Power lines, wiring, appliances and motors create low-frequency fields. What do we measure—and what does the health research actually show?
Power frequency is usually 50 or 60 cycles per second. It belongs to the extremely low-frequency (ELF) part of the spectrum. A number without a field type and unit is incomplete.
Volts per metre. The field can be present when a device is connected to power, even when it is switched off.
Microtesla (µT) and milligauss (mG) are two units for the same magnetic-field measurement. Fields depend on current, distance and how the source is arranged.
Diagrams are schematic, not measured field maps. NCI: electric and magnetic fields ↗
Low-frequency fields can induce electric fields inside the body and stimulate nerves and muscles. Exposure guidance addresses these established effects.
Childhood-leukemia studies have reported associations, but results vary. Bias, chance and other explanations remain under discussion; a causal link has not been established.
ICNIRP: low-frequency effects and protection ↗
An odds ratio compares two groups. A value of 1 means the estimated odds were the same. A value above 1 means higher estimated odds in the exposed group.
It does not tell you how many children developed the disease. To understand that, you also need to know how common the disease was.
The studies below use different exposure categories. Their numbers should not be read as a time trend.
Look at the estimate and the range beside it. That range is the 95% confidence interval: a way of showing statistical uncertainty. It does not account for every possible problem in the study.
Seven studies pooled by Kheifets and colleagues compared residential exposures ≥0.3 µT with <0.1 µT.
The estimate was elevated, but the interval includes 1. The result was not statistically significant.
Amoon and colleagues combined four studies: 24,994 cases and 30,769 controls.
For ≥0.4 µT versus <0.1 µT, the estimate was close to 1. The authors considered methodological differences and chance among possible explanations for the change from older findings.
Brabant and colleagues reviewed 38 studies. Twelve contributed to the measured-field comparison shown here. Published online in 2022; print edition 2023.
For >0.4 µT, the interval was above 1. This was a statistically significant association in observational data, not proof that the fields caused leukemia.
These analyses overlap in their underlying research and use different inclusion rules and exposure categories. They are not three independent experiments or a trend line. Study categories such as 0.3 or 0.4 µT are not proven safety boundaries.
Researchers used records to group apartments by their position relative to an electrical transformer, then compared cancer cases. Even across five countries, the highest category contained only three cases.
Evidence for an elevated risk was weak: RR 1.1, with a 95% interval from 0.3 to 3.8 in one model.
The wide interval permits very different interpretations. Location stood in for a direct exposure measurement, and different analyses gave different estimates.
| Vehicle group | Measured mean |
|---|---|
| Diesel · 3 models | 0.02 µT |
| Gasoline · 4 models | 0.04–0.05 µT |
| Hybrid · 3 models | 0.06–0.09 µT |
Fields differed by vehicle, seat and driving condition. Hybrid averages were higher in this sample.
No health outcomes were measured. Frequencies below 30 Hz were excluded, and these older models do not represent every vehicle today.
Electric or magnetic? Which frequency range does the instrument measure?
Distance, position, appliance or vehicle mode, and other sources nearby.
A momentary peak, an average over minutes or a longer personal record?
Instrument, units, calibration and what the measurement cannot resolve.
These are prompts for documenting evidence, not a diagnostic protocol. They help distinguish what was measured from what is being inferred.