Amazon smoke and forest loss now carry a death count in Brazil's own records

Four studies published between January and June 2026 have moved the health case against Amazon burning from the general to the quantified. The most concrete of them estimates that fine particulate pollution in Brazil's seven Amazonian state capitals was responsible for roughly 5,472 circulatory deaths and 2,621 respiratory deaths over six years [s1]. The most novel argues that deforestation itself — not only the smoke it produces — tracks with heart attack mortality across the region [s4].

The attributable-death estimate

The Toxicology study, published 9 January, quantified long-term PM2.5 exposure across Belém, Boa Vista, Macapá, Manaus, Palmas, Porto Velho and Rio Branco between 2018 and 2023, using CAMS reanalysis data for pollution and Brazilian Health Data Platform records for mortality, with health impacts estimated through the WHO's AirQ+ log-linear model [s1].

Annual mean PM2.5 ranged from 10.47 µg/m³ in Palmas in 2020 to 49.69 µg/m³ in Porto Velho in 2021 [s1]. Every city exceeded the WHO annual guideline of 5 µg/m³ in every year examined, and most capitals recorded over 100 days per year above the 15 µg/m³ daily guideline [s1]. Correlations between wildfire counts and PM2.5 were statistically significant in Porto Velho (r = 0.24) and Rio Branco (r = 0.21) [s1].

The attributable burden over 2018–2023 came to approximately 5,472 deaths from circulatory diseases (95% CI: 4,595–6,711) and 2,621 from respiratory diseases (95% CI: 864–4,278) [s1]. Broken out by condition: 627 COPD deaths (95% CI: 462–781), 427 lung cancer deaths (95% CI: 314–534), and 304 ischaemic heart disease deaths (95% CI: 150–467) [s1]. The heaviest burdens fell on Belém, Manaus and Porto Velho [s1].

Note the width of those intervals, particularly for respiratory mortality, where the lower bound of 864 and the upper bound of 4,278 differ by a factor of five. These are model-derived attributable fractions, not counted deaths, and they inherit the uncertainty of both the exposure reanalysis and the concentration–response function.

Fire counts and chronic respiratory burden

A BMC Public Health ecological study published 21 February approached the same question from the fire side, correlating annual wildfire focus counts from Brazil's National Institute for Space Research (INPE, Queimadas Program) against Global Burden of Disease estimates for chronic respiratory disease deaths and disability-adjusted life years across the states of the Legal Amazon from 2011 to 2021 [s3].

The result was heterogeneous rather than uniform. Pará carried the largest absolute burden of deaths and DALYs, followed by Maranhão and Mato Grosso, states which also showed high wildfire activity [s3]. Most states showed increasing temporal trends in deaths and DALYs, though Maranhão showed declining deaths and Rondônia declines in both outcomes [s3]. A pronounced peak in both wildfire activity and respiratory disease burden appeared in 2020 [s3].

Statistically significant positive correlations between fire foci and both deaths and DALYs turned up in four states — Acre, Amazonas, Mato Grosso and Rondônia — and not in the others [s3]. The authors are explicit that the findings rest on aggregated annual data, with only 11 annual observations per state, and that causal inference and individual-level risk estimation are not possible from this design [s3].

The deforestation–heart attack link

The newest of the four, published 30 June in IJERPH, asks a different question: whether deforestation is associated with cardiovascular death independently of any single fire season [s4]. It examined acute myocardial infarction mortality across health regions of the Brazilian Amazon from 2000 to 2023, fitting generalised additive models with a negative binomial distribution to test for nonlinear associations, plus Bayesian spatiotemporal models incorporating structured spatial effects and a smoothed temporal trend [s4].

The analysis identified a significant nonlinear association between deforestation and heart attack mortality, with risk intensifying progressively in areas subject to greater environmental degradation [s4]. Lagged models at one, two and three years showed the association persisting over time, which the authors read as suggesting cumulative effects of environmental exposure [s4]. Spatially, areas of elevated relative risk expanded over the study period, particularly within the region known as the Arc of Deforestation [s4].

The mechanism proposed is indirect: deforestation alters regional microclimatic dynamics, intensifies wildfires, and increases population exposure to cardiovascular risk factors [s4]. But this is an ecological study using data aggregated by health region and year, and the authors frame their conclusion carefully — deforestation may act as an important socioenvironmental determinant of cardiovascular health in the Brazilian Amazon [s4]. It cannot distinguish smoke exposure from the other things that change when forest is cleared, including who moves in, what work they do, and what health services exist where they live.

Who is most exposed

A PRISMA-based review in Critical Reviews in Toxicology, published 22 January, screened 302 articles on wildfire smoke exposure in the Amazon and included 30 [s2]. Two of its findings frame the rest of the evidence. The groups most susceptible to smoke effects were children up to five years old and adults aged 60 and above, along with individuals with pre-existing comorbidities [s2]. And all of the studies it found were conducted in the Brazilian Amazon [s2] — despite the review having searched for evidence across the wider Pan-Amazon, which extends well beyond Brazil's borders.

The review also notes that only one of the 30 included articles conducted a biomonitoring study in the region; the rest worked with hospital admission data [s2]. That is the methodological ceiling on this entire literature: almost everything known about Amazon smoke and human health is inferred from administrative records rather than measured in the people breathing it.

What this does and does not add up to

Taken together the four papers converge on a direction — more burning and more clearing, more cardiorespiratory death — while differing sharply in how firmly each can support it. None is a cohort study. Three of the four are ecological in design, which means every one of them is vulnerable to the possibility that fire-heavy regions differ from fire-light regions in ways that have nothing to do with smoke.

What is not in dispute is the exposure itself: seven Amazonian capitals exceeding the WHO annual PM2.5 guideline every year from 2018 to 2023, several of them for more than a hundred days a year above the daily guideline [s1].

What to watch

Whether biomonitoring studies appear — the review found essentially one [s2] — and whether the health literature ever extends beyond Brazil to the Peruvian, Colombian, Bolivian and Ecuadorian Amazon, where the same fires burn and no comparable evidence exists.

Sources

• Impact of PM2.5 on cardiorespiratory mortality: A study in the capitals of Brazilian amazon rainforest — Toxicology, 9 January 2026
• Wildfires in the Pan-Amazon: what is known about health impacts on affected populations? — Critical Reviews in Toxicology, 22 January 2026
• Linking fire incidence to chronic respiratory disease outcomes in Brazil's Legal Amazon: an ecological study — BMC Public Health, 21 February 2026
• Spatiotemporal Associations Between Deforestation and Acute Myocardial Infarction Mortality in the Brazilian Amazon — International Journal of Environmental Research and Public Health, 30 June 2026