ECG and 24-hour Holter HRV: how clinics measure it

In a clinic, HRV is measured from an ECG ECG ECG A recording of the heart's electrical activity, taken with electrodes on the skin. ECG-based RR intervals are the reference standard for measuring HRV. More about ECG , the recording of the heart’s electrical signal. A short test takes about 5 minutes at rest, sometimes followed by standing, tilting or deep breathing. A Holter monitor records every beat for 24 hours or more while you go about your day. Either way, a technician checks the beats before any HRV number is calculated, and a doctor reads the result alongside everything else they know about you.

The ECG: the reference every other sensor is judged against

Each heartbeat starts with an electrical impulse from the sinus node sinus node Sinus node The heart's natural pacemaker, a small patch of cells that starts every normal heartbeat. Left alone it fires about 100 times a minute; the vagus nerve holds it lower at rest. More about Sinus node , the heart’s natural pacemaker. Electrodes on the skin pick up that impulse as it spreads through the heart muscle, and the ECG draws it as a wave. The tallest part of each beat, the R peak, rises and falls within a few hundredths of a second, so software can place it in time very precisely. The gap from one R peak to the next is the RR interval RR interval RR interval The time from one heartbeat to the next, measured in milliseconds between the R peaks of an ECG. HRV is calculated from a series of these intervals. More about RR interval . When only beats that started in the sinus node are kept, the gaps are called NN intervals NN intervals NN interval An RR interval between two normal heartbeats, in milliseconds, after extra beats and measurement errors are removed. Most HRV metrics are calculated from NN intervals. More about NN interval , and every standard HRV measure is calculated from them (Task Force, 1996).

How precisely a beat is timed also depends on how often the recorder reads the signal. The 1996 standards call 250 to 500 samples a second, “or perhaps even higher”, the optimal range. A lower rate, never below 100 a second, “may behave satisfactorily only if” the software interpolates between samples to refine the position of the R peak (Task Force, 1996). Later reviews add that 500 samples a second may be needed when the breathing swing in heart rate is small, while 125 to 200 can be enough when it is normal (Shaffer and Ginsberg, 2017). The lesson from electrode to number works through the arithmetic.

Short HRV tests in a clinic

The 12-lead ECG is too short

The ECG most people have had is the resting 12-lead: ten electrodes, a few minutes on a couch, and a printout that captures about 10 seconds of heartbeats (Steinberg and colleagues, 2017). It is built to show the shape of each beat from twelve angles, which is what a doctor needs to see rhythm and conduction problems or signs of heart damage. Ten seconds is far shorter than the 5 minutes the HRV standards ask for, so a routine 12-lead ECG is not an HRV test.

The 5-minute standard

For short recordings, the Task Force asked for 5 minutes “made under physiologically stable conditions”, for example lying quietly and breathing normally, with nothing changing during the recording. Their two standard formats are this 5-minute recording, analysed mainly with frequency methods such as LF LF LF power The variance of the heartbeat intervals that falls in the 0.04 to 0.15 Hz band, in ms²: rhythms repeating every 7 to 25 seconds. Once read as sympathetic activity, which is now disputed. More about LF power and HF power HF power HF power The variance of the heartbeat intervals that falls in the 0.15 to 0.40 Hz band, in ms²: the range of normal breathing. It mainly reflects vagal activity when breathing is 9 to 24 breaths a minute. More about HF power , and a nominal 24-hour recording, analysed mainly with time methods such as SDNN SDNN SDNN The standard deviation of all NN intervals in a recording, in milliseconds. It measures overall variability, and its value depends strongly on how long the recording lasts. More about SDNN (Task Force, 1996). Shaffer and Ginsberg call 5 minutes the conventional short-term standard, and stress that short and 24-hour values are not interchangeable even when they share a formula (Shaffer and Ginsberg, 2017).

Tilt, standing and deep breathing

Some clinics go one step further and challenge the nervous system on purpose. On a tilt table the patient lies strapped to a table that is raised to near upright. In a healthy person, the overall variability falls and the slower LF rhythm takes a larger share of the spectrum during tilt (Task Force, 1996).

The best known set of bedside tests comes from David Ewing and colleagues. Over ten years they used five simple, non-invasive reflex tests on 774 people with diabetes and found that heart rate tests became abnormal first and blood pressure tests later. They also found that a single test, the heart rate response to deep breathing, could not show how far the damage had gone; the full set was needed (Ewing and colleagues, 1985). A 2026 systematic review lists the five as the Valsalva manoeuvre (straining against a closed airway), heart rate change during deep breathing at six breaths a minute, the heart rate response on standing up (the 30:15 ratio), the blood pressure response on standing, and a sustained handgrip (Barkoula and colleagues, 2026). The first three test the vagal brake on the heart and the last two the sympathetic sympathetic Sympathetic nervous system The "fight or flight" branch of the autonomic nervous system. It speeds up the heart and makes it pump harder, but it acts too slowly to drive beat-to-beat HRV. More about Sympathetic nervous system side. A diabetes consensus panel bases its criteria for nerve damage to the heart mainly on the heart rate tests (Spallone and colleagues, 2011).

These tests use short pieces of ECG and a few simple ratios. Normal values change with age and sex, and the 2026 review recommends that each laboratory set its own (Barkoula and colleagues, 2026). A home reading cannot stand in for them.

The 24-hour Holter

How it is worn

In 1961 Norman Holter described a portable “electrocardiocorder” that recorded the heart continuously for long periods, together with ways to analyse the large amount of data it produced (Holter, 1961). Today’s version is a small recorder, worn on a belt or a strap and wired to several sticky electrodes on the chest. You wear it at home and at work, keep a diary of what you do and how you feel, and press a button on the recorder if you have symptoms (Steinberg and colleagues, 2017). Standard recordings last 24 to 48 hours, and some newer recorders can run for up to 30 days. Holters record between 3 and 12 leads, and the consensus recommends 12 leads when the shape of the beats matters, for example to sort premature beats (Steinberg and colleagues, 2017).

1. Sticky electrodes on the chest pick up the heart’s electrical signal
2. The recorder, worn on a belt or strap, stores every beat for 24 hours or more
3. The R peak: the sharp spike of each beat that the software times
4. The RR interval: the gap from one R peak to the next

Two drawings. The first is the outline of a person's torso with five round electrodes stuck to the chest: two just below the collarbones, one on the breastbone and two on the lower ribs. Thin wires run from each electrode to a small box clipped to the belt on the person's left side. The second is a short ECG strip on grid paper with three heartbeats. Each beat has a small bump, a tall sharp spike called the R peak and a rounded wave after it. The gaps between the spikes are marked as 0.96 and 1.06 seconds, so they are not the same.

Why some HRV measures need 24 hours

A day contains rhythms that no 5-minute recording can capture: sleep and waking, meals, activity and rest. Several measures exist only to describe them (Task Force, 1996; Shaffer and Ginsberg, 2017):

• SDNN over 24 hours takes in every speed of change, from breath-to-breath swings to the day-night cycle. Shaffer and Ginsberg describe 24-hour SDNN as the gold standard for estimating cardiac risk.
• SDANN averages each 5-minute block of the day first, then measures how much those 288 averages vary. It keeps only the slow changes.
• The SDNN index is the opposite half: the average of the 288 five-minute SDNN values, which keeps the faster variation inside each block.
• The HRV triangular index is built from a histogram of all the intervals. The Task Force asks for at least 20 minutes, and preferably 24 hours, for this kind of geometric measure.
• ULF power, the slowest band of the spectrum, below 0.003 Hz, needs at least 24 hours (Shaffer and Ginsberg, 2017).

The classic findings after a heart attack

The modern clinical interest in HRV started with Holter tapes. In 1987 Kleiger and colleagues analysed 24-hour recordings made about 11 days after a heart attack in 808 survivors, and followed them for an average of 31 months. Of all the Holter measures, SDNN had the strongest link with death: the risk was 5.3 times higher in people with an SDNN under 50 ms than in those above 100 ms, and the link held after allowing for the pumping strength of the heart and other known risk factors (Kleiger and colleagues, 1987).

Other studies followed. In 715 patients after a heart attack, ULF power ULF power ULF power The variance of the heartbeat intervals below 0.003 Hz, in ms²: rhythms slower than one cycle every 5 minutes, including the day-night cycle. It can only be measured from a 24-hour recording. More about ULF power was the single strongest spectral predictor of death over four years (Bigger and colleagues, 1992). In ATRAMI, 1,284 patients with a recent heart attack had a 24-hour Holter, and an SDNN under 70 ms carried 3.2 times the risk of cardiac death after other factors were taken into account (La Rovere and colleagues, 1998). The Task Force concluded that SDNN under 50 ms or a triangular index triangular index HRV triangular index The number of NN intervals in a recording divided by the height of the tallest bar of their histogram. A unitless geometric measure of overall variability that a few bad beats barely move. More about HRV triangular index under 15 marks highly depressed HRV, and recommended 24-hour recordings for risk studies after a heart attack (Task Force, 1996).

Holter software can also report two newer measures: heart rate turbulence, the speed-up and slow-down after a premature beat premature beat Ectopic beat A heartbeat that starts outside the heart's normal pacemaker and usually comes early. Ectopic beats are common, but they distort HRV unless removed or corrected. More about Ectopic beat , which reflects the baroreflex baroreflex Baroreflex The reflex that steadies blood pressure from beat to beat by changing heart rate. Its built-in delay makes heart rate rise and fall about every 10 seconds, the 0.1 Hz rhythm in LF power. More about Baroreflex (Bauer and colleagues, 2008), and deceleration capacity, which predicted death after a heart attack better than SDNN or the heart’s pumping strength (Bauer and colleagues, 2006). Newer mathematical methods have had “rather limited” success as clinical tools so far (Sassi and colleagues, 2015).

All of these numbers describe risk in groups of patients. They were never designed to sort healthy people, and a single value does not predict what will happen to one person.

How a Holter differs from a wearable

A chest strap, a watch and a Holter all list the time between beats, but the work behind the list is different.

• Leads. A Holter records 3 to 12 leads (Steinberg and colleagues, 2017). A strap records one (see how chest straps measure HRV), and a watch usually times the pulse wave with light instead (see PPG: how your watch reads your pulse). How much that costs in accuracy, and for which measures, is covered in what PPG does to HRV numbers.
• Beat labels. Holter software detects every QRS complex and sorts it by shape into normal, premature and so on (Peltola, 2012). Several leads give it more than one view of each beat.
• A person checks the beats. The Task Force asks for the beat labels to be edited “to a very high standard”, with visual checks and manual corrections, and warns that automatic filters, such as dropping any interval more than 20% from the one before, “should not replace manual editing” (Task Force, 1996). Peltola’s review stresses the value of an experienced Holter analyst for reliable data (Peltola, 2012). A consumer app relies on automatic filters, and usually no one looks at the trace.
• The reader. The report is reviewed and interpreted by a doctor (Steinberg and colleagues, 2017). A wearable shows you a number and a colour.

What a Holter or HRV report shows

Reports vary between software makers and clinics, and many Holters are ordered to look for rhythm problems, so their reports may contain no HRV at all. When HRV is included, these are the parts you are likely to see. The aim here is to help you follow what is on the page; what it means for you is a question for the doctor who ordered the test.

• The summary. Total recording time, the number of beats, and the lowest, average and highest heart rate.
• The rhythm findings. Counts of premature beats from the upper chambers (often labelled SVE, PAC or APB) and the lower chambers (VE, PVC or VPB), runs and pauses, and the times they happened. These are the main reason most Holters are ordered.
• The time-domain table. SDNN, SDANN SDANN SDANN The standard deviation of the mean NN interval of each 5-minute segment of a 24-hour recording, in milliseconds. It measures the slow changes over a day and ignores everything faster. More about SDANN , SDNN index SDNN index SDNN index The mean of the SDNN of each 5-minute segment of a 24-hour recording, in milliseconds. It keeps the variability inside each segment and drops the slow drift between them. More about SDNN index , RMSSD RMSSD RMSSD Root mean square of successive differences: an HRV metric built from how much each heartbeat interval differs from the next, in milliseconds. It mainly reflects vagal activity. More about RMSSD , pNN50 pNN50 pNN50 The percentage of neighbouring NN intervals that differ by more than 50 milliseconds. It tracks fast, vagally driven changes, like RMSSD, but as a share rather than a size. More about pNN50 and the triangular index, usually for the whole recording, sometimes for day and night separately.
• The frequency-domain table. ULF, VLF VLF VLF power The variance of the heartbeat intervals in the 0.003 to 0.04 Hz band, in ms²: rhythms repeating every 25 seconds to 5 minutes. It needs at least a 5-minute recording and is best read over 24 hours. More about VLF power , LF and HF power in ms², and often LF and HF in normalised units normalised units Normalised units LF or HF power as a percentage of their sum (LF + HF), rather than in ms². The two always add up to 100, so they carry one piece of information between them. More about Normalised units or as a ratio.
• The hourly trend. Heart rate and one or two HRV measures for each hour. In healthy people the spectrum shifts across the day, with the breathing-related HF share larger at night and the LF share larger by day (Task Force, 1996).
• The diary entries and button presses, lined up with the trace at those times.

Three cautions. The numbers come from a 24-hour recording, so they cannot be compared with a ring or strap value, or with 5-minute norms (Shaffer and Ginsberg, 2017). The published cut-offs were set in heart patients, mostly after a heart attack, so they do not translate into a verdict for anyone else. And a value depends on what happened that day, including the medicines you took and how much of the recording could be used. Bring your questions to the doctor; can HRV detect heart problems? explains what HRV can and cannot say on its own.

Patch monitors and event recorders

A patch monitor is a sticky strip worn on the left side of the chest, with the electrodes built in and no wires. They record one or two leads, and they can stay on for up to two weeks, including in the shower (Steinberg and colleagues, 2017). In 146 patients who wore a 24-hour Holter and a 14-day patch at the same time, the patch found 96 rhythm events over its 14 days against 61 for the Holter over its 24 hours (Barrett and colleagues, 2014).

Event recorders capture only short stretches. Some are not worn all the time and are switched on when symptoms start. Loop recorders keep a rolling memory of the ECG, so when you press the button, or the device spots an abnormal rhythm, they can save the moments before and after (Steinberg and colleagues, 2017). They are built to catch rare symptoms. Their short stretches cannot give a 24-hour HRV, and patch monitors, with one or two leads and processing designed for rhythm, are a different setup from the multi-lead Holters behind the classic HRV studies.

When clinical HRV does not mean much

• Ectopic beats. A premature beat makes a short interval and usually a long one after it. These must be found and removed or corrected, and the effect on blood flow can spill into the next 10 to 30 beats (Shaffer and Ginsberg, 2017). A recording full of them leaves little clean data. See ectopic beat.
• Atrial fibrillation. The upper chambers quiver instead of beating, the beats no longer start in the sinus node, and the gaps between them become irregular. Peltola counts atrial fibrillation as a physiological artefact artefact Artifact An error in a heartbeat recording, such as a missed, extra or mistimed beat, often caused by movement or poor sensor contact. Artifacts can badly distort HRV numbers. More about Artifact that can make HRV analysis impossible (Peltola, 2012). The standard measures assume sinus beats, so during atrial fibrillation they stop describing what they were designed for (Task Force, 1996).
• Pacing. NN intervals are, by definition, intervals between beats started by the sinus node (Task Force, 1996). Beats started by a pacemaker fall outside that definition, so in a paced rhythm the standard measures lose their usual meaning.
• Medicines. Beta-blockers raise HRV slightly in heart attack survivors, and some antiarrhythmic drugs lower it (Task Force, 1996). Research guidelines list cardioactive medication, including some antidepressants, antipsychotics and blood pressure drugs, among the things to record before comparing HRV (Laborde and colleagues, 2017). A report read without the medication list can mislead; do medications affect HRV? has more.

What this means for you

• If you have an ECG or a Holter, HRV may or may not be part of the report. Ask the doctor who ordered it what it was for and what it showed.
• A 24-hour SDNN, a 5-minute clinic value and a morning RMSSD from a strap are three different numbers. Keep each in its own trend.
• The famous cut-offs come from heart attack survivors. They are not a target for a healthy adult, and they do not apply to a watch.
• For your own day-to-day trend, a chest strap gives ECG beat timing at home, and rings and watches come close at rest (fingertip HRV covers phone cameras and rings). How to measure HRV compares the options, and is wrist HRV as accurate as a chest strap? looks at the evidence.