Heart rate variability (HRV): what it tells you and what it doesn't

What it measures

The heart doesn't beat like a metronome: the time between one beat and the next keeps changing by a few thousandths of a second. Heart rate variability sums up how much it changes. There are many ways to calculate it; the one consumer devices usually show, RMSSD,¹⁰ is the measure of this kind used mainly to estimate the influence of the vagus nerve, the part of the nervous system that slows the heart down at rest.¹

Each device measures it in its own way: some throughout the night, others in blocks of a few hours, and one with a short test while you are awake.¹⁰ And values from short recordings aren't interchangeable with those from a 24-hour recording, so they can't be carried over to a medical table either.¹

An important clarification: higher isn't always better. Some heart rhythm disturbances also produce high variability.¹

Why your number can't be compared with anyone else's

A review of 44 studies, with 21,438 healthy adults, found very large differences between people.³ HRV also goes down with age: in a study with 24-hour recordings, RMSSD fell to 47% of its value in the second decade of life (ages 10 to 19) by the sixth (ages 50 to 59) and then leveled off.⁵

That's why, in research, the recommendation is to compare each person with themselves and not with others.² The useful question isn't “Is my HRV good?” but “How does it compare with what is usual for me?”

What moves it from day to day

• Alcohol. In 4,098 Finnish employees monitored over several days, drinking was associated with lower HRV in the first hours of sleep, and more so the more alcohol they drank. With 0.75 g of alcohol per kilogram of body weight, the study's model estimates an average drop in RMSSD of 10.9 ms at age 30 and of 4.7 ms at age 60.⁶
• Heart rate. When the heart beats faster, there is less time between beats for it to vary, and HRV goes down.¹
• Fitness. Over time, HRV tends to rise with better aerobic fitness.¹
• Sleep. A meta-analysis of trials found a drop in RMSSD after people were deprived of sleep, but with only four trials, and the effect was small, at the edge of what can be told apart from chance.⁷
• Stress. Stress induced in the laboratory usually lowers it.⁸ In daily life the relationship is less direct: according to that review, in a study with 909 adults, the number of stressful situations was not related to HRV, but how intensely each person reacted was.⁸
• Position, meals and coffee. The measurement is very sensitive to body position, and in research, meals, coffee and alcohol beforehand are controlled for.²

How to read it: your trend, not today's number

If so many things move it, a single night mixes too many causes. In a study with ten runners, the weekly average reflected how they were adapting to training better than the value of a single day.⁹ It is a very small sample, and of athletes, but the idea fits how researchers work: comparing each person with themselves over time.²

• Look at your average over several weeks, not each morning's number.
• Only compare data from the same device: each model measures differently.
• Before drawing conclusions, think of the obvious: alcohol, a bad night, a hard workout or not feeling well.

If a change holds and you want to know whether a habit has anything to do with it, it helps to frame it as a question and a simple test; we explain this in how to turn a wellness data point into a question and a test.

How accurate watches and rings are

Consumer devices calculate HRV with a light sensor, not with an electrocardiogram. In a sleep lab with 53 young adults, all the models tested came in below the electrocardiogram on average; the one that was furthest off, which measured with a short test while awake, came in 22.4 ms below, and in several of them low values tended to come out higher and high values lower.¹⁰ In another study, with 13 adults and 536 nights, the mean error was 6% for the most accurate model and 16% for the least accurate.¹¹

What your watch's HRV doesn't tell you

In population studies, lower HRV has been associated with higher cardiovascular risk and higher mortality.¹²¹³ These are associations in large groups, with other HRV measures and other recordings than those of a watch, and they don't make it possible to calculate one person's risk from the number on their watch.

The US Food and Drug Administration (FDA) may consider a sensor that displays HRV a wellness product if, among other conditions, it isn't intended to diagnose or treat and doesn't include diagnostic thresholds or alerts that call for medical action; it can display values, trends and personal baselines related to sleep, activity or recovery.¹⁴

If you notice palpitations, dizziness, shortness of breath or any symptom that worries you, see a health professional, whatever number your watch shows. A high HRV doesn't rule anything out either.