A news report says a chemical was detected in the urine of nearly every American tested. Another says a common ingredient is linked to a disease. A sign at the entrance to a parking garage warns you about cancer. None of those three statements tells you whether anything in your life needs to change.
That is not because they are false. All three can be perfectly accurate. They answer questions you were not asking.
Three separate questions sit under every chemical scare, and most confusing coverage depends on them being mashed together. Can this substance cause harm at all? At what dose does it do so? And how much of it do you get?
Our article on pesticides worked one case of this, explaining why EPA and the International Agency for Research on Cancer reached different conclusions about the same herbicide. This article generalizes the method and hands you the tools that go with it: biomonitoring results, concentration units, reference doses, cancer classifications and the warning signs that are now on everything.
Three questions, asked in one order
Federal risk assessment is not improvised. EPA’s human health risk assessment page, updated in January 2026, lays out four steps after a planning stage. Hazard identification asks whether a stressor can harm people and under what circumstances. Dose-response assessment works out the numerical relationship between exposure and effect. Exposure assessment examines what is known about the frequency, timing and levels of contact. Risk characterization combines the three into an estimate and a description of what it means, uncertainties included.
Collapse that into plain language and the order follows.
Hazard is a property of the substance. It travels with the chemical and does not depend on you.
Dose-response is a property of the biology. It asks at what amount something starts to happen, and that is a different question from whether anything can happen.
Exposure is a property of your life. Where you live, what you do for work, what you eat, how your home is ventilated.
A yes to question one therefore tells you nothing on its own. Water, oxygen, table salt and sunlight are all capable of killing a person, and nobody organizes their week around that. Most consumer confusion does not happen between hazard and risk. It happens between questions two and three, where a dose that produced an effect in a laboratory gets reported alongside an amount nobody receives.
What makes an exposure matter
Five things, and only the first is about the chemical.
Route. EPA’s exposure assessment toolbox separates inhalation, ingestion and skin contact, because the same substance behaves differently depending on how it gets in. A compound broken down rapidly after swallowing can act very differently when injected or inhaled.
Dose. How much reaches you, not how much is in the container.
Duration and frequency. A solvent used once a year in an open garage and the same solvent used twice a day in a closed bathroom are not the same exposure.
Timing. Development has windows when hormonal signals program tissues for decades. Our article on endocrine-disruptor claims covers that.
Who you are. Body size, pregnancy, age, kidney and liver function, existing illness and occupation all turn the same external exposure into a different internal dose.
Finding a chemical in your body is a measurement, not a diagnosis
This is the single most important sentence in the article, and the Centers for Disease Control and Prevention says it about its own data.
CDC’s National Report on Human Exposure to Environmental Chemicals measures chemicals in blood, serum and urine from participants in the National Health and Nutrition Examination Survey. Chemicals get into the report for practical reasons CDC publishes: evidence of exposure in the population, seriousness of the suspected health effects, availability of accurate laboratory methods, enough samples, and affordable analysis.
On its page explaining how to interpret those tables, CDC states: “The measurement of an environmental chemical in a person’s blood or urine does not mean, by itself, that the chemical causes disease.”
The agency gives its own reason. Analytical methods have improved to the point where very low levels can be measured, and separate research on varying exposure levels is needed to establish whether any particular blood or urine level is associated with disease. For some chemicals, lead among them, CDC says the health risks at different blood levels are well understood. For many others, more research is needed before anyone can say what a given blood or urine level means for health.
The mechanism behind that explains a whole genre of headline. CDC sets a limit of detection for each assay, defined as the point at which a reading has a 95 percent probability of landing above zero. The limit is specific to the survey period, and it moves between cycles as methods improve. A result reported as below the limit in an earlier cycle can come back as an actual number later, which is why CDC tells users to apply the limit belonging to the survey period rather than comparing across cycles carelessly.
So when a chemical goes from undetected to detected in nearly everyone, the most likely thing that changed is the instrument.
Three further cautions. Blood and urine levels reflect total exposure from all sources and do not identify which source or route produced them. CDC does not test individual patients, and most of these are research measurements unavailable in hospitals or commercial clinical laboratories. And for chemicals the body clears quickly, a single urine sample is a poor measure of ordinary long-term exposure, a limitation Duk-Hee Lee and David Jacobs named in a 2015 review in Reviews in Endocrine and Metabolic Disorders as one specific reason human studies in this field disagree.
A smaller unit is not a bigger problem
Modern instruments report concentrations at scales nobody has a feel for, and the words make it worse. Per million, per billion and per trillion look like three steps of similar size. A factor of a million separates the first from the last.
That spacing is worth absorbing once and then setting aside, because the unit itself carries no information about danger. Small concentrations can matter. Hormones work at these levels, and so do some toxicants.
The history is the giveaway. Over several decades routine environmental measurements moved from parts per million to parts per billion to parts per trillion. The chemicals did not get more dangerous. The detectors got better.
One more wrinkle for reading a biomonitoring table. CDC reports urine results both per litre and per gram of creatinine, because creatinine correction adjusts for how dilute the sample was. One person can therefore produce two different-looking numbers from one sample, and comparing a corrected result with an uncorrected one is a mistake.
The numbers that already have safety factors inside them
Agencies publish daily intake figures with names that sound like bright lines and are not.
EPA’s reference dose is a daily amount, expressed in milligrams per kilogram of body weight per day, that a person could take in every day for a lifetime without an appreciable risk of harm, people who are more sensitive than average included. EPA builds it from the highest tested dose that produced no statistically or biologically meaningful adverse effect, then divides.
What it divides by is the interesting part. Each uncertainty factor is usually tenfold. One covers variation in sensitivity among people, one covers extrapolating from animals to humans, one covers extrapolating from a shorter study to lifetime exposure, and one covers using a lowest-effect dose when no no-effect dose was found. A modifying factor based on professional judgment, covering things like how complete the database is, sits on top. A solid long-term animal study typically gets divided by 100, a shorter one by 1,000.
FDA uses the same logic under a different name for food. An acceptable daily intake, as FDA describes it, is how much of a substance a person could eat every day across a whole lifetime and still be considered safe. FDA starts from the highest animal dose that caused no adverse effect and then applies a safety factor to bring it down. FDA’s acceptable daily intake for aspartame is 50 milligrams per kilogram of body weight per day, and FDA notes that a 132-pound person would need about 75 tabletop packets in one day to reach it.
Two conclusions follow, and people usually draw only one.
Exceeding one of these numbers is a reason to look harder, not a declaration that harm has occurred, because a hundredfold or thousandfold margin was deliberately built in. And sitting under one is not a guarantee either, because the size of those factors is a judgment, and judgments get revisited.
When two regulators land far apart
Here is what that revisiting looks like, and it is the cleanest worked example available.
FDA’s page on bisphenol A in food contact applications, with content current as of April 2023, concludes that the amounts people now get from food are safe. FDA points to rodent studies dosed at roughly 100 to 1,000 times human food exposure and to swallowed BPA being rapidly converted to an inactive form.
The European Food Safety Authority published its re-evaluation on 19 April 2023 and set a tolerable daily intake of 0.2 nanograms per kilogram of body weight per day. EFSA’s own plain-language summary states that this is 20,000 times lower than the temporary figure of 4 micrograms per kilogram per day it had set in 2015, and that estimated dietary exposure runs two to three orders of magnitude above the new value. For a 70-kilogram adult, 0.2 nanograms per kilogram comes to about 14 billionths of a gram a day.
Neither body is lying, and neither number is a typo. The gap comes from the judgment layer: which effect was treated as adverse, and how much uncertainty was priced in. EFSA’s summary records an additional uncertainty factor of 2 applied for uncertainty about the toxicological reference point itself. Our article on bisphenols covers this chemical in full; what belongs here is the structure. When a tolerable intake moves by a factor of 20,000 while the molecule stays the same, what changed was the interpretation. Knowing that lets you read the next regulatory disagreement without concluding that science is broken.
“Probably carcinogenic” describes the evidence, not the danger
This is the most misreported scientific classification in consumer media, and the misreading is specific.
The International Agency for Research on Cancer, the World Health Organization’s cancer research arm, runs a long-standing program that reviews agents for cancer hazard. In a 2020 commentary in the Journal of the National Cancer Institute, Jonathan Samet and colleagues, including the IARC staff who run the program, described the 2019 overhaul of its procedures. More than 1,000 agents have been evaluated since 1971, and slightly more than half were placed in one of the categories indicating a carcinogenic hazard.
The authors explain why the series was renamed to foreground hazard identification. Hazard, in their framing, concerns how strong the evidence is that an agent can cause cancer. Risk concerns the probability that a given exposure will produce cancer. The program does the first job. It does not estimate how much cancer a given dose would cause.
Three streams of evidence get reviewed: cancer in humans, cancer in experimental animals, and mechanistic evidence, with human exposure characterized alongside. Each stream is graded for strength, and since 2019 they are integrated in a single step into one of four overall categories rather than five.
So a phrase like probably carcinogenic to humans is a statement about how much confidence the evidence supports. It is not a measure of potency, and it says nothing about how much of the agent anyone encounters. Two agents in the same category can differ by orders of magnitude in both.
Two more details change how the list should be read. Agents are selected for review only when there is evidence of human exposure and some evidence suggesting carcinogenicity, so the catalogue lists things worth checking rather than ranking dangers. Dinesh Barupal and colleagues, writing in Environment International in 2021, describe 119 agents recommended by an advisory group for evaluation across 2020 to 2024. And the lowest category is not a clean bill of health. Samet and colleagues state that it is not a determination of non-carcinogenicity or of overall safety, and often means unknown potential with prominent gaps in the research.
One honest limitation. IARC’s own website could not be opened while this article was researched, so no specific agent’s current classification appears here. For that, go to IARC’s site directly rather than to any secondary account.
Why there is a warning on the parking garage
California’s Proposition 65 is formally the Safe Drinking Water and Toxic Enforcement Act of 1986. Voters approved it in November 1986 by 63 percent to 37 percent. It requires a warning before people are exposed to a listed chemical, and it reaches any business with at least ten employees that operates in California or sells into it, catalog and internet sales included.
The enforcement design explains the signs. The state does not decide where warnings go. The Office of Environmental Health Hazard Assessment, which administers the program, says on its warnings website that each business has to work out for itself whether its own product or premises exposes people at a level that triggers the duty, and that nothing needs to be posted when a single product or location contributes too little to raise anyone’s risk meaningfully.
Put a business in that position, with litigation as the penalty for guessing wrong in one direction and a sticker as the cost of guessing wrong in the other, and you get a sticker.
OEHHA states the consequence plainly: “A Proposition 65 warning does not mean that a product or place violates safety or health standards.”
A warning tells you a business believes an exposure exists. It does not tell you which chemical, how much, or how large the risk is. OEHHA advises contacting the manufacturer or importer to ask what prompted it, and publishes roughly 75 fact sheets covering specific chemicals, products and places. Its general fact sheet, posted in September 2023, makes two points worth keeping. How much risk a chemical poses to you turns on how much of it you meet. And because these substances are everywhere, everyone carries some risk from them.
Proposition 65 is state regulation, which puts it well above an undefined marketing claim. It is still a disclosure rule rather than a safety standard, and treating it as one leads people to fear parking garages and ignore the exposures that fill their days.
Five checks for the next headline
- Animal or human? If animal, which species, and how was the substance administered? Injected, fed in the diet, or applied to cells in a dish?
- What dose, against what dose do people get? Ask for the ratio. If a study used a hundred times typical human intake, that is the story.
- Detection or effect? A chemical measured in blood is a detection. An outcome measured in people is an effect. Headlines routinely report the first as though it were the second.
- Association or experiment? Linked to almost always means an association found in an observational study, where people who differ in one exposure usually differ in a dozen other ways too. An association is where an investigation starts, not where it ends.
- Relative or absolute, and who funded it? A doubling of a rare outcome is still rare, and our article on reading health information covers that arithmetic. As for funding, note what the researchers disclosed. The IARC program bars working group members from consulting on the agents they reviewed until the volume is published, which is a useful standard to measure other disclosures against.
What to do with a frightening number
Before reacting to a chemical story, work out which of the three questions it answered. Most answer the first, a few answer the second, and almost none answer the third, which is the only one that describes you.
Hazard travels with the chemical and you cannot change it. Dose-response belongs to the biology and you cannot change that either. Exposure is yours, and it is where every useful decision lives, which is why route and duration are worth more attention than any ingredient list.
If you think someone has been exposed to something, do not spend the first ten minutes reading. Call Poison Help at 1-800-222-1222, 24 hours a day, and keep the container, because the label tells the responder what the product is. Call 911 if the person is unconscious or having trouble breathing.
If you want your own exposure assessed, such as a workplace chemical or something you encountered while pregnant, that is a conversation with your own clinician or an occupational medicine service. CDC does not test individuals, and most of the assays behind those national numbers are research methods rather than clinical tests, so there is no general-purpose chemical panel you can order to settle the question.
Sources
- Centers for Disease Control and Prevention. “Interpretation of the Data (National Report on Human Exposure to Environmental Chemicals).” cdc.gov
- Centers for Disease Control and Prevention. “Frequently Asked Questions (National Exposure Report).” cdc.gov
- Centers for Disease Control and Prevention. “Limit of Detection (National Exposure Report).” cdc.gov
- Centers for Disease Control and Prevention. “Units (National Exposure Report).” cdc.gov
- Centers for Disease Control and Prevention. “About the Chemicals (National Exposure Report).” cdc.gov
- Centers for Disease Control and Prevention. “About the National Exposure Report.” cdc.gov
- U.S. Environmental Protection Agency. “Human Health Risk Assessment.” epa.gov
- U.S. Environmental Protection Agency. “EPA ExpoBox.” epa.gov
- U.S. Environmental Protection Agency. “Reference Dose (RfD): Description and Use in Health Risk Assessments.” epa.gov
- U.S. Food and Drug Administration. “Aspartame and Other Sweeteners in Food.” fda.gov
- U.S. Food and Drug Administration. “Bisphenol A (BPA): Use in Food Contact Application.” fda.gov
- European Food Safety Authority. “Re-evaluation of the risks to public health related to the presence of bisphenol A (BPA) in foodstuffs (plain language summary).” efsa.europa.eu
- California Office of Environmental Health Hazard Assessment. “Proposition 65 Warnings Website.” p65warnings.ca.gov
- California Office of Environmental Health Hazard Assessment. “About Proposition 65.” p65warnings.ca.gov
- California Office of Environmental Health Hazard Assessment. “Frequently Asked Questions.” p65warnings.ca.gov
- California Office of Environmental Health Hazard Assessment. “Toxic Chemicals, Proposition 65 Warnings, and Your Health: The Big Picture.” p65warnings.ca.gov
- PubMed. “Samet JM, et al. "The IARC Monographs: Updated Procedures for Modern and Transparent Evidence Synthesis in Cancer Hazard Identification." Journal of the National Cancer Institute, 2020. DOI: 10.1093/jnci/djz169.” doi.org
- PubMed. “Barupal DK, et al. "Prioritizing cancer hazard assessments for IARC Monographs using an integrated approach of database fusion and text mining." Environment International, 2021. DOI: 10.1016/j.envint.2021.106624.” doi.org
- PubMed. “Lee DH, Jacobs DR. "Methodological issues in human studies of endocrine disrupting chemicals." Reviews in Endocrine and Metabolic Disorders, 2015. DOI: 10.1007/s11154-016-9340-9.” doi.org