Intake of vitamin E and other antioxidant nutrients in early life and the development of advanced ?-cell autoimmunity and clinical type 1 diabetes

Abstract

E-vitamiinin ja muiden antioksidantteina toimivien ravintoaineiden on oletettu suojaavan tyypin 1 diabeteksen kehittymiseltä. ETM Liisa Uusitalon väitöskirjassa selvitettiin veren E-vitamiinipitoisuuden yhteyttä tyypin 1 diabeteksen kehittymiseen lapsilla. Lisäksi tutkittiin äidin raskaudenaikaisen antioksidantteina toimivien ravintoaineiden saannin yhteyttä tyypin 1 diabeteksen esiasteeseen lapsella. Esiasteen määritelmä perustuu tyypin 1 diabetekseen kytkeytyvien autovasta-aineiden toteamiseen lapsen verestä, ja se merkitsee huomattavaa riskiä sairastua myöhemmin tyypin 1 diabetekseen. Tutkimuksen perusteella veren korkea E-vitamiinipitoisuus ei näytä suojaavan lapsia tyypin 1 diabeteksen esiasteelta. Vahvaa näyttöä ei saatu myöskään siitä, että veren korkea E-vitamiinipitoisuus suojaisi kliinisen tyypin 1 diabeteksen puhkeamiselta. Äidin runsas A-vitamiinin, beetakaroteenin, C-vitamiinin, E-vitamiinin, seleenin, sinkin tai mangaanin saanti odotusaikana ei pienentänyt lapsen riskiä saada esidiabetes ensimmäisten elinvuosien aikana. Odottavilla äideillä esiintyi suurta vaihtelua antioksidantteina toimivien ravintoaineiden saannissa. Nuori ikä, alhainen koulutustaso ja raskaudenaikainen tupakointi olivat tärkeimmät alhaista saantia ennustavat tekijät.

Väitöskirjatyö toteutettiin kahden suomalaisen seurantatutkimuksen aineistoilla. Lasten diabetes Suomessa (DiMe) tutkimukseen kuului tyypin 1 diabetekseen sairastuneiden lasten sisaruksia, ja Diabeteksen ennustaminen ja ehkäisy (DIPP) tutkimuksen aineisto muodostui tyypin 1 diabetekselle geneettisesti alttiista lapsista, joita seurattiin syntymästä lähtien.

Free radicals, highly reactive oxygen or nitrogen compounds at high concentrations are harmful to human health. They are believed to play a role in the development of all the major degenerative diseases, and may also be involved in the development of type 1 diabetes. Antioxidants are substances with an ability to neutralize free radicals. Antioxidant defense system of the body consists of antioxidative enzymes, other endogenously synthetized compounds and dietary antioxidants. Vitamin E, vitamin C and β-carotene are the classicl dietary antioxidants. Diet also provides the trace elements needed as cofactors of antioxidative enzymes. The effect of antioxidant nutrients on the development of type 1 diabetes has been studied mainly in animal models, and research has concentrated on vitamin E and zinc. Protective effects against diabetes have been reported for both. Epidemiological studies on the associations of antioxidant nutrients with type 1 diabetes are rare.

The present study falls into the field of nutritional epidemiology. It is based on epidemiological analyses of the associations of vitamin E and other antioxidant nutrients with the development of type 1 diabetes. The specific research objectives were: to analyze whether serum vitamin E concentrations are associated with the risk of pre-type 1 diabetes, defined as repeated positivity for ICA and at least one other disease-predictive autoantibody, and with clinical type 1 diabetes; to analyze whether maternal intake of vitamin E or other selected antioxidant nutrients or the consumption of their dietary sources during pregnancy is associated with pre-type 1 diabetes in the child; and to analyze the sociodemographic and lifestyle determinants of intake of antioxidant nutrients and consumption of their dietary sources among Finnish pregnant women.

The data are derived from two large Finnish cohort studies: the cohort of the Childhood Diabetes in Finland (DiMe) study, which consists of unaffected siblings of children with diabetes, and the Type 1 Diabetes Prediction and Prevention (DIPP) project, which is a birth cohort of children with HLA-conferred susceptibility to type 1 diabetes. The associations of serum vitamin E concentration and the risk of clinical type 1 diabetes were analyzed within the DiMe cohort. Two separate study designs were used. In a nested case-control study, α-tocopherol concentrations of cases with type 1 diabetes (n=16) were compared with the concentrations of matched controls (n=81). In the second study design, the association of α-tocopherol with type 1 diabetes was analyzed in a subcohort of seropositive siblings (n=80). During follow-up, 26 siblings progressed to type 1 diabetes. A conditional logistic regression model was employed in the case-control setting, and a Cox regression model in the cohort setting.

The associations of serum α- and γ-tocopherol concentrations and preclinical type 1 diabetes were analyzed within the DIPP cohort using a nested case-control study design. There were 108 cases and 216 matched controls with at least one serum sample available. For each case-control set, all the repeated serum samples, collected at 1-year intervals starting from the age of 1 year, were analyzed up to the age of seroconversion of the case. A conditional logistic regression model was used to compare vitamin E concentrations of cases and controls.

Maternal antioxidant intake during pregnancy and the consumption of their dietary sources, and their associations with preclinical type 1 diabetes in the child were studied on the DIPP study using a cohort design. The series comprised children with HLA-conferred diseases susceptibility who participated in the follow-up for the emergence of diabetes-associated autoantibodies and/or overt type 1 diabetes and their mothers. Dietary intake during pregnancy was assessed using a validated food frequency questionnaire. During follow-up, 165 children of the total of 4,297 children in the cohort were affected with preclinical and/or clinical type 1 diabetes. The associations of maternal diet with pre-type 1 diabetes in the child were analyzed by piece-wise exponential survival models. The associations of sociodemographic and lifestyle characteristics with antioxidant intake among the 3,730 mothers were analyzed by multiple linear regression models, and those with food consumption by multiple logistic regression models.

Serum α-tocopherol concentrations analyzed within the cohort of initially non-diabetic siblings showed an inverse association of borderline statistical significance with clinical type 1 diabetes both in the nested case-control design (p=0.08) and in the subcohort of seropositive children (p=0.09), adjusted for potential confounding factors. In the birth cohort of genetically susceptible children, serum - or γ-tocopherol concentrations were not significantly associated with the risk of pre-type 1 diabetes. Neither were the maternal intakes of vitamin E, retinol, β-carotene, vitamin C, selenium, zinc or manganese, or the consumption of their main dietary sources during pregnancy associated with the risk of pre-type 1 diabetes in the offspring. There was considerable variation in the intake of antioxidant nutrients among pregnant women, and all the sociodemographic and lifestyle factors studied (age, parity, own and partner s level of education, smoking during pregnancy, and BMI at the first antenatal visit) were independently associated with antioxidant intake. Young age, low level of education, and smoking during pregnancy were the most important predictors of a low intake of antioxidant nutrients.

As a conclusion, a protective effect of vitamin E against clinical type 1 diabetes seems possible, and further studies are warranted. No evidence was found of a protective effect of vitamin E against preclinical type 1 diabetes. In future research, the associations of dietary vitamin E intake with the development of type 1 diabetes should be analyzed. Methodological studies are needed to identify the optimal methods for the measurement of true long-term vitamin E intake. Maternal intake of antioxidant nutrients during pregnancy was also not associated with pre-type 1 diabetes in the offspring. The absence of association could result from the short period of time when the newborn infant is dependent on the maternal antioxidant intake, and therefore it is advisable to next analyze the associations of the child s own antioxidant intake with the development of clinical and preclinical type 1 diabetes.