Apple Watch Cardio Recovery: What a Good Number Is

Apple Watch Cardio Recovery is a modeled one-minute heart-rate drop (HRR1) after a workout. See age reference values and what moves the number.

Written by the Nuvo Longevity editorial team at Firefly Labs LLC.

NUVO wordmark in widely spaced white letters on a dark navy background with a cyan arc

Cardio Recovery on Apple Watch is a modeled estimate of heart-rate recovery in the first minute after exercise (HRR1). After a workout you start on the watch, Apple Watch keeps measuring heart rate for three minutes. The number it saves is a prediction of your 60-second heart-rate drop as if the effort had been submaximal work at 85 percent of predicted maximum heart rate. Apple caps the reported value at 55 beats per minute. A larger drop can indicate better cardiovascular health. It is the same idea clinicians call one-minute heart rate recovery, stored in HealthKit as heartRateRecoveryOneMinute, and shown in the Health app as Cardio Recovery.

Written by the Nuvo Longevity editorial team at Firefly Labs LLC, the developer of Nuvo. Measurement details are from Apple's November 2024 white paper, Using Apple Watch to measure heart rate, calorimetry, and activity, and from Apple's WWDC 2022 HealthKit session, which introduced the Cardio Recovery type. Mortality and training figures are from the peer-reviewed papers linked below. Checked 26 October 2026. This page is educational and is not medical advice. A Cardio Recovery number is not a diagnosis, and it is not a biological age.

What the Health app is showing

Heart-rate recovery is the fall in heart rate after effort stops. HRR1 is that fall over the first minute. The early drop is largely vagal reactivation: the parasympathetic nervous system putting the brake back on. Imai and colleagues showed that this early fall is vagally mediated, faster in athletes, and blunted in chronic heart failure (Journal of the American College of Cardiology, 1994).

Search interest in "cardio recovery apple watch" and "apple watch heart rate recovery" is the same object under two names. Apple's HealthKit identifier, available since iOS 16 and watchOS 9, is defined as the reduction from the peak exercise rate to the rate one minute after exercise ended. The watch's own estimate, described in the 2024 white paper, is narrower than a raw subtraction off the heart-rate graph. Apple designs it to stay comparable when one workout was easy and the next was hard, by predicting the 60-second recovery that would follow submaximal exercise at 85 percent of predicted maximum, with an upper bound of 55 bpm.

That distinction matters if you are 30 to 50 and already training. A 40 in the Health app is Apple's modeled value on a 0-to-55 scale. It is a different number from "my heart rate fell 40 bpm in the minute after I stopped." Both can be useful. They are useful for different comparisons.

How and when Apple Watch measures it

Cardio Recovery is tied to a user-initiated workout, including Fitness+ sessions. During the workout the watch uses its foreground heart-rate algorithm and tries to publish a heart rate about every five seconds, when the optical signal is good enough. In extended workout mode it duty-cycles the green LEDs, so the gap between samples can widen.

When you end that workout, the same foreground algorithm keeps running for three minutes to estimate cardio recovery. Keep the watch on, and keep the band snug, for those three minutes. End the workout when the effort ends. A long walk, stretch, or cooldown that is still inside the workout moves the moment Apple treats as the start of recovery.

On iPhone, the series lives in the Health app under Heart, labeled Cardio Recovery. WWDC 2022 says the type is readable and writable by apps, and that a workout built with HKLiveWorkoutBuilder saves a heart-rate recovery sample, with context, when the workout ends. Context can include the test type, the activity, the duration, and the maximum heart rate observed. A sample another app writes is still a one-minute recovery quantity. It is not automatically Apple's 85-percent model.

What good and typical values look like

Apple does not publish age bands, a "High" label, or a population average for Cardio Recovery. The white paper's guidance is directional: higher values can indicate better cardiovascular health, and 55 bpm is the top of the reported scale. Anyone promising an official Apple "good" cutoff is filling a blank Apple left empty.

The numbers people quote, including a drop of 12 bpm or less, come from clinical exercise tests. Those protocols are real, and they do not match the Health app scale one for one. Use them as a map of the physiology, then read your own trend on the watch.

Source Who and what protocol One-minute figure What it was used for
Apple Watch Cardio Recovery After a user-initiated workout. Three more minutes of optical heart rate. Modeled as 60-second recovery after submaximal work at 85% of predicted max. Cap of 55 bpm. The Health app value. Higher can indicate better cardiovascular health. No age grade is published. A within-person fitness signal that Apple tries to make comparable across workouts of different effort.
Cole et al., NEJM, 1999 2,428 adults referred for treadmill testing and thallium imaging (mean age 57). Two-minute cool-down at 1.5 mph and 2.5% grade. Median 17 bpm (IQR 12–23). Abnormal defined as a drop of 12 bpm or less. 26% of the cohort met that cutoff. All-cause death over six years: 19% vs 5%. Unadjusted relative risk 4.0 (95% CI 3.0–5.2). Adjusted relative risk 2.0 (95% CI 1.5–2.7).
Jouven et al., NEJM, 2005 5,713 asymptomatic Paris civil-service men, ages 42–53, bicycle test between 1967 and 1972. Recovery counted from peak to one minute after exercise stopped. A decrease of less than 25 bpm. Sudden death over 23 years. Relative risk 2.20 (95% CI 1.02–4.74), still present after adjustment. The same profile also tracked all-cause death.
Circulation abstract 12508, 2015 22,643 patients ages 30–79 without known cardiovascular disease. Non-imaging exercise tests, active recovery, first minute. A conference abstract, not a full methods paper. Median (IQR): ages 30–39, 20 (15–26); 40–49, 19 (14–25); 50–59, 18 (13–23); 60–69, 15 (10–20); 70–79, 11 (6–16). A 13 bpm drop sat at the 17th, 20th, 27th, 40th, and 60th percentiles across those age bands. The bpm that doubled mortality risk was lower at older ages.

For a data-driven reader in the 30–50 range, the clinical medians in that abstract sit around 18–20 bpm on an active-recovery treadmill test. Cole's referred patients, a decade older on average, had a median of 17. Jouven's healthy middle-aged men were flagged below 25 bpm, on a bicycle protocol without Cole's walking cool-down. The same body can produce a different one-minute drop when the cool-down changes. A slow walk keeps heart rate up. An abrupt stop lets it fall faster. That is why 12 and 25 both appear in textbooks, and why neither is an Apple Watch grade.

Training status moves the number at least as much as a birthday does. Darr, Bassett, Morgan, and Thomas compared young and older men, trained and untrained, after a maximal bike test (American Journal of Physiology, 1988). Each cell had only five men, so treat it as physiology, not a norm chart. Once fitness was matched, age did not slow the fast phase of recovery (15–120 seconds). Trained men recovered faster than untrained men. A fit 50-year-old can out-recover a sedentary 30-year-old on this metric.

Why it matters for longevity and cardiovascular risk

Cole's question was whether a slow first minute after exercise marked people who would die sooner. In 2,428 adults followed for six years, 213 died. A drop of 12 bpm or less carried an adjusted relative risk of death of 2.0 after accounting for age, sex, medications, thallium perfusion defects, standard risk factors, resting heart rate, the rise in heart rate during exercise, and the workload achieved. The signal was the recovery minute itself, on top of how hard the heart had worked and how fast it beat at rest.

Jouven's Paris Prospective Study asked a sharper outcome: sudden death in men who looked healthy at the start. Over 23 years, 81 sudden deaths occurred. A one-minute drop under 25 bpm carried a relative risk of 2.20. Resting heart rate above 75 bpm, and a small rise in heart rate during the test, were separate risks. Recovery was one piece of the heart-rate profile, and it added information.

Qiu and colleagues pooled prospective cohorts through August 2016 (Journal of the American Heart Association, 2017). Attenuated versus fast recovery was associated with a hazard ratio of 1.69 (95% CI 1.05–2.71) for cardiovascular events, across 1,061 events in 34,267 people, and 1.68 (95% CI 1.51–1.88) for all-cause mortality, across 2,082 deaths in 41,600 people. Each 10 bpm slower recovery lined up with a hazard ratio of 1.13 (95% CI 1.05–1.21) for cardiovascular events and 1.09 (95% CI 1.01–1.19) for death. In their subgroup checks, the all-cause mortality association was significant in studies that used maximal exercise testing.

Those hazard ratios belong to lab and clinic protocols. They are a reason to care about the direction of your Apple Watch series. They are not a calculator that turns a Health app value of 18 into Cole's relative risk of 2.0. Apple's number is submaximal and modeled. The mortality papers are mostly maximal tests with a defined cool-down. Slow recovery in both settings points at weaker vagal reactivation. The bpm scales differ.

For aging, that vagal story is the point. A cardiovascular system that can raise rate under load and then hand control back to the parasympathetic side is showing a capacity that declines, on average, with age and that training can hold onto. It sits next to cardiorespiratory fitness, which has its own longevity literature. How accurate the watch's VO2 max estimate is, and how to read it, is covered in VO2 max on Apple Watch. This page stays on the recovery minute.

What moves Cardio Recovery

Aerobic fitness. The trainable part is real. Kline and colleagues randomized overweight adults with untreated obstructive sleep apnea to 12 weeks of aerobic training or stretching (International Journal of Cardiology, 2013). Training built to 150 minutes a week at 60 percent of heart-rate reserve, four sessions a week. Heart-rate recovery at one, three, and five minutes improved versus control, and peak oxygen uptake rose about 11 percent. Apnea severity barely explained the recovery change. The adaptation was the training. How to place that work in a longevity week, using wearable zones, is the subject of zone 2 cardio and wearable data.

Sleep. One night changes the next test. Cincin and colleagues treadmill-tested 30 healthy adults (mean age 33) after a normal night and after a hospital night shift (Sleep and Breathing, 2015). One-minute recovery was 31.0 bpm after the shift and 36.1 bpm after regular sleep (p = 0.004). Exercise capacity also fell. A single bad night is a reason to ignore that day's Cardio Recovery as a fitness verdict.

Alcohol. The cleanest direct experiment is small and mixed with caffeine. Wiklund and colleagues had 10 healthy adults, ages 19–30, ride a maximal bike test after an energy drink mixed with alcohol, after the energy drink alone, and after no drink (Clinical Physiology and Functional Imaging, 2009). Post-exercise heart rate and heart-rate variability recovered more slowly after the energy drink plus alcohol than after exercise alone. That is alcohol plus a stimulant drink, in ten people, not a dose-response for wine at dinner. It is still a reason to keep a drinking night out of the set of workouts you use to judge a trend.

Illness. An acute illness shifts autonomic balance and resting heart rate, so the next workout's recovery minute is a sick-day sample. The sleep-apnea arm of the Kline trial is the chronic version of the same idea: the apnea group started with a blunted recovery versus matched adults without apnea. Treat a fever week as missing data for the trend line.

Heat. Heat and fluid loss raise heart rate at a given pace. Apple's model is built to absorb some differences in how hard a given workout was. A hot, under-hydrated session is still a different autonomic day from a cool one. Compare readings taken in similar conditions before you decide fitness moved.

Heavy training and overreaching. A bigger drop is the usual fitness direction, and it is the wrong direction to celebrate in the middle of an overload. Aubry and colleagues followed 31 experienced male triathletes through an overload and a taper (PLOS ONE, 2015). The ten who developed functional overreaching lost about 2 percent of performance and gained about 8 bpm of one-minute recovery during the overload. Both shifted back after the taper. Thomson and colleagues tested 11 male cyclists and triathletes after light training, heavy training, and two days of rest (Journal of Science and Medicine in Sport, 2015). A 5-minute time trial fell 2.2 percent after the heavy block, while most 60-second recovery indices got faster. Peak heart rate had dropped, and controlling for that peak weakened the link between recovery and performance. If you are in a hard block, read Cardio Recovery next to how the workout felt and whether performance is holding.

How to improve it

The lever with trial evidence is repeated aerobic work at a moderate heart-rate reserve, long enough for the autonomic system to adapt. Kline's dose is a concrete version: about four sessions a week, building over a month to 150 minutes a week at 60 percent of heart-rate reserve, held for the remaining eight weeks. Expect the interesting change over months, in the company of a higher sustainable pace, rather than after one heroic interval day. Zone boundaries, and how they show up on a watch, are already laid out in the zone 2 article linked above.

Protect the measurement while you train it. Sleep enough that the night shift result in Cincin's study is not your baseline. End the workout when the effort ends, then leave the watch on for the three-minute window. Retest in similar heat, at a similar time of day, and off an alcohol-plus-caffeine experiment. During a planned overload, a sudden jump up can be fatigue. The taper is the fairer retest, which is what Aubry's triathletes showed when the faster recovery reversed as performance returned.

Medications that limit heart rate, including beta blockers, change the raw curve. Cole's model adjusted for medication use. Stay on a prescribed drug and tell a clinician if recovery is a worry. This article is a way to read a wearable series.

Measurement caveats

  • The Health app value from Apple Watch is a prediction aimed at 85 percent of predicted maximum, capped at 55 bpm. Computing peak minus the rate one minute later, off the workout graph, answers a different question.
  • Optical heart rate depends on signal quality. A loose band, cold skin, or a tattoo under the sensor can drop samples. Extended workout mode spaces the green LEDs out to save battery.
  • Irregular rhythms are a weak use case for a one-minute optical drop. Apple's own validation of passive pulse rate shows lower coverage in several arrhythmias. Atrial fibrillation and frequent ectopy can make the minute uninterpretable.
  • A cool-down left inside the workout, or a third-party session that never runs the three-minute foreground window, can omit the sample or time it against the wrong moment.
  • Another app can write heartRateRecoveryOneMinute into Health. Check the source before you treat every sample as Apple's model.
  • One reading is a sample. The mortality papers used a standardized test, once, in thousands of people. Your series is the opposite design: many imperfect samples in one person. The trend across similar workouts is the quantity worth keeping.

A string of very low values, on workouts you end on time, in the absence of a rate-limiting drug, is a reason to talk with a clinician. It is material for that conversation. It is not Cole's diagnosis reproduced on a wrist.

How it relates to HRV, resting heart rate, and VO2 max

These four numbers share a cardiovascular system and answer different questions.

Heart-rate variability is the beat-to-beat scatter, usually an overnight RMSSD on Apple Watch. Cardio Recovery is the speed of the drop in the first minute after a workout. Both lean on parasympathetic function. They are sampled in different states: sleep versus the minute after effort. A low morning HRV with a stable Cardio Recovery trend is a different week from a falling recovery trend with a normal HRV. Training HRV on purpose is covered in heart rate variability training.

Resting heart rate is the tonic rate when you are still. Cole adjusted for it, and the recovery minute still predicted death. A lower resting rate often travels with fitness. It does not replace HRR1. Jouven found resting rate above 75 bpm and a small recovery drop were separate sudden-death markers in the same men.

VO2 max (Cardio Fitness on Apple Watch) estimates peak oxygen uptake. Kline's trainees improved recovery and VO2 peak together. Darr's trained subjects recovered faster at a given age. You can still split them: a person can hold a solid aerobic ceiling and have a sluggish vagal brake the morning after poor sleep, or the reverse during functional overreaching, when recovery looks faster while performance dips. Accuracy limits of the watch's VO2 max estimate are in the VO2 max article, not here.

Morning readiness scores from WHOOP, Oura, and Garmin fold overnight HRV, resting heart rate, sleep, and load into a daily number. Cardio Recovery does not exist until a workout ends. One asks whether today looks recovered. The other asks how the heart came down after the last effort. The readiness construct, and what the evidence supports for improving it, is in recovery readiness score.

Where this sits in a daily biological age

Nuvo is an iPhone and Apple Watch app. It reads Apple Health through HealthKit and calculates a daily biological age from 21 biomarkers: 16 wearable inputs plus optional lab results such as HbA1c, fasting glucose, HDL, CRP, and creatinine. Heart rate recovery is one of those wearable inputs, alongside VO2 max, HRV, and resting heart rate. The list is spelled out in the 21 longevity biomarkers Nuvo tracks. Nuvo does not publish a separate Cardio Recovery screen. The minute-after-exercise value is one HealthKit input in the daily score, computed by on-device AI. No account is required. The App Store privacy label is Data Not Collected.

Nuvo Premium is $4.99 a month or $24.99 a year after a 7-day free trial, which is $2.08 a month when billed annually. Labs stay optional. Wearable data alone produces a score. Adding labs fills in metabolic and inflammatory context the watch cannot see. How the broader set of Apple Watch signals connects to that score, beyond this one metric, is in lowering biological age with Apple Watch data and which Apple Watch metrics feed a biological age. Privacy of the on-device design is in why a health app should not need an account. Pricing detail is on the pricing page.

Download app

Frequently Asked Questions

What is Cardio Recovery on Apple Watch?
Cardio Recovery is Apple's name for a one-minute heart-rate recovery estimate (HRR1). After a user-initiated workout, the watch measures heart rate for three more minutes and saves a prediction of the 60-second drop you would show after submaximal exercise at 85 percent of predicted maximum heart rate. The reported value is capped at 55 bpm. HealthKit stores it as heartRateRecoveryOneMinute, a type added in iOS 16 and watchOS 9.

What is a good Apple Watch Cardio Recovery number?
Apple does not publish a good, high, or age-graded cutoff. Higher values can indicate better cardiovascular health, on a scale that stops at 55 bpm. Clinical one-minute medians are lower and protocol-specific: about 17 bpm in Cole's referred treadmill patients, and about 18–20 bpm for ages 30–50 in a 22,643-person exercise-test abstract that used active recovery. Those lab figures are context for the physiology. They are not grades for the Health app number.

How does Apple Watch measure heart rate recovery?
During a workout the watch uses its foreground optical heart-rate algorithm. When you end the workout it keeps that algorithm running for three minutes and estimates cardio recovery. End the session when the effort ends, and leave the watch on for those three minutes. A value written by another app into the same HealthKit type can be a raw peak-to-one-minute drop rather than Apple's 85-percent model. Check the source.

Does a Cardio Recovery under 12 bpm mean higher mortality risk?
A drop of 12 bpm or less was the abnormal cutoff in Cole and colleagues' 1999 New England Journal of Medicine study of 2,428 referred patients, using a treadmill cool-down at 1.5 mph and 2.5 percent grade. That group had an adjusted relative risk of death of 2.0 over six years. Jouven's healthy men were flagged below 25 bpm on a different bike protocol. Apple's Cardio Recovery is a modeled submaximal estimate with a different scale. A Health app value under 12 is a reason to look at the trend and, if it persists, to talk with a clinician. It is not Cole's result copied onto the wrist. This page is not medical advice.

How is Cardio Recovery different from HRV, resting heart rate, and VO2 max?
Cardio Recovery is the first minute after a workout. HRV is beat-to-beat variation, usually overnight. Resting heart rate is the rate while you are still, and Cole found recovery predicted death after resting heart rate was already in the model. VO2 max estimates peak oxygen uptake. They move together with aerobic training and they diverge after a bad night, in heat, or in a heavy training block, when a faster recovery can show up alongside worse performance.

What improves Cardio Recovery on Apple Watch?
Repeated aerobic training does. In a 12-week trial, about 150 minutes a week at 60 percent of heart-rate reserve improved one-minute heart-rate recovery and raised VO2 peak. Sleep affects the next reading: one night shift cut one-minute recovery from 36 bpm to 31 bpm in healthy adults around age 33. Keep test conditions similar, and treat a faster drop during an overload block as a possible fatigue sign until a taper confirms it. Nuvo folds heart rate recovery from HealthKit into a daily biological age from 21 biomarkers. It does not replace the training.

References