Pollen exposure and respiratory illness in a changing climate

Background: Pollen production is sensitive to other environmental factors like seasonal temperature, precipitation and carbon dioxide. Pollen with allergenic potential affects respiratory pathways of susceptible individuals and leads to a range of adverse health outcomes. With climate change, some of these environmental factors will change, with subsequent repercussions on pollen production and respiratory health outcomes.

Objective: The objectives of this research were to examine the (i) Effects of inter-annual seasonal climate variability on pollen production, using pollen concentration data observed daily from one station located in the metropolitan Atlanta area. (ii) Associations between three different pollen types (tree, grass, and ragweed) on outpatient visits for Allergic Rhinitis and allergy medication sales.

Methods: (i) We developed linear regression models for a known subset of allergic plants, to determine if significant year-to-year relationships exist between seasonal climate and 1) pollen count, 2) the timing of pollen season onset, 3) season peaking, 4) season ending, and 5) and the length of season. (ii) Daily concentrations for tree, grass, and ragweed pollen were obtained from monitors each in Atlanta and Baltimore metropolitan areas for 2007-09. Patient-level health data was obtained from an administrative health dataset (Marketscan®) for those years. Episodes and dates of outpatient visits for Allergic Rhinitis (ICD9 code: 477) and allergy medication refills (generic and brand name drugs) were extracted for patients who resided in any county within the metropolitan areas. We conducted a case-crossover analysis with a time-stratified approach to selecting controls. Pollen concentration on each case-day (episode of outpatient visit/medication refill) was compared with pollen concentration on control-day(s) chosen for the same day of week within the half-month period as the episode. We controlled for potential environmental confounders like daily average concentrations of O3, PM2.5, temperature, precipitation, and wind in order to identify any association between elevated pollen level and adverse health outcomes.

Results: (i) We found statistically significant associations between seasonal climate and pollen production, but results varied according to specific allergenic plant species. Generally, pollen production was preceded by cold season conditions (e.g. fall freezing days, winter rainfall), rather than the climatic conditions occurring during a typical pollen release (e.g. spring). (ii) We analyzed 192,436 and 137,755 episodes of Allergic Rhinitis related outpatient visits; 268,647 and 108,886 allergy medication refills. The different pollen seasons for the 3 types were identified based on the observed pollen data. Tree (3-day moving average), grass (5-day lag), and ragweed (7-day lag) pollen concentrations were most strongly associated with Allergic Rhinitis outpatient visits and refills of allergy medication. When pollen concentrations were classified into four categories based on percentile thresholds, we found a non-linear dose-response relationship between health episodes and pollen categories. Sensitivity analysis using a one-month period for selection of controls yielded consistent results.

Conclusion: Developing models to quantify the relationship between seasonal climate and allergenic pollen production can aid clinicians in formulating strategies (e.g. exposure prevention or immunotherapy) that are tailored to patients with specific pollen hypersensitivities (e.g. oak, birch, or ragweed), in addition to the development of more precise seasonal pollen projections. Using a health dataset and pollen data from independent monitors, the preliminary results suggest associations between allergy outcomes and elevated pollen levels.