Study & Participants
All visual investigations were part of an exploratory clinical study named CLIBIOMAR FXS. This exploratory study of clinical and biological markers focused on the visual phenotype as assessed by ERG using ISCEV (International Society of Clinical Electrophysiology of Vision) recommendations, contrast sensitivity test using the LEA SYMBOLS® low-contrast test and Short Sensory Profile (SSP) questionnaire (French version) for other neurosensorial abnormalities. A total of 20 FXS subjects (male, 18-45 years of age), with a previously confirmed molecular genetic diagnosis of the full Fragile X Mental Retardation mutation (≥200 CGG repetitions, fully methylated) (Table 1), and 20 male age-matched healthy control subjects were targeted to be enrolled in the study (Table 2). FXS subjects were identified from existing clinical populations on hospital files (Genetics Laboratory of Orléans Regional Hospital Center (Orléans, France), Biochemistry and Genetics Department of Angers University Hospital Center (Angers, France), Clinical Genetics Department of Rennes University Hospital Center (Rennes, France), and two French FXS patient organizations (Mosaïque and Fragile X France, France) and invited to participate. Control subjects, recruited as a separate cohort through the Eurofins OPTIMED (Grenoble, France) database, were non-smokers and considered as healthy based on a comprehensive clinical assessment including history and a normal clinical examination. Participants in either group were excluded if there was a family history of ocular disease, strabismus, or any history of epileptic seizures in the last year, any history of head/brain trauma or pathology. To be included, control subjects should be compliant to ophthalmologic recordings. The level of ID was not an eligibility criterion for FXS subjects but a study investigator (SB) screened FXS volunteers for their ability to be neurobehaviorally compliant to ophthalmologic recordings.
Ethics approval and consent to participate
Written informed consent was obtained from control subjects. For subjects with FXS, the informed consent was provided by the legal representative (parents in all cases) prior to any study procedure. Eligibility of FXS subjects for the study included company of a caregiver able to answer to the behavioral scale questionnaire and having a high probability for compliance with all examinations and completion of the study. The study adhered to the tenets of the Declaration of Helsinki and were performed in accordance to the protocols approved by the institutional review board (CPP Est 1, Dijon, France), along with proper notifications made to the Agence Nationale de Sécurité du Médicament et des Produits de Santé (The French National Agency for Medicines and Health Products Safety or ANSM) prior to the conduct of the study.
Procedures involving healthy controls were conducted at Eurofins OPTIMED’s clinical pharmacology unit (Grenoble, France), while procedures involving FXS subjects were conducted in the subject’s home in the presence of their primary caregiver (e.g., parent, family member) by the same investigators (SB, FL). After arrival at the FXS subject’s home, time was spent for discussion with the patient and the caregiver, study equipment was shown to build trust and confidence to ensure a smooth, stress-free conduct of the study. For FXS subjects, evaluations were initiated with the administration of the SSP questionnaire (see below) with the primary caregiver together with the FXS subject, whenever possible. Thereafter, after light-adaptation, ERG was assessed for both FXS and HVs using the RETeval® device (described below). Under LA conditions, ERG recordings began first with the single flash protocol, then followed by recordings under the LA-flicker protocol. After a 20-minute rest period, contrast sensitivity testing was performed using the LEA SYMBOLS® low-contrast test (described below). ERG and CS testing were performed by the same investigators for both cohorts. The behavioral status of FXS subjects was assessed with the Aberrant Behavior Checklist – Community in FXS: ABC-CFX scale  and the SSP and ABC-CFX scoring was done only for the FXS cohort.
Short Sensory Profile (SSP)
The Short sensory Profile (SSP) is a standardized questionnaire which permits to clinicians and researchers to quickly gather information about sensory processing problems that interfere with functional performance in children [54-56]. It can be used to signal a potential difference in children’s responses and behaviors to commonly occurring sensory events as compared to children without disability. The 38 items of the SSP are extracted from the 125 item long version of the Sensory Profile, which is based on factor analyses and correlation studies from two samples of 117 and 1037 children with a variety of problems . The SSP consists of 7 sections: (1) Tactile Sensitivity, (7 items, maximal sub-score 35), (2) Taste / Smell Sensitivity (4 items, maximal sub-score 20), (3) Movement Sensitivity (3 items, maximal sub-score 15), (4) Under-responsive / Seeks Sensation (7 items, maximal sub-score 35), (5) Auditory Filtering (6 items, maximal sub-score 30), (6) Low Energy / Weak (6 items, maximal sub-score 30), and (7) Visual / Auditory Sensitivity (5 items, maximal sub-score 25). Sub-scores from all 7 sections are added to obtain a total score and together describes the subject`s overall sensory profile. The profile is compared against a control database comprised of a national sample of children without disabilities and allows a comparison to determine if the score is likely different from the control population. For each section, the sub-score belonged to a classification called “typical performance”, “probable difference” or “definite difference” as compared to the control database sample.
ERGs were recorded in accordance with standards established by the International Society of Clinical Electrophysiology of Vision (ISCEV) under light-adapted (LA) conditions . Note that in standard ERG recording sessions, ERGs are obtained in both scotopic (dark-adapted) and photopic (light-adapted) conditions. In a pilot effort to see if scotopic recordings can be performed in FXS individuals, high anxiety was generated by 20 minutes of dark-adaptation and thus scotopic ERG was not recorded on FXS subjects. LA-ERG recordings were obtained using the FDA approved RETeval® device (LKC Technologies Inc, Gaithersburg, MD, USA), a hand-held, portable stimulus and recording instrument designed for performing ERGs in pediatric individuals. The RETeval® device has been validated in several studies generating results similar to the classical ERG device [58-60]. The instrument contains a normative reference range of values for ERG parameters [61, 62] and is ideally suited for the FXS study cohort with less compliant intellectually deficient individuals. Use of a single sticker recording electrode placed below the eye without requiring mydriasis greatly facilitated the challenges often encountered in clinical practice where corneal electrodes and mydriasis are required in classical ERG recordings . ERG responses were digitally recorded from a self-adhesive skin electrode positioned below the lower eyelid and protocols having constant retinal luminance without pupil dilation, which are described by the Troland unit (Td). In these protocols, the RETeval® device measures the pupil size in real time and continuously adjusts the flash luminance to deliver the targeted amount of light into the eye, regardless of the size of the pupil and according to the following formula: Troland = (pupil area in mm2) (luminance in cd/m2). It has been previously demonstrated that pupils do not need to be dilated to achieve consistent results [61, 64].
The participant was seated, and the skin electrode was placed 1 – 4 mm below the lower eyelid following skin preparation if required to reduce impedance to < 5 kΩ in accordance with the manufacturer’s instructions. The vertical and horizontal electrode position were recorded with an in-built infrared digital camera. Photographic images of each eye allowed post-analysis of electrode placement and fixation using a calibrated graticule. Electrode position can affect the amplitude of the ERG signal  and eyes for which the electrode position was > 4 mm were excluded. All images were inspected, and the position of the electrode measured with a weighting of 2 mm below the eye set at zero in the statistical model. Since iris color can affect the ERG response, iris color index was used to weight the amplitudes according to iris pigmentation which can reduce the b- wave amplitude in heavily pigmented individuals . Then, the participant was instructed to look steadily at a dim red LED located in the center of the Ganzfeld dome and to try to avoid blinking or eye movements. All recordings were performed under normal room lighting conditions, controlled by the RETeval® device. Retinal physiology was assessed using the Troland based ISCEV standard full-field white flash 3.0 cd s m−² (85 Td.s) on a 30 cd s m−² white background luminance (848 Td.s) at 2/s intervals made with 30 flashes averaged to generate the ‘single flash’ ERG waveform. This was followed by a 28.3 Hz series of repeated flashes (background luminance 848 Td) to generate the ‘flicker’ ERG. The right eye was always recorded first. Repeats of the recording were performed as required. ERG recordings typically lasted approximately 10 to 15 minutes to complete both eyes. Recordings were automatically stopped by the device if pupil tracking was lost (poor fixation), electrode impedance was > 5 kΩ (electrode placed improperly or came unstuck), or if pupil diameter was too small for the Troland protocol to provide the required flash strength for any specific retinal illuminance. In these cases, recordings were repeated up to two more attempts for each eye.
For this study the peak amplitudes and timings of LA-ERG a- and b-waves for the ISCEV standard light-adapted 3.0 flash (LA 3 cd s m−2, 2 Hz) and the 28.3 Hz flicker (LA 3 cd s m−2, 28.3 Hz) were compared. The ERG amplitude and time of the a- and b-waves were reported automatically by an in-built algorithm and checked manually for accurate placement. If the a-wave amplitude was < 1 µV the time and amplitude was ignored for that waveform. When repeated measurements were taken, the waveform with the largest b-wave amplitude was included. Raw data, video and image of the electrode on the eye and iris color index were all exported for analysis using the RFF extractor version 188.8.131.52 (LKC Technologies Inc, Gaithersburg, MD, USA).
Contrast Sensitivity Assessment using the LEA SYMBOLS® low-contrast test
Contrast sensitivity (CS) was assessed with the LEA SYMBOLS® low-contrast test, which assesses an individual’s ability to discriminate between symbols (e.g., square, circle, house, apple) printed at a fixed size (10M) onto flashcards at 5 sequentially decreasing contrast levels (25 %, 10 %, 5 %, 2.5 %, 1.25 %), and measured at three different distances (1, 3, and 5 meters). This test allows detection of contrast sensitivity in a population with potential language and/or cognitive deficits commonly observed in patients with neurodevelopmental disorders like in FXS patients [67-69]. To reduce known effects of luminance variation on threshold values, all tests were performed under controlled lighting conditions of 120 lux verified by measurement with a luminometer (Data Recorder PCE-VDL16l® device) for all distances assessed. Prior to the start of the test investigators assured that FXS subjects were able to identify the test symbols correctly using a flashcard with 100 % contrast level. Two investigators were required to perform the test. One investigator stood at the specified distance in front of the subject with the flashcard and pointed to different symbols. Contrast cards were presented from high to low. The second investigator stood behind the test subject assuring that the subject’s vision was not disturbed (e.g., by light reflection of the flashcards) noting the score of the test subject at each contrast level evaluated. The evaluation was done for each eye separately while the other eye was covered by hand. Decreasing contrast levels were tested first at a distance of 1 meter for the left eye while one hand covered the right eye. The procedure was repeated for the right eye while one hand covered the left eye. Thereafter, the test distance was increased to 3 meters and the sequence started again. Ultimately the distance was extended to 5 meters. At each contrast level, at least 3 symbols on the flashcard were pointed to the subject. In case of successful identification of all test symbols, a score of “5” was noted prior moving on to the next lower contrast level. If the test subject failed to correctly name one or two symbols, a score of "4" or "3", respectively was noted. If a test subject failed 3 times at a contrast level, the test was stopped, either the other eye was assessed or proceeded to the next distance. The result of the evaluation is the score of correct answers for each contrast level at each distance (score between 0 and 5). A total success score was calculated based on the sum of averaged values for each subject across contrast levels at the three distances, thus providing a more global assessment of CS (maximum score of 25). Two scores were calculated for subjects during each assessment. First, the average number of correctly identified symbols between both eyes of each subject was recorded to provide a total success score for each contrast level and at each distance (each with a maximum score of 5.0).
Analysis of descriptive and inferential statistics for ERG and CS assessments were performed using SAS® version V9.4 (Cary, NC, USA). Due to the exploratory focus of this study, point estimate (nominal) p-values from hypothesis tests and confidence intervals (CIs) obtained in the statistical analyses were not intended to make inferences, but rather to flag potential patterns and quantify the uncertainty of estimation (total variance). Imputation methods were not used in this study to replace missing data. No power calculation was performed as this was an exploratory study. Comparison between individuals with FXS and matched controls were performed using an analysis of variance (ANOVA), or covariance (ANCOVA) in cases where adjustments were made because of quantitative factors (e.g., adjustment on age factor for ERG analysis). The analysis consisted of the comparison of the mean parameters recorded in FXS and control subjects at 95 % CIs. Statistical tests were performed two-sided with an alpha risk fixed at 5 % (except for the Shapiro-Wilk test with an alpha risk set to 1%). Where residuals from the model lack normal distribution (Shapiro-Wilk test with an alpha risk of 1%), statistical tests were applied on ln-transformed or rank data. For models including interaction terms (i.e., interaction between two model factors), study group effect was analyzed for each factor level separately in case of interaction statistically significant.
Each eye (left and right) of a test subject served as an independent data source. ERG recordings were analyzed and compared between FXS and the age-matched healthy control group. When duplicate measures were available for ERG for each eye at a specific time, the ERG with the highest b-wave was used as the ERG parameter for the analysis. Only ERGs whose vertical height was below or equal to 4 mm were kept for the analysis. For each subject, there are 4 ERG measures, ISCEV LA flash 85 Td. for both eyes and LA flicker 28.3 Hz from both eyes. Guided by the animal model results [35, 40], both b-wave single flash and flicker amplitude measurements were considered primary outcome parameters; secondary outcomes were the time to peak of the b-wave, the a-wave amplitude and time to peak, the ratio of the b-wave amplitude to the a-wave amplitude (b:a ratio), amplitude and flicker implicit time (time to peak). The comparison between FXS subjects versus age-matched controls was made using a model including subject as random effect and study group, age, iris index (when appropriate) and electrode distance (vertical distance from lower end of the eye and the electrode (when appropriate) as fixed and interactions. Finally, to evaluate for potential influence of medications on the ERG waveform, a further sub-analysis was performed on the b-wave amplitude using data excluding those subjects on CNS active medications compared to the matched control set. It is known that some CNS active medications can alter the ERG waveform response.
Data from contrast sensitivity testing were analyzed by an ANOVA performed on success rate, i.e., the score obtained on the LEA SYMBOLS® low-contrast sensitivity test at each distance (1, 3 and 5 m) and each contrast (1.25, 2.5, 5, 10 & 25 %) for each eye using study group (FXS versus matched control), distance, contrast and interaction terms as fixed factors. The model included study group (FXS/matched control), eye (right and left), distance, contrast and interaction terms as fixed factors and subject as random effect. In case of significant interaction, the study group effect was assessed separately for each level of the factor included in the corresponding interaction. As a secondary analysis, this approach was replicated considering the FXS subjects not on CNS active medication only versus healthy age-matched control healthy subjects.