Dexcom EMF: What a CGM Sensor Actually Transmits

Most devices we write about are things you could walk away from. A continuous glucose monitor is not: it sits on the back of your arm for ten to fifteen days, and then another one goes on. For someone managing type 1 diabetes that is a decade-long habit, hundreds of small radios worn against skin.

So here is the number, from the manufacturer’s own manual. The Dexcom G7 sensor’s maximum output power is 1.0 mW EIRP, and it holds a Bluetooth connection for a few seconds out of every five minutes — one of the smallest transmit budgets of anything you own.

The rest of this piece is what that number means, where Dexcom and Abbott genuinely differ, and the one CGM safety issue that is actually worth acting on — which turns out not to be the radio at all.

What the Sensors Actually Transmit

This piece covers the four CGMs most people ask about: the prescription Dexcom G7 and FreeStyle Libre 3 Plus, and the over-the-counter Dexcom Stelo and Abbott Lingo from the same two manufacturers. (Medtronic’s sensors and the implantable Eversense are also sold in the US, but are outside the scope of this piece.) All four communicate over Bluetooth Low Energy in the 2.4 GHz band.

Device	Radio	Frequency	Max output	Cadence
Dexcom G7 sensor	Bluetooth LE (GFSK)	2.402–2.480 GHz	1.0 mW EIRP	Reading every 5 min
Dexcom receiver	Bluetooth LE	2.402–2.480 GHz	7.4 mW EIRP	Maintains sensor pairing
FreeStyle Libre 3 Plus	Bluetooth LE (GFSK)	2.402–2.480 GHz	2 dBm EIRP (≈1.6 mW)	Reading every 1 min
Libre sensor activation	NFC (ASK)	13.56 MHz	124 dBµV/m field strength	Seconds, at start-up

Those figures are not estimates. The Dexcom numbers come from the technical information appendix of the G7 user guide, which lists TX/RX frequencies of 2.402–2.480 GHz, 1.07 MHz bandwidth, Gaussian frequency-shift keying, and a 10-metre communication range. The Abbott numbers come from the FreeStyle Libre system specifications sheet, which states the sensor’s radio as “Near Field Communication (13.56 MHz RFID); ASK Modulation; 124 dBuV/m; 2.402-2.480 GHz BLE; GFSK; 2dBm EIRP.”

One milliwatt is a hard thing to picture. For scale, a consumer WiFi access point runs on the order of a hundred times that, continuously and in every direction. A phone’s cellular transmitter operates in a different league again — handset power classes put it two to three orders of magnitude above a CGM, and it climbs toward that ceiling exactly when reception is poor. Our guide to phone SAR ratings walks through why that number is the one people should be tracking.

The duty cycle is the part nobody quotes

Peak power tells you what a radio can do at its loudest. Duty cycle tells you how much of the time it is doing it, and for a CGM the gap between those two is enormous.

Dexcom describes it plainly in a troubleshooting section about the phone’s Bluetooth device list. The sensor, it notes, shows as not connected most of the time; its status “changes to connected during the few seconds every 5 minutes when it’s sending your sensor reading to your phone.”

A few seconds out of three hundred is somewhere in the low single-digit percentages. So the realistic picture of a G7 is a one-milliwatt transmitter that is silent roughly ninety-eight percent of the time. The Libre 3 Plus streams five times more often — every minute rather than every five — which is a real difference in transmission count, and still a rounding error next to the phone that receives it.

Scanner Versus Stream: Why the Two Brands Differ

Abbott’s product line changed shape over the last few years, and people who wore an early Libre often carry the old mental model into a conversation about the new one.

The original FreeStyle Libre was a flash monitor. The sensor stored readings and did not transmit at all until you held a reader or phone against it, at which point NFC — a 13.56 MHz near-field link with a range of a couple of centimetres — pulled the data across. Between scans, the standing RF from that sensor was essentially nothing.

Libre 2 added a Bluetooth link so the sensor could push high and low alarms without a scan. Libre 3 and 3 Plus went fully continuous: a reading streams to the phone every minute, with NFC now used mainly to wake and activate a fresh sensor.

That progression is a trade, and it is worth being clear about which side of it matters. The move from scan to stream bought real-time alarms — the feature that wakes someone at 3 a.m. before a low becomes dangerous. It cost a modest increase in low-power transmissions. Anyone weighing those against each other should understand that they are not weighing comparable quantities: one side is a documented safety benefit, the other is a 1.6 mW radio.

The Phone in Your Pocket Is the Bigger Variable

If you are wearing a CGM and want to reduce your RF exposure, the sensor is close to the last place to look. The receiver is the emitter worth thinking about.

A CGM display device — usually a phone — has to stay within about ten metres of the sensor to maintain the link. In practice that means a phone that lives closer to your body than it otherwise might, including overnight, because the alarms are the point. The phone is running cellular, WiFi, and Bluetooth simultaneously, at power levels far above the sensor’s.

This is the same pattern we found with Bluetooth trackers: the tag itself is a beacon operating at a fraction of a milliwatt, while the phone hunting for it is doing the actual transmitting. Our piece on AirTag Bluetooth radiation covers that asymmetry in more depth, and the intuition transfers cleanly here.

The practical consequences are all about phone placement, not sensor placement:

• Keep the phone off the mattress. A bedside table is within easy Bluetooth range of your arm — the ten-metre spec has plenty of headroom for a nightstand. There is no need to sleep with the phone under a pillow for the alarms to work.
• Use the dedicated receiver if you have one. It is a single-purpose device without a cellular modem. Its Bluetooth radio is rated higher than the sensor’s at 7.4 mW EIRP, but it is not doing everything else a phone does.
• Watch the stacking, not the sensor. A CGM plus a smartwatch plus earbuds plus a sleep tracker is four radios on one body. If you are auditing, audit the pile. Our sleep tracker guide covers the overnight subset of that stack.

What We Can and Cannot Say About Wearing One

Here is where the evidence gets thin, and pretending otherwise would not help anyone.

There is no body of research on the long-term RF exposure of continuous glucose monitors specifically. The devices are recent, the exposure is very small, and no one has run the multi-decade cohort study that would settle the question directly. What exists instead is the broader literature on low-power 2.4 GHz sources, which does not identify a mechanism by which single-digit milliwatts at a few percent duty cycle would cause harm, and which is also not the same thing as a study of this device on this population.

There is a second question that gets tangled with the first, and it is worth separating. Some research has looked at whether environmental EMF exposure influences glucose regulation itself — a genuinely interesting and genuinely unsettled area that we cover in our piece on EMF and blood sugar. That is a question about ambient fields affecting metabolism. It is not a question about whether the sensor measuring your glucose is a hazard, and the two should not be collapsed into each other.

What does not require any hedging: a CGM is a medical device that helps prevent severe hypoglycemia, and nothing in this article is a reason to wear one less. If the radio genuinely worries you, the productive conversation is with your endocrinologist about display-device options and alarm settings — not a decision made alone about whether to keep wearing it.

The Difference That Actually Matters: Medical Imaging

If you take one operational thing from this piece, take this one, because it is brand-specific and it has changed recently.

Dexcom G7 is contraindicated for MRI. The user guide is unambiguous: don’t wear any G7 component during magnetic resonance imaging or diathermy treatment, and for a CT scan, keep the sensor outside the scanned area and cover it with a lead apron. Dexcom’s reasoning is that the system has not been tested in those conditions and the fields and heat could damage components, producing inaccurate readings or missed alerts.

Abbott’s Libre sensors are no longer contraindicated for imaging. The FDA cleared the removal of the MRI, CT, and X-ray contraindication for the Libre 2 and Libre 3 families, making them the first patient-applied CGM sensors approved to stay on through those procedures. Abbott’s conditions are specific: scans between the pelvis and sternum on 3T scanners are limited to twelve minutes with a two-minute cooling period, other areas on 1.5T can run up to an hour, readings may be unreliable during the scan itself, and function returns to normal within an hour afterward. Diathermy remains contraindicated for both brands.

So two people with CGMs walking into the same MRI suite have genuinely different instructions depending on which sensor is on their arm. Tell the radiology team what you are wearing, and let them check the current labelling rather than assuming.

This is the same category of concern we cover for cardiac implants and electromagnetic interference: the risk is not ambient exposure, it is a specific high-field medical environment interacting with a specific device.

Measuring It Yourself

You can try, though set expectations first. A one-milliwatt burst lasting a couple of seconds every five minutes is a difficult target for a consumer RF meter. Most will show a brief spike if you hold the probe against the sensor at the right moment and nothing at all the rest of the time, which is an accurate picture but an unsatisfying one. Our guide to measuring EMF explains why peak-hold mode matters for intermittent sources like this.

The more useful measurement, if you own a meter, is the phone — and the room it sits in overnight. If you want a structured starting point for that rather than a spot reading, our exposure quiz walks through the sources in a typical home in order of how much they actually contribute.

Frequently Asked Questions

Does a Dexcom sensor emit radiation?

Yes, in the same sense a Bluetooth headphone does: it transmits low-power radio waves in the 2.4 GHz band. Dexcom’s G7 user guide lists the sensor’s maximum output power as 1.0 mW EIRP, and the sensor connects for only a few seconds every five minutes to send a reading. It is non-ionizing radio, not the ionizing kind produced by X-rays, and the power involved is a small fraction of what your phone transmits.

Is the FreeStyle Libre safer than Dexcom for EMF exposure?

Not in a way that should drive a decision. The Libre 3 Plus transmits at 2 dBm EIRP (about 1.6 mW) every minute; the Dexcom G7 transmits at 1.0 mW EIRP every five minutes. Dexcom’s total transmission time is lower, Abbott’s peak power is marginally higher, and both are far below any other radio you own. The meaningful difference between the two brands is medical imaging labelling, not RF output.

Can I turn off the Bluetooth on my CGM?

Not on the sensor itself, and it would defeat the device’s purpose — without the Bluetooth link there are no high and low alarms, which is the feature most responsible for the safety benefit. If you want fewer radios near you at night, move the phone off the bed rather than disabling the connection.

Can I wear my CGM during an MRI or CT scan?

It depends on the brand, and this is worth confirming rather than guessing. Dexcom G7 is MR-unsafe and must be removed before an MRI; for CT, Dexcom advises keeping the sensor out of the scanned area under a lead apron. Abbott’s Libre 2 and Libre 3 sensors have had the imaging contraindication removed by the FDA and can stay on for X-ray, CT, and MRI within specified scanner and duration limits. Tell your radiology team which device you wear and let them verify the current instructions.