An association has recently been described between increased birth weight and increased risk of childhood-onset type 1 diabetes. Whether this relationship is explained by genes associated with both increased birth weight and increased risk of type 1 diabetes is unknown. In the present study, we tested the association between birth weight and HLA-DQ genotypes known to confer risk for type 1 diabetes among 969 nondiabetic children randomly selected from the Norwegian population. We found that HLA genotypes previously shown to confer risk for type 1 diabetes were associated with reduced birth weight (the mean difference in birth weight between the DQB1*0602/DQB1*0602 and DQ8/DQ2 genotypes was 354 g [95% CI 105–604]), which was opposite of that expected if HLA genes explained the birth weight–type 1 diabetes association.

A positive association between birth weight and risk of type 1 diabetes has recently been described in large epidemiological studies, even after exclusion of maternal diabetes (1,2). Together with other lines of evidence, such studies have been taken as evidence that environmental risk factors for type 1 diabetes may play a role in utero or early in life (3). However, we cannot exclude the possibility that the above-mentioned association (1,2) is explained by genes associated with both increased birth weight and increased risk of type 1 diabetes.

The most important genetic contribution to the risk of type 1 diabetes is encoded within the HLA complex on chromosome 6p (4), and different HLA-DQA1 and -DQB1 alleles confer either strongly increased or reduced risk (5). The insulin gene region (INS) on chromosome 11p is probably the second most important genetic contribution to variation in susceptibility to type 1 diabetes (4,6). The class I allele at INS VNTR that confers susceptibility to type 1 diabetes (6) has been associated with reduced birth weight (7), indicating that INS VNTR is unlikely to explain the observed association between birth weight and type 1 diabetes (1,2).

Genes in the HLA complex encode molecules involved in rejection of nonself tissue, and antibodies to fetal HLA molecules have been found in a large proportion of parous women (8). Previous studies have focused on maternal-fetal HLA compatibility and pregnancy outcome. Because there are indications that genes in the HLA complex are associated with fetal growth (9), we tested whether HLA-DQ genotypes previously found to confer risk for type 1 diabetes were associated with increased birth weight among Norwegian children randomly selected from the official population registry.

A total of 1,484 children born in Norway between 1982 and 1998 and residing in Vest-Agder county were contacted by mail in 1999 and received a questionnaire and mouth brushes for extraction of DNA and HLA typing. Of the 1,056 who returned the buccal cell samples, 1,008 were linked to the Medical Birth Registry of Norway, which stores computerized information on birth weight, gestational age, maternal age at delivery, maternal parity, and other perinatal factors for all births in Norway since 1967 (10). The majority of those children not linked were born abroad. Five children with type 1 diabetes were excluded. Those who were part of a twin pregnancy, whose mother had gestational diabetes or diabetes diagnosed up to the time of study, or who had missing data for birth weight were also excluded, leaving 969 children for analysis.

The general perinatal characteristics of the study sample were as expected for the general population of Norway for the given period (Table 1). We decided a priori to categorize HLA-DQ genotypes into five groups according to the degree of risk for type 1 diabetes, as determined in previous studies (11,12). The high-risk DQA1*0301-DQB1*0302/DQA1*0501-DQB1*0201 (DQ8/DQ2) genotype, five other genotypes found by Rønningen (12) to confer significant risk of type 1 diabetes (DQA1*0301-DQB1*0302/DQA1*0401-DQB1*0402, DQA1*0301-DQB1*0302/DQA1*0301-DQB1*0302, DQA1*0301-DQB1*0302/DQA1*0101-DQB1*0501, DQA1*0301-DQB1*0302/DQA1*0102-DQB1*0604, or DQA1*0501-DQB1*0201/DQA1*0501-DQB1*0201), genotypes neutral with regard to type 1 diabetes susceptibility, genotypes with one copy of the strongly protective DQB1*0602 allele (11), and genotypes with two copies of the DQB1*0602 allele were compared (Fig. 1). The association was the opposite of that expected under the hypothesis that variation in the HLA-DQ genotype explains the previously observed positive association between birth weight and risk of type 1 diabetes (1,2). The overall association was statistically significant (P = 0.04; analysis of variance), and it was highly significant when we tested for trend by taking into consideration the ordering of the genotype categories (coded 1, 2, 3, 4, and 5) according to risk of type 1 diabetes determined in previous studies (P = 0.004; linear regression test for trend). Subjects homozygous for the DQB1*0602 allele, previously shown to confer protection against type 1 diabetes (11), had the highest birth weights, whereas subjects carrying the DQ8/DQ2 genotype, previously found to confer high risk for type 1 diabetes, had the lowest birth weights (mean difference 354 g [95% CI 105–604]). The strength of association was slightly reduced, but the conclusion remained unchanged after adjustment for gestational age, maternal smoking during pregnancy, maternal age at birth, birth order, cesarean section, maternal education, sex, and age (Table 2). The association was similar at different levels of the possible confounders adjusted for in Table 2 (data not shown). There was no significant association between HLA-DQ genotype and gestational age (Table 2).

Our data suggest that HLA genotype cannot explain the positive association between birth weight and risk of type 1 diabetes observed in previous studies (1,2). Our data supports previous indications that HLA genes or HLA-linked genes may be associated with variation in fetal growth (9), although the mechanisms that may be operating are unknown. We did not obtain DNA from the mothers, but maternal genotype is of less importance in regard to our research question, which is whether there were any associations between diabetes-associated HLA-DQ genes and birth weight among children. We did not have data on ethnicity in our study. Although it is theoretically possible that our results could be confounded by the presence of ethnic minorities with different birth weights and different frequencies of HLA-DQ alleles compared with ethnic Norwegians, this is unlikely because non-Western ethnic groups comprise a very small proportion of the Norwegian population. The six major non-Western immigrants to Norway comprised ∼3.2% of the Norwegian population (13), probably less in Vest-Agder county. We tested associations using two-sided tests, which are designed to detect deviations from the null hypothesis in either direction. It is probably premature to conclude that there is a causal link between HLA-DQ genes and birth weight, but our results strongly suggest that HLA-DQ genotypes previously shown to confer increased risk for type 1 diabetes are not associated with increased birth weight. Although contributions from other genes cannot be excluded, our data, together with the previously observed association between INS VNTR and birth weight (7), suggest that nongenetic factors are more likely explanations of the previously observed association between birth weight and risk of type 1 diabetes (1,2). Such factors are as yet unknown, but it could be speculated that increased growth in utero or in early childhood mediated by increased β-cell secretory activity could be part of the explanation. Future studies should include subjects with type 1 diabetes and assess the association between birth weight and type 1 diabetes after adjusting for HLA genotype.

Subjects were asked to take a buccal cell sample using two cytology brushes (Medscan Medical, Malmö, Sweden) and return them by mail. DNA was extracted using a method similar to that described by Parad (14). The validity of the mouth brush method had been tested previously (15,16). The HLA-DQA1 and -DQB1 loci were typed using polymerase chain reaction with biotinylated primers and reverse hybridization with sequence-specific oligonucleotide probes, as previously described (17). The DQA1 locus was only typed for the DQA1*0201, DQA1*03, and DQA1*0501 alleles, whereas other alleles were inferred from known linkage disequilibriums with DQB1 alleles (18). To ensure complete typing of the DQB1 locus, nine probes were used in addition to those described by Cinek et al. (17). The designations and sequences of these probes are available from the authors.

G. Joner and L.C. Stene were supported by a grant from the Norwegian Foundation for Health and Rehabilitation (1997/156) and grants from the Norwegian Diabetes Association. Funding was also kindly provided by TINE Norwegian Dairies and Novo Nordisk Pharma A/S.

The authors thank Unni Hinna, Jorunn Ulriksen, Per Vesterhus, Torleif Trydal, Liv Paltiel, Elisabet Witsø, Hang Lu, the Norwegian Childhood Diabetes Study Group, and the staff at the Medical Birth Registry of Norway.

1.
Stene LC, Magnus P, Lie RT, Søvik O, Joner G, the Norwegian Childhood Diabetes Study Group: Birth weight and childhood onset type 1 diabetes: population based cohort study.
BMJ
 
322
:
889
–892,
2001
2.
Dahlquist GG, Patterson C, Stoltesz G: Perinatal risk factors for childhood type 1 diabetes in Europe: the EURODIAB Substudy 2 Study Group.
Diabetes Care
 
22
:
1698
–1702,
1999
3.
Leslie RD, Elliott RB: Early environmental events as a cause of IDDM: evidence and implications (Review).
Diabetes
 
43
:
843
–850,
1994
4.
Concannon P, Gogolin-Ewens KJ, Hinds DA, Wapelhorst B, Morrison VA, Stirling B, Mitra M, Farmer J, Williams SR, Cox NJ, Bell GI, Risch N, Spielman RS: A second-generation screen of the human genome for susceptibility to insulin-dependent diabetes mellitus.
Nat Genet
 
19
:
292
–296,
1998
5.
Thorsby E, Rønningen KS: Particular HLA-DQ molecules play a dominant role in determining susceptibility or resistance to type 1 (insulin-dependent) diabetes mellitus.
Diabetologia
 
36
:
371
–377,
1993
6.
Bain SC, Prins JB, Hearne CM, Rodrigues NR, Rowe BR, Pritchard LE, Ritchie RJ, Hall JR, Undlien DE, Rønningen KS: Insulin gene region-encoded susceptibility to type 1 diabetes is not restricted to HLA-DR4-positive individuals.
Nat Genet
 
2
:
212
–215,
1992
7.
Dunger DB, Ong KK, Huxtable SJ, Sherriff A, Woods KA, Ahmed ML, Golding J, Pembrey ME, Ring S, Bennett ST, Todd JA: Association of the INS VNTR with size at birth: ALSPAC Study Team: Avon Longitudinal Study of Pregnancy and Childhood.
Nat Genet
 
19
:
98
–100,
1998
8.
Morin-Papunen L, Tiilikainen A, Hartikainen-Sorri AL: Maternal HLA immunization during pregnancy: presence of anti HLA antibodies in half of multigravidous women.
Med Biol
 
62
:
323
–325,
1984
9.
Ober C: The maternal-fetal relationship in human pregnancy: an immunogenetic perspective.
Exp Clin Immunogenet
 
9
:
1
–14,
1992
10.
Irgens LM: The medical birth registry of Norway: epidemiological research and surveillance throughout 30 years.
Acta Obstet Gynecol Scand
 
79
:
435
–439,
2000
11.
Pugliese A, Gianani R, Moromisato R, Awdeh ZL, Alper CA, Erlich HA, Jackson RA, Eisenbarth GS: HLA-DQB1*0602 is associated with dominant protection from diabetes even among islet cell antibody–positive first-degree relatives of patients with IDDM.
Diabetes
 
44
:
608
–613,
1995
12.
Rønningen KS: Genetics in the prediction of insulin-dependent diabetes mellitus: from theory to practice.
Ann Med
 
29
:
387
–392,
1997
13.
Vangen S, Stoltenberg C, Skrondal A, Magnus P, Stray-Pedersen B: Cesarean section among immigrants in Norway.
Acta Obstet Gynecol Scand
 
79
:
553
–558,
2000
14.
Parad RB: Buccal cell DNA mutation analysis for diagnosis of cystic fibrosis in newborns and infants inaccessible to sweat chloride measurement.
Pediatrics
 
101
:
851
–855,
1998
15.
Richards B, Skoletsky J, Shuber AP, Balfour R, Stern RC, Dorkin HL, Parad RB, Witt D, Klinger KW: Multiplex PCR amplification from the CFTR gene using DNA prepared from buccal brushes/swabs.
Hum Mol Genet
 
2
:
159
–163,
1993
16.
Gillespie KM, Valovin SJ, Saunby J, Hunter KM, Savage DA, Middleton D, Todd JA, Bingley PJ, Gale EA: HLA class II typing of whole genome amplified mouth swab DNA.
Tissue Antigens
 
56
:
530
–538,
2000
17.
Cinek O, Wilkinson E, Paltiel L, Saugstad OD, Magnus P, Rønningen KS: Screening for the IDDM high-risk genotype: a rapid microtitre plate method using serum as source of DNA.
Tissue Antigens
 
56
:
344
–349,
2000
18.
Rønningen KS, Spurkland A, Tait BD, Drummond B, Lopez-Larrea C, Baranda FS, Menedez-Diaz MJ, Caillat-Zucman S, Beaurain G, Garchon HJ, Ilonen J, Reijonen H, Knip M, Boehm BO, Rosak C, Loliger C, Ottenhoff T, Contu L, Carcassi C, Savi M, Zanelli P, Neri M, Hamaguchi K, Kimura A, Dong RP, Chikiba N, Nagataki S, Gorodezky C, Debaz H, Robles C, Coimbra HB, Martinho A, Ruas MA, Sachs JA, Garcia-Pachedo M, Biro A, Nikaein A, Dombrausky L, Gonwa T, Zmijewsky C, Monos D, Kamoun M, Layrisse Z, Magli MC, Balducci P, Thorsby E: HLA class II associations in insulin-dependent diabetes mellitus among blacks, Caucasoids and Japanese. In
HLA 1991
: Proceedings of the 11th International Histocompatibility Workshop and Conference, Yokohama, Japan,
1991
. Tsuji K, Aizawa M, Sasazuki T, Eds. Oxford, Oxford University Press,
1992
, p.
713
–722

Address correspondence and reprint requests to Lars C. Stene, Section of Epidemiology, National Institute of Public Health, P.O. Box 4404 Nydalen, N-0403 Oslo, Norway. E-mail: [email protected].

Received for publication 26 June 2001 and accepted in revised form 20 September 2001.

2001