PEDIATRICS Vol. 106 No. 4 October 2000, pp. 719-724

Different -Casein Fractions in Icelandic Versus Scandinavian Cow's Milk May Influence Diabetogenicity of Cow's Milk in Infancy and Explain Low Incidence of Insulin-Dependent Diabetes Mellitus in Iceland

Inga Thorsdottir, PhD^*, Bryndis Eva Birgisdottir, RD^*, Inga Maria Johannsdottir, MD, D. Paul Harris, PhD^§, Jeremy Hill, PhD^§, Laufey Steingrimsdottir, PhD, and Arni V. Thorsson, MD

From the ^* Unit for Nutrition Research, National University Hospital, Department of Food Science, and the  Department of Pediatrics, Reykjavik Hospital, University of Iceland, Reykjavik, Iceland; ^§ New Zealand Dairy Research Institute, Palmerston North, New Zealand; and  Icelandic Nutrition Council, Reykjavik, Iceland.

	ABSTRACT

Top Abstract Methods Results Discussion References

Objectives.  To compare children with insulin-dependent diabetes mellitus (IDDM) with controls in Iceland regarding their consumption of cow's milk in infancy, and to investigate the -casein fractions in Scandinavian and Icelandic cow's milk. The A1 variant of -casein has been shown to be diabetogenic in animal studies, and suggestions have been made that the B variant of -casein acts similarly. Differences in the relative proportions of -casein fractions might explain the lower incidence of IDDM in Iceland than in Scandinavia.

Methods.  A retrospective case-control study on IDDM patients and matching controls was performed in Iceland to compare their diets in infancy. Fifty-five children with IDDM born in Iceland over a 16-year period and randomly collected controls (n = 165) were recruited to the study. Mothers of the children answered questions on breastfeeding habits and on when cow's milk products were introduced. Samples of cow's milk from randomly selected milk batches from the largest consumption areas in Iceland and Scandinavia were collected. The milk samples were freeze-dried and their -casein fractions were analyzed using capillary electrophoresis.

Results.  No significant difference was found between IDDM patients and controls in the frequency and duration of breastfeeding or the first introduction of cow's milk products. The analyses of milk samples showed that the percentage of the A1 and B variants of -casein in Icelandic milk was significantly lower than in the milk from the Scandinavian countries.

Conclusions.  Cow's milk consumption in infancy is not related to IDDM in Iceland. The lower fraction of A1 and B -caseins in Icelandic cow's milk may explain why there is a lower incidence of IDDM in Iceland than in Scandinavia.  Key words:  insulin-dependent diabetes mellitus, cow, caseins, milk proteins, breastfeeding, primary prevention, diabetes, infants.

The incidence of insulin-dependent diabetes mellitus (IDDM) in Iceland is less than one half the IDDM incidence in the other Nordic countries, ie, Scandinavia (Norway, Denmark, Sweden, and Finland), despite a similar genetic background.^1-6 This difference cannot be explained by different breastfeeding habits because the duration of breastfeeding is similar and high in all these countries.⁷ However, studies from Europe and the United States have shown a relationship between short duration of breastfeeding and a higher incidence of IDDM.^8,9 The consumption of dairy products in infancy has also repeatedly been related to IDDM,^10,11 although not all studies have found this relationship.¹² This discrepancy indicates variations in diabetogenicity of cow's milk. However, a meta-analysis has pinpointed exposure to cow's milk as an important determinant of IDDM,¹³ and in 1994 the American Academy of Pediatrics found reason to strongly recommend breastfeeding. They also recommended that families with a strong history of IDDM should avoid products containing intact cow's milk protein during the first year of life.¹⁴

Recently a cow's milk -casein fraction, A1, has been found to be diabetogenic in animal research, whereas the -casein fraction A2 has not.¹⁵ These are the dominant variants of -casein in milk from most cattle breeds.¹⁵ Thus, it is possible that variations in -casein fractions could contribute to the apparent lack of agreement among studies on the diabetogenicity of cow's milk. Earlier, -casein, which is one of the primary proteins in cow's milk, was found to be diabetogenic in experimental animals (nonobese diabetic [NOD] mice),¹⁶ but hydrolyzed cow's milk proteins showed a highly protective effect against diabetes.^17,18 Increased lymphocyte response to -casein was also found in patients with recent onset of IDDM, compared with nondiabetic healthy subjects and patients with other autoimmune diseases.¹⁹ Another study showed higher antibody levels against A1 -casein in diabetic patients than in healthy controls.²⁰ A less frequent -casein variant, B -casein, has been suggested to act in a similar way as A1. A1 and B -caseins both have histidine at position 67 in the primary sequence of this protein, but A2 -casein has proline at this position.^15,21 The amino acid sequence of A1 and B -caseins causes a cleavage pattern by intestinal digestive enzymes that produces -casomorphin-7 (BCM), suggested to be diabetogenic.^15,21

Genetic research on the Icelandic and Scandinavian cattle breeds indicates that genes coding for A1 and B -caseins are less frequent in Icelandic cattle than in the predominant cattle breeds in Scandinavia.²² Icelandic cattle have been isolated for 1100 years,²² while cattle breeding has been performed within and between the Scandinavian countries.²³ In addition, in Scandinavia dairy cattle have also been imported from other countries.²³ Very recently a significant correlation was found between the calculated per capita consumption of A1 and B -casein and the incidence of IDDM in 10 countries, including Iceland and the 4 Scandinavian countries.²⁴ The consumption of dairy products in infancy and its relation to IDDM in Iceland have not been investigated. Also, the actual content of -casein variants in the milk supply in Scandinavia has not been analyzed.

The aims of this current study was to compare Icelandic infants who later became diabetic and healthy controls with regard to breastfeeding habits and consumption of cow's milk, and to determine the fractions of the A1, A2, and B -casein variants in Icelandic and Scandinavian cow's milk.

	METHODS

Top Abstract Methods Results Discussion References

Subjects

Fifty-five children with IDDM in Iceland, born during a 16-year period, were included in the study. For each patient the Statistical Bureau of Iceland chose 3 children randomly of the same gender and with the same date of birth as a control group. At the time of this study the children were between 3 and 19 years old, the average age being 12.5 years. The Icelandic Data Protection Commission approved the study. The subjects and their parents gave informed consent for participation.

Design of Infant Diet Study

A retrospective case-control method was used in which a letter introducing the study was sent to mothers of all the children, followed by a telephone interview. The questions asked the mothers whether and how long they breastfed their children and when the children received a cow's milk product for the first time. It is known that mothers are very precise in remembering the feeding of their children during the first months of life, especially regarding breastfeeding.²⁵ Telephone surveys on food habits have proven to be a good method to obtain dietary information.²⁶

Milk Sample Collection

A total of 15 different batches or consumer packages of whole consumer cow's milk were randomly collected from the largest consumption areas in Iceland (6 batches) and Scandinavia (9 batches), ie, Denmark (2 batches), Norway (2 batches), Sweden (3 batches), and Finland (2 batches). Milk batches were collected in 9 different areas within the countries surveyed; 3 in Iceland and 6 in Scandinavia. For 3 areas in Iceland and 3 areas in Scandinavia, samples were collected both in autumn and spring, with a 7-month interval between collections. To ensure that the proteins were not degraded before analysis, all samples were kept chilled before freezing. The samples were then freeze-dried, coded, and sent to New Zealand where they were analyzed blind.

Milk Sample Analysis

The content of A1, A2, and B -casein variants were determined in samples from randomly collected batches of consumers cow's milk. Method of analysis was capillary electrophoresis, described by Recio et al²⁷ with modifications. Freeze-dried samples were reconstituted to 130 mg/mL with water and mixed for 3 hours. Exactly .5 mL of reconstituted sample was added to 2.5 mL of sample buffer (pH 8.6) containing 6 mol/L urea, 42 mmol/L 4-morpholinepropanesulfonic acid, .05% methyl cellulose, and 167 mM Tris(hydroxymethyl)aminomethane. -Mercaptoethanol (30 µL) was added to the sample and left for 1 hour at room temperature before filtering (.45 µm). Capillary electrophoresis runs were performed on an Applied Biosystems 270A-HT system (Applied Biosystems Inc, San Jose, CA) with a hydrophilic coated capillary of 57 cm × 50 µm (Supelco, Bellefonte, PA). Approximately 10 nL of sample was injected onto the capillary. The running buffer contained 6 M urea, 20 mM sodium citrate, .32 mM citric acid, and .05% methyl cellulose at pH 3.0. The content of A1, A2, and B -casein variants were given as relative proportions, ie, fractions.

Statistical Analysis

The results from the case-control study were compared using the ² test. The results from the analyses of Icelandic milk batches versus Scandinavian milk batches are described using mean, range, and standard error (SE). The Mann-Whitney U test was used to test the difference between the results of analyses of milk samples from Iceland and Scandinavia. All statistical analysis was performed with the SPSS statistical program (SPSS, Chicago, IL). P < .05 was regarded as statistically significant.

	RESULTS

Top Abstract Methods Results Discussion References

No significant difference was found between infants later developing IDDM and controls in exclusive breastfeeding or cow's milk consumption (P > .1). The frequency and duration of breastfeeding were similar in both groups (P > .1). Figure 1 shows the percentage of both groups of infants getting cow's milk at 1 to 6 months of age. There was no connection between the age of the child at the time of the study and the mother's memory regarding how long she breastfed her child, and when the child got cow's milk product for the first time. The mothers seemed to remember this very well.

View larger version (37K): [in this window] [in a new window]   Fig. 1.   Percentage of infants consuming cow's milk and cow's milk products, children who later became diabetic (light shading), compared with controls (dark shading).

The results from analyses of -casein variants in cow's milk from the largest consumption areas in Iceland and Scandinavia are shown in Table 1. A1 -casein and the sum of A1 and B are shown as the mean and range of analyzed batches from Iceland (6 batches) and Scandinavia (9 batches) as well as from each country of Scandinavia, ie, Denmark (2 batches), Norway (2 batches), Sweden (3 batches), and Finland (2 batches). All the Icelandic milk samples that were analyzed had a lower fraction of A1 -casein and a lower fraction of the sum of A1 and B -caseins than did milk samples from Scandinavia. Table 1 also shows earlier published values, based on either chemical analyses or estimations from genetic research.²⁴ As seen in the table, the earlier published values showed a greater discrepancy between Iceland and Scandinavia, both for A1 -casein and for the sum of A1 and B -caseins. The fraction of the B -casein variant in the milk from all countries in this current study was higher compared with the earlier published values. Figure 2 shows that the -casein fractions A1, the sum of A1 and B, and A2 in milk, collected in 6 different areas with a 7-month interval, were fairly constant for both seasons.

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Changes in fractions of A1, B, and A2 -caseins in cow's milk samples taken from the same areas in Iceland (dark lines) and Scandinavia (light lines) within a 7-month interval.

Figure 3 shows the fractions of A1 and B -casein variants in Icelandic and Scandinavian milk samples analyzed in this study. The fraction of A1 -casein in Icelandic milk was 35.6 ± .7 (mean ± SE), which is significantly lower than the fraction of this milk protein variant in Scandinavian milk, ie, 42.2 ± .6 (P < .02). This was also the case for the sum of A1 and B -caseins, both with histidine in position 67 in the primary protein sequence, which was 38.4 ± .8 (mean ± SE) for the Icelandic milk and 48.7 ± 1.3 for the Scandinavian milk (P < .02).

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Fraction (mean and SE) of the sum of A1 (dark shading) and B (light shading) -caseins found in Icelandic and Scandinavian milk (P < .02).

	DISCUSSION

Top Abstract Methods Results Discussion References

Breastfeeding and cow's milk consumption in infancy were similar among the IDDM patients and controls, which is in accordance with some previous studies.¹² However, other studies have found a shorter duration of breastfeeding or an earlier consumption of cow's milk in infancy among IDDM patients than among controls,^8-11,13 suggesting a causative relationship between cow's milk consumption and IDDM. This relationship seems, therefore, to exist in some countries and not in others, indicating that the cow's milk itself might contain different amounts of diabetogenic factors in different countries.

Animal studies show that A1 -casein is diabetogenic,¹⁵ and suggestions have been made on a similar effect from B -casein.^15,24 The analysis of cow's milk in the present study showed a significantly lower fraction of A1 and the sum of A1 and B -caseins in Icelandic milk than in milk from Scandinavia. This may explain the lower incidence of IDDM in Iceland than in Scandinavia. The mechanism behind the putative diabetogenicity of A1 and B -caseins is not fully known. The peptide BCM, produced in digestion of A1 and B -caseins, behaves like an opioid receptor ligand and is absorbed through the gastrointestinal mucosa.²⁸ BCM has been shown to be inhibitory to immune cell function in NOD mice and prediabetic humans but not in Swiss mice or normal humans.¹⁵ A sequence homology between -casein and several -cell molecules has also been suggested as an immune response trigger.¹⁹

In the present study the lower fraction of A1 -casein and the sum of A1 and B -caseins found in consumers cow's milk from Iceland versus that from Scandinavia are in accordance with results from genetic studies on cattle breeds.²² The current results also support recently published values by Elliott et al²⁴ who found an even larger difference between the values for Icelandic and Scandinavian milk by chemical analysis or estimation from genetic research. The milk samples taken from the same areas within a 7-month interval in the present study had a relatively constant -casein variant composition. However, it is likely that over relatively long periods, or decades, cattle breeding programs will have resulted in some variation in -casein fractions.²³

For the development of IDDM, the timing, dosage, and duration of the exposure to cow's milk could be important,²⁹ and some studies suggest that exposure during infancy may be important.¹⁴ A possible explanation of the significance of early introduction could be that an infant's intestines can absorb larger protein fragments from food than can the intestines of children and adults. Infants given hydrolyzed formula had depressed cellular and humoral responsiveness to cow's milk proteins, compared with a control group.³⁰ An international, double-blind study is currently in progress in which infants at high risk for IDDM are given hydrolyzed infant formula for the first 6 months of life to avoid intact dietary protein.³¹ Scientists have suggested that continuous exposure is needed for those genetically susceptible to progress to IDDM, and results from animal studies show that food can have a diabetogenic effect through puberty.²⁹ However, it should be kept in mind that most IDDM patients are diagnosed before adulthood, and the consumption in childhood, therefore, seems of greatest importance.

Dahl-Jörgensen and coworkers³² found a strong correlation between total cow milk consumption per capita and the incidence of IDDM in 12 countries, including the 4 Scandinavian countries, Denmark, Norway, Sweden, and Finland. A similar relationship was found for 9 regions in Italy.³³ This obviously contradicts the Icelandic situation because the cow's milk consumption per capita in Iceland is one of the highest in the world, whereas the incidence of IDDM is low.³⁴ However, the recent paper by Elliott et al²⁴ found a strong correlation between the per capita consumption of A1 and B -caseins and the incidence of IDDM in 10 countries, including Iceland and the 4 Scandinavian countries.

In this context it is important to evaluate the total exposure to milk, or especially to -casein A1 and B, among infants and children. The high per capita consumption of cow's milk in Iceland is similarly found in the youngest population, while the consumption of formula is lower than in the other Scandinavian countries.^35-39 In childhood the consumption of cow's milk is similar in Iceland and in the 4 Scandinavian countries.^40-44 The total exposure to cow's milk in young age is, therefore, not less in Iceland than in Scandinavia, but the total exposure to -caseins A1 and B seems lower. This has to be verified with calculations adding exact information from chemical analysis on -caseins A1 and B in formulas.

Genetically predisposed people can develop IDDM, but the majority of those with the so-called high-risk genes do not.¹⁴ Nations that are genetically similar but have very different incidences of IDDM support the hypothesis that environmental factors are contributing to the risk of IDDM. The Icelandic population has the same genetic origin as the nations of Scandinavia.⁶ However, studies have shown that the Finnish nation has a special genetic susceptibility to IDDM.^45,46 No major differences in human leukocyte antigen allelic distribution between Icelandic and Norwegian diabetic patients can be found,⁴⁷ but the incidence of IDDM in Norway is approximately twice as high as in Iceland.^1,2

Today IDDM is the second most common chronic childhood disease.⁴⁸ Very recently it has been reported that the incidence of IDDM in Finland in 1996 was 44.8 per 100 000 for 0- to 14-year-old children.⁵ The incidence is especially increasing among young children 1 to 4 years old. The incidence is also increasing in the other countries used for comparison with Iceland in the present study, ie, Denmark, Norway, and Sweden. All of these nations are related to the Icelandic nation, which has a much lower and more stable incidence of IDDM. The importance of explaining this difference is large. The fact that IDDM is one of the fastest growing diseases in the world has drawn attention to the importance of identifying possible environmental triggers.⁴⁸ Whether there is a single trigger or whether multiple environmental insults accumulate to cause the destruction of the -cells is not known.¹⁴ To focus solely on breastfeeding and early exposure to cow's milk is most likely an oversimplification,²⁹ but other factors thought to be connected to IDDM⁴⁸ are similar in Iceland and the Scandinavian countries, such as the prevalence of infections among small children, consumption of nitrates, and insecticides, as well as the per capita consumption of coffee and tea.^49-52 Even more factors could be mentioned. Substantial evidence indicates that cow's milk is important for the development of IDDM and that variability in the casein composition of cow's milk could, in part, explain varying results from the studies on the relationship between IDDM and cow's milk consumption.

Cow's milk consumption in infancy has no relation to IDDM in Iceland. The fraction of A1 and B -caseins in Icelandic cow's milk is significantly lower than in Scandinavian cow's milk, which may explain the lower incidence of IDDM in Iceland than in Scandinavia.

	ACKNOWLEDGMENTS

We thank Olafur Reykdal for his assistance in preparing the milk samples for analysis; Carmen Norris for assistance with capillary electrophoresis; and Dr Mike Boland for his advice during the preparation of the manuscript.

	FOOTNOTES

Received for publication Sep 20, 1999; accepted Feb 11, 2000.

Reprint requests to (I.T.) Unit for Nutrition Research, National University Hospital, IS-101, Reykjavik, Iceland. E-mail: ingathor{at}rsp.is

	ABBREVIATIONS

IDDM, insulin-dependent diabetes mellitus; NOD, nonobese diabetic; BCM, -casomorphin-7; SE, standard error.

	REFERENCES

Top Abstract Methods Results Discussion References

1. Helgason T, Danielsen R, Thorsson AV Incidence and prevalence of type 1 (insulin dependent) diabetes mellitus in Icelandic children, 1970-1989. Diabetologia 1992; 35:880-883 [CrossRef][Medline]
2. Joner G, S¢vik O Increasing incidence of diabetes mellitus in Norwegian children 0-14 years of age, 1973-1982. Diabetologia 1989; 32:79-83 [CrossRef][Medline]
3. Green A, Gale EAM, Patterson CC, for the EURODIAB ACE study group Incidence of childhood-onset insulin-dependent diabetes mellitus: the EURODIAB ACE study. Lancet 1992; 339:905-909 [CrossRef][Medline]
4. Nyström L, Dahlquist G, Rewers M, Wall S The Swedish childhood diabetes study: an analysis of the temporal variation in diabetes incidence, 1978-1987. Int J Epidemiol 1990; 19:141-146 [Abstract/Free Full Text]
5. Tuomilehto J, Karvonen M, Pitkaniemi J, Record-high incidence of type I (insulin-dependent) diabetes mellitus in Finnish children. The Finnish Childhood Type I Diabetes Registry Group. Diabetologia 1999; 42:655-660 [CrossRef][Medline]
6. Wijsman EM Techniques for estimating genetic admixture and application to the problem of the origin of the Icelanders and the Ashkenasi Jews. Hum Genet 1984; 67:441-448 [CrossRef][Medline]
7. Nordic Committee of Ministers. The Nordic Nutrition Recommendations. Copenhagen, Denmark; 1996:28
8. Borch-Johnsen K, Joner G, Mandrup-Poulsen T, Relation between breast-feeding and incidence rates of insulin-dependent diabetes mellitus: a hypothesis. Lancet 1984; 2:1083-1086 [Medline]
9. Mayer EJ, Hamman RF, Gay EC, Lezotte DC, Savitz DA, Klingensmith GJ Reduced risk of IDDM among breast-fed children. Diabetes 1988; 37:1625-1632 [Abstract]
10. Virtanen SM, Rasanen L, Aro A, Infant feeding in Finnish children <7 years of age with newly diagnosed IDDM. Diabetes Care 1991; 14:415-417 [Abstract]
11. Kostraba JN, Cruickshanks KJ, Lawler-Heavner J, Early exposure to cow's milk and solid foods in infancy, genetic predisposition and risk of IDDM. Diabetes 1993; 42:288-295 [Abstract]
12. Bodington M, McNally PG, Burden AC Cows' milk and type I childhood diabetes: no increase in risk. Diabet Med 1994; 11:663-665 [Medline]
13. Gerstein HC Cow's milk exposure and type I diabetes mellitus: a critical overview of the clinical literature. Diabetes Care 1994; 17:13-19 [Abstract]
14. Drash AL, Kramer MS, Swanson J, Udall JN Infant feeding practices and their possible relationship to the etiology of diabetes mellitus: work group on cow's milk protein and diabetes mellitus. Pediatrics 1994; 94:752-754 [Abstract/Free Full Text]
15. Elliott RB, Wasmuth HE, Bibby NJ, Hill JP. The role of -casein in the induction of insulin-dependent diabetes in the non-obese diabetic mouse and humans. In: Seminar on Milk Protein Polymorphism: IDF Special Issue 9702. Brussels, Belgium: International Dairy Federation; 1997:445-453
16. Elliott RB Dietary prospects of prevention of type 1 diabetes. Diabetes Nutr Metab 1989; 2:67-71
17. Karges W, Hammond-McKibben D, Cheung RK, Immunological aspects of nutritional diabetes prevention in NOD mice: a pilot study for the cow's milk-based IDDM prevention trial. Diabetes 1997; 46:557-564 [Abstract]
18. Hermitte L, Atlan-Gepner C, Payan MJ, Mehelleb M, Vialettes B Dietary protection against diabetes in NOD mice: lack of a major change in the immune system. Diabete Metab 1995; 21:261-268 [Medline]
19. Cavallo MG, Fava D, Monetini L, Barone F, Pozzilli P Cell-mediated immune response to -casein in recent-onset insulin-dependent diabetes: implications for disease pathogenesis. Lancet 1996; 348:926-928 [CrossRef][Medline]
20. Elliott RB, Wasmuth H, Hill J, Songini M, Bottazzo GF, for the Sardinian IDDM Study Groups Diabetes and cows' milk. Lancet 1996; 348:1657 [Medline]
21. Hartwig A, Teschenmacher H, Gauly M, Erhardt G. Influence of genetic variants in bovine milk on the occurrence of bioactive peptides. In: Seminar on Milk Protein Polymorphism: IDF Special Issue 9702. Brussels, Belgium: International Dairy Federation; 1997:459-460. Abstract
22. Lien S, Kantanen J, Eythorsdottir E, Milk protein allele frequency in Noric cattle breeds. Anim Genet 1999; 30:85-91 [CrossRef][Medline]
23. Jacques F, Klemetsdal G Nordic dairy cattle comparison for milk production by using a mixed linear model approach. Livestock Prod Sci 1990; 26:91-106 [CrossRef]
24. Elliott RB, Harris P, Hill JP, Bibby NJ, Wasmuth HE Type 1 (insulin dependent) diabetes mellitus and cow milk: casein variant consumption. Diabetologia 1999; 42:292-296 [CrossRef][Medline]
25. Eaton-Evans J, Dugdale AE Recall by mothers of the birth weight and feeding of their children. Hum Nutr Appl Nutr 1986; 40:171-175 [Medline]
26. Fox TA, Heimendinger J, Block G Telephone surveys as a method for obtaining dietary information: a review. J Am Diet Assoc 1992; 92:729-732 [Medline]
27. Recio I, Amigo L, Lopez-Fandieo R Assessment of the quality of dairy products by capillary electrophoresis of milk proteins. J Chromatogr B 1997; 697:231-242 [CrossRef]
28. Elitsur Y, Luk GD Beta-casomorphin (BMC) and human colonic lamina propria lymphocyte proliferation. Clin Exp Immunol 1991; 85:493-497 [Medline]
29. Scott FW, Cloutier HE, Kleemann R, Potential mechanisms by which certain foods promote or inhibit the development of spontaneous diabetes in BB rats: dose, timing, early effect on islet area, and switch in infiltrate from Th1 to Th2 cells. Diabetes 1997; 46:589-598 [Abstract]
30. Vaarala O, Saukkonen T, Savilahti E, Klemonla T, Akerblom HK Development of immune response to cow's milk proteins in infants receiving cow's milk or hydrolysed formula. J Allergy Clin Immunol 1995; 96:917-923 [CrossRef][Medline]
31. Schatz DA, Rogers DG, Brouhard BH Prevention of insulin-dependent diabetes mellitus: an overview of three trials. Cleve Clin J Med 1996; 63:270-274 [Medline]
32. Dahl-Jörgensen K, Joner G, Hanssen KF Relationship between cows' milk consumption and incidence of IDDM in childhood. Diabetes Care 1991; 14:1081-1083 [Abstract]
33. Fava D, David R, Leslie G, Pozzilli P Relationship between dairy product consumption and incidence of IDDM in childhood in Italy. Diabetes Care 1994; 17:1488-1490 [Abstract]
34. International Dairy Federation. Consumption statistics for milk and milk products 1993. Bull Int Dairy Federation. 1995;301
35. Atladóttir H. Research on Infant Nutrition in Iceland. Reykjavik, Iceland: Department of Food Science, Faculty of Science, University of Iceland; 1998. Master's thesis
36. Kjærnes U, Botten G, Lande B, Nilson D Food intake and patterns of feeding of Norwegian infants. Eur J Clin Nutr 1988; 42:249-260 [Medline]
37. Michaelsen KF Nutrition and growth during infancy: the Copenhagen cohort study. Acta Pædiatr Suppl 1997; 420:1-36
38. Kylberg E, Hofvander Y, Sjölin S Diets of healthy Swedish children 4-24 months old. Scand J Nutr 1984; 28:123-130
39. Ministry of Social Affairs and Health. Feeding of Infants in Finland 1995. Helsinki, Finland: Ministry of Social Affairs and Health; 1996
40. Steingrimsdottir L, Thorgeirsdottir H, Aegisdottir S. Food Habits of School Children, 1992-1993. Reykjarik, Iceland: Icelandic Nutrition Council IV; 1994
41. Frost-Andersen L, Nes M, Bjørneboe GEA, Drevon CA Food habits among 13-year-old Norwegian adolescents. Scand J Nutr 1997; 41:150-154
42. Andersen NL, Fagt S, Groth MV, et al. Danish Food Habits 1995: Methods and Results. Copenhagen, Denmark: National Food Administration; 1996
43. National Food Administration. Swedish Food Habits, 1989 (HULKEN): Methods and Results. Uppsala, Sweden: National Food Administration; 1994
44. Räsänen L, Ahola M, Kara R, Uhari M Atherosclerosis precursors in Finnish children and adolescents: food consumption and nutrient intakes. Acta Paediatr Scand Suppl 1985; 318:135-153 [Medline]
45. Tuomilehto-Wolf E, Tuomilehto J, Cepaitis Z, Lounamaa R New susceptibility haplotype for type 1 diabetes. DIME Study Group. Lancet 1989; 2:299-302 [Medline]
46. Reijonen H, Ilonen J, Knip M, Akerblom HK HLA-DQB1 alleles and absence of Asp 57 as susceptibility factors of IDDM in Finland. Diabetes 1991; 40:1640-1644 [Abstract]
47. Backman VM, Gulcher JR, Fasquel AC, Childhood diabetes in Iceland: genotyping of the DRB gene. Horm Res 1998; 50:109
48. Akerblom HK, Knip M Putative environmental factors in type 1 diabetes. Diab Metab Rev 1998; 14:31-67 [Medline]
49. Kristjansson I, Ardal B, Jonsson JS, Sigurdsson JA, Foldevi M, Björksten B Adverse reactions to food and food allergy in young children in Iceland and Sweden. Scand J Prim Health Care 1999; 17:30-34 [CrossRef][Medline]
50. Reykdal O, Olafsson G. Chemical Analysis of Nutrients and Nitrates in Icelandic Garden Vegetables. Reykjavik, Iceland: The Icelandic Agricultural Research Institute; 1988
51. Johannesdottir H. Insecticides in Icelandic Vegetables. Reykjavik, Iceland: The Environmental and Food Agency of Iceland; 1995
52. Nordic Council of Ministers. Yearbook of Nordic Statistics 1999. Copenhagen, Denmark: Nordic Council of Ministers; 1999