Case Report: Staged alteplase administration under VA-ECMO support for high-risk pulmonary embolism with severe thrombocytopenia

Introduction

High-risk pulmonary embolism (PE), defined by hemodynamic instability, requires urgent evaluation for reperfusion (1, 2). Systemic thrombolysis is an option when bleeding risk is acceptable, but major bleeding and stroke remain concerns (1–3). Severe thrombocytopenia, extracorporeal membrane oxygenation (ECMO), and concurrent anticoagulation intensify this conflict (4–7). The PERT Consortium lists platelets below 100 × 10⁹/L as a relative contraindication to systemic thrombolysis (8), yet PE-specific evidence does not validate a safe threshold or regimen for severe thrombocytopenia (4, 9).

Venoarterial ECMO (VA-ECMO) supports perfusion while clinicians consider recovery or further reperfusion, but does not remove pulmonary arterial obstruction (5–7). Reduced-dose thrombolysis evidence largely concerns other PE populations (10, 11). This report describes staged alteplase administration, reassessment, and bleeding observations during VA-ECMO-supported treatment.

Case description

Patient information and initial presentation

A 28-year-old man (184 cm, 85 kg; body mass index 25.1 kg/m²) presented after a month of cough, five days of worsening cough and fever, and abrupt chest tightness and syncope four hours before transfer. He had received treatment for presumed pulmonary infection at another hospital. His history included allergic rhinitis and a recent pruritic rash. Day 0 denotes presentation to our hospital. At 08:21, blood pressure was 70/45 mmHg, heart rate 142 beats/min, and extremities were cold. Vasopressors and noninvasive ventilation were started (Figure 1A).

Table 1. Auditable treatment parameters and key monitoring data in this case.

Day 0 denotes presentation to our hospital; t0 is first alteplase initiation at 17:00. Times are medication-event or sampling times unless stated otherwise. Vasoactive preparations are from medication orders, pump-rate changes from nursing records, and 19:03 hemostasis from the progress note. UFH restart timing was clarified retrospectively; approximately 20:25 is derived, not an independently recorded timestamp. Platelet count at first alteplase initiation was unknown after intervening transfusion. Apheresis platelets are reported as institutional therapeutic doses and cryoprecipitate as charted units. Each alteplase stage lasted approximately 1 h; exact completion timestamps and volumetric rates were unavailable. The treatment sequence was individualized, not a validated dosing algorithm.

AEC, absolute eosinophil count; aPTT, activated partial thromboplastin time; BP, blood pressure; CTEPH, chronic thromboembolic pulmonary hypertension; CTPA, computed tomography pulmonary angiography; DIC, disseminated intravascular coagulation; FDP, fibrin degradation products; FEU, fibrinogen-equivalent units; FiO₂, fraction of inspired oxygen; Hb, hemoglobin; HR, heart rate; ICU, intensive care unit; INR, international normalized ratio; LVEF, left ventricular ejection fraction; PaCO₂, arterial carbon dioxide partial pressure; PaO₂, arterial oxygen partial pressure; PASP, pulmonary artery systolic pressure; PLT, platelet count; PT, prothrombin time; TAPSE, tricuspid annular plane systolic excursion; HIT, heparin-induced thrombocytopenia; UFH, unfractionated heparin; VA-ECMO, venoarterial extracorporeal membrane oxygenation; WBC, white blood cell count.

Platelets fell from 85 × 10⁹/L on Day −3 to 21 × 10⁹/L at 08:35 on Day 0, before ECMO. Further verification established that heparin had not been used during the recent outside-hospital admission. The platelet decline preceded the first heparin administration at our institution, a temporal sequence that does not support prior-heparin-associated heparin-induced thrombocytopenia (HIT) as the explanation for the initial thrombocytopenia. No formal 4Ts score was documented, and platelet factor 4 (PF4)/heparin antibody and functional testing were not performed; HIT was therefore not definitively excluded.

Diagnostic assessment

At 09:20, echocardiography showed right atrial and ventricular enlargement, left ventricular ejection fraction (LVEF) 48%, estimated pulmonary artery systolic pressure (PASP) 52 mmHg, moderate tricuspid regurgitation, and a 14 × 9-mm echogenic focus at the pulmonary artery bifurcation. Ultrasonography identified right femoral venous thrombosis. After bedside ECMO-team assessment, VA-ECMO was initiated at 10:20 for progressive circulatory failure. Computed tomography pulmonary angiography (CTPA) after initial stabilization at 11:58 showed filling defects in the main and branch pulmonary arteries (Figures 2A,D). The echogenic focus was not specifically reassessed later.

At 11:55, platelets were 21 × 10⁹/L, hemoglobin 150 g/L, prothrombin time (PT) 16.8 s, international normalized ratio (INR) 1.38, activated partial thromboplastin time (aPTT) 43.5 s, and fibrinogen 1.18 g/L. D-dimer exceeded 20 mg/L fibrinogen-equivalent units (FEU) and fibrin degradation products (FDP) exceeded 150 mg/L. Complete laboratory values and serial arterial blood gases are provided in Table 1.

Hematology reviewed the progressive thrombocytopenia on Day 0 and documented an uncertain cause. Acute thrombosis/critical-illness-related consumption, immune or infectious causes, and drug-related thrombocytopenia were considered in the differential. Available autoimmune screening was unrevealing, but did not exclude all alternative causes. Although the treating team retrospectively reported that disseminated intravascular coagulation (DIC) had not been favored, no formal score was documented. The 11:55 findings could meet the traditional International Society on Thrombosis and Haemostasis (ISTH) overt-DIC laboratory threshold if the D-dimer elevation were scored as strong: platelets contributed 2 points and fibrin-related markers would contribute 3. Fibrinogen of 1.18 g/L contributed 0 points; the PT component could not be assigned without the laboratory control value. DIC therefore could not be excluded. This sample was obtained after ECMO initiation, limiting interpretation of the pretreatment coagulation state but not invalidating the laboratory findings. The cause of thrombocytopenia remained unresolved, and ECMO cannot explain the decline that preceded cannulation.

Antinuclear, anti-double-stranded DNA, antineutrophil cytoplasmic, anti-glomerular basement membrane, anticardiolipin, and anti-β₂-glycoprotein I antibody testing was unrevealing. The normalized lupus anticoagulant ratio was 1.12. Protein C and antithrombin activity were within reference ranges; protein S was 74% (reference 75%–130%), with interpretation limited by acute thrombosis and anticoagulation. BCR::ABL1 testing was negative; bone marrow and tissue examinations were not performed. Absolute eosinophil count (AEC) was 3.16 × 10⁹/L on Day −3, 2.93 × 10⁹/L at 08:35 on Day 0, and 0.68 × 10⁹/L at 11:55 after glucocorticoids. Homocysteine was 54 μmol/L on Day 1.

Therapeutic intervention

Peripheral VA-ECMO used a 17-Fr femoral arterial and a 21-Fr femoral venous cannula, with distal perfusion. Cannulation unfractionated heparin (UFH) was charted as 25 mg; corresponding activity units were unavailable. The team selected UFH for anticoagulation during acute PE treatment and VA-ECMO support because it could be titrated to coagulation measurements and bleeding, had a short duration of action, and could be stopped and reversed with protamine if necessary. Methylprednisolone 40 mg was given at 08:21 and prescribed twice daily through Day 3. A separate dexamethasone entry prevented complete reconstruction of glucocorticoid exposure.

Vasoactive therapy had begun during emergency stabilization at 08:21. The ICU medication orders entered at 11:50, after VA-ECMO initiation at 10:20, listed dopamine hydrochloride 200 mg and norepinephrine bitartrate 8 mg, each added to 50 mL of 5% glucose. Nursing records documented dopamine/norepinephrine rates of 8/2 mL/h during this initial ICU treatment period, increasing to 12/5 mL/h before the first alteplase infusion. Final preparation volumes were not independently verified, so the recorded pump rates and preparation orders are reported without conversion to weight-adjusted doses. Post-ECMO lactate rose from 2.0 mmol/L at 11:55 to 3.2 mmol/L at 14:25. These findings raised concern about persistent support dependence.

The treating team retrospectively confirmed multidisciplinary consideration of surgical embolectomy, catheter-based thrombectomy, and systemic thrombolysis. Surgical trauma and perioperative bleeding were major concerns in the setting of severe thrombocytopenia. Catheter-based treatment required transfer to an operating room. The team recalled marked intolerance of even minor movement, with a sense of impending death, both before and after VA-ECMO initiation. Although CTPA had been completed after initial stabilization, concern about further circulatory deterioration during subsequent transfer remained. Bedside systemic thrombolysis was therefore selected after an individualized risk–benefit assessment.

The initial 20-mg alteplase dose was selected by individual clinical judgment because of the competing risks of persistent pulmonary arterial obstruction and bleeding. The intention was to assess the clinical response and bleeding before deciding whether to administer a further dose. This was not a guideline-defined regimen or a documented institutional protocol; expressing the total dose as 0.47 mg/kg describes exposure rather than the basis for dose selection.

Hematology recommended platelet transfusion according to bleeding and serial blood counts; the transfusion note documented cryoprecipitate for coagulation-factor replacement. Two institutional therapeutic doses of apheresis platelets were transfused from 14:50 to 15:10, with plasma, fibrinogen, and cryoprecipitate support before alteplase. No target platelet or fibrinogen threshold was documented, and no platelet count was obtained after transfusion but before the first infusion. The rise from 21 to 62 × 10⁹/L was interpreted primarily in relation to platelet transfusion, not as spontaneous recovery or an alteplase effect.

At 17:00 (t0), UFH was paused and the first 20-mg alteplase infusion began through right lower-limb deep venous access (Figures 1B). Arterial cannulation-site bleeding occurred at 17:20 and was managed with compression.

The 17:39 sample showed platelets of 62 × 10⁹/L, fibrinogen of 1.38 g/L, and aPTT of 47.5 s after hemostatic support. Nursing records documented a reduction in dopamine and norepinephrine rates to 8 and 2 mL/h at 18:10. The progress note documented effective compression and cessation of arterial cannulation-site bleeding at the 19:03 reassessment. Despite initial improvement, VA-ECMO and vasoactive support remained necessary. The team judged that further reperfusion might provide benefit after bleeding control and ordered another 20 mg at 19:03; infusion began at 19:25. No prespecified numerical escalation or stopping criteria were documented. Each stage lasted approximately 1 h; the total administered dose was 40 mg (0.47 mg/kg). Exact completion timestamps and volumetric alteplase rates were unavailable.

UFH was restarted at 6 mL/h after the second alteplase infusion, approximately 1 h after its 19:25 start, according to the treating team. The implied time of approximately 20:25 is not an independent execution timestamp. Preparation details and limitations are reported in Table 1. UFH was subsequently adjusted by aPTT through Day 10. Later orders included enoxaparin 4,000 anti-Xa IU every 12 h and fondaparinux 7.5 mg daily; discharge anticoagulation was edoxaban 60 mg daily. The patient-specific aPTT or anti-Xa target, exact transition times, and reasons for each switch could not be reconstructed. Additional plasma and cryoprecipitate were given; fibrinogen was 1.53 g/L at 00:41 on Day 1.

Follow-up and outcomes

The treating team confirmed that vasopressors were discontinued on Day 1; formal order discontinuation on Day 3 reflected subsequent order reconciliation. Repeat CTPA on Day 6 showed a lower pulmonary arterial thrombus burden than at baseline, although residual filling defects remained (Figures 2B,E). VA-ECMO was removed on Day 7 after a brief reduction and clamp assessment; the operative record documented 50 mL of blood loss.

The treating team reported no pleural or abdominal fluid collection or visible retroperitoneal hematoma on bedside ultrasonography during the bleeding assessment. The patient remained alert without neurological signs suggestive of intracranial hemorrhage; these observations did not exclude occult bleeding. Hemoglobin decreased from 150 g/L at 11:55 to 136 g/L at 17:39 and 117 g/L at 20:13 on Day 0. On Day 8, hemoglobin was 75 g/L, and two units of leukocyte-reduced red cells were transfused from 10:45 to 12:45 for a continuing hemoglobin decline in a critically ill patient.

Platelet counts fluctuated during continued UFH treatment: 65 and 64 × 10⁹/L on Days 1 and 2, 91 then 54 × 10⁹/L on Day 4, and 129, 173, and 160 × 10⁹/L on Days 8–10, respectively (Table 1). No additional platelet transfusions were documented in the available transfusion records and temporary medication orders from Day 1 through Day 10. This trajectory indicates eventual recovery, not a continuous rise or definitive exclusion of HIT.

Estimated PASP was 45 mmHg on Day 1 and 28 mmHg on Day 16. The patient left intensive care on Day 11 and was discharged on Day 19 on edoxaban 60 mg and folic acid 1.6 mg daily; no glucocorticoid was listed. AEC was 1.53 × 10⁹/L on Day 43 and 0.47 × 10⁹/L on Day 103. Day 103 CTPA showed no obvious residual filling defect; echocardiography and lower-limb ultrasonography were reported without obvious abnormality (Figures 2C,F). Ventilation–perfusion scanning and right-heart catheterization were not performed.

Discussion

VA-ECMO-supported reperfusion under competing risks

This case describes individualized reperfusion decisions during VA-ECMO support in a patient with severe thrombocytopenia. The platelet count of 21 × 10⁹/L preceded transfusion and must not be treated as the count at alteplase initiation.

VA-ECMO can stabilize perfusion and unload the failing right ventricle, but it does not directly remove pulmonary arterial obstruction. Evidence for adding reperfusion therapy during VA-ECMO remains observational and inconsistent. In a Japanese nationwide cohort of 1,220 patients, systemic thrombolysis within two days of VA-ECMO was not associated with lower in-hospital mortality; a sensitivity analysis restricted to monteplase showed more major bleeding (12). A three-center cohort likewise found no difference in 90-day survival or severe hemorrhage between patients cannulated after failed thrombolysis and those who had not received thrombolysis, although the comparison was strongly affected by treatment selection (13). The ELSO registry describes substantial mortality across ECMO modes but cannot determine the effect of a particular reperfusion strategy (14). More recently, early mechanical reperfusion during VA-ECMO was not associated with a statistically significant reduction in 90-day mortality in a 39-center cohort, and bleeding occurred in half of the patients (15). Taken together, these cohorts define the uncertainty within which individual decisions are made rather than a preferred treatment combination.

Baran et al. described a single-center series of 27 high-risk PE patients in which ECMO was considered alongside surgical or catheter-based reperfusion, with initiation tailored to the perioperative course (16). This experience illustrates ECMO as support for a treatment pathway rather than a substitute for selecting a reperfusion strategy. Mechanical clot removal, continued ECMO with anticoagulation, and bedside systemic thrombolysis entail different procedural, transport, and bleeding risks.

A randomized trial compared 50 mg with 100 mg over 2 h, but did not study severe thrombocytopenia during VA-ECMO (17). A retrospective comparison found more rescue treatment after 50 mg without a clear bleeding advantage (18). Other studies used heterogeneous regimens and populations (10, 11). Differences in populations and regimens mean that these studies provide only indirect context for the staged 20-mg doses used in this case.

A 2022 systematic review identified 11 cases of PE with thrombocytopenia, with inconsistent platelet sampling times and no comparative evidence (4). Later reviews reached similar conclusions (9). Selected reports involved malignancy, thrombotic thrombocytopenic purpura, HIT, lymphoma, and postpartum antiphospholipid syndrome (19–24). ECMO was not reported in the direct cases summarized in Table 2. These selected experiences cannot define a safe platelet threshold and are susceptible to publication and survivor bias (4, 9).

Table 2. Selected directly relevant reports of systemic thrombolysis for pulmonary embolism with thrombocytopenia.

Selected directly relevant adult reports were identified through a focused PubMed search and reference chaining updated through 21 August 2026. This narrative comparison places the present case in context and is not an exhaustive systematic review. Reports were prioritized when pulmonary embolism was objectively diagnosed, thrombocytopenia was present, and systemic thrombolysis was administered; pediatric cases, catheter-directed thrombolysis, and rt-PA used for indications other than pulmonary embolism were not considered. The reviews by Ata et al., Burns and McLaren, and Boca et al. had different scopes; review totals were not added to the individually listed cases. ECMO is shown as not reported when the source did not describe its use. Platelet counts were extracted at the nearest documented pre-thrombolysis timepoint where possible. In the present case, two institutional therapeutic doses of apheresis platelets were transfused between the last documented count and alteplase initiation; 21 × 10⁹/L must not be interpreted as the post-transfusion count at treatment initiation.

ECMO, extracorporeal membrane oxygenation; Hb, hemoglobin; HIT, heparin-induced thrombocytopenia; PE, pulmonary embolism; rt-PA, recombinant tissue-type plasminogen activator; VA-ECMO, venoarterial extracorporeal membrane oxygenation.

Treatment sequence and bleeding risk

The post-transfusion platelet rise does not establish spontaneous recovery or an alteplase effect. Neither platelet nor fibrinogen values were prespecified dose-escalation triggers. Support requirements decreased after the first stage, and the decision to administer the second stage depended on clinical reassessment rather than demonstrated failure of the first dose.

The Day 0 hemoglobin decline of 33 g/L exceeded the numerical ISTH major-bleeding threshold and fell within the BARC 3a hemoglobin range. However, the proportion attributable to bleeding and the relationship between the cannulation-site event and Day 8 transfusion could not be established; dilution, hemolysis, and phlebotomy were not quantified. We therefore report a potentially major bleeding event without assigning a definitive retrospective grade, rather than classifying it as confirmed non-major bleeding (25, 26).

Hypereosinophilia and alternative prothrombotic factors

The eosinophil counts of 3.16 and 1.53 × 10⁹/L, 46 days apart, met the numerical and confirmatory interval requirements for hypereosinophilia (27). Hypereosinophilic syndrome (HES) was not established because eosinophil-attributable organ injury and an adequate etiologic evaluation were lacking. BCR::ABL1 testing was negative, but this does not exclude all clonal eosinophilic disorders.

Homocysteine was 54 μmol/L on a single acute-phase measurement. The HOPE-2 VTE analysis found no reduction in VTE with homocysteine-lowering vitamin therapy (28), but did not resolve causality at this patient's concentration. Without repeat measurements and folate/vitamin B12 assessment, neither finding establishes the cause of PE.

Follow-up, limitations, and clinical implications

Day 103 imaging was reassuring but did not formally exclude chronic thromboembolic pulmonary hypertension because ventilation–perfusion imaging, right-heart catheterization, and structured symptom assessment were unavailable (1, 2).

Continuous ECMO blood flow, pump speed, sweep-gas flow, and oxygenator gas oxygen fraction records were unavailable, and serial quantitative right ventricular measurements around thrombolysis could not be reconstructed. Available lactate and blood gas values were obtained under extracorporeal and respiratory support. These gaps prevent reconstruction of changes in support intensity and limit attribution of physiological improvement to reperfusion.

This retrospective single-case report cannot separate the effects of alteplase from VA-ECMO, anticoagulation, and hemostatic support. Treatment rationales clarified retrospectively by the team are distinguished from contemporaneous observations. Missing continuous ECMO settings, incomplete anticoagulation details, and uncertain bleeding attribution limit interpretation. The case illustrates individualized reassessment under competing risks; it neither establishes the efficacy or safety of staged alteplase nor supports 20 mg followed by 20 mg as a general dosing regimen.

Patient perspective

At follow-up, the patient reported limited recollection of the initial resuscitation and VA-ECMO treatment. After learning how critical his condition had been, he expressed relief about his recovery. He described a gradual return to his usual daily activities after discharge and was satisfied with his progress at the three-month follow-up. He agreed to share his experience in the hope that it might help patients and clinicians facing similarly difficult treatment decisions.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. The authors declare that generative artificial intelligence was used solely for language editing during the preparation of this manuscript. OpenAI ChatGPT/Codex and Anthropic Claude were used only to improve the grammar, clarity, and readability of the text. These tools were not used to generate or analyze clinical data, make clinical decisions, conduct the literature search, select references, or create or modify medical images. All AI-assisted language edits were reviewed and approved by the authors, who take full responsibility for the final manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Luca Spiezia, University of Padua, Italy

Reviewed by: Arun Beeman, GOSH UCL Biomedical Research Centre, United Kingdom

Cagdas Baran, Ankara University, Türkiye