The normal resting pulse, measured on 93,334 people

Where the crowd actually stands

The middle adult measured here has a resting pulse near 70, not near 80. The median is 70 beats a minute for men and 74 for women, and nine adults in ten sit between 54 and 92 (men) or 56 and 96 (women). The lower fence at 60 cuts deep into the crowd: 15.3% of measured men and 7.9% of women have a resting pulse below it. Its upper fence at 100 stands almost clear of everyone: 2.6% are at 100 or above. One range, two very different fences.

Men above the line, women below it

In every adult age band in this survey, the women's middle value sits higher than the men's, by 2 to 6 beats. That is why a single range for everyone lets more men than women fall out at the bottom: of every hundred men measured, about 15 have a resting pulse under 60; of every hundred women, about 8. Nothing on this page says any one of those readings is a problem. It says where people are.

Young children measure fast. Adults differ little by age.

These are different people at each age, measured once, not the same people followed as they grow. The median measured pulse at one year old is 118 beats a minute; at nineteen it is 74. The difference between neighbouring ages is steepest in the first years of life and small by the teens. After twenty, the bands differ surprisingly little: the median for men is 70 in their twenties and 66 in their seventies. At three years old the middle measured child sits at 102 — above the adult range's top fence. One more reason a single “normal” range needs an age attached.

Two ranges, both put to the same crowd

The 60-to-100 limits were, in a cardiologist's words, “established by consensus and never formally examined”. Writing in 1996, David Spodick investigated the limits formally in 500 healthy people, set the results beside two large cohorts whose data pointed the same way, and proposed rounder fences: 50 to 90. MedlinePlus, the US National Library of Medicine's advice site, still prints 60 to 100 today. On the 50,249 adults here, Spodick's 1996 range holds 92.6% of men and 90.9% of women, against 82.6% and 89.1% for the conventional one. Holding more people does not by itself prove a better clinical limit; it shows which pair of fences describes this crowd. Neither range was built from these measurements. That is the test.

Switch the fences and watch the red move.

Your own pulse, placed among people your age

Sit down first, and stay seated for a few minutes — the survey rested people for about four minutes before counting, and a pulse taken just after climbing stairs is a pulse after effort. Then find your pulse at the wrist and tap the button with each beat for fifteen seconds. Treat the tapping result as an approximation; the count-for-thirty-seconds mode below matches the survey's own grain.

Twenty years of people, and the range did not drift

A reference range earns trust when it still describes people it was not built from. So: take the band that held the middle nine-tenths of adults measured in 1999–2008, and lay it over the adults of 2009–2018, whom it never saw. It contains 92.9% of the later men and 93.4% of the later women. The later median moved by 0 beats for men and 2 for women — for women, exactly one beat in a 30-second count, the smallest step this measurement can take. Year by year, the adult mean stays between 71.9 and 73.7. Whatever changed in American life across those twenty years, the measured crowd's resting pulse stayed where it was, to within one counted beat.

The limit to keep in mind: a band built to hold nine in ten of one decade's adults leaves about 7 in 100 of the next decade's outside it, and every recorded pulse here carries the grain of a 30-second count.

What shifts a reading in this crowd

About 12.1% of the adults report a prescription medicine from the classes used to slow the heart — beta blockers among them — matched here by the medicine's generic name. Their mean resting pulse is 68.1, against 73.2 for the other adults. Remove them, along with everyone pregnant or recorded with an irregular pulse, and the crowd barely moves: 87.5% inside 60–100 instead of 86%. The fences look the same without those people. And one reading is one reading: when the survey re-examined a sample of people on a second day in 2001–2002, the two readings gave an intraclass correlation of 0.69, so a single reading should not be read as exact.

Above 80: what follow-up studies found, stated carefully

In this crowd, 17% of men and 22.8% of women have a resting pulse between 82 and 98 — inside the conventional “normal”, and well above its middle. Cohort studies that follow people for years find an association at those levels: a 2016 meta-analysis of 46 studies and about 1.2 million people put the relative risk of death at 1.09 for each extra 10 beats a minute of resting pulse, and 1.45 for people resting above 80 against the lowest group. The same paper reports substantial differences between its studies and signs of publication bias; under one bias adjustment (a trim-and-fill estimate), the per-10-beats figure falls to 1.04. An association across a population is not a cause in a person, and this page gives no medical advice. What it can say is narrower: readings in the eighties and nineties sit inside the conventional range and well above where the typical measured adult in this crowd rests.

For advice about one person's pulse: a doctor. MedlinePlus's pulse page is linked below.

How these people were measured, and where the picture misleads

The National Health and Nutrition Examination Survey (NHANES) examines a sample chosen to stand for the civilian population of the United States. A trained examiner took each pulse by hand in a mobile examination centre, the person seated after resting quietly for about four minutes. From age five, thirty seconds counted at the wrist and doubled; up to age four, the heart counted through a stethoscope for thirty seconds and doubled — which is why every person's recorded pulse here is an even number (derived percentiles and means need not be). The 93,334 is the count of people with a recorded pulse between 1 and 199 in the public files. The dots are unweighted people; the percentages are weighted to the US population, so a printed share need not match the visible density of dots exactly. The unweighted adult mean differs from the weighted one by 0.03 beats. These are American readings, 1999–2018. They are not Romania's or Britain's, and a pulse at home in the morning is not a pulse in an examination centre. The percentiles printed here were first checked against the survey's own published reference tables for 1999–2008 (Ostchega and colleagues, 2011). Recomputed on the same years with similar exclusions, the male means land within half a beat of theirs and the percentiles within 3.4 beats, the tails sitting a little higher here; two of their laboratory exclusions (signs of current infection, thyroid tests) were not applied, which is the stated reason for the remaining gap. No second country's examiner-measured microdata with an open licence was found to test against; the replication on this page is decade against decade.

The measured ranges, by age and sex