Acute exposure to pollen and airway inflammation in adolescents

Summary/Abstract

Introduction:

Pollen exposure is known to exacerbate allergic asthma and allergic rhinitis symptoms, yet few studies have investigated if exposure to pollen affects lung function or airway inflammation in healthy children.

Methods:

We evaluated the extent to which higher pollen exposure was associated with differences in airway inflammation and lung function among 490 early adolescent participants (mean age of 12.9y) in Project Viva, a pre-birth cohort based in Massachusetts. We obtained regional daily total pollen counts, including tree, grass, and weed pollen, from a Rotorod pollen counter. We evaluated associations of 3- and 7-day moving averages of pollen with fractional exhaled nitric oxide (FeNO) and lung function using linear regression models and evaluated the linearity of associations with penalized splines. We tested if associations of pollen with FeNO and lung function were modified by current asthma diagnosis, history of allergic rhinitis, aeroallergen sensitivity, temperature, precipitation, and air pollution.

Results:

3- and 7-day median pollen concentrations were 19.0 grains/m³ (IQR:73.4) and 20.9 grains/m³ (IQR:89.7). In main models, higher concentrations of total pollen over the preceding 3 and 7 days were associated with a 4.6% (95% CI:0.1,9.2) and 7.4% (95% CI:0.9,14.3) higher FeNO per IQR of pollen, respectively. We did not find associations of pollen with lung function in main models. Asthma, allergic rhinitis, precipitation, and air pollution (nitrogen dioxide and ozone) modified associations of pollen with lung function (P_interaction<0.1), while temperature, sex, and aeroallergen sensitization did not.

Conclusion:

Short-term exposure to pollen was associated with higher FeNO in early adolescents, even in the absence of allergic sensitization and asthma.

Introduction

Climate change is causing longer pollen seasons and higher pollen concentrations, which has implications for respiratory and allergic diseases.¹ Increases in atmospheric carbon dioxide cause higher global temperatures, both of which cause plants to generate more pollen for longer periods of time.² This is particularly relevant in the Northeast United States (U.S.), which is experiencing greater temperature increases compared with the rest of the U.S.³ Children and adolescents are more likely to be exposed to these changes in pollen seasons because they spend more time outside than adults.⁴ Pollen grains induce an inflammatory response in the respiratory epithelium of sensitized mice⁵ and studies have shown that short-term pollen exposure increases allergic and asthmatic symptoms and increases medication use in susceptible populations.⁶

Despite the well-described relationship between pollen exposure and allergy and asthma symptoms,^7,8 the association between pollen exposure and lung function and airway inflammation is not well-established. Of the studies assessing the impact of pollen on lung function and airway inflammation, most have involved primarily adult cohorts in Australia⁹ or Europe¹⁰ where the plants producing pollen may be different than the U.S.

Our primary aim was to evaluate the extent to which short-term exposure to pollen, including tree, grass, and weed pollen, in Massachusetts, was associated with airway inflammation and lung function among early adolescents. Our hypothesis was higher pollen concentrations in the preceding 7 days would be associated with higher airway inflammation, as measured by fractional exhaled nitric oxide (FeNO), and lower function in generally healthy children. Secondarily, we also assessed whether associations of pollen with FeNO and lung function were modified by asthma diagnosis, allergic rhinitis, aeroallergen sensitivity, temperature, precipitation, sex, and air pollution.

Materials and Methods

Study Population

Study subjects were early adolescents in Project Viva, a longitudinal pre-birth cohort in Massachusetts that recruited women seen between April 1999 and November 2022 for prenatal care at eight urban and suburban practices of Atrius Harvard Vanguard Medical Associates, a multispecialty group practice in Eastern Massachusetts.¹¹ Exclusion criteria included multiple gestation, inability to answer questions in English, gestational age ≥22 weeks at recruitment and plans to move away before delivery. All mothers provided written informed consent for themselves and their child and early adolescents provided verbal assent. The study was approved by the institutional review boards of Beth Israel Deaconess Medical Center and Harvard Pilgrim Health Care.

Medical history and measurements of spirometry and FeNO were performed at the early adolescent study visit. Of the 2,128 infants who were enrolled in Project Viva, we included 490 early adolescent participants who had both exposure data and either spirometry or FeNO measurements and lived in Massachusetts. We confined our study population to those living in Massachusetts to align with our Massachusetts-based pollen measurements.

Demographic and questionnaire data

We obtained data on demographics and medical history from questionnaires and interviews. We use mother-reported responses to questions on race and ethnicity (and categorized them as non-Hispanic Asian, non-Hispanic Black, non-Hispanic White, Hispanic, or other, which includes responses by mothers identifying multiple races or ethnicities) to avoid concealing health disparities, acknowledging that race and ethnicity are social constructs without biological meaning. We use self-reported annual household income more than USD $70,000 and maternal college graduation (yes or no) ascertained at enrollment as measures of socioeconomic status in analyses. Household tobacco use was defined as the presence of anyone in the child’s home who smokes based on questionnaire response in early adolescence. We define current asthma as having a prior diagnosis of asthma by a doctor plus wheezing symptoms or use of asthma medications in the past year, which were obtained by maternal report. History of allergic rhinitis was obtained by questionnaire and defined as maternal report of the child having ever received a diagnosis of hay fever or allergic rhinitis by a healthcare professional (doctor, physician assistant, or nurse practitioner).

Exposure Assessment

We obtained regional daily total pollen counts, including tree, grass, and weed pollen, measured in grains per cubic meter (grains/m³) from Asthma and Allergy Affiliates, Inc., an allergy practice in Salem, Massachusetts, that records daily pollen concentrations using a Rotorod sampler. Total pollen concentrations include tree, grass, and weed pollen, as well as pollen that was unable to be classified as either tree, grass, or weed (Figure 1). We analyzed total pollen concentrations as well as pollen concentrations separated into tree, grass, and weed, similar to prior studies.⁷ Pollen collected over weekends and holidays was not counted until the next business day. Thus, for those days, concentrations represent an average of the total pollen counted per day (total pollen divided by the number of days the pollen was collected). For our analyses, we use the 3- and 7-day moving averages for pollen. Moving averages are defined as the total pollen collected over the specified time period preceding the study visit divided by the specified time period (e.g., the 7-day moving averages is the total amount of pollen over the preceding 7 days, divided by 7), similar to prior published work in this cohort.¹² The choice of 3- and 7-day moving averages reflects the exposure windows available for pollen, temperature, precipitation, relative humidity, and air pollution. The pollen season is defined as March to October, when pollen counting began and ended, which coincides with the typical start and end of pollen seasons in the Northeast U.S. We excluded early adolescents with study visits outside the pollen season. Data for days within the pollen season without pollen measurements were excluded.

Figure 1.

Smoothed (LOESS) pollen concentrations between April and October 2012 for tree pollen (solid line), weed pollen (dashed line), grass pollen (dotted line) in Massachusetts, United States.

Spatially and temporally resolved daily temperature, precipitation, and relative humidity measurements are based on the Parameter-elevation Relationships on Independent Slopes Model (PRISM) 800 meter resolution climate dataset for the U.S. maintained by Oregon State University.¹³ We defined temperature as the daily mean temperature. Daily air pollution data for PM_2.5, O₃, and NO₂ were estimated on a 1x1 kilometer grid spatial resolution based on modeling developed at the Harvard School of Public Health.^14-16 Weather and air pollution data were linked to geocoded home addresses for participants.

Lung function measurement in early adolescence

Trained research assistants obtained lung function measurements of forced vital capacity (FVC) and forced expiratory volume in 1 second (FEV₁) using the EasyOne Spirometer (NDD Medical Technologies, Andover, MA) in accordance with the American Thoracic Society guidelines for acceptability and reproducibility.¹⁷ Participants were required to produce at least three acceptable spirograms to be included, two of which had to meet criteria for reproducibility defined as a ≤0.150L difference between the two largest FEV₁ values and the two largest FVC values.

Exhaled NO measurement in early adolescence

Fractional concentration of exhaled nitric oxide (FeNO) was measured in early adolescents and calculated as the mean of two efforts.¹⁸ Study participants were instructed to inhale through an NO scrubbing filter and exhale into room air on two efforts to avoid ambient NO measurement. On the third breath, participants exhaled into the portable hand-held FeNO analyzer device, an electrochemical device (NIOX MINO) validated by the clinical standard chemiluminescence FeNO analyser.^19,20 Only the last three seconds of exhalation were used to ensure lower rather than upper airway measurements of NO. Nose clips were not used. The measurements were log-transformed due to non-normality, consistent with prior published work by this group.^12,21

Outdoor and indoor aeroallergen sensitization measurement in early adolescence

Trained research phlebotomists collected blood from participants at the early adolescent visit, which we centrifuged and stored at −80°C. Collection of IgE coincided with measurement of lung function and FeNO. Allergen extract-specific IgE antibodies were measured by ImmunoCap (Thermo Fisher Scientific/Phadia, Kalamazoo, Michigan).^22,23 329 participants provided blood samples with IgE results. We define outdoor aeroallergen sensitization as having any IgE >0.35 IU/mL against Aspergillus fumigatus (mold), Alternaria alternata (plant fungi) or common tree, weed, and grass species in the Northeast U.S.: Betula pendula (silver birch tree), Quercus (oak tree), Lolium (ryegrass), and Ambrosia artemisiifolia (common ragweed), similar to prior studies.^12,21 We define indoor aeroallergen sensitization as any IgE >0.35 IU/mL against Dermatophagoides farina (dust mite), cat dander, or dog dander. We define any aeroallergen sensitization as either an outdoor or indoor aeroallergen sensitization based on the definitions above.

Statistical Analysis

We analyzed associations of 3-day and 7-day moving averages for total pollen preceding the study visit with lung function and FeNO using multivariable linear regression models. We selected potential confounders and predictors of outcome a priori based on published and anticipated associations of pollen with lung function and FeNO. We adjusted primary models for early adolescent age, sex, height, weight, maternal education and household income at enrollment, any smoking in the home in early adolescence, race and ethnicity, and current asthma diagnosis. We additionally adjusted for date of study visit, season (as sine and cosine functions of the study visit date), as well as mean temperature, relative humidity, and air pollution (PM_2.5, O₃, and NO₂) in the same exposure window. We present effect estimates for log transformed FeNO as percent change, calculated as (e^β−1)*100. We evaluated the linearity of each association by plotting generalized additive models (GAM) with penalized splines.

By including cross product terms as well as the potential effect modifier, we tested if associations of total pollen with FeNO were modified by: (1) current asthma diagnosis, (2) any allergic sensitization to outdoor (tree, weed, or grass pollen) or indoor aeroallergens, (3) air pollution exposure as a continuous variable matching the pollen exposure time-window, (4) weather (mean temperature and precipitation as continuous variables) matching the pollen exposure time-window, and (5) sex. For effect modification analyses with P_interaction <0.1, we report the effects based on a dichotomized effect modifier (e.g., high or low air pollution) based on the medians for the 3- and 7-day moving averages.

Measures of association are reported as beta coefficients for lung function and % change for log-transformed FeNO with 95% confidence intervals. A two-sided P-value <0.05 will be used for statistical significance for main effects and <0.10 for assessing interaction in effect modification analyses. All analyses were performed using R (R Foundation for Statistical Computing, Vienna, Austria).²⁴

Results

Participant characteristics

Study participants had a mean age of 12.9 (SD 0.7), with a higher proportion of Black participants (16.8%) compared with overall Massachusetts demographics (Table 1). Participants were mostly from households with high maternal education and low household smoking.

Table 1.

Characteristics of early adolescent study participants

Characteristics (n=490)	Mean±SD or %
Female	51.4%
Age (years)	12.9 ± 0.7
Height (centimeters)	158.5 ± 8.5
Weight (kilograms)	52.7 ± 14.7
Race/ethnicity
Asian	2.9%
Black	16.8%
Hispanic	4.5%
White	64.4%
Other	11.5%
Active asthma diagnosis	14.4%
Allergic rhinitis/hay fever history	17.5%
Any aeroallergen sensitization*	59.3%
Indoor aeroallergen sensitization	49.2%
Outdoor aeroallergen sensitization	41.6%
Any smoking in the home	10.3%
Maternal college degree	73.4%
Annual household income >$70,000	65.5%
Spirometry
FEV1 (liters)	2.7 ± 0.6
FVC (liters)	3.2 ± 0.6
FeNO (parts per billion)	26.2 ± 26.6

^*Aeroallergen sensitization is defined as having IgE against common indoor and outdoor allergens using a cutoff of 0.35 IU/mL; 161 participants did not have aeroallergen data. FEV₁ = forced expiratory volume in 1 second; FVC = forced vital capacity; FeNO = fractional exhaled nitric oxide.

Pollen and weather exposures

Median (interquartile range) temperature, relative humidity, precipitation, air pollution, and pollen preceding the study visit are shown in Table 2. Tree pollen accounted for the majority of pollen in the study region, with concentrations occasionally reaching thousands of grains/m³ in a day (Figure 2). Over the study period, tree pollen peaked in late spring/early summer and weed pollen peaked in late summer/early fall. Grass pollen had two peaks, one in late spring/early summer and one in late summer/early fall.

Table 2.

3-day and 7-day moving averages for mean temperature, air pollution, relative humidity, precipitation, and pollen preceding study visits.

Exposure	3-day moving average Median (IQR)	7-day moving average Median (IQR)
Mean temperature (°C)	17.3 (10.6)	16.8 (11.0)
Air pollution
PM2.5 (μ/m3)	6.3 (3.3)	6.2 (2.6)
Ozone (ppb)	42.5 (9.0)	41.7 (7.7)
NO2 (μ/m3)	16.7 (11.1)	16.8 (9.7)
Relative humidity (%)	66.1 (15.8)	66.7 (11.4)
Precipitation (ml)	0.5 (3.3)	1.6 (3.7)
Pollen (grains/m3)
Total	19.0 (73.4)	20.9 (89.7)
Grass	0.9 (5.1)	0.9 (5.1)
Weed	1.4 (3.6)	1.6 (3.2)
Tree	3.0 (57.8)	2.8 (86.2)

PM_2.5: fine particulate matter with a mean aerodynamic diameter less than or equal to 2.5 microns; O₃: ozone; NO₂: nitrogen dioxide; IQR: interquartile range; ppb: parts per billion; ml: milliliters; m³: cubic meter.

Figure 2.

Generalized additive model^a with 3 degrees of freedom for the association of 7-day moving average for pollen (grains per cubic meter) and difference in log FeNO (log ppb). Dotted lines represent 95% CI. Black shading on x-axis reflects density of observations.

^aModel adjusted for early adolescent age, sex, height, weight, maternal education and household income at enrollment, any smoking in the home in early adolescence, race and ethnicity, and current asthma diagnosis, date of study visit, season (as sine and cosine functions of the study visit date), as well as precipitation, mean temperature, relative humidity, and air pollution (PM_2.5, O₃, and NO₂) in the same exposure window.

FeNO and lung function baseline characteristics

Mean FeNO was 26.2 ppb (SD 26.6), with 30% of participants with FeNO >35, and 22% with FeNO >50. For reference, FeNO >35 ppb is considered abnormal in children younger than age 12 and FeNO >50 is considered abnormal for anyone age 12 and older based on ATS guidelines.^18,21

Primary analyses

In fully adjusted models, 3- and 7-day moving average pollen were associated with higher FeNO in early adolescents (Table 3). We did not find associations of pollen with lung function (FEV₁ and FVC). There was a linear relationship between total pollen and natural log-transformed FeNO in the generalized additive model with 3 degrees of freedom (Figure 2). In subgroup analyses of the different types of pollen (tree, grass, weed), we found that all three types of pollen had similar direction in associations with FeNO, but the magnitude was greatest for tree pollen. By pollen type, associations of the 7-day moving averages with higher FeNO were 6.8% (95% CI 0.2, 13.8) for tree pollen, 2.2% (95% CI −0.8, 5.4) for grass pollen, and 1.6% (95% CI −2.0, 5.4) for weed pollen.

Table 3.

Associations of pollen (per IQR) with FeNO and lung function in early adolescence.

Pollen moving averages	Difference in FeNO (95% CI)	Difference in FEV1 (mL) (95% CI)	Difference in FVC (mL) (95% CI)
3-day	4.6% (0.1 to 9.2)*	−10.0 (−29.4 to 9.3)	2.8 (−16.8 to 22.3)
7-day	7.4% (0.9 to 14.3)*	−9.7 (−37.3 to 17.9)	7.0 (−20.8 to 34.8)

Linear regression models are adjusted for early adolescent age, sex, height, weight, maternal education and household income at enrollment, any smoking in the home in early adolescence, race and ethnicity, and current asthma diagnosis, date of study visit, season (as sine and cosine functions of the study visit date), as well as mean temperature, relative humidity, and air pollution (PM_2.5, O₃, and NO₂) in the same exposure window. IQR: interquartile range; CI = confidence interval; FEV₁ = forced expiratory volume in 1 second; FVC = forced vital capacity; FeNO = fractional exhaled nitric oxide.

^{* =}P value <0.05

Effect modification analyses

Associations of pollen with lung function were modified by allergic rhinitis history (P_interaction = 0.08) and current asthma diagnosis (P_interaction = 0.02) (E-Table 1 and E-Table 2). Among adolescents with asthma or allergic rhinitis, pollen exposure was associated with higher lung function (especially FVC), while associations were in the negative direction among adolescents without asthma and allergic rhinitis. Associations of total pollen with FEV₁ were modified by exposure to NO₂ for 3-day moving average pollen (P_interaction = 0.09) and by O₃ for 7-day moving average pollen (P_interaction = 0.05) (E-Table 3). With high NO₂ or O₃ air pollution, pollen exposure was associated with lower lung function compared to pollen exposure with low NO₂ or O₃ pollution. There was no evidence of effect modification by PM_2.5 for associations of total pollen with FeNO or lung function (P_interaction >0.1) (E-Table 3). We did not find that associations of pollen with FeNO differed by sex, mean temperature, or sensitization to indoor or outdoor aeroallergens, (P_interaction >0.1) (E-Tables 4-6). Lower precipitation augmented associations of pollen with lower lung function (P_interaction = 0.02 for 7-day pollen moving average and FEV₁) (E-Table 7).

Discussion

We found that short-term pollen exposure is associated with airway inflammation measured by FeNO in early adolescents in Massachusetts. This association was not isolated to those with asthma, allergic rhinitis, or allergic sensitization. Our findings argue that pollen exerts inflammatory effects on the airway even in the absence of conditions considered to confer higher susceptibility. These findings are supported by prior research demonstrating that exposure to pollen can stimulate nasal and serum inflammatory cytokines and nasal and respiratory symptoms even in nonatopic healthy individuals.^25-27

We found overall null results for associations between pollen and lung function. Testing for differential susceptibility based on asthma or allergic rhinitis diagnosis, we found paradoxical positive associations between pollen and lung function. We did not adjust for medication use, which may be an explanation for the apparent differential susceptibility. Adolescents with asthma or allergic rhinitis may use their inhalers, nasal sprays, and anti-allergy medications more frequently when pollen concentrations are high, thus inhibiting any bronchoconstrictive responses.²⁸

The literature on pollen and lung function is far from conclusive. A meta-analysis found pollen exposure by season, prior day exposure, and pollen load were not associated with changes in lung function,¹⁰ suggesting pollen can influence measurements of airway inflammation, such as FeNO, without influencing changes in spirometry.²⁹ The lack of association between short-term pollen exposure and lung function has been reported in healthy adolescents,²⁷ as well as in children with asthma.⁶ One explanation may be that pollen affects small to medium-sized airways preferentially, as demonstrated by associations between pollen and forced mid-expiratory flow (FEF_25-75%), which may be more sensitive at detecting airway inflammation and hyperresponsiveness than FEV₁.^9,30 It is also possible that the age of the individual and the type of pollen exposure influence responses measured by lung function. A study in Sweden found exposure to grass pollen in the prior 6 days was associated with lower lung function in children (age 8) but not in adolescents (age 16).³¹ In that study, they evaluated nine types of pollen, including species of tree and weed and found that the associations were isolated to grass pollen only. These differences in associations depending on the type of pollen (tree, grass, weed) may reflect differences in the allergenicity potential of different types of pollen.³²

We found that pollen exposure is associated with higher FeNO regardless of aeroallergen sensitization (i.e. aeroallergen sensitization to mold, plant fungi, and tree, grass, and weed species or common indoor aeroallergens did not modify associations between short-term pollen exposure and FeNO and lung function). Similarly, a German cohort study found that grass pollen was associated with higher FeNO in adolescents with and without aeroallergen sensitisation.²⁷ These findings suggest that pollen can trigger airway inflammation independent of the adaptive immune system, as noted in other studies.³³ Though pollen grains are typically thought to trigger the Th2-driven adaptive immune system response characteristic of atopic diseases, pollen grains also contain NADPH oxidases that can generate reactive oxygen species in the airways, triggering airway inflammation, regardless of atopic status.⁵ It is therefore plausible to see associations between pollen exposure and airway inflammation even in healthy adolescents, regardless of aeroallergen sensitization. This has implications for clinicians, who may consider healthy adolescents, that is, those without allergic rhinitis, asthma, or allergies, to be “immune” to pollen exposure. Together, this points to the complexity of differential pollen susceptibility by, for instance, the type of pollen, age, aeroallergen sensitization, and underlying health conditions (e.g., asthma).

A key difference between our study and prior studies is the emphasis on short-term pollen exposure in adolescents. Most research on airway inflammation has focused on the seasonal effects of pollen in populations considered higher risk, finding that FeNO is higher during the pollen season compared to outside the pollen season for adults with a diagnosis of asthma or allergic rhinitis.^29,34-36 A similar relationship between pollen season and higher FeNO was found in children with pollen allergies.³⁷ Yet, few studies have evaluated associations between short-term pollen and FeNO and lung function.^9,38,39 Understanding the short-term effects of pollen is relevant for public health because avoiding exposure to pollen over an entire season may not be practical. Instead, susceptible individuals might alter their behavior (e.g., wear a face mask or stay indoors with windows closed and air filtration devices) on days or weeks when pollen concentrations are high.

Our study has several limitations. The pollen concentrations assigned to participant addresses are based on a central pollen counter and may not reflect the actual pollen exposure of the participant. We included only participants living in Massachusetts to minimize differences in pollen concentrations, which have been shown to be similar within states and along latitudes.⁴⁰ Another limitation is missing pollen data, which occurred when the site was unable to record pollen counts (e.g., the designated reporter was on vacation or sick).

There are multiple strengths of this study. Our study includes a relatively large sample size of generally healthy adolescents. Additionally, our study is based on the Project Viva cohort, which is a well-characterized study population with aeroallergen-specific IgE, air pollution and weather data at a high resolution, as well as high quality FeNO data (best of 2 averages).

Conclusion

Exposure to pollen in the preceding week is associated with higher airway inflammation, as measured by FeNO, in generally healthy early adolescents. Our findings suggest that pollen promotes airway inflammation even in the absence of allergic sensitization and asthma.

Abbreviation list

FEF_25-75%

forced mid-expiratory flow

FeNO

fractional exhaled nitric oxide

FEV₁

forced expiratory volume in 1 second

FVC

forced vital capacity

GAMM

generalized additive mixed models

IQR

Interquartile range

NO₂

nitrogen dioxide

O₃

ozone

PM_2.5

fine particulate matter with a mean aerodynamic diameter less than or equal to 2.5 microns

PRISM

Parameter-elevation Relationships on Independent Slopes Model

US

United States

Data availability statement

The data underlying this article are available on reasonable request to the corresponding author and appropriate approval from the Project Viva team and Institutional Review Board. The data are not publicly available due to privacy or ethical restrictions.