The Fine Print issue 1
Escitalopram’s QTc effect varies with who takes it
The BMJ headline is +8.7 ms. The number that should change how you prescribe is on page six, and it changes with your patient’s age.
Pick an age. How much does escitalopram add to the QTc?
Clearly above placebo. Schematic ECG illustration, not a patient tracing. Only the QT interval is to scale, starting from the trials’ average baseline QTcF of 403 ms (dashed). Estimates are modeled at four set ages with every other risk factor at its mean (Fig 2).
The verdict
No increase in early cardiac events was detected, though the trials were too small to exclude rare ones. Escitalopram lengthens the QTc on average, and the modeled effect is larger in older adults. That fits existing caution about escitalopram in older patients; the study did not test dose limits.
Three guesses before the data
Tap an answer; the explanation appears underneath.
Which of these raised the QTc most versus placebo, on average?
In a 30-year-old, is escitalopram’s QTc effect distinguishable from placebo?
Across 52,398 trial participants, did antidepressants raise early cardiac events or non-suicidal sudden death?
524 trials went in. 35 could answer the question.
Most antidepressant trials never recorded a usable QTc. The authors chased individual patient data from every eligible trial; only a fraction had ECG measurements they could analyze patient by patient.
For the average patient, two drugs move the needle
Hold every risk factor at its mean and only escitalopram and amitriptyline separate from placebo. Both effects are small. The other estimates were closer to zero or had intervals that included zero.
Swipe the chart sideways to see the values.
QTcF change vs placebo, ms (95% credibility interval), all covariates at their mean. Redrawn from Fig 1, Ostinelli et al, BMJ 2026.
Age flips the picture
The same drug can have little detectable effect in one modeled patient profile and a much larger one in another. Pick a drug.
Swipe the chart sideways to see every age.
Exact values for this drug
| Age | ms | 95% CrI | |
|---|---|---|---|
| 30 | 2.66 | −7.78 to 13.14 | crosses zero |
| 45 | 8.03 | 0.90 to 15.14 | clear |
| 60 | 13.40 | 3.84 to 22.91 | clear |
| 75 | 18.77 | 3.43 to 33.90 | clear |
| Age | ms | 95% CrI | |
|---|---|---|---|
| 30 | −1.01 | −6.11 to 4.16 | crosses zero |
| 45 | 2.22 | −1.48 to 5.98 | crosses zero |
| 60 | 5.45 | 0.65 to 10.30 | clear |
| 75 | 8.68 | 1.38 to 16.08 | clear |
| Age | ms | 95% CrI | |
|---|---|---|---|
| 30 | 6.90 | 1.11 to 12.65 | clear |
| 45 | 3.30 | −0.38 to 7.00 | crosses zero |
| 60 | −0.30 | −7.20 to 6.65 | crosses zero |
| 75 | −3.90 | −15.49 to 7.81 | crosses zero |
| Age | ms | 95% CrI | |
|---|---|---|---|
| 30 | 9.22 | 1.68 to 16.76 | clear |
| 45 | 5.76 | 1.35 to 10.15 | clear |
| 60 | 2.30 | −5.28 to 9.87 | crosses zero |
| 75 | −1.16 | −14.26 to 11.96 | crosses zero |
QTcF change vs placebo, ms, at ages 30, 45, 60 and 75 with other covariates at their mean. Hollow grey points have an interval that crosses zero. Axis capped at −16 to 35 ms. Redrawn from Fig 2, Ostinelli et al, BMJ 2026.
Across 216 modeled risk-factor profiles, escitalopram had the largest QTc rise in more than half
The authors simulated every drug in 216 combinations of sex, age, baseline QTcF, BMI and potassium. The pale bar is how often a drug raised QTcF at all; the dark bar is how often it caused the largest rise of all ten.
Swipe the chart sideways to see the values.
Percent of 216 covariate combinations. Rises are counted regardless of size, so trazodone’s 91.7% means frequent but small. These percentages do not represent the proportion of real-world patients who develop QTc prolongation. Redrawn from Table 1, Ostinelli et al, BMJ 2026.
The QTc changed. An increase in cardiac events was not detected.
- 9 of 34,575 on an antidepressant
- 3 of 17,823 on placebo
- 0.01% risk difference (95% CrI −0.01% to 0.02%)
- 0 torsades among the 8,679 patient-level participants
Only 7 of 8,679 people (0.08%) passed 500 ms, and 56 (0.65%) rose more than 60 ms. The authors acknowledge these short trials can’t exclude rare events. Citing external European data, they put drug-induced torsades at roughly 0.8 to 1.2 cases per million person-years.
The fine print
We read past the abstract. Four things in this paper don’t add up.
- Escitalopram is the thinnest arm.186 participants, 2.1% of the network. Hence the wide intervals.
- Citalopram and sertraline are absent.The paper can’t rank either.
- Dose wasn’t modeled.Nothing here about 10 mg versus 20 mg.
- Trial patients aren’t your patients.Mean baseline QTcF 403 ms, 1.8% with ischemic heart disease, up to 8 weeks of treatment.
- Funding.NIHR, with a grant from Angelini Pharma. Disclosed; the funder had no stated role in the analysis. Our note, not the paper’s: Angelini markets trazodone, one of the drugs compared.
Three patients on Monday
Healthy, no cardiac history
- At age 30 the modeled increase is +2.66 ms, with wide uncertainty and no clear difference from placebo.
- In a healthy young adult with no other QTc risk factors, a routine ECG is generally not needed just to start escitalopram. That is clinical judgment from broader guidance, not something this study tested.
- The usual checks stand: other QT drugs, whether an electrolyte problem is plausible, family history.
Stable, never had an ECG on treatment
- The age-75 estimate is +18.77 ms (3.43 to 33.90).
- 20 mg/day exceeds the US-labeled recommended geriatric dose of 10 mg/day and the UK MHRA maximum of 10 mg/day over 65. These are regulatory positions, not thresholds established by this study. Revisit the dose, get an ECG, check potassium and magnesium, and review every co-prescribed QT drug.
- Long-term stability calls for a careful, individual review, not an abrupt switch. If you reduce, taper gradually and weigh the ECG, prior response and relapse risk. A clinically significant QTc abnormality may need faster action.
Coronary artery disease, no arrhythmia history
- In coronary artery disease, sertraline is often the preferred SSRI on evidence outside this study, including SADHART. Sertraline was not in this analysis. Long-QT syndrome, prior torsades or bradyarrhythmia would change the picture.
- Don’t assume fluoxetine is QTc-neutral in older adults: the model estimated +5.45 ms at 60 and +8.68 ms at 75. These are profile estimates, not a prediction for this patient.
- Venlafaxine and vortioxetine often showed lower modeled QTcF across the risk-factor profiles. That does not make them cardioprotective or the safer cardiac choice. Venlafaxine’s higher toxicity in overdose comes from other evidence and isn’t addressed here.
Five things to take into clinic
- Escitalopram lengthens the QTc a little on average: +8.7 ms vs placebo.
- Age modifies the effect. Escitalopram’s modeled increase is uncertain at 30 and larger at 75 (+18.77 ms).
- The model suggests different age patterns. Larger effects for mirtazapine and amitriptyline in younger adults, and for fluoxetine in older adults.
- No increase in early cardiac events was detected in 52,398 trial participants, though the trials were too small and short for rare events.
- Weigh the drug and the patient together. The authors caution against relying only on broad “proarrhythmic or not” drug labels.
Ostinelli EG, Capocci S, Li Z, et al. QTc changes and early major adverse cardiovascular events associated with antidepressant treatment for major depressive disorder in adults: individual participant data network meta-regression of double blind randomised trials. BMJ 2026;394:e100600. doi:10.1136/bmj-2026-100600. Peer-reviewed; open access (CC BY 4.0). Figures are redrawn from published values.
For clinician education. It supports, and does not replace, clinical judgment, current prescribing information and local protocols. Psychiatry Education Forum Academy.
Responses