点击显示 收起
1 From the Department of Nutrition (JSF) and the Rowe Program in Genetics (CHW), Department of Pediatrics and Section of Neurobiology, Physiology, and Behavior, University of California, Davis, Davis, CA
2 Address reprint requests to CH Warden, Rowe Program in Genetics, University of California, Davis, Davis, CA 95616. E-mail: chwarden{at}ucdavis.edu.
See corresponding article on page 646.
An epidemic of overweight and obesity is occurring in both industrial and developing societies. According to the National Health and Nutrition Examination Survey (NHANES), the prevalence of obesity doubled in the adult population, and the prevalence of overweight tripled in children and adolescents between 1982 and 2002 (Internet: www.cdc.gov/nchs/nhanes.htm). The most recent prevalence estimates from weight and height data collected by NHANES in 2003–2004 indicate that the prevalence of overweight in children and adolescents continues to increase and that 18.2% of boys and 16.0% of girls between the ages of 2 and 19 y are now overweight (1). Prevalence estimates differ by ethnicity, eg, 22.0% of Mexican American boys and 16.2% of Mexican American girls are overweight (1).
These recent rapid increases in obesity are due to environmental changes—including readily available palatable and energy-dense foods and a sedentary lifestyle (2). Comorbidities associated with obesity include the metabolic syndrome (hypertension, hyperlipidemia, and insulin resistance), type 2 diabetes, certain cancers, orthopedic problems, and sleep apnea. Because childhood overweight tracks with obesity in adulthood (3), the dramatic increase in the frequency of overweight in children and adolescents bodes ill for the health status of our population in future years. Thus, there is considerable interest in understanding the genetic and environmental contributions to obesity to aid in the development of more effective interventions.
As reviewed by Comuzzie and Allison (4), 40–70% of the within-population variation in obesity is due to genetic variation. The proportion of the total variance in a trait that is due to genetic factors is defined as the heritability coefficient (h2). Population or family studies tend to have lower and twin studies to have higher heritability values for body mass index (BMI), a common surrogate measure for obesity. Data from adult nondiabetic Mexican American participants in the San Antonio Diabetes Study showed the additive genetic heritability coefficient for BMI in that population to be 0.65 (5). Although there are few estimates of the heritability of overweight or obesity in children and adolescents, genetic influences contributed 0.75–0.80 of the variation in percentage body fat in a pediatric twin study (6).
New heritability estimates for childhood obesity–related phenotypes and risk factors for metabolic diseases are available in the Hispanic population from the Viva la Familia Study, as reported by Butte et al (7) in this issue of the Journal. In this study, the heritability coefficients (adjusted for sex, age, and environmental covariates) of anthropometric indexes range from 0.24 to 0.75, whereas the heritability coefficients for body composition are somewhat lower. For example, the h2 for fat mass is 0.18. Energy and macronutrient intakes have heritability coefficients ranging from 0.54 to 0.69, and indicators of physical activity have heritability coefficients ranging from 0.32 to 0.60. Risk factors for metabolic diseases have heritability coefficients of 0.25 to 0.73, the highest being that for serum total cholesterol concentrations. The report by Butte et al is the first to provide heritability estimates for a range of obesity phenotypes as well as for dietary intake and metabolic disease risk factors in children and adolescents.
Butte et al used variance component analysis to partition the total phenotypic variance into additive genetic and environmental components, for which the environmental component includes any genetic factor that is not additive (7). However, nonadditive genetic effects and interactions contribute considerably to the phenotypic variation of obesity (8). Segal and Allison (9) argue that the higher heritability estimates obtained in twin studies reflect the nonadditive nature of the genetic effects on obesity. Thus, it is likely that the heritability coefficients provided by Butte et al are conservative estimates.
Many of the heritability coefficients from the Viva la Familia Study are large, which suggests the presence of single major genes with large effects (7). In fact, genome scans of adult Mexican American populations identified major loci for obesity (5, 10), insulin concentrations and insulin resistance (5), leptin concentrations (10), and plasma triacylglycerol concentrations (11).
A powerful tool in the search for genetic contributions to obesity involves linkages of obese phenotypes to specific chromosomal regions from genome scans. Three genome-wide scans for childhood and adolescent obesity based on BMI in white populations were recently published (12-14). In 2 of these studies (13, 14), the families were studied on the basis of an overweight-child proband; in the third study (12), the families were unselected for obesity. One study (14) found only suggestive lod scores. The other 2 studies identified significant lods on chromosomes 12q (12) and 6q (13) and several other suggestive loci. These various loci confirm loci identified in genome scans of adult obesity, which suggests that the genetic bases of childhood and adult obesity may be similar (14).
The similarities between adults and children in the genetic contributions to BMI are supported by the recent identification, through a genome-wide scan, of a common genetic variant in participants of the Framingham Heart Study that was followed by replication of the association of the variant with BMI in 4 other populations of various ethnicities and ages (15). The study by Herbert et al (15) highlights the power of genome scans in identifying genetic variants that can then be verified in other populations.
The Viva la Familia Study was designed to genetically map childhood obesity and its comorbidities in the Hispanic population (7). A real strength of the design was the measurement of multiple phenotypes. When these mapping data become available, our knowledge of the genetic basis of both childhood and adult obesity should expand considerably.
REFERENCES