Pulmonary Embolism

In depth

The one-page summary above is what the PDF prints. These sections go further.

Pathophysiology: why the RV fails

Most emboli start in proximal leg or pelvic veins; ultrasound finds a DVT in 30–50% of patients with PE (ESC 2019). PA pressure rises once more than 30–50% of the pulmonary bed is occluded, and thromboxane A₂, serotonin and hypoxic vasoconstriction add to the block. A previously normal RV cannot generate a mean PA pressure above about 40 mm Hg (ESC 2019), so acute PE presents as RV failure rather than as very high PA pressures.

The dilated RV contracts longer and bows the septum into the LV in early diastole; right bundle branch block adds desynchrony. LV filling, output and BP fall, and the RV becomes ischemic (troponin release) and fails further. Inflammation can destabilize a patient again 24–48 h later (ESC 2019).

Hypoxemia comes from V/Q mismatch and low mixed venous saturation. About a third of patients shunt right to left through a patent foramen ovale, risking paradoxical embolism. Dead space widens the PaCO₂–EtCO₂ gap, PaCO₂ is usually low, and up to 40% have a normal SaO₂ (ESC 2019).

Risk stratification: ESC 2019 and the AHA/ACC 2026 categories

ESC 2019 first asks about hemodynamic instability: arrest, obstructive shock (SBP <90 mm Hg or pressors despite adequate filling, with hypoperfusion), or SBP <90 or a fall of 40 or more for over 15 min that arrhythmia, hypovolemia or sepsis does not explain. Stable patients are then graded by PESI or sPESI, RV imaging and troponin. RV dysfunction despite a low PESI still means intermediate risk (about four times the odds of early death).

The first AHA/ACC PE guideline (February 2026) retired massive and submassive. Its categories separate transient hypotension (D1) and normotensive shock (D2) and add a respiratory modifier.

PESI class I (65 points or less) carries 0–1.6% 30-day mortality and class V (over 125) 10.0–24.5%. RV/LV ≥1.0 on CT raises all-cause death (OR 2.5) and PE-related death (OR 5.0); a raised troponin carries an OR of 5.2 for death; high-sensitivity troponin T <14 pg/mL has a 98% negative predictive value for an adverse course; and lactate ≥2 mmol/L predicts complications (all ESC 2019). AHA/ACC 2026 makes lactate measurement Class 1 in categories C to E, and in C3 a MAP under 80 mm Hg flags patients who may need escalation (2a).

Diagnosis in the stable and the unstable patient

Symptoms are nonspecific, and 40% of patients have no risk factor; PE was found in 17% of patients admitted with syncope (ESC 2019).

In stable patients, testing follows pretest probability (two-level Wells: over 4 is PE likely). PERC applies only when gestalt probability is under 15%: age under 50, HR under 100, SaO₂ above 94%, and no hemoptysis, estrogen, prior VTE, unilateral leg swelling or recent surgery or trauma. In PROPER (2018), a PERC strategy cut CTPA use from 23% to 13% and missed 0.1%.

CTPA confirms PE with a segmental or more proximal defect when probability is intermediate or high; a normal CTPA with low or intermediate probability, or a normal perfusion scan, excludes it; and a proximal DVT on ultrasound is enough to treat (ESC 2019). V/Q suits contrast allergy, severe kidney disease and young women.

In the unstable patient, bedside echo asks two questions: is the RV overloaded (RV/LV >1.0, D-shaped septum, McConnell or 60/60 sign, right-heart thrombus), and does another cause explain the shock (tamponade, LV failure, dissection, hypovolemia)? Without RV overload, PE is practically excluded as the cause; clear overload with no other explanation justifies reperfusion if CTPA is not feasible (ESC 2019). In stable patients echo's negative predictive value is only 40–50%.

Anticoagulation: choice, dosing and timing

ESC 2019 prefers LMWH or fondaparinux (Class I, A) and keeps UFH for overt or imminent instability needing primary reperfusion, CrCl ≤30 and severe obesity. AHA/ACC 2026 recommends LMWH over UFH for categories C1 to E1 (Class 1, B-R), rates it reasonable when a catheter procedure is likely (2a) and prefers it after lysis or a procedure (Class 1); in E2 it allows LMWH or UFH, and UFH is usual when ECMO is likely. The 2019 PERT Consortium consensus favored UFH whenever advanced therapy was possible, so protocols vary.

Around alteplase, the label restarts anticoagulation once the aPTT or thrombin time is twice normal or less, and the PERT Consortium holds UFH during the infusion and restarts it without a bolus; ESC 2019 lets UFH run, and CMAJ 2026 continues LMWH on schedule. The order should say who restarts heparin and when.

ESC 2019 keeps intermediate-high patients on LMWH for 2–3 days, the window in which PEITHO's heparin-only patients decompensated (mean 1.8 days). Treatment lasts at least 3 months. AHA/ACC 2026 stops it after 3–6 months when a major reversible trigger caused the PE, extends it after a first PE without one (Class 1) or with a persistent risk factor (2a), and recommends half-dose apixaban or rivaroxaban for the extended phase (Class 1). IVC filters are for patients who cannot be anticoagulated or who recur despite it, never routine (ESC 2019; AHA/ACC 2026, 3: Harm).

Systemic thrombolysis: evidence, dosing and contraindications

Lysis clears clot, lowers PA pressure and shrinks the RV faster than heparin; benefit is greatest within 48 h of symptoms but persists to 6–14 days (ESC 2019). In a meta-analysis of 16 trials (Chatterjee 2014; 2,115 patients), lysis lowered mortality (2.17% vs 3.89%) but raised major bleeding (9.24% vs 3.42%) and intracranial hemorrhage (1.46% vs 0.19%).

PEITHO (2014) gave weight-based tenecteplase or placebo, both with heparin, to 1,005 normotensive patients with RV dysfunction and a positive troponin. Death or decompensation at 7 days fell from 5.6% to 2.6%, but 30-day mortality did not differ (2.4% vs 3.2%); major extracranial bleeding rose from 1.2% to 6.3% and stroke from 0.2% to 2.4%. Long-term mortality and CTEPH were unchanged (Konstantinides 2017), so full-dose lysis is kept for hypotension or rescue.

Lower doses are attractive but unproven for hard outcomes; AHA/ACC 2026 says they may be considered to reduce bleeding (2b). MOPETT (2013; 121 patients) found less pulmonary hypertension at 28 months with half-dose alteplase (16% vs 57%); STRATIFY (2026) found that 20 mg over 6 h cut clot more than heparin, equally IV or by ultrasound catheter. PEITHO-3 (reduced-dose alteplase vs placebo) is not yet published.

Relative contraindications (ESC 2019): TIA within 6 months, oral anticoagulation, pregnancy or first postpartum week, noncompressible punctures, traumatic CPR, SBP over 180, advanced liver disease, infective endocarditis, active peptic ulcer. In a dying patient they are weighed, not absolute.

In arrest from PE, ESC 2019 says to consider lysis and continue CPR for 60–90 min afterward. AHA 2025 rates lysis or surgical or catheter embolectomy reasonable for confirmed PE (2a), lysis possible for suspected PE (2b) and ECLS reasonable (2a), and calls the best CPR duration after a lytic unclear. Arrest doses vary by institution (CMAJ 2026): a 50 mg alteplase bolus is common, ESC lists the accelerated regimen, and tenecteplase can be pushed.

Catheter therapy, surgery and ECMO

Catheter-directed lysis (CDL) infuses low-dose alteplase into one or both pulmonary arteries, with or without ultrasound. In ULTIMA (2014; 59 patients), 10–20 mg over 15 h reduced the RV/LV ratio at 24 h by 0.30 vs 0.03 with heparin. HI-PEITHO (2026) was the first trial of CDL against anticoagulation alone with a clinical endpoint: in 544 patients with RV/LV ≥1.0, a positive troponin and at least two of SBP ≤110, HR ≥100 and RR >20, 7-day PE death, decompensation or recurrence was 4.0% vs 10.3%, with similar major bleeding and no intracranial hemorrhage. AHA/ACC 2026 keeps the standard 5–10 mg of alteplase per pulmonary artery (under 5 mg, 3: No Benefit) and recommends therapeutic LMWH or UFH, or low-dose UFH, during the infusion (Class 1).

Mechanical thrombectomy aspirates clot without a lytic. In PEERLESS (2025; 550 intermediate-risk patients), large-bore thrombectomy beat CDL on a win ratio of 5.01, driven by less deterioration or bailout (1.8% vs 5.4%) and less ICU use (41.6% vs 98.6%), with similar mortality and bleeding. In STORM-PE (2026; 100 patients), vacuum thrombectomy lowered RV/LV at 48 h more than anticoagulation (0.52 vs 0.24). Neither trial was powered for death.

Surgical embolectomy on bypass suits failed or contraindicated lysis (ESC 2019, Class I). In the STS database, in-hospital mortality was 12%, but 32% after preoperative arrest. AHA/ACC 2026 rates it reasonable in E1 but not over other therapies in E2 without mechanical support. For clot in transit in the right heart (C3 to E2), advanced therapy over anticoagulation alone is reasonable (2a).

VA-ECMO unloads the RV and bypasses the lungs while reperfusion is arranged. ESC 2019 calls stand-alone ECMO controversial and pairs it with embolectomy or catheter therapy (IIb); AHA/ACC 2026 rates it reasonable in E2 but calls added clot removal on ECMO of uncertain value (2b). Reported survival ranges from 36% to 95% (PERT Consortium 2019).

• After catheter therapy, check the access site and distal pulses with each set of vital signs.
• Label each catheter infusion at the pump and the hub; run alteplase and heparin on separate pumps.
• Report hemoptysis, hematoma, a falling hemoglobin or a new arrhythmia.

Supporting the failing RV: fluids, drugs, oxygen and the airway

ESC 2019 allows ≤500 mL over 15–30 min only when CVP is normal or low, as with a small collapsing IVC, and stops fluid once CVP is high, because volume overdistends the RV and worsens septal shift. A congested RV may improve with diuresis (Yuriditsky 2024).

Norepinephrine restores the coronary perfusion gradient without raising PVR (ESC 2019); above about 15 mcg/min it may raise PVR, so AHA/ACC 2026 suggests adding vasopressin or phenylephrine rather than pushing it higher. Dobutamine (2–20 mcg/kg/min, ESC 2019) can raise a low cardiac index once pressure holds but alone can drop BP; AHA/ACC 2026 allows it first in normotensive shock (D2). Milrinone usually needs a vasopressor. Inhaled nitric oxide or epoprostenol lowers PVR without hypotension (2b in C2 to E), but benefit in PE is unproven.

Give oxygen for SaO₂ <90% and prefer high-flow nasal cannula or NIV (ESC 2019; high-flow 2a in AHA/ACC 2026). AHA/ACC 2026 rates deep sedation and invasive ventilation harmful in C to E unless clinically needed (3: Harm). If intubation is needed:

1. Tell the PERT and ECMO teams and agree the reperfusion plan first.
2. Have pressors, inotropes or VA-ECMO ready (AHA/ACC 2026, Class 1): norepinephrine running, push-dose epinephrine drawn up and labeled.
3. Preoxygenate with high-flow nasal cannula or NIV; keep apnea short.
4. Most experienced operator; reduced-dose etomidate or ketamine (ketamine caused more collapse in the RSI trial).
5. VT about 6 mL/kg, plateau <30 cm H₂O, cautious PEEP (ESC 2019); avoid hypercapnia.

Nursing care by system, lysis safety and recovery

• Neuro: baseline exam before lysis; headache, vomiting, a new deficit or a BP surge is bleeding until CT says otherwise.
• Cardiovascular: continuous ECG (new RBBB, AF), arterial line, shock index, lactate trend; no bolus without a CVP or IVC check.
• Respiratory: SpO₂, rate, work of breathing and O₂ escalation; avoid sedatives that blunt drive.
• Hematology: aPTT or anti-Xa 6 h after heparin changes; platelets for HIT; CBC and fibrinogen after lysis.
• Renal: urine output and creatinine, since AKI marks category D2.
• Access: minimize punctures; avoid jugular and subclavian sticks during lysis (Activase label).
• Placement and transport: ICU or step-down during systemic or catheter lysis (Class 1); do not transfer category E before stabilizing, and do not delay IV lysis to transfer (AHA/ACC 2026, 3: Harm).
• Mobility: walk once anticoagulated and stable.

No PE guideline sets a vital-sign schedule after lysis; many units borrow the stroke schedule on the Neuro page (every 15 min for 2 h, every 30 min for 6 h, then hourly to 24 h). For bleeding: stop alteplase and heparin, call, send CBC, fibrinogen, aPTT and a crossmatch, and image the site. Give cryoprecipitate to fibrinogen ≥150 mg/dL, tranexamic acid 1 g IV and protamine for recent heparin (Bleeding, Reversal, HIT and DIC page). Back or flank pain with falling BP suggests retroperitoneal bleeding.

Before discharge, teach the anticoagulant schedule and bleeding signs. AHA/ACC 2026 asks for contact within about a week, a 3-month visit to decide duration, testing for chronic thromboembolic disease if dyspnea persists, and screening for anxiety, depression and PTSD.

Evidence and controversies

Unsettled: whether any advanced therapy lowers death or long-term disability in C3 and D (HI-PEITHO's benefit was prevented decompensation); catheter lysis versus thrombectomy versus low-dose IV lysis; arrest dosing; and validation of the new categories. PEITHO-3 and PEERLESS II had not been published by October 2026.

Exam pearls

• High risk (E) is hemodynamic: arrest, obstructive shock, or SBP <90 (or a fall of 40 or more) for over 15 min; clot size does not decide.
• Normotensive with RV dysfunction plus troponin is intermediate-high (C3): anticoagulate, monitor 2–3 days, rescue if it worsens.
• sPESI 0 is low risk and often goes home on a DOAC; any sPESI point carries about 11% 30-day mortality.
• Too unstable for CT: RV overload on bedside echo with no other cause justifies lysis.
• Alteplase for PE: 100 mg IV over 2 h; restart heparin at aPTT twice normal or less; CPR 60–90 min after a lytic in arrest.
• LMWH is the default (AHA/ACC 2026); UFH for arrest, refractory shock, ECMO or CrCl ≤30.
• Failing RV: small bolus only if CVP is low, norepinephrine first, avoid intubation.
• Headache or neuro change during lysis: stop the lytic and heparin, head CT, cryoprecipitate and tranexamic acid.
• ECG: sinus tachycardia most often (about 40%, ESC 2019); S1Q3T3, new RBBB and T inversion in V1–V4 mean RV strain.