[Summary text]
Study design: Randomized controlled trial
Study grouping: Crossover
Baseline Characteristics
Overall
Included criteria: Inclusion was based on a diagnosis of autistic disorder according to DSMIV criteria (American Psychiatric Association, 2000) and a score above cut-off on each symptom domain of the Autism Diagnostic Interview Revised (ADI-Dietary Intervention R).
Excluded criteria: Children were excluded from the study if their medical histories and/or physical examinations indicated that they had physical or sensory-impairments or significant medical problems including celiac disease.
Intervention Characteristics
Intervention
Control
Kernesymptomer, klinikerbedømt, ADOS social score, lower better
Sponsorship source: This work was supported by the University of Florida’s College of Nursing Biobehavioral NINR-funded research grant P20 NR 07791-03 and GCRC grant M01RR00082 from the National Institute of Research Resources, National Institutes of Health
Country: USA
Setting: Inclusion from: Center for Autism and Related Disabilities (CARD) and/or Child Psychiatry Services at the University of Florida’s Department of Psychiatry and Brain Institute. There were two settings: The General Clinical Research Center where the children were weighed weekly and the children’s homes.
Authors name: Jennifer Harrison Elder
Institution: College of Nursing, University of Florida
Email: elderjh@nursing.ufl.edu
Address: College of Nursing, University of Florida, HPNP Building, Box100187, Gainesville, FL, 32610, USA
12 ugers intervention, inkluderet resultater fra 6 uger
Study design: Randomized controlled trial
Study grouping: Crossover
Baseline Characteristics
Intervention
Control
Overall
Included criteria: The first inclusion criterion is being diagnosed with ASD. The criteria used for the diagnosis of ASD were those proposed by the tenth edition of the International Classi-fication of Diseases (ICD-10; World Health Organization 1992). The diagnosis was provided by child and adolescent psychiatrists with extensive experience in ASD. An age range between 2 and 18 years was set at the onset of the study as another required inclusion criterion.
Excluded criteria: Patients with disorders that do not clearly meet the ASD criteria of the ICD-10; patients diagnosed with allergy to gluten or casein, which could therefore not follow the normal diet; patients who had previously excluded gluten and/or casein from their diet; patients who were likely to not fol-low (not adhere to) the diet properly, for instance children with parents who were institutionalized or suffering from mental disorders.
Pretreatment: At the beginning of the study, group A had 20 participants (13 males, 65%), with ages between 3 and 16 years, being mean age (±SD) of 9.1 (±3.7) years. Group B consisted of 17 participants (16 males, 94%) with ages between 2 and 18 years, being mean age (±SD) of 8.8 year (±4.4).
Intervention Characteristics
Intervention
Control
Kernesymptomer, kliniker bedømt, ATEC, lower better
Vægt, BMI
Sponsorship source: NI
Country: Spain
Setting: child and adolescent psychiatry outpatient services, comprehensive care centers, and associations of parents of autistic children in the provinces of Jaen, Granada, Malaga and Almeria (southern Spain)
Authors name: Luis Gutiérrez-Rojas
Institution: Department of Psychiatry, University of Granada
Email: gutierrezrojasl@hotmail.com
Address: Department of Psychiatry, University of Granada, Granada, Spain
Study design: Randomized controlled trial
Study grouping: Parallel group
Baseline Characteristics
Intervention
Control
Included criteria: Study inclusion criteria were established via the following instruments:Autism Diagnostic Observation Schedule and the DSM-IV (APA2000) criteria for ASD to confirm the diagnosis. The diagnosis was based on both the observations during the ADOS as well as parent interview regarding history of symptoms and current concerns
Intervention Characteristics
Intervention
Control
Gastrointestinale gener
Adfærdsvanskeligheder, forældrebedømt, CBCL Aggression subscale, lower better
Bivirkninger, antal personer
Sponsorship source: Supported by the John F. & Nancy A. Emmerling Fund / The Pittsburgh Foundation. Financial DisclosuresDr. Benjamin Handen disclosed consulting fees for Forest, Bristol Myer Squibband has research funding from Forest, Bristol Myer Squibb, Johnson and Johnson, Neuropharm, and Curemark.
Country: USA
Setting: Participants were recruitedfrom a child developmental assessment clinic located in a metropolitan area, througharea developmental preschool programs, as well as via advertising in local AutismNewsletters and websites
Authors name: Cynthia R. Johnson
Institution: Children’s Hospital of Pittsburgh, Autism Center, University of Pittsburgh School of Medicine
Email: Cynthia.Johnson@chp.edu
Address: Children’s Hospital of Pittsburgh, Autism Center, University of Pittsburgh School of Medicine, 34205th Avenue, Pittsburgh, PA 15213, USA
12 ugers intervention
Study design: Randomized controlled trial
Study grouping: Parallel group
Baseline Characteristics
Intervention
Control
Included criteria: Participationincluded only children with both the diagnosis ofautism and abnormal urinary peptide patterns.
Intervention Characteristics
Intervention
Control
Kernesymptomer, klinikerbedømt, DIPAB autistic traits, lower better
Sponsorship source: NI
Country: Norway
Setting: Information on the project was distributed throughjournals and through the school psychologicalservice. Interested parents were given oral and written information. Structured interviews regarding autistic traits were done with the parents in their homes. formal testing of skills and abilitieswere done in the school or in the kindergartenthe child attended.
Authors name: A.M. KNIVSBERG
Institution: Center for Reading Research, Stavanger University College, P.O. Box 8002, Ullandhaug, N-4068 Stavanger, Norway
Email: ann-mari.knivsberg@slf.his.no
Address: Center for Reading Research, Stavanger University College, P.O. Box 8002, Ullandhaug, N-4068 Stavanger, Norway
Study design: Randomized controlled trial
Study grouping: Parallel group
Baseline Characteristics
Intervention
Control
Included criteria: Subjects with a Diagnostic and Statistical Manual ofMental Disorders (DSM)-IV (APA, 2000) diagnosisof Autistic Disorder (DSM-IV 299.90) recruited fromthe clinics associated with the division of PediatricGastroenterology and the division of Child andAdolescent Psychiatry whose parents were willing tofollow and maintain a gluten- and dairy-free (GDF)diet for 6 weeks, were included. The diagnosis ofautism was confirmed in each child based on theDSM-IV-TR, the Autism Diagnostic Interview(ADI-R),20the Autism Diagnostic ObservationSchedule (ADOS),21a clinical interview, clinic obser-vation, and record review. The confirmatory diagnosiswas made by a licensed psychologist (D.A.P.) who isexperienced in the assessment and diagnosis of ASDsand who is certified as meeting research reliability onboth the ADI-R and ADOS.
Excluded criteria: Children with food allergies, celiac disease, inflam-matory bowel disease, and infectious GI diseases thatmay increase IP were excluded. Children with neuro-logical problems that could interfere with proper evaluation of behavior or with parents unwilling toundertake the dietary challenge with milk and glutenor maintain a GDF diet during the study were alsoexcluded.
Pretreatment: The baseline characteristics in both groups,GD(−) and GD(+), were similar (Table 1). There were no significant differences in IP, behavior, or IQ between the groups at baseline
Intervention Characteristics
Intervention
Control
Sponsorship source: NI
Country: USA
Setting: University of TexasMedical School at Houston
Comments:
Authors name: Fernando Navarro
Institution: Department of Pediatrics, Sectionof Gastroenterology, Hepatology, and Nutrition, University of Texas Medical School at Houston
Email: Fernando.Navarro@uth.tmc.edu
Address: Department of Pediatrics, Sectionof Gastroenterology, Hepatology, and Nutrition, University of TexasMedical School at Houston, 6431 Fannin, Room 3.140, Houston, TX77030-1503
4 ugers intervention
Study design: Randomized controlled trial
Study grouping: Parallel group
Baseline Characteristics
Intervention
Control
Included criteria: Danish children aged between 4 years and 10 years 11 months formally diagnosed with PDD [ICD-10 code F84] at the Center for Autisme or other child psychiatric clinics were enrolled in the study.
Excluded criteria: Exclusion criteria were co-morbid diagnoses of epilepsy, fragile-X syndrome, tuberous sclerosis or a developmental age below 24 months.
Pretreatment: A side from a larger proportion of females in the dietary group, there were no obvious differences in group demographics or scores at baseline.
Intervention Characteristics
Intervention
Control
Kernesymptomer, klinikerbedømt, ADOS social score, lower better
Funktionsniveau, klinikerbedømt, VABS daily living, higher better
Sponsorship source: The study was supported by grants from the Center for Autisme, the Nils O. Seim Family Fund for Medical Research, the Eric Birger Christensen Fond, the Norwegian Protein Intolerance Association and the Robert Luff Foundation.
Country: Denmark
Setting: Center for Autisme, Denmark
Comments:
Authors name: Paul Whiteley
Institution: ESPA Research, The Robert Luff Laboratory, Unit 133I Business & Innovation Centre (BIC), Sunderland Enterprise Park, Wearfield, Sunderland SR5 2TA, UK
Email: paul.whiteley@espa-research.org
Address: ESPA Research, The Robert Luff Laboratory, Unit 133I Business & Innovation Centre (BIC), Sunderland Enterprise Park, Wearfield, Sunderland SR5 2TA, UK
Wrong intervention
Wrong intervention
Wrong study design
Wrong intervention
Wrong study design
Wrong intervention
Wrong dose
Selection bias (biased allocation to interventions) due to inadequate generation of a randomised sequence
Quote: "After obtaining parental consent and following a screening evaluation, the GCRC data manager ran- domly assigned participants who met inclusion crite- ria to either the GFCF or a placebo diet."
Judgement Comment: Insufficient information on sequence generation
Quote: "The order of the interven- tion—that is, beginning with normal diet or with GFCF diet—was assigned randomly, using a computer program."
Judgement Comment: Insufficient information on sequence generation
Judgement Comment: Insufficient information on sequence generation
Quote: "the parents at home daily. <b>Subjects were randomly assigned utilizing block randomization to two different dietary interventions:</b> gluten-dairy free (GD(−)) or gluten-dairy"
Quote: "A statistician, not involved in the study, generated a random allocation, stratified for age and VABS composite scores (n = 72)."
Judgement Comment: Insufficient information on sequence generation
Selection bias (biased allocation to interventions) due to inadequate concealment of allocations prior to assignment
Quote: "Children, parents, and all of the investigative team except for the data manager and dietician were blind to the dietary order."
Judgement Comment: Insufficient information on allocation concealment
Judgement Comment: Insufficient information on allocation concealment
Judgement Comment: Insufficient information on allocation concealment
Quote: "The research pharmacy prepared and documented both masked dietary supplements. The supplements were equivalent in appearance and taste; thus, they were undetectable to parents, study participants, and study personnel."
Judgement Comment: Insufficient information on allocation concealment
Performance bias due to knowledge of the allocated interventions by participants and personnel during the study
Quote: "received a 3- to 4-day supply of food twice each week from the GCRC. In addition, parents were given a list of allowed foods in case of emergencies, and they were asked to record their child’s diet intake to monitor compliance. Each participant’s regular diet was provided during the baseline period and for 6 weeks as the control diet. The experimental diet consisted of gluten- and casein-free counterparts food items in the control diet to insure that the parents and observers were blinded."
Quote: "The experimental diet consisted of gluten- and casein-free counterparts food items in the control diet to insure that the parents and observers were blinded."
Judgement Comment: Blinding of participants and personnel not feasible, and is likely to influence reporting of outcomes
Quote: "the current study did not have the resources to provide GFCF and placebo diets to families. Consequently, parents were aware of their child’s treatment assignment."
Judgement Comment: Open label. Blinding of participants and personnel not feasible, and is likely to influence reporting of outcomes
Quote: "and the data inspectorate, was <b>single blind and randomised. The single blind design was chosen because it has been reported that children with autism that had been on casein-free diet for 8 weeks deteriorated when they were given casein again for 2 weeks (Lucarelli et al., 1995).</b> Participation included only children with"
Judgement Comment: Single blinded. Participants were blinded, but not personnel
Quote: "This randomized double-blind placebo-controlled study was"
Quote: "The research pharmacy prepared and documented both masked dietary supplements. The supplements were equivalent in appearance and taste; thus, they were undetectable to parents, study participants, and study personnel."
Quote: "Aside from study nutritionists, all study members were blinded. Parents could not be blinded because they had to oversee the child’s food intake and consequently knew whether the child was following an exclusion diet."
Judgement Comment: Single-blinded
Detection bias due to knowledge of the allocated interventions by outcome assessors
Quote: "The experimental diet consisted of gluten- and casein-free counterparts food items in the control diet to insure that the parents and observers were blinded. Menus"
Quote: "Each participant’s regular diet was provided during the baseline period and for 6 weeks as the control diet. The experimental diet consisted of gluten- and casein-free counterparts food items in the control diet to insure that the parents and observers were blinded. Menus"
Judgement Comment: Likely blinded
Judgement Comment: Blinding of outcome assessors not feasible and is likely to influence reporting of outcome
Quote: "parents were aware of their child’s treatment assignment. Instead, this study was able to control for clinician attention (by having a dietician meet with control families to develop a healthy diet) as well as for the work that families needed to invest to establish and maintain a diet for their children. The study was single blinded in that the individuals who coded behavioral observations were blind both to treatment and order of videos."
Quote: "The study was single blinded in that the individuals who coded behavioral observations were blind both to treatment and order of videos."
Quote: "Structured interviews regarding autistic traits were done with the parents in their homes. As children with autistic syndromes habituate very slowly to people, settings, and tasks (Bernal and Miller, 1970), formal testing of skills and abilities were done in the school or in the kindergarten the child attended. The testing was carried out in the same room and at the same time of the day that the child usually had special one to one training. The special educator or the child’s special assistant was asked to be present during the testing."
Judgement Comment: Blinding of outcome assessors not feasible, and reporting of outcome is likely to influenced
Quote: "This randomized double-blind placebo-controlled study was performed at"
Quote: "All study participants (subjects and parents) and all investigators involved in assessing the children were blinded to the group assignment (GD(−) diet or GD(+) diet)."
Quote: "Aside from study nutritionists, all study members were blinded."
Quote: "At 8 months, all participants were re-tested by blinded investigators with baseline measures."
Attrition bias due to amount, nature or handling of incomplete outcome data
Quote: "We employed a missing at random model for the three of 15 subjects whose week 12 or week 6 data were incomplete on a major variable (CARS or ECOS)."
Judgement Comment: Loss to FU imbalance, but similar accross groups (4 out of 20 in control group, and 2 out of 17 in the intervention group) Per protocol analysis. No investigation of participants lossed to FU.
Quote: "Figure 1 summarizes average adherence rates for the GFCF and control groups over the 12 week study period. The GFCF group had significantly more dietary infractions at the 12 week period which rendered a significantly lower adherence rate (P-value=.009). However, there was missing data for several of the participants in the control group."
Judgement Comment: Apparrantly there were no missing data
Quote: "College Station, TX, USA). Results <b>Twenty-six parents of children with ASD were asked to participate, 17 gave informed consent, and 5 with- drew the consent after the initial psychological evalu- ation. The study was conducted on the remaining 12</b> Navarro et al. Are ‘leaky"
Judgement Comment: NI about the 5 subjects who withdrew after initial evaluation
Quote: "A per protocol (PP) model formed the basis for participant inclusion in the data analysis in contrast to an intention to treat (ITT) standard. Only observed values were included for repeated measures analyses, which took partially completed profiles into account."
Quote: "Fifteen children (21%) dropped out during the 12- month period (group A, n = 11; group B, n = 4; Fig. 1)."
Quote: "As per our a priori assumptions, the majority of participants who ceased intervention before 12 months primarily did so either because children did not want to engage in a dietary change or parents found the diet too difficult to maintain. This relates, in part, to the concept of a cost–benefit ratio, whereby cost (use of diet) was deemed greater than benefit (effect of diet). A per protocol (PP) model of statistical analysis on the basis of using data from surviving participants adhering to the experimental protocol was, therefore, chosen over an intention to treat (ITT) model. The reasons for this reflect a need to assess efficacy of intervention; that is, does diet lead to improved developmental outcome, alongside continuing debate on the suitability of ITT where data on performance and conformity are available to researchers. Whilst data missing completely at random (MCAR) is the preferred scenario, Feinman 50 illustrates that use of ITT specifically in dietary research can clearly be problematic as a function of the inclusion of non-participating data. Likewise, due to the potential risks associated with unwanted effects, 51,52 LOCF was also not included as part of the statistical plan. Given the small number of testing occasions employed and the use of a repeated measures model taking all measure- ments into account, we assume any bias based on non- inclusion of LOCF is minimal."
Reporting bias due to selective outcome reporting
Judgement Comment: No reference to study protocol, but appears to report on relevant outcomes
Judgement Comment: No reference to study protocol, but appears to report on relevant outcomes
Judgement Comment: No reference to study protocol, and no core symptoms or function outcomes as expected.
Judgement Comment: There is no reference to study protocol, and relevant outcomes are reported
Judgement Comment: No reference to studyp protocol, and no reporting of our prespecified relavant outcomes and no standardized reporting of variance (only raw scores and T scores).
Quote: "The study protocol was approved by the Danish National Committee on Biomedical Research Ethics (reference no. KA0503g). The study was registered with ClinicalTrials.gov, number NCT00614198."
Judgement Comment: There are reported on more outcomes, than prespecified in the protocol
Bias due to problems not covered elsewhere in the table
Judgement Comment: The study appears to be free from other sources of bias
Judgement Comment: The study appears to be free from other sources of bias
Judgement Comment: The study appears to be free from other sources of bias
Judgement Comment: The study appears to be free from other sources of bias
Judgement Comment: The study appears to be free from other sources of bias
Quote: "duration. Conflict of interest statement <b>PW and PS undertook commercial analysis of urine samples from people with ASD and related conditions as part of the University of Sunderland Enterprises Ltd whilst involved in the study. AMK and KLR are members of the Norwegian Protein Intolerance Association. KLR is a consultant to TipoGen, a biotechnology company involved in profiling psychiatric diseases. PS and JJ are parents of children with autism. SP has a sibling with autism.</b> Other authors (DH, ARS, MS,"
Judgement Comment: The study appears to be free from other sources of bias