Allergic Asthma in the Municipalities of the Palynological Network of the Community of Madrid and Its Interrelation with the Concentration of Tree Pollen and Atmospheric Pollutants

Abstract

Although the benefits of trees in cities are of great variety and value, attention must also be paid to the consequences for public health of the presence of pollen aeroallergens in the atmosphere, which are likely to interact with air pollutants, influencing the alteration of the immune system, facilitating allergic reactions, and enhancing the symptoms of asthmatic patients. This study analyses (using multiple linear regression calculations performed with the data analysis tool Statgraphics Centurion 19) the interaction of the concentration of six types of tree pollen (Cupressaceae, Olea, Platanus, Pinus, Ulmus, and Populus) and six atmospheric pollutants (O₃, PM₁₀ and PM_2.5, NO₂, CO, and SO₂), on asthma care episodes in the Community of Madrid (CAM). In most of the calculated equations, the adjusted R² value is higher than 30%, and in all cases, the P-value of the models obtained is lower than 0.0001. Therefore, almost all models obtained in the study period for asthma are statistically significant. Olea is the pollen type most frequently associated with asthma (followed by Pinus and Populus), in all the years studied. In the same period, O₃ is the most common air pollutant in the equations obtained for asthma. Stronger interrelations with asthma are generally found in more urban municipalities.

1. Introduction

The role played by plant species, and in particular, the trees that populate parks, gardens, and city roads, is crucial in terms of the very diverse and relevant benefits they provide. These may be, among others, of a biochemical and biophysical nature, such as those involved in photosynthetic processes, thanks to the plants’ chlorophyll and the improved thermoregulation brought about by the presence of the tree canopy. There are also social benefits, including an increase in the physical and psychological well-being of citizens, leading to an improvement in public health. Economic benefits include a reduction in morbidity caused by respiratory infections and asthma, as well as contributing to a reduction in hospital admissions due to these pathologies [1,2].

While the benefits of the presence of trees in cities are, therefore, undeniable, attention must also be paid to their less desirable and potentially public health-challenging consequences in terms of prevention and primary, hospital, and emergency care.

Airborne pollen is the main source of aeroallergens and is causing increasing sensitisation problems worldwide. It is also the main source, and the reason for the increasing clinical prevalence, of asthma and rhinoconjunctivitis, allergic pathologies that can affect 15–40% of the European population and are one of the most frequent, growing, and important public health problems in the Community of Madrid (CAM) and worldwide in recent decades [3,4].

According to several studies, tree species emit the highest percentage of pollen into the atmosphere of the CAM (73.5%) of the total spread by all anemophilous flora [5]. In addition, the aerobiological stations of the Palynological Network of the CAM have recorded, in recent decades, a clear evolution towards a higher pollen load in the air of the most abundant tree pollen taxa, Platanus, Cupressaceae, and Quercus (as well as Pinus, Olea, and others), which is explained by changes in the intensity and phenology of pollen dynamics in the air in the cities of the Autonomous Community of Madrid [4].

During the months of May and June, in the CAM, there is usually an epidemic increase in asthma episodes associated with pollen allergy, mainly to Gramineae pollen but also to other pollen types, including trees, such as Olea, Cupressaceae, and Platanus, among others. This epidemic upsurge should be taken into account in order to prevent exacerbations of asthma symptoms at that time and thus to reduce visits to hospital emergency departments due to asthma attacks [6].

Among aeroallergens, the most frequent cause of sensitisation in asthmatic patients of an allergic or extrinsic origin is pollen grains [7]. These wind-borne allergens are capable of provoking IgE-mediated type I hypersensitivity reactions, also known as immediate response, which can occur through the two distinct phases of sensitisation and effector and that occur in patients who are already asthmatic or have other diseases such as atopic dermatitis [8].

In Spain, a notable increase in sensitisation to pollen aeroallergens has been observed in the ten-year period between the 2005 and 2015 Allergológica publications (with 43.8% and 65.5%, respectively, positivity of tests performed with pollens in extrinsic asthma) [7]. Therefore, what happened in this time span seems to indicate that there was an increase in atopic diseases, such as asthma, caused by these airborne inhalant biological particles, taking into account the limitations derived from the methodology used, the criteria for patient inclusion, and the diverse geographical distribution of patients [7].

However, airborne pollen can be considered an underestimated component of air quality, despite its proven relevance for public health, as legislation does not adequately regulate the biological matter of air pollution, while it does so for the inorganic content [3].

These pollutants cause, firstly, due to nitrosative and oxidative stress, an increase in the permeability of the bronchial epithelium, which favours access of these pollutants to the submucosa, where they interact with fibroblasts and smooth muscle cells of the airways [9]. Secondly, they cause a delay in the clearance of inhaled irritants and aeroallergens, attributed to an inhibition of ciliary beating frequency. At the bronchial epithelial cellular level, it has been shown that airborne pollutants can modify several inflammatory parameters [9].

Therefore, air pollution, especially in urban areas, increases the risk of respiratory pathologies. Thus, exposure to SO₂, PM_2.5, and PM₁₀ particles may be associated with a higher frequency of respiratory distress in children [2,10]. And infants with bronchiolitis may be more likely to suffer from asthma if exposed to high levels of O₃, CO, and NO₂. In addition, exposure to NO₂ and CO is associated with an increased prevalence of childhood asthma [2,10].

In cities, the so-called urban heat island (UHI) effect can alter the indices of atmospheric pollutants and lead to increased aeroallergen production and concentration. In addition, pollen types may be different in urban and rural geographical areas [11].

As a single factor, air pollution is already a major public health problem, and there is increasing evidence of its important role in increasing the prevalence of asthma and other allergic diseases [12].

In addition, air pollutants can act as adjuvants, inducing epithelial damage and inflammation and chemically modifying, binding, and altering the immunogenicity of aeroallergen allergenic proteins. They can also enhance oxidative stress and alter the immune system, facilitating allergic reactions [13]. They also facilitate the dispersion of pollen allergens into smaller fractions due to their power to vary the allergenicity of aeroallergens [14].

Air pollution can cause chemical, biological, and physical modification of pollen grains, to a greater or lesser extent depending on the plant species from which they originate. Thus, with regard to their chemical composition, in particular, the alteration can manifest itself in the form of the nitration of proteins; also, with regard to the modification, by oxidative processes, of the properties and chemical composition of proteins, nucleic acids, and lipids, which influences the elongation of the pollen tube and alters germination, the modulation of the immune response in a multitude of molecular and cellular processes is affected [11,15].

In addition, the formation of adjuvant components in the allergic response, with pro-inflammatory properties, can be triggered. As a direct biological effect of atmospheric pollutants on pollen grains, two determining factors for the reproductive function of plants, i.e., the germination rate and viability, can be altered. As for the physical effect, this consists of the degradation of integrity and possible consequent breakdown of the exine, with the consequent release of pollen subparticles containing allergens, whose bioavailability increases as the smaller particles penetrate deeper into the respiratory tract [11,15].

Furthermore, due to the binding of certain air pollutants to airborne pollen grains, asthmatic patients may experience an exacerbation of their symptoms due to an inflammatory response in the airways and an increased susceptibility to aeroallergen exposure [2,9,12]. In addition, the combination of aeroallergens and particulate matter (PM_2.5 and PM₁₀) has the indirect effect of modifying the characteristics of aeroallergens [9].

Exposure of the surface layer of pollen grains to this particulate material in semi-arid, wind-driven cities can alter the physico-chemical characteristics of the exine, cause the release of pollen aeroallergens, and ultimately lead to an increase in sensitisation as well as a pollen allergenic potential [16].

Moreover, the risk of asthma may increase if there is exposure to air pollutants from road traffic, and the closer a person is to the road, the greater the risk. This is also true for the proximity of industrial sites, because these locations are exposed to higher indexes of atmospheric pollutants, including PM₁₀ and PM_2.5 particles as well as SO₂, NO₂, and CO [17,18].

This risk is particularly relevant during early and middle childhood, although more consistent associations between pollution and asthma (as well as reactive airway dysfunction syndrome) may occur in early childhood, given that younger children have less developed airways and are, therefore, more likely to wheeze and that they spend more time in the family home [17,18].

At the same time, it may be encouraging that air pollution can be attenuated more or less significantly, as empirically demonstrated by several studies [19,20]. Thus, suspended particulate matter, in addition to SO₂ and NO₂, can be reduced in the urban atmosphere by means of an appropriate arrangement of trees in green areas, that is, trees that have the capacity to absorb gaseous pollutants through the stomata of the leaves, by means of photosynthetic processes and plant respiration, and plants that have the capacity to intercept particulate pollutants [19,20].

Some of this material is removed by absorption, but most of it is retained on the surface and subsequently removed, after impaction, by washing away by rain or by leaves, branches, and even bark falling to the ground, as occurs in the case of Platanus x hispanica, a tree species that is highly resistant to atmospheric pollution, which undergoes an exfoliation, normally physiological, of the rhytidome of its trunk and branches [19,20].

An argument for the desirability of the appropriate selection of vegetation types as a health-promoting factor is provided by the study by Alcock, I. et al. [21], which examines how the presence of greenery and gardens (which are associated with reductions in asthma hospitalisations only when exposure to air pollutants is lower) may influence hospitalisations in asthma patients, in contrast to areas with a higher density of trees (which show an association with a lower rate of asthma hospitalisations only when air pollution exposures are higher) [21].

It is also interesting to analyse the influence of meteorological parameters on airborne pollen grains, on which they act synergistically, as well as on all types of atmospheric pollutants and aerobiological particles. This influence on all of them develops strongly, in very different procedures and at different time levels, in their production, concentration, presence, diffusion, and bioavailability in the atmosphere, leading in turn to the risk of sensitisation and exacerbation of the clinical symptoms associated with asthma and other allergic pathologies [11,22].

Both meteorological parameters and air pollutants are among the main stress factors for plants, especially in cities, and can impact the atmospheric concentration of allergenic bioparticles and the release of biogenic adjuvants and allergenic proteins [9,13]. In addition, both factors may bias physiological processes and the immune system towards the development of asthma and other allergic diseases, e.g., through alteration of related microbial communities (microbiomes), inflammation and oxidative stress, and disruption of protective epithelial barriers [13].

Thus, given what has been said so far, the increase in cases of extrinsic asthma, as well as other pathologies of the same aetiology, the interaction of atmospheric pollutants with airborne pollen, especially in urban areas, and the connection with the usually changing meteorology may lead to a need for investigations showing the interrelation of airborne pollen concentrations, both with the constituent elements of airborne pollution and with meteorological variables [22].

The most important pollen types that generate pollinosis in Spain are Cupressaceae (in January–March), Platanus (from March to April), and Gramineae and Olea (from April to June) [23]. Therefore, three of the six types present in this investigation are among the pollens with the highest allergenic potential, as is also discussed in another study [2].

2. Materials

This research was carried out in the period 2014–2017 and involved the collection of data on air concentrations of both pollen and pollutants, as well as episodes of asthma care.

While the data on atmospheric pollutants have been provided by the Air Quality Networks of both the Madrid City Council and the Department of Environment, Land Use Planning, and Sustainability of the CAM, the data on atmospheric pollen concentration have been provided by the Palynological Network of the CAM.

These two sets of data are the same as those provided in another study on rhinitis and allergic conjunctivitis [24], which were collected on a daily basis thanks to measurements made by the two types of sampling stations and were analysed and described in detail in another study on the interrelation of the pollen concentration of six tree pollen types and six atmospheric pollutants in the same geographical area of the CAM, which is also specified in the same study [2]. In that same study, in the period 2013–2017, tree biodiversity is also described, and the nomenclature of the two types of sampling stations is specified [2].

Of the 11 stations that make up the Palinocam Network, 3 of them are geographically located in the municipality of Madrid and the other 8 in different towns in the territory of the CAM. In order to carry out this study, the measurement stations of the CAM Air Quality Network closest to the pollen collection stations were selected. Thus, the Palinocam network station in Alcalá de Henares is approximately 985 m from the air pollution measurement station; 693 m separate the two stations in Alcobendas; 634 m in Aranjuez; 1957 m in Coslada; 1436 m in Getafe; 1586 m in Leganés; and 2648 m in Collado Villalba. The Palinocam Network station located in Las Rozas is 8784 m from the existing Air Quality station outside its territory, specifically in the neighbouring municipality of Majadahonda (this is the greatest distance between the two types of collectors in this study). As for the municipality of Madrid, the Palinocam Network stations of “Madrid: Ayuntamiento”, “Madrid Facultad de Farmacia”, and “Madrid Barrio de Salamanca” are located at respective distances of 1718 m, 3200 m, and 3450 m from the Air Quality stations closest to them, called, respectively, “Escuelas Aguirre”, “Farolillo”, and “Casa de Campo” [2].

Figure 1 shows the locations of both types of study stations, at the regional level and at the level of the municipality of Madrid, as reflected in previous research [2].

The procedures for collecting, analysing, and counting atmospheric pollen have been described in detail in other studies [2,5].

As for the CAM Air Quality Network, its current configuration is the result of the zoning study carried out by the CAM Regional Ministry of the Environment and Territorial Planning in 2005 and its subsequent revisions in 2010, 2014, and 2019. This study is carried out in accordance with the national and European legislation in force at any given time. Thus, the 2019 review took into consideration the provisions of Directive 2008/50/EC on Ambient Air Quality and Cleaner Air for Europe and Royal Decree 102/2011, of 28 January, on the Improvement of Air Quality [25].

In 2023, four new stations came into operation, including one located in the municipality of Las Rozas de Madrid, one of the areas of the present study. With their implementation, the objective is achieved that all municipalities in the CAM with a population of more than 75,000 inhabitants will have their own air quality station [25].

Accordingly, the Air Quality Network of the CAM is currently composed of 28 fixed measuring stations, distributed in six homogeneous zones of the Region’s territory. In addition, there is a seventh zone managed by the Madrid City Council, which has its own network of stations distributed throughout the municipality of Madrid [25].

The characterisation of the 28 fixed stations of the Madrid Air Quality Network is as follows: zoning for CO, SO₂, benzene, metals, and benzo (a)pyrene; zoning for particulate matter (PM₁₀ and PM_2.5) and NO₂; zoning for NOx (protection of vegetation and ecosystems); zoning for O₃. With regard to the first three zonings, the 28 stations of the network are classified as follows: 11 Traffic, 2 Industrial, and 15 Background. As for the type of area in relation to O₃, 10 are urban, 12 are suburban, and 6 are rural, of which 4 are remote and 2 are regional [25].

Each fixed air quality station consists externally of a meteorological tower equipped with various sensors (wind direction and speed, temperature, barometric pressure, humidity, solar radiation, and rainfall) as well as intakes for continuous ambient air sampling for the various automatic analysers (thus, depending on the case of the various stations, an intake for PM₁₀ equipment, an intake for PM_2.5 equipment, an intake for Black Carbon equipment, and an intake for gas analysers) [25].

There are automatic analysers inside the fixed measurement stations for the measurement of the different pollutants. In addition to those that are the subject of this study, these are PM₁; NOX (NO and NO₂); benzene, toluene and xylene; hydrocarbons (total, methane and non-methane) and Black Carbon (BCAR, measured at 800 nm, and HN, measured at 370 nm). Not all of them are measured at all stations but only those that are pre-designated. Some of the fixed measurement stations also have different manual equipment: high-volume samplers for the determination of heavy metals and polycyclic aromatic hydrocarbons; samplers for volatile organic compounds; and low-volume samplers for the gravimetric determination of PM₁₀ and PM_2.5 [25].

As a complement to the 28 fixed stations, the Air Quality Network of the Community of Madrid has 2 mobile units, with the same equipment as the most complete fixed station. Finally, as complementary equipment, both the fixed stations and the mobile units are equipped with the following systems: calibration (standard gases, calibrator, and zero air generator); data storage and transmission (computer, modem, etc.); air conditioning for the analysers (air-conditioning and heating) [25].

The data transmitted by the fixed stations and mobile units are received at the Data Processing Centre (DPC), where the measured data are reviewed and validated in accordance with current legislation and made available to the competent administrations and the general public. The DPC consists of a central server and the peripherals necessary for communication with the stations through the data transmission system in each of them [25].

The analytical techniques used by the CAM Air Quality Network to measure the different atmospheric pollutants correspond to the reference methods indicated in current legislation or intercomparison methods. Specifically, the techniques used to measure the air pollutants covered by this study are Chemiluminescence, for NO₂; ultraviolet photometry for O₃; beta absorption for PM_2.5 and PM₁₀ particles; infrared absorption for CO; and ultraviolet fluorescence for SO₂ [25].

Figure 2 shows the delimitations of the 179 municipalities of the CAM, among which those where the 11 pollen concentration measurement stations belonging to the Palynological Network of the CAM (Palinocam Network) are located have been highlighted. Three of them are located in the municipality of Madrid (highlighted in dark blue), and the other eight are located in the municipalities highlighted in white on the map.

The data on census population, territorial extension, population density, as well as the distances of the study municipalities from the capital of Madrid are specified in

Table 1. List of the municipalities that house the stations of the Palinocam Network, with population data and their size and distance from the capital of Madrid.

Municipalities	Census Population (Inhabitants Year 2024)	Surface Area of Municipality (Km2)	Population Density (Inhabitants/Km2)	Distance to the Capital of Madrid (Km)
Alcalá de Henares	199,804	87.7	2278.27	31
Alcobendas	121,446	45.0	2698.80	15
Aranjuez	62,508	201.1	310.83	47
Collado Villalba	67,323	26.5	2540.49	41
Coslada	80,688	12.0	6724.00	8
Getafe	191,560	78.4	2443.37	14
Las Rozas	99,193	58.3	1701.42	19
Leganés	193,934	43.1	4499.63	11
Madrid	3,422,416	607.1	5637.32	0

and have been obtained from the Statistical Institute of the CAM [26]. Thanks to this information, it is possible to classify the different municipalities as urban or rural, in accordance with different criteria established by different Spanish and European Union and Organisation for Economic Co-operation and Development (OECD) bodies.

Thus, in accordance with the provisions of Law 45/2007 on the Sustainable Development of the Rural Environment, a municipality is classified as rural if it has a population density of less than 100 inhabitants per km² and fewer than 30,000 inhabitants [27]. Meanwhile, rural areas are defined as those that are the object of interventions within the framework of the LEADER programme of the Strategic Plan of the Common Agricultural Policy (CAP) and that correspond to municipalities with less than 5000 inhabitants and less than 300 inhabitants per km² [28]. Finally, in accordance with OECD criteria, municipalities with less than 150 inhabitants per km² are included in the rural classification, without any reference to population size [29].

Therefore, according to the information in Table 1, none of the municipalities under study are classified as rural. However, despite the fact that all of them are considered urban, there is a gradation that brings some of them closer to the rural classification than others. In fact, the municipality of Aranjuez, with 62,508 inhabitants in the year 2024, is the least populated, the one with the lowest population density, close to the 300 inhabitants per square kilometre established as a limit by the Strategic Plan of the Common Agricultural Policy.

The census population variable is a first-order magnitude with high explanatory power for the socio-economic situation of the municipalities and is positively related to indicators of urban character and economic activity. Population density is an indicator of urban character, but it is also an indicator of economic activity and is highly correlated with urban and developable land and the price of housing [30].

The distance to the capital, measured by road from the urban centre of the municipality, is a criterion that generally implies greater depopulation and is positively correlated with the rurality variable as well as with ageing and the low socio-economic status of the population, among others [30]. As can be seen in Table 1, Aranjuez is also the furthest from the capital of Madrid, at 47 km. Collado Villalba has a population similar to that of Aranjuez, with 67,323 inhabitants, although they are concentrated in a much smaller area of only 26.5 km² (compared to 201.1 km² for Aranjuez), which is why its population density is more than 8 times higher, at 2540.49 inhabitants/km². Therefore, these two municipalities are the closest to being included in the rurality classification.

At the other extreme is the municipality of Madrid itself, and those closest to it, namely, Coslada, Leganés and Getafe.

Coslada is only 8 km from the capital of Madrid and has the highest population density in Table 1, with 6724 inhabitants/Km², largely due to its small area of 12 km². Leganés and Getafe are 11 and 14 km away from Madrid, respectively, and have very similar populations (slightly more than 190,000 inhabitants in both cases), as well as population densities proportional to the size of their territory. In fact, the population density of Leganés, with 4499.63 inhabitants/Km², is almost double that of Getafe, while the latter has a surface area of 78.4 Km², which almost doubles that of Leganés.

The municipality of Madrid itself is the one with the largest population in this study, with 3,422,416 inhabitants, the comparatively largest territory, with 607.1 Km², and the second highest population density, after Coslada, with 5637.32 inhabitants/Km².

The other 3 municipalities in this study, i.e., Alcalá de Henares, Alcobendas, and Las Rozas, are in an intermediate situation in terms of the variables analysed, with respect to the two extremes mentioned, which imply a greater or lesser scale of rurality.

On the other hand, thanks must be given to the General Subdirectorate of Epidemiology of the Department of Health of the CAM for providing diaries of asthma care episodes, which were collected from the consultations of the Primary Care Health Centres, geographically located in all the 11 areas of the Palinocam Network.

The designation of the Health Centres, as well as the details of their distances to the respective measuring stations of the Palinocam Network are specified in Table 1 of the aforementioned study on rhinitis and allergic conjunctivitis [24].

Asthma episodes of care refer to the number of people who have had an asthma-related consultation each day (recorded in the primary care electronic medical record with the code R96, according to the International Classification of Primary Care, ICPC-2).

Since primary care consultations are only open on working days, episodes of care cannot be collected on public holidays or weekends. From these, it is not possible to make a calculation of clinical prevalence, although an approximation of clinical prevalence can be made. This is because these episodes do not imply the existence of extrinsic asthma symptoms. Furthermore, they do not include the aetiology or confirmation of the pathology by allergy testing.

The total assigned population of each of the 89 Health Centres included in this study, and within the period 2014–2017, has been consulted in the 5th and 6th Reports of the Observatorio de Resultados del Servicio Madrileño de Salud [31]. Based on this information, we calculated the weighted average of the episodes of care, in order to carry out an accurate analysis of the health data provided by the Regional Ministry of Health of the CAM. For this purpose, a weight was assigned, which was calculated by dividing the total assigned population (comprising all age groups) of each Health Centre by the sum of the populations of all the Health Centres belonging to each of the 11 Palinocam Network study areas.

Both the population assigned to each Primary Care Centre and the weights calculated are included in

Table 2. List of the assigned populations (registered in the Observatorio de Resultados del Servicio Madrileño de Salud), in each of the study years, to each of the Health Centres of the Community of Madrid included in the research, together with the assigned weights (in red), according to the specified criteria. The total values are shown, shaded in light blue, as a result of the sum of the assigned population of each of the Primary Care Centres (PCC) in each of the 11 study areas, as well as the weights calculated.

PCC Alcalá de Henares	2014	2015	2016	2017	PCC Alcobendas	2014	2015	2016	2017
Carmen Calzado	15,507 0.078	15,605 0.078	15,842 0.078	16,002 0.079	La Chopera	30,050 0.270	30,335 0.269	30,726 0.267	31,009 0.265
Puerta de Madrid	13,969 0.070	13,889 0.069	13,880 0.069	13,834 0.068	Miraflores	23,398 0.210	23,149 0.205	23,187 0.202	23,358 0.200
Nuestra Señora del Pilar	19,275 0.097	19,339 0.096	19,417 0.096	19,404 0.095	Marqués de la Valdavia	16,883 0.152	16,738 0.148	16,693 0.145	16,726 0.143
Luis Vives	28,147 0.141	27,964 0.139	27,905 0.138	28,042 0.138	Arroyo de la Vega	21,171 0.190	21,893 0.194	22,631 0.197	23,317 0.199
Manuel Merino	12,162 0.061	12,199 0.061	12,234 0.061	12,320 0.060	Valdelasfuentes	19,773 0.178	20,721 0.184	21,665 0.189	22,508 0.193
Juan de Austria	32,703 0.164	32,683 0.163	32,690 0.162	32,786 0.161	TOTAL	111,275 1	112,836 1	114,902 1	116,918 1
María de Guzmán	21,342 0.107	21,242 0.106	21,326 0.106	21,400 0.105	PCC Aranjuez	2014	2015	2016	2017
Reyes Magos	28,463 0.143	28,714 0.143	28,997 0.144	29,105 0.143	Aranjuez	40,680 0.698	41,041 0.696	41,563 0.696	42,281 0.694
Miguel de Cervantes	22,378 0.112	22,932 0.114	23,617 0.117	24,505 0.120	Las Olivas	17,607 0.302	17,945 0.304	18,183 0.304	18,639 0.306
La Garena	5765 0.029	5981 0.030	6133 0.030	6288 0.031	TOTAL	58,287 1	58,986 1	59,746 1	60,920 1
TOTAL	199,711 1	200,548 1	202,041 1	203,686 1	PCC Madrid Facultad de Farmacia	2014	2015	2016	2017
PCC Madrid Ayuntamiento	2014	2015	2016	2017	Reina Victoria	29,981 0.169	30,378 0.171	31,017 0.172	31,454 0.173
Pacífico	33,858 0.091	33,742 0.090	33,899 0.089	34,099 0.088	Villaamil	22,937 0.130	23,272 0.131	23,837 0.133	24,218 0.133
Adelfas	25,382 0.068	25,815 0.069	26,325 0.069	26,825 0.069	María Auxiliadora	11,344 0.064	11,368 0.064	11,524 0.064	11,699 0.064
Las Cortes	26,994 0.072	27,415 0.073	27,888 0.073	28,233 0.073	Casa de Campo	12,390 0.070	12,542 0.071	12,733 0.071	12,844 0.071
Segovia	21,208 0.057	21,409 0.057	21,820 0.057	22,119 0.057	Argüelles	13,168 0.074	13,285 0.075	13,516 0.075	13,709 0.075
Lavapiés	23,194 0.062	23,814 0.063	24,464 0.064	24,898 0.064	Isla de Oza	20,580 0.116	20,258 0.114	20,340 0.113	20,434 0.112
Alameda	20,134 0.054	20,272 0.054	20,680 0.054	20,987 0.054	Andrés Mellado	22,408 0.127	22,296 0.126	22,371 0.124	22,521 0.124
Paseo Imperial	45,187 0.121	45,517 0.121	46,191 0.121	46,773 0.121	Cea Bermúdez	23,081 0.130	23,005 0.130	23,192 0.129	23,503 0.129
Martín de Vargas	16,961 0.045	16,961 0.045	17,157 0.045	17,465 0.045	Guzmán el Bueno	21,070 0.119	21,012 0.118	21,322 0.119	21,643 0.119
Párroco Julio Morate	20,851 0.056	21,042 0.056	21,367 0.056	21,833 0.056	TOTAL	176,959	177,416	179,852	182,025
Embajadores	19,251 0.052	19,402 0.052	19,372 0.051	19,450 0.050	PCC Collado Villalba	2014	2015	2016	2017
Cáceres	12,936 0.035	13,054 0.035	13,449 0.035	13,641 0.035	Collado Villalba Estación	44,135 0.497	44,214 0.493	44,490 0.488	44,994 0.485
Legazpi	31,035 0.083	31,743 0.084	32,494 0.085	32,961 0.085	Collado Villalba Pueblo	29,633 0.334	30,101 0.336	30,889 0.339	31,526 0.340
Quince de Mayo	15,621 0.042	15,451 0.041	156,05 0.041	15,835 0.041	Sierra de Guadarrama	14,999 0.169	15,339 0.171	15,758 0.173	16,201 0.175
Comillas	22,429 0.060	22,348 0.059	22,236 0.058	22,273 0.058	TOTAL	88,767 1	89,654 1	91,137 1	927,21 1
Las Calesas	28,266 0.076	28,467 0.076	29,000 0.076	29,492 0.076	PCC Coslada	2014	2015	2016	2017
Delicias	9748 0.026	9856 0.026	10,077 0.026	10,293 0.027	Doctor Tamames	22,629 0.258	22,762 0.260	22,816 0.260	22,901 0.261
TOTAL	373,055 1	376,308 1	382,024 1	387,177 1	Jaime Vera Coslada	13,703 0.156	13,529 0.154	13,420 0.153	13,423 0.153
PCC Madrid Barrio de Salamanca	2014	2015	2016	2017	Valleaguado	25,315 0.289	25,261 0.288	25,251 0.288	25,128 0.286
Ibiza	33,226 0.071	33,306 0.071	33,583 0.071	33,683 0.070	Ciudad San Pablo	13,026 0.149	13,036 0.149	12,946 0.147	12,877 0.147
Baviera	14,332 0.031	14,288 0.030	14,380 0.030	14,375 0.030	El Puerto	12,915 0.147	13,053 0.149	13,345 0.152	13,513 0.154
Goya	58,725 0.126	58,754 0.125	59,211 0.125	59,743 0.124	TOTAL	87,588 1	87,641 1	87,778 1	878,42 1
Montesa	24,112 0.052	24,616 0.052	25,294 0.053	25,780 0.054	PCC Getafe	2014	2015	2016	2017
Castelló	20,808 0.045	21,218 0.045	21,573 0.045	21,849 0.045	Juan de la Cierva	30,263 0.171	30,599 0.171	30,914 0.170	31,174 0.169
Lagasca	17,186 0.037	17,356 0.037	17,541 0.037	17,806 0.037	Las Margaritas	24,595 0.139	25,080 0.140	25,542 0.140	26,061 0.141
Londres	12,210 0.026	12,376 0.026	12,790 0.027	13,286 0.028	El Greco	21,408 0.121	21,331 0.119	21,461 0.118	21,554 0.117
Príncipe de Vergara	9337 0.020	9426 0.020	9541 0.020	9681 0.020	Las Ciudades	17,586 0.100	18,044 0.101	18,493 0.102	18,843 0.102
Prosperidad	18,449 0.040	18,676 0.040	18,845 0.040	19,015 0.040	Sector III	25,971 0.147	26,649 0.149	27,393 0.150	27,785 0.150
Santa Hortensia	16,661 0.036	16,771 0.036	17,026 0.036	17,300 0.036	El Bercial	12,477 0.071	12,978 0.072	13,495 0.074	13,891 0.075
Ciudad Jardín	18,036 0.039	18,133 0.039	18,421 0.039	18,623 0.039	Sánchez Morate	22,553 0.128	22,473 0.125	22,585 0.124	22,662 0.123
Segre	26,536 0.057	26,534 0.057	26,980 0.057	27,532 0.057	Getafe Norte	13,146 0.074	13,403 0.075	13,606 0.075	13,884 0.075
Potosí	26,798 0.058	27,549 0.059	28,094 0.059	28,651 0.060	Perales del Río	8527 0.048	8599 0.048	8706 0.048	8826 0.048
Daroca	52,770 0.113	53,174 0.113	53,685 0.113	54,046 0.113	TOTAL	176,526 1	179,156 1	182,195 1	184,680 1
Canal de Panamá	30,337 0.065	30,149 0.064	30,201 0.064	30,048 0.063	PCC Leganés	2014	2015	2016	2017
Espronceda	37,800 0.081	38,364 0.082	38,831 0.082	39,280 0.082	Huerta de los Frailes	12,638 0.067	12,863 0.068	13,189 0.069	13,530 0.070
Eloy Gonzalo	33,508 0.072	33,809 0.072	34,240 0.072	34,482 0.072	María Jesús Hereza	30,984 0.163	31,596 0.166	32,356 0.169	32,868 0.170
Justicia	14,281 0.031	14,500 0.031	14,735 0.031	15,149 0.032	Santa Isabel	33,042 0.174	33,099 0.174	33,343 0.174	33,497 0.173
TOTAL	465,112 1	468,999 1	474,971 1	480,329 1	M. Ángeles López Gómez	25,671 0.135	25,585 0.134	25,547 0.133	25,602 0.132
PCC Las Rozas	2014	2015	2016	2017	Jaime Vera	20,922 0.110	20,625 0.108	20,499 0.107	20,373 0.105
Las Rozas—El Abajón	44,103 0.492	44,731 0.492	45,595 0.493	46,640 0.497	María Montessori	15,147 0.080	15,131 0.080	15,095 0.079	15,086 0.078
Monterrozas	45,500 0.508	46,096 0.508	46,855 0.507	47,145 0.503	Marie Curie	12,685 0.067	12,769 0.067	12,903 0.067	13,113 0.068
TOTAL	89,603 1	90,827 1	92,450 1	93,785 1	Dr. Mendiguchía Carriche	25,081 0.132	25,004 0.131	25,151 0.131	25,281 0.131
					Leganés Norte	13,424 0.071	13,628 0.072	13,770 0.072	13,941 0.072
					TOTAL	189,594 1	190,300 1	191,853 1	193,291 1

.

Data on Asthma

The data on asthma are followed by a description of asthma, which is analysed in this study. Data are provided on the prevalence of this pathology, which are intended to provide a closer look at the reality of its influence on the Spanish population and, more specifically, where information is available, on the population of the CAM.

Asthma is a lung disease, which causes reversible obstruction of the airways (not only affecting the bronchial tubes but also the entire respiratory system), inflammation, and hypersensitivity to a series of stimuli. It is a systemic pathology, i.e., it affects the whole organism; in fact, it produces inflammatory mediators that can be measured in blood and breath. Its origin is due to several factors: hereditary, environmental, infectious, allergenic, socio-economic, and psychosocial. It can be stated that this disease is suffered by people with a genetic predisposition, under the influence of various environmental factors [5,32,33]. In addition, there is much evidence pointing to an increase in asthma indicators such as primary care demand, hospital admissions and prevalence [5,34,35].

This condition is often associated with other allergic diseases, such as allergic rhinitis (hay fever), atopic dermatitis, and eczema. In fact, one of the possible causes of the increase in asthma cases in the population in recent decades is the increase in allergic diseases. Thus, there is an overlap between asthma and allergic rhinitis of 50–83% and a 30–35% overlap between asthma and atopic dermatitis [5,36,37,38].

There is a functional complementarity between the lungs and the nasal airways, together with the paranasal sinuses (the latter two components of the upper airways are closely integrated with each other within the respiratory tract). The mucosa lining the nasal area is similar to the bronchial mucosa, so most asthmatic patients also have rhinitis, and the existence of rhinitis usually leads to an increase in hospital admissions and emergency visits for attacks due to asthma exacerbation [39].

In 2013, the prevalence of current asthma, in the CAM, was 6.3% (compared to 3.6% in the period 1996–1997), that of cumulative asthma was 13.5% (compared to 7.8% in the period 1996–1997), and that of asthmatic crisis in 2013 stood at 4.1% (compared to 2.0% in the period 1996–1997), for a population aged 18 to 64 years. Over the 18 years of the SIVFRENT-A (Sistema de Vigilancia de Factores de Riesgo de Enfermedades No Transmisibles de la CAM para Adultos) survey, in the study area, there has been an upward trend in the prevalence of self-perceived (or self-referred) asthma, defined as current asthma, cumulative asthma, and asthma attacks in the last year. This disease consumes approximately 2% of public health resources in the CAM, 70% of which are attributed to the severity and poor control of asthma. All this reveals the importance of epidemiological surveillance, aimed at improving prevention, reducing emergencies and hospital admissions, as well as deaths in patients with asthma [40].

In the prevalence survey carried out in the CAM in 1993, 62.5% of people with a history of asthma also suffer from spring rhinoconjunctivitis, and 27.5% of non-asthmatic people suffer from hay fever [5].

Based on the data obtained from the research carried out by Alergológica 2015 [7], the prevalence of the disease in the CAM stood at 24.2%, while in Alergológica 2005, it was 26.6%, which would imply a slight decrease in the prevalence of the disease in that 10-year period [7]. It is true that this study was carried out on patients attending Allergology consultations, so it is understood that, in the general population, the expected frequency of asthma will be much lower. This study shows that 63.2% of asthma patients in Spain live in urban areas, 17.8% in semi-urban areas, and 19.0% in rural areas. Also, at the national level, with regard to family history of atopic diseases, 39.9% of patients report having a first-degree relative with asthma, 46.8% with rhinitis, 18.8% with conjunctivitis, and 5.2% with atopic dermatitis [7].

According to the Rackeman classification, 82.2% of patients were diagnosed with extrinsic (allergic) asthma and 16.8% with intrinsic (non-allergic) asthma. The aeroallergens to which these patients are sensitised are mainly pollens (65.6%), followed by dust mites (46.5%), animal epithelia (21.3%), and fungal spores (10.1%). Specifically, in terms of the types of pollen that cause the greatest sensitisation of these patients seen in Allergology clinics in Spain, Gramineae predominates with 42.1%, closely followed by Olea with 36.9%, Cupressaceae with 12.6%, Salsola Kali with 7.3%, and Platanus with 7.0% so that, apart from herbaceous plants, it is the types of pollen belonging to tree species that cause the greatest allergic sensitization [7].

At the level of the CAM, the following stand out: Gramineae, with a 61% sensitisation frequency, Cupressaceae, with a 26% sensitisation frequency, and Platanus, with a 19% sensitisation frequency. Surprisingly, no information is provided on Olea in the study carried out by Alergológica 2015, although in the 2005 edition, Olea accounted for 34.7% of allergic sensitisation, compared to Gramineae, with 50.7%, Cupressaceae, with 24.7%, and Platanus, with 17.3%. With regard to diseases associated with asthma, the frequency of rhinitis, according to Allergológica 2015, at the national level, is 80.6% (and, specifically, it is more frequently associated with extrinsic asthma (87.2%) than with intrinsic asthma (12.1%) [7].

3. Methods

Information on the dependent variable, “asthma care episodes”, as well as on the two independent variables, “airborne concentration of pollen grains” and “concentration of air pollutants”, is composed of daily data. The data analysis tool Statgraphics Centurion 19 has been used to find the level of interrelationship between the predictor or explanatory variables and the response variable.

Specifically, multiple linear regression calculations have been performed, given the changing number of independent variables included in the adjusted models, due to the lack of concentration data of several atmospheric pollutants, at different stations of the Air Quality Networks and in the different years of study, as reflected and specified in other research [24].

This variation is detailed and justified in the aforementioned study on allergic rhinitis and conjunctivitis [24], since, as previously mentioned, all the information on these variables is common in both investigations. Likewise, in order to compare the calculated models, which present different numbers of independent variables, the R² statistic adjusted to the degrees of freedom has been used.

Note that values of the coefficient of determination between 0.3 and 0.7 (−0.3 and −0.7) express a moderate positive (or negative) linear relationship via a linear fuzzy-firm rule, and values between 0.7 and 1.0 (−0.7 and −1.0) reflect a strong positive (or negative) linear relationship via a linear firm rule [41]. Numerous studies have used the value of 0.3 as a threshold for a significant correlation. For example, with environmental variables [22,42] or health-related variables [43,44].

The procedures used for the calculation of the multiple linear regression models are the same as those executed in the case of the aforementioned study, in which all the details are described [24]. In summary, the degree to which the adjusted models express the variability of the dependent variable has been checked by the ordinary least squares method, thanks to the value of the coefficient of determination, R² adjusted to the degrees of freedom, as well as through the level of significance, P-value, of both the model (by means of Fisher’s F-ratio) and the explanatory variables (thanks to Student’s t statistic).

And, subsequently, the independent variables that do not appropriately explain the variability of the response variable have been discarded, by means of the method of the backward stepwise method.

The procedure for eliminating the rows of data corresponding to studentized residuals of more than three standard deviations, until achieving the best possible fit of the models, is also the same [24].

Table A1. Results of the calculations of the multiple linear regression fits performed for asthma (with CIAP-2 code R96), in each of the 11 areas of the Palinocam Network, which constitute the geographical reference of this study, and in the 4 years of study. The explanatory variables (air pollutants and pollen types) included in the different models are also shown. Statistically significant results are shaded in blue.

ASTHMA (R96)	YEAR 2014			YEAR 2015			YEAR 2016			YEAR 2017
Alcalá de Henares	42.0712 231/232 0.0000	O3 NO2 CO SO2	Olea Pinus Populus	47.5771 243/245 0.0000	O3 NO2 PM10 SO2	Olea Pinus Populus	30.8708 241/243 0.0000	O3 PM10 CO	Olea Pinus	37.8836 242/243 0.0000	O3 PM10 SO2	Cupressac. Olea Pinus
Alcobendas	29.6421 211/212 0.0000	O3 PM10 SO2	Olea Pinus Populus	42.1048 243/245 0.0000	NO2 PM10	Olea Pinus Populus	24.2444 241/243 0.0000	O3	Olea Pinus Ulmus	21.9464 230/232 0.0000	O3	Cupressac. Pinus
Aranjuez	28.8428 232/236 0.0000	O3 PM10	Olea Pinus Platanus Populus Ulmus	35.5350 239/242 0.0000	NO2 PM10	Cupressac. Olea	25.8380 225 0.0000	O3 PM10	Olea Pinus	22.7937 237/238 0.0000	O3 PM10	Pinus Populus Ulmus
Coslada	42.3669 237/242 0.0000	O3 PM10	Olea Platanus Populus	31.6809 231/233 0.0000	O3 NO2	Olea Pinus Populus	17.9732 207/209 0.0000	O3 PM10	Populus Ulmus	18.2183 230 0.0000	O3	Pinus Populus
Madrid: Barrio de Salamanca	49.1694 245/247 0.0000	O3 PM10 CO	Olea Pinus Platanus Populus	50.8117 245/247 0.0000	O3 NO2 CO SO2	Olea Pinus Populus	52.2529 245/247 0.0000	O3 PM2.5 SO2	Olea Pinus	31.9710 244 0.0000	CO SO2	Cupressac. Pinus Populus
Madrid: Ayunta- miento	59.8843 225/231 0.0000	PM10 CO SO2	Olea Pinus Platanus Populus	43.6201 247 0.0000	O3 PM10 SO2	Olea Pinus Populus	55.0366 237/242 0.0000	O3 PM10 CO SO2	Olea Pinus Ulmus	40.4575 213 0.0000	NO2 CO SO2	Olea Populus
Madrid: Facultad de Farmacia	50.6526 242/246 0.0000	O3 NO2 PM10 CO	Olea Pinus Populus	42.7070 242/245 0.0000	PM10 PM2.5	Cupressac. Olea Pinus	47.1049 235/244 0.0000	O3 PM10 SO2	Cupresac. Olea Pinus Populus	31.6296 243/244 0.0000	O3 PM2.5 CO	Olea Platanus Populus Ulmus
Getafe	51.1448 234/237 0.0000	O3 NO2	Olea Pinus Populus	49.3450 235/238 0.0000	O3 PM10	Cupressac. Olea Pinus Populus	48.0933 220/226 0.0000	O3 PM10	Olea Pinus Ulmus	45.2944 226/231 0.0000	O3 PM10	Cupressac. Olea Pinus Platanus Populus
Leganés	55.3196 240/244 0.0000	O3 NO2	Olea Pinus Platanus Populus	26.7813 227/231 0.0000	O3 PM10	Cupressac. Olea Pinus Platanus	40.8556 226/233 0.0000	O3 PM10	Olea Pinus Populus Ulmus	32.2166 234/235 0.0000	O3 PM10	Cupressac. Olea Pinus Platanus Populus
Las Rozas	43.4263 232/243 0.0000	O3 PM10	Olea Pinus Populus	41.4306 232/243 0.0000	O3	Olea Pinus Populus	36.7175 238/243 0.0000	O3 PM10	Cupresac. Olea	34.8693 233/240 0.0000	O3 NO2	Olea Pinus Populus
Collado Villalba	35.3715 193/194 0.0000	NO2 PM2.5 CO SO2	Cupressa. Pinus	25.4007 219/221 0.0000	O3 PM2.5	Pinus Populus	30.2347 194/195 0.0000	O3 NO2 SO2	Pinus Populus Ulmus	22.7516 238/239 0.0000	SO2	Cupressac. Olea Pinus Platanus Populus

R² adjusted for degrees of freedom. No. observations without outliers/No. total observations. p-value of the model.

in the appendix shows the number of data lines removed from the initially existing ones (the number of observations without outliers relative to the total number of observations).

In almost all the calculations performed, even after the elimination of the aforementioned outliers, the model has included several pollen concentration variables. In fact, they are the ones that, in general, have best explained the resulting model, in relation to the other type of explanatory variable, although this is not surprising, since all the possible data on pollen concentration are available for all the stations and in the 4 years of study, as mentioned above.

The equations of the calculated models are accompanied in

Table A2. Equations of the models calculated by means of multiple linear regression performed for asthma (with CIAP-2 code R96), in each of the 11 areas of the Palinocam Network, which constitute the geographical reference of this study, and in the 4 years of study. Equations with statistically significant results are shaded in light blue.

ASTHMA (R96)	YEAR 2014	YEAR 2015	YEAR 2016	YEAR 2017
Alcalá de Henares	Asthma Alcalá de Henares 2014 = 6.92088 − 0.0369873 × O3 − 0.0360638 × NO2 + 2.0679 × CO + 0.12524 × SO2 + 0.0546737 × Olea + 0.00559378 × Pinus + 0.0065671 × Populus	Asthma Alcalá de Henares 2015 = 9.79148 − 0.0492475 × O3 − 0.047496 × NO2 − 0.0309684 × PM10 + 0.315842 × SO2 + 0.00617946 × Olea + 0.00898512 × Pinus + 0.00458227 × Populus	Asthma Alcalá de Henares 2016 = 9.27489 − 0.0394726 × O3 − 0.0385437 × PM10 − 2.22202 × CO + 0.0105445 × Olea + 0.00427664 × Pinus	Asthma Alcalá de Henares 2017 = 6.79215 − 0.0269283 × O3 − 0.033623 × PM10 + 0.184509 × SO2 + 0.000981516 × Cupressaceae + 0.00371276 × Olea + 0.00609206 × Pinus
Alcobendas	Asthma Alcobendas 2014 = 4.97064 − 0.0201927 × O3 − 0.0262672 × PM10 + 0.202395 × SO2 + 0.0454947 × Olea + 0.0162579 × Pinus + 0.0070044 × Populus	Asthma Alcobendas 2015 = 3.78025 + 0.0419715 × NO2 − 0.0529594 × PM10 + 0.00506312 × Olea + 0.0144111 × Pinus + 0.00871629 × Populus	Asthma Alcobendas 2016 = 4.73145 − 0.0174153 × O3 + 0.0110092 × Olea + 0.00601298 × Pinus + 0.0298582 × Ulmus	Asthma Alcobendas 2017 = 4.67778 − 0.0158027 × O3 + 0.00167108 × Cupressaceae + 0.0328716 × Pinus
Aranjuez	Asthma Aranjuez 2014 = 14.2168 − 0.0518256 × O3 − 0.113777 × PM10 + 0.073782 × Olea + 0.0826072 × Pinus + 0.00538321 × Platanus + 0.070332 × Populus − 0.0163899 × Ulmus	Asthma Aranjuez 2015 = 10.4245 + 0.0936369 × NO2 − 0.158279 × PM10 + 0.00790116 × Cupressaceae + 0.0535221 × Olea	Asthma Aranjuez 2016 = 13.1755 − 0.0355478 × O3 − 0.103823 × PM10 + 0.00614581 × Olea + 0.0541875 × Pinus	Asthma Aranjuez 2017 = 11.921 − 0.0332926 × O3 − 0.0481715 × PM10 + 0.166509 × Pinus + 0.0172682 × Populus + 0.0222031 × Ulmus
Coslada	Asthma Coslada 2014 = 9.21838 − 0.0616481 × O3 − 0.037465 × PM10 + 0.121421 × Olea + 0.00498069 × Platanus + 0.0135303 × Populus	Asthma Coslada 2015 = 9.37723 − 0.050608 × O3 − 0.0216264 × NO2 + 0.0217098 × Olea + 0.00566132 × Pinus + 0.0240948 × Populus	Asthma Coslada 2016 = 7.8874 − 0.0258817 × O3 − 0.0439366 × PM10 + 0.0132475 × Populus + 0.0685507 × Ulmus	Asthma Coslada 2017 = 6.6213 − 0.0280946 × O3 + 0.0782881 × Pinus + 0.00668292 × Populus
Madrid: Barrio de Salamanca	Asthma Dr Subiza 2014 = 7.44622 − 0.0401219 × O3 − 0.0464036 × PM10 + 1.65598 × CO + 0.025379 × Olea + 0.0449605 × Pinus + 0.00146327 × Platanus + 0.0224213 × Populus	Asthma Dr. Subiza 2015 = 6.80539 − 0.0299026 × O3 − 0.0229315 × NO2 + 5.23247 × CO − 0.104121 × SO2 + 0.0121755 × Olea + 0.0196392 × Pinus + 0.0161604 × Populus	Asthma Dr. Subiza 2016 = 5.34591 − 0.0315159 × O3 − 0.0356518 × PM2.5 + 0.178614 × SO2 + 0.00689674 × Olea + 0.00812023 × Pinus	Asthma Dr. Subiza 2017 = 5.70672 + 2.04231 × CO − 0.153161 × SO2 + 0.00249545 × Cupressaceae + 0.0346903 × Pinus + 0.0152588 × Populus
Madrid: Ayunta− miento	Asthma Madrid Ayto. 2014 = 3.66188 − 0.0671616 × PM10 + 1.84835 × CO + 0.31889 × SO2 + 0.0762976 × Olea + 0.0154566 × Pinus + 0.000531998 × Platanus + 0.0121833 × Populus	Asthma Madrid Ayto. 2015 = 6.09369 − 0.0109032 × O3 − 0.0385109 × PM10 + 0.238251 × SO2 + 0.00548418 × Olea + 0.0194436 × Pinus + 0.0118667 × Populus	Asthma Madrid Ayto. 2016 = 6.34469 − 0.0160281 × O3 − 0.0445019 × PM10 + 4.92447 × CO − 0.242165 × SO2 + 0.0175723 × Olea + 0.00670987 × Pinus + 0.0058998 × Ulmus	Asthma Madrid Ayto. 2017 = 4.9028 − 0.0355397 × NO2 + 1.91881 × CO + 0.255125 × SO2 + 0.00628459 × Olea + 0.00603721 × Populus
Madrid: Facultad de Farmacia	Asthma C. Univ. 2014 = 4.87518 − 0.0299854 × O3 − 0.0391158 × NO2 − 0.0173987 × PM10 + 4.94642 × CO + 0.0455906 × Olea + 0.0146536 × Pinus + 0.00171875 × Populus	Asthma C. Univ. 2015 = 3.75549 − 0.0862543 × PM10 + 0.138403 × PM2.5 + 0.00227758 × Cupressaceae + 0.00577979 × Olea + 0.00982304 × Pinus	Asthma C. Univ. 2016 = 5.89473 − 0.0197887 × O3 − 0.021609 × PM10 − 0.345892 × SO2 + 0.00155281 × Cupressaceae + 0.0137827 × Olea + 0.00242409 × Pinus + 0.00174335 × Populus	Asthma C. Univ. 2017 = 3.70066 − 0.00643805 × O3 − 0.0715709 × PM2.5 + 3.58126 × CO + 0.00793662 × Olea + 0.000323787 × Platanus + 0.00343849 × Populus + 0.0142822 × Ulmus
Getafe	Asthma Getafe 2014 = 12.011 − 0.0951763 × O3 − 0.0239882 × NO2 + 0.0975832 × Olea + 0.0984703 × Pinus + 0.110834 × Populus	Asthma Getafe 2015 = 11.3369 − 0.0597383 × O3 − 0.0580158 × PM10 + 0.0114923 × Cupressaceae + 0.0134292 × Olea + 0.0516064 × Pinus + 0.047476 × Populus	Asthma Getafe 2016 = 10.5952 − 0.0633162 × O3 − 0.0562564 × PM10 + 0.0427731 × Olea + 0.0976613 × Pinus + 0.303252 × Ulmus	Asthma Getafe 2017 = 8.69819 − 0.0459755 × O3 − 0.0268317 × PM10 + 0.00488585 × Cupressaceae + 0.00750634 × Olea + 0.128874 × Pinus + 0.000909939 × Platanus + 0.0443417 × Populus
Leganés	Asthma Leganés 2014 = 11.2257 − 0.0843305 × O3 − 0.0190966 × NO2 + 0.165366 × Olea + 0.0993725 × Pinus + 0.0131861 × Platanus + 0.0469193 × Populus	Asthma Leganés 2015 = 11.2505 − 0.0444401 × O3 − 0.0686687 × PM10 + 0.01509 × Cupressaceae + 0.0174911 × Olea + 0.0280848 × Pinus + 0.00160732 × Platanus	Asthma Leganés 2016 = 10.6375 − 0.0615913 × O3 − 0.0546703 × PM10 + 0.0167692 × Olea + 0.0470158 × Pinus + 0.0400987 × Populus + 0.178237 × Ulmus	Asthma Leganés 2017 = 9.09893 − 0.0467436 × O3 − 0.0335593 × PM10 + 0.00557287 × Cupressaceae + 0.0105807 × Olea + 0.083035 × Pinus + 0.00332866 × Platanus + 0.0295104 × Populus
Las Rozas	Asthma Las Rozas 2014 = 12.6376 − 0.0668753 × O3 − 0.0470106 × PM10 + 0.274575 × Olea + 0.0448217 × Pinus + 0.0308489 × Populus	Asthma Las Rozas 2015 = 10.7132 − 0.0380397 × O3 + 0.0462633 × Olea + 0.029355 × Pinus + 0.0388175 × Populus	Asthma Las Rozas 2016 = 12.4233 − 0.0422013 × O3 − 0.0836115 × PM10 + 0.0175418 × Cupressaceae + 0.0894933 × Olea	Asthma Las Rozas 2017 = 12.6105 − 0.0618271 × O3 − 0.0302336 × NO2 + 0.0290067 × Olea + 0.0595487 × Pinus + 0.0357058 × Populus
Collado Villalba	Asthma Villalba 2014 = 4.5792 + 0.0310704 × NO2 − 0.154985 × PM2.5 + 3.68636 × CO + 0.378868 × SO2 + 0.00237073 × Cupresaceae + 0.0128953 × Pinus	Asthma Villalba 2015 = 5.62112 − 0.0157225 × O3 + 0.0891772 × PM2.5 + 0.00516212 × Pinus + 0.0444403 × Populus	Asthma Villalba 2016 = 9.6406 − 0.0544051 × O3 − 0.0636737 × NO2 + 0.490163 × SO2 + 0.0103773 × Pinus + 0.0855361 × Populus + 0.372528 × Ulmus	Asthma Villalba 2017 = 4.36117 + 0.631479 × SO2 − 0.00101313 × Cupressaceae + 0.0158547 × Olea + 0.0135319 × Pinus + 0.0329744 × Platanus + 0.0777578 × Populus

, for the different study years in the period 2014–2017.

4. Results

From the calculations of the multiple linear regression models, the values obtained for each year of the study are shown in Table A1, as well as the corresponding equations, which are shown in Table A2.

Accordingly, the following can be seen:

• The P-values derived from all the equations calculated in this study are all less than 0.0001, so they are all explanatory.
• The relationship between the variables is not only statistically significant, but it is a medium-high correlation. In fact, only 2 of the 11 models calculated for 2014 (in the stations of Alcobendas and Aranjuez) and another 2 (in Leganés and Collado Villalba) for 2015 present adjusted R² values lower than 30%. Again, in the stations of Alcobendas and Aranjuez, the fit coefficient of the calculated regression models presents statistically non-significant values for the years 2016 and 2017. In those same years, there are also non-significant values in the Coslada station and, in addition, in 2017, in Collado Villalba, so in that year there are 4 non-significant adjustments and the 3 indicated in 2016 (Table A1).
• In the equations obtained for asthma, for the year 2014 (Table A1), the highest adjusted R² value is given for the Madrid Ayuntamiento area, with 59.8843% (which represents the highest value in this study); and the lowest (within the models with adjusted R² > 30%) is given for the Collado Villalba station, with 35.3715%. In 2015, the adjusted R² value reaches its highest value, 50.8117%, for the Madrid Barrio de Salamanca station and the lowest, 31.6809%, for the Coslada area (Table A1). In 2016, the best fit is achieved with an adjusted regression coefficient R² value of 55.0366%, for the Madrid Ayuntamiento station, and it is in the Collado Villalba area, with 30.2347%, where the lowest value of this study is given (Table A1). Finally, for 2017, 45.2944% is the highest value of adjusted R², which corresponds to the Getafe area, and 31.6296% for the Madrid Facultad de Farmacia station is the value with the worst fit (Table A1).
• In general, the highest values of adjusted R² are found in the stations located in the most urban municipalities, i.e., the three located in Madrid, as well as Getafe and Leganés, with the exception of Coslada (Table A1). And the lowest values of this coefficient of determination are found in the municipalities closest to the rurality classification, i.e., Aranjuez and Collado Villalba, in accordance with the description in the Materials section.
• In all the equations obtained, there is an interrelation of asthma episodes of care with one or more types of air pollutants and with two or more pollen types.
• Although, in most of the statistically significant models (19 out of 33), pollen types outnumber air pollutants (Table A1), even taking into account the deprivation of the measurement of some atmospheric pollutants already mentioned, there is a coexistence of the two types of independent variables in five calculated equations; even in nine of them, there is a greater number of air pollutants than pollen types.

These exceptional cases are detailed as follows: for the Alcalá de Henares station, in 2014, PM_2.5 particles were not measured, and yet the calculated model has one more air pollution variable, i.e., four, than pollen types. Exactly the same situation is repeated for the same station in 2015 (Table A1).

In the Aranjuez station, also in 2015, the presence of both types of independent variables is equal, even with the lack of registration of PM_2.5, CO, and SO₂ values. In the area of Madrid: Ayuntamiento, the presence of both types of independent variables is equal in 2015 and exceeded the presence of pollen types in 2016 and 2017, the representation of atmospheric pollutants, despite not being quantified in the Farolillo Air Quality station as a PM_2.5 variable. Likewise, in 2016, and for the Las Rozas station, the presence of the two types of variables is equal, although, in the corresponding Air Quality station, the Majadahonda station, the values of the PM_2.5, CO, and SO₂ variables are not recorded. At the Collado Villalba station, the presence of the two types of independent variables is equal, with the number being three, in 2016, despite the lack of recording of CO and PM₁₀. Finally, for the year 2017, at the Alcalá de Henares station, the representation of the air pollution variables is equal, with respect to the types of pollen, although the PM_2.5 variable is not recorded at that station and in that year (Table A1).

Next, the degree of intervention of the explanatory variables in the calculated equations and, consequently, their interaction with the response variable is discussed, based on the Tables and Figure 3 and Figure 4.

As can be seen in

Table 3. Specification of the number of occurrences of the study pollen types in the equations with adjusted R² > 30%, calculated for asthma, in the study years and in the total for the period 2014–2017. The total sum of each pollen type is shown in the right-hand column, shaded in red. The bottom row reports the presence of pollen types for asthma (total asthma).

Types of Pollen	Year 2014	Year 2015	Year 2016	Year 2017	Period 2014–2017
Cupressaceae	1	3	2	4	10
Olea	8	9	7	6	30
Pinus	8	8	7	5	28
Platanus	4	-	-	3	7
Populus	8	7	3	6	24
Ulmus	-	-	4	1	5
Total Asthma	29	27	23	25	104

, the pollen types with the most interventions in the models obtained for the allergic pathology under study are, in 2014, Olea, Pinus, and Populus, on eight occasions, followed by Platanus, on four occasions. In 2015, Olea is the most present pollen type, with nine interventions, followed by Pinus, with eight, and Populus, with seven. In 2016, Olea equals its presence in the models calculated for asthma with Pinus, on seven occasions, followed by Ulmus, on four. In 2017, Olea equals its appearances with Populus, both with six, followed by Pinus, with five, and Cupressaceae, with four. Equally, in the 2014–2017 period as a whole (right-hand column), it can be seen that Olea is the most present in the equations obtained for asthma, on a total of 30 occasions, followed by Pinus, on 28, and Populus, on 24. Cupressaceae appears 10 times in total in the models, followed by Platanus, appearing 7 times, and Ulmus, appearing on only 5 occasions. Therefore, it can be stated that Olea is the pollen type most frequently associated with asthma in all the years studied.

Figure 3 shows graphically the interventions of the study airborne pollen types in the equations with adjusted R² > 30% for asthma, for each study year and for the whole period 2014–2017. As can be seen, Olea and Pinus are the pollen types with generally higher values of such presence.

Since the concentrations of the six pollen types are measured at the 11 stations of the Palinocam Network, they are all equally likely to participate as independent variables in the multiple linear regression equations.

Moreover, it is possible to check, in the last row of Table 3, that the overall relationship of the six pollen types with asthma occurs on 29 occasions in 2014, out of a total of 54 possible, excluding the two statistically non-significant models, corresponding to the stations of Alcobendas and Aranjuez (Table A1), followed by 2015, with 27 (also out of a total of 54 possible, removing the models with adjusted R² <30%, corresponding to the stations of Leganés and Collado Villalba), and 2017, with 25 (out of a total of 42 possible after excluding the non-significant models corresponding to the stations of Alcobendas, Aranjuez, Coslada, and Collado Villalba). Finally, 2016 has the lowest representation in the calculated models, with 23 interventions, out of a total of 48 possible, after discarding the models relating to the stations of Alcobendas, Aranjuez, and Coslada (Table A1). In the total study period, as can be seen in Table 3, the total number of occasions in which the six types of pollen are related to asthma is 104, out of a total of 198 possible, including all the equations with statistically significant results, taking into account the different values of the study period, as already noted.

Although PM₁₀ particles, as well as NO₂ and O₃, are the most frequently quantified air pollutants at the CAM Air Quality stations, they are not always the most strongly associated with asthma episodes of care throughout the study period.

In fact, CO (with five statistically significant appearances in the calculated models) is the second most frequent in 2014, together with NO₂ and PM₁₀, after O₃ (which appears seven times in the equations obtained); SO₂ is the third most frequent in 2016 (with four interventions), after O₃, with eight, and PM₁₀ particles (present six times), and ahead of NO₂, with only one intervention (also surpassed by CO with two). The same occurs in 2017 for SO₂ and CO, which this time, are the second most frequent air pollutants (with three appearances in the equations found for asthma, together with PM10, and after O₃, with five) and are ahead of NO₂, with only two. In 2015, PM₁₀ particles, together with O₃, both with six interventions, are the most frequently related to asthma, followed by NO₂, with five, and by SO₂, present on three occasions (

Table 4. Specification of the number of occurrences of the atmospheric pollutants in the equations with adjusted R² > 30%, calculated for asthma, in the study years and in the total for the period 2014–2017. The total sum of each air pollutant is shown in the right-hand column, shaded in red. The bottom row reports the presence of atmospheric pollutants for asthma (total asthma).

Atmospheric Pollutants	Year 2014	Year 2015	Year 2016	Year 2017	Period 2014–2017
O3	7	6	8	5	26
NO2	5	5	1	2	13
PM10	5	6	6	3	20
PM2.5	1	1	1	1	4
CO	5	1	2	3	11
SO2	3	3	4	3	13
Total Asthma	26	22	22	17	87

).

In the 2014–2017 period as a whole, O₃ is the atmospheric pollutant most present in the models calculated for asthma, with its intervention on 26 occasions, followed by PM₁₀, on 20 occasions. In third place are NO₂ and SO₂, which are equal in terms of their presence on 13 occasions, so once again, there is an exception to the aforementioned rule of a greater hypothetical presence of the most measured atmospheric pollutants in the Air Quality stations of the CAM and the City Council of the Spanish capital.

If we look at the last row of Table 4 (Total Asthma), it can be checked that, in 2014, all six air pollution variables are present in 26 models calculated for asthma. In 2015 and 2016, all six air pollution variables are present in 22 models calculated for asthma and in 17, in 2017 (taking into account the exclusions of statistically non-significant models, discussed in the case of pollen types and referring to Table A1). This means a total of 87 intervention occasions in the total study period, out of a total of 198 possible, within the significant models.

Figure 4 shows graphically the presence of the different air pollutants in the models with adjusted R² > 30% for asthma, for each study year and for the whole period 2014–2017. As can be seen, O₃ and PM₁₀ are the pollutants with generally higher values of such presence.

In the models obtained, pollen concentration has a positive influence on asthma episodes, except in a single case in 2014 in Aranjuez, where Ulmus pollen concentration has a negative influence (Table A2).

In the stations of Aranjuez, Getafe, and Madrid Barrio de Salamanca, the concentration coefficients of the Pinus pollen type are higher than the rest in at least 3 of the 4 years of study. In Las Rozas and Madrid Ayuntamiento, the highest coefficients correspond to the concentration of the Olea pollen type in most years (Table A2).

At the Alcobendas and Leganés stations, there are higher coefficients for Pinus in two of the years, compared to Olea and Ulmus, with higher coefficients in one of the years, respectively. The same occurs for Olea at the Madrid: Facultad de Farmacia station, with respect to Pinus and Ulmus. At the same time, at the Alcalá de Henares station, the presence of the highest coefficients for Pinus and Olea is equal in 2 years (Table A2).

As for the concentration of pollutants, O₃, CO (with the highest coefficients), and PM₁₀ are presented, in this order, as those with the greatest weight in the models (coefficients higher than the rest of the variables, in 15, 11, and 8 of the 44 equations found, respectively). Both O₃ and PM₁₀ have a negative influence on asthma episodes, while the presence of other pollutants has a different influence depending on the year and season.

In 2014, in 7 of the 11 stations, the highest coefficient corresponds to Olea, and in 2015, in 7 of the 11 stations, the Pinus coefficient is higher. In 2017, the Pinus coefficients are also the highest in 8 of the 11 models. In 2016, the number of stations where the highest coefficients, in the equations obtained, are those of Olea and Ulmus are the same at four stations (Table A2).

In 2014, in 5 of the 11 stations, the highest coefficient corresponds to CO, as is the case with O₃ in 2017. In 2015, O₃ and PM₁₀ have the highest coefficients at three of the stations, respectively. This triple equality at the station level occurs in exactly the same way in 2016 with O₃, PM₁₀, and SO₂ (Table A2).

In summary, the superiority in terms of weight in the models falls on Pinus, followed by Olea, with coefficients higher than the rest in 21 and 14, out of the 44 total calculated models, respectively. And in O₃ (in 15 of the 44 equations), CO (in 11) and PM₁₀ and SO₂ (in 8, in both cases) (Table A2).

5. Discussion

The quality of the atmosphere in cities is an issue of vital importance for public health, as well as for the realisation of urban design and development that is sensitive to the search for the well-being of citizens. Therefore, the urban planning policies developed in the town councils of large cities and towns are positively encouraged to promote the planting and conservation of vegetation, and trees in particular, which represent a real filter capable of partially mitigating the undesirable effects of atmospheric pollutants. Particularly striking is the fact that citizens living in areas of cities exposed to significantly higher levels of air pollutants can see their respiratory health improved by an increase in tree cover [21].

The presence of trees can lead to a reduction in the concentration of atmospheric pollutants and thus to an improvement in air quality, which is more noticeable for citizens living closer to green areas, among other reasons because of the reduction in wind turbulence, which reduces the dispersion of air pollution, as well as the capture and absorption of these unwanted components of the urban atmosphere [21].

Several studies indicate that certain types of allergenic pollens can be altered to a greater or lesser extent, depending on the species of origin and concentration, in terms of pollen allergen protein content and release, or in relation to pollen grain morphology, by atmospheric pollutants. This interaction between these two types of atmospheric components has been demonstrated by various field studies and laboratory experiments [21].

It is, therefore, very useful to determine the overall concentrations of air pollutants and pollen aeroallergens in order to determine their involvement in allergic pathologies, such as extrinsic asthma, and the quantification of the increase in allergenicity, due to people’s exposure to air pollution, of pollen types derived from plant species present in cities. Moreover, the effect of this interaction may be enhanced by other factors, such as meteorological factors [9,21].

Various investigations have analysed the interrelationship of short-term exposure to O₃, CO, NO₂, SO₂, and particulate matter (PM_2.5 and PM₁₀) and asthma-related hospitalisations. Some epidemiological studies have also found a significant correlation between increased incidence of asthma and long-term exposure to air pollutants. They further state that short-term exposure to O₃, NO₂, SO₂, PM_2.5, and road traffic pollutants cause an exacerbation of asthma symptoms, while long-term exposure to these pollutants, especially those related to vehicular traffic on streets and highways, is associated with the onset of asthma in adults and children [9].

Urban green spaces in general, without a high tree density, have a generally beneficial relationship with asthma hospitalisations. However, this positive relationship may be reduced due to the influencing effects of air pollution on enhancing the bioavailability of pollen aeroallergens and the consequent exacerbation of asthma caused by the synergy between air pollutants and airborne pollen grains [21].

Trees can generate potentially allergenic pollen, especially with specific types, that is specific to the areas of this study, such as Olea, Cupressaceae, and Platanus. However, these negative effects and those related to the influence of air pollution on pollen do not seem to affect the correlation between the existence of trees and the exacerbation of asthma, in accordance with the results obtained by other studies. On the contrary, trees contribute to the dispersion and removal of air pollutants, in proportion to how densely they are located in urban areas, and may lead to a reduction in asthma hospitalisations in urban areas that are both more polluted and more densely populated by trees [21].

In the study by Kim, H. et al., there is a significant overall relationship for allergic pathologies in the spring and autumn seasons between air pollutant concentrations, pollen concentrations, and the number of people who visited outpatient health clinics in Korea. Particularly significant is the relationship between NO₂ and asthma, as well as allergic rhinitis and atopic dermatitis, in spring. The same is true for SO₂ (with a high positive association), in relation to asthma and the other two pathologies mentioned, in autumn [45]. Whereas asthma, in the Korean study, shows a highly significant association only with herbaceous pollen in both seasons of the year, and with all types of air pollutants [45]. In the present study, as could be observed, statistically significant patterns have been obtained for asthma with all six types of air pollutants over the whole period 2014–2017, with O₃ being the most present, followed by PM₁₀, NO₂, and SO₂ (Table 3).

The study by Alcock, I. et al. concluded that there was a reduction in the number of asthma hospitalisations coincident with higher levels of SO, NO₂, and PM_2.5 and with more densely populated areas with trees as well as with lower levels of PM_2.5 and NO₂ in green areas and gardens with fewer trees [21].

In a study carried out in two Spanish cities, one of them with high levels of SO2, NO2, O3, and PM10 atmospheric concentrations due to its powerful industry, Puertollano, and the other, Ciudad Real, with less air pollution, the clinical evolution of asthma in patients allergic to Olea and Gramineae pollens during the pollination period, with positive skin sensitisation tests, and with a mild-to-moderate clinical evolution, was analysed. The evolution of these patients is found in this study to be worse in Puertollano, and this is evidenced by the consumption of medication, as well as by an increase in the severity of asthma symptoms. This study is of particular relevance given that it is usually asthmatic patients with a severe evolution who are the subject of research [46]. In this study conducted in the CAM, these four pollutants (headed by O₃, followed by PM₁₀, NO₂ and SO₂, the latter two at the same level of presence) are the most present in the models calculated for asthma. Furthermore, as we have already seen, Olea is the type of pollen most frequently associated with asthma, and it is also the type of tree that causes the greatest sensitisation in asthmatic patients (36.9% of cases) seen in Allergology clinics in Spain [7], as well as being one of the main causes of pollinosis in the CAM, although of lesser importance than grass pollen [5,47].

Several experimental and laboratory studies have led to evidence that exposure to O₃ results in a halving of the pollen dose required to develop a 15% decrease in forced expiratory volume (FEV1), compared to air free of that pollutant. In addition, exposure to PM₁₀ particles as well as NO₂ increases the asthmatic response in patients allergic to grass pollen. Not many studies have established an interrelationship between air pollutants, pollen grains, and visits to the emergency department due to asthma attacks, and the results obtained in these studies are controversial, possibly because they were not carried out during the pollination seasons with the types of pollen studied, which could imply a masking of the real interaction between pollutants and pollen by the action of the former, which have more stable concentration values throughout the year [46].

In the time series study conducted in 10 Canadian cities over 7 years, which analysed the interactions of tree, grass, and other herbaceous pollens and three classes of fungi, as well as NO₂, SO₂, and O₃ concentrations and the haze coefficient, there was a significant interaction between O₃, tree pollen counts, and the total number of hospital admissions for asthma in the analysis of all data [48,49]. In addition, another case-crossover study in New Jersey, USA, demonstrated an increase in asthma admissions of children to hospital emergency departments in response to increased O₃ and PM_2.5 particle counts on days with high tree, grass, and other herbaceous pollen concentrations during the warm season [48,50]. As already mentioned, in this investigation, O₃ is the atmospheric pollutant most present in the calculated equations for asthma, and PM_2.5 particles are the least frequent (for reasons already explained).

In another investigation carried out in the territory of the CAM, the results obtained are similar to the current study on asthma in terms of the pollen types, with a greater interrelation with allergic rhinitis and conjunctivitis, as these are, in first place, Pinus, Olea, and Populus, throughout the study period, also between 2014 and 2017, both years included [24]. In the period 2014–2017, O₃ is the most common air pollutant included in the equations considered for allergic conjunctivitis [24], as is the case for asthma in this study, while PM₁₀ particles are the most commonly present in the models obtained for allergic rhinitis, followed by O₃ [24]. PM₁₀ particles are the second most common air pollutants present in the models calculated for asthma.

6. Conclusions

-The relationship between the response variable (the number of primary care asthma episodes) and the explanatory variables (the concentration of airborne pollen grains of the tree types and air pollutants included in this study) is conditioned by the time and geographic location in which the measurements are made.

-The influence of both independent variables on asthma can be confirmed, given the statistically significant relationships between them in all areas and in each of the years of study.

-The association of both types of airborne agents acts on the pathology investigated in this study and explains more than 30% of the variability of asthma episodes of care in most of the models calculated. In all calculated equations, there is an interrelationship of asthma with at least one of the air pollutants and at least one of the types of pollen.

-In most of the statistically significant models obtained, the incorporation of pollen types is greater than that of air pollutants. This is largely due to the fact that the only pollutants that are quantified in the Air Quality stations and in the total number of stations and in this study’s period are O₃ and NO₂.

-The air pollutants analysed are significant in the calculated equations, with a greater intervention of O₃ and PM₁₀ and of NO₂ and SO₂ (the latter two to the same extent), so that the intervention of these pollutants influences the cases of asthma. Olea, Pinus, and Populus pollen concentrations are the most commonly existing ones in the interrelations with the pathology investigated in this study.

-The results of various investigations analysed in this study suggest, judging by the results obtained, that there is a joint interaction between atmospheric pollutants and pollen concentration in the atmosphere and asthma.

-It is interesting to consider carrying out a future study of meteorological factors, together with the variables analysed in this research, given their important influence on pollination and pollen grains. It would also be interesting to restrict the study to the main pollination periods in order to obtain equations that explain, if possible with greater precision, asthma episodes of care.

-Asthma cases, in general, interrelate more strongly with the independent variables in more urban municipalities and more weakly in municipalities with lower urban character indicators.

-Three measures can be proposed to counteract the possible increase in allergenicity of pollen aeroallergens as a consequence of their interaction with atmospheric pollutants: adequate epidemiological surveillance; green spaces planning and management in cities, taking into account a precise selection of anemophilous pollinating flora; and increasing control of air quality.