Gaining: Pediatric Patients and Use of Atypical Antipsychotics

In this edition of the Journal, Klein and co-workers present a treatment trial showing benefits of metformin for children and adolescents who experienced rapid weight gain, a risk factor for type II diabetes, while taking atypical antipsychotic medications. A placebo-controlled study in adults (1) and an open trial in children (2) have appeared, but this is the first double-blind, placebo-controlled trial in pediatric patients taking atypical antipsychotics.

The report is extremely timely. The prescription rate of atypical antipsychotics in children is following a familiar cycle—initial gradual use followed by accelerating, widespread practice. Awareness of cardiovascular and metabolic risks has yet to cool clinicians’ enthusiasm. Recent studies of overall pediatric use have shown a 6- to 20-fold increase in prescription of atypical antipsychotics in four state Medicaid programs (3) and, nationally, a sixfold increase in pediatric visits that included prescriptions of antipsychotic medication (4), more than 90% of which were prescriptions of atypical antipsychotics (4) . Certainly, there is good evidence that atypical antipsychotics have a place in the treatment of some pediatric symptoms—psychosis, mania, tics, and irritability or aggression in the context of autism-spectrum disorders. However, greater experience has also brought increasing evidence of and concern about weight gain and development of type II diabetes (5) . Data suggest that children may be particularly vulnerable to this ill effect (6), and several studies (5) now point to significant weight gain in 50%–60% of children treated with atypical antipsychotics. Rapid weight gain during treatment with atypical antipsychotics is unsafe and a barrier to adherence for children and adolescents. Medication-induced weight gain in children also must be viewed in the current climate of public health concern about obesity and its sequelae, including cardiovascular illness. Having developmental antecedents is a characteristic shared by psychiatric disorders and obesity. Recognition that adult obesity, cardiovascular illness, and type II diabetes have roots early in development is prompting extensive prevention efforts with children and adolescents.

Ultimately, in 2004, worrisome information on weight gain and development of type II diabetes from adult trials of atypical antipsychotics led the Food and Drug Administration (FDA) to require additional warnings stressing these risks in the package insert and inpatient instructions.

Can metformin stem type II diabetes in children related to atypical antipsychotics? The add-on, parallel-design study by Klein and colleagues enrolled 39 participants, 10–17 years old, who had gained more than 10% of their pretreatment body weight in less than 12 months. The participants were randomly assigned to receive metformin or placebo for 16 weeks while they continued their atypical antipsychotic at a stable dose. Eligibility was independent of diagnosis or specific antipsychotic agent. All participants and families received the same three sessions of generic dietary counseling. During the 16 weeks, those assigned to placebo gained an average of 9 lb and had an average increase in body mass index (BMI) of 1.1 kg/m ², compared to no change in weight and an average decline of 0.4 kg/m ² in BMI for the metformin recipients. At week 16, participants were referred for oral glucose tolerance testing, by a clinician who was blind to treatment status, if they had either an excessive BMI (higher than the 95th percentile) or a further increase of more than 10% in weight plus a high fasting serum glucose level. Ten (50%) of the 20 placebo-assigned participants required oral glucose tolerance testing, compared to two of the 18 metformin-treated participants. The results are notable for the clear differences between treatment groups in average weight gain, absolute BMI, corrected z score for BMI, and waist circumference. In addition, according to homeostasis model assessment, there were impressive changes in measures of insulin resistance—an average increase of 1.52 in the participants in the placebo group compared to an average decline of 0.71 in the metformin group.

The study has a number of notable strengths. The study participants were like patients clinicians see in consulting rooms and inpatient units each day, from different diagnostic groups, and taking combinations of medications in addition to atypical antipsychotics. The investigators used a randomized-allocation, placebo-controlled design and enrolled a number of participants that was sufficient, although not overwhelmingly large, to answer the question. The duration of treatment was adequate; the outcome measures of insulin resistance, BMI z score, and waist circumference were sensible; and the data were analyzed by including the last observation carried forward. The study limitations include a “dropout” rate of 21% (eight of 38 altogether; five of 20 placebo patients, three of 18 metformin patients), which may threaten validity, and the failure to assess compliance throughout the study. Finally, this was a short-term study, focused exclusively on this one pharmacological intervention.

This is a welcome investigation addressing a serious problem. Some pediatric patients who are responsive to atypical antipsychotics undergo a shocking increase in weight. These findings are good news for them and their families, although any zeal should be tempered. The study does not address the long-term safety of metformin treatment or whether weight loss is sustained. This is important because vulnerable children may require metformin treatment as long as they continue to take atypical antipsychotics, typically for symptoms that last many years. Metformin, a relative of pheformin, has long-term side effects that lead to B ₁₂ and folate deficiency (7) in up to 30% of chronic users. Fatal lactic acidosis with metformin has been seen in the context of uncompensated congestive heart disease, acute renal insufficiency, and liver disease (8) and may emerge with chronic use of nonsteroidal anti-inflammatory drugs (9) . It is reassuring that these conditions are relatively rare in children and adolescents. In addition, the study did not compare pharmacological treatment to specific, targeted treatments for dietary control that combine dietary counseling, exercise, and behavioral intervention, which have been shown to be effective in adults (10, 11) . Just as patients in the metformin trial missed doses of medication, behavioral programs also have problems with patients failing to adhere to their diet or exercise schedules on occasion. Finally, it is noteworthy for our field that such a high proportion of the cohort had diagnoses or symptoms for which there is no clear FDA-approved indication for treatment with atypical antipsychotics (12, 13) —17 (44%) of the 39 had “attentional disorders,” and seven (18%) had oppositional defiant disorder.

Further research may change views on the role of metformin with use of atypical antipsychotics, but there is nothing in the current study to indicate that metformin should be prescribed at the beginning of atypical antipsychotic treatment to prevent weight gain. One-half of the patients taking placebo, all of whom were enrolled because they exhibited initial rapid weight gain from atypical antipsychotics, did not need oral glucose tolerance testing or any further intervention 16 weeks later. Certainly, some of these individuals may continue to gain weight and require testing later on. But it is also possible that they will maintain their current weight and require no more than ongoing dietary counseling.

Specific recommendations emerge from this report. First, when prescribing atypical antipsychotics to children, physicians should monitor each patient’s weight for rapid increases or development of obesity (BMI >95th percentile). Second, insulin sensitivity and oral glucose tolerance testing is warranted for children, particularly those with a family history of type II diabetes, who develop these risk factors. Third, if rapid weight gain or obesity emerges and reduction in the dose of the atypical antipsychotic is impossible, gradual introduction and increment of metformin may be warranted. This should be done in close collaboration with the child’s regular pediatrician or family medical doctor. Final doses between 1500 and 2000 mg/day are recommended with monitoring of blood indices and liver and renal function and instruction about B ₁₂ and folate deficiency.

Klein and colleagues have provided a valuable study suggesting that metformin may be a worthwhile step for those patients who benefit from atypical antipsychotics but pay a cost of significant weight gain or insulin resistance. It is too soon to know whether long-term use of metformin will be a successful solution or lead one problem to become another. At this point, cautious application can be prudent after a thorough, systematic review of the dose and need for atypical antipsychotics.