- Email: [email protected]
Body Mass Index at the Time of Diagnosis of Autoimmune Type 1 Diabetes in Children
Body Mass Index at the Time of Diagnosis of Autoimmune Type 1 Diabetes in Children Brett M. Kaminski, MPH1, Georgeanna J. Klingensmith, MD2, Roy W. Beck, MD, PhD1, William V. Tamborlane, MD3, Joyce Lee, MD4, Krishna Hassan, MD5, Desmond Schatz, MD6, Craig Kollman, PhD1, and Maria J. Redondo, MD, PhD5, for the Pediatric Diabetes Consortium* Objectives To describe the body mass index (BMI) distribution of children developing autoimmune type 1 diabetes (T1D) compared with the general population and to assess factors associated with BMI at T1D onset. Study design Children age 2-<19 years enrolled in the Pediatric Diabetes Consortium at 7 US pediatric diabetes centers at T1D onset were included. Eligibility for analysis required a diagnosis of T1D, $1 positive diabetes autoantibody, and availability of BMI within 14 days of diagnosis. BMI at diagnosis was compared with the general population as described by the 2000 Centers for Disease Control. Regression analysis was used to assess the association between BMI and various participant characteristics. Results BMI scores for the 490 participants were slightly lower than the 2000 Centers for Disease Control population (P = .04). The median BMI percentile for age and sex was 48th, 11% of the children were overweight (BMI $85th and <95th percentile), 8% obese (BMI $95th and <99th percentile), and 2% severely obese ($99th percentile), percentages that were comparable across age and sex groups. Higher BMI Z-scores were associated with African American and Hispanic race/ethnicity (P = .001) and lower hemoglobin A1c (P < .001), and diabetic ketoacidosis, age, and Tanner stage were not associated. Conclusions Although the BMI distribution in children developing autoimmune T1D was lower than that of the general population, 21% of children were obese or overweight. Youth who are overweight, obese, racial/ethnic minority, and/or present without diabetic ketoacidosis should not be presumed to have type 2 diabetes because many patients with autoantibody-positive T1D present with the same clinical characteristics. (J Pediatr 2013;162:736-40).
T
he increasing prevalence of overweight and obesity in US children and adolescents has been a public health concern for the past 20 years and is predicted to nearly double by 2030.1 A similar trend has been observed in children with type 1 diabetes (T1D) with the percentage of overweight patients tripling to 37% since the 1980s.2 Moreover, recent studies have reported that 10%-34% of youth with obesity and clinically diagnosed with type 2 diabetes (T2D) phenotype have the immunologic markers of T1D,3,4 providing further evidence that children with T1D can no longer be characterized as slender individuals who are “rarely obese” at presentation.5 Although it is likely that differences in body mass index (BMI) influence the spectrum of initial clinical findings of youth with new onset T1D, large studies that have examined this question are scarce in the literature. The Pediatric Diabetes Consortium (PDC) is a collaborative research group From the Jaeb Center for Health Research, Tampa, FL; that is working together to improve the care of children with diabetes through Department of Pediatrics, Barbara Davis Center for 6 Childhood Diabetes, University of Colorado School of sharing of best practices. More than 1000 youth with T1D, ranging in age Medicine, Aurora, CO; Department of Pediatric from 0-18 years, have been enrolled in the initial PDC protocol, whose aim is Endocrinology, Yale University, New Haven, CT; Department of Pediatric Endocrinology, Mott Children’s to describe how diabetes is currently being treated in children and adolescents Hospital, University of Michigan, Ann Arbor, MI; Department of Pediatrics, Texas Children’s Hospital, with new onset T1D at pediatric diabetes specialty centers in the US. The chilBaylor College of Medicine, Houston, TX; and Department of Pediatrics, University of Florida, dren and adolescents who have enrolled in the PDC represent a large and geoGainesville, FL graphically diverse cohort of youth with T1D. The primary purpose of this *A list of members of the Pediatric Diabetes Consortium is available at www.jpeds.com (Appendix). article is to describe how the BMI distribution of children developing autoimThe Pediatric Diabetes Consortium and its activities are mune T1D compares with the general population, as described by the Centers supported by the Jaeb Center for Health Research Foundation through an unrestrictive grant from Novo for Disease Control (CDC). We also determined the influence of demographic 1
2
3
4
5
6
BMI CDC DKA HbA1c NHANES PDC T1D T2D
Body mass index Centers for Disease Control Diabetic ketoacidosis Hemoglobin A1c National Health and Nutrition Examination Survey Pediatric Diabetes Consortium Type 1 diabetes Type 2 diabetes
Nordisk, Inc. The University of Michigan center was supported by the Michigan Diabetes Research and Training Center funded by the National Institute of Diabetes and Digestive and Kidney Diseases (DK020572). Novo Nordisk and NIDDK were not involved in: (1) study design; (2) the collection, analysis, and interpretation of data; (3) the writing of the report; and (4) the decision to submit the manuscript for publication. The authors declare no conflicts of interest. Portions of the study were presented at International Society for Pediatric and Adolescent Diabetes, Miami Beach, FL, USA, October 19-22, 2011. 0022-3476/$ - see front matter. Copyright ª 2013 Mosby Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2012.09.017
736
Vol. 162, No. 4 April 2013 (ie, age, sex, race/ethnicity) and socioeconomic factors on BMI, as well as the potential impact of BMI on clinical presentation of T1D.
Methods The PDC has enrolled 1053 children and adolescents <19 years of age diagnosed with T1D from July 2009 to April 2011 and managed by 1 of the 7 consortium centers within 3 months of diagnosis. Written informed consent was obtained from participants 18 years of age and from parents of children <18 years of age. Written assent also was obtained from youth according to the local Human Subjects Investigational Review Board guidelines. To be included in this study, participants had to have a clinical diagnosis of autoimmune T1D, including at least 1 diabetes autoantibody present (islet cell autoantibody 512/insulinoma associated antibody 2, and/or glutamic acid decarboxylase autoantibody, within 91 days of diagnosis, and/or insulin autoantibody within 14 days of diagnosis), have height and weight data measured by health care provider within 14 days of diagnosis, and be $2 years (to determine BMI percentile), and <19 years of age. One participant was excluded for having a BMI that was an outlier because of failure to thrive and other severe medical conditions. Demographic, socioeconomic, and clinical data were retrieved from medical records and from interview of the participant and/or parent. Clinical characteristics at time of diagnosis included age at diagnosis, presence of diabetic ketoacidosis (DKA), weight, height, and Tanner stage (from examination within 91 days of diagnosis, imputed as stage 1 for girls <8 years of age and boys <10 years of age). The hemoglobin A1c (HbA1c) measurement within 14 days of diagnosis was used in analysis. DKA status at presentation was classified as confirmed DKA (either pH <7.3 or serum bicarbonate <15 mEq/l),7 unconfirmed DKA (provider determined only), or no DKA. Unconfirmed DKA cases were not included in the analyses. BMI was computed from the closest height and weight measured by the health care provider within 14 days of diagnosis (mean 3 days) and computed as weight (in kilograms) divided by the square of height (in meters). BMI percentile and SDS for age and sex was calculated using 2000 CDC population growth chart data.8 According to pediatric standards for ages 2-19 years, patients were classified as overweight when BMI percentile was $85% and <95%, obese when BMI percentile was $95% and <99%, and severely obese when BMI percentile was $99%.9 Statistical Analyses The BMI distribution of the T1D cohort was compared with the BMI distribution of the 2000 CDC population8 using Student t test. Comparisons also were performed in subgroups based on age, sex, and age-sex combined. Although results are reported using the entire range of BMI z-score data, analyses were performed with truncated data (3 SD) to verify that outliers did not have undue leverage.
Least squares regression was used to determine if BMI z-score at diagnosis was associated with DKA status, HbA1c at diagnosis, age, sex, race/ethnicity, number of diabetes-associated autoantibodies, and/or Tanner stage. A multivariate model was constructed using stepwise selection with P < .10 required to be included in the model. We had information on family income, parent education, and health insurance but elected to use parental education given collinearity among the variables. Only factors with P values <.01 were considered statistically significant because of multiple comparisons, although factors with P < .10 were included in the model to adjust for potential confounding. Interaction terms were tested for all variables included in the final multivariate model with P value <.01 required to be included. Continuous variables were examined for nonlinear trends by testing quadratic terms in the regression model. Multivariate model residuals were examined for an approximate normal distribution. All reported P values were 2-sided. All analyses were conducted using SAS v. 9.3 (SAS Institute, Cary, North Carolina).
Results For the 490 children included in the analysis, the mean age at the time of T1D diagnosis was 9.5 years (range 2-18 years); 48% were female, 61% non-Hispanic White, 26% Hispanic, 7% African American, and 6% other race. Of the 355 patients tested for all 3 diabetes autoantibodies, 134 (38%) were positive for all 3, 143 (40%) were positive for 2 out of 3, and 78 (22%) were positive for 1 only. A total of 147 children (31%) presented with DKA. By ethnicity, 31% of non-Hispanic White, 24% of Hispanic, and 47% of African American children presented with DKA. The majority (66%) of patients was prepubertal: 100% were 2-<5 years old, 99% were 5-<9 years old, 57% were 9-<12 years old, and 6% were 12-<19 years old. Family income and parental education were relatively high (53% with income $$75 000 and 53% with a college education) and most had some form of health insurance (65% private insurance and 30% Children’s Health Plan/ Medicaid/Medicare). A family history of T1D (parent or sibling) was present in 7% of participants, and 6% had a family history of T2D. Thirty-seven comorbidities were present in 36 (7%) patients, including thyroid disease (n = 21), celiac disease (n = 7), asthma (n = 7), autoimmune adrenal disease (n = 1), and alopecia areata totalis (n = 1). The median BMI percentile for age and sex was 48th overall and 40th, 49th, 55th, and 48th for patients 2-<5 year old, 5-<9 year old, 9-<12 years old, and 12-<19 years old, respectively. Overall, 11% were overweight, 8% obese, and 2% severely obese, percentages that were comparable across age and sex groups (Table I). The PDC new onset T1D cohort mean BMI z-score was slightly lower than the 2000 CDC population8 (mean = 0.13, 95% CI: 0.25 to 0.01, P = .04; Table I). The mean BMI z-score in new onset T1D girls was lower than 737
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 162, No. 4
Table I. Comparison of BMI distribution in PDC cohort and 2000 CDC population, by age, sex, and race/ethnicity
All Age group at diagnosis 2-<5 years 5-<12 years 12-<19 years Sex Female Male Sex by age group 2-<5 years Female Male 5-<12 years Female Male 12-<19 years Female Male Race White Hispanic African American Other
n
BMI z-score mean (95% CI)
Comparison of PDC and CDC BMI z-scores P value
Normal weight (<85th percentile)
Overweight (‡85th-<95th percentile)
Obese (‡95th-<99th percentile)
Severely obese (‡99th percentile)
490
0.13 (0.25, 0.01)
.04
79%
11%
8%
2%
71 289 130
0.20 (0.52, +0.12) 0.11 (0.27, +0.05) 0.14 (0.38, +0.10)
.21 .19 .26
76% 79% 81%
14% 11% 8%
7% 8% 8%
3% 2% 2%
236 254
0.30 (0.49, 0.12) +0.03 (0.13, +0.19)
.001 .69
79% 78%
11% 11%
8% 7%
1% 4%
31 40
0.13 (0.63, +0.36) 0.25 (0.69, +0.18)
.59 .25
71% 80%
19% 10%
6% 8%
3% 3%
151 138
0.39 (0.61, 0.16) +0.20 (0.02, +0.42)
<.001 .07
82% 75%
9% 14%
8% 7%
1% 4%
54 76
0.16 (0.58, +0.26) 0.12 (0.41, +0.17)
.45 .41
76% 84%
11% 7%
11% 7%
2% 3%
290 123 35 31
0.30 (0.46, 0.15) +0.26 (+0.03, +0.50) +0.14 (0.44, +0.72) 0.22 (0.69, +0.25)
<.001 .03 .62 .35
82% 72% 60% 90%
10% 13% 26% 3%
7% 10% 9% 3%
1% 6% 6% 3%
in girls from the CDC population (mean = 0.30, 95% CI: 0.49 to 0.12, P = .001). By age, the mean BMI z-score for girls was 0.13 for 2-<5 year olds, 0.39 for 5-<12 year olds, and 0.16 for 12-<19 year olds. This was significantly different from the CDC population for girls 5-<12 years old (P = .001). The mean BMI z-score in T1D boys did not differ from those in the CDC population (mean = +0.03, 95 % CI: 0.13 to +0.19, P = .69). The mean BMI z-score for boys was 0.25 for 2-<5 year olds, +0.20 for 5-<12 year olds, and 0.12 for 12-<19 year olds, which were not significantly different from the CDC population (Table I). In the multivariate analysis (Table II), higher BMI z-scores were associated with Hispanic and African American race/ ethnicity (P = .001) and lower HbA1c (P < .001). Compared with non-Hispanic White patients, Hispanics had a higher mean BMI z-score (+0.26 vs 0.30; Figure) and higher percentage of obese/severely obese patients (16% vs 8%). BMI z-scores and percent overweight and/or obese in African Americans were also higher than in nonHispanic White patients and appeared similar to Hispanics, although the African American sample size was small (n = 35; Table I). There were also trends towards higher BMI percentiles in boys, children with low parental education, and children positive for 3 diabetes autoantibodies that did not achieve our criterion for statistical significance. BMI z-score at diagnosis had a small inverse association with HbA1c (approximately 3 percentile decrease in BMI for every 1% HbA1c increase) after adjustment for clinic, race/ethnicity, sex, number of diabetes autoantibodies, and parent education (P < .001; Table II). The BMI z-score did not significantly vary based on age at diagnosis, Tanner pubertal stage, DKA at presentation, or family history of T1D or T2D (Table II). 738
Discussion In children with new onset autoimmune T1D, the average BMI adjusted for age and sex was lower than in the 2000 CDC general population,8 and yet 21% of the children were overweight or obese. The lower BMI of our T1D cohort is not surprising because of weight loss that often precedes T1D diagnosis. Our T1D cohort would most likely have even lower BMI compared with the current population because of the secular increase in BMI trend in the US population, which is not accounted for in 2000 CDC growth charts.10 The 2000 CDC growth charts were based on 5 surveys conducted between 1963 and 1994 but weight and BMI growth charts excluded data from the National Health and Nutrition Examination Survey (NHANES) III 1988-1994 survey for children ages 6 years and older. This exclusion was intended to avoid ‘normalizing’ overweight and obesity.10 Therefore, the 2000 CDC growth charts for weight and BMI, particularly in children 6 years and older, reflect the US population between 1963 and 1988. Analysis of the NHANES 2009-2010 reported that 31.8% of 2-<19 year olds were overweight or obese and 16.9% of the population were obese11 compared with 21% and 10%, respectively, in this new onset T1D cohort. When stratified by sex, we found new onset T1D girls to be lighter than girls in the 2000 CDC population and the boys did not differ. This finding seems to be a reflection of the differential BMI trend by sex reported by the NHANES Survey 2009-2010, which found an increase in obesity between 1999 and 2010 in boys ages 6-19, which was absent in girls,11 and may have reduced the comparative impact of weight loss among boys with newly diagnosed T1D. Kaminski et al
ORIGINAL ARTICLES
April 2013
Table II. Association of participant characteristics with BMI Z-score (n = 490)
All Age at diagnosis 2-<5 years 5-<12 years 12-<19 years Sex Male Female Race/ethnicity† White Hispanic or Latino Black/African American Other Tanner stage† 1 2 3 4 5 DKA at diagnosis† Yes No HbA1c at diagnosis† <10.0% 10.0-<13.0% 13.0-<15.0% $15.0% # Positive autoantibodies† 1 2 3 Parent education† <12 years 12 years Associate degree Bachelor’s degree Master/Professional degree Family history of T1D Parent and/or sibling with T1D No parent or sibling with T1D Family history of T2D Parent and/or sibling with T2D No parent or sibling with T2D
n
BMI Z-score mean (SD)
490
0.13 (1.37)
71 289 130
0.20 (1.35) 0.11 (1.38) 0.14 (1.39)
254 236
+0.03 (1.30) 0.30 (1.43)
290 123 35 31
0.30 (1.32) +0.26 (1.34) +0.14 (1.68) 0.22 (1.28)
172 27 16 16 29
0.31 (1.50) +0.15 (0.98) 0.71 (1.53) 0.15 (1.49) +0.02 (1.39)
147 331
0.25 (1.46) 0.07 (1.34)
135 207 90 44
+0.23 (1.33) 0.18 (1.38) 0.31 (1.43) 0.70 (1.11)
78 143 134
0.37 (1.35) 0.20 (1.40) +0.17 (1.28)
27 116 53 115 107
+0.33 (1.15) 0.03 (1.49) +0.05 (1.41) 0.26 (1.29) 0.49 (1.30)
35 455
+0.16 (1.29) 0.15 (1.38)
31 459
+0.25 (1.66) 0.15 (1.35)
Univariate P value
Multivariate* Difference (99% CI)
P value
.55
.007
.011 Reference 0.30 (0.61, 0.00)
<.001
.001 Reference +0.62 (+0.21, +1.03) +0.43 (0.20, +1.05) +0.30 (0.35, +0.94)
.36
.18 <.001z
<.001 Reference 0.42 (0.80, 0.09) 0.54 (1.02, 0.06) 0.68 (1.29, 0.04) .003 Reference +0.18 (0.30, +0.66) +0.42 (0.06, +0.92) <.001
.02z
.02z
+0.64 (0.14, +1.42) +0.32 (0.17, +0.81) +0.32 (0.25, +0.90) +0.16 (0.30, +0.61) Reference .20 .11
*To be considered statistically significant, a factor must have P < .01. However, factors with P < .10 were also included in the multivariate model to adjust for potential confounding. Factors with P > .10 were excluded. Model also included adjustment for clinic (P < .001). †Missing data for race (n = 11), Tanner stage (n = 230), DKA at diagnosis (n = 12), HbA1c at diagnosis (n = 14), number positive autoantibodies (n = 135), parent education (n = 72). zP value determined by analyzing factor as continuous/ordinal.
Similar to the general population,11,12 the BMI of our T1D cohort varied by race/ethnicity and socioeconomic status with non-Hispanic White having the lowest prevalence of overweight and obesity. The 2000 CDC growth charts do not report BMI reference data by race/ethnicity. However, consistent with our T1D cohort prevalence estimates, NHANES11 reported that in the African American youth, 39.1% were overweight or obese and 24.3% were obese; Hispanic youth also were more likely to be overweight than non-Hispanic White youth with rates similar to African American youth. In a T1D study, Keller et al found a prevalence of obesity similar to our study for African American children (16% vs 14%) and slightly lower (12% compared with 15%) for Hispanic children.12 The latter may reflect differences in study population, such as age, T1D definition (our study included autoantibody-positive only), or
Hispanic sub-groups. In our T1D cohort, BMI increased with lower socioeconomic status, which has been reported in the general US population as well.13 BMI percentile was not associated with age of onset in our cohort, which was similarly reported in a study on autoimmune T1D children from the SEARCH for Diabetes in Youth study.14 We observed a small but significant inverse relationship between HbA1c and BMI. Although the magnitude is fairly inconsequential for an individual (approximately 3 percentile decrease in BMI for every 1% HbA1c increase), an overall trend was observed. This inverse association may be caused by less weight loss prior to T1D diagnosis in children who had lesser degree or shorter time of hyperglycemia and, thus, higher BMI and lower HbA1c at diagnosis. It also may suggest that heavier children present with lower HbA1c as a reflection of greater b-cell function.
Body Mass Index at the Time of Diagnosis of Autoimmune Type 1 Diabetes in Children
739
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 162, No. 4 clinical implication of our findings is that health care providers should not assume that overweight and obese minority youth who present without DKA have T2D because many patients with antibody positive T1D present with the same clinical characteristics. n Submitted for publication Feb 17, 2012; last revision received Jul 24, 2012; accepted Sep 7, 2012. Reprint requests: Roy W. Beck, MD, PhD, Jaeb Center for Health Research, 15310 Amberly Drive, Suite 350 Tampa, FL. E-mail: [email protected]
References
Figure. BMI Z-score distribution by race/ethnicity (n = 490). Eleven patients with missing race/ethnicity data are not included. Bottom and top of each box denote the 25th and 75th percentiles. Horizontal line inside each box denotes the median and the C symbol denotes the mean.
In summary, our data indicate that the BMI distribution in children with autoimmune T1D is slightly lower than the 2000 CDC growth chart population8 and similar to that of the 2009-2010 general population.11 Despite the weight loss that often precedes T1D diagnosis, up to 21% of youth presenting with autoimmune T1D are obese or overweight. Furthermore, BMI distribution in children with new onset T1D is influenced by sex, race/ethnicity, and socioeconomic factors in a similar way as the general population. The finding that heavier children present with lower HbA1c may warrant further investigation. Our results suggest that although, as reported by Keller et al,12 overweight/obesity and racial/ethnic minority increase the odds of a diagnosis of T2D instead of T1D, T1D versus other types of diabetes cannot be definitely diagnosed or ruled out based on BMI alone or in combination with race/ethnicity. Overweight/obese children with T1D are more likely to be Hispanic or African American, and T2D is more common in these race/ethnicities than in nonHispanic Whites; a substantial proportion of children with autoimmune T1D have Hispanic or African American backgrounds and are overweight/obese. Finally, it is a minority of children who present with DKA at the onset of T1D. The
740
1. Wang Y, Beydoun MA, Liang L, Caballero B, Kumanyika SK. Will all Americans become overweight or obese? Estimating the progression and cost of the US obesity epidemic. Obesity (Silver Spring) 2008;16: 2323-30. 2. Libman IM, Pietropaolo M, Arslanian SA, LaPorte RE, Becker DJ. Changing prevalence of overweight children and adolescents at onset of insulin-treated diabetes. Diabetes Care 2003;26:2871-5. 3. Klingensmith GJ, Pyle L, Arslanian S, Copeland KC, Cuttler L, Kaufman F, et al. The presence of GAD and IA-2 antibodies in youth with a type 2 diabetes phenotype: results from the TODAY study. Diabetes Care 2010;33:1970-5. 4. Hathout EH, Thomas W, El-Shahawy M, Nahab F, Mace JW. Diabetic autoimmune markers in children and adolescents with type 2 diabetes. Pediatrics 2001;107:E102. 5. Diagnosis and classification of diabetes mellitus. Diabetes Care 2011; 34(Suppl 1):S62-9. 6. The pediatric diabetes consortium: improving care of children with type 1 diabetes through collaborative research. Diabetes Technol Ther 2010; 12:685-8. 7. Wolfsdorf J, Glaser N, Sperling MA. Diabetic ketoacidosis in infants, children, and adolescents: a consensus statement from the American Diabetes Association. Diabetes Care 2006;29:1150-9. 8. Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, Flegal KM, Guo SS, Wei R, et al. CDC growth charts: United States. Adv Data 2000;314:1-27. 9. Barlow SE. Expert committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: summary report. Pediatrics 2007;120(Suppl 4):S164-92. 10. Ogden CL, Kuczmarski RJ, Flegal KM, Mei Z, Guo S, Wei R, et al. Centers for Disease Control and Prevention 2000 growth charts for the United States: improvements to the 1977 National Center for Health Statistics version. Pediatrics 2002;109:45-60. 11. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of obesity and trends in body mass index among US children and adolescents, 19992010. JAMA 2012;307:483-90. 12. Keller N, Bhatia S, Braden JN, Gildengorin G, Johnson J, Yedlin R, et al. Distinguishing type 2 diabetes from type 1 diabetes in African American and Hispanic American pediatric patients. PLoS One 2012;7:e32773. 13. Levine JA. Poverty and obesity in the US. Diabetes 2011;60:2667-8. 13. 14. Dabelea D, D’Agostino RB Jr, Mayer-Davis EJ, Pettitt DJ, Imperatore G, Dolan LM, et al. Testing the accelerator hypothesis: body size, b-cell function, and age at onset of type 1 (autoimmune) diabetes. Diabetes Care 2006;29:290-4.
Kaminski et al
ORIGINAL ARTICLES
April 2013
Appendix The PDC include (personnel are listed as [PI] for Principal Investigator, [I] for Co-Investigator, and [C] for Coordinators): Baylor College of Medicine, Houston, TX: Morey Haymond, MD (PI), Maria J. Redondo, MD, PhD (I), Krishna Hassan, MD (C), and Kathy Shippy, RN, CCRP (C); Children’s Hospital of Los Angeles, Los Angeles, CA: Jamie Wood, MD (PI), Brian Ichihara, BA (C), Megan Lipton, MA, CCRP (C), and Marisa Cohen, MPH (C); Stanford University, Stanford, CA: Bruce Buckingham, MD (PI), Jennifer Block, BSRN, CDE (C), Breanne Harris, BS (C), and Satya Shanmugham, BS (C); Barbara Davis Center for Childhood Diabetes, University of Colorado School of Medicine, Aurora, CO: Georgeanna J. Klingensmith, MD (PI), Eric Cruz, BA (C), Heidi Haro, BA, BS (C), and Maria King, BA (C); University of Florida, Gainesville, FL: Desmond Schatz, MD (PI), Janet Silverstein, MD (I), Michael J. Haller, MD (I), and Erica Dougherty, BS (C); Yale University, New Haven, CT: William V. Tamborlane, MD (I), Eda Cengiz, MD (PI), Melody Martin, CCRP (C), Amy Steffen, BA (C), and Lori Carria, MS (C); University of Michigan, Ann Arbor, MI: Joyce Lee, MD, MPH (PI), and Surair Bashir, MPH (C); Coordinating Center: Jaeb Center for Health Research, Tampa, FL: Roy W. Beck, MD, PhD, Brett Kaminski, MPH, Craig Kollman, PhD, Siew Wong-Jacobson, MPH, Katrina J. Ruedy, MSPH, Callyn Hall, BS, and Beth Stevens, BA.
Body Mass Index at the Time of Diagnosis of Autoimmune Type 1 Diabetes in Children
740.e1
T
he increasing prevalence of overweight and obesity in US children and adolescents has been a public health concern for the past 20 years and is predicted to nearly double by 2030.1 A similar trend has been observed in children with type 1 diabetes (T1D) with the percentage of overweight patients tripling to 37% since the 1980s.2 Moreover, recent studies have reported that 10%-34% of youth with obesity and clinically diagnosed with type 2 diabetes (T2D) phenotype have the immunologic markers of T1D,3,4 providing further evidence that children with T1D can no longer be characterized as slender individuals who are “rarely obese” at presentation.5 Although it is likely that differences in body mass index (BMI) influence the spectrum of initial clinical findings of youth with new onset T1D, large studies that have examined this question are scarce in the literature. The Pediatric Diabetes Consortium (PDC) is a collaborative research group From the Jaeb Center for Health Research, Tampa, FL; that is working together to improve the care of children with diabetes through Department of Pediatrics, Barbara Davis Center for 6 Childhood Diabetes, University of Colorado School of sharing of best practices. More than 1000 youth with T1D, ranging in age Medicine, Aurora, CO; Department of Pediatric from 0-18 years, have been enrolled in the initial PDC protocol, whose aim is Endocrinology, Yale University, New Haven, CT; Department of Pediatric Endocrinology, Mott Children’s to describe how diabetes is currently being treated in children and adolescents Hospital, University of Michigan, Ann Arbor, MI; Department of Pediatrics, Texas Children’s Hospital, with new onset T1D at pediatric diabetes specialty centers in the US. The chilBaylor College of Medicine, Houston, TX; and Department of Pediatrics, University of Florida, dren and adolescents who have enrolled in the PDC represent a large and geoGainesville, FL graphically diverse cohort of youth with T1D. The primary purpose of this *A list of members of the Pediatric Diabetes Consortium is available at www.jpeds.com (Appendix). article is to describe how the BMI distribution of children developing autoimThe Pediatric Diabetes Consortium and its activities are mune T1D compares with the general population, as described by the Centers supported by the Jaeb Center for Health Research Foundation through an unrestrictive grant from Novo for Disease Control (CDC). We also determined the influence of demographic 1
2
3
4
5
6
BMI CDC DKA HbA1c NHANES PDC T1D T2D
Body mass index Centers for Disease Control Diabetic ketoacidosis Hemoglobin A1c National Health and Nutrition Examination Survey Pediatric Diabetes Consortium Type 1 diabetes Type 2 diabetes
Nordisk, Inc. The University of Michigan center was supported by the Michigan Diabetes Research and Training Center funded by the National Institute of Diabetes and Digestive and Kidney Diseases (DK020572). Novo Nordisk and NIDDK were not involved in: (1) study design; (2) the collection, analysis, and interpretation of data; (3) the writing of the report; and (4) the decision to submit the manuscript for publication. The authors declare no conflicts of interest. Portions of the study were presented at International Society for Pediatric and Adolescent Diabetes, Miami Beach, FL, USA, October 19-22, 2011. 0022-3476/$ - see front matter. Copyright ª 2013 Mosby Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpeds.2012.09.017
736
Vol. 162, No. 4 April 2013 (ie, age, sex, race/ethnicity) and socioeconomic factors on BMI, as well as the potential impact of BMI on clinical presentation of T1D.
Methods The PDC has enrolled 1053 children and adolescents <19 years of age diagnosed with T1D from July 2009 to April 2011 and managed by 1 of the 7 consortium centers within 3 months of diagnosis. Written informed consent was obtained from participants 18 years of age and from parents of children <18 years of age. Written assent also was obtained from youth according to the local Human Subjects Investigational Review Board guidelines. To be included in this study, participants had to have a clinical diagnosis of autoimmune T1D, including at least 1 diabetes autoantibody present (islet cell autoantibody 512/insulinoma associated antibody 2, and/or glutamic acid decarboxylase autoantibody, within 91 days of diagnosis, and/or insulin autoantibody within 14 days of diagnosis), have height and weight data measured by health care provider within 14 days of diagnosis, and be $2 years (to determine BMI percentile), and <19 years of age. One participant was excluded for having a BMI that was an outlier because of failure to thrive and other severe medical conditions. Demographic, socioeconomic, and clinical data were retrieved from medical records and from interview of the participant and/or parent. Clinical characteristics at time of diagnosis included age at diagnosis, presence of diabetic ketoacidosis (DKA), weight, height, and Tanner stage (from examination within 91 days of diagnosis, imputed as stage 1 for girls <8 years of age and boys <10 years of age). The hemoglobin A1c (HbA1c) measurement within 14 days of diagnosis was used in analysis. DKA status at presentation was classified as confirmed DKA (either pH <7.3 or serum bicarbonate <15 mEq/l),7 unconfirmed DKA (provider determined only), or no DKA. Unconfirmed DKA cases were not included in the analyses. BMI was computed from the closest height and weight measured by the health care provider within 14 days of diagnosis (mean 3 days) and computed as weight (in kilograms) divided by the square of height (in meters). BMI percentile and SDS for age and sex was calculated using 2000 CDC population growth chart data.8 According to pediatric standards for ages 2-19 years, patients were classified as overweight when BMI percentile was $85% and <95%, obese when BMI percentile was $95% and <99%, and severely obese when BMI percentile was $99%.9 Statistical Analyses The BMI distribution of the T1D cohort was compared with the BMI distribution of the 2000 CDC population8 using Student t test. Comparisons also were performed in subgroups based on age, sex, and age-sex combined. Although results are reported using the entire range of BMI z-score data, analyses were performed with truncated data (3 SD) to verify that outliers did not have undue leverage.
Least squares regression was used to determine if BMI z-score at diagnosis was associated with DKA status, HbA1c at diagnosis, age, sex, race/ethnicity, number of diabetes-associated autoantibodies, and/or Tanner stage. A multivariate model was constructed using stepwise selection with P < .10 required to be included in the model. We had information on family income, parent education, and health insurance but elected to use parental education given collinearity among the variables. Only factors with P values <.01 were considered statistically significant because of multiple comparisons, although factors with P < .10 were included in the model to adjust for potential confounding. Interaction terms were tested for all variables included in the final multivariate model with P value <.01 required to be included. Continuous variables were examined for nonlinear trends by testing quadratic terms in the regression model. Multivariate model residuals were examined for an approximate normal distribution. All reported P values were 2-sided. All analyses were conducted using SAS v. 9.3 (SAS Institute, Cary, North Carolina).
Results For the 490 children included in the analysis, the mean age at the time of T1D diagnosis was 9.5 years (range 2-18 years); 48% were female, 61% non-Hispanic White, 26% Hispanic, 7% African American, and 6% other race. Of the 355 patients tested for all 3 diabetes autoantibodies, 134 (38%) were positive for all 3, 143 (40%) were positive for 2 out of 3, and 78 (22%) were positive for 1 only. A total of 147 children (31%) presented with DKA. By ethnicity, 31% of non-Hispanic White, 24% of Hispanic, and 47% of African American children presented with DKA. The majority (66%) of patients was prepubertal: 100% were 2-<5 years old, 99% were 5-<9 years old, 57% were 9-<12 years old, and 6% were 12-<19 years old. Family income and parental education were relatively high (53% with income $$75 000 and 53% with a college education) and most had some form of health insurance (65% private insurance and 30% Children’s Health Plan/ Medicaid/Medicare). A family history of T1D (parent or sibling) was present in 7% of participants, and 6% had a family history of T2D. Thirty-seven comorbidities were present in 36 (7%) patients, including thyroid disease (n = 21), celiac disease (n = 7), asthma (n = 7), autoimmune adrenal disease (n = 1), and alopecia areata totalis (n = 1). The median BMI percentile for age and sex was 48th overall and 40th, 49th, 55th, and 48th for patients 2-<5 year old, 5-<9 year old, 9-<12 years old, and 12-<19 years old, respectively. Overall, 11% were overweight, 8% obese, and 2% severely obese, percentages that were comparable across age and sex groups (Table I). The PDC new onset T1D cohort mean BMI z-score was slightly lower than the 2000 CDC population8 (mean = 0.13, 95% CI: 0.25 to 0.01, P = .04; Table I). The mean BMI z-score in new onset T1D girls was lower than 737
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 162, No. 4
Table I. Comparison of BMI distribution in PDC cohort and 2000 CDC population, by age, sex, and race/ethnicity
All Age group at diagnosis 2-<5 years 5-<12 years 12-<19 years Sex Female Male Sex by age group 2-<5 years Female Male 5-<12 years Female Male 12-<19 years Female Male Race White Hispanic African American Other
n
BMI z-score mean (95% CI)
Comparison of PDC and CDC BMI z-scores P value
Normal weight (<85th percentile)
Overweight (‡85th-<95th percentile)
Obese (‡95th-<99th percentile)
Severely obese (‡99th percentile)
490
0.13 (0.25, 0.01)
.04
79%
11%
8%
2%
71 289 130
0.20 (0.52, +0.12) 0.11 (0.27, +0.05) 0.14 (0.38, +0.10)
.21 .19 .26
76% 79% 81%
14% 11% 8%
7% 8% 8%
3% 2% 2%
236 254
0.30 (0.49, 0.12) +0.03 (0.13, +0.19)
.001 .69
79% 78%
11% 11%
8% 7%
1% 4%
31 40
0.13 (0.63, +0.36) 0.25 (0.69, +0.18)
.59 .25
71% 80%
19% 10%
6% 8%
3% 3%
151 138
0.39 (0.61, 0.16) +0.20 (0.02, +0.42)
<.001 .07
82% 75%
9% 14%
8% 7%
1% 4%
54 76
0.16 (0.58, +0.26) 0.12 (0.41, +0.17)
.45 .41
76% 84%
11% 7%
11% 7%
2% 3%
290 123 35 31
0.30 (0.46, 0.15) +0.26 (+0.03, +0.50) +0.14 (0.44, +0.72) 0.22 (0.69, +0.25)
<.001 .03 .62 .35
82% 72% 60% 90%
10% 13% 26% 3%
7% 10% 9% 3%
1% 6% 6% 3%
in girls from the CDC population (mean = 0.30, 95% CI: 0.49 to 0.12, P = .001). By age, the mean BMI z-score for girls was 0.13 for 2-<5 year olds, 0.39 for 5-<12 year olds, and 0.16 for 12-<19 year olds. This was significantly different from the CDC population for girls 5-<12 years old (P = .001). The mean BMI z-score in T1D boys did not differ from those in the CDC population (mean = +0.03, 95 % CI: 0.13 to +0.19, P = .69). The mean BMI z-score for boys was 0.25 for 2-<5 year olds, +0.20 for 5-<12 year olds, and 0.12 for 12-<19 year olds, which were not significantly different from the CDC population (Table I). In the multivariate analysis (Table II), higher BMI z-scores were associated with Hispanic and African American race/ ethnicity (P = .001) and lower HbA1c (P < .001). Compared with non-Hispanic White patients, Hispanics had a higher mean BMI z-score (+0.26 vs 0.30; Figure) and higher percentage of obese/severely obese patients (16% vs 8%). BMI z-scores and percent overweight and/or obese in African Americans were also higher than in nonHispanic White patients and appeared similar to Hispanics, although the African American sample size was small (n = 35; Table I). There were also trends towards higher BMI percentiles in boys, children with low parental education, and children positive for 3 diabetes autoantibodies that did not achieve our criterion for statistical significance. BMI z-score at diagnosis had a small inverse association with HbA1c (approximately 3 percentile decrease in BMI for every 1% HbA1c increase) after adjustment for clinic, race/ethnicity, sex, number of diabetes autoantibodies, and parent education (P < .001; Table II). The BMI z-score did not significantly vary based on age at diagnosis, Tanner pubertal stage, DKA at presentation, or family history of T1D or T2D (Table II). 738
Discussion In children with new onset autoimmune T1D, the average BMI adjusted for age and sex was lower than in the 2000 CDC general population,8 and yet 21% of the children were overweight or obese. The lower BMI of our T1D cohort is not surprising because of weight loss that often precedes T1D diagnosis. Our T1D cohort would most likely have even lower BMI compared with the current population because of the secular increase in BMI trend in the US population, which is not accounted for in 2000 CDC growth charts.10 The 2000 CDC growth charts were based on 5 surveys conducted between 1963 and 1994 but weight and BMI growth charts excluded data from the National Health and Nutrition Examination Survey (NHANES) III 1988-1994 survey for children ages 6 years and older. This exclusion was intended to avoid ‘normalizing’ overweight and obesity.10 Therefore, the 2000 CDC growth charts for weight and BMI, particularly in children 6 years and older, reflect the US population between 1963 and 1988. Analysis of the NHANES 2009-2010 reported that 31.8% of 2-<19 year olds were overweight or obese and 16.9% of the population were obese11 compared with 21% and 10%, respectively, in this new onset T1D cohort. When stratified by sex, we found new onset T1D girls to be lighter than girls in the 2000 CDC population and the boys did not differ. This finding seems to be a reflection of the differential BMI trend by sex reported by the NHANES Survey 2009-2010, which found an increase in obesity between 1999 and 2010 in boys ages 6-19, which was absent in girls,11 and may have reduced the comparative impact of weight loss among boys with newly diagnosed T1D. Kaminski et al
ORIGINAL ARTICLES
April 2013
Table II. Association of participant characteristics with BMI Z-score (n = 490)
All Age at diagnosis 2-<5 years 5-<12 years 12-<19 years Sex Male Female Race/ethnicity† White Hispanic or Latino Black/African American Other Tanner stage† 1 2 3 4 5 DKA at diagnosis† Yes No HbA1c at diagnosis† <10.0% 10.0-<13.0% 13.0-<15.0% $15.0% # Positive autoantibodies† 1 2 3 Parent education† <12 years 12 years Associate degree Bachelor’s degree Master/Professional degree Family history of T1D Parent and/or sibling with T1D No parent or sibling with T1D Family history of T2D Parent and/or sibling with T2D No parent or sibling with T2D
n
BMI Z-score mean (SD)
490
0.13 (1.37)
71 289 130
0.20 (1.35) 0.11 (1.38) 0.14 (1.39)
254 236
+0.03 (1.30) 0.30 (1.43)
290 123 35 31
0.30 (1.32) +0.26 (1.34) +0.14 (1.68) 0.22 (1.28)
172 27 16 16 29
0.31 (1.50) +0.15 (0.98) 0.71 (1.53) 0.15 (1.49) +0.02 (1.39)
147 331
0.25 (1.46) 0.07 (1.34)
135 207 90 44
+0.23 (1.33) 0.18 (1.38) 0.31 (1.43) 0.70 (1.11)
78 143 134
0.37 (1.35) 0.20 (1.40) +0.17 (1.28)
27 116 53 115 107
+0.33 (1.15) 0.03 (1.49) +0.05 (1.41) 0.26 (1.29) 0.49 (1.30)
35 455
+0.16 (1.29) 0.15 (1.38)
31 459
+0.25 (1.66) 0.15 (1.35)
Univariate P value
Multivariate* Difference (99% CI)
P value
.55
.007
.011 Reference 0.30 (0.61, 0.00)
<.001
.001 Reference +0.62 (+0.21, +1.03) +0.43 (0.20, +1.05) +0.30 (0.35, +0.94)
.36
.18 <.001z
<.001 Reference 0.42 (0.80, 0.09) 0.54 (1.02, 0.06) 0.68 (1.29, 0.04) .003 Reference +0.18 (0.30, +0.66) +0.42 (0.06, +0.92) <.001
.02z
.02z
+0.64 (0.14, +1.42) +0.32 (0.17, +0.81) +0.32 (0.25, +0.90) +0.16 (0.30, +0.61) Reference .20 .11
*To be considered statistically significant, a factor must have P < .01. However, factors with P < .10 were also included in the multivariate model to adjust for potential confounding. Factors with P > .10 were excluded. Model also included adjustment for clinic (P < .001). †Missing data for race (n = 11), Tanner stage (n = 230), DKA at diagnosis (n = 12), HbA1c at diagnosis (n = 14), number positive autoantibodies (n = 135), parent education (n = 72). zP value determined by analyzing factor as continuous/ordinal.
Similar to the general population,11,12 the BMI of our T1D cohort varied by race/ethnicity and socioeconomic status with non-Hispanic White having the lowest prevalence of overweight and obesity. The 2000 CDC growth charts do not report BMI reference data by race/ethnicity. However, consistent with our T1D cohort prevalence estimates, NHANES11 reported that in the African American youth, 39.1% were overweight or obese and 24.3% were obese; Hispanic youth also were more likely to be overweight than non-Hispanic White youth with rates similar to African American youth. In a T1D study, Keller et al found a prevalence of obesity similar to our study for African American children (16% vs 14%) and slightly lower (12% compared with 15%) for Hispanic children.12 The latter may reflect differences in study population, such as age, T1D definition (our study included autoantibody-positive only), or
Hispanic sub-groups. In our T1D cohort, BMI increased with lower socioeconomic status, which has been reported in the general US population as well.13 BMI percentile was not associated with age of onset in our cohort, which was similarly reported in a study on autoimmune T1D children from the SEARCH for Diabetes in Youth study.14 We observed a small but significant inverse relationship between HbA1c and BMI. Although the magnitude is fairly inconsequential for an individual (approximately 3 percentile decrease in BMI for every 1% HbA1c increase), an overall trend was observed. This inverse association may be caused by less weight loss prior to T1D diagnosis in children who had lesser degree or shorter time of hyperglycemia and, thus, higher BMI and lower HbA1c at diagnosis. It also may suggest that heavier children present with lower HbA1c as a reflection of greater b-cell function.
Body Mass Index at the Time of Diagnosis of Autoimmune Type 1 Diabetes in Children
739
THE JOURNAL OF PEDIATRICS
www.jpeds.com
Vol. 162, No. 4 clinical implication of our findings is that health care providers should not assume that overweight and obese minority youth who present without DKA have T2D because many patients with antibody positive T1D present with the same clinical characteristics. n Submitted for publication Feb 17, 2012; last revision received Jul 24, 2012; accepted Sep 7, 2012. Reprint requests: Roy W. Beck, MD, PhD, Jaeb Center for Health Research, 15310 Amberly Drive, Suite 350 Tampa, FL. E-mail: [email protected]
References
Figure. BMI Z-score distribution by race/ethnicity (n = 490). Eleven patients with missing race/ethnicity data are not included. Bottom and top of each box denote the 25th and 75th percentiles. Horizontal line inside each box denotes the median and the C symbol denotes the mean.
In summary, our data indicate that the BMI distribution in children with autoimmune T1D is slightly lower than the 2000 CDC growth chart population8 and similar to that of the 2009-2010 general population.11 Despite the weight loss that often precedes T1D diagnosis, up to 21% of youth presenting with autoimmune T1D are obese or overweight. Furthermore, BMI distribution in children with new onset T1D is influenced by sex, race/ethnicity, and socioeconomic factors in a similar way as the general population. The finding that heavier children present with lower HbA1c may warrant further investigation. Our results suggest that although, as reported by Keller et al,12 overweight/obesity and racial/ethnic minority increase the odds of a diagnosis of T2D instead of T1D, T1D versus other types of diabetes cannot be definitely diagnosed or ruled out based on BMI alone or in combination with race/ethnicity. Overweight/obese children with T1D are more likely to be Hispanic or African American, and T2D is more common in these race/ethnicities than in nonHispanic Whites; a substantial proportion of children with autoimmune T1D have Hispanic or African American backgrounds and are overweight/obese. Finally, it is a minority of children who present with DKA at the onset of T1D. The
740
1. Wang Y, Beydoun MA, Liang L, Caballero B, Kumanyika SK. Will all Americans become overweight or obese? Estimating the progression and cost of the US obesity epidemic. Obesity (Silver Spring) 2008;16: 2323-30. 2. Libman IM, Pietropaolo M, Arslanian SA, LaPorte RE, Becker DJ. Changing prevalence of overweight children and adolescents at onset of insulin-treated diabetes. Diabetes Care 2003;26:2871-5. 3. Klingensmith GJ, Pyle L, Arslanian S, Copeland KC, Cuttler L, Kaufman F, et al. The presence of GAD and IA-2 antibodies in youth with a type 2 diabetes phenotype: results from the TODAY study. Diabetes Care 2010;33:1970-5. 4. Hathout EH, Thomas W, El-Shahawy M, Nahab F, Mace JW. Diabetic autoimmune markers in children and adolescents with type 2 diabetes. Pediatrics 2001;107:E102. 5. Diagnosis and classification of diabetes mellitus. Diabetes Care 2011; 34(Suppl 1):S62-9. 6. The pediatric diabetes consortium: improving care of children with type 1 diabetes through collaborative research. Diabetes Technol Ther 2010; 12:685-8. 7. Wolfsdorf J, Glaser N, Sperling MA. Diabetic ketoacidosis in infants, children, and adolescents: a consensus statement from the American Diabetes Association. Diabetes Care 2006;29:1150-9. 8. Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, Flegal KM, Guo SS, Wei R, et al. CDC growth charts: United States. Adv Data 2000;314:1-27. 9. Barlow SE. Expert committee recommendations regarding the prevention, assessment, and treatment of child and adolescent overweight and obesity: summary report. Pediatrics 2007;120(Suppl 4):S164-92. 10. Ogden CL, Kuczmarski RJ, Flegal KM, Mei Z, Guo S, Wei R, et al. Centers for Disease Control and Prevention 2000 growth charts for the United States: improvements to the 1977 National Center for Health Statistics version. Pediatrics 2002;109:45-60. 11. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of obesity and trends in body mass index among US children and adolescents, 19992010. JAMA 2012;307:483-90. 12. Keller N, Bhatia S, Braden JN, Gildengorin G, Johnson J, Yedlin R, et al. Distinguishing type 2 diabetes from type 1 diabetes in African American and Hispanic American pediatric patients. PLoS One 2012;7:e32773. 13. Levine JA. Poverty and obesity in the US. Diabetes 2011;60:2667-8. 13. 14. Dabelea D, D’Agostino RB Jr, Mayer-Davis EJ, Pettitt DJ, Imperatore G, Dolan LM, et al. Testing the accelerator hypothesis: body size, b-cell function, and age at onset of type 1 (autoimmune) diabetes. Diabetes Care 2006;29:290-4.
Kaminski et al
ORIGINAL ARTICLES
April 2013
Appendix The PDC include (personnel are listed as [PI] for Principal Investigator, [I] for Co-Investigator, and [C] for Coordinators): Baylor College of Medicine, Houston, TX: Morey Haymond, MD (PI), Maria J. Redondo, MD, PhD (I), Krishna Hassan, MD (C), and Kathy Shippy, RN, CCRP (C); Children’s Hospital of Los Angeles, Los Angeles, CA: Jamie Wood, MD (PI), Brian Ichihara, BA (C), Megan Lipton, MA, CCRP (C), and Marisa Cohen, MPH (C); Stanford University, Stanford, CA: Bruce Buckingham, MD (PI), Jennifer Block, BSRN, CDE (C), Breanne Harris, BS (C), and Satya Shanmugham, BS (C); Barbara Davis Center for Childhood Diabetes, University of Colorado School of Medicine, Aurora, CO: Georgeanna J. Klingensmith, MD (PI), Eric Cruz, BA (C), Heidi Haro, BA, BS (C), and Maria King, BA (C); University of Florida, Gainesville, FL: Desmond Schatz, MD (PI), Janet Silverstein, MD (I), Michael J. Haller, MD (I), and Erica Dougherty, BS (C); Yale University, New Haven, CT: William V. Tamborlane, MD (I), Eda Cengiz, MD (PI), Melody Martin, CCRP (C), Amy Steffen, BA (C), and Lori Carria, MS (C); University of Michigan, Ann Arbor, MI: Joyce Lee, MD, MPH (PI), and Surair Bashir, MPH (C); Coordinating Center: Jaeb Center for Health Research, Tampa, FL: Roy W. Beck, MD, PhD, Brett Kaminski, MPH, Craig Kollman, PhD, Siew Wong-Jacobson, MPH, Katrina J. Ruedy, MSPH, Callyn Hall, BS, and Beth Stevens, BA.
Body Mass Index at the Time of Diagnosis of Autoimmune Type 1 Diabetes in Children
740.e1