Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Birch pollen, air pollution and their interactive effects on airway symptoms and peak expiratory flow in allergic asthma during pollen season – a panel study in Northern and Southern Sweden

Birch pollen, air pollution and their interactive effects on airway symptoms and peak expiratory... Background: Evidence of the role of interactions between air pollution and pollen exposure in subjects with allergic asthma is limited and need further exploration to promote adequate preventive measures. The objective of this study was to assess effects of exposure to ambient air pollution and birch pollen on exacerbation of respiratory symptoms in subjects with asthma and allergy to birch. Methods: Thirty‑seven subjects from two Swedish cities (Gothenburg and Umeå) with large variation in exposure to both birch‑pollen and air pollutants, participated in the study. All subjects had confirmed allergy to birch and self‑reported physician‑ diagnosed asthma. The subjects recorded respiratory symptoms such as rhinitis or eye irrita‑ tion, dry cough, dyspnoea, the use of any asthma or allergy medication and peak respiratory flow (PEF), daily for five consecutive weeks during two separate pollen seasons and a control season without pollen. Nitrogen oxides (NO ), ozone (O ), particulate matter (PM ), birch pollen counts, and meteorological data were obtained from an urban 3 2.5 background monitoring stations in the study city centres. The data were analysed using linear mixed effects models. Results: During pollen seasons all symptoms and medication use were higher, and PEF was reduced in the subjects. In regression analysis, exposure to pollen at lags 0 to 2 days, and lags 0 to 6 days was associated with increased ORs of symptoms and decreased RRs for PEF. Pollen and air pollution interacted in some cases; during low pollen exposure, there were no associations between air pollution and symptoms, but during high pollen exposure, O concentra‑ tions were associated with increased OR of rhinitis or eye irritation, and PM concentrations were associated with 2.5 increased ORs of rhinitis or eye irritation, dyspnea and increased use of allergy medication. *Correspondence: hanne.krage.carlsen@amm.gu.se Section of Occupational and Environmental Medicine, School of Public Health and Community Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Medicinaregatan 16A, 40530 Gothenburg, Sweden Full list of author information is available at the end of the article © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Carlsen et al. Environmental Health (2022) 21:63 Page 2 of 13 Conclusions: Pollen and air pollutants interacted to increase the effect of air pollution on respiratory symptoms in allergic asthma. Implementing the results from this study, advisories for individuals with allergic asthma could be improved, minimizing the morbidities associated with the condition. Keywords: Birch, Betula, PM , O , Panel study, Allergic asthma, Pollen season 2.5 3 Background severe disease, as indicated by either steroid medica- Exposure to air pollution can cause a variety of adverse tion use, or have poorer asthma control. health effects such as respiratory illness, cardiovascu - lar disease and lung cancer, and it has been estimated Methods that approximately 9 million people die each year from Study population causes directly attributed to air pollution [1]. The prev - Forty-one non-smoking subjects (27 to 73 years of age) alence of asthma is reported to be 4.3% globally, and with self-reported physician diagnosed asthma and a as high as 10% in Sweden [2]. The prevalence has been confirmed birch allergy (positive skin-prick test, phadi - increasing in the last decade [3] and allergic asthma is atop ≥ 0.35 kU/L) were recruited to participate. 22 sub- the most commonly reported form of asthma in Swe- jects were recruited in Gothenburg (Lat. 57° N) in the den with an estimated prevalence of 7.3% [4]. southern part of Sweden and 17 from Umeå (Lat. 63° Exposure to air pollutants such as ozone, particulate N), in northern Sweden. Eventually, 22 and 15 had diary matter (PM), and nitric oxides (N O ) is associated with data and were included in the analysis. Subjects with increased rhinitis severity [5] and short term expo- cardiovascular disease or any chronic inflammatory dis - sure can induce airway inflammation, due to increased ease were excluded from participation. All participating oxidative stress, and further result in airway hyperre- subjects signed an informed consent agreement, and the sponsiveness in asthmatic individuals and sensitization study was approved by the Regional Ethical Review Board [6, 7] and poor asthma control [8]. Birch (betula) pollen at the University of Gothenburg (Dnr: 682 − 14). is a common allergen in allergic asthma [9]. Inhalation of pollen grains (who are themselves too large to reach Study design the small airways) results in the rupture of the grain The participating subjects filled out a health diary which and release of the allergenic content (Bet v1 protein) included questions about daily symptoms and medica- which activates an immune response, [10] and cause tion who had been used in previous studies [23, 24] for bronchoconstriction in individuals with allergic asthma five consecutive weeks (35 days) during two separate pol - [11]. Increased allergenicity of birch pollen following len-seasons (2015 and 2016) and once outside the pollen- exposure to ozone [12] and PM [13] has been reported season (November-December 2015). and it has also been speculated that pollution attaches The symptoms registered in the diary (Supplementary to the allergen [14]. Anthropogenic air pollution affects document D1) were cold, fever, rhinitis, dyspnea and dry the abundance and bioavailability of pollen, and future cough (yes or no), and allergy medication (yes or no), climate change may affect the duration and severity of bronchodilation medicines (compared to their regular pollen seasons [15–18]. Pollen and air pollution peaks dose)” (no, less, normal or more), and asthma symptoms often co-occur [19], therefore it is of interest to study (no, less, normal or more). Peak respiratory flow rate pollen and air pollution together. Although synergis- (PEF) measurements were made twice daily throughout tic effects or interactions between air pollution and the five weeks, in the morning before any medication and pollen on respiratory health outcomes in individuals before bedtime, with three registered measurements for with allergic asthma have been reported [14, 20], many each occasion. uncertainties remain, the estimated risks have a large The means for morning and evening PEF for each sub - range, and more studies on this topic are warranted [21, ject during the study waves  were first calculated (PEF mo 22].and PEF ). Individual deviations of daily performance ev The aim of the present study was to study associa - from each subject’s mean PE F and PE F per season mo ev tions and interactions between pollen and air pollut- were then calculated and averaged across the participants ants associated with worsening of daily self-reported to obtain daily mean deviations (∆) for PEF and PEF mo ev respiratory symptoms and peak expiratory flow (PEF). values, ∆PEF and ∆PEF . mo ev Furthermore, we wanted to assess if there are lagged To avoid error or bias due to effects of respiratory dis - effects of exposure to pollen and air pollution on symp - ease epidemics, we excluded days where participants toms reports, study effects in individuals with more reported having a “fever”. In total, 8 individuals reported C arlsen et al. Environmental Health (2022) 21:63 Page 3 of 13 “fever” on at least one occasion (46 observations with and pollen seasons. The proportion reporting symptoms, “fever”, 3133 observations with “no fever” reported). The medication use, and mean PEF was reported for each presence of cold was reported on at least one day by 22 of season. the participants, with a total of 297 observations of cold, The outcomes were analyzed with mixed models, 2855 observations with no cold), but as the symptoms of where subjects were included as a random effect, and cli - cold are somewhat similar to those of asthma and allergy, mate and air pollution variables were included as fixed they were retained in the data, but not analysed. effects. All subjects underwent a clinical examination once The mean of lag-intervals lag 0–2, and secondarily lag within each season and completed an Asthma Con- 0–6 were chosen based on previous literature. However, trol Questionnaire (ACQ) with six questions regarding the lag-association between outcomes and pollen and air symptoms and use of bronchodilators and inhaled ster- pollution exposure was also investigated using distrib- oids during the week before the clinical examination uted lag non-linear models (DLNM) [29] at lags 0 to lag and pre-bronchodilator forced expiratory volume in one 10 with different options for the shape of the lag-associ - second (FEV ) in categories [25]. The participants also ations and levels of adjustments. The models with best reported their medication usage in the previous month in (lowest) AIC were selected and plotted (Supplementary a questionnaire. Fig. S2) and showed that in most cases, the significant association between exposure and outcome did, in fact, Exposure occur at lag 0–2. Birch pollen counts were obtained from centrally located In the analysis of morning PEF values, same day pollen pollen monitoring stations using standard methods [26]. exposure was not included as they were deemed of less Ambient concentrations of the air pollutants NO, O interest, only exposure lags from the previous day and x 3 and PM , and meteorological data of relative humid- backwards in time were included (lag1, lag 1–2 and lag 2.5 ity and temperature were measured at centrally located 1–5). fixed monitoring stations in the two study centers and Logistic regression was used for analysis of the daily obtained from the environmental authorities as previ- symptoms in asthma-diary and linear regression for PEF. ously described [27]. In this study, we used P M as In the main analysis, data from all study seasons were 2.5 this was available in both locations for the study period. analyzed separately for each study center (reported in the In brief, routinely collected air pollution, pollen, and Supplement), then the results were pooled using random meteorology data were provided to us from the relevant effects meta-analysis. authorities (Gothenburg: Gothenburg municipality Single-exposure models included the covariates pol- h t t p s :// got eb org . s e/ w p s/ p or t al/ st ar t/ milj o/ milj o l age t -i- len, temperature and relative humidity, dual-exposure goteb org/ luft/ luftk valit eten-i- goteb org, and the Pollen models included pairwise combinations of pollen and air Laboratory at Gothenburg University https:// www. gu. pollutants O, NO , and PM , temperature, and rela- 3 x 2.5 se/ biolo gi- miljo veten skap/ pollen- och- aller gier), Umeå tive humidity. Multi-exposure models included pollen, all Municipality (https:// www. umea. se/ bygga booch miljo/ three pollutants, temperature, and relative humidity. b o end emilj obull er o c h luf t k v alit e t/ luf t e n ut om h u s/ luf t k Interaction analysis for pollen and air pollution were valit eteni umea.4. 250f9 65917 4ae4b 97941 ae7. html). Rel- first investigated in mixed models with id and study evant lags were calculated and used in the analyses. All center as random effects with an indicator variable for participating subject lived within, or in direct vicinity to the presence of pollen season (pollen season vs. con- the city centers in question. trol season) for the association between outcomes and The data were converted to 24-hour means and were air pollutants in all data. Then, using only pollen season assigned to each study date and lagged values were calcu- data, we investigated interaction between low and high lated [27]. The birch pollen season was defined as the first levels of pollen at lag 0–2 (using 100 pollen/m as a cut- occurrence of 5 consecutive days where 4 days had pollen off ) and air pollution values. counts > 0 [26] (not dissimilar to the recently proposed The sample size power calculation was based analyses pollen exposure threshold from recent position paper) of effects in respiratory biomarkers in analyses compar - [28]. ing effects in pollen season and control season. In the current study, each individual had up to 105 measure- Statistical analyses ments (days with diary data), the statistical power is suffi - Descriptive statistics for diary-reported outcome and cient. In order to estimate effect of disease severity on the exposure variables are presented as mean and stand- response from pollen exposure, stratified sensitivity anal - ard deviations (SD) and categorical variable frequencies yses were performed in individuals who reported using were compared with Chi2-tests between control season inhalation corticosteroids, nasal steroids, individuals with Carlsen et al. Environmental Health (2022) 21:63 Page 4 of 13 Table 1 Demographic characteristics of the participants at baseline Variables Gothenburg Umeå Total Subjects, N 22 15 37 Females, n (%) 12 (55%) 7 (47%) 19 (51%) Males, n (%) 10 (45%) 8 (53%) 18 (49%) Age, yr (mean ± SD) 44 ± 10 47 ± 13 48 ± 15 Height, cm (mean ± SD) 175 ± 9 172 ± 9 171 ± 8 Weight, kg (mean ± SD) 79 ± 18 78 ± 16 77 ± 15 Body mass index, kg/m2 (mean ± SD) 26 ± 4 26 ± 6 26 ± 5 Birch pollen allergy, n (%) 22 (100%) 15 (100%) 37 (100%) Asthma Control Questionnaire (mean ± SD) 1.3 ± 1.1 1.3 ± 1.2 1.3 ± 1.1 Self-reported medication Short‑acting bronchodilator, n (%) 18 (82%) 6 (40%) 24 (65%) Long‑acting medication, n (%) 2 (9%) 2 (13%) 4 (11%) Inhalation steroids medication, n (%) 19 (86%) 6 (40%) 25 (68%) Nasal steroid, n (%) 7 (32%) 2 (13%) 9 (24%) Steroid medication (Inhalation and combination steroids), n (%) 19 (86%) 8 (53%) 27 (73%) Allergy medicine, n (%) 22 (100%) 11 (73%) 33 (89%) Asthma control questionnaire score calculated as a mean of six questions about asthma symptoms in the last week Regular usage in the last month or year poor asthma control as measured by ACQ (poor asthma in the analysis. The proportion of females was 51%, the control being defined as a score of ACQ above 1.5) [25]. mean age was 48 (SD 15) years, and the mean BMI was To determine if drop-out rates affected the results, we 26 (SD 5) kg/m2. The participants’ mean ACQ score at also analyzed the individuals who participated in all three baseline was 1.3 (SD 1.1). The regular use of short-act - waves separately. Data were also analyzed stratified by ing bronchodilator during the last year were reported sex. by 65% of the participants, long-acting bronchodilator The models for PEF were adjusted for autocorrelation medication by 11% and inhalation corticosteroid use was by incorporating a first-order autoregressive component. reported by 68% of the participants (Table 1). This autocorrelation function is not defined for logistic Air pollutants and pollen concentrations from station- regression models. ary measuring stations varied between seasons and study All results are reported per 100-unit pollen and one cites. NO concentrations were lower in Umeå than in interquartile range (IQR) change in air pollution con- Gothenburg during spring in both 2015 and 2016, mean 3 3 3 3 centrations. The results for diary reported symptoms 9.2 µg/m and 14.6 µg/m vs. 25.6 µg/m and 28.7 µg/m and medication usage are reported as odds ratios (ORs), respectively, but higher during control season by 39.5 µg/ 3 3 which express the relative change in the odds associated m vs. 28.9  µg/m although the standard deviation for with change in exposure with 95% confidence intervals the control season in Umeå was high. O concentrations (CI). The coefficients for ∆PEF are reported with 95% were quite higher in Umeå than in Gothenburg during 3 3 CI, which express the increase in outcome per increment spring 2016, 86.8 µg/m vs. 53.9 µg/m. PM concentra- 2.5 change in exposure. The interaction terms are reported tions were two to three times as high in Gothenburg as in with their confidence intervals and p-values, and the pre - Umeå in all seasons (Table 2, Fig. S1). dicted values of the outcomes at low and high values of Pollen concentrations were higher in Gothenburg than the variable pollen at lag 0–2 were plotted. Analyses were in Umeå spring 2015 (maximum pollen value 1087 vs. performed using the packages “lme4”, “meta”, and “ggef- 210), but during spring 2016 pollen concentrations were fects” [30–32] in R. marginally higher in Umeå than Gothenburg (maximum pollen value 964 vs. 950) (Table  2; Fig.  1). Temperatures Results and relative humidity also differed between locations and Descriptive statistics seasons reflecting the geographic differences between the Initially, 41 individuals with allergic asthma were locations (Table 2). recruited for the study, and 37 individuals had informa- Data are measured at centrally located ambient meas- tion on all key variables of interest and were included uring stations. C arlsen et al. Environmental Health (2022) 21:63 Page 5 of 13 Table 2 Daily averages of exposure in the two study centers medication on 40.7% of the days. Using more bronchodi- during the study waves (mean ± standard deviation) lation medicines than normal was reported on 9.2% of days (Table 3). Mean morning PEF  was 436 mL (SD 88) mo Wave Gothenburg Umeå and mean PEF was 444 (SD 92). ev x 3 NO (µg/m ) 1 (Pollen season) 25.6 ± 14.7 9.2 ± 5.0 Rates of symptoms and medication usage were higher 2 (Control season) 28.9 ± 18.3 37.5 ± 30.7 during pollen season compared to the control season 3 (Pollen season) 28.7 ± 19.0 14.6 ± 6.6 (visualized in Fig.  2). PEF-values during pollen season 3 3 O (µg/m ) 1 (Pollen season) 66.0 ± 12.3 63.5 ± 7.7 were not statistically significant different from control 2 (Control season) 52.0 ± 15.7 40.0 ± 19.4 season, with p-value = 0.15 for PEF and p-value = 0.24 mo 3 (Pollen season) 53.9 ± 13.8 86.8 ± 29.0 for PEF (data not shown). ev PM (µg/m ) 1 (Pollen season) 9.0 ± 6.3 3.7 ± 2.7* 2.5 2 (Control season) 6.9 ± 4.3 2.1 ± 1.3 Eec ff ts on symptoms and medication 3 (Pollen season) 7.1 ± 3.9 4.6 ± 2.6 In the pooled analysis (Table  4), pollen exposure at Pollen (grains per m ) 1 (Pollen season) 180 ± 254 43 ± 49 lag 0–2 was statistically significantly associated with 2 (Control season) ‑ ‑ increased ORs of reporting rhinitis or eye irritation OR 3 (Pollen season) 176 ± 221 231 ± 270 1.44 (95% CI 1.36–1.52) per 100 pollen/m pollen in the Temperature (°C) 1 (Pollen season) 9.1 ± 1.9 1.0 ± 4.2 single-exposure model, and lower, but still statistically 2 (Control season) 7.5 ± 3.5 10.0 ± 3.6 significant ORs in dual-exposure and multi-exposure 3 (Pollen season) 10.8 ± 4.4 5.9 ± 5.0 models (City-specific ORs are shown in Table S2a). The Relative humidity (%) 1 (Pollen season) 72.5 ± 11.0 75.1 ± 11.2 same was true for pollen exposure at lag 0–6 which was 2 (Control season) 85.5 ± 6.9 95.1 ± 6.8 associated with increased ORs of rhinitis or eye irrita- 3 (Pollen season) 64.7 ± 12.9 68.9 ± 12.6 tion in single-exposure models by OR 1.67 (95% CI 1.57– 1.78). PM was significantly associated with increased 2.5 rhinitis and or irritation in multi-exposure models by OR 1.16 (95% CI 1.02; 1.32). After initially testing individual lags from 0 to 3 (data Dyspnea was only significantly associated with pollen not shown), lag 0–2, and lag 0–6, for both pollen and air exposure in single-exposure models by OR 1.17 (95% CI pollution, and afterwards investigating lag-associations 1.10–1.25) for exposure at lag 0–2 and OR 1.23 (95% CI with DLNM methods, it was found that the highest ORs 1.09–1.37) for exposure at lag 0–6. There were no statisti - for dry cough and allergy medication were at lag 0–2 and cally significant associations between air pollutants and lag 0–6, and these lags are reported onwards. For PE F , dyspnea. mo we report results for the mean of the previous day and Dry cough was associated with pollen exposure at the day before that, lag 1 to lag 2 instead of lag 0 to lag 2. lag 0–2 in dual-exposure models with NO by OR 1.10 After excluding observations where participants reported (95% CI 1.02–1.18) and with O3 by OR 1.09 (95% CI having a fever, there were 3112 observations. In the dia- 1.01; 1.17) and in multi-exposure models by OR 1.09 ries, the most reported symptoms were rhinitis or eye (95% CI 1.01–1.17). For pollen exposure at lag 0–6 irritation, reported on 27.5% of days during the study there were statistically significant associations in dual- period, followed by dry cough, reported on 26.4% of all exposure models adjusted for O, PM by 1.14 (95% CI 3 2.5 days in the study. Participants reported taking allergy 1.05–1.24) and in multi-exposure models by OR 1.15 Fig. 1 Pollen concentrations in the two study centres (3‑ day moving average) during the study period Carlsen et al. Environmental Health (2022) 21:63 Page 6 of 13 Fig. 2 Proportion of diary‑reported respiratory symptoms and medication use during the study period. For eye irritation/rhinitis, dyspnea, dry cough, allergy medication, proportion reporting “Yes” vs. “No”. For asthma medication, reporting “more than yesterday” vs. “No”, “Less”, or “Same” (95% CI 1.05–1.28). Exposure to O at lag 0–2 was sta- multi-exposure models was noted. Exposure to N O 3 x tistically significantly associated with dry cough by OR at lag 0–2 was associated with significantly reduced 1.15 (95% CI 1.00-1.31) per IQR O in models with pol- ORs of allergy medication usage in both dual-and len, and by 1.14 (95% CI 1.09–1.19) in multi-exposure multi-exposure models, with OR at 0.72 (95%CI 0.52- models. Dry cough rates were associated with PM 1.00) and OR 0.74 (95%CI 0.59–0.92) respectively. O 2.5 3 in pairwise models by OR 1.18 (95% CI 1.05–1.25) per exposure was associated with increased use of allergy IQR PM (Table 4). medication with OR 1.29 (95% CI 1.02; 1.62) in dual- 2.5 Exposure to pollen at lag 0–2 was associated with exposure models. PM was associated with increased 2.5 increased use of allergy medication in single-expo- allergy medication usage in dual-exposure model by sure models by OR 1.58 (95%CI 1.44–1.74). In dual- OR 1.15 (95% CI 1.02–1.31) and in the multi-exposure and multi-pollutant models, the ORs were reduced model by OR 1.25 (95% CI 1.07; 1.46) (Table 4). but remained significant. For lag 0–6 pollen exposure Pollen exposure was statistically significantly associated there was a significant increase in allergy medication with increased use of bronchodilating medication only in reports in single-exposure models by OR 1.85 (95%CI single-exposure models by OR 1.15 (95% CI 1.08–1.21) at 1.60–2.14) and a similar reduction in the dual-and lag 0–2 and lag 0–6 by OR 1.23 (95% CI 1.16–1.32). There C arlsen et al. Environmental Health (2022) 21:63 Page 7 of 13 Table 3 Prevalence of daily diary‑reported symptoms, medication use, and mean (SD) of peak expiratory flow in the study population (n = 37) during the three study waves Wave 1 (1068 observations) Wave 2 (1066 observations) Wave 3 (978 observations) Total (3112 observations) Symptoms n (%) n (%) n (%) n (%) Eye irritation/rhinitis 649 (60.9%) 68 (6.4%) 368 (37.7%) 851 (27.5%) Dyspnea 252 (23.8%) 112 (10.6%) 215 (22.9%) 579 (19.0%) Dry cough 296 (27.9%) 239 (22.7%) 272 (28.9%) 807 (26.4%) Allergy medication 547 (51.3%) 118 (11.2%) 581 (61.9%) 1246 (40.7%) Bronchodilating medication No 389 (36.5%) 474 (44.9%) 355 (37.6%) 1218 (39.7%) Less 46 (4.3%) 64 (6.1%) 14 (1.5%) 124 (4.0%) Normal 533 (50.0%) 455 (43.1%) 454 (48.1%) 1442 (47.0%) More 98 (9.2%) 62 (5.9%) 121 (12.8%) 281 (9.2%) Peak expiratory flow (PEF) Mean (SD) Mean (SD) Mean (SD) Mean (SD) PEF (mL) 434 (81) 441 (90) 434 (93) 436 (88) mo PEF (mL) 442 (85) 448 (93) 442 (97) 444 (92) ev PEF  Peak expiratory flow measured in the morning,  PEF  Peak expiratory flow measured in the evening mo  ev  For eye irritation/rhinitis, dyspnea, dry cough, allergy medication, “Yes” vs. “No”. For bronchodilating medication, reporting “more than yesterday” vs. “No”, “Less”, or “Same” were no associations between use of bronchodilating with PM concentrations, by ORs 1.56 (95% CI 1.10– 2.5 medication and pollen in dual- or multi pollution mod- 2.21) and 1.72 (95% CI 1.13–2.64) per IQR increase in els, nor with NO, O and PM air pollution (Table 4). pollutant. The predicted ORs at low and high pollen lev - x 3 2.5 els are illustrated in Fig. 3a-d. Eec ff ts on PEF PM exposure was negatively associated with decreased In the pooled analysis, there were no statistically signifi - allergy medication in the low-pollen exposure scenario, cant associations between pollen at lag 0–2, lag 0–6 and and we speculate. ΔPEFmo at any adjustment level, but all estimates were Additionally, there were indications of pollen-air pol- negative. lutant interactions (with p-values below 0.1 but above ΔPEF and pollen at lag 0–2 were statistically signifi - 0.05) for dyspnea, bronchodilating medication and O ev 3 cantly associated in single-exposure models by -1.24 (95% (Table S3, Fig. S3). CI -2.45- -0.04) and in dual-exposure models adjusted for The autocorrelation adjustment performed for PEF did PM by -0.93 (95% CI -1.70-0.17) and by -0.93 (95% CI not change effect estimates of lag 0–2 exposure. 2.5 -1.71-0.15) mL per 100 pollen/m in the multi-exposure model. For ΔPEFev, there were no statistically significant Sensitivity analysis associations with air pollution (Table 5, city-specific esti - In the sensitivity analyses, in observations from individ- mates are shown in Table S2b). ual with poor asthma control (ACQ > 1.5), the estimated association between exposure and rhinitis or eye irrita- Interactions between pollen and air pollutants tion was higher than in the total data (OR 1.29 (95% CI We observed no interactions between pollen season and 1.12–1.48) vs. OR 1.22 (95% CI 1.14–1.31). For other air pollution on symptoms, medication usage or PEF symptoms and PEF, the estimates were similar or lower when applying a crude measure of pollen season (yes or than in total data (Table S4) and for allergy and asthma no), but when categorizing pollen at lag 0–2 into low and medication, the estimates were lower, although the con- high pollen concentrations, there were statistically sig- fidence intervals overlapped with the estimates from the nificant interactions with air pollution for symptoms and total data. In individuals who used inhaled corticosteroid medication usage (all interaction term coefficients and medication (ICS users, n = 25), lower ORs were seen for p-values are found in Table S3). There were interactions all symptoms compared to the total population, although for rhinitis or eye irritation between pollen concentra- reports of rhinitis or eye irritation were still statistically tions and O by OR 1.45 (95% CI 1.14–1.84) per IQR and significantly associated with pollen exposure (Table S4). with PM by 1.41 (95% CI 1.04–1.92) per IQR. Also, for In ICS users, there were no associations with ΔPEFev 2.5 dyspnea and allergy medication pollen levels interacted or ΔPEFmo. In individuals who used nasal steroids (NS Carlsen et al. Environmental Health (2022) 21:63 Page 8 of 13 Table 4 OR of diary‑reported symptoms and medication use associated with exposure to pollen at lag 0–2 and lag 0–6 and air pollution at lag 0–2 exposure to pollen and air pollutants in single‑, pairwise ‑, and multi‑ exposure models with 95% confidence interval (pooled results). Significant results are indicated with bold font Model adjustment Pollen – Lag 0–2 Pollen – Lag 0–6 NO – Lag 0–2 O – Lag 0–2 PM – Lag 0–2 x 3 2.5 OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) Rhinitis or eye irritation Adjusted for Pollen 1.44 (1.36; 1.52) 1.67 (1.57; 1.78) + NO 1.17 (1.02; 1.35) 1.28 (1.08; 1.53) 1.06 (0.86; 1.29) + O 1.14 (1.02; 1.26) 1.23 (1.12; 1.35) 0.81 (0.56; 1.17) + PM 1.18 (1.01; 1.37) 1.35 (1.06; 1.72) 0.88 (0.78; 1.01) 2.5 + NO, O, PM 1.25 (1.16; 1.35) 1.39 (1.27; 1.51) 0.71 (0.42; 1.19) 1.02 (0.34; 3.04) 1.16 (1.02; 1.32) x 3 2.5 Dyspnea Adjusted for Pollen 1.17 (1.10; 1.25) 1.23 (1.09; 1.37) + NO 1.04 (0.96; 1.14) 1.12 (1.02; 1.23) 0.75 (0.54; 1.05) + O 1.03 (0.95; 1.12) 1.09 (1.00; 1.20) 1.06 (0.69; 1.61) + PM 1.01 (0.93; 1.10) 1.08 (0.98; 1.19) 0.91 (0.55; 1.50) 2.5 + NO, O, PM 1.03 (0.95; 1.13) 1.09 (0.99, 1.20) 0.79 (0.54; 1.15) 0.96 (0.66; 1.40) 1.03 (0.81; 1.33) x 3 2.5 Dry cough Adjusted for Pollen 1.05 (0.90; 1.24) 1.08 (0.85; 1.35) + NO 1.10 (1.02; 1.18) 1.14 (1.05; 1.24) 0.97 (0.72; 1.31) + O 1.09 (1.01; 1.17) 1.14 (1.05; 1.24) 1.15 (1.00; 1.31) + PM 1.07 (0.99; 1.16) 1.14 (1.05; 1.24) 1.18 (1.05; 1.33) 2.5 + NO, O, PM 1.09 (1.01; 1.17) 1.15 (1.05; 1.28) 0.99 (0.72; 1.35) 1.14 (1.09; 1.19) 1.01 (0.97; 1.06) x 3 2.5 Allergy medication Adjusted for Pollen 1.58 (1.44; 1.74) 1.85 (1.60; 2.14) + NO 1.23 (1.12 1.35) 1.16 (1.08; 1.24) 0.72 (0.52; 1.00) + O 1.19 (1.08; 1.30) 1.15 (1.07; 1.23) 1.29 (1.02; 1.62) + PM 1.15 (1.01; 1.26) 1.16 (1.07; 1.23) 1.15 (1.02; 1.31) 2.5 + NO, O, PM 1.17 (1.07; 1.29) 1.15 (1.06; 1.25) 0.74 (0.59; 0.92) 1.01 (0.78; 1.32) 1.25 (1.07; 1.46) x 3 2.5 Bronchodilating medication Adjusted for Pollen 1.15 (1.08; 1.21) 1.23 (1.16; 1.32) + NO 1.07 (0.87; 1.32) 1.10 (0.85; 1.42) 0.84 (0.81; 0.88) + O 1.06 (0.87; 1.30) 1.09 (0.83; 1.43) 1.02 (0.77; 1.34) + PM 1.07 (0.97; 1.30) 1.10 (0.82; 1.49) 0.97 (0.97; 0.99) 2.5 + NO, O, PM 1.07 (0.86; 1.33) 1.12 (0.84; 1.50) 0.81 (0.62; 1.07) 1.04 (0.72; 1.49) 1.09 (0.86; 1.37) x 3 2.5 For eye irritation/rhinitis, dyspnea, dry cough, allergy medication, “Yes” vs. “No”. For bronchodilating medication, reporting “more than yesterday” vs. “No”, “Less”, or “Same” Results from pooled with meta-analysis of mixed model results for pollen (single-exposure), pollen and pairwise combinations of pollen and either NO , O and PM , x 3 2.5 and multi-exposure models adjusted for pollen. All models were adjusted for relative humidity and temperature, with identification number as a random effect. 3 3 3 3 Associations for are reported per 100 grains/m for pollen and per city-specific IQR for pollutants (Göteborg: NO 20.0 µg/ m, O 22.9 µg/ m, PM 3.3 µg/ m , Umeå: x 3 2.5 3 3 3 NO 14.7 µg/m, O 24.1 µg/ m, PM 2.8 µg m ) x 3 2.5 values of ΔPEF were reduced in both males and females, users, n = 9), the ORs for rhinitis or eye irritation and but the reduction in ΔPEF did not reach statistical sig- allergy medication in association with pollen counts were mo nificance in males (Fig. S4). higher than in any other group. The results obtained from individuals who participated at all three waves were similar to the total data (Table S4). Stratifying by sex, pol- Discussion len exposure was only significantly associated with dysp - In this asthma-diary panel study with data from two nea, asthma medication usage in female subjects, but the different pollen-seasons, we found that symptom prev - only significant interaction between exposure and sex alence was increased during birch pollen season, and was observed for bronchodilating medication, where the that pollen exposure at lag 0–2 and lag 0–6 was asso- effects were stronger in females. Morning and evening ciated strongly with prevalence of most symptoms of C arlsen et al. Environmental Health (2022) 21:63 Page 9 of 13 Table 5 Associations (B) between peak expiratory flow (PEF) in the morning and evening (PEF and PEF ) and exposure to mo ev pollen (per 100 grains) and air pollutants in single‑, pairwise ‑, and multi‑ exposure models with 95% confidence interval (pooled results). Significant results are indicated with bold font Pollen lag 0–2 Pollen lag 0–6 NO O PM2.5 x 3 β (95% CI) 95% CI 95% CI 95% CI 95% CI ΔPEF (mL) mo Adjusted for Pollen ‑1.10 (‑2.86; 0.66) ‑1.00 (‑3.00; 1.01) ‑ ‑ ‑ + NO ‑1.06 (‑3.04; 0.92) ‑0.97 (‑3.20; 1.27) ‑0.16 (‑3.21; 2.88) ‑ ‑ + O ‑1.07 (‑2.90; 0.76) ‑0.97 (‑3.05; 1.11) ‑ 1.54 (‑0.92; 4.00) ‑ + PM ‑1.10 (‑2.67; 0.47) ‑1.20 (‑3.09; 0.70) ‑ ‑ 2.16 (0.23; 4.09) 2.5 + NO, O, PM -1.05 (-2.78;-0.68) ‑1.17 (‑3.19; 0.86) ‑0.43 (‑2.48; 1.63) 0.18 (‑1.70; 2.06) 2.14 (1.01; 3.26) x 3 2.5 ΔPEFev (mL) Adjusted for Pollen -1.24 (-2.45; -0.04) ‑1.24 (‑3.88; 1.40) ‑ ‑ ‑ + NO ‑1.25 (‑2.53; 0.03) ‑1.26 (‑4.01; 1.48) 0.40 (‑0.97; 1.77) ‑ ‑ + O ‑1.22 (‑2.48; 0.04) ‑1.24 (‑3.88; 1.40) ‑ 0.66 (‑1.24; 2.55) ‑ + PM -0.93 (-1.70; -0.17) ‑0.72 (‑2.30; 0.86) ‑ ‑ 2.88 (‑1.80; 7.56) 2.5 + NO, O, PM -0.93 (-1.71; -0.15) ‑0.73 (‑2.25; 0.80) 0.17 (‑1.65; 1.31) ‑0.27 (‑2.35; 1.81) 3.21 (‑1.27; 7.68) x 3 2.5 ΔPEF: deviation from each individual’s mean of each season. Exposure at lag 1–2 for PEF and lag 0–2 for PEF mo ev Results from pooled with meta-analysis of mixed model results for pollen (single-exposure), pollen and pairwise combinations of pollen and either NO , O and PM , x 3 2.5 and multi-exposure models adjusted for pollen. All models were adjusted for relative humidity and temperature, with identification number as a random effect. 3 3 3 3 Associations are reported per 100 grains/m for pollen and per city-specific IQR for pollutants (Göteborg: NO 20.0 µg/ m, O 22.9 µg/ m, PM 3.3 µg/ m , Umeå: NO x 3 2.5 x 3 3 3 14.7 µg/m, O 24.1 µg/ m, PM 2.8 µg m ) 3 2.5 Fig. 3 Predicted marginal effects (proportions) of significant interactions between birch pollen levels (below or above 100 grains /m ) and pollutants (per µg/m3) on symptoms and allergy medication usage. Footnote: For eye irritation/rhinitis, dyspnea and allergy medication, “Yes” vs. th th “No”. X‑axis from 5 to 95 percentile. Results are from mixed models with identification number and city as a random effect. In addition to pairwise combinations of air pollution and pollen concentration indicator at lag 0–2, the models are adjusted for temperature and relative humidity. Results 3 3 3 are reported per global pollutant IQR (NO 16.4 µg/m, O 15.2 µg/m, PM 4.7 µg/m ) x 3 2.5 Carlsen et al. Environmental Health (2022) 21:63 Page 10 of 13 allergic asthma as well as medication usage (Table  3). allergy and short-term air pollution exposure during pol- Regarding air pollutants, O was associated with dry len season were indicated in a panel studies of individuals cough and increased use of allergy medication in dual- with allergy, but the results did not reach statistical sig- exposure models. PM2.5 was associated with rhinitis nificance [36]. After exposure in a road tunnel, asthmatic or eye irritation and with allergy medication in multi- subjects had increased response to an allergen provoca- exposure models and with dry cough in dual-exposure tion test with significant correlations between increased models. Evening-measured values of PEF were associ- asthma symptoms and NO , but not PM exposure [37]. 2 2.5 ated with pollen, and PM was associated with morn- In a controlled study of exposure to diesel exhaust and 2.5 ing-measured PEF in multi-exposure models (Table  5). allergen, the effect of exposure to diesel exhaust was aug - An increase in NOx was in most analyses unexpect- mented with simultaneous allergen exposure [20]. edly associated with lower risks. This phenomenon can In our study, we observed only moderate sex-differ - sometimes be explained by the negative correlation ences in the relationship between pollen and symptoms, between NO and O , and the reduction in risk dis- where pollen exposure was associated with increased risk x 3 appears after adjusting for O .In this case the adjust- of asthma medication use in females, but not in males ment did not remove the tendency, and one explanation (Fig. S4). A similar observation was made in a study of could be that variations in NO levels are more local asthma symptoms and personal sampler-measured expo- around the central monitoring station than the fluctua - sure [38] where asthma symptoms during daytime was tions in ozone exposure. associated with exposure to oxidants (NO and O ) in 2 3 Our study was designed to investigate if pollen expo- women, but not in men, suggesting that females may be sure increased susceptibility to air pollution, however, more susceptible to effects of air pollution than males. we acknowledge that the causal mechanism could be For PEF, a statistically significant increase of 2.16 (95% opposite; that air pollution increased the susceptibility CI 0.23–4.09) mL per IQR (3  µg/m) PM in morning 2.5 to pollen. However, with the current study design, we are PEF was observed in our study, where a recent review unable to test this, as we have no scenario where there is found that in asthmatic individuals, PEF was reduced no air pollution. 0.56  L/m per 10  µg/m PM [39] and a recent review 2.5 We investigated effects at different lags with DLNM found reduction of 2% per 10 pollen/m , but the time methodology and found that lag 0–2 or lag 0–6 was rel- of day (morning or evening) was not taken into account evant in most cases, but that in some cases, for example [33]. The unexpected direction of association for PM 2.5 for PEF , there could be an additional effect at lag 4–5 and PEF could be due to coinciding pollen and PM mo 2.5 which could be investigated in future studies (Fig. S2). peaks [19] where pollen exposure compels individuals The associations between pollen exposure and lower, with asthma to use more bronchodilating medication upper respiratory symptoms, and rhinitis or eye irrita- which reduces airway restriction (Table S4). tion symptoms in our study are consistent with those In our study, statistically significant pollen – air pollu - from a recent review and meta-analysis of pollen expo- tion interactions (interaction p < 0.05), where air pollu- sure on symptoms in individuals with allergy or asthma. tion effects where significantly higher during moderate The authors report that 10 pollen/m3 increase in expo - to high levels of pollen compared to low or no pollen, sure was associated with increased lower respiratory were present for rhinitis or eye irritation and PM and 2.5 symptoms by 1%, upper respiratory or ocular symptoms O , asthma symptoms and NO . In a recent study of 3 x by 6.6%, and any symptom of allergy or asthma increased self-reported symptoms via a mobile phone app, Bédard 2%. PEF was not associated with exposure [33]. In a study and colleagues (2020) observed significant associations of mobile-phone app reported symptoms, birch pollen between rhinitis symptoms and O during grass pollen exposure was associated with 3% increase in respiratory season, but not birch pollen season. PM was statisti- 2.5 symptoms and 4% increase in eye/nose symptoms per 10 cally significantly associated with rhinitis during the pol - pollen/m3 [34]. len season of one year, but not the other year in the study In addition to previous reports of associations between [40]. In another study of mobile phone app-recorded short term air pollution exposure and severe outcomes symptoms, O concentration led to an increased symp- such as increased emergency room visits and hospi- tom severity during the birch pollen season, but the asso- tal admissions for asthma [21], or symptom severity ciation was not significant after adjusting for climate [41]. recorded during MD visits [35], previous studies report In a panel study with 15 individuals with allergy, there increased asthma symptoms in association with air pol- was an increase in symptom severity associated with lution: PM exposure was associated with increase in exposure to air pollution during pollen season, but the 2.5 mobile-phone app reported respiratory symptoms by 6% estimates did not reach statistical significance [36]. per 10 µg/m [34]. An association between symptoms of C arlsen et al. Environmental Health (2022) 21:63 Page 11 of 13 In a recent review of pollen or fungal spore-air pollu- levels obtained at fixed monitoring stations does not tion interactions, it was concluded that although interac- accurately reflect each subject’s personal exposure tions had been shown or indicated in time series studies, level, but are rather used as a proxy for exposure in most of the existing did not consider groups at risk, and this study. However, this study has a time series ele- there were no studies of adults with allergic asthma [22] ment, and so, the day-to-day fluctuations in exposure, making our study unique in that context. which can be expected to be similar across a city, are The generalizability of our results to the previous lit - similar at the monitoring station and the study partici- erature on pollen-air pollution interactions in asthma pants locations. Furthermore, studies have shown cen- is limited as much of it pertains to more severe out- tral monitoring stations to be a reasonable proxy for comes such as asthma emergency room visits or hos- personal exposure to birch pollen [42]. Also, monitor- pital admissions. Also, the few available panel studies ing station data correlated well with personal exposure apply heterogenous methodologies [21, 22]. Some pre- in the study [27]. The panel data collected both in, and vious research on pollen-air pollution interaction have outside, pollen season ensured a substantial exposure compared associations with air pollution during and contrast. The start of the pollen season measured in outside pollen season [40, 41], which did not yield sig- the study were determined by the first 5-day period nificant results in our study (data not shown), where we with 4 days with pollen counts over 0 [26]. Newer defi- entered short-term pollen categories in the interaction nitions are available which could improve the exposure models. We speculate that the lag 0–2 pollen exposure assessment [28]. However, using this definition may is a more precise measure, and that exposure during have caused us to choose a slightly premature date for the whole pollen season was too variable (especially the pollen season in 2015, which was also very mild during wave 1 in Umeå, see Fig. 1) to obtain significant (Fig.  1; Table  1). We speculate that pollen counts dur- results. ing that season were too low to have an effect, but no In our study, the main results are presented as pooled universally accepted thresholds of effect have been results (with random effect meta-analyses) from sepa - reported [33]. rate analyses of each study center, as we knew of dif- We observed negative, or protective associations ferences in air pollution levels, temperature, as well as between NO exposure and symptoms. However, the recruitment procedures in the two study locations. In often-negative correlation between NO and O could x 3 the interaction analysis, we entered city as a random skew the analysis. However, we calculated a combina- effect in the model. For PEF, we tested for autocorre - tion value for oxides, O , and used that in the analy- lation by entering an autoregressive term, but for the sis, which did not change the results for pollen. The reported lag 0–2 results, this procedure did not alter the coefficients associated with O was statistically sig- results. There is no definition of this term for logistic nificant for bronchodilating medication and morning regression, so this could not be tested in the analysis of PEF and in both cases indicated a positive or protec- symptom scores. tive effect of O . Our study population was well-defined at base- line, with allergy confirmed with skin-prick tests, but Conclusions asthma was not confirmed with formal testing as in Our results show that a substantial proportion of allergic some other studies [36]. There are indications that asthma symptoms can be attributed to pollen exposure. some of the cases in our study population had mild Also, in addition to direct effects of pollen on symptoms disease, e.g., not using any asthma medication which in allergic asthma, pollen exposure increases susceptibil- could reduce the strength of our results (Table 1). The ity to adverse respiratory effects of exposure to PM and 2.5 results from the sensitivity analysis which was strati- O . fied by medication status underlines that our sam- Implementing the results into advisories could improve ple had sizable heterogeneity with respect to disease their predictive power, which could minimize the mor- severity. This fact has some consequences for the bidities associated with allergic asthma and allergy. Con- observed effect size. The sample size deduced based sidering the number of affected individuals, this could be on power calculation of respiratory biomarkers, but as of substantial benefit to public health. Also, reduction of the power in this analysis was increased by the many pollution levels should be priority to improve health in measures of each individual, the study is unlikely to be this susceptible population. underpowered. In our study, exposure to both pollen and air pollu- Abbreviations tion was measured at central urban background sta- ACQ: Asthma Control Questionnaire; NO : Nitrogen dioxides; O : Ozone; PEF: x 3 tions, which induces some uncertainty as exposure Peak Expiratory Flow; PM: Particle matter. Carlsen et al. Environmental Health (2022) 21:63 Page 12 of 13 3. Borna E, Nwaru BI, Bjerg A, Mincheva R, Rådinger M, Lundbäck B, et al. Supplementary Information Changes in the prevalence of asthma and respiratory symptoms in west‑ The online version contains supplementary material available at https:// doi. ern Sweden between 2008 and 2016. Allergy. 2019;74(9):1703–15. org/ 10. 1186/ s12940‑ 022‑ 00871‑x. 4. Backman H, Räisänen P, Hedman L, Stridsman C, Andersson M, Lindberg A, et al. Increased prevalence of allergic asthma from 1996 to 2006 and Additional file 1. further to 2016—results from three population surveys. Clin Experimental Allergy. 2017;47(11):1426–35. 5. Burte E, Leynaert B, Marcon A, Bousquet J, Benmerad M, Bono R, et al. Acknowledgements Long‑term air pollution exposure is associated with increased severity of The authors would like to acknowledge Marianne Andersson, Helene Friberg, rhinitis in 2 European cohorts. J allergy Clin Immunol. 2020;145(3):834–42. Annika Claesson (GU) and Helen Bertilsson and Chatrin Wahlgren (Umeå U), e6. who gathered the patient data and measurements. The authors would also 6. Guarnieri M, Balmes JR. Outdoor air pollution and asthma. Lancet. like to thank the participants. The authors acknowledge code from Hedi Katre 2014;383(9928):1581–92. Kriit and Johan Nilsson Sommar used in the DLNM analysis. 7. D’Amato M, Cecchi L, Annesi‑Maesano I, D’Amato G. News on Climate Change, Air Pollution, and Allergic Triggers of Asthma. J Investig Allergol Authors’ contributions Clin Immunol. 2018;28(2):91–7. HKC managed the data, analyzed, wrote the original draft, reviewed and 8. Jacquemin B, Kauffmann F, Pin I, Le Moual N, Bousquet J, Gormand F, et al. edited. SLH Performed the investigation, curated the data, write and edited Air pollution and asthma control in the Epidemiological study on the the manuscript, DO contributed to the methodology and analysis, AFB Genetics and Environment of Asthma. J Epidemiol Community Health. contributed to the discussion, writing and editing of the manuscript. LM 2012;66(9):796–802. contributed to the methodology, writing and review of the manuscript. KM 9. Biedermann T, Winther L, Till SJ, Panzner P, Knulst A, Valovirta E. Birch pol‑ managed and curated data. BF conceptualized, acquired funding, wrote and len allergy in Europe. Allergy. 2019;74(7):1237–48. reviewed the manuscript. A‑ C conceptualized the study, acquired funding, 10. Gilles S, Blume C, Wimmer M, Damialis A, Meulenbroek L, Gökkaya M, investigated, wrote and reviewed the study. The author(s) read and approved et al. Pollen exposure weakens innate defense against respiratory viruses. the final manuscript. Allergy. 2020;75(3):576–87. 11. Cockcroft DW, Davis BE. Mechanisms of airway hyperresponsiveness. J Funding Allergy Clin Immunol. 2006;118(3):551–9. Open access funding provided by University of Gothenburg. The study was 12. Beck I, Jochner S, Gilles S, McIntyre M, Buters JT, Schmidt‑ Weber C, et al. supported by FORMAS (A Swedish Research Council for sustainable develop‑ High environmental ozone levels lead to enhanced allergenicity of birch ment, dnr: 210‑2013‑805), the Swedish Heart and Lung Foundation (dnr: pollen. PLoS ONE. 2013;8(11):e80147. 2013 − 0279 and 2016 − 0250), the Swedish Cancer and Allergy Foundation 13. Baldacci S, Maio S, Cerrai S, Sarno G, Baïz N, Simoni M, et al. Allergy and and the Asthma and Allergy Association. asthma: Eec ff ts of the exposure to particulate matter and biological allergens. Respir Med. 2015;109(9):1089–104. Availability of data and materials 14. D’Amato G, Liccardi G, D’Amato M, Cazzola M. Outdoor air pollu‑ The data are available for other researchers upon reasonable request and tion, climatic changes and allergic bronchial asthma. Eur Respir J. pending ethics approval. 2002;20(3):763–76. 15. Reinmuth‑Selzle K, Kampf CJ, Lucas K, Lang‑ Yona N, Fröhlich‑Nowoisky J, Shiraiwa M, et al. Air pollution and climate change effects on allergies in Declarations the anthropocene: abundance, interaction, and modification of allergens and adjuvants. Environ Sci Technol. 2017;51(8):4119–41. Ethics approval and consent to participate 16. Zhu C, Farah J, Choel M, Gosselin S, Baroudi M, Petitprez D, et al. Uptake of The study was approved by the Regional Ethical Review Board at the Uni‑ ozone and modification of lipids in Betula Pendula pollen. Environ Pollut. versity of Gothenburg (Dnr: 682 − 14), and the participants signed informed 2018;242:880–6. consent forms for analysis and publication of the data. 17. Hong Q, Zhou S, Zhao H, Peng J, Li Y, Shang Y, et al. Allergenicity of recombinant Humulus japonicus pollen allergen 1 after combined expo‑ Competing interests sure to ozone and nitrogen dioxide. Environ Pollut. 2018;234:707–15. The authors declare no competing interests. 18. Rojo J, Oteros J, Picornell A, Maya‑Manzano JM, Damialis A, Zink K, et al. Eec ff ts of future climate change on birch abundance and their pollen Author details load. Glob Change Biol. 2021;27(22):5934–49. Section of Occupational and Environmental Medicine, School of Public 19. Ørby PV, Peel RG, Skjøth C, Schlünssen V, Bønløkke J, Ellermann T, et al. An Health and Community Medicine, Institute of Medicine, Sahlgrenska Academy, assessment of the potential for co‑ exposure to allergenic pollen and air University of Gothenburg, Medicinaregatan 16A, 40530 Gothenburg, Sweden. pollution in Copenhagen, Denmark. Urban Clim. 2015;14:457–74. Department of Public Health and Clinical Medicine, University Hospital, 20. Carlsten C, Blomberg A, Pui M, Sandstrom T, Wong SW, Alexis N, et al. Sustainable Health, Umeå University, Building 1A, 4st, 901 87 Umeå, Sweden. Diesel exhaust augments allergen‑induced lower airway inflamma‑ Section of Medicine, Department of Public Health and Clinical Medicine, tion in allergic individuals: a controlled human exposure study. Thorax. University Hospital, Umeå University, Building 1A, 4st, 901 87 Umeå, Sweden. 2016;71(1):35–44. Department of Statistics, USBE, Social Sciences Building Level 2 (ground 21. Anenberg SC, Haines S, Wang E, Nassikas N, Kinney PL. Synergistic health floor), Umeå University, 90187 Umeå, Sweden. effects of air pollution, temperature, and pollen exposure: a systematic review of epidemiological evidence. Environ Health. 2020;19(1):1–19. Received: 26 January 2022 Accepted: 9 June 2022 22. Lam HCY, Jarvis D, Fuertes E. Interactive effects of allergens and air pollution on respiratory health: A systematic review. Sci Total Environ. 2021;757:143924. 23. Forsberg B, Stjernberg N, Falk M, Lundback B, Wall S. Air pollution levels, meteorological conditions and asthma symptoms. Eur Respir J. References 1993;6(8):1109–15. 1. Landrigan PJ, Fuller R, Acosta NJ, Adeyi O, Arnold R, Baldé AB, 24. Forsberg B, Stjernberg N, Linne R, Segerstedt B, Wall S. Daily air pol‑ et al. The Lancet Commission on pollution and health. Lancet. lution levels and acute asthma in southern Sweden. Eur Respir J. 2018;391(10119):462–512. 1998;12(4):900–5. 2. To T, Stanojevic S, Moores G, Gershon AS, Bateman ED, Cruz AA, et al. 25. Juniper E, O’byrne P, Guyatt G, Ferrie P, King D. Development and Global asthma prevalence in adults: findings from the cross‑sectional validation of a questionnaire to measure asthma control. Eur Respir J. world health survey. BMC Public Health. 2012;12(1):204. 1999;14(4):902–7. C arlsen et al. Environmental Health (2022) 21:63 Page 13 of 13 26. Dahl Å, Galán C, Hajkova L, Pauling A, Sikoparija B, Smith M, et al. The onset, course and intensity of the pollen season. Allergenic pollen: Springer; 2013. pp. 29–70. 27. Haga SL, Hagenbjörk A, Olin A‑ C, Forsberg B, Liljelind I, Carlsen HK, et al. Personal exposure levels to O3, NOx and PM10 and the associa‑ tion to ambient levels in two Swedish cities. Environ Monit Assess. 2021;193(10):674. 28. Pfaar O, Bastl K, Berger U, Buters J, Calderon M, Clot B, et al. Defining pollen exposure times for clinical trials of allergen immunotherapy for pollen‑induced rhinoconjunctivitis–an EAACI position paper. Allergy. 2017;72(5):713–22. 29. Gasparrini A, Armstrong B, Kenward MG. Distributed lag non‑linear mod‑ els. Stat Med. 2010;29(21):2224–34. 30. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed‑ effects models using lme4. arXiv preprint arXiv:14065823. 2014. 31. Balduzzi S, Rücker G, Schwarzer G. How to perform a meta‑analysis with R: a practical tutorial. Evid Based Ment Health. 2019;22(4):153–60. 32. Lüdecke D. ggeffects: Tidy Data Frames of Marignal Eec ff ts from Regres‑ sion Models. J Open Source Softw. 2018;3(26):772. 33. Kitinoja MA, Hugg TT, Siddika N, Yanez DR, Jaakkola MS, Jaakkola JJ. Short‑term exposure to pollen and the risk of allergic and asthmatic manifestations: a systematic review and meta‑analysis. BMJ Open. 2020;10(1):e029069. 34. Jones PJ, Koolhof IS, Wheeler AJ, Williamson GJ, Lucani C, Campbell SL, et al. Can smartphone data identify the local environmental drivers of respiratory disease? Environ Res. 2020;182:109118. 35. Annesi‑Maesano I, Rouve S, Desqueyroux H, Jankovski R, Klossek J‑M, Thibaudon M, et al. Grass pollen counts, air pollution levels and allergic rhinitis severity. Int Arch Allergy Immunol. 2012;158(4):397–404. 36. Riediker M, Monn C, Koller T, Stahel WA, Wüthrich B. Air pollutants enhance rhinoconjunctivitis symptoms in pollen‑allergic individuals. Ann Allergy Asthma Immunol. 2001;87(4):311–8. 37. Svartengren M, Strand V, Bylin G, Jarup L, Pershagen G. Short‑term expo ‑ sure to air pollution in a road tunnel enhances the asthmatic response to allergen. Eur Respir J. 2000;15(4):716–24. 38. Chambers L, Finch J, Edwards K, Jeanjean A, Leigh R, Gonem S. Eec ff ts of personal air pollution exposure on asthma symptoms, lung function and airway inflammation. Clin Experimental Allergy. 2018;48(7):798–805. 39. Edginton S, O’Sullivan DE, King WD, Lougheed MD. The effect of acute outdoor air pollution on peak expiratory flow in individuals with asthma: A systematic review and meta‑analysis. Environ Res. 2021;192:110296. 40. Bédard A, Sofiev M, Arnavielhe S, Antó JM, Garcia‑Aymerich J, Thibaudon M, et al. Interactions Between Air Pollution and Pollen Season for Rhinitis Using Mobile Technology: A MASK‑POLLAR Study. J Allergy Clin Immunol Pract. 2020;8(3):1063‑73.e4. 41. Berger M, Bastl K, Bastl M, Dirr L, Hutter H‑P, Moshammer H, et al. Impact of air pollution on symptom severity during the birch, grass and ragweed pollen period in Vienna, Austria: Importance of O3 in 2010–2018. Environ Pollut. 2020;263:114526. 42. Bastl M, Bastl K, Karatzas K, Aleksic M, Zetter R, Berger U. The evaluation of pollen concentrations with statistical and computational methods on rooftop and on ground level in Vienna–How to include daily crowd‑ sourced symptom data. World Allergy Organization J. 2019;12(5):100036. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : lished maps and institutional affiliations. fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Environmental Health Springer Journals

Birch pollen, air pollution and their interactive effects on airway symptoms and peak expiratory flow in allergic asthma during pollen season – a panel study in Northern and Southern Sweden

Download PDF
13 pages

Loading next page...
 
/lp/springer-journals/birch-pollen-air-pollution-and-their-interactive-effects-on-airway-aRDThYbbNk

References (46)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2022
eISSN
1476-069X
DOI
10.1186/s12940-022-00871-x
Publisher site
See Article on Publisher Site

Abstract

Background: Evidence of the role of interactions between air pollution and pollen exposure in subjects with allergic asthma is limited and need further exploration to promote adequate preventive measures. The objective of this study was to assess effects of exposure to ambient air pollution and birch pollen on exacerbation of respiratory symptoms in subjects with asthma and allergy to birch. Methods: Thirty‑seven subjects from two Swedish cities (Gothenburg and Umeå) with large variation in exposure to both birch‑pollen and air pollutants, participated in the study. All subjects had confirmed allergy to birch and self‑reported physician‑ diagnosed asthma. The subjects recorded respiratory symptoms such as rhinitis or eye irrita‑ tion, dry cough, dyspnoea, the use of any asthma or allergy medication and peak respiratory flow (PEF), daily for five consecutive weeks during two separate pollen seasons and a control season without pollen. Nitrogen oxides (NO ), ozone (O ), particulate matter (PM ), birch pollen counts, and meteorological data were obtained from an urban 3 2.5 background monitoring stations in the study city centres. The data were analysed using linear mixed effects models. Results: During pollen seasons all symptoms and medication use were higher, and PEF was reduced in the subjects. In regression analysis, exposure to pollen at lags 0 to 2 days, and lags 0 to 6 days was associated with increased ORs of symptoms and decreased RRs for PEF. Pollen and air pollution interacted in some cases; during low pollen exposure, there were no associations between air pollution and symptoms, but during high pollen exposure, O concentra‑ tions were associated with increased OR of rhinitis or eye irritation, and PM concentrations were associated with 2.5 increased ORs of rhinitis or eye irritation, dyspnea and increased use of allergy medication. *Correspondence: hanne.krage.carlsen@amm.gu.se Section of Occupational and Environmental Medicine, School of Public Health and Community Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Medicinaregatan 16A, 40530 Gothenburg, Sweden Full list of author information is available at the end of the article © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Carlsen et al. Environmental Health (2022) 21:63 Page 2 of 13 Conclusions: Pollen and air pollutants interacted to increase the effect of air pollution on respiratory symptoms in allergic asthma. Implementing the results from this study, advisories for individuals with allergic asthma could be improved, minimizing the morbidities associated with the condition. Keywords: Birch, Betula, PM , O , Panel study, Allergic asthma, Pollen season 2.5 3 Background severe disease, as indicated by either steroid medica- Exposure to air pollution can cause a variety of adverse tion use, or have poorer asthma control. health effects such as respiratory illness, cardiovascu - lar disease and lung cancer, and it has been estimated Methods that approximately 9 million people die each year from Study population causes directly attributed to air pollution [1]. The prev - Forty-one non-smoking subjects (27 to 73 years of age) alence of asthma is reported to be 4.3% globally, and with self-reported physician diagnosed asthma and a as high as 10% in Sweden [2]. The prevalence has been confirmed birch allergy (positive skin-prick test, phadi - increasing in the last decade [3] and allergic asthma is atop ≥ 0.35 kU/L) were recruited to participate. 22 sub- the most commonly reported form of asthma in Swe- jects were recruited in Gothenburg (Lat. 57° N) in the den with an estimated prevalence of 7.3% [4]. southern part of Sweden and 17 from Umeå (Lat. 63° Exposure to air pollutants such as ozone, particulate N), in northern Sweden. Eventually, 22 and 15 had diary matter (PM), and nitric oxides (N O ) is associated with data and were included in the analysis. Subjects with increased rhinitis severity [5] and short term expo- cardiovascular disease or any chronic inflammatory dis - sure can induce airway inflammation, due to increased ease were excluded from participation. All participating oxidative stress, and further result in airway hyperre- subjects signed an informed consent agreement, and the sponsiveness in asthmatic individuals and sensitization study was approved by the Regional Ethical Review Board [6, 7] and poor asthma control [8]. Birch (betula) pollen at the University of Gothenburg (Dnr: 682 − 14). is a common allergen in allergic asthma [9]. Inhalation of pollen grains (who are themselves too large to reach Study design the small airways) results in the rupture of the grain The participating subjects filled out a health diary which and release of the allergenic content (Bet v1 protein) included questions about daily symptoms and medica- which activates an immune response, [10] and cause tion who had been used in previous studies [23, 24] for bronchoconstriction in individuals with allergic asthma five consecutive weeks (35 days) during two separate pol - [11]. Increased allergenicity of birch pollen following len-seasons (2015 and 2016) and once outside the pollen- exposure to ozone [12] and PM [13] has been reported season (November-December 2015). and it has also been speculated that pollution attaches The symptoms registered in the diary (Supplementary to the allergen [14]. Anthropogenic air pollution affects document D1) were cold, fever, rhinitis, dyspnea and dry the abundance and bioavailability of pollen, and future cough (yes or no), and allergy medication (yes or no), climate change may affect the duration and severity of bronchodilation medicines (compared to their regular pollen seasons [15–18]. Pollen and air pollution peaks dose)” (no, less, normal or more), and asthma symptoms often co-occur [19], therefore it is of interest to study (no, less, normal or more). Peak respiratory flow rate pollen and air pollution together. Although synergis- (PEF) measurements were made twice daily throughout tic effects or interactions between air pollution and the five weeks, in the morning before any medication and pollen on respiratory health outcomes in individuals before bedtime, with three registered measurements for with allergic asthma have been reported [14, 20], many each occasion. uncertainties remain, the estimated risks have a large The means for morning and evening PEF for each sub - range, and more studies on this topic are warranted [21, ject during the study waves  were first calculated (PEF mo 22].and PEF ). Individual deviations of daily performance ev The aim of the present study was to study associa - from each subject’s mean PE F and PE F per season mo ev tions and interactions between pollen and air pollut- were then calculated and averaged across the participants ants associated with worsening of daily self-reported to obtain daily mean deviations (∆) for PEF and PEF mo ev respiratory symptoms and peak expiratory flow (PEF). values, ∆PEF and ∆PEF . mo ev Furthermore, we wanted to assess if there are lagged To avoid error or bias due to effects of respiratory dis - effects of exposure to pollen and air pollution on symp - ease epidemics, we excluded days where participants toms reports, study effects in individuals with more reported having a “fever”. In total, 8 individuals reported C arlsen et al. Environmental Health (2022) 21:63 Page 3 of 13 “fever” on at least one occasion (46 observations with and pollen seasons. The proportion reporting symptoms, “fever”, 3133 observations with “no fever” reported). The medication use, and mean PEF was reported for each presence of cold was reported on at least one day by 22 of season. the participants, with a total of 297 observations of cold, The outcomes were analyzed with mixed models, 2855 observations with no cold), but as the symptoms of where subjects were included as a random effect, and cli - cold are somewhat similar to those of asthma and allergy, mate and air pollution variables were included as fixed they were retained in the data, but not analysed. effects. All subjects underwent a clinical examination once The mean of lag-intervals lag 0–2, and secondarily lag within each season and completed an Asthma Con- 0–6 were chosen based on previous literature. However, trol Questionnaire (ACQ) with six questions regarding the lag-association between outcomes and pollen and air symptoms and use of bronchodilators and inhaled ster- pollution exposure was also investigated using distrib- oids during the week before the clinical examination uted lag non-linear models (DLNM) [29] at lags 0 to lag and pre-bronchodilator forced expiratory volume in one 10 with different options for the shape of the lag-associ - second (FEV ) in categories [25]. The participants also ations and levels of adjustments. The models with best reported their medication usage in the previous month in (lowest) AIC were selected and plotted (Supplementary a questionnaire. Fig. S2) and showed that in most cases, the significant association between exposure and outcome did, in fact, Exposure occur at lag 0–2. Birch pollen counts were obtained from centrally located In the analysis of morning PEF values, same day pollen pollen monitoring stations using standard methods [26]. exposure was not included as they were deemed of less Ambient concentrations of the air pollutants NO, O interest, only exposure lags from the previous day and x 3 and PM , and meteorological data of relative humid- backwards in time were included (lag1, lag 1–2 and lag 2.5 ity and temperature were measured at centrally located 1–5). fixed monitoring stations in the two study centers and Logistic regression was used for analysis of the daily obtained from the environmental authorities as previ- symptoms in asthma-diary and linear regression for PEF. ously described [27]. In this study, we used P M as In the main analysis, data from all study seasons were 2.5 this was available in both locations for the study period. analyzed separately for each study center (reported in the In brief, routinely collected air pollution, pollen, and Supplement), then the results were pooled using random meteorology data were provided to us from the relevant effects meta-analysis. authorities (Gothenburg: Gothenburg municipality Single-exposure models included the covariates pol- h t t p s :// got eb org . s e/ w p s/ p or t al/ st ar t/ milj o/ milj o l age t -i- len, temperature and relative humidity, dual-exposure goteb org/ luft/ luftk valit eten-i- goteb org, and the Pollen models included pairwise combinations of pollen and air Laboratory at Gothenburg University https:// www. gu. pollutants O, NO , and PM , temperature, and rela- 3 x 2.5 se/ biolo gi- miljo veten skap/ pollen- och- aller gier), Umeå tive humidity. Multi-exposure models included pollen, all Municipality (https:// www. umea. se/ bygga booch miljo/ three pollutants, temperature, and relative humidity. b o end emilj obull er o c h luf t k v alit e t/ luf t e n ut om h u s/ luf t k Interaction analysis for pollen and air pollution were valit eteni umea.4. 250f9 65917 4ae4b 97941 ae7. html). Rel- first investigated in mixed models with id and study evant lags were calculated and used in the analyses. All center as random effects with an indicator variable for participating subject lived within, or in direct vicinity to the presence of pollen season (pollen season vs. con- the city centers in question. trol season) for the association between outcomes and The data were converted to 24-hour means and were air pollutants in all data. Then, using only pollen season assigned to each study date and lagged values were calcu- data, we investigated interaction between low and high lated [27]. The birch pollen season was defined as the first levels of pollen at lag 0–2 (using 100 pollen/m as a cut- occurrence of 5 consecutive days where 4 days had pollen off ) and air pollution values. counts > 0 [26] (not dissimilar to the recently proposed The sample size power calculation was based analyses pollen exposure threshold from recent position paper) of effects in respiratory biomarkers in analyses compar - [28]. ing effects in pollen season and control season. In the current study, each individual had up to 105 measure- Statistical analyses ments (days with diary data), the statistical power is suffi - Descriptive statistics for diary-reported outcome and cient. In order to estimate effect of disease severity on the exposure variables are presented as mean and stand- response from pollen exposure, stratified sensitivity anal - ard deviations (SD) and categorical variable frequencies yses were performed in individuals who reported using were compared with Chi2-tests between control season inhalation corticosteroids, nasal steroids, individuals with Carlsen et al. Environmental Health (2022) 21:63 Page 4 of 13 Table 1 Demographic characteristics of the participants at baseline Variables Gothenburg Umeå Total Subjects, N 22 15 37 Females, n (%) 12 (55%) 7 (47%) 19 (51%) Males, n (%) 10 (45%) 8 (53%) 18 (49%) Age, yr (mean ± SD) 44 ± 10 47 ± 13 48 ± 15 Height, cm (mean ± SD) 175 ± 9 172 ± 9 171 ± 8 Weight, kg (mean ± SD) 79 ± 18 78 ± 16 77 ± 15 Body mass index, kg/m2 (mean ± SD) 26 ± 4 26 ± 6 26 ± 5 Birch pollen allergy, n (%) 22 (100%) 15 (100%) 37 (100%) Asthma Control Questionnaire (mean ± SD) 1.3 ± 1.1 1.3 ± 1.2 1.3 ± 1.1 Self-reported medication Short‑acting bronchodilator, n (%) 18 (82%) 6 (40%) 24 (65%) Long‑acting medication, n (%) 2 (9%) 2 (13%) 4 (11%) Inhalation steroids medication, n (%) 19 (86%) 6 (40%) 25 (68%) Nasal steroid, n (%) 7 (32%) 2 (13%) 9 (24%) Steroid medication (Inhalation and combination steroids), n (%) 19 (86%) 8 (53%) 27 (73%) Allergy medicine, n (%) 22 (100%) 11 (73%) 33 (89%) Asthma control questionnaire score calculated as a mean of six questions about asthma symptoms in the last week Regular usage in the last month or year poor asthma control as measured by ACQ (poor asthma in the analysis. The proportion of females was 51%, the control being defined as a score of ACQ above 1.5) [25]. mean age was 48 (SD 15) years, and the mean BMI was To determine if drop-out rates affected the results, we 26 (SD 5) kg/m2. The participants’ mean ACQ score at also analyzed the individuals who participated in all three baseline was 1.3 (SD 1.1). The regular use of short-act - waves separately. Data were also analyzed stratified by ing bronchodilator during the last year were reported sex. by 65% of the participants, long-acting bronchodilator The models for PEF were adjusted for autocorrelation medication by 11% and inhalation corticosteroid use was by incorporating a first-order autoregressive component. reported by 68% of the participants (Table 1). This autocorrelation function is not defined for logistic Air pollutants and pollen concentrations from station- regression models. ary measuring stations varied between seasons and study All results are reported per 100-unit pollen and one cites. NO concentrations were lower in Umeå than in interquartile range (IQR) change in air pollution con- Gothenburg during spring in both 2015 and 2016, mean 3 3 3 3 centrations. The results for diary reported symptoms 9.2 µg/m and 14.6 µg/m vs. 25.6 µg/m and 28.7 µg/m and medication usage are reported as odds ratios (ORs), respectively, but higher during control season by 39.5 µg/ 3 3 which express the relative change in the odds associated m vs. 28.9  µg/m although the standard deviation for with change in exposure with 95% confidence intervals the control season in Umeå was high. O concentrations (CI). The coefficients for ∆PEF are reported with 95% were quite higher in Umeå than in Gothenburg during 3 3 CI, which express the increase in outcome per increment spring 2016, 86.8 µg/m vs. 53.9 µg/m. PM concentra- 2.5 change in exposure. The interaction terms are reported tions were two to three times as high in Gothenburg as in with their confidence intervals and p-values, and the pre - Umeå in all seasons (Table 2, Fig. S1). dicted values of the outcomes at low and high values of Pollen concentrations were higher in Gothenburg than the variable pollen at lag 0–2 were plotted. Analyses were in Umeå spring 2015 (maximum pollen value 1087 vs. performed using the packages “lme4”, “meta”, and “ggef- 210), but during spring 2016 pollen concentrations were fects” [30–32] in R. marginally higher in Umeå than Gothenburg (maximum pollen value 964 vs. 950) (Table  2; Fig.  1). Temperatures Results and relative humidity also differed between locations and Descriptive statistics seasons reflecting the geographic differences between the Initially, 41 individuals with allergic asthma were locations (Table 2). recruited for the study, and 37 individuals had informa- Data are measured at centrally located ambient meas- tion on all key variables of interest and were included uring stations. C arlsen et al. Environmental Health (2022) 21:63 Page 5 of 13 Table 2 Daily averages of exposure in the two study centers medication on 40.7% of the days. Using more bronchodi- during the study waves (mean ± standard deviation) lation medicines than normal was reported on 9.2% of days (Table 3). Mean morning PEF  was 436 mL (SD 88) mo Wave Gothenburg Umeå and mean PEF was 444 (SD 92). ev x 3 NO (µg/m ) 1 (Pollen season) 25.6 ± 14.7 9.2 ± 5.0 Rates of symptoms and medication usage were higher 2 (Control season) 28.9 ± 18.3 37.5 ± 30.7 during pollen season compared to the control season 3 (Pollen season) 28.7 ± 19.0 14.6 ± 6.6 (visualized in Fig.  2). PEF-values during pollen season 3 3 O (µg/m ) 1 (Pollen season) 66.0 ± 12.3 63.5 ± 7.7 were not statistically significant different from control 2 (Control season) 52.0 ± 15.7 40.0 ± 19.4 season, with p-value = 0.15 for PEF and p-value = 0.24 mo 3 (Pollen season) 53.9 ± 13.8 86.8 ± 29.0 for PEF (data not shown). ev PM (µg/m ) 1 (Pollen season) 9.0 ± 6.3 3.7 ± 2.7* 2.5 2 (Control season) 6.9 ± 4.3 2.1 ± 1.3 Eec ff ts on symptoms and medication 3 (Pollen season) 7.1 ± 3.9 4.6 ± 2.6 In the pooled analysis (Table  4), pollen exposure at Pollen (grains per m ) 1 (Pollen season) 180 ± 254 43 ± 49 lag 0–2 was statistically significantly associated with 2 (Control season) ‑ ‑ increased ORs of reporting rhinitis or eye irritation OR 3 (Pollen season) 176 ± 221 231 ± 270 1.44 (95% CI 1.36–1.52) per 100 pollen/m pollen in the Temperature (°C) 1 (Pollen season) 9.1 ± 1.9 1.0 ± 4.2 single-exposure model, and lower, but still statistically 2 (Control season) 7.5 ± 3.5 10.0 ± 3.6 significant ORs in dual-exposure and multi-exposure 3 (Pollen season) 10.8 ± 4.4 5.9 ± 5.0 models (City-specific ORs are shown in Table S2a). The Relative humidity (%) 1 (Pollen season) 72.5 ± 11.0 75.1 ± 11.2 same was true for pollen exposure at lag 0–6 which was 2 (Control season) 85.5 ± 6.9 95.1 ± 6.8 associated with increased ORs of rhinitis or eye irrita- 3 (Pollen season) 64.7 ± 12.9 68.9 ± 12.6 tion in single-exposure models by OR 1.67 (95% CI 1.57– 1.78). PM was significantly associated with increased 2.5 rhinitis and or irritation in multi-exposure models by OR 1.16 (95% CI 1.02; 1.32). After initially testing individual lags from 0 to 3 (data Dyspnea was only significantly associated with pollen not shown), lag 0–2, and lag 0–6, for both pollen and air exposure in single-exposure models by OR 1.17 (95% CI pollution, and afterwards investigating lag-associations 1.10–1.25) for exposure at lag 0–2 and OR 1.23 (95% CI with DLNM methods, it was found that the highest ORs 1.09–1.37) for exposure at lag 0–6. There were no statisti - for dry cough and allergy medication were at lag 0–2 and cally significant associations between air pollutants and lag 0–6, and these lags are reported onwards. For PE F , dyspnea. mo we report results for the mean of the previous day and Dry cough was associated with pollen exposure at the day before that, lag 1 to lag 2 instead of lag 0 to lag 2. lag 0–2 in dual-exposure models with NO by OR 1.10 After excluding observations where participants reported (95% CI 1.02–1.18) and with O3 by OR 1.09 (95% CI having a fever, there were 3112 observations. In the dia- 1.01; 1.17) and in multi-exposure models by OR 1.09 ries, the most reported symptoms were rhinitis or eye (95% CI 1.01–1.17). For pollen exposure at lag 0–6 irritation, reported on 27.5% of days during the study there were statistically significant associations in dual- period, followed by dry cough, reported on 26.4% of all exposure models adjusted for O, PM by 1.14 (95% CI 3 2.5 days in the study. Participants reported taking allergy 1.05–1.24) and in multi-exposure models by OR 1.15 Fig. 1 Pollen concentrations in the two study centres (3‑ day moving average) during the study period Carlsen et al. Environmental Health (2022) 21:63 Page 6 of 13 Fig. 2 Proportion of diary‑reported respiratory symptoms and medication use during the study period. For eye irritation/rhinitis, dyspnea, dry cough, allergy medication, proportion reporting “Yes” vs. “No”. For asthma medication, reporting “more than yesterday” vs. “No”, “Less”, or “Same” (95% CI 1.05–1.28). Exposure to O at lag 0–2 was sta- multi-exposure models was noted. Exposure to N O 3 x tistically significantly associated with dry cough by OR at lag 0–2 was associated with significantly reduced 1.15 (95% CI 1.00-1.31) per IQR O in models with pol- ORs of allergy medication usage in both dual-and len, and by 1.14 (95% CI 1.09–1.19) in multi-exposure multi-exposure models, with OR at 0.72 (95%CI 0.52- models. Dry cough rates were associated with PM 1.00) and OR 0.74 (95%CI 0.59–0.92) respectively. O 2.5 3 in pairwise models by OR 1.18 (95% CI 1.05–1.25) per exposure was associated with increased use of allergy IQR PM (Table 4). medication with OR 1.29 (95% CI 1.02; 1.62) in dual- 2.5 Exposure to pollen at lag 0–2 was associated with exposure models. PM was associated with increased 2.5 increased use of allergy medication in single-expo- allergy medication usage in dual-exposure model by sure models by OR 1.58 (95%CI 1.44–1.74). In dual- OR 1.15 (95% CI 1.02–1.31) and in the multi-exposure and multi-pollutant models, the ORs were reduced model by OR 1.25 (95% CI 1.07; 1.46) (Table 4). but remained significant. For lag 0–6 pollen exposure Pollen exposure was statistically significantly associated there was a significant increase in allergy medication with increased use of bronchodilating medication only in reports in single-exposure models by OR 1.85 (95%CI single-exposure models by OR 1.15 (95% CI 1.08–1.21) at 1.60–2.14) and a similar reduction in the dual-and lag 0–2 and lag 0–6 by OR 1.23 (95% CI 1.16–1.32). There C arlsen et al. Environmental Health (2022) 21:63 Page 7 of 13 Table 3 Prevalence of daily diary‑reported symptoms, medication use, and mean (SD) of peak expiratory flow in the study population (n = 37) during the three study waves Wave 1 (1068 observations) Wave 2 (1066 observations) Wave 3 (978 observations) Total (3112 observations) Symptoms n (%) n (%) n (%) n (%) Eye irritation/rhinitis 649 (60.9%) 68 (6.4%) 368 (37.7%) 851 (27.5%) Dyspnea 252 (23.8%) 112 (10.6%) 215 (22.9%) 579 (19.0%) Dry cough 296 (27.9%) 239 (22.7%) 272 (28.9%) 807 (26.4%) Allergy medication 547 (51.3%) 118 (11.2%) 581 (61.9%) 1246 (40.7%) Bronchodilating medication No 389 (36.5%) 474 (44.9%) 355 (37.6%) 1218 (39.7%) Less 46 (4.3%) 64 (6.1%) 14 (1.5%) 124 (4.0%) Normal 533 (50.0%) 455 (43.1%) 454 (48.1%) 1442 (47.0%) More 98 (9.2%) 62 (5.9%) 121 (12.8%) 281 (9.2%) Peak expiratory flow (PEF) Mean (SD) Mean (SD) Mean (SD) Mean (SD) PEF (mL) 434 (81) 441 (90) 434 (93) 436 (88) mo PEF (mL) 442 (85) 448 (93) 442 (97) 444 (92) ev PEF  Peak expiratory flow measured in the morning,  PEF  Peak expiratory flow measured in the evening mo  ev  For eye irritation/rhinitis, dyspnea, dry cough, allergy medication, “Yes” vs. “No”. For bronchodilating medication, reporting “more than yesterday” vs. “No”, “Less”, or “Same” were no associations between use of bronchodilating with PM concentrations, by ORs 1.56 (95% CI 1.10– 2.5 medication and pollen in dual- or multi pollution mod- 2.21) and 1.72 (95% CI 1.13–2.64) per IQR increase in els, nor with NO, O and PM air pollution (Table 4). pollutant. The predicted ORs at low and high pollen lev - x 3 2.5 els are illustrated in Fig. 3a-d. Eec ff ts on PEF PM exposure was negatively associated with decreased In the pooled analysis, there were no statistically signifi - allergy medication in the low-pollen exposure scenario, cant associations between pollen at lag 0–2, lag 0–6 and and we speculate. ΔPEFmo at any adjustment level, but all estimates were Additionally, there were indications of pollen-air pol- negative. lutant interactions (with p-values below 0.1 but above ΔPEF and pollen at lag 0–2 were statistically signifi - 0.05) for dyspnea, bronchodilating medication and O ev 3 cantly associated in single-exposure models by -1.24 (95% (Table S3, Fig. S3). CI -2.45- -0.04) and in dual-exposure models adjusted for The autocorrelation adjustment performed for PEF did PM by -0.93 (95% CI -1.70-0.17) and by -0.93 (95% CI not change effect estimates of lag 0–2 exposure. 2.5 -1.71-0.15) mL per 100 pollen/m in the multi-exposure model. For ΔPEFev, there were no statistically significant Sensitivity analysis associations with air pollution (Table 5, city-specific esti - In the sensitivity analyses, in observations from individ- mates are shown in Table S2b). ual with poor asthma control (ACQ > 1.5), the estimated association between exposure and rhinitis or eye irrita- Interactions between pollen and air pollutants tion was higher than in the total data (OR 1.29 (95% CI We observed no interactions between pollen season and 1.12–1.48) vs. OR 1.22 (95% CI 1.14–1.31). For other air pollution on symptoms, medication usage or PEF symptoms and PEF, the estimates were similar or lower when applying a crude measure of pollen season (yes or than in total data (Table S4) and for allergy and asthma no), but when categorizing pollen at lag 0–2 into low and medication, the estimates were lower, although the con- high pollen concentrations, there were statistically sig- fidence intervals overlapped with the estimates from the nificant interactions with air pollution for symptoms and total data. In individuals who used inhaled corticosteroid medication usage (all interaction term coefficients and medication (ICS users, n = 25), lower ORs were seen for p-values are found in Table S3). There were interactions all symptoms compared to the total population, although for rhinitis or eye irritation between pollen concentra- reports of rhinitis or eye irritation were still statistically tions and O by OR 1.45 (95% CI 1.14–1.84) per IQR and significantly associated with pollen exposure (Table S4). with PM by 1.41 (95% CI 1.04–1.92) per IQR. Also, for In ICS users, there were no associations with ΔPEFev 2.5 dyspnea and allergy medication pollen levels interacted or ΔPEFmo. In individuals who used nasal steroids (NS Carlsen et al. Environmental Health (2022) 21:63 Page 8 of 13 Table 4 OR of diary‑reported symptoms and medication use associated with exposure to pollen at lag 0–2 and lag 0–6 and air pollution at lag 0–2 exposure to pollen and air pollutants in single‑, pairwise ‑, and multi‑ exposure models with 95% confidence interval (pooled results). Significant results are indicated with bold font Model adjustment Pollen – Lag 0–2 Pollen – Lag 0–6 NO – Lag 0–2 O – Lag 0–2 PM – Lag 0–2 x 3 2.5 OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) Rhinitis or eye irritation Adjusted for Pollen 1.44 (1.36; 1.52) 1.67 (1.57; 1.78) + NO 1.17 (1.02; 1.35) 1.28 (1.08; 1.53) 1.06 (0.86; 1.29) + O 1.14 (1.02; 1.26) 1.23 (1.12; 1.35) 0.81 (0.56; 1.17) + PM 1.18 (1.01; 1.37) 1.35 (1.06; 1.72) 0.88 (0.78; 1.01) 2.5 + NO, O, PM 1.25 (1.16; 1.35) 1.39 (1.27; 1.51) 0.71 (0.42; 1.19) 1.02 (0.34; 3.04) 1.16 (1.02; 1.32) x 3 2.5 Dyspnea Adjusted for Pollen 1.17 (1.10; 1.25) 1.23 (1.09; 1.37) + NO 1.04 (0.96; 1.14) 1.12 (1.02; 1.23) 0.75 (0.54; 1.05) + O 1.03 (0.95; 1.12) 1.09 (1.00; 1.20) 1.06 (0.69; 1.61) + PM 1.01 (0.93; 1.10) 1.08 (0.98; 1.19) 0.91 (0.55; 1.50) 2.5 + NO, O, PM 1.03 (0.95; 1.13) 1.09 (0.99, 1.20) 0.79 (0.54; 1.15) 0.96 (0.66; 1.40) 1.03 (0.81; 1.33) x 3 2.5 Dry cough Adjusted for Pollen 1.05 (0.90; 1.24) 1.08 (0.85; 1.35) + NO 1.10 (1.02; 1.18) 1.14 (1.05; 1.24) 0.97 (0.72; 1.31) + O 1.09 (1.01; 1.17) 1.14 (1.05; 1.24) 1.15 (1.00; 1.31) + PM 1.07 (0.99; 1.16) 1.14 (1.05; 1.24) 1.18 (1.05; 1.33) 2.5 + NO, O, PM 1.09 (1.01; 1.17) 1.15 (1.05; 1.28) 0.99 (0.72; 1.35) 1.14 (1.09; 1.19) 1.01 (0.97; 1.06) x 3 2.5 Allergy medication Adjusted for Pollen 1.58 (1.44; 1.74) 1.85 (1.60; 2.14) + NO 1.23 (1.12 1.35) 1.16 (1.08; 1.24) 0.72 (0.52; 1.00) + O 1.19 (1.08; 1.30) 1.15 (1.07; 1.23) 1.29 (1.02; 1.62) + PM 1.15 (1.01; 1.26) 1.16 (1.07; 1.23) 1.15 (1.02; 1.31) 2.5 + NO, O, PM 1.17 (1.07; 1.29) 1.15 (1.06; 1.25) 0.74 (0.59; 0.92) 1.01 (0.78; 1.32) 1.25 (1.07; 1.46) x 3 2.5 Bronchodilating medication Adjusted for Pollen 1.15 (1.08; 1.21) 1.23 (1.16; 1.32) + NO 1.07 (0.87; 1.32) 1.10 (0.85; 1.42) 0.84 (0.81; 0.88) + O 1.06 (0.87; 1.30) 1.09 (0.83; 1.43) 1.02 (0.77; 1.34) + PM 1.07 (0.97; 1.30) 1.10 (0.82; 1.49) 0.97 (0.97; 0.99) 2.5 + NO, O, PM 1.07 (0.86; 1.33) 1.12 (0.84; 1.50) 0.81 (0.62; 1.07) 1.04 (0.72; 1.49) 1.09 (0.86; 1.37) x 3 2.5 For eye irritation/rhinitis, dyspnea, dry cough, allergy medication, “Yes” vs. “No”. For bronchodilating medication, reporting “more than yesterday” vs. “No”, “Less”, or “Same” Results from pooled with meta-analysis of mixed model results for pollen (single-exposure), pollen and pairwise combinations of pollen and either NO , O and PM , x 3 2.5 and multi-exposure models adjusted for pollen. All models were adjusted for relative humidity and temperature, with identification number as a random effect. 3 3 3 3 Associations for are reported per 100 grains/m for pollen and per city-specific IQR for pollutants (Göteborg: NO 20.0 µg/ m, O 22.9 µg/ m, PM 3.3 µg/ m , Umeå: x 3 2.5 3 3 3 NO 14.7 µg/m, O 24.1 µg/ m, PM 2.8 µg m ) x 3 2.5 values of ΔPEF were reduced in both males and females, users, n = 9), the ORs for rhinitis or eye irritation and but the reduction in ΔPEF did not reach statistical sig- allergy medication in association with pollen counts were mo nificance in males (Fig. S4). higher than in any other group. The results obtained from individuals who participated at all three waves were similar to the total data (Table S4). Stratifying by sex, pol- Discussion len exposure was only significantly associated with dysp - In this asthma-diary panel study with data from two nea, asthma medication usage in female subjects, but the different pollen-seasons, we found that symptom prev - only significant interaction between exposure and sex alence was increased during birch pollen season, and was observed for bronchodilating medication, where the that pollen exposure at lag 0–2 and lag 0–6 was asso- effects were stronger in females. Morning and evening ciated strongly with prevalence of most symptoms of C arlsen et al. Environmental Health (2022) 21:63 Page 9 of 13 Table 5 Associations (B) between peak expiratory flow (PEF) in the morning and evening (PEF and PEF ) and exposure to mo ev pollen (per 100 grains) and air pollutants in single‑, pairwise ‑, and multi‑ exposure models with 95% confidence interval (pooled results). Significant results are indicated with bold font Pollen lag 0–2 Pollen lag 0–6 NO O PM2.5 x 3 β (95% CI) 95% CI 95% CI 95% CI 95% CI ΔPEF (mL) mo Adjusted for Pollen ‑1.10 (‑2.86; 0.66) ‑1.00 (‑3.00; 1.01) ‑ ‑ ‑ + NO ‑1.06 (‑3.04; 0.92) ‑0.97 (‑3.20; 1.27) ‑0.16 (‑3.21; 2.88) ‑ ‑ + O ‑1.07 (‑2.90; 0.76) ‑0.97 (‑3.05; 1.11) ‑ 1.54 (‑0.92; 4.00) ‑ + PM ‑1.10 (‑2.67; 0.47) ‑1.20 (‑3.09; 0.70) ‑ ‑ 2.16 (0.23; 4.09) 2.5 + NO, O, PM -1.05 (-2.78;-0.68) ‑1.17 (‑3.19; 0.86) ‑0.43 (‑2.48; 1.63) 0.18 (‑1.70; 2.06) 2.14 (1.01; 3.26) x 3 2.5 ΔPEFev (mL) Adjusted for Pollen -1.24 (-2.45; -0.04) ‑1.24 (‑3.88; 1.40) ‑ ‑ ‑ + NO ‑1.25 (‑2.53; 0.03) ‑1.26 (‑4.01; 1.48) 0.40 (‑0.97; 1.77) ‑ ‑ + O ‑1.22 (‑2.48; 0.04) ‑1.24 (‑3.88; 1.40) ‑ 0.66 (‑1.24; 2.55) ‑ + PM -0.93 (-1.70; -0.17) ‑0.72 (‑2.30; 0.86) ‑ ‑ 2.88 (‑1.80; 7.56) 2.5 + NO, O, PM -0.93 (-1.71; -0.15) ‑0.73 (‑2.25; 0.80) 0.17 (‑1.65; 1.31) ‑0.27 (‑2.35; 1.81) 3.21 (‑1.27; 7.68) x 3 2.5 ΔPEF: deviation from each individual’s mean of each season. Exposure at lag 1–2 for PEF and lag 0–2 for PEF mo ev Results from pooled with meta-analysis of mixed model results for pollen (single-exposure), pollen and pairwise combinations of pollen and either NO , O and PM , x 3 2.5 and multi-exposure models adjusted for pollen. All models were adjusted for relative humidity and temperature, with identification number as a random effect. 3 3 3 3 Associations are reported per 100 grains/m for pollen and per city-specific IQR for pollutants (Göteborg: NO 20.0 µg/ m, O 22.9 µg/ m, PM 3.3 µg/ m , Umeå: NO x 3 2.5 x 3 3 3 14.7 µg/m, O 24.1 µg/ m, PM 2.8 µg m ) 3 2.5 Fig. 3 Predicted marginal effects (proportions) of significant interactions between birch pollen levels (below or above 100 grains /m ) and pollutants (per µg/m3) on symptoms and allergy medication usage. Footnote: For eye irritation/rhinitis, dyspnea and allergy medication, “Yes” vs. th th “No”. X‑axis from 5 to 95 percentile. Results are from mixed models with identification number and city as a random effect. In addition to pairwise combinations of air pollution and pollen concentration indicator at lag 0–2, the models are adjusted for temperature and relative humidity. Results 3 3 3 are reported per global pollutant IQR (NO 16.4 µg/m, O 15.2 µg/m, PM 4.7 µg/m ) x 3 2.5 Carlsen et al. Environmental Health (2022) 21:63 Page 10 of 13 allergic asthma as well as medication usage (Table  3). allergy and short-term air pollution exposure during pol- Regarding air pollutants, O was associated with dry len season were indicated in a panel studies of individuals cough and increased use of allergy medication in dual- with allergy, but the results did not reach statistical sig- exposure models. PM2.5 was associated with rhinitis nificance [36]. After exposure in a road tunnel, asthmatic or eye irritation and with allergy medication in multi- subjects had increased response to an allergen provoca- exposure models and with dry cough in dual-exposure tion test with significant correlations between increased models. Evening-measured values of PEF were associ- asthma symptoms and NO , but not PM exposure [37]. 2 2.5 ated with pollen, and PM was associated with morn- In a controlled study of exposure to diesel exhaust and 2.5 ing-measured PEF in multi-exposure models (Table  5). allergen, the effect of exposure to diesel exhaust was aug - An increase in NOx was in most analyses unexpect- mented with simultaneous allergen exposure [20]. edly associated with lower risks. This phenomenon can In our study, we observed only moderate sex-differ - sometimes be explained by the negative correlation ences in the relationship between pollen and symptoms, between NO and O , and the reduction in risk dis- where pollen exposure was associated with increased risk x 3 appears after adjusting for O .In this case the adjust- of asthma medication use in females, but not in males ment did not remove the tendency, and one explanation (Fig. S4). A similar observation was made in a study of could be that variations in NO levels are more local asthma symptoms and personal sampler-measured expo- around the central monitoring station than the fluctua - sure [38] where asthma symptoms during daytime was tions in ozone exposure. associated with exposure to oxidants (NO and O ) in 2 3 Our study was designed to investigate if pollen expo- women, but not in men, suggesting that females may be sure increased susceptibility to air pollution, however, more susceptible to effects of air pollution than males. we acknowledge that the causal mechanism could be For PEF, a statistically significant increase of 2.16 (95% opposite; that air pollution increased the susceptibility CI 0.23–4.09) mL per IQR (3  µg/m) PM in morning 2.5 to pollen. However, with the current study design, we are PEF was observed in our study, where a recent review unable to test this, as we have no scenario where there is found that in asthmatic individuals, PEF was reduced no air pollution. 0.56  L/m per 10  µg/m PM [39] and a recent review 2.5 We investigated effects at different lags with DLNM found reduction of 2% per 10 pollen/m , but the time methodology and found that lag 0–2 or lag 0–6 was rel- of day (morning or evening) was not taken into account evant in most cases, but that in some cases, for example [33]. The unexpected direction of association for PM 2.5 for PEF , there could be an additional effect at lag 4–5 and PEF could be due to coinciding pollen and PM mo 2.5 which could be investigated in future studies (Fig. S2). peaks [19] where pollen exposure compels individuals The associations between pollen exposure and lower, with asthma to use more bronchodilating medication upper respiratory symptoms, and rhinitis or eye irrita- which reduces airway restriction (Table S4). tion symptoms in our study are consistent with those In our study, statistically significant pollen – air pollu - from a recent review and meta-analysis of pollen expo- tion interactions (interaction p < 0.05), where air pollu- sure on symptoms in individuals with allergy or asthma. tion effects where significantly higher during moderate The authors report that 10 pollen/m3 increase in expo - to high levels of pollen compared to low or no pollen, sure was associated with increased lower respiratory were present for rhinitis or eye irritation and PM and 2.5 symptoms by 1%, upper respiratory or ocular symptoms O , asthma symptoms and NO . In a recent study of 3 x by 6.6%, and any symptom of allergy or asthma increased self-reported symptoms via a mobile phone app, Bédard 2%. PEF was not associated with exposure [33]. In a study and colleagues (2020) observed significant associations of mobile-phone app reported symptoms, birch pollen between rhinitis symptoms and O during grass pollen exposure was associated with 3% increase in respiratory season, but not birch pollen season. PM was statisti- 2.5 symptoms and 4% increase in eye/nose symptoms per 10 cally significantly associated with rhinitis during the pol - pollen/m3 [34]. len season of one year, but not the other year in the study In addition to previous reports of associations between [40]. In another study of mobile phone app-recorded short term air pollution exposure and severe outcomes symptoms, O concentration led to an increased symp- such as increased emergency room visits and hospi- tom severity during the birch pollen season, but the asso- tal admissions for asthma [21], or symptom severity ciation was not significant after adjusting for climate [41]. recorded during MD visits [35], previous studies report In a panel study with 15 individuals with allergy, there increased asthma symptoms in association with air pol- was an increase in symptom severity associated with lution: PM exposure was associated with increase in exposure to air pollution during pollen season, but the 2.5 mobile-phone app reported respiratory symptoms by 6% estimates did not reach statistical significance [36]. per 10 µg/m [34]. An association between symptoms of C arlsen et al. Environmental Health (2022) 21:63 Page 11 of 13 In a recent review of pollen or fungal spore-air pollu- levels obtained at fixed monitoring stations does not tion interactions, it was concluded that although interac- accurately reflect each subject’s personal exposure tions had been shown or indicated in time series studies, level, but are rather used as a proxy for exposure in most of the existing did not consider groups at risk, and this study. However, this study has a time series ele- there were no studies of adults with allergic asthma [22] ment, and so, the day-to-day fluctuations in exposure, making our study unique in that context. which can be expected to be similar across a city, are The generalizability of our results to the previous lit - similar at the monitoring station and the study partici- erature on pollen-air pollution interactions in asthma pants locations. Furthermore, studies have shown cen- is limited as much of it pertains to more severe out- tral monitoring stations to be a reasonable proxy for comes such as asthma emergency room visits or hos- personal exposure to birch pollen [42]. Also, monitor- pital admissions. Also, the few available panel studies ing station data correlated well with personal exposure apply heterogenous methodologies [21, 22]. Some pre- in the study [27]. The panel data collected both in, and vious research on pollen-air pollution interaction have outside, pollen season ensured a substantial exposure compared associations with air pollution during and contrast. The start of the pollen season measured in outside pollen season [40, 41], which did not yield sig- the study were determined by the first 5-day period nificant results in our study (data not shown), where we with 4 days with pollen counts over 0 [26]. Newer defi- entered short-term pollen categories in the interaction nitions are available which could improve the exposure models. We speculate that the lag 0–2 pollen exposure assessment [28]. However, using this definition may is a more precise measure, and that exposure during have caused us to choose a slightly premature date for the whole pollen season was too variable (especially the pollen season in 2015, which was also very mild during wave 1 in Umeå, see Fig. 1) to obtain significant (Fig.  1; Table  1). We speculate that pollen counts dur- results. ing that season were too low to have an effect, but no In our study, the main results are presented as pooled universally accepted thresholds of effect have been results (with random effect meta-analyses) from sepa - reported [33]. rate analyses of each study center, as we knew of dif- We observed negative, or protective associations ferences in air pollution levels, temperature, as well as between NO exposure and symptoms. However, the recruitment procedures in the two study locations. In often-negative correlation between NO and O could x 3 the interaction analysis, we entered city as a random skew the analysis. However, we calculated a combina- effect in the model. For PEF, we tested for autocorre - tion value for oxides, O , and used that in the analy- lation by entering an autoregressive term, but for the sis, which did not change the results for pollen. The reported lag 0–2 results, this procedure did not alter the coefficients associated with O was statistically sig- results. There is no definition of this term for logistic nificant for bronchodilating medication and morning regression, so this could not be tested in the analysis of PEF and in both cases indicated a positive or protec- symptom scores. tive effect of O . Our study population was well-defined at base- line, with allergy confirmed with skin-prick tests, but Conclusions asthma was not confirmed with formal testing as in Our results show that a substantial proportion of allergic some other studies [36]. There are indications that asthma symptoms can be attributed to pollen exposure. some of the cases in our study population had mild Also, in addition to direct effects of pollen on symptoms disease, e.g., not using any asthma medication which in allergic asthma, pollen exposure increases susceptibil- could reduce the strength of our results (Table 1). The ity to adverse respiratory effects of exposure to PM and 2.5 results from the sensitivity analysis which was strati- O . fied by medication status underlines that our sam- Implementing the results into advisories could improve ple had sizable heterogeneity with respect to disease their predictive power, which could minimize the mor- severity. This fact has some consequences for the bidities associated with allergic asthma and allergy. Con- observed effect size. The sample size deduced based sidering the number of affected individuals, this could be on power calculation of respiratory biomarkers, but as of substantial benefit to public health. Also, reduction of the power in this analysis was increased by the many pollution levels should be priority to improve health in measures of each individual, the study is unlikely to be this susceptible population. underpowered. In our study, exposure to both pollen and air pollu- Abbreviations tion was measured at central urban background sta- ACQ: Asthma Control Questionnaire; NO : Nitrogen dioxides; O : Ozone; PEF: x 3 tions, which induces some uncertainty as exposure Peak Expiratory Flow; PM: Particle matter. Carlsen et al. Environmental Health (2022) 21:63 Page 12 of 13 3. Borna E, Nwaru BI, Bjerg A, Mincheva R, Rådinger M, Lundbäck B, et al. Supplementary Information Changes in the prevalence of asthma and respiratory symptoms in west‑ The online version contains supplementary material available at https:// doi. ern Sweden between 2008 and 2016. Allergy. 2019;74(9):1703–15. org/ 10. 1186/ s12940‑ 022‑ 00871‑x. 4. Backman H, Räisänen P, Hedman L, Stridsman C, Andersson M, Lindberg A, et al. Increased prevalence of allergic asthma from 1996 to 2006 and Additional file 1. further to 2016—results from three population surveys. Clin Experimental Allergy. 2017;47(11):1426–35. 5. Burte E, Leynaert B, Marcon A, Bousquet J, Benmerad M, Bono R, et al. Acknowledgements Long‑term air pollution exposure is associated with increased severity of The authors would like to acknowledge Marianne Andersson, Helene Friberg, rhinitis in 2 European cohorts. J allergy Clin Immunol. 2020;145(3):834–42. Annika Claesson (GU) and Helen Bertilsson and Chatrin Wahlgren (Umeå U), e6. who gathered the patient data and measurements. The authors would also 6. Guarnieri M, Balmes JR. Outdoor air pollution and asthma. Lancet. like to thank the participants. The authors acknowledge code from Hedi Katre 2014;383(9928):1581–92. Kriit and Johan Nilsson Sommar used in the DLNM analysis. 7. D’Amato M, Cecchi L, Annesi‑Maesano I, D’Amato G. News on Climate Change, Air Pollution, and Allergic Triggers of Asthma. J Investig Allergol Authors’ contributions Clin Immunol. 2018;28(2):91–7. HKC managed the data, analyzed, wrote the original draft, reviewed and 8. Jacquemin B, Kauffmann F, Pin I, Le Moual N, Bousquet J, Gormand F, et al. edited. SLH Performed the investigation, curated the data, write and edited Air pollution and asthma control in the Epidemiological study on the the manuscript, DO contributed to the methodology and analysis, AFB Genetics and Environment of Asthma. J Epidemiol Community Health. contributed to the discussion, writing and editing of the manuscript. LM 2012;66(9):796–802. contributed to the methodology, writing and review of the manuscript. KM 9. Biedermann T, Winther L, Till SJ, Panzner P, Knulst A, Valovirta E. Birch pol‑ managed and curated data. BF conceptualized, acquired funding, wrote and len allergy in Europe. Allergy. 2019;74(7):1237–48. reviewed the manuscript. A‑ C conceptualized the study, acquired funding, 10. Gilles S, Blume C, Wimmer M, Damialis A, Meulenbroek L, Gökkaya M, investigated, wrote and reviewed the study. The author(s) read and approved et al. Pollen exposure weakens innate defense against respiratory viruses. the final manuscript. Allergy. 2020;75(3):576–87. 11. Cockcroft DW, Davis BE. Mechanisms of airway hyperresponsiveness. J Funding Allergy Clin Immunol. 2006;118(3):551–9. Open access funding provided by University of Gothenburg. The study was 12. Beck I, Jochner S, Gilles S, McIntyre M, Buters JT, Schmidt‑ Weber C, et al. supported by FORMAS (A Swedish Research Council for sustainable develop‑ High environmental ozone levels lead to enhanced allergenicity of birch ment, dnr: 210‑2013‑805), the Swedish Heart and Lung Foundation (dnr: pollen. PLoS ONE. 2013;8(11):e80147. 2013 − 0279 and 2016 − 0250), the Swedish Cancer and Allergy Foundation 13. Baldacci S, Maio S, Cerrai S, Sarno G, Baïz N, Simoni M, et al. Allergy and and the Asthma and Allergy Association. asthma: Eec ff ts of the exposure to particulate matter and biological allergens. Respir Med. 2015;109(9):1089–104. Availability of data and materials 14. D’Amato G, Liccardi G, D’Amato M, Cazzola M. Outdoor air pollu‑ The data are available for other researchers upon reasonable request and tion, climatic changes and allergic bronchial asthma. Eur Respir J. pending ethics approval. 2002;20(3):763–76. 15. Reinmuth‑Selzle K, Kampf CJ, Lucas K, Lang‑ Yona N, Fröhlich‑Nowoisky J, Shiraiwa M, et al. Air pollution and climate change effects on allergies in Declarations the anthropocene: abundance, interaction, and modification of allergens and adjuvants. Environ Sci Technol. 2017;51(8):4119–41. Ethics approval and consent to participate 16. Zhu C, Farah J, Choel M, Gosselin S, Baroudi M, Petitprez D, et al. Uptake of The study was approved by the Regional Ethical Review Board at the Uni‑ ozone and modification of lipids in Betula Pendula pollen. Environ Pollut. versity of Gothenburg (Dnr: 682 − 14), and the participants signed informed 2018;242:880–6. consent forms for analysis and publication of the data. 17. Hong Q, Zhou S, Zhao H, Peng J, Li Y, Shang Y, et al. Allergenicity of recombinant Humulus japonicus pollen allergen 1 after combined expo‑ Competing interests sure to ozone and nitrogen dioxide. Environ Pollut. 2018;234:707–15. The authors declare no competing interests. 18. Rojo J, Oteros J, Picornell A, Maya‑Manzano JM, Damialis A, Zink K, et al. Eec ff ts of future climate change on birch abundance and their pollen Author details load. Glob Change Biol. 2021;27(22):5934–49. Section of Occupational and Environmental Medicine, School of Public 19. Ørby PV, Peel RG, Skjøth C, Schlünssen V, Bønløkke J, Ellermann T, et al. An Health and Community Medicine, Institute of Medicine, Sahlgrenska Academy, assessment of the potential for co‑ exposure to allergenic pollen and air University of Gothenburg, Medicinaregatan 16A, 40530 Gothenburg, Sweden. pollution in Copenhagen, Denmark. Urban Clim. 2015;14:457–74. Department of Public Health and Clinical Medicine, University Hospital, 20. Carlsten C, Blomberg A, Pui M, Sandstrom T, Wong SW, Alexis N, et al. Sustainable Health, Umeå University, Building 1A, 4st, 901 87 Umeå, Sweden. Diesel exhaust augments allergen‑induced lower airway inflamma‑ Section of Medicine, Department of Public Health and Clinical Medicine, tion in allergic individuals: a controlled human exposure study. Thorax. University Hospital, Umeå University, Building 1A, 4st, 901 87 Umeå, Sweden. 2016;71(1):35–44. Department of Statistics, USBE, Social Sciences Building Level 2 (ground 21. Anenberg SC, Haines S, Wang E, Nassikas N, Kinney PL. Synergistic health floor), Umeå University, 90187 Umeå, Sweden. effects of air pollution, temperature, and pollen exposure: a systematic review of epidemiological evidence. Environ Health. 2020;19(1):1–19. Received: 26 January 2022 Accepted: 9 June 2022 22. Lam HCY, Jarvis D, Fuertes E. Interactive effects of allergens and air pollution on respiratory health: A systematic review. Sci Total Environ. 2021;757:143924. 23. Forsberg B, Stjernberg N, Falk M, Lundback B, Wall S. Air pollution levels, meteorological conditions and asthma symptoms. Eur Respir J. References 1993;6(8):1109–15. 1. Landrigan PJ, Fuller R, Acosta NJ, Adeyi O, Arnold R, Baldé AB, 24. Forsberg B, Stjernberg N, Linne R, Segerstedt B, Wall S. Daily air pol‑ et al. The Lancet Commission on pollution and health. Lancet. lution levels and acute asthma in southern Sweden. Eur Respir J. 2018;391(10119):462–512. 1998;12(4):900–5. 2. To T, Stanojevic S, Moores G, Gershon AS, Bateman ED, Cruz AA, et al. 25. Juniper E, O’byrne P, Guyatt G, Ferrie P, King D. Development and Global asthma prevalence in adults: findings from the cross‑sectional validation of a questionnaire to measure asthma control. Eur Respir J. world health survey. BMC Public Health. 2012;12(1):204. 1999;14(4):902–7. C arlsen et al. Environmental Health (2022) 21:63 Page 13 of 13 26. Dahl Å, Galán C, Hajkova L, Pauling A, Sikoparija B, Smith M, et al. The onset, course and intensity of the pollen season. Allergenic pollen: Springer; 2013. pp. 29–70. 27. Haga SL, Hagenbjörk A, Olin A‑ C, Forsberg B, Liljelind I, Carlsen HK, et al. Personal exposure levels to O3, NOx and PM10 and the associa‑ tion to ambient levels in two Swedish cities. Environ Monit Assess. 2021;193(10):674. 28. Pfaar O, Bastl K, Berger U, Buters J, Calderon M, Clot B, et al. Defining pollen exposure times for clinical trials of allergen immunotherapy for pollen‑induced rhinoconjunctivitis–an EAACI position paper. Allergy. 2017;72(5):713–22. 29. Gasparrini A, Armstrong B, Kenward MG. Distributed lag non‑linear mod‑ els. Stat Med. 2010;29(21):2224–34. 30. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed‑ effects models using lme4. arXiv preprint arXiv:14065823. 2014. 31. Balduzzi S, Rücker G, Schwarzer G. How to perform a meta‑analysis with R: a practical tutorial. Evid Based Ment Health. 2019;22(4):153–60. 32. Lüdecke D. ggeffects: Tidy Data Frames of Marignal Eec ff ts from Regres‑ sion Models. J Open Source Softw. 2018;3(26):772. 33. Kitinoja MA, Hugg TT, Siddika N, Yanez DR, Jaakkola MS, Jaakkola JJ. Short‑term exposure to pollen and the risk of allergic and asthmatic manifestations: a systematic review and meta‑analysis. BMJ Open. 2020;10(1):e029069. 34. Jones PJ, Koolhof IS, Wheeler AJ, Williamson GJ, Lucani C, Campbell SL, et al. Can smartphone data identify the local environmental drivers of respiratory disease? Environ Res. 2020;182:109118. 35. Annesi‑Maesano I, Rouve S, Desqueyroux H, Jankovski R, Klossek J‑M, Thibaudon M, et al. Grass pollen counts, air pollution levels and allergic rhinitis severity. Int Arch Allergy Immunol. 2012;158(4):397–404. 36. Riediker M, Monn C, Koller T, Stahel WA, Wüthrich B. Air pollutants enhance rhinoconjunctivitis symptoms in pollen‑allergic individuals. Ann Allergy Asthma Immunol. 2001;87(4):311–8. 37. Svartengren M, Strand V, Bylin G, Jarup L, Pershagen G. Short‑term expo ‑ sure to air pollution in a road tunnel enhances the asthmatic response to allergen. Eur Respir J. 2000;15(4):716–24. 38. Chambers L, Finch J, Edwards K, Jeanjean A, Leigh R, Gonem S. Eec ff ts of personal air pollution exposure on asthma symptoms, lung function and airway inflammation. Clin Experimental Allergy. 2018;48(7):798–805. 39. Edginton S, O’Sullivan DE, King WD, Lougheed MD. The effect of acute outdoor air pollution on peak expiratory flow in individuals with asthma: A systematic review and meta‑analysis. Environ Res. 2021;192:110296. 40. Bédard A, Sofiev M, Arnavielhe S, Antó JM, Garcia‑Aymerich J, Thibaudon M, et al. Interactions Between Air Pollution and Pollen Season for Rhinitis Using Mobile Technology: A MASK‑POLLAR Study. J Allergy Clin Immunol Pract. 2020;8(3):1063‑73.e4. 41. Berger M, Bastl K, Bastl M, Dirr L, Hutter H‑P, Moshammer H, et al. Impact of air pollution on symptom severity during the birch, grass and ragweed pollen period in Vienna, Austria: Importance of O3 in 2010–2018. Environ Pollut. 2020;263:114526. 42. Bastl M, Bastl K, Karatzas K, Aleksic M, Zetter R, Berger U. The evaluation of pollen concentrations with statistical and computational methods on rooftop and on ground level in Vienna–How to include daily crowd‑ sourced symptom data. World Allergy Organization J. 2019;12(5):100036. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub‑ Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : lished maps and institutional affiliations. fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions

Journal

Environmental Health – Springer Journals

Published: Jul 6, 2022

Keywords: Birch; Betula; PM2.5; O3; Panel study; Allergic asthma; Pollen season

There are no references for this article.