Is a Microgram of Smoke Particulate Matter with an Aerodynamic Diameter ≤2.5 μm Worse for Respiratory Health? Interpreting New Evidence

Air quality has long been recognized as important for respiratory health. But when sources of particulate matter with an aerodynamic diameter ≤2.5 μm (PM_2.5) change, the same concentration can carry different risks because composition and origin also matter. Over the past two decades, large declines in transportation and industrial emissions have been partially offset, and in many places reversed, by emissions from increasingly frequent and severe wildfires (1). A growing share of the PM_2.5 we now breathe comes from smoke. As that mix shifts, the question for practice, policy, and public health is not only whether smoke worsens respiratory health (it does), but whether the same PM_2.5 concentration carries different clinical risk when the mass is smoke-dominated.

Some have argued that wildfire PM_2.5 should be less toxic per microgram because biomass smoke is more carbonaceous and metal poor than fossil fuel PM_2.5 (2). Yet field and laboratory evidence shows that ambient smoke is not “clean.” Fresh and aged smoke contains oxidized organics and quinones with measurable oxidative potential, ultrafine particles that reach small airways, and soil- and ash-derived metals (3, 4); smoke aging can alter oxidative potential (3). Wildland–urban interface fires add plastics, treated wood, electronics, and other synthetics, introducing additional metals and halogenated organics and plausibly increasing hazard (5). Toxicology supports mechanisms for acute airway irritation, inflammatory signaling, and systemic oxidative stress from both smoke and nonsmoke PM_2.5 (6) but does not yield a definitive per-microgram toxicity ranking. Epidemiology agrees: both total PM_2.5 and wildfire smoke increase short-run respiratory morbidity and healthcare use, with especially consistent signals for asthma (7–9).

In this issue of the Journal, Wang and colleagues (pp. 2086–2095) add evidence to this conversation using an individual-level, time-stratified case-crossover analysis of more than 6 million respiratory-related emergency department (ED) visits across six western U.S. states during wildfire seasons (10). Each encounter is compared with itself on nearby referent days, thereby controlling for fixed patient characteristics and slow-moving time trends. Combined with daily, spatially resolved, source-specific PM_2.5 concentrations, the design estimates marginal, per-microgram associations for smoke and nonsmoke PM_2.5.

For asthma ED visits, the authors report 1.016 times higher odds per 1 μg/m³ smoke PM_2.5 (three-day average) versus about 1.002 for nonsmoke PM_2.5. ED visits for chronic obstructive pulmonary disease, bronchitis, and upper respiratory infections show smaller, positive associations. The nonsmoke asthma estimate sits within the range reported by all-source PM_2.5 meta-analyses (7), while the smoke-specific asthma estimate is clearly larger.

How should readers interpret the smoke-versus-nonsmoke contrast? Because both components are modeled together, each estimate answers the question: if smoke increases by 1 μg/m³ while nonsmoke stays the same (and vice versa), how do the odds of an ED visit change? The larger smoke effect means that an additional microgram is associated with higher odds of a visit compared with an additional microgram of nonsmoke PM_2.5; it does not, by itself, mean that smoke causes more total visits: that also depends on how much of each type people inhale.

The key interpretive issue is whether this steeper slope reflects intrinsic toxicity or the context in which smoke occurs. Consider the hypothetical scenario that the PM_2.5–risk curve is steeper at the concentrations when and where smoke tends to occur. Here, a 1 μg/m³ smoke increase would appear more toxic than a 1 μg/m³ nonsmoke increase without actually being more toxic. Even with equal toxicity, though, in this scenario smoke PM_2.5 would increase healthcare use more than comparable nonsmoke PM_2.5 increments.

Measurement and scope matter too. First, in the context of higher healthcare use from smoke PM_2.5, it is worth emphasizing that ED visits are a proxy for morbidity: smoke is salient—people smell it, see it, and receive warnings—so care seeking may differ on smoky days for the same underlying symptoms. Second, wildfire smoke concentrations are not measured directly; they must be inferred from satellites, models, and monitors. If some true smoke mass is misclassified as nonsmoke, the per-microgram smoke effect can be biased upward because each labeled microgram stands in for more than a microgram in reality. That could amplify a real difference, but is unlikely to explain the large differences found by Wang and colleagues (10), particularly given their use of a carefully designed smoke PM_2.5 product. Third, this analysis is focused on May to October (“wildfire season”), when meteorology, chemistry, copollutants, and people’s behavior (staying indoors, using filtration, masking) can differ from other seasons; therefore, these estimates should be interpreted as season specific.

A strength of this study is its stratified, individual-level analysis. Across groups, smoke PM_2.5 is associated with higher respiratory ED use, with the largest responses for asthma. Female subjects show a modestly higher asthma association than male subjects; by age, adults tend to have slightly higher odds for some outcomes, but children are neither uniquely spared nor uniquely at risk on a per-microgram basis. Race/ethnicity patterns are mixed. State estimates vary—larger point estimates in Arizona and Nevada, null effects for select outcomes in Utah and Oregon—but precision is limited outside California’s much larger sample.

Taken together, Wang and colleagues (10) present one of very few carefully designed studies to estimate the differential impacts of smoke and nonsmoke PM_2.5. Given this still limited body of evidence (11), more work that conducts like-for-like comparisons across sources and applies quasi-experimental approaches is needed to confirm these findings in other settings. By pooling data from multiple states, Wang and colleagues provide credible evidence that wildfire smoke increases short-term respiratory ED visits across a wider range of smoke conditions than prior single-state work (12–15). Their finding that impacts are worse for smoke PM_2.5 is important. Whether that higher effect reflects intrinsic toxicity, exposure dynamics, behavioral responses, some other factors, or a combination remains uncertain yet is increasingly important as fires intensify and wildland–urban interface burns produce more complex mixtures.