Medical Cannabinoids in Children and Adolescents: A Systematic Review
- Shane Shucheng Wong, MD and
- Timothy E. Wilens, MD
Dr Wong conceptualized and designed the study, collected the data, conducted the initial analyses, and drafted the initial manuscript; Dr Wilens conceptualized and designed the study and supervised data collection; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
CONTEXT: Legalization of medical marijuana in many states has led to a widening gap between the accessibility and the evidence for cannabinoids as a medical treatment.
OBJECTIVE: To systematically review published reports to identify the evidence base of cannabinoids as a medical treatment in children and adolescents.
DATA SOURCES: Based on Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, a search of PubMed, Medline, and the Cumulative Index to Nursing and Allied Health Literature databases was conducted in May 2017.
STUDY SELECTION: Searching identified 2743 citations, and 103 full texts were reviewed.
DATA EXTRACTION: Searching identified 21 articles that met inclusion criteria, including 22 studies with a total sample of 795 participants. Five randomized controlled trials, 5 retrospective chart reviews, 5 case reports, 4 open-label trials, 2 parent surveys, and 1 case series were identified.
RESULTS: Evidence for benefit was strongest for chemotherapy-induced nausea and vomiting, with increasing evidence of benefit for epilepsy. At this time, there is insufficient evidence to support use for spasticity, neuropathic pain, posttraumatic stress disorder, and Tourette syndrome.
LIMITATIONS: The methodological quality of studies varied, with the majority of studies lacking control groups, limited by small sample size, and not designed to test for the statistical significance of outcome measures. Studies were heterogeneous in the cannabinoid composition and dosage and lacked long-term follow-up to identify potential adverse effects.
CONCLUSIONS: Additional research is needed to evaluate the potential role of medical cannabinoids in children and adolescents, especially given increasing accessibility from state legalization and potential psychiatric and neurocognitive adverse effects identified from studies of recreational cannabis use.
- CBD —
- CINV —
- chemotherapy-induced nausea and vomiting
- FDA —
- Food and Drug Administration
- PRISMA —
- Preferred Reporting Items for Systematic Reviews and Meta-Analyses
- PTSD —
- posttraumatic stress disorder
- RCT —
- randomized controlled trial
- THC —
Cannabis is a plant that produces pharmacologically active cannabinoids, of which the constituents cannabidiol (CBD) and tetrahydrocannabinol (THC) are the most studied.1 CBD may function via a variety of mechanisms, including indirect antagonism and potentiation of cannabinoid receptors, whereas THC acts primarily as a partial agonist to cannabinoid receptors.
Within the THC class of cannabinoids, δ-9-THC is the primary form found in cannabis, whereas δ-8-THC is prepared by cyclization and has less psychotropic potency.
Currently, there are 2 synthesized cannabinoids that the Food and Drug Administration (FDA) has approved as medications in the United States, dronabinol and nabilone, both of which mimic δ-9-THC. These 2 medications are the only current cannabinoids available by physician prescription. For pediatric populations, dronabinol dosage for chemotherapy-induced nausea and vomiting (CINV) is the same as for adults. However, dronabinol use for AIDS-related anorexia as approved in adults is not recommended in children because of a lack of pediatric studies, with further caution recommended because of psychoactive effects.2 Similarly, the use of nabilone is cautioned in pediatric patients because of psychoactive effects and a lack of established safety and effectiveness.3
On the other hand, naturally derived products from cannabis include marijuana (dried leaves and flowers that are most commonly smoked) and oral cannabinoid extracts, and such products have varying concentrations of cannabinoids (eg, CBD and THC) depending on the strain of the plant. There are also 2 plant-derived cannabinoid medications with standardized THC and CBD content currently undergoing FDA-regulated clinical trials, nabiximols and a CBD oral solution (See Table 1 for a summary of cannabis products).
Because of state legalization of medical marijuana, the medical use of naturally derived products from cannabis, including marijuana and oral cannabinoid extracts, is now legal in more than half of US states via physician certification. All states with operational medical marijuana programs allow use by minors but require consent from a legal guardian and certification from a physician.4 Certain states require the consenting guardian to control the acquisition, dosage, and frequency of use (ie, AK, AZ, HI, ME, MI, NH, NM, NY, OR, RI, and Washington, DC). Additionally, some states require a second physician for the certification of a minor’s use (ie, CT, CO, DE, FL, IL, MA, ME, MI, MT, NH, and NJ), including 4 states that require specific certification from a pediatrician (ie, MA and NH), pediatric subspecialist (ie, DE), or pediatrician and psychiatrist (ie, NJ).
The legalization of medical marijuana has led to a widening gap between its accessibility and the limited evidence base for medical cannabinoids as a treatment of pediatric populations. Currently, the American Academy of Pediatrics opposes dispensing medical cannabis to children and adolescents outside the regulatory process of the US FDA, although the Academy does recognize that cannabis may currently be an option for cannabinoid administration for children with life-limiting or severely debilitating conditions and for whom current therapies are inadequate.5 The purpose of this review is to systematically examine the current evidence for using cannabinoids as a medical treatment in children and adolescents.
A systematic review of the literature on medical cannabinoids in children and adolescents was performed according to the Preferred Reporting Items for Systematic Reviews and Meta- Analyses (PRISMA) statement.6 Medline, PubMed, and the Cumulative Index to Nursing and Allied Health Literature were searched for studies published from 1948 to 2017 and indexed by May 2017 by using the following medical subject heading terms and keywords (listed alphabetically): “cannabinoids,” “cannabis,” “CBD,” “δ-8-THC,” “dronabinol,” “marijuana,” “marijuana smoking/therapy,” “marijuana smoking/therapeutic use,” “medical marijuana,” “nabilone,” and “THC-CBD combination.” Each was cross-referenced with child, adolescent, or pediatric keywords (see Fig 1 for a sample search strategy with Boolean search parameters).
Given the preliminary stage of research in this area, only minimal exclusion criteria were used. Studies were included if they were primary research that reported original data, examined the benefits of cannabis for a clinical indication (ie, all medical disorders), in English, and comprised of a child and adolescent patient sample. Studies were excluded if the majority of the sample was older than 18 years or if age and/or data for children and adolescents were not reported separately.
One independent reviewer (S.S.W.) assessed study eligibility by screening the titles, abstracts, and full-text articles in a standardized manner. Both investigators for final inclusion then reviewed the resulting full-text articles, with summarized information focusing on details such as clinical indication, cannabinoid type, sample characteristics, methodological design, and outcome. For cases in which primary outcomes were not specified, only the most frequently reported outcome was reported.
Medline, PubMed, and the Cumulative Index to Nursing and Allied Health Literature searches yielded 2743 citations. After adjusting for duplicates (n = 132), 2611 citations remained. Of these, 2508 were excluded, with the most common reasons for exclusion being an article without information about clinical use (n = 1832), an article without original data (n = 574), an article not relating to cannabis (n = 78), and an article not available in English (n = 24). The remaining 103 citations were assessed for eligibility by reviewing the full-text articles, and 82 were excluded because of the majority of the sample being older than 18 years or the data for children and adolescents were not reported separately. A total of 21 articles describing 22 studies were identified for final inclusion. A flow diagram is provided in Fig 2.
The 21 articles identified dated from 1979 to 2017, with 14 of the studies published within the last 5 years. Five randomized controlled trials (RCTs), 5 retrospective chart reviews, 5 case reports, 4 open-label trials, 2 parent surveys, and 1 case series were identified. The total number of participants across all studies was 795. Of the 5 medical conditions studied, the most common indication was for seizures (n = 11) and CINV (n = 6), followed by spasticity (n = 2), tics (n = 1), posttraumatic stress disorder (PTSD) (n = 1), and neuropathic pain (n = 1). Data abstraction followed the PRISMA guidelines. Table 2 summarizes the studies by clinical indication, sample characteristics, cannabinoid type, measures, and outcomes. Table 3 presents additional clinical descriptions of findings from each study, including cannabinoid dosage, frequency, formulation, secondary outcomes, and side effects.
Medical Cannabinoids for CINV
There have been 6 studies of cannabinoids for the treatment of CINV in children and adolescents. Dalzell et al10 showed that nabilone decreased nausea severity and frequency of vomiting in comparison with domperidone in a double-blind, crossover RCT of 23 children. Over a 5-day cycle of chemotherapy, patients treated with nabilone had an average of 6 episodes of emesis in comparison with 17 episodes of emesis among patients given domperidone. Nabilone also reduced nausea severity rated as 1.5 on a 5-point scale in comparison with a 2.5 severity rating in the domperidone treatment group. In a subsequent double-blind, cross-RCT of 30 children, Chan et al9 reported nabilone improved retching and emesis by 70% compared with 30% with prochloperazine. Over a cycle of chemotherapy, patients experienced 13 episodes of retching or emesis in comparison with 27 episodes when given prochloperazine. In an article reporting on 2 double-blind RCTs, Ekert et al11 showed that δ-9-THC reduced nausea and vomiting in comparison with metoclopramide as well as prochloperazine.
In an open-label trial, Abrahamov et al8 reported that δ-8-THC prevented vomiting during 480 cycles of chemotherapy among 8 children when given 2 hours before chemotherapy and repeated every 6 hours. In a more recent retrospective chart review of 95 children, Elder and Knoderer7 reported that dronabinol treatment given a median of 3 times over the course of chemotherapy led to a positive response in 60% of children (0–1 bouts of emesis). Notably, 95% of patients received lower dosing than was guideline referred (5 mg/m2), with the most common dose given being 2.5 mg/m2 every 6 hours as needed. Two-thirds of patients received repeated courses, and 62% received outpatient prescriptions, suggesting good tolerability of the medication.
Medical Cannabinoids for Epilepsy
There have been 11 studies of medical cannabinoids for the treatment of seizures in children and adolescents. In a recent RCT, Devinsky et al12 found that CBD significantly reduced convulsive seizure frequency in children with treatment-resistant epilepsy in Dravet syndrome as compared with a placebo. Among 61 participants who received CBD, the median frequency of monthly convulsive seizures decreased from 12.4 seizures per month to 5.9 seizures per month, as compared with a decrease of 14.9 to 14.1 in the placebo group. This represented an adjusted reduction in median seizure frequency by 22.9% with CBD in comparison with a placebo. Fifteen percent of those in the CBD group discontinued treatment before the 14 weeks as compared with 5% of those in the placebo group.
In a previous open-label trial, Devinsky et al16 reported that CBD reduced seizure frequency in a pediatric population with childhood-onset treatment-resistant epilepsies from a range of different causes. In the efficacy analysis of 137 completers over the 12-week treatment period, CBD led to a clinically relevant reduction in seizures with a median decrease in monthly motor seizures of 37%, from a baseline median of 30 motor seizures monthly to 16 motor seizures monthly. There was a low rate of patient discontinuation of CBD because of poor efficacy (3%). CBD also had acceptable tolerability, with only 3% of patients discontinuing treatment because of an adverse event. Notably, 24% of enrolled patients were not included in the safety analysis because of <12 weeks of treatment or follow-up.
In a small open-label case series of CBD for patients with treatment-refractory epilepsy in Sturge-Weber syndrome, Kaplan et al14 reported that seizures were reduced in 3 of the 5 patients. In a similar open-label case series of CBD for patients diagnosed with febrile infection-related epilepsy syndrome, Gofshteyn et al13 reported that seizures were reduced in 6 of the 7 patients.
In a retrospective chart review of 119 pediatric patients with epilepsy, Treat et al15 reported oral cannabis extracts improved seizures in 49% of the cohort, with 24% of the patients considered responders as defined by a >50% reduction in seizure burden. In a second retrospective chart review from the same institution, Press et al19 found that oral cannabis extracts reduced seizures in 57% of the 75 patients with treatment-refractory seizures. Tzadok et al17 conducted a retrospective chart review of 74 children and adolescents with treatment-resistant epilepsy and reported that CBD-enriched medical cannabis reduced seizures in 89% of patients.
In a small survey of 19 parents of children with treatment-resistant epilepsy, Porter and Jacobson21 found that CBD-enriched cannabis reduced seizure frequency in 84% of patients. In a follow-up on this early report, Hussain et al18 further surveyed 117 parents of children with epilepsy and found that CBD-enriched cannabis reduced seizures in 85% of patients. In a case series of 6 children with epilepsy, Lorenz22 reported that dronabinol reduced seizures in 2 of the patients. Saade and Joshi20 reported that CBD reduced seizure frequency in a 10-month-old patient with malignant migrating partial seizures of infancy.
Medical Cannabinoids for Spasticity
In a retrospective chart review of 12 children, Kuhlen et al25 described the effects of dronabinol for treatment-refractory spasticity related to developmental disorders at a palliative care setting. Dronabinol solution given twice daily reduced spasticity and was continued for a median of 181 days with no habituation observed. In a case report, Lorenz26 reported that dronabinol reduced spasticity and myoclonus in a toddler with a neurodegenerative disease called neuronal ceroid lipofuscinosis.
Medical Cannabinoids for Other Indications
In a case report of 2 adolescents with neuropathic pain and comorbid major depressive disorder, Rudich et al23 reported that dronabinol reduced the affective component of pain by 40% and improved psychosocial functioning after 4 months, although there was a gradual dissipation of effectiveness after 6 months that led to discontinuation. Shannon and Opila-Lehman24 reported that CBD improved anxiety and sleep in a case report of a 10-year-old girl with PTSD from early childhood trauma. Hasan et al27 reported that δ-9-THC decreased tic severity and improved quality of life in a case report of a 16-year-old patient with treatment-refractory Tourette syndrome.
This systematic review based on PRISMA guidelines identified 22 studies that evaluated the therapeutic benefits of medical cannabinoids in 795 children and adolescents, although 2 sets of studies (ie, 2 retrospective chart reviews and 2 parent surveys) may have had overlap in their sample groups. The study methods were heterogeneous, with only a minority of studies designed and powered for efficacy analysis (6 of 22). Of the double-blind RCTs (n = 5), all reported statistically significant postintervention reductions in the primary outcomes of CINV (n = 4) and convulsive seizures (n = 1). An open-label trial for treatment-refractory epilepsy also reported statistically and clinically significant postintervention reductions in seizure frequency, although the lack of a blinded control group limits the strength of the conclusion. Although the remaining reports suggested that cannabinoids were associated with improvements in CINV (n = 2), seizures (n = 9), spasticity (n = 2), tics (n = 1), PTSD (n = 1), and neuropathic pain (n = 1), the publications were not designed to evaluate the statistical significance of outcomes. In comparison with the paucity of pediatric studies on medical cannabinoids, the adult literature is relatively more substantive. Therefore, to facilitate an interpretation of the findings of this review, the identified pediatric studies are interpreted in context of a larger adult literature.
Although several of the RCTs investigating CINV date back to the 1980s, there is quality evidence that cannabinoids are effective as an antiemetic in children undergoing chemotherapy. Of note, all 6 studies used a THC cannabinoid, including δ-8-THC, δ-9-THC, dronabinol, and nabilone. The studies demonstrate that THC is more efficacious than antiemetics such as prochloperazine, metoclopramide, and domperidone, although side effects of drowsiness and dizziness were common. This evidence parallels the adult literature. In a Cochrane review of 23 adult trials, Smith et al28 reported that cannabinoids are more efficacious than a placebo and are similar to conventional antiemetics in the treatment of CINV.
The research in cannabinoids as a seizure treatment in children has grown rapidly over the past decade, with the number of studies investigating it as an antiepileptic equaling the number of studies for all other pediatric conditions combined. The 11 studies suggest cannabinoids may have a therapeutic benefit for seizures from different etiologies, including treatment-refractory epilepsy as studied in 8 of the studies. CBD is the cannabinoid that appears to have more evidence for efficacy as used in 8 of the 11 studies, including the only RCT and all 3 prospective open-label studies. However, most studies lacked a placebo control group, and the resulting potential for regression to the mean greatly reduces the strength of conclusions. Furthermore, the 2 survey studies recruited parents from online forums and a parent interest group, both of which are at high risk of sampling bias. In contrast to other diagnoses, the pediatric literature on cannabinoids for epilepsy informs the adult literature in this area, not vice versa.
Researchers in 2 studies, which are at high risk of bias because of a lack of controls and blinding, examined dronabinol for the treatment of spasticity in children with developmental disabilities. This evidence, albeit limited, parallels the adult literature. In summarizing 2 systematic reviews and an additional RCT of adult patients, the National Academy of Sciences29 concluded there was substantial evidence that oral cannabinoids benefit patient-reported spasticity symptoms, although the evidence is primarily from populations with multiple sclerosis. Nabiximols, an oromucosal spray containing an ∼1:1 ratio of THC to CBD, is a medication approved in Canada and multiple European countries for the treatment of adult patients with spasticity from multiple sclerosis and remains in phase 3 of FDA trials in the United States.
Reseachers in only 1 case report of 2 adolescents that lacked controls and blinding examined dronabinol for treatment of neuropathic pain; therefore, conclusions are limited. However, these preliminary findings tentatively align with findings in the adult literature. A systematic review by Whiting et al30 identified 28 RCTs of adults with chronic pain, of which 17 trials were related to a neuropathy; the resulting analysis suggested that cannabinoids lead to greater improvement in pain. In addition, Andreae et al31 conducted a subsequent systematic review of inhaled cannabis for peripheral neuropathy, which demonstrated pain relief with a possible dose-dependent effect.
Researchers in only 1 case report at high risk of bias given a lack of controls and blinding have examined CBD for the treatment of PTSD; therefore, conclusions are also limited. The limited adult literature is conflicting in regard to the association between cannabinoids and PTSD. In the only RCT, Jetly et al32 reported that nabilone improved nightmares, global clinical state, and general well-being compared with a placebo in a crossover design. However, this single study contrasts with nonrandomized literature that shows limited evidence of an association between cannabis use and increased PTSD symptom severity.31,33
Researchers in only 1 case report at high risk of bias given a lack of controls and blinding investigated the benefits of δ-9-THC in Tourette syndrome. In this case study, THC was associated with a reduction in tic severity. In the adult literature, 2 small controlled trials suggested a benefit of THC on tic severity in Tourette syndrome, although the reports are at similarly high risk of bias given the lack of an adequate description of randomization, allocation concealment, and incomplete outcome data.34,35
The literature on medical cannabinoids in children and adolescents is constrained by several important limitations, including between-study heterogeneity in the studied cannabinoid form and dosage (ie, CBD and THC content), indication, and ages of the sample. The sample sizes in many studies were small, with 13 of the 22 studies containing <20 participants. Notably, 17 of the 22 studies lacked a control group, and 16 of the 22 studies were not designed to test the statistical significance of changes in outcome measures. Finally, most studies lacked long-term follow-up to test for potential adverse neurocognitive and psychiatric side effects that have been demonstrated in recreational cannabis studies.
Risks of Cannabinoids
Although there is evidence for potential benefits in pediatric populations, pediatricians, families, and patients must balance the decision to use medical cannabinoids with the associated risks. In controlled trials, THC most commonly led to side effects of drowsiness and dizziness, with severity associated with higher doses. However, no major side effects were reported with dose reduction. The most common side effects with CBD were somnolence, diarrhea, and decreased appetite. In the controlled trial, although 75% of patients receiving CBD experienced side effects, only 13% withdrew from the trial because of the side effects. This parallels a systematic review of adult side effects from medical cannabinoids, which found dizziness and somnolence as the most commonly reported adverse events, followed by muscle spasm, pain, and dry mouth; notably, there was no evidence of a higher incidence of serious adverse events.36 Of note, accidental overdose of cannabis has been associated with multiple adverse effects, including reports of seizures among toddlers, which may be because of the toxicity of high-dose THC.37
The paucity of the studies limits our understanding of long-term risks associated with medical cannabinoids in pediatric populations. In the absence of substantive quality data from literature on medical cannabinoids, we highlight the findings of harms from recreational cannabis literature. There are important differences between recreational and medical cannabinoid use, including frequency, dosing, and potency, as well as significant confounds in the recreational use population, such as comorbid substance use and psychiatric illness. Although the applicability of the findings from the recreational cannabis literature to medical cannabinoids remains uncertain, pediatricians and families should understand the potential risks because it directly informs the decision for medical cannabinoid treatment.
The brain, including the endocannabinoid system, undergoes active development during adolescence,38 which may confer increased vulnerability to adverse long-term outcomes from cannabinoid use before adulthood. Cannabinoid receptors type 1 are particularly concentrated in brain regions that are critical for executive functioning, reward processing, and memory, including the prefrontal cortex, anterior cingulate cortex, basal ganglia, hippocampus, amygdala, and cerebellum.39 Neuroimaging studies show that individuals who begin using cannabis regularly in adolescence tend to have differences in cortical and subcortical volumes, white matter integrity, and functional connectivity compared with nonusers.40 The structural and functional neuroimaging differences appeared to correlate with cognitive impairments, such as attention deficits associated with right-hippocampus activation,41 verbal memory deficits associated with frontoparietal circuitry,42 and poorer executive functioning associated with prefrontal cortex volume.43
In a large, prospective study, long-term cannabis use in adolescents was associated with lower-than-expected IQ scores at follow-up,44 although this finding is confounded by familial environment, genetic liability, and sociodemographic factors, such as school dropout.45,46 Studies have demonstrated a dose-response relationship between cannabis use (ie, frequency, quantity, and duration) and cognitive impairments, including deficits in verbal learning and memory,47 psychomotor performance,48 and attention.49 Converging lines of evidence showed that the onset of cannabis use before age 16 years, compared with later onset, is associated with poorer attention,50 executive functioning,51 memory performance,47 and verbal IQ.52
Notably, recreational cannabis users in controlled settings have shown a preference for certain types of medical cannabinoids, including dronabinol and high-dose nabiximol, in comparison with a placebo, suggesting an abuse liability in at-risk populations.53 Long-term recreational use of cannabis is associated with risk of cannabis use disorder, which is characterized by impaired control over cannabis use and difficulty in ceasing use despite its harms. An estimated 8.9% of cannabis users escalate use to meet cannabis use disorder criteria.54 For those who initiate use in adolescence, the risk of cannabis use disorder rises to 1 in 6,55 with peak risk appearing at ∼17 years of age.56 Furthermore, twin studies reported that adolescent cannabis users have an elevated risk of developing other substance use disorders.57
One study has reported that recreational, frequent cannabis use during adolescence before age 15 years has been associated with an increased risk of depression.58 However, subsequent longitudinal studies reported contradicting results,59 with baseline depression associated with future initiation of cannabis use and suggesting confounds, such as sociodemographic factors and comorbidities, that limit conclusions regarding simple causality. Twin studies showed early-onset cannabis use and depression likely reflect shared genetic and environmental vulnerabilities.60 Adolescent cannabis use, particularly earlier onset and regular use, has also been associated with later suicidality.60,61 Cannabis use in early adolescence is further linked to earlier onset of psychotic disorders among at-risk populations.62 Adolescents who use cannabis regularly subsequently reported higher levels of subclinical psychotic symptoms, such as paranoia and hallucinations, and the effect persisted despite 1 year of abstinence.63
A review of prospective longitudinal studies reported that early cannabis use increases risk of poor school performance, particularly leaving school early.64 Adolescent cannabis use is also linked to externalizing problems, such as delinquent and aggressive behavior.65 Finally, increasing levels of cannabis use before age 21 years was associated with higher unemployment and welfare dependence and lower levels of income and relationship and life satisfaction by age 25 years.66
This review raises an important methodological issue in the field. Although we found a larger number (n = 2743) of citations that invoked terms related to cannabinoids in children and adolescents, we identified only 22 studies that examined cannabinoids for clinical indications in the pediatric population. Under the Controlled Substances Act of 1970, cannabis remains a Schedule I drug, and restrictive regulations continue to limit the research of medical cannabinoids. Concurrently, medical cannabinoids are becoming increasingly available to populations because of state legalization, of which cannabis plant products that are available in dispensaries may have highly variable cannabinoid concentrations. Finally, potential neurocognitive and psychiatric harms have been identified in the recreational cannabis literature. In this context, pediatricians, families, patients, and policy makers continue to lack urgently needed information to make balanced decisions regarding the use of medical cannabinoids in children and adolescents.
In summary, the objective of this systematic review was to synthesize the current state of the research on medical cannabinoids in children and adolescents. Beyond studies of CINV and epilepsy, the findings provided limited evidence of variable quality supporting the use of cannabinoids for different clinical indications. Additional larger, prospective, and controlled studies are required to better delineate the medical utility of cannabinoids in different pediatric disorders. This body of evidence has important implications in identifying the risks and benefits of medical cannabinoids in children and adolescents, especially in the context of psychiatric and neurocognitive adverse effects that have been identified from pediatric studies of recreational cannabis use.
- Accepted August 3, 2017.
- Address correspondence to Shane Shucheng Wong, MD, Division of Child and Adolescent Psychiatry, Massachusetts General Hospital, YAW 6A, Boston, MA 02114. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Copyright © 2017 by the American Academy of Pediatrics