Long QT syndrome

• The QT interval is the time the ventricles take to depolarise and then recover. In long QT syndrome (LQTS) recovery is delayed, so the heart is vulnerable to a twisting, polymorphic ventricular tachycardia called torsades de pointes, which causes faints, seizure-like episodes and sudden death.
• LQTS is inherited (about 1 in 2,000 to 2,500 people; mostly KCNQ1, KCNH2 and SCN5A) or acquired (drugs, low potassium or magnesium, slow heart rates, other illnesses).^[13][2]
• The 2022 European Society of Cardiology (ESC) guideline diagnoses LQTS from a corrected QT interval (QTc) of 480 ms or more on repeated ECGs, or a diagnostic score above 3, or a pathogenic gene variant whatever the QT looks like.^[2]
• Treatment of the inherited form is built on a non-selective beta-blocker (nadolol or propranolol), avoidance of triggers and QT-prolonging drugs, and in selected people mexiletine, left cardiac sympathetic denervation or an implantable defibrillator (ICD).^[2]
• Torsades is treated by stopping the offending drug, giving intravenous magnesium whatever the serum level, correcting potassium, raising the heart rate, and defibrillating if the patient is unstable.^[1][4]

By the end of this page you should be able to:

1. Explain which ion currents repolarise the ventricle and how a fault in each lengthens the QT interval.
2. Measure a QT interval correctly, correct it for heart rate, and say why the correction formula matters.
3. Quote the diagnostic thresholds for LQTS and apply the modified Schwartz score.
4. Describe the inheritance of Romano-Ward and Jervell and Lange-Nielsen syndromes and compare LQT1, LQT2 and LQT3 (gene, trigger, ECG, treatment response).
5. List the causes of acquired LQTS and the risk factors that make torsades more likely.
6. Explain how torsades de pointes starts and give emergency treatment with doses.
7. Plan long-term management: lifestyle, drug avoidance, beta-blockers, mexiletine, left cardiac sympathetic denervation and ICD.
8. Advise on sport, pregnancy and the postpartum period, children, and family screening.

This section recalls how a heart muscle cell recovers after each beat, because every cause of long QT is a story about that recovery being slowed.

A ventricular cell fires in five phases: a fast sodium upstroke (phase 0), a brief early repolarisation (1), a plateau in which slow calcium entry is balanced by potassium leaving (2), the main repolarisation driven by potassium leaving (3), and rest (4).^[4] The QT interval, from the start of the QRS complex to the end of the T wave, spans phases 0 to 3 of all the ventricular cells together.^[4] Two potassium currents do most of the work of phase 3: the rapid current IKr (channel encoded by KCNH2, formerly HERG) and the slow current IKs (KCNQ1).^[3][4] A small fraction of sodium channels (SCN5A) inactivate slowly and leak sodium into the cell during the plateau, the late sodium current.^[2]

The QT lengthens whenever outward (repolarising) current falls or inward (depolarising) current rises.^[3][4] Loss of potassium-channel function (LQT1, LQT2) and a gain of sodium-channel function that enlarges the late sodium current (LQT3) are the commonest inherited examples; drugs that block IKr are the commonest acquired example.^[3][4]

Why is slow recovery dangerous? When repolarisation is very prolonged, inward currents can re-activate during phase 2 or 3, producing small oscillations called early after-depolarisations (EADs). If an EAD reaches threshold it triggers an extra beat. Different layers of the ventricular wall also recover at different speeds, creating zones of unidirectional block that favour re-entry.^[4] A premature beat then falls on a vulnerable, uneven myocardium and starts torsades.

A wrongly measured QT is the commonest reason for both missed and over-diagnosed LQTS. This section covers how to measure, how to correct for heart rate, and the limits of each formula. For basic ECG reading, see ECG interpretation.

• Start the measurement at the first deflection of the QRS and stop where the T wave ends; take the mean of three to five beats in a single lead.^[4] The T wave is usually best seen in leads II and V5 in one source,^[3] while another finds the longest intervals in the precordial leads, V3 or V4.^[4] Choose the lead in which the end of the T wave is clearest.
• Find the end of the T wave with the tangent method: draw a tangent along the steepest downward slope of the T wave and take the point where it meets the baseline.^[3] The threshold method, in which the end is where the T wave returns to the baseline, gave QT values on average about 10 ms longer in a study of 1,484 people with and without LQTS gene variants, although both methods were reproducible and had good diagnostic accuracy.^[6] The practical lesson: use one method consistently and compare like with like.
• Low-amplitude U waves are left out. A large U wave fused to the T wave is included.^[3][4]
• In children and in anyone with sinus arrhythmia, use the average QT and average preceding RR interval, and look at several beats and take the longest.^[3]
• Automated machine readings are a starting point only. Check the measurement by hand when the value is borderline or when a decision depends on it.

The QT shortens as the heart rate rises, so a QT measured at 100 beats per minute cannot be compared with one at 60. The corrected QT (QTc) removes this effect. With QT and RR (the interval between two R waves) in seconds:^[5]

In healthy people the 99% upper limit of normal for Bazett was 472 ms in men and 482 ms in women, higher than usual clinical standards.^[5] The paediatric review states that Bazett is still the most widely used in studies and in practice and advises repeating the ECG after mild exercise if the resting rate is below 50.^[3]

Pragmatic advice: at heart rates between about 50 and 90 the differences between formulae are small. Outside that range, or when a decision depends on a value near a threshold, calculate Fridericia as well, and always state which formula was used. No formula is exact for every person, because each individual has their own relationship between QT and heart rate, and rate-correction may be especially unreliable when assessing small drug-induced changes.^[4]

Thresholds differ by purpose, and this causes much confusion.

• Diagnosis of inherited LQTS (2022 ESC): QTc of 480 ms or more on repeated 12-lead ECGs, with or without symptoms, or a diagnostic score above 3 (Class I, level C).^[2] In someone with an unexplained arrhythmic faint or cardiac arrest, a QTc of 460 ms or more is enough to consider the diagnosis.^[3]
• Borderline range (older review): 440 to 460 ms in men and 440 to 470 ms in women.^[4] Routine drug-safety monitoring has used cut-offs of 450 ms for men and 470 ms for women.^[5]
• High arrhythmic risk: the dangerous arrhythmias cluster in people whose QTc reaches 500 ms or longer.^[4] A QTc above 500 ms is a high-risk feature in congenital LQTS, and above 600 ms is regarded as extremely high risk.^[3] The ESC drug table advises discontinuing a QT-prolonging drug if QTc exceeds 500 ms.^[1]
• Sex and age: adult women have longer QTc than men, because the QT shortens in men after puberty.^[4]
• Natural variation: QT varies with the way the ECG is acquired, electrolytes, autonomic tone and observer; over a day it can swing by 75 to 100 ms.^[4] This is why repeated ECGs are required for the diagnosis.
• Normal does not exclude: About 25% of people with gene-confirmed LQTS have a QTc in the normal range.^[3]

Inherited LQTS (also called congenital LQTS) is a channelopathy: a disease of the ion channels. This section covers how common it is, how it is inherited, the genes, the three main genotypes and the less common syndromes.

In a study of 44,596 newborns (43,080 of them white) who had an ECG at 15 to 25 days of age, with genetic testing when the QTc was prolonged, at least 1 in 2,534 white infants had LQTS (the authors suggested the true figure may be near 1 in 2,000).^[13] In that study about half (51%) of the relatives of genotype-positive infants also carried the disease-causing variant.^[13] Most cases are inherited, and roughly one in ten arises as a new (de novo) mutation.^[12]

Two classic patterns are recognised.^[14]

• Romano-Ward syndrome is the common form, with normal hearing. It is inherited as an autosomal dominant trait: one altered copy of the gene is enough, and a child of an affected parent has a one in two chance of inheriting it.^[3][12]
• Jervell and Lange-Nielsen syndrome (JLNS) combines marked QT prolongation with congenital sensorineural deafness. It is autosomal recessive: both copies of the gene are altered, either the same change from each parent (homozygous) or two different changes (compound heterozygous). Most mutations (90.5%) are in KCNQ1, which makes IKs; the rest are in KCNE1, which encodes a small subunit of the same channel.^[3][10] The KCNQ1/KCNE1 channel generates IKs, which is needed both for the cardiac action potential and for endolymph production in the inner ear, which explains the pairing of long QT and deafness.^[25]

LQTS shows incomplete penetrance: some people with the gene change have a normal QT. About 25% of the at-risk population have a normal resting QTc despite a confirmed gene change ("concealed" LQTS).^[3] This is why a normal ECG in a sibling does not prove they are unaffected.^[14]

At least 17 genes have been linked to LQTS, though only seven have strong evidence of causing it.^[3] KCNQ1, KCNH2 and SCN5A account for about 75% of clinically definite cases and about 90% of those with a positive genotype.^[2][3]

How to reason it out instead of memorising it: in LQT1 the slow potassium current that normally speeds up repolarisation during adrenergic stimulation is missing, so exercise and emotion, which release adrenaline, cannot shorten the QT and the heart is exposed. In LQT3 the excess late sodium current is greatest when the heart rate is slow, so rest and sleep are the danger periods and a sodium blocker is the logical antidote.^[2][3] In one cohort 62% of LQT1 events occurred during exercise and 3% at rest or asleep, against 13% and 39% in LQT3.^[7] Low potassium is particularly dangerous in LQT2.^[3]

Genotype also predicts the course. In 647 genotyped patients, the chance of a first faint, arrest or sudden death before age 40 was 30% in LQT1, 46% in LQT2 and 42% in LQT3, and both the gene and the QTc (but not sex overall) independently predicted events.^[8] In the largest study of event rates on treatment, a QTc above 500 ms on therapy and a first cardiac event before age 7 each raised the risk of further events (relative risks 2.0 and 4.3).^[9]

• Andersen-Tawil syndrome (LQT7, KCNJ2) combines ventricular arrhythmias with episodes of periodic muscle weakness and facial and bodily dysmorphism.^[3]
• Timothy syndrome (LQT8, CACNA1C) includes congenital heart malformations, fused fingers or toes, autism spectrum features and dysmorphism; it is one of the most malignant forms.^[3]
• LQT4 (ANK2), LQT5 (KCNE1) and LQT6 (KCNE2) are each rare (below 3% of cases), and the KCNE1 form can also cause JLNS when both copies are affected.^[3]
• JLNS is the most severe common variant. In a study of 186 patients, 86% had cardiac events, half were symptomatic by age 3, the mean QTc was 557 ms, and beta-blockers had only partial effect: 51% had events despite treatment and 27% had a cardiac arrest or sudden death. Higher-risk features were a QTc above 550 ms and a faint in the first year of life; females and carriers of KCNE1 mutations did better. The authors advised considering early ICD implantation.^[10]

Acquired LQTS is a reversible prolongation of the QT after an exposure (a stressor), returning to normal after the stressor is removed; drugs are the commonest stressor.^[4] This section covers the mechanisms, the causes and the risk factors.

The defining mechanism is block of the IKr channel (KCNH2/hERG) by a drug. This channel is unusually easy to block because its central cavity contains aromatic amino acids that bind many drugs, and it lacks the two proline residues that restrict drug access in other potassium channels.^[4] That is why such varied drugs (antibiotics, antipsychotics, antiemetics) share the side effect.

The idea of repolarisation reserve explains why most people on such drugs never get torsades. The heart repolarises through several redundant mechanisms, and a single hit is usually tolerated; several hits at once (a drug, low potassium, a slow heart rate, a female sex hormone milieu, a silent gene variant) exhaust the reserve and allow torsades.^[4] In about 100 patients with drug-induced LQTS, 5 to 10% carried a congenital LQTS gene mutation, so a drug reaction can be the first sign of an unrecognised inherited disease.^[4] Relatives of people who had torsades with a drug show more repolarisation abnormality than relatives of people who tolerated the same drug.^[4]

Further points to remember about drugs.

• There is no safe prediction for an individual. In reviews of acquired LQTS, most patients who developed drug-induced torsades had at least one other risk factor besides the drug.^[4]
• Class IA drugs cause torsades at low or even subtherapeutic blood levels, whereas for most other drugs the risk rises with the dose.^[4]
• Amiodarone prolongs the QT without much torsades risk (below 1% at usual doses), in contrast to sotalol (0.8 to 3.8%) and ibutilide (3.6 to 8.3%).^[4] A possible explanation is that amiodarone prolongs repolarisation evenly across the wall, which reduces the dispersion that fuels re-entry.^[4]
• Interactions often act by raising drug levels, for example through inhibitors of the liver enzyme CYP3A4 (clarithromycin, erythromycin, diltiazem, verapamil, grapefruit juice and others).^[4]
• Rapid infusion of a QT-prolonging drug raises the risk.^[4]

Female sex (torsades is two to three times commoner in women), low potassium, bradycardia, recent conversion from atrial fibrillation (the QT effect of dofetilide was much greater after cardioversion), heart failure, left ventricular hypertrophy, high drug concentrations, a prolonged baseline QT, a silent congenital gene variant and severe low magnesium.^[4] A 2025 consensus statement adds age over 65, a history of syncope or torsades, renal or liver failure and use of diuretics.^[16]

Low extracellular potassium paradoxically reduces IKr and also increases the block of IKr by drugs, so hypokalaemia is important in two ways at once. Pauses, especially after an ectopic beat, precipitate drug-induced torsades, and in Holter recordings the sinus rate often rose in the minutes before an event, suggesting that a pause during heightened sympathetic activity is especially dangerous.^[4]

The first step in any patient with long QT is a drug history, including over-the-counter and herbal products. CredibleMeds (QTdrugs) is a free-to-search database maintained by AZCERT that lists drugs by their risk of torsades and also provides a list of drugs to avoid in congenital LQTS.^[21][3] Check it before prescribing any new drug to someone with a long QT or a family history of LQTS.

Torsades de pointes (French for "twisting of the points") is the characteristic arrhythmia of long QT. This section covers how it looks, how it starts and why it is sometimes self-limiting.

It is a polymorphic ventricular tachycardia: the QRS complexes vary in shape and in height, twisting around the baseline in a sinusoidal pattern, in a patient whose baseline QT is prolonged.^[4][1] Dessertenne described it in 1966 in an elderly woman with complete heart block and fainting spells.^[4] It is usually self-limiting, which is why the patient faints and recovers (and may be thought to have had a fit), but it can degenerate into ventricular fibrillation.^[4][3]

Torsades is typically preceded by a "short-long-short" pattern. A premature ventricular beat (short coupling) is followed by a compensatory pause (long). The next sinus beat has an especially long QT and often a deformed T or U wave. A further premature ventricular beat then lands on that vulnerable T wave and sets off the tachycardia.^[4] The pause lengthens the action potential of the next beat, and the longer the action potential the more likely an EAD. That is why drugs that slow the heart, bradycardia and pauses all promote torsades, and why increasing the heart rate prevents it.

Other warning signs on the ECG include T-wave alternans (beat-to-beat variation in T-wave shape or polarity), giant T-U waves, and a slow upstroke of the QRS in the triggering premature beat.^[4] T-wave alternans is regarded as a direct sign of electrical instability.^[3]

Torsades is a polymorphic ventricular tachycardia; ordinary monomorphic ventricular tachycardia and ventricular fibrillation are covered in ventricular arrhythmias, and the general approach to a fast rhythm in approach to tachyarrhythmias. The key discriminator is the QT interval in the sinus beats before or after the run. If the baseline QT is long, think torsades and treat it as below. Class III antiarrhythmics such as sotalol can make torsades worse, and the 2024 European Heart Rhythm Association (EHRA) consensus says class III drugs should be avoided in an LQTS storm.^[15]

LQTS can be silent for years or announce itself with a catastrophe. This section covers the presentations and how to recognise them.

• Syncope is the most typical symptom, often during exercise, emotion, a sudden noise, or in a swimming pool, depending on genotype (see the table above). Palpitations may precede it.^[3]
• Seizure-like events. Brief loss of consciousness from torsades can include jerking movements. Epilepsy is misdiagnosed in up to one in five people referred with epilepsy, and one published LQTS case went undiagnosed in an epilepsy clinic for 15 years until a near-fatal arrhythmia revealed it.^[20] Any "epilepsy" that begins with exertion, emotion or a sudden noise, does not respond to treatment, or runs in a family with sudden deaths needs a 12-lead ECG with a careful QTc.
• Cardiac arrest or sudden death. Sudden cardiac death is the first presentation in around 12% of cases.^[3] Congenital LQTS is one of the leading causes of sudden death under the age of 20.^[3] The untreated symptomatic patient is at high risk: annual sudden-death risk is about 5% and 10-year mortality about 50% in symptomatic patients with congenital LQTS (figures quoted by the commentary on the 2022 guideline).^[2]
• Family history of faints, cot deaths, drownings, unexplained single-vehicle accidents or sudden death before age 30. LQTS mutations account for about 10% of sudden infant death syndrome cases.^[3]
• Age. The first symptoms most often appear around 11 to 12 years, though events can occur from birth.^[3]
• Extra-cardiac clues: deafness (JLNS), periodic muscle weakness (Andersen-Tawil), syndactyly and developmental features (Timothy).^[3]
• Acquired LQTS is silent until torsades occurs, often in a hospital patient on several drugs with a low potassium. The commonest warning is a long QTc on a routine ECG.

Epilepsy and reflex (vasovagal) faints are far commoner than LQTS and are separated by triggers, family history and the ECG (see clinical features); if in doubt, check the QTc. Other mimics follow.

• Acquired QT prolongation in a person with a normal gene. The history, drug list and electrolytes decide; a normal QT after correction points to an acquired cause, but a silent congenital variant may remain.^[4]
• Brugada syndrome overlap. A few SCN5A variants produce both a long QT and ST elevation in V1 to V3, which matters when choosing a sodium-channel blocker.^[3]
• Catecholaminergic polymorphic ventricular tachycardia (CPVT). This inherited arrhythmia also causes exertional faints but has a normal QT; it is triggered by adrenaline and managed by an inherited-arrhythmia service (see ventricular arrhythmias).^[15]

The aims are to confirm the diagnosis, identify the genotype, define the risk and screen the family. This section covers the tests in order.

1. Resting 12-lead ECG, repeated. Measure the QTc carefully (above). Look at the T-wave shape (broad, notched, late), the heart rate (sinus bradycardia for age is a minor clue), T-wave alternans and any ventricular ectopy. Record previous ECGs.
2. Exclude secondary causes before diagnosing the inherited form: drugs, electrolytes, bradycardia and acquired cardiac conditions.^[3] Check serum potassium, magnesium and calcium, thyroid function if suspected and other relevant blood tests.
3. Provocation tests that record the QT response. Rising quickly from lying to standing provokes an abnormal lack of QT shortening in LQTS and helps the diagnosis.^[3] An exercise test is also helpful: a QTc of 480 ms or more in the fourth minute of recovery scores a point, and the QT behaviour during exercise differs by genotype (it fails to shorten in LQT1).^[3] The adrenaline (epinephrine) test is not recommended routinely, because of its false-positive rate and the usefulness of exercise testing.^[1][3]
4. Ambulatory (Holter) ECG records spontaneous ventricular ectopy, bradycardia and T-wave alternans, and signs of electrical instability on 24-hour recordings help to guide the choice of therapy.^[14]
5. Echocardiogram to exclude structural heart disease (also relevant to acquired LQTS).
6. Genetic testing and counselling are recommended in everyone with a clinical diagnosis of LQTS (Class I, level C).^[2] The yield is about 75%: in the remaining 20 to 25% no variant is found, and this does not mean milder disease.^[3] A negative test does not exclude LQTS in someone who meets the clinical criteria.^[12] Finding the family variant allows cascade testing of relatives (see the end of this page).

The score combines ECG, history and genetics into a probability. The version in the 2022 ESC guideline gives LQTS when the total is above 3.^[3][2] Earlier versions included congenital deafness; the modified score omits it and adds the genetic finding.^[3]

A pathogenic variant or a QTc of 480 ms or more scores above 3 on its own, which is why either can establish the diagnosis independently.^[2]

The aim is to stop torsades from ever starting. This section covers lifestyle, drugs to avoid, specific drug therapy, the role of the ICD and sympathetic denervation, and risk stratification. Management is led by a specialist in inherited arrhythmia.

• Avoid QT-prolonging drugs, using the CredibleMeds list of drugs to avoid in congenital LQTS (Class I, level C).^[3][21]
• Avoid and correct electrolyte disturbances, for instance during diarrhoea, vomiting, restrictive diets or bulimia (Class I, level C).^[3] Hypokalaemia is particularly dangerous in LQT2, and potassium supplements, with or without a potassium-sparing diuretic, are reasonable.^[14][3]
• Avoid genotype-specific triggers. In LQT1: strenuous exercise and competitive sport, and in particular swimming (99% of swimming-related episodes in one review occurred in LQT1).^[14] In LQT2: remove telephones and alarm clocks from the bedroom and wake children gently.^[14]

The cornerstone of treatment is a non-selective beta-blocker, ideally nadolol or propranolol, for everyone with a documented long QT (Class I, level B).^[2] They reduce the effects of adrenaline and prevent EADs, and are more effective than beta-1-selective drugs such as metoprolol.^[12][15] In the 2022 guideline a beta-blocker is also reasonable (Class IIa) for a pathogenic variant carrier with a normal QTc, because of the 'concealed' risk.^[3] In the Italian genotyped cohort, LQT1 patients responded best (events in 10%, versus 23% for LQT2 and 32% for LQT3).^[9]

Dosing. No trial has identified the best dose, so the practice is the full dose the patient tolerates, set by a specialist; check the local formulary.^[3]

• Children: propranolol 2 to 3 mg/kg/day (higher in very severe cases) or nadolol 1 to 1.5 mg/kg/day in one or two doses.^[3]
• Adults: weight-based, nadolol 0.75 to 2 mg/kg/day or propranolol 2 to 3 mg/kg/day. Below 0.75 mg/kg/day of nadolol or 2 mg/kg/day of propranolol is regarded as submaximal.^[23]

Practical points.

• Nadolol is mainly cleared by the kidney, so renal function matters; propranolol is lipophilic, crosses into the brain and is associated with more central nervous system side effects. The more favourable pharmacokinetics of nadolol are suggested to improve adherence, which may partly explain why it performs better in some comparisons.^[16]
• Never stop a beta-blocker suddenly: abrupt withdrawal raises the risk of events.^[3]
• Active asthma is the commonest contraindication, though treatment is usually tolerated.^[3]
• In patients whose QT lengthens at slow rates (bradycardia-dependent), a beta-blocker may be harmful or insufficient and mexiletine is an alternative.^[2]

Mexiletine, a class IB sodium-channel blocker that shortens the QT by blocking the late sodium current, is indicated for LQT3 patients with a prolonged QT (Class I, level C).^[2] A response test is advised: give an oral dose and check that the QTc shortens by 40 ms or more before committing to long-term use.^[2][3] The guideline does not say whether mexiletine should be used alone or added to a beta-blocker; in practice it is usually added.^[2] It also shortened the QTc in LQT2 patients in small studies.^[2] The general oral dose for ventricular arrhythmia is 200 to 300 mg every 8 hours (maximum 1.2 g/day, plasma level 0.8 to 2.0 micrograms per mL).^[16] At high doses mexiletine can block the peak sodium current, with a risk of bradycardia or arrhythmia, so use the lowest effective dose and be cautious in critical illness.^[2] In patients with an overlap of LQT3 and Brugada features, mexiletine does not provoke ST elevation, unlike flecainide, which is itself on the list of drugs to avoid in congenital LQTS.^[3] Ranolazine has also been used as a late sodium blocker in some LQT3 patients.^[12]

LCSD removes the first three or four thoracic sympathetic ganglia on the left, reducing adrenaline release to the heart.^[3] The 2022 ESC guideline indicates it (Class I, level C) in symptomatic LQTS when an ICD is contraindicated or declined, or when a patient with an ICD on beta-blocker and genotype-specific therapy has multiple shocks or faints due to ventricular arrhythmia.^[2][3] It is also useful for breakthrough events on treatment (Class IIa), and when beta-blockers are not tolerated.^[3] The evidence is from small studies; it is not an alternative to an ICD in people at high risk.^[2]

An ICD terminates a lethal arrhythmia but does not prevent it. The recommendations are the following.^[2][3]

• Cardiac arrest survivors: ICD in addition to beta-blocker (Class I, level B), because the risk of recurrence is high (14% within 5 years even on treatment).^[3]
• Symptomatic despite beta-blocker and genotype-specific therapy: ICD (Class I, level C).
• High-risk asymptomatic patients: an ICD may be considered in addition to genotype-specific medical therapy (Class IIb, level B).
• Asymptomatic patients who have not yet had a beta-blocker: an ICD is generally not advised, except in special circumstances (Class III in the 2013 HRS/EHRA/APHRS consensus, as cited in the review, and in the 2021 paediatric device consensus, Class III, level C).^[3][24] The 2022 ESC guideline instead allows an ICD to be considered (Class IIb) in asymptomatic patients judged high risk, for example by the 1-2-3-LQTS-Risk calculator.^[3]
• A shock itself can trigger an adrenaline surge and electrical storm, so an ICD is combined with beta-blockade. In children there are risks of inappropriate shocks, infection and lead problems with growth; epicardial systems are preferred in small children, transvenous or subcutaneous systems in older or larger ones.^[2][3]

The factors that identify those at higher risk are:^[3][8][9]

• QTc above 500 ms (extremely high above 600 ms);^[3]
• two pathogenic variants (as in JLNS) or Timothy syndrome;
• obvious T-wave alternans;
• a faint or arrest before age 7, or in the first year of life;
• events despite full medical therapy;
• genotype: LQT2 and LQT3 carry more events than LQT1 in untreated and treated cohorts;
• sex, where the effect varies by age and genotype.

With correct diagnosis and treatment the outlook in most people is good; Timothy syndrome, JLNS with KCNQ1 mutations and LQT3 with 2:1 AV block and very early arrhythmias do worse.^[14] Untreated symptomatic disease carries about a 5% annual risk of sudden death.^[2] Treatment has its own complications: beta-blocker side effects and, for ICDs, inappropriate shocks, infection and lead problems in growing children.^[3] Acquired LQTS recovers when the cause is removed, but the patient stays at risk if re-exposed.

A "1-2-3 LQTS risk calculator" has been proposed to weigh arrhythmic risk against the burden of an ICD.^[3] In silent carriers with a normal QTc, aborted arrest or sudden death was less common than in those with a long QT (4% versus 15%) but still higher than in controls, and a transmembrane missense variant and a LQT1 or LQT3 genotype raised the risk.^[3]

This is the section to know by heart. Torsades is an emergency, but most episodes stop on their own, so treatment is aimed at ending an ongoing run and stopping it from coming back.^[4]

1. Unstable or pulseless: follow the advanced life support algorithm and defibrillate (a sustained run needs immediate cardioversion).^[4] For the wider approach to shockable rhythms, see ventricular arrhythmias.
2. Stop every QT-prolonging drug and look for the trigger.^[4]
3. Give intravenous magnesium sulphate, even if the serum magnesium is normal.^[1][4] Give 1 to 2 g (2 g is usual) over about 5 minutes.^[4][16] Published maintenance infusions differ widely: 2 to 4 mg per minute in one review,^[4] 3 to 20 mg per minute for 7 to 48 hours in the original case series,^[22] and 0.5 to 1 g per hour in the 2025 EHRA compendium.^[16] There is no agreed rate, so use your local protocol and check your formulary. The evidence is small uncontrolled series, not randomised trials.^[22] Flushing is the commonly reported effect of the bolus.^[4]
4. Correct potassium to the high normal range;^[4] a figure often quoted for congenital LQTS is 4.0 mmol/L or more, though no trial defines the target.^[22] If it is low, the EHRA storm consensus gives oral potassium chloride 40 to 100 mmol, or intravenous potassium chloride 20 to 40 mmol in 500 mL saline at no more than 20 mmol/h (faster only in an emergency, through a central line). Correct low calcium too.^[15][16]
5. Raise the heart rate if torsades is pause- or bradycardia-dependent or keeps recurring. Overdrive pacing, by a temporary transvenous wire, at 90 to 110 beats per minute, shortens the QT and prevents pauses; it is highly effective.^[4] Isoprenaline can be used while a pacing wire is being placed (the compendium gives 10 micrograms as a bolus, then 2 to 20 micrograms per minute),^[16] and the ESC advises either approach for refractory recurrent torsades in acquired LQTS.^[1] Permanent pacing may be indicated when sick sinus syndrome or AV block with bradycardia underlies it.^[4]
6. Avoid antiarrhythmics that prolong the QT. Class III drugs should be avoided.^[15] Lidocaine, phenytoin and atropine have been reported to help, but their effectiveness is uncertain.^[4]
7. Treat the cause: thyroid hormone for severe hypothyroidism, correction of bradycardia and the electrolyte or nutritional disorder. In one case of torsades from hypothyroid crisis, pacing at 90 per minute abolished the arrhythmia until thyroxine replacement took effect.^[18]

• Acquired LQTS: no sport until the cause is corrected.^[3]
• Congenital LQTS with symptoms: disqualification from competitive sport.^[3]
• Genotype-positive with a normal QT: an individual decision, in a shared discussion of the risks (Class IIa, level C); a normal exercise test does not predict safety.^[3]
• QTc above 500 ms, or above 470 ms in men and 480 ms in women with genetically confirmed LQTS: high-intensity sport is not recommended even on a beta-blocker. After cardiac arrest or arrhythmic syncope, competitive sport is not recommended.^[3]
• Athletes can have a slow rate that makes Bazett under-read, and intense training can produce repolarisation changes that look like LQTS but regress after 3 to 4 months of detraining.^[3]
• Swimming is particularly dangerous in LQT1.^[14]

Hormones and heart rate change risk across pregnancy. In 391 women with LQTS, cardiac events (faint, aborted arrest, sudden death) were less frequent during pregnancy (hazard ratio 0.28 compared with before conception) but more frequent in the 9 months after delivery (hazard ratio 2.7), especially in LQT2; after 9 months the risk returned to baseline.^[11] Beta-blockers reduced events in the postpartum period (hazard ratio 0.34), and a later review states a fall in major events from 3.7% to 0.8% in that period with treatment.^[11][12] Pregnancy is generally safe and not contraindicated.^[12]

Practical advice.

• Plan beforehand: a multidisciplinary team (cardiologist with inherited-arrhythmia expertise, obstetrician, anaesthetist, geneticist); ECG and echocardiogram before conception; ICD implantation can be considered before pregnancy in women at high risk.^[12]
• Continue the beta-blocker through pregnancy and for at least 40 weeks after delivery (Class I, level C in the AHA/ACC/HRS and ESC pregnancy guidance cited by the review).^[12] Non-selective propranolol and nadolol were more effective than metoprolol.^[12] Beta-blockers may be associated with lower birth weight; the review reports no clear excess of malformations or miscarriage, and the drugs are secreted in breast milk with a low risk of adverse effects in babies with normal kidney and liver function.^[12]
• Avoid amiodarone (further QT prolongation; harm to the fetus).^[12]
• Check electrolytes in hyperemesis, and avoid QT-prolonging antiemetics such as ondansetron.^[12][3]
• Labour: vaginal delivery is considered safest when there is no obstetric reason against it, with continuous ECG monitoring and a defibrillator available in higher-risk women; Caesarean delivery is advised for high-risk LQTS. Combined spinal-epidural anaesthesia limits sympathetic surges; single-shot spinal anaesthesia is avoided because of abrupt haemodynamic changes.^[12]
• After delivery: early cardiology review in the first weeks and monthly review for the first 9 months, with attention to sleep disturbance, noise and the dose of beta-blocker.^[12]

• LQTS is a leading cause of sudden death under 20, and events are triggered by swimming (LQT1) or noises (LQT2).^[3]
• Dosing is by weight (see above). A syncope or arrest before age 7 or in infancy marks a higher-risk child.^[3][9]
• Exercise, school sport, swimming and wake-up alarms should be reviewed with the family; teachers and coaches should know what to do if the child collapses.
• Children share the same drug-avoidance rules, and the commonest culprits in children are macrolides, fluoroquinolones and azole antifungals; ondansetron and domperidone are also on the list.^[3]
• Newborn ECG screening can detect LQTS, but is not routine; the QT peaks at 6 to 11 weeks of age, and the Bazett correction is unreliable at the fast heart rates of infants.^[3][13]
• In JLNS (deafness) early ICD is considered because beta-blockers alone are often insufficient.^[10]

• The proband (first person identified) should be genotyped whenever possible, because success allows rapid screening of relatives and finds the roughly one in four carriers who have a normal QT and are still at risk if undiagnosed.^[3]
• Cascade screening: first-degree relatives are offered a clinical assessment (history, 12-lead ECG, echocardiogram) and testing for the family variant. After an unexplained sudden death in a family, all first-degree relatives, including children, should be evaluated, bearing in mind that a normal ECG does not exclude LQTS.^[3]
• Inheritance counselling: in the autosomal dominant forms each child of an affected parent has a one in two chance of inheriting the variant; in JLNS both parents of an affected child are carriers (each with one altered copy) and each further pregnancy carries a one in four risk of an affected child. Prenatal and reproductive options should be discussed. A new mutation (about 1 in 10) can occur with no family history.^[12]
• Result interpretation: a variant in KCNQ1, KCNH2 or SCN5A that is known to be pathogenic establishes the diagnosis whatever the QT.^[2] A negative test does not rule LQTS out clinically.^[12] Variants of uncertain significance need specialist interpretation.
• Practical advice to carriers: carry a card naming the condition and the drugs to avoid, and tell every prescriber and pharmacist.

Whether someone with long QT syndrome may drive depends on whether the rhythm has caused, or is likely to cause, sudden incapacity, and on the national licensing authority. Advise the patient not to drive after any collapse or near-collapse until the cause is assessed and treated, and tell them to check the rules that apply to them. The UK DVLA rules for arrhythmia are set out in Supraventricular tachycardias (Driving section) and apply in the same way here.