|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 A COMPARISION OF BODY ADIPOSITY INDEX AND BODY MASS INDEX TO BODY FAT PERCENTAGE IN YOUNG ADULT NON ATHLETES AND ATHLETES By BLAKE BARTHOLOMEW A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUI REMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2013
2 2013 Blake Bartholomew
3 To everyone that has helped and guided me along this long winding road that has been my college career
4 ACKNOWLEDGMENTS I thank my parents for all the love and support they have given me. Thank you Dr. Shelnutt, you have trea t have made it this far without your guidance. Thank you Sweet pea for being by my side for these past two years.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST O F ABBREVIATIONS ................................ ................................ ............................. 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 LITERATURE REVIEW AND RESEARCH RATIONALE ................................ ........... 12 Trends in Obesity ................................ ................................ ................................ ... 12 Health Consequences of Obesity ................................ ................................ ........... 12 Measuring Adiposity ................................ ................................ ............................... 13 The Gold Standards ................................ ................................ ......................... 14 Hydrostatic weighing ................................ ................................ .................. 14 Dual energy X ray absorptiometry (DEXA) ................................ ................ 15 Air displacement plethysmography ................................ ............................ 16 Imaging techniques ................................ ................................ .................... 17 Other Measures of Adip osity ................................ ................................ ............ 18 Waist circumference ................................ ................................ .................. 18 Waist to hip ratio ................................ ................................ ........................ 20 Skinfold meas urement ................................ ................................ ............... 21 Sagittal abdominal diameter ................................ ................................ ....... 22 Body mass index ................................ ................................ ........................ 23 Body a diposity index ................................ ................................ .................. 29 2 METHODS AND PROCEDURE ................................ ................................ ................. 34 Study Design and Subjects ................................ ................................ ..................... 34 Physical Assessments ................................ ................................ ............................ 34 Statistical Analysis ................................ ................................ ................................ .. 36 3 RESULTS ................................ ................................ ................................ ................... 38 Subject Characteristics ................................ ................................ ........................... 38 ................................ ................................ ............................ 42 All Subjects ................................ ................................ ................................ ....... 42 Non athletes ................................ ................................ ................................ ..... 45 Athletes ................................ ................................ ................................ ............ 49
6 4 DISCUSSION ................................ ................................ ................................ ............. 53 APPENDIX A INFORMED CONSENT ................................ ................................ ............................. 57 B DATA COLLECTION SHEETS ................................ ................................ .................. 63 LIST OF REFERENCES ................................ ................................ ............................... 67 BIOGRA PHICAL SKETCH ................................ ................................ ............................ 73
7 LIST OF TABLES Table page 1 1 Weight status classification within BMI ranges ................................ ................... 24 1 2 Body fat percentages within BMI classifications* ................................ ................ 25 1 3 Mean percentage BF% according to BMI categories among US adults from NHANES 1999 2004 ................................ ................................ .......................... 26 3 1 Overall subject characteristics. All values are mean SD ................................ .. 38 3 2 Nonathlete subject characteristics. All values are mean SD ............................ 39 3 3 Athlete subject characteristics. All values are mean SD ................................ ... 40 3 4 subjects ................................ ................................ ................................ .............. 42 3 5 nonathletes ................................ ................................ ................................ ......... 46 3 6 BF% and, BMI to BF% in athletes ................................ ................................ ................................ ............... 49
8 LIST OF FIGURES Figure page 2 1 Placement of tape during measurement of hip circumference ........................... 36 3 1 Correlation of BMI and BAI to BF% in all male subjects ................................ .... 43 3 2 Correlation of BMI and BAI to BF% in all female subjects ................................ 43 3 3 Correlation of BMI and BAI to BF% in all subjects ................................ ............. 44 3 4 Bland Altman limits of agreement plot between BAI and BF% in all subjects. ... 45 3 5 Correlation of BMI and BAI to BF% in male non athletes ................................ .. 46 3 6 Correlation of BMI and BAI to BF% in female non athletes ............................... 47 3 7 Correlation of BMI and BAI to BF% in nonathletes ................................ ............ 48 3 8 Bland Altman limits of agreement plot between BAI and BF% in non athl etes .. 48 3 9 Correlation of BMI and BAI to BF% in male athletes ................................ ......... 50 3 10 Correlation of BMI and BAI to BF% in female athletes ................................ ...... 50 3 11 Correlation of BMI and BAI to BF% in all athletes ................................ ............. 51 3 12 Bland Altman limits of agreement plot between BAI and BF% in a thletes ......... 52
9 LIST OF ABBREVIATIONS ADP Air displacement plethysmography BAI Body adiposity index BIA Bioelectrical impedance analysis BF% Body fat percentage BMI B ody mass i ndex CT Computed tomography DEXA Dual energy x r ay absorptiometry IRB University of Florida Institutional Review Board 02 MONICA Monitoring Trends and Determinants in Cardiovascular Disease Augsburg study MRI Magnetic resonance imaging NHANES National H ealth and Nutrition Examination Survey NHANES III T he third National Health and Nutrition Examination Survey NHANES 99 04 National Health and Nutrition Examination Survey from 1999 to 2004 NHLBI National Heart, Lung, and Blood Institute SAD S agittal abdominal diameter TARA Triglyceride and Cardiova scular Risk in African American study UAA University Athletic Association WC Waist circumference WHO World Health Organization WHR Waist to hip ratio
10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial F ulfillment of the Requi rements for the Degree of Master of Science A COMPARISION OF BODY ADIPOSITY INDEX AND BODY MASS INDEX TO BODY FAT PERCENTAGE IN YOUNG ADULT NON ATHLETES AND ATHLETES By Blake Bartholomew May 2013 Chair: Karla Shelnutt Major: Food Science and Human Nutrition In this study we wanted to determine whether body adiposity index (BAI) more strongly correlated t o body fat percentage (BF%) than body mass index (BMI) in young adult athletes and non athletes. University of Florida athl etes (N=53; 29 men and 24 women) and non athletes (N=195; 64 men and 131 women) aged 18 24 participated in physical assessments (height, weight, hip circumference and BF% measured by air displacement plethysmography) in a clinical lab. The relationship bet ween BF% and BAI ding 95% confidence intervals. Bland Altman limits of agreement plots were used to visually compare the existence of any differences between BAI and BF% For all participants BAI was more strongly correlated to BF% than BMI was correlated to BF% [r=0.73; (0.67, 0.79) vs. r=0.31; (0.19, 0.41)]. When separated by sex BAI and BMI correlated similarly to BF% with no significant difference between the two measures. When separated by athletic status, BAI correlated more strongly with BF% than BMI in non athletes [r=0.76; (0.70, 0.81) vs. 0.38; (0.25, 0.49) ], but not in athletes [BAI BF% r=0.41; (0.16, 0.61) vs. BMI BF% r=0.29; (0.0 2, 0.52)]. When separated by sex and the combina tion of sex and
11 athletic status BAI and BMI correlated similarly to BF% with no significant difference between the two measures. T hese results suggest that BAI is a better measure of adiposity than BMI in young adult non athletes.
12 CHAPTER 1 LITERATURE REVIEW AND R ESEARCH RATIONALE Trends in Obesity The most recent National Health and Nutrition Examination Survey (NHANES) data from 2009 2010 indicate that 35.7% of adult men and women in the U.S, are obese (1). Obesity is defined as excess adiposity and with this exc ess body adiposity there is an increased risk of mortality and medical comorbidities such as type 2 diabetes, some cancers, and cardiovascular disease (2 4). Obese adults also experience a lower quality of life, increased medical expenses, and they miss mo re work than non obese adults (5 6). Young adults are at a particularly high risk of weight gain. Many young adults leave home without the proper knowledge to prepare healthy food and do not exercise on a regular basis ( 7 9). In addition, young adults may adopt unhealthy behaviors once they leave home, such as becoming more sedentary, developing unhealthy eating and sleeping habits (9 11), and drinking excessively (12), which may lead to weight gain. Trends in body weight of young adults from ages 18 30 reflect this and are marked by weight increases and an increased prevalence of overweight and obesity (13, 14) Health Consequences of Obesity Obesity has health related consequences. The excess fat mass in obesity is associated with multiple comorbidities (15) such as coronary heart disease, cardiovascular disease (16), liver and gallbladder disease, osteoarthritis and muscle skeletal problems, respiratory problems (17) insulin resistance and increased mortality (18). Subjects with a BF% c lassified as obese h ad higher levels of cardio metabolic risk factors (inflammatory markers, insulin resistance, dyslipidemia, systolic blood pressure, low HDL cholest erol) than subjects with a BF% classified as lean (19) An increase in fat
13 mass accumulation can further be linked with an increased occurrence of metabolic syndrome (19,20). Metabolic syndrome is a name for a cluster of risk factors that occur together and increase the risk for coronary artery disease, stroke, and type 2 diabetes. The two most important risk factors for metab olic syndrome are increased abdominal adiposity (i.e. "apple shaped") and insulin resistance. Other risk factors include aging, hormone changes, and lack of exercise. According to the American Heart Association and the National Heart, Lung, and Blood Insti tute (NHLBI), metabolic syndrome can be diagnosed if three or more of the following signs are present in the patient: high blood lesterol (men <40 mg/dL, Measuring Adiposity A variety of methods have been developed to assess body adiposity, each va rying by ease of use and cost. The gold standards are those methods that produce the most accurate measure of adiposity, but this increased accuracy comes at a price. These techniques are limited by their cost, inconvenience of use, and degree of training required to use them. Anthropometric measures use body measu rements such as height, weight, and circumferences or ratios of body measurements to estimate body adiposity. These measures have their strengths such as minimal cost, convenience for the patient, and portability but are not as accurate as the gold standards This section will describe the various measures of adiposity and their strengths and weaknesses
14 The Gold Standards Hydrostatic weighing, dual energy x ray absorptiometry ( DEXA ) air displacement plethysmography ( ADP ) a nd imaging techniques are considered the gold standards for body composition analysis Hydrostatic w eighing Hydrostatic weighing also known as underwater weighing is the o ldes t of the gold standards, and relies on Archimedes principle to calculate BF%. The first step in underwater weighing is determining body density. To det ermine body density, the subject is first weighed outside of the water. Next the subject is weighed while submerged. From these two weights the specific gravity of the subject is determi ned by dividing the weight of the subject out side of water by the loss in weight of the subject when submerged in water (weight outside of the water minus weight while submerged). Once the specific gravity of the subject has been determined, it is mult iplied by the density of water (1g/cm3). This yields the subject s body density. Human body densities vary between approximately 1.08 g/cm3 (very lean) and 1.00 g/cm3 (severely obese). Roughly sixty percent of the human body is composed of water, which exp lains why body density is very close to the density of water (22). Because adipose tissue is less dense than water, lower body density equals a greater BF%. After body density is calculated, BF% is estimated using a standard formula based on a two compartm ent model. In a two compartment model the body is divided into two parts. One is composed of body fat and the other is composed of all other fat free mass. The most commonly used two compartment model was developed by Siri (23). Other two compartment model s have been proposed to account for body composition differences of fat free mass (muscle and bone) associated with age, sex, ethnicity and race. (24)
15 Underwater weighing has been a gold standard for a multitude of years but the measurement involves distin ct requirements from subjects being measured. Subjects must entirely submerge their body underwater, exhale maximally, then hold their breath and maintain their body position until a weight measurement is acquired. The subject must exhale maximally to mini mize the buoyant effect of air in the lungs (23). Young children, older adults, and individuals with cardio pulmonary diseases may not be able to perform this part of the measurement sufficiently. Special procedures may be required when performing underwat er weighing of obese subjects. Obese subjects have a tendency to float, so it may be difficult for these subjects to fully submerge their bodies. A weight belt may be necessary to completely submerge the subject. T he weights must then be measured and subtr acted from the recorded underwater weight to obtain a true underwater weight (23). Dual e nergy X ray a bsorptiometry (DEXA) Although DEXA is primarily used to measure bone density, it provides an accurate measure of body adiposity. A DEXA scan is obtained b y a DEXA scanner, which is a machine with a mechanical arm that passes over the subject while they lay on a specialized table. The scan usually takes ten minutes to complete. DEXA determines body composition by using X ray beams at two energy levels (one i s high energy and the other is low energy) to differentiate between three types o f tissue (bone, lean soft tissu e and fat) (25). At any one time a DEXA scan can only differentiate between two types of tissues. When bone is present, soft tissues and bone are ea sily distinguished from each other. When only soft tissue is present adipose and lean tissue fractions can be partitioned from each other, with the assumption that lean tissue is composed primarily of water whereas adipose tissue contains little water. The adipose
16 and lean fractions of the soft tissue are then extrapolated to the soft tissue around the bone this produces estimates of total body fat and lean tissue mass. The algorithms used for these extrapolations are considered proprietary information and differ between DEXA scanner manufacturers (25). Validation studies have shown that body fat assessment by DEXA aligns well with the body fat measurements derived using the four compartment model (24). The four compartment model estimates body fat from meas urements of underwater weighing (body density), total body water, and bone mineral values. A concern for the use of DEXA in body composition analysis is the reported differences among scanners from different manufacturers (25). These differences may aris e because each manufacturer uses a different formula for calculating body composition. Air displacement p lethysmography Air displacement plethysmography ( ADP ) calculates BF% by measuring body volume through air d s law. The measureme nt of body volume involves three steps. Before ADP can be used, a calibration process is performed whereby the ADP chamber calibrated to a known volume. The subject is then weighed by the electric scale connected to the ADP chamber wearing only spandex shorts and spandex sports bra ADP chamber. Subjects sit in the ADP wearing spandex and a skull cap. This minimizes sources of isothermal air which could affect the calculation. The ADP chamber calculates body volume by subtracting t he volume of air in the closed ADP chamber that contains the subject from the volume of air in an empty chamber. This measure is repeated twice to verify the agreement between the two volume measures. In the third step adjustments to the volume calculation s are made to account for air in the lungs.
17 This can be done by measuring lung volume directly (while pinching their nose the subject breathes in a disposable tube connected to the machine), or it can be estimated with an equation provided by the ADP syste m. The predictive equation is used when the subject is unable to perform the lung volume measure. Once the system calculates body volume, body density is calculated by use of the subject's weight. Body density is then inserted into a standard formula to ca lculate BF% based on a two compartment model, such as Siri (same model used in underwater weighing) for the general population and Ortiz for African Americans. The Ortiz formula accounts for the higher bone density of African Americans. The reliability and validity of ADP measurements have been published at length (26). Imaging t echniques Imaging techniques have been reserved for research purposes. The two most common imaging techniques are computed tomography (CT) and magnetic resonance imaging (MRI). Comp uted tomography uses X rays in a fan shaped beam to produce two dimensional cross sectional slices of the body. Three dimensional images are produced from a series of these two dimensional cross sections taken aro und a single axis of rotation. Adipose, mus cle, skin, or bone tissue can be identified by their differing densities. Total body mass determined using scans along the length of the body at 10 cm intervals has been shown to be highly accurate when estimating the amount of different body tissues (22). One advantage to CT is that it can be used to separate total adipose tissue into subcutaneous and visceral components. One disadvantage of CT is the high dose of radiation required for the scan. Human cadaver studies have validated the accuracy of CT esti mates of adipose tissue (27). Magnetic resonance imaging (MRI) is based on the interaction of protons in tissues and magnetic fields generated and
18 controlled by the MRI machine. Whole body images are created according to the rate at which protons in hydrog en atoms from various tissues, such as fat, and muscle, return to their normal state after exposure to various magnetic fields. Multiple scans along the length of the body are needed for whole body measurements, this process may require the subject to be i n the MRI machine for 30 min or longer. Like CT images MRI im ages, allow for the separation of subcutaneous adipose tissue from visceral adipose tissue. Human cadaver studies have also validated the accuracy of MRI for estimating adiposity. An advantage of MRI over CT is the absence of radiation. Although imaging techniques like MRI and CT are the most accurate of the gold standards they are limited by their high cost, and need of a radiologist for interpretation (28). Other Measures of Adip osity While the gold standards discussed above provide the most accurate measurement of adiposity, cost and availability to clinicians and researchers limit their use. Instead other anthropometric methods have been developed to estimate adiposity, includin g waist circumference (WC), the waist to hip ratio (WHR), skin fold thickness, bioelectrical impedance analysis (BIA), sagittal abdominal diameter (SAD), body mass index (BMI), and body adiposity index (BAI). These measures are simple and inexpensive to pe rform but have their limitations. This section will describe these anthropometric measures. Waist c ircumference ( WC) Waist circumference is used to assess central adiposity or visceral adipose tissue and has been shown to be highly associated with cardiovascula r disease, metaboli c syndrome and mortality (29). The World Health Organization (WHO ) recommends cutoffs for WC of >102 cm for men and > 88 cm for women. These cutoffs
19 are derived by recognizing WC values that correspond to BMI cutoffs for obesity Waist circumference cutoffs that correspond to a BMI class ification of overweight are >90 cm for men and >80 cm for women (30). Although WC cutoffs were derived from BMI cutoffs, studies have indicated that subjects with normal BMI but higher WC can be at increased risk for cardiovascular dis ease and mortality (31). Pischon et al (32) examined the association of BMI, WC, and WHR with the risk of mortality in 359,387 subjects (25 70 years old) from nine European countries (Denmark, France, Germany, Greece, Italy, the Netherlands, Norway, Spain and the United Kingdom). They reported that high BMI, WC and WHR were each associated with an increased risk of death. In addition there was positive association between high WC and increased mortality among participants with normal BMI. This highlights the importance of assessing abdominal adiposity especially among subjects with normal BMI. Even though WC has been shown to correlate strongly with cardiovascular disease, and mortality it does not provide an estimate of total fat mass. Differences in vis ceral adiposity within WC vary significantly by sex, age, and race (31). Compared to Caucasians, Asian populations have greater visceral adipose tissue, and African populations have less visceral and/or total BF % at a given WC (30). Ford et al. (34) exam ined WC trends in U.S. adults using data from the third National Health and Nutrition Examination Survey (NHANES III) and NHA NES from 1999 to 2000 (NHANES 99 04). They concluded that WC increases with age across both sexes. Waist circumference measurement methods have not been standardized. There are currently eight different locations for measuring WC. Although there has not been standardization for a site measuring WC, most of the eight sites have very high reproducibility (35). The NHLBI recommends WC be measured at the
20 top of the iliac crest whereas the WHO recommends WC measurements should be made at the midpoint between the lowest palpable rib and the top of the iliac crest (36). This is a problem because each measurement does not provide the same meas urement estimate, which may affect results from one person to the next. An advantage of WC over BMI is its ability to detect chan ges in adiposity. Velthuis et al. (37) investigated the effect of a moderate to vigorous exercise program on body composition, among 189 sedentary, postmenopausal women. They measured BF% using DEXA, BMI, WC and hip before and after participation in this exercise program to determine the effect of weight loss on these measures. The authors reported that the exercise program reduce d fat mass, increased lean body mass, and reduced WC, although weight and BMI were unchanged. Waist to h ip ratio ( WHR ) Waist to hip ratio is calculated by dividing waist circumference by hip circumference and is used to predict the risk of metabolic disorders and cardiovascular disease (30). Like WC, WHR is a measure of central adiposity and visceral adipose suggested using cu toffs specific to ethnic groups (30). Koning et al (29) performed a meta analysis of fifteen prospective cohort studies and randomized clinical trials of cardiovascular di sease that measured WC or WHR. They concluded that WHR and WC were significantly a ssociated with the risk of cardiovascular disease and that these measures should be incorporated into cardiovascular disease risk assessments. In addition in the Monitoring Trends and Determinants in Cardiovascular Disease Augsburg (MONICA) study, BMI, WC, and WHR were all strongly and independently
21 related to incident type 2 diabetes in both men and women ages 35 74 (38). Increased WHR has also been linked to increased mortality. Seidell (39) reviewed the relationship of WC and WHR to all cause mortality i n twenty pros pective epidemiological studies T he auth or concluded that increased WC and WHR are related to increased all cause mortality. Th e author also concluded that increased WC and WHR are stronger predictors of increased mortality in younger adults compared to older adu lts. Skinfold m easurement The skinfold measurement involves gripping the skin in between the thumb and forefinger, pulling it away from the body slightly, and placing the fold in between specialized calipers designed to measure skinfold thickness. This mea sures the thickness of two layers of skin and the underlying subcutaneous fat. Skinfold measures are taken at three sites (chest, abdomen, and thigh) on the right side of the body. The skinfold measurements from all three sites are then summed and plugged into a formula to estimate body density. The most commonly used equations for estimating body density were developed by Jackson and Pollock for men and Jackson, Pollock, and Ward for women (40). Body density is then plugged into a formula to estimate BF%. Because of gender, age, racial and ethnic differences in body composition there are over 100 BF% prediction equations reported in the literature and each of these equations are restricted to the population from which the predict ion equation was derived (4 1). The measurement of skinfolds is a popular method of estimating body composition but its inaccuracies have been described. Lohman (42) reported standard errors from skinfold measurements to be 2.6 kg for fat free mass and 3.5% for percent body fat. Some of the potential sources of error found in the skinfolds method included variation in subcutaneous in relation to total fat, variation in skinfold thickness in relation
22 to subcutaneous fat, and technical error in the skinfold measurement (i.e. intra and i nter observer variability, difficulty in obtaining reliable and accurate readings on older participants with loose connective tissue or obese individuals with large folds) (43). Bioelectrical impedance analysis (BIA) Bioelectrical impedance measures the resistan ce of body tissues t o the flow of a small, electrical current. D ifferent tissues conduct electricity differently based on their water and dissolved electrolyte content Seventy three percent free mass is water and conducts the electrical current F at and b one are composed of a relatively small amount of water and are thus nonconductive and resist the flow of the electrical curren t (23). During BIA a small electrical current passes through electrodes attached to either the wrists or ankles Th e resistance of the electrical current is used to estimate total body lean mass and BF% using a standardized equation. There are a multitude of population specific BF% estimation equations reported in the literature (44 ). The equipment needed for BIA is portable, relativ ely easy to use, affordable, and pose little risk to subjects ( BIA is not recommended for subjects with a pacemaker). Studies have shown that body composition estimated using BIA is influenced by sex, age, disease state race and ethnicity level of fatness environment, and phase of menstrual cycle (23). Sagittal abdominal d iameter (SAD) S agittal abdominal diameter is a simple measurement that may be even better than WC for predicting metabolic syndrome (4 5 49 ), dyslipidemia, and cardiovascular disease (50 54 ), although this is still under investigation. Gustat et al (55 ) examined the relationship between SAD and other measures of adiposity (Triceps skinfold, WC, BMI, WHR, and Coincity Index) and cardiovascular risk factors in young adults (20 38 years
23 old). T he sample was biracial and consisted of 409 African Americans and 1,011 Caucasians form Bogalusa, Louisiana. They concluded SAD correlated more strongly with the cardiovascular risk factors than other obesity measures. Sagittal abdominal diameter can be me asured directly on a patient, generally in the supine position, as the distance between the examining table and the apex of the abdominal girth or the largest diameter between the bottom of the sternum and the umbilicus. Although SAD may be a promising mea sure of abdominal adiposity, the measure needs to be standardized and normal th resholds identified. Risrus (47 ) suggested a SAD cutoff for increased cardiometabolic risk to be 22cm in men and 20cm in women. These cutoffs were developed from 4032 subjects (1936 men and 2096 women) older (60 years old) Europeans and may not generalize well to other populations. Sampaio et al (50 ) evaluated the validity of SAD as a predictor of visceral adipose tissue and identified SAD cutoffs for increased cardiovascular d isease risk in 92 subjects (57 women and 35 men ages 20 83 years old ) They concluded there was a high correlation between SAD and visceral adipose tissue and recommended SAD cutoffs for increased cardiometabolic risk to be 20.5 cm in men and 19.3 cm in wom en. These cutoffs were based on a small sample of subjects that included a large age range. Age specific standards are needed due to changes in body composition with age. Body mass i ndex (BMI) Currently clinicians and researchers commonly use body mass index (BM I) as an indicator of adiposity. Body mass index is calculated as weight in kilograms divided by height in meters squared. This measure was first proposed by Adolphus Quetelet in 1832 based on the observation that body weight was proportional to the square of the height in adults with normal body frames. When developing the index that would bear
24 his name, Quetelet had no interest in obesity. His concern was defining the into a Gaussian curve (56 ). A Gaussian curve is often referred to as a bell or normal curve. It is a symmetrical curve that represents the normal distribution of a trait in a pop ulation. In 1972 Keys et al. (57 ) examined how different relative weight ratios (weight to height, the Po nderal index, and the Quetelet index) correlated to obesity in 7,426 male subjects from the ages of 18 60 (the majority of the sample was aged 40 59 years old). The Ponderal index is found by dividing the cube root of weight by height (57 ). The subjects w ere composed of 12 cohorts from 5 countries (United States, Japan, South Africa, Finland and Italy). Keys et al (57 ) among the ratios of relative weight Keys went on to rename 1998 the NHLBI released the first guidelines for the ass essment of obesity using BMI (58). Table 1 1 lists the different BMI categories and the corresponding weight status classification set forth by th e NHLBI (59 ): Table 1 1 W eight status classification within BMI ranges BMI Weight Status < 18.5 Underweight 18.5 24.9 Normal 25.0 29.9 Overweight Obese Despite its wide spread use in obesity research, it has been debated whether BMI represent s body adiposity adequately ( 60 63). Gmez Ambrosi et al. (63 ) compared the diagnosis of obesity in 6 123 Caucasian subjects (924 lean, 1 637 overweight and 3 562 obese classified according to BMI) aged 18 80 years old, using BMI and BF %. Body fat percentage was determined using air dis placement plethysmography. Gmez Ambrosi
25 et al. (63 ) reported that 29% of subjects with a normal BMI and 80% of subjects with a BMI that classified them as overweight had a BF% classified as obese. These results suggest that BMI does not reliably reflect BF% and that BF% should be considered when at all possible when estimating disease risk. Body Mass Index does not directly generalize among different age groups, between the sexes, different ethnic groups, and different racial groups (64,65 ). The Dietary Reference Intakes for Energy contains a T able 1 2 that lists typical BF% ranges (obtained from bioelectrical impendence analysis) wi thin BMI classifications Table 1 2. Body fat percentages within BMI classifications* BF% BMI Range (kg/m 2 ) Classification Men Women 18.5 24.99 Normal 13 21 23 31 25 29.99 Overweight 21 25 31 37 30 34.99 Obese 25 31 37 42 35 or higher Clinically obese > 31 > 42 Rep roduced from IOM, 2005. These ranges were based on body composition data from the NHANES III (1) Although gender was considered when evaluating normal BF% ranges within BMI classifications and a large sample size was assessed (15,000 participants), the ranges have limitations. Theses ranges do not take into account, age, race, and ethni city (65,66). Li et al. (64 ) developed BF% means that align with BMI categories stratified by sex, age, race, and eth nicity using data from NHANES 99 04. Body fat percentage was calculated using DEXA. Table 1 3. s ummarize s some interesting patterns, revea led in the study.
26 Table 1 3. Mean percentage BF% according to BMI categories among US adults from NHANES 1999 2004 BMI (kg/m 2 ) Classification Total BF% < 25 25 29 30 34 Mean Mean Mean Mean Mean Men BF% Crude 28.1 22.7 28.2 32.3 36.9 Adjusted 28.2 22.9 28 32.1 37 Race ethnicity Non hispanic white 28.3a 22.9a 28.3a 32.6a 37.2a Non hispanic black 25.8b 19.7b 26.2b 29.9b 35.8a Mexican American 28.9a 23.6a 28.8a 32.3a 37.2a Other 27.9a 23.6a 28.1a 32.2a 36.1a Age 20 39 y 26.1a 21.0a 27.0a 31.4a 36.7a 40 59 y 28.7b 23.6b 28.0b 31.8a 36.7a 60 79 y 30.9c 25.8c 30.2c 34.5b 38.0b 30.6c 27.5d 31.9d 35.8b 38.8a,b Women Crude 40.05 34.05 40.85 44.25 48.25 Adjusted 39.95 34.15 40.65 44.15 48.35 Race ethnicity Non hispanic white 39.7a 33.8a 41.1a 44.4a 48.6a Non hispanic black 40.9b.c 32.4b 39.1b 43.1b 47.2b Mexican American 41.6b 36.0c 41.1a 44.4a 47.6b Other 39.9a,c 34.8a,c 40.7a 43.8a,b 47.8b
27 Table 1 3 Continued BMI (kg/m 2 ) Classification Total BF% < 25 25 29 30 34 Mean Mean Mean Mean Mean Women BF% Age 2 0 39 y 37.8a 32.2a 39.5a 43.5a 48.0a,b 40 59 y 40.6b 34.4b 40.8b 43.9a,c 48.1a 60 79 y 42.5c 36.9c 42.3c 45.2b 48.7b 40.6b 36.9c 42.0c 45.0b,c 47.7a,b Values within categories of race ethnicity or age groups in the same column with di fferent superscript letters are significantly different. Across all BMI classifications regardless of age and sex non Hispanic blacks had a lower BF% compared to non Hispanic whites, and Mexican Americans. Secondly BF% cutoffs were higher in women than me n for any given BMI irrespective of age, race, and ethnicity. Lastly the oldest group in their study (50 84 years old) had the highest BF% ranges for the same BMI classification regardles s of sex, race, and ethnicity. Li et al. (64 ) concluded that BF% rang es that corresponded to BMI classifications noticeably varied depending on age, sex, race and ethnicity. Gallagher et al. (66 ) used DEXA to measure and determine BF% ranges that align with BMI categories of 1626 subjects (ages 20 79 years old) from three ethnic groups (Caucasian, African American, and Asian) The study also included cutoffs for a young adult (20 39 year old) sub group. Gallagher et al. (66 ) proposed the following BF% cutoffs corresponding to BMI classifications in the young adults (Table 1 4)
28 Table 1 4. Recommended BF% cut offs by ethnicity for 20 39 year olds* Women Men BMI African American Asian White African American Asian White < 18.5 20 25 21 8 13 8 32 35 33 20 23 21 38 40 39 26 28 26 Reproduced From Gall agher et al. (2) Table 1 4 demonstrates and reinforces the notion of the variability of BF% within BMI classification in regards to ethnicity. Asians had higher BF% at lower BMIs, than the other two ethnic groups (Cauc a sians and African Americans). Table 1 4 also shows a slight difference between African Americans and Caucasian men and women. dings of Li et al. Li et al. (64 ) reported small yet significant differences in BF% between non Hispanic whites and no n Hispanic blacks across all BMI categories. These differences may be due to the use different DEXA scanners. As stated earlier, DEXA scanners from different manufactures can produce difference in BF%. The differences in BF% ranges that corresponded to BMI classifications depending on age, sex, race and ethnicity arise because distribution of body fat for a gi ven BMI differs by ethnicity (67 ). between fat or lean body mass it is particularly inaccurate in persons wi th elevated lean body mass such as athletes (68,69 ) and persons with normal weight but high visceral adipose tissue (70,71 ). This disparity occurs because the numerator in BMI is total weight and does not distinguish between fat and lean mass. Thus, indivi duals with higher weight due to increased lean body mass will have a higher BMI and falsely diagnosed as overweight or obese, and individuals with normal weight but excess body
29 fat may not be diagnosed as overweight or obese when in fact they should. Ode e t al. (68 ) used air displacement plethysmography ( ADP BF% in collegiate athletes and non athletes. They determined that the BF% ranges within BMI classifications put forth in the Dietary Reference Intakes for Energy mis classified male college athletes with normal BF%, female college athletes with normal BF% and male non athletes with normal BF% as having a BMI classification of overweight or obese. A significant percentage of female nonatheletes classified as having a no rmal BMI had a higher BF%. Ode et al. (68 ) concluded that BMI was not a good measure of BF% in this population. Body a diposity i ndex (BAI) A new anthropometric measure, BAI has recently been developed and has been found to better reflect BF% than BMI for adult m en and women of different ethnicities. Bergman et al. (60 ) evaluated existing data from the BetaGene study of Mexican Americans to find a trait or combination of traits most strongly correlated to DEXA measured adiposity to develop a new index of BF%. They analyzed demographic and anthropometric measures, including sex, age, weight, height, waist and hip circumference, from 1733 Mexican American adults between the ages of 18 67. They also examined the covariance among variables to select a combination of va riables that independently correlated with adiposity so that each variable would contribute independent information to the predi ction of BF%. Bergman et al. (60 ) determined that hip circumference and height correlated the strongest to BF%, and used these t wo measures as the basis for the new index they named BAI. The correlation between hip circumference and BF% was positive, and there was an inverse correlation between height and BF%, so they determined that the base for BAI would be hip circumference
30 divi ded by height (hip/hei ght). Quetelet (56 ) established that the relationship between body fat and height is nonlinear, so height in the proposed BAI had to be raised to a power term. The power term for the proposed BAI was determined by fitting the correlat ion between BF% and the BAI for all subjects on a parabola to a range of exponent values (1.2 1.8). Once fitted to the parabola derivation was used to find that an exponent of 1.5 in the BAI formula (hip/height 1.5 ) produced the maximal correlation to BF%. The authors then determined that the relationship between BF% and BAI (hip/height1.5) had a slope similar to 1 (0.934).This relationship allowed for the identification of a y intercept of 18 that maximized the correlation between BF% and BAI. The final BAI formula is Hip Circum ference/(Height 1.5 ) 18 and is numerically approximately equal to the percentage of body fat. Although Bergman et al. (60) concluded that BAI predicted body adiposity better than BMI they also reported that BAI overestimated BF% at lower levels of adiposity and underestimated BF% a t higher levels of adiposity Once they developed the for mula for BAI, Bergman et al. (60 ) validated the measure using cross sectional data of 223 African American adults between the ages of 18 67 who participat ed in the Triglyceride and Cardiovascular Risk in African American (TARA) study. Demographic and anthropometric information collected in the TARA study included sex, age, height, weight, BMI, waist and hip circumference and DEXA measured BF%. They calculat ed the percent difference between the BAI estimates of BF% of the BetaGene and TARA studies at specific ranges of DEXA derived BF%. They found that BAI predicted BF% similarly in both studies, only differing at the BF % range <10%. Vinknes et al. (72 ) compa red the relationships of BMI and BAI with BF%
31 assessed by DEXA, in 5,193 middle aged (47 49 years old) and elderly (71 74 years old) Caucasian, European subjects. They reported that the correlation between BAI and BF% was stronger (r = 0.78) than the corre lation between BMI and BF% (r = 0.56) with similar results in the middle aged and elderly groups. When separated by sex BMI was more strongly correlated with BF% in men, (r = 0.76), and women (r = 0.81) than BAI in men, (r = 0.57) and women (r = 0.72). The y also reported that BAI overestimated BF% in lean subjects (particularly in men) and underestimate it in those with higher proportio ns of body fat. Lopez et al. (73 ) compared BAI and BMI measurements to BF% using bioelectrical impedance analysis (BIA) in Spanish subjects from Mallorca, Spain (1,726 women and 1,474 men). They found that although BAI and BMI were positively correlated, BAI was more strongly correlated to BF% than BMI was correlated to BF% for all subjects. When separated by sex the correlat ions between BMI and BF% were higher than the ones obtained between BAI and BF% for both men and women. Lopez et al (73 ) reported BAI overestimated BF% in men and slightly underestimated BF% in women. Although the researchers used BIA to assess body adip osity, which has its limitations, a strength of this study is that it inclu ded a large number of subjects. Johnson et al. (74 ) evaluated whether BAI reflected adiposity better than BMI in 623 European Americans adults, using DEXA measured adiposity. Their findings were simi lar to Bergman et al. (60 ). The correlation between BAI and BF% was significantly stronger than that between BMI and BF% for all subjects. When separated by sex BAI overestimated BF% in men and underestimated BF% in women. They also repo rted that BAI overestimated BF% at lower levels of adiposity and underestimated BF% at higher levels of adiposity. Barreira et al (75 ) investigated the sex specific relationship between
32 BAI and BF% and BMI and BF% in a large (n=3851) biracial (Caucasian a nd African American) sample ages 18 69. They concluded that BMI and BAI correlated similarly to BF% across both sex and race groups. Freedman et al (76 ) compared the relationship of BF%, as assessed by DEXA, to WC, hip circumference, BAI, and BMI, in 1151 adults ages 18 110 years old (mean age was 45), in five ethnic categories (Caucasian, African American, Hispanic, Asian and other). They reported that BF% was correlated similarly to BAI, BMI, WC, and hip circumference with BAI having the strongest correl ation to BF% for all subjects W hen controlled for both sex and age BMI correlated more strongly to BF%. Additionally BAI in general underestimated BF% by 2.5% in women and o verestimated BF% by 4.0% in men T his was similar to the findin gs of Johnson et al (74 ). From these results they concluded BAI is not more accurate than BMI, waist circumference, or hip circumference when estimating BF%. BAI appears to be a promising new anthropometric measure that may more accurately assess body ad iposity than BMI in a variety of populations. The accuracy of the BAI has been evaluated in college athletes. To date only one study has re ported BAI in athletes. Esco (77 ) evaluated whether BAI reflected DEXA measured BF% better than BMI in 30 collegiate female athletes (age = 20.0 1.3) from three sports (so ccer, tennis, and basketball). The author reported that neither BAI nor BMI correlated strongly to BF%, but that BMI was more strongly correlated to BF% than BAI (rBMI BF% =0.49, rBAI BF% =0.28). Esco also reported that BAI underestimated BF% at higher levels of adiposity and overestimated BF% at lower levels of adiposity in these women. More research is needed in athletes to determine if BAI has the same
33 limitations as BMI in reflecting body adiposity of athletes. The current study has one specific aim: To determine wheth er BAI more strongly correlates to BF% than BMI in undergraduate non athletes and athletes ages 18 24 years. Hypothesis : BAI will correlate more strongly to BF% than BMI in undergraduat e non athletes and athletes between the ages of 18 24
34 CHAPTER 2 METHODS AND PROCEDURE Study Design and S ubjects A cross sectional observational study was conducted using a convenience sample of u ndergraduate students between the ages of 18 24 years Approval for the study was granted by the University of Florida Institutional Review Board 02 (IRB) starting fall 2011 and ending fall 2012 All subjects were volunteers. Undergraduate students were recruited by visiting classes, advertising in the school newspaper, and r ecruiting from other studies. University athletes were recruited from a convenience sample of athletes who visited the University Athletic Association (UAA) nutrition office for BF% measurements. Students from participating Family, Youth & Community Scienc es classes received five points of extra credit for participation. No compensation was given to the other participants. Non athletes interested in the study contacted the study coordinator via email to schedule an appointment for participation in the study Physical Assessments Nonathletes completed a 20 minute physical assessment in the clinical lab (Room 227) of the Food Science and Human Nutrition building. University athletes who participated in the study completed the physical assessment in the UAA nut rition office during their BF% measurement. Subjects were fasting for two hours before their scheduled appointments. Upon arrival to their appointment, informed consent was obtained (Appendix A) from each subject and the subjects were asked to change into approved clothes for the physical assessment (i.e., spandex shorts, sports bra). Physical assessment measurements were obtained by trained research staff and included height, weight, hip circumference using the NHANES protocol described below
35 and BF% using ADP M easures were recorded into an approved data collection form (Appendix B). Height was measured in centimeters (cm) to the nearest 0.1 cm using a wall mounted stadiometer (Heightronic 235) or a portable stadiometer (SECA 213). Subjects were measured withou t shoes. Height was obtained twice to ensure accuracy. If the measures varied by greater than 0.5 cm they were repeated until two values within 0.5 cm were obtained. Weight was obtained during body composition assessment using the ADP calibrated electroni c scale to the nearest 0.1 kg. Hip circumference was measured in centimeters (cm) using NHANES protocol to the nearest 0.5 cm using a Gulick tape (Patterson Medical). A Gulick tape is a self retracting measuring tape that allows the researcher to keep a co nstant tension when measuring body dimensions. Attached to the end of the Gulick tape is a tension indicator composed of a compression spring attached to two beads spaced slightly apart. When there is no tension on the tape measure the beads are covered by a sleeve. When taking measurements enough tension should be applied so that only one bead can be observed. This is what allows the researcher to keep constant tension. Keeping constant tension when measuring body dimensions is important because the measur ement will vary if tension is changed. If tension is increased tissue compression will increase thus decreasing the measured circumference. Subjects were measured in spandex shorts or a swim suit with weight distributed on both feet. The tape was placed a t the maximum extension of the buttocks (Figure 2 1).
36 Figure 2 1. Placement of tape during measurement of hip circumference To ensure accuracy hip circumference was obtained twice. If the measures varied by greater than 0.5 cm they were repeated unti l two values within 0.5 cm were obtained. Body fat percentage was measured using air displacement plethysmography ( ADP Life Measurement, Inc.). In order to obtain an accurate measurement subjects removed all jewelry, and wore a swim cap. The swim cap was used to minimize a source of isothermal air (air trappe d i n hair around the head) that could affect the calculation Statistical Analysis correlation and corresponding 95% confidence int ervals. This assesses the strength of the relationship between two anthropometric measures. Comparing the confidence intervals for each correlation coefficient allows for the determination of whether a correlation coefficient is different from another corr elation coefficient. Bland Altman limits of agreement plots were used to visually compare t he existence of any differences between BAI and BF% In this method the differences between the two measures for each subject were plotted against the averages of the two measures for each subject.
37 Statistical analysis was carried out using SAS 9.0 (SAS Institute, Cary, North Carolina) and GraphPad Prism 6m (GraphPad Inc., La Jolla, California) statistical software.
38 CHAPTER 3 RESULT S Subject Characteristics Overall s ubject characte ristics are presented in Table 3 1 Nonathlete and a thlete subject characteristics are presented in Table 3 2 and Table 3 3 respectively. Table 3 1 Overall subject characteristics. All values are mean SD Total Men Women Total (n) 249 94 155 Height (m) 1.7 0.1 1.8 0.1 1.6 0.7 Weight (kg) 68.1 15.8 78.8 15.0 61.8 14.4 BMI(kg/m 2 ) 23.4 4.2 24.8 4.2 22.5 3.8 Hip Circumference (cm) 97.5 8.3 98.3 8.1 96.9 8.0 BAI, mean 26.2 4.6 23.2 3.3 27.9 4.1 BF%, mean 22.9 9.2 15.6 6.4 27.1 7.5 Race ethnicity Non Hispanic white 127 51 76 Non Hispanic African American 23 5 18 Hispanic white 23 10 13 Hispanic African American 0 0 0 Asian 14 10 4 Other 5 1 4 Not reported 57 17 40 In the o verall sample women substantially outnumbered men. Two thirds of the population was composed of women. The majority of the overall sample was composed of non Hispanic whites (n=127). The rest of the sample identified themselves as non Hispanic African Amer ican (n=23), Hispanic whites (n=23), Asian (n=14), other (n=5) or did not report their race ethnicity (n=57). Weight, height, hip circumference, and BMI were higher in men compared to women. When comparing average BAI and BF%
39 between the sexes in the overa ll population, both BAI and BF% were higher in women compared to men. Subject characteristics for non a thletes are presented in Table 3 2 Table 3 2 Nonathlete subject characteristics. All values are mean SD Total Men Women Total (n) 195 64 131 H eight (m) 1.7 0.1 1.8 0.1 1.6 0.1 Weight (kg) 65.6 14.7 76.5 14.2 60.3 11.7 BMI(kg/m 2 ) 23.1 4.2 24.5 4.4 22.4 3.9 Hip Circumference (cm) 70.0 8.2 96.7 8.2 97.1 28.4 BAI 26.7 4.6 23.2 3.6 28.4 4.1 BF% 24.4 8.9 16.3 6 .4 28.6 7.1 Race ethnicity Non Hispanic white 88 27 61 Non Hispanic African American 18 4 14 Hispanic white 19 8 11 Hispanic African American 0 0 0 Asian 13 10 3 Other 3 0 3 Not reported 54 15 39 The sex and racial ethnic mak e up of the nonathlete sample was similar to the overall sample in that women substantially outnumbered men, and it was primarily composed of non Hispanic whites. When comparing average BMI across all the non athlete compared to women. Men also on were taller and weighed more than women. When comparing average hip circumference across sex women had slightly larger hip circumferences compared to men. BAI and BF% were lower in men compared to women.
40 Athlete subject ch aracteristics are presented in Table 3 3 Table 3 3 Athlete subject characteristics. All values are mean SD Total Men Women Total (n) 54 30 24 Height (m) 1.8 0.1 1.8 0.1 1.7 0.1 Weight (kg) 76.4 16.6 84.1 15.7 65.6 10.2 BMI(kg/m 2 ) 28.4 11.4 25.3 3.5 22.8 2.9 Hip Circumference (cm) 99.0 7.3 101.7 6.9 95.6 6.1 BAI, mean 24.1 3.2 23.1 2.8 25.4 3.2 BF%, mean 17.0 6.5 14.0 6.2 20.0 5.8 Race ethnicity Non Hispanic white 39 24 15 Non Hispanic African Ameri can 5 1 4 Hispanic white 4 2 2 Hispanic African American 0 0 0 Asian 1 0 1 Other 2 1 1 Not reported 3 2 1 Sport Baseball 12 12 0 Basketball 4 1 3 Crew 7 4 3 Football 1 1 0
41 Table 3 3 Continued Total Men Women Sport Golf 2 1 1 Gymnastics 3 0 3 Lacrosse 1 0 1 Tennis 7 7 0 Soccer 4 0 4 Softball 1 0 1 Swimming 1 1 1 Track 6 1 6 Volleyball 1 0 1 Cross country/track 2 2 0 The majority of the athlete sample was composed of non Hispanic whites (n=39). The rest of the sample identified themselves as non Hispanic African American (n=5), Hispanic whites (n=4), Asian (n=1), other (n=2) or did not report their race ethnicity (n=3). All athletes, male athletes, and female athletes were taller, heavie r and had a higher BMI than nonathletes. All athletes, female athletes, and male athletes had lower B AI and BF% than non athletes. All athletes, female athletes, and male athletes had larger hip circumferences than their non athletes counterparts. Female a thletes had smaller hip circumferences than female nonathletes. The athletes that participated in this study represented a wide gamut of sports, the majority coming from baseball (n=12), track/cross country (n=8), crew (n=7), and tennis (n=7).
42 rrelations All Subjects for all s ubjects are presented in Table 3 4 Table 3 4 subjects B MI BAI Men 0.62 (0.52, 0.71) 0.65 (0.55, 0.73) Women 0.62 (0.48, 0.73) 0.55 (0.39, 0.68) Total 0.31 (0.19, 0.41) 0.73 (0.67, 0.79) The correlations between BAI to BF% and BMI to BF% were not different for men [r BAI BF% =0.65; (0.55, 0.73), r BMI BF% =0.62; (0.5 2, 0.71)] and women [r BAI BF% =0.55; (0.39, 0.68), r BMI BF% =0.62; (0.48, 0.73)]. These relationships for men and women are illustrated in Figures 3 1 and 3 2 respectively.
43 Figure 3 1 Correlation of BMI and BAI to BF% in all male subjects Figure 3 2 Correlation of BMI and BAI to BF% in all female subjects For all participants BAI was more strongly correlated to BF% than BMI was correlated to BF% [r BAI BF% =0.73; (0.67, 0.79) vs. r BMI BF% =0.31; (0.19, 0.41)]. The confidence
44 intervals for all subjects do not overlap, which indicates that the correlations are different. This relationship is illustrated in F igure 3 3. Figure 3 3 Correlation of BMI and BAI to BF% in all subjects A Bland Altman limits of agreement between BAI and BF% for all subjects is pres ented in Figure 3 4
45 Figure 3 4 Bland Altman limits of agreement plot between BAI and BF% in all subjects. The limits of agreement (95% confidence intervals) between the BAI and BF% ranged between 9.5 and 16.2 Body adiposity index underestimated BF% at lower body fat percentages and overestimated BF% at higher BF%. The plot also showed that BAI predicted BF% well for those whose BF% was in the 20 30% range Non athletes BF% for non a thletes are presented in Table 3 5
46 Table 3 5 nonathletes BMI BAI Men 0.66 (0.50, 0.78) 0.63 (0.45, 0.76) Women 0.72 (0.63,0.80) 0.66 (0.55, 0.74) Total 0.38 (0.25, 0.49) 0.76 (0.70, 0.81) In nonathletes, the correlations between BAI to BF% and BMI and BF% were not different for men [ r BAI BF% =0.63; (0.45 0.76), r BMI BF% =0.66; (0 .50, 0.78)] and women [r BAI BF% =0.66; (0.55, 0.74), r BMI BF% =0.72; (0.63, 0.80)]. The correlations are for men and wo men are illustrated in Figures 3 5 and 3 6 respectively. Figure 3 5 Correlati on of BMI and BAI to BF% in male non athletes
47 Figure 3 6. Correlation of BMI and BAI to BF% in fe male non athletes BAI correlated more strongly with BF% than BMI in non athletes [r BAI BF% =0.76; (0.70, 0.81) vs. r BMI BF% =0. 38; (0.25, 0.49)] Since the confidence intervals do not overlap the correlations are different. The correlation of BAI to BF% and the correlation of BMI to BF% in non ath letes is illustrated in Figure 3 7
48 Figure 3 7 Correlati on of BMI and BAI to BF% in nonathletes The Bland Altman limits of agreement plot in non athletes is illustrated in Figure 3 8 Figure 3 8 Bland Altman limits of agreement plot between BAI and BF% in non athletes
49 The limits of agreement (95% confidence intervals) between the BAI and BF% ranged between 9.8 and 14.4. According to this plot BAI overestimated BF% at lower levels of adiposity and underestimated BF% at higher levels of adiposity. Athletes coefficients and 95% confidence intervals for BAI, BMI, and BF% in athletes are listed in Table 3 5. Table 3 6 r BAI to BF% and, BMI to BF% in athletes BMI BAI Men 0.62 (0.33, 0.81) 0.35 (0.02, 0.62) Women 0.43 (0 .04, 0.71) 0.24 (0.18, 0.59) Total 0.29 (0.02, 0.52) 0.41(0.16, 0.61) In athletes BMI was more strongly correlated to BF% than BAI in men r BAI BF% =0.35; (0.02, 0.62), r BMI BF% =0.62; (0.33, 0.81) and women [r BAI BF% =0.24; (0.18, 0.59), r BMI BF% = 0.43; (0.04, 0.71)]. Although numerically higher the 95% confidence intervals overlapped for both the values with the estimate falling within the bounds. This indicates that two measures are not different. The correlations for men and women are illustrated in Figures 3 9 and 3 1 0 respectively.
50 Figure 3 9 Correlation of BMI and BAI to BF% in male athletes Figure 3 10. Correlation of BMI and BAI to BF% in fe male athletes
51 For all athletes BAI and BMI moderately to weakly correlated with BF%, [r BAI BF % =0.41; (0.16, 0.61) vs. r BMI BF% =0.29; (0.02, 0.52)]. This relation ship is illustrated in Figure 3 11 F igure 3 11 Correlation of BMI and BAI to BF% in all athletes Bland Altman limits of agreement plot between BAI and BF% in ath letes is illustrated in Figure 3 12
52 Figure 3 12 Bland Altman limits of agreement plot between BAI a nd BF% in athletes The limits of agreement (95% confidence intervals) between the BAI and BF% ranged between 4.5 and 18.8. According to this plot BAI overestimated BF% at all levels of adiposity
53 CHAPTER 4 DI S CU SS ION Body adiposity index is a new anthro pometric measure that has been shown to better reflect BF% than BMI in a variety of populations (60,72 77 ). The scientific literature is limited for BAI because of its relative newness. The purpose of this research was to determine whether BAI is an approp riate measure of adiposity in young non athletes and athletes ages 18 24. When separated by sex both BAI and BMI correlated similarly to BF%. This is consistent with the findings of Barreira et al (75 ). They investigated the sex specific relationship betw een BAI and BF% and BMI and BF% in a large biracial sample and found that in each sex and race group that the correlations with BF% were similar for BMI and BAI. Lopez et al (73 ) Johnson et al.(74 ), and Freedman et al.(76) showed the correlation between B AI and BF% was stronger than that of BMI and BF%, but BAI overestimated BF% in men and slightly underestimated BF% in women. The f indings by Lopez et al. (73 ), Johnson et al. (74 ) and Freedman et al.(76) that BAI overestimated BF% in men and underestimate d BF% in women could be explained by the fact that men tend to have lower BF%, whereas women tend to have higher BF% and BAI overestimates BF% at lower levels of adiposity and underestimate BF% at higher levels of adiposity. The findings by Lopez et al. (7 3 ), Johnson et al. (74 ) and Freedman et al.(76) could also be explained by the differences in hip circumferences of men and women Bergman et al.(60 ) stated that hip circumference has the potential to introduce error when estimating BF% using BAI. A 10% change in hip circumference could produce a 10% change in BAI predicted BF% because it is in the numerator of the fraction defining BAI
54 The correlation betw een BAI and BF% was stronger than the correlation between BMI and BF% for all subjects These resu lts agree with those reported in other studies of African American s and Mexican Americans (60), Caucasians(74 ), and Europeans of all different ages(7 2,73 ). Although there was not enough subjects in the current study to analyze by race/ethnicity, BAI appear s to reflect adiposity better than BMI in young adults of mixed ethnicities. According to the Bland Altman limits of agreement plot, BAI overestimate d BF% at lower levels of adiposity by as much as 16.2 % and underestimated BF% at higher levels of adiposity by as much as 9.5 % for all subjects The same relationship of BAI and BF% was observed by other stud ies in different populations (60, 72, 74 76). Vinknes et al (72 ) suggested the underestimation of BF% by BAI at higher levels of adiposity may be explai n ed by, as weight increases, abdominal adiposity increases and this increased abdominal obesity is not captured well by hip circumference Three subjects from this study have been chosen that illustrate this finding. Subject A had a lower BF% of 10.2% and B AI estimated BF% to be 22.1%. BAI overestimated BF% by 11.9%. Subject B had a high BF% of 33.1% and BAI estimated BF% to be 25.5%. BAI underestimated BF% by 7.6%. Subject C had a BF% to be 26.7% and BAI estimated BF% of 26.1%. When athletes were separated from non athletes a strong correlation was found between BAI and BF% in non athletes which was stronger than the correlation between BMI and BF%. The Bland Altman limits of agreement plot in non athletes also showed that BAI overestimate d BF% at lower lev els of adiposity by as much as 14.4 % and underestimated BF% at higher levels of adiposity by as much as 9.9 %. In addition,
55 although neither BAI nor BMI correlated strongly to BF% in athletes, the correlation was higher for BAI. This finding coupled with th e finding that BAI overestimates BF% at lower levels of adiposity suggest that BAI suffers from the limitation s of BMI in this population. BMI does not correlate strongly with BF% in athletes because it only accounts for weight and has the inability to dis tinguish between fat and lean body mass and athletes have elevated lean body mass. Even though BAI does not take into account weight when estimating BF% it does not correlate strongly with BF% in athletes because of th eir increased hip circumferences. In athletes the muscles around the hip region are very active and tend to be larger than the non athlete population. When comparing the correlation between BAI and BF% to that of BMI and BF% in the sex by athletic status sub populations BMI was more strong ly c orrelated to BF% than BAI in male athletes and female athletes. Although numerically higher the 95% confidence intervals overlapped for both the values with the estima te falling within the bounds. This indicates that two measures are not different The se results are similar to those of Esco (77 ) wh o evaluated whether BAI reflected BF% better than BMI in 30 collegiate female athletes, but did not explore this same relationship in male athletes or in non athletes. They found that neither BMI nor BAI cor related strongly to BF% but the correlation was higher for BMI Although BAI and BF% correlated more strongly than the correlation between BMI and BF% in the overall sample and the nonathlete sample, BAI and BMI correlated similarly to BF% when divided int o sex, non athlete, athlete sex by non athlete, and sex by athlete populations The lack of differe ntiation in these sample populations
56 between BAI and BMI in this study could be a result of lost statistical power due to the division of the overall sampl e population into smaller and smaller subpopulations. In conclusion, the results of this study have been consistent with other studies comparing B AI, BMI and BF%. Body adiposity index not BMI correlates more strong ly with BF% in the overall and non athl ete sample but BMI and BAI correlated similarly to BF% when the overall sample was divided into sex, non athlete, athlete, sex by non athlete, and sex by athlete populations Body adiposity index may still suffer from the same limitations of BMI, such as o verestimating BF% at lower levels of adiposity and underestimating BF% at higher adiposity. B ody adiposity index appears t o be good m easure of adiposity in large epidemiological studies because of its convenience and low cost. When estimating BAI, a scale is not needed this could be an invaluable advantage of BMI especially when evaluating body composition in populations were a scale is not readily accessible. Although BAI offers advantages over BMI, it still remains inconclusive whether BAI is a more usefu l predictor of obesity related morbidity and mortality compared with BMI.
57 APPENDIX A INFORMED CONSENT Nonathletes
63 APPENDIX B DATA COLLECTION SHEETS BIRTHDATE (mo/day/yr): Year in School (circle) First Second T hird Fourth Major: Race: (Please Circle) American Indian or Alaska Native Asian Black or African American Native Hawaiian or Other Pacific Islander White Ethnicity: (Please Circle) Hispanic or Latino Not Hispanic or Latino A certain amount of body fat is absolutely necessary for good health. Fat plays an important role in protecting internal organs, providing energy, and regulating hormones. The minimal amount 5% for men. If too much accum ulates over time, health may be compromised. 1. Have you ever had your body fat percentage measured by a trained professional? yes no 2. What is your current body fat percentage range (circle one)? <5% 5 8% 9 12% 13 20% 21 30% >30% 3. Which of the following best describes your current body fat percentage (circle one)? Risky (Too Low) Ultra Lean Lean Moderately Lean Excess Fat Risky (Too H igh) 4. What is your current Height? __________ 5. What is your Current weight? __________ 6. Do you know what body mass index is? Yes No ID______________ DATA FORM FOR MEN
64 7. Which of the following best describes your current weight (circle one)? underweight normal weight overweight obese Data collector initials ____________________ Weight (pounds) Height (inches) Hip Circumference (cm) Waist Circumference (cm) Body Fat Percentage Sagittal Abdominal Di ameter
65 BIRTHDATE (mo/day/yr): Year in School (circle) First Second Third Fourth Major: Race: (Please Circle) American Indian or Alaska Native Asian Black or African American Native Hawaiian or Other Pacific Islander White Ethnicit y: (Please Circle) Hispanic or Latino Not Hispanic or Latino Are you currently pregnant or lactating? A certain amount of body fat is absolutely necessary for good health. Fat plays an important role in protecting internal organ s, providing energy, and regulating hormones. The minimal amount 15% for women. If too much accumulates over time, health may be compromised. 1. Have you ever had your body fat percentage measured by a trained professio nal? yes no 2. What is your current body fat percentage range (circle one)? <15% 15 18% 18 22% 22 30% 30 40% >40% 3. Which of the following best describes your current body fat percen tage (circle one)? Risky (Too Low) Ultra Lean Lean Moderately Lean Excess Fat Risky (Too High) 4. What is your current Height? __________ 5. What is your Current weight? __________ 6. Do you know what body mass index is? Yes No 7. Which of the following best describes your current weight (circle one)? underweight normal weight overweight obese ID______________ DATA FORM FOR WOMEN
66 Data collector initials ____________________ Weight (pound s) Height (inches) Hip Circumference (cm) Waist circumference (cm) Body Fat Percentage Sagittal Abdominal Diameter
67 LIST OF REFERENCES 1. ved April 23, 2012, from: http://www.cdc.gov/nchs/data/databriefs/db82.pdf 2. Calle, E. et al. Overweight, obesity and mortality from cancer in a prospectively studied cohort of U.S. adults. New England Journal of Medicine 2003;348(17):1625 38. 3. Berenson G. Cardiovascular risk factors in children: the early natural history of atherosclerosis and essential hypertension. New York: Oxford University Press, 1980. 4. Center for Disease Prevention. 29, 2012 5. Finkelstein E, et al. Economic causes and consequences of obesity. Annual Review of Public Health 2005; 26:239 45. 6. Bungum T, et al. The relationship of body mass index, medical cost, and job absenteeism. American Journal of Health Behavior 2003; 27(4):456 62. 7. Suminski R, et al. Physical activity among ethnically diverse college students. Journal of American College Health 2002; 51(2) :75 80. 8. Arnett J. Young people's conceptions of the transition to adulthood. Youth and Society 1997; 29(3):3 23. 9. Arnett J. Conceptions of the transition to adulthood: perspectives from adolescence through midlife. Journal of Adult Development 2001; 8(2): 133 43. 10. Bell S, Lee C. Does timing and sequencing of transitions to adulthood make a difference? Stress, smoking, and physical activity among young Australian women. International Journal of Behavioral Medicine 2006;13(3):265 74 11. Brown W, Trost S. Life tr ansitions and changing physical activity patterns in young women. American Journal of Preventative Medicine 2003; 25(2). 12. Kelly Weeder S. Binge drinking and disordered eating in college students. Journal of the American Academy of Nurse Practitioners 2011; 23(1):33 41. 13. Lewis C, et al. Seven year trends in body weight and associations with lifestyle and behavioral characteristics in black and white young adults: The CARDIA study. American Journal of Public Health 1997; 87(4):635 42.
68 14. Gordon Larsen P. Longit udinal trends in obesity in the U.S. from adolescence to the third decade of life. Obesity 2010; 18(9):1801 4. 15. Dervaux N, et al. Comparative associations of adiposity measures with cardiometabolic risk burden in asymptomatic subjects. Atherosclerosis 2008 ; 201(2):413 7. 16. Katzmarzyk P, et al. Fitness, fatness, and estimated coronary heart disease risk: the Heritage Family Study. Medicine and Science in Sports Exercise 2001; 33(4):585 90. 17. Malnick S, Knobler H. The medical complications of obesity. QJM 2006; 99(9):565 79. 18. Heitmann B, et al. Mortality associated with body fat, fat free mass and body mass index among 60 year old men a 22 year follow up. The study of men born in 1913. International Journal of Obesity and Related Metabolic Disorders 2000; 24(1): 33 7. 19. Ferreria I, et al. Development of fatness, fitness and lifestyle from adolescence to the age of 36 years: determinants of the metabolic syndrome in young adults: the Amsterdam Growth and Health Longitudinal Study. Archives of Internal Medicine 2005; 165(1):42 8. 20. Mattsson N, et al. The prevalence of the metabolic syndrome in young adults. The Cardiovascular Risk in Young Finns Study. Journal of Internal Medicine 2007; 261(2):159 69. 21. Metabolic syndrome. ADAM medical encyclopedia 22. Ellis K. Human body composition: in vivo methods. Physiological Reviews 2000; 80(2):649 80. 23. Cornier M, et al. Assessing adiposity: a scientific statement from the american heart association. Circulation 2011; 124. doi: 10.1161/CIR.0b013e318233. 24. Genton L, et al. Dual energy x ray absorptiometry and body composition: differences between devices and comparison with reference methods. Nutrition 2002; 18(1):66 70. 25. Plank L. Dual x ray absorptiometry and body composition. Current Opinion in Clinical Nutrition and Metabolic Care 20 05; 8(3):305 9. 26. Fields D, Goran M, McCrory M. Body composition assessment via air displacement plethysmography in adults and children: a review. American Journal of Clinical Nutrition 2002; 75(3):453 67.
69 27. Mitsiopoulos N, et al. Cadaver validation of skelet al muscle measurement by magnetic resonance imaging and computerized. Journal of Applied Physiology 1998; 85(1):305 9. 28. Lee S, Gallagher D. Assessment methods in human body composition. Current Opinion in Clinical Nutrition and Metabolic Care 2008; 11(5):5 66 72. 29. Koning L, et al. Waist circumference and waist to hip ratio as predictors of cardiovascular events; meta regression analysis of prospective studies. European Heart Journal 2007; 28(7):850 6. 30. WHO. Waist circumference and waist hip ratio: report of a WHO expert consultation. Geneva: World Health Organization, 2008. 31. Lear S, et al. Appropriateness of waist circumference and waist to hip ratio cutoffs for different ethnic groups. European Journal of Clinical Nutrition 2010; 64(42 61). 32. Pischon T, et al General and abdominal adiposity and risk of death in europe. The New England Journal of Medicine 2008; 359:2105 20. 33. Martorell R, et al. Body proportions in three ethnic groups: children and youths 2 17 years in NHANES II and HHANES. Human Biology 1988; 60(2):205 22. 34. Ford E, Mokdad A, Giles W. Trends in waist circumference among U.S. adults. Obesity Research 2003; 11(10):1223 31. 35. Mason C, Katzmarzyk P. Effect of the site of measurement of waist circumference on the prevalence of the metabolic syndrome. American Journal of cardiology 2009; 103(12):1716 20. 36. Ross R, et al. Does the relationship between waist circumference, morbidity and mortality depend on measurement protocol for waist circumference? Obesity Review 2008; 9(4):312 25. 37. Velthuis MJ, et al. Exercise program affects body composition but not weight in postmenopausal women. Menopause 2009; 16(4): 777 84. 38. Meisinger C, et al. Body fat distribution and the risk of type 2 diabetes in the general population: are there differences between men and wom en? The MONICA/KORA Augsburg cohort study. American Journal of Clinical Nutrition 2006; 84(3):483 9. 39. Seidell J. waist circumference and waist/hip ratio in relation to all cause mortality, cancer, and sleep apnea. European Journal of Clinical Nutrition 200 8; 64:35 41.
70 40. Rodriguez G, Moreno L. Body fat measurement in adolescents: comparison of skinfold thickness equations with dual energy x ray absorptiometry. European Journal of Clinical Nutrition 2005; 59:1158 66. 41. Zillikens M, Conway J. Anthropometry in bl acks: applicability of generalized skinfold equations and differences in fat patterning between blacks and whites. American Journal of Clinical Nutrition 1990; 52(1):45 51. 42. Lohman T, et al. Current issues in exercise sciences series (Monograph 3). Champai gn, IL: Human Kinetics, 1992. 43. Jackson A, Pollock M. Practical assessment of body composition. The Physician and Sports medicine 1985; 13:76 90. 44. Kyle U, et al. Bioelectrical impedance analysis part 1: review of principles and methods. Clinical Nutrition 2 004; 23(5):1226 43 45. Pouliot M, et al. Waist circumference and abdominal sagittal diameter: best simple anthropomorphic indexes of abdominal visceral adipose tissue accumulation and related cardiovascular risk in men and women. American Journal of Cardiolog y 1994; 73(7):460 8. 46. Iribarren C, et al. Value of the sagittal abdominal diameter in coronary heart disease risk assessment: cohort study in a large, multiethnic population. American Journal of Epidemiology 2006; 164(12):1150 9. 47. Riserus U, et al. Sagitta l abdominal diameter as a screening tool in clinical research: cutoffs for cardiometabolic risk. Journal of Obesity 2010; 2010. doi: 10.1155/2010/757939. 48. Hoenig M. MRI sagittal abdominal diameter is a stronger predictor of metabolic syndrome than visceral fat area or waist circumference in a high risk vascular cohort. Vascular Health and Risk Management 2010; 6:629 33. 49. Nakata K, et al. Stronger associations of sagittal abdominal diameter with atherogenic lipoprotein than waist circumference in middle aged U.S. white and Japanese men. Metabolism 2010; 59(12):1742 51. 50. Sampaio L, et al. Validity and reliability of the sagittal abdominal diameter as a predictor of visceral abdominal fat. Arquivos Brasileiros de Endocrinologia e Metabologia 2007; 51(6):980 6. 51. Ohrvall M, Berglund L, Vessby B. Sagittal abdominal diameter compared with other anthropometric measurements in relation to cardiovascular risk. International Journal of Obesity and Related Metabolic Disorders 2000; 24(4):497 501.
71 52. Hwu C, et al. Sagittal abdominal diameter is associated with insulin sensitivity in Chinese hypertensive patients and their siblings. Journal of Human Hypertension 2003; 17(3):193 8. 53. Onat A, et al. Measures of abdominal obesity assessed for visceral adiposity and relation to cor onary risk. International Journal of Obesity and Related Metabolic Disorders 2004; 28(8):1018 25. 54. Mukkuddem Peterson J, et al. Sagittal abdominal diameter: no advantage compared with other anthropometric measures as a correlate of components of the metabo lic syndrome in elderly from the Hoorn Study. American Journal of Clinical Nutrition 2006; 84(5):995 1002. 55. Gustat J, et al. Relation of abdominal height to cardiovascular risk factors in young adults. American Journal of Epidemiology 2000; 151(9):885 91. 56. Ekonoyan G. Adolphe Quetelet (1796 1874) the average man and indices of obesity. Nephrology Dialysis Transplantation 2008; 23(1):47 51. 57. Keys A, et al. Indices of relative weight and obesity. Journal of Chronic Disease 1971; 25: 329 43. 58. National Heart, L ung, and Blood Institute. First federal obesity clinical guidelines released. In: Health NIo, ed. Bethesda, MD: National Institutes of Health, 1998 59. National Heart L, and Blood Institute. The practical guide: identification, evaluation, and treatment of ov erweight and obesity in adults. Bethesda, MD: National Institutes of Health, 1998. 60. Bergman R, et al. A better index of body adiposity. Obesity 2011; 19(5):1083 9. 61. Lopez Jimenez F, et a. Diagnostic performance of body mass index to identify obesity as de fined by body adiposity: a systematic review and meta analysis. International Journal of Obesity 2010; 34(5):791 9. 62. Green D. Is body mass index really the best measure of obesity in individuals? author reply. Journal of the American College of Cardiology 2009; 53(6):527 8. 63. Gomez Ambosi J, et al. Clinical usefulness of a new equation for estimating body fat. Diabetes Care 2012; 35(2): 383 88. 64. Li C, et al. Estimates of body composition with dual energy x ray absorptiometry in adults. American Journal of C linical Nutrition 2009; 90(6):1457 65. 65. Heymsfield SB, et al. How useful is the body mass index for comparison of body fatness across age, sex, and ethnic groups? American Journal of Epidemiology 1996; 143: 228 39.
72 66. Gallagher D, et al. Healthy percentage body fat ranges: an approach for developing guidelines based on body mass index 2000; The American Journal of Clinical Nutrition. 72:694 701 67. Heymsfield SB, et al. Percentage of body fat cutoffs by sex, age, and race ethnicity in the US adult population from NHANES 1999 2004. 2012 American Journal of Clinical Nutrition. 95:594 602 68. Ode JJ, et al. Body mass index as a predictor of percent fat in college athletes and non athletes 2007: Medicine Science and Sports Exercise. 39:403 9. 69. Roche E, et al. Correlation between body mass index and body composition in elite athletes 2009 Journal of Sports Medicine and Physical Fitness. 49:278 84. 70. Bell JD, et al. The missing risk: MRI and MRS phenotyping of abdominal adiposity and ectopic fat 2012 Obe sity 20(1), 76 87. doi:10.1038/oby.2011.142 71. Di Renzo L, et al. Normal weight obese syndrome: early inflammation? 2007 American Journal of Clinical Nutrition. 85:40 45. 72. Vinknes K, et al. Evaluation of the body adiposity index in a Caucasian population: th e hordaland health study. American Journal of Epidemeiology 2013;177(5). doi: 10.1093/aje/kws271. 73. Lopez A, et al. Body adiposity index utilization in a spanish mediterranean population: comparison with the body mass index. Public Library of Science 2012; 7(4). doi: 10.1371/journal.pone.0035281. 74. Johnson W, et al. Concordance of the recently published body adiposity index with measured body fat percentage in european american adults. Obesity 2012; 20(4):900 3 75. Barreira V, et al. Research letter: Body adipos ity index, body mass index, and body fat in white and black adults. Journal of the American Medical Association 2011;306(8):828 30 76. Freedman. et al. The body adiposity index (hip circumference/ height (1.5) 18) is not a more accurate measure of adiposity than BMI, waist circumference, or hip circumference. Obesity 2012; 20(12):2438 44. 77. Esco, M. The accuracy of the body adiposity index for predicting body fat percentage in collegiate female athletes. Journal of Strength and Conditioning Published Ahead of Print 2012. doi: 10.1519?JSC.0b013e3182712714.
73 BIOGRAPHICAL SKETCH Blake Bartholomew received his Bachelor of Science in food science and human n utrition from the University of Florida in the spring of 2010. After graduating Blake taught at Vero Beach Hi gh School for one year In the sp ring of 2013 he received his Master of Science in human n utrition from the University of Florida and hopes to become a Registered Dietician.