Diabetes

Bi-directional association between allergic rhinitis and diabetes mellitus from the national representative data of South Korea

ParticipantsData was extracted from a cross-sectional study, Korean National Health and Nutrition Examination Survey (KNHANES) 2007–2018 which was a nationwide population-based self-questionnaire survey conducted by the Korean Ministry of Health and Welfare. Participants were recruited by a two-stage stratified cluster sampling. Sampling frame was the latest population and housing census, conducted by Statistics Korea and supplemented by the data of public housing price conducted by Ministry of Land, Infrastructure and Transport, Korea. Stratification variables were administrative district, sex age, type of housing, and area of living space. Trained interviewers performed the interviews using structured questionnaires to obtain information including sociodemographic factors, health-related factors, lifestyle factors, the use of medical services and female reproductive factors. Data on lifestyle were recorded by self-reported questionnaire. HbA1c levels were measured via high performance liquid chromatography (HLC-723G7; Tosoh, Tokyo, Japan). All data of the participants was weighted based on a sociodemographic character and significance because KNHANES data represented total population of Korea. To be more specific, the Rao-scott chi‐square test for complex samples was used to compare the differences in the exposure group (PROC SURVEYFREQ) and the complex sample multiple logistic regression (PROC SURVEYLOGISTIC) was used to determine the bi-directional association between AR and DM.Among KNHANES 2007-2018, total number of study participants are 97,622. To determine the association between AR and DM, we excluded the data of 68,376 participants. First of all, from the 2010–2012 data, 25,534 participants were excluded due to absence of assessment of AR. Secondly, among participants aged under 30 years, 23,366 subjects were excluded to increase the homogeneity of the DM group, as near to type 2 DM. Lastly, patients with missing value on AR and DM were excluded (17,527 and 1949 participants, respectively). Subsequently, 29,246 participants (12,466 men and 16,780 women, respectively), aged over 30 years, were included (Fig. 1).Definitions of allergic rhinitis and diabetes mellitusIn the KNHANES study, AR and DM were reported via an interview questionnaire. AR and DM patients were identified using self-reported history of a clinical diagnosis of AR and DM. Cases of “DM incidence after AR diagnosis” were defined according to age at DM and AR diagnosis. Cases of “AR incidence after DM diagnosis” were also defined in the same way. For sensitivity analysis, individuals with HbA1c levels above 6.5% (48 mmol/mol) or with current DM treatment was defined as DM patients. (Glycated hemoglobin (HbA1c) is a form of hemoglobin bonded to a sugar and the target analyze of a blood test for diagnosing and monitoring DM.) The reason why we conducted the main analysis by defining DM through self-reporting, rather than HbA1c, was based on our main goal to evaluate the temporal sequence between DM and AR. In this study, to clarify the temporal sequence of each association between AR → DM and DM → AR, we utilized the data regarding the age of diagnosis for both DM and AR. HbA1c was not proper for comparing temporality, since it only reflected the state at the latest measurement. Therefore, we chose self-reported morbidities as the main exposure and outcome measurements. For the AR → DM association evaluation, we additionally conducted sensitivity analysis by defining DM based on HbA1c and medication history.CovariatesSince the goal of this study was to assess the risk of DM in AR patients and vice versa, we considered several confounders known to affect both AR and DM occurrence, such as age, family income, education, marital status, ever smoking status, drinking frequency, BMI, sleep deprivation, dyslipidemia, the number of comorbidities, residence, and menopause. “Family income” was defined by monthly earnings, and it was divided into four groups: income under 150, 150–316, 316–518, and over 518 Korean-won (unit). “Education level” was assigned to seven categories, including pre-school, graduating from elementary school, middle school, high school, college, university, graduate school. “Marital status” was defined by four categories as follows: married and living with spouse, married but not living together or divorced, bereaved, and unmarried. “Ever smoking” was divided into three categories, including never, quit now, and still smoking. “Drinking frequency” was assigned to three categories, including less than once per month, two to four times per month, and more than two times per week. BMI (kg/m2) was divided into two categories, underweight/normal (under 23.0 kg/m2) and overweight/obese (above 23.0 kg/m2), according to the WHO-WPRO (Western Pacific Regional Office) guideline; normal range was 18.5 to 2320. Overweight and underweight were over 23.5 and under 18.5 respectively. Our database included a small population of underweight group, which could lack statistical power due to the small sample; therefore, we did not divide the under-23 group into two categories. We defined sleep disturbance as a few days’ experience of waking from sleep and difficulty going to sleep with lethargy in the daytime. “Sleep deprivation” was based on the weekly frequency of sleep disturbance: not at all, a few days, more than a week, and every day. The “number of comorbidities” was the sum of comorbidities assessed in the KNHANES, including liver cirrhosis, hepatitis C, hepatitis B, chronic renal failure, macular degeneration, glaucoma, cataract, otitis media, sinusitis, atopic dermatitis, depressive disorder, cancer (thyroid, lung, cervix, breast, colon, liver, gastric, and others), benign thyroid disease, asthma, tuberculosis, arthritis, myocardial ischemia or angina, stroke, and hypertension. Lastly, “residence” was divided into urban and rural: urban areas included Seoul, other metropolitan cities, and areas whose administrative district was larger than “dong” in Gyeonggi province; rural included all non-urban areas. All covariates were coded as categorical variables. In case of missing values in these variables, we coded them as a separate category; this allowed us to include the information on exposure and outcome variables from the entire individual in the major statistical analysis without omitting the individuals with any missing values present in the covariates. (e.g. missing/unknown was coded as “9”; family income, n = 104; education, n = 189; marital status, n = 24; smoking status, n = 249, drinking frequency, n = 4027, dyslipidemia, n = 73; BMI category, n = 71; sleep deprivation, n = 14,514).Statistical analysisTo assess the risk of DM in AR patients and vice versa, we excluded some participants group from all of them for each case. To be more specific, by using the value of ‘age of diagnosis’ in database, we compared diagnosis order between AR and DM for each participant. All participants were divided into 5 types. Type A was the subject who had both diseases and AR was diagnosed earlier than DM. Type B was the subject who had both, and DM was diagnosed earlier than AR. Type C was the people only diagnosed AR and type D was only DM. Type E was the people who was not diagnosed both AR and DM. There were two main analyses in this study (Fig. 1). Firstly, in main analysis 1, to evaluate the effects of AR on DM occurrence, we used all participants except type B group. Likewise, in main analysis 2, we used all participants except type A, to evaluate the effects of DM on AR. By considering temporality, we could evaluate the association between AR and DM. Since men and women have different physiologies, we analyzed the data of men and women separately.Multivariable logistic regression models were used to determine the associations between AR and DM (Figs. 2 and 3). We calculated the odds ratios (ORs) and 95% CI (confidence intervals) of DM prevalence in subjects with AR. In the counter direction, we also calculated the ORs and 95% CI of AR prevalence in subjects with DM. In each direction of analysis, we added the following covariates to the final model: age, family income, education, marital status, ever smoking status, drinking frequency, BMI (cut-off: 23.0 kg/m2), sleep deprivation, dyslipidemia, and the number of comorbidities. Stratified analysis was also performed by age group, residence, presence of comorbidity, BMI group, and menopause (-in women only). All analyses were performed using the statistical software package SAS version 9.4 (SAS Institute Inc., Cary, NC, USA), and weight distribution of the population was considered. Since both AR and DM had sex difference in many aspects of the diseases21,22, to evaluate precisely, all analyzing process had done in each sex, respectively.Figure 2Assumed structure of the association regarding the effects of allergic rhinitis (AR; exposure X) with covariates (C) on diabetes mellitus (DM; outcome Y) with stratification (S).Figure 3Assumed structure of the association regarding the effects of diabetes mellitus (DM; exposure X) with covariates (C) on allergic rhinitis (AR; outcome Y) with stratification (S).

Via Source link

Most Popular

Safety Health News Provides information on HEALTH NEWS, HEALTH CARE, WOMEN’S HEALTH, KID’S HEALTH, MEN’S HEALTH, DIABETES & DEALS NEWS

All About information only. Please consult your Doctor for any illness.

Copyright © 2020 Safety Health News. Powered by Wordpress.

To Top