Body temperature is one of the oldest signals in medicine and one of the newest signals on your wrist. Almost every recent smartwatch — Withings, Apple, Oura, Fitbit, Garmin — now measures skin temperature overnight. Very few people know what to do with the number, and the marketing does not help. This is a walk-through of what a continuous temperature reading can honestly tell you about your sleep, your cycle, and your immune system, and where it is still mostly a rounding error dressed up as an insight.
Distal skin, core, and why your wrist is not a thermometer
The temperature reading you care about clinically — the one your doctor takes when you feel unwell — is core body temperature. That is the temperature of your internal organs and deep blood, held tightly around 36.5 to 37.5 degrees Celsius (roughly 97.7 to 99.5 degrees Fahrenheit) by the hypothalamus. Core temperature is what fever is measured against. It is measured accurately only in the mouth under the tongue with a proper thermometer, in the ear canal with an infrared probe, or invasively via rectal or esophageal probes.
A smartwatch does not measure core temperature. It measures distal skin temperature — the temperature at the back of your wrist, at the fingertips, at the surface of your body. Distal skin temperature drifts wildly across the day. It is pushed around by the air temperature in the room, whether your sleeve is covering the sensor, the blanket you kicked off at three a.m., a hot shower an hour ago, and blood flow to the skin driven by everything from stress to alcohol to being cold.
This is why the absolute number on your wrist is essentially useless. If your ScanWatch says your skin was 32.4 degrees Celsius last night, that is not a fever, that is not hypothermia, that is just a data point that means almost nothing on its own.
What is meaningful is the deviation from your own personal baseline. Withings, Apple, Fitbit, and Oura all take the same approach: for the first few weeks of wear they build a baseline of your typical overnight temperature curve, and after that they report each night as a delta — plus or minus a fraction of a degree against your personal norm. The absolute value is thrown away. The trend is kept.
The overnight temperature curve you did not know you had
Even if you have never seen a hypnogram, your body follows a stereotyped temperature curve every 24 hours. Core temperature peaks in the late afternoon or early evening, roughly 4pm to 7pm, and hits its lowest point — the circadian nadir — two to four hours before your natural wake time. For most adults that is roughly 3am to 5am. The drop from evening peak to overnight trough is around 0.5 to 1.0 degrees Celsius, or roughly 1 to 2 degrees Fahrenheit.
The sleep-related drop is not incidental. Falling asleep requires that core temperature drop. Your body accomplishes this by dilating blood vessels in the hands and feet, dumping heat to the skin surface, which is why cold feet keep people awake and why a warm bath 60 to 90 minutes before bed helps you fall asleep — the warm bath triggers exactly the vasodilation your brain uses as its "time to sleep" signal, and by the time you get into bed you are cooling.
What your wearable actually sees at the wrist is the inverse of core temperature during the sleep-onset window. As core drops, blood flow to the skin rises, and distal skin temperature rises. That warm wrist late at night is a working sleep-onset system. Studies have shown that people who fall asleep quickly have larger and earlier distal skin warming than people with insomnia. The wearable is watching the mechanism, not the temperature itself.
Sleep quality clues from nightly temperature deviation
Once you have a stable personal baseline, three deviations tend to matter.
- Elevated overnight skin temperature. A sustained rise across the night, especially in the first half, is a common signature of alcohol. Ethanol dilates peripheral blood vessels and interferes with the normal cooling curve. Two drinks with dinner can move overnight wrist temperature up by 0.3 to 0.6 degrees Fahrenheit against your baseline, and it will usually track with a matching drop in deep sleep and HRV.
- Flattened temperature curve. When the difference between your evening peak and overnight nadir is small, sleep quality is often poor. Aging, sleep apnea, chronic stress, and heavy late-night eating can all blunt the curve. This one is subtle and only shows up on a monthly view.
- Sudden departure from the trend. This is the one that matters most day-to-day, and we will return to it below in the illness section.
Cycle tracking: the honest story
Body temperature and the menstrual cycle have a well-documented relationship that predates smartwatches by about a century. During the follicular phase — from the first day of your period through ovulation — basal body temperature runs a few tenths of a degree lower. Immediately after ovulation, progesterone rises, and progesterone is a mild thermogenic. Core temperature climbs by roughly 0.2 to 0.5 degrees Celsius, holds through the luteal phase, and drops again at the start of the next period.
This is the biphasic pattern that the traditional fertility awareness method relies on, using an oral thermometer every morning before getting out of bed. The wearable version is much more comfortable and much noisier.
Peer-reviewed validation work — including work from research groups in Kyoto and elsewhere published around 2019 to 2021 — has shown that continuous wrist-based skin temperature can reproduce the biphasic pattern in most cycles, and can retrospectively identify ovulation by looking for the sustained temperature shift that follows it. The word doing the work in that sentence is retrospectively.
By the time your wearable is confident an ovulatory temperature shift has occurred, the ovulation itself is already two to three days in the past. That is fine for someone tracking cycles who wants to understand the general shape of their month. It is not sufficient for time-sensitive fertility decisions. For those, a luteinizing hormone (LH) urine strip caught the day before ovulation still beats wrist temperature by a mile.
The wearable temperature signal earns its place as a longitudinal record. Six months of nightly deviations plotted against your cycle days will show you whether your luteal phase is a consistent length, whether the temperature rise is a clean 0.4 degrees Fahrenheit or an unclear half-signal, and whether anovulatory cycles (no temperature shift) are showing up. That is genuinely useful clinical information you can bring to a physician, and it does not require a $500 device.
To be blunt about limits: none of the wearables in our catalog, or in the broader consumer market, are cleared as contraception. The FDA-cleared exception is the Natural Cycles app paired with a traditional oral basal thermometer, and even that has failure rates several times higher than a copper IUD or hormonal methods. Do not treat wrist temperature as birth control.
Illness forewarning: the most interesting signal
The application that got the most academic attention is fever forewarning. In 2020 and 2021, during the acute phase of the pandemic, several groups looked at whether wrist-worn wearables could detect an incoming illness before the person felt sick.
The most cited work came from research collaborators using Oura ring data and separately using Fitbit data. In both cases, roughly 60 to 70 percent of confirmed COVID-19 cases showed a detectable deviation in resting heart rate, respiratory rate, HRV, and skin temperature 24 to 72 hours before the participant reported symptoms. Skin temperature by itself was not the star of the show, but temperature deviation was a meaningful contributor to the composite signal.
A few honest qualifiers on that finding:
- Detection rates were much higher for symptomatic infections than asymptomatic ones, which is not surprising and also not that useful.
- False positives are real. Alcohol, hard exercise the previous evening, hot rooms, and hormonal cycle phase can all move the same metrics, and the wearable cannot distinguish a coming cold from a Thursday-night pinot noir.
- The specific algorithm and the confidence interval matter. Off-the-shelf apps that flag "you might be getting sick" are guessing.
The pragmatic use of this signal is: if two or three metrics move in the same wrong direction on the same morning — HRV down, resting heart rate up, skin temperature elevated — it is a legitimate cue to reduce training load, sleep more, hydrate more, and pay closer attention. It is not a diagnosis. It is a hint from your own physiology.
Which wearables in our catalog actually give you temperature
Within the Vyvata catalog, wrist-based skin temperature is available on the Withings ScanWatch 2 (both 38mm and 42mm cases), the Fitbit Sense, and the Fitbit Sense 2. Skin temperature reporting is a headline sensor on all three. Withings publishes more of its validation methodology than most competitors — one of the reasons the ScanWatch 2 lands at 85 Verified in our system.
What the specs pages will not tell you is that all four devices give you the same type of number: a nightly deviation against your own baseline, once the baseline stabilizes over roughly two to four weeks of continuous wear. If you take the watch off for a week, the baseline decays. Wear consistency is more important than which brand you chose.
How to actually use nightly temperature data
Here is a simple framework that keeps you out of the weeds.
- Wear the device continuously for four weeks before you look at any temperature graph. Anything you see before the baseline stabilizes is noise around zero.
- Look at the temperature panel weekly, not nightly. Single-night deviations of 0.2 to 0.4 degrees Fahrenheit are within normal variation and mean nothing on their own.
- Correlate with three things: cycle day (if applicable), alcohol, and training load. You will see that most of your "unexplained" deviations line up with one of those.
- Treat sudden two-signal deviations as a hint, not a diagnosis. Elevated skin temperature + suppressed HRV two mornings in a row before a scheduled hard workout is a reasonable case for switching to easy mobility work that day.
- Do not treat wrist temperature as fever monitoring for anyone who is actually sick. Use a real thermometer.
The honest bottom line
Continuous body temperature is one of the more underrated signals a consumer wearable can give you. It is also one of the most misunderstood. The absolute number is not a temperature reading in any medical sense — it is a proxy for skin blood flow at the wrist, useful mainly as a delta against your own baseline. Its clearest applications are cycle tracking (retrospectively), sleep quality context, and illness forewarning as one input to a composite signal.
Used the wrong way, it turns a healthy adult into an anxious data monitor watching a thin line wobble around zero. Used the right way, it is a quiet second opinion on how your body is actually doing in the background — a signal that has been valuable in medicine for a hundred and fifty years, finally made continuous.
Wear the watch, ignore the first month, and start looking at the monthly view.