Large-bore Aspiration Thrombectomy with the FlowTriever System for the Treatment of Pulmonary Embolism: A Large Single-Center Retrospective Analysis

Electronic health records of patients treated for PE at a large university medical center between September 2019 and January 2023 were retrospectively analyzed. By September 2019, LBAT had largely supplanted catheter-directed thrombolysis (CDT) as the primary interventional PE treatment method at our institution due to immediacy of clinical improvement following a single-session procedure without the need to administer thrombolytics. To assess outcomes with this updated approach, all patients treated with LBAT using the FlowTriever System were included in this analysis, and those treated with CDT were excluded. No other criteria were applied. Patients were treated using an institutional PE treatment protocol which included: 1) imaging via CT, 2) treatment determination by a multidisciplinary Pulmonary Embolism Response Team (PERT) composed of an emergency medicine physician, interventional radiologist, and pulmonary critical care physician, 3) interventional reperfusion with LBAT in appropriate patients per PERT decision, and 4) admission to the progressive care unit following LBAT unless intensive care unit (ICU) admission was needed.

PE risk stratification was performed using the current European Society of Cardiology (ESC) guidelines [5]. Generally, the PERT recommended intermediate-risk PE patients with central bulky thrombus to be treated via an interventional approach, and patients with low-risk or peripheral PE to receive anticoagulation. Risk stratification markers like troponin, brain natriuretic peptide (BNP), and right heart strain, determined independently by diagnostic radiologists using commonly defined imaging findings such as right ventricular-to-left ventricular (RV/LV) ratio > 0.9, were also considered by the PERT in making treatment determinations.

At our institution, a typical LBAT procedure course is as follows. The patient is on anticoagulation before arrival to the procedure room. Additional heparin is given to achieve activated clotting time of 250–300 s, and the patient is placed under conscious sedation. LBAT is initiated by gaining vascular access through the femoral vein. A balloon-tipped or pigtail catheter is navigated through the heart into the pulmonary arteries, and pressure measurements may be taken. Once the pulmonary vasculature has been accessed, the catheter is removed over a stiff guidewire to allow for insertion of the Triever aspiration catheter (16 Fr, 20 Fr or 24 Fr). This is positioned immediately proximal to the embolic debris and aspiration is initiated with a 60 mL syringe attached to the Triever catheter side port. Aspiration is repeated as needed. Mechanical disruption of embolic debris with the FlowTriever disks is not used at our institution. Once it became available (August 2021), the FlowSaver blood return system (Inari Medical) was implemented in the majority of LBAT procedures to filter aspirated blood in the syringe and return it to the patient through the access site. The decision of when to terminate the LBAT procedure is left to the discretion of the interventional radiologist performing the procedure based on thrombus removal and observed improvement in clinical status. At the conclusion of the procedure, hemostasis is achieved using a pre-close technique with two Perclose ProGlide devices (Abbott, Plymouth, MN) or a purse-string suture technique.

The primary safety and effectiveness outcomes assessed included 7- and 30-day all-cause mortality, major bleeding, procedure-associated clinical decompensation, pulmonary vascular or cardiac injury, and pulmonary artery pressure (PAP) reductions. Major bleeding was determined using the Society of Interventional Radiology (SIR) reporting standards, i.e., intracranial, intraocular, or retroperitoneal hemorrhage or any hemorrhage requiring transfusion and/or resulting in a hematocrit decrease ≥ 15% or hemoglobin decrease ≥ 5 g/dL [15]. All clinical decompensation events, defined as an acute physiological worsening of clinical status that posed an immediate increased risk of serious harm or death [16], occurring during hospitalization whether before, during, or after the thrombectomy procedure were recorded. PAPs were measured immediately prior to and following thrombectomy via standard invasive hemodynamic assessment techniques. Additional outcomes included technical success per the SIR reporting standards, i.e., the ability to deliver the device into the pulmonary artery, operate the device to aspirate thrombus, and remove the device at the end of the procedure [15], change in hemodynamic measurements, and postprocedural hospital length of stay (LOS).

Data values are reported as counts (%) for categorical variables and mean ± standard deviation (SD) or median [interquartile range (IQR)] for continuous variables. For comparative analysis of changes in hemodynamics and oxygenation status following thrombectomy, p-values were calculated using two-sided paired Student t tests for continuous data and McNemar tests for categorical data. Change in PAP following thrombectomy was analyzed in patients with paired pre-procedure and post-procedure measurements, and p-values were calculated using one-sided paired Student t tests. A p-value < 0.05 was considered statistically significant. Statistical analysis was performed in R, version 4.3.2.

This research activity was granted exemption from full Institutional Review Board review and waived informed consent by a qualified staff member of the Human Research Protection Program of the affiliated university in accordance with 45 CFR 164.512(i)(2)(ii).

During the study period, 319 PE patients were treated with catheter-directed therapy, including 33 who underwent CDT and 286 who underwent LBAT. All 286 patients treated with LBAT were included in the analysis population, while the 33 CDT patients were excluded. As depicted in Table 1, the analysis population was 53.5% male, with a mean age of 60.5 ± 16.2 years and a mean BMI of 35.2 ± 10.6 kg/m². Concurrent deep vein thrombosis was present in 62.9% of patients, 18.9% had a history of malignancy, and 16.1% were diagnosed with PE within 3 months of having surgery. Intermediate-high-risk PE was diagnosed in 75.2% of patients, while 15.7% presented with intermediate-low-risk PE, 8.0% with high-risk PE, and 1.0% with low-risk PE. Right ventricle dilatation on CT was identified in 87.4% of the cohort, while 95.5% had elevated cardiac biomarkers.

Table 2 shows the procedural characteristics and LOS data for the cohort. The mean total procedure time from pre-procedure time out to access site closure was 90.5 ± 32.3 min. No patients were treated with thrombolytic therapy intraprocedurally, but 4.5% (n = 13) received thrombolytic therapy at an outside hospital prior to the procedure. The mean post-procedural hospital LOS was 7.3 ± 9.1 days and the mean post-procedural ICU LOS was 2.0 ± 4.1 days, with 49.3% of patients not requiring ICU admittance. Among 145 (50.7%) patients with post-procedure ICU admittance, the mean LOS was 3.9 ± 5.1 days.

Technical success was achieved in 277 (96.9%) cases. Of the 9 technical failures, 7 involved emboli that could not be aspirated, 1 procedure could not be completed for undocumented reasons, and 1 procedure was terminated due to a pulmonary artery pseudoaneurysm, details of which are described below. In the 7 cases involving emboli that could not be aspirated, 6 were due to intractable chronic thrombus and 1 was due to very large central thrombus size. Of the 6 patients with intractable chronic thrombus, 1 expired during the procedure and 5 continued medical management with anticoagulation and survived through 30 days. The patient with central thrombus that was too large to aspirate died of uterine hemorrhage in the setting of endometrial cancer 18 days following the procedure.

Table 3 shows changes in hemodynamics and oxygenation status at presentation and post-procedure. Prior to thrombectomy, 183 (64.0%) of patients were tachycardiac (heart rate ≥ 100 beats per minute [bpm]). Of these tachycardiac patients, 54.1% experienced resolution of their tachycardia immediately following the procedure. The mean reduction in heart rate among the entire cohort was 13.5 bpm (p < 0.0001). There was not a statistically significant change in the number of patients who required supplemental oxygen prior to and immediately following thrombectomy (n = 180 [62.9%] vs n = 165 [57.7%]; p = 0.0997), though 27.2% of patients with supplemental oxygen requirements at presentation no longer required oxygen support following the procedure.

Figure 1 depicts the change in PAPs among patients with paired pre- and post-thrombectomy measurements. The mean reduction in mean and systolic PAP immediately following thrombectomy was 6.8 mmHg (28.7 ± 9.0 to 21.9 ± 8.0 mmHg; p < 0.0001) and 11.7 mmHg (46.0 ± 14.0 to 34.3 ± 12.0; p < 0.0001), respectively.

Safety and mortality outcomes for the cohort are shown in Table 4. There were 2 (0.7%) major bleeds consisting of 1 case of abdominal bleeding detected post-procedure likely due to rebleed of a retroperitoneal hematoma and 1 case resulting from intraprocedural laceration of the corona mortis which occurred during sheath insertion and was treated with coil embolization. Two (0.7%) pulmonary vascular injuries occurred: 1 case of contrast extravasation into the mediastinum observed on postprocedural angiogram and 1 pulmonary artery pseudoaneurysm. The pulmonary artery pseudoaneurysm was identified intraprocedurally when a 1-cm outpouching was noted along the main pulmonary artery immediately distal to the pulmonic valve. Immediate CTA of the chest confirmed the diagnosis, and no corrective procedure was performed. At 1-month follow-up the pseudoaneurysm was measured and documented stable, and at 1-year follow-up, the pseudoaneurysm was not appreciable on CTA. All four major bleeding or pulmonary vascular injury events observed in the study occurred in patients with intermediate-risk PE.

There were 4 (1.4%) procedure-associated decompensation events, consisting of 3 (1.0%) intraprocedural cardiac arrests and 1 (0.4%) stroke. These 4 events appeared to be exacerbations of ongoing decompensations and not new events that began during or after the procedure. Of the 3 intraprocedural cardiac arrests, 2 occurred in high-risk patients who also experienced cardiac arrest prior to the procedure (1 expired during the procedure and 1 expired in the ICU following the procedure). The remaining intraprocedural cardiac arrest occurred in an intermediate-low-risk patient who had multiple seizures prior to the procedure and an intraprocedural seizure requiring resuscitation; this patient survived to discharge. The stroke occurred in a high-risk patient with arrest, an impaired neurological exam, and CT suggestive of anoxic brain injury prior to the procedure. PE thrombectomy was performed expediently and thereafter, assessment revealed the patient had suffered a stroke. The stroke was considered procedure-associated because the occurrence of intraprocedural paradoxical embolism via a patent foramen ovale could not be ruled out. Twenty patients had clinical decompensation events not associated with the procedure: 5 additional high-risk patients experienced cardiac arrest prior to the procedure, 11 intermediate-risk patients experienced pre-procedural decompensation related to other conditions (cancer n = 4, pneumonia n = 4, congestive heart failure n = 1, sepsis n = 1, severe lactic acidosis n = 1), and 4 intermediate-risk patients decompensated following the procedure for unrelated reasons (cardiac arrest in the setting of palliative care, hemorrhagic shock, stroke, and adrenal insufficiency; n = 1 each). Pre-procedural decompensation in the 4 patients presenting with known cancer diagnosis was attributable to superimposed pneumonia, sepsis, metabolic acidosis, and hemorrhage.

At 7 days post-procedure, all-cause mortality for the full cohort was 2.4% (n = 7), including 21.7% (n = 5) for patients with high-risk PE, 0.9% (n = 2) with intermediate-high-risk PE, and 0% with intermediate-low- or low-risk PE. No deaths were attributed to the device, relationship to the procedure could not be ruled out. All 5 high-risk patients who expired within 7 days had experienced cardiac arrest prior to the procedure and expired while hospitalized. One additional high-risk patient died within 30 days after hospitalization while in hospice; this patient also experienced cardiac arrest prior to the procedure. At 30 days post-procedure, all-cause mortality for the full cohort was 6.7% (n = 19), including 26.1% (n = 6) for patients with high-risk PE, 4.7% (n = 10) with intermediate-high-risk PE, and 6.3% (n = 3) with intermediate-low- or low-risk PE. PE-related mortality at 30 days post-procedure was 3.5% (n = 10). These PE-related deaths through 30 days included all 6 deaths in high-risk patients, 3 deaths in intermediate-high-risk patients, and 1 death in an intermediate-low-risk patient.