Fragmented QRS on the admission ECG marks a worse pulmonary embolism

Design

Systematic review and random-effects meta-analysis of observational studies

Population

Five studies, 1,383 adults with confirmed pulmonary embolism, classified by fragmented QRS on admission ECG

Primary outcome

All-cause in-hospital and long-term mortality; cardiogenic shock as secondary

Effect

In-hospital mortality OR 3.40 (95% CI 1.72–6.71, I² 51%); long-term mortality OR 3.90 (1.80–8.47); cardiogenic shock OR 4.88 (1.70–14.01)

Five observational studies totalling 1,383 adults with confirmed pulmonary embolism were pooled to test whether fragmented QRS on the admission ECG predicts outcome. It was associated with in-hospital mortality (odds ratio 3.40, 95% CI 1.72–6.71, I² 51%), long-term mortality (3.90, 1.80–8.47, I² 47%) and cardiogenic shock (4.88, 1.70–14.01, I² 65%). No association was found with thrombolysis, embolectomy or mechanical ventilation.

The appeal is obvious: the ECG is already done, and fragmented QRS costs nothing to look for. The problems are equally clear and the authors state them. Every pooled estimate came from unadjusted effect sizes, so none of this accounts for the clinical severity that would also predict death. The funnel plot was asymmetric, suggesting publication bias, though with five studies no formal test is meaningful. And there is no standardised definition of fragmented QRS in this setting.

So what this supports is noticing the finding, not scoring with it. A patient with fragmented QRS is probably a sicker patient — but the validated tools, PESI and simplified PESI plus right ventricular assessment, are what should drive the disposition decision.

• Use PESI or simplified PESI with right ventricular assessment for risk stratification; fragmented QRS is not a validated addition.
• Note fragmented QRS on the admission ECG as a reason to look harder at the right ventricle, not as a score in itself.
• Assess right ventricular strain with echocardiography or CT rather than inferring it from the ECG.
• Be sceptical of any ECG sign whose pooled estimates are all unadjusted — severity confounds everything here.

Why it matters

An ECG sign that is free to look for is exactly the kind of marker that enters practice before it has been properly adjusted.

Don't overread it

These are unadjusted observational estimates with funnel plot asymmetry — the association may reflect severity rather than add to it.

The statistics, in plain English

Unadjusted odds ratios are the key limitation: they compare outcomes in patients with and without the sign, without accounting for the fact that sicker patients are more likely to have both the sign and the outcome. An odds ratio of 4.88 for cardiogenic shock with an interval from 1.70 to 14.01 is also extremely imprecise — the data are compatible with a modest association or an enormous one. I² of 51% to 65% means the studies disagree substantially.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.