Sun Exposure and Asthma Prevalence in Spanish Schoolchildren and Adolescents in the Global Asthma Network (GAN) Phase I Study: A Semi-Individual Cross-Sectional Study Authors: Alberto Arnedo-Pena, Saeed Fattahi, Inés Aguinaga-Ontoso, Alberto Bercedo-Sanz, Carlos González-Díaz, Angel López-Silvarrey Varela, Antonia Elena Martínez-Torres, Javier Pellegrini-Belinchón, Manuel Sánchez-Solís, Luis García-Marcos, the Spanish GAN Group Published: 2026 ABSTRACT Background/Objective: In the multifactorial etiology of asthma, environmental factors, including sun exposure, measured either by sunshine hours (SHs) or global solar surface irradiance (GSSI), are critical. The objective of this study is to estimate the association between sun exposure and the prevalence of asthma in Spanish schoolchildren (6–7 years old) and adolescents (13–14 years old) from six geographic centers in Spain (A Coruña, Bilbao, Cantabria, Cartagena, Pamplona and Salamanca) included in the Global Asthma Network (GAN) study during 2016–2019. Methods: Measurements of SHs and GSSI were obtained from the Spanish “Agencia Estatal de Meteorología”. A semi-individual cross-sectional design was used, and multilevel logistic regression models were employed, with adjustments for age, sex, temperature, relative humidity and gross domestic product, taking the center as the reference level. Permutation tests were performed to account for the low number of centers. Results: Increases in both SHs and GSSI were associated with a lower prevalence of “asthma ever” in the 6–7 and 13–14 years age groups. An increase of 100 annual sunshine hours was associated with a lower prevalence of “asthma ever” among schoolchildren (adjusted odds ratio [aOR] = 0.82; 95% confidence interval [CI], 0.79–0.85); from the crude model, permutation p = 0.08. In adolescents, the corresponding estimates were aOR = 0.90 (95% CI, 0.87–0.94); from the crude model, permutation p = 0.02. An increase of 1 kWh m−2 day−1 in GSSI was associated with a lower prevalence of “asthma ever” among schoolchildren (aOR = 0.15; 95% CI, 0.10–0.23); from the crude model, permutation p = 0.25. In adolescents, the corresponding estimates were aOR = 0.37 (95% CI, 0.25–0.54); from the crude model, permutation p = 0.01. Conclusions: The results of this study suggest that sun exposure, measured either as SHs or GSSI, may be protective against asthma in adolescents. In contrast, the association between sun exposure and the prevalence of “asthma ever” in schoolchildren was not significant, although it followed a similar trend. Confirmation of these findings in other geographical settings could improve our understanding of the role of sun exposure in asthma prevention and control. 1. Introduction According to asthma’s multifactorial etiology, some effects of climate on its prevalence have already been proven; studied meteorological variables include temperature, humidity, and thunderstorms [1,2,3,4,5]. More recently, it has been hypothesized that vitamin D deficiency may be associated with the increase in childhood asthma in Western countries [6,7]. This hypothesis has led to some observational studies on the prevalence of asthma and its association with sun exposure, in which geographical latitude or sunshine hours per year were used as proxies of vitamin D exposure [8,9]. These studies found an inverse relationship between childhood asthma prevalence and sun exposure in different countries, supporting this hypothesis. However, the effects of vitamin D on asthma incidence or prevalence continue to be disputed, and there is no consensus on the mechanisms, factors, or associated comorbidities [10,11,12]. Prenatal supplementation of vitamin D has been found to have no significant, or very limited, effect on asthma and allergies in children [13,14], despite some clinical trials observing significant decreases [15,16]. On the other hand, it was found that vitamin D supplementation in infants did not result in a reduction in the prevalence of wheezing or asthma after several years of follow-up [17,18,19]. Additionally, no effect on asthma control has been observed [20], and in a Mendelian randomization study, the level of vitamin D was not associated with allergic asthma [21]. However, vitamin D concentration was associated with decreased asthma incidence in a prospective cohort of adults in China [22], and an ecological study in England showed that in areas with the highest ultraviolet radiation, weighted by the pre-vitamin D action spectrum, there was a reduced prevalence of asthma in adults [23]. Epidemiological studies on the prevalence of asthma and geographical asthma variations have seen renewed interest in the context of climate change [24], with studies measuring the effect of sun exposure in new populations and distinguishing between sunshine hours (SHs) (a measure of the time of solar radiation) and global solar surface irradiance (GSSI) (total energy received from the sun). Sun exposure can cause diseases, such as skin cancer, and dermatological damage, but can also have health benefits [25,26]. Asthma-related benefits include increased vitamin D production, decreased transmission of infectious diseases, moderation of pollen seasons, and positive immunomodulation [27,28]. A worldwide study called the Global Asthma Network (GAN) study [29] measured the prevalence of asthma, eczema and allergic rhinitis in schoolchildren and adolescents in many countries, following the methods of the International Study of Asthma and Allergies in Childhood (ISAAC) established between 1994 and 2004 [30]. In Spain, five centers fully participated in this study from 2016 to 2019: Bilbao, Cantabria, and A Coruña on the Atlantic coast; Salamanca in the central east; and Cartagena on the Mediterranean coast. Pamplona, in the interior north, was the site of a survey during the same period, but this only included partial application of the GAN questionnaire and thus was not incorporated in the global dataset. The GAN study provides an opportunity to better study the effect of sun exposure on the prevalence of asthma in different geographical areas, expanding the ISAAC results obtained in 2011 [9]. The objective of the present study was to estimate the association between sun exposure, measured in sunshine hours (SHs) or global solar surface irradiation (GSSI), and the prevalence of asthma in schoolchildren and adolescents in Spain. 2. Materials and Methods 2.1. Study Design and Variable Definitions The study has a semi-individual design [31]. This design has been used in studies on environmental pollution [32] and has advantages over an ecological design, as outcomes are measured at the individual level and predictors are measured as aggregate data. Adjustment for aggregated variables can reduce ecological bias. The dependent variable was “asthma ever” based on the questionnaire used in the GAN study. In children aged 6–7 years old, “asthma ever” was defined as a positive answer (from parents) to the question “Has this child ever had asthma? (Yes/No)”, while in adolescents, the question “Have you ever had asthma? (Yes/No)” was answered by the adolescents themselves. For validation purposes, we also performed a sensitivity analysis to explore the consistency between the question on “asthma ever” and an “asthma diagnosis”. This was defined as a positive answer to the question “Was this child’s asthma confirmed by a doctor? (Yes/No)”, from the parents of schoolchildren, and “Was asthma confirmed by a doctor? (Yes/No)”, from adolescents themselves. The data on the prevalence of asthma were retrieved from the studies of the Spanish Group of the Global Asthma Network from 2016 to 2019 [33,34]. SHs and GSSI were used as predictor variables. SHs (from 1981 to 2010) were obtained from the Spanish “Agencia Estatal de Meteorología (AEMET)”, and the daily mean GSSI (period: 1983–2005) for the capital cities of each center’s province was acquired via satellite measurement [35]. Other variables included as potential confounding factors were the age and sex of the participant; annual average of temperature and relative humidity (period: 1981–2010), also taken from the AEMET [36]; and the annual averages of gross domestic product (GDP) per capita of the province where each center is located (2016–2019), obtained from the Spanish “Instituto Nacional de Estadística (INE)” [37]. 2.2. Statistical Analysis Asthma prevalence was calculated by dividing the number of positive answers to the aforementioned questions by the total number of participants for the two age groups for each center, expressed as percentages. Logistic regression was used to compare the asthma prevalence rates of each center using crude odds ratios (cORs) with 95% confidence intervals (CIs). Temperature was measured in degrees centigrade, relative humidity was taken as the percentage of water saturation in the air, and GDP was measured in Euros (€). Annual sunshine hours were divided by 100 to facilitate interpretation of the results, and associations between asthma prevalence (“asthma ever” and physician-diagnosed asthma) and sun exposure (SHs and GSSI) in schoolchildren and adolescents were estimated using multilevel logistic regression models, with odds ratios (ORs) as the measure of association. Multilevel logistic regression models were used in the multivariable analysis, and the center (Bilbao, Cantabria, A Coruña, Pamplona, Salamanca and Cartagena) was considered the reference level of the six analyzed zones and was also the random effect component. Sun exposure (SHs and GSSI), temperature, relative humidity and GDP were the fixed-effect components. This allows for comparing variations between centers and between participants while controlling for variables at multiple levels [38,39]. Table 1 presents the parameters of multilevel models and the null model, with covariates for “asthma ever” and “asthma diagnosis”, considering SHs and GSSI. It is worth noting that the inter-class correlation coefficient (ICC) of the null model for asthma diagnosis was 7.6%, which explains the effect of the center, with a median odds ratio (MOR) of 1.64, which quantifies the effect of center for schoolchildren; the ICC was 3.41% and MOR was 1.38 for adolescents. For “asthma ever”, these values were 6.9% and 1.60 for schoolchildren and 2.45% and 1.32 for adolescents, respectively. The full models with predictor variables exhibited substantially reduced variances, ICCs, and MORs for both SHs and GSSI. Model comparison via Akaike’s information criterion (AIC), Bayesian information criterion (BIC), log likelihood ratio (LLR), and deviance indicated the full models gave a better fit, as also evidenced by different models’ graphs of predicted and observed values [40]. Crude and adjusted odds ratios (cORs; aORs) with a 95% CI were calculated. These models allowed for adjustment for potential confounding factors such as age, sex, temperature, relative humidity, and GDP. We followed the Directed Acyclic Graph (DAG) method [41] to control for potential confounding factors, using DAGitty software version 3.1 to create a DAG diagram (Figure 1) [42]. Given the small number of centers included in the study, permutation tests were performed for the crude SH and GSSI multilevel models using the “ritest” command developed by S. Heß, with 100 replications [43]. In addition, a leave-one-center-out sensitivity analysis was conducted by sequentially removing one center at a time from the crude SH and GSSI analysis to identify centers that might disproportionately influence the associations between asthma prevalence and sun exposure or the estimated odds ratios (ORs). Statistical analyses were performed using Stata™ version 14.2 (StataCorp LLC, College Station, TX, USA). 2.3. Ethics The GAN study in Spain was approved by the Clinical Research Ethical Committee of the “Hospital Clínico Universitario de la Arrixaca”, Murcia. Passive informed consent was applied to questionnaires, which were anonymized. 3. Results The GAN study took place in Spain in 2016–2019, with a total of 17,215 schoolchildren, aged 6–7 years old, and 19,943 adolescents, aged 13–14 years old, participating. The overall response rate was 62.6% (17,215/27,497) for schoolchildren and 81.3% (19,943/24,516) for adolescents. Table 2 shows the asthma prevalence for “asthma ever” and “asthma diagnosis” in the six centers for both age groups, and includes a comparison, conducted by means of logistic regression analyses, using the center with the lowest prevalence as the reference. There are large differences between the six centers: the highest prevalence of “asthma ever” and “asthma diagnosis” in schoolchildren was observed in Bilbao, at 22.7% (OR = 4.34; 95% CI 3.59–5.24) and 21.0% (OR = 4.49; 95% CI 3.68–5.47), respectively. In adolescents, the highest prevalence was again in Bilbao at 29.9% (OR= 2.43; 95% CI 2.16–2.74) for “asthma ever” and 26.9% (OR = 2.78; 95% CI 2.45–3.15) for “asthma diagnosis”, followed by Cantabria, at 17.2% (OR = 3.07; 95% CI 2.54–3.73) and 15.1% (OR = 3.01; 95% CI 2.46–3.67) in schoolchildren and 23.6% (OR = 1.76; 95% CI 1.56–1.98) and 20.6% (OR = 1.96; 95% CI 1.73–2.22) in adolescents, respectively. A low asthma prevalence was observed both in schoolchildren and adolescents in A Coruña and Salamanca. Pamplona exhibited the lowest prevalence of “asthma ever” and “asthma diagnosis” in schoolchildren at 6.3% and 5.6%, respectively, while Cartagena had the lowest in adolescents, at 14.9% and 11.7%, respectively. Table 3 shows SHs, GSSI, temperature (°C), relative humidity (%), and GDP per capita (€) in the six centers. Bilbao and Cantabria had the lowest number of sunshine hours/year and daily solar irradiance, while Salamanca and Cartagena had the highest, and the mean annual temperatures were lowest in Salamanca and Pamplona and highest in Cartagena. Relative humidity exhibited small differences between centers, and per capita GDP was the highest in Pamplona and Bilbao and the lowest in Salamanca and Cartagena. Table 4 shows the results of the multilevel logistic regression models for “asthma ever”, including SHs and GSSI, in the two age groups and the six centers. In the adjusted models, an increase of 100 sunshine hours per year was associated with a significant decrease in the prevalence of “asthma ever” in schoolchildren (aOR = 0.82; 95% CI 0.79–0.85) and adolescents (aOR = 0.90; 95% CI 0.87–0.94). On the other hand, an increase of one kWh m−2 day−1 in daily GSSI significantly reduced the prevalence of “asthma ever” in schoolchildren (aOR = 0.15; 95% CI 0.10–0.23) and adolescents (aOR = 0.37; 95% CI 0.25–0.54). Thus, the reducing effect of GSSI on the prevalence of “asthma ever” was higher than that of SHs in both groups. During the sensitivity analysis (Table 5), the prevalence of “asthma diagnosis” and sun exposure was also analyzed using multilevel logistic regression. In the adjusted models, an increase of 100 h in mean annual SHs was significantly associated with a lower prevalence of asthma diagnoses both in schoolchildren (aOR = 0.81; 95% CI 0.77–0.84) and adolescents (aOR = 0.88; 95% CI 0.85–0.92). Similarly, an increase of one kWh m−2 day−1 in daily GSSI was also associated with a decrease in the prevalence of asthma diagnoses, again both in schoolchildren (aOR = 0.13; 95% CI 0.08–0.20) and adolescents (aOR = 0.32; 95% CI 0.22–0.47). The effects of SHs were lower than those of GSSI in both groups. Table 6 presents the results of the permutation tests. Among schoolchildren, neither SHs (p = 0.08) nor GSSI (p = 0.25) was significantly associated with the prevalence of “asthma ever” or the prevalence of physician-diagnosed asthma SHs (p = 0.07 and p = 0.13, respectively). In contrast, among adolescents, both SHs (p = 0.02) and GSSI (p = 0.01) were significantly associated with the prevalence of “asthma ever” and the prevalence of physician-diagnosed asthma (p = 0.04 and p = 0.02, respectively). Table 7 shows the results of the leave-one-center-out sensitivity analysis, in which one center was sequentially removed from the crude SH and GSSI analyses. Among schoolchildren, removal of the Bilbao center led to non-significant associations between both SHs and GSSI and the prevalence of “asthma ever”, whereas removal of the Cartagena center led to significant associations. A similar pattern was observed for physician-diagnosed asthma: associations became significant after exclusion of the Cartagena center but were no longer significant after exclusion of the Bilbao center. Considerable variations in the estimated odds ratios (ORs) were observed for GSSI. Among adolescents, significant associations between both SHs and GSSI and the prevalence of “asthma ever” remained after the exclusion of each individual center. Likewise, significant associations with physician-diagnosed asthma persisted regardless of the center removed, with only minor variations in the estimated ORs. 4. Discussion The present study reveals a significantly lower prevalence of “asthma ever” with higher sun exposure, measured by both SHs and GSSI, among adolescents from six Spanish centers. In addition, the sensitivity analyses showed that the prevalence of physician-diagnosed asthma also decreased with increasing sun exposure in this age group. In contrast, among schoolchildren, neither SHs nor GSSI was significantly associated with the prevalence of “asthma ever” or physician-diagnosed asthma, although the direction of the association was similar to that observed in adolescents. Multilevel models based on a small number of centers may yield spuriously low p-values; permutation tests and leave-one-center-out sensitivity analyses are useful to identify this phenomenon. The lack of significant associations in schoolchildren may be explained by the limited between-center variability in asthma prevalence and the disproportionate influence of two centers on the results. However, the effect of sun exposure may be weaker in schoolchildren than in adolescents, or other age-related risk factors may play a more important role in younger children [44,45]. These results are consistent with those of several epidemiological studies that used different proxies for sun exposure, such as latitude, studied in the USA and Australia [8]; sunshine hours, studied in West Europe in the ISAAC study [46]; and, more recently, ultraviolet radiation weighted by the pre-vitamin-D action spectrum, which was investigated in England [23]. The results of the present study are consistent with those of a study by Arnedo-Pena and co-authors [9] in comparable geographical areas, with decreases in asthma prevalence of 0.6% for the 6–7-year-old group and 1.1% for the 13–14-year-old group following a 100 h increase in mean annual sunshine. As found in the present study, centers located in the coastal area of northern Spain with an Atlantic climate (Bilbao, Cantabria, and A Coruña) had higher asthma prevalence rates and lower sun exposure. Pamplona, with a transitional Atlantic–Mediterranean climate, had the lowest asthma prevalence rates in schoolchildren, and Cartagena, with high sun exposure, had the lowest asthma prevalence in adolescents. The association between GSSI and asthma prevalence was stronger than that observed for SHs in adolescents, suggesting that the reduction in asthma prevalence associated with greater sun exposure may extend beyond the potential effects of sunlight on vitamin D production [47]. Additional mechanisms may include increased resistance to respiratory infections, protection of the mucosal barrier, and enhancement of immune function through multiple biological pathways [21,48,49]. The immunomodulation effect of GSSI, triggered by ultraviolet radiation, produces anti-inflammatory, anti-allergic, and anti-autoimmunity effects via activation of the innate immune system, which would help to reduce asthma susceptibility [28,44,50,51]. In addition, GSSI decreases the presence of house-dust mites, mold and fungal spores, and other allergens through increasing temperature and reducing humidity. Conversely, low solar irradiation increases the risk of allergic reactions to house-dust mites [52,53]. Additionally, solar irradiation can decrease the transmission of viral agents [54] and reduce food sensitization and eczema in infants [47,55], and exposure during the second trimester of pregnancy has been associated with reduced aeroallergen sensitization in two groups of children aged 6 and 12 years old, while irradiation in the third trimester has been associated with increased sensitization to these allergens [56]. In a recent prospective cohort study performed in Paris [57], solar irradiation during both the prenatal and postnatal periods and higher levels of vitamin D in the prenatal period reduced the risk of sensitization of children aged 8–9 years. Periods of intense solar and geomagnetic activity have been shown to improve asthma control, probably by decreasing airway inflammation, in a cohort of asthmatic children in Boston, further supporting the benefits of solar irradiation [58]. Regarding asthma prevalence in different climates, areas with a sunny climate exhibit lower prevalence compared with humid areas [59,60,61,62]. However, it remains unknown whether the potential effects of vitamin D on asthma depend on the asthma phenotype. Multi-omics studies may help to better characterize this association [63]. The strengths of this study include the following: (1) The high number of participants in each center and (2) a particularly high participation rate in the 13–14-year-old group and a good participation rate in the 6–7-year-old group. (3) The use of the validated asthma symptom questionnaire from the ISAAC and GAN studies [29,34] and (4) the use of the Spanish official meteorology agency-measured sun exposure data, with a satellite-based determination of GSSI. (5) The multilevel logistic regression models allowed us to examine the individual prevalence of asthma while controlling for center characteristics and other potential confounding factors. (6) Finally, the sensitivity analysis considering doctor-confirmed asthma diagnoses strengthened the initial results for “asthma ever”. However, this study has some limitations: Firstly, the semi-individual design reduces the ecological fallacy, but this cannot be fully ruled out considering that exposure variables are aggregated measurements and they may have considerable individual variations. Secondly, the cross-sectional approach prevents us from inferring a causal relationship between exposure and outcome, and only associations can be concluded. Thirdly, the small number of centers included in our multilevel models may have resulted in spuriously low p-values. Fourthly, exposure variables were measured in the capital cities of provinces, and there may be certain differences for centers not located in these capitals (for example, Cantabria) or for those that are located in the capital city but for which the measured area of sun exposure does not coincide with the GAN-surveyed areas. Fifthly, self-reporting of asthma prevalence by participants or their parents may produce some information bias. Sixthly, potential confounding factors such as air pollution, lifestyle, food or food supplement intake, and indoor environmental and social determinants were not included in the analysis [64,65,66]. Finally, some residual confounding could remain; we also used the means of meteorological variables from long time periods. Nevertheless, the results of the present study suggest a protective effect of sun exposure on the prevalence of asthma in adolescents, with GSSI having a greater effect than SHs. Solar irradiation may have a beneficial effect against several diseases, and increasing research in this area suggests revising the recommendations for sun exposure [26,44,67,68]. However, the harmful effects of sun irradiation should also be taken into account. Further epidemiological observational studies in different geographical areas would be useful to increase knowledge on the causes of sun-exposure- and climate-related asthma; these results may help improve asthma prevention and control. Additionally, prospective cohort studies beginning during pregnancy and following participants through infancy and childhood would be advantageous to investigate the causal relationship between sun exposure and asthma incidence. 5. Conclusions The results of the present study suggest that sun exposure, measured either by SHs or GSSI, may be protective against asthma in adolescents, while in schoolchildren, the associations between sun exposure and asthma prevalence were not significant, although they followed a similar trend. Confirmation of these findings in other geographical settings would improve our understanding of the role of sun exposure in asthma prevention and control.