Which swaps lower your chemical exposure the most?

This page ranks swaps by what trials measured in urine or blood. Biomarkers show how much a swap changes your dose, not how much illness it prevents.

1. Food and packaging: biggest and fastest. Three days of fresh, unpackaged food cut urinary BPA by 66% and DEHP phthalate metabolites by 53% to 56% in five families¹. In a crossover trial, a daily serving of canned soup for five days raised urinary BPA by 1221% compared with fresh soup². Fresh is not automatic: one dietary trial raised DEHP when the "fresh" ingredients were themselves contaminated⁷.

2. Personal care: fast and moderate. Three days of low-chemical products lowered monoethyl phthalate by 27%, methyl and propyl parabens by 44% to 45%, triclosan by 36%, and benzophenone-3 by 36% in 100 teenage girls⁴.

3. Produce: large for pesticide residues. An organic diet cut 13 pesticide markers in four families, including clothianidin by 83% and a chlorpyrifos metabolite by 61%⁸. Whether that improves health is untested.

4. Contaminated water: large and lasting. A point-of-use arsenic filter program cut urinary arsenic by 47% over two years³. After clean water arrives, PFAS fall slowly, with half-lives of 2.7 to 5.3 years, so early action matters⁵. Real-world PFAS filter performance is poorly characterized, and short-chain PFAS may pass through⁹.

5. Dust and furniture: small and slow. Children in clean-looking housing had 12% to 17% lower phthalate metabolite levels¹⁰, but dust control did not lower blood lead in a Cochrane review⁶. Replacing foam furniture with newer flame-retardant-free pieces shortened the median time for blood PBDEs to halve to 1.4 years, versus 2.6 to 5.2 years without replacement¹¹.

6. Air and cookware. Replacing an unvented gas stove with an electric one cut kitchen nitrogen dioxide by 51%, while a range hood alone did not¹². We found no human trials measuring body-level change from swapping cookware, so it ranks last.

Go deeper.

• Food and plastics: BPA, phthalates, microplastics, organic food.
• Personal care: hidden chemicals, shampoo, cosmetics, deodorant.
• Water and cookware: choosing a filter, what's in your water, avoiding PFAS, safer cookware.
• Cleaning and laundry: cleaning products, bleach and quats, vinegar and baking soda, laundry.
• Bugs and pets: pest control, repellents, flea and tick treatments.
• Air, dust, and furniture: house dust, gas stoves, air purifiers, diffusers, candles and air fresheners, flame retardants, furniture, clothing, outdoor air.
• Other home hazards: mold, radon, lead in water, humidity and dust mites, stale air, wildfire smoke.

Fresh food (Rudel 2011). Twenty participants in five families ate their usual diet, then three days of food that was not canned or packaged in plastic, then returned to usual; evening urine was pooled for each period¹. Geometric mean BPA fell from 3.7 to 1.2 ng/mL and the DEHP metabolite MEHHP from 57 to 25 ng/mL, with maximum values down 76% for BPA and 93% to 96% for DEHP metabolites¹. The sample was small and the families were chosen for heavy packaged-food use¹. In the Harvard crossover trial, 75 of 84 volunteers ate canned and fresh soup for five days each; geometric mean BPA was 1.1 µg/L after fresh and 20.8 µg/L after canned, a 22.5 µg/L difference (95% CI 19.6 to 25.5), while the authors called the peak likely transient².

A trial that backfired (Sathyanarayana 2013). Ten families were randomized to five days of complete dietary replacement (21 people) or written advice (19)⁷. In the replacement arm, DEHP metabolites rose from a median 283.7 to 7,027.5 nmol/g, and ingredient testing found 21,400 ng/g of DEHP in ground coriander and 673 ng/g in milk⁷. Estimated child intake exceeded the EPA reference dose⁷.

Personal care (Harley 2016). In 100 Latina girls, three days of products labeled free of the target chemicals lowered monoethyl phthalate 27.4% (95% CI 13.2 to 39.3), methyl paraben 43.9%, propyl paraben 45.4%, triclosan 35.7%, and benzophenone-3 36.0%⁴. Two phthalate metabolites did not change, and ethyl and butyl parabens rose, mostly at low levels⁴.

Organic diet (Hyland 2019). Sixteen people in four US families gave 158 urine samples before and after an organic diet⁸. Thirteen metabolites or parent compounds fell, including clothianidin (-82.7%), a malathion metabolite (-95.0%), and a chlorpyrifos metabolite (-60.7%)⁸.

Dust (Sears 2020; Cochrane 2016). In the HOME Study, 355 pregnant women were randomized to lead-hazard or injury-prevention interventions, and 288 children were followed¹⁰. The lead-hazard group had 11% to 12% lower DEHP metabolites; in observational analyses, children in clean-looking homes had 12% to 17% lower levels¹⁰. The Cochrane review of 14 trials in 2,643 children found education ineffective for blood lead and dust control showing little or no difference (mean difference -0.15, 95% CI -0.42 to 0.11), with low-quality evidence⁶. In a 2026 study, 25 people who replaced foam or upholstered furniture with items made after California's updated flammability standard were compared with 28 who did not; the time for serum BDE-47, -99, and -100 to halve was 1.9 to 3.9 times longer without replacement, while short-lived organophosphate flame-retardant metabolites varied too much to judge¹¹.

Water. In a cluster randomized trial, both arms installed a point-of-use arsenic filter and urinary arsenic fell from 13.2 to 7.0 µg/g creatinine (47%) over two years, with no difference between phone-only and home-visit support³. After Ronneby, Sweden switched to clean water, 106 residents' serum fell 13%, 20%, and 26% per year for PFHxS, PFOS, and PFOA, equal to half-lives of 5.3, 3.4, and 2.7 years⁵. A review found only five real-life studies of tap-water PFAS removal and warned that poorly maintained carbon systems might raise PFAS⁹.

Air. In a three-arm randomized trial in homes with unvented gas stoves, stove replacement cut median NO2 by 51% in the kitchen and 42% in the bedroom at three months; a purifier cut it by 20% in the kitchen, and a hood made no significant difference¹².

Null and conflicting results. Fresh food backfired in one trial⁷, dust control did not lower blood lead⁶, hoods did not cut NO2¹², and two phthalate markers did not change with personal-care swaps⁴. BDE-153 fell no faster after furniture replacement¹¹. None of these trials measured disease.

Bottom line (confidence 3 of 5). Food packaging, personal care, and contaminated water give the largest measured drops; dust and air give smaller ones. Pick one or two swaps, not all of them.