At the EPA Ozone Limit, an RCT Found Airway Harm

The U.S. ozone standard is meant to protect public health with an adequate margin of safety. Yet a controlled human trial at that exact limit found measurable effects in healthy young adults who exercised during exposure. The result does not show that one afternoon outdoors causes clinical lung disease. It does show that compliance with the standard does not make ozone biologically inert.

In the 2026 Pennington et al. trial, 0.07 ppm ozone for 6.6 hours produced a net decline in forced expiratory volume in one second (FEV1) of about 2.07 percentage points compared with filtered air. The study also found more airway neutrophils 16–18 hours later in a small sputum subsample. For runners, cyclists, outdoor workers, and other people who spend long afternoons exerting themselves outside, that is a more useful definition of an ozone-advisory day than a reassuring word such as "compliant."

The experiment tested the regulatory boundary

Pennington and colleagues conducted a double-blind, randomized crossover trial in 38 healthy adults aged 19 to 34. Each participant completed both conditions: 6.6 hours in 0.07 ppm ozone and 6.6 hours in filtered clean air, with intermittent moderate exercise. Because participants served as their own controls, the comparison avoids many differences that complicate ordinary observational studies, such as baseline fitness or lung size.

The concentration matters. Ozone advisory days are triggered at an 8-hour average of at least 0.070 ppm, the same concentration used in the experiment. Intermittent exercise increased ventilation, so participants inhaled more ozone than they would while sitting still. That setup resembles a long warm-weather ride, an extended training session, or much of an outdoor workday more closely than a brief walk to the mailbox.

After ozone exposure, mean FEV1 changed by −1.24% ± 0.92%, compared with +0.83% ± 0.50% after clean-air exposure. The resulting net difference of roughly −2.07 percentage points was statistically significant (p=0.017). Forced vital capacity (FVC) moved in the same direction but did not reach statistical significance (p=0.148). The trial therefore supports an acute FEV1 effect; it does not establish a broader restrictive pattern of impaired spirometry.

Symptoms also rose late in the exposure period, at 5.6 and 6.6 hours. That timing fits a practical concern: a person may feel fine at the start of a session and accumulate dose as heat, sunlight, and exertion continue.

The delayed inflammatory result deserves attention—and caution

In 14 participants who provided induced sputum, neutrophils made up 33.4% ± 6.9% of cells 16–18 hours after ozone exposure, versus 18.6% ± 5.6% after clean air (p=0.013). That is an absolute increase of 14.8 percentage points, or roughly an 80% relative increase. Neutrophils are immune cells, so this endpoint indicates airway inflammation after exposure; it is not a diagnosis and does not by itself predict a person's long-term health outcome.

The delay also matters. Immediate post-exposure spirometry cannot capture every downstream response if inflammatory changes emerge the next day. But the sputum result comes from only 14 people, and the sizable standard errors show substantial variation between participants. It is statistically significant, but its precise average size remains uncertain.

Statistical significance is not clinical impairment

The average FEV1 change was small. It falls below the individual-level threshold commonly used in respiratory practice for clinically important change—typically at least 10–12% predicted or 200 mL. The trial did not show that a single 6.6-hour exposure would leave a healthy participant with a diagnosable lung disorder.

That distinction should temper the headline, not erase the finding. Clinical thresholds help physicians judge an individual patient. Air-quality standards must also consider whether repeated small effects across many people, including those more vulnerable than the trial participants, add up to a meaningful public-health burden. The trial answers the narrower causal question: under its conditions, ozone at 0.07 ppm altered lung function and an inflammatory biomarker in healthy young adults.

Population studies add context, not threshold proof

A 2022 ELAPSE analysis by Brunekreef et al. pooled European prospective cohorts and linked long-term ambient ozone exposure, across broader real-world concentration ranges, with respiratory morbidity and mortality. Those associations are consistent with the controlled trial's direction of effect. They cannot identify whether an 8-hour concentration of exactly 0.07 ppm caused the observed outcomes, and observational associations do not establish causation as cleanly as a randomized exposure comparison.

Likewise, Macintyre et al.'s UK modeling study projected that ozone would continue to contribute to respiratory hospital admissions under current emissions policies. It describes a population burden across projected concentrations; it does not measure the effect of the U.S. standard's precise boundary.

Evidence in vulnerable groups points in the same general biological direction but cannot be pasted onto this trial. A 2025 multi-city case-crossover study by Wang et al. associated short-term ambient ozone exposure with pediatric pneumonia admissions in China. The study involved children, a hospital outcome, and concentrations above the U.S. standard. It supports concern about ozone's respiratory effects in a vulnerable population, not a claim that 0.07 ppm causes pediatric pneumonia.

Why heavily polluted-city data cannot settle this question

A 2022 ecological analysis from Ahvaz, Iran examined health endpoints alongside ozone, particulate matter, nitrogen oxides, carbon monoxide, and sulfur dioxide in a heavily polluted industrial setting. It neither overturns nor confirms the controlled trial. Ozone concentrations and co-pollutant exposures were far above conditions relevant to a U.S. compliance day, and an area-level ecological design cannot separate ozone's individual contribution. The study illustrates a recurring problem in air-pollution epidemiology: outside an exposure chamber, pollutants travel in packs.

What the trial cannot tell us

The participants were healthy adults ages 19 to 34. The findings cannot be directly extrapolated to children, older adults, people with asthma, or people with cardiopulmonary disease. Those groups are central to public-health protection, but this experiment did not test them.

It also tested one acute exposure. The researchers did not report 24-hour or 72-hour recovery, repeated-exposure effects, seasonal accumulation, or a dose-response curve below and above 0.07 ppm. Without controlled studies across nearby concentrations, the available evidence cannot identify a threshold below which these effects disappear. The available abstract also does not disclose a funding source, so readers cannot assess possible sponsor influence from that publication alone.

A sensible training response

For a healthy person, this evidence supports risk management rather than panic. On warm, sunny days with an ozone advisory, use local real-time air-quality data before scheduling a long outdoor workout. Ozone commonly accumulates as sunlight and heat drive its formation, making early morning a practical lower-exposure option than afternoon.

• Move a prolonged session earlier when local ozone is lower.
• Shorten or reduce the intensity of an outdoor workout on an advisory day.
• Train indoors when a long, hard session cannot be shifted.
• Treat air quality as a training condition alongside heat and humidity.

The strongest conclusion is narrow but consequential: in a randomized, blinded crossover experiment, healthy young adults had a small statistically significant FEV1 decrement and delayed airway neutrophilia after exercising for 6.6 hours at the current 0.07 ppm ozone standard. "Within the limit" describes a regulation. It does not describe a no-effect dose.