Coenzyme Q10, Vitamin E and Polyvitamin B: an exploratory double-blind randomized cross-over study in Phelan-McDermid Syndrome

INTRODUCTION

Phelan-McDermid syndrome (PMS, OMIM#606232) is a rare genetic neurodevelopmental disorder mainly characterized by global developmental delay, severe deficits in expressive language, and intellectual disability (ID)^[1-4]. The prevalence of PMS has been estimated in the range of 2.5-10 cases per million births^[5]. However, the prevalence of PMS is likely to be underestimated and over 3,000 individuals self-identified as PMS patients worldwide are registered in the “Phelan-McDermid Syndrome International Registry” of the PMS Foundation (https://www.pmsf.org/international/)^[5]. Autism Spectrum Disorder is present in as many as 30%-80% of PMS patients, while at least 0.5%-2.0% of autistic individuals carry a genetic abnormality related to PMS in chr. 22q13^[6-8]. A wide range of other neurological and behavioral signs/symptoms may also be present, such as muscle hypotonia and deficits in motor coordination, seizures, structural brain abnormalities, gastrointestinal symptoms, renal malformations, and minor dysmorphic features^[1-4,7,8]. Recently, other psychiatric disorders were found to be associated with PMS, including catatonia, mood, and psychotic disorders^[9]. Lasting regressions in communication, self-care, and motor function beginning within 3 years of the onset of psychiatric illness were found to exert a very negative impact on clinical prognosis^[10].

The pathophysiology of PMS is due to a spectrum of genetic anomalies in the terminal region of the long arm of chromosome 22, ranging from disruptive single-nucleotide variants to ring chromosomes and large deletions affecting multiple genes^[2]. SHANK3 (OMIM *606230), located in the distal portion of the long arm of chromosome 22, is the strongest candidate gene to determine PMS^[3,11]; however, genotype-phenotype correlations in PMS are complex, as even interstitial deletions sparing SHANK3 have been described in some PMS patients^[12]. The size of PMS-causing deletions can be extremely variable, ranging from 0.22 to 9.22 Mb^[8], and a multitude of genes distributed along this terminal region may contribute to phenotypic heterogeneity and clinical severity^[13]. Furthermore, rare autosomal recessive forms of mitochondrial disorders have been found to emerge from the coincidence of a large PMS-causing deletion on one chr. 22q13 allele and of a missense mutation or small deletion not detectable by array-CGH located on the other allele^[14]. The larger chr.22 segment deleted in PMS hosts several genes involved in mitochondrial function (SCO2, CHKB, TYMP, TRMU, NDUFA6, SAMM50, SULT4A1)^[13,15]. Abnormalities in electron transport chain (ETC) complex activity have been detected in 30/51 (59%) individuals with PMS^[15], supporting the hypothesis that mitochondrial dysfunction could contribute to the pathological process and phenotypic expression of PMS in many patients.

Behavioral and psychological/physical treatments are considered the first line intervention, because there is no pharmacological compound yet proven effective in ameliorating the core signs/symptoms of PMS^[16]. Pharmacological treatment strategies adopted to improve psychiatric comorbidities and sleep disorders are not different from those prescribed in the general population^[17], but responses to traditional psychiatric intervention are variable and sensitivity to the development of side effects is seemingly high^[9].

Coenzyme Q10 (CoQ10, ubiquinone) is a lipid-soluble quinone present in the highest concentrations in brain, heart, kidney, and liver tissue^[18]; tissue levels reflect the extent of metabolic activity and energy of each tissue and organ^[19]. CoQ10 can be found in three distinct redox states: oxidized (ubiquinone), semiquinone (ubisemiquinone), and reduced (ubiquinol)^[18]. It is synthesized by human cells and can also be partly obtained from dietary sources (on average 5 mg/day)^[20]. In the latter case, Q10 ubiquinone is reduced to Q10 ubiquinol in the liver following gut absorption. Both ubiquinone and ubiquinol are currently available as dietary supplements^[21]. CoQ10 plays an important role in supporting energy production in living cells^[19], and orally administered CoQ10, in the presence of partial tissue deficiency, can boost electron transfer and ATP synthesis^[22-24]. More specifically, CoQ10 is present in two intracellular pools, a protein-bound pool (30%) and a free-cytosolic pool^[21]. The former plays a pivotal role in the electron chain transport system responsible for the synthesis of ATP, whereby CoQ10 transfers electrons from complex I and II to complex III by shuttling in its redox cycle^[21]. Through this mechanism, Q10 supplementation can boost energy levels and reduce the risk of mitochondrial damage due to excessive oxidative stress^[21-25]. The cytosolic pool participates in other metabolic processes, most importantly playing an additional antioxidant role by protecting lipids and other cell components, while also restoring reduced levels of other antioxidants^[21]. Overall, increased energy production and antioxidant capacity are predicted to limit the damage generated by the neuroinflammation and excitotoxicity well documented in ASD brains, ultimately leading to excessive neuritic pruning and/or cell apoptosis^[26,27]. Interestingly, serum coenzyme Q10 levels were found to interact with genetic predisposition to a variety of neuronal diseases in a genome-wide association study^[28]. Moreover, converging evidence strongly supports its neuroprotective effects^[22,29,30] and numerous disease processes can benefit from oral administration of CoQ10, including neurodegenerative diseases^[24,31-34].

In an attempt to transfer this knowledge “from bench to bedside”, we designed a “metabolic support therapy” (MST) encompassing Q10 ubiquinol or CoQ10 (50-100 mg b.i.d.) + vit. E (30-60 mg/die) + poly vitamin B (B50 complex, 1/2 or 1 caps/die), under the hypothesis that its metabolic, antioxidant and neurochemical effects could at least partly improve the clinical signs/symptoms of neurodevelopmental disorders. Recently, we performed a retrospective chart review reporting small-to-moderate improvement in cognitive and adaptive functioning, social interaction and motor coordination in 45/59 (76.3%) patients with different neurodevelopmental disorders, with excellent tolerability^[35]. These results were especially promising in five PMS patients, all displaying some measurable degree of clinical response^[35].

Based on the evidence summarized above, we have now designed and performed an exploratory 32-week randomized, double-blind, placebo-controlled, cross-over trial (RCT), with (A) the primary aim to evaluate the efficacy and safety of MST in PMS patients and to begin dissecting the selective contribution of CoQ10 to efficacy and tolerability, as compared to vitamins alone; (B) the secondary aim to identify a panel of psychodiagnostic tools endowed with sufficient sensitivity to reliably detect a small-to-moderate effect size in a severely-compromised patient population. This information will then be used to design a confirmatory RCT applying a targeted strategy.

METHODS

Study design

This double-blind, randomized cross-over, placebo-controlled study assessed the efficacy and safety of an MST with CoQ10, vitamin E and complex-B vitamins in PMS. All procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration. The Ethical Committee of the University of Messina (Messina, Italy) approved the study (prot. n. 15/18, approved on May 2nd, 2018). The parents of all patients enrolled in this study were informed about potential adverse reactions. Written informed consent was obtained from all caregivers for their child's participation. This clinical trial adheres to CONSORT guidelines and has been registered at https://clinicaltrials.gov/ with NCT n. 04312152.

The experimental design of this trial is displayed in Figure 1. Briefly, the total duration of the trial was 8 months. Each patient received an active compound [CoQ10 + Vit. E + polyvit. B] for 4 months and active comparator [Vit. E + polyvit. B] for 4 months. Half of the sample received the active compound during period I (months 1-4) and the active comparator during period II (months 5-8), while the remaining half assumed the active comparator during period I and active compound during period II [Figure 1]. More specifically, following a medical screening visit, enrolled patients were double-blindly randomized according to a permuted two-block design, to receive either active compound or active comparator (Period I) and then were switched to the other treatment condition (Period II); each treatment period lasted for 4 months. Initially eligible subjects underwent a complete clinical and psychometric assessment (T0), which was repeated at the end of each 4-month treatment period [T1 and T2 in Figure 1]. Three investigators (L.T., G.C., T.DB.), never involved in clinical assessments nor in direct contact with patients and caregivers, were in charge of treatment allocation. To monitor compliance, unused capsules were retrieved and counted at each time point. One investigator (A.R.), not involved in clinical assessments, interacted with caregivers as needed throughout the study, for any medical need and for issues regarding the trial itself. Families were requested to refrain from commenting on their experience and purported treatment efficacy on social media.

Figure 1. Experimental design of the trial. Patients were assessed at baseline (T0), 4 months (T1), and 8 months (T2), to monitor the safety and efficacy of active compound [CoQ10 + Vit. E+ Polyvit. B] vs. active comparator [Vit. E+ Polyvit. B].

Participants

Inclusion and Exclusion Criteria

Participants were required to hold a diagnosis of PMS due to a genetically documented deletion of human chromosome 22q13.3 or a disruptive mutation in the SHANK3 gene. Other inclusion criteria were: (1) age between 2 and 40 years old; (2) Children's Global Assessment Scale (CGAS) score between 45 and 59 at baseline; (3) medication regimen stable for at least three months prior to enrolment and kept constant throughout the 8-month duration of the trial; (4) behavioral intervention started at least 3 months prior to enrolment and kept unchanged throughout the 8-month duration of the trial; (5) written informed consent provided by both parents or a legally authorized patient representative; (6) parents and guardian able to understand and comply with the experimental protocol.

Potential participants were excluded in the presence of any of the following: (1) known genetic syndromes other than PMS (for example, Rett syndrome, fragile-X syndrome, etc.); (2) serious medical illnesses (chronic renal disease, severe liver disease, cardiovascular disorders, uncontrolled hypertension with systolic pressure values > 170 mm Hg and diastolic pressure > 100 mm Hg, malignant tumors, HIV infection); (3) history of acute cerebrovascular disease; (4) history of stomach bleeding or active peptic ulcer; (5) documented allergies, hypersensitivity or intolerance to one of the excipients of the experimental compound or comparative product; (6) treatment with anticoagulants. Seizures were not a cause of exclusion and were categorized as “frequent” if occurring more often than once every 6 months, “rare” if occurring at a lower frequency, or otherwise “absent”.

Trial interruption criteria include (1) the onset of a severe medical condition during the trial; (2) changes in psychopharmacological or behavioral treatment lasting longer than two weeks during the 8-month duration of the trial; (3) medical conditions requiring anticoagulant treatment started during the trial.

Patient sample

Sixty-six PMS patients were screened for eligibility at the Interdepartmental Program “Autism 0-90” of the “G. Martino” University Hospital in Messina, Italy, between March 2019 and September 2021. The CONSORT flow-chart of the study is depicted in Figure 2. Five patients did not meet inclusion criteria, five already received open treatment with CoQ10/ubiquinol and vitamins^[34], and twenty-one refused participation. The remaining 35 PMS patients were recruited. Comorbid neurodevelopmental disorders were diagnosed based on DSM-5 criteria^[36]. The total sample was stratified by sex and randomly assigned to receive the active compound [CoQ10+VitE+VitB] either during the first or the second administration period. Two patients, one per group, dropped out due to non-compliance; hence a total of 33 PMS patients completed the entire trial and were included in the statistical analysis [Figure 2].

Figure 2. CONSORT flow-chart. Sixty-six PMS patients were assessed for eligibility. Thirty-five of them were recruited and randomized into groups receiving either active compound (n = 18) or active comparator (n = 17) during the first 4 months. Thirty-three patients completed the entire study, whereas one patient dropped out of each treatment group during the first 4-month period.

RESULTS

The study was completed by 33 patients whose demographic, genetic, and clinical characteristics are summarized in Table 4. The sample was randomized to receive the active compound either during the first (n = 18, 51.4%) or during the second administration period (n = 17, 48.6%). Almost all patients displayed Intellectual Disability (ID) or Global Developmental Delay (GDD), either alone (n = 19, 57.6%) or associated with Autism Spectrum Disorder (ASD) (n = 13, 39.4%). All 33 patients also displayed Developmental Motor Coordination Disorder, while 7 (21.2%) have a history of Bipolar Disorder. Severe Language Developmental Disorder was present in 31 (93.9%) patients. Thirteen (42.9%) patients were taking psychopharmacological treatment or nutraceuticals at baseline (T0), which were kept constant throughout the study [Table 4].

Regarding primary outcome measures, several measures reached nominal significance in favor of the active compound [CoQ10 + Vit. E + polyvitamin B] producing greater improvement than the active comparator [Vit. E + polyvitamin B] [Table 5A for summary statistics and Figure 3, depicting single patients]. Specifically, greater positive responses were recorded mainly in:

Figure 3. Single-patient line graphs for primary outcome measures reaching significantly greater improvement with active compound (thick lines) vs. active comparator (thin lines) [see Table 5A]: (A) VABS Motor Skills scores; (B) VAS Enjoyment in shared play; (C) VAS Stereotyped behaviors (parent-rated); (D) VAS Stereotyped behaviors (clinician-rated). The patient legend of panel B also refers to panels C and D. Higher scores in panel A indicate better motor skills; in panels B-D, higher scores indicate greater symptom severity.

(a) Motor skill deficits, with VABS motor skill displaying a sizable improvement (MD +6.81, 95%CI: +1.89; +11.72, P = 0.0147). This statistic includes only 10/33 patients whose age falls within the range for VABS motor skill (0-6 years), but it is paralleled by non-significant improvements in VAS gross- and fine-motor skills collected for all 33 patients, reaching a P = 0.058 in clinician-scored VAS for fine-motor skills. These results are also consistent with clinical observation of widespread small-to-moderate improvement, especially in fine-motor function. This positive effect was surprisingly most evident in three highly compromised patients initially wheelchair-bound due to very large deletions spanning from SHANK3 to TCF20 and beyond: all three became able to hold up and use fork and spoon during meals; for one, it became feasible to sit up and undertake sessions of mild horse-assisted intervention.

(b) "Enjoyment in shared play", i.e., participation in play activities with others, employing exchanges of glances and smiles (VAS, P = 0.0199). This variable exemplifies one of the benefits most consistently reported by parents, namely increased responsiveness to environmental stimulation and greater social motivation;

(c) Repetitive behaviors were significantly reduced according to parents (VAS, P = 0.0198), with a similar trend also observed by clinicians (VAS, P = 0.0770). This observation describes a consistent subgroup of patients and not a small subset because at T0, parents reported the presence of motor stereotypies in 27 (81.8%) patients, and during the intake visit, motor stereotypic behaviors were directly observed by the clinician in 15 (45.5%) patients.

No effect modification by age was detected for enjoyment in shared play (P = 0.2896) and repetitive behaviors (med P = 0.5297, par P = 0.5656), whereas the number of observations for VABS motor skills was not sufficient to allow reliable testing of age effects.

CGI-I scores were consistent with both active compound and active comparator producing overall improvement in a sizable number of cases, with a slight, non-significant advantage for active compound. However, CGI-I scores and parent-rated VAS “Restricted Interests” scores yielded significant Time × Treatment interactions [Table 5B, Figure 4]. In fact, both variables displayed no difference when the active compound was administered during the first four months (Period I), whereas significantly greater efficacy was observed when active compound was administered during Period II, following four months of administration of vitamins E and B complex [Table 5B, Figure 4]. Finally, secondary outcome measures revealed a very significant positive influence of active compound on parental quality of life, with over four points of mean increase in maternal WHOQOL scores (MD +4.72, 95%CI: +1.34; +8.11, P = 0.0079), and paternal scores displaying a similar trend (MD +3.61, 95%CI: -0.40; +7.63, P = 0.0755). None of the single WHOQOL items reached statistical significance. Subjective parental reports describe their child as becoming more alert, more responsive to parental requests possibly due to improved receptive language skills, displaying a better mood and greater social motivation, and overall better adapted to family life and in school.

Figure 4. Time x Treatment interaction for changes in (A) CGI-Improvement rated by clinicians; and (B) VAS “Restricted interests” rated by parents. For both variables, active compound produces significantly greater improvement when administered during Period II, but not during Period I. CGI-I scores: 1 = Very much improved, 2 = Much improved, 3 = Minimally improved, 4 = No change, 5 = Minimally worse, 6 = Very much worse.*P< 0.05.

Side effects were rare, minor, transient and did not significantly differ between active compound and active comparator [Table 6]. The most frequent were hyperactivity and sleep disturbances, irritability, and increased appetite. Among six patients with a history of seizures [Table 4], none displayed a clinically meaningful change in seizure frequency and/or severity during either treatment period compared to baseline. No serious adverse event occurred and no participant had to stop treatment because of intolerable side effects.

DISCUSSION

This is the first clinical trial on the effects of an MST, encompassing CoQ10, Vitamin E, and Polyvitamin B, in Phelan-McDermid Syndrome (PMS). In addition to being a relatively frequent cause of intellectual disability, PMS represents one of the most important syndromic forms of ASD. In fact, our sample includes 13 (39.4%) patients who also satisfy DSM-5 diagnostic criteria for ASD, in agreement with the 30%-80% range reported in the Literature^[6-8]. The present study represents a timely follow-up of our recent retrospective chart review, documenting the potential efficacy of MST in a variety of neurodevelopmental disorders, and includes five PMS patients who, surprisingly, all display some degree of clinical improvement^[35]. Despite the exploratory nature of this RCT, the comparison between full MST vs. "vitamins only" provides further evidence pointing towards the beneficial effects of CoQ10 as part of MST in PMS. Clinical improvement in our PMS cohort appears to mostly occur in the domains of motor skills, cognition, responsiveness to environmental stimuli (adaptive functioning), and stereotypic behaviors [Table 5]. Given the severity and pathophysiology of this genetic syndrome and of neurodevelopmental disorders in general, it is not surprising that most responders to MST appear "minimally improved", with only some patients occasionally appearing "much improved". We did not detect significant effects on language and communication skills, in contrast to our retrospective study^[35]. This discrepancy may reflect the more severe language impairment present in PMS compared to other neurodevelopmental disorders, with its documented structural underpinnings^[11,56]. Importantly, the positive influence recorded here on parental quality of life with MST compared to vitamins only, especially for mothers, provides converging support to MST efficacy from a different point of observation. Finally, the extreme safety of this therapy is documented by the rare, mild, and easily manageable side effects reported by parents, equally distributed between the active compound and active comparator [Table 6].

Mitochondrial function plays a pivotal role in brain development and central nervous system (CNS) function. Most brain energy supply is generated by mitochondrial oxidative phosphorylation, and each different neuronal cell type presents local mitochondrial adaptation processes reflecting functional heterogeneity^[57]. In addition to energy production, mitochondria are involved in neurogenesis and apoptosis, neuronal migration and differentiation, and neuroplasticity by regulating redox homeostasis, protein and lipid mediators, and intracellular calcium levels^[58]. Therefore, it is not surprising that mitochondrial dysfunction and oxidative stress appear to be involved in the pathogenesis of many neurodevelopmental disorders and that antioxidants produce clinical improvement in many patients^[59-62]. More specifically, Autism Spectrum Disorder (ASD) is frequently diagnosed in PMS patients and often represents their first clinical diagnosis, formulated long before genetic testing^[6-8]. ASD represents a wide collection of multiple genetic and/or epigenetic disorders (the "autisms"), sharing socio-communication deficits, repetitive behaviors, insistence on sameness and abnormal sensory processing as their clinical expression end-point^[63]. Enhanced oxidative stress and mitochondrial dysfunction represent one of the most replicated abnormalities detected both systemically and in the Central Nervous System (CNS) of autistic individuals^[64]. Targeted and untargeted studies of post-mortem brains, plasma, CSF, urines, and peripheral blood mononuclear cells (PBMCs) consistently documented significant oxidative damage to proteins, lipids and nucleotides, reductions in GSH/GSSG ratio, reduced activity levels of redox protection enzymes, including superoxide dismutase 2 (SOD2), glutathione peroxidase (GPx), glutathione-S-transferase (GST), and glutamate-cysteine ligase, reduced activity of respiratory chain complex I and complex IV due to mitochondrial oxidative damage^[65-73]. Abnormalities in these redox parameters have been found to be significantly correlated with the severity of autistic behaviors^[74,75] and gastrointestinal symptoms^[76,77], which are frequent among autistic children^[78]. Not surprisingly, the activity of complex I and IV is significantly reduced in as many as 59% of PMS patients, whose clinical features often overlap with those of mitochondrial diseases^[79]. In addition to non-specific impairment of energy metabolism, PMS-causing chr. 22q11.3 deletions often span genes playing relevant roles in mitochondrial function^[13]. Although oxidative stress response may represent an evolutionarily preserved strategy aimed at boosting resilience^[80] and usually represents the consequence and not the primary cause of ASD-associated genetic syndromes^[65], reduced ATP production and excessive oxidative damage can potentially exacerbate the dysfunction caused directly by ASD-causing genetic or epigenetic defects, thereby imposing an additional burden. Importantly, redox abnormalities have also been detected in young autistic children and are not correlated with age^[69]. Therefore, enhanced oxidative stress and mitochondrial dysfunction represent a "trait-dependent" characteristic present in a consistent number of PMS patients and autistic individuals, regardless of their age and their specific underlying pathogenetic underpinnings. Since it is not yet possible to directly correct the genetic alteration underlying PMS or its major downstream consequences, sustaining mitochondrial function while controlling redox imbalance may represent a viable indirect therapeutic approach, potentially able to ameliorate cognition and adaptive functioning, motor skills and stereotypic behaviors in many patients, albeit with small-to-moderate effects and with large interindividual differences.

This is the first clinical trial to assess the efficacy and safety of CoQ10, vit. E and polyvit. B in PMS. In reference to ASD, three RCTs have evaluated the effects of over 30 supplements including CoQ10, multiple vitamins, minerals such as lithium, N-acetylcysteine, choline, acetyl-L-carnitine and inositol in 20, 55, and 67 autistic children and adults, respectively^[81-83]. Several blood biomarkers were measured, including nutritional (vitamins, minerals, and aminoacids) and metabolic parameters (oxidative stress, methylation, sulfation, glutathione, and neurotransmitters). At baseline, the ASD sample showed significantly higher levels of oxidative stress compared to typically developing controls. After 3 months of treatment, these parameters were either improved or even normalized. Additionally, a behavioral improvement was recorded for several symptoms, as well as in mean Parental Global Impressions-Revised scores, especially for the hyperactivity, oppositional and receptive language subscales^[82], and nonverbal intellectual ability^[83]. These domains largely overlap with the areas of greatest improvement reported in our retrospective chart review of neurodevelopmental patients^[35]. However, the administration of over thirty active compounds hampers the possibility of detecting with reasonable certainty specific contributions by CoQ10 to the observed improvement in biochemical and behavioral parameters. In contrast, more focused RCTs demonstrated a sizable reduction in autistic symptoms following 30 wks of high-dose Vitamin C (110 mg/kg)^[84] or after high-dose Vitamin B6 (500-1,000 mg)^[85]. To our knowledge, only one RCT has selectively supplemented 78 ASD children with placebo, 30 or 60 mg/day of CoQ10 (n = 26 per sample), recording with the higher CoQ10 dose an improvement in several peripheral redox parameters, paired with a mild reduction ASD severity, sleep disorders and gastrointestinal symptoms^[86]. This RCT differs from the present study in experimental design (case-control vs. cross-over), duration (3 mo vs. 4 mo of active compound/comparator), CoQ10 dosage (not based on body weight or age in ref. 86), and in the composition of active compound and comparator (inactive placebo in ref. 86, vitamins E and B here). Nonetheless, its clinical outcome appears convergent with our past^[35] and present observations.

From a methodological standpoint, multiple features of this RCT define its exploratory nature. On the one hand, they may represent the limitations of the present study, but on the other hand, they will enable investigators to precisely design targeted confirmatory trials in neurodevelopmental disorders. These points can be summarized as follows:

(A) Initially, this study was undertaken assuming that vitamins would have been entirely inactive on the severe clinical picture of PMS. However, this turned out not to be the case. On the one hand, the Time x Treatment interactions observed on CGI-I and VAS “Restricted Interests” scores demonstrate that in many patients, a four-month course of vitamin E and B seemingly paves the path to a greater response to CoQ10 [Figure 4]. On the other hand, a slight but detectable amelioration was recorded after the administration of vitamin E and polyvitamin B to most patients [Figure 4A]. In some cases, the convergent observation by parents, clinicians, and therapists blind to the treatment of a "much improved" therapeutic effect by vitamins E and B is unlikely to represent a placebo effect in a severe syndrome such as PMS. For the above-mentioned reasons, contrary to its initial expectation, this study has become an "active comparator-controlled", randomized, double-blind, cross-over trial, as it would be inaccurate to define it as "placebo-controlled"^[87].

(B) It is not customary to apply as many as four different primary outcome measures in an RCT. This strategy was chosen to verify which outcome measures would display the greatest sensitivity in detecting clinical change over time. The low sensitivity of psychodiagnostic scales and questionnaires in detecting clinically meaningful changes not large enough to bring patients outside the ASD spectrum has been one of the major hurdles encountered in experimental pharmacology of genetic syndromes such as Fragile X^[39]. Based on our results, clinician-rated CGI-I and parent-rated VAS appear to be the most sensitive measures, followed by the VABS. For example, the outcome of RBS-R and VAS for stereotypic behaviors appears inconsistent. However, mean pre-/post-treatment changes in VAS for stereotypic behaviors were -0.79 and -0.97 for parent- and clinician-rated scores, respectively [Table 5]. VAS ranges from 0 to 10, whereas RBS-R is essentially a Likert scale ranging from 0 to 3. It is not surprising that the RBS-R may not be sufficiently sensitive to change in this context. Nonetheless, care must be placed in ensuring that CGI-I and VAS be administered by the same clinician throughout the study to ensure reliability (notice that also "parent-rated VAS" were explained and presented to parents by the same clinician and were not self-administered). The CARS instead failed to detect changes recorded by these other scales, likely due to difficulty in ensuring reliability at 4-month time intervals even when administered by the same clinician. In reference to secondary measures, we find especially promising that maternal quality of life improves by an average of +4.72 points at the WHOQOL, reaching the highest nominal P-value recorded in this study (P = 0.0079). This improvement with CoQ10 over vitamins alone is confirmed by the similar trend observed in fathers (see Results). This effect size is not at all marginal, considering that a 6-point improvement in maternal WHOQOL scores is produced by the complete normalization of child sleep patterns (Persico AM, Cucinotta F, et al., manuscript in preparation).

(C) Placebo effects in behavioral pharmacology can be large enough to obscure the therapeutic effects exerted by the active compound. This is another critical issue, considering that the therapeutic effect size of MST can be realistically expected in the small-to-moderate range^[35]. Several strategies potentially able to minimize the size of placebo effects were applied. Clinicians involved in medical visits and psychodiagnostic testing were screened from interactions with parents outside of T0, T1, and T2. Throughout the study, families were free to contact a dedicated investigator not involved in experimental measures whenever necessary. Using this strategy, the number of contacts between families and experimenters was nil or limited to extremely rare and pre-filtered major issues. Participating families were also generally compliant with our request not to comment on the study on social media. Overall, these strategies appeared to yield satisfactory results, although the lack of an inactive placebo here does not allow us to reliably quantify the placebo effect size and estimate their efficacy.

(D) The duration of RCTs is often dictated more by the available budget and other practical considerations than by preliminary efficacy and safety data. The Time x Treatment interactions depicted in Figure 4 suggest that, while 4 months represent a time interval of sufficient duration to produce therapeutic effects, a slightly longer trial duration, if possible, may yield an even greater therapeutic effect size and/or improvements in more areas.

(E) Cross-over studies typically include a one-week wash-out period between the "active compound" and the "active comparator" periods [Figure 1]. This strategy was not applied here under the hypothesis that neurodevelopmental plasticity changes, and not acute pharmacological effects, would underlie behavioral changes. Neuroplastic effects should be expected to carry over for weeks and may be more reliably detected by searching for differential effects, depending on whether the active compound is administered during the first or the second period, rather than expecting one week of wash-out to truly bring all CNS parameters back to baseline. Our results provide some support to this line of thought; nonetheless, a one-week wash-out period could be applied in future studies.

(F) Our sample size was dictated by the availability of cases affected with a rare disorder, like PMS. It was sufficient to yield nominal significance for several relevant variables but not enough for a superiority trial design, given the small-to-moderate therapeutic effect (i.e., 0.2-0.3). Future studies should benefit from a greater gap in efficacy between MST and a truly inactive placebo, but given these premises, a sample size of n = 50 appears to be the minimum requirement for a cross-over study of MST. Multicenter studies may be necessary if single rare disorders are targeted.

(G) The vast majority of patients recruited for this study were unable to swallow capsules. In these patients, both active compound and active comparator were administered by opening capsules and dissolving their content in a small quantity of juice or soft drink. We allowed this procedure, because we were indeed aware that severe swallowing difficulties are frequent in PMS patients^[88] and also because, in our retrospective study^[35], we did not observe a loss of efficacy following this protocol. Nonetheless, we cannot entirely exclude that differences in drug administration strategy may have contributed to the interindividual variability in response recorded in the present trial.

In summary, this study provides further promising evidence of the positive effects of CoQ10 in PMS when administered in association with Vitamin E and Polyvitamin B, which also appear to provide independent synergistic contributions. Mild-to-moderate improvement was recorded in 24/31 (77.4%) PMS patients and concerned primarily with the domains of cognition, responsiveness to environmental stimuli and adaptive functioning, motor skills, and stereotypic behaviors. In addition, the low incidence and mildness of side effects encourage further studies of this therapeutic approach, given the severity of PMS and the lack of available treatments. From a different perspective, the small but tangible relief provided to parental quality of life by the addition of CoQ10 to vitamins appears to be an important added value in a severe genetic syndrome with a major impact on the daily living of family members^[89] and on parental stress^[90]. Defining the most sensitive and reliable outcome measures represents a major hurdle in the experimental psychopharmacology of neurodevelopmental disorders^[91]. This trial has been instrumental in defining the outcome measures most sensitive to small-to-moderate clinical change in this severely-affected population, the optimal duration, and many other methodological aspects of the experimental design, setting the stage for confirmatory targeted RCTs for PMS, ASD, and other neurodevelopmental genetic syndromes. These studies will also benefit from the assessment of biochemical parameters of oxidative stress to explore whether and to what extent they predict clinical response.