Short-Term Exposure to PM2.5 and Emergency Department Visits for Asthma and Chronic Obstructive Pulmonary Disease Authors: Anagh Kulkarni, Yohei Okada, Joel Aik, Fahad Javaid Siddiqui, Mariko Siyue Koh, Fintan Hughes, Tetsu Ohnuma, Vijay Krishnamoorthy, Marcus Eng Hock Ong Published: 2026 ABSTRACT This cohort study evaluates the association between short-term fine particulate matter (PM2.5) exposure and asthma- and chronic obstructive pulmonary disease (COPD)–related emergency department visits in Singapore. Introduction Fine particulate matter (PM2.5) is consistently linked to acute cardiopulmonary morbidity, but evidence on emergency department visits (EDVs), which capture earlier clinical deterioration than mortality or admission outcomes, remains comparatively limited.1,2 This evidence base is also geographically uneven, concentrated in temperate settings with different climates, pollution sources, and health care access relative to tropical Southeast Asia.3,4 Singapore’s environment involves chronic low-level urban PM2.5 exposure with episodic transboundary haze; average PM2.5 concentrations oscillate near current World Health Organization (WHO) guideline thresholds with intermittent spikes rather than the higher baseline levels typically seen in the existing literature pool. We examined the association between short-term PM2.5 exposure and asthma- and chronic obstructive pulmonary disease (COPD)–related EDVs in Singapore. Methods This retrospective, nationwide, ecological time-series cohort study included all eligible EDVs and PM2.5 data (February 2016 to December 2022) available at time of analysis (January to June 2026). Daily counts of all eligible asthma or COPD-related EDVs, same-day hospital admission, and 30-day mortality among adult Singapore citizens and permanent residents (aged ≥20 years) were extracted from all public hospitals via a national government administrative and health data platform using primary International Classification of Diseases, Ninth Revision (ICD-9) and Tenth Revision (ICD-10) diagnosis codes. A single national daily mean PM2.5 value, accessed through the NEA monitoring network,5 was referenced to 15 μg/m3 (2021 WHO 24-hour guideline), with missing values (4.3% of days) linearly interpolated. Temperature, humidity, and rainfall data were obtained from the same source.5 All analyses used fully deidentified administrative data; the requirement for informed consent was waived accordingly and the study was approved by the National University of Singapore institutional review board (NUS-IRB-2024-1078). The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies. Counts were modeled using quasi-Poisson regression with a distributed lag nonlinear model (DLNM)6 cross-basis for PM2.5 (3 df, lag 0-7) to estimate relative risks (RRs) and 95% CIs, adjusting for calendar time (cubic spline, 9 df per year), month, day-of-week, public holidays, and the COVID-19 pandemic period (2020-2021), with temperature and humidity as cross-basis terms. A single-pollutant model was used, consistent with standard practice given high correlation among ambient pollutants. A lag of 0 to 2 days was the primary window, based on observed risk stabilization. Effect modification by sex, age, income, and comorbidity burden was assessed via stratification. Analyses were conducted in R using the dlnm package (R Project for Statistical Computing). Statistical significance was determined based on whether the 95% CI included the null value (1.0). Methodological details are found in the supplemental documentation (eMethods 1 and 2 in Supplement 1). Results A total of 74 839 asthma and/or COPD-related EDVs were identified (49 318 [66.1%] asthma and 25 327 [33.9%] COPD). The median (IQR) age of patients was 59.0 (39.0-73.0) years, 44 870 (59.9%) were male, 73 189 (97.8%) had documented comorbidities, and 56 320 (75.3%) had a monthly household income of less than 2000 Singaporean dollars. Temporal trends in asthma and COPD EDVs and exposure distribution are described in Figure 1. Hospital admission occurred in 43 446 presentations (58.1%); 30-day mortality occurred in 1369 (1.8%). Line Graphs of Daily Temporal Trends of Environmental Variables and Emergency Department Presentations in Singapore From February 2016 to December 2022 Spikes in late 2016 and 2019 in panel A correspond with transboundary haze periods thought to be driven by Indonesian wildfires. The drop in counts observed in 2020 and 2021 in all outcomes was likely due to COVID-19 effects; counts trended back upward toward those seen in pre-COVID years in 2022. COPD indicates chronic obstructive pulmonary disease; ED, emergency department; PM2.5, fine particulate matter. PM2.5 showed a positive, nonlinear association with combined, asthma-only, and COPD-only EDVs, with risk increasing continuously above the 15 μg/m3 reference and cumulative risk strengthening from lag 0 to lag 0 to 2 before plateauing (Figure 2). At a 10 μg/m3 increment above reference, combined asthma and/or COPD EDV risk increased nonsignificantly at lag 0 (RR, 1.01; 95% CI, 0.99-1.03) and significantly by 3.2% at lag 0 to 2 (RR, 1.03; 95% CI, 1.00-1.06). Asthma-only EDVs showed a nonsignificant same-day increase (RR, 1.02; 95% CI, 0.99-1.04) and significant 4.4% cumulative lag 0 to 2 increase (RR, 1.04; 95% CI, 1.01-1.08). COPD-only EDVs showed nonsignificant increases at lag 0 (RR, 1.00; 95% CI, 0.97-1.04) and lag 0 to 2 (RR, 1.01; 95% CI, 0.96-1.06), with both becoming distinguishable from reference at approximately 55 μg/m3 (RR, 1.11; 95% CI, 1.01-1.23; RR, 1.15; 95% CI, 1.00-1.31, respectively). Hospital admission risk increased nonsignificantly at lag 0 (RR, 1.01; 95% CI, 0.99-1.04) and lag 0 to 2 (RR, 1.03; 95% CI, 0.99-1.06). No association was observed with 30-day mortality and no evidence of effect modification by sex, age, income, or comorbidity burden was identified. Line Graphs of Distributed Nonlinear Lag Models Demonstrating Immediate (Lag 0) and 0- to 2-Day Cumulative Lagged Associations Between Fine Particulate Matter (PM2.5) and Respiratory Emergency Department Visits (EDVs) Shaded areas indicate 95% CIs. COPD indicates chronic obstructive pulmonary disease; RR, relative risk. Discussion This nationwide cohort study found that short-term increases in PM2.5 were associated with higher asthma- and COPD-associated EDVs in Singapore, with cumulative risk concentrated in the 2 days following exposure, consistent with the early lag pattern reported in prior literature.7 These findings extend short-term PM2.5-EDV evidence to a tropical Southeast Asian setting with a distinct pollution source profile and comparatively low ambient concentrations (mean 16.2 μg/m3), addressing a geographic gap in an evidence base concentrated in temperate, high-pollution regions, and are consistent with emerging evidence that respiratory risk remains detectable well below historically studied exposure levels.4,8,9 Associations for asthma and COPD were directionally similar to prior meta-analytic estimates, although COPD estimates were less precise.1,2,10 Hospital admissions showed a similar pattern; the absence of a mortality association was expected given the low baseline event rate. There were several limitations. First, exposure was assigned at the population level, precluding individual-level assessment. Second, analyses were restricted to public EDs (primary and urgent-care visits excluded), although these capture most of Singapore’s emergency care. Third, a single-pollutant model was used, consistent with standard practice, but we cannot exclude copollutant confounding (eg, nitrogen dioxide or sulfur dioxide). Lastly, a binary COVID-19 indicator may not have fully captured pandemic-related disruption; residual confounding cannot be excluded. Overall, these findings support integrating real-time air quality data into respiratory ED surge planning and patient-facing alerts, even in lower-pollution settings.