Conflict of interest: Dr. Hoover-Fong is a paid consultant to BioMarin. This arrangement has been reviewed and approved by the Johns Hopkins University in accordance with its conflict of interest policies.
Body mass index (BMI): The case for condition-specific cut-offs for overweight and obesity in skeletal dysplasias
Article first published online: 24 JUN 2013
Copyright © 2013 Wiley Periodicals, Inc.
American Journal of Medical Genetics Part A
Volume 161, Issue 8, pages 2110–2112, August 2013
How to Cite
2013. Body mass index (BMI): The case for condition-specific cut-offs for overweight and obesity in skeletal dysplasias. Am J Med Genet Part A. 161A:2110–2112., , , .
- Issue published online: 24 JUL 2013
- Article first published online: 24 JUN 2013
- Manuscript Accepted: 23 FEB 2013
- Manuscript Received: 28 AUG 2012
TO THE EDITOR:
It has come to our attention that medical insurance companies have cited excessive body mass index (BMI, kg/m2) as rationale to deny coverage to short stature skeletal dysplasia patients. Though there is strong evidence to link BMI to mortality in average stature adults, the application of the current BMI guidelines to adults with dwarfism is not accurate. Three years ago we addressed a similar issue in children and adolescents with dwarfism who were assessed at school and by healthcare providers as morbidly obese when compared to BMI-for-age reference data established in average stature children [Hoover-Fong et al., 2008]. In children with dwarfism, as in adults, the height component of BMI is drastically reduced compared to weight, rendering the ratio extremely high and therefore uninterpretable against conventional cutoffs. Until BMI is correlated with body composition and related to long-term health outcomes in short stature skeletal dysplasia patients, it is inaccurate to use it as a surrogate of body fat or a predictor of health outcomes.
Weight and height have historically been used as indicators of health. At the beginning of the twentieth century, the life insurance industry began producing tables of “desirable” weights for men and women by height and frame size. In the early 1970s, at the outset of the U.S. obesity epidemic, the concept of the BMI was introduced as a means to use weight and height to capture information on body build and fatness with a single statistic [Keys et al., 1972]. Its use began with the demonstration of two key criteria: First, BMI is strongly correlated with body fat. Although weight captures both fat and non-fat mass (e.g., muscle, fluids, bone), fat mass is most variable and therefore the major contributor to differences in BMI among people. Second, BMI is statistically independent of height in adults and thus common cutoffs can be applied, generally speaking, to adults regardless of their height. (This is in contrast to expressing weight per unit height alone, where taller people have a greater expected body weight.)
The World Health Organization advocates BMI-cutoffs to characterize nutritional status (BMI < 18.5, underweight; 18.5–24.9, normal; 25.0–29.9, overweight; ≥30.0, obese) [WHO, 2006]. Studies have also associated increasing BMI with worse mortality outcome and chronic disease risk [NTFPTO, 2000]. Therefore, it has become increasingly common to characterize an individual according to BMI and make weight loss recommendations based on his/her value. BMI also has been adopted for use in children, although its interpretation is more complicated because BMI changes with age as children grow [Freedman and Sherry, 2009]. Nonetheless, BMI-for-age percentile charts published in 2000 by the Center for Disease Control [CDC, 2000] have been generally adopted in the United States to monitor overweight and obesity during childhood. Although limitations of using BMI as a proxy for body fatness are somewhat intuitive (i.e., individuals with lower body fat but abundant musculature can have a high BMI), the role of body proportion in BMI interpretation is underappreciated [Norgan, 1994]. For example, BMI overestimates body fat among average stature groups with longer trunk relative to extremity length as in some Asian and Inuit peoples [Charbonneau-Roberts et al., 2005]. Conversely, in some aboriginal populations with greater extremity to trunk length, BMI underestimates body fat [Norgan, 1994].
People with skeletal dysplasias typically have body types that are even more disproportionate relative to the extremes observed in the average stature population. Among the myriad skeletal dyplasias currently recognized [Warman et al., 2011], there are a variety of body types, coarsely divided into those with short stature and a relatively long trunk versus those with a short trunk. Achondroplasia is the most common skeletal dysplasia, occurring in over 1 in 20,000 births. It is characterized by rhizomelia and a relatively long trunk, such that the upper:lower segment remains at ∼1.5 in achondroplasia from late childhood through adult years [Hoover-Fong et al., 2008] versus a decline to ∼1 in average stature by 6–8 years of age. The type 2 collagen conditions causing short stature (e.g., spondyloepiphyseal dysplasia congenital, Kneist dysplasia) are likely the most common short trunk dysplasias, but are far less common than achondroplasia with a birth prevalence of ∼1 in 100,000 births. Upper:lower segment ratios have not been studied in populations with type 2 collagen conditions; thus, the influence of body proportions on BMI in this group cannot be compared with those with achondroplasia or average stature. However the relative influence on BMI of a short versus long trunk in short stature individuals pales in comparison to the dramatic deficit in height overall as compared to weight; this is the most likely cause of the distorted BMI values of short stature skeletal dysplasia patients.
Using achondroplasia as an example, Table I illustrates the fallibility of using average-stature BMI cutoffs to categorize nutritional status in short stature skeletal dysplasia women. Note the short stature values are based on actual data from a recent pilot study with minor simplifications to incorporate “normal” and “overweight” body fat percentages in theoretical average stature women. While the average stature women (1 and 2) are the same height (1.62 m) with the same lean mass (48 kg), woman 1 has a normal percentage body fat (25%) while that of woman 2 is considered unhealthy (40%). BMI of woman 1 is in the “normal” category (24.2 kg/m2), while that of woman 2 is correctly categorized as “obese” (30.3 kg/m2). The short stature skeletal dysplasia women (3 and 4) have the same percent body fat as the average stature women, 25% and 40%, respectively. Although women 3 and 4 weigh less than the average stature women, they are also considerably shorter (1.16 m). Hence, the BMI of woman 3 is 29.3 kg/m2, 5 units higher than her average stature counterpart with comparable percent body fat (25%). Similarly, woman 4 has a BMI of 36.6 kg/m2, >6 units greater than woman 2 with the same percent body fat (40%).
|Average stature women||1||2|
|Height, m (ft, in)||1.62 (5′4′′)||1.62 (5′4′′)|
|Weight, kg (lb)||64 (141)||80 (176)|
|% Body fat||25%||40%|
|Short stature women||3||4|
|Height, m (ft, in)||1.16 (3′10′′)||1.16 (3′10′′)|
|Weight, kg (lb)||40 (88)||50 (110)|
|% Body fat||25%||40%|
|Assessment||Same % body fat as normal average stature but BMI 5 units higher, considered “obese”||Same % body fat as obese average stature, but BMI 6+ units higher, considered “obese”|
The BMI is certain to be associated with body fat in people with skeletal dysplasias, and obesity is a concern in those with skeletal dysplasias as in the general population—perhaps even more so, given that there is a smaller body frame on which to accumulate excess weight. But more information is needed to properly interpret calculated BMI values in skeletal dysplasias, or to determine whether other measures, such as skinfolds, body circumferences, or more direct measures of body composition, are more relevant. Moreover, the link between the degree of body fatness and morbidity and mortality has been entirely unstudied in this population. Without such an evidence base, average stature BMI recommendations must not be applied to short stature individuals as an indicator of fatness, disease, or mortality risk. Studies to address these issues are underway.
We would like to acknowledge the Medical Advisory Board of the Little People of America for fruitful discussion about this topic as we prepared this manuscript.
- Centers for Disease Control. 2000. CDC Clinical Growth Charts. Last accessed September 28, 2011. Available at: http://www.cdc.gov/growthcharts/
- 2005. Body mass index may overestimate the prevalence of overweight and obesity among the Inuit. Int J Circumpolar Health 64:163–169. , , , .
- 2009. The validity of BMI as an indicator of body fatness and risk among children. Pediatrics 124:S23–S34. , .
- 2008. Age-appropriate body mass index in children with achondroplasia: Interpretation in relation to indexes of height. Am J Clin Nutr 88:364–371. , , , , .
- 1972. Indices of relative weight and obesity. J Chronic Dis 25:329–343. , , , , .
- 1994. Relative sitting height and the interpretation of the body mass index A. Ann Hum Biol 21:79–82. .
- 1994. B Interpretation of low body mass indices: Australian aborigines. Am J Phys Anthropol 94:229–237. .
- National Task Force on the Prevention and Treatment of Obesity. 2000. Overweight obesity, and health risk. Arch Intern Med 160:898–904.
- 2011. Nosology and classification of genetic skeletal disorders: 2010 revision. Am J Med Genet Part A 155A:943–968. , , , , , , , , , , , , , , , , , , , .
- World Health Organization. 2006. Last accessed September 28, 2011. Available at: http://apps.who.int/bmi/