Type I (insulin-dependent) diabetes mellitus results Type I diabetes is a multifactorial disease but the diabetes before the age of 15 have been …
Diabetologia 2000 43: 10931098 Ó Springer-Verlag 2000
Articles
Use of cod liver oil during pregnancy associated with lower risk of Type I diabetes in the offspring
L C Stene1, J Ulriksen2, P Magnus1, G Joner3
1 2
Section of Epidemiology, National Institute of Public Health, Oslo, Norway Department of Paediatrics, Vest-Agder Central Hospital, Kristiansand, Norway 3 Aker Diabetes Research Centre, Aker University Hospital, Oslo, Norway
Abstract Aims/hypothesis To test whether cod liver oil or vitamin D supplements either taken by the mother during pregnancy or by the child in the first year of life is associated with lower risk of Type I insulin-dependent diabetes mellitus in children Methods We carried out a population-based case control study in Vest-Agder county of Norway, evaluating the use of supplements by a mailed questionnaire We received responses from 85 diabetic subjects and 1071 control subjects Odds ratios OR with 95 confidence intervals CI were estimated using logistic regression analyses Results When mothers took cod liver oil during pregnancy their offspring had a lower risk of diabetes The unadjusted OR was 030, 95 CI: 012 to 075, p 001 This association changed
very little and was still significant after adjusting for age, sex, breast-
feeding and maternal education Mothers taking multivitamin supplements during pregnancy [adjusted OR 111, 95 CI: 069 to 177], infants taking cod liver oil in the first year of life [adjusted OR 082, 95 CI: 047 to 142 and the use of other vitamin D supplements in the first year of life [adjusted OR 127, 95 CI: 070 to 231] was significantly associated with the risk of diabetes Conclusion/interpretation We found that cod liver oil taken during pregnancy was associated with reduced risk of Type I diabetes in the offspring This suggests that vitamin D or the n-3 fatty acids eicosapentaenoic acid and docosahexaenoic acid in the cod liver oil, or both, have a protective effect aginst Type I diabetes [Diabetologia 2000 43: 10931098] Keywords Type I insulin-dependent diabetes mellitus, diet, vitamin D, case control study, aetiology
Type I insulin-dependent diabetes mellitus results from the destruction of the insulin-producing beta cells of the pancreas, a disease process that can last several years before the clinical onset of the disease [1] Type I diabetes is a multifactorial disease but the factors
initiating and propagating the process leading to clinical disease are mostly not known Genetic factors in the HLA complex play an important part in its
Received: 19 April 2000 and in revised form: 6 June 2000 Corresponding author: L C Stene, Section of Epidemiology, National Institute of Public Health, P O Box 4404 Torshov, N-0403 Oslo, Norway Abbreviations: OR, Odds ratio; EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid
pathogenesis, probably together with environmental factors such as diet or virus infections [1,2] The hypothesis proposing that breastfeeding could reduce the risk of Type I diabetes [3] has stimulated studies focusing on the possible causal role of infant diet in Type I diabetes The role of breastfeeding and early introduction of cows milk is still controversial [46] but there are good reasons to believe that environmental risk factors could operate early in life, possibly in utero [7] Recently, knowledge of untraditional effects of vitamin D has been applied to Type I diabetes [8] Vitamin D has several effects on the immune system which could be of potential relevance in the pathogenesis of Type I diabetes [811] and has been shown to protect against
autoimmune diabetes in experimental animals [12,13] Furthermore, a previous
1094
L C Stene et al: Cod liver oil and Type I diabetes responded Written informed consent was obtained from the parents of all participants and from all participants able to write The study protocol was approved by the Regional Ethics Committee and by the National Data Inspectorate Evaluation of dietary supplements and potential confounders Diabetic and control subjects received a four-page questionnaire by mail The questionnaire included general questions on sex, age, and on the childrens health during childhood and the mothers health during pregnancy The participants were informed that the causes of childhood diabetes are mostly not known and that we were studying possible causes of diabetes The mother of the index child was asked to complete the questionnaire, preferably together with the father and the index child The questionnaire asked whether the participants mother took cod liver oil or vitamin/mineral supplements during pregnancy with the participant A negative or affirmative answer could be ticked for cod liver oil, multivitamin and other to be specified In a section devoted to the participants
dietary habits in the first year of life, the questionnaire asked whether the participant received cod liver oil or other vitamin D supplements during the first year of life A negative or affirmative answer could be ticked for cod liver oil, multivitamin with examples of the most common brand names and other to be specified Those who ticked the affirmative for cod liver oil or vitamin supplement were considered exposed to the respective factor Those who specified another preparation containing vitamin D were also considered exposed to vitamin D supplement The questionnaire also contained questions on the duration of exclusive breastfeeding and maternal educational level at the time of birth of the index child Data analysis The data from all questionnaires were optically scanned and checked for errors We used SPSS for Windows, release 90 for data analysis Information was missing for one diabetic 12 and 24 control subjects 22 on the mothers use of cod liver oil or multivitamin supplements during pregnancy and for 7 diabetic 82 and 91 control subjects 85 on the use of cod liver oil or other vitamin D supplements in the first year of life These were excluded from the respective
analyses Odds ratios OR with 95 confidence intervals CI associated with each exposure were estimated using logistic regression, adjusting for age and other potential confounding factors [26] An OR less than 10 indicates a negative association, an OR more than 10 a positive association and an OR of 10 a lack of association The relation between age at the time of data collection and risk of Type I diabetes was approximately logit-linear and age was entered into the logistic model as a continuous variable Separate analyses stratified by age group in four categories were also done Breastfeeding practices and maternal education were adjusted for by entering them into the logistic model as binary variables exclusive breastfeeding 3 months vs 3 months and maternal education 12 years vs 12 years For evaluation of one supplement e g cod liver oil, use of other supplements e g multivitamin was also evaluated as a confounder Analyses were also adjusted for sex A two-sided significance level less than 005 was regarded as statistically significant With the given sample size, we had 52 power to obtain a significant result for exposures with a prevalence of 20 among control subjects if the
true OR was 05, and 84 power if the true OR was 03
multicentre case control study has shown that vitamin D supplements taken in the first year of life is associated with reduced risk of Type I diabetes [14] Vitamin D is produced in the skin when exposed to ultra-violet radiation from sunlight [15] Exposure to sunlight in northern countries is, however, below optimal for infants and possibly for pregnant women [16] In Norway, dietary vitamin D supplement is recommended from infancy, preferably in the form of cod liver oil [17] This is mainly to prevent vitamin D deficiency that causes rickets Cod liver oil is specifically recommended because it also contains the long-chain n-3 fatty acids eicosapentaenoic acid EPA and docosahexaenoic acid DHA that have been regarded to be essential for infants [17,18] Dietary long-chain n-3 fatty acids are incorporated in cell membranes and have anti-inflammatory effects which could be potentially relevant in the aetiology of Type I diabetes [19,20] Because breast-milk is a relatively poor source of vitamin D, the vitamin D status of the newborn child is dependent on stored vitamin D acquired from the mother in utero or dietary supplements before
weaning [15] The vitamin D intake of the mother during pregnancy is reflected by the vitamin D status of the newborn child [16,21,22] Therefore, if vitamin D has a protective effect against Type I diabetes in children, it is possible that the mothers use of vitamin D supplements during pregnancy is associated with a reduced risk of Type I diabetes in their offspring The objective of our study was to estimate the associations between the use of cod liver oil or other vitamin D supplements, either by the mother during pregnancy or by the child in the first year of life and the risk of childhood-onset Type I diabetes in the children
Subjects and methods
Subjects In Norway, all patients newly diagnosed with Type I diabetes before the age of 15 have been registered in the National Childhood Diabetes Registry with a high level of ascertainment between 1973 and 1982 [23] Since 1 January 1989, all new cases of diabetes have been prospectively registered, still with a high level of ascertainment [24] We followed the diagnostic criteria of EURODIAB [25] Based on this prospective registration, we designed a population-based case control pilot study in Vest-Agder county of Norway Vest-Agder
is the southernmost county of Norway, with a population of 154 000 in 1998 This corresponds to 35 of the population of Norway All patients diagnosed with Type I diabetes before the age of 15 in Vest-Agder county between 1989 and 1998 and who were born between 1982 and 1998 were eligible to participate Control subjects were selected at random from the official population register among residents in Vest-Agder county at the end of 1998 who were born between 1982 and 1998 All participants were contacted by mail in May 1999 After one reminder, 85 diabetic subjects 944 of those invited and 1071 control subjects 727 of those invited had
L C Stene et al: Cod liver oil and Type I diabetes
1095
Table 1 Distribution of age, sex, breastfeeding practice and maternal education for subjects with Type I diabetes and population controls selected at random in Vest-Agder county, Norway SD standard deviation
Variable Age mean, SD Male sex Exclusive breastfeeding 3 monthsb Maternal education 12 yearsc
a b
a
Cases n 85 111 36 54 635 22 259 33 393
c
Controls n 1071 8 5 47 522 487 258 242 521 494
Odds ratio 95 CI 18 12, 29 11 07, 18 07 04, 10
Age at data collection There were four
controls with missing information for breastfeeding practice
Highest attained education at birth of the index child there were one case and 17 controls with missing information for maternal education
Table 2 Estimated odds ratios for Type I diabetes in children associated with mothers use of dietary supplements during pregnancy
Type of supplement used by mother in pregnancy Cod liver oil Yes No Multivitamin Yes No
a c
Cases n 84 5 79 32 52
Controls n 1047 182 865 388 659
Odds ratio 95 CI Unadjusted 030 012, 075a 105 066, 165 Adjustedc 036 014, 090b 111 069, 177
p 001; b p 003; Adjusted for age, sex, breastfeeding, maternal education and other supplement use of cod liver oil adjusted for use of
multivitamin and vice versa Six individuals were excluded due to missing information on maternal education or breastfeeding
Results
Background variables On average the diabetic were older than the control subjects mean age 111 vs 85 years There were more males among diabetic than among control subjects Table 1 The proportion of participants who had mothers with a high level of education was slightly but not statistically significantly lower among diabetic than among control
subjects Maternal use of supplements in pregnancy We found a strong negative association between mothers taking cod liver oil during pregnancy and the risk of Type I diabetes in their children, with an OR of 03 Table 2 This association was only slightly changed and still significant after adjustment for age, sex, maternal use of multivitamins, duration of exclusive breastfeeding and mothers educational level There was no significant association between the mothers use of multivitamins during pregnancy and the risk of Type I diabetes, neither crude nor after adjustment for possible confounders Odds ratios were homogeneous in all age groups data not shown Because a few mothers had used both cod liver oil and multivitamins during pregnancy, an analysis considering the single and combined effects of cod liver oil and multivitamins was done The odds ratio for each type of supplement alone and that for using both types of supplements together were estimated relative to using neither Table 3 This showed that the
association between taking of cod liver oil during pregnancy and the risk of diabetes was independent of taking multivitamin supplements duringpregnancy Use of supplements in the
first year of life In the first year of life neither use of cod liver oil nor use of other vitamin D supplements showed statistically significant associations with the risk of diabetes Table 4 Because a few children had used both cod liver oil and other vitamin D supplements in the first year of life, an analysis was done to estimate the single and combined associations Table 5 This indicated a possible negative association with cod liver oil alone adjusted OR 045 and had a complete lack of association with other vitamin D supplements adjusted OR 100 There was some variation in the size of the odds ratios in different age groups, but this was not systematic and not statistically significant data not shown
Discussion
In Norway cod liver oil is an important dietary source of vitamin D [27] and the long-chain n-3 fatty acids eicosapentaenoic acid EPA and docosahexaenoic acid DHA [28] The status of vitamin D [15,16,22] and of EPA and DHA [18] in the mother during pregnancy is reflected by the status of these nutrients in the newborn child
1096
L C Stene et al: Cod liver oil and Type I diabetes
Table 3 Analysis of single and joint effects of mothers taking cod liver oil and
multivitamin supplements during pregnancy on the risk of childhood-onset Type I diabetes in the children
Type of dietary supplement used by mother in pregnancy None Cod liver oil Multivitamin Cod liver oil and multivitamin Cases n 84 49 3 30 2 Controls n 1047 551 108 314 74 Odds ratio 95 CI Unadjusted 100 reference 031 010, 102 107 067, 173 030 007, 128 Adjusteda 100 reference 038 012, 127 112 069, 183 036 008, 153
a Adjusted for age, sex, breastfeeding and maternal education Six individuals were excluded from the adjusted analysis due to missing information on maternal education or breastfeeding
Table 4 Estimated odds ratios for Type I diabetes associated with taking cod liver oil and other vitamin D supplements in the first year of life
Type of supplement used in the first year of life Cod liver oil Yes No Vitamin Db Yes No Cases n 78 Controls n 980 Odds ratio 95 CI Unadjusted 19 59 63 15 330 650 709 271
b
Adjusteda 082 047, 142 127 070, 231
063 037, 108 160 090, 286
a Adjusted for age, sex, breastfeeding, maternal education and other supplement use of cod liver oil adjusted for use of multivitamin and vice versa
May contain other vitamins in addition to vitamin
D
We did not observe a statistically significant association between the mothers use of multivitamin supplements during pregnancy and the risk of diabetes in their children Assuming that vitamin D was responsible for the association between the mothers taking cod liver oil and the risk of diabetes in children, this could seem inconsistent A possible explanation is that the bioavailability of vitamin D could be better from cod liver oil than from other supplements In a prospective study of more than 300 women and their children born in Oslo, it was found that newborn children of mothers who had taken cod liver oil during pregnancy had statistically significantly higher concentrations of 25-OH-vitamin D in the cord blood than children of mothers who had taken other vitamin D supplements during pregnancy Gry Hay et al, Institute for Nutrition Research, University of Oslo, unpublished data It should be added, however, that the pregnant women taking cod liver oil in that study took amounts providing statistically significantly higher doses of vitamin D than that provided from other vitamin D supplements in those who took other supplements during pregnancy Another possible explanation
is that there was some misclassification of intake of vitamin D based on the multivitamin information Although most multivitamin supplements commercially available in Norway contain vitamin D, not all the participants in our study specified the type of multivitamin supplement they had used According to nutritionists experienced in interview-based assessments of dietary intake, the
concept of multivitamin supplements is not clear to some people, whereas the use of cod liver oil is always easy to estimate In addition, because our data are compatible with an effect of cod liver oil and not with other vitamin D supplements, we must consider the possibility that constituents in cod liver oil other than vitamin D are responsible for the negative association with Type I diabetes in children, perhaps in combination with vitamin D Both EPA and DHA have anti-inflammatory properties of potential relevance in the pathogenesis of Type I diabetes For instance, supplementation with dietary long-chain n-3 fatty acids has been shown to decrease the production of interleukin-1 and tumour necrosis factor by human mononuclear cells [19] and to decrease the expression of HLA class II molecules on
activated human monocytes [20] An effect of marine fatty acids is also compatible with the observation of a statistically significantly lower incidence of Type I diabetes in municipalities in Norway which are fisheries than in the rest of the country [29] Although there is little direct evidence from previous studies in humans or animals to support a protective role of long-chain n-3 fatty acids in the pathogenesis of Type I diabetes, this possibility should be considered We found that the associations between taking cod liver oil or other vitamin D supplements in the first year of life and the risk of Type I diabetes were not statistically significant Stratification by use of other supplements Table 5 indicated a tendency for a negative association between taking cod liver oil in the
L C Stene et al: Cod liver oil and Type I diabetes
1097
Table 5 Analysis of single and joint effects of use of cod liver oil and other vitamin D supplements in the first year of life on the risk of childhood-onset Type I diabetes
Type of supplement used in the first year of life None Cod liver oil Vitamin Da Cod liver oil and vitamin D
a b
Cases n 78 11 4 48 15
Controls n 980 139 132 511
198
Odds ratio 95 CI Unadjusted 100 reference 038 012, 123 119 060, 235 096 043, 215 Adjustedb 100 reference 045 014, 147 100 050, 200 098 043, 223
May contain other vitamins in addition to vitamin D Adjusted for age, sex, breastfeeding and maternal education Six individuals were excluded from the adjusted analysis
due to missing information on maternal education or breastfeeding
first year of life and the risk of diabetes and a lack of such an association for other vitamin D supplements Although speculative, this suggests a similar relation between cod liver oil and other supplements taken by the child in the first year of life to that of the mother taking those supplements during pregnancy, and the above mentioned possible explanations could also be relevant here A large multicentre case control study in seven European countries, found that vitamin D supplements taken in the first year of life was significantly associated with reduced risk for childhood Type I diabetes OR 067 but with some differences between the centres [14] No conclusion could, however, be drawn from our data on a possible relation between the use of vitamin D supplements in the first year of life and the
risk of Type I diabetes A number of precautions are warranted when making inferences from the data of this study Although the negative association between the mothers use of cod liver oil and diabetes in their children was strong and statistically significant, the absolute number of those exposed to the supplements was small and the confidence intervals for the odds ratios were relatively wide The assessment of vitamin D and cod liver oil intake was relatively crude We did not attempt to estimate the amount or duration of supplementation Because the mean dietary intake of vitamin D among pregnant women when excluding supplements is approximately a third of that provided by the recommended daily dose of cod liver oil and other supplements [30], there is good reason to believe that those who took supplements had a higher intake of vitamin D than those who did not A similar argument is probably valid for the intake of the n-3 fatty acids EPA and DHA in cod liver oil In our study, we selected two exposure windows for the children, in utero and in the first year of life, respectively The first year of life was also used in the EURODIAB study [14] and roughly corresponds to the period
when children are most prone to vitamin D deficiency [15] In addition, our selected exposure windows are in accordance with the concept that the effect of environmental factors in the development of Type I diabetes is important early in life, possibly in utero [7]
The time period of exposure to vitamin D, EPA and DHA that is biologically most relevant in our context is not known but our data suggest that exposure in utero could be relevant for the development of Type I diabetes There could have been some degree of misclassification of exposure due to inaccurate recall and the possibility of recall bias should be considered If recall bias was responsible for the observed association between the mothers use of cod liver oil and the risk of diabetes, we would expect the association to be stronger among those whose length of recall was longest assuming homogeneous effect over age groups The strength of association was, however, homogeneous in different age groups Furthermore, we would expect the extent of differential recall between mothers of diabetic and those of control children to be similar for use of cod liver oil and multivitamins during pregnancy but there was no association
with multivitamins In addition, the possible association of vitamin D or cod liver oil with Type I diabetes is not known by the public It therefore seems unlikely that recall bias is responsible for the majority of the observed association between the mothers use of cod liver oil and the risk of diabetes in their children Although the analyses were adequately adjusted for age, sex, duration of breastfeeding and maternal education, we cannot rule out the possibility that other extraneous factors are associated with both the use of cod liver oil and the low risk of Type I diabetes It is, however, difficult to envisage other factors that could have a sufficiently strong association with both exposure and disease to account for the observed relation Our study showed a negative association between use of cod liver oil in pregnancy and the subsequent risk of Type I diabetes in the children before they reach 15 years of age Plausible biological mechanisms could explain a potential causal effect of vitamin D or long-chain n-3 fatty acids present in the cod liver oil, or both
Acknowledgements G Joner and L C Stene were supported by a grant from the Norwegian Foundation for Health and
Re-
1098 habilitation grant no 1997/156 Funding was also kindly provided by TINE Norwegian Dairies and Novo Nordisk Pharma A/S We are grateful for the assistance and support from research nurse U T Hinna, Dr P Vesterhus, Professor O Sùvik and the Norwegian Childhood Diabetes Study Group
L C Stene et al: Cod liver oil and Type I diabetes 16 Markestad T 1983 Effect of season and vitamin D supplementation on plasma concentrations of 25-hydroxyvitamin D in Norwegian infants Acta Paediatr Scand 72: 817821 17 National Nutrition Council 1994 National Nutrition Councils recommendations for infant feeding in Norwegian National Nutrition Council, Oslo 18 Nettleton JA 1993 Are n-3 fatty acids essential nutrients for fetal and infant development? J Am Diet Assoc 93: 5864 19 Endres S, Ghorbani R, Kelley VE et al 1989 The effect of dietary supplementation with n-3 polyunsaturated fatty acids on the synthesis of interleukin-1 and tumor necrosis factor by mononuclear cells N Engl J Med 320: 265271 20 Hughes DA, Pinder AC 2000 n-3 polyunsaturated fatty acids inhibit the antigen-presenting function of human monocytes Am J Clin Nutr 171 [Suppl]: 357S-360S 21 Delvin EE, Salle BL, Glorieux FH,
Adeleine P, David LS 1986 Vitamin D supplementation during pregnancy: effect on neonatal calcium homeostasis J Pediatr 109: 328334 22 Zeghoud F, Vervel C, Guillozo H et al 1997 Subclinical vitamin D deficiency in neonates: definition and response to vitamin D supplements Am J Clin Nutr 65: 771778 23 Joner G, Sùvik O 1989 Increasing incidence of diabetes mellitus in Norwegian children 014 years of age 19731982 Diabetologia 32: 7983 24 EURODIAB ACE Study Group 2000 Variation and trends in incidence of childhood diabetes in Europe Lancet 355: 873876 25 Green A, Gale EA, Patterson CC 1992 Incidence of childhood-onset insulin-dependent diabetes mellitus: the EURODIAB ACE study Lancet 339: 905908 26 Breslow NE, Day NE 1980 Statistical methods in cancer research, Vol 1, the analysis of case-control studies IARC Scientific Publications No 32 International Agency for Research on Cancer, Lyon 27 Johansson L, Solvoll K, Bjùrneboe GE, Drevon CA 1997 Dietary habits among Norwegian men and women Scand J Nutr 41: 6370 28 Johansson L, Solvoll K, Bjùrneboe GE, Drevon CA 1998 Intake of very-long-chain n-3 fatty acids related to social status and lifestyle Eur J Clin Nutr 52: 716721 29 Joner G 1992
The epidemiology of insulin-dependent diabetes mellitus in Norway: incidence, prevalence, microvascular complications and mortality [Thesis] Haukeland University Hospital, and Aker University Hospital, Oslo 30 Trygg K, Lund-Larsen K, Sandstad B et al 1995 Do pregnant smokers eat differently from pregnant non-smokers? Paediatr Perinat Epidemiol 9: 307319
References
1 Atkinson MA, Maclaren NK 1994 The pathogenesis of insulin-dependent diabetes mellitus N Engl J Med 24: 14281436 2 Nerup J, Mandrup-Poulsen T, Helqvist S et al 1994 On the pathogenesis of IDDM Diabetologia 37 [Suppl 2]: S82-S89 3 Borch-Johnsen K, Joner G, Mandrup-Poulsen T et al 1984 Relation between breast-feeding and incidence rates of insulin-dependent diabetes mellitus: a hypothesis Lancet ii: 10831086 4 kerblom HK, Knip M 1998 Putative environmental factors in Type 1 diabetes Diabetes Metab Rev 14: 3167 5 Atkinson MA, Ellis TM 1997 Infants diets and insulindependent diabetes: evaluating the cows milk hypothesis and a role for anti-bovine serum albumin immunity J Am Coll Nutr 16: 334340 6 Scott FW, Kolb H 1996 Cows milk and insulin-dependent diabetes mellitus Lancet 348: 613 7 Leslie DG, Elliott RB 1994 Early
environmental events as a cause of IDDM Diabetes 43: 843850 8 Mauricio D, Mandrup-Poulsen T, Nerup J 1996 Vitamin D analogues in insulin-dependent diabetes mellitus and other autoimmune diseases: a therapeutic perspective Diabetes Metab Rev 12: 5768 9 Rigby WF, Yirinec B, Oldershaw RL, Fanger MW 1987 Comparison of the effects of 1,25-dihydroxyvitamin D3 on T lymphocyte subpopulations Eur J Immunol 17: 563566 10 Rigby WF 1988 The immunobiology of vitamin D Immunol Today 9: 5458 11 Rabinovitch A, Skyler JS 1998 Prevention of type 1 diabetes Med Clin North Am 82: 739755 12 Mathieu C, Laureys J, Sobis H et al 1992 1,25-Dihydroxyvitamin D3 prevents insulitis in NOD mice Diabetes 41: 14911495 13 Mathieu C, Waer M, Laureys J, Rutgeerts O, Bouillon R 1994 Prevention of autoimmune diabetes in NOD mice by 1,25 dihydroxyvitamin D3 Diabetologia 37: 552558 14 EURODIAB Substudy 2 Study Group 1999 Vitamin D supplement in early childhood and risk for Type I insulindependent diabetes mellitus Diabetologia 42: 5154 15 Fraser DR 1995 Vitamin D Lancet 345: 104107
Source:mathematica-mpr.com
