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

Learn More →

The connection of pollen concentrations and crowd-sourced symptom data: new insights from daily and seasonal symptom load index data from 2013 to 2017 in Vienna

The connection of pollen concentrations and crowd-sourced symptom data: new insights from daily... Background: Online pollen diaries and mobile applications nowadays allow easy and fast documentation of pollen allergy symptoms. Such crowd-sourced symptom data provides insights into the development and the onset of a pollen allergy. Hitherto studies of the symptom load index (SLI) showed a discrepancy between the SLI and the total pollen amount of a season, but did not analyze the daily data. Methods: The Patient’s Hayfever Diary (PHD) was used as data pool for symptom data. Symptom data of Vienna (Austria) was chosen as a large and local sample size within the study period of 2013 until 2017. The city was divided into three different areas based on equal population densities and different environmental factors. Correlation factors, regression lines, locally weighted smoothing (LOESS) curves and line plots were calculated to examine the data. Results: Daily SLI and pollen concentration data correlates well and the progress of the SLI within a pollen season is mirrored by the pollen concentrations. The LOESS curves do not deviate much from the regression line and support the linearity of the symptom-pollen correlation on a daily basis. Seasonal SLI data does not follow the same pattern as the respective seasonal pollen indices. Results did not vary in the three areas within Vienna or when compared with the Eastern region of Austria showing no significant spatial variation of the SLI. Discussion: Results indicate a linear relationship of the SLI and pollen concentrations/seasonal polllen index (SPIn) on a daily basis for both in general and throughout the season, but not on a seasonal basis. These findings clarify the frequent misinterpretation of the SLI as index that is tightly connected to pollen concentrations, but reflects as well the seasonal variation of the burden of pollen allergy sufferers. Conclusion: More than just the seasonal pollen index has to be considered when the SLI of a selected pollen season has to be explained. Cross-reactivity to other pollen types, allergen content and air pollution could play a considerable role. The similar behavior of the SLI in Vienna and a whole region indicate the feasibility of a possible symptom forecast in future and justifies the use of a single pollen monitoring station within a city of the size of Vienna. Keywords: Symptom data- SLI, Patient’s Hayfever diary, Pollen allergy, Pollen concentrations * Correspondence: katharina.bastl@meduniwien.ac.at Aerobiology and Pollen Information Research Unit, Department of Oto-Rhino-Laryngology, Medical University of Vienna, Währinger Gürtel 18-20, 1090 Vienna, Austria Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 2 of 8 Background It is essential to understand the onset and progress of Allergies are recognized by the World Allergy Organization allergic symptoms for more adequate pollen forecasts [7] (WAO) as a global public health issue [1] – including and for the development of symptom forecasts. First pollen allergies. Allergic rhinitis makes up a significant pro- steps were already taken in this direction: the personal- portion of 10–30% of the global population [1] and pollen ized pollen information classifies users based on their allergies are still on the rise causing a considerable entries and compare them to other users of the same socioeconomic impact ([2, 3]). One million people of home region which is already reaching beyond a pollen eight million inhabitants are affected by a pollen allergy forecast [12]. However, a true symptom forecast is not on in Austria [4]. Therefore, pollen information plays a key the horizon yet and certainly requires a deep understand- role in allergen avoidance [5] and is strongly requested ing of the relationship and behavior of crowd-sourced by the population [6]. However, the generation of pollen symptom data. forecasts became more complicated and now has to rely The goal of this study was to evaluate how the SLI on good fundamental data including symptom data in the performs/behaves with a focus to (1) repeat a compari- best case [7] since pollen measurements alone do not son of the felt burden within a certain pollen season (by automatically infer the burden that has to be expected the means of a mean SLI) with the pollen data of a cer- from pollen allergy sufferers. tain pollen type (by the means of the seasonal pollen It has been known for some time now, that the rela- integral (SPIn)), (2) to examine if and in which extent tionship of allergic symptoms and allergen/pollen expos- correlation between daily SLI and pollen concentrations ure is not a direct one: Ellis et al. [8] showed that the occur and (3) gain insight into the temporal behavior of skin test reactivity to ragweed did neither correlate with the SLI and pollen concentrations throughout a pollen the rate of symptom development nor the symptom season (e.g., if peaks occur simultaneously). severity. A panel study on grass pollen allergy sufferers revealed that the linear relationship between grass pollen Methods concentrations and grass-sensitized persons is followed Pollen data by a plateau [9]. One of the early documented causes that The pollen data for this study was evaluated with an could explain at least a part of the observed patterns is the automatic volumetric pollen and spore trap of the Hirst “priming of the end organ” (short: priming effect), which design [18] situated at the rooftop of the Zentralanstalt comprises the observed decreased nasal threshold for aller- für Meteorologie und Geodynamik in Vienna. Pollen gic rhinitis after exposure/challenges and its reversibility data were collected following the minimum recommen- [10]. Up to now, the relationship and behavior of symptom dations of the European aerobiology community ([19, patterns and exposure is still not clarified in detail. 20]), also adhered to in the European Aeroallergen Net- Pollen diaries are easy to fill in for users and constitute work (EAN; https://ean.polleninfo.eu/Ean/). The follow- a useful pool of symptom data nowadays if handled care- ing seasons were calculated following the standard fully. They provide crowd-sourced data that directly relate season definition in the EAN database starting with 1% to the pollen concentrations and impact on persons con- of the cumulative annual pollen amount and ending with cerned. A number of tools, proven valuable, are available 95% of the annual pollen amount: alder, birch and grass (e.g. [11–13]). pollen season. We followed the recently published ter- The symptom load index (SLI) was developed as an minology for aerobiological studies by [21]. Thus, the index based on a pollen diary (Patient’s Hayfever diary sum of the average daily (alder, birch or grass) pollen (PHD); www.pollendiary.com;orinthe “Pollen” App) that amounts over the defined respective pollen season was provides information on the burden of pollen allergy suf- calculated as SPIn (formerly also named SPI = seasonal ferers [14]. As such it is a tool with a direct connection to pollen index) and the defined pollen season refers to the those who suffer from pollen allergies and provides crucial MPS (main pollen season in [21]). information. The SLI proved to be a robust index through different pollen seasons or years [15] and is applicable Symptom data even on a local level [16]. Pollen measurements are The symptom data used for this study was retrieved needed for the definition of a pollen season and for com- from the Patient’s Hayfever Diary (www.pollendiary.com) parisons with the SLI (e.g. [15, 17]). Thus, the SLI as a which is available as well as a mobile application number on its own is independent of pollen data and cal- (“Pollen”). Data for analyses were downloaded on 29th culated only on symptom data, but is in need for pollen of November 2017. Data were selected and retrieved data when a pollen season has to be defined as a pre- only for districts in Vienna (Austria) for the years 2013 requisite for a mean SLI in a certain period. Other issues until 2017 since user numbers are especially high from concern the misconception that SLI and pollen data do 2013 onwards. More than 50 entries per season allow a not correlate. This point will be clarified in detail herein. statistical analysis of a pollen season in terms of Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 3 of 8 crowd-sourced symptom data [17]. This recommendation a more detailed spatio-temporal analysis of crowd-sourced was far exceeded by daily mean entries of above 25 (see symptom data. Area classification). We applied the symptom load index In addition, the region “Pannonian Lowlands” comprised (SLI) described in [14], which comprises in summary the of Vienna, parts of Lower Austria and Burgenland was average of entered symptoms (eyes, nose, lungs) and included to examine the SLI pattern for a larger area in medication intake within a defined time period of a de- comparison. fined user sample. Users are granted anonymity and the services concerned were adapted to fulfill the latest EU Statistics regulation on data privacy (Regulation EU 2016/679) The graphs and computations were performed using the resulting in pooling only the absolute minimum of data statistical software R 3.4.3 [23]. We used descriptive sta- such as symptom data and an email address in this case. tistics to explore the behavior of the SLI in comparison Thus, personal data such as birthday or medical conditions to the daily pollen concentrations. Data were analyzed are were not collected. Personal and symptom datasets are for different regions (see Area classification). The rela- saved on separate servers to grant high security to avoid tionship of the SLI and the square root of the daily any unauthorized connection between them. pollen concentrations is figured in a scatter plot (Fig. 1a). A linear regression line (grey) and a locally weighted Area classification smoothing (LOESS) curve (black) were calculated. The Vienna was divided into three areas to explore the vari- linear regression line models the relationship of a ation of the SLI within a city. The division was based on dependent variable (SLI) and the explanatory variable population density and environmental factors. We estab- (daily pollen concentrations) via the formula y = a + bx. lished three areas: (1) “Vienna Center” (zip codes: 1010, In addition, the correlation factor of the SLI including 1040, 1050, 1060, 1070, 1080, 1090, 1100, 1120, 1150), the p-values and significance levels as well as the daily (2) “Vienna East” (zip codes: 1020, 1030, 1110, 1200, 1210, pollen concentrations in the respective pollen season 1220) and (3) “Vienna West” (zip codes: 1130, 1140, 1160, were calculated (see Table 2). LOESS [24] is a modern, 1170, 1180, 1190, 1230). Each area within Vienna thus in- non-parametric regression method and was chosen due cludes dense and less dense populated districts (see [22]). to its advantages as a flexible and simple tool applied to This was cross-checked with the sample size of the PHD complex processes without theoretical model. Besides, and resulted in a very similar sample size for each of the LOESS finds the curve of best fit without assuming the three areas throughout the different pollen seasons and data have to fit to a certain distribution shape. years. A comparable sample size is important to exclude The temporal development of the SLI and the daily possible influences of an unreliable small dataset. As an pollen concentration for specific pollen seasons were ex- example, the mean samples sizes for the year 2017 are plored in a line plot (Fig. 1b). The birch and the grass noted here for “Vienna Center” (alder/birch/grass pollen pollen season 2014 was chosen as an example. The SLI season: 26/43/23), “Vienna East” (alder/birch/grass pollen and the square root of the daily pollen concentrations season: 36/60/30) and “Vienna West” (alder/birch/grass were scaled to mean zero and standard deviation 1 to be pollen season: 38/50/28). The three areas correspond also comparable. The line chart was chosen to visualize the to areas of a different vegetation influence: “Vienna behavior of the SLI and the daily pollen concentrations Center” can be characterized as the most urbanized area over a certain time period (pollen season). with mostly parks and alley trees as green areas. This area is characterized by less grass vegetation (especially natural Results meadows) and fallow lands (preferred by Artemisia and The SLIs attained in different areas within Vienna show Ambrosia). “Vienna East” is influenced by the river little variation in different years, but the same pollen Danube and the Danube valley. This area is characterized season (Table 1). In general, the linear relationships and by alluvial vegetation and taxa preferring proximity to thus the correlation is very similar between the different water, e.g., species of the genus Alnus, Populus, Salix and areas in Vienna and the Pannonian lowlands and is Fraxinus. Grass areas and fallow lands are frequent due to observed due to the parallelism of the linear regression the agricultural influence in certain districts (1110, 1210 lines (Fig. 1a). Deviations from linearity are shown in the and 1220). “Vienna West” is influenced by the Vienna for- LOESS curve. While the SLI is almost perfectly linear ests. Therefore, natural meadows can be found in this area over the whole spectrum of observed grass pollen con- as well as forest elements, such as Corylus, Carpinus, centrations, the SLI deviates from the general trend for Fagus, Quercus and Pinus. Fallow lands occur in this area high birch pollen concentrations (Fig. 1a). The temporal due to the industrial area in certain districts (1130 and development of the SLI and the square root of the pollen 1230). Vienna was chosen for this study due to the high concentrations throughout a pollen season is shown in user numbers making it the only location so far possible for the line plots (Fig. 1b) and supports the relationship of Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 4 of 8 A B Fig. 1 Behavior of the SLI data and the pollen data as square root of the respective aeroallergen is displayed. a, Linear regression (grey) and LOESS curve (black) in the three areas within Vienna and the Pannonian Lowlands for the respective pollen seasons (2013–2017) of alder (upper), birch (middle) and grass (lower). Note the similar behavior of the LOESS for all areas/regions within the same pollen season. b, Temporal development of square root of the daily pollen concentrations (grey) and the respective SLI data (black). The birch pollen season 2014 (upper) and the grass pollen season 2014 (lower) were chosen as examples to show the similar progress of the datasets. Peaks of the pollen data are reflected in the SLI data SLI and pollen concentrations (Table 2). The three season when compared to the birch and grass pollen pollen seasons analyzed are characterized in the follow- season. The pattern of increase/decrease in SPIn of alder ing separately. pollen is consistent from 2014 until 2016, but is inter- rupted in 2017 (Table 1). Daily alder pollen concentra- Alder pollen seasons tions correlate with daily symptom data during the alder The mean SLI during the alder pollen season varies from pollen season. The correlation factors attain values a minimum of 2.4 to a maximum of 4.1 from 2013 to above 50% for the areas within Vienna up to around 2017. The three areas within Vienna show a variation 70% for the Pannonian Lowlands (Table 2). The behavior from 0.2 to 0.7 between each other in the observed time of the SLI during the alder pollen seasons 2013–2017 period. The mean SLI is lowest during the alder pollen shows the same progress within the three Vienna areas Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 5 of 8 Table 1 SLI values in different years (2013–2017) through Vienna ranges from 0.1 to 0.7 in the observed time different pollen seasons (alder, birch and grass) are shown for period. The mean SLI is highest during the birch pollen the three selected areas within Vienna (“Vienna East”, “Vienna season when compared with the alder and grass pollen Center” and “Vienna West”) in combination with their respective season. The pattern of increase/decrease in SPIn of birch average for Vienna pollen is inconsistent over most seasons (2014–2016) Region Year SLI/SPIn alder SLI/SPIn birch SLI/SPIn grasses and only consistent in 2017 (Table 1). Daily birch pollen Vienna Average 2013 2.4/2679 6.4/4517 3.8/3107 concentrations correlate with daily symptom data during Vienna East 2.3 6.2 3.9 the birch pollen season. The correlation factors attain values above 60% (Table 2). The behavior of the SLI dur- Vienna Center 2.4 6.4 3.7 ing the birch pollen seasons 2013–2017 shows the same Vienna West 2.6 6.6 3.8 progress in the three Vienna areas and the Pannonian Vienna Average 2014 ⇑ 3.6/2883⇑⇓ 5.1/6916⇑⇑ 4.1/2031 ⇓ lowlands (Fig. 1a). The progress within a single birch Vienna East 3.4 4.7 4.1 pollen season (Fig. 1b) shows that the peaks of the Vienna Center 3.5 5.3 4.0 square root of the daily mean pollen concentrations is Vienna West 4.0 5.3 4.1 mirrored in the SLI. The three areas in Vienna and the Pannonian lowlands show overall the same pattern as Vienna Average 2015 ⇓ 2.8/760⇓⇑ 5.6/4992⇓⇑ 4.3/3264 ⇑ well. Line plots of birch pollen seasons provided similar Vienna East 2.6 5.6 4.3 results. Vienna Center 2.5 5.6 4.2 Vienna West 3.2 5.5 4.3 Grass pollen seasons Vienna Average 2016 ⇑ 3.9/2685⇑⇓ 5.2/9233⇑⇔ 4.2/2522 ⇓ The mean SLI during the grass pollen season varies from Vienna East 3.7 5.1 4.2 a minimum of 3.8 to a maximum of 4.3 from 2013 to 2017. The three areas within Vienna show a variation Vienna Center 4.0 5.3 4.3 from 0.1 to 0.5 in the observed time period and thus the Vienna West 4.0 5.2 4.0 lowest variation. The mean SLI attains a value between Vienna Average 2017 ⇑ 4.1/1477⇓⇓ 4.8/3497⇓⇓ 4.0/2517 ⇔ those of the alder and the birch pollen season. The pattern Vienna East 4.0 4.8 3.9 of increase/decrease in SPIn of grass pollen is mostly in- Vienna Center 3.9 4.3 3.9 consistent over most seasons (2014; 2016–2017) with the Vienna West 4.3 5.0 4.4 exception of 2015 (Table 1). Daily grass pollen concentra- tions correlate with daily symptom data in the grass pollen The SPIn was calculated for the observed time period and is shown for Vienna with the average SLI. The arrows indicate the change of the SLI (left) and the season. The correlation factors are comparable to those SPIn (right) compared to the previous year. Note the inconsistency between during the birch pollen season and attain as well values the increase/decrease of SLI and SPIn in most of the cases, especially in the birch and the grass pollen seasons above 60% (Table 2). The LOESS for the grass pollen sea- sons 2013–2017 shows the same pattern within the three and the Pannonian lowlands (see LOESS curve in areas of Vienna and the Pannonian lowlands (Fig. 1a). The Fig. 1a). The data scatters more than for the birch and LOESS curve is here most similar to the regression line the grass pollen season and the LOESS curve deviates when compared with the alder and the birch pollen sea- from the regression line in some extent. sons. The progress within a single grass pollen season (Fig. 1b) shows the reflection of the peaks in the square Birch pollen seasons root of the daily mean pollen concentrations in the SLI The mean SLI during the birch pollen season varies even better than for the birch pollen season. Only from a minimum of 4.8 to a maximum of 6.4 from 2013 minor differences between the three regions in Vienna to 2017. The variation between the three areas within and the Pannonian lowlands can be observed (e.g. the Table 2 The correlation factors including p-values (and significance levels) between the daily SLI values of the defined pollen season (alder, birch and grass) in the observed time period (2013–2017) and the daily mean pollen concentrations are displayed for each of the areas in Vienna as well as for a whole region (“Pannonian Lowlands”) in comparison Corr.-Factor sqrt/p-value SLI alder SLI birch SLI grasses − 13 − 8 − 28 Pannonian Lowlands 69.1/2.02e *** 69.3/1.90e *** 69.1/1.85e *** − 12 − 7 − 21 Vienna East 59.4/9.15e *** 66.1/1.68e *** 62.2/1.77e *** − 10 − 7 − 21 Vienna Center 56.4/1.71e *** 65.9/2.45e *** 61.5/8.58e *** − 10 − 8 − 21 Vienna West 55.3/5.43e *** 67.9/4.96e *** 61.9/1.47e *** Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 “1 Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 6 of 8 double-headed peak during the main peak of the grass for the birch pollen season. This is in accordance with pollen season for “Vienna Center” compared to “Vienna the highest SLIs found in the birch pollen season, des- East” and “Vienna West”). Line plots of other grass pollen pite the lower frequency of birch pollen allergy. User seasons provided similar results. numbers and thus sample size may be connected with a high burden and triggered by the occurrence of severe Discussion symptoms. This is reflected as well in pollen informa- Daily and seasonal SLIs were analyzed for three different tion consumption (high user numbers and web site pollen seasons and for three different areas in Vienna to visits) during the birch pollen season [6]. examine the relationship of the SLI and pollen concen- trations/SPIn. The SLI data shows a significant correl- Allergen content ation with the daily pollen concentrations in general Significant positive correlations between major aeroaller- (Table 2) and throughout the season (Fig. 1b). This find- gens such as Bet v 1 and Phl p 5 with the SLI were ing has to be emphasized since the SLI has been misin- found for Vienna [26]. It is known that the allergen con- terpreted since its development [14] in the scientific tent may vary from pollen grain to pollen grain [27]. field and is thwarted by the new and further evidence of Pollen allergens can be present in the air also as small its inconsistency as mean value for a pollen season with respirable particles ([28–30]), thus resulting in a differ- the respective SPIn of that pollen season (Table 1). The ent occurrence when compared to pollen concentrations complexity of the results indicate a more or less linear [31]. Allergen content evaluation is highly time and cost relationship of SLI and pollen concentrations/SPIns on a consuming. Additionally, not only one major allergen is daily, but not on a seasonal basis. This can be explained sufficient to monitor and a range of major and minor al- by a variety of factors of importance for the development lergens are of importance for the immune response (e.g. and onset of allergic symptoms. [32, 33]) besides panallergens such as profilins and pol- calcins [34]. Therefore, it remains currently unknown, if Cross-reactivity a large range of allergens analyzed would explain the Cross-reactivities and other allergic triggers will play a role symptom burden. Allergen content in any case is a for individual pollen allergy sufferers. Hazel (Corylus)is major factor influencing the SLI. cross-reactive to alder (Alnus)[2] and its occurrence/ab- sence in different seasons may increase or decrease the Air quality and pollution mean SLI during an alder pollen season. There are several Air quality in general has an impact on the allergenicity candidates for a possible cross-reactivity during the of pollen grains ([35, 36]) and thus the burden of pollen birch pollen season (all flowering times of possible allergy sufferers e.g. the occurrence of atopic diseases cross-reactive aeroallergens are based on unpublished (asthmatic bronchitis, hay fever, eczema, sensitization; long term pollen data from Vienna and personal obser- [37]) and asthma admissions [38]. Allergenicity changes vation in the following): The hazel and alder pollen sea- in birch and grass pollen under air pollution such as ni- son sometimes overlaps with the beginning of the birch trogen dioxide (NO ), ammonia (NH ) and ozone (O ) 2 3 3 pollen season. Beech (Fagus)and oak(Quercus) flower and decreases microbial diversity [36]. Majd et al. [35] later during the birch pollen season and are possible found that air pollution causes the release of pollen pro- triggers for cross-reactivity [2]. Besides, ash (Fraxinus) teins resulting in a higher allergenicity besides a changed flowers about the same time with birch (Betula), but is shape of the pollen and its tectum. Fine particles shall not closely related. It belongs to the olive family (Olea- be mentioned as well, since diesel exhaust carbon parti- ceae) and may cause as well pollen allergy during spring cles are able to bind to major allergens under in vitro time. Poly-sensitized pollen allergy sufferers may react conditions (e.g. Lol p 1; [39]). during the analyzed time period to ash and birch D´Amato et al. [40] summarized the possible relation- pollen. However, the frequency of ash pollen allergy is ships between air pollution and allergens as follows: (1) with 17.7% much lower than birch pollen allergy with modified allergenicity, (2) interaction with microscopic 41.7% in Eastern Austria [25]. A number of plants have allergen-carrying particles reaching to the lower airways, (3) the potential to irritate pollen allergy sufferers during inflammatory effect with increased epithelial permeability the grass pollen season such as plantain (Plantago), and (4) adjuvant immunologic effect on IgE synthesis in dock (Rumex), the nettle family (Urticaceae), the goose- atopic persons. Therefore, it has to be assumed that air pol- foot family (Amaranthaceae) or mercury (Mercurialis) lution could explain observed SLI patterns in addition. [2]. Grass pollen allergy is the most frequent pollen al- lergy in Eastern Austria with a frequency of 56.3% [25]. Remarks and limitations It is noteworthy here to point to the sample size (see The SLI calculations performed herein rely on an excel- Symptom data and Area classification) which is largest lent sample size and allow to split Vienna in three areas Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 7 of 8 to observe possible spatio-temporal differences. The ob- (3) pollen monitoring stations do not have to be too servation period of five years and the comparison with a numerous. Results indicate that a second or more whole bio-geographical region (the Pannonian lowlands) pollen traps in Vienna would not provide any adds to the robustness of the outcomes and allows to benefit, since the SLI patterns are comparable in compare insights in different years and different pollen the whole city of Vienna. A certain density of pollen seasons. A noteworthy impact of the vegetation (see Area monitoring stations across a country/region has of classification) could not be observed in the three areas, course to be guaranteed and depends on vegetation, e.g. the SLI was not highest in “Vienna East”, where the topography and other factors. Danube has a strong influence. The results from [14] Abbreviations could be reproduced (seasonal SLI and SPIn do not in- EAN: European Aeroallergen Network (https://ean.polleninfo.eu/Ean); crease/decrease with each other). This is contrasted by the LOESS: Locally weighted smoothing; PHD: Patient’s Hayfever Diary (www.pollendiary.com); SLI: Symptom load index; SPIn: Seasonal Pollen Integral correlation factors line plots that were produced to show a correlation concerning the daily SLI and daily mean pollen Acknowledgements concentrations (Table 2 and Fig. 1a). Moreover, the results We thank Christoph Jäger for his constant efforts concerning the Pollen Diary. We are grateful to Alexander Kowarik for support in the statistical analyses. do not match with those from Berlin [41] – the only other Furthermore, we are grateful for the input of two anonymous reviewers. study that analyzed the spatio-temporal differences of crowd-sourced symptom data within a city to our know- Availability of data and materials Users were guaranteed anonymity, so now raw data or individual user data ledge. However, the observed differences therein could is available. have occurred due to the much lower sample size in this study in combination with the larger study area. Authors’ contributions Care has to be taken concerning the interpretation of KB drafted the main body of the manuscript, supervised the analyzes and prepared the figures. MK prepared the Tables, took part in the analysis of the the results. Pollen allergy is a complex disease and a range results and evaluated the pollen datasets. MB took part in the analysis of the of key factors might play an important role as discussed results and prepared the raw datasets of pollen and symptom data. UB took concerning cross-reactivities, allergenicity and air pollu- part in drafting the manuscript and supervised the study. All authors designed the study and revised and approved the final version of the tion. Therefore, important limitations are comprised of (1) manuscript. the SLI is calculated for a certain pollen season, but might be caused also by other aeroallergens and (2) users can Ethics approval and consent to participate Only crowd-sourced and anonymized symptom data was used, so no ethics not be characterized directly as patients due to the miss of approval was needed. personal information (although the crowd-sourced nature of the data assures significance; e.g., [15] and works also Consent for publication All authors confirm their consent for publication. for local phenomena e.g., [16]). Competing interests Conclusions Uwe Berger developed the PHD and the “Pollen” App. The services are free and there is no financial interest. The in-depth analysis of daily and seasonal SLIs of Katharina Bastl and Maximilian Kmenta were involved in the development Vienna revealed that the SLI and pollen concentrations and/or improvements of these services. correlate on a daily basis, but not on the seasonal level Markus Berger reports no competing interests. and that the SLI behavior and pattern do not vary on a local level. In fact, the SLI behavior of a large city like Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in Vienna is more or less the same as for a whole region such published maps and institutional affiliations. as the Pannonian lowlands. These results lead to far-reaching consequences: Author details Aerobiology and Pollen Information Research Unit, Department of Oto-Rhino-Laryngology, Medical University of Vienna, Währinger Gürtel (1) a symptom forecast based on crowd-sourced 18-20, 1090 Vienna, Austria. Paracelsus Medizinische Privatuniversität, symptom data could be feasible in the future. Strubergasse 21, 5020 Salzburg, Austria. The development of the SLI could be calculated Received: 19 July 2018 Accepted: 23 August 2018 since the SLI shows a widespread, consistent pattern during the season as shown herein. (2) the overall severity of a season is impacted by References 1. Pawankar R, Canonica GW, Holgate ST, Lockey RF, Blaiss MS. White Book on additional factors than just the major aeroallergen Allergy: Update 2013: WAO World Allergy Organization; 2013. http://www. in the air during this period. The mean SLI during worldallergy.org/wao-white-book-on-allergy. the pollen season usually does not correlation with 2. D’Amato G, Cecchi L, Bonini S, Nunes C, Annesi-Maesano I, Behrendt H, et al. Allergenic pollen and pollen allergy in Europe. Allergy. 2007;62:976–90. the SPIn. This should be considered for a range of https://doi.org/10.1111/j.1398-9995.2007.01393.x. activities such as pollen forecasting and clinical 3. Asher MI, Montefort S, Björksten B, Lai CK, Strachan DP, Weiland SK, et al. studies. Worldwide time trends in the prevalence of symptoms of asthma, allergic Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 8 of 8 rhinoconjunctivitis, and eczema in childhood: ISAAC phases one and three 25. Hemmer W, Schauer U, Trinca AM, Neumann C. Endbericht 2009 zur Studie: repeat multicountry cross-sectional surveys. Lancet. 2006;368:733–43. Prävalenz der Ragweedpollen-Allergie in Ostösterreich. Amt der NÖ https://doi.org/10.1016/S0140-6736(06)69283-0. Landesregierung, Landesamtsdirektion, Abteilung Gebäudeverwaltung, 4. Dorner T, Rieder A, Lawrence K, Kunze M. Österreichischer Allergiebericht. Amtsdruckerei: St. Pölten; 2010. Vienna: Verein Altern mit Zukunft; 2006. 26. Bastl K, Kmenta M, Pessi AM, Prank M, Saarto A, Sofiev M, et al. First comparison of allergen content (Bet v 1 and Phl p 5 measurements) with 5. Kiotseridis H, Cilio CM, Bjermer L, Tunsäter A, Jacobsson H, Dahl A. Grass symptom and pollen data from four European regions during 2009–2011. pollen allergy in children and adolescents-symptoms, health related quality Sci Total Environ. 2016;548–549:229–35. https://doi.org/10.1016/j.scitotenv. of life and the value of pollen prognosis. Clin Transl Allergy. 2013;3:19. 2016.01.014. https://doi.org/10.1186/2045-7022-3-19. 27. Buters JTM, Thibaudon M, Smith M, Kennedy R, Rantio-Lehtimäki A, Albertini 6. Kmenta M, Zetter R, Berger U, Bastl K. Pollen information consumption as an R, et al. Release of bet v 1 from birch pollen from 5 European countries. indicator of pollen allergy burden. Wien Klin Wochenschr. 2015;128(1–2):59–67. Results from the HIALINE study. Atmos Environ. 2012;55:496–505. https://doi.org/10.1007/s00508-015-0855-y. https://doi.org/10.1016/j.atmosenv.2012.01.054. 7. Bastl K, Berger M, Bergmann KC, Kmenta M, Berger U. The medical and 28. Spieksma FT, Emberlin JC, Hjelmroos M, Jäger S, Leuschner RM. scientific responsibility of pollen information services. Wien Klin Atmospheric birch (Betula) pollen in Europe: trends and fluctuations in Wochenschr. 2017;129:70–4. https://doi.org/10.1007/s00508-016-1097-3. annual quantities and the starting dates of the seasons. Grana. 1995;34:51–7. 8. Ellis AK, Ratz JD, Day AH, Day JH. Factors that affect the allergic rhinitis https://doi.org/10.1080/00173139509429033. response to ragweed allergen exposure. Ann Allergy Asthma Immunol. 29. Schäppi G, Supphioglu C, Taylor PE, Knox RB. Concentrations of the major 2010;104(4):293–8. https://doi.org/10.1016/j.anai.2010.01.012. birch tree allergen bet v 1 in pollen and respirable fine particles in the 9. Caillaud DM, Martin S, Segala C, Besancenot JP, Thibaudon M. French atmosphere. J Allergy Clin Immunol. 1997;100:656–61. https://doi.org/10. aerobiology network. Nonlinear short-term effects of airborne Poaceae 1016/S0091-6749(97)70170-2. levels on hay fever symptoms. J Allergy Clin Immunol. 2012;130(3):812–4. 30. Schäppi GF, Taylor PE, Pain MC, Cameron PA, Dent AW, Staff IA, et al. https://doi.org/10.1016/j.jaci.2012.04.034. Concentrations of major grass group 5 allergens in pollen grains and atmo- 10. Connell JT. Quantitative intranasal pollen challenge, II: effect of daily pollen spheric particles: implications for hay fever and allergic asthma sufferers challenge, environmental pollen exposure, and placebo challenge on the nasal sensitized to grass pollen allergens. Clin Exp Allergy. 1999;29(5):633–41. membrane. Allergy. 1986;41:123–39. https://doi.org/10.1016/0021-8707(68)90053-1. https://doi.org/10.1046/j.1365-2222.1999.00567.x. 11. Costa C, Menesatti P, Brighetti MA, Travaglini A, Rimatori V, Di Rienzo BA, et al. 31. Buters JTM, Weichenmeier I, Ochs S, Pusch G, Kreyling W, Boere AJ, et al. Pilot study on the short-term prediction of symptoms in children with hay The allergen bet v1 in fractions of ambient air deviates from birch pollen counts. fever monitored with e-health technology. Eur Ann Allergy Clin Immunol. Allergy. 2010;65(7):850–8. https://doi.org/10.1111/j.1398-9995. 2009.02286.x. 2017;46(6):216–25. 32. Andersson K, Lidholm J. Characteristics and immunobiology of grass pollen 12. Kmenta M, Bastl K, Jäger S, Berger U. Development of personal pollen aller- gens. Int Arch Allergy Immunol. 2003;130(2):87–107. https://doi.org/10. information – the next generation of pollen information and a step forward 1159/ 000069013. for hay-fever sufferers. Int J Biometeorol. 2014;58:1721–6. https://doi.org/10. 33. Karamloo F, Schmitz N, Scheurer S, Foetisch K, Hoffmann A, Haustein D, et 1007/s00484-013-0776-2. al. Molecular cloning and characterization of a birch pollen minor allergen, 13. de Weger LA, Hiemstra PS, Op den Buysch E, van Vliet AJ. Spatiotemporal bet v 5, belonging to a family of isoflavone reductase-related proteins. J monitoring of allergic rhinitis symptoms in the Netherlands using citizen Allergy Clin Immunol. 1999;104(5):991–9. https://doi.org/10.1016/S0091- science. Allergy. 2014;69(8):1085–91. https://doi.org/10.1111/all.12433. 6749(99)70080-1. 14. Bastl K, Kmenta M, Jäger S, Bergmann KC, Berger U. Development of a 34. Hauser M, Roulias A, Ferreira F, Egger M. Panallergens and their impact on symptom load index – enabling temporal and regional pollen season the allergic patient. Allergy Asthma Clin Immunol. 2010;6(1):1. https://doi. comparisons and pointing out the need for personalized pollen information. org/10.1186/1710–1492-6-1. Aerobiologia. 2014;30:296–80. https://doi.org/10.1007/s10453-014-9326-6. 35. Majd A, Chehregani A, Moin M, Gholami M, Kohno S, Nabe T, et al. The 15. Bastl K, Kmenta M, Geller-Bernstein C, Berger U, Jäger S. Can we improve effects of air pollution on structures, proteins and allergenicity of pollen pollen season definitions by using the symptom load index in addition to grains. Aerobiologia. 2004;20:111–8. pollen counts? Environ Pollut. 2015;204:109–16. https://doi.org/10.1016/j. 36. Obersteiner A, Gilles S, Frank U, Beck I, Häring F, Ernst D, et al. Pollen- envpol.2015.04.016. associated microbiome correlates with pollution parameters and the 16. Bastl K, Kmenta M, Berger U. Unusually early flowering of alder in Austria: allergenicity of pollen. PLoS One. 2016;11(2):e0149545. https://doi.org/10. first report of Alnus x spaethii in Austria, combined LM and SEM study of 1371/journal.pone.0149545. alder species and impact on pollen allergy sufferers. Aerobiologia. 2015; 37. Morgenstern V, Zutavern A, Cyrys J, Brockow I, Koletzko S, Krämer U, et al. 31(4):515–21. https://doi.org/10.1007/s10453-015-9383-5. Atopic diseases, allergic sensitization, and exposure to traffic-related air 17. Kmenta M, Bastl K, Berger U, Kramer MF, Heath MD, Pätsi S, et al. The grass pollution in children. Am J Respir Critical Care Med. 2008:1331–7. pollen season 2015: a proof of concept multi-approach study in three https://doi.org/10.1164/rccm.200701-036OC. different European cities. World Allergy Organ J. 2017;10:31. https://doi.org/ 38. Sunyer J, Spix C, Quénel P, Ponce-de-Léon A, Pönka A, Barumandzadeh T, et 10.1186/s40413-017-0163-2. al. Urban air pollution and emergency admissions for asthma in four 18. Hirst JM. An automatic volumetric spore trap. Ann Appl Biol. 1952;39:257–65. European cities: the APHEA project. Thorax. 1997;52(9):760–5. https://doi.org/10.1111/j.1744-7348.1952.tb00904.x. 39. Knox RB, Suphioglu C, Taylor P, Desai R, Watson HC, Peng JL, et al. Major 19. GalánC,Smith M,ThibaudonM, FrenguelliG,Oteros J,GehrigR,etal. grass pollen allergen lol p 1 binds to diesel exhaust particles: Pollen monitoring: minimum requirements and reproducibility of implications for asthma and air pollution. Clin Exp Allergy. 1997;27(3):246–51. analysis. Aerobiologia. 2014;30(4):385–95. https://doi.org/10.1007/ https://doi.org/10.1111/j.1365-2222.1997.tb00702.x. s10453-014-9335-5. 40. D’Amato G, Cecchi L, D’Amato M, Liccardi G. Urban air pollution and 20. Jäger S, Mandroli P, Spieksma F, Emberlin J, Hjelmroos M, Rantio-Lehtimaki climate change as environmental risk factors of respiratory allergy: an A. News. Aerobiologia. 1995;11:69–70. update. J Invest Allerg Clin. 2010;20(2):95–102. 21. Galán C, Ariatti A, Bonini M, Clot B, Crouzy B, Dahl A, et al. Recommended 41. Werchan B, Werchan M, Bergmann KC. Studie zum Vergleich von terminology for aerobiological studies. Aerobiologia. 2017;33:293–5. allergischen Symptomdaten aus einer Pollen-App mit 22. Vienna Government. EinwohnerInnen-Dichte 2017. https://www.wien.gv.at/ Pollenkonzentrationen aus 2014 in Berlin. Umwelt und Gesundheit 02/ stadtentwicklung/grundlagen/stadtforschung/karten/bevoelkerung.html.2017. 2017, Umweltbundesamt. 2017. https://www.umweltbundesamt.de/sites/ Accessed 20 Mar 2018. default/files/medien/1410/publikationen/2017-11-06_umwelt-gesundheit_ 23. R Core Team. R: A language and environment for statistical 02-2017_pollenstudie-2016.pdf. Accessed 26 Mar 2018. computing. R Foundation for Statistical Computing 2017; Vienna, Austria https://www.R-project.org/ 24. Cleveland WS, Grosse E, Shyu WM. Local Polynomial Regression Fitting Local regression models. In: Chambers JM, Tj H, editors. Statistical models in S. California: Wadsworth & Brooks/Cole; 1992. p. 309–76. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png World Allergy Organization Journal Springer Journals

The connection of pollen concentrations and crowd-sourced symptom data: new insights from daily and seasonal symptom load index data from 2013 to 2017 in Vienna

Download PDF
8 pages

Loading next page...
 
/lp/springer-journals/the-connection-of-pollen-concentrations-and-crowd-sourced-symptom-data-n7mQhAE4U0

References (45)

Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s).
Subject
Medicine & Public Health; Allergology; Immunology
eISSN
1939-4551
DOI
10.1186/s40413-018-0203-6
Publisher site
See Article on Publisher Site

Abstract

Background: Online pollen diaries and mobile applications nowadays allow easy and fast documentation of pollen allergy symptoms. Such crowd-sourced symptom data provides insights into the development and the onset of a pollen allergy. Hitherto studies of the symptom load index (SLI) showed a discrepancy between the SLI and the total pollen amount of a season, but did not analyze the daily data. Methods: The Patient’s Hayfever Diary (PHD) was used as data pool for symptom data. Symptom data of Vienna (Austria) was chosen as a large and local sample size within the study period of 2013 until 2017. The city was divided into three different areas based on equal population densities and different environmental factors. Correlation factors, regression lines, locally weighted smoothing (LOESS) curves and line plots were calculated to examine the data. Results: Daily SLI and pollen concentration data correlates well and the progress of the SLI within a pollen season is mirrored by the pollen concentrations. The LOESS curves do not deviate much from the regression line and support the linearity of the symptom-pollen correlation on a daily basis. Seasonal SLI data does not follow the same pattern as the respective seasonal pollen indices. Results did not vary in the three areas within Vienna or when compared with the Eastern region of Austria showing no significant spatial variation of the SLI. Discussion: Results indicate a linear relationship of the SLI and pollen concentrations/seasonal polllen index (SPIn) on a daily basis for both in general and throughout the season, but not on a seasonal basis. These findings clarify the frequent misinterpretation of the SLI as index that is tightly connected to pollen concentrations, but reflects as well the seasonal variation of the burden of pollen allergy sufferers. Conclusion: More than just the seasonal pollen index has to be considered when the SLI of a selected pollen season has to be explained. Cross-reactivity to other pollen types, allergen content and air pollution could play a considerable role. The similar behavior of the SLI in Vienna and a whole region indicate the feasibility of a possible symptom forecast in future and justifies the use of a single pollen monitoring station within a city of the size of Vienna. Keywords: Symptom data- SLI, Patient’s Hayfever diary, Pollen allergy, Pollen concentrations * Correspondence: katharina.bastl@meduniwien.ac.at Aerobiology and Pollen Information Research Unit, Department of Oto-Rhino-Laryngology, Medical University of Vienna, Währinger Gürtel 18-20, 1090 Vienna, Austria Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 2 of 8 Background It is essential to understand the onset and progress of Allergies are recognized by the World Allergy Organization allergic symptoms for more adequate pollen forecasts [7] (WAO) as a global public health issue [1] – including and for the development of symptom forecasts. First pollen allergies. Allergic rhinitis makes up a significant pro- steps were already taken in this direction: the personal- portion of 10–30% of the global population [1] and pollen ized pollen information classifies users based on their allergies are still on the rise causing a considerable entries and compare them to other users of the same socioeconomic impact ([2, 3]). One million people of home region which is already reaching beyond a pollen eight million inhabitants are affected by a pollen allergy forecast [12]. However, a true symptom forecast is not on in Austria [4]. Therefore, pollen information plays a key the horizon yet and certainly requires a deep understand- role in allergen avoidance [5] and is strongly requested ing of the relationship and behavior of crowd-sourced by the population [6]. However, the generation of pollen symptom data. forecasts became more complicated and now has to rely The goal of this study was to evaluate how the SLI on good fundamental data including symptom data in the performs/behaves with a focus to (1) repeat a compari- best case [7] since pollen measurements alone do not son of the felt burden within a certain pollen season (by automatically infer the burden that has to be expected the means of a mean SLI) with the pollen data of a cer- from pollen allergy sufferers. tain pollen type (by the means of the seasonal pollen It has been known for some time now, that the rela- integral (SPIn)), (2) to examine if and in which extent tionship of allergic symptoms and allergen/pollen expos- correlation between daily SLI and pollen concentrations ure is not a direct one: Ellis et al. [8] showed that the occur and (3) gain insight into the temporal behavior of skin test reactivity to ragweed did neither correlate with the SLI and pollen concentrations throughout a pollen the rate of symptom development nor the symptom season (e.g., if peaks occur simultaneously). severity. A panel study on grass pollen allergy sufferers revealed that the linear relationship between grass pollen Methods concentrations and grass-sensitized persons is followed Pollen data by a plateau [9]. One of the early documented causes that The pollen data for this study was evaluated with an could explain at least a part of the observed patterns is the automatic volumetric pollen and spore trap of the Hirst “priming of the end organ” (short: priming effect), which design [18] situated at the rooftop of the Zentralanstalt comprises the observed decreased nasal threshold for aller- für Meteorologie und Geodynamik in Vienna. Pollen gic rhinitis after exposure/challenges and its reversibility data were collected following the minimum recommen- [10]. Up to now, the relationship and behavior of symptom dations of the European aerobiology community ([19, patterns and exposure is still not clarified in detail. 20]), also adhered to in the European Aeroallergen Net- Pollen diaries are easy to fill in for users and constitute work (EAN; https://ean.polleninfo.eu/Ean/). The follow- a useful pool of symptom data nowadays if handled care- ing seasons were calculated following the standard fully. They provide crowd-sourced data that directly relate season definition in the EAN database starting with 1% to the pollen concentrations and impact on persons con- of the cumulative annual pollen amount and ending with cerned. A number of tools, proven valuable, are available 95% of the annual pollen amount: alder, birch and grass (e.g. [11–13]). pollen season. We followed the recently published ter- The symptom load index (SLI) was developed as an minology for aerobiological studies by [21]. Thus, the index based on a pollen diary (Patient’s Hayfever diary sum of the average daily (alder, birch or grass) pollen (PHD); www.pollendiary.com;orinthe “Pollen” App) that amounts over the defined respective pollen season was provides information on the burden of pollen allergy suf- calculated as SPIn (formerly also named SPI = seasonal ferers [14]. As such it is a tool with a direct connection to pollen index) and the defined pollen season refers to the those who suffer from pollen allergies and provides crucial MPS (main pollen season in [21]). information. The SLI proved to be a robust index through different pollen seasons or years [15] and is applicable Symptom data even on a local level [16]. Pollen measurements are The symptom data used for this study was retrieved needed for the definition of a pollen season and for com- from the Patient’s Hayfever Diary (www.pollendiary.com) parisons with the SLI (e.g. [15, 17]). Thus, the SLI as a which is available as well as a mobile application number on its own is independent of pollen data and cal- (“Pollen”). Data for analyses were downloaded on 29th culated only on symptom data, but is in need for pollen of November 2017. Data were selected and retrieved data when a pollen season has to be defined as a pre- only for districts in Vienna (Austria) for the years 2013 requisite for a mean SLI in a certain period. Other issues until 2017 since user numbers are especially high from concern the misconception that SLI and pollen data do 2013 onwards. More than 50 entries per season allow a not correlate. This point will be clarified in detail herein. statistical analysis of a pollen season in terms of Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 3 of 8 crowd-sourced symptom data [17]. This recommendation a more detailed spatio-temporal analysis of crowd-sourced was far exceeded by daily mean entries of above 25 (see symptom data. Area classification). We applied the symptom load index In addition, the region “Pannonian Lowlands” comprised (SLI) described in [14], which comprises in summary the of Vienna, parts of Lower Austria and Burgenland was average of entered symptoms (eyes, nose, lungs) and included to examine the SLI pattern for a larger area in medication intake within a defined time period of a de- comparison. fined user sample. Users are granted anonymity and the services concerned were adapted to fulfill the latest EU Statistics regulation on data privacy (Regulation EU 2016/679) The graphs and computations were performed using the resulting in pooling only the absolute minimum of data statistical software R 3.4.3 [23]. We used descriptive sta- such as symptom data and an email address in this case. tistics to explore the behavior of the SLI in comparison Thus, personal data such as birthday or medical conditions to the daily pollen concentrations. Data were analyzed are were not collected. Personal and symptom datasets are for different regions (see Area classification). The rela- saved on separate servers to grant high security to avoid tionship of the SLI and the square root of the daily any unauthorized connection between them. pollen concentrations is figured in a scatter plot (Fig. 1a). A linear regression line (grey) and a locally weighted Area classification smoothing (LOESS) curve (black) were calculated. The Vienna was divided into three areas to explore the vari- linear regression line models the relationship of a ation of the SLI within a city. The division was based on dependent variable (SLI) and the explanatory variable population density and environmental factors. We estab- (daily pollen concentrations) via the formula y = a + bx. lished three areas: (1) “Vienna Center” (zip codes: 1010, In addition, the correlation factor of the SLI including 1040, 1050, 1060, 1070, 1080, 1090, 1100, 1120, 1150), the p-values and significance levels as well as the daily (2) “Vienna East” (zip codes: 1020, 1030, 1110, 1200, 1210, pollen concentrations in the respective pollen season 1220) and (3) “Vienna West” (zip codes: 1130, 1140, 1160, were calculated (see Table 2). LOESS [24] is a modern, 1170, 1180, 1190, 1230). Each area within Vienna thus in- non-parametric regression method and was chosen due cludes dense and less dense populated districts (see [22]). to its advantages as a flexible and simple tool applied to This was cross-checked with the sample size of the PHD complex processes without theoretical model. Besides, and resulted in a very similar sample size for each of the LOESS finds the curve of best fit without assuming the three areas throughout the different pollen seasons and data have to fit to a certain distribution shape. years. A comparable sample size is important to exclude The temporal development of the SLI and the daily possible influences of an unreliable small dataset. As an pollen concentration for specific pollen seasons were ex- example, the mean samples sizes for the year 2017 are plored in a line plot (Fig. 1b). The birch and the grass noted here for “Vienna Center” (alder/birch/grass pollen pollen season 2014 was chosen as an example. The SLI season: 26/43/23), “Vienna East” (alder/birch/grass pollen and the square root of the daily pollen concentrations season: 36/60/30) and “Vienna West” (alder/birch/grass were scaled to mean zero and standard deviation 1 to be pollen season: 38/50/28). The three areas correspond also comparable. The line chart was chosen to visualize the to areas of a different vegetation influence: “Vienna behavior of the SLI and the daily pollen concentrations Center” can be characterized as the most urbanized area over a certain time period (pollen season). with mostly parks and alley trees as green areas. This area is characterized by less grass vegetation (especially natural Results meadows) and fallow lands (preferred by Artemisia and The SLIs attained in different areas within Vienna show Ambrosia). “Vienna East” is influenced by the river little variation in different years, but the same pollen Danube and the Danube valley. This area is characterized season (Table 1). In general, the linear relationships and by alluvial vegetation and taxa preferring proximity to thus the correlation is very similar between the different water, e.g., species of the genus Alnus, Populus, Salix and areas in Vienna and the Pannonian lowlands and is Fraxinus. Grass areas and fallow lands are frequent due to observed due to the parallelism of the linear regression the agricultural influence in certain districts (1110, 1210 lines (Fig. 1a). Deviations from linearity are shown in the and 1220). “Vienna West” is influenced by the Vienna for- LOESS curve. While the SLI is almost perfectly linear ests. Therefore, natural meadows can be found in this area over the whole spectrum of observed grass pollen con- as well as forest elements, such as Corylus, Carpinus, centrations, the SLI deviates from the general trend for Fagus, Quercus and Pinus. Fallow lands occur in this area high birch pollen concentrations (Fig. 1a). The temporal due to the industrial area in certain districts (1130 and development of the SLI and the square root of the pollen 1230). Vienna was chosen for this study due to the high concentrations throughout a pollen season is shown in user numbers making it the only location so far possible for the line plots (Fig. 1b) and supports the relationship of Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 4 of 8 A B Fig. 1 Behavior of the SLI data and the pollen data as square root of the respective aeroallergen is displayed. a, Linear regression (grey) and LOESS curve (black) in the three areas within Vienna and the Pannonian Lowlands for the respective pollen seasons (2013–2017) of alder (upper), birch (middle) and grass (lower). Note the similar behavior of the LOESS for all areas/regions within the same pollen season. b, Temporal development of square root of the daily pollen concentrations (grey) and the respective SLI data (black). The birch pollen season 2014 (upper) and the grass pollen season 2014 (lower) were chosen as examples to show the similar progress of the datasets. Peaks of the pollen data are reflected in the SLI data SLI and pollen concentrations (Table 2). The three season when compared to the birch and grass pollen pollen seasons analyzed are characterized in the follow- season. The pattern of increase/decrease in SPIn of alder ing separately. pollen is consistent from 2014 until 2016, but is inter- rupted in 2017 (Table 1). Daily alder pollen concentra- Alder pollen seasons tions correlate with daily symptom data during the alder The mean SLI during the alder pollen season varies from pollen season. The correlation factors attain values a minimum of 2.4 to a maximum of 4.1 from 2013 to above 50% for the areas within Vienna up to around 2017. The three areas within Vienna show a variation 70% for the Pannonian Lowlands (Table 2). The behavior from 0.2 to 0.7 between each other in the observed time of the SLI during the alder pollen seasons 2013–2017 period. The mean SLI is lowest during the alder pollen shows the same progress within the three Vienna areas Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 5 of 8 Table 1 SLI values in different years (2013–2017) through Vienna ranges from 0.1 to 0.7 in the observed time different pollen seasons (alder, birch and grass) are shown for period. The mean SLI is highest during the birch pollen the three selected areas within Vienna (“Vienna East”, “Vienna season when compared with the alder and grass pollen Center” and “Vienna West”) in combination with their respective season. The pattern of increase/decrease in SPIn of birch average for Vienna pollen is inconsistent over most seasons (2014–2016) Region Year SLI/SPIn alder SLI/SPIn birch SLI/SPIn grasses and only consistent in 2017 (Table 1). Daily birch pollen Vienna Average 2013 2.4/2679 6.4/4517 3.8/3107 concentrations correlate with daily symptom data during Vienna East 2.3 6.2 3.9 the birch pollen season. The correlation factors attain values above 60% (Table 2). The behavior of the SLI dur- Vienna Center 2.4 6.4 3.7 ing the birch pollen seasons 2013–2017 shows the same Vienna West 2.6 6.6 3.8 progress in the three Vienna areas and the Pannonian Vienna Average 2014 ⇑ 3.6/2883⇑⇓ 5.1/6916⇑⇑ 4.1/2031 ⇓ lowlands (Fig. 1a). The progress within a single birch Vienna East 3.4 4.7 4.1 pollen season (Fig. 1b) shows that the peaks of the Vienna Center 3.5 5.3 4.0 square root of the daily mean pollen concentrations is Vienna West 4.0 5.3 4.1 mirrored in the SLI. The three areas in Vienna and the Pannonian lowlands show overall the same pattern as Vienna Average 2015 ⇓ 2.8/760⇓⇑ 5.6/4992⇓⇑ 4.3/3264 ⇑ well. Line plots of birch pollen seasons provided similar Vienna East 2.6 5.6 4.3 results. Vienna Center 2.5 5.6 4.2 Vienna West 3.2 5.5 4.3 Grass pollen seasons Vienna Average 2016 ⇑ 3.9/2685⇑⇓ 5.2/9233⇑⇔ 4.2/2522 ⇓ The mean SLI during the grass pollen season varies from Vienna East 3.7 5.1 4.2 a minimum of 3.8 to a maximum of 4.3 from 2013 to 2017. The three areas within Vienna show a variation Vienna Center 4.0 5.3 4.3 from 0.1 to 0.5 in the observed time period and thus the Vienna West 4.0 5.2 4.0 lowest variation. The mean SLI attains a value between Vienna Average 2017 ⇑ 4.1/1477⇓⇓ 4.8/3497⇓⇓ 4.0/2517 ⇔ those of the alder and the birch pollen season. The pattern Vienna East 4.0 4.8 3.9 of increase/decrease in SPIn of grass pollen is mostly in- Vienna Center 3.9 4.3 3.9 consistent over most seasons (2014; 2016–2017) with the Vienna West 4.3 5.0 4.4 exception of 2015 (Table 1). Daily grass pollen concentra- tions correlate with daily symptom data in the grass pollen The SPIn was calculated for the observed time period and is shown for Vienna with the average SLI. The arrows indicate the change of the SLI (left) and the season. The correlation factors are comparable to those SPIn (right) compared to the previous year. Note the inconsistency between during the birch pollen season and attain as well values the increase/decrease of SLI and SPIn in most of the cases, especially in the birch and the grass pollen seasons above 60% (Table 2). The LOESS for the grass pollen sea- sons 2013–2017 shows the same pattern within the three and the Pannonian lowlands (see LOESS curve in areas of Vienna and the Pannonian lowlands (Fig. 1a). The Fig. 1a). The data scatters more than for the birch and LOESS curve is here most similar to the regression line the grass pollen season and the LOESS curve deviates when compared with the alder and the birch pollen sea- from the regression line in some extent. sons. The progress within a single grass pollen season (Fig. 1b) shows the reflection of the peaks in the square Birch pollen seasons root of the daily mean pollen concentrations in the SLI The mean SLI during the birch pollen season varies even better than for the birch pollen season. Only from a minimum of 4.8 to a maximum of 6.4 from 2013 minor differences between the three regions in Vienna to 2017. The variation between the three areas within and the Pannonian lowlands can be observed (e.g. the Table 2 The correlation factors including p-values (and significance levels) between the daily SLI values of the defined pollen season (alder, birch and grass) in the observed time period (2013–2017) and the daily mean pollen concentrations are displayed for each of the areas in Vienna as well as for a whole region (“Pannonian Lowlands”) in comparison Corr.-Factor sqrt/p-value SLI alder SLI birch SLI grasses − 13 − 8 − 28 Pannonian Lowlands 69.1/2.02e *** 69.3/1.90e *** 69.1/1.85e *** − 12 − 7 − 21 Vienna East 59.4/9.15e *** 66.1/1.68e *** 62.2/1.77e *** − 10 − 7 − 21 Vienna Center 56.4/1.71e *** 65.9/2.45e *** 61.5/8.58e *** − 10 − 8 − 21 Vienna West 55.3/5.43e *** 67.9/4.96e *** 61.9/1.47e *** Significance codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 “1 Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 6 of 8 double-headed peak during the main peak of the grass for the birch pollen season. This is in accordance with pollen season for “Vienna Center” compared to “Vienna the highest SLIs found in the birch pollen season, des- East” and “Vienna West”). Line plots of other grass pollen pite the lower frequency of birch pollen allergy. User seasons provided similar results. numbers and thus sample size may be connected with a high burden and triggered by the occurrence of severe Discussion symptoms. This is reflected as well in pollen informa- Daily and seasonal SLIs were analyzed for three different tion consumption (high user numbers and web site pollen seasons and for three different areas in Vienna to visits) during the birch pollen season [6]. examine the relationship of the SLI and pollen concen- trations/SPIn. The SLI data shows a significant correl- Allergen content ation with the daily pollen concentrations in general Significant positive correlations between major aeroaller- (Table 2) and throughout the season (Fig. 1b). This find- gens such as Bet v 1 and Phl p 5 with the SLI were ing has to be emphasized since the SLI has been misin- found for Vienna [26]. It is known that the allergen con- terpreted since its development [14] in the scientific tent may vary from pollen grain to pollen grain [27]. field and is thwarted by the new and further evidence of Pollen allergens can be present in the air also as small its inconsistency as mean value for a pollen season with respirable particles ([28–30]), thus resulting in a differ- the respective SPIn of that pollen season (Table 1). The ent occurrence when compared to pollen concentrations complexity of the results indicate a more or less linear [31]. Allergen content evaluation is highly time and cost relationship of SLI and pollen concentrations/SPIns on a consuming. Additionally, not only one major allergen is daily, but not on a seasonal basis. This can be explained sufficient to monitor and a range of major and minor al- by a variety of factors of importance for the development lergens are of importance for the immune response (e.g. and onset of allergic symptoms. [32, 33]) besides panallergens such as profilins and pol- calcins [34]. Therefore, it remains currently unknown, if Cross-reactivity a large range of allergens analyzed would explain the Cross-reactivities and other allergic triggers will play a role symptom burden. Allergen content in any case is a for individual pollen allergy sufferers. Hazel (Corylus)is major factor influencing the SLI. cross-reactive to alder (Alnus)[2] and its occurrence/ab- sence in different seasons may increase or decrease the Air quality and pollution mean SLI during an alder pollen season. There are several Air quality in general has an impact on the allergenicity candidates for a possible cross-reactivity during the of pollen grains ([35, 36]) and thus the burden of pollen birch pollen season (all flowering times of possible allergy sufferers e.g. the occurrence of atopic diseases cross-reactive aeroallergens are based on unpublished (asthmatic bronchitis, hay fever, eczema, sensitization; long term pollen data from Vienna and personal obser- [37]) and asthma admissions [38]. Allergenicity changes vation in the following): The hazel and alder pollen sea- in birch and grass pollen under air pollution such as ni- son sometimes overlaps with the beginning of the birch trogen dioxide (NO ), ammonia (NH ) and ozone (O ) 2 3 3 pollen season. Beech (Fagus)and oak(Quercus) flower and decreases microbial diversity [36]. Majd et al. [35] later during the birch pollen season and are possible found that air pollution causes the release of pollen pro- triggers for cross-reactivity [2]. Besides, ash (Fraxinus) teins resulting in a higher allergenicity besides a changed flowers about the same time with birch (Betula), but is shape of the pollen and its tectum. Fine particles shall not closely related. It belongs to the olive family (Olea- be mentioned as well, since diesel exhaust carbon parti- ceae) and may cause as well pollen allergy during spring cles are able to bind to major allergens under in vitro time. Poly-sensitized pollen allergy sufferers may react conditions (e.g. Lol p 1; [39]). during the analyzed time period to ash and birch D´Amato et al. [40] summarized the possible relation- pollen. However, the frequency of ash pollen allergy is ships between air pollution and allergens as follows: (1) with 17.7% much lower than birch pollen allergy with modified allergenicity, (2) interaction with microscopic 41.7% in Eastern Austria [25]. A number of plants have allergen-carrying particles reaching to the lower airways, (3) the potential to irritate pollen allergy sufferers during inflammatory effect with increased epithelial permeability the grass pollen season such as plantain (Plantago), and (4) adjuvant immunologic effect on IgE synthesis in dock (Rumex), the nettle family (Urticaceae), the goose- atopic persons. Therefore, it has to be assumed that air pol- foot family (Amaranthaceae) or mercury (Mercurialis) lution could explain observed SLI patterns in addition. [2]. Grass pollen allergy is the most frequent pollen al- lergy in Eastern Austria with a frequency of 56.3% [25]. Remarks and limitations It is noteworthy here to point to the sample size (see The SLI calculations performed herein rely on an excel- Symptom data and Area classification) which is largest lent sample size and allow to split Vienna in three areas Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 7 of 8 to observe possible spatio-temporal differences. The ob- (3) pollen monitoring stations do not have to be too servation period of five years and the comparison with a numerous. Results indicate that a second or more whole bio-geographical region (the Pannonian lowlands) pollen traps in Vienna would not provide any adds to the robustness of the outcomes and allows to benefit, since the SLI patterns are comparable in compare insights in different years and different pollen the whole city of Vienna. A certain density of pollen seasons. A noteworthy impact of the vegetation (see Area monitoring stations across a country/region has of classification) could not be observed in the three areas, course to be guaranteed and depends on vegetation, e.g. the SLI was not highest in “Vienna East”, where the topography and other factors. Danube has a strong influence. The results from [14] Abbreviations could be reproduced (seasonal SLI and SPIn do not in- EAN: European Aeroallergen Network (https://ean.polleninfo.eu/Ean); crease/decrease with each other). This is contrasted by the LOESS: Locally weighted smoothing; PHD: Patient’s Hayfever Diary (www.pollendiary.com); SLI: Symptom load index; SPIn: Seasonal Pollen Integral correlation factors line plots that were produced to show a correlation concerning the daily SLI and daily mean pollen Acknowledgements concentrations (Table 2 and Fig. 1a). Moreover, the results We thank Christoph Jäger for his constant efforts concerning the Pollen Diary. We are grateful to Alexander Kowarik for support in the statistical analyses. do not match with those from Berlin [41] – the only other Furthermore, we are grateful for the input of two anonymous reviewers. study that analyzed the spatio-temporal differences of crowd-sourced symptom data within a city to our know- Availability of data and materials Users were guaranteed anonymity, so now raw data or individual user data ledge. However, the observed differences therein could is available. have occurred due to the much lower sample size in this study in combination with the larger study area. Authors’ contributions Care has to be taken concerning the interpretation of KB drafted the main body of the manuscript, supervised the analyzes and prepared the figures. MK prepared the Tables, took part in the analysis of the the results. Pollen allergy is a complex disease and a range results and evaluated the pollen datasets. MB took part in the analysis of the of key factors might play an important role as discussed results and prepared the raw datasets of pollen and symptom data. UB took concerning cross-reactivities, allergenicity and air pollu- part in drafting the manuscript and supervised the study. All authors designed the study and revised and approved the final version of the tion. Therefore, important limitations are comprised of (1) manuscript. the SLI is calculated for a certain pollen season, but might be caused also by other aeroallergens and (2) users can Ethics approval and consent to participate Only crowd-sourced and anonymized symptom data was used, so no ethics not be characterized directly as patients due to the miss of approval was needed. personal information (although the crowd-sourced nature of the data assures significance; e.g., [15] and works also Consent for publication All authors confirm their consent for publication. for local phenomena e.g., [16]). Competing interests Conclusions Uwe Berger developed the PHD and the “Pollen” App. The services are free and there is no financial interest. The in-depth analysis of daily and seasonal SLIs of Katharina Bastl and Maximilian Kmenta were involved in the development Vienna revealed that the SLI and pollen concentrations and/or improvements of these services. correlate on a daily basis, but not on the seasonal level Markus Berger reports no competing interests. and that the SLI behavior and pattern do not vary on a local level. In fact, the SLI behavior of a large city like Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in Vienna is more or less the same as for a whole region such published maps and institutional affiliations. as the Pannonian lowlands. These results lead to far-reaching consequences: Author details Aerobiology and Pollen Information Research Unit, Department of Oto-Rhino-Laryngology, Medical University of Vienna, Währinger Gürtel (1) a symptom forecast based on crowd-sourced 18-20, 1090 Vienna, Austria. Paracelsus Medizinische Privatuniversität, symptom data could be feasible in the future. Strubergasse 21, 5020 Salzburg, Austria. The development of the SLI could be calculated Received: 19 July 2018 Accepted: 23 August 2018 since the SLI shows a widespread, consistent pattern during the season as shown herein. (2) the overall severity of a season is impacted by References 1. Pawankar R, Canonica GW, Holgate ST, Lockey RF, Blaiss MS. White Book on additional factors than just the major aeroallergen Allergy: Update 2013: WAO World Allergy Organization; 2013. http://www. in the air during this period. The mean SLI during worldallergy.org/wao-white-book-on-allergy. the pollen season usually does not correlation with 2. D’Amato G, Cecchi L, Bonini S, Nunes C, Annesi-Maesano I, Behrendt H, et al. Allergenic pollen and pollen allergy in Europe. Allergy. 2007;62:976–90. the SPIn. This should be considered for a range of https://doi.org/10.1111/j.1398-9995.2007.01393.x. activities such as pollen forecasting and clinical 3. Asher MI, Montefort S, Björksten B, Lai CK, Strachan DP, Weiland SK, et al. studies. Worldwide time trends in the prevalence of symptoms of asthma, allergic Bastl et al. World Allergy Organization Journal (2018) 11:24 Page 8 of 8 rhinoconjunctivitis, and eczema in childhood: ISAAC phases one and three 25. Hemmer W, Schauer U, Trinca AM, Neumann C. Endbericht 2009 zur Studie: repeat multicountry cross-sectional surveys. Lancet. 2006;368:733–43. Prävalenz der Ragweedpollen-Allergie in Ostösterreich. Amt der NÖ https://doi.org/10.1016/S0140-6736(06)69283-0. Landesregierung, Landesamtsdirektion, Abteilung Gebäudeverwaltung, 4. Dorner T, Rieder A, Lawrence K, Kunze M. Österreichischer Allergiebericht. Amtsdruckerei: St. Pölten; 2010. Vienna: Verein Altern mit Zukunft; 2006. 26. Bastl K, Kmenta M, Pessi AM, Prank M, Saarto A, Sofiev M, et al. First comparison of allergen content (Bet v 1 and Phl p 5 measurements) with 5. Kiotseridis H, Cilio CM, Bjermer L, Tunsäter A, Jacobsson H, Dahl A. Grass symptom and pollen data from four European regions during 2009–2011. pollen allergy in children and adolescents-symptoms, health related quality Sci Total Environ. 2016;548–549:229–35. https://doi.org/10.1016/j.scitotenv. of life and the value of pollen prognosis. Clin Transl Allergy. 2013;3:19. 2016.01.014. https://doi.org/10.1186/2045-7022-3-19. 27. Buters JTM, Thibaudon M, Smith M, Kennedy R, Rantio-Lehtimäki A, Albertini 6. Kmenta M, Zetter R, Berger U, Bastl K. Pollen information consumption as an R, et al. Release of bet v 1 from birch pollen from 5 European countries. indicator of pollen allergy burden. Wien Klin Wochenschr. 2015;128(1–2):59–67. Results from the HIALINE study. Atmos Environ. 2012;55:496–505. https://doi.org/10.1007/s00508-015-0855-y. https://doi.org/10.1016/j.atmosenv.2012.01.054. 7. Bastl K, Berger M, Bergmann KC, Kmenta M, Berger U. The medical and 28. Spieksma FT, Emberlin JC, Hjelmroos M, Jäger S, Leuschner RM. scientific responsibility of pollen information services. Wien Klin Atmospheric birch (Betula) pollen in Europe: trends and fluctuations in Wochenschr. 2017;129:70–4. https://doi.org/10.1007/s00508-016-1097-3. annual quantities and the starting dates of the seasons. Grana. 1995;34:51–7. 8. Ellis AK, Ratz JD, Day AH, Day JH. Factors that affect the allergic rhinitis https://doi.org/10.1080/00173139509429033. response to ragweed allergen exposure. Ann Allergy Asthma Immunol. 29. Schäppi G, Supphioglu C, Taylor PE, Knox RB. Concentrations of the major 2010;104(4):293–8. https://doi.org/10.1016/j.anai.2010.01.012. birch tree allergen bet v 1 in pollen and respirable fine particles in the 9. Caillaud DM, Martin S, Segala C, Besancenot JP, Thibaudon M. French atmosphere. J Allergy Clin Immunol. 1997;100:656–61. https://doi.org/10. aerobiology network. Nonlinear short-term effects of airborne Poaceae 1016/S0091-6749(97)70170-2. levels on hay fever symptoms. J Allergy Clin Immunol. 2012;130(3):812–4. 30. Schäppi GF, Taylor PE, Pain MC, Cameron PA, Dent AW, Staff IA, et al. https://doi.org/10.1016/j.jaci.2012.04.034. Concentrations of major grass group 5 allergens in pollen grains and atmo- 10. Connell JT. Quantitative intranasal pollen challenge, II: effect of daily pollen spheric particles: implications for hay fever and allergic asthma sufferers challenge, environmental pollen exposure, and placebo challenge on the nasal sensitized to grass pollen allergens. Clin Exp Allergy. 1999;29(5):633–41. membrane. Allergy. 1986;41:123–39. https://doi.org/10.1016/0021-8707(68)90053-1. https://doi.org/10.1046/j.1365-2222.1999.00567.x. 11. Costa C, Menesatti P, Brighetti MA, Travaglini A, Rimatori V, Di Rienzo BA, et al. 31. Buters JTM, Weichenmeier I, Ochs S, Pusch G, Kreyling W, Boere AJ, et al. Pilot study on the short-term prediction of symptoms in children with hay The allergen bet v1 in fractions of ambient air deviates from birch pollen counts. fever monitored with e-health technology. Eur Ann Allergy Clin Immunol. Allergy. 2010;65(7):850–8. https://doi.org/10.1111/j.1398-9995. 2009.02286.x. 2017;46(6):216–25. 32. Andersson K, Lidholm J. Characteristics and immunobiology of grass pollen 12. Kmenta M, Bastl K, Jäger S, Berger U. Development of personal pollen aller- gens. Int Arch Allergy Immunol. 2003;130(2):87–107. https://doi.org/10. information – the next generation of pollen information and a step forward 1159/ 000069013. for hay-fever sufferers. Int J Biometeorol. 2014;58:1721–6. https://doi.org/10. 33. Karamloo F, Schmitz N, Scheurer S, Foetisch K, Hoffmann A, Haustein D, et 1007/s00484-013-0776-2. al. Molecular cloning and characterization of a birch pollen minor allergen, 13. de Weger LA, Hiemstra PS, Op den Buysch E, van Vliet AJ. Spatiotemporal bet v 5, belonging to a family of isoflavone reductase-related proteins. J monitoring of allergic rhinitis symptoms in the Netherlands using citizen Allergy Clin Immunol. 1999;104(5):991–9. https://doi.org/10.1016/S0091- science. Allergy. 2014;69(8):1085–91. https://doi.org/10.1111/all.12433. 6749(99)70080-1. 14. Bastl K, Kmenta M, Jäger S, Bergmann KC, Berger U. Development of a 34. Hauser M, Roulias A, Ferreira F, Egger M. Panallergens and their impact on symptom load index – enabling temporal and regional pollen season the allergic patient. Allergy Asthma Clin Immunol. 2010;6(1):1. https://doi. comparisons and pointing out the need for personalized pollen information. org/10.1186/1710–1492-6-1. Aerobiologia. 2014;30:296–80. https://doi.org/10.1007/s10453-014-9326-6. 35. Majd A, Chehregani A, Moin M, Gholami M, Kohno S, Nabe T, et al. The 15. Bastl K, Kmenta M, Geller-Bernstein C, Berger U, Jäger S. Can we improve effects of air pollution on structures, proteins and allergenicity of pollen pollen season definitions by using the symptom load index in addition to grains. Aerobiologia. 2004;20:111–8. pollen counts? Environ Pollut. 2015;204:109–16. https://doi.org/10.1016/j. 36. Obersteiner A, Gilles S, Frank U, Beck I, Häring F, Ernst D, et al. Pollen- envpol.2015.04.016. associated microbiome correlates with pollution parameters and the 16. Bastl K, Kmenta M, Berger U. Unusually early flowering of alder in Austria: allergenicity of pollen. PLoS One. 2016;11(2):e0149545. https://doi.org/10. first report of Alnus x spaethii in Austria, combined LM and SEM study of 1371/journal.pone.0149545. alder species and impact on pollen allergy sufferers. Aerobiologia. 2015; 37. Morgenstern V, Zutavern A, Cyrys J, Brockow I, Koletzko S, Krämer U, et al. 31(4):515–21. https://doi.org/10.1007/s10453-015-9383-5. Atopic diseases, allergic sensitization, and exposure to traffic-related air 17. Kmenta M, Bastl K, Berger U, Kramer MF, Heath MD, Pätsi S, et al. The grass pollution in children. Am J Respir Critical Care Med. 2008:1331–7. pollen season 2015: a proof of concept multi-approach study in three https://doi.org/10.1164/rccm.200701-036OC. different European cities. World Allergy Organ J. 2017;10:31. https://doi.org/ 38. Sunyer J, Spix C, Quénel P, Ponce-de-Léon A, Pönka A, Barumandzadeh T, et 10.1186/s40413-017-0163-2. al. Urban air pollution and emergency admissions for asthma in four 18. Hirst JM. An automatic volumetric spore trap. Ann Appl Biol. 1952;39:257–65. European cities: the APHEA project. Thorax. 1997;52(9):760–5. https://doi.org/10.1111/j.1744-7348.1952.tb00904.x. 39. Knox RB, Suphioglu C, Taylor P, Desai R, Watson HC, Peng JL, et al. Major 19. GalánC,Smith M,ThibaudonM, FrenguelliG,Oteros J,GehrigR,etal. grass pollen allergen lol p 1 binds to diesel exhaust particles: Pollen monitoring: minimum requirements and reproducibility of implications for asthma and air pollution. Clin Exp Allergy. 1997;27(3):246–51. analysis. Aerobiologia. 2014;30(4):385–95. https://doi.org/10.1007/ https://doi.org/10.1111/j.1365-2222.1997.tb00702.x. s10453-014-9335-5. 40. D’Amato G, Cecchi L, D’Amato M, Liccardi G. Urban air pollution and 20. Jäger S, Mandroli P, Spieksma F, Emberlin J, Hjelmroos M, Rantio-Lehtimaki climate change as environmental risk factors of respiratory allergy: an A. News. Aerobiologia. 1995;11:69–70. update. J Invest Allerg Clin. 2010;20(2):95–102. 21. Galán C, Ariatti A, Bonini M, Clot B, Crouzy B, Dahl A, et al. Recommended 41. Werchan B, Werchan M, Bergmann KC. Studie zum Vergleich von terminology for aerobiological studies. Aerobiologia. 2017;33:293–5. allergischen Symptomdaten aus einer Pollen-App mit 22. Vienna Government. EinwohnerInnen-Dichte 2017. https://www.wien.gv.at/ Pollenkonzentrationen aus 2014 in Berlin. Umwelt und Gesundheit 02/ stadtentwicklung/grundlagen/stadtforschung/karten/bevoelkerung.html.2017. 2017, Umweltbundesamt. 2017. https://www.umweltbundesamt.de/sites/ Accessed 20 Mar 2018. default/files/medien/1410/publikationen/2017-11-06_umwelt-gesundheit_ 23. R Core Team. R: A language and environment for statistical 02-2017_pollenstudie-2016.pdf. Accessed 26 Mar 2018. computing. R Foundation for Statistical Computing 2017; Vienna, Austria https://www.R-project.org/ 24. Cleveland WS, Grosse E, Shyu WM. Local Polynomial Regression Fitting Local regression models. In: Chambers JM, Tj H, editors. Statistical models in S. California: Wadsworth & Brooks/Cole; 1992. p. 309–76.

Journal

World Allergy Organization Journal – Springer Journals

Published: Oct 16, 2018

There are no references for this article.