STORM-PE: early right ventricular unloading in intermediate-high risk pulmonary embolism Authors: Hassan Saleh, Firas Amil, Luke Domaleski Published: 2026 Intermediate-high risk pulmonary embolism (PE) has been a therapeutic gray zone in contemporary medicine. These patients are normotensive yet demonstrate right ventricular (RV) dysfunction and biomarker elevation, placing them at measurable risk for early decompensation and mortality (1). The central question is whether anticoagulation alone is sufficient for these elevated risk patients, or whether they benefit from early thrombectomy. STORM-PE represents the first randomized controlled trial directly comparing thrombectomy plus anticoagulation versus anticoagulation alone in this population (2). Its findings therefore, warrant careful interpretation. STORM-PE deserves credit for accomplishing randomization in a high-acuity condition where clinicians still struggle with the best approach for patients who appear stable but are high risk for clinical deterioration. In practice, the decision to pursue advanced therapy is often made within Pulmonary Embolism Response Teams (PERT), which can make equipoise difficult to preserve at the bedside. The large screening funnel (767 screened to enroll 100) and the number of eligible patients who declined participation underscores how difficult it is to run a pragmatic randomized controlled trial in this setting (2,3). At the same time, that same funnel signals an important limitation, the randomized cohort represents a selected subset of intermediate-high risk PE patients treated at experienced centers with structured PE pathways, which as described by the authors, can influence generalizability to hospitals without an established PERT infrastructure (2). The manuscript’s reporting around cancer warrants clarification. In the Results, “active cancer” is listed as one of the most common eligibility criteria not met during screening (14%), yet the main Methods section does not list active cancer among the key inclusion/exclusion criteria, and enrolled patients include a small proportion with active cancer (2.1% in the thrombectomy arm and 5.7% in the anticoagulation arm) (1). This creates confusion for readers trying to understand who the results apply to. At minimum, the trial should define “active cancer” explicitly (e.g., recent systemic therapy, metastatic disease, hematologic malignancy) and reconcile whether active cancer was an operational screening consideration that evolved during enrollment and the rationale behind this. This is not a trivial distinction, patients with active malignancy represent a substantial proportion of PE presentations and represent a distinct population with higher thrombus burden and competing risks for adverse outcomes including bleeding, prolonged hospitalization, escalation of care and overall morbidity and mortality. Ambiguity in how these patients were screened or excluded complicates interpretation of both safety and efficacy outcomes. Although baseline characteristics were broadly similar, the small sample size makes modest imbalances worth acknowledging when interpreting secondary clinical signals. For example, the anticoagulation arm had numerically higher rates of coronary artery disease and congestive heart failure and a higher proportion of patients with both elevated troponin and natriuretic peptides (2). In patients with heart failure and coronary artery disease, detectable cardiac biomarkers may partly reflect underlying myocardial disease with the added burden of PE rather than isolated PE-related RV injury alone, complicating interpretation when biomarker positivity is both an entry criterion and a descriptor of baseline severity (1). This does not negate the imaging core-lab primary endpoint, but it strengthens the argument that larger trials should ensure tighter balance for comorbidities that influence biomarkers and for the distribution of “dual-positive” biomarker profiles. The trial reports access site use and procedural times; however, more procedural granularity would improve interpretability for readers. Specifically, details around periprocedural anticoagulation management (which anticoagulant was used, whether anticoagulation was paused or continued during large-bore venous instrumentation, targets used, timing of heparin restart), venous closure strategy for 16 French access, and standardized post-procedure monitoring protocols would help readers reproduce the approach and contextualize bleeding and hemoglobin trends (2). This matters as the study reports a greater hemoglobin nadir in the thrombectomy arm, and in device-based trials, procedural and access-site practices can materially influence safety outcomes independent of thrombus extraction itself (1,4). The reported median thrombectomy time of 25 minutes and procedure time of 56 minutes suggest a technically efficient intervention and notably much lower than seen in PEERLESS with catheter time of 47.9 minutes and procedure time of 93.2 minutes (2,5). That said, while readers naturally compare procedural metrics with other contemporary PE trials, it is important to acknowledge that cross-trial comparisons can mislead because of differences in patient selection, endpoints, and comparator arms. For example, PEERLESS evaluated large-bore mechanical thrombectomy versus catheter-directed thrombolysis within a population already selected for an advanced-therapy strategy, whereas STORM-PE uniquely includes an anticoagulation-only comparator arm (2,5). Within the bounds of STORM-PE’s design, the procedural efficiency supports feasibility, but it should not be interpreted as superiority versus other devices or strategies without head-to-head trials with shared entry criteria and outcomes. The most convincing contribution of STORM-PE is its randomized demonstration that computer-assisted vacuum thrombectomy (CAVT) produces a larger early reduction in RV/left ventricle (LV) ratio and thrombus burden than anticoagulation alone, measured by a blinded core laboratory (2). The key unresolved question is not whether RV/LV ratio is prognostic at baseline, it likely is, but whether an intervention driven improvement at 24–48 hours reliably translates into fewer hard clinical events or better longer-term recovery (6-9). Whether a 24–48-hour change in RV/LV ratio reflects meaningful recovery or simply a transient hemodynamic unloading remains uncertain. Early mechanistic success is a necessary step toward demonstrating clinical outcome benefit in larger trials; however, it is not synonymous with it. Importantly, the absence of longer-term functional or clinical outcomes limits the ability to determine whether these early imaging improvements translate into sustained benefit. STORM-PE was not powered for mortality and observed few clinical events; its imaging and physiologic signals are best interpreted as mechanistic and hypothesis-generating—strengthening equipoise and providing a biologic rationale for larger, outcome-focused studies rather than settling the question of definitive clinical benefit. Observational data suggesting longer-term outcome separation with thrombectomy, particularly among higher-risk strata and after accounting for confounders such as metastatic cancer (10). These observations underscore the hypothesis while highlighting why randomized outcomes trials remain essential. Overall safety was notable with the thrombectomy group experiencing no procedural related complications (2). However, reporting would benefit from additional information, particularly regarding the BARC 5b bleeding event in the thrombectomy arm. Because this event clusters with other major adverse events in a single patient, readers will want to understand timing relative to the procedure, suspected bleeding source, anticoagulation exposure, and whether there were procedural or access-site issues (2). Even if adjudication concluded that outcomes were not device-related, a clearer narrative would help clinicians contextualize risk, counsel patients, and refine best practices around access management and periprocedural anticoagulation in large-bore venous thrombectomy. It is important to state that large-bore aspiration in the pulmonary arteries is not physiologically neutral, decreased cardiac output from catheters altering RV geometry, tricuspid regurgitation, blood loss secondary to aspiration, and rapid shifts in RV afterload can occur during the case (4). These dynamics may help explain why some patients improve immediately while others experience procedural instability despite technically successful thrombus extraction. Future trials and registries should incorporate standardized intraprocedural physiologic endpoints that capture oxygen delivery and cardiopulmonary reserve in real-time, rather than relying on pre and post imaging. Measurement of the pulmonary artery (PA) oxygen saturation recorded at three standardized time points: (I) at the start of the procedure using a 7 French or smaller catheter representing a baseline saturation; (II) shortly after the thrombectomy device is first positioned in the PA and prior to aspiration to understand how these devices effect output; and (III) after completion of aspiration and after the device is removed representing post-device saturation. A drop in PA sat during the “device-in” phase could signal transient impairment in forward flow and oxygen delivery—whether from partial PA occlusion by the catheter, tricuspid regurgitation or hemodynamic shifts. Capturing this signal would allow operators to quantify a phenomenon that many recognize clinically but rarely measure, that the procedure itself can quickly lead to deterioration in patients with normotensive shock. Imaging modalities such as transthoracic echocardiography, to measure treatment effect in real-time, have not been systematically studied in this setting but may have value, particularly in the highest risk patients. It is important to interpret STORM-PE in the context of the 2026 AHA/ACC/ACCP/ACEP/CHEST/SCAI/SHM/SIR/SVM/SVN guidelines for the evaluation and management of acute PE. The guidelines introduced a new clinical classification system which integrates hemodynamics, RV dysfunction, and biomarkers to guide treatment decisions, effectively reframing what has historically been termed intermediate-low and intermediate-high risk PE into a spectrum ranging from C1 to D2. Within this framework, anticoagulation remains the cornerstone of therapy, while catheter directed therapy is a class 2a or 2b recommendation in select patients (11). The guidelines highlight the heterogeneity of intermediate-high risk PE and the lack of robust randomized data supporting routine early intervention, underscoring the relevance of STORM-PE as the first randomized trial to directly compare thrombectomy plus anticoagulation versus anticoagulation alone. However, these findings remain surrogate in nature and align with the cautious stance of the guidelines regarding routine upfront thrombectomy. Importantly, STORM-PE emphasizes key considerations emphasized by the new guidelines, including careful patient selection, bleeding risk assessment, the role of PERTs and most importantly the endpoint that STORM-PE measured, being clinical deterioration and not mortality. As such, STORM-PE both supports current guideline restraint and strengthens the rationale for larger, outcomes-driven trials that may ultimately better define the role of catheter-based therapies in this high-risk yet clinically heterogeneous population. A reasonable way to interpret STORM-PE is that it facilitates early normalization of both the RV/LV ratio along with hemodynamics. In practice normalization of the RV size and patient hemodynamics can help form a framework for which patients are effectively unloaded. Future trials should explicitly test whether early (<6 hours) improvements across these domains predict durable functional recovery, reduced dyspnea and functional limitations, and fewer hard events, and should define standardized response thresholds that can guide escalation, de-escalation, and discharge planning. Supplementary The article’s supplementary files as Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0751/prf