Campbell et al. (1989) administered naltrexone on an open basis to 10 hospitalized children aged 3.42 to 6.5 years (mean age, 5.04 years). The study lasted 6 weeks. Following a 2-week baseline, single doses of 0.5, 1, and 2 mg/kg/day were administered at 1-week intervals. Ratings were made 1, 3, 5, 7, and 24 hours after each dose and 1 week after the last dose. Subjects showed diminished withdrawal at all three dose levels. Verbal production increased at 0.5 mg/kg/day, and stereotypies decreased following the 2 mg/kg/day dose. Symptoms such as aggressiveness and “self-aggressiveness” showed little improvement. The major untoward effect was mild sedation, which occurred in 70% of the subjects. Laboratory measurements, including liver function tests and electrocardiograms (ECGs), showed no significant change from baseline. Overall, raters considered 80% of the children to be positive responders for some symptoms (Campbell et al., 1989).

Campbell et al. (1990) subsequently conducted a double-blind, placebocontrolled study of naltrexone in 18 children, aged 3 to 8 years, diagnosed with autistic disorder. The study consisted of a 2-week placebo baseline phase, random assignment to placebo or naltrexone for 3 weeks, and a posttreatment 1-week
placebo phase. The initial naltrexone dose was 0.5 mg/kg/day; this was increased to 1 mg/kg/day, if no adverse effects occurred. Nine children received naltrexone; the optimal dose was 1 mg/kg/day. Six subjects receiving naltrexone were rated moderate (five) or marked (one) in improvement on Clinical Global Consensus Ratings, whereas only one child on placebo achieved a moderate rating and none was markedly improved. The difference was significant (P = .026). In contrast, no reduction in symptoms occurred on the Children’s Psychiatric Rating Scale or Clinical Global Impressions. Naltrexone did not appear to affect discrimination learning in an automated laboratory. The authors also reported that overall symptom reduction seemed better in older autistic children than in younger ones.

Although there are case studies and open studies with some encouraging data, the 1993 report of Campbell et al.—an 8-week double-blind study in which 41 hospitalized children (2.9 to 7.8 years of age; mean, 4.9 years) diagnosed with autistic disorder were treated with naltrexone or placebo—did not support the efficacy of naltrexone in this population. All their subjects received placebo during the first 2 weeks while baseline data were obtained. Following this phase, subjects were randomly assigned to naltrexone or placebo for the next 3 weeks. During the final week, all subjects again received placebo. Twenty-three patients were assigned to the naltrexone group and 18 to the placebo group. The initial dose was 0.5 mg/kg/day of either placebo or active drug given in the morning; dose was increased to 1.0 mg/kg/day after 1 week and maintained at that level because untoward effects were minimal and did not require a reduction in dose. Naltrexone did not improve the core symptoms of autism. The only significant finding was a modest decrease in hyperactivity on three different measures. It did not improve discrimination learning significantly more than placebo. Naltrexone was no better than placebo in reducing self-injurious behavior, but six of eight subjects who had a severity rating of mild or above on the Aggression Rating Scale who received naltrexone experienced rebound (increase) in symptoms during the final placebo period; only one child in the placebo group exhibited worsening of self-injurious behavior during that time. The authors concluded that it remains to be determined whether naltrexone is efficacious in treating moderate-to-severe self-injurious behavior and that its use cannot be recommended as a first-line treatment for patients diagnosed with either autistic disorder or self-injurious behavior (Campbell et al., 1993).

In a 7-week, double-blind, placebo-controlled, crossover study, Feldman et al. (1999) evaluated the efficacy of naltrexone in improving communication skills, a core deficit, in 24 children (mean age, 5.1 years; range, 3 to 8.3 years) diagnosed with autistic disorder by DSM-III-R criteria (American Psychiatric Association [APA], 1987) who had previously shown modest behavioral improvement on naltrexone in previous studies by the authors (Kolman et al., 1995, 1997). Communication skills of the subjects at baseline ranged from preverbal to nearly normal for age. During the active drug phase, 1 mg/day of naltrexone was administered.

There was no significant improvement in communication skills with naltrexone treatment, including number of utterances, total number of words, number of different words, or reduction in echolalia in these subjects who had shown some behavioral improvement on naltrexone. Also, the authors reported that use of parental language with the patient did not change according to whether the child was receiving naltrexone. The authors suggested that medications that improve core deficits and target symptoms of autistic disorder should be preferred over those that improve only associated symptoms.