Predisposition to Atopic Symptoms to Inhaled Antigens May Protect From Childhood Type 1 Diabetes

A questionnaire was sent to 819 probands with type 1 childhood diabetes (aged ≤14 years at the time of diagnosis), their 1,082 siblings, and 819 age- and sex-matched unaffected population control subjects in 1998. They had previously served as a study cohort in another study on diabetes epidemiology (8). The ages of the probands and unaffected population control subjects ranged from 10 to 27 years at the time the questionnaire was mailed. Altogether, 347 (42%) probands, 620 (57%) siblings of the probands, and 483 (59%) unaffected population control subjects returned completed questionnaires. At the time of the survey, the mean age of the probands was 18.4 years (ranging from 11 to 26 years), and the mean age of the siblings and control subjects was 19.2 years (ranging from 2 to 43 years) and 18.3 years (ranging from 10 to 27 years), respectively. In the analysis, those siblings (n = 19) who had been diagnosed to have diabetes later than the proband, as verified by the records of the National Diabetes Registry, were included in the analysis as a proband.

The following questions on asthma, allergy to inhaled antigens, and adenoidectomy were asked: 1) Have you been diagnosed to have asthma by a physician? 2) At what age was your asthma diagnosed? 3) Have you ever been diagnosed to have any allergy to the following airborne allergens: a) animal dust (dog, cat, horse, etc.) and b) pollen? 4) Have your adenoids been removed (adenoidectomy)? and 5) If adenoidectomy was performed, at what age was it done?

The questions were followed by multiple-choice answer fields including “no,” “yes,” and “do not know” choices in applicable places. The questionnaire also included a picture with explanatory text showing the anatomical location of the adenoids and tonsils to help the respondent understand the questions concerning adenoidectomy. The survey was approved by the Ethical Review Committee of the Finnish National Health Institute, Helsinki. Multivariate logistic regression analysis was performed to evaluate independent risk factors for diabetes.

The frequencies of asthma and allergy to the inhaled antigens, animal dust and pollen, among the probands, siblings, and unaffected population control subjects are given in Table 1. In addition, the frequencies of asthma and allergy to inhaled antigens are given separately for those study subjects who had undergone adenoidectomy before the age of 4 years and for those who had not undergone adenoidectomy (Table 1). The frequency of asthma and allergy was somewhat lower among the probands than among the unaffected control population. It was higher among those who had undergone adenoidectomy at an early age (Table 1).

To test risk factors for diabetes (dependent variable), an analysis was performed, first, among the probands as compared with the population control subjects and, second, among the probands as compared with the unaffected siblings, with asthma, sex, and adenoidectomy status entered as independent variables (Table 2). Another set of analyses was carried out, with sex, allergy to animal dust, and adenoidectomy status entered as independent variables. Yet another set of analyses was carried out, with sex, allergy to pollen, and adenoidectomy status entered as independent variables (Table 2). These three sets of analyses, each with three independent variables, were performed instead of one with all of the five variables (i.e., asthma, allergy to animal dust, allergy to pollen, sex, and adenoidectomy status). Asthma and allergy to animal dust were very strongly associated with each other (odds ratio [OR] 17.6) as well as asthma and allergy to pollen (8.9) and allergy to pollen and animal dust (36.2).

The analysis revealed that the risk of diabetes was inversely associated with allergy to animal dust when the probands were compared with the population control subjects (OR 0.67 [95% CI] 0.45–0.99, P = 0.0450) (Table 2) but not when the probands were compared with their siblings (0.99 [0.66–1.47]) (Table 2). On the other hand, the risk of diabetes was inversely associated with asthma when the probands were compared with the population control subjects (0.49 [0.24–1.00], P = 0.0505) (Table 2), and there was also a tendency for the diabetic probands to have a lower frequency of asthma than their unaffected siblings (0.54 [0.27–1.09], P = 0.0862) (Table 2).

The analysis showed an association between sex and diabetes when the probands were compared with their siblings, because the female siblings of the probands returned the questionnaire more often than the male ones (data not shown).

In the current study, the diagnosis of diabetes or no diabetes was verified by the records of the National Diabetes Registry (8), which allowed an objective diagnosis. The frequency of asthma was evaluated by the question “Have you been diagnosed as having asthma by a physician?” Such an approach presumably results in a rather specific and sensitive diagnosis of asthma because self-reported physician-diagnosed asthma has been shown to correlate well with objective findings of asthma diagnosis (9,10). Furthermore, the diagnosis of asthma by physicians in Finland is made by uniform criteria because it entitles the patient to a reimbursement for asthma medication granted by the Finnish Social Insurance Institute. The criteria for asthma diagnosis are based on the recommendations of the American Thoracic Society.

Self-reported atopic symptoms may not always have a good correlation with objective findings of atopy, such as skin-prick tests or serum allergen specific or total IgE (9). Although symptoms of respiratory allergy have been shown to have a better correlation with serum allergen specific or total IgE than symptoms of atopic dermatitis (11), the evaluation of self-reported symptoms of allergy to inhaled antigens is restricted in that it may reflect a broader range of hypersensitivity reactions that may not completely parallel atopy confirmed by skin-prick tests or specific IgE measurements. We attempted to minimize this bias in the self-reported atopic symptoms in the present survey by asking the participants whether they had ever been “diagnosed” as having an allergy to animal dust or pollen.

We have previously found that adenoidectomy at early age is strongly associated with asthma (P.S.M., J.T., H.S., J.P., M.K., J.T., unpublished observations). This could be seen primarily among the unaffected control population. To account for the possible confounding effect of adenoidectomy status, it was entered as an independent variable in the multiple logistic regression analysis. The analysis revealed that diabetes was inversely associated with asthma independently of adenoidectomy status or sex.

Our observation of the decreased frequency of asthma and respiratory allergic symptoms among diabetic children is in keeping with the results of the two previous reports (4,5). In a multicenter study involving eight European centers with 1,028 diabetic and 2,744 nondiabetic individuals, the risk of self-reported asthma (OR 0.66 [95% CI 0.49–0.90]) and self-reported eczema or allergic rhinoconjunctivitis (0.79 [0.64–0.97]) was decreased among children with type 1 diabetes (5). In another study involving 157 diabetic children, 173 siblings of the diabetic children, and 2,324 control children, the history of ever wheezing was decreased among diabetic subjects as compared with the control subjects (0.36 [0.25–0.52]) (4). However, there was no difference between the groups in response to the question “Have you ever been told that you have asthma?” (4), which is in contrast to the results of the above multicenter study (5) and to our results.

In this survey, complete information was available from 306 probands, 506 unaffected siblings, and 406 unaffected control subjects. Thus far, this study is the largest on asthma and diabetes that also involves the unaffected siblings. As compared with the unaffected population control subjects, many, though not all, unaffected siblings of the probands presumably have genetic predisposition to type 1 diabetes and/or have been exposed to environmental factors that may trigger type 1 diabetes. Therefore, the analysis using the unaffected siblings as a reference enabled us to indirectly estimate the contribution of these risk factors. The analysis revealed that although the risk of diabetes was inversely associated with asthma when the probands were compared with the population control subjects, it also appeared to be decreased when the probands were compared with their unaffected siblings. However, the risk of diabetes was inversely associated with atopic symptoms to animal dust when the probands were compared with the population control subjects but not when the probands were compared with their siblings; moreover, the OR between the probands and their siblings was close to 1. This suggests that the probands and their unaffected siblings apparently share a genetic and/or environmental background that reduces the risk of developing animal dust allergy.

The results suggest that the risk of type 1 diabetes is inversely associated with asthma and atopic symptoms to inhaled antigens. Factors predisposing to atopic symptoms to inhaled antigens may protect against type 1 diabetes.

This work received financial support from the Paulo Foundation, Helsinki, Finland; Research Funds of Helsinki University Central Hospital; and the Academy of Finland grants 3837 and 46558.