Resting Heart Rate

A slow pulse at rest goes with living longer, and it costs nothing to measure — but the trials that tried to lower it on purpose mostly failed, so it is a number worth watching and a poor thing to chase.

Resting heart rate is the cheapest measurement in this entire site. It needs no lab and no appointment — though a wearable that averages it over weeks turns out to carry most of the signal. Across cohorts totalling well over a million people it predicts who dies, and in the one study that set the two side by side it held its own against blood pressure. It is also the clearest worked example on this site of a marker that is not a lever: three large randomised trials have now lowered it deliberately with a drug that does nothing else, and only one of them improved anything.

What the number predicts — Strong

The association with dying is one of the most consistently replicated in cardiovascular epidemiology, and it is close to linear: across the ordinary range, every step up carries a little more risk.

The largest synthesis pooled 87 prospective studies and found all-cause mortality about 17% higher for each 10 beats per minute of resting heart rate, drawing on 48 cohorts for that outcome alone^[1]. That figure should be read as a direction rather than a precise quantity — the studies behind it disagree with each other enormously, which the authors quantify as a heterogeneity statistic of 94%, close to the maximum. A more conservative meta-analysis covering more people — 46 studies and 1,246,203 participants — puts it at 9% per 10 beats, and detected both substantial heterogeneity and evidence of publication bias, meaning small studies with null results are probably missing from the literature^[2].

In that second analysis the risk is concentrated at the top of the range rather than spread evenly. Above 80 beats per minute, all-cause mortality ran 45% higher than in the lowest category (relative risk 1.45, 95% confidence interval 1.34–1.57). That bracketed range is where the true effect most plausibly lies; because it sits entirely above 1.0, the increase is very unlikely to be chance, whereas a range spanning 1.0 would be compatible with no effect at all. The 60–80 band carried a much smaller excess — 12% higher (relative risk 1.12, 1.07–1.17) — and cardiovascular mortality only became clearly elevated around 90 beats per minute.

Two features of the association are more informative than the headline number.

It predicts deaths that have nothing to do with the heart. In the Melbourne Collaborative Cohort Study — 41,386 adults, 9,846 deaths — each 10 beats per minute carried an 11% higher cardiovascular mortality, a 10% higher cancer mortality, and a 20% higher mortality from all remaining causes, the largest of the three^[3]. Breast, colorectal, kidney and lung cancer deaths were each individually elevated. A marker that predicts kidney cancer about as well as it predicts heart disease is not describing a cardiac mechanism. Part of what it is picking up is almost certainly illness that has not been diagnosed yet — an inflammatory or malignant process raises resting heart rate long before it declares itself. That reverse arrow is a permanent caveat on every figure in this section, and the Melbourne analysis makes it worse by noting it did not adjust for atrial fibrillation or other irregular-rhythm disorders.

Change matters, not just level. In the same cohort, people whose resting heart rate rose by more than 15 beats per minute over roughly a decade had higher mortality than those whose rate held steady. A single reading is a snapshot; a trend is closer to a signal.

How it compares to blood pressure — Moderate (a single pooled analysis of two observational cohorts)

A 2025 analysis pooled the Taiwan MJ cohort and Norway's HUNT cohort — 692,217 adults followed for about 25 years — and set a raised resting heart rate (80–99 beats per minute) directly against hypertension (≥140/90 mmHg)^[4]. Three findings are worth carrying:

People with normal blood pressure but a high heart rate died at a higher rate than people with hypertension but a normal heart rate.
Under age 40, hypertension was not a statistically significant predictor of death at all. A raised heart rate was significant at every age.
The prevalence of a high resting heart rate barely moved across the lifespan — 21% in the twenties, 25% past 70 — while hypertension climbed from 5% to 57%. It is a signal that shows up early, in a group where the standard risk factor has not yet appeared.

The authors also estimated life expectancy: about 10.3 years lower (8.1–12.5) for normal blood pressure with a high heart rate, against 5.5 years (3.6–7.6) for hypertension with a normal heart rate. Read that as a statement about how much predictive weight the number carries, not as a bill the reader will be charged. These are observational cohorts; nothing here shows that lowering the number buys the years back — and the next section is about exactly why it probably doesn't.

Is it just a proxy for being unfit? — Moderate

The obvious objection is that fit people have slow hearts, so resting heart rate might be measuring fitness in disguise — and cardiorespiratory fitness is the most replicated mortality predictor in the field.

Only two cohorts have properly tested this, and both say no.

The Copenhagen Male Study is the cleaner design. It estimated maximal oxygen uptake from a stationary-bike test in 1970–71 and measured resting heart rate by electrocardiogram fifteen years later, then followed 2,798 healthy middle-aged men for 16 years — of whom 1,082, close to four in ten, died. After adjusting for that fitness measure, leisure-time physical activity and conventional cardiovascular risk factors, mortality still rose 16% per 10 beats per minute (10–22%), and men above 90 beats per minute had roughly triple the mortality of men at or below 50^[5]. That threefold figure compares two small groups at opposite extremes of the distribution, so the per-10-beat number is the more useful one. The point estimate was higher in smokers (20% per 10 beats) than non-smokers (14%), but the two overlap heavily and the difference was not statistically clear (p=0.07).

The second test is smaller but asks a sharper question. The Belgian Physical Fitness Study put a 24-hour ambulatory heart rate and a seated clinic reading into the same model, adjusted for fitness, occupational and leisure activity, and the usual risk factors, and followed 439 men for a mean of 16.5 years^[6]. The 24-hour average predicted death strongly — roughly three times the risk in the highest third versus the lowest (hazard ratio 3.21, 1.22–8.44). The clinic reading, once the ambulatory average was accounted for, predicted nothing at all: a hazard ratio of 1.26 with a range running from 0.48 to 3.31, which comfortably includes no effect. With 439 men and an interval that wide, this is one suggestive study rather than a settled result, but it points somewhere useful: what predicts is the rate your heart actually runs at, not the number a nurse writes down.

Fitness and resting heart rate are related, of course. In Norway's HUNT cohort, a baseline resting heart rate above 80 beats per minute predicted a maximal oxygen uptake 4.6 mL per kilogram per minute lower in men 23 years later, compared with a baseline below 60 — though the same gap in women (1.4) was indistinguishable from no difference. Notably, being physically active largely cancelled the disadvantage of starting with a fast heart^[7].

Does lowering it help? — Strong, and the answer is mostly no

This is the question that decides whether resting heart rate is worth acting on directly, and it has a real answer, because a drug exists that lowers heart rate and does almost nothing else. Ivabradine slows the sinus node — the heart's own pacemaker — without affecting blood pressure or the force of contraction. That makes it close to a clean experiment on the number itself. It has been run three times, in three populations.

In heart failure, lowering it works. SHIFT randomised 6,558 people with symptomatic heart failure, an ejection fraction — the share of blood the main pumping chamber pushes out with each beat, healthy being roughly 55–70% — of 35% or less, and a heart rate of at least 70 beats per minute. Cardiovascular death or hospital admission for worsening heart failure fell from 29% to 24% — about an 18% relative reduction, or roughly 5 fewer people affected per 100 over a median of 23 months (hazard ratio 0.82, 0.75–0.90)^[8]. The benefit came almost entirely from fewer heart-failure hospitalisations and fewer heart-failure deaths.

In coronary disease without heart failure, it does nothing. SIGNIFY was much larger — 19,102 people with stable coronary artery disease and no clinical heart failure — and achieved a bigger reduction, dropping the average heart rate from 70.6 to 60.7 beats per minute^[9]. Cardiovascular death or non-fatal heart attack occurred in 6.8% on the drug and 6.4% on placebo: no benefit, and if anything a nudge the wrong way (hazard ratio 1.08, 0.96–1.20). This is not a small trial that missed a real effect. Two further details matter. Among the subgroup with angina that limited their activity, the drug was associated with more events, while among those without such angina it was not (interaction p=0.02). A subgroup interaction inside a null trial is a reason to look again, not a finding in its own right — this trial did not show that lowering heart rate is harmful, it showed that lowering it did not help. And 18.0% developed a heart rate slow enough to be flagged as bradycardia, against 2.3% on placebo. The trial was funded by the drug's manufacturer.

In coronary disease with impaired pumping, it also does nothing. BEAUTIFUL lowered heart rate by 6 beats per minute in 10,917 people with coronary disease and an ejection fraction below 40%, most of them already on beta-blockers, and moved the primary endpoint not at all — a hazard ratio of exactly 1.00 (0.91–1.10)^[10]. Two secondary endpoints did fall in a prespecified faster-heart-rate subgroup, which is where the SIGNIFY hypothesis came from — and SIGNIFY then tested it properly and found nothing.

In hypertension, the correlation runs backwards. Across nine randomised trials of beta-blockers as first-line treatment for high blood pressure — 34,096 people on beta-blockers — the trials that achieved the lowest heart rates were the trials with the worst outcomes: all-cause mortality, cardiovascular mortality, heart attack and heart failure all tracked in the wrong direction^[11]. This is a comparison across trials rather than a randomised test of heart-rate targets, so it cannot show that the slowing caused the harm. But it is the opposite of what a simple lever model predicts.

In healthy adults, it has never been tested at all. No randomised trial has lowered resting heart rate in people without cardiovascular disease to see what happens.

The pattern is coherent. In heart failure, a fast heart is part of the disease — the failing ventricle compensates by beating more often, and slowing it gives it time to fill. Everywhere else, a fast heart appears to be a readout of something else: deconditioning, an over-active stress-response system, inflammation, illness. Lower the readout without changing what produced it and nothing improves. That is what a marker looks like when you push on it.

What's normal — and why the textbook range is nearly useless — Moderate

The familiar 60–100 beats per minute band describes a population, and applying it to one person throws away almost everything informative.

The best evidence on this comes from continuous wearable data rather than clinic visits. A retrospective study followed 92,457 US adults wearing wrist trackers for at least 35 weeks each, yielding roughly 33 million daily resting-heart-rate values^[12]. The findings:

The average was 65 beats per minute, but individual personal averages ran from 40 to 109 — a spread of about 70 beats between one healthy adult and another.
Personal averages differed systematically by age, sex, body mass index and habitual sleep duration.
There is a seasonal cycle: lowest in July, highest in January.
Within a person the value was fairly stable — but one in five had at least one week where it shifted by 10 beats per minute or more.

So two people can both be entirely normal and 40 beats apart, and the same person can drift 10 beats in a week without anything being wrong. A single reading compared against a population band is close to uninformative. A reading compared against your own rolling average, built over weeks, is a real measurement.

What actually moves it

Training — Strong, but modest and slow. A systematic review of 191 studies covering 215 samples of healthy people found that every form of exercise studied lowered resting heart rate, though only endurance training and yoga reached statistical significance in both sexes^[13]. That is a statement about which effects were reliably detected, not a claim that yoga and endurance training are equivalent — the review did not compare them head to head. Two patterns are more useful than the pooled average: the fall is larger the higher you start, and smaller the older you are.

In sedentary adults over 60, pooled controlled trials put the effect at about 6 beats per minute (range 2–12), an 8.4% reduction — and found that training programmes shorter than about 30 weeks generally did not produce a significant fall in this age group^[14]. This is a slow adaptation, not a response you will see in a fortnight. See Zone 2.

Alcohol — Strong, and the most visible thing in your own data. A controlled dose-response study found pre-sleep alcohol raised heart rate through the night by about 4% at a low dose and 14% at a high one, with the effect strongest in the first half of the night. See Alcohol.

Omega-3, specifically DHA — Moderate. Pooling 51 randomised trials in roughly 3,000 people, supplementation lowered heart rate by 2.2 beats per minute (1.4–3.1). Splitting fish oil's two main fatty acids apart, the DHA fraction produced the whole effect and the EPA fraction alone did nothing^[15]. See Omega-3.

Nicotine — Caution. E-cigarette use raises resting heart rate by roughly 1.4 beats per minute alongside blood pressure and measurable stiffening of the artery lining. See Smoking and nicotine.

Sauna — Weak-to-null. Pooling the eleven randomised passive-heating trials that measured it, a week or more of heat produced no significant change in resting heart rate. See Sauna.

Sleep timing — real but tiny. Across 255,736 nights in 557 students, going to bed just 30 minutes later than usual raised heart rate through the night by 0.18 beats per minute (0.11–0.26)^[16]. Statistically unambiguous and practically negligible — a useful calibration on how small most single-night influences really are, and a reason not to over-interpret a bad morning.

Beta-blockers and other rate-slowing drugs — Strong (that they decouple the number). These decouple the number from everything it normally reflects. On a beta-blocker, resting heart rate stops being a readout of autonomic state and becomes a readout of the dose.

When a single reading means nothing — Weak / qualitative

Any of the following makes one measurement uninterpretable: a recent meal, caffeine, alcohol the night before, standing rather than sitting, time of day, the month of the year, an infection you are incubating, dehydration, a poor night's sleep, anxiety in a clinic room, or a thyroid that is over- or under-active — see Thyroid. None of these has a well-quantified pooled effect size, and no attempt is made here to invent one — which is itself the point: the list is long, the magnitudes are unknown, and the sensible response is to stop reading single numbers.

When a low heart rate is a problem — consensus criteria; mechanism contested

For most people a slow pulse is good news, and endurance athletes routinely sit in the 40s. But the bottom of the range is not unconditionally benign.

The consensus criteria for interpreting athletes' electrocardiograms treat sinus bradycardia above 30 beats per minute in an asymptomatic trained athlete as a normal training-related finding needing no further evaluation. At or below 30 beats per minute, or with pauses of 3 seconds or longer, the response of heart rate to exercise should be assessed^[17].

The discriminator is not the resting number but the response. A trained heart accelerates normally on demand and causes no symptoms. A sinus node that is failing does not — and the giveaways are dizziness, fainting, breathlessness or exhaustion on exertion out of proportion to fitness. Those symptoms with a slow pulse warrant evaluation regardless of how much you train.

It is worth knowing that why athletes' hearts slow is genuinely unsettled. The traditional explanation is high vagal tone — the parasympathetic brake described in heart rate variability. The competing account, argued in a formal exchange in The Journal of Physiology, is that endurance training structurally remodels the sinus node itself, downregulating the ion channel that drives its pacemaker current^[18]^[19]. If the second account is even partly right, training-induced bradycardia is a lasting structural adaptation rather than a reversible shift in tone — which, if right, would help explain the long-standing reports that lifelong endurance athletes need pacemakers more often than expected — a claim this article notes but does not have a source for.

Resting heart rate versus heart rate variability — a comparison, not a finding

These are different measurements of the same beat sequence, and they are easy to conflate. Resting heart rate is how often the heart beats; heart rate variability is how much the interval between beats fluctuates. They move in opposite directions in health — a slower heart usually varies more — and that inverse relationship is strong enough that an uncorrected variability figure partly just reflects the rate.

For practical purposes, resting heart rate is the more robust of the two: it is easier to measure, and its association with mortality rests on far larger cohorts. In the one nocturnal validation against an electrocardiogram chest strap — 13 adults over 536 nights — resting heart rate came out accurate even in a device whose variability did not, and the details are in heart rate variability. Variability carries information that rate does not, but it is noisier and less comparable between devices and between people. Between the two, rate is the one to track if you are only going to track one.

Practical guidance

Measure against yourself, not a chart. The only useful comparison is your own rolling average over weeks. A population "normal range" spanning 60 beats is not a target.
Prefer continuous or overnight measurement to a spot check. A 24-hour average predicted mortality in the one study that tested both; the clinic reading did not. Wearables measure the rate more reliably than they measure variability, so this is the number your device is best at.
If you take it manually, do it the same way each time: seated, at rest for several minutes, before caffeine, at a consistent time of day. Count for a full 60 seconds.
Watch the trend, not the day. A rise sustained over weeks or months is the signal. A bad morning is noise, and the noise is larger than most single influences on it.
Treat a rise as a prompt to look for a cause, not a number to correct. New medication, poor sleep, drinking more, an untreated thyroid problem, a developing illness, deconditioning. The number is the smoke.
There is no case for taking a drug to lower it if you are otherwise well. That was tested in 19,102 people with coronary disease and did not help. There is no version of this experiment in healthy people, and no reason from the existing trials to expect a different answer.
The thing that lowers it durably is aerobic training, and it works slowly — plan in months, not weeks, and expect more movement if you are starting from a fast baseline.
A slow pulse with symptoms is not athletic. Dizziness, fainting or breathlessness on exertion alongside a low heart rate deserves evaluation.

What's overrated

Chasing a lower number. The trials that lowered it directly failed everywhere except heart failure, and one of them produced bradycardia in nearly a fifth of participants.
The 60–100 range as a personal verdict. Healthy individual averages span 40 to 109; the band tells you almost nothing about one person.
A single clinic reading. In the one head-to-head test, it added nothing once a continuous average was available.
Reading it as a fitness score. It survives adjustment for a directly assessed fitness measure, so it is not simply fitness — and conversely, an active person with a fast heart largely closes the fitness gap anyway.
Precision about the effect size. "17% per 10 beats" is quoted everywhere; the studies behind it disagree so violently that the honest claim is a direction and a rough magnitude, not a figure to two significant digits.