Association between ambient air pollutants and major infectious diseases in Singapore

Introduction

Infectious diseases remain a major cause of morbidity and mortality worldwide, posing significant challenges to public health, especially in low- and middle-income countries. These diseases are caused by a variety of pathogens, including bacteria, viruses, and parasites, and can be exacerbated by environmental factors. Among these factors, air pollution has been identified as a significant risk. It is however unknown how mixtures of ambient air pollutants affect different infectious diseases with different transmission pathways. To address this gap, this study investigates the nonlinear and potential interactive association between ambient air pollutants mixtures and multiple infectious diseases. Infectious diseases chosen were those which had the highest reported burden in Singapore and were plausibly affected by ambient air pollutants.

Methods

We harmonized weekly data on ambient air pollutants (PM_2.5, PM₁₀, SO₂, NO₂, O₃, and CO), environmental exposures such as rainfall, absolute humidity and mean temperature as well as weekly disease surveillance data from 2012 to 2019. We utilized generalized linear models (GLMs) and generalized additive models (GAMs) to examine both linear and non-linear associations between pollutants and disease incidences, adjusting for confounders, lagged effects, and autocorrelation. Incidence rate ratios (IRRs) and excess incidence ratios (EIRs) were derived to interpret exposure–response relationships. Additionally, we conducted a sensitivity analysis using Gaussian Process (GP) regression with various kernel functions and five-fold cross-validation to assess model robustness and potential interactive effects among pollutants.

Results

Our analyses revealed significant associations between pollutant concentrations and several disease EIRs. High PM₁₀ levels were linked to an immediate increase in the incidence rates compared to the reference level for acute conjunctivitis and acute upper respiratory infections. Elevated SO₂ concentrations were associated with higher contemporaneous incidence rates for acute conjunctivitis and varying effects for Hand, food, and mouth disease (HFMD) depending on concentration levels and the time lag. NO₂ concentrations had delayed effects on HFMD at 1-week and 4-week lags, and the effects were such that as concentration increased EIR decreased. CO and O₃ showed minor effects on the infectious diseases studied. No significant interactive effects between pollutants were found.

Conclusion

Specific pollutant concentration thresholds influence the incidence of various infectious diseases. Targeted air quality management strategies are essential to mitigate public health risks. The absence of interactive effects simplifies the design of policies aimed at reducing individual pollutant levels.