Cold Plunge and Insulin Sensitivity: What the Glucose Research Shows
There is a genuinely impressive human trial behind the claim that cold improves insulin sensitivity. It used cold air for six hours a day, not cold water for three minutes, and that difference changes what you can honestly conclude from it.
The strongest evidence comes from Hanssen et al. (2015, Nature Medicine): 10 days of mild cold air exposure at 57-59°F (14-15°C) for several hours daily raised peripheral insulin sensitivity by roughly 43% in eight men with type 2 diabetes. The mechanism was GLUT4 translocation in skeletal muscle, not brown fat. That protocol is not a cold plunge. The only trial that tested repeated brief cold water immersion directly found glucose tolerance temporarily got worse. Cold plunging is not an established glucose-lowering intervention, and acutely it tends to raise blood sugar rather than lower it.
The key study, and what it actually did
Almost every claim you will read about cold and blood sugar traces back to one paper. Hanssen and colleagues at Maastricht University published it in Nature Medicine in 2015 under the title "Short-term cold acclimation improves insulin sensitivity in patients with type 2 diabetes mellitus." Eight men with type 2 diabetes spent 10 consecutive days in a climate chamber held at 57-59°F (14-15°C) for roughly six hours per day. Insulin sensitivity was measured with a hyperinsulinemic-euglycemic clamp, which is the reference method rather than a surrogate index. Peripheral insulin sensitivity increased by about 43%.
A 43% improvement in 10 days is a large effect by the standards of metabolic research. It is comparable in magnitude to what months of structured exercise training produce. That is why the finding travelled so far. What travelled less well is the dose. Six hours a day for 10 days is roughly 60 hours of cold exposure. A person plunging for three minutes, five days a week, accumulates one hour of cold in a year.
The study also had the limitations you would expect from an early proof of concept: eight participants, all men, no randomised control arm, and a short follow-up. It establishes that human glucose metabolism is cold-responsive. It does not establish a treatment.
The GLUT4 mechanism, and why it is not brown fat
GLUT4 is the insulin-responsive glucose transporter in muscle and fat cells. In a healthy state, insulin binding sets off a signaling cascade that moves GLUT4 from internal storage vesicles to the cell membrane, where it lets glucose in. In insulin resistance, that translocation step works poorly, so glucose stays in the blood despite adequate insulin.
The mechanistic result in Hanssen et al. is the interesting part. Basal GLUT4 at the skeletal muscle membrane increased markedly after cold acclimation, but insulin signaling itself did not improve and AMPK activation did not change. In other words, cold appeared to move transporters into position through a route that bypasses the broken insulin signal, rather than by repairing it. That is conceptually similar to how muscle contraction recruits GLUT4 during exercise without needing insulin.
Brown adipose tissue glucose uptake rose only minimally in the same participants. This matters because most popular coverage assumes brown fat is doing the work. Skeletal muscle accounts for a large share of body mass and handles the bulk of post-meal glucose disposal, while adult brown fat depots are small and concentrated around the collarbone and spine. Brown fat is genuinely cold-activated and worth understanding on its own terms, which we cover in the guide to cold plunging and brown fat, but it is not the tissue that explains this particular result.
Why shivering seems to be the active ingredient
The most informative follow-up was a null result. Remie, Sellers and colleagues published a 2021 trial in Nature Communications in which nine patients with type 2 diabetes underwent 10 days of mild cold acclimation at 61-63°F (16-17°C), this time with overt shivering deliberately prevented. Insulin sensitivity did not improve. Neither did postprandial glucose or lipid metabolism. The authors connected the null finding to the absence of self-reported shivering and to the absence of any upregulation in muscle activation and muscle contraction gene pathways.
The mirror image arrived in 2024. Sellers et al., writing in Nature Metabolism, exposed 15 adults with overweight or obesity, most with impaired glucose tolerance, to intermittent cold designed to produce about one hour of shivering per day for 10 days. Oral glucose tolerance improved, as did fasting glucose, triglycerides, non-esterified fatty acids and blood pressure.
This fits the mechanistic literature. Blondin and colleagues (2015) showed that skeletal muscle glucose uptake rises during shivering in cold-exposed humans, and that this uptake is insulin-independent. Later work from the same lab, including a 2025 Cell Metabolism paper, found that shivering intensity scales proportionally with falling mean skin temperature while adipose tissue non-shivering thermogenesis does not follow the same pattern. The practical reading is blunt: contracting muscle, not the sensation of being cold, looks like what moves glucose.
Most cold plunge advice treats reduced shivering as a marker of successful adaptation. If shivering is the mechanism behind the glucose effect, then adapting away from it removes the stimulus. The same tension shows up in the calorie expenditure data, where adapted plungers burn less per session precisely because they shiver less.
Water is not air, and the plunge data points the other way
Water conducts heat roughly 25 times faster than air, so 57°F water and 57°F air are not the same stimulus at all. Immersion produces a rapid, intense sympathetic response in the first minute, then a steep core temperature drop that limits how long anyone can stay in. Mild cold air produces a low-grade, sustained thermogenic load for hours. These are different interventions that happen to share a temperature reading.
Direct evidence on repeated cold water immersion and glucose control is thin, and what exists is not encouraging. Solianik, Jarutiene and Brazaitis published a 2025 study in the Journal of Thermal Biology in which participants completed 16 daily sessions of 10-minute whole-body immersion at 14°C (57°F). Glucose tolerance and insulin sensitivity temporarily decreased, and the measurements returned to baseline after a week without immersions. That is one small study and it should not be over-read in the negative direction either, but it is the closest thing we have to a test of the actual practice, and it did not reproduce the cold air findings.
A plausible explanation is dose and stress balance. Ten minutes of immersion delivers only a few minutes of meaningful shivering, far short of the hour per day that produced benefit in the 2024 trial, while delivering a much larger acute sympathetic and cortisol load. If you are choosing water temperature for other goals, the temperature guide and the frequency guide cover the trade-offs, but neither should be optimised on the assumption of a glucose benefit.
Acute versus chronic: why your CGM spikes after a plunge
People who wear continuous glucose monitors often notice their glucose climbing after a plunge and conclude something has gone wrong. It has not. Acute cold is a sympathetic stressor. Catecholamine release drives hepatic glycogenolysis, and cold also produces a rapid rise in plasma glucagon, documented in humans and rats since the early 1980s, which pushes gluconeogenesis. The liver is deliberately releasing glucose to fuel thermogenesis. A short-term rise is the system working.
Abramoff et al. (2023, Journal of the Endocrine Society) measured this carefully in eight fasted adults with type 1 diabetes across three 60-minute conditions: cool water at 73°F (23°C), thermoneutral water at 92°F (33.5°C), and thermoneutral air. Blood glucose did not change during any of the immersions. It rose during on-land recovery after the cool condition only, and the authors attributed part of that to a concurrent fall in plasma insulin.
The lesson is that acute readings tell you almost nothing about chronic adaptation, and the two can point in opposite directions. An acute glucose rise is compatible with a long-term improvement in insulin sensitivity, exactly as a hard interval session can raise glucose during the workout while improving glucose control over months. Judging a cold routine by the reading 20 minutes after the plunge is the wrong measurement at the wrong timescale.
Evidence grading
Every claim in this area deserves a confidence label. Here is how the individual propositions stand up against the published human data as of mid-2026.
Notice that nothing reaches "strong." The supportive trials involve eight to fifteen participants each, run for 10 days, and mostly lack randomised control arms. That is early-stage evidence for a real physiological phenomenon, not a settled clinical finding.
Safety notes if you have diabetes
This section is general information, not medical advice, and it does not substitute for the judgement of the clinician who manages your care. Cold immersion introduces specific interactions with diabetes management that are worth knowing before you get in.
The hypoglycemia awareness point deserves emphasis. Cold immersion produces tremor, pallor, elevated heart rate and a degree of mental fog. Those overlap closely with the early warning signs of low blood sugar, and in cold water you have every reason to attribute them to the water. Anyone using insulin should treat that overlap as the main hazard, not the cold itself. The general risk profile of immersion, independent of diabetes, is covered in the benefits and risks overview.
Practical takeaways
If glucose control is your actual goal, the honest ranking is straightforward. Resistance training and aerobic exercise have decades of randomised evidence behind them, produce GLUT4 recruitment through the same insulin-independent route, and work at doses people can sustain. Sleep and dietary pattern come next. Cold exposure is an interesting adjunct with a plausible mechanism and thin human data.
If you want to lean into the cold angle specifically, the trials point toward longer mild cold rather than shorter extreme cold. Underdressing outdoors in winter, keeping the house at 61-64°F, and tolerating a cool room for an hour resemble the protocols that worked far more closely than a three-minute plunge does. That is an unglamorous conclusion, and it is what the evidence supports.
If you plunge anyway, and there are plenty of good reasons to, judge it on the outcomes it has better evidence for rather than on your glucose monitor. Keep expectations about body composition separate too, since the weight loss evidence has its own set of caveats. If you do want to track a possible glucose effect, change one variable at a time, keep training and diet stable, and compare fasting values and post-meal curves across several weeks rather than reacting to single sessions.
If you are running a structured cold protocol and want repeatable water temperature rather than ice-and-guesswork, a chiller-based unit removes the largest variable. Read the full breakdown before buying.
