TABLE 1

Results of Systematic Literature Review

Study and GradeSampleIntervention and DurationMeasures Used to Arrive at ConclusionResultsConclusions
Pharmacologic interventions
 Aman et al, 200541Included: 101 children aged 5 to 17 y from Research Units on Pediatric Psychopharmacology Network with ASD16 wk (nonresponders) to 6 mo (responders) given placebo or risperidoneParent report sleep logSleep problems and anxiety less common in risperidone group (P = .02; P = .05). Average sleep duration increased short-term (17 min; 40 min) but not beyond 6 mo (29 min). Adverse events scored as moderate (placebo versus risperidone, respectively): somnolence (12% vs 37%), enuresis (29% vs 33%), excessive appetite (10% vs 33%), rhinitis (8% vs 16%), difficulty waking (8% vs 12%), and constipation (12% vs 10%)Risperidone improves sleep latency in children with ASD but not sleep duration; high rates of adverse outcomes
  RCT, level IIExcluded: Mental age <18 mo, positive β-human chorionic gonadotropin test result for girls, significant medical condition, previous trial with risperidone, history of neuroleptic malignant syndrome, and/or weight <15 kg20 to 44.9 kg = 0.5 mg up to maximum of 2.5 mg/dayAbnormal Involuntary Movement Scale
< 20 kg = similar, except slower dosing ≥45 kg started at 0.5 mg nightly then titrated to a maximum of 3.5 mg/day in divided dosesAdverse events scored as mild or moderate/severe (parent report)
 Honomichl et al, 200242Included: 17 children aged 3 to 8 y with autism recruited as subset of larger study (California)4 wk placebo and 4 wk intervention 2 CU/kg porcine secretin intravenously; no washoutCSHQSecretin did not significantly affect CSHQ scores: Bedtime resistance (Pre, mean: 9 [range: 6–17]; Post, mean: 8 [range: 6–13])Secretin does not significantly improve sleep-onset delay, duration, bedtime resistance, and night wakings in children with ASD
  RCT, level IIIExcluded: not specifiedSleep diary (bedtime, sleep onset, time/duration night waking, morning rise)Sleep-onset delay (Pre, mean: 2.1 [range: 1–3]; Post, mean: 1.6 [range: 1–3])
Night wakings (Pre, mean: 4.1 [range: 3–7]; Post, mean: 3.6 [range: 3–5])
Sleep duration (Pre, mean: 4.7 [range: 3–7]; post, mean: 4.3 [range: 3–8])
 Ellaway et al, 200143Included: experimental group consisted of 21 females aged 7 to 41 y; recruited from previous 8-wk trial; control group included 62 females aged 4 to 30 y recruited from Australian Rett Syndrome Register100 mg/kg liquid l-carnitine twice a day for 6 moRett syndrome Symptoms Severity IndexCompared with Rett syndrome controls, l-carnitine significantly improved sleep efficiency (P < .03) but not duration (P = .57), latency (P = .15), daytime sleep (P = .55), or night wakings (P = .25)l-carnitine supplementation improves sleep efficiency but not duration, latency, daytime sleep, or number of night wakings in children with Rett syndrome
  Pre/post-control, level IIIExcluded: Not specifiedSF-36 Health Survey
Hand apraxia scale
7-d sleep diary
TriTrac-R3D Ergometer (Hemokinetics, Inc, Madison Wisconsin)
 Posey et al, 200144Included: 26 children and young adults, aged 3 to 23 y with ASD (Indiana); 20 had comorbid intellectual disability. Mirtazapine was prescribed to target symptoms of aggression, self-injury, irritability, anxiety, depression, insomnia, and interfering repetitive behaviorMirtazapine starting dose of 7.5 mg daily with dosage increases made in 7.5-mg increments up to a maximum of 45 mg daily in divided doses, based on response of target symptoms and side effects (range: 7.5–45 mg). Treatment duration ranged from 11 to 368 d (mean duration: 150 ± 103 d)Clinicians used the CGI scale to rate severity and improvement. A modified CGI improvement item assessing sleep quality was also includedNine participants responded (defined according to a CGI score of much improved or very much improved). Five of 9 responders showed improved sleep. Of 17 nonresponders, 3 showed improved sleep. Statistically significant improvement was seen on the CGI severity ratings of sleep (P = .002)Mirtazapine is effective in improving sleep quality in children with ASD
  Pre/post-no control, level IIIExcluded: Not specified
 Rossi et al, 199945Included: 25 children aged 2 to 20 y with ASD (University of Bologna) with continuous presence (≥1 y) of severe mood disorders, aggressiveness, and hyperkinesiaNiaprazine 1 mg/kg per day TID for 60 dBehavioral Summarized Evaluation (included sleep disorders; difficulty falling asleep, night wakings, and early waking. Scored 0 to 4 (absent to very severe)Compared with start of the treatment, Behavioral Summarized Evaluation showed improvement in sleep disorders by the end of the trial (P < .001)Niaprazine is effective in improving sleep latency and night wakings in children with ASD
  Pre/post-no control, level IIIExcluded: children nosographically defined congenital or acquired encephalopathy
 Ming et al, 200846Included: 19 children aged 4 to 16 y with ASD (New Jersey) and sleep and behavioral disordersClonidine 1 time per day, initial dose 0.05 mg and gradually advanced to 0.1 mg; oral tablet. Duration for 6 mo to 2 yCaregiver report (average bedtime, latency, number of wakings) before and during treatmentSixteen of 17 children experienced improved sleep initiation (2–5 h before treatment, 0.5–2 h during treatment); 16 of 17 with sleep maintenance disorders experienced improvements; 5 still experienced night wakingsClonidine improves sleep latency and wakings in children with developmental disorders
  Retrospective case control, level IIIExcluded: not specified
Other biologic agents
 Adams and Holloway, 200447Included: 25 children aged 3 to 8 y with ASD with no changes in treatment therapies in previous 2 mo, and no previous use of multivitamin supplements other than standard children’s multivitamin (Arizona)Days 0 to 24: 1/8 dose of SS-II (standard multivitamin), increased linearly to maximumCGISleep and gastrointestinal symptoms as evidenced by CGI score improved compared with placebo (P = .03)Moderate-dose multivitamin (including vitamin B6 and vitamin C) may have a positive impact on CGI sleep score in children with ASD
  RCT level IIExcluded: not specifiedDays 25 to 34: held maximum dose of SS-IIBlood and urine sample (plasma B6, pyridoxine, pyridoxal, pyridoxamine, α−lipoic acid, vitamin C)
Days 35 to 50: gradual transition to SS-III
Days 50 to 90: continued SS-III; full dose 1 ml/5 lb TID with food for a total daily intake of 3 ml/5 lb body weight
 Dosman et al, 200748Included: 33 children aged 2 to 5 y diagnosed with autism with ferritin measured previously (Toronto)6 mg elemental iron/kg per day for 8 wkSleep Disturbance Scale for ChildrenPer the sleep disturbance scale scores, restless sleep improved significantly (29%; P = .04) and 35% had improved sleep latency; no relation was found with ferritin (P = .61), irritability (P = .83), or Periodic Leg Movements During Sleep Scale (P = .82)Oral iron supplementation improves restless sleep in children with ASD but does not affect irritability, sleep latency, or periodic leg movements
  Pre/post-control, level IIExcluded: currently taking iron supplementationIf not tolerated, received 60 mg/day of microencapsulated powdered elemental ironPeriodic Leg Movements During Sleep ScaleMean ferritin and MCV increased significantly (16 to 29 μg/L)
Blood sample (MCV, mercury, albumin, vitamin B12, serum ferritin, transferrin)
 Wright et al, 201149Included: 20 children, aged 4 to 16 y with ASD and prolonged sleep latency, excessive night waking, and/or reduced total sleep time. Children had undergone behavior management that was not successful and were free of psychotropic medicationsMelatonin 2 mg, 30 to 40 min before planned sleep, or placebo and then crossed over to the other agent. The dose was increased by the parent every 3 nights by 2 mg to a maximum dose of 10 mg until “good” sleep was achieved, defined as an improvement of ≥50%. Treatment phase lasted 3 mo for each agent with a 1 mo washoutDaily sleep diaries (time melatonin taken, bedtime, sleep time, night wakings, time awake) were collected every month for 9 moMean sleep latency was lower (P = .004) and total sleep time was longer (P = .002) during melatonin treatment compared with placebo. Night wakings did not improve (P = .2)Melatonin improves sleep latency and total sleep time in children with ASD but not night wakings
  RCT, level IIIExcluded: children previously or currently on melatonin, those currently on psychotropic medications, and those with other neurodevelopmental disorders such as Fragile X or Rett syndromeSleep difficulties questionnaireDyssomnia subscale improved with treatment (P = .04) but not parasomnia, sleep apnea, or other sleep disorders
 Garstang and Wallis, 200627Included: 7 children aged 4 to 16 y with autism and sleeping difficulty (at least 1 h sleep latency and/or night waking 4 times per week for the last 6 mo) (United Kingdom)Placebo or 5 mg melatonin for 4 wk, washout 1 wk, then reversed for 4 wkParental sleep charts (total sleep time, sleep latency, night wakings, morning awakening)Baseline, placebo, and melatonin values respectively: Sleep latency improved 1.54 h versus baseline (2.6 h [CI: 2.28–2.93]; 1.91 h [CI: 1.78–2.03]; 1.06 h [CI: 0.98–1.13])Melatonin is an effective treatment to improve sleep latency number of night wakings and sleep duration
  RCT, level IIExcluded: children previously/currently using melatonin and/or taking sedative medications for <4 wkWakings per night decreased by 0.27 per night versus baseline (0.35 [CI: 0.18–0.53]; 0.26 [CI: 0.20–0.34]; 0.08 [CI: 0.04–0.12])
Duration increased by 1.79 h versus baseline (8.05 h [CI: 7.65–8.44]; 8.75 h [CI: 8.56–8.98]; 9.84 h [CI: 9.68–9.99])
 Wirojanan et al, 200928Included: 18 children aged 2 to 15 y with autism and/or Fragile X syndrome who reported sleep problems (California)Placebo or 3 mg melatonin given 30 min before bedtime for 2 wk, then reversed for 2 wk (no washout)Sleep diary (sleep latency, duration, sleep onset, number of night wakings)Compared with placebo, sleep duration improved in 9 of 12 participants by 21 min (P = .057; effect size: 2.12), sleep latency improved in 11 of 12 participants by 28 min (P = .10; effect size: 1.79), sleep onset improved by 42 min (P = .017; effect size: 2.80), and awakenings were improved by 0.07 but were insignificant (P = .73; effect size: 0.3540)Melatonin is an effective treatment to improve sleep duration, sleep latency, and sleep onset time but not night wakings
  RCT level IIExcluded: not specifiedActiwatch (Philips Respironics, Bend, Oregon)
 Paavonen et al, 200350Included: 15 children 5 to 17 y diagnosed with Asperger’s syndrome and severe sleep problems in last 3 mo (Helsinki)Received 3 mg of melatonin 30 min before bed for 14 dChildren's Self-Report for Sleep ProblemsMelatonin improved nocturnal activity (31.39 ± 7.86 to 18.74 ± 4.99; P = .041) and sleep latency (from 40.02 ± 24.09 to 21.82 ± 9.64; P = .002) but increased number of wakings (from 15.14 ± 6.12 to 17.85 ± 6.25; P = .048). No significant changes were found in sleep efficiency (85.13 ± 5.57 to 86.03 ± 4.62; P = .331) or duration (477.40 ± 55.56 to 480.48 ± 50.71; P = .572).Melatonin is an effective treatment to improve nocturnal activity and sleep latency but not night wakings, sleep efficiency, or duration in children with Asperger’s syndrome
  Pre/post-control, Level IIExcluded: children taking psychotropic medications or with major psychiatric comorbiditySleep Disturbance Scale for Children
CBCL
Karolinska Sleepiness Scale
Teacher’s Daytime Sleepiness Questionnaire
Actigraphy
 Giannotti et al, 200651Included: 25 children aged 2.6 to 9.6 y with autism and CARS score <29.5 with sleep disorder (>45 min sleep latency, >3 times per week night wakings, and waking before 5 am >3 times per week) (Italy)6- to 24-mo (n = 25 and n = 16, respectively) studyCSHQCSHQ values from baseline to 6 mo were significantly improved (65 vs 43; P < .001). CSHQ values regarding latency, duration, resistance, anxiety, night wakings, and daytime sleepiness were significantly improved (P < .001). Sleep-disordered breathing also improved (P < .01). Sleep diaries showed improvements in duration by 2.6 h (P < .001), awake time after sleep onset by from 70 to 10 min (P < .001), cosleeping in 55% (P < .001), parental presence at bedtime in 63% (P < .001), and bedtime irregularity in 61% (P < .001). Those who continued 24 mo showed significant improvement in CSHQ scores in years 1 and 2 (63 vs 44 vs 44; P < .001)Melatonin is an effective treatment to improve sleep duration, latency, number of night wakings, and bedtime resistance in children with ASD
  Pre/post-control, level IIExcluded: children with autism diagnosis but with CARS score below cutoff, coexisting conditions, and/or taking medications for 6 mo previously3 mg controlled-release melatonin 30 to 40 min before bedtime (1 mg FR and 2 mg CR)Sleep diary (bed time, rise time, duration, night wakings and duration, day naps)
Children advised to give 2 mg FR is children awoke for >15 min during the night
Physician could increase dose to maximum of 4 mg in children aged <4 y and up to 6 mg in children aged >4 y
 Malow et al, 201152Included: 24 children aged 3 to 9 y with ASD and sleep-onset delay of ≥30 min on ≥3 nights per weekTwo-week acclimation phase: inert liquid 30 min before bedtimeActigraphy; CSHQWith melatonin treatment (compared with acclimation phase), statistically significant improvements were seen in sleep latency (42.9 to 21.6 min; P < .0001) but not total sleep time or wake time after sleep onsetMelatonin is an effective treatment of sleep latency, and improves aspects of daytime behavior and parenting stress in children with ASD. It is safe and well tolerated
  Pre/post-control, Level IIExcluded: children with epilepsy or taking psychotropic medications. Before melatonin treatment, medical comorbidities were addressed, and parents received sleep education trainingOptional escalating dose protocol based on 3-wk periods. Dose was escalated from 1 mg to 3 mg to 6 mg based on response (if a satisfactory response occurred, defined as falling asleep within 30 min on ≥5 nights per week by actigraphy, the dose was not escalated)CBCL, RBS, PSI. Laboratory findings (CBC, metabolic profile including liver and renal function, corticotropin, cortisol, estrogen, testosterone, FSH, LH, and prolactin). Hague Side Effects ScaleSignificantly significant improvements were also noted in CSHQ subscales of sleep-onset delay and sleep duration, CBCL subscales of withdrawn, attention-deficit/hyperactivity, and affective, RBS stereotyped and compulsive, and the Difficult Child subscale on the PSI. No change in laboratory findings. Loose stools in 1 child; no other adverse effects
 Andersen et al, 200830Included: 107 children aged 2 to 18 y diagnosed with autism previously recommended to take melatonin sleep disorder (Tennessee)Aged <6 y, .75-1 mg melatonin 30 to 60 min before bed. After 2 wk, if no response, increased by 1 mg every 2 wk up to 3 mgChart review of clinic notes (including sleep hygiene, other psychiatric conditions, severity of ASD, use of medications)After initiation, 25% with sleep problems no longer a concern; 60% had improved sleep but still had concerns; for 13% sleep problems remained a major concern, and 1% reported worse sleep after treatment. Three children had adverse effects (morning sleepiness and/or increased enuresis)Melatonin is a safe, effective treatment to reduce sleep problems in children with ASD
  Pre/post-No control, level IIIExcluded: children with a diagnosis of bipolar disorderAged ≥6 y, 1.5 mg melatonin 30 to 60 min before bed time If no response, increased dose to 3 mg after 2 wkSleep diary
All children: if no response after 4 wk, increase dose to 6 mg
Behavioral/educational interventions
 Reed et al, 200931Included: 20 families of children aged 3 to 10 y with clinical diagnosis of ASD with sleep concerns (Tennessee)• Three 2-h sessions for 3 consecutive weeks; follow-up 1 mo after endCSHQEducational intervention showed improved CSHQ scores for bedtime resistance (P = .001), latency (P = .004), duration (P = .003), and sleep anxiety (P = .022) but not night wakings (P = .508), parasomnia (P = .607), sleep-disordered breathing (P = .625), and daytime sleepiness (P = .096)Educational intervention with parents improves bedtime resistance, latency, duration, and sleep anxiety but not night wakings, early morning waking, parasomnias, sleep disordered breathing, or daytime sleepiness in children with autism
  Pre/post-no control, level IIExcluded: children with primary sleep disorders such as sleep apnea, narcolepsy, and neurologic/medical conditions that may contribute to disordered sleepSession 1: established daytime and night time habits and a bedtime routine based on FISH and CSHQActigraphyActigraphy showed improved sleep latency (62.2 ± 33.33 min vs 45.6 ± 27.6; P = .039) but not waking after sleep onset (24.5± 9.8 vs 32.2 ± 24.7 min; P = 1.0)
Session 2: Strategies to minimize night waking and early morning wakingLatency, duration, night wakings71% of parents reported fewer nights cosleeping; 33% reported improvement in early morning waking
Session 3: address individual sleep concerns
 Weiskop et al, 200532Included: 13 children (across 13 families) aged 5 y with either Fragile X syndrome (n = 7) or ASD (n = 6) with perceived sleep difficulty (Australia)Five sessions over 7 wk with weekly telephone calls; follow-up at 3 and 12 moSleep Diary (behavior, lights out, sleep onset, waking, cosleeping, morning wake time)Baseline compared with intervention: 4.6% moderate deterioration of sleep behaviors, 25% no change, 29.7% moderate improvement, and 40.6% substantial improvementsEducational intervention with parents improved sleep latency and night wakings but not duration in children with developmental disabilities
  Case series, level IIIExcluded: excluded if diagnosed with epilepsy and/or if diagnosed with ASD, were not to be taking medicationSession 1: goal settingBaseline compared with 3 mo: 1.6% substantial deterioration, 4.8% moderate deterioration, 27% no change, 23.8% moderate improvement, and 41.3% substantial improvement
Session 2: learning theory; antecedents and consequences. Created intervention with reinforcement and visual representationBaseline compared with 12 mo: 7.7% moderate deterioration, 19.2% no change, 26.9% moderate improvement, and 46.2% substantial improvement
Session 3: effective instructions and partner support strategiesSleep latency improved in 6 of 10 (60%)
Session 4: extinction techniquesNight waking improved for 7 of 10 (70%)
Session 5: review sessionDuration was variable and unchanged
Cosleeping was also addressed100% of parents reported improved sleep but 50% still considered sleep an issue
Complementary and alternative medicine
 Piravej et al, 200933Included: 60 children aged 3 to 10 y diagnosed with autism; 30 were put into control group (SI) and 30 were experimental (SI + TTM) (Thailand)Two 60-min standard SI or Two 60-min TTM and SI for 8 wkSleep diaryMassage improved sleep scores (11.5 vs 5.3; P < .001). Standard SI improved sleep behavior (13.9 vs 8.2; P < .001). The difference between pre- and post-sleep behavior scores of the control and massage groups were not statistically significant (5.7 vs 6.3 respectively; P = .85)TT, before bedtime improves sleep in children with ASD but not more than compared with standard SI
  RCT, level IIExcluded: those with contraindications for TTM and those unable to complete 80% of treatment (13 massages)Two 60-min TTM and SI for 8 wk
 Williams, 200634Included: 12 children aged 12 to 15 y diagnosed with ASD from a residential school for children with autism (United Kingdom)Three administrations of aroma therapy (2% lavender oil in grape seed oil) over 24 d via massage of foot and leg ∼2 hours before bedSleep diary recorded by staff on 30-min interval rounds (sleep onset, duration, wakings)Participants did not demonstrate statistically significant sleep-onset time (F = 1.27; df = 4.15, 41.5; P = .30). Night wakings did not differ with and without aromatherapy (χ2 = 20.19; df = 16; P = .21). Sleep duration was not affected by aromatherapy (F = 0.59; df = 16, 160; P < .89)Aroma therapy is not an effective treatment to affect sleep onset, duration, and night wakings
  Case series level IIExcluded: none specified
 Escalona et al, 200135Included: 20 children aged 3 to 6 y recruited from school for children with autism (Florida)For 1 mo, received 15-min massage therapy by parents (trained by therapist) or 15 min of reading before bedtimeSleep diaries (fussing, restlessness, crying, self-stimulating behavior, and number of times child left the bed)Greater declines for the massage therapy group with regard to fussing/restlessness, crying, self-stimulating behavior, and getting out of bed (actual numbers not provided; only provided statistics for day-time behavior)Massage before bed decreases fussing/restlessness, crying, self-stimulating behavior, and getting out of bed in children with ASD more than reading
  RCT, level IVExcluded: not specified
  • CARS, Childhood Autism Rating Scale; CBC, complete blood cell count; CBCL, Child Behavior Checklist; CGI, Clinical Global Impression; CI, confidence interval; CR, controlled release; FISH, Family Inventory of Sleep Habits; FR, fast release; FSH, follicle-stimulating hormone; LH, luteinizing hormone; MCV, mean corpuscular volume; PSI, Parental Stress Index; RBS, Repetitive Behavior Scale; SI, Sensory Integration; SS, Spectrum Support; TTM, Thai Traditional Massage.