QTc Calculator

Why the QT interval needs correcting

QTc is the QT interval corrected for heart rate to the value it would have at 60 beats per minute; Bazett’s formula, the default here, is QTc = QT / √RR, with the RR interval in seconds. The QT interval spans ventricular depolarisation and repolarisation, so it is roughly the duration of the ventricular action potential as seen from the body surface. It shortens as the heart speeds up, because repolarisation has less time available and the ion channel kinetics adapt. A raw QT of 400 ms is unremarkable at 60 beats per minute and clearly prolonged at 120.

Comparing raw intervals against a fixed threshold therefore mislabels people in both directions: tachycardic patients look normal when they are not, and bradycardic patients look prolonged when they are fine. Every correction answers the same question, which is what this QT would have been had the rate been 60 beats per minute.

That gives a check worth remembering. At a rate of exactly 60 the RR interval is exactly one second, and every formula below collapses to the raw QT. If a correction returns something other than the measured QT at a rate of 60, it has been applied wrongly.

The four corrections

All of them are functions of the RR interval in seconds, where RR = 60 / heart rate.

• Bazett, 1920. QTc = QT / √RR. The one machines print and examinations ask for.
• Fridericia, 1920. QTc = QT / RR^(1/3). Cube root instead of square root.
• Framingham, 1992. QTc = QT + 154 × (1 − RR). A linear correction fitted to the Framingham cohort.
• Hodges. QTc = QT + 1.75 × (heart rate − 60). The only one written in terms of rate rather than interval.

For a QT of 400 ms at 80 beats per minute the RR interval is 0.75 s, and the four give 461.9, 440.3, 438.5 and 435.0 ms. That is a spread of 27 ms from the same tracing, which straddles the 450 ms threshold. The spread is shown as a readout above precisely because it tells you when the answer is being driven by the choice of formula rather than by the patient.

Which one to use

Bazett is the default here because it is what you will meet on a machine printout and in an examination, not because it is the best. Its flaw is systematic and predictable: the square root over-corrects above 60 beats per minute and under-corrects below.

The reason is arithmetic. Above 60 the RR interval is below 1, and the square root of a number below 1 is smaller than its cube root, further from 1, so Bazett divides by a smaller number and returns a larger QTc than Fridericia. At 80 beats per minute the square root of 0.75 is 0.866 and the cube root 0.909. Below 60 the ordering reverses. In practice this means Bazett exaggerates prolongation in tachycardia, which is exactly the situation in which someone is most likely to be worried about a drug.

Fridericia holds up better at both extremes, which is why regulators and drug safety studies prefer it. If a patient’s rate is between about 50 and 90 the difference rarely matters. Once the rate is outside that band, quote the rate and the raw QT alongside whichever corrected value you report, so that the reader can see how much work the formula did.

Measuring the interval properly

The correction cannot rescue a bad measurement, and the measurement is the harder half.

• Measure from the earliest onset of the QRS to the end of the T wave, in the lead where the interval is longest. Lead II and V5 are conventional.
• At the standard paper speed of 25 mm/s, one small square is 40 ms and one large square is 200 ms.
• For the end of the T wave, take the tangent to the steepest part of its downslope and use the point where that tangent crosses the baseline. Guessing where the T wave flattens out systematically overestimates.
• Do not include a U wave. If the T and U waves merge, the measurement is unreliable and should be reported as such rather than as a number.
• Machine measurements are reasonable most of the time and unreliable exactly when it matters: flat T waves, prominent U waves, wide QRS complexes and artefact. A QTc near a decision threshold is worth measuring by hand.

Thresholds

• Above 450 ms in men and above 460 ms in women counts as prolonged.
• Above 500 ms is where the risk of torsades de pointes rises steeply, and it is usually the threshold for stopping a culprit drug.
• An increase of more than 60 ms from a patient’s own baseline is significant in itself.

The sex difference is real and appears after puberty, which is thought to reflect the effect of testosterone on repolarisation. These are population thresholds, so a single borderline value in a well person matters far less than a value that has clearly lengthened from a previous tracing on the same patient. That is the argument for a baseline electrocardiogram before starting any QT prolonging drug, and it is why the comparison tracing is worth more than the calculator.

Common mistakes

• Correcting in atrial fibrillation from one beat. Every formula depends on a single RR interval, and in atrial fibrillation that changes beat to beat, so the answer depends on which beat you picked. Average the QT and RR over ten or more consecutive beats, or accept that no single value is meaningful.
• Correcting a wide QRS without adjusting for it. In bundle branch block or a paced rhythm, part of the long QT is prolonged depolarisation rather than prolonged repolarisation. A common approach is to subtract the excess QRS duration above 120 ms, and at minimum the QRS width should be reported alongside.
• Entering the RR interval in milliseconds. Every formula here wants seconds. Using 750 instead of 0.75 in Bazett gives a QTc of 14.6 ms.
• Reading a single QTc as the risk. Torsades risk depends on the substrate as well as the interval: hypokalaemia, hypomagnesaemia, hypocalcaemia, bradycardia, female sex, structural heart disease and multiple interacting drugs all matter. A QTc of 480 with a potassium of 2.8 is a different situation from the same interval with normal electrolytes.