Dirty Air Drives Rio’s Respiratory Hospital Surges, Decade-Long Study Finds

A sweeping ten-year analysis of Rio de Janeiro has delivered one of the clearest pictures yet of how air pollution and weather combine to send residents of a tropical megacity to the hospital. Researchers examined daily respiratory admissions in the Brazilian city from 2015 to 2024, alongside measurements of fine and coarse particulate matter, temperature, humidity, and rainfall. Their conclusion is stark: particulate matter, even at concentrations well below Brazil’s national limits, acts as an acute trigger for respiratory hospitalizations, while the effects of heat unfold on a surprising delay of up to two weeks.

The study, published in the journal Air Quality, Atmosphere & Health, applied a statistical framework known as a distributed lag nonlinear model, or DLNM. Unlike conventional linear regression, this approach captures two dimensions at once: how the health risk changes across the range of an exposure, and how that risk is distributed over the days and weeks that follow. Each environmental variable was represented by a cross-basis function, and the models were fitted with a negative binomial distribution to handle the overdispersed count data. Long-term trends and seasonality were controlled with natural cubic splines of time, and day of the week was included to account for systematic weekly rhythms in hospital admissions.

The scale of the dataset is formidable. Over the decade, the city recorded an average of 284 respiratory hospitalizations per day, with daily counts ranging from zero to 1,417. Women accounted for nearly 58 percent of admissions, and the burden fell disproportionately on the youngest and oldest residents: children aged 14 and under and adults aged 65 and over together made up more than 70 percent of the daily average. The researchers caution, however, that these descriptive patterns do not demonstrate greater susceptibility to pollution by age, because the modeling was performed on total daily counts rather than age-stratified analyses.

Particulate matter emerged as the most consistent and robust risk factor. When concentrations of PM2.5, particles smaller than 2.5 micrometers, rose to their 90th percentile of about 20 micrograms per cubic meter compared with the median, the cumulative risk of respiratory hospitalization over the following 14 days increased by roughly 26 percent. Coarse PM10 particles, at a 90th percentile of 28 micrograms per cubic meter, carried an even stronger cumulative effect of about 30 percent. Crucially, these effects were acute: the risk spikes appeared primarily within the first three to seven days after exposure and then attenuated, a temporal signature consistent with particulate matter acting as an immediate irritant and inflammatory trigger rather than a slow-burning hazard.

Perhaps the most provocative finding is that PM10 outperformed PM2.5 in the strength of its association, inverting the usual assumption that fine particles are the dominant threat. The authors point to Rio’s local emission profile. Coarse particles are typically generated by resuspended road dust, construction activity, and traffic, all pervasive in a densely built coastal city. The measured PM2.5 to PM10 ratio of 0.69 confirms that coarse particles make up a substantial share of the suspended load. Coarse particles can also carry a heterogeneous mixture of biological material, crustal elements, and other irritants capable of provoking airway inflammation, and their predominance in short-lag effects suggests they act as fast-acting triggers of acute respiratory events.

The regulatory implications are hard to ignore. Average concentrations during the study period were 13 micrograms per cubic meter for PM2.5 and 19 for PM10, both below Brazil’s current national standards of 20 and 40 micrograms respectively under CONAMA Resolution 506/2024. Yet both exceed the World Health Organization’s updated guidelines, which recommend annual means of just 5 micrograms for PM2.5 and 15 for PM10. Maximum daily values frequently breached the WHO 24-hour limits outright. Detecting significant hospitalization effects at concentrations deemed acceptable by national law, the authors argue, suggests that current Brazilian standards may not adequately protect public health, and they call for closer alignment with international guidelines.

Temperature told a more intricate story. The exposure-response curve was U-shaped, with elevated risk at both cold and hot extremes. Low temperatures at the 10th percentile, around 20.8 degrees Celsius, raised cumulative risk by about 24 percent. High temperatures near 27 degrees produced a cumulative increase of roughly 20 percent, but the striking feature was timing: heat effects were delayed, peaking at lags of 8 to 14 days rather than appearing immediately. At the 99th percentile of temperature, 29 degrees, the cumulative risk more than doubled. The authors speculate that this lag reflects indirect pathways, including heat-driven photochemical formation of tropospheric ozone, prolonged exposure to poorly ventilated air-conditioned indoor environments, and the exacerbation of underlying chronic conditions over consecutive warm days.

Humidity and rainfall, by contrast, appeared protective. Higher relative humidity was associated with a 16 percent reduction in cumulative hospitalization risk at the 90th percentile, and heavy precipitation cut risk by 28 percent, with the strongest effects in the first week. Extreme rainfall at the 99th percentile was associated with a dramatic 83 percent reduction. The pattern supports a well-known atmospheric cleansing mechanism: rain scavenges particles from the air through wet deposition, temporarily lowering exposure. Low humidity, meanwhile, showed a borderline increase in risk, hinting that dry conditions may favor both pollutant persistence and pathogen survival.

The team subjected the results to extensive robustness checks. Varying the maximum lag window from 7 to 21 days left the particulate matter estimates essentially unchanged, and altering the flexibility of the lag-response function had little effect. A sensitivity analysis excluding the COVID-19 pandemic years of 2020 through 2022, when hospitalization patterns and pollution levels were disrupted, attenuated the estimates modestly but preserved both their direction and statistical significance. The one specification that mattered was the flexibility of the exposure-response curve itself: with four degrees of freedom the particulate associations lost significance, a reminder that model choices in nonlinear time-series work can shape conclusions.

The authors acknowledge the limits inherent in an ecological design. Exposure was measured as a citywide average from ten fixed monitoring stations, which cannot capture the sharp spatial gradients created by Rio’s massifs, coastal boundaries, and microclimates, nor the personal exposure differences shaped by commuting and daily activity. Gaseous pollutants such as nitrogen dioxide and ozone were not included in the final models, and interactions between pollution and weather were not directly tested. Even so, the study stands as the first of its kind in the region with such an extended time series, and its message for policymakers is concrete: short-term tools such as air quality alerts and emission controls could measurably reduce hospital burden, while the delayed heat effects argue for heat-health warning systems as climate change intensifies extreme weather in tropical cities.

Subject of Research: Nonlinear and delayed effects of particulate matter and meteorological conditions on respiratory hospitalizations in Rio de Janeiro

Article Title: Associations of particulate matter and meteorological factors with respiratory hospitalizations in Rio de Janeiro: A distributed lag nonlinear model study

Article References: de Souza Pedreira, M. F., Lorenz, C., Ynoue, R. Y., Nogueira, T., & Gioda, A. (2026). Associations of particulate matter and meteorological factors with respiratory hospitalizations in Rio de Janeiro: A distributed lag nonlinear model study. Air Quality, Atmosphere & Health, 19(10), Article 232. https://doi.org/10.1007/s11869-026-02124-z

Image Credits: AI Generated

DOI: 10.1007/s11869-026-02124-z

Keywords: air pollution, particulate matter, PM2.5, PM10, respiratory hospitalizations, distributed lag nonlinear model, Rio de Janeiro, temperature, humidity, precipitation, tropical megacity, WHO air quality guidelines