Commentary: RCT of optimal dose methylphenidate in children and adolescents with severe ADHD and ID - a reply to Arnold (2013)



Emily Simonoff, Department of Child and Adolescent Psychiatry, King's College London, Institute of Psychiatry, DeCrespigny Park London SE5 8AF, UK; Email:


The commentary by Arnold (2013) raises a number of interesting issues and speculations about the action of methylphenidate in children with intellectual disability (ID) and associated neurodevelopmental disorders, such as autism spectrum disorders. In our article (Simonoff et al., 2013), we were careful to stick closely to the statistical analysis plan drawn up (and approved by the Data Monitoring Committee) during data collection and prior to any exploratory analysis. However, a number of the issues raised by Arnold warrant further response, with the aim of clarifying the lessons that can reasonably be drawn for clinical practice.

Arnold highlights that the rate of intolerable adverse effects was lower in our trial than that of the Research Units on Pediatric Psychopharmacology (RUPP) Autism Network (2005) and speculates whether this is due to the lower rate of autism spectrum disorders (ASD) in our trial or to the different method of dose titration used in the two trials. We observe that all five participants who withdrew because of intolerable adverse effects (all of whom were on methylphenidate) had scores on the Social Communication Questionnaire (SCQ; Rutter, Bailey, & Lord, 2003) above the suggested cut-off of 15 for pervasive developmental disorder, or ASD. Furthermore, these five children are amongst 32 with SCQ scores ≥15 who received methylphenidate, i.e. 16% of those with SCQ scores in the ASD range receiving methylphenidate had intolerable adverse effects, a rate very similar to the 18% described by RUPP. This suggests that different titration regimes contributed little to the occurrence of adverse effects.

In the present study, structured assessments for ASD were not undertaken and we noted that many participating children who demonstrated clinical features of autism did not, by parental report, have a clinical diagnosis of ASD. We include further information on the proportion of children scoring above the two SCQ cut-offs of 15 and 21 and above for ASD and autism, respectively, in Table 1 below. These findings are consistent with those from a previous UK epidemiological study (SNAP) of the prevalence of ASD, which found that only 58% of 12-year-old children meeting research criteria had a clinical diagnosis of ASD, although all children with ASD had received a clinical diagnosis of at least one psychiatric or developmental disorder, likely due in part at least to diagnostic overshadowing(Baird et al., 2006). This highlights the importance of a thorough assessment of children with ID prior to commencing treatment.

Table 1. Participant baseline characteristics
CharacteristicTreatment group
Overall (N = 122)Placebo (N = 61)Methylphenidate (N = 61)
SCQ score ≥ 15 (N, %)65 (58.5)33 (50.0)32 (57.1)
SCQ score ≥ 21 (N, %)37 (33.3)21 (38.2)16 (28.7)

Arnold cites the difference in parent-reported effect sizes for our trial (0.39) and that for RUPP (0.88), while noting that the teacher-reported effect sizes are strikingly similar. He raises the important question as to whether autism moderates the effect of medication in children with ID. We tested this by using SCQ scores as a continuous, rather than categorical variable, to extract the greatest power, and found no evidence of an interaction between SCQ score and treatment effect. Arnold also points out that the different diagnostic criteria (DSM-IV ADHD in the RUPP study, ICD-10 hyperkinetic disorder in the present trial) would be expected to show a greater effect size in our trial than that of RUPP.

An alternative explanation for the difference in parent-reported effect sizes could be that parents in the RUPP study were less likely to be blind to treatment arm. RUPP used a crossover design, which increases the chance of participants (and their parents/carers) noticing a difference between the two treatment conditions and contrasting, or exaggerating the difference between the two arms. Teachers, who are less likely to know when the changeover between treatments occurred, are more likely to remain blind.

The clinical implications suggested by Arnold (2013) provide a useful guide to treating children with ADHD/hyperkinetic disorder and intellectual and developmental disorders.


The Simonoff et al. (2013) study was funded by The Health Foundation, formerly the PPP Foundation, reference number 1978, and sponsored by King's College. E.S. received support from the Biomedical Research Centre for Mental Health, National Institute for Health Research. The authors of this Reply have declared that they have no competing or potential conflicts of interest.