The left hand rests on the hospital blanket, fingers slightly curled, not quite obeying. Yesterday it held a coffee cup. The chart reads cerebral infarction. ICD-10 code I63.9: unspecified. The attending notes hypertension, atrial fibrillation, age. Medications adjusted, follow-up scheduled, rehabilitation referral. Everything documented is accurate.
The patient lived in the Sacramento Valley for eleven years. During those years, wildfire smoke settled into the valley with increasing regularity each summer and fall, pooling between the Sierra Nevada and the Coast Range the way cold air pools in winter, except this was particulate matter fine enough to pass through walls. The patient breathed that air. The air entered the blood. The blood carried what the air carried to the vessels feeding the brain, and this happened not once but across seasons, across years, below the threshold of any symptom, until the vessel failed.
The chart has no field for this.
The crossing
A wildfire-smoke particle smaller than 2.5 micrometers across is roughly thirty times narrower than a human hair. At that scale, the nose and throat offer nothing. The particle rides inhaled air past the trachea, past the branching bronchi, into the terminal bronchioles, and arrives at the alveolar membrane: two cells thick, stretched across a surface area roughly the size of a tennis court. Oxygen crosses here. Carbon dioxide crosses here. Particles this small cross here too.
The AHA's scientific statement on particulate matter and cardiovascular disease describes three pathways from that crossing. Pulmonary nerve receptors fire, shifting heart rhythm and blood pressure. Oxidative stress in lung tissue triggers inflammatory signals that spill into systemic circulation. And ultrafine particles, or soluble chemical constituents stripped from larger particles, translocate directly across the alveolar membrane into the bloodstream.
Wildfire smoke is not generic particulate matter. An editorial in the European Heart Journal notes that wildfire smoke carries a higher proportion of ultrafine and submicron particles than many other PM2.5 sources, along with more oxidative and proinflammatory constituents: oxygenated polycyclic aromatic hydrocarbons, quinones, aldehydes. These properties facilitate deeper alveolar penetration and more efficient systemic translocation. Smaller particles, more reactive chemistry. They go further in.
What follows is slow, distributed, and produces no sensation a person would recognize as illness. The body's inflammatory machinery shifts. C-reactive protein rises. A meta-analysis of 40 studies found long-term PM2.5 exposure associated with an 18% increase in circulating CRP per 10 µg/m³. Interleukin-6 rises. Fibrinogen, the protein that forms the scaffolding of blood clots, increases in circulation, thickening the blood's readiness to seal a wound that hasn't happened. The endothelial cells lining the arteries, which regulate vascular tone and resist clot formation, begin to dysfunction under chronic inflammatory pressure. Arterial walls stiffen. Existing plaques destabilize. There is no cough that means vascular damage. No ache that means the endothelium is losing its hold.
A person living through this feels nothing specific. Or feels what could be anything: the fatigue attributed to poor sleep, the slight heaviness behind the eyes on a haze morning, the headache that arrives in September and leaves in October and seems, reasonably, to belong to the smoke itself, temporary and self-limiting. The air clears. The headache lifts. Rain comes. The particulate count drops. The body seems to return to what it was.
But the vascular changes do not fully reverse before the next season begins. The arterial wall that stiffened does not entirely soften. The endothelial cells that lost function do not entirely recover. The body carries the residue of one season into the next, and the next season adds its own, and the accumulation is so gradual, so far below the threshold of feeling, that the person living inside this body has no reason to suspect that ordinary is no longer what ordinary was three summers ago.
The measurement
In January 2026, researchers led by Hao et al. published a national cohort study in the European Heart Journal tracking 24.7 million Medicare fee-for-service beneficiaries across the contiguous United States from 2007 to 2018. Over 132.7 million person-years of follow-up, they identified nearly 3 million incident stroke events.
The study's technical achievement was separating wildfire-smoke PM2.5 from all other sources of fine particulate matter. Using NOAA smoke-plume maps, chemical transport modeling, and a machine-learning framework trained on satellite data, meteorology, and ground monitors including low-cost PurpleAir sensors, the researchers estimated daily wildfire-smoke PM2.5 at 1 km² resolution, then averaged exposures over one-, two-, and three-year windows at each participant's residential ZIP code.
The mean three-year wildfire-smoke PM2.5 exposure across the cohort was 0.51 µg/m³. A fraction of the EPA's 24-hour PM2.5 standard of 35 µg/m³. Not a smoke event. The accumulated average of many smoke events and many clear days, compressed into a single number representing years of breathing.
Each 1 µg/m³ increase in three-year average wildfire-smoke PM2.5 was associated with a 1.3% increase in stroke risk. The concentration-response curve was linear with no apparent threshold.
The researchers estimated this translated to roughly 17,200 additional stroke cases per year among older adults in the United States. The per-unit stroke risk from wildfire-smoke PM2.5 was nearly double that of non-smoke PM2.5 (1.3% versus 0.7% per µg/m³), though the authors caution that direct toxicity comparisons are difficult because the two pollutant types differ in distribution patterns and chemical composition. Medicaid-eligible beneficiaries, a proxy for lower socioeconomic status, showed consistently stronger associations.
One study. Observational, not experimental. Exposure modeled at the area level, not measured in individual lungs. Individual stroke risk factors incompletely captured. A separate county-level analysis of Medicare cardiovascular hospitalization rates found minimal chronic effects specifically for stroke, using different methods, outcomes, and exposure products. A 2026 California comparison of wildfire-smoke PM2.5 exposure models cautioned that no gold-standard exposure surface exists and that estimates can differ substantially across datasets.
What is settled, from the broader literature, is the biological plausibility. The pathway from chronic fine-particulate exposure to vascular damage to cerebrovascular events is supported by the AHA's scientific statement, by mechanistic reviews, by the inflammatory-biomarker evidence. Hao et al. did not discover a new mechanism. They measured, at population scale, what the mechanism predicts.
The record
A primary care physician practicing in a smoke-exposed region of the western United States sees cardiovascular patients year-round. Strokes, heart failure, hypertension that resists management. She lives in the same air her patients breathe. During smoke season she checks AirNow on her phone each morning before opening the blinds, runs a HEPA filter in the bedroom, keeps her windows sealed. Her own sinuses tighten in August. Her own throat catches on mornings when the sky turns the wrong color of white, when the light flattens and the mountains disappear and the air has a quality she has learned not to call a taste because it is too faint. More a texture at the back of the palate. Dry, mineral, slightly sweet.
She reads a stroke chart in September. Outside the clinic window, the air quality index has been above 150 for three days. Her own chest feels tight, a shallow constriction she notices when she inhales to speak. The chart says cerebral infarction. The chart says what it is built to say.
Her electronic health record does not ask about wildfire smoke. A representative intake form from a Boise clinic asks about chronic conditions, allergies, medications, tobacco use, alcohol, family history. No field for air-quality exposure. No field for smoke-season duration or years lived in a wildfire-affected region. One form, not a universal standard, but it reflects a structural pattern.
The ATSDR recommends incorporating exposure-history questionnaires into clinical practice and notes that environmental and occupational diseases typically manifest as common medical problems. Hazardous exposures, ATSDR states, rarely enter the clinician's differential diagnosis unless an exposure history is actively pursued.
ICD-10-CM does contain a code. Z77.110: contact with and suspected exposure to air pollution. It exists. It is billable. But it does not modify the stroke diagnosis itself. The stroke remains I63, classified by mechanism: thrombosis, embolism, occlusion. The organ-system code and the exposure code occupy different categories in the record. They can coexist on a chart, but in practice, the exposure code requires a clinician to elicit the history, judge it relevant, and document it separately from the presenting event. For a patient whose stroke arrives years after the smoke seasons that contributed to it, in a clinical encounter focused on acute stabilization and discharge planning, this almost never happens.
The record was designed for a world in which strokes have medical causes and environmental exposures have environmental codes. The two categories occupy separate architectures. They were not built to converge on the same patient at the same moment.
The latency
A three-year exposure window means the smoke seasons that contributed to a stroke in 2020 were the ones in 2017, 2018, 2019. In the Boise corridor, a study of the 2017 smoke season documented 42 days with smoke overhead out of 61 days between August and September. By the time a vessel fails in 2020, that smoke is finished, historical, no longer present in the body as particulate but only as what particulate left behind: arterial stiffness, endothelial dysfunction, inflammatory priming that made a clot more likely on an ordinary Tuesday in clean air.
This temporal gap is not unique to stroke. A 2025 study in JACC found that each 1 µg/m³ increase in two-year average wildfire-smoke PM2.5 was associated with a 1.4% increase in incident heart failure among Medicare beneficiaries, estimating roughly 20,200 additional cases annually. A 2026 JACC study of 65.2 million Medicare beneficiaries found significant associations between three-year wildfire-smoke PM2.5 and cardiovascular hospitalization, including ischemic heart disease and arrhythmias. Each finding arrives in a different journal, under a different organ-system heading, studied by a different research team. The stroke researchers publish in cardiology journals. The kidney researchers publish in nephrology journals. A single environmental exposure producing damage across organ systems, each diagnosis coded and treated in its own clinical compartment.
An earlier piece in this publication, "The Night Is Breaking," described a version of this problem in heat-related mortality:
"The science and the death certificate describe the same event in different languages."
The same mismatch operates here, extended across years rather than hours. The shared cause visible only from the epidemiological altitude where someone thinks to look for it.
What the body carries
The study's linear, no-threshold finding means there is no concentration below which the association disappears. For someone who has lived through five smoke seasons in the Sacramento Valley, or ten in western Montana, or the summer of 2017 in Boise when the sky stayed white for weeks, this is not abstract. Those seasons are not behind you the way weather is behind you. The body metabolized the smoke. The vessels recorded the exposure in a language the body cannot read back as sensation, only as risk, accumulating quietly, shifting the odds by increments too small to feel and too persistent to ignore.
The physician cannot solve this with better charting. Even if she documented every patient's smoke-exposure history, she would be adding a circumstance code to records organized around organ-system diagnoses, in a reimbursement system that pays for treating the stroke, not for tracing its environmental contributors. The infrastructure decisions that shape air quality over years, the wildfire management and filtration standards and land-use choices that determine who breathes what for how long, are cardiovascular decisions. No cardiovascular chart will record them as such.
The smoke clears. The season ends. The body continues its work in the quiet that follows, repairing what it can, carrying what it cannot repair into the next season, the next year, the next encounter with a clinician who will document what the record was built to hold.
The patient in the Sacramento Valley is in rehabilitation now. The left hand closes around a foam ball with effort that used to be automatic. The fingers tighten, release, tighten again. The chart is complete. The chart is accurate.
The stroke does not say smoke.
-
Smoke reversing ozone progress: Deng et al. found in Science that wildfire emissions shifted U.S. ozone trends from declining to rising after 2015, offsetting nearly four years of air-quality mitigation gains.
-
Annual wildfire-smoke mortality burden: Zhang et al. estimated in Science Advances that wildfire-smoke PM2.5 was responsible for roughly 24,000 deaths per year in the contiguous United States from 2006 to 2020, a figure that frames stroke as one outcome within a broader chronic toll.
-
Fires burning through the night: Luo et al. found that total annual potential burning hours across North American wildfires rose about 36% from 1975 to 2024, weakening the nighttime pause that fire crews and communities have historically relied on for recovery.
-
Who controls the indoor air: EPA's 2026 wildfire-smoke factsheets include updated guidance on cleaner-air rooms and indoor filtration, resources that assume household-level control over building sealing and filter quality that renters and residents of older housing often lack.