Your Patient's QTc Came Back at 480 — Do You Stop the Antipsychotic?

The message arrives from the medical team: QTc 512, please stop the quetiapine. Or it's 480 on a routine ECG in clinic, and the reflex is to hold the drug until someone works out whether it's safe. Both instincts feel responsible, and both usually reach for the one thing on the list that is hardest to replace. The antipsychotic is rarely the most modifiable contributor to a prolonged QTc — and there is an order in which to check things before you touch it.

🎯️ So — do I stop it?

At 480, in a patient without syncope and without other risk stacked on top: almost certainly not yet. Correct what's reversible first and repeat the ECG.

A QTc of 480 sits in the band where the number is abnormal but the absolute risk in an otherwise low-risk patient is modest. What that band calls for is a search — for hypokalemia, for the three other QT-prolonging drugs that arrived during a medical admission, for the interaction quietly raising the antipsychotic's blood level — not for discontinuation of a drug the patient may have been stable on for years.

Where 480 does change things is if it represents a jump. A rise of more than 60 ms from the patient's own baseline is a warning in its own right, even when the total looks acceptable. That is the single best argument for taking a baseline ECG before starting a high-liability agent: without it, you cannot tell a drug effect from a lifelong finding.

📏 Is 480 even 480?

Not necessarily. Verify the number manually before you make a decision on it — particularly if the patient was tachycardic.

Most ECG machines print a Bazett-corrected QTc. Bazett is reasonable at normal rates but over-corrects at tachycardia and under-corrects at bradycardia. An agitated patient with a heart rate of 110 can be handed a QTc that looks alarming and isn't. Fridericia's correction performs better when the rate sits well outside 60–100.

Then look at the tracing yourself: measure in lead II or V5, from the start of the QRS to the end of the T wave — where its steepest downslope crosses the isoelectric line — excluding U waves, and average across a few beats. In atrial fibrillation, average several. When the QRS is wide, from bundle branch block or pacing, the standard thresholds don't apply cleanly and the JT interval is the more meaningful measure.

🔍 If it isn't the antipsychotic, what is it?

Four things, and they're worth working through in this order — the first two are faster and safer to fix than any change to psychiatric treatment.

Step 3 is where most of the yield is, because these drugs arrive quietly and from other teams. Worth knowing the usual suspects by the situation that brings them:

That last one deserves more than a passing thought. A meaningful share of patients who develop drug-induced QT prolongation turn out to carry a long-QT gene variant, so a brief family screen — sudden death under 40, unexplained drowning, unexplained seizures, known long QT syndrome in relatives — is high-yield and takes thirty seconds.

⚖️ Which antipsychotics actually matter here?

Liability varies enormously — but the amount an agent prolongs QTc is not the same thing as its actual torsades risk, and that distinction matters more than the ranking.

Now the part that catches people out. Ziprasidone prolongs QTc measurably, yet is classified as only a conditional torsades risk — meaningful under specific circumstances such as overdose or interaction. Haloperidol prolongs less on average but carries a known torsades risk classification. Reasoning from milliseconds alone will mislead you. CredibleMeds maintains the authoritative free classification, and it's the single most useful tool for this problem — check it when you're stacking agents rather than relying on recall.

Droperidol deserves a footnote of its own. It sits in the notable tier on paper, but its 2001 black box is now widely regarded as disproportionate to its real-world risk, and it remains in routine emergency and anaesthetic use. Read its position on that table as historical as much as pharmacological.

Clozapine is its own argument. Its traditional "intermediate" label may understate things — one real-world cohort found it associated with a longer QTc than any other agent studied — but other analyses place it lower, and the question isn't settled. What isn't in doubt is that clozapine patients often carry several cardiac risk factors already.

🚨 When does it stop being a search and become urgent?

At 500 ms, at a rise of more than 60 ms from baseline, or at any syncope — and syncope outranks both numbers.

Above 500 ms, act: correct electrolytes urgently, review and remove other QT drugs, reduce or switch the antipsychotic, involve cardiology, repeat the ECG and monitor. But the symptom matters more than the number. Syncope or near-syncope in a patient on QT-prolonging drugs is a cardiac event until proven otherwise — not a faint, not dehydration, and not anxiety.

Torsades de pointes itself is a medical emergency: stop the offending drugs, give intravenous magnesium (effective even when serum magnesium is normal), correct potassium aggressively, manage bradycardia or pauses, defibrillate if it degenerates, and get cardiology and critical care involved immediately.

One differential worth holding, and the reason the new section has two chapters rather than one: collapse in a patient on an antipsychotic is not automatically arrhythmic. Orthostatic hypotension produces the same presentation and is managed in a completely different way — and it is far more common. If the patient stood up before they went down, and especially if their dose was recently increased or restarted after an interruption, you may be looking at the other chapter.

⚠️ What's the catch?

The thresholds are less settled than they look, and the relationship between a long QTc and dying is looser than the anxiety around it suggests.

The cut-points vary by source. The 2017 AHA ECG-monitoring standard gives 450 ms for men and 460 for women; the 2010 AHA/ACCF torsades statement uses 470 and 480 as 99th-percentile values; Schwartz's long-QT work puts the Bazett upper limits at 440 and 460. A QTc of 465 in a woman is a different finding depending on which document you open. That is not a reason to ignore the number — it's a reason to treat these as gradients rather than cliffs, and to weigh the whole clinical picture rather than a single value.

The deeper caveat is that QTc predicts risk only modestly at the individual level, and autopsy data suggest much of the excess mortality associated with QT-prolonging drugs may be driven in part by non-arrhythmic causes. In other words, a long QTc is often a marker of how unwell the patient is overall, rather than a direct route to a fatal arrhythmia. Nothing dramatic happens at exactly 500 ms, and torsades can occur below it — particularly in congenital long QT syndrome.

Which brings the argument back where it started: the number should inform your decision, not make it.