Long-term air-pollution research is strong enough to support major public-health action, but it is often communicated badly. Population associations become personal predictions, relative risks are reported without a baseline, and an annual city average is treated as if it were one person's measured dose.
This review separates established findings from emerging evidence and explains what the numbers can—and cannot—tell you.
What “Long-Term Exposure” Means
Long-term studies usually estimate average outdoor pollution at a home address over months or years. Researchers link those estimates to health records while adjusting for factors such as age, smoking patterns, income, temperature, and pre-existing disease.
The exposure is not the same as personal dose. People move, work indoors, use different ventilation and filtration, and encounter pollution during travel. Outdoor models also have measurement error. Well-designed cohort studies address many confounders, but they are not randomized trials.
That is why evidence is evaluated across many studies, locations, methods, outcomes, and biological mechanisms rather than from one dramatic paper.
The Strongest Evidence: PM2.5 and Mortality
The systematic review commissioned to support the 2021 WHO guidelines screened more than 3,000 abstracts and included 107 studies, mostly cohorts. Its pooled estimate associated each 10 µg/m³ higher long-term PM2.5 concentration with an 8% higher relative risk of natural-cause mortality (risk ratio 1.08; 95% confidence interval 1.06–1.09). The review rated the evidence high certainty for most PM2.5 mortality outcomes and moderate certainty for respiratory mortality. The methods, outcome definitions, and DOI are available in Chen and Hoek, Environment International (2020).
This is a relative population estimate, not a statement that eight of every 100 exposed people will die because of pollution. Absolute risk depends on age, baseline health, exposure duration, competing risks, and the population being studied.
A large U.S. cohort provides a useful example rather than a universal conversion factor. Di and colleagues followed almost 61 million Medicare beneficiaries from 2000 through 2012. A 10 µg/m³ higher annual PM2.5 estimate was associated with 7.3% higher all-cause mortality after adjustment for individual and area-level variables. The original article is Air Pollution and Mortality in the Medicare Population (DOI: 10.1056/NEJMoa1702747).
The cohort was composed of older U.S. adults, so its percentage should not be pasted onto a child, a resident of another country, or a specific year's AQI. Its importance is that the association persisted in analyses restricted to concentrations below the U.S. standard in effect at the time.
Evidence by Health Outcome
Cardiovascular disease: high confidence
The evidence links long-term PM2.5 exposure with cardiovascular mortality, ischemic heart disease, and stroke. The WHO guideline review incorporates epidemiologic evidence alongside plausible pathways including inflammation, oxidative stress, altered blood-vessel function, autonomic effects, and clot formation. EPA's current scientific summary explains those pathways and the evidence for long- and short-term cardiovascular effects on its PM2.5 cardiovascular research page.
This does not mean PM2.5 is the only cause of a heart attack or stroke. It means higher population exposure shifts the probability of these outcomes.
Respiratory disease: strong, with outcome-specific uncertainty
Long-term particle exposure is associated with respiratory mortality and chronic respiratory disease. Short-term pollution can also aggravate asthma and other lung conditions, but an acute asthma attack and the development of chronic disease are different research questions. The Chen and Hoek review rated PM2.5 evidence for respiratory mortality moderate rather than high, illustrating why every outcome should not be presented with the same certainty.
Lung cancer: established hazard
The International Agency for Research on Cancer classified outdoor air pollution as carcinogenic to humans (Group 1) after an expert review, with sufficient evidence for lung cancer. IARC states that conclusion directly in its official 2013 announcement.
“Carcinogenic” identifies a hazard supported by evidence; it does not mean every exposed person will develop cancer. Individual risk depends on cumulative exposure and many other factors, especially tobacco smoking.
Pregnancy, birth, and child development: important associations
WHO summarizes links between long-term particle exposure and adverse perinatal outcomes. Study designs, exposure windows, and outcomes vary, so a city annual mean cannot predict an individual pregnancy result. Public-health guidance appropriately treats pregnant people and children as groups for whom exposure reduction can be especially important; individual clinical questions belong with a qualified health professional.
Diabetes, cognition, and neurodegenerative disease: developing evidence
Research has reported associations with metabolic and neurological outcomes, but certainty differs by endpoint and study design. These topics should be described as developing evidence unless a current systematic review or authoritative causal assessment supports a stronger statement. It is not responsible to claim that a particular AQI causes dementia, Parkinson's disease, depression, or a quantified loss of intelligence in an individual.
Is There a Safe Threshold?
The evidence does not identify a concentration below which population risk is proven to be zero. The 2020 WHO review found similar or larger pooled relative risks in subsets of studies conducted at lower mean PM2.5 concentrations. That supports continued reductions even in comparatively clean regions.
However, “no observed zero-risk threshold” is not the same as “every tiny exposure produces measurable harm in every person.” Epidemiologic estimates become more uncertain at the edges of the observed exposure distribution, and unavoidable background particles exist.
WHO's annual PM2.5 guideline is 5 µg/m³, with interim targets of 35, 25, 15, and 10 µg/m³ to support stepwise improvement. The values and evidence process are documented in the 2021 WHO Global Air Quality Guidelines. The guideline is health-based advice, not a guarantee of zero risk.
Why “Years of Life Lost” Is Not a Personal Forecast
Burden-of-disease models can estimate changes in population life expectancy under explicit assumptions. They combine concentration surfaces, baseline mortality, demographic data, and concentration-response functions. Those models are useful for policy, but their result is not an appointment date for an individual.
A claim such as “living in this city will remove exactly six years from your life” hides several uncertainties:
- Future pollution may rise or fall.
- A city average is not personal exposure.
- Population concentration-response estimates are not deterministic.
- Smoking, healthcare, occupation, housing, and underlying health differ.
- Models may use different counterfactual concentrations and endpoints.
For the same reason, multiplying one day's AQI by 365 is not a valid chronic-risk model.
Annual Mean, Daily AQI, and Personal Exposure
These measures should remain separate:
- Annual mean PM2.5 (µg/m³): a long-term concentration used in cohort studies, guidelines, and standards.
- Daily AQI: a pollutant-specific communication index for short-term conditions and activity guidance.
- Personal exposure: the concentration along a person's time-location pattern.
- Inhaled dose: affected by breathing rate, duration, particle behavior, and protective equipment.
AQI is useful for deciding when to move a workout or reduce prolonged outdoor exertion. It cannot diagnose cumulative disease, and it should not be averaged into a personal mortality prediction.
What Reduces Long-Term Risk?
The largest durable benefits come from population-level emissions reduction: cleaner power and heating, industrial controls, low-emission transport, reduced open burning, and policies addressing inequitable exposure. Personal actions can reduce some exposures, especially during episodes, but they cannot replace clean-air policy.
Evidence-aligned personal steps include:
- Use the responsible local authority's pollutant-specific current data and forecast.
- Avoid adding indoor particles through smoking, incense, candles, and unvented combustion.
- Use appropriately sized particle filtration when outdoor smoke or chronic local PM2.5 is a concern.
- Move prolonged heavy activity away from high-traffic roads and high-pollution periods when practical.
- Follow a clinician's advice if you have heart or lung disease, are pregnant, or develop symptoms.
These steps reduce exposure; they do not promise a specific percentage reduction in disease.
Where Cigarette Equivalents Fit—and Where They Do Not
The cigarette-equivalent model on this site is a risk-communication heuristic for a U.S. PM2.5 AQI and a stated duration. It is not a biological dose comparison, a mortality model, or a prediction of long-term disease.
Do not use it with annual PM2.5, ozone AQI, PM10 AQI, an unknown dominant pollutant, or an AQI from another national system. Review the methodology and keep the result separate from clinical advice.
Bottom Line
The long-term PM2.5 evidence is strongest for mortality, cardiovascular disease, respiratory outcomes, and lung cancer. Systematic reviews find associations across a wide concentration range, including relatively low levels. The responsible interpretation is population-level: reducing sustained exposure lowers risk. The evidence does not support sensational claims that one AQI reading proves organ damage or that a city average can predict an individual's exact disease or lifespan.