomized, double-blind, placebo-controlled, crossover trial.	CGI-I and VAS 1 score (severity of target behavior 1).	Minocycline treatment for 3 months was associated with modest but significant benefits in global functioning when compared with placebo.
Berry-Kravis et al. [30]	2016	Mavoglurant	N = 314 Sex: 288; female 26. Age range 12–45 y; Further details: Cytogenic diagnosis of FXS, CGI-S score of ≥4, ABC-C score of >20, and an IQ score during screening that was lower than 2 SDs below the median	Two phase 2b, randomized, double-blind, placebo- controlled, parallel-group trials of mavoglurant in adult (18–45 y) and adolescent patients (12–17 y) with FXS 12 weeks.	Aberrant Behavior Checklist-Community Edition, FXS-specific (ABC-CFX) [31,32]; CGI-I scale.	No significant difference with placebo in primary outcomes measures in either study.
Hess et al. [33]	2016	Sertraline	N = 57 Sex: male 48, female 9; Age range: 2–6 y	Randomized, double-blind, placebo-controlled trial	MSEL expressive language raw score, MSEL expressive language standard score [34], and CGI-I.	No difference between placebo and sertraline for primary outcome measures.
Berry-Kravis et al. [35]	2017	Arbaclofen	N = 278 (N = 119 adult/adolescent group, N = 159 child group) Sex: 243 male, 41 female (adult adolescent group, 99 male, 26 female; child group, 144 male, 15 female); age range: 5–50 y (12–50 y adolescent/adult group, 5–11 y child group); Further details: DNA-confirmed FMR1 full mutation	Two phase 3 placebo-controlled trials were conducted, flexible dosing.	Social Avoidance subscale of the ABC-CFX. Secondary outcomes included other ABC-CFX subscale scores, CGI-I, CGI-Severity, and VAB, Second Edition (Vineland-II) Socialization domain score.	No significant difference to placebo in adolescent/adult group. In the child group, highest dose group improved on irritability (p = 0.03) and parenting stress Index (p = 0.03).
Youssef et al. [36]	2017	Basimglurant	N = 183 Sex: 149 male, 34 female; age range: 14–50 y (mean 23.4 y)	Phase 2, double-blind, randomized, placebo-controlled, 12-week study.	ADAMS total score [37], a 28-item caregiver-completed instrument assessing aspects of anxiety, social interaction, compulsiveness, communication, concentration, depression, and mood.	No improvement over placebo.
Bruno et al. [38]	2019	Donepezil	N = 41 Sex: 26 male, 15 female; age range: 12–29 y (mean 19.6 y); Further details: cytogenic diagnosis of FXS; Verbal IQ 50 to 75 Tanner pubertal development stage 3 or higher	Randomized placebo-controlled trial study; 12 weeks.	Executive functioning: Contingency Naming Task (CNT) [39] ABC, neural processing assessed using functional MRI.	No significant difference to placebo on any outcome measure.
Berry-Kravis et al. [40]	2020	Trofinetide	N = 70 Sex: All males Age range: 12-41y (mean 23.5y)	Phase 2, multi-center, double-blind, placebo-controlled, parallel group study.	Investigator developed the fragile X syndrome rating scale (FXSRS) and the fragile X syndrome domain-specific concerns (FXSDSC), VAS, CGI-I, ABC-CFX.	Trofinetide at the 70 mg/kg dose level demonstrated no worsening of symptoms. A trend toward improvement in cores symptoms of FXS.
Berry-Kravis et al. [41]	2021	BPN14770	N = 30 Sex: all males Age range: 18–41 y	Phase II, 24-week randomized, placebo-controlled, two-way crossover study.	National Institutes of Health Toolbox Cognition Battery assessments of Oral Reading Recognition, Picture, Cognition Crystallized Composite score, visual analogue caregiver rating scale [42].	Significant improvement across cognitive domains with collaborating caregiver rating on VAS.
Berry-Kravis et al. [43]	2022	Transdermal gel Cannabidiol (ZYN002)	N = 212 Sex: 159 male, 53 female; Age range: 3–<18 y (mean 9.7 y); further details: 79.7% more than 90% methylation of the FMR1 promotor and full mutation of FMR1; body mass index of 12–30 kg/m^2	Double‐blind, randomized phase 3 trial assessing efficacy and safety of ZYN002.	Change in social avoidance (SA) measured by the ABC‐CFXS SA subscale.	No significant difference for overall cohort. For those with over 90% methylation of the FMR1 promotor significant improvements in primary outcome plus isolation, irritability, disruptive behaviors and social interactions.

1. 63 participants were enrolled; 7 patients dropped out, 56 completed the one-year treatment, and 51 were included in the statistical analysis.

2. 66 randomized; 55 completed first period, 48 completed full trial.

3.1.1. Folic acid

Level of Evidence: 2b

Risk of Bias: Small sample size, no ITT, crossover study design.

Tolerability: Not formally measured

Efficacy: No significant difference to placebo

Disorders of the metabolism of folate and folate deficiency is known to be associated with a range of neurological problems, not specific to FXS. It was hypothesized that FXS was associated with folic acid deficiency/problems with metabolism. This review identified one randomized, double-blind, placebo-controlled trial including 25 males with FXS [10]. This was a six-month cross over study. The dose of Folic acid was 5 mg twice daily. The results showed no significant difference to placebo in performance overall on any of the validated psychological or behavioral outcome measures. Two participants were removed from the study due to significant deterioration in functioning when folic acid was removed (treatment then continued). Four pre-pubertal participants in each treatment group completed psychological testing (IQ score) and showed significant improvement with folic acid compared to placebo. This suggests a possible response in younger children with FXS, however the magnitude/relevance of this IQ change is not clearly reported. Furthermore, the low IQ scores (average IQ 27) reduce the accuracy of testing.

3.1.2. Methylphenidate

Level of Evidence: 2b

Risk of Bias: Small sample size, no ITT, cross over design, sort intervention period.

Tolerability: No difference on side effects scores between methylphenidate and placebo.

Efficacy: Significant improvement on attention and social skills factor domains.

This review identified one randomized, double-blind, placebo-controlled trial [18]. This was a crossover study design with three phases (placebo/methylphenidate/dextroamphetamine). Each intervention was assigned randomly for a one-week period. The methylphenidate dose was 0.3 mg/kg twice daily. The total number of participants was 15 (13 males, 2 females). The only statistically significant findings were improvement on the domains of attention (p = 0.025) and social skills factor (p = 0.034) improved on methylphenidate. There were no significant differences between dextroamphetamine and placebo. Ten of the 15 participants were considered responders and continued with intervention. However, there was a strong trend for improvement across all other variables.

3.1.3. Folinic acid (leucovorin)

Level of Evidence: 2b

Risk of Bias: Small sample size, crossover design.

Tolerability: No side effect reported with folinic acid

Efficacy: No significant difference with placebo on any psychological or behavioral assessment tool

Folinic acid is the active metabolite of folic acid. As a result, it skips the stages of folic acid activation. It is hypothesized that it may have therapeutic advantages over taking folic acid. This review identified one randomized, double-blind, placebo-controlled trial [21]. A cross over design, with intervals of 12 weeks. The study included 21 male participants. Only 16 participants completed both phases of the study. There were no statistically significant differences between folinic acid and placebo on any of the validated assessment tools.

3.1.4. AMPA-Modulating CX516

Level of Evidence: 2b

Risk of Bias: Small sample size, no ITT, short intervention period.

Tolerability: No significant difference to placebo. No serious adverse events. One withdrawal on CX516 due to potentially serious rash.

Efficacy: No significant difference to placebo.

CX516 is an Ampakine compound with evidence to support potential utility for cognitive deficits. AMPA-receptor activity may be deficient in FXS. Ampakine’s may enhance AMPA-related neurotransmission and promote BDNF mediated synaptic maturation. Therefore, CX516 is a targeted treatment that may impact upon the behavioral phenotype of FXS [23]. This review identified one study including 49 participants (38 males, 11 females). This study was a 4-week randomized, double-blind placebo-controlled trial. There were no significant differences between CX516 and placebo across any of the validated assessment tools.

3.1.5. L-acetylcarnitine

Level of Evidence: 2b

Risk of Bias: No ITT

Tolerability: No significant side effects identified.

Efficacy: Significant improvement across primary outcome measures (clinical global impression)

L-acetylcarnitine (LAC) has been considered as a possible alternative to stimulant treatment for ADHD symptoms in FXS. LAC has been shown to reduce hyperactivity, an impulsive behavior in association with ADHD, including within a small male FXS population [44,45]. LAC is thought to act on the FMR1 promotor. This review identified one study including 56 males with FXS and ADHD [26]. This was a multicentre, randomized, double-blind, parallel, placebo-controlled trial over 12 months. Only 51 were included in the final analysis (24 LAC, 27 placebos). The dose of LAC was 1000mg/day. The results show significant improvement in the treatment group for the primary outcome measures. This study builds on a previous review from the same group; this was published as a letter to the editor and so did not meet the inclusion criteria for this review [44].

3.1.6. Melatonin

Level of Evidence: 2b

Risk of Bias: No ITT, mixed population, crossover design, accuracy of technology, sleep measures

Tolerability: No significant side effects identified.

Efficacy: Significant improvement in sleep-onset time.

Melatonin is a pharmacological treatment for sleep disorders. Melatonin is secreted from the pineal gland at night and regulates sleep in accordance with circadian rhythm. Autism and FXS sleep disorders are common, and melatonin levels can vary resulting in abnormal circadian rhythm, due to a range of postulated mechanisms. This review identified one study including 18 participants (16 males, 2 females) with clinically reported sleep problem [27]. This was a randomized, double-blind, placebo-controlled study, with a 4-week cross over design. The melatonin dose was 3 mg. The results show a significant difference in sleep onset time, although not necessarily to a clinically relevant level.

3.1.7. Donepezil

Level of Evidence: 1b

Risk of Bias: No ITT, short duration

Tolerability: No significant side effects identified.

Efficacy: No significant difference to placebo on cognitive or behavioral measures

It has been postulated that cholinergic channels may play a role in some of the phenotypic features of FXS. It has been demonstrated in animal models with FX that there is cholinergic dysfunction in memory and learning regions of the limbic system [46]. This review identified two studies with Donepezil as the treatment.

The first study included 20 participants, all male [28]. This was a randomized, double-blind, placebo controlled, parallel group study. The duration was 12 weeks. The dose of Donepezil was 2.5 mg daily for 4 weeks, increasing to 5 mg daily for 8 weeks. There was no difference to placebo on cognitive or behavioral measures.

The second study included 41 participants (27 males, 15 females), with a focus on basic cognitive functions and including functional MRI imaging [38]. Participants were randomized to receive Donepezil (2.5–10 mg) or placebo for 12 weeks. There was no significant difference to placebo on cognitive or behavioral outcomes.

3.1.8. Minocycline

Level of Evidence: 2b

Risk of Bias: Blinding inconsistent, crossover design

Tolerability: No serious adverse events, no significant difference to placebo

Efficacy: Significant improvement in primary outcome measure based on ITT, although difference modest and may not be clinically relevant.

Minocycline is thought to be a neuroprotective agent. In FXS, the lack of FMRP is associated with alterations in the expressions of a number of other proteins including MMP9 (matrix metalloproteinase 9). MMP9 plays a role in the synaptic function and plasticity in the hippocampus. Levels may be raised in association with FXS. Minocycline treatment has been found to lower these levels in animal models [47]. This review identified one study including 55 participants [29]. This was a randomized, double-blind, placebo-controlled, crossover trial, 3 months for each phase. The results showed a statistically significant improvement in clinical global impression; however, results may not be clinically relevant.

3.1.9. Mavoglurant

Level of Evidence: 2b

Risk of Bias: Short duration, validity of assessment tools

Tolerability: Adverse events were higher in the treatment group, and at higher dosing. Discontinuation was higher in the treatment group. The most common adverse effects were psychiatric.

Efficacy: No significant difference to placebo in any primary outcome measure in adolescent or adult cohort.

Metabotropic glutamate receptor (mGluR) signaling may be overactive in FXS and may lead to increased long-term depression in the hippocampus. This is a postulated theory for the FXS phenotype [48]. This review identified one article including two studies or cohorts with similar design (adolescent cohort and adult cohort) [30]. The studies included 175 adults (age 18–45 years) and 139 adolescents (age 12–17 years). The duration of both studies was 12 weeks. The dose of Mavoglurant was 25, 50, or 100 mg twice daily. There was no significant difference to placebo in either study in any of the primary outcome measures.

3.1.10. Sertraline

Level of Evidence: 2b

Risk of Bias: Inconsistent blinding, not all outcome measures completed at baseline, impact of normal development

Tolerability: No serious adverse events. No significant difference with placebo.

Efficacy: No significant difference to placebo in any primary outcome measure.

Selective Serotonin Re-uptake Inhibitors (SSRIs) such as Serotonin are commonly prescribed in association with FXS. This review identified one study that met the inclusion criteria. Sertraline was investigated in younger children, not with specific focus on the treatment of comorbid symptoms of psychiatric illness (anxiety or mood disorder). Serotonin levels may be lower in children with autism, and SSRIs may be beneficial to combat the lack of synaptic maturation in FXS and may also stimulate BDNFs’ role in the early development period. This review identified one study, a randomized double-blind, placebo-controlled trial, including 52 children (age 2–6 years) [33]. The duration was 6 months. The dose of sertraline was low (between 2.5 and 5 mg per day, depending on age). There was no significant difference to placebo on primary outcome measures.

3.1.11. Arbaclofen

Level of Evidence: 2b

Risk of Bias:

Tolerability: No serious adverse events, similar in both placebo and treatment groups.

Efficacy: No significant difference to placebo in primary outcome measure in child or adolescent/adult cohort. Reduction in irritability measure in child group at the highest dose.

As we have discussed inhibition of the mGLuR signaling may correct the developmental phenotype associated with FXS in animal models (mGLuR5). The use of GABA agonists may be a mechanism to reduce glutaminergic activate through regulation. Arbaclofen is a potent GABA-B agonist with observed behavioral benefits in autistic people, and preliminary data suggest such benefits also have application in FXS [49]. The review identified one article including two studies or cohorts with similar design (adolescent/adult cohort, 12–50 years and child cohort, 5–11 years) [35]. The studies included 119 in the adolescent/adult study and 159 in the child study. These studies were multisite, randomized, double-blind, placebo controlled, parallel group in both cohorts. The dose of Arbaclofen was titrated up though trial (5 mg twice daily to 15 mg three times daily). There was no significant benefit to placebo on primary outcome measures (social avoidance). In the child group, the highest dose court had significant benefit on other secondary measures including irritability, with an effect size similar to SSRIs.

3.1.12. Basmimglurant

Level of Evidence: 2b

Risk of Bias: Short duration of study, validity of assessment tools

Tolerability: Well tolerated, but a far higher rate of adverse effects (including psychiatric effects) with treatment (dose-dependent).

Efficiency: No significant difference to placebo in primary outcome measures in either group. Outcomes were better in the placebo group (possible side-effect related).

Basmimglurant is an mGluR5 Negative Allosteric Modulator, again considered based upon the underlying ‘mGluR theory’ of FXS phenotype being driven by excessive mGluR-dependent protein synthesis. This review identified one study, including 183 adults/adolescents (14–50 years mean age 23.4) [36]. This was a 12-week, multi-site, randomized, double-blind, placebo-controlled trial using two doses of Basmimglurant (0.5 mg and 1.5 mg). There was no difference to placebo in either group in primary outcome measures.

3.1.13. Trofinetide

Level of Evidence: 2b

Risk of Bias: Small sample size, short duration, new unvalidated assessment tools

Tolerability: No serious adverse events, similar in both placebo and treatment group.

Efficacy: No significant difference to placebo in primary outcome measure in child or adolescent/adult cohort. Reduction in irritability measure in child group at the highest dose.

Trofinetide is an analog of insulin-like growth factor 1 (IGF-1). Levels of IGF-1 have been found to be low in FMR1 knock out mice. Treatment with Trofinetide normalized the behavioral phenotype (and raised IGF-1) [50]. This review identified one study including 70 participants, all male [40]. This was a randomized, double-blind, placebo-controlled trial, over 28 days. Participants were randomly assigned to placebo, trofinetide 35 mg/kg, or trofinetide 70 mg/kg. There was no significant improvement compared to placebo on any of the outcome measures. Trend of improvement with treatment observed.

3.1.14. BPN14770

Level of Evidence: 2b

Risk of Bias: Small sample size, recruitment, cross over design

Tolerability: No serious adverse events, no significant difference to placebo

Efficacy: Significant improvement on secondary outcome measures (language, communication, level of functioning)

BPN14770 is a phosphodiesterase-4D (PDE4D) allosteric inhibitor. PDE4D is a key modulator of cAMP activity, related to learning and memory. cAMP metabolism may be altered in FXS, based upon animal models. PDE4D missense mutations are associated with a neurodevelopmental disorder phenotype with ID. This review identified one study including 30 adult male patients from a single center [41]. The dose of BPN14770 was 25 mg twice daily. Significant and clinically relevant improvement on secondary outcomes include language and daily functioning (parent/caregiver observation only).

3.1.15. Transdermal cannabidiol gel

Level of Evidence: 2b

Risk of Bias: Validity of assessment tools, Post-Hoc analysis (small sample size)

Tolerability: No serious adverse events, safe and well tolerated, application site pain main adverse event.

Efficacy: Significant improvement on secondary outcome measures (language, communication, level of functioning)

FXS is associated with dysregulation in endocannabinoid signaling. It is therefore postulated that treatment with cannabidiol may be effective. Preliminary clinical results with transdermal cannabidiol, get (ZYN002) have shown positive impact on anxiety and behavior [51]. This review identified one study including 212 participants, 75% male, age 3 to <18 years [43]. This was a randomized, double-blind, placebo controlled, multi-site trial, over 12 weeks. Treatment doses were 250 mg or 500 mg daily. No significant improvement with placebo in primary outcome measures. There was significant improvement in the subgroup with ≥90% methylation (generally more severe phenotype).

4. Discussion

A previous systematic review of the pharmacological treatments in FXS employed less restrictive inclusion criteria [52]. The studies were controlled but not restricted to requiring randomization. The review identified 10 studies, with a crossover to those eligible for this review. In this updated, more restricted systematic review we can identify that the quality of methodology in the studies and level of evidence is improving over time. We can observe the development from controlled designs through to RCTs by the same research groups in those molecules identified as safe and tolerable. This review has also identified new pharmacological treatment options, and these have been investigated using a parallel design.

There are limitations to the literature search based upon the inclusion criteria and sources searched. A previous literature review conducted by experts in the field provides an overview of the evolving status of clinical trials focused on interventions in FXS [53–56]. The authors indicate over 20 clinical trials completed or in progress. If we cross-reference there is significant overlap with the results of this systematic review. A small number of studies included would not have met the inclusion criteria for this review based upon methodology (no randomization, open-label trials), outcome measures (no assessment of global clinical impact), or no results data available. This review also only included peer reviewed journal articles (excluding conference abstracts). If we consider the additional studies, then there is no significant new information provided for the compounds already included in this review.

However, the review by the expert research team did identify one study that would have met the inclusion criteria for this review. A randomized double blind, placebo-controlled trial of Ganaxolone (GABA-A agonist) in children and adolescents (age 6–17 years) with FXS (n=59, male 50, female 9). The results found Ganaxolone to be safe and tolerable; however, there was no significant difference between treatment group and placebo group for the primary outcome measure (CGI). A post-hoc subgroup analysis suggests the potential for benefit for those with anxiety and lower IQ [57].

Researchers remain restricted by the prevalence of FXS, which has led to the use of crossover studies of short duration that introduces the significant risk of bias. There is a need for multi-center collaborative project to ensure enough participants to demonstrate any significant effects. On observing the development of the research base, there are specific research centers with a focus on identifying treatment options for FXS, for which the FMR1 mutation was only identified around 35 years ago. There has been progression through the research pathway to more robust study designs with improved power that is improving the evidence rapidly.

One challenge posed to researchers is how to measure success. If we consider a model of behavior, there are multi-factorial influencers [58]. Behavior is a crude representation, and broad characteristics (e.g. aggression) may be influenced by a plethora of genetic variants that then lead a range of consequences to different proteins and their function. There are a number of validated assessment tools that consider global impact and level of function [55]. We can also observe that more specific validated assessment measures have been developed by the expert researchers in this field. Being a single gene disorder, FXS provides the opportunity to focus on interventions targeted at the behavioral phenotype, particularly for those with complete or near-complete methylation, associated with a more severe phenotype. Moreover, the strong association with autism and autistic traits may provide inside into treatment of core autistic features. This has led to the development of treatment strategies based upon our molecular understanding of the FMR1 gene and the effect on FMRP and wider impacts including synapse function, maturation, and plasticity. Disease-specific treatments targeted, and these changes have been successful in animal models (FMR1 KO mice). However, the findings from this review indicate that these results are not easily replicated in human studies. Perhaps highlighting the challenges in recruiting enough participants and in study methodology.

In this review, there has been a trend toward improvement with intervention driven by a focus on precision medicine. We can also learn that interventions may need to be targeted at particular populations. This includes a focus on those with complete or near-complete methylation, and the recurring theme that outcomes are better for younger children. This correlates with earlier intervention in the neurodevelopmental period but poses ethical considerations when developing study protocols.

4.1. Neurodevelopmental/Psychiatric Co-morbidities

There are a range of lower-level studies that consider treatment options for the most common co-morbid symptoms/diagnoses in FXS. This evidence base and clinical utility are summarized by Hagerman et al. (2009), as part of the leading research collaborative in FXS treatments [6]. The focus of pharmacological treatments covers a range of key domains (Table 2).

Table 2. Evidence base for treatment considerations by domain in FXS [6].

Behavioral Phenotype	1. mGluR5 antagonists. 2.Ampakines (CX516) 3. Lithium
ADHD	1. Stimulants (methylphenidate). 2. ά-Adrenergic receptor agonists-Clonidine/Guanfacine-(profile: more severe phenotype, younger children, unable to tolerate stimulants). 3. L-acetyl-carnitine
Anxiety	SSRIs-Fluoxetine, Sertraline
BtC	Antipsychotic’s (second generation) – Aripiprazole, Risperidone

(Aggression,/SIB/mood dysregulation).

*Please note that these treatment considerations are aligned to the current limited evidence base. This includes evidence that does not meet the criteria for the review. These associations are not treatment recommendations, and many of these options and may not be available to clinicians. Treatment should be person centered and overseen by a clinician with relevant expertise.

5. Conclusion

This systematic review has identified 16 studies that focus on pharmacological interventions for FXS across a range of indications. This review finds that there is limited higher level evidence available confirming the efficacy of any of the molecules tested. However, there is a clear research pathway of developing compounds based upon molecular understanding driving precision medicine, which must be the focus for the future and FXS researchers may be leading the way for other complex gene disorders.

6. Expert opinion

The evidence base for pharmacological treatments in FXS is developing rapidly, driven by increased understanding of the molecular impact of the FMR1 mutation. FXS provide a single gene disorder with a defined phenotype to target treatment considerations that may potentially be useful in wider populations. Given changes to neuroplasticity [59] over time and the impact on planning families [60], there may well be possible benefit to early diagnosis as well as social, educational, and/or pharmaceutical intervention but access to diagnostics and care may be inconsistent or be associated with waiting lists.

Based on this review to date, there is some encouraging data but no specific findings to support the newer pharmacological treatment options, although the data for cannabinoids are encouraging in those with full expansions and there is hope that gene therapy may be developed using antisense oligonucleotides [61]. We are currently reliant on the lower level evidence available in combination with expert clinical opinion to guide treatment choice. The treatment of co-morbidities has an evidence base specifically in FXS, and we need to extrapolate from wider populations and be mindful of the possibility of carriers suffering from under-diagnosed anxiety, ataxia, or premature ovarian insufficiency [62].

This review identified a range of potential treatment options that have been investigated through methodologies that did not meet the strict inclusion criteria. In the main, the studies represented were open-label cohort studies. Many of these initial investigations show promising results and are being investigated through randomized placebo-controlled trials now or will be in the near future.

One drug of particular note for discussion is Metformin. Metformin is well studied in many settings, generally well tolerated, low cost, has a range of properties, and currently mainly utilized for the treatment of diabetes. Metformin also has anti-cancer properties and has been found to have beneficial effects in a number of neurological conditions including Alzheimer’s Dementia, depression, and FXS [63].

Alongside other known effects of Metformin, it likely plays a role in regulation of the mTOR and AMPK signaling pathways and potentially has a range of direct and indirect neuroprotective effects [63,64]. Metformin has indicated positive results in mice models and a case series of seven people with FXS who were treated with Metformin clinically [65]. Alongside weight loss, the results demonstrated consistent improvements in irritability, social responsiveness, hyperactivity, and social avoidance. There were no significant side-effects. Currently, a multi-center controlled trial is underway.

Any approach to treatment must be person-centered and holistic. This should involve a multidisciplinary team to undertake in-depth assessments of an individual’s needs given the combination of medical, genetic, cognitive, and emotional challenges. Databasing of research interests [66], working with patient and family organizations and adoption of internationally approved metrics for tracking cognitive improvements [67] could well be beneficial in improving access to future interventional research.

Article highlights

• Mutation of the FMR1 (Fragile X Messenger Ribonucleoprotein 1) gene leading to Fragile X Syndrome (FXS) provides an opportunity to consider pharmacological interventions on a molecular level.

• This systematic review of Randomised Controlled Trials identified 16 studies including 15 treatments specifically focused on FXS including specific symptoms and co-morbidities associated with the broad phenotype.

• Many of the treatments still being developed show positive findings, but results are not always translated to a significant clinical benefit.

• Further research will focus on instruments that measure outcomes specific to FXS, and require collaboration between research centres for larger cohorts.

• There is promise in a number of the pharmacological treatments identified including Cannabidiol based products and the versatile drug Metformin.

• FXS pharmacological research is leading the way in precision-based medicine, developments in gene therapy will likely be the future.

Declaration of interest

R Shankar has received institutional and research support from LivaNova, UCB, Eisai, Veriton Pharma, Bial, Angelini, UnEEG, and Jazz/GW pharma outside the submitted work. He holds grants from NIHR AI, SBRI, and other funding bodies all outside this work.

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Author contributions

All authors satisfy the ICMJE guidance by substantially contributing to the design, analysis and interpretation of the work, drafting of the manuscript, final approval of the manuscript, and all agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Data statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.

Appendix 1

Search Strategy

1. (Fragile X or FragileX).ab,ti,kf. 9998

2. exp fragile X syndrome/ 9736

3. ((pharmacological or clinical or medical) adj2 (treatment* or management)).ab,kf,ti. 326665

4. pharmacotherapy.ab,kf,ti. 57391

5. ‘medication*.’ab,kf,ti. 649360

6. exp drug therapy/ 3414714

7. sertraline.ab,kf,ti. 8037

8. zoloft.ab,kf,ti. 177

9. exp sertraline/ 29425

10. metformin.ab,kf,ti. 45982

11. exp metformin/ 81466

12. cannabidiol.ab,kf,ti. 6845

13. CBD.ab,kf,ti. 16534

14. exp cannabidiol/ 8236

15. acamprosate.ab,kf,ti. 1312

16. exp acamprosate/ 2718

17. lovastatin.ab,kf,ti. 5325

18. minocycline.ab,kf,ti. 10716

19. exp minocycline/ 27761

20. anti-depressant*.ab,ti,kf. 4459

21. antidepressant.ab,ti,kf. 72485

22. exp antidepressant agent/ 553449

23. ‘SSRI*.’ab,ti,kf. 19868

24. selective serotonin reuptake inhibitor.ab,ti,kf. 6609

25. ‘SNRI*.’ab,ti,kf. 3232

26. Serotonin norepinephrine reuptake inhibitors.ab,ti,kf. 1198

27. (Noradrenergic and specific serotonergic antidepressants).ab,ti,kf. 52

28. NaSSa.ab,ti,kf. 187

29. ‘tricyclic anti-depressant*.’ab,kf,ti. 252

30. tricyclic antidepressant.ab,kf,ti. 4602

31. ‘anti-psychotics*.’ab,kf,ti. 792

32. antipsychotic.ab,ti,kf. 53568

33. exp neuroleptic agent/ 284094

34. ‘anti-seizure medication*.’ab,kf,ti. 1275

35. ‘antiseizure medication*.’ab,kf,ti. 1803

36. ASMs.ab,kf,ti. 1203

37. ‘anti-epileptic medication*.’ab,kf,ti. 802

38. ‘antiepileptic medication*.’ab,kf,ti. 2831

39. AEMs.ab,kf,ti. 488

40. exp anticonvulsive agent/ 472103

41. ‘mood stabiliser*.’ab,kf,ti. 737

42. ‘Mood stabilizer*.’ab,kf,ti. 6755

43. exp mood stabilizer/ 136894

44. ‘adhd medication*.’ab,kf,ti. 1436

45. ‘Benzodiazepine*.’ab,kf,ti. 55426

46. exp benzodiazepine derivative/ 253811

47. 1 or 2 12,569

48. 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33 or 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46 4,877,314

49. 47 and 48 1546