Skip to main content

Trial Summaries & Critiques

HI-PEITHO

Image: Shutterstock / BelezaPoy

Pulmonary angiogram showing a pulmonary embolism

Publication

  • Title: Ultrasound-Facilitated, Catheter-Directed Fibrinolysis for Acute Pulmonary Embolism
  • Acronym: HI-PEITHO
  • Year: 2026
  • Journal published in: New England Journal of Medicine
  • Citation: Rosenfield K, Klok FA, Piazza G, et al. Ultrasound-Facilitated, Catheter-Directed Fibrinolysis for Acute Pulmonary Embolism. N Engl J Med. 2026 Mar 28.

Context & Rationale

  • Background
    • For haemodynamically unstable acute pulmonary embolism (PE), reperfusion is accepted; for haemodynamically stable but right-heart–strained PE, the balance between preventing collapse and causing major bleeding remained unresolved.
    • The original PEITHO trial showed that full-dose systemic fibrinolysis reduced early haemodynamic collapse in intermediate-risk PE, but at the cost of substantially more major bleeding and intracranial haemorrhage, which prevented routine first-line adoption.1
    • Post-hoc PEITHO work suggested that adding simple clinical distress variables to RV dysfunction and troponin elevation identifies a subgroup at more immediate risk of deterioration, thereby offering a more rational target for reperfusion trials.2
    • Ultrasound-facilitated catheter-directed thrombolysis had biologic plausibility, because ultrasound may enhance local alteplase activity and clot penetration, and earlier studies suggested favourable RV remodelling with lower-dose alteplase; however, those data were mechanistic, surrogate-based, dose-finding, or single-arm rather than definitive clinical-outcome evidence.3456
  • Research Question/Hypothesis
    • In adults with an enriched, higher-risk intermediate-risk PE phenotype — proximal clot burden, RV dilation, troponin elevation, and at least two markers of cardiorespiratory distress — does protocolised ultrasound-facilitated catheter-directed fibrinolysis plus anticoagulation reduce PE-related death, cardiorespiratory decompensation/collapse, or symptomatic recurrent PE within 7 days, as compared with anticoagulation alone?4
  • Why This Matters
    • Before HI-PEITHO, catheter-directed fibrinolysis was widely discussed and increasingly used, but lacked a large randomised trial against anticoagulation alone with a clinically meaningful composite end point.
    • A positive trial would support an invasive, lower-dose reperfusion strategy for carefully selected patients before overt shock develops; a negative or unsafe trial would strengthen anticoagulation alone as default management.4

Design & Methods

  • Research Question: Whether ultrasound-facilitated, catheter-directed fibrinolysis plus anticoagulation is superior to anticoagulation alone for preventing 7-day PE-related death, cardiorespiratory decompensation/collapse, or symptomatic recurrent PE in an enriched intermediate-risk acute PE cohort.
  • Study Type: Post-market, multinational, multicentre, adaptive-design, open-label, randomised controlled PROBE trial with blinded adjudication of the primary outcome; 59 enrolling sites across Europe and the United States; acute hospital/ED PE pathways with access to interventional angiography-capable services.
  • Population:
    • Inclusion criteria:
      • Age 18–80 years.
      • Objectively confirmed acute PE on computed tomographic pulmonary angiography, with clot in at least one main or proximal lobar pulmonary artery.
      • RV:LV end-diastolic diameter ratio ≥1.0 on computed tomographic pulmonary angiography.
      • Troponin I or T above the upper limit of normal.
      • At least two new-onset clinical distress criteria: systolic blood pressure ≤110 mm Hg for at least 15 minutes; heart rate ≥100 beats/min; respiratory rate >20 breaths/min or resting hypoxaemia.
      • Written informed consent.
      Exclusion criteria:
      • Persistent haemodynamic instability, including cardiac arrest/CPR, ECMO, PE-related shock, or persistent hypotension <90 mm Hg.
      • Need for ICU admission for a reason other than the index PE episode.
      • Symptom duration >14 days or logistical inability to provide rapid interventional therapy.
      • Major bleeding risk, including active bleeding, prior intracranial or intraocular bleeding, recent stroke/TIA, recent major surgery/trauma, thrombocytopenia, CNS neoplasm, or severe coagulopathy.
      • Recent thrombolytic or glycoprotein IIb/IIIa exposure; disallowed recent or chronic DOAC/VKA use; dual antiplatelet therapy.
      • Pregnancy or lactation, prior study inclusion, known hypersensitivity to study drugs, or life expectancy <6 months.
  • Intervention:
    • Ultrasound-facilitated catheter-directed fibrinolysis with the EkoSonic system plus anticoagulation.
    • Alteplase regimen: 2 mg bolus per catheter, then 1 mg/hour per catheter for 7 hours; total 9 mg for unilateral treatment or 18 mg for bilateral treatment.
    • Bilateral proximal thrombus generally prompted bilateral catheter placement.
    • Before the procedure: therapeutic LMWH twice daily or therapeutic UFH.
    • During and for up to 4 hours after the procedure: IV UFH 300–600 U/hour.
    • After the procedure: return to full-dose parenteral anticoagulation, then transition to oral anticoagulation no sooner than 24 hours after fibrinolytic infusion; anticoagulation for at least 3 months.
  • Comparison:
    • Anticoagulation alone with therapeutic LMWH or UFH according to protocol and local practice.
    • Transition to oral anticoagulation no sooner than 24 hours after randomisation.
    • No planned interventional procedure.
    • Rescue reperfusion or escalation therapy was permitted only after documented objective decompensation/collapse.
  • Blinding: Patients and treating clinicians were unblinded; the clinical-events committee adjudicating primary and major safety outcomes, and the echocardiographic core laboratory, were blinded to treatment assignment.
  • Statistics: Originally, 406 patients were required to detect an absolute reduction in the 7-day primary outcome from 15% to 5% (odds ratio 0.298) with 90% power at one-sided α=0.025; after the prespecified 50% interim analysis, the sample size was increased to 544 to preserve 90% power for detecting 15% versus 6%, with a final two-sided α boundary of 0.02938. The primary analysis was intention-to-treat; per-protocol analysis was supportive; safety analyses used the treated population. Fisher’s exact test was used for the primary outcome because Cochran–Mantel–Haenszel assumptions were violated by sparse strata.
  • Follow-Up Period: Primary end point at 7 days (or discharge); acute safety and functional results reported through 30 days; prespecified 6- and 12-month follow-up ongoing.

Key Results

This trial was not stopped early. A planned interim analysis at 50% enrolment led to a prespecified sample-size increase from 406 to 544 patients; no early efficacy or futility stopping boundary was crossed.

Outcome USCDT + anticoagulation Anticoagulation alone Effect p value / 95% CI Notes
Primary composite outcome at 7 days 11/273 (4.0%) 28/271 (10.3%) RR 0.39 95% CI 0.20 to 0.77; P=0.005 Primary intention-to-treat analysis.
PE-related death at 7 days 3/273 (1.1%) 1/271 (0.4%) RR 3.0 95% CI 0.3 to 28.5; P not reported Very low event counts.
Cardiorespiratory decompensation/collapse at 7 days 10/273 (3.7%) 28/271 (10.3%) RR 0.4 95% CI 0.2 to 0.7; P not reported Main driver of the primary effect; isolated NEWS ≥9 events occurred in 1 vs 14 patients.
Symptomatic recurrent PE at 7 days 1/273 (0.4%) 1/271 (0.4%) RR 1.0 95% CI 0.1 to 15.8; P not reported No separation for recurrence.
ISTH major bleeding at 7 days 11/271 (4.1%) 6/271 (2.2%) RR 1.8 95% CI 0.7 to 4.9; P=0.32 Treated population.
ISTH major bleeding at 30 days 11/271 (4.1%) 8/271 (3.0%) RR 1.4 95% CI 0.6 to 3.4; P=0.64 Treated population.
GUSTO moderate/severe bleeding at 7 days 9/271 (3.3%) 4/271 (1.5%) RR 2.3 95% CI 0.7 to 7.2; P=0.26 Treated population.
Ischaemic stroke at 30 days 1/271 (0.4%) 0/271 (0.0%) NE P=1.00 Treated population.
Intracranial haemorrhage at 30 days 0/271 (0.0%) 0/271 (0.0%) NE P=1.00 No intracranial haemorrhage in either group.
All-cause mortality at 30 days 5/271 (1.8%) 3/271 (1.1%) RR 1.7 95% CI 0.4 to 6.9; P not reported Treated population; no mortality signal.
Serious adverse events at 30 days 40/271 (14.8%) 44/271 (16.2%) RR 0.9 95% CI 0.6 to 1.3; P=0.64 Treated population.
    • The clinical benefit was almost entirely a decompensation-prevention signal rather than a mortality or recurrence signal.
    • Harms were not statistically increased, but bleeding counts numerically favoured control: ISTH major bleeding at 7 days 4.1% versus 2.2%, at 30 days 4.1% versus 3.0%; there was 1 ischaemic stroke in the intervention arm and no intracranial haemorrhage in either arm.
    • Exploratory subgroup analyses were hypothesis-generating: age <75 years RR 0.29 (95% CI 0.13 to 0.66), CTPA RV:LV ≥1.5 RR 0.15 (95% CI 0.04 to 0.48), age ≥75 years RR 0.96 (95% CI 0.27 to 3.47), and CTPA RV:LV <1.5 RR 1.10 (95% CI 0.41 to 2.95). Early core-lab RV:LV change at 48 hours was −0.38±0.24 versus −0.27±0.23, and 30-day median 6-minute walk distance was 405.0 m versus 393.0 m.

Internal Validity

    • Randomisation and allocation: Central computer randomisation in a 1:1 ratio, stratified by age (<75 vs ≥75 years) and CTPA RV:LV ratio (<1.5 vs ≥1.5), is a major strength. Operational details of allocation concealment were not exhaustively described, but the EDC-based process appears robust.
    • Drop out or exclusions: Early follow-up was excellent. The 7-day/discharge visit was completed in 266/273 (97.4%) versus 266/271 (98.2%); 30-day visits in 253/273 (92.7%) versus 250/271 (92.3%). Only 2 patients had missing primary-outcome data. Twenty patients (3.7%) had major protocol deviations and were excluded from the per-protocol analysis: 15/273 in the intervention arm and 5/271 in control.
    • Post-randomisation treatment fidelity: 271/273 (99.3%) patients assigned to intervention actually underwent device treatment. Two did not start fibrinolysis; one died before treatment began, and one had a major protocol deviation.
    • Performance/detection bias: The trial was open-label, so clinician behaviour and escalation thresholds could in principle bias a composite dominated by decompensation. This was mitigated by prespecified objective rescue rules, mandatory confirmation of NEWS ≥9 after 15 minutes, blinded CEC adjudication, and blinded core-lab echocardiography.
    • Protocol adherence: Delivery of the experimental protocol was generally good: bilateral catheters were used in 251/273 (91.9%), mean total alteplase dose was 16.32±3.95 mg, and mean infusion duration was 7.33±1.39 hours. Treatment began within 120 minutes of randomisation in 200/271 treated patients (73.3%), but 71/271 (26.0%) started later than 2 hours, so the preferred time target was not universal.
    • Baseline characteristics: Groups were broadly comparable for age (58.2 vs 58.2 years), female sex (41.8% vs 43.5%), BMI (32.9±8.5 vs 33.4±8.2), baseline NEWS (6.0±1.9 vs 6.0±1.9), and concomitant ultrasound-confirmed DVT (39.2% vs 40.2%). However, several standardised differences exceeded 10%, including symptom duration (3.5 vs 3.9 days), dyspnoea (93.0% vs 89.3%), tachypnoea (43.2% vs 52.4%), bilateral PE (91.9% vs 96.7%), and CTPA RV:LV ratio (1.6±0.5 vs 1.5±0.4).
    • Heterogeneity: A 59-centre international device trial inevitably introduces operator and centre variability, and randomisation was not stratified by site. The protocolised device regimen and blinded adjudication reduce this concern but do not eliminate centre effects.
    • Timing: Randomisation and treatment initiation were required within 6 hours of diagnostic confirmation. That is clinically sensible for an early deterioration-prevention strategy. The 73.3% within-2-hour initiation rate suggests good, but not complete, operational success.
    • Dose: The alteplase dose was deliberately low relative to systemic fibrinolysis (9 or 18 mg total rather than full systemic-dose regimens). This seems biologically and clinically plausible for reducing bleeding risk, but HI-PEITHO did not compare alternative catheter-delivered doses.
    • Separation of the variable of interest: The intervention and control arms were clearly separated procedurally. At 48 hours, the core-lab RV:LV ratio was 0.92±0.18 in the intervention arm versus 1.00±0.20 in control; change from baseline was −0.38±0.24 versus −0.27±0.23. Rescue therapy was required in 8/273 (2.9%) versus 25/271 (9.2%), showing real clinical separation after randomisation.
    • Crossover/rescue therapy: Protocol-allowed rescue reperfusion occurred more often in control. However, 7 of 33 rescue escalations did not satisfy prespecified decompensation criteria, and 5 intervention-arm and 3 control-arm patients received adjuvant/additional therapy without meeting criteria, contributing to per-protocol exclusions.
    • Adjunctive therapy use: UFH was more common in the intervention arm (71.6% vs 55.7%), which reflects the procedural protocol rather than contamination, but means the trial compares treatment strategies rather than an isolated device effect.
    • Outcome assessment: PE-related death, recurrence, major bleeding, and serious adverse events were independently adjudicated. The main vulnerability is that the composite included a score-based deterioration criterion (NEWS), which is more open to ascertainment nuance than death, CPR, ECMO, or intubation.
    • Statistical rigour: The adaptive design was executed as planned. The sample size increase from 406 to 544 followed the prespecified promising-zone rule; the final two-sided alpha boundary was preserved. The primary effect remained consistent in the per-protocol analysis (7/258 [2.7%] vs 28/266 [10.5%]; RR 0.26; 95% CI 0.12 to 0.59) and in the stratification-adjusted sensitivity analysis (OR 0.36; 95% CI 0.18 to 0.75).

Conclusion on Internal Validity: Internal validity is moderate-to-strong. Randomisation, excellent early follow-up, blinded adjudication, and a tightly protocolised intervention support a credible efficacy signal, but open-label care, a composite heavily influenced by decompensation rather than death, site/operator variability, and some baseline imbalances prevent the trial from being methodologically definitive.

External Validity

    • Population representativeness: This was a highly selected cohort: 544 of 4313 screened patients were randomised. Patients had to be 18–80 years old, troponin-positive, have RV:LV ≥1.0, proximal clot burden, and at least two clinical distress markers. Mean age was 58.2 years, only 10.1% were ≥75 years, 42.6% were women, and 15.8% identified as non-White.
    • Clinical phenotype: The findings are most applicable to a higher-risk intermediate-risk or intermediate-high-risk PE phenotype with central clot burden and objective RV injury, not to lower-risk intermediate PE, isolated distal PE, or overt shock.
    • Setting dependence: Generalisability is strongest for centres with PERT-like systems, rapid CT diagnosis, 24/7 interventional expertise, angiography-capable suites, and familiarity with EKOS-style catheter-directed fibrinolysis. Applicability is lower in resource-limited systems or hospitals without rapid interventional availability.
    • Treatment platform: The trial specifically tested ultrasound-facilitated catheter-directed alteplase delivery with a standardised low-dose regimen. Extrapolation to other catheter-directed lytic systems, other lytic doses, or large-bore thrombectomy devices should be cautious.
    • Excluded groups: Older frailer patients, pregnant/lactating patients, patients already receiving many anticoagulant regimens, and those with major bleeding risk were largely excluded; these are precisely the groups in whom bedside equipoise is often hardest.

Conclusion on External Validity: External validity is moderate. HI-PEITHO is highly relevant to contemporary advanced PE centres treating carefully selected, haemodynamically stable but clinically stressed patients with RV injury, but it should not be overgeneralised to broader “intermediate-risk” PE populations, older/frailer patients, or settings without mature endovascular capability.

Strengths & Limitations

  • Strengths:
    • First large randomised clinical-outcome trial comparing ultrasound-facilitated catheter-directed fibrinolysis directly against anticoagulation alone.
    • International multicentre design with 59 sites and broad acute-care recruitment pathways.
    • Clinically enriched population that makes biological sense for reperfusion testing.
    • Near-complete early follow-up with only 2 missing primary-outcome observations.
    • Blinded CEC adjudication and blinded echocardiographic core laboratory.
    • Standardised lytic dose, procedural approach, and anticoagulation protocols.
    • Adaptive design that preserved power when the originally assumed treatment effect proved uncertain.
  • Limitations:
    • Open-label care in a device trial, with inevitable risk of expectation bias despite adjudication safeguards.
    • Primary benefit driven by decompensation/collapse rather than mortality or recurrent PE.
    • NEWS contributed materially to the event difference and is not a universally adopted decompensation metric.
    • Bleeding estimates were imprecise; the trial was not powered to exclude clinically important bleeding excess.
    • Older patients were underrepresented, and site was not a stratification factor.
    • Some baseline imbalances and post-randomisation protocol deviations complicate interpretation.
    • Long-term functional, quality-of-life, CTEPH, and resource-use outcomes are not yet available.
    • The comparison was against anticoagulation alone, not against reduced-dose systemic fibrinolysis or mechanical thrombectomy.

Interpretation & Why It Matters

  • Clinical meaning
    • HI-PEITHO shows that, in a carefully selected enriched intermediate-risk PE population, early low-dose ultrasound-facilitated catheter-directed fibrinolysis can reduce early clinical deterioration beyond anticoagulation alone.
    • In practical terms, the signal is prevention of collapse rather than rescue of established shock or reduction in short-term mortality.
    • This refines the lesson from PEITHO: reperfusion may matter before shock develops, but how reperfusion is delivered determines whether the efficacy–safety trade-off becomes clinically acceptable.18

Controversies & Other Evidence

    • The main contemporaneous critique is that the composite outcome was dominated by cardiorespiratory decompensation rather than death, and that NEWS — although prespecified, objective, and validated — is not universally embedded in acute PE care pathways, particularly outside systems that routinely use it. The accompanying editorial also highlights that 321 screened patients were excluded for “other” or “unknown” reasons and that only 78.8% of rescue escalations fulfilled the prespecified decompensation definition, both of which complicate interpretation and generalisability.7
    • The editorial’s second major caution is safety precision: the absence of intracranial haemorrhage is reassuring, but the event counts for major bleeding still numerically favoured control, so HI-PEITHO does not prove procedural equivalence for bleeding — it only failed to show a statistically significant difference.7
    • Set against prior systemic lysis evidence, HI-PEITHO appears to improve the acute benefit–harm profile: PEITHO reduced early collapse with systemic tenecteplase but caused substantially more major bleeding and intracranial haemorrhage, while long-term PEITHO follow-up did not show durable mortality or functional advantage. HI-PEITHO therefore supports the concept that route, dose, and local delivery may matter as much as reperfusion itself.18
    • Parallel endovascular evidence remains complementary rather than contradictory. PEERLESS showed superiority of large-bore mechanical thrombectomy over catheter-directed thrombolysis on a hierarchical composite in intervention-selected intermediate-risk PE, but it did not test either strategy against anticoagulation alone. A 2025 meta-analysis of catheter-based therapies versus anticoagulation suggested better short- and mid-term outcomes without clear safety penalty, but that literature was still dominated by non-randomised data; HI-PEITHO is therefore the crucial randomised clinical-outcome upgrade for catheter-based reperfusion evidence.910
    • Recent guidance already moved toward cautious acceptance of endovascular reperfusion in selected intermediate-high-risk and high-risk PE, especially when systemic fibrinolysis is contraindicated or deterioration occurs. The 2026 multisociety AHA/ACC/ACCP/ACEP/CHEST/SCAI/SHM/SIR/SVM/SVN guideline and contemporary CIRSE standards therefore provide a receptive framework, but they were issued essentially contemporaneously with HI-PEITHO and cannot yet be viewed as formal post-HI-PEITHO guideline uptake.1112
    • The key unresolved therapeutic comparison is reduced-dose systemic fibrinolysis. HI-PEITHO does not establish whether local low-dose catheter delivery is superior to lower-dose systemic alteplase, and PEITHO-3 remains the most relevant randomised programme designed to address that adjacent question.13

Summary

    • HI-PEITHO randomised 544 patients with an enriched intermediate-risk / intermediate-high-risk acute PE phenotype to ultrasound-facilitated catheter-directed fibrinolysis plus anticoagulation or anticoagulation alone.
    • The 7-day primary composite outcome fell from 10.3% to 4.0% (RR 0.39; 95% CI 0.20 to 0.77; P=0.005), mainly because cardiorespiratory decompensation/collapse was reduced.
    • No mortality benefit was shown, and recurrent PE was rare in both groups.
    • Major bleeding and stroke were infrequent; no intracranial haemorrhage occurred, but bleeding event counts numerically remained higher with intervention.
    • The trial is highly important because it provides the first robust randomised clinical-outcome evidence that catheter-directed fibrinolysis can prevent early deterioration in a carefully selected PE subgroup, while leaving long-term benefit, bleeding precision, and comparative effectiveness versus other reperfusion strategies unresolved.

Overall Takeaway

HI-PEITHO is a practice-shaping trial rather than the final word. It demonstrates that protocolised low-dose ultrasound-facilitated catheter-directed fibrinolysis can reduce early clinical deterioration in a carefully selected higher-risk intermediate PE population, but it does not yet establish a mortality benefit, definitive bleeding neutrality, or long-term superiority; those questions will determine whether the trial becomes fully practice-changing.

Overall Summary

  • HI-PEITHO provides the first randomised clinical-outcome evidence that ultrasound-facilitated catheter-directed fibrinolysis can outperform anticoagulation alone for preventing early decompensation in carefully selected non-shock PE.

Bibliography