Corrected QT in Atrial Fibrillation Calculator

In atrial fibrillation one beat’s QTc means little. This takes the QT at the longest and the shortest RR interval of a 10-second ECG and averages them, one of the two approaches the American Heart Association sets out.

Last updated 8 Oct 2026

Averaged QTc in atrial fibrillation

2 beats, averaged

QT at the longest RR interval (ms)Earliest QRS onset to latest T-wave offset, in the lead where the QT is longest (usually II, V5 or V6), by the tangent method. Do not include a U wave; where T and U merge the measurement is unreliable and so is everything below it. Take it from the beat FOLLOWING the longest RR interval on the strip — the beat whose repolarisation that long cycle determined.

Longest RR interval (ms)In milliseconds, from the same 10-second recording. At 25 mm/s each millimetre is 40 ms, so a 27.5 mm cycle is 1,100 ms.

QT at the shortest RR interval (ms)The QT of the beat following the shortest RR interval on the same strip.

Shortest RR interval (ms)In milliseconds. If this exceeds the longest RR interval the two pairs have been entered the wrong way round — the averaged answer is the same either way, which is why the mistake is easy to miss.

427ms (Fridericia, averaged)Example

QT 440 ms at an RR of 1,100 ms; QT 360 ms at an RR of 600 ms

Formula

QTc = ½ × [ QT_long / RR_long^1/3 + QT_short / RR_short^1/3 ]
QT in ms, RR in seconds

why two beats: the AHA’s practice standards for in-hospital ECG monitoring set out two approaches in atrial fibrillation: the average of the longest and shortest QTc on a 10-second ECG, or the average of every QTc on it. This page implements the first, the one that can be done by hand
Fridericia rather than Bazett: Musat’s series compared the three corrections in the same patients in atrial fibrillation and in sinus rhythm. Bazett over-estimated (464 ± 34 against 445 ± 38 ms, P = 0.008) and Framingham under-estimated (385 ± 48 against 431 ± 40 ms, P less than 0.001), while Fridericia showed no significant difference (435 ± 33 against 440 ± 35 ms, P = 0.46)
the symmetry, and the trap in it: the average is symmetric in the two beats, so swapping the long and short pairs leaves the answer unchanged. A transposition is therefore invisible in the headline and visible only in the two individual values beside it
what no formula fixes: the QT needs at least a minute to settle after a change in rate, so in an irregular rhythm every beat carries the memory of earlier cycles. Some authorities go further: where QT variability is large, as in atrial fibrillation, rate correction “should not be applied at all”. In Musat’s dofetilide cohort, Bazett produced 24 dose reductions, “avoided in 33% of patients had Fridericia been used”

Worked example

QT 440 ms at an RR of 1,100 ms; QT 360 ms at an RR of 600 ms
Long cycle: 440 / 1.100^1/3 = 440 / 1.0323 = 426 ms
Short cycle: 360 / 0.600^1/3 = 360 / 0.8434 = 427 ms
Average = 427 ms. The two beats agree to within 1 ms, which is what a correction that is working looks like: a 500 ms difference in cycle length has been removed
By Bazett the same beats give 420 and 465 ms, averaging 442 ms — a 45 ms spread between beats and 15 ms higher overall. That is Bazett's rate dependence appearing as apparent beat-to-beat variability
Swap the two pairs and the average is unchanged, because the formula is symmetric in the two beats. Only the individual values beside the answer reveal the transposition

The three corrections in atrial fibrillation against the same patients’ sinus rhythm

Musat et al, Am J Cardiol 2013: patients studied in atrial fibrillation and again in sinus rhythm, so each is their own control. The cohort was small and taking dofetilide, which prolongs the QT, so the absolute values do not generalise — but a separate 90-patient series found the same ordering of residual rate dependence (Bazett r = −0.32, P = 0.002; Fridericia r = 0.20, P = 0.06).
Correction	QTc in atrial fibrillation	QTc in sinus rhythm	Difference
Bazett	464 ± 34 ms	445 ± 38 ms	Over-estimates by 19 ms, P = 0.008
Fridericia	435 ± 33 ms	440 ± 35 ms	No significant difference, P = 0.46
Framingham	385 ± 48 ms	431 ± 40 ms	Under-estimates by 46 ms, P less than 0.001

How the published approaches differ on the same irregular rhythm

These are genuinely different recommendations, not restatements of one. The last row is the strongest and least often followed: it says the quantity this page computes does not exist as a stable property of an irregular rhythm. Worth holding in mind before a number from here is written in a chart.
Approach	Source	What it asks for
Average the longest and the shortest QTc	AHA practice standards for in-hospital ECG monitoring	Two beats from a 10-second ECG — what this page implements
Average every QTc on the strip	AHA practice standards for in-hospital ECG monitoring	All available beats; more robust and not practical by hand
Measure the QT of the beat after the longest RR	Hospital-medicine guidance, as a practical alternative	A single beat, chosen because the long cycle gives the longest repolarisation
Do not rate-correct at all	Polish Cardiac Society working group on non-invasive electrocardiology	Where QT variability is large, “correction of QT intervals relative to heart rate should not be applied at all”, because the QT adjusts over a minute or more and no beat-by-beat relationship holds

Averaging a moving target

Every rate correction assumes one thing that atrial fibrillation denies: that the QT interval of a beat is determined by the cycle preceding it. It is not, or not only. Repolarisation adapts to a change in heart rate slowly — over at least a minute — so in an irregular rhythm each beat’s QT carries the memory of several earlier cycles as well as the last one. Correct a single beat and you get a number that depends on which beat you happened to choose.

The American Heart Association’s practice standards for in-hospital monitoring offer two ways round this: average the longest and shortest QTc on a 10-second ECG, or average every QTc on it. The first is what can be done on a printed strip and is what this page implements. It is a compromise and should be read as one. The spread between the two beats, printed beside the answer, is the useful diagnostic: where the two corrected values nearly agree, the correction has absorbed the cycle-length difference and the average means something; where they are tens of milliseconds apart, it has not.

Which correction to average matters more here than in sinus rhythm, and there is direct evidence. Musat and colleagues measured the same patients in atrial fibrillation and again in sinus rhythm. Bazett over-estimated the QTc in fibrillation by about 19 ms; Framingham under-estimated it by about 46 ms; Fridericia showed no significant difference. The consequence was not academic — in that dofetilide cohort, Bazett produced 24 dose reductions, a third of which would not have happened on Fridericia. This page therefore averages Fridericia values and prints the Bazett average beside them; the four corrections are compared directly on a regular rhythm.

One further position deserves airing because it contradicts the whole exercise. The Polish Cardiac Society’s working group on non-invasive electrocardiology advises that where QT variability is large, as it characteristically is in atrial fibrillation, rate correction “should not be applied at all”. The argument is sound: a quantity that depends on a relationship which does not hold cannot be measured by computing it more carefully. The practical middle ground is to average, to look at the spread, to prefer a longer recording when the answer will change anything, and not to quote a QTc from fibrillation to the nearest millisecond. This supports a clinician’s judgement rather than replacing it. An ECG interval is read alongside the patient, the drug chart and the electrolytes, and never instead of them.

Frequently asked questions

How do you measure QTc in atrial fibrillation?

The American Heart Association’s practice standards give two approaches: average the longest and the shortest QTc on a 10-second ECG, or average every QTc on it. A single beat is unreliable because repolarisation adapts to a change in rate over a minute or more, so each beat’s QT depends on several preceding cycles rather than just the last one.

Which QT correction is best in atrial fibrillation?

Fridericia, on the available direct evidence. Musat and colleagues measured the same patients in fibrillation and in sinus rhythm: Bazett over-estimated by about 19 ms (P = 0.008), Framingham under-estimated by about 46 ms (P less than 0.001), and Fridericia showed no significant difference (P = 0.46).

Can the QTc be corrected at all in an irregular rhythm?

Some authorities say no. The Polish Cardiac Society’s working group on non-invasive electrocardiology advises that where QT variability is large, as in atrial fibrillation, rate correction “should not be applied at all”, because the QT adjusts over at least a minute and no beat-by-beat relationship holds. The averaging approaches are a compromise, not a solution.

Which beat should the QT be measured on?

For the averaging method, the beat following the longest RR interval and the beat following the shortest, from the same 10-second recording. A practical single-beat alternative described in hospital-medicine guidance is the beat after the longest RR interval, because the long cycle produces the longest repolarisation.

Does it matter if I swap the long and short beats?

Not to the averaged answer, which is symmetric in the two beats — and that is the trap. A transposition shows up only in the two individual values beside the result, where the beat labelled as the long cycle will have the shorter corrected QT. The page flags it when the entered intervals are the wrong way round.

Related calculators

References

Musat DL, Adhaduk M, Preminger MW, et al. Correlation of QT interval correction methods during atrial fibrillation and sinus rhythm. Am J Cardiol. 2013;112(9):1379–1383.
Sandau KE, Funk M, Auerbach A, et al. Update to practice standards for electrocardiographic monitoring in hospital settings: a scientific statement from the American Heart Association. Circulation. 2017;136(19):e273–e344.
Correlation of QT interval correction method in patients with atrial fibrillation. Journal of the Department of Medical Services (Thailand). 90 patients. Accessed via he02.tci-thaijo.org, October 2026.
Baranowski R, et al. Guidelines of the Polish Cardiac Society Working Group on Noninvasive Electrocardiology on QT interval measurement and interpretation. Polish Heart Journal. Accessed via journals.viamedica.pl, October 2026.

Not medical advice. For healthcare professionals and education. Reference intervals vary by laboratory and assay — always use your own laboratory's. Never base a dose or a treatment decision on this page alone. Full disclaimer at calcengines.com/disclaimer/