- Email: [email protected]
Childhood obesity in a population at high risk for type 2 diabetes
C
Childhood obesity in a population at high risk for
type 2 diabetes T. Kue Young, MD, PhD, Heather J. Dean, MD, Bertha Flett, and Pauline Wood-Steiman
Objectives: To determine the prevalence of obesity and investigate its association with fasting glucose and insulin among children and adolescents in a population at high risk for type 2 diabetes. Design: A cross-sectional screening survey involving anthropometry and fasting serum levels of glucose and insulin. Setting: A remote aboriginal (Ojibwa-Cree) community in northern Manitoba, Canada. Participants: All children aged 4 to 19 years in the community were invited to participate, with a response rate of 82% (n = 719). Main outcome measures: Obesity is defined as body mass index exceeding the 85th percentile of the National Center for Health Statistics reference data. The diagnosis of diabetes and impaired fasting glucose is based on the new criteria of the American Diabetes Association. Results: There is a high prevalence of obesity, with 64% (female) and 60% (male) exceeding the 85th percentile and 40% (female) and 34% (male) exceeding the 95th percentile. Body mass index is a significant predictor of both glucose and insulin in both sexes, independent of age. Obese children are at increased risk of being classified as having diabetes or impaired fasting glucose (odds ratio 5.1, 95% CI 1.51, 17.0). Conclusions: The early onset of type 2 diabetes in childhood is increasingly observed in many populations. Childhood obesity is a strong risk factor. Early detection and intervention directed at obesity are potential strategies to avert the long-term consequences of type 2 diabetes. (J Pediatr 2000;136:365-9)
The epidemic of type 2 diabetes among many Aboriginal/Native American populations in Canada and the United States during the past several decades
has been recognized for some time.1-3 A more recent development is the rising prevalence among children, associated with a decreasing age of onset. In
From the Department of Community Health Sciences and the Department of Pediatrics and Child Health, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba, Canada; and the Island Lake Tribal Council, Winnipeg, Manitoba, Canada. Funding for this study was provided by the National Health Research Development Program of Health Canada (Grant No. 6607-1665-ND). Dr Young is a Senior Scientist of the Medical Research Council of Canada. Submitted for publication Apr 29, 1999; revision received Aug 30, 1999; accepted Oct 1, 1999. Reprint requests: T. Kue Young, Department of Community Health Sciences, University of Manitoba, 750 Bannatyne Ave, Winnipeg, MB, Canada, R3E 0W3. Copyright © 2000 by Mosby, Inc. 0022-3476/2000/$12.00 + 0 9/21/103504 doi:10.1067/mpd.2000.103504
the past decade clinicians have diagnosed this disease in Aboriginal children as young as 7 years of age.4-6 This is a serious concern in view of the increased duration of diabetes and the likelihood of microvascular and macrovascular complications appearing during young adulthood. The cause of the epidemic is not completely understood. Based on studies in adults, obesity has been demonstrated to be one of the most important risk factors for type 2 diabetes.7 It is likely that the emerging pattern and burden of childhood type 2 diabetes is associated with an increasing prevalence of obesity among Aboriginal children. BMI IFG
Body mass index Impaired fasting glucose
We report on a study to determine the prevalence of obesity and its association with metabolic markers of glucose metabolism in a Canadian Aboriginal community known to be at high risk for type 2 diabetes in the adult population. This study was initiated at the request of the community of St Theresa Point in northern Manitoba, whose members have become concerned with the high rate of diabetes in the population and its implications for the future health of their children. A preliminary report on the clinical features of new cases of diabetes detected by screening has been published.5
METHODS Study Setting St Theresa Point First Nation, with a population of 2430, is located 467 km 365
YOUNG ET AL
THE JOURNAL OF PEDIATRICS MARCH 2000 increasing dependence on government subsidies. Storebought food has become predominant in the people’s diet, and the level of physical activity has declined.
Data Collection
Fig 1. Prevalence of obesity by age-sex group based on percentile of National Center for Health Statistics reference population data.
Fig 2. Mean fasting glucose and insulin by age group and sex. by air northeast of Winnipeg, Manitoba, Canada. It is one of 4 communities that make up the Island Lake Tribal Council. In 1875 the Ojibwa-Cree people of the Island Lake area signed Treaty No 5 with the Canadian government. The traditional lifestyle of 366
hunting, trapping, and fishing remained relatively intact well into the 1960s. However, in the decades since then, there has been rapid social change characterized by reduced dependence on the land for subsistence, involvement in the wage economy, and
A project office was established in the school, which had an enrollment of 873 children from nursery school to grade 11 in the 1995 to 1996 and 1996 to 1997 academic years. The children were instructed to arrive at school at 8:00 AM before breakfast on their scheduled day, usually in family groupings. A venous blood sample was taken followed by a nutritious breakfast provided by the project nurse. Standing height was measured with a wallmounted tape, and weight was measured with a standard balance scale. Samples were taken from approximately 5 to 10 children 4 mornings per week between April 1996 and July 1997 except during school holidays. Fasting blood samples were taken to the local health center for centrifuge and storage at 4°C. At the end of each month, samples were packed in dry ice and shipped to the Health Sciences Center in Winnipeg, where they were batched for serum glucose and insulin assays. Children with high fasting glucose levels were referred to the health center for further clinical evaluation by the visiting physician. All results were explained to the parents in person or in writing by the project nurse. Serum glucose was measured with a Hitachi 705 Analyzer and insulin by radioimmunoassay with the Pharmacia kit at the Clinical Chemistry Laboratory of the Health Sciences Center, Winnipeg. This study was approved by the Committee on the Use of Human Subjects in Research of the Faculty of Medicine, University of Manitoba, the Island Lake Diabetes Committee, and the St. Theresa Point First Nation Council. Informed consent was obtained from the parents of all participating children.
YOUNG ET AL
THE JOURNAL OF PEDIATRICS VOLUME 136, NUMBER 3 Table I. Association of fasting glucose and insulin levels with categories of body mass index
Fasting glucose (mmol/L) NCHS percentile Male <85th 85th-94th 95th+ Female <85th 85th-94th 95th+
Fasting insulin (pmol/L)
No.
Mean
95% CI
Mean
95% CI
143 96 122
4.96 5.06 5.10
4.89, 5.02 4.96, 5.17 5.01, 5.19
60.13 66.17 93.61
54.23, 66.03 58.35, 73.99 81.16, 106.06
126 86 140
4.83 4.92 5.06
4.73, 4.93 4.79, 5.06 4.92, 5.20
80.08 100.69 119.11
67.59, 92.57 74.48, 126.90 103.84, 134.38
Results of analysis of variance comparing differences in fasting glucose/insulin across the 3 BMI categories: males, glucose (F = 3.244, P = .04); males, insulin (F = 15.5, P < .001); females, glucose (F = 3.6, P = .028); females, insulin (F = 5.64, P = .004). To convert from SI units to conventional units for glucose, divide by 0.05551; for insulin, divide by 6.0. NCHS, National Center for Health Statistics.
Table II. Prevalence of diabetes and impaired fasting glucose by categories of body mass index
Diabetes NCHS percentile Male <85th 85th-94th 95th+ Female <85th 85th-94th 95th+
Impaired fasting glucose
No.
No cases
Prevalence (%)
No cases
Prevalence (%)
143 96 122
0 0 0
0 0 0
0 6 5
0 6.3 4.1
126 86 140
1 1 6
0.8 1.2 4.3
2 3 3
1.6 3.5 2.1
Diagnostic Criteria A recent World Health Organization Expert Committee on anthropometry recommended the body mass index (weight in kilograms per height in meters squared) as the indicator of choice for overweight among adolescents, because it has been validated by studies on total body fat.8 The recommended reference data are those of the U.S. all-race population derived from the U.S. National Health and Nutrition Examination Survey of 1971 to 1974 published by the National Center for Health Statistics. Individuals above the 85th percentile are considered to be overweight.9 For diabetes the new diagnostic criteria approved by the American Diabetes Association10 were used, which define diabetes in individuals with a fasting
serum glucose ≥7 mmol/L (126 mg/dL). There is also a category of “impaired fasting glucose,” which encompasses individuals with values <7 mmol/L (126 mg/dL) but ≥6.1 mmol/L (110 mg/dL).
RESULTS A total of 719 children (354 girls and 365 boys) ranging in age from 4 to 19 years were screened, a response rate of 82%. All children were attending school in the community at the time of study. Two children did not undergo venipuncture, and they were excluded from further analyses. The prevalence of obesity among children in Island Lake is considerably higher than that in the National Center
for Health Statistics reference population, because 64% of girls and 60% of boys exceed the 85th percentile of the age-sex all-race reference values for BMI, and 40% of girls and 34% of boys exceed the 95th percentile (Fig 1). The mean fasting serum glucose and insulin levels by age and sex are shown in Fig 2. Among normoglycemic individuals girls tend to have higher fasting insulin levels than boys, whereas for glucose, age-sex differences are not pronounced. BMI is correlated with both fasting glucose and insulin, with partial correlation coefficients of 0.224 and 0.406, respectively, controlling for age and sex. Study participants were divided into 3 categories on the basis of their BMI percentiles: <85th, 85th to 94th, and 367
YOUNG ET AL
THE JOURNAL OF PEDIATRICS MARCH 2000
Table III. Association of BMI and fasting glucose levels: multiple linear regression analyses
Model R2
Standardized coefficient (β)
P value
(1) Male only, dependent variable: fasting glucose BMI, age BMI, age, insulin
0.022 0.092
BMI (2.778) BMI (–0.013) Insulin (0.310)
.006 .838 <.001
(2) Female only, dependent variable: fasting glucose BMI, age BMI, age, insulin
0.080 0.145
BMI (0.528) BMI (0.193) Insulin (0.280)
<.001 .001 <.001
Independent variables
95th and above, corresponding to low, moderate, and high levels of obesity. Table I indicates that the 3 categories differ in terms of fasting insulin and glucose levels, with individuals below the 85th percentile reporting the lowest levels of fasting glucose and insulin. Females have significantly higher insulin levels than males in each BMI category, whereas for glucose the sex difference is not significant. With the American Diabetes Association criteria for diabetes and IFG, no cases of diabetes were detected among boys in any BMI category and no IFG below the 85th percentile (Table II). For girls the prevalence of diabetes is highest among those above the 95th percentile for BMI. Children with BMI exceeding the 85th percentile are at increased risk of being classified as having diabetes or IFG (OR 5.1, 95% CI 1.51, 17.0). In multiple linear regression analyses, BMI is a significant predictor of both fasting glucose and fasting insulin in both sexes, independent of age. However, when insulin is also included in the independent variable list, BMI is a predictor of fasting glucose in girls but not in boys (Table III).
DISCUSSION This study demonstrates a high prevalence of obesity among children and adolescents in a northern Canadian Aboriginal community. For the past 2 368
decades the early onset of obesity among Native Americans has been recognized in North America11-14 and in specific tribes across the continent such as the Pima,15 Cherokee,16 Mescalero Apache,17 Onondaga,18 Navajo,19 and urban Natives in Minneapolis.20 Evidence that this obesity is a recent phenomenon is available from a limited number of populations with historical data series.21-23 For Pima children it was possible to compare weight-for-height data with those collected by the anthropologist Hrdlicka in the early 1900s.15 Our study indicates that the obesity is associated with high fasting glucose and fasting insulin levels. Although the prevalence of obesity is comparable between boys and girls, new cases of diabetes are detected only among girls. The association between obesity and fasting glucose level is independent of fasting insulin only among girls and not among boys in this population. The prevalence of obesity is lower in the teen years. It should be noted that obesity in this study was defined relative to the U.S. National Center for Health Statistics “standard” population. In the teen years obesity in our population is still substantial, but relative to American teens, among whom obesity is also widespread, the excess is less than that observed among the younger age groups. This is a drawback of the use of a distribution-based definition of obesity rather than one based on physiologic (eg, body fat) or predictive criteria.
In addition to the short-term impact of an earlier onset of type 2 diabetes, there are important long–term health consequences of childhood obesity. Prospective cohort studies in the nonaboriginal population have shown that childhood obesity is a predictor of allcause mortality during adulthood, particularly mortality from cardiovascular diseases, which is independent of adult weight.24,25 This study highlights the critical importance of early identification and intervention of childhood obesity in populations at high risk for type 2 diabetes to avert the long-term health consequences. There is an urgent need to follow cohorts of at-risk children to investigate cofactors acting in concert with obesity that are responsible for the early onset of type 2 diabetes during childhood, a phenomenon increasingly being observed in many populations. We acknowledge the support by the community leaders and people of the St Theresa Point First Nation and Island Lake Tribal Council, without whom this study would not have been initiated, implemented, and completed.
REFERENCES 1. Gohdes D. Diabetes in North American Indians and Alaska Natives. In: National Diabetes Data Group. Diabetes in America. 2nd ed. Bethesda, MD: National Institutes of Health; 1995. p. 683-701 (NIH Publ No 951468). 2. In: Joe JR, Young RS, editors. Diabetes as a disease of civilization: the
YOUNG ET AL
THE JOURNAL OF PEDIATRICS VOLUME 136, NUMBER 3
3.
4.
5.
6.
7.
8.
9.
10.
impact of culture change on indigenous peoples. Berlin: Mouton de Gruyter; 1994. Young TK. Diabetes mellitus among Native Americans in Canada and the United States: an epidemiological overview. Am J Hum Biol 1993;5:399413. Dean HJ, Mundy RL, Moffatt M. Non-insulin-dependent diabetes mellitus in Indian children in Manitoba. Can Med Assoc J 1992;147:52-7. Dean HJ, Young TK, Flett B, WoodSteiman P. Screening for type 2 diabetes in aboriginal children in northern Canada. Lancet 1998;352:1523-4. Harris SB, Perkins BA, WhalenBrough E. Non-insulin-dependent diabetes mellitus among First Nation children: new entity among First Nation people of northwestern Ontario. Can Fam Physician 1996;42:869-76. Felber JP, Acheson KJ, Tappy L. From obesity to diabetes. Chichester: John Wiley and Sons; 1993. World Health Organization. Physical status: the use and interpretation of anthropometry. Report of a WHO Expert Committee. Geneva: WHO; 1995 (Tech Rep Ser 854). Must A, Dallal GE, Dietz WH. Reference data for obesity: 85th and 95th percentiles of body mass index (wt/ht2) and triceps skinfold thicknesses. Am J Clin Nutri 1991;53:839-46, 54:773. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the Expert Committee on the Diagnosis and Classification
11.
12.
13.
14.
15.
16.
17.
of Diabetes Mellitus. Diabetes Care 1997;20:1183-58. Young TK. Obesity among aboriginal peoples in North America: epidemiological patterns, risk factors and metabolic consequences. In: Angel A, Anderson C, Bouchard C, Lau D, Leiter L, Mandelson R, editors. Progress in obesity research: 7. London: John Libbey; 1996. p. 337-42. Harrison GG, Ritenbaugh CK. Obesity among North American Indians. In: Björntorp P, Brodoff BN, editors. Obesity. Philadelphia: JB Lippincott; 1992. p. 610-8. Broussard BA, Johnson A, Himes JH, Story M, Fichter R, Hauck F. Prevalence of obesity in American Indians and Alaska Natives. Am J Clin Nutr 1991;53:1535S-42. Jackson MY. Height, weight, and body mass index of American Indian schoolchildren, 1990-1991. J Am Diet Assoc 1993;93:1136-40. Knowler WC, Pettitt DJ, Saad MF, Charles MA, Nelson RG, Howard BV, et al. Obesity in the Pima Indians: its magnitude and relationship with diabetes. Am J Clin Nutr 1991;53:1543S-51. Story M, Tompkin RA, Bass MA, Wakefield LM. Anthropometric measurements and dietary intakes of Cherokee Indian teenagers in North Carolina. J Am Diet Assoc 1986;86:1555-60. Gallaher MM, Hauck FR, Yang-Oshida M, Serdula MK. Obesity among Mescalero preschool children: association with maternal obesity and birthweight. Am J Dis Child 1991;145: 1262-5.
18. Botash AS, Kavey RW, Emm N, Jones D. Cardiovascular risk factors in Native American children. NY State J Med 1992;92:378-81. 19. Gilbert TJ, Percy CA, Sugarman JR, Benson L, Percy C. Obesity among Navajo adolescents. Relationship to dietary intake and blood pressure. Am J Dis Child 1992;146:289-95. 20. Johnston FE, McKigney JI, Hopwood S, Smelker J. Physical growth and development of urban Native Americans: a study in urbanization and its implications for nutrition. Am J Clin Nutr 1978;31:1017-27. 21. Hauck FR, Gallaher MM, YangOshida M, Serdula MK. Trends in anthropometric measurements among Mescalero Apache Indian preschool children, 1968 through 1988. Am J Dis Child 1992;146:1194-8. 22. Malina RM. Ethnic variation in the prevalence of obesity in North American children and youth. Crit Rev Food Sci Nutr 1993;33:389-96. 23. Sugarman JR, White LL, Gilbert TJ. Evidence for a secular change in obesity, height, and weight among Navajo Indian school children. Am J Clin Nutr 1990;52:960-6. 24. Javier-Nieto F, Szklo M, Comstock GW. Childhood weight and growth rate as predictors of adult mortality. Am J Epidemiol 1992;136:201-3. 25. Must A, Jacques PF, Dallal GE, Bajema CJ, Dietz WH. Long term morbidity and mortality of overweight adolescents. A follow-up of the Harvard Growth Study of 1922 to 1935. N Engl J Med 1992;327:1350-5.
369
Childhood obesity in a population at high risk for
type 2 diabetes T. Kue Young, MD, PhD, Heather J. Dean, MD, Bertha Flett, and Pauline Wood-Steiman
Objectives: To determine the prevalence of obesity and investigate its association with fasting glucose and insulin among children and adolescents in a population at high risk for type 2 diabetes. Design: A cross-sectional screening survey involving anthropometry and fasting serum levels of glucose and insulin. Setting: A remote aboriginal (Ojibwa-Cree) community in northern Manitoba, Canada. Participants: All children aged 4 to 19 years in the community were invited to participate, with a response rate of 82% (n = 719). Main outcome measures: Obesity is defined as body mass index exceeding the 85th percentile of the National Center for Health Statistics reference data. The diagnosis of diabetes and impaired fasting glucose is based on the new criteria of the American Diabetes Association. Results: There is a high prevalence of obesity, with 64% (female) and 60% (male) exceeding the 85th percentile and 40% (female) and 34% (male) exceeding the 95th percentile. Body mass index is a significant predictor of both glucose and insulin in both sexes, independent of age. Obese children are at increased risk of being classified as having diabetes or impaired fasting glucose (odds ratio 5.1, 95% CI 1.51, 17.0). Conclusions: The early onset of type 2 diabetes in childhood is increasingly observed in many populations. Childhood obesity is a strong risk factor. Early detection and intervention directed at obesity are potential strategies to avert the long-term consequences of type 2 diabetes. (J Pediatr 2000;136:365-9)
The epidemic of type 2 diabetes among many Aboriginal/Native American populations in Canada and the United States during the past several decades
has been recognized for some time.1-3 A more recent development is the rising prevalence among children, associated with a decreasing age of onset. In
From the Department of Community Health Sciences and the Department of Pediatrics and Child Health, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba, Canada; and the Island Lake Tribal Council, Winnipeg, Manitoba, Canada. Funding for this study was provided by the National Health Research Development Program of Health Canada (Grant No. 6607-1665-ND). Dr Young is a Senior Scientist of the Medical Research Council of Canada. Submitted for publication Apr 29, 1999; revision received Aug 30, 1999; accepted Oct 1, 1999. Reprint requests: T. Kue Young, Department of Community Health Sciences, University of Manitoba, 750 Bannatyne Ave, Winnipeg, MB, Canada, R3E 0W3. Copyright © 2000 by Mosby, Inc. 0022-3476/2000/$12.00 + 0 9/21/103504 doi:10.1067/mpd.2000.103504
the past decade clinicians have diagnosed this disease in Aboriginal children as young as 7 years of age.4-6 This is a serious concern in view of the increased duration of diabetes and the likelihood of microvascular and macrovascular complications appearing during young adulthood. The cause of the epidemic is not completely understood. Based on studies in adults, obesity has been demonstrated to be one of the most important risk factors for type 2 diabetes.7 It is likely that the emerging pattern and burden of childhood type 2 diabetes is associated with an increasing prevalence of obesity among Aboriginal children. BMI IFG
Body mass index Impaired fasting glucose
We report on a study to determine the prevalence of obesity and its association with metabolic markers of glucose metabolism in a Canadian Aboriginal community known to be at high risk for type 2 diabetes in the adult population. This study was initiated at the request of the community of St Theresa Point in northern Manitoba, whose members have become concerned with the high rate of diabetes in the population and its implications for the future health of their children. A preliminary report on the clinical features of new cases of diabetes detected by screening has been published.5
METHODS Study Setting St Theresa Point First Nation, with a population of 2430, is located 467 km 365
YOUNG ET AL
THE JOURNAL OF PEDIATRICS MARCH 2000 increasing dependence on government subsidies. Storebought food has become predominant in the people’s diet, and the level of physical activity has declined.
Data Collection
Fig 1. Prevalence of obesity by age-sex group based on percentile of National Center for Health Statistics reference population data.
Fig 2. Mean fasting glucose and insulin by age group and sex. by air northeast of Winnipeg, Manitoba, Canada. It is one of 4 communities that make up the Island Lake Tribal Council. In 1875 the Ojibwa-Cree people of the Island Lake area signed Treaty No 5 with the Canadian government. The traditional lifestyle of 366
hunting, trapping, and fishing remained relatively intact well into the 1960s. However, in the decades since then, there has been rapid social change characterized by reduced dependence on the land for subsistence, involvement in the wage economy, and
A project office was established in the school, which had an enrollment of 873 children from nursery school to grade 11 in the 1995 to 1996 and 1996 to 1997 academic years. The children were instructed to arrive at school at 8:00 AM before breakfast on their scheduled day, usually in family groupings. A venous blood sample was taken followed by a nutritious breakfast provided by the project nurse. Standing height was measured with a wallmounted tape, and weight was measured with a standard balance scale. Samples were taken from approximately 5 to 10 children 4 mornings per week between April 1996 and July 1997 except during school holidays. Fasting blood samples were taken to the local health center for centrifuge and storage at 4°C. At the end of each month, samples were packed in dry ice and shipped to the Health Sciences Center in Winnipeg, where they were batched for serum glucose and insulin assays. Children with high fasting glucose levels were referred to the health center for further clinical evaluation by the visiting physician. All results were explained to the parents in person or in writing by the project nurse. Serum glucose was measured with a Hitachi 705 Analyzer and insulin by radioimmunoassay with the Pharmacia kit at the Clinical Chemistry Laboratory of the Health Sciences Center, Winnipeg. This study was approved by the Committee on the Use of Human Subjects in Research of the Faculty of Medicine, University of Manitoba, the Island Lake Diabetes Committee, and the St. Theresa Point First Nation Council. Informed consent was obtained from the parents of all participating children.
YOUNG ET AL
THE JOURNAL OF PEDIATRICS VOLUME 136, NUMBER 3 Table I. Association of fasting glucose and insulin levels with categories of body mass index
Fasting glucose (mmol/L) NCHS percentile Male <85th 85th-94th 95th+ Female <85th 85th-94th 95th+
Fasting insulin (pmol/L)
No.
Mean
95% CI
Mean
95% CI
143 96 122
4.96 5.06 5.10
4.89, 5.02 4.96, 5.17 5.01, 5.19
60.13 66.17 93.61
54.23, 66.03 58.35, 73.99 81.16, 106.06
126 86 140
4.83 4.92 5.06
4.73, 4.93 4.79, 5.06 4.92, 5.20
80.08 100.69 119.11
67.59, 92.57 74.48, 126.90 103.84, 134.38
Results of analysis of variance comparing differences in fasting glucose/insulin across the 3 BMI categories: males, glucose (F = 3.244, P = .04); males, insulin (F = 15.5, P < .001); females, glucose (F = 3.6, P = .028); females, insulin (F = 5.64, P = .004). To convert from SI units to conventional units for glucose, divide by 0.05551; for insulin, divide by 6.0. NCHS, National Center for Health Statistics.
Table II. Prevalence of diabetes and impaired fasting glucose by categories of body mass index
Diabetes NCHS percentile Male <85th 85th-94th 95th+ Female <85th 85th-94th 95th+
Impaired fasting glucose
No.
No cases
Prevalence (%)
No cases
Prevalence (%)
143 96 122
0 0 0
0 0 0
0 6 5
0 6.3 4.1
126 86 140
1 1 6
0.8 1.2 4.3
2 3 3
1.6 3.5 2.1
Diagnostic Criteria A recent World Health Organization Expert Committee on anthropometry recommended the body mass index (weight in kilograms per height in meters squared) as the indicator of choice for overweight among adolescents, because it has been validated by studies on total body fat.8 The recommended reference data are those of the U.S. all-race population derived from the U.S. National Health and Nutrition Examination Survey of 1971 to 1974 published by the National Center for Health Statistics. Individuals above the 85th percentile are considered to be overweight.9 For diabetes the new diagnostic criteria approved by the American Diabetes Association10 were used, which define diabetes in individuals with a fasting
serum glucose ≥7 mmol/L (126 mg/dL). There is also a category of “impaired fasting glucose,” which encompasses individuals with values <7 mmol/L (126 mg/dL) but ≥6.1 mmol/L (110 mg/dL).
RESULTS A total of 719 children (354 girls and 365 boys) ranging in age from 4 to 19 years were screened, a response rate of 82%. All children were attending school in the community at the time of study. Two children did not undergo venipuncture, and they were excluded from further analyses. The prevalence of obesity among children in Island Lake is considerably higher than that in the National Center
for Health Statistics reference population, because 64% of girls and 60% of boys exceed the 85th percentile of the age-sex all-race reference values for BMI, and 40% of girls and 34% of boys exceed the 95th percentile (Fig 1). The mean fasting serum glucose and insulin levels by age and sex are shown in Fig 2. Among normoglycemic individuals girls tend to have higher fasting insulin levels than boys, whereas for glucose, age-sex differences are not pronounced. BMI is correlated with both fasting glucose and insulin, with partial correlation coefficients of 0.224 and 0.406, respectively, controlling for age and sex. Study participants were divided into 3 categories on the basis of their BMI percentiles: <85th, 85th to 94th, and 367
YOUNG ET AL
THE JOURNAL OF PEDIATRICS MARCH 2000
Table III. Association of BMI and fasting glucose levels: multiple linear regression analyses
Model R2
Standardized coefficient (β)
P value
(1) Male only, dependent variable: fasting glucose BMI, age BMI, age, insulin
0.022 0.092
BMI (2.778) BMI (–0.013) Insulin (0.310)
.006 .838 <.001
(2) Female only, dependent variable: fasting glucose BMI, age BMI, age, insulin
0.080 0.145
BMI (0.528) BMI (0.193) Insulin (0.280)
<.001 .001 <.001
Independent variables
95th and above, corresponding to low, moderate, and high levels of obesity. Table I indicates that the 3 categories differ in terms of fasting insulin and glucose levels, with individuals below the 85th percentile reporting the lowest levels of fasting glucose and insulin. Females have significantly higher insulin levels than males in each BMI category, whereas for glucose the sex difference is not significant. With the American Diabetes Association criteria for diabetes and IFG, no cases of diabetes were detected among boys in any BMI category and no IFG below the 85th percentile (Table II). For girls the prevalence of diabetes is highest among those above the 95th percentile for BMI. Children with BMI exceeding the 85th percentile are at increased risk of being classified as having diabetes or IFG (OR 5.1, 95% CI 1.51, 17.0). In multiple linear regression analyses, BMI is a significant predictor of both fasting glucose and fasting insulin in both sexes, independent of age. However, when insulin is also included in the independent variable list, BMI is a predictor of fasting glucose in girls but not in boys (Table III).
DISCUSSION This study demonstrates a high prevalence of obesity among children and adolescents in a northern Canadian Aboriginal community. For the past 2 368
decades the early onset of obesity among Native Americans has been recognized in North America11-14 and in specific tribes across the continent such as the Pima,15 Cherokee,16 Mescalero Apache,17 Onondaga,18 Navajo,19 and urban Natives in Minneapolis.20 Evidence that this obesity is a recent phenomenon is available from a limited number of populations with historical data series.21-23 For Pima children it was possible to compare weight-for-height data with those collected by the anthropologist Hrdlicka in the early 1900s.15 Our study indicates that the obesity is associated with high fasting glucose and fasting insulin levels. Although the prevalence of obesity is comparable between boys and girls, new cases of diabetes are detected only among girls. The association between obesity and fasting glucose level is independent of fasting insulin only among girls and not among boys in this population. The prevalence of obesity is lower in the teen years. It should be noted that obesity in this study was defined relative to the U.S. National Center for Health Statistics “standard” population. In the teen years obesity in our population is still substantial, but relative to American teens, among whom obesity is also widespread, the excess is less than that observed among the younger age groups. This is a drawback of the use of a distribution-based definition of obesity rather than one based on physiologic (eg, body fat) or predictive criteria.
In addition to the short-term impact of an earlier onset of type 2 diabetes, there are important long–term health consequences of childhood obesity. Prospective cohort studies in the nonaboriginal population have shown that childhood obesity is a predictor of allcause mortality during adulthood, particularly mortality from cardiovascular diseases, which is independent of adult weight.24,25 This study highlights the critical importance of early identification and intervention of childhood obesity in populations at high risk for type 2 diabetes to avert the long-term health consequences. There is an urgent need to follow cohorts of at-risk children to investigate cofactors acting in concert with obesity that are responsible for the early onset of type 2 diabetes during childhood, a phenomenon increasingly being observed in many populations. We acknowledge the support by the community leaders and people of the St Theresa Point First Nation and Island Lake Tribal Council, without whom this study would not have been initiated, implemented, and completed.
REFERENCES 1. Gohdes D. Diabetes in North American Indians and Alaska Natives. In: National Diabetes Data Group. Diabetes in America. 2nd ed. Bethesda, MD: National Institutes of Health; 1995. p. 683-701 (NIH Publ No 951468). 2. In: Joe JR, Young RS, editors. Diabetes as a disease of civilization: the
YOUNG ET AL
THE JOURNAL OF PEDIATRICS VOLUME 136, NUMBER 3
3.
4.
5.
6.
7.
8.
9.
10.
impact of culture change on indigenous peoples. Berlin: Mouton de Gruyter; 1994. Young TK. Diabetes mellitus among Native Americans in Canada and the United States: an epidemiological overview. Am J Hum Biol 1993;5:399413. Dean HJ, Mundy RL, Moffatt M. Non-insulin-dependent diabetes mellitus in Indian children in Manitoba. Can Med Assoc J 1992;147:52-7. Dean HJ, Young TK, Flett B, WoodSteiman P. Screening for type 2 diabetes in aboriginal children in northern Canada. Lancet 1998;352:1523-4. Harris SB, Perkins BA, WhalenBrough E. Non-insulin-dependent diabetes mellitus among First Nation children: new entity among First Nation people of northwestern Ontario. Can Fam Physician 1996;42:869-76. Felber JP, Acheson KJ, Tappy L. From obesity to diabetes. Chichester: John Wiley and Sons; 1993. World Health Organization. Physical status: the use and interpretation of anthropometry. Report of a WHO Expert Committee. Geneva: WHO; 1995 (Tech Rep Ser 854). Must A, Dallal GE, Dietz WH. Reference data for obesity: 85th and 95th percentiles of body mass index (wt/ht2) and triceps skinfold thicknesses. Am J Clin Nutri 1991;53:839-46, 54:773. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the Expert Committee on the Diagnosis and Classification
11.
12.
13.
14.
15.
16.
17.
of Diabetes Mellitus. Diabetes Care 1997;20:1183-58. Young TK. Obesity among aboriginal peoples in North America: epidemiological patterns, risk factors and metabolic consequences. In: Angel A, Anderson C, Bouchard C, Lau D, Leiter L, Mandelson R, editors. Progress in obesity research: 7. London: John Libbey; 1996. p. 337-42. Harrison GG, Ritenbaugh CK. Obesity among North American Indians. In: Björntorp P, Brodoff BN, editors. Obesity. Philadelphia: JB Lippincott; 1992. p. 610-8. Broussard BA, Johnson A, Himes JH, Story M, Fichter R, Hauck F. Prevalence of obesity in American Indians and Alaska Natives. Am J Clin Nutr 1991;53:1535S-42. Jackson MY. Height, weight, and body mass index of American Indian schoolchildren, 1990-1991. J Am Diet Assoc 1993;93:1136-40. Knowler WC, Pettitt DJ, Saad MF, Charles MA, Nelson RG, Howard BV, et al. Obesity in the Pima Indians: its magnitude and relationship with diabetes. Am J Clin Nutr 1991;53:1543S-51. Story M, Tompkin RA, Bass MA, Wakefield LM. Anthropometric measurements and dietary intakes of Cherokee Indian teenagers in North Carolina. J Am Diet Assoc 1986;86:1555-60. Gallaher MM, Hauck FR, Yang-Oshida M, Serdula MK. Obesity among Mescalero preschool children: association with maternal obesity and birthweight. Am J Dis Child 1991;145: 1262-5.
18. Botash AS, Kavey RW, Emm N, Jones D. Cardiovascular risk factors in Native American children. NY State J Med 1992;92:378-81. 19. Gilbert TJ, Percy CA, Sugarman JR, Benson L, Percy C. Obesity among Navajo adolescents. Relationship to dietary intake and blood pressure. Am J Dis Child 1992;146:289-95. 20. Johnston FE, McKigney JI, Hopwood S, Smelker J. Physical growth and development of urban Native Americans: a study in urbanization and its implications for nutrition. Am J Clin Nutr 1978;31:1017-27. 21. Hauck FR, Gallaher MM, YangOshida M, Serdula MK. Trends in anthropometric measurements among Mescalero Apache Indian preschool children, 1968 through 1988. Am J Dis Child 1992;146:1194-8. 22. Malina RM. Ethnic variation in the prevalence of obesity in North American children and youth. Crit Rev Food Sci Nutr 1993;33:389-96. 23. Sugarman JR, White LL, Gilbert TJ. Evidence for a secular change in obesity, height, and weight among Navajo Indian school children. Am J Clin Nutr 1990;52:960-6. 24. Javier-Nieto F, Szklo M, Comstock GW. Childhood weight and growth rate as predictors of adult mortality. Am J Epidemiol 1992;136:201-3. 25. Must A, Jacques PF, Dallal GE, Bajema CJ, Dietz WH. Long term morbidity and mortality of overweight adolescents. A follow-up of the Harvard Growth Study of 1922 to 1935. N Engl J Med 1992;327:1350-5.
369