The control arm in this trial was the same air purifier with the filter pulled out. Same box, same hum (58 decibels instead of 60), same breeze at the vent. Participants lived with it for a month and could not tell. The blinding index says the disguise held.

So for one month, about a hundred households near a Massachusetts interstate ran the most expensive fans in Somerville in their bedrooms and living rooms. Then they got the real thing. The question was whether their blood pressure noticed.

HAFTRAP (the HEPA Filtration for Traffic Related Air Pollution trial) enrolled adults 30 and older living within 200 meters of I-93 or Route 38. Each household got one month of true HEPA filtration and one month of sham, in random order, with a one month washout between. Blood pressure was measured at home on weekday mornings at the start and end of each period.

The exclusion list is long and sensible for a mechanism study. No smokers or vapers, no prior heart attack or stroke, nobody on blood pressure or anti-inflammatory medication. The group averaged 41 years old with a baseline pressure of 118.8/76.5, and most were college educated with household incomes above $85,000. These were healthy people who happen to live next to a highway. That matters for everything that follows. The people most likely to benefit, those already on blood pressure pills, were not in the trial.

Every positive result here is a subgroup. The overall blood pressure effect is zero. The overall cognition effect is zero. Slice by starting pressure and by age and a signal appears, each one just under the p = 0.05 line. Read strictly, this trial is negative.

I keep reading it anyway, for three reasons. The subgroups make biological sense (people with some pressure to spare had room to fall). The mechanism is old and well described. And the intervention is a box on the floor with no side effects beyond noise and an electric bill. When a finding is plausible, cheap, and harmless, the bar for acting on it is lower than the bar for publishing it.

That is different from calling it proven. It is a hypothesis with a good pedigree and a quiet fan.

The American Heart Association laid out the pathway in a 2010 scientific statement. Inhaled particles irritate the airway and alveoli, which triggers a reflex toward sympathetic tone. Oxidative stress and inflammatory mediators spill into the circulation, nitric oxide becomes less available, and arterioles constrict. Blood pressure rises within hours of exposure.

The brain story is less settled. Ultrafine particles can travel along the olfactory nerve or cross a leaky blood brain barrier, and circulating cytokines can prime microglia. Plausible, and nobody has shown the dose that matters. HAFTRAP is best read as a test of turning the dial down at home. It turned a little.

The senior author is Doug Brugge (UConn Health, formerly Tufts), who has studied near highway pollution in Somerville for more than fifteen years and wrote a book subtitled The Deadliest Pollutant Is One You Breathe Every Day. Two coauthors come from Somerville community groups, and the paper says the trial answered neighborhood requests. NIEHS funded it. Austin Air built the units and sold them at a discount, with no role in the analysis. I see no financial conflict. The lean is intellectual and civic, and it shows when the authors recommend purifiers for vulnerable groups on the strength of a subgroup.

On blood pressure, HAFTRAP lands where the field already was. A 2023 meta-analysis of 17 true versus sham trials (about 880 people) found a pooled systolic reduction of 2.35 mmHg. The trials were small, and the biggest effects came from Chinese cities with far dirtier air than Somerville.

On infection the data are softer. A 2023 systematic review of 32 studies found no convincing evidence yet that air cleaning reduces respiratory infection in real world settings. A HEPA trial in Australian aged care rooms missed significance in its main analysis (odds ratio 0.57, confidence interval 0.32 to 1.04).

And yet, during COVID, the observational signal and the physics pointed the same way. Georgia elementary schools that improved ventilation had 39% lower incidence, and those adding HEPA filtration to dilution had 48% lower. Observational and confounded, and consistent with everything we know about aerosols. More air exchanges were better than fewer. The purifiers in my house never went back in the closet.