Epidemiology of allergic rhinitis and associated risk factors in Asia

This review article aims to present the epidemiology and associated risk factors of allergic rhinitis (AR) in Asia. AR-related literature published on Asia was systematically reviewed and the associated risk factors were investigated. The prevalence of AR in Asia varied considerably depending on the geographical location, study design and population involved. Several risk factors were observed to have strong association with disease presentation across multiple studies. Among these, family income, family size, daily personal computer usage time, personal and parental education attainment, and stress level have shown some level of biological gradient influence when multiple risk levels were analyzed. This suggests that AR manifestation and presentation possibly might be strongly affected by various personal and family factors. These findings are beneficial as they may provide insights into modifiable factors that may influence AR presentation. In addition, these results indicate that strategies to reduce personal and family-related risk factors have to be developed in order to alleviate the odds of AR expression. Electronic supplementary material The online version of this article (10.1186/s40413-018-0198-z) contains supplementary material, which is available to authorized users.


Risk factors affecting the presentation of allergic rhinitis
Apart from the demographic factors, smoking and drinking habits, pet adoption, education attainment, and family history were the risk factors of AR, commonly studied in Asian countries [7][8][9][10][11]. Conversely, Western countries focus more on the effects of pollens, drugs, pets, and family history on the presentation of AR [12][13][14]. The differences between the risk factors analyzed could be culturally induced or due to the climatic differences between Asian and Western countries. However, it was observed that pet adoption and family history are the common risk factors studied in both regions, suggesting their pervasiveness in inducing AR manifestation worldwide.

Disease diagnosis
While AR is influenced by genetic predisposition, the symptom presentation also depends on environmental exposures [15]. In addition, the disease can co-present with other diseases, such as asthma and other infectious diseases, which could further complicate the disease diagnosis. A robust association of rhinitis was found among individuals with allergic and non-allergic asthma [16]. Among patients with persistent and severe rhinitis, asthma was found prevailing [17].
Moreover, patients can experience adverse effects on social life, productivity at work and performance in school, especially for those who suffer from a more severe form of AR [16]. The use of suboptimum pharmacotherapy and antihistamines with sedative effects can further exacerbate the situation. This incurs a financial burden from both direct and indirect costs which adversely affects society [18]. Therefore, a prompt and accurate diagnosis, followed by appropriate disease management and awareness of the exacerbation risk factors, would be crucial to ease this burden.
Diagnosis of the disease is usually based on medical history of the patient in addition to skin prick test or blood test. However, misinterpretation can occasionally occur and this delays the golden treatment period which can result in other unexpected consequences, such as paying unnecessary medical expenditure and missing work [18].

The aim of the study
This review article aims to study the epidemiology of AR in Asia and identify significant modifiable risk factors associated with disease presentation. Several criteria have been employed to establish association between triggering factors and disease manifestation.

Search strategy and selection criteria
The epidemiology and potential factors associated with AR manifestation were obtained from the Web of Science using the search terms of 'rhinitis' , 'risk' and Asian countries. The list of Asian countries and independent territories used in the search is listed in Additional file 1. 'Rhinitis' is used as it represents a general form of the disease which serves to capture as many risk factors, including both modifiable and non-modifiable, as possible. As Asian and Western countries are known to have different cultural and social backgrounds, our study only evaluated articles published on Asia and this articles serves to provide a detailed list of triggering risk factors associated with AR in Asia.
Using these search terms, 56 articles were first identified. The articles were carefully reviewed and those with unclear study design or disease definition or which were conducted in a narrow pool of individuals were excluded. Apart from these 14 articles, additional 6 cross-referencing articles were also included. These 20 articles, published between 1994 and 2017, were evaluated closely for their study design, disease prevalence, disease definition, and the AR risk factor analyzed.

Establishing the association link
The factors investigated in the 20 articles were further classified either as a potential risk factor or a co-morbidity. The association between potential modifiable risk factors and AR manifestation were evaluated using several important criteria established in literature. These criteria include the strength of association, consistency of the observed association, specificity, biologic gradient, biologic plausibility, coherence, analogy, and temporality. In addition, meta-analysis was conducted using the software/program-Stata/SE 11.2 with random effects model to evaluate the influence of modifiable risk factors with replicative results reported in at least three independent AR publications.

AR epidemiology in Asia
Based on the methodology described, different articles published in Asia were reviewed. The reviewed articles have variable study design, disease definition and adopt different analysis parameters as shown in Table 1. The population size also varies from study to study, ranging from 200 in Kidoni et al. [19] to 30,000 in An et al. [2] Though similar parameters were used to study the epidemiology of AR, a larger population group will help to further establish the prevalence of the disease as it better represents the targeted population. In addition, apart from the country of study, the disease prevalence differs depending on the disease definition and the study population. In the study conducted by Min et al. [20], AR prevalence is 1.14% among Korean residents; while in a retrospective study published by Alsowaidi et al., 2010 [3], 32% of United Arab Emirates residents are AR patients.

Risk factors and co-morbidities of AR
Apart from the general demographic factors, many modifiable risk factors for allergic diseases, such as smoking and drinking habits were investigated as summarised in Table 1. Furthermore, cultural-or socioeconomic-related factors specific to an individual country have been explored in some studies to identify their association with AR presentation. For instance, heavy traffic and individual stress level are two factors investigated in a Singapore [21] and Korea [2] study, respectively. These factors were identified worrying elements in the respective countries, thus finding their association with AR presentation is crucial.
We further classified these factors into a potential risk factors or co-morbidities category based on the following definitions. A typical risk factor is a demographical, physical, sociological or environmental component which potentially increases the risk of presenting a disease or is protective against the expression of an illness. On the other hand, if AR manifestation is linked to another disease occurrence, it will be known as a co-morbid of AR. As listed in Table 2, most of the factors analysed are in the risk factor category. However, diseases such as asthma are co-morbidities which can possibly induce AR expression as shown in Table 3. In this article, only the modifiable risk factors were evaluated for their relationship with AR manifestation.

Demographical factors affecting the AR presentation
Multiple papers have suggested the importance of age, gender, race, and nationality in affecting AR presentation ( Table 4). The association of race and nationality on the disease expression could signify the difference in social and cultural backgrounds, as well as genetics, which can potentially influence the presentation of AR. However, as these factors are non-modifiable, they are only useful in evaluating the risk of presenting AR, but not for prevention. In Li et al. [22], the odds of AR have shown to increase with the rise of household income when different household income groups are compared. For the household with an income of > 2500 RMB/month, the odds of AR is 2.88 times of those with an income of 800 RMB/ month. A similar trend is observed in another two independent studies. A pooled odds ratio of 2.75 has been obtained which suggests the significant role of household income in affecting AR expression (Fig. 1).
Moreover, being married, a large number of members in the household, and parity were indicated to be beneficial for protecting one against AR. However, their influences towards protection of AR are likely to be interrelated as married individuals are usually with children and are therefore likely to report an increased parity number and household members.

Personal risk factors affecting AR presentation
Apart from the demographical factors that are usually non-changeable to an individual, one's behaviours, attitude, and encounters might have direct and indirect    influences to the disease presentation. These factors are highly varied from one person to another and are often affected by their background and the social group they interact with. As stated in Table 5, like the case with many other infectious diseases [23,24], alcohol consumption and smoking habits have shown to increase the odds of presenting AR. This is especially true for the smoking habit; which shows higher odds of expressing AR among present smokers, past smokers, and even passive smokers as compared to non-smokers and those who are not exposed to passive smoking. The result is consistent across four independent articles and a pooled odds ratio of 1.34 was obtained indicating smoking habit does associate with the increased AR manifestation (Fig. 2).
Coincidentally, people with more computer usage, higher education, higher stress level and lesser sleeping time were presented with higher AR susceptibility. Though several pathways were speculated for such association, the effects of confounders and bias could not be ruled out and further study is required to establish the direct association link between these factors.
Stress might be one of the critical risk factors for AR presentation. Studies have shown the association between the level of stress in individuals with more frequent drinking and smoking habits, having higher daily computer usage, and higher education levels but with less sleeping time [25][26][27]. Being in a stressful situation can trigger the expression of cortisol which can induce allergic responses and enhance AR expression (Table 8) [28]. In addition, literature has suggested the possibility of dust trapped on the computer [7] and higher indoor allergen exposure [2] to explain the higher odds of AR manifestation among office workers who usually have higher education qualifications. Dose-response effects were also observed in computer usage, education attainment and stress level as odds of AR increase with higher level of risk exposure, with the exception for AR odds of college students to illiterate individuals in Min et al. [20].
In contrast, people with parasitic or past respiratory infections were reported to have higher odds of AR presentation. The results are contradictory with biological plausibility discussed in other literature. Phathammavong et al. [9], proposed that AR and other respiratory infections compete for immune responses, resulting in a higher odds of presenting AR among the respiratory infection patients. This hypothesis is supported by the reported odds of AR for individuals with either parasitic infection or past respiratory infection are exceptionally high (3.41 and 4.06 respectively). However, this factor has only been studied in Phathmmavong et al. among the articles reviewed and further analysis is essential to confirm the effects of these infections on AR presentation, which could be one of the most important factors in predicting AR risk.

Family risk factors affecting AR presentation
In Table 6, mother depression and cesarean delivery are positively correlated with the odds of AR presentation. As stated in Li et al. [22], pre-and postnatal depression stimulates the production of cortisol, and this secretion affects the immune development of a fetus and increases the odds of presenting AR. Apart from this, cesarean delivery might further exacerbate this situation as unlike vaginal delivery, the infants are not exposed to the mother's birth canal microflora, which has shown to be protective against AR expression [29] as illustrated in Table 8. Conversely, inconsistent results are observed for the association of breastfeeding with AR presentation across multiple studies [30,31]. This refutes the commonly accepted hypothesis which states breastfeeding as protective through the antibodies present in the milk and the additional nutrients from the mother's diet transferred to the milk [32,33]. In contrast, parental education and awareness encourages a hygienic environment which is unfavorable for AR protection as this reduces the chance of exposing their children to a larger variety of allergens in early life. Similarly, for gruel consumption, the subtle protection might be due to the effect of gruel to stimulate inflammatory cytokines which suppress the allergic reaction [34].
On the other hand, genetic factor is long established to play an influential role in AR presentation [15] and a family history of atopy and allergic diseases might predispose children to AR. Multiple studies have shown that family history is a key risk factor associated with the increased risk of AR expression. This is particularly true for children with a family history of AR as high odds ratios of 6.08 and 3.51 have been reported in studies conducted by Alsowaidi et al. [3] and Li et al. [7], respectively. However, genetic factor is non-modifiable and hence, it needs to be complemented with other preventive measures in order to reduce the risk of presenting the disease.

Environmental risk factors affecting AR presentation
As suggested in multiple studies investigated, environmental factors are highly important in triggering AR. For instance, Table 7 has shown that the presence of allergens such as fungi, molds, insects and house dust mites could increase the odds of presenting AR. Among the allergens studied, the presence of fungi and molds were reported to have very high odds of association to AR with 3.44 for fungi in Norbäck et al. [35] and 9.40 for molds in Kidoni et al. [19] Moreover, insect exposure and house dust mite have been identified as two of the most important risk factors for AR as indicated in Table 7. These common indoor allergens, such as mold and fungal spores, insect wastes and house dust mite fecal proteins can induce Type I hypersensitivity reaction by promoting the expression of a range of allergic-causing mediators, thus increasing the odds of expressing AR (Table 8). In addition, the utilization of carpets, which trap dust, and home renovation, which introduces a variety of allergic-causing renovation materials, further exacerbate the situation.
Similarly, outdoor exposures to heavy traffic, air pollution, and fume exposure were also reported to be positively correlated with AR manifestation. These factors are especially crucial for those whose occupations expose them to the allergens [36]. Constant outdoor encounters with pollutants released from motor vehicles and heavy fumes during work promote AR presentation by changing a person's susceptibility towards allergens [2,10,21,36].

Evaluation of risk factors associated with AR manifestation using several criteria
Various risk factors have shown strong association with AR presentation. Results are consistent for several risk  In addition to the ORs, criteria such as biological gradient, biological plausibility and temporality are important in evaluating the association between risk factors and AR. The biological gradient of the factor can be established especially when it is studied in a continuous manner or in multiple exposure levels. This was demonstrated in various demographical factors such as in family income, family size, personal factors like computer usage, education attainment, stress levels and even in parental education attainment. Moreover, the association between the risk factors and AR manifestation are further strengthened when factors with similar roles in AR presentation, such as the common allergens like house dust mites, fungi, and molds, display comparable results.
Furthermore, the listed factors can only be considered as a potential risk if its exposure is reasonably affected or altered the risk of AR development. Its biologic plausibility must also be coherent to the study results found. However, with reference to Table 7, breastfeeding, parasitic infection and past respiratory infections show contradictory results as to what is hypothesized and further analysis and interpretation is thus needed.
Last but not least, with reference to Fig. 3, the two risk factors, family income and smoking, analyzed using meta-analysis are consistently being identified as significant AR risk factor before and after 2010. In addition, education attainment and occupational exposure are two other significant modifiable risks that appeared in AR publications before and after 2010 in Asia. In contrast, it was observed that after the year 2010, more family-related risks were analyzed and shown to be significant AR risk factors, such as the age of gestation and breastfeeding. This suggests a shift in focus to consider more family-related risk factors among the Asian population.

Limitations and conclusion
The studies chosen for this review are limited to articles published in Asia. Thus, the result might not be relevant and applicable to other nations outside Asia. In addition, the analysis might still be biased though several criteria have been used in establishing the significance of the potential AR risk factor in triggering or protecting against AR presentation. The analysed data could be affected by personal viewpoints in addition to errors occurred when translating data from primary  Fig. 2 Individual and combined odds ratio and 95% confidence intervals for smokers/past-smokers in association with Allergic Rhinitis presentation literature to the review summaries, such as misrepresentation and misinterpretation of the original data. Thus, it is highly recommended for readers to refer to the original articles before extracting any information from this article. Furthermore, as most of the studies used in this review are observational studies, confounding effects cannot be ruled out and the association of a particular risk factor with the disease presentation might not be as straightforward as what is illustrated here. From the articles reviewed, family income, family size, computer usage, personal and parental education attainment and stress level are identified as risk factors with the greatest potential to influence AR presentation, and   Exposure of the fetus to the mother microflora during birth is an advantage to protect them against allergic sensitization [29,51]. In contrast, cesarean birth is associated with higher AR risk [51]. Parasitic infection might have some effects to a person's gut microbiota, which could later offer some protection against allergic sensitization as stated in hygiene hypothesis [52]. However, some literature also show that parasitic infection influences the allergy development due to its competition with human immune response [9].
Phathammavong et al., 2008 [9] 37 Parity 2 1 Being allergic might cause reduced reproductivity in females, resulting in a lower parity which is associated with AR presentation [29]. 38 Past measles infection 1 0 The association of measles with AR is not clear, but it was hypothesized that measles infection might protect against AR development or could promote allergic sensitization [52].
Phathammavong et al., 2008 a [9] 39 Past respiratory infection 1 1 Evidence shows that past respiratory infection, such as tuberculosis caused by Mycobacterium tuberculosis could be protective against AR, possibly through reduction of allergy sensitization [52]. In contrast, some studies have shown that past respiratory infection is directly associated with AR development [9]. Phathammavong et al., 2008 [9] 40 Paternal education 1 1 Educated parents are more likely to keep a hygienic living environment, thus possibly increasing the incidence of allergic conditions in their children [45].
Li et al., 2015 [22] Fig. 3 Years in which significant AR risk factors were identified when compared to other factors, they fulfill most of the criteria listed. In contrast, more considerations are required in interpreting the effects of breastfeeding, parasitic infections and past respiratory infections to AR presentation. These factors show incoherent biological plausibility and more in-depth investigation and analysis is thus required. The results obtained from this review article can be used to improve the diagnosis of AR in clinical settings by identifying patients with risk factors strongly associated with AR manifestation. In addition, as personal and family-related modifiable factors are found to be strong AR triggering factors, strategies to alleviate personal stress levels and increase the awareness of allergy risk in a hygienic environment have to be developed.

Additional file
Additional file 1: List of countries and dependent territories used in the literature review search. (PDF 322 kb) Abbreviations AR: Allergic rhinitis; OR: Odds ratio; PR: Prevalence ratio