ContentServer (1)

Randomized controlled double-blind trial ofoptimal dose methylphenidate in children and

adolescents with severe attention deficithyperactivity disorder and intellectual disability

Emily Simonoff,1 Eric Taylor,1GillianBaird,2 SarahBernard,3OliverChadwick,1

Holan Liang,1 Susannah Whitwell,1 Kirsten Riemer,1 Kishan Sharma,1 SantvanaPandey Sharma,1 Nicky Wood,1 Joanna Kelly,4 Ania Golaszewski,1 Juliet

Kennedy,1 Lydia Rodney,1 Nicole West,1 Rebecca Walwyn,4 and Fatima Jichi41Department of Child and Adolescent Psychiatry, Kings College London, Institute of Psychiatry, London, UK; 2Guys& St Thomas NHS Foundation Trust, London, UK; 3South London & Maudsley NHS Foundation Trust, London, UK;

4Department of Biostatistics, Kings College London, Institute of Psychiatry, London, UK

Background: Attention deficit hyperactivity disorder is increased in children with intellectual disability.Previous research has suggested stimulants are less effective than in typically developing children butno studies have titrated medication for individual optimal dosing or tested the effects for longer than4 weeks. Method: One hundred and twenty two drug-free children aged 715 with hyperkineticdisorder and IQ 3069 were recruited to a double-blind, placebo-controlled trial that randomizedparticipants using minimization by probability, stratified by referral source and IQ level in a one to oneratio. Methylphenidate was compared with placebo. Dose titration comprised at least 1 week each of low(0.5 mg/kg/day), medium (1.0 mg/kg/day) and high dose (1.5 mg/kg/day). Parent and teacherAttention deficit hyperactivity disorder (ADHD) index of the Conners Rating Scale-Short Version at16 weeks provided the primary outcome measures. Clinical response was determined with the ClinicalGlobal Impressions scale (CGI-I). Adverse effects were evaluated by a parent-rated questionnaire,weight, pulse and blood pressure. Analyses were by intention to treat. Trial registration: ISRCTN68384912. Results: Methylphenidate was superior to placebo with effect sizes of 0.39 [95% confidenceintervals (CIs) 0.09, 0.70] and 0.52 (95% CIs 0.23, 0.82) for the parent and teacher Conners ADHDindex. Four (7%) children on placebo versus 24 (40%) of those on methylphenidate were judged im-proved or much improved on the CGI. IQ and autistic symptoms did not affect treatment efficacy. Activemedication was associated with sleep difficulty, loss of appetite and weight loss but there were nosignificant differences in pulse or blood pressure. Conclusions: Optimal dosing of methylphenidate ispractical and effective in some children with hyperkinetic disorder and intellectual disability. Adverseeffects typical of methylphenidate were seen and medication use may require close monitoring in thisvulnerable group. Keywords: Attention deficit disorder with hyperactivity, Randomized controlledtrial, autism, mental retardation, intellectual disability, methylphenidate, stimulants.

IntroductionAttention deficit hyperactivity disorder (ADHD) andthe more severe hyperkinetic disorder (HD) arecharacterized by levels of overactivity, inattentionand impulsivity that are developmentally inappro-priate and pervasive across different situations suchas home, school and leisure activities. An associa-tion with subaverage intelligence is consistentlyidentified; one epidemiological study reported aneightfold increase in the rate of HD among childrenwith global learning problems (Emerson, 2003) andanother reported a 16-point decrease in IQ of thosemeeting a questionnaire algorithm for HD (Simonoff,Pickles, Wood, Gringras, & Chadwick, 2007). Thelatter study demonstrated that the psychiatric and

cognitive correlates of HD were similar in those withIQ 4069 compared with those of normal ability andshowed both groups experienced similar psychoso-cial impact of HD.Guidelines on medical management recommend

methylphenidate as the first line (Pliszka, et al.,2006; Taylor et al., 2004). The MTA trial demon-strated that individually titrated medication wassignificantly better than either treatment as usual(mostly medication), suggesting that individual doseselection is important in optimizing behaviouralimprovement (The MTA Cooperative Group, 1999).Most treatment trials for ADHD have excluded

children with intellectual disability (ID) and the evi-dence pertaining to this group is restricted to ahandful of studies. (Aman, Buican, & Arnold, 2003)aggregated the studies from his group, concludingthat there was consistent evidence for a moderateeffect size (0.57 for teachers, 0.39 for parents) butthat this was considerably less than reported in

Author deceased - see Acknowledgements.

Conflict of interest statement: No conflicts declared - see

Acknowledgment for others contributing to the trial.

Journal of Child Psychology and Psychiatry 54:5 (2013), pp 527535 doi:10.1111/j.1469-7610.2012.02569.x

2012 The Authors. Journal of Child Psychology and Psychiatry 2012 Association for Child and Adolescent Mental Health.Published by Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA 02148, USA

studies with typically developing children, e.g.0.80.9 in the MTA titration trial (Greenhill et al., 2001).Other groups (Handen, Feldman, Lurier, & Murray,1999; Pearson et al., 2003) consistently find similarmoderate effects of methylphenidate. Amans aggre-gated analysis further suggested that lower IQ/mental age predicted a poorer medication response.However, design limitations in all these studies affectthe interpretation and resulting confidence in theevidence base. First, all recent studies were cross-over in design; the success of blinding was not con-sistently reported and there is often low power todetect order effects. Second, all trials were short-term in duration (24 weeks follow-up) and do notassess the medium- to long-term efficacy of ADHDtreatment. Third, dosing regimens in some oldertrials did not follow current practice; all usedimmediate-release formulation in twice rather thanthree times daily dosing and none used individualdose titration. Finally, sample sizes were very small(1135 participants), producing imprecise estimatesand not allowing tests of moderator effects. The latterare particularly important as both clinical reports(Campbell, 1978) and theoretical grounds (Barkley,McMurray, Edelbrock, & Robbins, 1992) suggestthat those with autism may be particularly sensitiveto adverse effects and prone to develop more tics,rituals/compulsions, stereotyped behaviours, irri-tability and social withdrawal (Handen, Johnson, &Lubetsky, 2000; Quintana et al., 1995). A recentlarger randomized controlled crossover trial in chil-dren with autism spectrum disorders (ASDs) andhigh levels of hyperactivity ((Research Units of Pedi-atric Psychopharmacology (Autism Network), 2005)reported beneficial effects of stimulants amongstthose tolerating medication but 13 of 72 (18%)withdrew due to adverse effects.The present trial of immediate-release methyl-

phenidate versus placebo in children with ID andhyperkinetic disorder aimed to answer the question,as set out in the protocol: (a) What is the efficacy andsafety of stimulant medication over 16 weeks, underconditions of individual optimal dose titration; (b)What is the profile of adverse effects in this popula-tion; and (c) Is efficacy moderated by the level of ID,the presence of autistic symptoms or the severity ofADHD symptoms?

MethodsThiswasa randomized controlled double-blind trial ofimmediate-release methylphenidate versus placeboindividually titrated over three ormoreweeks followedby optimal dose until 16 weeks post randomization.Participants were allocated to treatments through theMental Health and Neuroscience Clinical Trials Unit(CTU-London, UK). Allocation was undertaken inde-pendently of the trial team. The first 15 participantswereallocatedwith simple randomization. Thereafter,allocation was stratified by source of referral (clinical

referral or community screening) and ID (full-scale IQscore

substance abuse disorder. Only one child per familywas enrolled in the double-blind phase at any onetime. Written informed consent was obtained prior torandomization from parents/legal guardians of par-ticipants with assent of children as appropriate fortheir level of understanding.

Medication

Immediate-release methylphenidate supplied as Equ-asym in 5, 10 and 20 mg tablets (UCB pharma, Slough,Berkshire, UK) comprised the active treatment; amatching placebo in identical doses was manufacturedby the Guys and St. Thomas Pharmaceutical Produc-tion Laboratory, which is approved by the UK MHRA.

Optimal dose titration and selection

This procedure aimed to assess participants on threedaily doses: 0.5 (low dose). 1.0 (medium) and 1.5 (high)mg/kg, given in increasing dose and delivered threetimes daily at breakfast, lunchtime and after school. Atthe end of titration, two senior medical investigators(ES, ET, GB and SB) independently judged optimal dosefor each participant using parent, teacher and clinicianratings on adverse events and behavioural improve-ment on the parent and teacher Conners ADHD indexand hyperactivity scale. This dose was then prescribedfor the remainder of the 16 week trial (Appendix S1).

Measures of efficacy

The primary outcome measures for this study were theparent and teacher Conners ADHD index at 16 weeks.Secondary outcome measures at 16 weeks includedscores on the parent and teacher Conners hyperactivityscale and hyperactivity subscale of the AberrantBehavior Checklist (ABC). The Clinical Global Impres-sions-Improvement rating (CGI-I; Sharkey, 1985) wascompleted by the clinician with the parent at the16-week assessment; ratings of improved, muchimproved and completely recovered were categorizedpost hoc as a clinical response.

Safety

Possible adverse effects were monitored during dosetitration and at the 8, 12 and 16 weeks post randomi-zation visits, as described in the Appendix S1 (Hill &Taylor, 2001). Because the optimal dosing design in-volved the trial of doses that may evoke short-term ad-verse effects, the primary interest was in adverse effectspresent at 16 weeks, on optimal dose, as measured bythe parent questionnaire. Behaviours of a priori interestfor individual analysis were: parent-rated sleep diffi-culties, poor appetite, looks sad/miserable, cryingspells, looks anxious, meaningless repetitive behaviourand social withdrawal (talks less than usual with otherchildren), with significant worsening classified as a two-point deterioration from baseline. In addition, twoaggregate scores were calculated post hoc. The targetadverse effects score summed the behaviours listedabove and aimed to identify smaller differences acrossseveral adverse effects. The other adverse effects score

comprised the other behaviours of the adverse effectsquestionnaire: stares a lot or daydreams; not interestedin other children; irritable; complains of stomach ache;complains of headache; drowsy; seems unsteady; ex-cited; angry; has nightmares; displays twitches (tics);and shaking or tremor of hands. Changes in weight,pulse and blood pressure were also considered. Weobserved children who were withdrawn frommedicationbecause of adverse effects.

Statistical analysis

No formal interim analyses were performed. Sample sizecalculations were based on an effect size of 0.5 in thedifferential reduction in symptoms between baselineand 16 weeks (Research Units of Pediatric Psycho-pharmacology (Autism Network), 2005). A sample sizeof 64 per group gives 80% (a = .05, two-tailed) power.The analyses, using Stata version 11, followed a pre-specified plan and were performed on the intention totreat sample. Effect sizes were calculated as the esti-mated adjusted treatment effect divided by the pooledwithin group standard deviation of the outcome. Linearregression models were fitted for each outcome. Allmodels included a fixed contrast for the treatmenteffect, the randomization stratification factors, age, gen-der and baseline SCQ score. The screening and baselinevalues of the outcome were included as fixed covariates.The moderator effects specified a priori were: severity

of ID (baseline full-scale IQ), autistic symptoms (base-line SCQ score) and ADHD severity (baseline parent andteacher Conners ADHD index). Moderator analyseswere performed on both primary outcomes by includingan interaction between treatment and moderator andthe results considered exploratory.The adverse events measured on a continuous scale

were examined using linear regression while binaryevents were examined using exact logistic regression,with contrasts for treatment, source and severity ineach case. The two adverse effects scores were exam-ined using separate regression analyses with contrastsfor treatment, severity, source, gender and the covari-ates age and baseline SCQ. All other safety data arereported descriptively.As a sensitivity analysis, the primary analyses were

repeated including only compliers who took most or allmedication doses as determined by parent and teacherreport and clinician judgement. All statistical tests aretwo-sided.Missing data were treated as missing at random and

imputed using multiple imputation (Rubin, 1987) bychained equations (Van Buuren, Boshuizen, & Knook,1999) using the ice package [version 1.1.4 (Royston,2005)]. The results of 200 imputed datasets were com-bined using Rubins Rules (Rubin, 1987), implementedusing the mim command [version 1.1.8 (Carlin, Galati,& Royston, 2008)]. For the adverse events analysis,multiple imputation was not used.

ResultsWe assessed 890 children (764 through communityscreening, 129 through clinical referral) for eligibilityand 122 (73 community, 49 clinical) were randomized

doi:10.1111/j.1469-7610.2012.02569.xMethylphenidate for hyperkinetic disorder in children with intellectual disability 529

2012 The Authors. Journal of Child Psychology and Psychiatry 2012 Association for Child and Adolescent Mental Health.

between June 2005 and July 2008 (Figure 1). Par-ticipant characteristics by treatment group areshown in Table 1. Sixteen withdrew from the trialbefore week 16; 5 of the withdrawals, all on activemedication, were due to adverse events (details inAppendix S1). We followed 122 participants until16 weeks and all are included in the intention totreat analysis. Eighty-seven (47 active, 40 placebo)were rated as having good or excellent medicationcompliance and are included in a sensitivity analy-sis. The doses being taken (including none for thosewithdrawn) are shown in the Appendix S1.

Behavioural outcome

Table 2 shows that participants receiving activemedication experienced a significantly greaterreduction in parent- and teacher-rated symptoms onthe Conners ADHD index. Significant group differ-ences were also seen for the parent and teacherConners hyperactivity scale and the ABC hyperac-tivity scale. On the CGI-I, 24 of 61 (40%) participantsreceiving methylphenidate were rated as improved orbetter, compared to four of 57 (7.1%) of those onplacebo, giving a number needed to treat (NNT) ofthree. Sensitivity analysis including compliers onlyreported similar effect sizes (0.330.74) to those seenin the intention to treat analysis (Table 2).

Moderators of response

None of the candidate moderators (IQ, autisticsymptoms, parent- and teacher-rated ADHD sever-ity) had an effect on parent- or teacher-rated Con-ners ADHD index at 16 weeks with p-values rangingfrom 017 to .94.

Adverse effects

Children taking methylphenidate were more likely tohave parent reports of sleep difficulties and poorappetite but the other parent-reported adverseevents showed nonsignificant group differences(Table 3). There were no group differences in meanchange in pulse or blood pressure. Exploratoryanalyses to detect a subgroup with a large increasein blood pressure or pulse revealed the followingincreases of 5 points: for diastolic blood pressure,20 in the methylphenidate arm versus 17 in theplacebo arm; for systolic blood pressure, 23 meth-ylphenidate versus 19 placebo; for pulse 23 partici-pants in both arms (all nonsignificant; numbers bygroup above 95th centile at week 16 given in theAppendix S1). Weight change from baseline differedsignificantly with a group difference in mean differ-ence from baseline to week 16 of )2.70 (95% CI)3.72, )1.67) (Table 4). The target adverse effects

Table 1 Participant baseline characteristics

Characteristic

Treatment group

Overall(N = 122)

Placebo(N = 61)

Methylphenidate(N = 61)

Type of referral (n, % clinical referral)a 73 (60) 36 (59) 37 (61)Severity of intellectual disability (N, % IQ

score decreased for both groups between baselineand week 16. There was a marginally significanttrend for the active treatment group to decrease lessover the course of treatment. The other adverseeffects score also showed a reduction for both groupsbetween baseline and week 16 with no group differ-ence.Comparing the means and standard errors (SEs) of

all 16 withdrawals (from both arms) to the non-withdrawals revealed a M (SE) IQ of 51.8 (2.6) versus53.7 (1.0) and M (SE) SCQ scores of 15.1 (1.9) versus17.4 (0.8) respectively. A similar comparisonamongst only those receiving active medication andcomparing the five withdrawn because of likely/

definite adverse effects versus the 51 remaining onactive medication revealed a M (SE) IQ of 47.2 (4.7)versus 54.1 (1.3) and SCQ scores of 18.4 (2.0) versus16.9 (1.1).

ConclusionsMethylphenidate is effective in reducing ADHDsymptoms in children with intellectual disability; theresults reveal moderate effect sizes of 0.390.52 forthe primary outcome of parent and teacher ADHDindex. Some researchers have argued that thehyperactivity scale of the ABC is a more suitablemeasure for children with ID and it is interesting tonote the parent (but not the teacher) effect size islarger and more consistent than seen for the Con-ners scales. However, the Conners scales providecomparability both to other studies of children withID and also to the wider literature on children ofaverage ability. A responder/nonresponder classifi-cation using the CGI shows a highly significant(p < .001) group difference but this is partly drivenby the relative absence of a placebo effect, with only7% of those receiving placebo being responders.Other studies with a similar participant group haveused a crossover design and placebo effects are lar-gely unreported.The present findings, including the effect sizes, are

consistent with the previous literature. Amanssummary of the aggregated crossover trials of

Table 2 Effect of treatment on primary and secondary ADHD measures, intention to treat

Methylphenidate M(SE)

Placebo M (SE) Group differencein the mean

difference betweenthe time-points

(95% CI)a p-value

Adjustedeffect size unadjustedSD (95% CI)

Baseline 16 weeks Baseline 16 weeks

Intention to treatParent ADHD Index 27.4 (0.8) 19.1 (1.4) 27.8 (0.7) 22.4 (1.1) )3.8 ()6.7, )0.9) 0.011 0.39 (0.09, 0.70)Teacher ADHD Index 21.5 (1.2) 14.5 (1.2) 20.0 (1.3) 18.6 (1.3) )5.1 ()7.9,)2.2) 0.001 0.52 (0.23, 0.82)Parent HyperactivityScale

12.0 (0.5) 7.7 (0.7) 12.1 (0.5) 9.2 (0.7) )1.8 ()3.4, )0.2) 0.03 0.33 (0.03, 0.63)

Teacher HyperactivityScale

10.1 (0.8) 6.4 (0.7) 9.8 (0.8) 9.0 (0.8) )3.2 ()4.9, )1.5)

methylphenidate versus placebo, which reported amean effect size of 0.57 for the four teacher-reportedhyperactivity measures and 0.39 for the parallelparent scales (Aman et al., 2003). However, becauseof the design differences in the present study, itcannot be inferred that the finding of similar effectsizes indicate that the use of individually titratedoptimal three times daily dosing does not achievebetter results than group-standardized, twice-dailydosing. Furthermore, the trial end-point is signifi-cantly longer than that reported in previous trials.This is the first trial to formally examine the role of

treatment moderators such as IQ, autistic symptomsand ADHD severity. In the aggregate studies, Amansuggested that lower IQ/mental age might predictpoorer treatment response. In our larger sample, wefound no support for this. The studies aggregated byAman included individuals with IQs up to 90 and itmay be that the suggestive difference reported is dueto greater response in those with IQ >70, rather thana difference among those in the ID range. The pres-ent finding of no subgroup difference according toautistic symptoms is in line with the findings inchildren with an ASD and high levels of hyperactiv-ity, in which effect sizes for methylphenidate rangedfrom 0.20 to 0.54 (Research Units of Pediatric Psy-chopharmacology (Autism Network), 2005). Takentogether, these two studies support the use of stim-ulant medication in children with combined ADHDand ASD symptoms. The current criteria do not allowthe diagnoses of ADHD/hyperkinetic disorder whenan ASD has been diagnosed. This is likely to changein the revisions of the DSM and ICD criteria and thecurrent findings support this alteration in showingthat children who likely meet both sets of criteria canbenefit from stimulant treatment. We found thatADHD symptom severity did not predict responseamongst those receiving a diagnosis of hyperkineticdisorder. In the present trial, 60% of the participantswere not clinically recognized or treated for their

ADHD symptoms; these findings suggest that clini-cians should be more aware of the increased rate ofADHD among children with ID and the potentialbenefits of stimulant treatment, even amongst thosewith milder symptoms.The rate of withdrawals, 13%, is similar to that

reported in other studies of similar populations(Aman, Kern, McGhee, & Arnold, 1993; ResearchUnits of Pediatric Psychopharmacology (AutismNetwork), 2005); furthermore, it should be notedthat only 4% were withdrawn due to adverse effects.Nevertheless, treatment with methylphenidate pro-duced significantly more adverse effects than seenwith placebo. The adverse effects profile was typicalof methylphenidate, with sleep and appetite diffi-culties along with failure to gain weight. It is notablethat these difficulties were seen even when partici-pants were prescribed an optimally selected dose, inwhich acceptability of any adverse effects were con-sidered. This highlights the need to regularly reviewadverse effects even when an initial dose has beencarefully selected. However, it is interesting to notethat the target adverse effects score decreasedfor both groups, indicating overall improvement,between baseline and week 16. Explanations for thisinclude the fact that many adverse effects can alsobe indicators of comorbidity or alternatively thatinitial high scores show regression to the mean.The current study has a number of strengths. The

sample size is more than three times that of the nextlargest randomized double-blind study of methyl-phenidate in this population (Aman et al., 1993). Thelarger sample size increases the precision of the ef-fect size estimates and also allows the moderatingeffects of IQ, autistic symptoms and ADHD severityto be tested. The length of follow-up (16 weeks) issubstantially longer than any other trial and there-fore provides the first data on medium-term effects ofmedication. The study employed an intention to treatanalysis and all but one participant was followed up

Table 4 Mean changes indicative of possible adverse events

Methylphenidate M (SD) Placebo M (SD) Group differencein the mean differencebetween the time-points (95% CI)a p-valueBaseline Week 16 Baseline Week 16

Weight1 41.3 (19.9) 39.6 (17.3) 41.0 (14.9) 43.8 (16.8) )2.70 ()3.72,)1.67)

until 16 weeks. All diagnoses of ADHD/hyperkineticdisorder were carefully considered using a researchclinical instrument, the CAPA, as well as teacherConners and direct behavioural observation; thecombined information was reviewed in the context ofparticipants IQ results by at least one experiencedclinician. The recruitment strategy included bothclinically identified cases and also children in whomADHD had not been previously recognized. There-fore, the present findings are generalizable to thewider population of children with ADHD and ID. Thedose optimization strategy employed in the currentstudy was close to current clinical practice.A potential limitation of the present study is the

number of participants (33, 27%) with poor medi-cation compliance. However, this proportion broadlyreflects the experience in UK health care and shouldbe considered a reflection of the representative nat-ure of the sample. It is interesting to note that theeffect sizes from the sensitivity analysis that in-cluded only good compliers were similar to thosefrom the full intention to treat analysis. Although the

current trial is much larger than any previous study,the sample size is powered for the main effect ofmedication on ADHD symptoms and not for themoderating influences that were tested. Although thefailure to identify any significant interaction musttherefore be viewed with caution, it should also benoted that none of these effects approached signifi-cance. This suggests that any such moderatinginfluence is likely to be subtle.The study has three main findings. First, methyl-

phenidate is effective in about 40% of children withhyperkinetic disorder and ID. Second, the adverseeffects experienced by this population are similar tothose seen in typically developing children withADHD and can be identified with careful but routinemonitoring of medication. The current study cannotexclude the presence of rare, atypical adverse events,but these apply to any population of children treatedwith medication. Nevertheless, because childrenwith ID often have complicating medical problemsand are less likely to be able to communicate sub-jectively experienced adverse events, particular care

Excluded from the Trial pre Interview (n = 519; 58%) Primary Eligibility Reason (n = 227; 26%) No diagnosis of hyperkinetic disorder (HD) (n = 5) IQ not between 30 and 69 at randomization (n = 7) Not aged between 7 and 15 at randomization (n = 81) Other primary eligibility reason (n = 134) Primary Consent Reason (n = 292; 33%) Parent declined (n = 208) Local clinician declined (n = 1) Repeated appointments missed (n = 1) Child declined (n = 0) Other (n = 82)

Assessed for Eligibility (n = 890)

Excluded from Trial at Interview (n = 249; 28%) Primary Eligibility Reason (n = 187; 21%)

No diagnosis of hyperkinetic disorder (HD) (n = 49) IQ not between 30 and 69 at randomization (n = 106) Not aged between 7-15 years at randomization (n = 2) Other primary eligibility reason (n = 30)

Primary Consent Reason (n = 62; 7%) Parent declined (n = 39) Local clinician declined (n = 1) Repeated appointments missed (n = 12) Child declined (n = 2) Other (n = 8)

Randomized (n = 122)

Allocated to placebo (n = 61) Received allocated intervention (n = 60)

Allocated to methylphenidate (n = 61) Received allocated intervention (n = 58)

Lost to follow-up (n = 1) Reasons: Uncontactable (n = 1)

Discontinued intervention (n = 6) Reasons: Poor adherence (n = 3) Consent withdrawn (n = 2) Unable to attend appointments (n = 1)

Lost to follow-up (n = 0)

Discontinued intervention (n = 10) Reasons: Adverse events (n = 5) Consent withdrawn (n = 3) Delay in optimal dose (n = 1) Eligibility not met (n = 1)

Analyzed (n = 61) Excluded from analysis (n = 0)

Analyzed (n = 61) Excluded from analysis (n = 0)

Figure 1 Diagram of participant selection and enrolment to the trial



should be exercised in prescribing and monitoringstimulant medication. Finally, we found no evidenceto support moderating effects of IQ, autistic symp-toms or ADHD severity. Therefore, stimulants shouldbe considered in children with hyperkinetic disorderand ID regardless of their profile with respect tothese characteristics.

Supporting informationAdditional supporting information is provided alongwith the online version of this article.Appendix S1 Doses taken at 16 weeks

AcknowledgementsTheworkwas fundedbyTheHealthFoundation, formerlythe PPP Foundation, reference number 1978, and spon-sored by Kings College. The independent Trial SteeringCommittee comprised Professor Patricia Howlin (Chair20042005), Professor Chris Oliver (Chair 20052008),Professor JeremyTurk,Dr FrankBesag (20062008),MrBrian McGinnis (representing carer views 20042006),Dr Michael Fitzpatrick (representing carer views 20062008) and Dr Gill Dale (representing the Sponsor). Theindependent Data Monitoring Committee comprised DrRuth Pickering (Chair), Dr David Coghill and Dr ClaireSturge. The Mental Health and Neurosciences ClinicalTrials Unit, Kings College London undertook randomi-zation, database development and assisted with projectmanagement. Clare Rutterford contributed to the devel-opmentof thestatisticalanalysisplan,dataoversightandinitial analyses. Professor AndrewPickles supervised thestatistical analyses for this paper. We thank all the doc-

tors who referred participants to the trial and the partici-pants and their families. None of the authors of themanuscripthasanyconflictof interest, includingspecificfinancial interests, relationshipsoraffiliationsrelevanttothe subject of themanuscript.One of the authors, Dr Oliver Chadwick, very sadly

passed away before the final submission of the manu-script. Dr Chadwick, prior to his death, made a sub-stantial intellectual and practical contribution to thisstudy and his involvement was essential to the otherauthors. Dr Chadwick was involved from the inceptionof the study, including in writing the research grant thatfunded the study, and played a key role in the runningof the study for the first 3 years. His contributions in-cluded both theoretical ideas about the role of medica-tion in children with intellectual disability and thediagnosis of ADHD in children with severe to profoundintellectual disability. Most importantly, he had thesenior responsibility for assessment of intellectualability in the study. Dr Chadwicks involvement in for-mulating the conceptual and practical framework forselecting measures of intellectual ability and followingthis over time was essential to the study. The journaleditors have kindly confirmed the appropriateness ofincluding Dr Chadwick as a listed author for the article.

Correspondence toEmily Simonoff, Department of Child and AdolescentPsychiatry, Kings College London, Institute of Psychi-atry, DeCrespigny Park London SE5 8AF, UK; Tel:+20 7848 5312; Fax: +20 7708 5800; Email: [email protected]

What is known

ADHD and hyperkinetic disorder (HD) are increased about eightfold in children with intellectual disability (ID)but often goes unrecognized

Methylphenidate is the first line medication for children with ADHD/HD but the large, two-arm trials haveexcluded children with ID

What is new

Treatment with methylphenidate was superior to placebo in reducing ADHD symptoms and producing clinicalimprovement

Common adverse effects of insomnia, poor appetite and weight loss were associated with methylphenidatetreatment but overall safety was good

What is clinically relevant

ADHD/HD is important to identify and treat in children with intellectual disability

ReferencesAman, M.G., Buican, B., & Arnold, L.E. (2003). Methylphe-nidate treatment in children with borderline IQ andmental retardation: Analysis of three aggregated studies.Journal of Child & Adolescent Psychopharmacology, 13,2940.

Aman, M.G., Kern, R.A., McGhee, D.E., & Arnold, L.E. (1993).Fenfluramine and methylphenidate in children with mentalretardation and ADHD: Clinical and side effects. Journal of

American Academy of Child and Adolescent Psychiatry, 32,851859.

Angold, A., Prendergast, M., Cox, A., Harrington, R., Simonoff,E., & Rutter, M. (1995). The Child and AdolescentPsychiatric Assessment (CAPA). Psychological Medicine, 25,739753.

Barkley, R., McMurray, M.B., Edelbrock, C.S., & Robbins, K.(1992). Side effects of methylphenidate in children withattention deficit hyperactivity disorder: A systemic, placebo-controlled evaluation. Pediatrics, 86, 184192.

534 Emily Simonoff et al. J Child Psychol Psychiatry 2013; 54(5): 52735


Campbell, M. (1978). Pharmacotherapy. In M. Rutter, & E.Schopler (Eds.), Autism: A reappraisal of concepts andtreatment (pp. 337355). New York: Plenum Press.

Carlin, J.B., Galati, J.C., & Royston, P. (2008). A new frame-work for managing and analyzing multiply imputed data inStata. Stata Journal, 8, 4967.

Conners, C.K. (1989). Conners rating scales manual. NewYork: Multi-Health Systems Inc.

Emerson, E. (2003). Prevalence of psychiatric disorders inchildren and adolescents with and without intellectualdisabilities. Journal of Intellectual Disability Research, 47,5158.

Greenhill, L.L., Swanson, J.M., Vitiello, B., Davies, M., Cle-venger, W., Wu, M., ... & Wigal, T. (2001). Impairment anddeportment responses to different methylphenidate doses inchildren with ADHD: The MTA titration trial. Journal of theAmerican Academy of Child & Adolescent Psychiatry, 40,180187.

Handen, B.L., Feldman, H.M., Lurier, A., & Murray, P.J.(1999). Efficacy of methylphenidate among preschool chil-dren with developmental disabilities and ADHD. Journal ofthe American Academy of Child & Adolescent Psychiatry, 38,805812.

Handen, B.L., Johnson, C.R., & Lubetsky, M. (2000). Efficacyof methylphenidate among children with autism and symp-toms of attention-deficit hyperactivity disorder. Journal ofAutism & Developmental Disorders, 30, 245255.

Hill, P., & Taylor, E. (2001). An auditable protocol for treatingattention deficit/hyperactivity disorder. Archives of Diseasein Childhood, 84, 404409.

Mullen, E.M. (1997). Mullen scales of early learning. LosAngeles, CA: Western Psychological Services.

Novacheck, T.F., Stout, J.L., & Tervo, R. (2000). Reliability andvalidity of the Gillette Functional Assessment Questionnaireas an outcome measure in children with walking disabilities.Journal of Pediatric Orthopaedics, 20, 7581.

Pearson, D.A., Santos, C.W., Roache, J.D., Casat, C.D.,Loveland, K.A., Lachar, D., ... & Cleveland, L.A. (2003).Treatment effects of methylphenidate on behavioral adjust-ment in children with mental retardation and ADHD. Jour-nal of the American Academy of Child & AdolescentPsychiatry, 42, 209216.

Pliszka, S.R., Crismon, M.L., Hughes, C.W., Corners, C.K.,Emslie, G.J., Jensen, P.S., ... & the Texas ConsensusConference Panel on Pharmacotherapy of Childhood Atten-tion Deficit Hyperactivity Disorder. (2006). The Texas Chil-

drens Medication Algorithm Project: Revision of thealgorithm for pharmacotherapy of attention-deficit/hyper-activity disorder. Journal of the American Academy of Child& Adolescent Psychiatry, 45, 64257.

Quintana, H., Birmaher, B., Stedge, D., Lennon, S., Freed, J.,Bridge, J., & Greenhill, L.. (1995). Use of methylphenidate inthe treatment of children with autistic disorder. Journal ofAutism and Developmental Disorders, 25, 283294.

Research Units of Pediatric Psychopharmacology (Autism Net-work). (2005). Randomized, controlled, crossover trial ofmethylphenidate in pervasive developmental disorders withhyperactivity.ArchivesofGeneral Psychiatry,62, 12661274.

Royston, D.B. (2005). Multiple imputation of missing values:Update. Stata Journal, 5, 188201.

Rubin, D.B. (1987). Multiple imputation for nonresponse insurveys. New York: John Wiley and Sons.

Rutter, M., Bailey, A., & Lord, C. (2003). The social communi-cation questionnaire (1st edn). Los Angeles: Western Psy-chological Services.

Sharkey, F.H. (1985). Clinical Global Impressions Scale.Psychopharmacology Bulletin, 21, 839843.

Simonoff, E., Pickles, A., Wood, N., Gringras, P., & Chadwick,O. (2007). ADHD symptoms in children with mild intellec-tual disability. Journal of the American Academy of Child &Adolescent Psychiatry, 46, 591600.

Taylor, E., Dopfner, M., Sergeant, J., Asherson, P., Banas-chewski, T., Buitelaar, J., ... & Zuddas, A. (2004). Europeanclinical guidelines for hyperkinetic disorder - first upgrade.European Child & Adolescent Psychiatry, 13(Suppl. 1), I7I30.

The MTA Cooperative Group. (1999). A 14-month randomizedclinical trial of treatment strategies for attention-deficithyperactivity disorder. Archives of General Psychiatry, 56,10731086.

Van Buuren, S., Boshuizen, H.C., & Knook, D.L. (1999).Multiple imputation of missing blood pressure covariatesin survival analysis. Statistics in Medicine, 18, 681694.

Wechsler, D. (2004). Wechsler Preschool and Primary SchoolIntelligence Scale (3rd edn). London: Harcourt Assessment.

Wechsler, D., Rust, J., & Golombok, S. (2004). WechslerIntelligence Scale for children Fourth UK edition (WISC-IVUK). London: Harcourt Assessment.

Accepted for publication: 6 February 2012Published online: 7 June 2012



Copyright of Journal of Child Psychology & Psychiatry is the property of Wiley-Blackwell and its content maynot be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express writtenpermission. However, users may print, download, or email articles for individual use.

Documents

ContentServer (1)