Women Who Cleaned Lost Lung Function at a Pace Compared to Smoking 10 to 20 Cigarettes a Day

I’ve always thought of the danger under the kitchen sink as a swallowing problem. A kid gets the cap off, somebody calls poison control, and the fix is a plastic latch on the cabinet door. So I wasn’t ready for a 20-year study suggesting the bottle does slower work on the adult using it exactly as directed: spraying, wiping, and breathing whatever drifts up off the counter a few times a week.

That study came out of the European Community Respiratory Health Survey, which followed 6,235 adults from 22 study centers for about two decades and tested their breathing at three points along the way. In 2018, Øistein Svanes and colleagues reported in the American Journal of Respiratory and Critical Care Medicine that women who cleaned at home, or cleaned for a living, lost lung function faster than women who did neither.

Everyone’s lungs decline with age, so this is a question of pace. FEV1 is the air you can force out in the first second of a hard breath, and women who didn’t clean lost 18.5 milliliters of it a year, while women who cleaned at home lost 22.1 and occupational cleaners lost 22.4. FVC, the total volume you can blow out, slid 8.8 milliliters a year in the women who didn’t clean, 13.1 in the home cleaners and 15.9 in the professionals. Cleaning sprays and other cleaning agents were each tied to faster FEV1 decline.

My first reaction was a shrug, because a few milliliters a year sounds like rounding error. Then I multiplied the home cleaners’ extra 3.6 milliliters by 20 years and got roughly 72 milliliters of breath gone, on average, across thousands of women. Kristin Cummings and M. Abbas Virji, scientists at the National Institute for Occupational Safety and Health, sized it more bluntly in an accompanying editorial: the effect “was comparable to smoking 10 to 20 cigarettes daily during the study period.” Federal scientists don’t usually write slogans, and they ended with one anyway: “The time to act is now.”

They also laid out the weak spots, and they’re worth knowing. Cleaning habits were self-reported. Only 57 men reported occupational cleaning, too few to say much about men, and in the men studied cleaning wasn’t significantly linked to decline. The odds of chronic airway obstruction ran higher in the women who cleaned, but on few cases and without statistical significance, and because spirometry changes are nonspecific, the study can’t name the disease underneath. It’s observational, so something else cleaners share could be feeding the signal. What it does have is two decades of repeated lung measurements, and no ethics board will ever randomize women to 20 years of spray bottles. Waiting for that trial is a choice, and the NIOSH authors argued against making it.

Wait, though. How does a mist that’s gone in thirty seconds show up on a lung test years later? The editorial names what’s in it: “chemicals that are known sensitizers and irritants,” others with poorly characterized respiratory effects, and mixtures of both. A spray takes a liquid you’d otherwise only touch and floats it onto the wet lining of your airways. The working idea, and it’s an idea rather than something this study tested, is that each session is a small irritation the airway mostly repairs, and that “mostly” piles up over a thousand ordinary Tuesdays. It’s the kind of question I’d have wanted answered before the trigger nozzle ever went on the shelf.

A bouquet nobody had to list

The editorial only gestures at what’s in the mist. Anne Steinemann’s team measured it. Testing emissions from 25 scented products, including air fresheners, laundry products, cleaners and personal care items, they found 133 different volatile organic compounds, an average of 17 per product, as Environmental Health Perspectives reported in 2011. Every product gave off between 1 and 8 chemicals classed as toxic or hazardous. About 44% released at least one of 24 carcinogenic hazardous air pollutants, a list that includes acetaldehyde, 1,4-dioxane, formaldehyde and methylene chloride, pollutants that, the journal noted, the EPA says “have no safe exposure level.”

Of those 133 compounds, ethanol was the only one named on any label, and it showed up on just 2 products. The safety data sheets did barely better, listing ethanol and 2-butoxyethanol and nothing else. Products sold as “green,” “natural” or “organic” released as many hazardous chemicals as the standard ones. And limonene, the citrus note I’d always read as the smell of a clean kitchen, reacts with ozone in indoor air to form secondary pollutants, formaldehyde among them. I think about that now every time a counter smells like lemons.

These were emissions from the products, not doses measured in anyone’s lungs, and the team didn’t track health outcomes. But you can’t weigh a risk nobody tells you about, and in 2011 nobody had to. The Consumer Product Safety Commission did not require manufacturers to disclose any ingredients on cleaning supplies, air fresheners or laundry products, fragrances included, and the FDA required no fragrance disclosure on personal care products either. A fix, the Household Product Labeling Act, was under review in the Senate. Congress never passed it. When disclosure finally came, it came from a state: California’s Cleaning Product Right to Know Act required intentionally added and fragrance ingredients to be posted online starting in 2020, according to the Retail Industry Leaders Association. Everywhere Washington had the say, shoppers were left to read the word “fragrance” and guess.

Safe alone, tested alone

Ronald Kostoff, a researcher at Georgia Tech’s School of Public Policy, returned to this subject this week in TrialSite News, and he’s been making the core argument for years. In a 2018 Toxicology Reports editorial with Marina Goumenou and Aristidis Tsatsakis of the University of Crete, he went through thousands of Medline abstracts on combined exposures and concluded that when agents act together, “less of each component stimulus is required to cause damage.” In rats, a mixture of 13 chemicals, each dosed below its own no-observed-adverse-effect level, caused biochemical and blood changes. In mice, bisphenol A plus the plasticizer DEHP significantly reduced uterine epithelial cadherin, a protein that holds the cells of the uterine lining together, when the same doses given separately did not. Zinc oxide nanoparticles, low in toxicity on their own, “significantly enhanced” the insecticide dimethoate’s oxidative damage to mouse livers.

I had always assumed “no adverse effect” meant no adverse effect. It means none when that chemical is the only one in the body. Here’s how I picture the gap, and it’s my illustration, not their finding: if two substances lean on the same defense, say the liver’s antioxidant reserve, each one spends part of the margin, and a threshold measured for one assumed the other wasn’t there spending its share. Kostoff’s group wrote that there was “no reason to believe today” that single-stressor limits “are fully protective against serious adverse health effects in all real life exposure scenarios.”

Most of that evidence is animal and cell work, and nobody has measured the full mix under an American sink, in people, over decades. That missing study should count against the companies that put the mixture on the market, not against the people asking what’s in it.

Phthalates are where this runs straight into my own beat. In a cross-sectional study of U.S. teenagers published in Pediatrics in 2013, Leonardo Trasande’s group linked urinary metabolites of the phthalates commonly found in food to insulin resistance, an association that held after controlling for bisphenol A, while the lower-molecular-weight phthalates typical of cosmetics weren’t significantly associated. In 2021, Trasande, Buyun Liu and Wei Bao used urine samples from more than 5,000 adults aged 55 to 64 to estimate that phthalate exposure was associated with roughly 100,000 premature deaths a year among older Americans and $40 billion to $47 billion in lost productivity. The most exposed adults had higher heart disease mortality, and cancer deaths didn’t rise. It’s an extrapolated, NIH-funded estimate, and Trasande said plainly that it doesn’t establish cause and effect. Cross-sectional, modeled, associational: these are exactly the signals a system that demands proof one chemical at a time can always find a reason to wait on.

The latch was never wrong. Poisonings are sudden, and they put children in emergency rooms. But that’s the part of the hazard packaging can fix, which may be why it’s the part that got fixed. The slow exposure, daily and mixed and mostly unlabeled, is the one that showed up in 20 years of lung tests.

I’d still put a latch on any cabinet a child can reach. But I’d do everyday counters with a damp cloth and plain soap, save the disinfectant for the cutting board and counter after raw chicken, and I won’t buy another spray whose only listed scent ingredient is “fragrance.”

Sources

1. American Journal of Respiratory and Critical Care Medicine: Svanes et al., cleaning at home and at work and lung function decline, ECRHS (2018)
2. American Journal of Respiratory and Critical Care Medicine: Cummings and Virji (NIOSH), editorial on cleaning and lung function (2018)
3. Environmental Health Perspectives: “Scented Products Emit a Bouquet of VOCs,” on Steinemann et al. (2011)
4. Toxicology Reports: Kostoff, Goumenou and Tsatsakis, “The role of toxic stimuli combinations in determining safe exposure limits” (2018)
5. NYU Langone Health: Trasande, Liu and Bao, phthalates and attributable mortality in Environmental Pollution (2021)
6. MedicalResearch.com: Trasande on urinary phthalates and insulin resistance in adolescents, Pediatrics (2013)
7. Retail Industry Leaders Association: deadlines for the California Cleaning Product Right to Know Act (2019)
8. TrialSite News: Kostoff, “From Household Hazards to Disease: An Integrated Study” (2026)