Pesticides as an overlooked exposomic association in allergic asthma exacerbations: A nationwide database study

Abstract

Occupational pesticide exposure (PE) is a causal factor for asthma, but there is only low certainty of evidence for non-occupational PE, considered a minor exposomic risk.^1-3 Pesticides share the One Health pathophysiology of exposome-related diseases: leaky barrier due to epithelial lesions and inflammation driven by epithelium-derived alarmins. Leaky barriers allow further penetration of noxious molecules while the vicious circle of chronic inflammation, most often type 2-skewed, fosters and exacerbates adaptive allergic responses to common airborne allergens.^3-5 Direct toxicity, allergenicity, and epigenetic effects have also been described.^{4, 6-9} Thus, the association between PE and asthma can be assessed as “incidence” (asthma development in exposed populations) or as “impact on asthma outcomes” (exacerbation, lung function deterioration etc).^{1, 2} Our study belongs to the second category. The lack of general population cohorts addressing the relationship between PE and asthma outcomes prompted us to build a nationwide dataset (F20210106180024/2203674v, French healthcare data studies registry) using four French databases (Figure 1; Supporting Methods—Data S1). Institutional Review Board approval was waived for this retrospective observational study with anonymized data.

Cases (n = 21,050) were defined as 2014–2020 stays with an ICD-10 primary diagnosis of allergic or mixed asthma (J45.0, J45.8). Controls (n = 8782) were defined as stays with a primary diagnosis of non-allergic asthma (J45.1) during the same period (Figure S1). Cases were matched 1:1 with controls on age, sex, comorbidities, month and year of the hospital stay (Supporting Methods—Data S1). Patients' residence place was encoded from ZIP code (Table S1). PE was estimated as the index of non-occupational annual pesticide purchases per ZIP code area. Matched (n = 12,716) cases and controls were compared according to exposomic variables: PE, lifestyle, and socio-economic indicators (Figure 1; Table S2). The comparison between the two groups: cases (allergic and mixed asthma) and controls (non-allergic asthma) is presented using odds-ratio (OR) to estimate the association between exposure and outcome.

Univariate analysis revealed higher exposure to any pesticide in cases: 7.81 ± 18.6 kg/ha mean ± SD in cases versus 6.50 ± 18.1 kg/ha in controls, OR 1.07, p < .001 (Table 1). The association was significant for each subgroup of pesticides, with higher exposure in cases than in controls: insecticides (OR 1.09, p = .007), fungicides (OR 2.45, p = .002) and herbicides (OR 1.84, p = .002).

In multivariate analysis, after adjustment on age, sex, and comorbidities (Table 1), and on social deprivation (Table S3), significantly higher exposure persisted in cases compared to controls for any pesticide (OR = 1.06, p = .009) and for pesticide subgroups, with highest OR of 1.96 for fungicides. However, adjusting on rural residence canceled PE significance (Table S3). Rural residence as a risk factor for high non-occupational PE and possible links to respiratory health outcomes was consistent with previous reports.^{3, 10} Since urban versus rural residence appeared as an effect modifier, urban residents were analyzed separately. Among urban residents, exposure to any pesticide was significantly higher in cases than in controls (3.94 ± 15.81 kg/ha versus 2.35 ± 6.50 kg/ha, OR = 1.41, p = .006, Table 1). Herbicide and fungicide exposure ORs were 18 (p < .001) and 58 (p = .03) in urban residents, although wide CI95% intervals warrant confirmation.

Our findings challenge the current view of harmless non-occupational PE^1-3 by exposing the association between PE and hospital stays for allergic versus non-allergic asthma. They support a contribution of PE to hospital stays for allergic asthma. Moreover, higher PE may be an unrecognized determinant of hospital stays for allergic asthma in urban residents, potentially related to the increased prevalence of allergic diseases and asthma in socially disadvantaged urban neighborhoods.

The strengths of our study include the nationwide and unbiased design of the study population, subsequently matched as allergic versus non-allergic asthma hospital stays, combined with PE and lifestyle data, and the methodological validation of exposomic health risk studies combining open data from available repositories.

Limitations include selection bias toward hospitalized patients, non-differential misclassification bias due to ICD-10, and indirect PE evaluation based on the amount of non-occupational purchases. Missing data on smoking and occupation are major methodological limitations. While smoking exposure is considered a significant confounder for assessing asthma outcomes, neither smoking exposure nor occupational PE are known to preferentially associate with allergic rather than non-allergic asthma outcomes.^{1, 2, 11} In a large study assessing lung function decline, a long-term characteristic of asthma, in over 17,000 participants, smoking status did not influence the effects of occupational exposure.¹² Moreover, smoking is closely associated with low socio-economic conditions, which were addressed in this study. The most significant occupational PE occurs in agricultural workers, who make only 2.5% of the French adult population (https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Farmers_and_the_agricultural_labour_force_-_statistics).

Considering our identification of higher PE in patients hospitalized for allergic asthma compared to patients hospitalized for non-allergic asthma, and as an exposomic risk in urban populations, further studies at the personal and neighborhood level (asthma endotype, actual PE at the chemical level, housing characteristics, green spaces proximity and quality, pest infestation) are warranted in addition to cross-referencing administrative, socio-economic, environmental and hospital databases. From a One Health perspective, further characterization of the pesticide exposomic risk is an obvious unmet need for improved personalized strategies for asthma prevention, diagnosis and management.

JV: conceptualization, project supervision, validation, writing—first draft; writing—final draft revision and approval, LB: data curation, formal analysis, writing—first draft, BNP: conceptualization, validation, writing—final draft revision and approval, CB: conceptualization, methodology, writing—final draft revision and approval SS: conceptualization, methodology, project supervision, writing—first draft; writing—final draft revision and approval.

No specific funding was allocated for this study. All authors are employees of the University of Reims Champagne-Ardenne, Reims, France.

JV reports speaker and consultancy fees in the past 5 years from Astra Zeneca, HpVac, L'Oréal, Novartis, Sanofi, Thermo Fisher Scientific, and travel support from Stallergènes-Greer, outside the submitted work. The other authors declare no competing interests in relation to this study.