How fast should your heart rate drop after exercise?

Short answer

Your one-minute heart rate recovery number is meaningless until you know how you recovered. The famous cutoff — a fall of 12 beats or fewer — comes from a study where people kept walking after the treadmill stopped. When people lay down instead, the equivalent cutoff was 18 beats. Same heart, different number, because posture changes how fast your pulse falls.

That single fact explains most of the confusion around this metric. Heart rate recovery, or HRR, is how far your pulse falls in the minute after you stop exercising, and it is genuinely one of the most useful numbers your body will give you for free. It reflects how quickly your parasympathetic nervous system — the brake — comes back online after effort. Faster is better, it improves with training, and in large studies a slow recovery predicts mortality independently of how fit you are, how fast your heart beats at rest, and how much exercise you managed on the test.

But it is also one of the most misreported numbers in consumer health writing. You will find tables online claiming that 15 to 20 beats is normal, 25 to 30 marks an athlete, and anything under 12 is a warning sign. We went looking for the primary research behind those tiers and could not find any. There is no age-and-sex normative table for heart rate recovery the way there is for VO₂max. What exists instead is a set of study-specific cutoffs, each tied to the exact protocol that produced it — and that turns out to be more useful than a fake norm, once you know how to use it.

It's not just you, and you weren't imagining the contradiction. If you compared your recovery number to a chart online and came away confused or frightened, the chart was the problem — not your reading comprehension and not your heart. Most published thresholds never state which recovery posture they assume, and that single omission can move the number by more than the diagnostic cutoff itself. Being unable to make sense of your own data here is not a failure of understanding. It is what happens when a real physiological measurement gets repackaged as a scary table with no protocol attached.

A note on the data: Welltory measures heart rate continuously, which means it can capture how fast your pulse falls after effort and follow that trend over weeks. The trend is the part worth watching. A recovery that is speeding up is one of the earliest visible signs that your fitness is genuinely improving — it often moves before your resting heart rate does and long before anything shows up on a scale.

Why does the "normal" number change depending on where you read it?

Because the number depends on what you do in the minute after you stop, and different studies did different things.

The landmark paper is Cole and colleagues in the New England Journal of Medicine in 1999. They followed 2,428 adults, average age 57, for six years, during which 213 died. They defined abnormal recovery as a fall of 12 beats per minute or fewer by the end of the first minute after peak exercise. Of the group, 639 people — 26% — fell into that category. Looking at the number on its own, those people were about four times more likely to die during follow-up. After adjusting for age, sex, medications, perfusion defects on thallium imaging, standard risk factors, resting heart rate, the rise in heart rate during exercise, and the workload achieved, the risk was still twice as high.

Both of those numbers are true and they say different things. Fourfold is what the raw association looks like. Twofold is what remains once you account for everything else about the person — and that is the honest figure to hold onto, because it is the part that heart rate recovery contributes on its own.

Two years later, Watanabe and colleagues published in Circulation on 5,438 patients who had exercise echocardiography. Those patients did not walk it off after the treadmill; they had to lie down immediately so their hearts could be imaged. In that setting the prognostic cutoff was not 12 beats but 18. Nothing about the patients changed. The posture changed.

The physiology behind this is simple. Lying down, blood returns to the chest more easily, the pressure receptors in your arteries respond, and vagal tone comes back faster. Standing still is slower. Sitting upright is the slowest of the common positions. So a person who sat on a bench and recorded a 14-beat drop and a person who lay flat and recorded a 14-beat drop are not in remotely the same situation — the first is on the safer side of the walking cutoff, the second is well below the lying-down one.

What is actually happening in that first minute?

The fall in heart rate after exercise is not one smooth curve. It has two distinct phases, and they are driven by different systems.

Imai and colleagues showed this in the Journal of the American College of Cardiology in 1994. Plotted on a log scale, the decline is steep for roughly the first 30 seconds and then flattens out. When they gave participants atropine, which blocks parasympathetic signalling, the steep early phase disappeared entirely. That is about as clean a demonstration as physiology offers: the first half-minute of recovery is parasympathetic reactivation, sometimes called vagal rebound. What comes after is a slower mix of sympathetic withdrawal, changes at the sinus node, and central regulation.

In the same work, patients with heart failure had no steep phase to begin with. Their recovery curve was flat from the start.

This is why the one-minute window carries so much information despite being so crude. You are not measuring your fitness in some general sense. You are measuring how quickly a specific nerve pathway reasserts control over your heart — and that pathway is sensitive to conditioning, to illness, to sleep debt, and to stress.

It is also why heart rate recovery and heart rate variability tell overlapping stories. Both are windows onto vagal function. If your HRV has been drifting down and your recovery has been slowing at the same time, those are not two findings. They are one finding, seen twice.

Does a slow recovery matter if I am healthy?

This is the fair objection, and it has been tested directly. Cole's original study was done in people referred to a cardiology centre for exercise testing — not the general population.

So look at the follow-up work. In 2000, the same group published in the Annals of Internal Medicine on 5,234 adults with no evidence of cardiovascular disease, followed for 12 years, during which 312 died. Here the measure was the two-minute recovery, with a cutoff of 42 beats or fewer, and the relative risk of death was 2.58 with a confidence interval of 2.06 to 3.20. It held after adjusting for risk factors, for fitness, and for resting and peak heart rate.

Jouven and colleagues in the New England Journal of Medicine in 2005 followed 5,713 asymptomatic working men aged 42 to 53 for an average of 23 years — 1,516 died, 400 of cardiac causes, 81 suddenly. A fall of fewer than 25 beats after exercise was one of three heart-rate predictors of sudden death, alongside a resting rate at or above 75 and a rise during exercise of 89 beats or fewer.

And the cleanest summary comes from a meta-analysis of nine prospective general-population cohorts by Qiu and colleagues in the Journal of the American Heart Association in 2017. Instead of a cutoff, they treated recovery as a continuous scale. For every 10 beats per minute of slower recovery, the hazard ratio was 1.13 for cardiovascular events and 1.09 for death from any cause. Comparing slow to fast recovery outright gave hazard ratios of 1.69 and 1.68.

Notice how modest those per-10-beat figures are. That is the correct scale of the thing. A slow recovery is a real signal in a population of thousands. It is not a verdict on one person on one Tuesday.

There is one more finding worth knowing, from a 2018 Mayo Clinic analysis in the same journal of 19,551 people without cardiovascular disease. Recovery predicted mortality consistently across sex, age, obesity, hypertension and diabetes — but it worked less well in people taking beta-blockers, in smokers, and in people who already had normal cardiorespiratory fitness. If you are fit, this number tells you less than it tells someone who is not.

What is a good heart rate recovery, honestly?

Here is the part where most articles give you a table. We are going to give you the study cutoffs instead, because the table does not exist.

Every widely repeated consumer figure — "healthy people drop 15 to 20," "athletes drop 25 to 30," "under 12 is concerning regardless" — traces back to sites that do not cite a primary source. The 12-beat cutoff itself was not a physiological constant. It was the bottom quartile of Cole's specific cohort, chosen because that is where the risk separated in that group of people.

So the practical answer to "what is a good number" is: a number that is bigger than your own number was three months ago, measured the same way. That is not a dodge. Given that the norms do not exist and the protocol shifts the value by six beats or more, your own trend is the only comparison that is actually valid.

How do I measure mine so the number means something?

The measurement takes one minute. Making it comparable takes discipline about four things.

1. Finish a genuinely hard effort. Recovery from a moderate effort is not comparable to recovery from a maximal one. Get your heart rate genuinely high — the end of a hard run, ride, or interval.

2. Note your heart rate the instant you stop. That is your peak value.

3. Choose one recovery posture and never change it. Walking slowly, standing still, sitting, or lying down — any of them works as a personal baseline. Mixing them destroys the comparison. If you want your number to be comparable to the 12-beat literature, keep walking.

4. Check again at exactly 60 seconds and subtract. Finish at 170, sit at 150 a minute later, and your one-minute recovery is 20.

One methodological point makes a real difference. A 2019 systematic review of reproducibility found that the reliability of recovery indices ranges enormously: intraclass correlation from 0.12 for the T30 index up to 0.99 for the one-minute value, with coefficients of variation from 3.4% to 90.1%. The one-minute figure is the only index reliable enough to act on from a single measurement. The others need averaging across several sessions before they mean anything.

What can throw the number off?

More than you would expect, and by more than the diagnostic cutoff.

Heat is the biggest one. In a study of passive heating, one-minute recovery fell from 44 ± 12 beats to 27 ± 9 — a difference of roughly 17 beats, which is larger than the entire gap between the walking and lying-down cutoffs. If you train outdoors in summer, your recovery will look worse than it did in March, and nothing about your heart has changed.

Posture, as covered above. Lying fastest, sitting upright slowest. Active recovery in a seated position is slower than passive recovery in the same position.

Caffeine slows parasympathetic recovery after both aerobic and resistance exercise. One study after anaerobic exercise found the opposite, so the picture is not perfectly clean, but the direction in most work is toward slower recovery.

Beta-blockers are genuinely unsettled, and it is worth saying so plainly. One analysis found the prognostic value of recovery unaffected by them. The Mayo work found it reduced. A third found the recovery itself blunted. If you take a beta-blocker, treat your recovery number as a personal trend and nothing more.

Blood glucose shows a pattern most people would not guess: the proportion of people with abnormal recovery rose steadily with fasting glucose while still inside the normal range — from the 70s to the 80s to the 90s in milligrams per decilitre.

Time of day appears not to matter much for the one-minute figure specifically, though other recovery indices do shift.

Hydration is often listed as a factor. The study most cited on it found the between-group differences were not statistically significant and the effect sizes small, and the authors themselves asked for caution. We are not going to tell you it matters when the data say it probably does not.

Can I trust the number on my watch?

Partly, and it is worth understanding exactly where the limit is.

We could not find a single validation study of heart rate recovery measured by a wrist-worn optical sensor. Not one. And there is a specific reason to expect problems: wrist photoplethysmography loses accuracy precisely during rapid transitions in heart rate — which is exactly what the first minute of recovery is. In a validation against ECG, a Fitbit Charge 2 showed a mean difference of −3.47 beats per minute with a mean absolute error of 5.96% in general use.

An error of three to six beats sits inside the range that separates the study cutoffs from each other. So the honest position is this: use your watch to watch your own trend, and do not use it to place yourself relative to a published threshold. If your watch says 11 one week and 14 the next, that is noise. If it says 11 in March and 19 in June, measured the same way, that is signal.

Does it improve if I train?

Yes, and this is the reassuring part of the whole topic.

A systematic review of randomised trials — eight studies, 449 patients with established heart disease — concluded there is level 1A evidence that aerobic training increases heart rate recovery. That is the strongest grade of evidence available. The review was narrative, so it did not produce a pooled figure in beats per minute, and we are not going to invent one: no reliable "expect plus X beats in Y weeks" number exists for healthy people.

What we do know is the timing. In a study of ten untrained men doing eight weeks of cycling, the fast-decay constant shortened significantly by week four and stayed improved at week eight. After stopping, it lengthened significantly within two weeks and was back to baseline by four.

That is a useful pair of numbers to carry. Roughly a month to see it move. Roughly two to four weeks to lose it. It is one of the more responsive things your body will show you, which also means a single bad reading after a stressful fortnight is not a finding.

How does this fit with my other numbers?

Recovery is most informative read alongside the rest of your pulse data, because the four numbers describe different parts of the same system. Your resting heart rate is the floor. Your maximum is the ceiling, largely set by age. Your heart rate variability reflects day-to-day recovery and load. And your heart rate recovery shows how fast you get back from effort.

When these move together — resting rate drifting down, variability trending up, recovery quickening — you have coherent evidence that fitness is genuinely improving, and you can trust it more than any one metric alone. When they diverge, that is informative too: a recovery that suddenly slows while your resting heart rate ticks up is a recognisable early pattern of under-recovery, illness, or accumulated stress.

If you are building a picture from scratch, the floor is covered in what's a normal resting heart rate — worth reading first, because the same problem of missing context applies there and the reference range turns out to be far wider than most charts admit. For the day-to-day layer, why HRV fluctuates from day to day explains how much variation is normal before you should read anything into a single reading, and how to improve HRV covers what actually moves that number and what does not. Sleep is the other half of the picture, since a recovery that has slowed for a fortnight often traces back to short nights rather than to training — HRV during sleep covers what your overnight numbers can and cannot tell you.

When is this not a self-help project?

Some heart-rate responses to exercise need a doctor rather than a training plan. These are the patterns that appear in the research as independent predictors of risk, which means they deserve a real conversation rather than another month of tracking.

• Your heart rate barely rises during exercise. Failing to reach about 85% of your age-predicted maximum is called chronotropic incompetence and is an independent predictor of mortality. Not being able to raise your rate by more than 89 beats was one of the sudden-death predictors in the Paris cohort.

• Extra beats appear during the recovery phase. In a study of more than 29,000 patients, frequent ventricular ectopy during recovery strongly predicted death — while the same ectopy during exercise did not. The recovery phase is where it counts.

• Your recovery is very slow and it is not explained by heat, caffeine, illness, or a change of posture. One reading means little. A consistent pattern across weeks, measured the same way, is worth showing someone.

• Any symptom during or after effort — chest pain or pressure, disproportionate breathlessness, feeling faint, an irregular or racing pulse that does not settle. Symptoms outrank numbers, always.

How to bring this up with your doctor — and what to ask for

Bringing a wearable number to an appointment can go badly if it arrives as a number without context. It goes much better when you arrive with a pattern and a specific request.

Bring the trend, not the reading. Three months of recovery values measured the same way after comparable efforts is a clinical observation. Yesterday's number is not. Say how you measured it — which posture, how hard the effort — because as this article has laboured to explain, that is what makes the value interpretable.

Ask specifically for an exercise stress test if the pattern warrants it, and ask what your recovery was. Heart rate recovery is routinely captured during standard exercise testing but is not always reported back. You can ask for the number and for the recovery protocol used, which lets you compare it to the right cutoff.

Say the words "my heart rate barely rises when I exercise" if that is what is happening. This is a different concern from slow recovery, it has a name — chronotropic incompetence — and naming it changes what gets checked.

Ask whether any of your medications affect this. Beta-blockers are the obvious case and the evidence on them is genuinely mixed, so it is a reasonable question rather than a challenge. Rate-limiting calcium channel blockers and some other drugs are relevant too.

Mention symptoms out loud even if they seem minor, and mention them separately from your data. Breathlessness that is out of proportion to the effort, chest discomfort, light-headedness or palpitations after exercise change the urgency of everything else in the conversation.

A worked example

Say you are 46, you have started running again after several years off, and your watch reports a one-minute recovery of 13 beats. You have read online that under 12 is concerning and 18-plus is good, so 13 feels alarming.

Work through what you actually know. First, what posture — you sat down on a wall to catch your breath, which is the slowest of the common positions, so the 12-beat walking cutoff does not apply to your measurement at all. Second, how hard was the effort — it was your first hard run in years, so your peak was probably closer to your true maximum than it will be in two months, which tends to make recovery look worse. Third, what was the weather — it was late July and 29 degrees, and heat alone has been shown to cost around 17 beats. Fourth, the device — a wrist sensor during rapid heart-rate change, with a plausible error of several beats.

None of that tells you your recovery is fine. It tells you that this particular number cannot bear the weight you were putting on it. What it can do is become your baseline. Measure the same way, same posture, at a similar effort, once a week. If you are at 13 now and 19 in October, you have learned something real about your fitness. If you are still at 13 in October and you have been training consistently, that is a pattern worth a conversation — and now you have something concrete to bring.

Caveats worth keeping in mind

Heart rate recovery is a good signal, not an oracle, and a few limits are worth stating.

The cutoffs come from specific populations on specific protocols and do not transfer cleanly to your Tuesday evening. The prognostic studies measured groups, so the risk figures describe populations rather than predicting individuals. The metric tells you less if you are already fit. Beta-blockers and other rate-limiting drugs complicate it. No age-and-sex norms exist. Wrist devices are unvalidated for this specific measurement. And a single reading is heavily influenced by heat, posture, caffeine, sleep, and how hard you actually went.

What survives all of that is still worth having: measured consistently, your own recovery trend is one of the earliest and clearest signs that training is working.

How Welltory helps

Welltory measures your heart rate continuously, so the fall after exercise is captured without you having to stand still watching a clock. That matters most for the thing this article keeps coming back to — the trend. One measurement is noise. A line over three months, from efforts of similar intensity, is the part that tells you whether your fitness is genuinely moving.

Because the app also tracks resting heart rate, heart rate variability, and sleep, you can see recovery in context rather than in isolation. A recovery that slows in the same week your variability drops and your resting rate climbs is a coherent picture of accumulated load. A recovery that quickens while the others improve is a coherent picture of adaptation. And when recovery looks worse for a week, you can usually see why in the same view — a hot spell, three short nights, or a training block that got ahead of you.

Discounts for blog readers: up to 36% off

See what affects your energy, stress, sleep, and daily state with Welltory

Explore your data