Documented differences in measures of insulin secretion and action between African Americans and European Americans may be due to either genetic or environmental factors. This study used genetic admixture (ADM), determined from ∼20 ancestry informative markers, and a questionnaire designed to assess socioeconomic status (SES) to examine potential genetic and environmental contributions to minimal model–derived measures of insulin sensitivity (SI), fasting insulin, and the acute insulin response to glucose (AIRg) in 125 children residing in Birmingham, Alabama. The study was longitudinal in design and yielded multiple outcome measures on each subject. Mixed models analysis was used to determine if ADM and SES were independently related to SI, fasting insulin, and AIRg after adjusting for confounding factors (pubertal status, adiposity, age) and for repeated testing of individuals. In this cohort, African ADM ranged from 0% (individuals with no markers reflecting African ancestry) to 100% (individuals with all 20 markers reflecting African ancestry). Results indicated that ADM was independently related to SI (P < 0.001) and fasting insulin (P < 0.01), with individuals having greater African ADM having a lower SI and a higher fasting insulin concentration. Both ADM (P < 0.001) and SES (P < 0.05) were independently related to AIRg; children with greater African ADM or lower SES had a higher AIRg, even after adjusting for SI. These observations suggest that use of ADM can replace assignment of individuals to categorical racial groups; that lower SI and higher fasting insulin among African Americans compared with European Americans may have a genetic basis; and that higher AIRg among African Americans may be due to both genetic factors and to environmental factors that remain to be identified.

Differences in measures of insulin secretion and action between African Americans and European Americans have been identified in the literature (1–6). These differences were identified by using self-identified “race” as an independent variable in statistical models. However, because race is a combination of both genetic and environmental factors, the basis for racial differences in aspects of physiology and metabolism is not clear.

One way to investigate putative genetic factors contributing to racial differences in complex physiological traits such as insulin sensitivity (SI) is by use of genetic admixture (ADM) analysis. ADM analysis is an extension of the “Mapping by Admixture Linkage Disequilibrium” approach. In general terms, the admixture approach takes advantage of the allelic variation produced by the intermixing of long-separated parental populations. In the U.S., this process was initiated during the European colonial period, which was extended to the New World in 1492, and resulted in a number of admixed populations (African Americans, and Hispanics and Latinos of North and South America). The combination of allelic variants in an individual that corresponds to each parental population considered is translated into a quantitative value. The ADM value reflects a portion of the biological ancestral determinants that influence an individual’s racial identity. It can also be thought of as the estimated proportion of an individual’s genome that is of a given ancestral origin (e.g., African) based on the specific markers used for the analysis.

The ADM approach provides a useful means to avoid assignment of individuals of mixed ethnic descent into categorical racial groups. The ADM approach also has potential value in identifying genes that confer disease risk. If a given allelic variant is associated with altered risk for a given disease, then this variant can ultimately be used as a mapping tool to identify genes associated with physiological outcomes that affect disease risk. The application of the ADM approach to mapping genes associated with variation in complex traits has been discussed elsewhere (7–10).

In the present study, individual values of ADM were used to explore the extent to which ancestral biological determinants explained racial differences in insulin-related phenotypes previously reported in the literature (1–6). A quantitative value representing ADM was obtained after genotyping African-American and European- American children at ∼20 markers informative for parental ancestry. These ADM values were then examined for their association with measures of SI, fasting insulin concentration, and the acute insulin response to glucose (AIRg). The specific hypothesis tested was that ADM, which represents the genetic contribution of race, and socioeconomic status, an index of the environmental contribution of race, would explain unique portions of variance in SI, fasting insulin, and AIRg. This approach—i.e., quantitative use of ADM values—has not been used previously to examine racial/ethnic differences in aspects of insulin secretion and action.

Subjects.

Of the 125 children included in the study, 52 identified themselves as African American (23 boys; 29 girls), and 73 as European American (26 boys; 47 girls). Individuals were considered to be of a given ethnic/racial group if both parents and all grandparents were of the same ethnicity/race as the child. The age range of the subjects was 5–16 years. All subjects were recruited as part of a longitudinal study exploring the contribution of body fat distribution to disease risk. Recruitment of study subjects has been explained elsewhere (6). Only individuals for whom ADM could be determined (from whom genetic material was obtained) were included in the study. The study was in compliance with the regulations of the Institutional Review Board at the University of Alabama at Birmingham.

Study design.

The study design was longitudinal and observational. Subjects visited the University of Alabama at Birmingham for testing annually, at approximately the same time each year. Recruitment began in 1995 and continued until enough subjects were recruited. Thus, not all subjects joined the study during the same year; likewise, not all subjects completed all possible visits. Insulin sensitivity testing was incorporated into the protocol in 1996. Data used for the present study are from 1996 through 2001.

Protocol.

Subjects were admitted into the General Clinical Research Center for an overnight stay during which SI, fasting insulin, and AIRg were obtained during a tolbutamide-modified, frequently sampled intravenous glucose tolerance test, as previously described (6). SI was derived from minimal model analysis; AIRg was calculated as the incremental area under the curve for insulin during the first 10 min after glucose injection; and fasting insulin was determined as the average insulin concentration in three baseline (pre–glucose injection) blood draws. Total body fat was determined by dual-energy X-ray absorptiometry (Lunar DPX-L, software version 1.5e; Lunar Radiation Corp., Madison, WI); body composition was not available on two boys at Tanner stage I, one girl at Tanner stage II, and one boy at Tanner stage V. Socioeconomic status (SES) was measured with the Hollingshead four-factor index of social class (11), which combines the educational attainment and occupational prestige for the number of working parents in the child’s family. Scores ranged from 8 to 66, with the higher score indicating higher theoretical social status. SES information was not available on one subject. Tanner stage was determined through physical examination by a pediatrician. Stage I denotes prepubertal status, whereas stage V denotes young adult status; stages II–IV denote stages of pubertal transition. Total body fat and Tanner stage were included in all statistical models for SI because both have been shown to influence this outcome measure in this cohort (6,12).

Determination of genetic admixture.

Genotyping was performed at Pennsylvania State University. Molecular techniques used for the genotyping include the melting curve analysis of single nucleotide polymorphisms (McSNP) method described by Akey et al. (13) and agarose gel electrophoresis. Markers and techniques used for the identification of the ancestry-informative DNA sequences have previously been described by Parra et al. (8) and are available through dbSNP (http://www.ncbi.nlm.nih.gov/SNP/) using the handle PSA-ANTH. Markers used for this study, their chromosomal and centimorgan location, and their allelic difference between European and African parental populations are described in Table 1. All genotype information was translated into an individual value for ADM using a maximum likelihood method (14), where the likelihood that a particular multilocus genotype is from each of 100 different combinations of the two parental populations is calculated as the probability of the genotype given the allele frequencies in each parental population. The parental population combination where the specific multilocus genotype for an individual has the highest probability is thus the most likely and represents the ADM estimate for that person.

Statistical analyses.

Because multiple phenotypic measures were available for each individual due to repeated, annual testing, a mixed linear model (PROC MIXED) was used to examine the relationship of ADM with SI, fasting insulin, and AIRg. To improve normality of the distribution of phenotypic variables, log transformations were used on total body fat, SI, fasting insulin, and AIRg for all statistical analyses; thus, all results shown were derived from log-transformed data. Tanner stage was modeled as a categorical variable in the statistical models, with Tanner = 5 as the reference value. Because some children remained at a given Tanner stage throughout several consecutive years of testing, an “age-nested-in-Tanner stage” component was included in the models. Although there was a high degree of colinearity between the Tanner stage and age variables, use of the nested term was necessary to account for changes in the dependent variables that occurred with age, within a given Tanner stage (12).

Exploratory analyses, including a quadratic and linear component of Tanner stage, and the interaction of these variables with ADM, were used to identify the most parsimonious model explaining variation in the dependent variables (data not shown). For SI and fasting insulin, the model was defined by Tanner stage, age nested in Tanner stage, total fat, SES, and ADM. For AIRg, the model was defined by Tanner stage, age nested in Tanner stage, SI, SES, and ADM. Pearson correlation analysis was used to examine the association between ADM and SES. All statistical analyses were performed using SAS version 10.0 (SAS Institute, Cary, NC).

Descriptive statistics for the subjects by Tanner stage are shown in Table 2. Sample size is given for each combination of sex and Tanner stage and ranged from a high of 90 (boys, Tanner stage I) to a low of 7 (boys, Tanner stage V).

The distribution of African ADM was from 0 to 100%. Among children who identified themselves as European American, African ADM ranged from 0 to 21%, with the exception of one individual with a score of 61; among those who identified themselves as being African American, African ADM ranged from 50 to 100%. SES of the subjects ranged from 8 to 66. African admixture was significantly correlated with SES among African-American children (r = −0.134; P < 0.05) but not among Caucasian children (r = −0.08; P = 0.220).

In exploratory analyses for the dependent variable SI, the quadratic term for Tanner stage was not significant. This result indicated that the decrease and subsequent rebound in SI that occurs during puberty (15) was not statistically detectable in these data, perhaps due to the relatively small number of subjects at Tanner stages IV and V. Thus, only the linear component of Tanner stage was used in subsequent models. The independent effect of sex also was examined in exploratory analysis. Because sex was not significant in the model for SI (P = 0.126), AIRg (P = 0.365), or fasting insulin (P = 0.339; data not shown), it was not evaluated in the analyses.

Table 3 shows the results of the statistical model with SI as the dependent variable, as predicted by Tanner stage, age nested in Tanner stage, total fat, SES, and ADM. ADM was independently related to SI (P < 0.001). SES was not uniquely related to SI (P = 0.052).

Table 4 shows the results of the model for fasting insulin, as predicted by Tanner stage, age nested in Tanner stage, total fat, SES, and ADM. ADM was independently related to fasting insulin (P < 0.01); SES was not (P = 0.074).

Table 5 shows the results of the model for AIRg, where Tanner stage, age nested in Tanner stage, SI, SES, and ADM were used as predicting variables. Both ADM (P < 0.001) and SES (P < 0.05) were uniquely related to AIRg.

In the models for both SI and AIRg, the overall effect of Tanner stage was not significant, but the overall effect of age nested in Tanner stage was significant (P < 0.05).

The results of this study demonstrated that ADM explained a significant portion of the variance in SI, fasting insulin, and AIRg, whereas SES was independently related to AIRg. These observations suggest that use of ADM can replace assignment of individuals to categorical racial groups; that lower SI and higher fasting insulin among African Americans compared with European Americans may have a genetic basis; and that higher AIRg among African Americans may be due to both genetic and environmental factors.

As reported in cross-sectional observations (6), body fat was a primary determinant of SI, with more obese individuals having lower SI. In contrast to reports from cross-sectional studies (12), Tanner stage was not independently related to SI. This discrepancy may relate to the relatively small number of subjects available at the higher Tanner stages, and to the high correlation between Tanner stage and age. However, age nested within Tanner stage was significantly related to SI, indicating that within Tanner stages I and III, SI increased with age (12). Likewise, age nested within Tanner stage was uniquely and inversely related to AIRg at Tanner stages I through III, perhaps reflecting an age-related decrease in β-cell function (12). The lack of significance of age nested within Tanner stage at Tanner stages IV and V may have been due to the relatively low sample size in these groups, or to the collinearity between Tanner stage and age; further study in children of higher pubertal stages will be needed to clarify this issue.

The model for AIRg contained SI as an independent variable. Thus, variables in the AIRg model that emerged as significant reflect a contribution to AIRg that is in addition to compensatory pancreatic phenomena (16). It has previously been reported that, at least in children, the disproportionately greater AIRg among African Americans results from a combination of greater first-phase insulin secretion and lower hepatic insulin extraction (4). In the present study, the significant contribution of genetic admixture to AIRg, independent of SI and SES, suggests that genetic factors may influence pancreatic function, the ability of the liver to clear insulin, or both. Furthermore, results indicated that individuals with lower SES had relatively higher AIRg, independent of SI. Neither dietary factors (17) nor physical activity (18), both of which are highly associated with measures of SES, explained racial differences in AIRg in cross-sectional analyses of this cohort during the subjects’ prepubertal and early pubertal years. However, due to the limitations of lifestyle measures obtained by questionnaire, the possible contribution of diet and physical activity, as well as other components of SES, to AIRg remains an open question.

In the model for fasting insulin, as in that for SI, ADM explained a unique portion of the variance. This result could reflect the well-documented inverse association between SI and fasting insulin (16). Age nested in Tanner stage showed a significant inverse relationship at Tanner stage I, indicating that fasting insulin declined with age in prepubertal children. This observation agrees with the observed increase in SI with age during Tanner stage I (Table 3).

Limitations of this study include the small number of individuals with African ADM in the central range of the distribution; the small number of children at Tanner stages IV and V; and the use of ∼20 molecular markers to characterize ADM. Use of additional markers, when available, will likely improve the genetic characterization of individuals, and increase the likelihood of successfully mapping genes associated with SI or AIRg.

In conclusion, results suggest that use of ADM can replace assignment of individuals to categorical racial groups. A further benefit of using ADM is the likelihood that, because it has more variability than a dichotomous race variable, it will explain more variance in SI, AIRg, and other physiological variables (19,20). The independent association of ADM with SI and AIRg indicated that lower SI and higher AIRg among African Americans may have a genetic basis. Higher AIRg among African Americans may be due to social/environmental factors; identifying these factors will be an important step toward understanding and preventing racial and ethnic disparities in type 2 diabetes and cardiovascular disease.

TABLE 1

Markers used for this study, their chromosomal and centimorgan location, and their allelic difference (frequency in population 1 minus frequency in population 2) between European and African parental populations.

MarkerLocationcMAfrican vs. European
MID 575 1p34.3 ∼64 0.130 
MID 187 1p34.1 ∼75 0.370 
FY-null 1q23.2 ∼165 0.999 
AT3 1q25.1 ∼191 0.575 
WI-11392 1q42.2 ∼252 0.444 
WI-16857 2p16.1 ∼79 0.536 
WI-11153 3p12.3 ∼106 0.652 
GC*1F 4q13.3 79 0.697 
GC*1S 4q13.3 79 0.538 
SGC30055 5q23.1 ∼120 0.457 
CYP3A4 7q22.1 ∼111 0.761 
LPL 8p21.3 ∼39 0.479 
D11S429 11q11 ∼70.9 0.429 
DRD2-Taq I “D” 11q23.1 ∼105 0.535 
APOA1 11q23.3 ∼113 0.505 
GNB3 12p13.31 ∼15 0.463 
OCA2 15q13.1 ∼16 0.631 
MC1R314 16q24.3 ∼133 0.350 
WI-14867 17p13.2 ∼10 0.448 
WI-7423 17p12 ∼16 0.476 
Sb19.3 19p13.11 ∼49 0.488 
MID 154 20q11.22 ∼50 0.444 
MarkerLocationcMAfrican vs. European
MID 575 1p34.3 ∼64 0.130 
MID 187 1p34.1 ∼75 0.370 
FY-null 1q23.2 ∼165 0.999 
AT3 1q25.1 ∼191 0.575 
WI-11392 1q42.2 ∼252 0.444 
WI-16857 2p16.1 ∼79 0.536 
WI-11153 3p12.3 ∼106 0.652 
GC*1F 4q13.3 79 0.697 
GC*1S 4q13.3 79 0.538 
SGC30055 5q23.1 ∼120 0.457 
CYP3A4 7q22.1 ∼111 0.761 
LPL 8p21.3 ∼39 0.479 
D11S429 11q11 ∼70.9 0.429 
DRD2-Taq I “D” 11q23.1 ∼105 0.535 
APOA1 11q23.3 ∼113 0.505 
GNB3 12p13.31 ∼15 0.463 
OCA2 15q13.1 ∼16 0.631 
MC1R314 16q24.3 ∼133 0.350 
WI-14867 17p13.2 ∼10 0.448 
WI-7423 17p12 ∼16 0.476 
Sb19.3 19p13.11 ∼49 0.488 
MID 154 20q11.22 ∼50 0.444 
View Large
TABLE 2

Descriptive statistics by Tanner stage

VariableTanner ITanner IITanner IIITanner IVTanner V
Boys      
 n 90 52 19 10 7 
 Age (years) 9.3 ± 1.5 11.0 ± 1.3 12.0 ± 0.9 13.3 ± 1.0 14.1 ± 0.8 
 BMI (kg/m2) 19.3 ± 3.6 23.2 ± 6.4 25.4 ± 5.8 24.6 ± 5.6 24.5 ± 4.2 
 Total fat (kg) 9.2 ± 6.7* 16.8 ± 12.1 21.4 ± 14.6 17.4 ± 14.6 16.7 ± 12.5† 
 SI (× 10−5 · min−1 · pmol−1 · l−1) 11.8 ± 6.3 7.7 ± 5.7 5.3 ± 3.2 7.2 ± 5.7 8.5 ± 7.8 
 Fasting insulin (pmol/l) 67.2 ± 34.2 87.0 ± 48.0 106.8 ± 63.0 80.4 ± 16.8 59.4 ± 28.8 
 AIRg (pmol/l × 10 min) 6,174 ± 7,176 9,780 ± 8,460 16,050 ± 23,466 10,944 ± 12,324 7,314 ± 5,202 
Girls      
 n 73 67 62 30 18 
 Age (years) 9.2 ± 1.6 10.9 ± 1.3 12.3 ± 1.11 13.5 ± 1.1 14.3 ± 1.0 
 BMI (kg/m2) 19.0 ± 4.7 21.7 ± 5.8 22.0 ± 4.7 23.4 ± 5.6 25.5 ± 6.1 
 Total fat (kg) 10.1 ± 7.6 15.2 ± 10.9‡ 16.8 ± 10.1 20.2 ± 11.0 23.9 ± 11.9 
 SI (× 10−5 · min−1 · pmol−1 · l−1) 10.5 ± 7.2 7.5 ± 5.5 6.3 ± 4.5 6.8 ± 3.8 5.8 ± 3.5 
 Fasting insulin (pmol/l) 70.2 ± 40.2 97.2 ± 45.0 97.2 ± 52.8 88.2 ± 44.4 81.0 ± 33.6 
 AIRg (pmol/l × 10 min) 6,288 ± 4,524 6,948 ± 4,746 6,612 ± 4,908 7,110 ± 3,510 5,514 ± 3,132 
VariableTanner ITanner IITanner IIITanner IVTanner V
Boys      
 n 90 52 19 10 7 
 Age (years) 9.3 ± 1.5 11.0 ± 1.3 12.0 ± 0.9 13.3 ± 1.0 14.1 ± 0.8 
 BMI (kg/m2) 19.3 ± 3.6 23.2 ± 6.4 25.4 ± 5.8 24.6 ± 5.6 24.5 ± 4.2 
 Total fat (kg) 9.2 ± 6.7* 16.8 ± 12.1 21.4 ± 14.6 17.4 ± 14.6 16.7 ± 12.5† 
 SI (× 10−5 · min−1 · pmol−1 · l−1) 11.8 ± 6.3 7.7 ± 5.7 5.3 ± 3.2 7.2 ± 5.7 8.5 ± 7.8 
 Fasting insulin (pmol/l) 67.2 ± 34.2 87.0 ± 48.0 106.8 ± 63.0 80.4 ± 16.8 59.4 ± 28.8 
 AIRg (pmol/l × 10 min) 6,174 ± 7,176 9,780 ± 8,460 16,050 ± 23,466 10,944 ± 12,324 7,314 ± 5,202 
Girls      
 n 73 67 62 30 18 
 Age (years) 9.2 ± 1.6 10.9 ± 1.3 12.3 ± 1.11 13.5 ± 1.1 14.3 ± 1.0 
 BMI (kg/m2) 19.0 ± 4.7 21.7 ± 5.8 22.0 ± 4.7 23.4 ± 5.6 25.5 ± 6.1 
 Total fat (kg) 10.1 ± 7.6 15.2 ± 10.9‡ 16.8 ± 10.1 20.2 ± 11.0 23.9 ± 11.9 
 SI (× 10−5 · min−1 · pmol−1 · l−1) 10.5 ± 7.2 7.5 ± 5.5 6.3 ± 4.5 6.8 ± 3.8 5.8 ± 3.5 
 Fasting insulin (pmol/l) 70.2 ± 40.2 97.2 ± 45.0 97.2 ± 52.8 88.2 ± 44.4 81.0 ± 33.6 
 AIRg (pmol/l × 10 min) 6,288 ± 4,524 6,948 ± 4,746 6,612 ± 4,908 7,110 ± 3,510 5,514 ± 3,132 
*

n = 88;

†

n = 6;

‡

n = 66.

View Large
TABLE 3

Parameter estimates for mixed model with the dependent variable SI

EffectEstimateSEP
Intercept 5.4065 1.5183 0.0005 
Tanner I 1.1113 1.4108 0.4322 
Tanner II 1.5288 1.4353 0.2886 
Tanner III 0.6601 1.4854 0.6574 
Tanner IV 1.0127 1.6365 0.5370 
Tanner V 0 — — 
Age(Tanner) I 0.09295 0.02596 0.0004 
Age(Tanner) II 0.03337 0.03621 0.3576 
Age(Tanner) III 0.09830 0.04727 0.0385 
Age(Tanner) IV 0.08261 0.06902 0.2323 
Age(Tanner) V 0.1458 0.09770 0.1367 
Total fat −0.5723 0.05487 <0.0001 
African ADM −0.00762 0.001292 <0.001 
SES −0.00723 0.003694 0.0525 
EffectEstimateSEP
Intercept 5.4065 1.5183 0.0005 
Tanner I 1.1113 1.4108 0.4322 
Tanner II 1.5288 1.4353 0.2886 
Tanner III 0.6601 1.4854 0.6574 
Tanner IV 1.0127 1.6365 0.5370 
Tanner V 0 — — 
Age(Tanner) I 0.09295 0.02596 0.0004 
Age(Tanner) II 0.03337 0.03621 0.3576 
Age(Tanner) III 0.09830 0.04727 0.0385 
Age(Tanner) IV 0.08261 0.06902 0.2323 
Age(Tanner) V 0.1458 0.09770 0.1367 
Total fat −0.5723 0.05487 <0.0001 
African ADM −0.00762 0.001292 <0.001 
SES −0.00723 0.003694 0.0525 
View Large
TABLE 4

Parameter estimates for mixed model with the dependent variable fasting insulin

EffectEstimateSEP
Intercept −0.09500 1.2176 0.9379 
Tanner I −0.6287 1.1730 0.5928 
Tanner II −0.8173 1.1929 0.4944 
Tanner III −0.2747 1.2412 0.8252 
Tanner IV −0.05419 1.3682 0.9685 
Tanner V 0 — — 
Age(Tanner) I −0.05972 0.01998 0.0030 
Age(Tanner) II −0.02531 0.02732 0.3551 
Age(Tanner) III −0.06882 0.03878 0.0771 
Age(Tanner) IV −0.08737 0.05677 0.1249 
Age(Tanner) V −0.1004 0.08142 0.2184 
Total fat 0.3769 0.03340 <0.001 
African ADM 0.002211 0.000747 0.0037 
SES 0.003764 0.002090 0.0743 
EffectEstimateSEP
Intercept −0.09500 1.2176 0.9379 
Tanner I −0.6287 1.1730 0.5928 
Tanner II −0.8173 1.1929 0.4944 
Tanner III −0.2747 1.2412 0.8252 
Tanner IV −0.05419 1.3682 0.9685 
Tanner V 0 — — 
Age(Tanner) I −0.05972 0.01998 0.0030 
Age(Tanner) II −0.02531 0.02732 0.3551 
Age(Tanner) III −0.06882 0.03878 0.0771 
Age(Tanner) IV −0.08737 0.05677 0.1249 
Age(Tanner) V −0.1004 0.08142 0.2184 
Total fat 0.3769 0.03340 <0.001 
African ADM 0.002211 0.000747 0.0037 
SES 0.003764 0.002090 0.0743 
View Large
TABLE 5

Parameter estimates for mixed model with the dependent variable AIRg

EffectEstimateSEP
Intercept 8.1235 1.1355 <0.0001 
Tanner I −0.1827 1.1234 0.8710 
Tanner II −0.1134 1.1455 0.9213 
Tanner III 0.8785 1.1804 0.4580 
Tanner IV 0.07426 1.3024 0.9546 
Tanner V 0 — — 
Age(Tanner) I −0.04863 0.01951 0.0132 
Age(Tanner) II −0.05602 0.02918 0.0559 
Age(Tanner) III −0.1417 0.03795 0.0002 
Age(Tanner) IV −0.06465 0.05554 0.2453 
Age(Tanner) V −0.07767 0.07811 0.3209 
SI −0.3913 0.03792 <0.0001 
African ADM 0.006891 0.001377 <0.0001 
SES −0.00817 0.003848 0.0358 
EffectEstimateSEP
Intercept 8.1235 1.1355 <0.0001 
Tanner I −0.1827 1.1234 0.8710 
Tanner II −0.1134 1.1455 0.9213 
Tanner III 0.8785 1.1804 0.4580 
Tanner IV 0.07426 1.3024 0.9546 
Tanner V 0 — — 
Age(Tanner) I −0.04863 0.01951 0.0132 
Age(Tanner) II −0.05602 0.02918 0.0559 
Age(Tanner) III −0.1417 0.03795 0.0002 
Age(Tanner) IV −0.06465 0.05554 0.2453 
Age(Tanner) V −0.07767 0.07811 0.3209 
SI −0.3913 0.03792 <0.0001 
African ADM 0.006891 0.001377 <0.0001 
SES −0.00817 0.003848 0.0358 
View Large

This research was supported by NIH Grants R01DK58278, R01HD/HL33064, R01HG02154, R01DK53958, M01RR00032, and P30 DK56336. The efforts of study coordinator Tena Hilario, and the participation of the subjects and their families, are gratefully acknowledged.

1.
Arslanian S, Suprasongsin C: Differences in the in vivo insulin secretion and sensitivity of healthy black versus white adolescents.
J Pediatr
 
129
:
440
–443,
1996
2.
Haffner SM, D’Agostino R Jr, Saad MF, Rewers M, Mykkanen L, Selby J, Howard G, Savage PJ, Hamman RF, Wagenknecht LE, Bergman RN: Increased insulin resistance and insulin secretion in nondiabetic African-Americans and Hispanics compared with non-Hispanic whites. The Insulin Resistance Atherosclerosis Study.
Diabetes
 
45
:
742
–748,
1996
3.
Lovejoy JC, de la Bretonne JA, Klemperer M, Tulley R: Abdominal fat distribution and metabolic risk factors: effects of race.
Metabolism
 
45
:
1119
–1124,
1996
4.
Gower BA, Granger WM, Franklin F, Shewchuk RM, Goran MI: Contribution of insulin secretion and clearance to glucose-induced insulin concentration in African-American and Caucasian children.
J Clin Endocrinol Metab
 
87
:
2218
–2224,
2002
5.
Osei K, Schuster DP: Ethnic differences in secretion, sensitivity, and hepatic extraction of insulin in black and white Americans.
Diabet Med
 
11
:
755
–762,
1994
6.
Gower BA, Nagy TR, Goran MI: Visceral fat, insulin sensitivity, and lipids in prepubertal children.
Diabetes
 
48
:
1515
–1521,
1999
7.
Shriver MD, Smith MW, Jin L, Marcini A, Akey JM, Deka R, Ferrell RE: Ethnic affiliation estimation by use of population-specific DNA markers.
Am J Hum Genet
 
60
:
957
–964,
1997
8.
Parra EJ, Marcini A, Akey J, Martinson J, Batzer MA, Cooper R, Forrester T, Allison DB, Deka R, Ferrell RE, Shriver MD: Estimating African American admixture proportions by use of population-specific alleles.
Am J Hum Genet
 
63
:
1839
–1851,
1998
9.
McKeigue PM: Mapping genes that underlie ethnic differences in disease risk: methods for detecting linkage in admixed populations by conditioning on parental admixture.
Am J Hum Genet
 
63
:
241
–251,
1998
10.
Collins-Schramm HE, Phillips CM, Operario DJ, Lee JS, Weber JL, Hanson RL, Knowler WC, Cooper R, Li H, Seldin MF: Ethnic-difference markers for use in mapping by admixture linkage disequilibrium.
Am J Hum Genet
 
70
:
737
–750,
2002
11.
Hollingshead AB:
Four Factor Index of Social Status.
Yale University,
1975
12.
Goran MI, Gower BA: Longitudinal study on pubertal insulin resistance.
Diabetes
 
50
:
2444
–2450,
2002
13.
Akey JM, Sosnoski D, Parra E, Dios S, Hiester K, Su B, Bonilla C, Jin L, Shriver MD: Melting curve analysis of SNPs (McSNP): a gel-free and inexpensive approach for SNP genotyping.
Biotechniques
 
30
:
358
–367,
2001
14.
Hanis CL, Chakraborty R, Ferrell RE, Schull WJ: Individual admixture estimates: disease associations and individual risk of diabetes and gallbladder disease among Mexican-Americans in Starr County, Texas.
Am J Physiol Anthropol
 
70
:
433
–441,
1986
15.
Moran A, Jacobs DR Jr, Steinberger J, Hong, C-P, Prineas R, Luepker R, Sinaiko AR: Insulin resistance during puberty: results from clamp studies in 357 children.
Diabetes
 
48
:
2039
–2044,
1999
16.
Kahn SE, Prigeon RL, McCulloch DK, Boyko EJ, Bergman RN, Schwartz MW, Neifing JL, Ward WK, Beard JC, Palmer JP, Porte D Jr: Quantification of the relationship between insulin sensitivity and beta cell function in human subjects: evidence for a hyperbolic function.
Diabetes
 
42
:
1663
–1672,
1993
17.
Lindquist CH, Gower BA, Goran MI: Role of dietary factors in ethnic differences in early risk of cardiovascular disease and type 2 diabetes.
Am J Clin Nutr
 
71
:
725
–732,
2000
18.
Ku C-Y, Gower BA, Hunter GR, Goran MI: Racial differences in insulin secretion and sensitivity in prepubertal children: role of physical fitness and physical activity.
Obes Res
 
8
:
506
–515,
2000
19.
Cohen J: The cost of dichotomization.
Appl Psychol Meas
 
7
:
249
–253,
1983
20.
Peters CC, Von Voorhis WR:
Statistical Procedures and Their Mathematical Bases.
New York, McGraw Hill,
1940

Address correspondence and reprint requests to Barbara A. Gower, Department of Nutrition Sciences, University of Alabama at Birmingham, 423 Webb Bldg., 1675 University Blvd., Birmingham, AL 35294-3360. E-mail: [email protected].

Received for publication 9 July 2001 and accepted in revised form 2 December 2002.

ADM, genetic admixture; AIRg, acute insulin response to glucose; SES, socioeconomic status; SI, insulin sensitivity.

DIABETES
2003