Long-Term PM2.5 Mortality Studies

The science of PM2.5 and premature death rests on a handful of very large cohort studies that follow people for years, measure their long-term exposure to fine particles, and count who dies and why. After the Harvard Six Cities Study opened the field in 1993, the cohorts grew bigger and richer, and the central finding kept getting more confident: long-term exposure to fine particles shortens lives, with no clear safe level.

Background

A cohort study follows a defined group of people forward in time. For air pollution, the key variable is long-term average PM2.5 at each person's home address, which researchers estimate using monitor data plus satellite retrievals and land-use models. The endpoint is usually all-cause or cause-specific mortality. Three cohort families have done most of the heavy lifting: the Harvard Six Cities cohort, the American Cancer Society's Cancer Prevention Study II (CPS-II), and the U.S. Medicare cohort. Europe has its parallel in the ELAPSE consortium.

What the studies found

American Cancer Society CPS-II (Pope et al., JAMA 2002 and Turner et al., AJRCCM 2016). Roughly 1.2 million adults enrolled in 1982, tracked for decades. Each 10 microgram per cubic meter increase in long-term PM2.5 was associated with about a 6 percent increase in all-cause mortality, a 9 percent increase in cardiopulmonary mortality, and a 14 percent increase in lung cancer mortality. The 2016 extension found similar effects with finer-resolution exposure data.

Medicare cohort (Di et al., NEJM 2017). Joel Schwartz, Francesca Dominici, and colleagues at Harvard linked 60 million Medicare beneficiary records to address-level PM2.5 estimates from 2000 to 2012. Each 10 microgram per cubic meter rise in annual PM2.5 was associated with a 7.3 percent increase in all-cause mortality among adults 65 and older. The slope was steepest at low concentrations: the risk per microgram was higher below 12 micrograms per cubic meter than above it.

ELAPSE (Strak et al., BMJ 2021, and HEI report 2022). Bert Brunekreef, Maciej Strak, and a European consortium pooled eight cohorts covering 325,000 people and analyzed seven administrative cohorts covering 28 million more. Even after restricting to people exposed below the old WHO PM2.5 guideline of 10 micrograms per cubic meter, mortality rose with exposure. The 2022 results were cited explicitly in the European Commission's case for tightening EU air quality standards.

Canadian Census Health and Environment Cohort (CanCHEC) (Pinault et al. and others). Follow-up of millions of Canadians at PM2.5 levels often below 8 micrograms per cubic meter also showed continued mortality risk, with non-linear concentration-response that steepens at low exposures.

Why it matters

These studies are the empirical backbone for the U.S. National Ambient Air Quality Standards, the WHO Global Air Quality Guidelines, and the new EU Ambient Air Quality Directive. The repeated finding of no observable threshold reshaped how regulators think about PM2.5: it is now treated more like radiation or lead than like a classical toxin with a safe dose.

The Burnett Global Exposure Mortality Model (PNAS 2018), which feeds the Global Burden of Disease estimates, pulled these cohorts together with Chinese and other non-Western data to produce the curve used for the 8 million annual deaths estimate.

Open questions

How much of the residual mortality signal at very low PM2.5 (below 5 micrograms per cubic meter) is real versus statistical noise from exposure misclassification is still debated. Researchers including Daniel Krewski (Ottawa) and Doug Dockery have argued the low-end signal is robust; others note that satellite-derived exposure estimates have larger relative error at low concentrations. Causal inference methods (instrumental variables, regression discontinuity around regulatory actions) are increasingly used to strengthen the case.

Sources

• Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution (Pope et al., JAMA 2002)
• Long-term ozone exposure and mortality in a large prospective study (Turner et al., AJRCCM 2016)
• Air Pollution and Mortality in the Medicare Population (Di et al., NEJM 2017)
• Long term exposure to low level air pollution and mortality in eight European cohorts within the ELAPSE project: pooled analysis (Strak et al., BMJ 2021)
• Mortality and Morbidity Effects of Long-Term Exposure to Low-Level PM2.5, BC, NO2, and O3 (HEI ELAPSE Report, 2022)
• Global estimates of mortality associated with long-term exposure to outdoor fine particulate matter (Burnett et al., PNAS 2018)
• Extended Follow-Up and Spatial Analysis of the American Cancer Society Study (HEI 2009)