Prevalence of Overweight and Obesity in a Large Clinical Sample of Children With Autism Sarabeth Broder-Fingert, MD; Karissa Brazauskas, MD; Kristen Lindgren, MD, PhD; Dorothea Iannuzzi, MSW; Jeanne Van Cleave, MD From the Division of General Pediatrics, Center for Child and Adolescent Health Research and Policy (Dr Broder-Fingert, Ms Iannuzzi, and Dr Van Cleave), and MassGeneral Hospital for Children (Dr Brazauskas and Dr Lindgren), Boston, Mass The authors declare that they have no conflict of interest. Address correspondence to Sarabeth Broder-Fingert, MD, Division of General Pediatrics, Center for Child and Adolescent Health Research and Policy, MassGeneral Hospital for Children, 100 Cambridge St, Room 1542, Boston, MA 0211 (e-mail: sbroder-finger@ partners.org). Received for publication August 24, 2013; accepted April 26, 2014.
ABSTRACT BACKGROUND: Overweight and obesity are major pediatric public health problems in the United States; however, limited data exist on the prevalence and correlates of overnutrition in children with autism. METHODS: Through a large integrated health care system’s patient database, we identified 6672 children ages 2 to 20 years with an assigned ICD-9 code of autism (299.0), Asperger syndrome (299.8), and control subjects from 2008 to 2011 who had at least 1 weight and height recorded in the same visit. We calculated age-adjusted, sex-adjusted body mass index and classified children as overweight (body mass index 85th to 95th percentile) or obese ($95th percentile). We used multinomial logistic regression to compare the odds of overweight and obesity between groups. We then used logistic regression to evaluate factors associated with overweight and obesity in children with autism, including demographic and clinical characteristics. RESULTS: Compared to control subjects, children with autism and Asperger syndrome had significantly higher odds of over-
weight (odds ratio, 95% confidence interval: autism 2.24, 1.74–2.88; Asperger syndrome 1.49, 1.12–1.97) and obesity (autism 4.83, 3.85–6.06; Asperger syndrome 5.69, 4.50–7.21). Among children with autism, we found a higher odds of obesity in older children (aged 12–15 years 1.87, 1.33–2.63; aged 16–20 years 1.94, 1.39–2.71) compared to children aged 6 to 11 years. We also found higher odds of overweight and obesity in those with public insurance (overweight 1.54, 1.25–1.89; obese 1.16, 1.02–1.40) and with co-occurring sleep disorder (obese 1.23, 1.00–1.53). CONCLUSIONS: Children with autism and Asperger syndrome had significantly higher odds of overweight and obesity than control subjects. Older age, public insurance, and cooccurring sleep disorder were associated with overweight or obesity in this population.
KEYWORDS: autism; obesity; overweight ACADEMIC PEDIATRICS 2014;14:408–414
WHAT’S NEW
among children with developmental disorders, such as autism spectrum disorders (ASD). ASD comprise a collection of developmental disorders characterized by restricted and repetitive behaviors, difficulty with social interaction, and deficient communication.6 The estimated prevalence of ASD is 1 in 68 children.7 Therefore, the co-occurrence of both ASD and obesity represents a potential important and unique public health burden. Children with ASD have several issues that may put them at greater risk for obesity than children without ASD. Children with ASD often have restricted food preferences8 or specialized diets9 that may make healthy dietary interventions less effective. Children with ASD are also less physically active than their non-ASD counterparts10–12 and may have difficulty participating in standard obesity prevention initiatives because of their behavioral problems. Finally, antipsychotic medications,13,14 which
The prevalence of overweight and obesity in a large population of children with autism is significantly higher than for children without autism. A limited set of autism-related medications and co-occurring conditions were not associated with overweight or obesity.
CHILDHOOD OVERWEIGHT AND obesity are major pediatric public health concerns in the United States,1 with a prevalence in the US pediatric population of 17%.2 Obesity increases a child’s risk for both short-term and long-term health problems, such as diabetes, heart disease, and psychosocial issues.3,4 Although there is a national effort to address the epidemic of pediatric obesity in the United States,5 little is understood about this disease
ACADEMIC PEDIATRICS Copyright ª 2014 by Academic Pediatric Association
408
Volume 14, Number 4 July–August 2014
ACADEMIC PEDIATRICS
are known to be associated with weight gain, are more commonly prescribed to children with ASD than to other children.15,16 Previous studies suggest higher prevalence of overweight and obesity in children with ASD, but these studies are limited by small samples sizes, lack of control comparison groups,17–22 and lack of objectively measured anthropomorphic data.17 Whether ASD truly confer additional risk for obesity remains unclear. On the basis of the unique risk factors in this population, we hypothesized that the prevalence of overweight and obesity would be higher in children with autism than in the general population. We performed a retrospective analysis of electronic medical record data collected on children with a diagnosis of autism or Asperger syndrome from a large integrated health care system. We compared the prevalence of overweight and obesity in these children to those without ASD, who acted as control subjects, and we investigated potential factors associated with overweight and obesity in children with ASD, including gender, age, race, insurance type, medication use, and co-occurring conditions. These data may inform recommendations for patientfocused interventions for prevention and treatment of overweight and obesity in this vulnerable population.
METHODS We evaluated prevalence of overweight and obesity for patients with autism, Asperger syndrome, and control subjects assessed at a large integrated health care system in eastern Massachusetts and compared prevalence among children with autism and Asperger syndrome to children without ASD. We also documented factors associated with overweight and obesity in children with autism. SAMPLE Patients for the cohort were identified through the Partners HealthCare System Research Patient Database Repository (RPDR). The RPDR is an extensive administrative and clinical care database that contains information on over 2.5 million patients and 550 million records from patient encounters.23 The RPDR gathers data from various health care settings (inpatient, outpatient, and emergency visits) and stores it in a single database. The database includes demographic information (including patient/ parent-reported race and ethnicity, insurance, and patient home address), all diagnoses billed for each visit, location of the visit (specific clinic name or inpatient unit), medications, procedures, vital signs, and Evaluation and Management codes. We identified all patients aged 2 to 20 years at the MassGeneral Hospital for Children (MGH) and its affiliate primary care centers who received care between January 1, 2008, and December 30, 2011, where an International Classification of Diseases, Ninth Revision (ICD-9), diagnosis of autism (299.0) or Asperger syndrome (299.8) was assigned. We chose 2008 as our cutoff because all MGH primary care practices were required to have adopted the institution electronic medical record system during or
OVERWEIGHT AND OBESITY IN AUTISM
409
before 2007. Inclusion criteria included age between 2 and 20 years at the time of clinical encounter and ICD-9 code of 299.0 or 299.8 at any encounter. We excluded any patient who had an ICD-9 diagnosis of 299.9 (Pervasive Developmental Disorder) because we were concerned only with the cohort of children with autism or Asperger syndrome in this analysis. For children who had both 299.0 and 299.8 listed in their medical record (n ¼ 28), we used the most recent code listed for categorization. A total of 5303 children met our inclusion criteria (3501 with autism, 1802 with Asperger syndrome) out of 337,091 patients aged 2 to 20 years seen at MGH from 2008 to 2011. Of these, 2075 children with autism and 901 children with Asperger syndrome had both a weight and a height recorded at a primary care visit. We included height and weight taken from any primary care encounter (well-child, urgent, and follow-up visits). Only subjects with both weight and height documented at the same visit were included in our sample. We chose the most recent visit at which a weight and height were recorded because ASD is diagnosed later in some populations than others (ie, Hispanic children are diagnosed at an older age), and we felt that using the most recent visit at which both weight and height were recorded would help minimize bias.24 A control sample of 3696 patients without autism or Asperger syndrome who had height and weight recorded at a primary care visit were then selected randomly from the larger database of 337,091 total patients aged 2 to 20 years seen at MGH from 2008 to 2011. CALCULATION OF BODY MASS INDEX Our main outcomes of interest were prevalence of overweight and obesity. We selected each patient’s most recent primary care visit in which both height and a weight were recorded. We then calculated body mass index (BMI) using the following equation: BMI ¼ weight (kg)/[height (m)]2. MAIN OUTCOME: OVERWEIGHT AND OBESITY We defined overweight as age-adjusted, sex-specific BMI between the 85th and 95th percentiles and obesity as a BMI at or above the 95th percentile.25 Weight, height, age at time of measurement, and gender, required to determine BMI percentile, were derived from the medical record. DEMOGRAPHIC CHARACTERISTICS We categorized race as black, white, and other. Ethnicity was categorized as Hispanic or other. Age in months and years was coded at the time of visit as a continuous variable. Sex was coded dichotomously. For the purpose of analyses, we used insurance type at the time of clinical encounter. In the original data pull, 65% of autism subjects and 22% of Asperger syndrome subjects had missing data for insurance type. Therefore, we used a structured chart review to extract their current insurance listed from their medical records. Mirroring the high prevalence of insurance among Massachusetts children generally,26 only 17 of the identified patients were uninsured, and we collapsed
410
BRODER-FINGERT ET AL
uninsured and public insurance into a single group. We then categorized the encounters as public insurance/no insurance and private insurance. MEDICATION USE The RPDR provided prescription medication history. We selected medications that are known to be used more frequently in children with ASD27 for our analysis. These included medications prescribed in both the inpatient and outpatient settings. Medications were grouped into 5 categories: antiepileptic, antipsychotic, antidepressant, sleep agent, and stimulant. Each of the 5 medication categories were coded dichotomously. Because we had only prescription data and not pharmacy records, we were not able to adjust for the length of time each child had been receiving the medication or whether prescriptions were filled. CO-OCCURRING CONDITIONS Co-occurring health conditions were identified by ICD-9 code. We selected conditions that are known to be associated with ASD for our analysis.28 These included congenital anomalies (740–759), perinatal conditions (760–779), sleep disorders (780), and metabolic disorder (277.9). We counted a patient as having a co-occurring condition if that condition was billed for at any time in their medical record. Co-occurring conditions were coded dichotomously in the 4 categories. ANALYSIS We initially determined the number of children with autism, Asperger syndrome, and control subjects who had overweight or obesity. By means of multinomial logistic regression, we calculated unadjusted and adjusted odds of overweight and obesity for children with autism and children with Asperger syndrome compared to control subjects. We then determined the prevalence of overweight and obesity in children with autism by age, gender, race/ ethnicity, autism-specific medication use, and cooccurring conditions, testing for significant associations by the chi-square test. We chose to look at only children with autism because this was the larger group of subjects (compared to the Asperger syndrome group) and therefore was better powered to detect any differences. We used adjusted multivariate logistic regression to determine what factors were associated with overweight and obesity while controlling for other factors. We used backward selection to determine which variables to include in our model. Our final model included gender, age, race/ ethnicity, insurance type, and receipt of antipsychotic medication, as well as co-occurring sleep disorder, perinatal conditions, and congenital anomalies. We also tested the interaction of sex and race with overweight and obesity. Odds ratios (OR) and 95% confidence intervals (CI) are reported. After identification of our cohort, we performed a sensitivity analysis on the autism cohort using patients with an ICD-9 code of 299.0 listed at least twice in their medical record (n ¼ 1412). The prevalence of overweight and obesity were not significantly different between these 2 populations, and therefore these data are not shown.
ACADEMIC PEDIATRICS
RESULTS CHARACTERISTICS OF THE STUDY POPULATION We identified 2976 children aged 2 to 20 years with an ICD-9 code of autism (299.00) or Asperger syndrome (299.8) seen at Partners HealthCare and who had a weight and height listed at the same primary care visit (Table 1). Subjects were mostly white (80.8%) and male (79.3%), with 5% black and 7% Hispanic. Controls were more evenly distributed by gender (50.1% male) and race/ ethnicity (51.1% white, 24.3% Hispanic). Age distribution varied by study group. Control children were younger than both children with autism and Asperger syndrome (Table 1). OBESITY AND OVERWEIGHT AMONG CHILDREN WITH AUTISM ASPERGER SYNDROME COMPARED TO CONTROL SUBJECTS At the most recent visit, 14.8% of children with autism and 11.1% with Asperger syndrome were overweight and 23.2% of children with autism and 25.3% of children with Asperger syndrome were obese, which differed significantly from control subjects (10.9% overweight, 6.3% obese, P < .01; Table 2). After adjusting for multiple factors, we found that children with autism and Asperger syndrome had significantly higher odds of overweight (OR, 95% CI: autism 2.24, 1.74–2.88; Asperger syndrome 1.49, 1.12–1.97) and obesity (autism 4.83, 3.85–6.06; Asperger syndrome 5.69, 4.50–7.21) than control children.
AND
OVERWEIGHT AND OBESITY BY AGE We determined the percentage of children with overweight and obesity by age category. We found significant differences emerged at the youngest age category (2 to 5 years) and persisted to the oldest age category (Figure). RISK FACTORS FOR OVERWEIGHT AND OBESITY We sought to determine which factors were associated with overweight or obesity among children with autism. We found that the odds of overweight or obesity were higher in multiple categories using bivariate analysis (Table 3). After adjusting for age, race/ethnicity, sex, medications, and co-occurring conditions, these findings persisted (Table 3). Compared to children aged 6 to 11 years, odds of obesity were higher in children aged 12 to 15 years (OR, 95% CI: 1.87, 1.33–2.63) and aged 16 to 20 years (1.94, 1.39–2.71). Higher odds of overweight were seen in children with public insurance (overweight 1.55, 1.25–1.89) compared to those with private insurance, and higher odds of obesity were found in children with cooccurring sleep disorder (obese 1.23, 1.00–1.53). Children who were female (0.71, 0.55–0.93) had a lower odds of obesity than those who were male. Because of the known association between race and sex in overweight and obesity, we also tested the interaction between race and sex in this population.29 The interaction was not significant (P ¼ .52). None of the autism-related co-occurring conditions evaluated were associated with overweight or obesity.
ACADEMIC PEDIATRICS
OVERWEIGHT AND OBESITY IN AUTISM
411
Table 1. Demographics of Children Aged 2 to 20 With Autism (ICD-9 299.0), Asperger Syndrome (299.8), and Controls Seen at Primary Care Visits, 2008–2011 Characteristic Gender Male Female Age 2–5 y 6–11 y 12–15 y 16–20 y Race/ethnicity White Black Hispanic Other Insurance Private Public Medication use Antiepileptics Antidepressants Antipsychotics Stimulants Co-occurring conditions 740–759 (Congenital Anomalies) 760–779 (Perinatal Diagnosis) 780 (Sleep Disorder) 277.9 (Metabolic Disorder)
Autism n (%)
Asperger Syndrome n (%)
Control n (%)
1650 (79.5) 425 (20.5)
709 (78.7) 192 (21.3)
1852 (50.1) 1844 (49.9)
406 (19.6) 871 (42.0) 356 (17.2) 436 (21.0)
2 (0.2) 222 (24.7) 318 (35.3) 359 (39.8)
633 (17.1) 1218 (40.0) 1063 (28.8) 782 (21.2)
1670 (80.7) 114 (5.5) 153 (7.4) 132 (6.4)
733 (81.3) 33 (3.6) 58 (6.4) 77 (8.6)
1888 (51.1) 601 (16.3) 897 (24.3) 310 (8.4)
931 (44.9) 1144 (55.1)
497 (55.2) 403 (44.8)
1738 (47.0) 1958 (53.0)
46 (2.0) 6 (0.3) 36 (1.6) 8 (0.4)
7 (0.8) 19 (21.1) 10 (1.1) 29 (3.2)
2 (0.03) 31 (0.8) 0 (0) 62 (1.7)
872 (38.4) 296 (13.0) 905 (43.6) 87 (4.2)
0 (0) 3 (0.2) 38 (2.8) 8 (0.6)
5 (0.1) 2 (0.1) 7 (0.2) 2 (0.1)
Additionally, none of the medication categories we assessed were associated with overweight or obesity (OR, 95% CI: antiepileptic overweight 0.81, 0.36–1.79, obese 0.66, 0.29–1.48; antidepressant obese 1.30, 0.22–7.67; antipsychotic overweight 1.16, 0.49–2.79, obese 0.66, 0.26–1.69; stimulant overweight 0.51, 0.06–4.32).
DISCUSSION In this study, we found a high prevalences of overweight and obesity in a cohort of patients with autism and Asperger syndrome compared to control subjects. Of children with autism and Asperger syndrome, 23.2% and 25.3% were obese, respectively. To our knowledge, this is the first study showing higher prevalence of overweight and obesity in children with autism or Asperger syndrome compared to control children using anthropomorphic measurements. These findings are consistent with previous work looking at overweight and obesity in children with ASD. Anal-
ysis of the National Survey of Children’s Health using parent-reported height and weight found 30.4% of children with ASD were overweight or obese, versus 23.6% of other children.17 In a review of clinical data, Curtin et al18 found the prevalence of overweight and obesity in patients with ASD to be quite high as well at 35.7% and 19%, respectively. Using the National Health Interview Survey, Phillips et al22 looked at overweight and obesity in adolescents with developmental disability. In this population, they found that children with autism had the highest rates of obesity compared to the other developmental disabilities examined. An explanation for the high prevalence of overweight and obesity in children with ASD remains to be elucidated. In their review, Curtain et al20 suggest multiple possible causal factors, including limited food preferences, increased sedentary behavior, and psychotropic medication use. Interestingly, our data are consistent with a recent study by Phillips et al22 that found higher prevalence of
Table 2. Overweight and Obesity in Children With Autism and Asperger Syndrome Compared to Controls, 2008–2011† Characteristic Overweight (85th to 95th percentile) Prevalence, n (%) Unadjusted OR (95% CI) Adjusted OR (95% CI)
Autism (n ¼ 2075)
Asperger Syndrome (n ¼ 901)
Controls (n ¼ 3696)
307 (14.8%) 2.30 (1.94–2.73)* 2.24 (1.74–2.88)*
100 (11.1%) 1.67 (1.31–2.13)* 1.49 (1.12–1.97)*
403 (10.9%) Ref Ref
481 (23.2%) 5.03 (4.25–5.96)* 4.83 (3.85–6.06)*
228 (25.3%) 5.36 (4.37–6.57)* 5.69 (4.50–7.21)*
233 (6.3%) Ref Ref
CI ¼ confidence interval; OR ¼ odds ratio. *P < .01. †Adjusted for sex, age, race/ethnicity, insurance type, autism-related medication use, and select co-occurring conditions.
BRODER-FINGERT ET AL
412
100
ACADEMIC PEDIATRICS
Control
90 Percent Overweight and Obese
21%) but lower in children with Asperger syndrome (22% and 10%). Current explanations for the mechanisms driving the high rates of overweight and obesity in children with ASD remain speculative. Multiple theories exist, including genetic susceptibility, difference in dietary or activity patterns, and inability to participate in obesity prevention initiatives, or many of these factors in combination. A key to understanding the high prevalence of overweight and obesity in children with ASD could be determining at what age overweight and obesity emerge in this population. We examined overweight and obesity by age category and found significant differences in prevalence of overweight and obesity in every age category between children with autism, Asperger syndrome, and control subjects. This implies that overweight and obesity begin at an early age for children with ASD. We were not able to examine individual weight trajectories in this cohort. It would be a valuable future endeavor to determine at what point children with autism are most likely to become overweight or obese. Finally, we were surprised to find that psychotropic medication use was not associated with overweight or obesity in our study population. Psychotropic medications, specifically risperidone, are a known risk factor for overweight or obesity.15 It is possible that our medication data were incomplete, given that it was based on a prescription history only within our health system. Alternatively,
*
Aspergers
80
AuƟsm
70 60
*
*
50
*
40
*
*
*
*
30 20
10 0 2 to 5
6 to 11
12 to 15
16 to 20
Age Category in Years
Figure. Percentage of control subjects, Asperger syndrome patients, and autism patients with either overweight or obesity by age category. *P < .01 compared to control subjects by t test.
obesity than overweight in adolescents with ASD. These authors hypothesized that dietary and physical activity patterns among those with the most severe limitations lead to more obesity than overweight. The hypothesis of Phillips et al22 is supported by a recent study by Zuckerman et al,21 which showed that the degree of ASD-specific symptoms was associated with degree of overweight and obesity in a clinical population. Our study contrasts with a study by Egan et al19 that found rates of overweight and obesity to be high in children with autism (39% and
Table 3. Risk Factors for Overweight and Obesity Within Children With Autism, 2008–2011 OR (95% CI)† Univariate Characteristic Gender Male Female Age 2–5 y 6–11 y 12–15 y 16–20 y Race/ethnicity White Black Hispanic Other Insurance Private Public Medication use Antiepileptics Antidepressants Antipsychotics Stimulants Co-occurring conditions 740–759 (Congenital Anomalies) 760–779 (Perinatal Diagnosis) 780 (Sleep Disorder) 277.9 (Metabolic Disorder)
Multivariate
Overweight
Obese
Overweight
Obese
Ref 0.77 (0.68–0.93)*
Ref 0.69 (0.51–0.89)*
Ref 1.06 (0.81–1.39)
Ref 0.71 (0.55–0.93)*
1.01 (0.79–1.55) Ref 1.28 (0.99–2.10) 11.11 (1.01–1.77)*
1.25 (0.91–1.77) Ref 1.98 (1.42–2.87)* 2.10 (1.98–2.86)*
1.09 (0.80–1.47) Ref 1.21 (0.84–1.74) 1.06 (0.74–1.51)
1.24 (0.93–1.66) Ref 1.87 (1.33–2.63)* 1.94 (1.39–2.71)*
Ref 1.01 (0.72–1.39) 1.80 (1.19–2.65)* 1.33 (0.82–2.10)
Ref 1.22 (0.89–2.11) 1.79 (1.10–2.26)* 0.98 (0.45–1.01)
Ref 0.83 (0.53–1.30) 0.89 (0.60–1.33) 1.31 (0.82–2.10)
Ref 1.19 (0.75–1.89) 0.89 (0.61–1.28) 0.70 (0.49–1.02)
Ref 2.01 (1.34–2.22)*
Ref 1.41 (1.02–1.99)*
Ref 1.55 (1.25–1.93)*
Ref 1.22 (0.99–1.49)
0.80 (0.35–1.77) NA 1.27 (0.61–2.99) 0.50 (0.06–3.33)
0.69 (0.22–1.49) 1.31 (0.21–7.11) 0.81 (0.39–1.72) NA
0.81 (0.36–1.79) NA 1.16 (0.49–2.79) 0.51 (0.06–4.32)
0.66 (0.29–1.48) 1.30 (0.22–7.67) 0.66 (0.26–1.69) NA
1.21 (0.93–1.44) 0.55 (0.40–0.81)* 1.25 (1.00–1.98)* 0.81 (0.50–1.44)
0.61 (0.45–0.88)* 1.21 (0.99–1.82) 1.39 (1.11–1.73)* 0.99 (0.65–1.61)
1.13 (0.89–1.43) 0.59 (0.40–0.86)* 1.23 (0.98–1.56) 0.85 (0.51–1.43)
0.68 (0.54–0.86)* 1.27 (0.93–1.73) 1.23 (1.00–1.53)* 0.95 (0.54–1.65)
OR ¼ odds ratio; CI ¼ confidence interval. *P < .01. †Adjusted for sex, age, race/ethnicity, insurance type, autism-related medication use, and select co-occurring conditions.
ACADEMIC PEDIATRICS
autism may convey such a large risk for overweight and obesity that the impact of medications is obscured. Finally, this study may be underpowered to detect a difference based on medication use as a result of the small number of patients prescribed these medications. LIMITATIONS It is important to note several limitations of our study. First, this was a cross-sectional study, and therefore, we could not address the time course of overweight or obesity in this population. Second, our ability to evaluate factors associated with overweight and obesity in this population was limited. The factors we examined were few and crudely measured. For example, medications use was measured by prescription data rather than actual documented use or duration of use. Additionally, medications may have been missed if a physician outside of the Partners HealthCare system prescribed a medication. There is also concern for misclassification. First, racial and ethnic minority groups tend to receive a diagnosis of autism later than similar white patients.30 This may mean that some of the younger nonwhite patients in our cohort who were classified as control subjects may have undiagnosed autism. At the same time, some physicians may be hesitant to label a patient with autism as a result of the possible stigma associated with the diagnosis and could lead to additional misclassification. Both misclassifications would bias toward the null hypothesis. One possible explanation for the extremely high rates of overweight and obesity in this cohort is that the population of children with autism is not demographically reflective of the overall autism population. Although this concern cannot be ruled out entirely, our data do closely resemble published data on the population of patients with ASD in Massachusetts.31 Published statewide data from the 2001 to 2005 birth cohort found that 75.3% of children with ASD were white (vs 77.5% white in our cohort) and 82.0% male (vs 79.6% male in our cohort). Our control population was limited in that it was not matched by age, race, or gender. We chose not to match control subjects in order to have a sufficient population derived from primary care clinics. We instead adjusted for age, race, and gender in our model. This also allowed us to look at a control population more representative of the general population. Finally, our data were collected from a single medical center database. Although Partners HealthCare services a diverse population of patients, this may still limit the generalizability to the broader nationwide community of children with autism. Despite these limitations, the results presented here represent the largest study to date with objectively collected height and weight examining the question of overweight and obesity among children with autism.
CONCLUSIONS The prevalence of overweight and obesity in a population of children with autism is much higher than those
OVERWEIGHT AND OBESITY IN AUTISM
413
without autism. These findings are important to future public health initiatives addressing overweight and obesity for several reasons. First, development of future prevention and treatment interventions should take into account this special and growing population7 of patients and consider how they can best be included. Second, because differences in overweight and obesity emerged in the youngest age group evaluated, prevention and anticipatory guidance may be most effective early in a child’s life. Finally, identification of specific environmental factors contributing to overweight and obesity in this population, such as time spent in sedentary activity, access to physical activity programs, poor social support from peers, or use of food as a behavioral reward, should be a priority for future work.
ACKNOWLEDGMENTS Supported in part by the APA Resident Investigator Award.
REFERENCES 1. Ebbeling CB, Pawlak DB, Ludwig DS. Childhood obesity: publichealth crisis, common sense cure. Lancet. 2002;360:473–482. 2. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of childhood and adult obesity in the United States, 2011–2012. JAMA. 2014;311: 806–814. 3. Han JC, Lawlor DA, Kimm SY. Childhood obesity. Lancet. 2010;375: 1737–1748. 4. Biro FM, Wien M. Childhood obesity and adult morbidities. Am J Clin Nutr. 2010;91:1499S–1505S. 5. Lumeng JC, Castle WP, Lumeng CN. The role of pediatricians in the coordinated national effort to address childhood obesity. Pediatrics. 2010;126:574–575. 6. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed., text revision ed. Washington, DC: American Psychiatric Association; 2000. 7. Blumberg SJ, Bramlett MD, Kogan MD, et al. Changes in Prevalence of Parent-Reported Autism Spectrum Disorder in School-Aged US Children: 2007 to 2011–2012. Hyattsville, Md: National Center for Health Statistics; 2013. National Health Statistics Report 65. 8. Schreck KA, Williams K, Smith AF. A comparison of eating behaviors between children with and without autism. J Autism Dev Disord. 2004;34:433–438. 9. Marcason W. What is the current status of research concerning use of a gluten-free, casein-free diet for children diagnosed with autism? J Am Diet Assoc. 2009;109:572. 10. Pan CY. Objectively measured physical activity between children with autism spectrum disorders and children without disabilities during inclusive recess settings in Taiwan. J Autism Dev Disord. 2008 Aug;38:1292–1301. 11. Hinckson EA, Dickinson A, Water T, et al. Physical activity, dietary habits and overall health in overweight and obese children and youth with intellectual disability or autism. Res Dev Disabil. 2013;34: 1170–1178. 12. Rimmer JH, Yamaki K, Lowry BM, et al. Obesity and obesity-related secondary conditions in adolescent with intellectual/developmental disabilities. J Intellect Disabil Res. 2010;54:787–794. 13. Esbensen AJ, Greenberg JS, Seltzer MM, Aman MG. A longitudinal investigation of psychotropic and non-psychotropic medication use among adolescents and adults with autism spectrum disorders. J Autism Dev Disord. 2009;39:1339–1349. 14. Barnard L, Young AH, Pearson J, et al. A systematic review of the use of atypical antipsychotics in autism. J Psychopharmacol. 2002;16: 93–101. 15. Maayan L, Correll CU. Weight gain and metabolic risks associated with antipsychotic medications in children and adolescents. J Child Adolesc Psychopharmacol. 2011;21:517–535.
414
BRODER-FINGERT ET AL
16. Sachs GS, Guille C. Weight gain associated with use of psychotropic medications. J Clin Psychiatry. 1999;60(Suppl 21):16–19. 17. Curtin C, Anderson SE, Must A, Bandini L. The prevalence of obesity in children with autism: a secondary data analysis using nationally representative data from the National Survey of Children’s Health. BMC Pediatr. 2010;10:11. 18. Curtin C, Bandini LG, Perrin EC, et al. Prevalence of overweight in children and adolescents with attention deficit hyperactivity disorder and autism spectrum disorders: a chart review. BMC Pediatr. 2005;5:48. 19. Egan AM, Dreyer ML, Odar CC, et al. Obesity in young children with autism spectrum disorders: prevalence and associated factors. Child Obes. 2013;9:125–131. 20. Curtin C, Jojic M, Bandini LG. Obesity in children with autism spectrum disorder. Harv Rev Psychiatry. 2014;22:93–103. 21. Zuckerman KE, Hill AP, Guion K, et al. Overweight and obesity: prevalence and correlates in a large clinical sample of children with autism spectrum disorder. J Autism Dev Disord. In press. 22. Phillips KL, Schieve LA, Visser S, et al. Prevalence and impact of unhealthy weight in a national sample of US adolescents with autism and other learning and behavioral disabilities. Matern Child Health J. In press. 23. Nalichowski R, Keogh D, Chueh HC, et al. Calculating the benefits of a research patient data repository. AMIA Annu Symp Proc. 2006;1044. 24. Zuckerman KE, Mattox K, Donelan K, et al. Pediatrician identification of Latino children at risk for autism spectrum disorder. Pediatrics. 2013;132:445–453.
ACADEMIC PEDIATRICS 25. Kuczmarski RJ, Ogden CL, Guo SS, et al. 2000 CDC growth charts for the United States: methods and development. National Center for Health Statistics. Vital Health Stat. 2002;11. 26. Long SK, Cook A, Stockley K. Health Insurance Coverage in Massachusetts: Estimates from the 2008 Massachusetts Health Insurance Survey. Boston, Mass: Massachusetts Division of Health Care Finance and Policy, Commonwealth of Massachusetts, and Urban Institute; 2008. Available at: http://www.urban.org/UploadedPDF/411815_ Massachusetts_Insurance.pdf. Accessed May 13, 2014. 27. Murray ML, Hsia Y, Glaser K, et al. Pharmacological treatments prescribed to people with autism spectrum disorder (ASD) in primary health care. Psychopharmacology (Berl). 2014;231: 1011–1021. 28. Bauman ML. Medical comorbidities in autism: challenges to diagnosis and treatment. Neurotherapeutics. 2010;7:320–327. 29. Ogden CL, Carroll MD, Curtin LR, et al. Prevalence of overweight and obesity in the United States, 1999–2004. JAMA. 2006; 295:1549–1555. http://www.urban.org/UploadedPDF/411815_ Massachusetts_Insurance.pdf. Accessed May 13, 2014. 30. Mandell DS, Wiggins LD, Carpenter LA, et al. Racial/ethnic disparities in the identification of children with autism spectrum disorders. Am J Public Health. 2009;99:493–498. 31. Manning SE, Davin CA, Barfield WD, et al. Early diagnoses of autism spectrum disorders in Massachusetts birth cohorts, 2001–2005. Pediatrics. 2011;127:1043–1051.
ABSTRACT BACKGROUND: Overweight and obesity are major pediatric public health problems in the United States; however, limited data exist on the prevalence and correlates of overnutrition in children with autism. METHODS: Through a large integrated health care system’s patient database, we identified 6672 children ages 2 to 20 years with an assigned ICD-9 code of autism (299.0), Asperger syndrome (299.8), and control subjects from 2008 to 2011 who had at least 1 weight and height recorded in the same visit. We calculated age-adjusted, sex-adjusted body mass index and classified children as overweight (body mass index 85th to 95th percentile) or obese ($95th percentile). We used multinomial logistic regression to compare the odds of overweight and obesity between groups. We then used logistic regression to evaluate factors associated with overweight and obesity in children with autism, including demographic and clinical characteristics. RESULTS: Compared to control subjects, children with autism and Asperger syndrome had significantly higher odds of over-
weight (odds ratio, 95% confidence interval: autism 2.24, 1.74–2.88; Asperger syndrome 1.49, 1.12–1.97) and obesity (autism 4.83, 3.85–6.06; Asperger syndrome 5.69, 4.50–7.21). Among children with autism, we found a higher odds of obesity in older children (aged 12–15 years 1.87, 1.33–2.63; aged 16–20 years 1.94, 1.39–2.71) compared to children aged 6 to 11 years. We also found higher odds of overweight and obesity in those with public insurance (overweight 1.54, 1.25–1.89; obese 1.16, 1.02–1.40) and with co-occurring sleep disorder (obese 1.23, 1.00–1.53). CONCLUSIONS: Children with autism and Asperger syndrome had significantly higher odds of overweight and obesity than control subjects. Older age, public insurance, and cooccurring sleep disorder were associated with overweight or obesity in this population.
KEYWORDS: autism; obesity; overweight ACADEMIC PEDIATRICS 2014;14:408–414
WHAT’S NEW
among children with developmental disorders, such as autism spectrum disorders (ASD). ASD comprise a collection of developmental disorders characterized by restricted and repetitive behaviors, difficulty with social interaction, and deficient communication.6 The estimated prevalence of ASD is 1 in 68 children.7 Therefore, the co-occurrence of both ASD and obesity represents a potential important and unique public health burden. Children with ASD have several issues that may put them at greater risk for obesity than children without ASD. Children with ASD often have restricted food preferences8 or specialized diets9 that may make healthy dietary interventions less effective. Children with ASD are also less physically active than their non-ASD counterparts10–12 and may have difficulty participating in standard obesity prevention initiatives because of their behavioral problems. Finally, antipsychotic medications,13,14 which
The prevalence of overweight and obesity in a large population of children with autism is significantly higher than for children without autism. A limited set of autism-related medications and co-occurring conditions were not associated with overweight or obesity.
CHILDHOOD OVERWEIGHT AND obesity are major pediatric public health concerns in the United States,1 with a prevalence in the US pediatric population of 17%.2 Obesity increases a child’s risk for both short-term and long-term health problems, such as diabetes, heart disease, and psychosocial issues.3,4 Although there is a national effort to address the epidemic of pediatric obesity in the United States,5 little is understood about this disease
ACADEMIC PEDIATRICS Copyright ª 2014 by Academic Pediatric Association
408
Volume 14, Number 4 July–August 2014
ACADEMIC PEDIATRICS
are known to be associated with weight gain, are more commonly prescribed to children with ASD than to other children.15,16 Previous studies suggest higher prevalence of overweight and obesity in children with ASD, but these studies are limited by small samples sizes, lack of control comparison groups,17–22 and lack of objectively measured anthropomorphic data.17 Whether ASD truly confer additional risk for obesity remains unclear. On the basis of the unique risk factors in this population, we hypothesized that the prevalence of overweight and obesity would be higher in children with autism than in the general population. We performed a retrospective analysis of electronic medical record data collected on children with a diagnosis of autism or Asperger syndrome from a large integrated health care system. We compared the prevalence of overweight and obesity in these children to those without ASD, who acted as control subjects, and we investigated potential factors associated with overweight and obesity in children with ASD, including gender, age, race, insurance type, medication use, and co-occurring conditions. These data may inform recommendations for patientfocused interventions for prevention and treatment of overweight and obesity in this vulnerable population.
METHODS We evaluated prevalence of overweight and obesity for patients with autism, Asperger syndrome, and control subjects assessed at a large integrated health care system in eastern Massachusetts and compared prevalence among children with autism and Asperger syndrome to children without ASD. We also documented factors associated with overweight and obesity in children with autism. SAMPLE Patients for the cohort were identified through the Partners HealthCare System Research Patient Database Repository (RPDR). The RPDR is an extensive administrative and clinical care database that contains information on over 2.5 million patients and 550 million records from patient encounters.23 The RPDR gathers data from various health care settings (inpatient, outpatient, and emergency visits) and stores it in a single database. The database includes demographic information (including patient/ parent-reported race and ethnicity, insurance, and patient home address), all diagnoses billed for each visit, location of the visit (specific clinic name or inpatient unit), medications, procedures, vital signs, and Evaluation and Management codes. We identified all patients aged 2 to 20 years at the MassGeneral Hospital for Children (MGH) and its affiliate primary care centers who received care between January 1, 2008, and December 30, 2011, where an International Classification of Diseases, Ninth Revision (ICD-9), diagnosis of autism (299.0) or Asperger syndrome (299.8) was assigned. We chose 2008 as our cutoff because all MGH primary care practices were required to have adopted the institution electronic medical record system during or
OVERWEIGHT AND OBESITY IN AUTISM
409
before 2007. Inclusion criteria included age between 2 and 20 years at the time of clinical encounter and ICD-9 code of 299.0 or 299.8 at any encounter. We excluded any patient who had an ICD-9 diagnosis of 299.9 (Pervasive Developmental Disorder) because we were concerned only with the cohort of children with autism or Asperger syndrome in this analysis. For children who had both 299.0 and 299.8 listed in their medical record (n ¼ 28), we used the most recent code listed for categorization. A total of 5303 children met our inclusion criteria (3501 with autism, 1802 with Asperger syndrome) out of 337,091 patients aged 2 to 20 years seen at MGH from 2008 to 2011. Of these, 2075 children with autism and 901 children with Asperger syndrome had both a weight and a height recorded at a primary care visit. We included height and weight taken from any primary care encounter (well-child, urgent, and follow-up visits). Only subjects with both weight and height documented at the same visit were included in our sample. We chose the most recent visit at which a weight and height were recorded because ASD is diagnosed later in some populations than others (ie, Hispanic children are diagnosed at an older age), and we felt that using the most recent visit at which both weight and height were recorded would help minimize bias.24 A control sample of 3696 patients without autism or Asperger syndrome who had height and weight recorded at a primary care visit were then selected randomly from the larger database of 337,091 total patients aged 2 to 20 years seen at MGH from 2008 to 2011. CALCULATION OF BODY MASS INDEX Our main outcomes of interest were prevalence of overweight and obesity. We selected each patient’s most recent primary care visit in which both height and a weight were recorded. We then calculated body mass index (BMI) using the following equation: BMI ¼ weight (kg)/[height (m)]2. MAIN OUTCOME: OVERWEIGHT AND OBESITY We defined overweight as age-adjusted, sex-specific BMI between the 85th and 95th percentiles and obesity as a BMI at or above the 95th percentile.25 Weight, height, age at time of measurement, and gender, required to determine BMI percentile, were derived from the medical record. DEMOGRAPHIC CHARACTERISTICS We categorized race as black, white, and other. Ethnicity was categorized as Hispanic or other. Age in months and years was coded at the time of visit as a continuous variable. Sex was coded dichotomously. For the purpose of analyses, we used insurance type at the time of clinical encounter. In the original data pull, 65% of autism subjects and 22% of Asperger syndrome subjects had missing data for insurance type. Therefore, we used a structured chart review to extract their current insurance listed from their medical records. Mirroring the high prevalence of insurance among Massachusetts children generally,26 only 17 of the identified patients were uninsured, and we collapsed
410
BRODER-FINGERT ET AL
uninsured and public insurance into a single group. We then categorized the encounters as public insurance/no insurance and private insurance. MEDICATION USE The RPDR provided prescription medication history. We selected medications that are known to be used more frequently in children with ASD27 for our analysis. These included medications prescribed in both the inpatient and outpatient settings. Medications were grouped into 5 categories: antiepileptic, antipsychotic, antidepressant, sleep agent, and stimulant. Each of the 5 medication categories were coded dichotomously. Because we had only prescription data and not pharmacy records, we were not able to adjust for the length of time each child had been receiving the medication or whether prescriptions were filled. CO-OCCURRING CONDITIONS Co-occurring health conditions were identified by ICD-9 code. We selected conditions that are known to be associated with ASD for our analysis.28 These included congenital anomalies (740–759), perinatal conditions (760–779), sleep disorders (780), and metabolic disorder (277.9). We counted a patient as having a co-occurring condition if that condition was billed for at any time in their medical record. Co-occurring conditions were coded dichotomously in the 4 categories. ANALYSIS We initially determined the number of children with autism, Asperger syndrome, and control subjects who had overweight or obesity. By means of multinomial logistic regression, we calculated unadjusted and adjusted odds of overweight and obesity for children with autism and children with Asperger syndrome compared to control subjects. We then determined the prevalence of overweight and obesity in children with autism by age, gender, race/ ethnicity, autism-specific medication use, and cooccurring conditions, testing for significant associations by the chi-square test. We chose to look at only children with autism because this was the larger group of subjects (compared to the Asperger syndrome group) and therefore was better powered to detect any differences. We used adjusted multivariate logistic regression to determine what factors were associated with overweight and obesity while controlling for other factors. We used backward selection to determine which variables to include in our model. Our final model included gender, age, race/ ethnicity, insurance type, and receipt of antipsychotic medication, as well as co-occurring sleep disorder, perinatal conditions, and congenital anomalies. We also tested the interaction of sex and race with overweight and obesity. Odds ratios (OR) and 95% confidence intervals (CI) are reported. After identification of our cohort, we performed a sensitivity analysis on the autism cohort using patients with an ICD-9 code of 299.0 listed at least twice in their medical record (n ¼ 1412). The prevalence of overweight and obesity were not significantly different between these 2 populations, and therefore these data are not shown.
ACADEMIC PEDIATRICS
RESULTS CHARACTERISTICS OF THE STUDY POPULATION We identified 2976 children aged 2 to 20 years with an ICD-9 code of autism (299.00) or Asperger syndrome (299.8) seen at Partners HealthCare and who had a weight and height listed at the same primary care visit (Table 1). Subjects were mostly white (80.8%) and male (79.3%), with 5% black and 7% Hispanic. Controls were more evenly distributed by gender (50.1% male) and race/ ethnicity (51.1% white, 24.3% Hispanic). Age distribution varied by study group. Control children were younger than both children with autism and Asperger syndrome (Table 1). OBESITY AND OVERWEIGHT AMONG CHILDREN WITH AUTISM ASPERGER SYNDROME COMPARED TO CONTROL SUBJECTS At the most recent visit, 14.8% of children with autism and 11.1% with Asperger syndrome were overweight and 23.2% of children with autism and 25.3% of children with Asperger syndrome were obese, which differed significantly from control subjects (10.9% overweight, 6.3% obese, P < .01; Table 2). After adjusting for multiple factors, we found that children with autism and Asperger syndrome had significantly higher odds of overweight (OR, 95% CI: autism 2.24, 1.74–2.88; Asperger syndrome 1.49, 1.12–1.97) and obesity (autism 4.83, 3.85–6.06; Asperger syndrome 5.69, 4.50–7.21) than control children.
AND
OVERWEIGHT AND OBESITY BY AGE We determined the percentage of children with overweight and obesity by age category. We found significant differences emerged at the youngest age category (2 to 5 years) and persisted to the oldest age category (Figure). RISK FACTORS FOR OVERWEIGHT AND OBESITY We sought to determine which factors were associated with overweight or obesity among children with autism. We found that the odds of overweight or obesity were higher in multiple categories using bivariate analysis (Table 3). After adjusting for age, race/ethnicity, sex, medications, and co-occurring conditions, these findings persisted (Table 3). Compared to children aged 6 to 11 years, odds of obesity were higher in children aged 12 to 15 years (OR, 95% CI: 1.87, 1.33–2.63) and aged 16 to 20 years (1.94, 1.39–2.71). Higher odds of overweight were seen in children with public insurance (overweight 1.55, 1.25–1.89) compared to those with private insurance, and higher odds of obesity were found in children with cooccurring sleep disorder (obese 1.23, 1.00–1.53). Children who were female (0.71, 0.55–0.93) had a lower odds of obesity than those who were male. Because of the known association between race and sex in overweight and obesity, we also tested the interaction between race and sex in this population.29 The interaction was not significant (P ¼ .52). None of the autism-related co-occurring conditions evaluated were associated with overweight or obesity.
ACADEMIC PEDIATRICS
OVERWEIGHT AND OBESITY IN AUTISM
411
Table 1. Demographics of Children Aged 2 to 20 With Autism (ICD-9 299.0), Asperger Syndrome (299.8), and Controls Seen at Primary Care Visits, 2008–2011 Characteristic Gender Male Female Age 2–5 y 6–11 y 12–15 y 16–20 y Race/ethnicity White Black Hispanic Other Insurance Private Public Medication use Antiepileptics Antidepressants Antipsychotics Stimulants Co-occurring conditions 740–759 (Congenital Anomalies) 760–779 (Perinatal Diagnosis) 780 (Sleep Disorder) 277.9 (Metabolic Disorder)
Autism n (%)
Asperger Syndrome n (%)
Control n (%)
1650 (79.5) 425 (20.5)
709 (78.7) 192 (21.3)
1852 (50.1) 1844 (49.9)
406 (19.6) 871 (42.0) 356 (17.2) 436 (21.0)
2 (0.2) 222 (24.7) 318 (35.3) 359 (39.8)
633 (17.1) 1218 (40.0) 1063 (28.8) 782 (21.2)
1670 (80.7) 114 (5.5) 153 (7.4) 132 (6.4)
733 (81.3) 33 (3.6) 58 (6.4) 77 (8.6)
1888 (51.1) 601 (16.3) 897 (24.3) 310 (8.4)
931 (44.9) 1144 (55.1)
497 (55.2) 403 (44.8)
1738 (47.0) 1958 (53.0)
46 (2.0) 6 (0.3) 36 (1.6) 8 (0.4)
7 (0.8) 19 (21.1) 10 (1.1) 29 (3.2)
2 (0.03) 31 (0.8) 0 (0) 62 (1.7)
872 (38.4) 296 (13.0) 905 (43.6) 87 (4.2)
0 (0) 3 (0.2) 38 (2.8) 8 (0.6)
5 (0.1) 2 (0.1) 7 (0.2) 2 (0.1)
Additionally, none of the medication categories we assessed were associated with overweight or obesity (OR, 95% CI: antiepileptic overweight 0.81, 0.36–1.79, obese 0.66, 0.29–1.48; antidepressant obese 1.30, 0.22–7.67; antipsychotic overweight 1.16, 0.49–2.79, obese 0.66, 0.26–1.69; stimulant overweight 0.51, 0.06–4.32).
DISCUSSION In this study, we found a high prevalences of overweight and obesity in a cohort of patients with autism and Asperger syndrome compared to control subjects. Of children with autism and Asperger syndrome, 23.2% and 25.3% were obese, respectively. To our knowledge, this is the first study showing higher prevalence of overweight and obesity in children with autism or Asperger syndrome compared to control children using anthropomorphic measurements. These findings are consistent with previous work looking at overweight and obesity in children with ASD. Anal-
ysis of the National Survey of Children’s Health using parent-reported height and weight found 30.4% of children with ASD were overweight or obese, versus 23.6% of other children.17 In a review of clinical data, Curtin et al18 found the prevalence of overweight and obesity in patients with ASD to be quite high as well at 35.7% and 19%, respectively. Using the National Health Interview Survey, Phillips et al22 looked at overweight and obesity in adolescents with developmental disability. In this population, they found that children with autism had the highest rates of obesity compared to the other developmental disabilities examined. An explanation for the high prevalence of overweight and obesity in children with ASD remains to be elucidated. In their review, Curtain et al20 suggest multiple possible causal factors, including limited food preferences, increased sedentary behavior, and psychotropic medication use. Interestingly, our data are consistent with a recent study by Phillips et al22 that found higher prevalence of
Table 2. Overweight and Obesity in Children With Autism and Asperger Syndrome Compared to Controls, 2008–2011† Characteristic Overweight (85th to 95th percentile) Prevalence, n (%) Unadjusted OR (95% CI) Adjusted OR (95% CI)
Autism (n ¼ 2075)
Asperger Syndrome (n ¼ 901)
Controls (n ¼ 3696)
307 (14.8%) 2.30 (1.94–2.73)* 2.24 (1.74–2.88)*
100 (11.1%) 1.67 (1.31–2.13)* 1.49 (1.12–1.97)*
403 (10.9%) Ref Ref
481 (23.2%) 5.03 (4.25–5.96)* 4.83 (3.85–6.06)*
228 (25.3%) 5.36 (4.37–6.57)* 5.69 (4.50–7.21)*
233 (6.3%) Ref Ref
CI ¼ confidence interval; OR ¼ odds ratio. *P < .01. †Adjusted for sex, age, race/ethnicity, insurance type, autism-related medication use, and select co-occurring conditions.
BRODER-FINGERT ET AL
412
100
ACADEMIC PEDIATRICS
Control
90 Percent Overweight and Obese
21%) but lower in children with Asperger syndrome (22% and 10%). Current explanations for the mechanisms driving the high rates of overweight and obesity in children with ASD remain speculative. Multiple theories exist, including genetic susceptibility, difference in dietary or activity patterns, and inability to participate in obesity prevention initiatives, or many of these factors in combination. A key to understanding the high prevalence of overweight and obesity in children with ASD could be determining at what age overweight and obesity emerge in this population. We examined overweight and obesity by age category and found significant differences in prevalence of overweight and obesity in every age category between children with autism, Asperger syndrome, and control subjects. This implies that overweight and obesity begin at an early age for children with ASD. We were not able to examine individual weight trajectories in this cohort. It would be a valuable future endeavor to determine at what point children with autism are most likely to become overweight or obese. Finally, we were surprised to find that psychotropic medication use was not associated with overweight or obesity in our study population. Psychotropic medications, specifically risperidone, are a known risk factor for overweight or obesity.15 It is possible that our medication data were incomplete, given that it was based on a prescription history only within our health system. Alternatively,
*
Aspergers
80
AuƟsm
70 60
*
*
50
*
40
*
*
*
*
30 20
10 0 2 to 5
6 to 11
12 to 15
16 to 20
Age Category in Years
Figure. Percentage of control subjects, Asperger syndrome patients, and autism patients with either overweight or obesity by age category. *P < .01 compared to control subjects by t test.
obesity than overweight in adolescents with ASD. These authors hypothesized that dietary and physical activity patterns among those with the most severe limitations lead to more obesity than overweight. The hypothesis of Phillips et al22 is supported by a recent study by Zuckerman et al,21 which showed that the degree of ASD-specific symptoms was associated with degree of overweight and obesity in a clinical population. Our study contrasts with a study by Egan et al19 that found rates of overweight and obesity to be high in children with autism (39% and
Table 3. Risk Factors for Overweight and Obesity Within Children With Autism, 2008–2011 OR (95% CI)† Univariate Characteristic Gender Male Female Age 2–5 y 6–11 y 12–15 y 16–20 y Race/ethnicity White Black Hispanic Other Insurance Private Public Medication use Antiepileptics Antidepressants Antipsychotics Stimulants Co-occurring conditions 740–759 (Congenital Anomalies) 760–779 (Perinatal Diagnosis) 780 (Sleep Disorder) 277.9 (Metabolic Disorder)
Multivariate
Overweight
Obese
Overweight
Obese
Ref 0.77 (0.68–0.93)*
Ref 0.69 (0.51–0.89)*
Ref 1.06 (0.81–1.39)
Ref 0.71 (0.55–0.93)*
1.01 (0.79–1.55) Ref 1.28 (0.99–2.10) 11.11 (1.01–1.77)*
1.25 (0.91–1.77) Ref 1.98 (1.42–2.87)* 2.10 (1.98–2.86)*
1.09 (0.80–1.47) Ref 1.21 (0.84–1.74) 1.06 (0.74–1.51)
1.24 (0.93–1.66) Ref 1.87 (1.33–2.63)* 1.94 (1.39–2.71)*
Ref 1.01 (0.72–1.39) 1.80 (1.19–2.65)* 1.33 (0.82–2.10)
Ref 1.22 (0.89–2.11) 1.79 (1.10–2.26)* 0.98 (0.45–1.01)
Ref 0.83 (0.53–1.30) 0.89 (0.60–1.33) 1.31 (0.82–2.10)
Ref 1.19 (0.75–1.89) 0.89 (0.61–1.28) 0.70 (0.49–1.02)
Ref 2.01 (1.34–2.22)*
Ref 1.41 (1.02–1.99)*
Ref 1.55 (1.25–1.93)*
Ref 1.22 (0.99–1.49)
0.80 (0.35–1.77) NA 1.27 (0.61–2.99) 0.50 (0.06–3.33)
0.69 (0.22–1.49) 1.31 (0.21–7.11) 0.81 (0.39–1.72) NA
0.81 (0.36–1.79) NA 1.16 (0.49–2.79) 0.51 (0.06–4.32)
0.66 (0.29–1.48) 1.30 (0.22–7.67) 0.66 (0.26–1.69) NA
1.21 (0.93–1.44) 0.55 (0.40–0.81)* 1.25 (1.00–1.98)* 0.81 (0.50–1.44)
0.61 (0.45–0.88)* 1.21 (0.99–1.82) 1.39 (1.11–1.73)* 0.99 (0.65–1.61)
1.13 (0.89–1.43) 0.59 (0.40–0.86)* 1.23 (0.98–1.56) 0.85 (0.51–1.43)
0.68 (0.54–0.86)* 1.27 (0.93–1.73) 1.23 (1.00–1.53)* 0.95 (0.54–1.65)
OR ¼ odds ratio; CI ¼ confidence interval. *P < .01. †Adjusted for sex, age, race/ethnicity, insurance type, autism-related medication use, and select co-occurring conditions.
ACADEMIC PEDIATRICS
autism may convey such a large risk for overweight and obesity that the impact of medications is obscured. Finally, this study may be underpowered to detect a difference based on medication use as a result of the small number of patients prescribed these medications. LIMITATIONS It is important to note several limitations of our study. First, this was a cross-sectional study, and therefore, we could not address the time course of overweight or obesity in this population. Second, our ability to evaluate factors associated with overweight and obesity in this population was limited. The factors we examined were few and crudely measured. For example, medications use was measured by prescription data rather than actual documented use or duration of use. Additionally, medications may have been missed if a physician outside of the Partners HealthCare system prescribed a medication. There is also concern for misclassification. First, racial and ethnic minority groups tend to receive a diagnosis of autism later than similar white patients.30 This may mean that some of the younger nonwhite patients in our cohort who were classified as control subjects may have undiagnosed autism. At the same time, some physicians may be hesitant to label a patient with autism as a result of the possible stigma associated with the diagnosis and could lead to additional misclassification. Both misclassifications would bias toward the null hypothesis. One possible explanation for the extremely high rates of overweight and obesity in this cohort is that the population of children with autism is not demographically reflective of the overall autism population. Although this concern cannot be ruled out entirely, our data do closely resemble published data on the population of patients with ASD in Massachusetts.31 Published statewide data from the 2001 to 2005 birth cohort found that 75.3% of children with ASD were white (vs 77.5% white in our cohort) and 82.0% male (vs 79.6% male in our cohort). Our control population was limited in that it was not matched by age, race, or gender. We chose not to match control subjects in order to have a sufficient population derived from primary care clinics. We instead adjusted for age, race, and gender in our model. This also allowed us to look at a control population more representative of the general population. Finally, our data were collected from a single medical center database. Although Partners HealthCare services a diverse population of patients, this may still limit the generalizability to the broader nationwide community of children with autism. Despite these limitations, the results presented here represent the largest study to date with objectively collected height and weight examining the question of overweight and obesity among children with autism.
CONCLUSIONS The prevalence of overweight and obesity in a population of children with autism is much higher than those
OVERWEIGHT AND OBESITY IN AUTISM
413
without autism. These findings are important to future public health initiatives addressing overweight and obesity for several reasons. First, development of future prevention and treatment interventions should take into account this special and growing population7 of patients and consider how they can best be included. Second, because differences in overweight and obesity emerged in the youngest age group evaluated, prevention and anticipatory guidance may be most effective early in a child’s life. Finally, identification of specific environmental factors contributing to overweight and obesity in this population, such as time spent in sedentary activity, access to physical activity programs, poor social support from peers, or use of food as a behavioral reward, should be a priority for future work.
ACKNOWLEDGMENTS Supported in part by the APA Resident Investigator Award.
REFERENCES 1. Ebbeling CB, Pawlak DB, Ludwig DS. Childhood obesity: publichealth crisis, common sense cure. Lancet. 2002;360:473–482. 2. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of childhood and adult obesity in the United States, 2011–2012. JAMA. 2014;311: 806–814. 3. Han JC, Lawlor DA, Kimm SY. Childhood obesity. Lancet. 2010;375: 1737–1748. 4. Biro FM, Wien M. Childhood obesity and adult morbidities. Am J Clin Nutr. 2010;91:1499S–1505S. 5. Lumeng JC, Castle WP, Lumeng CN. The role of pediatricians in the coordinated national effort to address childhood obesity. Pediatrics. 2010;126:574–575. 6. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed., text revision ed. Washington, DC: American Psychiatric Association; 2000. 7. Blumberg SJ, Bramlett MD, Kogan MD, et al. Changes in Prevalence of Parent-Reported Autism Spectrum Disorder in School-Aged US Children: 2007 to 2011–2012. Hyattsville, Md: National Center for Health Statistics; 2013. National Health Statistics Report 65. 8. Schreck KA, Williams K, Smith AF. A comparison of eating behaviors between children with and without autism. J Autism Dev Disord. 2004;34:433–438. 9. Marcason W. What is the current status of research concerning use of a gluten-free, casein-free diet for children diagnosed with autism? J Am Diet Assoc. 2009;109:572. 10. Pan CY. Objectively measured physical activity between children with autism spectrum disorders and children without disabilities during inclusive recess settings in Taiwan. J Autism Dev Disord. 2008 Aug;38:1292–1301. 11. Hinckson EA, Dickinson A, Water T, et al. Physical activity, dietary habits and overall health in overweight and obese children and youth with intellectual disability or autism. Res Dev Disabil. 2013;34: 1170–1178. 12. Rimmer JH, Yamaki K, Lowry BM, et al. Obesity and obesity-related secondary conditions in adolescent with intellectual/developmental disabilities. J Intellect Disabil Res. 2010;54:787–794. 13. Esbensen AJ, Greenberg JS, Seltzer MM, Aman MG. A longitudinal investigation of psychotropic and non-psychotropic medication use among adolescents and adults with autism spectrum disorders. J Autism Dev Disord. 2009;39:1339–1349. 14. Barnard L, Young AH, Pearson J, et al. A systematic review of the use of atypical antipsychotics in autism. J Psychopharmacol. 2002;16: 93–101. 15. Maayan L, Correll CU. Weight gain and metabolic risks associated with antipsychotic medications in children and adolescents. J Child Adolesc Psychopharmacol. 2011;21:517–535.
414
BRODER-FINGERT ET AL
16. Sachs GS, Guille C. Weight gain associated with use of psychotropic medications. J Clin Psychiatry. 1999;60(Suppl 21):16–19. 17. Curtin C, Anderson SE, Must A, Bandini L. The prevalence of obesity in children with autism: a secondary data analysis using nationally representative data from the National Survey of Children’s Health. BMC Pediatr. 2010;10:11. 18. Curtin C, Bandini LG, Perrin EC, et al. Prevalence of overweight in children and adolescents with attention deficit hyperactivity disorder and autism spectrum disorders: a chart review. BMC Pediatr. 2005;5:48. 19. Egan AM, Dreyer ML, Odar CC, et al. Obesity in young children with autism spectrum disorders: prevalence and associated factors. Child Obes. 2013;9:125–131. 20. Curtin C, Jojic M, Bandini LG. Obesity in children with autism spectrum disorder. Harv Rev Psychiatry. 2014;22:93–103. 21. Zuckerman KE, Hill AP, Guion K, et al. Overweight and obesity: prevalence and correlates in a large clinical sample of children with autism spectrum disorder. J Autism Dev Disord. In press. 22. Phillips KL, Schieve LA, Visser S, et al. Prevalence and impact of unhealthy weight in a national sample of US adolescents with autism and other learning and behavioral disabilities. Matern Child Health J. In press. 23. Nalichowski R, Keogh D, Chueh HC, et al. Calculating the benefits of a research patient data repository. AMIA Annu Symp Proc. 2006;1044. 24. Zuckerman KE, Mattox K, Donelan K, et al. Pediatrician identification of Latino children at risk for autism spectrum disorder. Pediatrics. 2013;132:445–453.
ACADEMIC PEDIATRICS 25. Kuczmarski RJ, Ogden CL, Guo SS, et al. 2000 CDC growth charts for the United States: methods and development. National Center for Health Statistics. Vital Health Stat. 2002;11. 26. Long SK, Cook A, Stockley K. Health Insurance Coverage in Massachusetts: Estimates from the 2008 Massachusetts Health Insurance Survey. Boston, Mass: Massachusetts Division of Health Care Finance and Policy, Commonwealth of Massachusetts, and Urban Institute; 2008. Available at: http://www.urban.org/UploadedPDF/411815_ Massachusetts_Insurance.pdf. Accessed May 13, 2014. 27. Murray ML, Hsia Y, Glaser K, et al. Pharmacological treatments prescribed to people with autism spectrum disorder (ASD) in primary health care. Psychopharmacology (Berl). 2014;231: 1011–1021. 28. Bauman ML. Medical comorbidities in autism: challenges to diagnosis and treatment. Neurotherapeutics. 2010;7:320–327. 29. Ogden CL, Carroll MD, Curtin LR, et al. Prevalence of overweight and obesity in the United States, 1999–2004. JAMA. 2006; 295:1549–1555. http://www.urban.org/UploadedPDF/411815_ Massachusetts_Insurance.pdf. Accessed May 13, 2014. 30. Mandell DS, Wiggins LD, Carpenter LA, et al. Racial/ethnic disparities in the identification of children with autism spectrum disorders. Am J Public Health. 2009;99:493–498. 31. Manning SE, Davin CA, Barfield WD, et al. Early diagnoses of autism spectrum disorders in Massachusetts birth cohorts, 2001–2005. Pediatrics. 2011;127:1043–1051.
