Aeroallergens, Allergic Disease, and Climate Change: Impacts and Adaptation

Changes in meteorological factors can affect aeroallergen production, which, in turn, impacts the prevalence or severity of allergic illness via sensitivity and response pathways. For example, greater pollen concentrations in the air increase the chance of sensitization and subsequent exacerbation of illness. Exposures to higher concentrations of aeroallergens or allergen proteins may lead to more severe allergic responses (Nielsen et al., 2002; Singer et al., 2005). Whereas land-use change can have an impact on the amounts and types of pollen present at any location, climate also can play a role.

There are several examples where regional weather patterns, such as increased precipitation and temperature, have led to enhanced pollen production. Maximum pollen counts were higher during a warmer, wetter El Niño year than in the year before or after it in New England (Freye et al., 2001). Oak pollen counts in the San Francisco Bay Area have shown strong correlations with total rainfall from the previous year (Weber, 2003). However, increasing trends of total pollen production in cedar, oak, hickory, pecan, and birch in Oklahoma were not significantly correlated with increasing winter temperatures (Levetin, 2001), but the increasing pollen counts of cypress family pollen in Tulsa showed significant positive correlations with daily minimum, mean, and maximum temperatures and significant negative correlations with precipitation (Lo and Levetin, 2007).

A study in Denver, Colorado showed correlations between grass pollen counts and meteorological variables with 1-day time lags that were not consistent from year to year, suggesting that the intraseasonal meteorological conditions that determine pollen counts may vary from year to year or that preseason conditions are more important than day-to-day variations (Glassheim et al., 1995). An analysis of three cities in England found different trends in start date and total grass pollen count by city, signifying the need for regional assessments of climatic trends on pollen counts. Although meteorological factors explained some of the variation in grass pollen counts, changes in the amount of grassland in a region also played a large role (Emberlin et al., 1999).

Studies of ragweed in controlled environments and field studies show that pollen production can be expected to increase with increased temperature and CO₂ levels. The Wan et al. (2002) experiment demonstrated that ragweed plants grown in warmer soil exhibited increases in number of stems, total biomass, percent coverage, pollen diameter, and total pollen production. Other experimental results have demonstrated that doubling CO₂ levels from current to projected future levels would result in a 30–90% increase in ragweed pollen production (Ziska and Caulfield, 2000; Wayne et al., 2002; Rogers et al., 2006). Field studies of differences between rural and urban growth patterns also clearly showed that ragweed flowers earlier and produces greater amounts of pollen at urban locations where there are higher CO₂ concentrations and temperatures compared with rural locations (Ziska et al., 2003).

Short-term forecasts indicate that although temperature and humidity can be strong predictors of mold concentrations, the effect varies by mold species and geography. A study in Denver, Colorado observed an upward trend in Cladosporium spore counts during 8 years that was significantly positively correlated with average temperature and humidity and significantly negatively correlated with precipitation; however, no significant trends were found for Alternaria or Epicoccum (Katial et al., 1997). Temperature and dew point were found to be important predictors of mold spore counts in Tulsa; however, meteorological factors alone could not predict all of the variability in spore counts (Troutt and Levetin, 2001). A study in New England observed higher mold counts after a warmer and wetter El Niño event (Freye et al., 2001). Overall, the relationship between climate factors and mold species, extent, and geography suggests a complex multifactorial mechanism.

Changes in the timing and length of the pollen season lead to changes in human exposure, which can affect sensitization as well as exacerbation of allergic illnesses. Shifts in phenology are among the most consistent findings in studies of plant pollen production. A recent meta-analysis found that the current rate of spring phenological advance is 5 days per decade for numerous species of non-tree plants and 3 days earlier per decade for numerous species of trees (Root et al., 2003). Alterations in the timing of aeroallergen production in response to weather variables have been clearly demonstrated for certain tree species, but less so for grass and weed pollens and mold (Katial et al., 1997; Emberlin et al., 2002; Clot, 2003). Evidence is mixed for grass pollens, with trend studies showing substantial differences by region in England (Emberlin, 1994), earlier start dates in Switzerland (Clot, 2003), but no apparent effect after an El Niño event in New England (Freye et al., 2001).

Ragweed has been shown to flower earlier in urban environments where temperature and CO₂ concentrations were higher compared with rural areas (Ziska et al., 2003). There is limited evidence of changes in the start dates for the emergence of mold, except for the El Niño event in New England, which indicated an earlier start (Freye et al., 2001). Some molds, such as Alternaria, are associated with agriculture and therefore the timing of production is associated with the harvest (Corden and Millington, 2001), which, if the timing of the harvest changes with climate change, could mean changes to the timing and duration of the mold season.

Studies in Spain have used the evidence of a linkage between temperatures and pollen season start to project future phenological changes based on output from climate models. These studies project that the Olea europaea L. pollen season in Andalusia will advance by 1 to 3 weeks by the end of the century (Galán et al., 2005) and that the Quercus pollen season may start a month earlier and have up to a 50% increase in pollen throughout Spain (Garcia-Mozo et al., 2006).

Long-term responses to climate change (during 50–100 years) are likely to involve range or distribution shifts in species and, in some cases, extinctions (Davis and Shaw, 2001; Joyce et al., 2001; Melillo et al., 2001). Ecological models indicate that the potential habitats for many tree species in the United States are likely to change, in some cases dramatically, by the end of the 21^st century. Potential habitats for trees favored by cool environments are likely to shift poleward and upward (Fischlin et al., 2007). The habitats of alpine, subalpine spruce/fir, and aspen communities will likely contract dramatically in the United States and largely shift into Canada. Potential habitats are likely to increase in the United States for oak/hickory, oak/pine, ponderosa pine, and arid woodland communities. Projections for vegetation redistribution suggest that savannahs and grasslands are likely to expand in the United States, especially in the Southeast, where hot and dry climate conditions are projected (Joyce et al., 2001). Climate change may act alone or in concert with other forces to influence dispersion or distribution of aeroallergens, which could result in significant impacts on allergic illnesses through exposure to novel or unusually high concentrations of aeroallergens.

Shifts in vegetation distribution are likely to occur during relatively long periods of time. Furthermore, the range of a species (e.g., birch) could contract or move northward and another (e.g., white oak) could take its place without any appreciable difference in allergic illness if those two species are cross-reactive. However, there is some indication that the introduction of new pollen into a region can lead to increased allergic illnesses in the population, even among people who did not have previous allergies to other pollens. The increase in pollen allergies to birch and ragweed pollens in the suburbs north of Milan has been attributed to the introduction of birch trees in the 1970s and 1980s and the natural spread of ragweed into the area in the 1980s. An investigation of people in Milan with new allergies solely to birch or ragweed pollen found that they had an older age of onset of allergies and were less likely to have a family history of allergic illnesses compared with people allergic to other pollens naturally found in the area (Asero, 2002). Although not directly related to changes in climate, this provides an example of how, if species ranges shift due to climate change, new cases of allergies in people who exhibit no previous symptoms could emerge.

There has been only limited research on how climate change could affect the dispersal of pollen and mold. Dispersion has the potential, via shifts in long-term weather patterns and extreme weather events, to expose and thus sensitize populations to novel allergens. There are cases of both pollen and dust being dispersed long distances from their release sites. For example, long distance dispersion of Juniperus ashei pollen in Tulsa, Oklahoma has been routinely observed and is associated with allergic illness in that community (Van de Water et al., 2003). Transcontinental transport of dust particulates also has been observed. During April 1998, the dust plume from two large dust storms over the Gobi desert crossed the Pacific Ocean and resulted in strong spikes in particulate matter concentrations 10 days later along the west coast of the United States (Husar et al., 2001). Although it is unclear whether there were health impacts of this dust event, other long-range transport events have transported pollen grains and fungal spores across long distances in desert dust (Shinn et al., 2003; Kellogg and Griffin, 2006). There also is evidence of pollen grains being transported long distances by wind without accompanying dust, for example, birch pollen from Poland and Germany to Denmark (Skjoth et al., 2007), and beech pollen from Italy and Germany to northeastern Spain (Belmonte et al., 2008).

Allergic symptoms are related to pollen in a dose-response manner (Singer et al., 2005). Whereas pollen concentration has been taken as the indicator of potential dose, the underlying mechanism for allergic symptoms comes in part from the protein allergens in the pollen (Ahlholm et al., 1998; Beggs, 1998; Singer et al., 2005). Recent research has examined the influence of meteorological variables, such as temperature and precipitation, and air pollutants, such as CO₂, on the concentration of allergen protein, or the allergenicity, of pollen.

Singer et al. (2005) evaluated how Amb a 1 allergen (the main allergenic protein in ragweed pollen) concentrations changed in response to rising CO₂ concentrations. Although total pollen protein remained unchanged, concentrations of Amb a 1 increased as a function of CO₂ concentrations. After controlling for descendant group, another study found that IgE-immunoblotting responses were stronger in human sera exposed to pollen grown at higher temperatures (Ahlholm et al., 1998). The authors suggest soil temperature, which limits nutrient uptake and thus the rate of allergen synthesis, as a possible mechanism for the increased allergic response. Although research on allergen content of pollen/mold is limited, it does suggest that even if pollen production remained unchanged, allergic illness could increase because of increasing levels of allergenic protein due to higher CO₂ concentrations.

There are very few examples of observational studies that investigate the linkages between weather, aeroallergens, and health outcomes. These studies provide limited evidence of the seemingly obvious, but difficult to demonstrate, link between weather, aeroallergen production, and subsequent illness.

Epton et al. (1997) conducted a 1-year prospective study to explore relationships between weather, fungal spore counts, pollen counts, and peak expiratory flow rates (PEFR) and asthma in a group of subjects with asthma. A small positive association was found between PEFR and mean temperature. The study also found an association between days with high basidiospore counts and nocturnal wakening and medication use to relieve asthma. The authors concluded that the effects of weather and aeroallergens on PEFR and asthma symptoms in the studied population were small and that other currently unknown causes may explain variations in asthma severity and exacerbations; however, these minimal effects also could be due to the high rate of anti-inflammatory medications being used by the cases and the lack of control subjects in this study.

Another prospective study of asthma admissions in Israel found that daily emergency room (ER) visits for asthma were positively correlated with NO_x (nitrogen oxides), SO₂ (sulfur dioxide), and high barometric pressure and negatively correlated with ozone (O₃), and minimum and maximum temperature. They found no correlation between asthma-related ER visits and pollen, spores, humidity, or particulates (Garty et al., 1998). However, a time-series study of the link between grass and weed pollen and emergency department (ED) visits for asthma in Montreal found significant correlations between increases in grass pollen concentrations and increases in first-time ED visits for asthma at lags of 3 days and 5 days, and that readmissions for asthma increased 4 days after an increase in grass pollen concentrations. Adding meteorological or air quality variables did not significantly alter the relationships observed between pollen and asthma (Héguy et al., 2008).

Whereas most epidemiological studies that investigate the relationship between meteorological variability and aeroallergens focus on asthma, there is some evidence of an impact of meteorological factors on the prevalence and intensity of atopic dermatitis and allergic rhinitis. Another Montreal study investigated the links between weather, Ambrosia (ragweed) pollen, and medical consultations for allergic rhinitis. Ambrosia pollen concentrations were significantly positively correlated with maximum, mean, and minimum temperatures but negatively correlated with precipitation. Logistic regression showed strong correlations between pollen count and higher medical consultations for allergic rhinitis on the day of and 1, 2, 3, and 5 days after elevated pollen exposure (Breton et al., 2006). A cross-sectional population-based study of 6- to 7-year-olds in urban areas of three climatic regions in Spain found not only different rates of eczema in these regions, but also prevalence rates that were positively associated with precipitation and humidity and negatively associated with temperature and number of sunny hours (Suárez-Varela et al., 2008). This implies that as climate changes for a given region, rates of atopic dermatitis also could change; however, they did not specifically assess the role of allergen concentrations in this study. Further investigations of the relationship between meteorological factors, aeroallergens, and allergic diseases are needed to better understand whether the inconsistent findings of these studies are due to study design, location, population, or other factors.

Severe weather events, such as thunderstorms and hurricanes, provide intriguing evidence of an association between weather, aeroallergens, and allergic illness. There have been many documented cases of higher rates of asthma attacks during thunderstorms, not just for asthmatics but also for people who have previously only had allergic rhinitis. Thunderstorm winds and associated precipitation produce the correct conditions for pollen grains to burst and release allergen particles small enough to enter the lower airways (D’Amato et al., 2007). In England there were high correlations between the timing of thunderstorms, concentrations of Didymella, Cladosporium, and broken Alternaria spores, and asthma hospital admissions. This study also found that the odds of being sensitive to Alternaria and/or Cladosporium spores was significantly increased for the thunderstorm-induced asthma admissions compared with controls (Pulimood et al., 2007).

A 6-year time-series analysis of aeroallergens, air pollutants, and asthma admissions in a Birmingham, England children’s hospital found that the fungal spore concentration explained more of the increase in asthma rates on thunderstorm versus non-thunderstorm days than did pollens or air pollutants (Dales et al., 2003). Grundstein et al. (2008) found a significant correlation between asthma ED visits and thunderstorms in Atlanta, Georgia, and there was significant effect modification by rainfall and wind speed, in that there was only an effect in thunderstorms with rainfall and this effect increased with moderate and high wind speeds.

Hurricanes also have been shown to impact allergic diseases. The most widely studied of these is Hurricane Katrina, where large sections of New Orleans were flooded for weeks, resulting in extensive mold growth in buildings. Of 112 households in the New Orleans area assessed by the CDC, almost half the homes had “visible mold growth” and 17% had “heavy mold coverage,” and allergy-causing molds were found both indoors and outdoors (Ratard et al., 2006). Solomon et al. (2006) also found high mold spore counts in October and November 2005 in New Orleans. Higher mold spore concentrations were found inside compared with outside and much higher spore concentrations were found in flooded areas compared with unflooded areas, which were higher than background concentrations.

Hospitals in New Orleans have reported an increased number of patients with allergy and cold symptoms, and doctors have suggested that allergy to the mold and dust circulating in New Orleans made residents susceptible to respiratory illness (Wilson, 2006). There have been reports of a nagging cough throughout New Orleans that has been nicknamed “Katrina cough” or “shelter cough,” which may be partially due to the higher spore or other particulate concentration in the air (Manuel, 2006). Overall rates of asthma in children from the New Orleans area increased post-Katrina from 14% (2003) to 18% (2006) and was possibly higher for minority and underprivileged children or children residing in certain geographical areas that were affected by post-Katrina flooding (Epstein et al., 2005).

Current climate change projections show likely increases in heavy downpours and hurricane intensity but not necessarily frequency (Climate Change Science Program, 2008). Although an individual event like Hurricane Katrina cannot be attributed solely to climate change, and the post-Katrina mold could be attributed more to the response and not the storm itself, this example does highlight what could happen if more intense precipitation events occur without an increase in our ability to adapt or respond appropriately.

Some recent studies have shown that the interactions between air pollution, aeroallergens, and allergic diseases are complex and that air pollution may not only play a significant role in the etiology of some allergic diseases but that it also may modify the effect of aeroallergens on these diseases. D’Amato et al. (2002) hypothesize that the reason for the increase in urban allergic disease may be due to the role that air pollutants play in mediating the health effects of aeroallergens. The authors summarize the scientific literature that demonstrates these interactions: the inflammatory effects of O₃, particulate matter (PM), and SO₂ allow for easier penetration of pollen allergens into the airways; air pollutants can increase the release of antigens in pollen grains that lead to allergic responses; and pollutants also can absorb pollen grains and thus prolong the retention of the pollen grains in the body once inhaled. Given that increases in temperature from climate change may increase ground-level ozone as well as aeroallergen concentrations, the interactions between these two airway irritants are important for further study.

Knowlton et al. (2007) placed the locations of ragweed pollen prevalence and areas with at least one ozone exceedance day per year on the same map because both of these exposures are associated with asthma. Notably, 14 of the 15 cities listed by the Asthma and Allergy Foundation of America as the “most challenging places to live with asthma” were in areas that had overlapping risks of ozone and ragweed. These regions of overlap highlight regions of vulnerability that may increase in size with climate change.

In observational studies of asthma, where aeroallergens and air pollutants could both be the cause, it is important to analyze the potential for confounding of one by the other. A few examples highlight the inconclusive results found in studies that investigate these confounding effects. It is important to note that these studies controlled for meteorological factors but did not specifically analyze the impacts of meteorological variability on the outcome. In a study that investigated the link between air pollution and ED visits for asthma in Alberta, Canada, adding aeroallergen data to the analysis did not alter risk estimates for air pollution on ED visits for asthma (Villeneuve et al., 2007). However, a study by Erbas et al. (2007), which focused on the relationship between grass pollen and ED visits for asthma in Melbourne, Australia, found that adding NO₂ (nitrogen dioxide), but not O₃ or SO₂, made an impact on their model. They also found that the air particle index and NO₂ were independently associated with asthma ED visits, and that grass pollen was still associated with asthma ED visits even when all air pollutants were added to the model (Erbas et al., 2007).

Two studies of ER visits for asthma in Madrid show the complexity of the roles of aeroallergens and air pollutants on asthma. Galán et al. (2003) reported that pollen did not modify the association between O₃, NO₂, and PM₁₀ and asthma ER visits, but Tobías et al. (2004) found a relationship between three pollen types and asthma ER admissions that remained constant even when controlling for air pollutants. This study also found two epidemics of asthma that were correlated with peaks in pollen releases.

A study in Finland, however, found no relationship between asthma attacks and pollen or meteorological factors except for temperature, which had a low inverse correlation with asthma attacks, and air pollutants, specifically NO₂, SO₂, TSP (Total Suspended Particles) and H₂S (hydrogen sulfide), were significantly correlated with asthma attacks (Rossi et al., 1993). A time-series analysis of 7 years in ten cities in Canada found that controlling for air pollution and other allergens, such as mold spores and grass pollens, did not confound the relationship between tree pollen and asthma hospitalizations (Dales et al., 2008).

The interrelationships between climate variability, air pollution, and aeroallergens are complex and the different findings in the literature may be due to the specific aspects of a given study, such as the species of pollen, air pollutants, geography, current climate, and sensitization of the population. A recent review of these issues by D’Amato and Cecchi (2008) explains this complexity and demonstrates how projected climatic changes could have both positive and negative effects on respiratory allergies because of the multiple ways in which climatic and other global changes, such as urbanization and increased energy use, may impact air pollutants, aeroallergens, and ultimately allergic illnesses.