Long-term wildfire smoke was linked to slightly more dementia in 40 million older Americans

Wildfires in the United States have grown larger, more intense and longer-lasting in recent decades, and their smoke now drifts over much of the country. Most of the health concern centres on PM2.5, tiny particles less than 2.5 thousandths of a millimetre across that can travel deep into the lungs. Long-term exposure to PM2.5 in general has already been linked to dementia. Whether smoke from wildfires carries a risk of its own, and which part of the smoke might be responsible, has been much less clear.

Wildfire smoke is rich in carbon: organic carbon from partly burned plants and elemental carbon, the soot made when burning is incomplete. A new preprint, not yet peer-reviewed, asks whether long-term exposure to smoke particles, and to their carbon in particular, goes together with new diagnoses of Alzheimer's disease and related dementias in older Americans.

What they did

The team used records from Medicare, the US public health insurance for people aged 65 and over. They included 39,678,630 people in the traditional fee-for-service plan living in the contiguous US (all states but Alaska and Hawaii), and followed them from 2005 to 2019. To count only new cases, everyone had to go three years without a dementia diagnosis before follow-up began.

• Outcomes: two overlapping ones, taken from insurance claims: Alzheimer's disease alone, and the broader group of Alzheimer's disease and related dementias.
• Exposure: a computer model combining satellite data, air chemistry simulations and thousands of ground monitors estimated daily pollution on a 1-km grid, split into smoke and non-smoke particles and their carbon. Each person got the average for their ZIP code over the three previous years.
• Fair comparison: the analysis accounted for age, sex, race, low income (shown by also qualifying for Medicaid), neighbourhood characteristics such as poverty, weather, distance to a hospital, county-level smoking and body weight, region and calendar year. Smoke and non-smoke pollution were included together.

What they found

Over the follow-up there were 3,501,414 new Alzheimer's diagnoses and 8,377,416 new diagnoses in the broader dementia group. The average three-year level of smoke particles was low, about 0.44 micrograms per cubic metre of air, compared with 9.40 for particles from other sources.

• Smoke particles: each extra microgram per cubic metre was linked to a hazard ratio (the relative rate of new diagnoses) of 1.015 for Alzheimer's and 1.029 for the broader dementia group, meaning rates a few percent higher.
• Smoke carbon: the links were stronger per microgram, with hazard ratios of 1.063 and 1.090.
• Per typical difference: compared across the spread of exposures people actually experienced, the smoke links were small (hazard ratios of 1.005 and 1.009), smaller than those for ordinary non-smoke pollution (1.086 and 1.048), because smoke levels varied much less.
• Carbon seemed key. When the team split the smoke up, both kinds of carbon were linked to more dementia, while the rest of the smoke particles showed only weak links.
• Uneven map. The links were strongest in the Northeast and generally absent or reversed in the Midwest and West.
• Estimated toll. The researchers estimate that smoke particles accounted for about 2,284 new Alzheimer's cases and 11,142 new cases in the broader group each year, less than 2% of all new cases. As other pollution fell over the years, smoke's share of the pollution-related burden grew, from 10.6% to 42.1% for the broader dementia group.

The results held up in most extra checks, such as using longer or shorter exposure windows, looking only at people who never moved, or also accounting for recorded tobacco, alcohol and drug use disorders.

What to keep in mind

• The authors' own limits. Diagnoses came from insurance claims, not from doctors' assessments or brain examinations, and Alzheimer's-specific codes are known to miss cases. There was no individual information on smoking history, exercise or education, so other differences between people could explain part of the link. Pollution levels were modelled, not measured where people breathed, and say nothing about time indoors or air filters. Gases in wildfire smoke, such as ozone, were not accounted for. And only people in traditional Medicare were included.
• Big unit, small reality. The authors warn that one extra microgram of smoke per cubic metre is a very large and unusual step, more than twice the average level. The per-microgram numbers are a standard way to compare, not what most people experienced.
• Links, not causes. This is an observational study. It shows that smoke and dementia diagnoses went together, not that the smoke brought on dementia.
• Puzzling patterns. Why the links were strongest in the Northeast and absent in the West, where most smoke is, isn't known; the authors offer possible reasons but call the regional findings descriptive.
• Not yet peer-reviewed. As a preprint, the work has not yet been checked by independent experts.

Why it matters

Air pollution from cars and power plants has been falling in the US, while wildfire smoke has been growing. This is, according to the authors, the first national, long-term study to look at wildfire smoke and its carbon in relation to new dementia diagnoses. Each person's added risk looks small, but because so many people breathe the smoke, the estimated number of cases is in the thousands each year. If the link is confirmed, the authors argue, it would help to track wildfire smoke and its make-up separately in air-quality and health planning.