Main

Inhalation of airborne pollutants such as ozone, nitrogen dioxide, sulfur dioxide, carbon monoxide, and particulate matter of 2.5 and 10 micrometers or less (PM2.5 and PM10) has been connected to a wide range of health problems, including increased morbidity and mortality from respiratory, cardiovascular, and neurologic disease1. PM2.5 is of particular concern because of its ability to penetrate the lungs deeply and enter the bloodstream, and PM2.5 exposure is a known driver of oxidative stress and cytokine-mediated systemic inflammation that can affect multiple levels of biologic function, including DNA integrity, gene expression, and cell viability2,3. Specific to neurovascular disease, short- and long-term exposure to elevated PM2.5 levels have been correlated with ischemic stroke and spontaneous intracerebral hemorrhage risk, likely from inflammatory and reactive oxygen species–mediated endothelial cell dysfunction4. Interestingly, this has not been replicated in aneurysmal subarachnoid hemorrhage (aSAH)5,6, a highly morbid subset of hemorrhagic stroke that occurs as a result of rupture of an intracranial aneurysm, despite the formation, growth, and rupture of these aneurysms being linked to local endothelial cell and vessel wall inflammatory cascades7.

In the Intermountain West region of the United States, and in particular along Utah’s densely populated Wasatch Front, cyclic elevations in air pollution are common in cities bordered by high-elevation mountains as a result of trapped wildfire smoke in the summer (often originating from distant locations) and thermal inversions in the winter (where warm air paradoxically layers at higher elevations and traps air pollutants)8,9. Prior studies have leveraged this distinctive microenvironment to further our understanding of the detrimental health effects of PM2.5 exposure10,11,12. In this brief communication, we examined the effects of PM2.5 levels on aSAH rates in an exposed Intermountain West population and report a potential link between aSAH risk and exposure to cyclic PM2.5 elevations.

We performed a single-center retrospective analysis of adult patients admitted to the University of Utah Hospital with aSAH from the Salt Lake and neighboring valleys over a 5-year period (Fig. 1). We also collected publicly available PM2.5 data from sampling stations closest to each patient’s home address over the entire study period (Fig. 2). Given their potential influence on PM2.5 levels and aneurysm rupture, temperature, season, and barometric pressure data were also collected13. Multivariable Poisson regression models were then used to assess the relationship between PM2.5 levels and daily risk of incident aSAH presentation, adjusting for season and daily mean atmospheric pressure and temperature.

Fig. 1: Patient inclusion flowchart.
figure 1

SAH subarachnoid hemorrhage.

Full size image
Fig. 2: Map of geographic study area.
figure 2

Included patients had aSAH and a recorded residence within predefined zip codes within the Salt Lake and surrounding valleys (approximated by red shading). The average altitude was 4000–5000 feet within the valleys and 8000–12,000 feet in the surrounding mountains. Air pollution and meteorologic data were collected from six sampling stations located throughout the study area (marked with yellow stars), with patients manually assigned to the closest sampling station.

Full size image

Seventy patients with aSAH were included (mean age 58.2 ± 13.8 years, 72% female) (Table 1). There was no association between aSAH risk and day 0, -1, -2, -3, or -7 PM2.5 levels, temperature, or season (all p > 0.5). However, PM2.5 levels in the months preceding aSAH were significantly associated with hemorrhage risk, with a positive association seen 90–180 days before aSAH (incident risk ratio [IRR] 2.03, 95% confidence interval [CI] 1.05–3.93, p = 0.035) (Fig. 3). Over the entire study period, recurring spikes of aSAH were often seen 3–6 months after PM2.5 cyclic elevations (Fig. 4), suggesting a delayed adverse effect of PM2.5 exposure. Accordingly, spikes in aSAH incidence were often concurrent with cyclical nadirs in PM2.5 exposure, with PM2.5 inversely associated with aSAH from day 0 to -90 (IRR 0.38, 95% CI 0.18–0.81, p = 0.012). These patterns were consistent using multiple time windows on sensitivity analyses for robustness but were most clear with 90-day intervals. Independent of PM2.5 levels, day 0 atmospheric pressure was also associated with approximately double the aSAH risk (IRR 1.95–2.10 across PM2.5 time windows, 95% CI 1.57–2.68, p < 0.001).

Fig. 3: Delayed impact of PM2.5 exposure on aSAH risk.
figure 3

Multivariable Poisson regression analysis demonstrating daily incident risk ratios (IRRs) for 90-day intervals before aneurysmal subarachnoid hemorrhage (aSAH). Error bars represent 95% confidence intervals. Statistical significance is denoted as * (p < 0.05). Mean particulate matter 2.5 (PM2.5) levels demonstrate a 3- to 6-month delayed positive association with aSAH incidence (red line). Mean PM2.5 levels within 3 months before aSAH were inversely associated with daily SAH incidence (blue line), possibly because of cyclic PM2.5 levels and the delayed effects of PM2.5 exposure. There was no association at 6–9 months (green line).

Full size image
Fig. 4: Cyclic PM2.5 exposures precede aSAH peaks.
figure 4

Plot of 90-day moving average of daily particulate matter 2.5 (PM2.5) levels (green line) and 90-day moving sum of aneurysmal subarachnoid hemorrhage (aSAH) incidence (blue). Peaks in PM2.5 often appear to precede peaks in aSAH incidence.

Full size image
Table 1 Baseline patient characteristics
Full size table

These data suggest that exposure to elevated PM2.5 levels may influence the risk of intracranial aneurysm rupture, independent of season, temperature, and barometric pressure. This is the first reported significant association between PM2.5 levels and aSAH. Prior investigations, including those by Czernych et al.5 and Kim et al.6, did not identify such a relationship, possibly because their analyses were restricted to PM2.5 levels within the week of rupture. By widening the temporal lens, however, any longer-term impact of PM2.5 levels on aSAH risk can be seen more clearly. Mechanistically, the reason for this delayed effect is unclear, because PM2.5 exposure can immediately affect systemic inflammation14. We nonetheless hypothesize that the longer-term effects of PM2.5 exposure on aSAH risk are likely multifactorial and may include the sensitization of inflammatory cytokine cascades and upregulation of responses to secondary stimuli15 or the accumulation of cellular DNA damage from reactive oxygen species16 combined with impaired DNA repair processes17. These pathways may affect intracranial endothelial cell and vessel wall integrity, potentially promoting existing local inflammatory cascades involved in aneurysm formation, growth, and ultimately rupture7, with the prolonged timeframe analogous to the known carcinogenic potential of PM2.5 exposure18. The sensitivity of intracranial aneurysm rupture risk to seemingly small variables is consistent with the alteration in observed risk seen with use of daily over-the-counter anti-inflammatory medications19, exposure to the presumed anti-inflammatory effects of ultraviolet radiation from sunlight20, and changes in atmospheric pressure as seen herein and in other reports13. Future studies on how PM2.5 exposure affects local inflammatory or cell-level changes are needed to further define this association.

The generalizability of this preliminary study is limited by its single-center, retrospective design, with larger-scale studies needed to confirm the correlative exposure-risk patterns observed, as well as to demonstrate causation. Additionally, although a sampling bias cannot be excluded because we sampled only a portion of aSAH within the specified geographic area, the stability of regional admission patterns over the study timeframe supports this as a representative analysis of regional aSAH incidence. Although patients relocating to the study area or traveling in the months preceding an aSAH could have had different exposure patterns, this is likely an uncommon situation and was not controlled for. The relative effects of cyclical versus non-cyclical PM2.5 exposure, or PM2.5 source, on aSAH risk were also not assessed in this work and remain unclear, as are any potential thresholds for safe versus unsafe PM2.5 exposures in relation to aSAH risk. The effects of other airborne pollutants on aSAH risk were also not examined, with PM2.5 selected as a proxy for overall airborne pollutant levels and hypothesized to most directly affect aneurysm integrity because of the known endothelial cell and systemic effects of such exposure2,3. Additional studies to answer such questions are warranted. Finally, although not the focus of this work, the observed association of atmospheric pressure and aSAH risk has been previously reported13. Aggregate data on the impact of meteorological variables on aneurysm risk are nonetheless mixed and require additional exploration21,22.

Given the high morbidity and mortality of aSAH, we hope these and other data on the deleterious effects of PM2.5 exposure23,24,25 encourage public health initiatives that combat air pollution. At a state level, potential policies include incentives for public transportation use, stricter regulations around daily pollution quotas, creation of statewide pollution mitigation plans for peak days, and closer coordination between state environmental quality agencies and the federal government. At the national level, the federal government could subsidize state-level efforts to curb pollution, set stricter nationwide standards on daily pollution quotas, and increase funding for studies on the environmental impact of air pollution from all sources (human emissions, wildfires, etc.) on disease.

Methods

Patient and environmental data collection

The records of all adult patients (≥18 years old) admitted to the University of Utah Hospital with a diagnosis of aSAH from 1/2018 to 12/2023 were reviewed retrospectively. Included patients had a recorded residence within 36 predesignated zip codes within the Salt Lake, Ogden, Provo, and surrounding valleys. Patients admitted to the University of Utah with aSAH with a home address outside the specified geographic area were excluded (Fig. 1). Mean daily PM2.5 levels (micrograms per cubic meter), temperature, and barometric pressure data were collected from six representative sampling stations located throughout the included geographic area from publicly available datasets (https://air.utah.gov/). Each patient was manually assigned to the sampling stations closest to their home address (Fig. 2).

Data analysis

Each day at each collection site was considered a distinct datapoint, yielding 12,948 total datapoints. Exposure data were standardized using annual standardization, in which each daily reading was expressed relative to the mean and standard deviation measured from its respective year and location26. Means and standard deviations were calculated for each year at each sampling station. Daily values were then transformed by subtracting the mean and then dividing by the standard deviation. Standardized data for each station cluster were treated as separate data points. Univariable statistical tests were conducted using nonstandardized exposure data.

Each independent variable followed a non-normal distribution, so nonparametric statistical tests (i.e., Wilcoxon rank-sum tests for continuous variables) were used to assess for univariable analysis of differences between groups. Data were analyzed as time series. Multivariable Poisson regression models were used to assess the relationship between PM2.5 and daily risk of incident aSAH presentation, controlling for daily mean atmospheric pressure and temperature13. Timeframes analyzed included days 0, -1, -2, -3, and -7, as well as 30- and 90-day rolling intervals preceding aSAH up to 270 days. All Poisson models used standardized exposure data and were clustered by season of the year (Fall: September-November, Winter: December-February, Spring: March-May, Summer: June-August) to adjust for the possible relationship between season and aSAH27. Significance was defined as α = 0.05. Statistical analyses were performed using Stata MP version 18.0 (StataCorp LLC, College Station, Texas, USA).