Resting heart rate predicts risk across a population. It says less about one person.

A higher resting heart rate is associated with higher risk of several diseases, and the association is linear: pooling 45 prospective cohort studies, each 10 beats per minute increment carried a relative risk of 1.12 for coronary artery disease (95% CI 1.09 to 1.14), 1.05 for stroke (95% CI 1.01 to 1.08) and 1.16 for non-cardiovascular disease (95% CI 1.12 to 1.21) [s1]. What that does not establish is what any individual's number means, because the same measurement varies between healthy people by as much as 70 beats per minute [s2].

Those two findings are both true and they pull in opposite directions. Together they describe what a resting heart rate is good for and what it is not.

The population-level association

The 2016 meta-analysis, published in CMAJ, searched PubMed, Embase and MEDLINE from inception to 5 March 2016 and pooled 45 non-randomised prospective cohort studies using a random-effects model, with restricted cubic splines to test the shape of the relationship [s1].

Every association it found was linear rather than threshold-shaped [s1]. Per 10 beats per minute, the multivariable-adjusted relative risk was 1.12 for coronary artery disease, 1.05 for stroke, 1.12 for sudden death (95% CI 1.02 to 1.24), 1.09 for all types of cancer (95% CI 1.06 to 1.12), 1.16 for non-cardiovascular disease, and 1.25 for non-cardiovascular disease excluding cancer (95% CI 1.17 to 1.34) [s1].

Analysed by category rather than increment — with under 60 beats per minute as the reference — the relative risk of coronary artery disease was 0.99 at 60–70 (95% CI 0.93 to 1.04), 1.08 at 70–80 (95% CI 1.01 to 1.16) and 1.30 above 80 (95% CI 1.19 to 1.43) [s1]. For non-cardiovascular disease the same categories gave 1.17, 1.31 and 1.57 [s1].

One result did not survive scrutiny. When the analysis was restricted to studies of general populations, excluding those enrolling people with hypertension or diabetes, the associations with coronary artery disease, stroke and non-cardiovascular disease persisted — but the association with sudden death disappeared [s1].

Why "normal" is a weak concept here

The population data says the risk gradient is continuous, which already implies there is no clean cut-off. Consumer wearables have since shown how wide the healthy spread is.

A retrospective longitudinal study analysed 92,457 de-identified adults across all 50 US states who wore a wrist-based heart rate tracker consistently — at least two days a week, at least 20 hours a day, for at least 35 weeks between March 2016 and February 2018 — producing nearly 33 million daily resting heart rate values over a median 320 days per person [s2].

Mean daily resting heart rate across the cohort was 65 beats per minute. The range across individuals ran from 40 to 109 [s2]. Mean values differed significantly by age, sex, body-mass index and average sleep duration, and there was a seasonal pattern, with a minimum in July and a maximum in January [s2].

The authors' conclusion is the operative one: individuals have a daily resting heart rate that is normal for them, and one person's normal can differ from another's by as much as 70 beats per minute [s2]. Within a person, though, the measure was comparatively stable — although 20% of participants experienced at least one week in which their resting heart rate fluctuated by 10 beats per minute or more [s2].

What the number tracks

If a resting heart rate is not a standalone verdict, it is still correlated with things clinicians care about. The Project Baseline Health Study compared resting heart rate from a study watch against simultaneous ECG measurement in 875 participants with a mean age of 50.9 years, and found excellent agreement, with an intraclass correlation coefficient of 0.946 [s3]. Mean values were 66.6 beats per minute in female participants and 64.4 in male participants [s3].

In univariate analyses, the same characteristics trended with a higher resting heart rate in both sexes: higher body-mass index, higher C-reactive protein, the presence of type 2 diabetes, a higher score on the World Health Organization Disability Assessment Schedule, and lack of health insurance [s3]. On regression analysis, which characteristics remained associated differed between female and male participants [s3].

That list is a reminder of what an observational association contains. Resting heart rate moves with fitness, body composition, inflammation, metabolic disease, medication, illness, sleep, stress, caffeine, alcohol and time of year — several of which independently cause the outcomes the cohort studies measured. The meta-analysis adjusted for conventional risk factors, but residual confounding in 45 non-randomised cohorts is not something adjustment fully removes, and the authors describe the measure as an independent predictor rather than a cause [s1].

The useful reading

The evidence supports treating resting heart rate as a trend line rather than a score. A stable personal baseline is the informative quantity; a comparison against someone else's, or against a textbook range spanning 60 to 100, is much less so given a healthy spread of 40 to 109 [s2].

A sustained, unexplained change from an individual's own baseline is the pattern the wearable literature can actually detect [s2]. Whether such a change matters, and what might be causing it, is a clinical question — the same measurement can reflect improved fitness, a developing infection, a new medication, or nothing at all.