the diabetes alert study: growth, fatness, and fat patterning, adolescence through adulthood in...

AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 64:389-399 (1984)

The Diabetes Alert Study: Growth, Fatness, and Fat Patterning, Adolescence Through Adulthood in Mexican Americans

WILLIAM H. MUELLER, SANDRA K. JOOS, CRAIG L. HANIS, ANTHONY N. ZAVALETA, JUNE EICHNER, AND WILLIAM J. SCHULL The University of Texas School of Public Health N H M . , S.K.J., J.E.) and The Uniuersity of Texas Graduate School of Biomedical Sciences, Center for Demographic and Population Genetics (C.L. H., W J.S), Houston, Texas 77025 and South Texas Institute of Latin and Mexican American Research, Texas Southmost College, Brownsville, Texas 78520 (A. N.Z.)

KEY WORDS Americans, Anthropometry

Growth, Diabetes, Fat patterning, Obesity, Mexican

ABSTRACT Diabetes Alert is a multidisciplinary genetic and epidemiological study of Type I1 (non insulin-dependent) diabetes in Texas Mexican Americans. We report the anthropometry of 1,155 individuals 10 to 70 or more years with particular reference to overweight, fatness, and anatomical fat patterning in the sample. Children ages 10-18 of both sexes are growing at the 50th percentile of the reference data (U.S. Health and Nutrition Examination Survey-1) for height, Wt/Ht2, and triceps and subscapular skinfolds. Adults are well below median height but well above median Wt/Ht2 and skinfolds. Preva- lence of obesity (Wt/Ht2 2 30) among adults is typically 30% or higher by age 30. Diabetics compared to agehex-matched non diabetics have shorter sitting heights, have more upper body fat (subscapular skinfold), have less lower body fat (lateral calf skinfold), and were heavier a t maximum weight and at age 18. The ratio of lower to upper body fat distribution decreases over the life cycle, being highest a t adolescence and lowest a t ages 40-50 in both sexes. Our results show a precipitous weight gain after maturity in the sample and an association of diabetes with differences in anatomical fat patterning. The age- related changes in fat patterning need to be explained in terms of their ecological and genetic influences.

Diabetes Alert is a multidisciplinary genetic and epidemiologic study of Type I1 (non insulin-dependent) diabetes in Texas Mexi- can Americans. The research has two phases. In the first, which is now complete, a ran- domly selected sample of approximately 10% of a Texas county was taken to determine the prevalence of diabetes (Hanis et al., 1983). In the second, diabetic probands identified in the first phase and their extended families were studied for their social, demographic, medical, genetic, and physical characteris- tics. This second phase is currently on-going, and this is a report concerning 1,155 subjects studied up to January 1983. We report the anthropometry of individuals 10 to 70 or more years with particular reference to over

weight, fatness, and anatomical fat patterning in the sample. We compare diabetics and their non diabetic relatives in an effort to confirm or reject the hypothesis that Type I1 diabetes is associated with concentration of fatness on the upper part of the body (Vague, 1956; Feldman et al., 1969; Kissebah et al., 1982; Szathmary and Holt, 1983). Then we construct an index of fat patterning by tech- niques developed earlier to identify periods in the life cycle of rapid development of fat toward the central and upper body and away from the extremities (Mueller and Reid, 1979;

Received November 28,1983; revised March 19,1984; accepted March 20, 1984.

0 1984 ALAN R. LISS, INC

390 W.H. MUELLER ET AL.

Ramirez and Mueller, 1980; Mueller and Stallones, 1981; Mueller and Wohlleb, 1981; Mueller, 1982; Malina et al., 1982; Stallones et al., 1982; Mueller et al., 1982; Mueller 1983).

MATERIALS AND METHODS Study population

The sample of diabetic probands was ob- tained by random sampling of three major population centers in Starr County, Texas. This county is about 95% Mexican American in ethnic origin. The county is essentially rural and most inhabitants are employed in agricultural labor and/or work in the oil fields in the county. Details of the study sample and the prevalence of Type I1 diabetes (10-15% among those over age 30), can be found in Hanis et al. (1983).

The sample The sample consists of 1,175 individuals

examined for the study up to January 1983. Twenty individuals were excluded from the analysis due to abnormalities (e.g., Down syndrome or other congenital problems), missing records, or pregnancy. The sample analyzed thus consists of 1,155 individuals: 141 preadult males (<20 years), 296 adult males, 187 preadult females (< 20 yearsland 531 adult females (see Table 1).

A nthropometry Fifteen body measurements were taken fol-

lowing standard guidelines (Weiner and Lourie, 1969): weight with a Detect0 beam- balance scale; height with a wall-mounted GPM anthropometer, the subject stretched to maximum height; sitting height with a GPM anthropometer, the subject with back straight and stretched up to maximum height; maximum calf, mid-arm and abdomen (navel) circumferences with a steel tape;

triceps, subscapular, waist, and medial and lateral calf skinfolds with a Lange caliper; and biepicondylar, biacromial, bicristal, and bitrochanteric breadths of the body with a sliding caliper and a GPM anthropometer. Firm pressure was applied in taking the breadths to reduce inclusion of soft tissue in the measurement. Subjects wore a light cloth smock over underwear. The smock was opened to permit measurement of skinfolds. Most measurements were taken by a single observer (W.H.M.) with the assistance of S.K.J. Two other observers were occasionally involved (A.N.Z. and J.E.). With the height and sitting height measurements, S.K.J. low- ered the bar of the anthropometers and read the measurement while W.H.M. positioned and stretched subjects upward. For the skinfold measurements S.K.J. took up the pinch of subcutaneous fat using both hands, one pinch 1 cm above, the other pinch 1 cm below the area, to which the caliper was applied. It was felt that this method would assure get- ting a more parallel fold of skin and subcutaneous fat, especially on obese subjects. The data were screened for outliers and mistakes corrected or data set to missing where appro- priate. Weight history data were also taken: the subjects were asked to recall the most ever weighed (maximum weight), the age at which it occurred, their weight at age 18, and the age at which weight gain began to occur (most adult subjects gained weight from age 18 onward).

Analyses Means of height and medians of relative

weight (body mass index: Wt/ht2 in kg/m2) and subscapular and triceps skinfolds were calculated and compared to the loth, 50th, and 90th percentiles of the Health and Nu- trition Examination Survey (HANES) of the United States (Hamill et al., 1977; Abraham

TABLE 1. Age and sex distribution of the 1,155 individuals studied

Age Males Females Age Males Females group br) N N group (yr) N N

9 or less 14 7 20-24 32 66 10 13 19 25-29 25 39 11 15 16 30-34 34 67 12 14 11 35-39 29 59 13 19 18 40-44 23 51 14 10 27 45-49 28 55 15 9 20 50-54 26 57 16 16 27 55-59 29 49 17 15 16 60-64 22 33 18-19 16 26 65-69 26 25

70 + 22 30

ANTHROPOMETRY AND DIABETES IN MEXICAN AMERICANS 391

et al., 1981; Cronk and Roche, 1982). We also calculated the prevalence of obesity in the sample by age group and sex, setting a body mass index of 30 or above as “obese.” This is a functional rather than a statistical definition of obesity as mortality increases significantly in adults with a body mass index of 30 or greater (Select Committee, 1977).

Comparison of diabetics and nondiabetics was made by creating a separate file in which each diabetic was matched with a nondiabetic of the same age and sex. This was necessary as most diabetics in the sample are over age 40, and we did not want to confound effects of age and diabetic status. Those above age 70 were excluded as most were diabetic, but this diabetes may be physiologically different from that which is so prevalent in middle age. Thus, we constructed a file of 336 subjects 30-69 years. Diabetes was de- fined following guidelines set forth by the national Diabetes Data Group: diabetics were individuals who were being treated for diabetes or had high fasting capillary whole blood glucose levels on one or more occasions and an abnormal glucose tolerance test (zero hour > 140 or 2 hours > 200). Nondiabetics were those with no diabetic history and with fasting whole blood glucose levels of 120 mgl dl or less. This conservative definition of non- diabetes was adopted to reduce the likelihood of including prediabetics or those with impaired glucose tolerance in the nondiabetic category.

Univariate one-way analyses of variance between groups were carried out for all anthropometric data. The median relative fat pattern, the graphic approach used by Garn (1955), was plotted in the two groups. The statistical significance of fat pattern differences between diabetics and nondiabetics was tested by discriminant function analysis. Three other discriminant functions were computed to discriminate diabetic status based on other measurements, all attempts to get at androgyny, body shape, or fat patterning differences between diabetics and nondiabetics: 1) arm, calf, and abdomen circumferences; 2) biacromial, bicristal, and bitrochanteric breadths; and 3) four skinfolds (excluding lateral calf because the latter was often found too tight to be measured). Lastly, from the discriminant function analysis for the five skinfolds, we calculated the median discriminant function value by age, sex, and diabetic status group in order to identify the age changes in the anatomical distribution of fat that occur over the life cycle.

RESULTS Age trends in growth and fatness

Figure 1 shows the average height in the sample compared to HANES data. Individu- als of both sexes 15 years and younger are close to the 50th percentile. Older teens tend to be shorter than average, and the medians of all adult groups are well below median U.S. height.

Figure 2 gives the median body mass index by sedage groups. Individuals 19 years or less of both sexes have median indices close to the reference median. In contrast all adult medians, except the 70 years or older group, are well above the reference median. This trend occurs somewhat earlier in males than females.

Figure 3 gives the median subscapular skinfold in males and the median triceps skinfold in females by age group. These two skinfolds are the most valid subcutaneous indicators of body fatness in males and females, respectively (Cronk and Roche, 1982; Roche et al., 1982). They confirm the trends for the body mass index seen in Figure 2.

Prevalence of extreme overweight in adults Table 2 presents the percentage of individ-

uals in the sample who are obese by the criteria Body Mass Index (BMD greater than or equal to 30. by age 25 years a body mass index of 30 corresponds to about the 90th percentile of the reference distribution (Fig. 2). Thus, we would expect about 10% of the adults to be above this value. In fact, by age 20 the percentage of individuals above a BMI of 30 ranges from 22 to 53% in females and from 27 to 40% in males. There is thus a high prevalence of overweight in the sample.

Comparisons of diabetics and nondiabetics Univariate analyses of variance of anthro-

pometric variables are shown in Table 3. Di- abetics compared to nondiabetics are shorter (sitting height), have more upper (subscapular) fat, and have less lower extremity fat (lateral calf skinfold). The weight history data show that diabetics were significantly heavier than nondiabetics at maximum weight and at age 18. The anthropometric differences between groups are generally more pronounced in females than males, but trends are similar.

Figure 4 shows the median relative fat pattern in the two groups. Diabetics are different from nondiabetics in the relative distribution of body fat but not in the amount of subcutaneous fat. The discriminant func-

392 W.H. MUELLER ET AL

Fig. 1. Mean heights in the sample by sex and age groups compared to HANES percentiles. Data from Hamill et al. (1977) and Abraham et al. (1981).


0 0

0

10 1

I ' " - 1 " I I I , I * l . l . 1 10 15 19 22 32 42 52 62 70+

AGE (In yeen) (Log Xale)

Fig. 2. Median body mass index (Wt/ht2) in the sample by sex and age groups compared to HANES percentiles. Data from Cronk and Roche (1982).

394 W.H. MUELLER ET AL

SUBSCAPULAR SKINFOLD MEDIANS -

1 / 0

35

30

25

20

15

10

5

0

0

50,

0 0

0 0

0

0 0 0

0

I ' " ' 1 ' ' I ' I I I I I , 1 , 1 10 15 19 22 32 42 52 62 70+

AGE (In yearr) (Lea Scale)

Fig. 3. Median subscapular (male) and triceps (female) skinfolds in the sample compared to HANES percentiles. Data from Crank and Roche (1982).


TABLE 2. Percentage of obese individuals (body mass index 2 30)

Males Females Total Total

Age % N' % N'

15-19 12.5 56 13.5 89 20-29 28.1 57 23.8 105 30-39 39.7 63 22.2 126 40-49 37.3 51 38.7 106 50-59 36.4 55 52.8 106 60 + 27.1 70 27.3 88

'Denominator.

tion in both sexes was highly significant (P < 0.01) when five skinfolds were used. It was also significant when four skinfolds (excluding lateral calf) were used, but less so in females than when five skinfolds were used. Circumference and body breadth discriminant functions did significantly (P < 0.01) discriminate groups in females, but not in males.

The skinfold discriminant functions (D, for males and Df for females) are as follows: D, = +.35 (tricep) - 1.33 (subscap) + .60 (waist) - .24 (M calf) + .42 (L calf)-canonical correlation = 0.41 (P < 0.01) (N = 103); Df = -.03 (tricep) - .82 (subscap) + .08 (waist) - .16 (M calf) + .80 (L calf)-canonical correlation = 0.41 (P < 0.01) (N = 172). The canonical correlation indicates that about 16% of

the variation in the discriminant function is explained by diabetic status. Thus, the statistical significance of these functions is high but the discriminating power low.

Life cycle changes in fat patterning Using the above discriminant functions we

calculated the medians per age, sex, and diabetic status group from 10 to 69 years (Fig. 5). The resulting variable is based on skinfolds standardized on sex-specific means and standard deviations. It thus has an expected mean of zero and standard deviation of 1.0. The variable is best conceptualized as the ratio of lower to upper body subcutaneous fat, hence a high value should be associated with less diabetic risk, a low value with high risk. The 1ower:upper body fat ratio function decreases significantly (P < 0.01) over the life cycle. In this way it behaves like an extremity-trunk fat patterning principal component identified elsewhere (Mueller, 1982). In males these changes appear to start in adolescence and continue till 40 to 50 years of age. In females the changes seem to start after age 20. Diabetics are, of course, significantly lower than nondiabetics in both sexes.

DISCUSSION

Studies on the epidemiology of impaired glucose tolerance have focused on adults (Va- gue, 1956; Feldman et al., 1969; Kissebah et

TABLE 3. One-way analysis of variance testing for univariate differences between diabetics and age-matched nondiabetics (30-69 years).

Means and (standard errors)

Body measurement

Weight (kg) Body mass index (kg/m2) Maximum weight (kg) Weight at age 18 (kg) Height (cm) Sitting Ht (cm) Elbow breadth (cm) Biacromial br (cm) Bicristal br (cm) Bitrochanteric br (cm) Arm circumference (cm) Abdomen cir (cm) Calf cir (cm) Triceps skinfold (mm) Subsc. skinfold (mm) Waist skinfold (mm) M calf skinfold (mm) L calf skinfold (mm)

*Between-group difference, P < 0.05. **Between-group difference, P < 0.01. tvariances not homogenous, P < 0.05.

Males Diabetic Nondiabetic (N = 59) (N = 59)

83.0 (1.87) 29.2 (0.54) 94.7 (2.11) 67.2 (2.11)

168.3 (0.87) 89.7 (0.46)

7.5 (0.10) 40.2 (0.26) 30.4 (0.26) 32.6 (0.27) 34.1 (0.40)

101.8 (1.47) 37.2 (0.43) 15.2 (0.86) 24.7 (1.13) 25.7 (1.22)

82.2 (2.01) 28.6 (0.59) 86.6 (2.02)** 61.7 (1.50)*

169.8 (0.93) 91.0 (0.46)*

7.4 (0.12) 40.6 (0.27) 30.1 (0.28) 32.9 (0.25) 34.1 (0.45)

100.0 (1.57) 37.3 (0.39) 15.5 (0.99) 20.2 (1.12)** 27.5 (1.45) ~

10.0 (0.84) 10.2 (0.62)t 7.3 (0.36) 7.9 (0.36)

Females Diabetic Nondiabetic

(N = 109) (N = 109)

76.1 (1.46) 31.2 (0.55) 85.2 (1.68) 55.7 (1.20)

155.9 (0.64) 83.9 (0.37) 6.6 (0.06)

36.8 (0.21) 30.7 (0.21) 32.6 (0.22) 35.1 (0.45)

106.3 (1.19) 36.5 (0.34) 33.0 (0.85) 34.7 (1.09) 36.1 (0.98) 23.2 (0.82) 12.6 (0.60)

72.9 (1.45) 29.3 (0.56)* 76.3 (1.57)** 51.7 (0.95)**7

84.8 (0.29)*t 6.6 (0.09)t

36.4 (0.17) 30.2 (0.24) 33.0 (0.22) 33.6 (0.46)*

100.4 (1.17)** 37.2 (0.38) 31.3 (0.90) 28.9 (l.ll)** 33.1 (1.04)* 25.9 (1.03)*t 15.8 (0.76)**

157.3 (0.52)


+.5 1 MEDIAN RELATIVE FAT PATTERN IN Dlabetlcr (- - -) and Non Dlabstlcs( -)

MALES N=59

FEMALES c----\

- . 5 J - v) ti I- n a

Y 5: m 0 i- u 3

v) v)

U U J A

- a a 2 a a

c!

LL U J -1

rn ti I- a n 0

U m I I- 3

Lu s z s 0 0 $ c! w

SKINFOLD SITE

Fig. 4. Median relative fat patterns in diabetic and agekex-matched nondiabetic subjects.

al., 1982; Szathmary and Holt, 1983). Be- cause of its gentic epidemiological approach, the Diabetes Alert study is seeking the char- acteristics associated with diabetes in extended families. Because of the genetic relationship between adults and preadults in the Diabetes Alert study, we are able to look for patterns of growth, fatness, and fat patterning among adolescents and relate these to patterns of body composition later in the life cycle. Given a high percentage of overweight adults (Table 21, we would have expected to see a fair number of overweight adolescents. Yet, as is evident in Figures 2 and 3, the prevalence of overweight and fatness in this age group appears to be no greater than that found in the general population of the United States. This is not a cohort effect due to the cross-sectional nature of the data, because adults had reported body mass indices at age 18 that averaged close to the medians for HANES-1 seen in Figure 2 (Joos et al., 1984). The excessive fatness associated with adult onset or Type I1 diabetes appears to be acquired after the achievement of physical maturity, There is thus a high prevalence of adult weight gain in this sample. Adult weight gain has been implicated

in chronic diseases of the cardiovascular sys- tem (Abraham et al., 1971).

Diabetes Alert adolescents and their elders differ as well in linear growth. Youths appear to be growing on average at median height for the United States (Fig. 1). In contrast adults are markedly shorter. Ten years ago Brownsville (about 100 miles from Starr County) Mexican-American children were growing on average at about the 25th percentile of the United States Health Examination Survey reference values (Zavaleta and Mal- ina, 1980). We speculate the pattern of height growth in Starr County may reflect recent ecological changes toward positive energy balance in families in which late-onset diabetes has occurred. The children’s increased growth in stature reflects this secular change. This also suggests that poverty when followed by improved environmental circum- stances, may predispose small-framed adults to impaired glucose tolerance. This hypothesis is consistent with the observation in Ta- ble 3 that diabetic adults tended to be shorter than nondiabetic persons of similar age, al- though why this is reflected mainly in sitting height remains a mystery. There appears to be a mild secular increase in height of the


+ I

1 LOWER UPPER BODY FAT DISCRIMINANT FUNCTION

FEMALE MEDIANS ---- Dlrbrtlc - Non.Dlrbotlc

r

J . ' , I . . I I 10 15 18 22 32 42 52 82 70+

1 I I I I . l . 1

AQE On yearn) (Lag k d 0 )

Fig. 5. Median fat patterning index 0ower:upper body fat ratio) derived from discriminant function analysis of the data in Figure 4, by age, sex, and diabetic status.


Mexican-American population in Texas (Mal- ina and Zavaleta, 1980). In families of diabetics such a change may be more evident and may point out fundamental ecological factors a t work in producing this disease.

Vague (1956) first suggested that diabetes and other cardiovascular problems were associated with upper body fat distribution rather than obesity per se. Our study also confirms this. We have constructed an index of fat patterning by using discriminant function. Examination of the discriminant function coefficients associated with each of five skinfold sites, as well as examination of the median relative fat pattern in diabetics and age-matched nondiabetics in Figure 4, allows us to interpret the aspect of skinfold variation that distinguishes diabetics from nondiabetics. The function is somewhat different in males and females. In males it involves the contrast of subscapular with waist and leg fat. In females subscapular and waist sites together are contrasted with leg fat. It is notable that the widely used triceps skinfold offers little information concerning this aspect of fat patterning. Leg fat and possibly forearm fat may be the most important extremity sites for the description of individual differences in fat patterning (Mueller and Stallones, 1981; Szathmary and Holt, 1983).

The discriminant function for skinfolds is also somewhat different from the extremity- trunk index we have derived from principal components analysis (Mueller and Wohlleb, 1981). The discriminant function more clearly contrasts lower and upper body fat distribution. The principal component, however, seems not to vary by sex and involves a contrast of extremity sites, particularly those of the lower limb, with trunk sites, upper and lower trunk together. Hence, there may be other aspects of fat patterning in human pop- ulations besides the extremity-trunk component which we have emphasized previously as a major component of variance in fat patterning.

The discriminant function changes with age in a way consistent in both sexes: Adoles- cents have high ratios of lower to upper body fat; adults of ages 40 to 60 have the lowest ratios. These changes with age are quite large, about one standard deviation difference in the medians between adolescence and middle age. In males already there is a decline in the median index during adolescence; in females the ratio of lower to upper body fat appears to begin its decline after age

20. Diabetics of both sexes already have ratios significantly lower than their nondiabetic contemporaries. How did they get there? Were they born with genetically deter- mined low ratios? Were they average in their fat patterning scores at adolescence and did they then plunge a t a faster rate with adult onset of overweight? It has been suggested that upper body obesity may be more respon- sive to weight loss, because enlarged fat cells are characteristic of this type of obesity (Kis- sebah et al., 1982). The possibility follows that such an obesity is due to a weight gain after maturity. However, Joos et al. (1984) reported no association of fat patterning with history of weight gain (early versus late onset of overweight) in the sample, but a consistent association of fat patterning and diabetic status. If anything, those who were already overweight by age 18 had the more pronounced upper body fat patterning (among diabetics). Deutsch et al. (1984) report an association of overweight and upper body fat patterning in US. adolescents that in females is independent of physiological maturity. The evidence seems to point to overweight early in life as a correlate of upper body obesity. In this respect, it is noteworthy that diabetics were already heavier than nondiabetics by age 18 (Table 3). The genetic and ecological determinants of early fatness and of changes in fat patterning across the life cycle (Fig. 5) need to be sought in future studies on the origin of individual differences in the relative distribution of body fat.

ACKNOWLEDGMENTS

This research was supported by grant 5- R01-AM27582 from the National Institutes of Health. This paper was originally pre- sented at the annual meeting of the Ameri- can Association of Physical Anthropologists, in Indianapolis, Spring, 1983. The authors wish to thank Drs. Linda Adair and Francis Johnston, the organizers of the symposium of which this paper was a part. We are most grateful to Gay Robertson and the staff of the word processing center of the University of Texas School of Public Health for prepara- tion of the manuscript,

LITERATURE CITED

Abraham, S, Collins, G, and Nordsieck, F (1971) Rela- tionship of childhood weight status to morbidity in adults. HSMHA Health Rep. 86t273-284.

Abraham, S, Johnston, CL, and Najjar, MF (1981) Weight and height of adults 18-74 years of age, United States, 1971-1974. Vital and Health Statistics, Series 11, No,


211, Washington: U.S. Department of Health, Educa- ton and Welfare.

Cronk, CE, and Roche, AF (1982) Race- and sex-specific reference data for triceps and subscapular skinfolds and weight/stature2. Am. J. Clin. Nutr. 35:347-354.

Deutsch, MI, Mueller, WH, and Malina, RM (1984) Bio- logical maturity, adolescent fatness and relative fat patterning. (Submitted for publication).

Feldman, R, Sender, AJ, and Siegelaub, AB (1969). Dif- ference in diabetic and non-diabetic fat distribution patterns by skinfold measurements. Diabetes 18t478- 486.

Garn, SM (1955) Relative fat patterning, an individual characteristic. Hum. Biol. 27t75-89.

Hamill, PVV, Drizd, TA, Johnson, CL, Reed, RB, and Roche, AF (1977) NCHS growth curves for children birth to 18 years United States. Vital and Health Sta- tistics, Series ll, No. 165, Washington: U.S. Depart- ment of Health, Education and Welfare.

Hanis, CL, Ferrell, RE, Barton, SA, Aguilar, L, Garza- Ibarra, A, Tulloch, BR, Garcia, CA, and Schull, WJ (1983) Diabetes among Mexican-Americans in Texas. Am. J. Epidemiol. 118t659-672.

Joos, SK, Mueller, WH, Hanis, CL, and Schull, WJ (1984) Diabetes Alert study: Weight history and upper body obesity in diabetic and non-diabetic Mexican American adults. Ann. Hum. Biol. lltl67-171.

Kissebah, AH, Vydelingum, N, and Murray, R (1982) Relation of body fat distribution to metabolic compli- cations of obesity. J. Clin. Endocrinol Metabol. 54:254- 260.

Malina, RM, and Zavaleta, AN (1980) Secular trend in the stature and weight of Mexican American children in Texas between 1930 and 1970. Am. J. Phys. Anthro- pol. 52t453-461.

Malina, RM, Mueller, WH, Bouchard, C, Shoup, RF, and Lariviere, G (1982) Fatness and fat patterning among athletes at the Montreal Olympic Games, 1976. Med. Sci. Sports Exerc. 15t445-452.

Mueller, WH (1982) The changes with age of the anatomical distribution of fat. SOC. Sci. Med. 16: 191-196.

Mueller, WH (1983) The genetics of human fatness. Yrbk. Phys. Anthropol. 26:215-230.

Mueller, WH, and Reid, RM (1979) A multivariate analysis of fatness and relative fat patterning. Am. J. Phys. Anthropol. 50t199-208.

Mueller, WH, Shoup, RF, and Malina, RM (1982) Fat patterning in athletes in relation to ethnic origin and sport. Ann Hum Biol. 9:371-376

Mueller, WH, and Stallones, L (1981) Anatomical distribution of subcutaneous f a t Skinfold site choice and construction of indices. Hum. Biol. 53~321-335.

Mueller, WH, and Wohlleb, JC (1981) (1981) Anatomical distribution of subcutaneous fat and its description by multivariate methods: How valid are principal components? Am. J. Phys. Anthropol. 54t25-35.

Ramirez, ME, and Mueller, WH (1980) The development of obesity and fat patterning in Tokelau children. Hum. Biol. 52t675-687.

Roche, AF, Siervogel, RM, Chumlea, WC, and Webb, P (1982) Grading of body fatness from limited anthropometric data, Am. J. Clin. Nutr. 36:162-171.

Select Committee on Nutrition and Human Needs (1977) Diet Related to Killer Diseases, Part 11, Obesity. Wash- ington: US . Government Printing Offce, p. 11.

Stallones, L, Mueller, WH, and Christensen, BL (1982) Blood pressure, fatness, and fat patterning among USA adolescents from two ethnic groups. Hypertension, 4t483-486.

Szathmary, EJE, and Holt, N (1983) Hyperglycemia in Dogrib Indians of the Northwest Territories, Canada: Association with age and a centripetal distribution of body fat. Hum. Biol. 55t493-515.

Vague, J (1956) The degree of masculine differentiation of obesities: A factor determining predisposition to diabetes, atherosclerosis, gout and uric calculous disease. Am. J. Clin. Nutr. 4t20-34.

Weiner, JS, and Lourie, JA (1969) Human Biology: A Guide to Field Methods. Oxford Blackwell Scientific.

Zavaleta, AN, and Malina, RM (1980) Growth, fatness, and leanness in Mexican-American children. Am. J. Clin. Nutr. 33t2008-2020.

the diabetes alert study: growth, fatness, and fat patterning, adolescence through adulthood in...

Documents