2026 AHA/ACC Multi-Society Guideline for Acute Pulmonary Embolism

Authors, Journal, Affiliations, Type, DOI

• Authors: Creager MA, Barnes GD, Giri J, Mukherjee D, Jones WS, Burnett AE, Carman T, Casanegra AI, Castellucci LA, Clark SM, Cushman M, de Wit K, Eaves JM, Fang MC, Goldberg JB, Henkin S, Johnston-Cox H, Kadavath S, Kadian-Dodov D, Keeling WB, Klein AJP, Li J, McDaniel MC, Moores LK, Piazza G, Prenger KS, Pugliese SC, Ranade M, Rosovsky RP, Russo F, Secemsky EA, Sista AK, Tefera L, Weinberg I, Westafer LM, Young MN
• Journal: Circulation. 2026;153:e977–e1051
• Affiliations: AHA/ACC Joint Committee; co-developed with ACCP, ACEP, CHEST, SCAI, SHM, SIR, SVM, SVN
• Type: De novo clinical practice guideline; evidence review February 2024–October 2024; select key studies added to April 2025
• DOI: 10.1161/CIR.0000000000001415
• Recommendation grading: COR (Class of Recommendation) indicates strength of recommendation, encompassing estimated benefit in proportion to risk; LOE (Level of Evidence) rates the quality of supporting data by type, quantity, and consistency. Cited throughout this page as COR x/LOE y

Overview

This de novo AHA/ACC multi-society guideline provides comprehensive, evidence-based recommendations for the evaluation, management, and follow-up of adults (≥18 years) with acute pulmonary embolism (PE). Its central innovation is the introduction of the AHA/ACC Acute PE Clinical Categories (A–E) — a five-category classification that replaces prior risk schemes (AHA 2011 massive/submassive/low-risk; ESC 2019 four-tier) by integrating clinical, hemodynamic, biomarker, and imaging parameters to guide severity classification, prognosis, and therapeutic decision-making. The guideline covers the full continuum from diagnosis through post-acute follow-up including CTEPD surveillance.

Keywords

Acute pulmonary embolism · anticoagulant · diagnosis · chronic thromboembolic pulmonary hypertension · CTEPD · DOAC · heparin · risk assessment · risk stratification · thrombectomy · thrombolytic therapy · venous thromboembolism

Key Takeaways

Top Take-Home Messages

The guideline's own 10 designated take-home messages:

1. A new classification scheme is presented, Acute Pulmonary Embolism Clinical Categories, with 5 categories (A–E) and subcategories ranging from low to high risk, to enhance precision of severity classification, prognosis assessment, and evidence-based therapeutic decision-making
2. Patients who are asymptomatic (Category A) can safely be discharged home from the emergency room and do not need hospitalization
3. Early hospital discharge is generally recommended for symptomatic patients with a low clinical severity score (Category B)
4. Symptomatic patients with an elevated severity score (Category C), incipient cardiopulmonary failure (Category D), and cardiopulmonary failure with persistent hypotension (Category E) should be hospitalized to optimize treatment strategies
5. Advanced therapies — systemic thrombolysis, catheter-based thrombolysis, mechanical thrombectomy, surgical embolectomy — are reasonable in Category E1 and can be considered in Category D1–2
6. PE response teams (PERTs) are recommended to improve timeliness of care
7. For initial parenteral anticoagulation, LMWH is recommended over UFH
8. For patients eligible for oral anticoagulation, DOACs are recommended over VKAs unless contraindicated, to prevent recurrent VTE and reduce major bleeding
9. Continuing anticoagulation into the extended phase is recommended for a first acute PE without a major reversible risk factor, and for those with a persistent risk factor
10. Patients who have had acute PE should be asked about PE-related symptoms and functional limitations at every visit for at least 1 year, to screen for CTEPD or other causes of dyspnea and functional limitation

Definitions and Treatment Phases

• Guideline-directed medical therapy — encompasses clinical evaluation, diagnostic testing, and both pharmacological and procedural treatments; recommendations are limited to drugs, devices, and treatments approved for clinical use in the United States
• Initial Treatment Phase — the first 3 to 6 months of anticoagulation therapy after an acute PE
• Extended Treatment Phase — more than 6 months of anticoagulation therapy after an acute PE; "anticoagulation beyond the initial 3 to 6 months without an anticipated stop date"
• ACC/AHA Acute PE Clinical Categories — a system categorizing PE severity and prognosis by integrating clinical, laboratory, and imaging parameters

Evaluation and Diagnosis

Clinical Assessment

• Targeted history and physical exam recommended (COR 1/A) to establish pre-test probability using validated tools (Wells score, Revised Geneva Score, PERC)
• Age-adjusted D-dimer (age × 10 μg/L for FEU assays) safely excludes PE in patients with low/intermediate pre-test probability (<50%), with <2% missed VTE at 3 months (COR 2a/B-R)
• YEARS algorithm (D-dimer threshold 500 μg/L if ≥1 criterion; 1000 μg/L if none) useful to avoid imaging (COR 2a/B-R); pregnancy-adapted YEARS criteria reasonable in pregnant patients (COR 2b/B-NR)

Diagnostic Imaging

• CTPA recommended over V/Q scan as standard imaging modality for PE confirmation (COR 1/B-R); preferred for its accessibility, diagnostic performance, and ability to assess RV
• V/Q SPECT preferred over planar V/Q (COR 2a/B-R)
• In CTPA-confirmed PE, reporting the numerical RV/LV ratio (≥1.0 cut-point: sensitivity 85%, specificity 72%) is recommended for risk stratification (COR 1/B-R)
• Echocardiography is NOT recommended to confirm or refute PE diagnosis (COR 3:No Benefit) — most patients with acute PE will not show evidence of RV dysfunction, which is why it cannot serve as a diagnostic test. When TTE is obtained, parameters of RV dysfunction and strain should be reported as comprehensively as possible based on local expertise — all eight of: RV/LV end-diastolic ratio, RV end-diastolic diameter, TAPSE, estimated RVSP, McConnell's sign, tricuspid systolic velocity, paradoxical septal motion, and IVC respirophasic collapse (COR 1/B-NR). The cut-off for each is in Table 4, under RV Imaging for Risk Stratification below
• Venous duplex ultrasound not useful after negative CTPA or normal V/Q SPECT (COR 3:No Benefit)
• Routine CTV as adjunct to CTPA: not recommended (COR 3:No Benefit)

PE Outcomes Risk Stratification — AHA/ACC Acute PE Clinical Categories

The New 5-Category System

• Respiratory modifier (R) added when: O₂ requirement >6 L/min nasal cannula or nonrebreather, or RR >25 (Cat B–D); NIV/invasive ventilation required (Cat E-R)
• D1/D2 definitions (Figure 2 footnotes): D1 = systolic BP <90 mmHg or a decrease >40 mmHg lasting <15 min, or responding to IV fluids. D2 = that plus any marker of reduced perfusion/end-organ dysfunction — lactate >2 mmol/L; acute kidney injury; urine output <0.5 mL/kg/hr; mental status change; cardiac index <2.2 L/min/m²; mean arterial pressure <60 mmHg; or increased shock score/stage (SCAI stage, CPES score)
• Clinical severity score gate (Figure 2 footnote): Low = PESI ≤85 or sPESI = 0 or Bova ≤4; Elevated = PESI >85 or sPESI ≥1 or Bova >4. Bova is an accepted alternative to PESI/sPESI for this gate and is quietly absent from most summaries of the scheme — including, until this revision, this page

Clinical Risk Scores (§3.2.1 — Table 6)

The guideline tabulates six scores used in the acute care setting. All were developed from factors measured at presentation or diagnosis; the guideline notes as an "intriguing avenue for investigation" whether longitudinally measured dynamic variables (as in NEWS2) better distinguish patients who will deteriorate.

PE-specific scores

Generic scores

How the guideline says to use them

• Hestia, PESI and sPESI all identify low-risk patients suitable for outpatient management (COR 1/B-R). HOME-PE randomized sPESI-guided vs Hestia-guided home treatment: comparable proportions discharged, no significant difference in 30-day adverse outcomes, but sPESI recommendations were overruled by clinicians more often than Hestia's
• PESI Class I–II or Hestia-negative → 30-day mortality 0.30%, recurrent VTE 0.57%, major bleeding 0.45%
• The PE-specific scores — Bova, CPES, PESI, sPESI — have all been validated in haemodynamically stable acute PE as approaches to identifying higher risk of adverse outcomes such as all-cause mortality, PE-related mortality and clinical decompensation (COR 2a/B-NR)
• No score is substantially superior to another, so "clinicians can choose to apply any of the validated PE-specific risk scores in hemodynamically stable PE to identify potentially higher risk individuals"
• The caveat that matters most: the scores "generally show modest to good discriminatory ability in separating patients by their probability of developing adverse outcomes", but none had consistently strong predictive ability in identifying which individual patients would have adverse outcomes — so "many patients with hemodynamically stable PE who are categorized as higher risk do not develop adverse outcomes." An ideal score would identify patients who benefit from different monitoring or management; none does
• Generic scores perform comparably in some studies. Shock Index, NEWS and NEWS2 have shown performance comparable to PE-specific scores, with the advantage of being calculable from routinely obtained, repeatable vital signs. Which cut-points to use in PE, and whether repeated longitudinal measures help, is flagged as an evidence gap

CPES — Composite Pulmonary Embolism Shock

CPES is worth separating out because it is the score most directly tied to the guideline's novel Category D ("incipient cardiopulmonary failure" / normotensive shock) — and because it is easily mis-transcribed. Its definition appears in Table 6 and it is referenced twice more; the guideline gives it no dedicated narrative section and quotes no performance statistics of its own.

• Components — six items, 1 point each: elevated cardiac troponin; elevated B-type natriuretic peptide; moderately or severely reduced RV function; central thrombus burden (saddle PE); concomitant deep vein thrombosis; heart rate ≥100 bpm
• Range: 0–6 points
• Risk split: 0–5 points = lower risk, 6 points = higher risk for normotensive shock, defined here as a cardiac index ≤2.2 L/min/m²
• The cut-point is the top of the range, not a threshold with headroom. Because the score is unweighted and capped at 6, "higher risk" requires all six factors to be present — a deliberately demanding gate. A patient with five of six factors scores 5 and is classed lower risk. This is the detail most often lost when the score is summarised, and it is why CPES will classify the great majority of intermediate-risk patients as lower risk
• CPES is itself a Category D2 criterion. Figure 2's footnote lists the markers of reduced perfusion/end-organ dysfunction as lactate >2 mmol/L, acute kidney injury, urine output <0.5 mL/kg/hr, mental status change, cardiac index <2.2 L/min/m², MAP <60 mmHg, and "increased shock score/stage (SCAI stage, CPES score)". So CPES is used both to predict normotensive shock and, once elevated, as an instance of it — the same recursive structure as SCAI staging
• Validation status in this guideline: listed alongside Bova, PESI and sPESI as validated in haemodynamically stable PE for identifying higher risk (COR 2a/B-NR), subject to the same caveat — modest-to-good discrimination, no reliable individual-level prediction, no superiority over the other scores
• The supporting evidence sits outside the guideline. The score is derived and validated in Najarro M, Briceno W, Rodriguez C, et al. "Shock score for prediction of clinical outcomes among stable patients with acute symptomatic pulmonary embolism." Thromb Res. 2024;233:18–24 (guideline reference 16). Two further references support the same recommendation: a validation of a fast prognostic score for risk stratification of normotensive PE (Clin Res Cardiol. 2020;109:1008–1017) and the FLASH registry analysis of cardiogenic shock prevalence in intermediate-risk PE (JACC Cardiovasc Interv. 2023;16:958–972)

Hemodynamic Assessment of Categories C1–D (§3.2.2)

• The first 24–72 hours are the critical window. In recent trials this is when haemodynamic collapse or laboratory change is most commonly observed. In PEITHO, the mean time from randomization to the primary endpoint of haemodynamic collapse or escalation to lysis was 1.5–1.79 days ± 1.5 days — the evidentiary basis for close monitoring in this period. The PEITHO population (RV dysfunction on echocardiography or CT plus positive troponin) correlates with Category C3
• End-organ hypoperfusion outweighs hypotension alone. In the SHOCK trial, markers of end-organ malperfusion (elevated lactate, acute kidney injury or reduced urine output, and/or reduced cardiac index) were associated with higher in-hospital mortality than hypotension alone — the conceptual basis for separating Category D from Category E
• MAP has real discriminative value in Category C3. Observational data suggest MAP >80 mmHg carries very low risk of in-hospital death or adverse outcomes; in one retrospective series of 122 intermediate-high-risk patients, those with MAP 80–90 mmHg had few adverse events. The Italian Pulmonary Embolism Registry ROC analysis established 81.5 mmHg as the optimal cut-off for predicting 48-hour clinical deterioration (AUC 0.77; sensitivity 77.5%; specificity 95.0%; PPV 63.2%; NPV 97.7%). Hence COR 2a/B-NR for using MAP <80 mmHg in Category C3 to identify patients who may require escalation of therapy
• FLASH registry — the prevalence anchor for Category D: isolated hypoperfusion was present in 34% of intermediate-risk PE patients undergoing mechanical thrombectomy, and roughly one-third of thrombectomy patients had normalisation of the cardiac index — which the guideline reads as a way to refine stratification toward Category D2
• The guideline's own thresholds are stated inconsistently. Normotensive shock is defined in §3.2.2 as urine output <720 mL in 24 hours and cardiac index ≤2.2 L/min/m², whereas the Figure 2 footnote for the D2 criteria uses urine output <0.5 mL/kg/hr and cardiac index <2.2 L/min/m². These are not equivalent — 720 mL/24 h ≈ 30 mL/h irrespective of body weight — and the inequality differs at exactly 2.2. Any implementation of Category D must pick one

Biomarkers (§3.2.3)

• Troponin and BNP recommended in Cat C for short-term risk stratification (COR 1/B-NR); troponin elevation: OR 4.33 for all-cause mortality
• Lactate (venous or arterial) recommended in Cat C–E (COR 1/B-NR); elevated lactate: OR 5.13 for all-cause mortality in unselected PE; 4.54 in normotensive PE
• Lactate should be measured whenever cardiac biomarkers are assessed for risk stratification, with the threshold for an elevated level set by the local assay rather than a single universal cut-off
• Biomarkers are incremental to, not a substitute for, haemodynamic assessment — the normotensive-shock markers and the FLASH prevalence data are above (§3.2.2)

RV Imaging for Risk Stratification (§3.2.4)

• COR 1/A: in acute PE with an elevated clinical severity score but without evidence of shock (Categories C–D), RV imaging is recommended for short-term risk stratification
• COR 2a/B-NR: in the same population, echocardiography is preferred over CT for short-term risk stratification
• Prognostic weight of RV dysfunction, which varies by the baseline risk of the cohort studied — the page previously carried only the single figure in bold below, without its population or confidence interval:

• CT and echocardiography are not interchangeable measures of RV dysfunction. Compared with echocardiographic measurement, septal deviation on CT had sensitivity 0.31 (0.25–0.38) and specificity 0.98 (0.90–1.00), while increased RV/LV ratio on CT had sensitivity 0.83 (0.78–0.87) and specificity 0.75 (0.66–0.82) — i.e. septal deviation is highly specific but insensitive, whereas the RV/LV ratio trades specificity for sensitivity
• Point-of-care ultrasound may substitute for formal TTE when the latter is unavailable; in a prospective study of critically ill patients with suspected PE it showed acceptable accuracy for identifying RV dysfunction
• Thrombus burden quantification is NOT recommended for risk stratification in Cat A–C (§3.2.5, COR 3:No Benefit)

Echocardiographic RV Assessment — Table 4

The guideline's Table 4, "Optimal Methods of RV Dysfunction Assessment on Echocardiogram", is adapted from the American Society of Echocardiography right-heart guidance (reference 44, Rudski et al.). It is the operational counterpart to the reporting recommendation below — the cut-offs are the ones a sonographer needs, and none of them appeared on this page before this pass.

• The reporting recommendation (COR 1/B-NR, §3.1 Recommendation 12): in patients with acute PE who undergo TTE, RV dysfunction should be assessed and reported by all eight of the following — RV/LV end-diastolic ratio; RV end-diastolic diameter; TAPSE; estimated right ventricular systolic pressure; RV free wall hypokinesis with sparing of the apex (McConnell's sign); tricuspid systolic velocity; paradoxical septal motion; and IVC respirophasic collapse — to assist with risk stratification. Note: this page previously listed only six of the eight, omitting RV end-diastolic diameter and tricuspid systolic velocity.
• Reporting should be "as comprehensively as possible based on local expertise." Physicians with specialised echocardiography training — whether goal-directed or formal TTE — can detect RV dysfunction parameters in acute PE with good performance.
• The practical rule: more than one parameter raises accuracy. The supportive text states that "the sensitivity and specificity of performance increases if >1 parameter of RV dysfunction is present" (Table 4). A single abnormal measurement is therefore a weaker signal than a concordant set — which is the operational reason the recommendation asks for all eight to be reported rather than a single index.
• The echo and CTPA thresholds disagree, and the guideline does not reconcile them. Table 4 sets the echocardiographic cut-off at RV/LV >0.9, whereas the CTPA reporting recommendation (§3.1 Recommendation 11) and this page's diagnostic section both use RV/LV ≥1.0 as the headline. The §3.1 supportive text gives both CTPA thresholds with their performance — ≥1.0: sensitivity 85% (81–89%), specificity 72% (67–77%); ≥0.9: sensitivity 92% (89–95%), specificity 56% (46–66%) — so the same numerical index carries a different operating point depending on modality and on whether sensitivity or specificity is being optimised. Values are not transferable between the two, and the guideline notes that the degree of RV enlargement may be more predictive than a binary normal/abnormal call

Acute Management

Hospitalization Decisions

• Cat A and B: outpatient management reasonable (COR 2a/B-R) using Hestia/PESI/sPESI; requires immediate access to anticoagulation and reliable follow-up
• Patients on thrombolytics (systemic or CDL): ICU or intermediate care required (COR 1/C-EO)
• Unstable patients (Cat E): must NOT be transferred before stabilisation (COR 3:Harm)

PE Response Team (PERT)

• PERT is recommended for Cat C–E (COR 1/B-NR) to improve timeliness of care, reduce time-to-anticoagulation, decrease IVC filter use, and reduce length of stay
• PERT functions like code stroke/STEMI team; members vary by institution

Anticoagulation

• Anticoagulation recommended for all without absolute contraindication (COR 1/B-R)
• Parenteral: LMWH > UFH for Cat C1–E1 (COR 1/B-R); lower recurrent VTE without increased bleeding; once/twice daily dosing; no routine anti-Xa monitoring needed (COR 3:No Benefit/A)
• Oral: DOACs > VKAs (COR 1/B-R); lower major bleeding including ICH; lower fatal bleeding and all-cause mortality vs warfarin
• VKA preferred in thrombotic antiphospholipid syndrome (COR 1/A) due to arterial thrombosis risk with DOACs
• Pregnancy: LMWH or UFH only (COR 1/C-LD); DOACs and warfarin contraindicated (COR 3:Harm)
• Breastfeeding: LMWH, UFH, or warfarin over DOACs (COR 1/C-LD)
• Obesity (BMI >30): DOACs preferred over VKA (COR 2a/B-NR); dose reduction of LMWH may be reasonable in class III obesity (COR 2b/B-NR)
• CKD stage 2–3: DOAC over VKA (COR 1/A)
• Liver disease: DOAC reasonable in Child-Pugh A (COR 2a/C-LD) and B (COR 2b/C-LD); Child-Pugh C contraindicated (COR 3:Harm)

Hemodynamic Support

• Vasopressors/inotropes: recommended in Cat D2–E2 (COR 1/C-LD); norepinephrine first-line (SVR increase, modest inotropy, no PVR effect ≤15 μg/min); dobutamine adjunct for low CO
• Cautious fluid resuscitation (≤500 mL): reasonable in Cat D1–2 with reduced preload (COR 2b/B-R); avoid excessive volume (RV overload risk)
• Inhaled pulmonary vasodilators may reduce RV afterload in Cat C2–E (COR 2b/B-R); systemic vasodilators not effective
• Deep sedation and mechanical ventilation: should be avoided in Cat C–E unless clinically necessary (COR 3:Harm); associated with catastrophic hemodynamic collapse
• HFNC preferred over standard nasal cannula for moderate-severe hypoxia (COR 2a/C-LD)

Mechanical Circulatory Support

• VA-ECMO: reasonable for Cat E2 refractory cardiogenic shock (COR 2a/B-NR); continue systemic anticoagulation on ECMO unless bleeding (COR 1/B-NR)
• Additional advanced therapies while on VA-ECMO: not well established (COR 2b/C-LD)

IVC Filters

• Retrievable IVC filters over permanent when anticoagulation contraindicated (COR 1/B-R); retrieve as soon as safe (FDA recommends 29–54 days)
• Routine IVC filter in anticoagulated patients: not recommended (COR 3:Harm/A); increases DVT risk without mortality benefit (PREPIC, PREPIC2)
• Structured follow-up program to improve retrieval rates (COR 2a/C-LD)

Advanced Management (by Category)

Advanced Therapy by Category — Summary Table

Systemic Thrombolysis

• Standard dose: rt-PA 100 mg over 2 hours; tenecteplase has been studied but not FDA-approved for PE
• Cat E1–2: reasonable over anticoagulation alone (COR 2a/C-LD)
• Cat C3–D2 (PEITHO trial, n=1006): prevented cardiovascular collapse but increased ICH (2.4% vs 0.2%) and major extracranial bleeding (6.3% vs 1.2%)
• Lower-dose thrombolytics (25–50 mg rt-PA) may reduce bleeding with similar efficacy (COR 2b/C-LD); PEITHO-3 trial ongoing
• NOT recommended in Cat A1–C2 (COR 3:Harm/B-R)

Catheter-Directed Thrombolysis (CDL)

• rt-PA infused via multi-side-hole catheter directly into PA; standard dose 5–10 mg per PA; reduced dose (<5 mg) not recommended over standard (COR 3:NB/B-NR)
• Cat E1: reasonable (COR 2a/C-LD); Cat D1–2: may consider (COR 2b/B-NR)
• PEERLESS trial (n=550 Cat C2–D2): CDL vs MT — no difference in 30-day mortality or major bleeding; MT had fewer protocol-defined clinical deterioration events requiring bailout
• CDL may be preferred over systemic thrombolysis to reduce major/intracranial bleeding (COR 2b/C-LD)

Mechanical Thrombectomy (MT)

• Large-bore percutaneous suction thrombectomy; no lytic agent; allows non-ICU post-procedure care in most patients
• Cat E1: reasonable (COR 2a/B-NR); FLAME study (n=115 Cat E1): MT primary endpoint 17% vs context arm 63.9%; in-hospital mortality 1.9% vs 29.5%
• Cat D1–2: may consider (COR 2b/B-NR); FLARE and EXTRACT-PE trials: significant RV/LV ratio reduction at 48 hours
• FLASH registry (n=1000): mean PA pressure and cardiac index significantly improved post-procedure; major adverse events 1.8%
• MT preferred over systemic thrombolysis to reduce major bleeding (COR 2b/B-NR)

Surgical Embolectomy

• Cat E1: reasonable (COR 2a/B-NR); survival >97% in modern series among patients without pre-operative CPR
• Cat E2: NOT recommended over other advanced therapies (COR 3:NB/B-NR); most post-operative deaths in this group
• Cat A–C3: NOT recommended (COR 3:NB/C-EO)
• Compared to systemic thrombolysis: lower ICH risk; better RV recovery; equivalent long-term survival in non-randomised comparisons
• Reduces RV size, PA systolic pressure, and improves perfusion better than thrombolysis

Follow-Up and Long-Term Management

Post-Discharge Follow-Up

• Within 1 week of discharge: patient education, adherence, bleeding assessment (COR 1/C-LD)
• At/before 3 months: discuss anticoagulation duration, further testing, symptom assessment (COR 1/C-EO)
• Every visit for ≥1 year: screen for PE-related symptoms and functional limitations to detect CTEPD (COR 1/C-LD)
• Screen for anxiety and depression at follow-up visits (COR 2a/B-NR)
• Specialist PE clinic referral for complex patients (COR 2a/B-NR)
• Cancer screening (thorough history, exam, age-appropriate screening) for unprovoked PE (COR 1/A); routine CT/PET-CT NOT recommended (COR 3:NB/A)
• Thrombophilia testing: reasonable in patients <55 or with family history if results will change management (COR 2b/C-LD)

Anticoagulation Duration

• First PE without major reversible risk factor: extend beyond 3–6 months (COR 1/A); 30–40% recurrence risk at 10 years without extended treatment
• Major reversible risk factor: stop at 3–6 months (COR 1/B-NR); <1%/year recurrence risk post-surgery
• Persistent risk factor (cancer, autoimmune disease, chronic immobility): continue extended phase (COR 1/C-LD)
• Minor reversible risk factor: shared decision-making on stopping vs continuing (COR 2a/B-NR)
• Extended phase agent: DOAC over VKA (COR 1/A); half-dose apixaban 2.5 mg BID or rivaroxaban 10 mg daily recommended to reduce bleeding while maintaining efficacy (COR 1/A); RENOVE and API-CAT trials confirm half-dose safety even in cancer-associated VTE
• Active cancer: for the extended phase, either a DOAC or LMWH is recommended over VKA to reduce recurrent VTE (COR 1/A)
• VKA for patients unable to take DOAC or refusing it (COR 1/B-R); aspirin reasonable if anticoagulation refused/contraindicated (COR 2a/B-R)

Recurrent PE

• Confirm with CTPA or V/Q scan (COR 1/C-EO); compare with prior imaging
• Investigate for nonadherence, subtherapeutic dosing, cancer, antiphospholipid antibodies, drug interactions (COR 1/C-LD)
• Switch to alternative drug class if recurrence on therapeutic anticoagulation (COR 2a/C-EO)
• If recurrence on reduced-dose DOAC: escalate to full dose (COR 2a/C-EO)
• Cancer with recurrence on LMWH: dose escalation 20–25% (COR 2a/B-NR)

CTEPD Surveillance

• If persistent dyspnea/functional impairment ≥3 months after PE: evaluate for CTEPD with TTE + lung perfusion scan V/Q or SPECT/CT (COR 1/B-NR); CPET useful to exclude CTEPD (COR 2a/B-NR)
• CTEPD with PH: referral to expert centre (COR 1/C-LD); without PH: consider referral (COR 2a/C-LD)
• Pulmonary rehabilitation for persistent impairment when CTEPD excluded (COR 1/B-R)
• Continue anticoagulation during CTEPD evaluation (COR 1/B-NR)
• Routine imaging for resolved, asymptomatic patients: NOT recommended (COR 3:NB/B-NR)
• If the echocardiogram was normal at the time of the acute PE, a repeat echocardiogram at 3–6 months is low yield and may be omitted

Evidence Gaps and Future Directions

The guideline's own declared gaps (its Table 11) — distinct from the methodological limitations below:

• Risk stratification — validating the AHA/ACC clinical categories and integrating novel predictors such as thrombus burden and RV enlargement metrics to better guide therapeutic decisions
• Anticoagulation and thrombolysis — need for safer agents, improved strategies for patients already on treatment, and clearer guidance for travel-related prophylaxis
• Advanced therapies — robust algorithms and comparative-effectiveness data, particularly for catheter-based interventions and surgical embolectomy in high-risk populations
• Long-term outcomes — remain poorly understood, especially in chronic thromboembolic disease and in special populations such as rural patients, those requiring hormone therapy, and pregnant individuals
• Emerging technology — artificial intelligence and radiomics offer promising avenues for diagnosis and personalisation of care but require further validation

Limitations of the Document

Evidence base

• Literature search concluded October 2024; key studies to April 2025 added by committee; contemporary ongoing trials (PEITHO-3, PEERLESS-2) not included
• Advanced therapy recommendations (CDL, MT, surgical embolectomy) largely based on single-arm prospective studies and registries rather than randomised controlled trials vs anticoagulation alone
• Long-term functional outcomes and quality of life after advanced therapies are not well characterised beyond 30–90 days
• CTEPD prevalence without PH is unknown; optimal diagnostic workup and natural history not established
• CDL vs systemic thrombolysis has not been directly compared in randomised trials
• Most DOAC trials excluded severe renal impairment and severe hepatic disease; recommendations in these groups based on registry data

Risk scores and thresholds

• CPES is presented without any performance data. The guideline lists it among validated PE-specific scores but quotes no discrimination statistic, no derivation or validation cohort size, and no cut-point performance — those sit in the primary studies it cites (Najarro et al., Thromb Res 2024). Its 0–5 vs 6 split is unusually demanding for a six-item unweighted score, and nothing in this document indicates how often real patients reach 6, or what proportion of Category D2 rests on CPES versus the other six markers
• No risk score predicts for the individual patient. The guideline states plainly that none of the PE-specific scores had consistently strong predictive ability for identifying which individual patients would have adverse outcomes, so many patients stratified as higher risk do not deteriorate. This is a limitation of the entire risk-score apparatus, and it applies to CPES as much as to PESI or Bova — the scores are best read as triage aids, not prognoses
• The severity-score gate has three alternative definitions in the same figure. Low severity is defined as PESI ≤85 or sPESI = 0 or Bova ≤4 (elevated: PESI >85 or sPESI ≥1 or Bova >4), so the same patient can land in Category B or Category C depending on which score is applied — the guideline gives no rule for choosing between them, and states elsewhere that no score is superior
• Normotensive-shock thresholds are internally inconsistent. §3.2.2 defines urine output <720 mL/24 h and cardiac index ≤2.2 L/min/m²; the Figure 2 footnote for the D2 criteria uses urine output <0.5 mL/kg/hr and cardiac index <2.2 L/min/m². These are not equivalent, and the inequality differs at exactly 2.2

Imaging and echocardiographic thresholds

• Table 4's cut-offs are borrowed, not derived in PE. The table is adapted from the American Society of Echocardiography's general right-heart assessment guidance (reference 44, Rudski et al.), so the thresholds for RV dimension, TAPSE, tricuspid velocity and Doppler PH evidence are general-population echocardiographic norms rather than PE-specific validated cut-points. The guideline does not state their sensitivity or specificity in acute PE
• The same index carries two different operating points, and the guideline does not reconcile them. Echocardiographic RV/LV is abnormal above 0.9 (Table 4); CTPA RV/LV is reported at ≥1.0 as the headline (§3.1 Recommendation 11). The two modalities therefore disagree about the same ratio, on the same patient, and the values are not transferable — a point made sharper by the guideline's own CTPA data showing that moving the cut-point from 1.0 to 0.9 shifts sensitivity from 85% to 92% while dropping specificity from 72% to 56%
• No guidance is given on how many abnormal parameters should change management. The guideline states that accuracy improves when more than one parameter is abnormal, and asks for all eight to be reported — but does not specify which combination, or what count, should move a patient between Categories C1, C2 and C3. The reporting recommendation is therefore more concrete than the decision rule it feeds
• The RIETE prose overstates its own effect size. The §3.2.4 supportive text describes RV hypokinesis as "associated with a 4-fold increased odds of PE-related mortality" while reporting OR 3.11 (95% CI 1.85–5.21) — approximately 3-fold. The confidence interval does not comfortably reach 4
• One prognostic estimate is too imprecise to use. The individual-patient-data meta-analysis reports PE-related death at OR 22.9 (95% CI 2.89–181) — a lower bound near 3 and an upper bound near 181. It is cited on the page for completeness but cannot support a specific risk estimate

Key Concepts Mentioned

• concepts/Acute-PE-Clinical-Categories — central classification system (A–E) driving all management decisions; also carries the summary of the risk scores below
• CPES (Composite Pulmonary Embolism Shock) — six-item unweighted score (range 0–6; 0–5 lower risk vs 6 higher risk) predicting normotensive shock (cardiac index ≤2.2 L/min/m²); the score most closely tied to Category D, and itself one of the D2 hypoperfusion markers alongside SCAI stage
• Bova score — the third accepted alternative to PESI/sPESI for the Category B-vs-C severity gate (low = Bova ≤4), frequently omitted from summaries of the scheme
• Shock Index and NEWS/NEWS2 — generic scores that the guideline notes perform comparably to PE-specific scores in some studies, with the advantage of being recalculable from routine vital signs; NEWS2 ≥9 = high risk
• Normotensive shock — isolated hypoperfusion without hypotension; the defining concept of Category D, and the outcome CPES is designed to predict
• Echocardiographic RV dysfunction parameters (Table 4) — five parameter groups with cut-offs: RV EDD >30 mm and RV basal EDD >42 mm; RV/LV >0.9; TAPSE <1.6 cm; pulmonary acceleration time <90 ms or RV/atrial gradient >30 mm Hg (Doppler evidence of PH); tricuspid systolic velocity >2.6 m/sec. Accuracy improves when more than one parameter is abnormal, which is why all eight reporting parameters are requested
• McConnell's sign — RV free wall hypokinesis with apical sparing; one of the eight parameters Recommendation 12 requires to be reported
• YEARS algorithm — D-dimer based diagnostic tool
• PERT — PE response team model
• CTEPD — chronic thromboembolic pulmonary disease (without PH); distinct from CTEPH
• Half-dose DOAC for extended anticoagulation
• Initial vs Extended Treatment Phase — the guideline's own phase definitions (initial = first 3–6 months; extended = >6 months, without an anticipated stop date)

Key Entities Mentioned

• entities/Pulmonary-Embolism — primary entity
• entities/CTEPH — post-PE sequelae; CTEPD distinct entity formalised
• concepts/Pulmonary-Hypertension — CTEPD/CTEPH relationship

Wiki Pages Updated

• Created: wiki/sources/acute-pe-aha-2026.md
• Created: wiki/entities/Pulmonary-Embolism.md
• Created: wiki/concepts/Acute-PE-Clinical-Categories.md
• Updated: wiki/entities/CTEPH.md
• Updated: wiki/sourceindex.md
• Updated: wiki/wikiindex.md

Revision history

• 2026-04-21 — initial ingest. Source page, entities/Pulmonary-Embolism.md and concepts/Acute-PE-Clinical-Categories.md created; entities/CTEPH.md, sourceindex.md and wikiindex.md updated.
• 2026-10-15 — > improve pass (source fidelity). Re-converted raw/Acute-PE-AHA-2026.md from the source PDF with the pinned pymupdf4llm engine — restoring 1,172 citation superscripts the previous conversion had dropped and normalising all image paths to /raw/assets/. Added sections absent from the first ingest: Top Take-Home Messages, Definitions and Treatment Phases (§2.1), Evidence Gaps and Future Directions (§7); plus the §5.2 active-cancer recommendation, the §6.1 repeat-echocardiogram footnote, and the COR/LOE definition.
• 2026-10-15 — > improve pass (risk scores / CPES). The page carried no mention of CPES, Bova, Shock Index, NEWS, or Table 6 — the entire risk-score apparatus beyond PESI/sPESI/Hestia was missing. Added: the full Table 6 with all six scores, their components, ranges and cut-offs; a dedicated CPES subsection (six 1-point components, range 0–6, the 0–5 vs 6 split, its role as the operational score for Category D and as itself a D2 marker, and its primary citation Najarro et al. Thromb Res 2024); the complete D1/D2 definitions and D2 hypoperfusion marker list; the Bova-based severity gate; a Hemodynamic Assessment of Categories C1–D subsection (24–72 h window and PEITHO timing, SHOCK trial, MAP ROC data, FLASH normalisation); and four new Limitations entries. Corrected in passing: concepts/Acute-PE-Clinical-Categories.md described CPES's range as 0–5, conflating the range with the risk cut-off — it is 0–6.
• 2026-10-15 — > improve pass (layout). Structure tidied without changing any finding: the three bold pseudo-headings (**PE-specific scores**, **Generic scores**, **How the guideline says to use them**) promoted to real ##### headings; the two risk-score tables aligned to a common Risk categories column and their component lists wrapped with <br> (the PESI row was a single 1,015-character line); all four table separator rows normalised to |---|---|; §-references added to the four §3.2.x headings; the duplicated FLASH/normotensive-shock bullet removed from Biomarkers and replaced with a cross-reference and the guideline's own "measure lactate alongside biomarkers, local assay threshold" point; ## Limitations split into Evidence base and Risk scores and thresholds; the changelog moved out of this section's bullet list into Revision history above; #### Summary Table renamed to Advanced Therapy by Category.
• 2026-10-15 — > improve pass (RV dysfunction assessment on echocardiography). The page carried no version of the guideline's Table 4, "Optimal Methods of RV Dysfunction Assessment on Echocardiogram" — no cut-offs, no recommended views, and no TAPSE/RV-dimension/Doppler thresholds. Added: the full Table 4 (five parameter groups, each with recommended technique and definition); the complete COR 1/B-NR reporting recommendation with all eight parameters, where the page had previously listed only six (RV end-diastolic diameter and tricuspid systolic velocity were absent, both here and in the diagnostic-imaging section); the guideline's practical rule that accuracy improves when more than one parameter is abnormal; a prognostic-performance table replacing the page's bare "OR 4.01" with all seven ORs, their populations and their confidence intervals; the CT-vs-echocardiography sensitivity/specificity comparison showing septal deviation (sens 0.31, spec 0.98) and CT RV/LV ratio (sens 0.83, spec 0.75) are not interchangeable with echo; the point-of-care ultrasound alternative; and the rationale that most patients with acute PE show no RV dysfunction, which is why echo cannot be a diagnostic test. New Imaging and echocardiographic thresholds limitation group added (5 entries), including the unreconciled RV/LV 0.9 (echo) vs 1.0 (CTPA) discrepancy and the RIETE prose reporting "4-fold" against its own OR 3.11.