A practical, evidence-based approach to pulmonary embolism diagnosis in the ED

Pulmonary embolism diagnosis in the emergency department starts with clinical probability and an assessment of stability. This adult-focused guide covers PERC, D-dimer, imaging and initial treatment decisions. Pregnancy needs a dedicated diagnostic pathway. Drug doses and local escalation arrangements should be checked against the relevant prescribing information and the treating service’s protocol.

The 2026 AHA/ACC multisociety guideline pushes a very important message:

PE is not a “CT first” diagnosis. It is a probability-first diagnosis.

That is the right ER mindset.

The first principle: before you diagnose PE, understand what PE does

A pulmonary embolus is not just “a clot in the lung.”

Physiologically, it does 3 main things:

That physiology explains the whole bedside spectrum:

• pleuritic pain and dyspnea, which do not reliably indicate clot size or clinical severity
• tachycardia and hypoxemia across a range of PE severities
• syncope, hypotension, lactate elevation, RV strain, and collapse in severe PE

So when I evaluate PE, I ask 2 questions at the same time:

That second question is where good emergency medicine matters.

Step 1: Decide whether this is a PE workup at all

Not every dyspneic or tachycardic patient needs a PE pathway.

Start with a real bedside assessment:

• sudden unexplained dyspnea
• pleuritic chest pain
• syncope
• unexplained tachycardia
• unexplained hypoxemia
• signs or symptoms of DVT
• recent surgery, immobility, travel, cancer, estrogen exposure, pregnancy/postpartum state, prior VTE

But do not stop thinking broadly. The patient may have:

• pneumonia
• ACS
• heart failure
• COPD/asthma
• pneumothorax
• pericardial tamponade
• sepsis
• anemia
• metabolic acidosis

Before ordering a D-dimer, ask whether the history, examination and competing diagnoses justify a PE workup.

Keep the differential open throughout the workup.

Step 2: First separate the sick patient from the stable patient

Before scores, before algorithms, before D-dimer:

Is the patient unstable?

Look at:

• blood pressure
• MAP
• mental status
• lactate
• work of breathing
• oxygen requirement
• signs of RV strain or shock
• whether this feels like impending decompensation

This matters because the stable and unstable PE workups are not the same.

If unstable

In the crashing patient, PE becomes a resuscitation diagnosis, not just a radiology diagnosis.

Shock with a pulse: support oxygenation and perfusion while arranging urgent definitive treatment. Avoid indiscriminate fluid loading: an overdistended RV can further reduce LV filling. Give fluid cautiously only when reduced preload is plausible, and reassess. Norepinephrine is commonly used for hypotensive shock. Intubation and positive-pressure ventilation can precipitate collapse; when clinically necessary, prepare hemodynamic support and an experienced airway team. ESC guidance.

Cardiac arrest is a separate pathway: continue high-quality CPR and activate the local PE-arrest response. AHA permits consideration of systemic fibrinolysis for suspected PE arrest; confirmed PE offers additional specialist reperfusion options. ECLS may be reasonable in selected suspected or confirmed cases. The optimal CPR duration after fibrinolysis is uncertain. AHA resuscitation guidance.

During arrest, RV dilation alone does not establish PE. Integrate the history and other findings, and do not prolong compression pauses for ultrasound. RCUK ultrasound guidance.

You may not have the luxury of a perfect outpatient-style algorithm. In that patient, bedside echo, shock physiology, ECG clues, ultrasound evidence of DVT, and overall context become much more important. The guideline is clear that echocardiography alone should not be used to confirm or exclude PE in routine diagnosis, but in hemodynamic compromise it becomes a critical tool for rapid physiologic assessment and treatment direction.

In plain ER language:

In stable patients, PE diagnosis is a probability pathway.
In unstable patients, PE diagnosis becomes a physiology-and-time pathway.

When imaging is delayed or unavailable, therapeutic anticoagulation may be beneficial in suspected acute PE with anticipated AHA/ACC severity Category C2 or higher and low bleeding risk. Diagnostic probability and clinical severity are different assessments; intermediate pretest probability alone is not this indication. AHA emergency physician guidance.

Step 3: Estimate pretest probability before ordering tests

This is the center of the whole strategy.

Use:

• Wells
• revised Geneva
• or honest clinical gestalt

The guideline allows all three. What matters is not which score makes you feel sophisticated. What matters is that you categorize the patient before testing.

I simplify it into 3 buckets:

• Low probability
• Intermediate probability
• High probability

Once you know the bucket, the next step becomes obvious.

Step 4: In low-risk patients, ask whether you can stop with PERC

If the patient is truly low risk, use PERC.

This is where many clinicians go wrong.

PERC is not a PE test.
PERC is a test-avoidance tool.

It is only for patients who are already low probability.

Use PERC only after identifying a low pretest probability of PE (usually a clinical estimate below 15%) in a hemodynamically stable adult ED population. All eight rule-out criteria must be met: age under 50 years; heart rate under 100/min; oxygen saturation at least 95%; no unilateral leg swelling; no hemoptysis; no previous DVT or PE; no estrogen use; and no surgery or trauma requiring hospitalization in the preceding four weeks. If all are met, further PE testing is usually unnecessary. Use a pregnancy-specific pathway during pregnancy. PROPER excluded patients already receiving anticoagulants; do not assume that its validation extends to this group.

PERC evidence: PROPER trial; 2026 PE guideline, Table 3.

That is powerful. It avoids unnecessary cascades.

Practical PERC message

Use PERC only when your honest bedside impression is:

“PE is possible, but already unlikely.”

Do not use PERC to talk yourself out of a PE you are actually worried about.

That is one of the most common cognitive errors in the ED.

Step 5: In low-to-intermediate risk patients, use D-dimer properly

D-dimer is useful, but only when you respect what it is.

It is a marker of fibrin formation and breakdown.
So it tells you that clot turnover is happening somewhere.

That gives it good sensitivity for PE.
But poor specificity.

D-dimer goes up in:

• infection
• inflammation
• recent surgery
• cancer
• trauma
• pregnancy
• aging
• hospitalization
• many other inflammatory states

So D-dimer is a rule-out test, not a rule-in test.

When it helps

It helps in:

• low-risk patients who are not PERC negative
• intermediate-risk patients

When it does not help

It does not help when the patient is high probability.

Once pretest probability is high, you should stop asking a screening blood test to rescue the situation. That patient needs imaging.

Step 6: Use age-adjusted D-dimer or YEARS to reduce unnecessary CT

This is one of the most practical updates from contemporary PE diagnosis.

In bedside terms:

Choose one validated diagnostic pathway before interpreting the result. Age-adjustment and YEARS are alternative approaches; do not switch between thresholds simply to obtain a negative result. These pathways require the correct pretest probability, assay and population, and should not be extrapolated casually to patients already receiving therapeutic anticoagulation. YEARS study; 2026 PE guideline, section 3.1.1.

That is good emergency medicine.

Pregnancy

Pregnancy is different.

The guideline says pregnancy-adapted YEARS may be reasonable in pregnant patients, and this is one of the most important updates because it helps reduce unnecessary chest imaging in a population where indiscriminate CT use is especially undesirable.

Pregnancy-adapted YEARS includes compression ultrasound when there are symptoms of DVT. A positive proximal-DVT study can establish the need for anticoagulation without chest imaging in an appropriate stable patient. If PE is not excluded by the complete pathway, obtain indicated chest imaging promptly. The pregnancy study does not validate extending this rule to the postpartum period. Pregnancy-adapted YEARS study.

Step 7: High probability means imaging, not more hedging

If the patient is high clinical probability, go to diagnostic imaging.

Do not hide behind D-dimer.

This is a major pitfall in real life. People sometimes order D-dimer because it feels faster, cleaner, or less confrontational than committing to CT. But that is not the right logic.

When the probability is high, the question is no longer:

“Can I cheaply screen this out?”

It becomes:

“Can I confirm or exclude this quickly enough to change outcome?”

That usually means imaging.

Step 8: Choose the right imaging test

CTPA

The guideline continues to favor CT pulmonary angiography as the preferred imaging test in most patients.

Why?

Because CTPA:

• is widely available
• has strong diagnostic performance
• can assess clot burden
• gives RV/LV information
• often identifies alternate diagnoses

That last point matters in the ED.
Sometimes the patient does not have PE — but CT still shows you pneumonia, malignancy, pleural disease, or another explanation.

V/Q scanning

Use V/Q when CTPA is not the right test, especially if contrast is a problem or CT is not feasible.

The guideline favors V/Q SPECT over planar V/Q when available because diagnostic yield is better.

Pregnancy

In pregnant patients with symptoms, suggestive YEARS criteria, and a normal chest x-ray, the guideline says low-radiation-dose CTPA is reasonable over low-dose perfusion scanning. In practice, this is still a center-dependent workflow issue, but the important point is that modern low-dose protocols have changed the old reflexive assumptions.

Step 9: Know what not to use as routine PE diagnostics

This is just as important as knowing what to order.

Echocardiography and Point-of-Care Ultrasound (POCUS)

Echo is not a routine rule-in or rule-out test for PE.

It is useful for:

• unstable patients
• assessing RV strain
• helping with risk stratification
• guiding urgency and escalation

But it should not be used as a stand-alone diagnostic substitute for proper imaging in the stable patient.

Leg ultrasound

Do not routinely add leg duplex solely to look for PE after an adequate negative CTPA or normal V/Q SPECT. A technically limited or nondiagnostic scan, new DVT symptoms, or a major mismatch between the clinical assessment and imaging requires reassessment with the radiologist or treating specialist; this is a different situation from routine add-on testing. 2026 PE guideline, section 3.1.2; ASH diagnostic guideline.

Leg ultrasound is useful when:

• the patient has leg symptoms
• confirming DVT would change management
• in pregnancy or other settings where a positive DVT result may spare chest imaging

That is a focused, purposeful use. Not fishing.

Step 10: Once PE is diagnosed, immediately switch from diagnosis to severity assessment

This is where the 2026 guideline is especially useful.

The 2026 guideline introduces AHA/ACC acute PE clinical categories A through E.

You do not have to memorize every label to use the concept well.

That is the real bedside upgrade.

A PE patient is not just “PE positive.”
He may be:

• low risk and suitable for discharge
• stable but vulnerable
• normotensive but already hypoperfusing
• or on the edge of collapse

A positive PE study is the start of severity assessment.

Choose the care setting by support needs, deterioration risk and available monitoring; reassess as physiology changes.

The practical ER severity frame

After diagnosis, I think in 4 bedside questions:

That is a very important update.

Step 11: Start anticoagulation early when appropriate

For confirmed PE, anticoagulation is the foundation unless there is a true contraindication.

For suspected PE awaiting delayed imaging, apply the Category C2-or-higher and low-bleeding-risk qualification described above. Do not use diagnostic probability alone as a substitute for severity assessment.

That is clinically important.

In real ED practice, the dangerous mistake is not only missing PE.
It is also recognizing likely PE but waiting too long to treat because the scan is delayed.

Anticoagulant selection requires a bleeding-risk assessment and review of renal and hepatic function, pregnancy or breastfeeding, interacting medicines and likely procedures. For confirmed PE, DOACs are preferred over vitamin K antagonists in eligible patients; this does not make every patient suitable for an oral agent or outpatient care. 2026 PE guideline, section 4.2.1.

Deterioration or shock requires immediate PERT or local specialist assessment for reperfusion and circulatory support. Treatment selection depends on bleeding risk, contraindications and available expertise; RV strain alone does not mandate thrombolysis. 2026 guideline.

The PEITHO trial illustrates the trade-off in normotensive PE with RV dysfunction and positive troponin: tenecteplase reduced hemodynamic decompensation but increased major bleeding and stroke, without a significant mortality difference. This does not support treating every patient with RV strain with thrombolysis. PEITHO trial.

What is genuinely updated in this guideline?

A few practical updates stand out.

1. More structured use of D-dimer strategies

Age-adjusted D-dimer and YEARS are reasonable imaging-sparing options in appropriately selected patients; neither replaces pretest probability assessment.

2. Pregnancy is no longer a diagnostic dead zone

The guideline incorporates pregnancy-adapted YEARS and modern imaging thinking rather than defaulting to vague fear of radiation.

3. A more nuanced severity model

Use the definitions in Step 10 when assigning a category.

4. More emphasis on RV, perfusion, and early deterioration

This is a big one. The guideline does not treat PE as binary. It pushes clinicians to identify who is heading toward trouble even before frank hypotension appears.

5. More explicit role for multidisciplinary care

In patients at increased risk of adverse outcomes, PERT involvement is recommended where available. For the ED, that means earlier coordination, not later rescue.

Pitfalls to avoid

These errors can distort the diagnostic and treatment decisions.

My practical ED approach

This is the bedside script.

That is how you evaluate PE like an emergency physician.

Not by memorizing boxes.
By understanding the physiology inside the boxes.

Bottom line

Suspected PE is not mainly a radiology problem. It is a bedside probability-and-physiology problem.

The right approach is:

• think of PE when the story fits
• assess stability first
• estimate pretest probability before ordering tests
• use PERC only in truly low-risk patients
• use D-dimer intelligently, not reflexively
• use YEARS or age-adjusted thresholds to reduce unnecessary imaging
• image high-probability patients
• do not misuse echo or ultrasound
• once PE is diagnosed, switch immediately to severity assessment
• treat early when appropriate, especially if physiology is deteriorating

That is the practical, evidence-based, physiology-based way to do it.

For confirmed PE, continue with pulmonary embolism severity, disposition and escalation. If discharging a suitable patient, confirm immediate access to anticoagulation, bleeding and deterioration advice, and clinical follow-up within one week. AHA discharge guidance.

Sources and further reading

• 2026 AHA/ACC multisociety acute pulmonary embolism guideline
• July 14, 2026 correction to the PE guideline
• PROPER randomized PERC trial
• ADJUST-PE prospective management study
• YEARS prospective multicentre cohort
• Pregnancy-adapted YEARS, ARTEMIS