The effects of weather and air pollution on the severity and persistence of atopic dermatitis (AD) are important issues that have not been investigated in detail. The objective of our study was to determine the short-term effects of meteorological variables and air pollution on AD symptoms in children.
We enrolled 177 AD patients with 5 years or younger from the Seoul Metropolitan Area, Korea, and followed for 17 months between August 2013 and December 2014. Symptoms records of 35,158 person-days, including itching, sleep disturbance, erythema, dry skin, oozing, and edema, were obtained. We estimated the effect of meteorological variables including daily mean temperature, relative humidity (RH), diurnal temperature range (DTR), rainfall and air pollutants including particulate matter with an aerodynamic diameter ≤10 μm (PM10), nitrogen dioxide (NO2), and tropospheric ozone (O3) on AD symptoms using a generalized linear mixed model with adjustment for related confounding factors.
A 5°C increase in outdoor temperature and a 5% increase in outdoor RH was associated with 12.8% (95% confidence intervals (CI): 10.5, 15.2) and 3.3% (95% CI: 1.7, 4.7) decrease in AD symptoms, respectively, on the same day. An increase of rainfall by 5 mm increased AD symptoms by 7.3% (95% CI: 3.6, 11.1) for the days with <40 mm rainfall. The risk of AD symptoms increased by 284.9% (95% CI: 67.6, 784.2) according to a 5°C increase in DTR when it was >14°C. An increase in PM10, NO2, and O3 by 10 units increased the risk of AD symptoms on the same day by 3.2% (95% CI: 1.5, 4.9), 5.0% (95% CI: 1.4, 8.8), and 6.1% (95% CI: 3.2, 9.0), respectively.
Citation: Kim Y-M, Kim J, Han Y, Jeon B-H, Cheong H-K, Ahn K (2017) Short-term effects of weather and air pollution on atopic dermatitis symptoms in children: A panel study in Korea. PLoS ONE 12(4): e0175229. https://doi.org/10.1371/journal.pone.0175229
Editor: Yu Ru Kou, National Yang-Ming University, TAIWAN
Received: July 19, 2016; Accepted: March 22, 2017; Published: April 6, 2017
Copyright: © 2017 Kim et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The dataset supporting the conclusions of this article is available from the Zenodo database [https://zenodo.org/deposit/124411/].
Funding: This study was funded by the Korea Ministry of Environment (MOE) as “the Environmental Health Action Program” (2013001360002) and was contracted to Kangmo Ahn at the Department of Pediatrics, Samsung Medical Center, Sungkyunkwan University School of Medicine. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: AD, atopic dermatitis; CI, confidence interval; CO, carbon monoxide; DOW, day of week; DTR, diurnal temperature range; GAMM, generalized additive mixed model; GLMM, generalized linear mixed model; IPCC, Intergovernmental Panel on Climate Change; KMA, Korean Meteorological Administration; MA, Moving average; NO2, nitrogen dioxide; O3, tropospheric ozone; PM10, particulate matter with an aerodynamic diameter ≤ 10 μm; PM2.5, particulate matter with an aerodynamic diameter ≤ 2.5 μm; RCP, representative concentration pathways; RH, relative humidity; SCORAD, SCORing Atopic Dermatitis index
Atopic dermatitis (AD) is a chronic inflammatory skin disease mostly occurring in early childhood and the prevalence of AD in children is still increasing in both developing and developed countries [1, 2]. In Korea, a nationwide cross-sectional study in 2010 demonstrated that the prevalence of ‘itchy eczema, ever’ was 27.0% for children aged 6–7 years and 19.9% for adolescents aged 12–13 years . Because of its high prevalence and progression into respiratory allergies , the management of ongoing AD is a critical issue in public health. One of the basic principles in the management of AD is to avoid aggravating factors; these may include microbes, allergens, emotional stress, meteorological factors and pollutants in both indoor and outdoor environments [1, 4, 5].
Concerns regarding the impact of climatic factors on atopic diseases are growing as climate change is now widely recognized as a major environmental problem [6–11]. One group found that the annual variation of temperature and relative humidity (RH) were negatively associated with both asthma and eczema symptoms in children . The prevalence of eczema in the United States is lowest in parts of the country where there is abundant sun exposure, high humidity, and elevated temperatures . In contrast, a prospective cohort study of American children found that poorly controlled eczema was influenced by long-term weather patterns, including high temperatures and increased sun exposure . Most previous studies, however, evaluated long-term weather effect on AD. Few studies have focused on the short-term effects of weather on flares in children with preexisting AD. The acute or subacute effects of meteorological variables on the severity and persistence of eczema symptoms must be assessed to determine the appropriate management of AD.
Air pollutants are also considered as risk factors for AD. In a study involving 4,907 French children residing at their current address for 3 years or longer, lifetime eczema was significantly associated with 3-year averaged concentrations of particulate matter with an aerodynamic diameter ≤ 10 μm (PM10), nitrogen dioxide (NO2), and carbon monoxide (CO) . Several studies have reported that exposure to outdoor and indoor air pollutants is associated with worsening AD [14–17]. In our previous study, elevated outdoor concentrations of PM10 were significantly associated with increased AD symptoms on the following day . However, investigating the impact of air pollution on AD is still challenging. Moreover, there are few prospective studies that have evaluated the acute effect of air pollution on AD symptoms.
In this study, we investigated whether outdoor air pollution and meteorological variables including temperature, humidity, diurnal temperature range (DTR), and rainfall affect AD symptom flares in children with pre-existing AD by measuring daily symptom scores and exposure levels to outdoor environment.
Subjects and AD symptom assessment
A total of 177 children with AD (110 boys and 67 girls) were enrolled. They were infants and young children with the age of 5 years or younger who lived in the Seoul Metropolitan Area, Korea. The patients were followed for 17 months between August 2013 and December 2014. The diagnosis of AD was determined by two pediatric allergists according to the Hanifin and Rajka criteria . The severity of AD was assessed using the SCORing Atopic Dermatitis (SCORAD) index . The total IgE and specific IgE against common food and inhalant allergens in the peripheral blood were measured using ImmunoCAP (Thermo Fisher Scientific Inc., Waltham, MA, USA), and considered positive if >0.35 kU/L.
The parents were instructed to record AD symptoms on a daily basis. Parents used a smartphone-adjusted symptom diary that we developed to describe the extent of itching, sleep disturbance, erythema, dryness, oozing, and edema on a scale of 0 to 4. The presence of AD symptoms was determined when a symptom score was 2 points or more for both itching and sleep disturbance scores, accompanied by at least 2 points of the sum of the following symptoms: erythema, dryness, edema, or oozing. For the statistical analysis, the presence of AD symptoms was coded as a binary variable (0 or 1).
Written informed consent was obtained from the parents or guardians of all participating children. The study protocols were reviewed and approved by the Institutional Review Board (IRB) at Samsung Medical Center (IRB No. 2013-05-009).
Meteorological variables and air pollution
Meteorological data including hourly outdoor temperature, hourly outdoor RH and daily rainfall were obtained from the Korean Meteorological Administration (KMA) which runs 76 automatic weather stations in the Seoul Metropolitan Area. Based on the hourly temperature data, we calculated the daily mean temperature. DTR was calculated as the difference between the daily maximum and minimum temperatures. The hourly concentrations for PM10, NO2, and tropospheric ozone (O3) were collected from the National Institute of Environmental Research, which monitors the ambient air pollution at 111 sites in the study area. The daily 24 hour-average values of PM10 and NO2, as well as the maximum daily 8 hour-average value of O3 were calculated. We matched the daily levels of meteorological variables and air pollutants with the daily AD symptom data based on residential address.
Considering the repeated measurement of the AD symptoms based on the panel of AD patients, we adopted a generalized linear mixed model (GLMM) with binomially distributed errors to estimate the effects of the meteorological variables and air pollutants on the AD symptoms.
Ambient air pollution is strongly associated with daily variations in meteorology [20,21]. For example, O3 is a secondary air pollutant that strongly depends on meteorological conditions. Tai et al.  reported that 50% of particulate matter with an aerodynamic diameter of ≤2.5 μm (PM2.5) could be explained by temperature, RH, precipitation, and circulation in the United States. As a result, meteorological variables and air pollution may affect AD symptoms with interaction. All of the predicting variables, including meteorological variables and air pollutants were, therefore, included in a GLMM analysis. Age, sex, SCORAD at enrollment, the presence of fever (0 or 1, as a proxy of infection), and the day of week (DOW) were controlled in the analysis. We treated temperature, DTR, RH, rainfall, PM10, NO2, and O3 as fixed effects and each participant as a random effect in the models. The model specifications are as follows: (1) where E(Y) is the expected expression of AD symptoms; Temp is daily mean temperature; DTR is diurnal temperature range; RH is relative humidity; Rain is rainfall; PM10 is particulate matter with an aerodynamic diameter ≤ 10 μm; NO2 is nitrogen dioxide, and O3 is tropospheric ozone; SCORAD is the SCORing Atopic Dermatitis index at enrollment; fever is the presence of fever on the day of record; DOW is the day of week; and γ is the random effect for subjects.
We also used penalized regression curves of a generalized additive mixed model (GAMM) to examine the linearity of the relationship between the meteorological variables and air pollutants with AD symptoms . Like the GLMM, all of the predicting variables were included in the GAMM analysis, adopting smoothing splines with adjustment for the same confounding factors as the GLMM.
Moving averages (MAs) up to 6 days of meteorological variables and air pollution were used to examine the cumulative effects of weather and air pollution on AD symptoms. For instance, the MA0-5 of PM10 indicates an average concentration of PM10 over 6 days (including the previous 5 days and the current day). We also analyzed the effects of meteorological variables and air pollution on AD symptoms according to sex. The percent change in risk and 95% confidence intervals (CI) were calculated using a regression coefficient. These factors were measured according to 5-unit increases in temperature (°C), RH (%), DTR (°C), rainfall (mm/day) and10-unit increases in PM10 (μg/m3), O3 (ppb) and NO2 (ppb).
All procedures were conducted using the R version 3.0.0 (The Comprehensive R Archive Network: http://cran.r-project.org) with the “lme4” package (version 3.1–2) for GLMM and “mgcv” package (version1.8–3) for GAMM model fitting. All tests were two-sided. An alpha level <0.05 was considered significant.
During the 17 months of study period, the symptom records of 35,158 person-days were obtained. A total of 23,454 (66.7%) and 11,704 (33.3%) person-days were recorded by boys and girls, respectively. The average age of the study subjects was 2.0 years. AD symptoms were present in 49.9% of boys and 39.4% of girls (p < .0001). The average SCORAD at the time of enrollment was 31.1 (Table 1).
The average daily mean temperature was 15.0°C and RH was 64.9% (Table 2). The average values of DTR and rainfall were 8.9°C and 2.7 mm/day, respectively. The mean concentrations of PM10, NO2 and O3 were 45.2 μg/m3, 32.4 ppb and 38.1 ppb, respectively.
The spline curves shown in Fig 1 demonstrate the relationship of meteorological variables and air pollution with the presence of AD symptoms on the same day (MA0). These results were obtained using GAMM fitting, controlling for SCORAD at enrollment, age, sex, fever, and DOW. The outdoor temperature showed negative linear relationships with AD symptoms. In contrast, the relationships of RH, rainfall, and DTR with AD symptoms were nonlinear. Interestingly, DTR and AD symptoms showed a positive relationship when DTR was above 14°C and daily rainfall was also positively related with AD symptoms when rainfall was less than approximately 40 mm/day. There was a negative relationship between RH and AD symptoms overall although the relationship was not linear.
Each figure shows the spline curve (solid lines) with a 95% confidence intervals (two dashed lines). All variables are based on the same model on the same day, without using moving averages for the exposure to each risk factor. The model was controlled for the severity score at the initial visit, age, sex, presence of fever, and day of week (DOW). RH: relative humidity; DTR, diurnal temperature range; OR: odds ratio.
NO2 showed positively linear relationship with AD symptoms for the whole range, suggesting that it has hazardous effects on AD. PM10 had positive relationship with AD symptoms when it was approximately over 35 μg/m3. Relationship between O3 and AD symptoms was not linear. However, the relationship was linear within the range of 30–70 ppb of O3 concentration.
Fig 2 shows the effect of meteorological variables and air pollution on AD symptoms as a result of GLMM fitting after controlling for SCORAD at enrollment, age, sex, fever, and DOW. An increase in daily mean temperature by 5°C was significantly associated with a 12.8% (95% CI: 10.5, 15.2) decrease in AD symptoms on the same day (Table 3). A statistically significant association was also found between a decrease in AD symptoms and an increase in various moving averaged temperatures. Similarly, an increase in outdoor RH by 5% decreased the risk of AD symptoms from 3.3% (95% CI: 1.7, 4.8) with MA0 to 5.4% (95% CI: 2.9, 7.8) with MA0-5. This finding reflects the protective effect of elevated outdoor RH on AD. The effect was stronger as the number of moving average days increased, indicating that there was a lag effect. An increase in DTR of MA0-4, i.e. moving average of the previous 4 days and the current day, by 5°C increased AD symptoms by 10.6% (95% CI: -3.1, 26.2), but not significantly. Of note, we found that AD symptoms increased by 284.9% (95% CI: 67.6, 784.2) per 5°C increase in DTR when we fitted for the data >14°C DTR in the GLMM after controlling for the same confounders. A daily rainfall increase by 5 mm/day was significantly associated with an increase in AD symptoms by 2.2% (95% CI: 0.4, 4.1) on the same day. When we fit the same model for the daily rainfall data <40 mm/day, the effect size of rainfall increased to 7.3% (95% CI: 3.6, 11.1) per 5 mm increase in daily rainfall. However, the adverse effect of rainfall was not significant in the moving averages of daily rainfall after 2 days (MA0-2 through MA0-5).
Data represent percent changes and 95% confidence intervals in AD symptoms per 5-unit increase in daily mean temperature (°C), relative humidity (%), diurnal temperature range (°C) and 10-unit increase in PM10 (μg/m3), NO2 (ppb), and O3 (ppb). RH: relative humidity; DTR: diurnal temperature range; MA: moving average.
The concentrations of ambient PM10, NO2, and O3 were significantly associated with AD symptoms. An increase in the PM10 concentration by 10 μg/m3 increased the risk of AD symptoms by 3.2% (95% CI: 1.5, 4.9) with MA0 and 6.1% (95% CI: 3.6, 8.6) with MA0-5 for the whole range of PM10 concentration. When we fit the GLMM model for data with PM10 concentration over 35 μg/m3 (Fig 1), the risk of AD symptoms increased by 3.6% (95% CI: 1.5, 5.7) as PM10 increased by 10 μg/m3 on the same day, which is slightly higher than that for the whole range. The effects of the moving averages of PM10 from MA0 to MA0-5 on AD symptoms were statistically significant. This effect gradually increased as the number of averaging days increased. Similarly, the O3 concentration was positively associated with AD symptoms for all moving average levels of O3 (MA0 to MA0-5). Considering the relationship between O3 and AD symptoms was linear within the range of 30–70 ppb of O3 concentration (Fig 1), we selected data within this range. When the GLMM was fitted and the same confounders were controlled, the AD symptoms increased by 9.1% (95% CI: 3.6,14.9) per 10 ppb increase in O3 concentration. This effect was greater than 6.1% (95% CI:3.2, 9.0) for the whole range (Table 3). An increase in NO2 concentration by 10 ppb increased the risk of AD symptoms on the same day by 5.0% (95% CI: 1.4, 8.8) and there was no lag effect of NO2 on AD symptoms.
Fig 3 shows the effects of meteorological variables and air pollution on AD symptoms by sex. In both boys and girls, the positive association of temperature with AD symptoms was found. A similar result was observed in RH (Table 3). There was a significantly harmful effect of PM10 in girls, who showed a 5.2% (95% CI: 2.5, 8.0) change per 10 μg/m3 change in PM10 on the same day (MA0), whereas there was no such significant effect in boys. Boys exhibited 9.2% (95% CI: 4.4, 14.3) and 10.2% (95% CI: 6.4, 14.1) increases in AD symptoms by 10 ppb increase in NO2 and O3 concentration, respectively. However, there was no significant association of NO2 and O3 with AD symptoms in girls. These divergent effects between boys and girls more clearly observed when the moving average levels (MA0 to MA0-5) were fitted to the GLMM (Fig 3).
Data represent percent changes and 95% confidence intervals in AD symptoms per 5-unit increase in daily mean temperature (°C), relative humidity (%), diurnal temperature range (°C), and rainfall (mm) and 10-unit increase in PM10 (μg/m3), NO2 (ppb), and O3 (ppb). DTR: diurnal temperature range; RH: relative humidity; MA: moving average.
Our results demonstrated that short-term exposure to meteorological factors and air pollution are strongly associated with AD symptoms in children. Among four meteorological factors, increases in outdoor temperature and RH significantly reduced AD symptom scores. In contrast, increases in daily rainfall and DTR aggravated AD symptoms. Ambient PM10, NO2, and O3 increased the risk of AD symptoms. The effects of PM10 and O3 increased as the number of days of moving average increased, indicating that there were lag effects.
Our finding of negative associations between outdoor temperature and humidity and AD symptoms is consistent with a previous study in which a dry environment increased mast cell numbers and histamine content in the dermis of hairless mice . Lower temperature typically results in greater use of indoor heating and lower indoor humidity, both of which aggravate eczema symptoms . Consequently, our result supports the recommendation that outdoor temperature and humidity should be considered to alleviate AD symptoms, although the evidence level is not strong . In contrast, several studies have reported opposing results. Langan et al.  found that increased humidity and heat were associated with a higher incidence of disease flares in a 28-day observational study of 25 Irish children. In a prospective cohort study of American children, Sargen et al.  demonstrated that warm and humid climates are associated with poorly controlled eczema. The associations between weather and AD symptoms might vary according to region and season. More studies are required to elucidate the role of temperature and humidity in AD.
DTR as well as the absolute level of ambient temperature influences our health. DTR has an adverse effect on mortality and morbidity in cardiovascular and respiratory diseases [26–29]. An increase in DTR >10°C was associated with increased emergency room admissions in asthmatic children . However, the relationship between daily temperature variation and AD was not reported so far. In this study, we found that AD symptoms increased by 284.9% (95% CI: 67.6, 784.2) per 5°C increase in DTR when it was >14°C. These results suggest that children with AD should be more careful on days with elevated DTR, even on the days with high daily mean temperatures. To the best of our knowledge, this is the first study addressing the adverse effect of daily temperature fluctuation on AD symptom.
The Intergovernmental Panel on Climate Change (IPCC)  projected that in the future, extreme weather patterns (including cold/hot temperatures, drought, and flood) will increase in frequency as a result of climate change, particularly in temperate regions. According to a report on future climate change in the Korean peninsula published by the KMA , the average annual precipitation of the Korean Peninsula will exceed the natural variation and show a clear increase after the mid-21st century. This change is projected to occur in both representative concentration pathways (RCP) 4.5 and RCP 8.5, which are climate change scenarios adopted by the IPCC since 2011. This increase will be approximately 3.9 times the global average. It implies that the adverse effect of rainfall on AD symptoms found here would increase in the future, leading to poor control of preexisting AD symptoms.
Based on the large data set derived from daily symptom records, we demonstrated that air pollutants, including PM10, NO2, and O3, are related to the aggravation of AD symptoms; this finding supports those of previous studies [14, 15, 32]. The effects of PM10 and O3 were robust; the effect size increased as the MA0 went up to MA0-5 (Fig 2). Of interest, we found that the ambient O3 concentration had a greater effect on AD symptoms in children who were exposed within the range of 30–70 ppb than over the entire range. Although the mechanism is not clearly understood, behavioral changes might be involved. There is a possibility that patients with AD tried to avoid exposure to O3 when high levels were detected and announced by the local government. As O3 will increase by the climate change, particularly during warm regions such as the Korean peninsula [33–36], it is necessary to reduce O3 level for the management of AD.
Another interesting finding in this study was that there were differences in the harmful effect of air pollutants on AD symptoms according to sex. There was a significantly harmful effect of PM10 on AD symptoms in girls, whereas NO2 and O3 in boys. With regards to AD, it has been known that there are gender differences in prevalence and quality of life [37–39]. In addition, immune development and behavioral patterns such as outdoor play are different between boys and girls [40–42]. Infection and gut microbiota affects males and females differently . However, there is no evidence to explain the sex-specific response to air pollutants. Although boys have a different immune development in early childhood compared to girls, leading to a higher sensitization rate, total IgE level and peripheral eosinophil counts , the exact reason needs to be investigated.
In this study, we analyzed the relationships between AD symptoms and weather, and air pollution using GAMM and GLMM. GAMM contains a random effect, in addition to penalized spline curves, which the generalized additive model (GAM) contains. GAM and GAMM allow for nonparametric adjustments to nonlinear confounding effects on trends and other factors, such as weather variables [44, 45]. GAMM is appropriate to determine whether relationships are linear and, if so, whether positive or negative. However, GAM and GAMM are not used to estimate the effect sizes of risk factors because they are nonparametric models. In contrast, GLMM is an extension to the generalized linear model, while containing random effects in addition to the usual fixed effects . To estimate the effect size of the risk factors, we adopted a panel study using daily-repeated measurements for AD symptoms and environmental variables. This is the reason why we adopted two statistical models for data analysis in our study. GAMM was used to determine whether the relationships were linear or not, and GLMM to estimate the effect sizes of the risk factors on AD symptoms.
One of the strengths of this panel study is that we collected individual symptom records on a daily basis using smartphone-adjusted symptom diary. We also analyzed a large data set of 35,158 person-days by fitting GAMM and GLMM after controlling for potential confounders. By these efforts, we were able to obtain more objective and reliable results.
However, our study also has several limitations. For instance, socioeconomic factors such as household income and parental education levels were not controlled in the models. Although the study participants were enrolled in a single center and therefore their socioeconomic statuses were likely similar, it is important to consider such confounding factors in the analysis. Another limitation is that we did not evaluate the effect of indoor environment. It is well known that children spend most of their time indoors. If we, therefore, consider indoor environment in the exposure assessments, our results about the air pollutants and weather may be different from the current results. In case of air pollution, previous studies addressed that indoor environment such as PM10 and NO2 in Seoul are very dependent on outdoor condition . For example, indoor to outdoor ratios for PM10 and NO2 were 0.69 and 1.05, respectively [47–48]. This indicates that even if we considered indoor levels for PM10 and NO2, the effect of these factors would be similar as shown in the present study. In contrast, indoor temperature and RH are quite different from outdoor temperature and RH. If indoor levels of temperature and RH were controlled for the analysis of the present study, our results would be different. Further studies on the relationships between AD symptoms and indoor environmental factors including indoor temperature and RH are required.
This study demonstrates that the short-term exposure to ambient temperature, RH, DTR, and rainfall are significantly associated with AD symptoms in infants and young children living in a temperate region. Low ambient temperature and RH, large DTR, and high rainfall may aggravate AD symptom in children. In addition, PM10, NO2, and O3 showed significantly adverse effects on AD symptoms with gender differences. These findings could be applied to assist in establishing a management strategy of AD symptoms such as early warning system based on weather and air quality forecasts.
The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the Korea Ministry of Environment (MOE). We are grateful to all of those who recorded symptom diary data.
- Conceptualization: YMK KA JK YH HKC.
- Formal analysis: YMK BHJ.
- Funding acquisition: KA.
- Methodology: KA JK YMK HKC.
- Writing – original draft: YMK KA.
- 1. Ahn K. The role of air pollutants in atopic dermatitis. J Allergy Clin Immunol. 2014;134(5): 993–9; discussion 1000. pmid:25439225
- 2. Williams H, Stewart A, von Mutius E, Cookson W, Anderson HR, International Study of A, et al. Is eczema really on the increase worldwide? J Allergy Clin Immunol. 2008;121(4): 947–54 e15. pmid:18155278
- 3. Park YM, Lee SY, Kim WK, Han MY, Kim J, Chae Y, et al. Risk factors of atopic dermatitis in Korean schoolchildren: 2010 International study of asthma and allergies in childhood. Asian Pac J Allergy Immunol. 2016;34: 65–72. pmid:26994628
- 4. Bieber T. Atopic dermatitis. N Engl J Med. 2008;358(14): 1483–94. pmid:18385500
- 5. Boguniewicz M, Leung DY. The ABC's of managing patients with severe atopic dermatitis. J Allergy Clin Immunol. 2013;132(2): 511–2 e5. pmid:23905920
- 6. Byremo G, Rod G, Carlsen KH. Effect of climatic change in children with atopic eczema. Allergy. 2006;61(12): 1403–10. pmid:17073869
- 7. Hashimoto M, Fukuda T, Shimizu T, Watanabe S, Watanuki S, Eto Y, et al. Influence of climate factors on emergency visits for childhood asthma attack. Pediatr Int. 2004;46(1): 48–52. pmid:15043664
- 8. Lee YL, Su HJ, Sheu HM, Yu HS, Guo YL. Traffic-related air pollution, climate, and prevalence of eczema in Taiwanese school children. J Invest Dermatol. 2008;128(10): 2412–20. pmid:18449213
- 9. Silverberg JI, Hanifin J, Simpson EL. Climatic factors are associated with childhood eczema prevalence in the United States. J Invest Dermatol. 2013;133(7): 1752–9. pmid:23334343
- 10. Suarez-Varela MM, Garcia-Marcos Alvarez L, Kogan MD, Gonzalez AL, Gimeno AM, Aguinaga Ontoso I, et al. Climate and prevalence of atopic eczema in 6- to 7-year-old school children in Spain. ISAAC phase III. Int J Biometeorol. 2008;52(8): 833–40. pmid:18779981
- 11. Weiland SK, Husing A, Strachan DP, Rzehak P, Pearce N, Group IPOS. Climate and the prevalence of symptoms of asthma, allergic rhinitis, and atopic eczema in children. Occup Environ Med. 2004;61(7): 609–15. pmid:15208377
- 12. Sargen MR, Hoffstad O, Margolis DJ. Warm, humid, and high sun exposure climates are associated with poorly controlled eczema: PEER (Pediatric Eczema Elective Registry) cohort, 2004–2012. J Invest Dermatol. 2014;134(1): 51–7. pmid:23774527
- 13. Penard-Morand C, Raherison C, Charpin D, Kopferschmitt C, Lavaud F, Caillaud D, et al. Long-term exposure to close-proximity air pollution and asthma and allergies in urban children. Eur Respir J. 2010;36: 33–40. pmid:20075054
- 14. Kim J, Kim EH, Oh I, Jung K, Han Y, Cheong HK, et al. Symptoms of atopic dermatitis are influenced by outdoor air pollution. J Allergy Clin Immunol. 2013;132(2): 495–8 e1. pmid:23763977
- 15. Eberlein-Konig B, Przybilla B, Kuhnl P, Pechak J, Gebefugi I, Kleinschmidt J, et al. Influence of airborne nitrogen dioxide or formaldehyde on parameters of skin function and cellular activation in patients with atopic eczema and control subjects. J Allergy Clin Immunol. 1998;101(1 Pt 1): 141–3. pmid:9449520
- 16. Herbarth O, Fritz GJ, Rehwagen M, Richter M, Roder S, Schlink U. Association between indoor renovation activities and eczema in early childhood. Int J Hyg Environ Health. 2006;209(3): 241–7. pmid:16490398
- 17. Morgenstern V, Zutavern A, Cyrys J, Brockow I, Koletzko S, Kramer U, et al. Atopic diseases, allergic sensitization, and exposure to traffic-related air pollution in children. Am J Respir Crit Care Med. 2008;177(12): 1331–7. pmid:18337595
- 18. Hanifin J, Rajka G. Diagnostic features of atopic dermatitis. Acta Derm Venereol (Stockh). 1980;92(Suppl): 44–7.
- 19. European Task Force on Atopic Dermatitis: Severity scoring of atopic dermatitis. the SCORAD index. Consensus Report of the European Task Force on Atopic Dermatitis. Dermatology. 1993;186(1): 23–31. pmid:8435513
- 20. Jacob DJ, Winner DA. Effect of climate change on air quality. Atmos Environ. 2009;43: 51–63.
- 21. Tai APK, Mickley LJ, Jacob DJ. Correlations between fine particulate matter (PM2.5) and meteorological variables in the United States: Implications for the sensitivity of PM2.5 to climate change. Atmos Environ. 2010;44: 3976–84.
- 22. Wood SN, Augustin NH. GAMs with integrated model selection using penalized regression splines and applications to environmental modeling. Ecol Model. 2002;157: 157–77.
- 23. Kramer U, Weidinger S, Darsow U, Mohrenschlager M, Ring J, Behrendt H. Seasonality in symptom severity influenced by temperature or grass pollen: results of a panel study in children with eczema. J Invest Dermatol. 2005;124(3): 514–23. pmid:15737191
- 24. Schneider L, Tilles S, Lio P, Boguniewicz M, Beck L, LeBovidge J, et al. Atopic dermatitis: a practice parameter update 2012. J Allergy Clin Immunol. 2013;131(2): 295–9 e1-27. pmid:23374261
- 25. Langan SM, Bourke JF, Silcocks P, Williams HC. An exploratory prospective observational study of environmental factors exacerbating atopic eczema in children. Br J Dermatol. 2006;154(5): 979–80. pmid:16634905
- 26. Chen G, Zhang Y, Song G, Jiang L, Zhao N, Chen B, et al. Is diurnal temperature range a risk factor for acute stroke death? Int J Cardiol. 2007;116(3): 408–9. pmid:16863670
- 27. Lim YH, Hong YC, Kim H. Effects of diurnal temperature range on cardiovascular and respiratory hospital admissions in Korea. Sci Total Environ. 2012;417–418: 55–60. pmid:22281041
- 28. Song G, Chen G, Jiang L, Zhang Y, Zhao N, Chen B, et al. Diurnal temperature range as a novel risk factor for COPD death. Respirology. 2008;13(7): 1066–9. pmid:18922144
- 29. Xu Z, Huang C, Su H, Turner LR, Qiao Z, Tong S. Diurnal temperature range and childhood asthma: a time-series study. Environ Health. 2013;12: 12. pmid:23374669
- 30. Intergovernmental Panel on Climate Change (IPCC). Climate change 2013: the physical science basis. Seoul. 2013. https://ipcc.ch/pdf/assessment-report/ar5/wg1/WG1AR5_Chapter12_FINAL.pdf. Accessed 1 March 2016.
- 31. Korea Meteorological Administration (KMA). Korean peninsula climate change outlook report. KMA: Seoul. 2012. http://www.climate.go.kr/home/cc_data/2013/korea_climate/13-0311_korea_climate.pdf. Accessed 1 March 2016. Accessed 1 March 2016.
- 32. Song S, Lee K, Lee YM, Lee JH, Lee SI, Yu SD, et al. Acute health effects of urban fine and ultrafine particles on children with atopic dermatitis. Environ Res. 2011;111(3): 394–9. pmid:21367405
- 33. Kim M, Park R, Ho CH, Woo JH, Choi KC, Song CK, et al. Future ozone and oxidants change under the RCP scenarios. Atmos Environ. 2015;101: 103–15.
- 34. Ebi KL, McGregor G. Climate change, tropospheric ozone and particulate matter, and health impacts. Environ Health Perspect. 2008;116(11): 1449–55. pmid:19057695
- 35. Gao Y, Fu JS, Drake JB, Lamarque JF, Liu Y. The impact of emission and climate change on ozone in the United States under representative concentration pathways (RCPs). Atmos Chem Phys. 2013;13: 9607–21.
- 36. Kim YM, Zhou Y, Gao Y, Fu JS, Johnson BA, Huang C, et al. Spatially resolved estimation of ozone-related mortality in the United States under two Representative Concentration Pathways (RCPs) and their uncertainty. Clim Change. 2015;128(1–2): 71–84. pmid:25530644
- 37. Chernyshov PV. Gender differences in health-related and family quality of life in young children with atopic dermatitis. Int J Dermatol 2012;51: 290–4. pmid:22348564
- 38. Harris JM, Cullinan P, Williams HC, Mills P, Moffat S, White C, et al. Environmental associations with eczema in early life Br J Dermatol 2001;144: 795–802. pmid:11298539
- 39. Yu JS, Lee CJ, Lee HS, Kim J, Han Y, Ahn K, et al. Prevalence of atopic dermatitis in Korea: Analysis by using national statistics. J Korean Med Sci 2012;27: 681–5. pmid:22690101
- 40. Uekert SJ, Akan G, Evans MD, Li Z, Roberg K, Tisler C, et al. Sex-related differences in immune development and the expression of atopy in early childhood. J Allergy Clin Immunol. 2006;118(6): 1375–81. pmid:17157669
- 41. Stone MR, Faulkner GE. Outdoor play in children: associations with objectively-measured physical activity, sedentary behavior and weight status. Prev Med 2014;65: 122–7. pmid:24836417
- 42. Tagin MA, Woolcott CG, Vincer MJ, Whyte RK, Stinson DA. Hypothermia for neonatal hypoxic ischemic encephalopathy: an updated systematic review and meta-analysis. Arch Pediatr Adolesc Med 2012;166: 707–12.
- 43. Fish EN. The X-files in immunity: sex-based differences predispose immune responses. Nat Rev Immunol 2008;8:737–44. pmid:18728636
- 44. Dominici F, McDermott A, Zeger SL, Samet JM. On the use of generalized additive models in time-series studies of air pollution and health. Am J Epidemiol. 2002;156(3): 193–203. pmid:12142253
- 45. Lin X, Zhang D. Inference in generalized additive mixed models by using smoothing splines. J R Statist Soc B. 1999;61: 381–400.
- 46. Breslow NE, Clayton DG. Approximate inference in generalized linear mixed models. J Amer Statist Assoc. 1993;88(421): 9–25.
- 47. Lee JY, Ryu SH, Lee G, Bae GN. Indoor-to-outdoor particle concentration ratio model for human exposure analysis. Atmos Environ. 2016;127: 100–106.
- 48. Lee JY, Ryu SH, Kim C, Bae GN. Indoor-to-outdoor pollutant concentration ratio modeling of CO2, NO2, and lung-deposited nanoparticles. Atmos Pollut Res. 2016;7: 664–670.