| Study characteristics |
| Methods |
8‐phase, 70‐week, multi‐centre trial (phase III trial) including:
phase 1: screening: varying time phase 2: uncontrolled parent training: 10 weeks phase 3: baseline: 2‐4 weeks phase 4: open‐label, safety lead‐in: 1 week phase 5: random‐sequence, double‐blind, placebo‐controlled, cross‐over titration: 5 weeks. Optional pharmacogenetics trial simultaneously phase 6: randomised, optimal dose, double‐blind, placebo‐controlled, parallel trial: 4 weeks phase 7: open‐label, uncontrolled maintenance: 10 months phase 8: randomised, double‐blind, placebo discontinuation: 6 weeks
If parents requested and clinicians agreed that participants were severely symptomatic, children could be moved directly into the medication phase that was concurrent with parent training. If participants did not tolerate the dosing in phase 4, they could enter the open‐label maintenance phase if they tolerated lower doses (e.g. 1.25 mg, 2.5 mg). If they tolerated all doses except 7.5 mg, they were eligible to enter phase 5, the cross‐over titration, with the planned week on a 7.5‐mg dose replaced by an additional 5‐mg week. After phase 5, cross‐over:
if the child showed the greatest clinical benefit during 1 of the 5 weeks of the cross‐over trial, with no room for improvement, the blind was broken and the child was randomly assigned to that MPH dose or placebo in phase 6 if the child was a placebo responder, phase 6 was skipped and the child was allowed to enter phase 7 while taking no medication and with monthly monitoring by the treating physician if a particular week was deemed best, with ongoing room for improvement, a 2‐week double‐blind trial of 7.5 mg and 10 mg 3 times/d, each for 1 week, was implemented, and teacher and parent ratings and side effects data were subsequently blindly reviewed to determine the best dose, or participants with no clinical benefit any week were not eligible to continue in phase 6 or 7 but were given 1‐month follow‐up and then were referred for community treatment
|
| Participants |
Number of participants screened: 1915
Number of participants included in the trial: 303 (229 boys, 74 girls)
Number of participants who completed the last phase of the trial: 8
Number of withdrawals during the trial: 295
Diagnosis of ADHD: DSM‐IV (combined (75%), hyperactive‐impulsive (25%), inattentive (0%))
Age: mean 4.41 years (range not reported)
IQ: > 70 (mean 99.06)
MPH‐naive: 100%
Ethnicity: white (63%), African American (19%), Asian (2%), Hispanic or Latino (16%), American Indian or Alaskan Native (0.7%)
Country: USA
Setting: outpatient clinic
Comorbidity: ODD (52%), communication disorder (22%), elimination disorder (i.e. encopresis, enuresis; 8%), specific phobia (8%), anxiety disorder (8%), developmental co‐ordination disorder (3%), CD (2%), pica (2%), adjustment disorder (1%), reactive attachment disorder (1%), OCD (0.7%), sleepwalking disorder (0.3%)
Comedication: no
Other sociodemographics: double‐parent family (76%), single‐parent family (18%), mean Hollingshead socioeconomic status 47.20 (SD 9.56)
Phase 4, open‐label, safety lead‐in
Number of participants included: 183
Number of participants followed up: 169
Number of withdrawals: 12
Number of participants leaving the trial phase and entering maintenance (phase 7): 2
Phase 5, cross‐over
Number of participants included: 165 (122 boys, 43 girls)
Number of participants followed up: 147
Number of withdrawals: 14
Number of participants leaving the trial phase and entering maintenance (phase 7): 4
Diagnosis of ADHD: DSM‐IV (combined (76%), hyperactive‐impulsive (24%), inattentive (0%))
Age: mean 4.74 years (range 3‐5.5)
IQ: > 70 (mean 97.93)
MPH‐naive: 100%
Ethnicity: white (63%), African American (18%), Asian (1%), Hispanic or Latino (18%), American Indian or Alaskan Native (0.6%)
Country: USA
Comorbidity: ODD (55%), communication disorder (20%), elimination disorder (i.e. encopresis, enuresis; 8%), specific phobia (7%), anxiety disorder (10%), developmental co‐ordination disorder (4%), CD (3%), pica (2%), adjustment disorder (0.6%), reactive attachment disorder (2%), OCD (0.6%), sleepwalking disorder (0.6%)
Comedication: no
Sociodemographics: double‐parent family (79%), single‐parent family (21%), mean Hollingshead socioeconomic status 47.01 (SD 9.58)
Phase 6, parallel
Number of participants included: 114 (85 boys, 29 girls)
Number of participants followed up: 77
Number of withdrawals: 1
Number of participants leaving the trial phase and entering maintenance (phase 7): 36
Diagnosis of ADHD: DSM‐IV (combined (75%), hyperactive‐impulsive (25%), inattentive (0%))
Age: mean 4.76 years (range not reported)
IQ: > 70 (mean 97.45)
MPH‐naive: 100%
Ethnicity: white (65%), African American (17%), Asian (0.9%), Hispanic or Latino (17%), American Indian or Alaskan Native (0.9%)
Country: USA
Comorbidity: ODD (53%), communication disorder (22%), elimination disorder (i.e. encopresis, enuresis; 7%), specific phobia (7%), anxiety disorder (11%), developmental co‐ordination disorder (5%), CD (3%), pica (0.9%), adjustment disorder (0.9%), reactive attachment disorder (2%), OCD (0.9%), sleepwalking disorder (0.9%)
Comedication: no
Sociodemographics: double‐parent family (80%), single‐parent family (19%), mean Hollingshead socioeconomic status 47.61 (SD 9.45)
Phase 7, open‐label maintenance
Number of participants included: 140 (104 boys, 36 girls)
Number of participants followed up: 95
Number of withdrawals: 45
Diagnosis of ADHD: DSM‐IV (combined (76.4%), hyperactive‐impulsive (23.6%), inattentive (0%))
Age: mean 4.4 years
IQ: > 70
MPH‐naive: 100%
Ethnicity: white (65.0%), African American (17.1%), Asian (1.4%), Hispanic (15.7%), American Indian (0.7%)
Country: USA
Comorbidity: ODD (52.9%), communication disorder (19.3%), anxiety disorder (11.4%)
Comedication: no
Sociodemographics: double‐parent family (81.4%), single‐parent family (18.6%), mean Hollingshead socioeconomic status 47.2 (SD 9.5)
Inclusion criteria
36‐65 months (3‐5.5 years) DSM‐IV criteria for ADHD, hyperactive/impulsive or combined subtype, on Parent DISC‐4, and clinical interview by experienced clinician; symptoms were required to be present for a minimum of 9 months Age‐ and sex‐adjusted T score ≥ 65 on the Hyperactive‐Impulsive subscale of CPRS and CTRS Score < 55 on the Children's Global Assessment of Functioning Scale IQ > 70 as on the Differential Abilities Scale; children scoring < 70 were considered for inclusion if their composite score from the Vineland Adaptive Behavior Scale was > 70 Enrolled in some type of day programme: day care, pre‐school, nursery school, kindergarten, for ≥ 2 half‐days/week. School‐type programme, in which class included ≥ 8 same‐age peers; if children had been expelled from an eligible programme in the 3 months before screening, they could be considered for enrolment Teachers willing to complete rating scale Residing with primary carer for ≥ 6 months before screening Patients and parents willing to attend all visits required by the trial Otherwise generally healthy, and SBP and DBP < 95th percentile for age and sex Stimulant‐naive
Additional inclusion criteria for phase 4, open‐label lead‐in
Not showing substantial ADHD improvement after parent training (phase 2) (continued impairment, operationalised as < 30% reduction on CPRS or CTRS, or a rating of less than "improved" by at least 2 of the 3 raters (parent, teacher, clinician) completing the CGI Scale Parental consent to a medication trial
Additional inclusion criteria for phase 5, cross‐over
Exclusion criteria
Children or their parent(s) could not understand or follow instructions given in the trial Evidence of moderate to severe AEs or evidence of a much improved response to any dose of MPH or another stimulant > 5 weeks of exposure to ≥ 30 mg/d of MPH or equivalent doses of other stimulants Use of any other psychotropic medication or an investigational drug in the past 30 days; episodic use of sympathomimetic decongestants for the common cold under the trial physician's supervision was allowed History of motor or vocal tics or Tourette's syndrome Major medical conditions that would interfere with involvement in a long‐term trial or could be affected negatively by MPH Current evidence of adjustment disorder, pervasive developmental disorders, autism, psychosis, significant suicidality or other psychiatric disorder, in addition to ADHD that requires treatment with additional medication Evidence of current physical, sexual or emotional abuse Living with anyone who currently abuses stimulants or cocaine History of bipolar disorder in both biological parents
All cases were presented to a cross‐site panel of clinicians, and only patients for whom consensus indicated that all inclusion (and no exclusion) criteria were met could be enrolled |
| Interventions |
Phases 1‐3
Enrolment, parent training; baseline: no medical intervention
Phases 4 to 8
Medical intervention (SA‐IR‐MPH)
Phase 4, open‐label, safety lead‐in
Titration: starting dose of 1.25 mg twice daily; increased to 7.5 mg 3 times daily
Duration: 1 week
Treatment compliance: not stated
Phase 5, cross‐over
Randomly assigned sequence of doses of 1.25 mg, 2.5 mg, 5.0 mg or 7.5 mg IR‐MPH and placebo administered 3 times daily for a week
Medication‐free period between interventions: none
Mean MPH dose: not stated
Duration: 5 weeks
Phase 6, parallel
After a 24‐h medication washout, 4 weeks of randomly assigned treatment with a participant's optimal MPH dose as determined in phase 5, or placebo
No of participants randomised to each group: MPH 61, placebo 53
Mean MPH dose: 14.22 mg/d, 0.7 mg/kg/d
Duration: 4 weeks
Treatment compliance: not stated
Phase 7, open‐label, maintenance
Maintenance starting doses were based on the best dose decision from cross‐over titration. Phase 5 placebo responders were maintained without medication for ≥ 4 weeks, unless their condition deteriorated, in which case open‐label treatment could be initiated. For any participant whose condition deteriorated, the MPH dose was gradually titrated for optimal response. The dosing regimen was adjusted to minimise some AEs to 3 times/d (at breakfast, around noon after lunch and at 3:30 pm), 7 days a week
Mean MPH dose: increased from 14.04 mg/d (0.71 mg/kg/d) at month 1 to 19.94 mg/d (0.92 mg/kg/d) at month 10
Duration: 10 months
Treatment compliance: not stated
Phase 8, discontinuation
Randomised, double‐blind, placebo discontinuation trial, in which an abrupt medication replacement consisted of placebo for half of the children, while others continued on their best MPH dose from the end of phase 7. Children returned to active medication if they met relapse criteria, in other words, CPRS or CTRS scores on the Hyperactivity/Impulsivity Index > 1.5 SD above age‐ and sex‐adjusted norms, or a clinician rating of 4 on the CGI‐S
Duration: 6 weeks
Treatment compliance: not stated. Those who opted out of the double‐blind phases would be allowed to continue on open‐label maintenance therapy. This greatly reduced incentive for families to remain in the double‐blind phases, especially if there was reason to suspect that a child had been randomly assigned to placebo |
| Outcomes |
ADHD symptoms
Conners', Loney and Milich Rating Scale, and SKAMP: teacher‐ and parent‐rated weekly in phase 5 SNAP‐IV, average of parent and teacher ratings: at the end of the last week (4th) of phase 6
Serious AEs
General behaviour
SWAN, Early Childhood Inventory: teacher‐ and parent‐rated at baseline and at the end of phase 6 CBCL (home functioning) Hillside Behaviour Rating Scale (school functioning)
Non‐serious AEs
Height and weight were measured without shoes or heavy clothes at each trial visit. (Growth charts provided by the CDC were used to transform absolute units of measurement into Z‐scores.) Laboratory tests were performed at each trial visit General clinician inquiry regarding the child's health problems at each trial visit. AEs monitored by telephone, and parent‐ and teacher‐rated in phase 4. Side Effects Rating Scale, parent‐rated, weekly in phases 5, 6 and 8; monthly in phase 7. Side Effects Rating Scale, teacher‐rated, weekly in phases 5 and 8, and in the last week of phase 6, as well as in first and tenth months of phase 7 AEs Checklist was based on the Pittsburgh Side Effects Rating Scale. The 4‐point (none, mild, moderate or severe) teacher‐rated scale included the following: buccal‐lingual movements; picking at skin or finger; lip or cheek chewing; other abnormal motor movements; worried, anxious appearance; dull, tired, listless appearance; headaches; stomachaches; crabby, irritable behaviour; tearful, sad, depressed behaviour; appetite loss; prone to crying; and uninterested in others with social withdrawal. The Parent Adverse Events Checklist also included trouble sleeping. Only AEs rated as moderate or severe were counted as reportable in the open lead‐in, titration and parallel phases of the trial, whereas AEs rated as mild, moderate or severe were reported from the maintenance phase BP and pulse, at each trial visit. Cardiovascular AEs were based on age‐adjusted normative values. Tachycardia was defined as 2 measurements of resting heart rate > 120 beats per minute (bpm) at the same visit. Hypertension was defined as 2 BP readings at the same visit that were > 95th percentile for age and sex (SBP or DBP), ranging from 110/72 for 3‐year‐olds to 115/74 for 6‐year‐olds. BP was checked again within 7‐14 days. If the reading remained above the cut‐off limit, an AE for hypertension was reported. Hypertension was rated as mild if < 10 mmHg, moderate if 11 mmHg‐20 mmHg and severe if > 20 mmHg above the limit
|
| Notes |
Sample calculation: yes. Target sample size stated in the online protocol 165. Sample to be randomly assigned 120
Ethics approval: yes; by institutional review boards at each trial site. The trial was monitored by the Data and Safety Monitoring Board of the NIMH
Comments from trial authors
Design: the trial included only IR‐MPH, not an ER preparation. The trial may have included an order effect in the cross‐over phase, particularly for children who were randomly assigned to receive higher doses first. The PATS protocol did not provide a stimulant‐untreated clinical control group in non‐controlled phases. Comparisons of height and weight before and after treatment were made against population norms (i.e. children without an ADHD diagnosis and MPH‐exposure). The follow‐up period was not sufficient for evaluation of the critical issue of long‐term effects of initial growth suppression observed in the first year of treatment Dose: doses used in the PATS were relatively low and homogeneous, and the high end of the proposed dose range was truncated. This may have masked dose‐related effects Population: the rigorous procedures for diagnosis of ADHD enhanced the validity of the diagnostic process at the expense of excluding some children likely to have met the conventional criteria for ADHD in other settings. Children who showed substantial ADHD improvement after parent training were not eligible for medication phases. The sample was too small for the medication to be declared safe for this age group. Failure to meet remission criteria may be caused by severity of ADHD symptoms, not by possible ineffectivity of MPH in pre‐schoolers with ADHD. A high attrition rate and differential attrition rates in the allocated groups were possibly due to a delayed trial start, repeated consent procedures, an always‐available option to skip directly into maintenance and greater nervousness among parents about medication side effects in pre‐schoolers. Parents’ experience during the titration phase presumably heightened their awareness of behavioural differences associated with active and placebo medication Data: missing data in general. AE data entry procedures may have inflated AE rates. Parents were told when double‐blind switches between the drug took place. This may have contributed to confounding of negative expectancy and reporting of AEs Exploratory moderator analyses of efficacy data from phase 5 (cross‐over): no adjustments were made for multiple comparisons. Findings should be considered preliminary. The DISC‐4, Parent Version, has not been validated in pre‐school children. Sample sizes across different moderator categories/subgroups were relatively small. Data on compliance were missing
Key conclusions of trial authors
Phase 5 (cross‐over): IR‐MPH, delivered in 2.5‐mg, 5‐mg and 7.5‐mg doses 3 times/d, produced significant reductions in ADHD symptom scale scores in pre‐schoolers compared with placebo, although effect sizes (0.4‐0.8) were smaller than those cited for school‐aged children taking the same medication Pharmacogenetics trial in phase 5 (cross‐over): emerging evidence suggests the potential for understanding individual variability of responses to and side effects of ADHD medications through the trial of genetics, although additional research is required before these findings can be proven to have clinical utility Exploratory moderator analyses of efficacy data from phase 5 (cross‐over): of the 14 variables examined as potential moderators, only 1 (number of concurrent comorbid disorders) served as a moderator of MPH dose response. In pre‐schoolers with ADHD, the presence of no or 1 comorbid disorder (primarily ODD) predicted a large treatment response at the same level as has been found in school‐aged children, and 2 comorbid disorders predicted moderate treatment response; whereas the presence of ≥ 3 comorbid disorders predicted no treatment response to MPH Phase 6 (parallel): medication effects varied by informant and outcome measure. Parent measures and teacher scores on the SWAN did not differentially improve with MPH. Parent‐rated depression (P value < 0.02) and dysthymia (P value < 0.001) on the Early Childhood Inventory worsened with MPH, but scores were not in the clinical range. Significant medication effects were found on the clinician CGI‐S (P value < 0.0001) and in teacher ratings on the Social Competence Scale (P value < 0.03). Pre‐schoolers with ADHD treated with MPH for 4 weeks improve in some aspects of functioning. Additional improvements might require longer treatment, higher doses and/or intensive behavioural treatment in combination with medication Phase 7 (maintenance): with careful monitoring and a gradual medication dose increase, most pre‐schoolers with ADHD maintained improvement during long‐term IR‐MPH treatment. Variability in effective and tolerated dosing was substantial -
Entire trial duration
AEs during the trial: 11% of pre‐schoolers discontinued treatment because of intolerable MPH AEs. Of the serious AEs reported, 1 occurred at baseline, 2 at lead‐in, 3 in titration, 1 in parallel and 1 in maintenance. Only 1 was possibly related to MPH Growth during the trial: average relative size at baseline was significantly greater than 0 for z height and z weight, indicating greater than expected height (by 2.04 cm) and weight (by 1.78 kg) for participants. During treatment, slopes were significantly less than 0 for z height and z weight, indicating reduction in growth rates. For 95 children who remained on medication, annual growth rates were 20.3% less than expected for height (−1.38 cm/year) and 55.2% less (−1.32 kg/year) for weight. Risks of reduced growth rates should be balanced against expected benefits when pre‐school‐aged children are treated with stimulant medication
Exclusion of MPH non‐responders/children who have previously experienced AEs while taking MPH before randomisation: yes, participants were selected for specific phases of the trial/interventions depending on response in earlier phases, as described in methods
Withdrawals due to AEs: number of withdrawals due to AEs during medication phases: 21 (i.e. 11%)
Number of participants leaving trial phase and entering maintenance (phase 7): 4
Funding source:
Phase 5 (cross‐over): sponsored by the NIMH, Columbia/New York State Psychiatric Institute, Johns Hopkins University, Columbia University, University of California Irvine, Duke University Medical Center, New York University Child Study Center and University of California Los Angeles, Arizona Institute of Mental Health Research to J.K.G. Generic MPH was purchased by grant funds Phase 6 (parallel‐group): sponsored by the NIMH, Columbia/New York State Psychiatric Institute, Johns Hopkins University, Columbia University, University of California Irvine, Duke University Medical Center, New York University Child Study Center and University of California Los Angeles, Arizona Institute of Mental Health Research to J.K.G. Generic MPH was purchased by grant funds. Phase 8 (discontinuation): sponsored by the NIMH, Columbia/New York State Psychiatric Institute, Johns Hopkins University, Columbia University, University of California Irvine, Duke University Medical Center, New York University Child Study Center and University of California Los Angeles, Arizona Institute of Mental Health Research to J.K.G. Generic MPH was purchased by grant funds
Email correspondence with trial authors: June 2014. We obtained supplemental information/data from the trial authors. |
| Risk of bias |
| Bias |
Authors' judgement |
Support for judgement |
| Random sequence generation (selection bias) |
Low risk |
Phase 5 (cross‐over): randomisation was done centrally at the co‐ordinating site, using a computerised stratified randomisation, a 1:1:1:1 starting dose allocation ratio and a randomised, balanced, cross‐over protocol designed to avoid order effects Phase 6 (parallel‐group): a second randomisation to active MPH or to placebo was performed before entry into the parallel‐design, placebo‐controlled phase Phase 8 (discontinuation): centralised randomisation used a computer programme; each child was allocated 1:1 to continuing MPH or switching to placebo under double‐blind conditions |
| Allocation concealment (selection bias) |
Low risk |
Phase 5 (cross‐over): central randomisation using a computer programme Phase 6 (parallel‐group): central randomisation using a computer programme Phase 8 (discontinuation): central randomisation using a computer programme |
| Blinding of participants and personnel (performance bias)
All outcomes |
Low risk |
Phase 5 (cross‐over): double‐blind. Placebo pills were identical to pills containing active medication capsules Phase 6 (parallel‐group): double‐blind. Placebo pills were identical to pills containing active medication capsules Phase 8 (discontinuation): double‐blind. Placebo pills were identical to pills containing active medication capsules |
| Blinding of outcome assessment (detection bias)
All outcomes |
Low risk |
Phase 5 (cross‐over): double‐blind. Except in emergencies, clinicians remained blind to dose sequences. Blinding was maintained for primary dependent measures until after the best dose was determined, or as needed. Parent and teacher dose‐response rating scale graphs were prepared and were blindly evaluated by 2 trial clinicians Phase 6 (parallel‐group): double‐blind. Except in emergencies, clinicians remained blind to dose sequences. Blinding was maintained for primary dependent measures until after the best dose was determined, or as needed. Parent and teacher dose‐response rating scale Phase 8 (discontinuation): double‐blind |
| Incomplete outcome data (attrition bias)
All outcomes |
Low risk |
Phase 5 (cross‐over): all analyses were run using the ITT principle (i.e. each observation obtained for the child was used in the analysis, including those from children who entered each of the 2 phases under consideration and did not complete the phase) Phase 6 (parallel‐group): all analyses were run using the ITT principle (i.e. each observation obtained for the child was used in the analysis, including those from children who entered each of the 2 phases under consideration and did not complete the phase) Phase 8 (discontinuation): ITT
Selection bias (e.g. titration after randomisation → exclusion of MPH non‐responders or placebo responders): no |
| Selective reporting (reporting bias) |
Low risk |
Phase 5 (cross‐over): outcome measures reported in accordance with the published protocol Phase 6 (parallel‐group): outcome measures reported in accordance with the published protocol Phase 8 (discontinuation): not relevant |
