Do Dexcom and Libre Stay Accurate at Altitude?
The chemistry behind a continuous glucose monitor depends on oxygen, which raises a real question for anyone using one somewhere the air is thinner.
Evidence rating for the proposition “CGM sensors remain accurate under low oxygen conditions”: Uncertain
Accuracy drops, and by different amounts per sensor. In a randomized trial of 30 people breathing 14.5% oxygen, mean absolute relative difference against venous plasma glucose was 21.2% for FreeStyle Libre and 41.8% for Dexcom G6 in healthy volunteers, and 11.8% and 17.5% in participants with diabetes. In the same 30 people, on consensus error grid analysis, 97% of FreeStyle and 87% of Dexcom readings still fell in zones A or B. The authors concluded both systems were less accurate under hypoxia, most of all during rapid glucose changes.
A sensor built on a gas most people never think about
A continuous glucose monitor gets treated like any other wearable sensor, something that just reads a number off the body without much thought to how. What that framing skips over is that the two most widely used CGM systems work through an enzymatic reaction that needs oxygen to function, the same glucose-oxidase chemistry used in a lot of glucose testing technology generally.
That detail matters more in some settings than others. Altitude, air travel, and certain medical conditions all reduce how much oxygen is actually available in the body. And until recently, nobody had run a controlled test of what that does to a CGM's accuracy.
1 study
- The two glucose-oxidase-based CGM systems most commonly used, FreeStyle Libre and Dexcom G6, are described as well established for accuracy under normal oxygen conditions, but the study notes that little evidence existed on their performance during hypoxia, the kind encountered at altitude, during air travel, or in certain medical conditions. A randomized controlled trial was designed specifically to close that gap, exposing 30 participants, 15 healthy volunteers and 15 with diabetes, to standardized normobaric hypoxia while wearing both sensor types simultaneously, with venous plasma glucose as the reference measurement and induced glycemic swings from a standardized meal and moderate cycling exercise.
What the trial was built to answer
The design itself tells you what the open question actually was: not whether these sensors work in everyday, sea-level conditions, which the study describes as well established, but whether the same accuracy holds once oxygen availability drops in a controlled, standardized way. Using normobaric hypoxia rather than a real ascent let researchers isolate the oxygen variable specifically, rather than mixing in cold, exertion, or altitude sickness as confounds.
Running both healthy volunteers and participants with diabetes through the same testing approach, and comparing both sensor brands against a lab-grade venous blood reference rather than each other, was built to produce a genuine answer to whether hypoxia degrades CGM accuracy and by how much. The mean absolute relative difference between sensor and reference glucose served as the primary measure.
What the trial does not settle
Thirty participants, one hypoxia level, one session each. The trial used normobaric hypoxia at 14.5% inspired oxygen, which is a controlled stand-in for altitude rather than altitude itself, and it does not follow anyone across days of acclimatisation or up a mountain.
It also does not report a matched sea-level comparison for the same participants, so the size of the drop relative to normal use is not established here, only that the authors judged accuracy reduced. Whether the same holds for newer sensors, or for the closed-loop insulin systems that read from them, was explicitly named as needing separate work. For the wider question of what a CGM shows outside diabetes, see what a CGM shows if you do not have diabetes.
This piece is built from a summary of the FOX study's design, population, and methodology. It does not include the trial's reported accuracy results, mean absolute relative difference figures, or error-grid findings, so no claim about whether CGM accuracy actually changes under hypoxia can be made from the material available here.
Common questions
Why would altitude affect a CGM's accuracy?
The two most common CGM systems, FreeStyle Libre and Dexcom G6, use glucose-oxidase chemistry that depends on oxygen availability, which is well established as how these sensors work, and altitude reduces how much oxygen is available in the body.
Has CGM accuracy actually been tested under low-oxygen conditions?
A randomized controlled trial was specifically designed to test this, exposing 30 participants, both healthy volunteers and people with diabetes, to standardized hypoxia while wearing both sensor types and comparing readings against a lab-grade blood glucose reference.
Does this mean CGMs are inaccurate at altitude?
This piece cannot answer that. It describes the trial's design and the mechanism behind the concern, but the trial's actual reported accuracy results are not part of the evidence available here.
Does this apply to air travel as well as mountain altitude?
The trial used standardized normobaric hypoxia to simulate reduced oxygen availability broadly, which the source material frames as relevant to altitude exposure, air travel, and certain medical conditions collectively, rather than testing each setting separately.
Sources
- Accuracy of Dexcom G6® and FreeStyle Libre® Sensors in Standardized Hypoxia Conditions: A Randomized Controlled Trial-FOX Study. [conflict statement: two authors have received honoraria or other payments from Abbott]