HHS Public Access Author manuscript Author Manuscript
Dysmorphol Clin Genet. Author manuscript; available in PMC 2016 September 14. Published in final edited form as: Dysmorphol Clin Genet. 1992 ; 6(1): 9–14.
HLA ASSOCIATIONS IN OBESE WHITE AND BLACK ADULTS Merlin G. Butler, Dominique Walton, Weitong Zhu, and Gary Niblack* Division of Genetics, Departments of Pediatrics and Pathology, Vanderbilt University Medical Center, Nashville, Tennessee *Ren
Laboratory, Nashville, Tennessee
Author Manuscript
Abstract We summarized HLA-A and -B data from 1095 black and white adult men and women with or without obesity to determine if specific HLA tissue types are overrepresented in obese individuals compared with nonobese. None of the three HLA types (Aw30, B18, Bw35) previously reported to relate to obesity was overrepresented in obese subjects in our study. However, B14 and B41 haplotypes were overrepresented in obese white men compared with nonobese men, and B7 was overrepresented in obese black men compared with nonobese men. Additional research will be required to confirm the HLA associations we found and to determine if methodologic differences could account for the differences among the previous studies.
Author Manuscript
Associations between human leukocyte antigens (HLA) and diseases or clinical conditions including ankylosing spondylitis, Hodgkin disease, and type I diabetes mellitus, to name a few, have been confirmed, but associations with obesity have not been consistently reported. For example, Fumeron and Apfelbaum1 found an association of HLA-B18 with familial obesity, and Fabitz et al.2 found an association with HLA-Aw30 and obesity. Beck et al.3, Digby et al.4, and Bouchard et al.5 did not confirm these findings. However, Digby et al.4 did find significant associations with Bw35 and Cw4 in a comparison of obese women with nonobese men and women controls. We conducted a large survey of black and white men and women from the general population with and without obesity, with HLA typing previously obtained for paternity testing, to determine if specific HLA tissue types are overrepresented in obese compared with nonobese adults.
Author Manuscript
MATERIALS AND METHODS We examined black and white front-view photographs obtained from 1095 normal, healthy, unrelated adults (370 white men, 367 white women, 179 black men, 179 black women) between 18 and 65 years of age from the general population in whom HLA-A and -B tissue types were determined previously for paternity testing. The obesity status was determined photoanthropometrically without knowledge of the HLA data by categorizing each individual into one of five groups by four independent observers who ranked each
Address reprint requests to Merlin G. Butler, MD, PhD, Division of Genetics, Department of Pediatrics, Vanderbilt University School of Medicine, DD-2205 Medical Center North, Nashville, TN 37232-2578.
Butler et al.
Page 2
Author Manuscript
photograph (obese = 5, borderline obese = 4, normal = 3, borderline thin = 2, thin = 1). The rankings were averaged for each subject by the four observers. An individual was judged to be in the obese range if his or her average score was greater than 4. To monitor quality control and observer reliability of determination of obesity by photoanthropometric analysis, several photographs of known obese (i.e., ideal body weight >130%) and nonobese individuals were evaluated by the observers. Agreement among them was reasonable in the assignment of obesity (score >4) for a specific individual who was known to be obese by physical measurements. The chi-square test with Yates’s correction was used for statistical analysis of the HLA and obesity data.
RESULTS
Author Manuscript
We determined that 190 adult individuals were photoanthropometrically obese: 11.4% of white men, 21.5% of white women, 14.0% of black men, and 24.6% of black women.
Author Manuscript
To determine if the three HLA types (Aw30, B18, Bw35) previously reported to be associated with obesity, or to determine if previously unrecognized associations exist, chisquare values with Yates’s correction were calculated for all subjects. The values for the three reported HLA types were not significant, and ranged from 0.03 to 1.4 for white men, 0.01 to 0.75 for white women, 0.1 to 0.3 for black men, and 0.0 to 1.2 for black women. Therefore, we found no evidence of a higher frequency (chi-square test; P > 0.05) in obese adult individuals for either sex or race of any of the three types. However, significant chisquare values (P < 0.05) of 7.8 and 4.5 were found for B14 and B41, respectively, in the comparison of obese and nonobese white men. A significant value of 8.7 was also calculated for B7 for obese black men compared with nonobese. There was no evidence of a higher frequency or overrepresentation of any of the types in obese compared with nonobese women for either whites or blacks (Tables 1 through 4).
DISCUSSION
Author Manuscript
Analyses of HLA-A and -B data from 1095 black and white men and women did not confirm previously reported associations, particularly of Aw30, B18, and Bw35, with obesity. However, B14 and B41 were overrepresented in obese white men compared with nonobese white men, and B7 was overrepresented in obese black men compared with nonobese black men. Of interest, Beck et al.3 found the frequency of B14 in their obese subjects to be nearly twice that of nonobese individuals, but possibly due to a small sample size a significant difference was not observed in their study, although it was significant in our study. Our study differs from others in the determination of obesity, although the percentages of obese individuals in our study (ranges from 11% for white men to 25% for black women) are comparable to the frequency of obesity in the general population. Obesity is heterogeneous, with the presence or absence of concomitant disorders such as diabetes and dyslipoproteinemia, but genetic factors may play a role in familial obesity.3 Although we did not record other conditions (e.g., diabetes mellitus) in our subjects or obtain family histories, our results did not confirm previous associations of HLA and obesity. We also examined black and white adults only, but of both sexes. Additional
Dysmorphol Clin Genet. Author manuscript; available in PMC 2016 September 14.
Butler et al.
Page 3
Author Manuscript
research will be required to confirm the HLA associations with obesity in the analysis of obese and nonobese adults.
Acknowledgments We thank Pamela Grimm for expert preparation of the manuscript. This research was partially supported by the March of Dimes Birth Defects Foundation grant 6–571.
References
Author Manuscript
1. Fumeron F, Apfelbaum M. HLA-B18 and the familial-obesity syndrome. N Engl J Med. 1981; 305:645. [PubMed: 6943425] 2. Fabsitz RR, Nam J, Gart J, Stunkard A, Price RA, Wilson PWF. HLA associations with obesity. Hum Hered. 1989; 39:156–164. [PubMed: 2591979] 3. Beck MD, Svejcar J, Kuhnl P, Baur MP. HLA and obesity. Biomed Pharmacother. 1984; 38:209– 211. [PubMed: 6594167] 4. Digby, JP.; Raffoux, C.; Pointel, JP.; Perrier, P.; Drouin, P.; Mejean, L.; Streiff, F.; Debry, G. HLA and familial obesity: Evidence for a genetic origin. In: Hirsch, J.; VanItalie, T., editors. Recent advances in obesity research. Vol. 4. London: Libbey Press; 1983. p. 171-175. 5. Bouchard C, Perusse C, Rivest J, Morissette J, Allard C, Theriault G, LeBlanc C, Tremblay A. HLA system, body fat and fat distribution in children and adults. Int J Obesity. 1985; 9:411–422.
Author Manuscript Author Manuscript Dysmorphol Clin Genet. Author manuscript; available in PMC 2016 September 14.
Author Manuscript
Author Manuscript
Author Manuscript
4
3
8
2
1
4
1
5
1
1
0
0
0
23
24
25
26
28
29
30
31
32
33
34
36
0
3
3
23
10
26
17
14
12
14
47
14
32
84
174
114
Nonobese White Men (n = 328)
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
Pa
1
0
0
2
2
6
7
8
6
1
11
1
9
24
37
23
Obese White Women (n = 79)
NS = not significant.
P value determined by two tailed chi-square test with Yates’s correction.
a
13
3
19
11
13
2
Obese White Men (n = 42)
1
HLA-A
0
0
4
16
9
28
19
15
13
14
30
8
45
72
151
90
Nonobese White Women (n = 288)
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
Pa
Comparison of HLA-A Haplotypes Between Obese and Nonobese White Men and Women
Author Manuscript
Table 1 Butler et al. Page 4
Dysmorphol Clin Genet. Author manuscript; available in PMC 2016 September 14.
Author Manuscript
Author Manuscript
Author Manuscript
7
3
6
3
1
4
1
0
3
4
8
0
0
0
4
2
8
2
0
1
1
0
0
0
2
4
8
14
15
16
17
18
21
22
27
35
37
38
39
40
41
44
45
49
50
51
52
53
55
57
60
8
13
2
7
Obese White Men (n = 42)
5
HLA-B
Dysmorphol Clin Genet. Author manuscript; available in PMC 2016 September 14. 26
8
1
0
1
18
5
8
6
83
1
34
9
4
10
39
25
14
3
16
20
10
30
11
11
76
89
15
Nonobese White Men (n = 328)
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
Dysmorphol Clin Genet. Author manuscript; available in PMC 2016 September 14. Published in final edited form as: Dysmorphol Clin Genet. 1992 ; 6(1): 9–14.
HLA ASSOCIATIONS IN OBESE WHITE AND BLACK ADULTS Merlin G. Butler, Dominique Walton, Weitong Zhu, and Gary Niblack* Division of Genetics, Departments of Pediatrics and Pathology, Vanderbilt University Medical Center, Nashville, Tennessee *Ren
Laboratory, Nashville, Tennessee
Author Manuscript
Abstract We summarized HLA-A and -B data from 1095 black and white adult men and women with or without obesity to determine if specific HLA tissue types are overrepresented in obese individuals compared with nonobese. None of the three HLA types (Aw30, B18, Bw35) previously reported to relate to obesity was overrepresented in obese subjects in our study. However, B14 and B41 haplotypes were overrepresented in obese white men compared with nonobese men, and B7 was overrepresented in obese black men compared with nonobese men. Additional research will be required to confirm the HLA associations we found and to determine if methodologic differences could account for the differences among the previous studies.
Author Manuscript
Associations between human leukocyte antigens (HLA) and diseases or clinical conditions including ankylosing spondylitis, Hodgkin disease, and type I diabetes mellitus, to name a few, have been confirmed, but associations with obesity have not been consistently reported. For example, Fumeron and Apfelbaum1 found an association of HLA-B18 with familial obesity, and Fabitz et al.2 found an association with HLA-Aw30 and obesity. Beck et al.3, Digby et al.4, and Bouchard et al.5 did not confirm these findings. However, Digby et al.4 did find significant associations with Bw35 and Cw4 in a comparison of obese women with nonobese men and women controls. We conducted a large survey of black and white men and women from the general population with and without obesity, with HLA typing previously obtained for paternity testing, to determine if specific HLA tissue types are overrepresented in obese compared with nonobese adults.
Author Manuscript
MATERIALS AND METHODS We examined black and white front-view photographs obtained from 1095 normal, healthy, unrelated adults (370 white men, 367 white women, 179 black men, 179 black women) between 18 and 65 years of age from the general population in whom HLA-A and -B tissue types were determined previously for paternity testing. The obesity status was determined photoanthropometrically without knowledge of the HLA data by categorizing each individual into one of five groups by four independent observers who ranked each
Address reprint requests to Merlin G. Butler, MD, PhD, Division of Genetics, Department of Pediatrics, Vanderbilt University School of Medicine, DD-2205 Medical Center North, Nashville, TN 37232-2578.
Butler et al.
Page 2
Author Manuscript
photograph (obese = 5, borderline obese = 4, normal = 3, borderline thin = 2, thin = 1). The rankings were averaged for each subject by the four observers. An individual was judged to be in the obese range if his or her average score was greater than 4. To monitor quality control and observer reliability of determination of obesity by photoanthropometric analysis, several photographs of known obese (i.e., ideal body weight >130%) and nonobese individuals were evaluated by the observers. Agreement among them was reasonable in the assignment of obesity (score >4) for a specific individual who was known to be obese by physical measurements. The chi-square test with Yates’s correction was used for statistical analysis of the HLA and obesity data.
RESULTS
Author Manuscript
We determined that 190 adult individuals were photoanthropometrically obese: 11.4% of white men, 21.5% of white women, 14.0% of black men, and 24.6% of black women.
Author Manuscript
To determine if the three HLA types (Aw30, B18, Bw35) previously reported to be associated with obesity, or to determine if previously unrecognized associations exist, chisquare values with Yates’s correction were calculated for all subjects. The values for the three reported HLA types were not significant, and ranged from 0.03 to 1.4 for white men, 0.01 to 0.75 for white women, 0.1 to 0.3 for black men, and 0.0 to 1.2 for black women. Therefore, we found no evidence of a higher frequency (chi-square test; P > 0.05) in obese adult individuals for either sex or race of any of the three types. However, significant chisquare values (P < 0.05) of 7.8 and 4.5 were found for B14 and B41, respectively, in the comparison of obese and nonobese white men. A significant value of 8.7 was also calculated for B7 for obese black men compared with nonobese. There was no evidence of a higher frequency or overrepresentation of any of the types in obese compared with nonobese women for either whites or blacks (Tables 1 through 4).
DISCUSSION
Author Manuscript
Analyses of HLA-A and -B data from 1095 black and white men and women did not confirm previously reported associations, particularly of Aw30, B18, and Bw35, with obesity. However, B14 and B41 were overrepresented in obese white men compared with nonobese white men, and B7 was overrepresented in obese black men compared with nonobese black men. Of interest, Beck et al.3 found the frequency of B14 in their obese subjects to be nearly twice that of nonobese individuals, but possibly due to a small sample size a significant difference was not observed in their study, although it was significant in our study. Our study differs from others in the determination of obesity, although the percentages of obese individuals in our study (ranges from 11% for white men to 25% for black women) are comparable to the frequency of obesity in the general population. Obesity is heterogeneous, with the presence or absence of concomitant disorders such as diabetes and dyslipoproteinemia, but genetic factors may play a role in familial obesity.3 Although we did not record other conditions (e.g., diabetes mellitus) in our subjects or obtain family histories, our results did not confirm previous associations of HLA and obesity. We also examined black and white adults only, but of both sexes. Additional
Dysmorphol Clin Genet. Author manuscript; available in PMC 2016 September 14.
Butler et al.
Page 3
Author Manuscript
research will be required to confirm the HLA associations with obesity in the analysis of obese and nonobese adults.
Acknowledgments We thank Pamela Grimm for expert preparation of the manuscript. This research was partially supported by the March of Dimes Birth Defects Foundation grant 6–571.
References
Author Manuscript
1. Fumeron F, Apfelbaum M. HLA-B18 and the familial-obesity syndrome. N Engl J Med. 1981; 305:645. [PubMed: 6943425] 2. Fabsitz RR, Nam J, Gart J, Stunkard A, Price RA, Wilson PWF. HLA associations with obesity. Hum Hered. 1989; 39:156–164. [PubMed: 2591979] 3. Beck MD, Svejcar J, Kuhnl P, Baur MP. HLA and obesity. Biomed Pharmacother. 1984; 38:209– 211. [PubMed: 6594167] 4. Digby, JP.; Raffoux, C.; Pointel, JP.; Perrier, P.; Drouin, P.; Mejean, L.; Streiff, F.; Debry, G. HLA and familial obesity: Evidence for a genetic origin. In: Hirsch, J.; VanItalie, T., editors. Recent advances in obesity research. Vol. 4. London: Libbey Press; 1983. p. 171-175. 5. Bouchard C, Perusse C, Rivest J, Morissette J, Allard C, Theriault G, LeBlanc C, Tremblay A. HLA system, body fat and fat distribution in children and adults. Int J Obesity. 1985; 9:411–422.
Author Manuscript Author Manuscript Dysmorphol Clin Genet. Author manuscript; available in PMC 2016 September 14.
Author Manuscript
Author Manuscript
Author Manuscript
4
3
8
2
1
4
1
5
1
1
0
0
0
23
24
25
26
28
29
30
31
32
33
34
36
0
3
3
23
10
26
17
14
12
14
47
14
32
84
174
114
Nonobese White Men (n = 328)
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
Pa
1
0
0
2
2
6
7
8
6
1
11
1
9
24
37
23
Obese White Women (n = 79)
NS = not significant.
P value determined by two tailed chi-square test with Yates’s correction.
a
13
3
19
11
13
2
Obese White Men (n = 42)
1
HLA-A
0
0
4
16
9
28
19
15
13
14
30
8
45
72
151
90
Nonobese White Women (n = 288)
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
Pa
Comparison of HLA-A Haplotypes Between Obese and Nonobese White Men and Women
Author Manuscript
Table 1 Butler et al. Page 4
Dysmorphol Clin Genet. Author manuscript; available in PMC 2016 September 14.
Author Manuscript
Author Manuscript
Author Manuscript
7
3
6
3
1
4
1
0
3
4
8
0
0
0
4
2
8
2
0
1
1
0
0
0
2
4
8
14
15
16
17
18
21
22
27
35
37
38
39
40
41
44
45
49
50
51
52
53
55
57
60
8
13
2
7
Obese White Men (n = 42)
5
HLA-B
Dysmorphol Clin Genet. Author manuscript; available in PMC 2016 September 14. 26
8
1
0
1
18
5
8
6
83
1
34
9
4
10
39
25
14
3
16
20
10
30
11
11
76
89
15
Nonobese White Men (n = 328)
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
