Factors correlating with serum birch pollen IgE status in pregnant women in Hokkaido, Japan: The Japan Environment and Children's Study (JECS)

Background Birch pollen allergy affects pregnant women, and such allergy may affect the development of allergic diseases in their children. Using nationwide birth cohort data, this study aimed to investigate the prevalence of birch pollen IgE positivity and to identify correlating factors in pregnant women in Hokkaido, Japan, a high-latitude island. Methods Participants included 6856 pregnant women. Participants responded to questionnaires regarding lifestyle factors and history of allergies. Data regarding parity, height, and pre-pregnancy weight were collected from medical records. Blood samples were obtained from participants during the first trimester of pregnancy, and serum allergen-specific IgE titers were determined. Results The serum of 30.2% participants was positive for birch pollen IgE (≥0.35 UA/mL). Such positivity significantly correlated with a history of other allergic diseases, particularly food allergy and allergic rhinitis/hay fever. In multivariate logistic regression analysis, pre-pregnancy high body mass index (BMI ≥ 25) significantly correlated with birch pollen IgE positivity [odds ratio (OR), 1.24; 95% CI, 1.05–1.47; reference BMI, 18.5–24.9] and higher income (≥10 million yen per year; OR,0.55; 95% CI, 0.37–0.81; reference, household income < 2 million yen per year), and second quintile level physical activity (OR,0.75; 95% CI, 0.63–0.88; reference, the first quintile of physical activity) had significant protective effects. Conclusions Birch pollen IgE positivity in pregnant women was positively associated with food allergy, allergic rhinitis, pre-pregnant high BMI, and was negatively associated with light exercise and high household income in Hokkaido. Trial registration UMIN000030786.


INTRODUCTION
Birch pollen, which causes springtime-allergyrelated diseases in high-latitude countries, is one of the major causes of allergic rhinitis and a cause of asthma. Cross-reactivity with birch pollen allergens extends to plant food allergens, resulting in pollen-food allergy syndrome. 1 Pollen-related allergic rhinitis is increasing in Japan, with Japanese cedar as the dominant pollen in much of Japan. 2 However, its northernmost island, Hokkaido, is located at high latitude and thus has fewer Japanese cedar trees and many birch trees. 3 The major cause of springtime allergic rhinitis on Hokkaido is birch pollen. 4 The symptoms of allergic rhinitis can negatively affect quality of life, and they are associated with sleep disturbances. Workers with allergies may be less productive while at work, a circumstance known as presenteeism. 5 In addition, about 70% of birch pollen allergic patients experience hypersensitivity reactions caused by IgE crossreaction to food items, 1 also affecting their quality of life. 6 Serum levels of specific IgE are biomarkers for allergic reactions and can be used to assess allergy, 7 including birch pollen sensitization. 8 During pregnancy, allergic rhinitis and asthma can adversely affect both maternal quality of life, and, in the case of maternal asthma, perinatal outcomes. 9 This study aims to investigate the prevalence of birch pollen IgE positivity and to identify factors that correlate with IgE status among pregnant women in a high-latitude region of Japan. Data were retrieved from the Japan Environment and Children's Study (JECS), an ongoing nationwide birth cohort study. 10 Though JECS is a nation-wide cohort, birch pollen specific IgE among pregnant women were measured only in Hokkaido because it is the only region with a high prevalence of birch pollen allergy in Japan.

Participants
The JECS is a prospective birth cohort study of participants located throughout Japan, including Hokkaido, 10,11 and we performed a crosssectional analysis that included pregnant women. Women were recruited in the early stages of pregnancy, and a total of 103, 099 pregnancies throughout Japan were covered in this study between January 1, 2011 and March 31, 2014; discounting pregnancies in the same woman, the study involved 97, 454 unique mothers. The Hokkaido unit is one of 15 regional centers of the JECS, administrating participants from 3 areas: Sapporo (Kita-and Toyohira-ward of Sapporo city; total population, 489 000), Asahikawa (Asahikawa city; population, 351 000), and Kitami area (Kitami city, Okedo town, Kunneppu town, Tsubetsu town, and Bihoro town; total population, 160, 000). In Hokkaido, 8365 pregnancies (8441 fetal records) had been registered at the time of the study. After excluding second or later participation, withdrawal of consent, or missing data, we analyzed data from 6856 pregnant women ( Fig. 1).

Data collection
Questionnaires were administered to enrolled mothers during the first (T1) and second/third trimester (T2) of pregnancy. The T1 questionnaire included questions regarding age and past history of doctor-diagnosed allergy-related diseases. Participants responded to the question, "Have you ever been diagnosed by a physician for asthma, allergic rhinitis/hay fever, atopic dermatitis, allergic conjunctivitis, or food allergy (including oral allergy syndrome)?". The mothers were also asked by research coordinators about their medication use, including steroid medications for allergy (oral, inhaled, or injected) between 12 weeks of pregnancy and submission of the T1 questionnaire. Physical activity before pregnancy was evaluated using the Japanese version (short and selfadministered) of the international physical activity questionnaire (IPAQ) in the T1 questionnaire, and the physical activity in terms of Met$min/day (metabolic equivalent of a task measured as the number of minute per day) was calculated. [12][13][14] Physical activity as defined in the IPAQ includes all time spent being physically active, including work-related activities, housework, and leisuretime activities. Physical activity was quintilized for categorical analysis.
The T2 questionnaire included questions related to smoking habit, drinking habit, marital status, educational attainment, household income, dog or cat residing in the home, exposure to organic solvent during pregnancy (either yes or no), and exposure to dust during pregnancy (either yes or no).
The following information was also collected from medical records: parity, gestational period, and height and pre-pregnancy weight, from which body mass index (BMI) was calculated.

Statistical analysis
The association between birch pollen IgE positivity and age and pregnancy term (at blood collection), pre-pregnancy BMI, smoking habits, alcohol consumption, parity, marital status, education, household income, dog and/or cat in the home, organic solvent exposure, dust exposure, steroid medication, physical activity, other allergen-specific IgEs, and past history of allergy was investigated using the chi-square test. In crude logistic regression analysis, the odds ratios (ORs) of birch pollen IgE positivity with other allergenspecific IgEs were calculated.
For participants with missing data (19.8% of the cohort), the information was replaced using multiple imputations (25 imputed datasets) based on the assumption that data were missing at random. Variables included in the imputation model were as follows: age, pregnancy term, BMI, smoking habits, alcohol consumption, parity, marital status, education, household income, dog and/or cat in the house, organic solvent exposure, dust exposure, steroid medication use, physical activity, other allergen-specific IgEs, and past history of allergy. Using the imputed datasets, the crude ORs of birch pollen IgE positivity or past history of allergy were calculated. Multiple logistic regression analysis was conducted to correlations with factors with p < 0.2 in the crude analysis of birch pollen IgE positivity.
Significance was defined as p < 0.05. All analyses were conducted using IBM SPSS Statistics 25.0 for Windows (SPSS Inc., Chicago, IL, USA) based on the dataset jecs-ag-20160424 released in June 2016 and revised in October 2016.

RESULTS
The distribution of each antigen-specific IgE titer is shown in Table 1. Overall, 30.2% of the cohort was positive for birch pollen IgE (0.35 UA/mL).
Participant characteristics are compared between those who were positive and negative for birch pollen IgE in Table 2. Significant differences were found in household income and physical activity. Participant characteristics are also compared between birch pollen IgE levels (Table S1). Positivity for birch pollen IgE correlated significantly with all other antigen-specific IgEs, and Der p 1 positive had highest percentage (73.1%) of birch pollen IgE positive participants (Table 3). Positivity for birch pollen IgE correlated significantly with history of allergy of all types, and allergic rhinitis/hay fever had the highest percentage (50.0%) of birch pollen IgE positive participants (Table 4,  Table S2).
The ORs of positivity for birch pollen IgE with other allergen-specific IgEs are shown in Table S3. All antigen-specific IgEs had significantly raised ORs. Positivity for animal dander IgE had the highest OR (5.28; 95% confidence interval (CI), 4.69-5.94), followed by that of Der 1 (OR, 3.96; 95% CI, 3.54-4.43). Table S4 shows the ORs for birch pollen IgE positivity with history of allergic diseases. All allergic diseases had significantly raised ORs. The highest OR was observed for food allergy (4.20; 95% CI, 3.52-5.00), followed by that of rhinitis/hay fever (OR, 3.26; 95% CI, 2.92-3.64).
Crude ORs (Table 5) were determined for birch pollen IgE positivity and the investigated variables; multivariate adjusted ORs (

DISCUSSION
In our cohort of pregnant women in Hokkaido, 30.2% were positive for birch pollen IgE. This positivity was significantly related to history of allergic diseases, especially food allergy and allergic rhinitis/hay fever. In addition, in multivariate analysis, pre-pregnancy high BMI was significantly related to birch pollen IgE positivity, and higher income and second quintile level physical activity had significant protective effects. To our knowledge, this study is the first to report the prevalence of birch pollen IgE positivity and its related factors in a relatively large population in Japan, though our participants were limited to pregnant women. In Northern and Central Europe, over the last few decades, levels of birch pollen have risen, and the prevalence of birch pollen sensitization has also increased. 1 A recent review reports that in general populations in Europe, the prevalence of birch pollen sensitization ranges from approximately 8%-16%. 1 A previous study using the JECS data of participants throughout Japan reports that 55.6% were positive for Japanese cedar IgE. 15  seasonal allergic rhinitis caused by Japanese cedar pollen, and the number has significantly increased in the last 2 decades. 17 Because the prevalence of allergic rhinitis to birch pollen is unknown in the general population in Hokkaido and the prevalence of birch pollen IgE positivity was relatively high in our cohort, further studies are needed to investigate the prevalence of rhinitis and its relationship to birch pollen IgE positivity.
Sensitization to multiple antigens is common, especially among people with allergic symptoms. 18  Oral allergy syndrome, also called pollen-food allergy syndrome, is a type of food allergy. People with birch pollen hypersensitivity may experience oral symptoms after the ingestion of apricot, peach, apple, carrot, almond, plum, hazelnut, pear, celery, fennel, parsley, aniseed, coriander, soybean, caraway, or peanut. 19 We observed that 17.4% of the birch pollen IgE positive participants had a history of food allergy. Among allergy patients in Korea, 20% of those with a positive response to birch pollen allergen on the skin prick test had oral allergy syndrome 20 ; of the participants positive for birch pollen IgE, 49.6% had a history of allergic rhinitis/hay fever, and 23.6% of participants negative for birch pollen had a history of allergic rhinitis/hay fever. Thus, positivity for a specific IgE does not necessarily indicate clinically evident allergic disease. Moreover, major causes of allergic rhinitis include other pollens, house dust, and mold, 2 which account for a higher proportion of participants with a history of allergic rhinitis/hay fever among those negative for birch pollen.
In our present study, pre-pregnancy high BMI had a significantly higher OR for birch pollen IgE positivity. A review article reported that no clear association was found between obesity and the prevalence of allergic rhinitis or allergic conjunctivitis or increased sensitization to food allergens. 21 However, studies in Canada and Denmark observed that obesity was significantly related to atopy sensitization (skin prick test in the Canadian study, skin prick test or specific IgE in the Danish study), especially among women; 22,23 the later study proposed the possible influence of sex hormones on the development and expression of atopy and atopic disorders. 23 Our multivariate analysis showed a negative correlation between the highest household income category and positivity for birch pollen IgE and no correlation between educational attainment and birch pollen IgE positivity. A systematic review reported that low socioeconomic status had a significantly higher OR for asthma, but significantly lower ORs for allergies in general, atopic dermatitis, and allergic rhinoconjunctivitis. 24 Low educational attainment correlated significantly to atopy. 23,25 Since the higher prevalence of allergies in higher social groups is considered consistent with the hygiene hypothesis, the protective effect of higher household income may occur by chance. Further, household income did not necessarily reflect lifetime socioeconomic status for relatively young women, and women of higher socioeconomic status may tend to have lower birch pollen exposure chance due to higher quality housing that has air conditioning, a factor known to reduce indoor pollen levels. 26 Thus, these factors may contribute to the observed protective effect of high household income.
In our present study, the second quintile of physical activity had a significantly lower OR for birch pollen IgE positivity. Performing moderate (3-6 Mets) physical activity for 150 min/week, which is equivalent to 64 to 128 Mets$min/day, is recommended globally for substantial health benefits, 27 and the second quintile of physical activity of our present study ranged from 28.4 to 94.3 Mets$min/ day. A cohort study of children in Norway reported that low levels of physical activity in preschool children were positively associated with later atopic sensitization. 16 However, a study of German adolescents reported that low physical activity was associated with asthma and rhinitis in boys but not girls, and that atopic sensitization was not associated with physical activity. 28 Further, a Danish study reported that physical activity was not related to atopy sensitization. 22 Thus, the effect of physical activity on atopic sensitization may depend on age, sex, and environmental factors. The lower OR of second quintile level physical activity may indicate that lack of physical activity is bad for atopy but that higher physical activity, particularly outdoors, increases pollen exposure.
This study has several limitations. First, since the study design is cross-sectional, we were unable to infer cause-effect relationships. Second, due to the exploratory nature of the analyses, corrections for multiple comparisons were not performed. However, all variables had significances even after Bonferroni corrections (multiplied five). Third, the participants were limited to pregnant women because we used birth cohort data. Moreover, the timing of blood sampling varied between the first and the second trimester, but the birch pollen IgE positivity had no significant difference between the first and second trimester in our present study. A study reported a statistically significant increase in birch pollen IgE over time, from preconception to postpartum, but the change was relatively small. 29 Therefore, we believe that our result can be generalized to women of the same age group. Fourth, changes in the location of residence before participation were not considered. If someone lived in a region of Japan other than Hokkaido for many years, birch exposure would have been rare. Inclusion of many such participants would underestimate the prevalence of birch pollen IgE positivity compared to that of lifelong Hokkaido residents. Fifth, only six allergen-specific IgEs were measured, and other major allergens such as Der f 1, dog and cat fur, pollens, fungi were not evaluated because of budget limitation. However, they were selected form main asthma allergens (Der p1 and animal dander), main allergic rhinitis/ hay fever allergens (cedar (all participants) and birch (only in Hokkaido), a main food allergen for children (egg white, measured as a confounder of child food allergies), and a main insect allergen in Japan (moth). 30,31 Finally, the symptoms did not necessarily present at the study time, because allergic diseases were defined as a history of each of the allergic diseases, and seasonal variations of symptoms and specific causes of food allergies were not observed.

CONCLUSIONS
We observed that in Hokkaido, a high-latitude island with many birch trees, 30.2% of pregnant women were positive for birch pollen IgE. Such positivity correlated significantly to a history of allergic diseases, especially food allergy (OR ¼ 4.20) and allergic rhinitis/hay fever (OR ¼ 3.26). Birch pollen IgE positivity correlated significantly and positively with pre-pregnancy high BMI (OR ¼ 1.24) and negatively with higher income (OR ¼ 0.55) and second quintile level of physical activity (OR ¼ 0.75).