Participants and design
The participants were 7 to 12 years of age with an IQ between 80 and 120, measured using the Wechsler Intelligence Scale for Children (WISC IIIR or WISC IV). A child neurologist, a child psychiatrist and a psychologist enrolled children in two clinical sites, in Lyon and Paris, if they met these inclusion criteria:
1.
NF1 diagnosed using two or more of the following National Institutes of Health (NIH) criteria [
22]: (1) six or more café au lait macules over 5 mm at the greatest diameter in prepubertal children and over 15 mm in postpubertal children, (2) two or more neurofibromas of any type or one plexiform neurofibroma, (3) freckling in the axillary or inguinal regions, (4) two or more Lisch nodules, (5) optic glioma, (6) a distinctive bone lesion such as sphenoid dysplasia or thinning of long bone cortex with or without pseudarthrosis, or (7) a first-degree relative with NF1 as defined by the above criteria;
2.
School difficulties pointed out by parents or teachers;
3.
Attention difficulties as defined by anamnesis.
We had the following exclusion criteria:
1.
80 < IQ > 120 measured using the WISC III or WISC IV;
3.
Unwillingness to participate;
4.
Patients with NF1 cerebral complications (chiasma tumor, moya-moya, cerebral glioma) detected using MRI;
5.
Participation in another interventional study.
The NF1 Attention Study was a randomised, double-blind, placebo-controlled, crossover trial. Total follow-up lasted 9 weeks including 4 weeks for each period and a 1-week wash out period.
The main outcome was the improvement of the simplified Conners’ Parent Rating Scale [
23] because it is largely used in children and a French version was available.
The short version comprises 10 questions and provides evaluation of the key areas of inattention, hyperactivity/impulsivity, learning problems, behavioural disorders, anxiety, and peer relations. Specifically developed for detecting hyperactivity, it allows quantifying the intensity of hyperactivity and assessing its various dimensions: hyperactivity, inattention and impulsivity. The score varies from 0 to 30, scores >15 are prognostic of hyperactivity.
The secondary outcomes were the improvement of scores on the Conners’ Teacher Rating Scale. The short version comprises 10 questions provides evaluation of the key areas of (behavioural disorders, hyperactivity/impulsivity, immaturity and passivity). The interpretation of scores is similar to Conner’s Parent Rating Scale.
The Children’s Depression Rating scale (CDRs-R) [
24], the Children’s Depression Inventory (CDI) [
25], and the State-trait Anxiety Inventory for Children (STAIc) [
26] were used in order to monitor anxiety and depression symptoms of participants during the trial.
CDRs-R is a 17-item measure used to determine the severity of depression in children 6–12 years of age. 14 items are based on parent, child and schoolteacher interviews, 3 items from the direct observation of children (depression affects, language time, and hypoactivity). Items are measured on 7 points or 5 points scale. The CDRS is derived from the Hamilton Rating Scale for Depression (HAM-D); a score of 15 on the CDRS is equivalent to a score of 0 on the HAM-D. A score ≥40 indicates depression but is not enough to confirm the diagnosis. A score between 40 and 60 characterize light or moderate and scores > 60 severe depression. Score < 30 are normal.
The CDI is adapted from Beck inventory depression to determine the severity of depression in children for children 7 to 17. It’s a psychological assessment that rates the severity of symptoms related to depression and/or dysthymic disorder in children and adolescents, is a 27-item scale that is self-rated and symptom-oriented. Each item is score from 0 to 2. The higher the score, the higher the severity of depression. A score above 15 indicates depression.
The STAIC consists of two 20-item scales that measure state and trait anxiety in children between the ages of 8 and 14. The A-State scale examines the shorter-term state anxiety that is commonly specific to situations and the trait anxiety measures general anxiety. It prompts the child to rate 20 statements from hardly ever true (1point) to often true (3 points). Scores vary from 20 to 60. Score > 34 indicate anxiety.
Randomisation
Patients were enrolled at the Department of Paediatric Neurology (Lyon and Paris Trousseau teaching hospitals). Two physicians, a child neurologist, and a child psychiatrist, asked for Children’s assent and informed consent from both parents. The identity of eligible patients was transmitted by the Department of Paediatric Neurology to the Clinical Investigation Centre (CIC) of Lyon and the MPD prescription was faxed to the central pharmacy where a masked randomisation list according to a computer generated randomization allowing concealed allocation was available. Randomisation was performed by a computer generated random number list prepared by the department of biostatistics of the coordination center with no clinical involvement in the trial. The random list was created using SAS (version 8.2) statistical software with a 1:1 allocation using block size of 4.
Patients received their treatment directly from the central pharmacy. Neither patients nor investigators knew which treatment was given.
The CIC and the Department of Biostatistics of Lyon Teaching Hospitals, Lyon, France, conducted data management and statistical analyses. All data were recorder in case report forms and entered by two independent technicians blindly in a central secured database under the responsibility of Clininfo. Site monitoring was carried out by the CIC. A neurofibromatosis association and the regional healthcare network for neurofibromatosis informed potential eligible patients about the trial.
Intervention
MPD or matched placebo was started at a dose of 0.5 mg/kg/day and increased to 0.8 mg/kg/day if symptoms did not improve after a one-week check-up [
13]. There was a one-week washout between the two treatment periods to avoid a carry-over effect due to MPD’s short half-life (2 hours) [
24]. Treatment adherence was measured by counting returned pills [
25]. Six visits were organized for the inclusion of eligible participants, randomisation, dose-adjustment at week 1, end of first period at week 4, debut of the second period at week 5, the dose-adjustment at week 6, and end of study visit at week 9. Efficacy measures and adverse events were collected by the investigators.
The efficacy of MPD on ADHD disorders has been proven using a minimum dose of 0.3 mg/kg/day. According to the guidelines, the patient should not be given more than 2 or 3 doses of 1 mg/kg/day. Treatment should begin with 5 mg, 2 times per day. The dose should be gradually increased by 5-10 mg per week, without exceeding the maximum dose of 60 mg per day (source: leaflet from VIDAL).
Statistical analysis
The efficacy analyses adhered to the intention-to-treat approach regardless of treatment status at the time of analysis. To compare efficacy between MPD and placebo, we performed a two-factor analysis of variance (ANOVA for repeated measures) on the 2 N measurements with period and treatment factors (MPD versus placebo). Interactions between patient characteristics and treatment effect were also examined. Missing data were replaced only when less than 50% of answers were missing to compute a score. Imputation was performed for each dimension (multidimensional scales) or directly (one-dimensional scale) with the mean for completed items [
26]. The global score was then computed as the sum of scores for each dimension (multidimensional scales) or directly (one-dimensional scale).
The results presented in the tables are based on raw data (score variations) but the p-values in the text were calculated using non-parametric ANOVA (ANOVA performed on normal scores from ranks computed with the Van der Waerden method) [
27]. Data from each patient were used to estimate the treatment effect. Missing measurements (7/78) were considered as missing by chance, with the same mean as non-missing measurements. None of the patients were excluded from the analysis.