EMF and Health: What the Evidence Actually Shows

An honest review of EMF health research — IARC classifications, NTP study, childhood leukemia data, and EHS. What's established, what's uncertain, and what to do.

You've got a WiFi router in the hallway, a smart meter on the side of the house, a phone on the nightstand, and a question: should I actually be worried about any of this? Maybe you've read something alarming, or maybe someone told you it's all nonsense. Either way, you want a straight answer.

The honest answer is that the evidence is more interesting, and more uncertain, than either side usually admits. This page lays out what the research shows, with citations you can check. We'll be direct about what's strong, what's uncertain, and what's simply unknown.

The Basics: Ionizing vs. Non-Ionizing Radiation

The electromagnetic spectrum runs from extremely low frequency (ELF) fields, produced by power lines and household wiring at 50/60 Hz, through radio frequencies (RF), microwaves, infrared, visible light, ultraviolet, X-rays, and gamma rays. The dividing line that matters for biology falls in the ultraviolet range: everything above that frequency is ionizing radiation, meaning it carries enough energy per photon to strip electrons from atoms and directly break chemical bonds in DNA. X-rays and gamma rays are ionizing. Their cancer risk is well established and not debated.

Everything below that line, power-frequency fields, radio waves, microwaves, infrared, visible light, is non-ionizing radiation. Individual photons don't carry enough energy to break molecular bonds. This is the basis for the common claim that non-ionizing EMF "can't cause biological effects." The reasoning is straightforward, but as we'll see, the experimental evidence doesn't entirely cooperate with that conclusion.

The EMFs relevant to building biology fall into two categories:

Extremely low frequency (ELF) fields, produced by power lines, household wiring, electrical panels, and appliances. These operate at 50 Hz (Europe) or 60 Hz (North America). They include both electric fields (from voltage) and magnetic fields (from current flow).
Radiofrequency (RF) radiation, produced by wireless devices, cell towers, WiFi routers, smart meters, Bluetooth, and cordless phones. Frequencies typically range from about 700 MHz to 6 GHz for common consumer devices, with 5G millimeter-wave extending into the 24–100 GHz range.

IARC Classifications: What "Possibly Carcinogenic" Actually Means

Two of the most frequently cited data points in this field are the classifications issued by the International Agency for Research on Cancer (IARC), part of the World Health Organization.

In 2002, IARC classified ELF magnetic fields as Group 2B, possibly carcinogenic to humans. The basis was epidemiological evidence showing a consistent statistical association between residential magnetic field exposure above 3–4 milligauss (300–400 nanotesla) and an approximately twofold increase in childhood leukemia risk.

In 2011, IARC classified RF electromagnetic fields as Group 2B as well. This was based on limited evidence of an increased risk for glioma (a type of brain cancer) associated with heavy, long-term wireless phone use, drawing primarily from the Interphone study and research by the Hardell group in Sweden.

Group 2B is widely misunderstood, often overstated by one side and dismissed by the other. Here is what it means in IARC's own framework:

Group 1. Carcinogenic to humans (sufficient evidence). Examples: tobacco, asbestos, processed meat.
Group 2A. Probably carcinogenic (strong evidence, not quite conclusive). Examples: red meat, night shift work, glyphosate.
Group 2B. Possibly carcinogenic (limited evidence of carcinogenicity in humans, or sufficient evidence in animals, or strong mechanistic evidence). Examples: ELF magnetic fields, RF radiation, pickled vegetables, talcum powder (body), some gasoline exhaust components.
Group 3. Not classifiable (inadequate evidence).

Group 2B means the evidence is strong enough that cancer risk cannot be ruled out, but not strong enough to confirm it. IARC didn't say EMF causes cancer. It didn't say EMF is safe. It said the evidence warrants attention and further study.

Some researchers, including members of IARC advisory panels, have argued that accumulated evidence since 2011 warrants upgrading RF to Group 2A or even Group 1. As of this writing, IARC has not issued an updated classification, though a re-evaluation has been discussed.

Key Studies: What Was Found and What It Means

Childhood Leukemia and ELF Magnetic Fields

The association between power-frequency magnetic fields and childhood leukemia is one of the most studied environmental health questions of the last forty years. It began with Wertheimer and Leeper's 1979 epidemiological study in Denver and has been followed by dozens of studies across multiple countries.

Two large pooled analyses. Ahlbom et al. (2000) and Greenland et al. (2000), combined data from multiple studies and found a statistically significant, approximately twofold increase in childhood leukemia risk at average magnetic field exposures above 3–4 milligauss. These pooled analyses formed the primary basis for IARC's 2002 Group 2B classification.

What makes this association notable:

It has been observed consistently across studies in different countries with different methodologies.
It is specific, it appears for leukemia but not for most other childhood cancers.
A dose-response relationship has been observed: higher exposure correlates with higher risk.

What limits the conclusion:

No biological mechanism has been firmly established to explain how ELF magnetic fields at these levels could cause leukemia.
Selection bias and confounding factors cannot be entirely excluded in epidemiological studies.
The absolute number of cases attributable to magnetic field exposure (if the association is causal) is small, because few children are exposed above the 3–4 mG threshold.

The SBM-2008 standard sets its "No Concern" threshold for AC magnetic fields at less than 0.2 mG (20 nT) in sleeping areas, well below the level associated with increased leukemia risk in the epidemiology.

The NTP Study (2018)

The National Toxicology Program (NTP) study was a $30 million, decade-long investigation conducted by the U.S. government, the largest and most rigorous animal study on cell phone radiation ever undertaken. Rats and mice were exposed to whole-body RF radiation at 900 MHz (GSM and CDMA modulations) for two years, with exposure levels of 1.5, 3, and 6 W/kg specific absorption rate (SAR).

Key findings in male rats:

"Clear evidence" of malignant schwannomas (tumors of the nerve sheath) in the heart. NTP's highest confidence rating.
"Some evidence" of malignant gliomas in the brain.
A dose-response relationship: higher exposure levels produced more tumors.
Female rats and both sexes of mice showed weaker or equivocal results.

The study matters because of its rigor, government sponsorship, and clear dose-response findings. It is also legitimately controversial:

The lowest exposure level (1.5 W/kg) was roughly four times the maximum SAR permitted for consumer cell phones in the U.S. (though the highest tier overlapped with levels that can occur in body tissue during phone use close to the head).
Exposed rats lived longer than unexposed controls, which some critics argue complicates the tumor findings (longer-lived animals have more time to develop tumors).
The whole-body exposure protocol differs from the localized exposure pattern of actual phone use.

Supporters note that the tumor types found (schwannomas and gliomas) are the same types flagged in the human epidemiology, and that dose-response relationships are a classical indicator of a real effect, not a statistical artifact.

The Ramazzini Institute Study (2018)

Shortly after NTP's results were released, Italy's Ramazzini Institute published findings from an independent, large-scale animal study. Researchers exposed over 2,400 rats to RF radiation at frequencies simulating cell tower emissions (1.8 GHz), at exposure levels far lower than the NTP study, specifically 0.001, 0.03, and 0.1 W/kg. These levels fall within the range of actual environmental exposures from cell towers.

The Ramazzini study found a statistically significant increase in heart schwannomas in male rats at the highest exposure level, the same rare tumor type identified in the NTP study.

Two independent labs, different frequencies, different exposure levels, different continents, same uncommon tumor. That convergence is hard to dismiss, and many toxicologists don't. Critics point to the small number of tumors and methodological differences between the two studies.

The BioInitiative Report (2012)

The BioInitiative Report is a review of over 1,800 peer-reviewed studies on the biological effects of EMF, compiled by a group of independent scientists and researchers. It concludes that existing public safety limits are inadequate and recommends substantially lower exposure guidelines.

The report is frequently cited by those advocating for stricter EMF standards. It is also criticized:

Critics (including some national health agencies) argue the report selectively emphasizes studies that found effects while underweighting null results, and that some contributors have advocacy positions.
Defenders argue the report applies a weight-of-evidence approach appropriate for emerging risks and that dismissing 1,800 studies requires its own justification.

Whether or not you find the BioInitiative Report persuasive as a whole, the primary studies it catalogs remain part of the scientific record and are worth reading on their own terms.

What Remains Uncertain

The Mechanism Question

Classical physics predicts that non-ionizing radiation at the power levels we encounter in daily life should not have enough energy to damage DNA or disrupt biological processes. This is the single strongest argument against EMF health effects, and it is a real argument, not a dismissal.

But biological systems are not simple physics problems. Several research groups have documented oxidative stress, an increase in reactive oxygen species (free radicals), in cells exposed to non-ionizing EMF. A 2015 meta-analysis by Yakymenko et al. found that 93 out of 100 peer-reviewed studies reported oxidative effects from low-intensity RF radiation. Oxidative stress is a known pathway to DNA damage and is implicated in cancer, neurodegeneration, and cardiovascular disease.

Other proposed mechanisms include effects on voltage-gated calcium channels (the Pall hypothesis), disruption of cellular signaling, and interference with melatonin production. None of these is universally accepted, and some remain speculative. But the oxidative stress findings, at minimum, suggest that "non-ionizing means non-biological" is an oversimplification.

Association vs. Causation

Epidemiological studies, no matter how consistent, show correlation, not causation. The childhood leukemia association could, in theory, be explained by some unmeasured confounding variable that correlates with living near power lines. This is a standard limitation of observational research, and it applies equally to many accepted public health findings (the link between secondhand smoke and lung cancer, for example, was initially epidemiological).

Generalizability

Most studies test specific frequencies, modulations, power levels, and exposure durations. Real-world exposure is a mixture of multiple simultaneous sources with varying characteristics. Whether findings for 900 MHz GSM signals in rats translate to human exposure from a WiFi router operating at 2.4 GHz with different modulation is a genuinely open question.

Replication

Some EMF bioeffects studies have been replicated; others have not. This is partly because controlling all relevant variables, exact frequency, modulation, power density, exposure duration, biological model, environmental conditions, is genuinely difficult. Inconsistent replication doesn't prove the original finding was wrong, but it does mean caution is warranted.

Electromagnetic Hypersensitivity (EHS)

Some people report symptoms, headaches, fatigue, difficulty concentrating, sleep disturbance, skin tingling, that they attribute to EMF exposure. This condition is commonly called electromagnetic hypersensitivity or EHS.

The World Health Organization acknowledges that the symptoms are real and can be disabling, but states that "current scientific evidence does not support a link between these symptoms and EMF exposure." This conclusion is based largely on double-blind provocation studies, in which subjects are exposed to EMF or sham (no EMF) conditions without knowing which is which. In most of these studies, self-identified EHS individuals cannot reliably detect whether EMF is present.

There are legitimate criticisms of the provocation study design, they typically test single frequencies for short durations in laboratory settings, which may not replicate the chronic, multi-source exposures people experience at home. Some researchers also note that the stress of a lab environment could mask detection ability.

The building biology approach to EHS is pragmatic: measure the person's actual environment, identify and reduce the dominant sources, and observe whether symptoms improve. This sidesteps the causation debate entirely. If someone feels better after reducing EMF in their bedroom, the practical outcome matters regardless of whether a double-blind study would confirm the mechanism. If they don't feel better, that's useful information too.

This approach doesn't require believing EMF causes the symptoms. It only requires that reducing exposure is low-cost and low-risk, and that the outcome, improvement or no change, tells you something either way.

The Precautionary Principle

The precautionary principle holds that when an activity raises threats of harm to health or the environment, precautionary measures should be taken even if some cause-and-effect relationships are not fully established scientifically. It does not require proof of harm, it requires reasonable evidence of possible harm combined with feasible protective measures.

This is not a radical concept. It underpins much of environmental and public health law in the European Union, food safety, chemical regulation, workplace exposure limits. We mandated seatbelts before crash fatality data looked the way it does today. We restricted lead in paint and gasoline based on evidence that accumulated over decades, not overnight.

The SBM-2008 standard embodies this principle explicitly. Its thresholds are not based on proven harm at specific exposure levels. They reflect what long-term precaution looks like: keeping sleeping area exposures as close to natural background as technology and reasonable effort allow. The standard acknowledges this openly, it doesn't claim that exceeding its thresholds will make you sick, only that staying below them represents a sensible margin of safety given what we know and what we don't.

The alternative is waiting for conclusive proof, which, for chronic low-level exposures in complex biological systems, may take decades to arrive if it arrives at all. Whether that wait is acceptable depends on your reading of the evidence and your tolerance for uncertainty.

What to Do: A Practical Framework

If you've read this far, here's what acting on it looks like.

1. Measure Your Environment

You can't make informed decisions about EMF without data. Our home EMF assessment guide walks through the full testing protocol step by step. If you're starting from scratch, the meters buying guide covers what equipment you need and what you can skip. A professional building biologist can do a full assessment if you'd prefer expert help.

2. Prioritize the Bedroom

You spend a third of your life in bed, and sleep is when the body does most of its repair work. The SBM-2008 standard focuses on sleeping areas for this reason. If you're going to address EMF in one room, this is the room.

3. Reduce the Biggest Sources First

Most EMF reduction is neither expensive nor complicated. Common high-impact steps include:

Moving your bed away from a breaker panel on the other side of the wall.
Switching your WiFi router off at night or relocating it away from the bedroom.
Removing cordless DECT phones (or replacing with wired phones).
Switching off the circuit breaker(s) feeding bedroom wiring at night (a demand switch can automate this).
Keeping phones on airplane mode while sleeping.

These changes cost little or nothing and address the most common sources of elevated bedroom exposure. For more targeted guidance based on which types of EMF are elevated in your space, see our guides on RF radiation and magnetic fields.

4. Don't Panic

If the evidence on this page concerns you, that's a reasonable response. But concern should lead to measurement and action, not anxiety. The epidemiological risks, even if real, are modest. The steps to reduce exposure are straightforward. And the precautionary principle is about doing what's reasonable, not about achieving zero exposure in a world that runs on electricity.

Start with the easy changes. Measure again. Adjust. That's the building biology approach: undramatic, and effective.