Abstract
Objective
This study aimed to evaluate how the choice of intervention modality for pulmonary embolism (PE) is influenced by patient and hospital characteristics. We compared catheter-directed thrombolysis (CDT), percutaneous mechanical thrombectomy (PMT), and surgical embolectomy.
Methods
Utilizing the Maryland statewide database from the Health Services Cost Review Commission (HSCRC), we analyzed interventions for PE over a 6-year period (fiscal years 2019-2024), focusing on CDT, PMT, and surgical embolectomy. Key hospital factors, including trauma center status, hospital size, Leapfrog safety ratings, extracorporeal membrane oxygenation (ECMO) availability, and PE volume, were assessed for their association with intervention choice. Patient characteristics were assessed for their association with intervention choice including demographics, comorbidities, and health insurance status. Univariate and multivariate statistics were performed to assess the relationship between patient and hospital characteristics and PE interventions.
Results
Over time the utilization of PMT increased with a concomitant decrease in utilization of CDT. CDT was more commonly performed in trauma centers, smaller hospitals, Leapfrog B to D rated hospitals, and hospitals with lower PE volumes. In contrast, PMT was more frequently performed in non-trauma centers, non-ECMO programs, larger hospitals, Leapfrog A rated hospitals, and hospitals with higher PE volumes. Surgical embolectomy was primarily performed in high-volume centers equipped with cardiac surgery and ECMO capabilities. African American patients and those with higher social vulnerability index were more likely to undergo CDT. Comorbidity profiles increased progressively from CDT to PMT to surgical embolectomy. Surgical embolectomy patients were generally younger and more likely to be on Medicaid, whereas PMT/CDT patients were more commonly covered by Medicare. Patients undergoing surgery had higher rates of transfer from other facilities. There were no significant differences in rates of non-routine discharge or mortality across the interventions. In the multivariate model, presence of a cardiac surgery program was associated with increased odds of PMT.
Conclusions
PMT has been adopted at a higher rate than CDT in Maryland over the past 6 years, potentially due to benefits such as reduced length of stay and intensive care unit requirements. Our findings demonstrate that both patient and hospital characteristics influence the modality of PE intervention. These results highlight significant disparities based on race, social vulnerability, and hospital characteristics that warrant systemic attention.
Keywords: Administrative data/registries, Catheter-directed therapies, Pulmonary embolism, Social Determinants of Health
Article Highlights.
-
•
Type of Research: Analysis of statewide registry data: Maryland Health Services Cost Review Commission
-
•
Key Findings: Utilization of percutaneous mechanical thrombectomy (PMT) utilization increased from 20% to 96%, whereas catheter-directed thrombolysis (CDT) decreased from 51.7% to 1.7%. Trauma designation, higher social vulnerability index, and African American race were associated with CDT. Larger pulmonary embolism volume, larger hospitals, and White race were associated with PMT.
-
•
Take Home Message: Hospital and patient characteristics significantly affect the treatment modality for pulmonary embolism. These findings highlight significant disparities based on race, social vulnerability, and hospital characteristics that warrant systemic attention.
Acute pulmonary embolism is the third leading cause of cardiovascular death in the United States.1 Over the past 15 years, there has been an increase in the incidence of symptomatic acute pulmonary embolism (PE) as well as increased use of novel interventions aside from systemic anticoagulation or thrombolysis to treat patients with acute PE, including percutaneous mechanical thrombectomy (PMT) and catheter-directed thrombolysis (CDT).2 Despite the increased utilization of advanced therapies for PE, there has not been a significant change in the rate of in-hospital mortality.3
The safety of catheter-directed interventions (CDIs) for PE has been demonstrated in a variety of studies.4, 5, 6 Despite this favorable safety profile demonstrated in industry-sponsored registries, there is still a lack of consensus regarding ideal patient selection for intervention vs anticoagulation and/or systemic thrombolysis. Considering this uncertainty there has been an increase in the presence of pulmonary embolism response teams (PERTs) of various iterations in hospitals—some more structured than others. Based on data from a single institution, the implementation of PERTs has been shown to improve clinical outcomes, and interestingly, also leads to an increase in the use of catheter-directed interventions.7 Other studies have shown that implementation of PERTs led to shorter hospital stays but did not substantially improve outcomes.8 Along with variability in PERTs, there are studies examining variability in use of advanced therapies for PE.3 These reports do not differentiate between large bore mechanical thrombectomy vs CDT. Additional studies demonstrate disparate application of advanced therapies for PE based on patient demographics, again not differentiating between PMT and CDT.9
In the state of Maryland, there are a variety of hospitals including academic, community, and regional hospitals that have access to advanced interventions for PE, very similar to other statewide variability. Given the lack of guidelines or consensus on which catheter-based therapies are superior and which patients should be treated, we sought to further explore the rates of advanced intervention for PE, including PMT, catheter-directed lysis, and surgical embolectomy. How these treatment modalities may be associated with variability in hospital and patient factors was also examined. A variety of hospital factors were analyzed by variability in PMT vs CDT. Similarly, patient features were analyzed by type of intervention performed. Analyzing these factors may reveal how strategies for PE management are utilized across systems and why some patients may undergo variable procedures for PE intervention.
Methods
Study population
We performed a retrospective review of the Maryland Health Services Cost Review Commission (HSCRC) database from July 1, 2018, to June 30, 2024 (fiscal years 2019-2024; fiscal year defined as July 1 through June 30). Established in 1971, the HSCRC is an independent state agency responsible for regulating rates for inpatient, outpatient, and emergency services at Maryland hospitals.10 As part of a waiver from national Centers for Medicare and Medicaid Services (CMS) reimbursement policy, Maryland hospitals are reimbursed using all-payer rates established by the HSCRC.11 Annual submission of hospital- and patient-level data is required for ongoing operational and quality monitoring, and data is subjected to rigorous auditing and compliance measurement; de-identified patient level datasets are publicly available upon request.12,13 As of the time of writing, 47 acute-general hospitals are regulated by the HSCRC and included in the state registry. Use of the database was approved by the HSCRC data review board prior to initiation of the study.
All patients included in the study presented to an acute care hospital in Maryland, with acute PE with or without cor pulmonale within the study window. All study patients were admitted for inpatient management and underwent advanced PE intervention—CDT, PMT, or surgical embolectomy. HSCRC uses 5-year age groups to protect patient information. Given this, patients <20 years of age were excluded as opposed to the traditional 18 years of age to exclude pediatric patients. Additionally, patients presenting with chronic PE, and those treated at hospitals without publicly available characteristics (as described below) were excluded. In total 906 patients met the inclusion criteria. Over the study period, 52,910 total patients were discharged with a PE diagnosis code, indicating advanced interventions were performed in 1.7% of patients. Of note, this represents all PE patients and not only those with intermediate or high-risk PE. The full list of International Classification of Disease 10th Edition (ICD-10) codes used to identify the patient population, intervention groups, comorbidities, and outcomes are presented in the Supplementary Appendix (online only).
Outcome measures
The primary outcome of interest was the type of advanced PE intervention performed (CDT, PMT, or surgical embolectomy). Secondary outcomes assessed included concomitant treatments, in-hospital complications, length of stay (LOS), non-routine discharge, in-hospital mortality, and total hospitalization charges.
Independent variables
The independent variables included patient and hospital characteristics. Patients’ age (grouped in 10-year intervals), sex, race, and insurance type were recorded. Comorbidities of interest included history of myocardial infarction (MI), congestive heart failure (CHF), peripheral vascular disease (PVD), chronic obstructive pulmonary disease (COPD), renal disease, malignancy, and metastatic solid tumor. Patient acuity was measured using the severity of illness (SOI) score assigned to each hospitalization, measured on a scale of 1 (least severe disease) to 4 (most severe).14 Socioeconomic status was measured using the Centers for Disease Control and Prevention and Agency for Toxic Substances and Disease Registry (CDC/ATSDR) Social Vulnerability Index (SVI). Patients were assigned SVI scores based on the average SVI score of the census-tracts within their ZIP code of residence. Overall social vulnerability and the four dimensions of the SVI (ie, socioeconomic status, household characteristics, racial and ethnic minority status, and housing type and transportation) were assessed.15 Scores range from 0 (least vulnerability) to 1 (most vulnerability) and depict the region’s percentile ranking nationally. Lastly, arrival source (transfer from other acute care hospital vs other) and the presence of a saddle PE were recorded.
Hospital characteristics were collected from publicly available data sources. The number of licensed beds were categorized into quartiles based on all Maryland hospitals. Advanced PE intervention volume was categorized into quartiles based on hospital volume in the HSCRC dataset. Hospitals were then classified by the presence or absence of a general surgery residency program, extracorporeal membrane oxygenation (ECMO) program, cardiac surgery program, trauma center designation, trauma center level (levels 1 to 3), and 2024 LeapFrog Hospital Safety Grade. LeapFrog grades range from A to F and are assigned using 22 CMS safety measures, a proprietary safety survey, and supplemental data.16,17
Data analysis
The rates of advanced intervention use by fiscal year were depicted graphically. Univariate analyses, including χ2 tests and one-way analysis of variance (ANOVA), were performed to compare patient and hospital characteristics between patients undergoing CDT, PMT, and surgical embolectomy. Fisher’s exact test was used when the assumptions of χ2 testing were not met. Outcomes were also compared between treatment types using univariate analyses. Post-hoc between-group comparisons were performed using Bonferroni adjustment. Multivariable logistic regression was then performed to assess the relationship between patient and hospital characteristics and CDT or PMT intervention, with separate models were created for each intervention. Results were reported as odds ratios (ORs) with 95% confidence intervals (CIs). Statistical analysis was performed in SPSS v28.0 (IBM Corp); statistical significance was assessed at P < .05.
Results
During the study period, there was a significant increase in the total number of advanced PE interventions from 60 in 2019 to 360 in 2024. There was a notable increase in the utilization of PMT from 20% of total interventions in 2019 to 96% in 2024, with a concomitant decrease in the utilization of CDT from 51.7% to 1.7% (Fig).
Fig.
Six-year advanced pulmonary embolism (PE) intervention trends.
When comparing patient characteristics on univariate analysis, there was a stark difference in the use of CDT vs PMT between African American (60% vs 41%) and White patients (35% vs 54%; both P < .001). Patients undergoing surgical embolectomy were generally younger and more likely to be covered by Medicaid (P < .001), whereas PMT and CDT patients were more commonly covered by Medicare (P < .001). Overall SVI scores were also significantly different across groups (P < .001), with the highest level of vulnerability observed in CDT patients. Within the SVI dimensions, significant differences were observed for minority status and language, again with CDT patients displaying the highest vulnerability (Table I). The prevalence of MI, CHF, and PVD varied across interventions (all P < .05), with the highest rates observed among surgical embolectomy patients. Similarly, SOI scores varied across interventions (P < .001), again with the highest severity in the surgical embolectomy group. A greater proportion of patients undergoing surgical embolectomy were transferred from another hospital (52.2%) vs 15.1% and 18.4% of CDT and PMT patients, respectively (P < .001) (Table II).
Table I.
Patient demographics by intervention type
| Patient demographics | CDT (n = 119) | PMT (n = 718) | Surgical embolectomy (n = 69) | P-value |
|---|---|---|---|---|
| Age, years | ||||
| 20-29 | 1 (0.8) | 20 (2.8) | 7 (10.1) | .004a |
| 30-39 | 7 (5.9) | 57 (7.9) | 12 (17.4) | .015 |
| 40-49 | 17 (14.3) | 73 (10.2) | 12 (17.4) | .103 |
| 50-59 | 28 (23.5) | 129 (18.0) | 11 (15.9) | .297 |
| 60-69 | 29 (24.4) | 180 (25.1) | 16 (23.2) | .935 |
| 70-79 | 30 (25.2) | 172 (24.0) | 10 (14.5) | .183 |
| 80+ | 7 (5.9) | 87 (12.1) | 1 (1.4) | .005 |
| Female sex | 68 (57.1) | 359 (50.0) | 35 (50.7) | .352 |
| Race | ||||
| White | 42 (35.3) | 387 (53.9) | 29 (42.0) | <.001 |
| African American | 72 (60.5) | 297 (41.4) | 36 (52.2) | <.001 |
| Other | 2 (1.7) | 24 (3.3) | 4 (5.8) | .345a |
| Unknown | 3 (2.5) | 10 (1.4) | 0 (0.0) | .493a |
| Insurance type | ||||
| Commercial | 26 (21.8) | 203 (28.3) | 20 (29.0) | .333 |
| Medicare | 56 (47.1) | 373 (51.9) | 17 (24.6) | <.001 |
| Medicaid | 14 (11.8) | 102 (14.2) | 24 (34.8) | <.001 |
| Uninsured | 3 (2.5) | 8 (1.1) | 1 (1.4) | .274a |
| Other | 20 (16.8) | 32 (4.5) | 7 (10.1) | <.001 |
| Overall SVI | 0.51 ± 0.21 | 0.46 ± 0.19 | 0.46 ± 0.21 | .025 |
| SVI – Socioeconomic Status | 0.49 ± 0.21 | 0.45 ± 0.20 | 0.46 ± 0.21 | .185 |
| SVI – Household Composition and Disability | 0.51 ± 0.15 | 0.48 ± 0.16 | 0.47 ± 0.18 | .179 |
| SVI – Minority Status and Language | 0.72 ± 0.25 | 0.59 ± 0.28 | 0.63 ± 0.25 | <.001 |
| SVI – Housing Type and Transportation | 0.46 ± 0.16 | 0.44 ± 0.17 | 0.42 ± 0.15 | .263 |
CDT, Catheter-directed thrombolysis; PMT, percutaneous mechanical thrombectomy; SVI, Social Vulnerability Index.
Data presented as number (%) or mean ± standard deviation.
Denotes Fisher’s exact test.
Table II.
Patient comorbidities and clinical presentation by intervention type
| Patient characteristic | CDT (n = 119) | PMT (n = 718) | Surgical embolectomy (n = 69) | P-value |
|---|---|---|---|---|
| Comorbidity | ||||
| MI | 10 (8.4) | 114 (15.9) | 6 (8.7) | .037 |
| CHF | 26 (21.8) | 139 (19.4) | 24 (34.8) | .010 |
| PVD | 2 (1.7) | 34 (4.7) | 9 (13.0) | .002 |
| COPD | 20 (16.8) | 124 (17.3) | 17 (24.6) | .297 |
| Renal disease | 9 (7.6) | 83 (11.6) | 6 (8.7) | .361 |
| Malignancy | 10 (8.4) | 68 (9.5) | 4 (5.8) | .576 |
| Metastatic solid tumor | 6 (5.0) | 46 (6.4) | 3 (4.3) | .736 |
| APR SOI | 3.1 ± 0.8 | 3.2 ± 0.8 | 3.7 ± 0.5 | <.001 |
| Transferred from other ACH | 18 (15.1) | 132 (18.4) | 36 (52.2) | <.001 |
| Saddle PE | 54 (45.4) | 342 (47.6) | 22 (31.9) | .043 |
ACH, Acute care hospital; APR SOI, All Patient Related Diagnosis Groups Severity of Illness (range 0-4); CDT, catheter-directed thrombosis; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; MI, myocardial infarction; PE, pulmonary embolism; PMT, percutaneous mechanical thrombectomy; PVD, peripheral vascular disease.
Data presented as number (%) or mean ± standard deviation.
In the univariate analysis of hospital characteristics, significant differences in interventions were observed across all measures (all P < .001). Nearly all surgical embolectomies (98.6%) were performed in the largest hospitals (bed size quartile 4), whereas bed size quartiles 2 and 4 accounted for the majority of CDT and PMT interventions. Similarly, 97.1% of surgical embolectomies were performed in hospitals with the highest advanced PE intervention volume (quartile 4). These highest PE intervention volume hospitals performed 63.9% and 61.3% of CDTs and PMTs, respectively. Over 95% of surgical embolectomies were performed in hospitals with surgery residency programs, trauma designation (primarily level 1), ECMO programs, and cardiac surgery programs. Greater variability in these hospital characteristics was observed for CDT and PMT. LeapFrog safety grades were also significantly different across interventions. LeapFrog B hospitals accounted for the greatest proportion of surgical embolectomies (97.1%), CDTs (48.7%), and PMTs (46.0%) performed (Table III).
Table III.
Hospital characteristics by intervention type
| Hospital characteristics | CDT (n = 119) | PMT (n = 718) | Surgical embolectomy (n = 69) | P-value |
|---|---|---|---|---|
| Hospital bed size quartile | <.001 | |||
| 1 (median, 100; range, 18-114) | 7 (5.9) | 41 (5.7) | 0 (0.0) | |
| 2 (median, 189; range, 125-198) | 43 (36.1) | 297 (41.4) | 0 (0.0) | |
| 3 (median, 240; range, 221-240) | 20 (16.8) | 91 (12.7) | 1 (1.4) | |
| 4 (median:, 424; range, 307-1105) | 49 (41.2) | 289 (40.3) | 68 (98.6) | |
| PE with advanced intervention volume quartile | <.001 | |||
| 1 (median, 1; range, 1-2) | 5 (4.2) | 7 (1.0) | 0 (0.0) | |
| 2 (median, 10; range, 4-19) | 16 (13.4) | 46 (6.4) | 0 (0.0) | |
| 3 (median, 34; range, 20-50) | 22 (18.5) | 225 (31.3) | 2 (2.9) | |
| 4 (median, 88; range, 51-115) | 76 (63.9) | 440 (61.3) | 67 (97.1) | |
| Surgery residency program | 40 (33.6) | 198 (27.6) | 68 (98.6) | <.001 |
| Trauma designation | 54 (45.4) | 285 (39.7) | 69 (100.0) | <.001 |
| Trauma level | <.001 | |||
| Level 1 | 24 (44.4) | 96 (33.7) | 66 (95.7) | |
| Level 2 | 30 (55.6) | 82 (28.8) | 3 (4.3) | |
| Level 3 | 0 (0.0) | 107 (37.5) | 0 (0.0) | |
| ECMO program | 24 (20.2) | 96 (13.4) | 66 (95.7) | <.001 |
| Cardiac surgery program | 61 (51.3) | 455 (63.4) | 69 (100.0) | <.001 |
| LeapFrog grade | <.001 | |||
| A | 9 (7.6) | 194 (27.0) | 0 (0.0) | |
| B | 58 (48.7) | 330 (46.0) | 67 (97.1) | |
| C | 41 (34.5) | 169 (23.5) | 0 (0.0) | |
| D | 11 (9.2) | 25 (3.5) | 2 (2.9) |
CDT, catheter-directed thrombosis; ECMO, extracorporeal membrane oxygenation; PE, pulmonary embolism; PMT, percutaneous mechanical thrombectomy.
Data presented as number (%).
In the multivariate model evaluating the association of patient and hospital characteristics with CDT performance, trauma designation was associated with increased odds of CDT (OR, 9.509; P < .001) after controlling for all other factors. Conversely, MI (OR, 0.440; P = .026), increased SOI scores (OR, 0.735; P = .027), presence of a cardiac surgery program (OR, 0.093; P < .001), LeapFrog grade A (OR, 0.199; P = .007), and PE volume quartiles 3 (OR, 0.037; P = .001) and 4 (OR, 0.068; P = .006) were all associated with decreased odds of CDT performance (Table IV).
Table IV.
Multivariate model of patient and hospital characteristics associated with catheter-directed thrombolysis (CDT)
| Characteristic | OR | 95% CI (lower bound) | 95% CI (upper bound) | P-value |
|---|---|---|---|---|
| Age 40+ years | 2.012 | .883 | 4.585 | .096 |
| White race | .744 | .439 | 1.261 | .272 |
| Medicare | .822 | .507 | 1.332 | .426 |
| Medicaid | .602 | .299 | 1.216 | .157 |
| MI | .440 | .213 | .907 | .026 |
| CHF | 1.258 | .734 | 2.155 | .404 |
| PVD | .321 | .070 | 1.467 | .143 |
| APR SOI | .735 | .559 | .965 | .027 |
| Overall SVI | 2.089 | .640 | 6.821 | .223 |
| Trauma designation | 9.509 | 2.731 | 33.110 | <.001 |
| Surgery residency | 2.304 | .841 | 6.310 | .104 |
| ECMO program | .566 | .174 | 1.840 | .344 |
| Cardiac surgery program | .093 | .035 | .248 | <.001 |
| LeapFrog grade A | .199 | .062 | .637 | .007 |
| Bed size quartile 2 | 2.880 | .552 | 15.016 | .209 |
| Bed size quartile 3 | 2.218 | .361 | 13.641 | .390 |
| Bed size quartile 4 | 1.127 | .200 | 6.341 | .892 |
| PE volume quartile 2 | .245 | .040 | 1.501 | .128 |
| PE volume quartile 3 | .037 | .005 | .279 | .001 |
| PE volume quartile 4 | .068 | .010 | .459 | .006 |
APR SOI, All Patient Related Diagnosis Groups Severity of Illness (range 0-4); CHF, congestive heart failure; CI, confidence interval; ECMO, extracorporeal membrane oxygenation; MI, myocardial infarction; PE, pulmonary embolism; PVD, peripheral vascular disease; SVI, Social Vulnerability Index.
In the multivariate model evaluating the association of patient and hospital characteristics with PMT performance, MI (OR, 1.992; P = .027), the presence of a cardiac surgery program (OR, 10.793; P < .001), Leapfrog grade A (OR, 5.049; P = .005), and PE volume quartiles 3 (OR, 26.573; P = .001) and 4 (OR, 13.047; P = .008) were associated with increased odds of PMT, controlling for all other factors. Trauma designation (OR, 0.088; P < .001) and the presence of a surgery residency program (OR, 0.362; P = .044) were associated with decreased odds of PMT performance (Table V).
Table V.
Multivariate model of patient and hospital characteristics associated with percutaneous mechanical thrombectomy (PMT)
| Characteristic | OR | 95% CI (lower bound) | 95% CI (upper bound) | P-value |
|---|---|---|---|---|
| Age 40+ years | 1.085 | .604 | 1.951 | .785 |
| White race | 1.313 | .833 | 2.070 | .240 |
| Medicare | 1.492 | .967 | 2.302 | .070 |
| Medicaid | .978 | .561 | 1.704 | .936 |
| MI | 1.992 | 1.082 | 3.665 | .027 |
| CHF | .742 | .470 | 1.171 | .200 |
| PVD | .781 | .329 | 1.855 | .575 |
| APR SOI | 1.090 | .846 | 1.404 | .506 |
| Overall SVI | .642 | .229 | 1.798 | .399 |
| Trauma designation | .088 | .026 | .302 | <.001 |
| Surgery residency | .362 | .134 | .975 | .044 |
| ECMO program | .488 | .163 | 1.465 | .201 |
| Cardiac surgery program | 10.793 | 4.065 | 28.657 | <.001 |
| LeapFrog grade A | 5.049 | 1.615 | 15.789 | .005 |
| Bed size quartile 2 | .463 | .089 | 2.424 | .362 |
| Bed size quartile 3 | .589 | .096 | 3.626 | .568 |
| Bed size quartile 4 | 1.129 | .199 | 6.398 | .891 |
| PE volume quartile 2 | 3.782 | .615 | 23.267 | .151 |
| PE volume quartile 3 | 26.573 | 3.578 | 197.354 | .001 |
| PE volume quartile 4 | 13.047 | 1.961 | 86.785 | .008 |
APR SOI, All Patient Related Diagnosis Groups Severity of Illness (range 0-4); CHF, congestive heart failure; CI, confidence interval; ECMO, extracorporeal membrane oxygenation; MI, myocardial infarction; PE, pulmonary embolism; PVD, peripheral vascular disease; SVI, Social Vulnerability Index.
With regards to hospital treatments and outcomes, there were no significant differences in rates of non-routine discharge or mortality across the interventions. Complication rates were also similar across interventions, with the exception of adverse events (AEs) of anticoagulants, cardiogenic shock, and cardiac arrest. Rates of intensive care unit (ICU) admission were significantly different across groups, ranging from 45.8% in PMT, to 85.7% in CDT, and 97.1% in surgical embolectomy (P < .001; all post-hoc between group differences P < .0167). Transfusion, vasopressor use, and ECMO rates were also significantly different across interventions, and most frequently performed in surgical embolectomy patients. Average hospital length of stay (LOS) was also significantly different across interventions (P < .001) and was longer in surgical patients (21.5 ± 21.1 days; P < .0167) but similar between CDT and PMT patients (9.3 ± 24.9 vs 6.4 ± 10.0 days). Average hospitalization charges were significantly different across interventions (P < .001), and were also higher for surgical embolectomy patients ($182,491 ± $167,473; P < .0167), but similar between CDT and PMT patients ($56,756 ± $59,152 vs $55,352 ± $55,092) (Table VI).
Table VI.
Hospitalization treatments and outcomes by intervention type
| Outcome | CDT (n = 119) | PMT (n = 718) | Surgical embolectomy (n = 69) | P-value |
|---|---|---|---|---|
| Treatments | ||||
| ICU admission | 102 (85.7)a | 329 (45.8)b | 67 (97.1)c | <.001 |
| ICU LOS days | 3.7 ± 4.8a | 3.5 ± 5.0a | 10.6 ± 14.6b | <.001 |
| Transfusion | 15 (12.6)a | 68 (9.5)a | 26 (37.7)b | <.001 |
| Vasopressor use | 11 (9.2)a,b | 42 (5.8)b | 10 (14.5)a | .015 |
| CPR | 3 (2.5) | 18 (2.5) | 4 (5.8) | .241d |
| ECMO | 6 (5.0)a | 14 (1.9)a | 17 (24.6)b | <.001 |
| Complications | ||||
| Any bleeding complication | 12 (10.1) | 52 (7.2) | 9 (13.0) | .164 |
| Intracranial hemorrhage | 3 (2.5) | 10 (1.4) | 1 (1.4) | .589d |
| AE of thrombolytics | 2 (1.7) | 1 (0.1) | 0 (0.0) | .063d |
| AE of anticoagulants | 3 (2.5)a | 1 (0.1)b | 0 (0.0)a,b | .010d |
| Cardiogenic shock | 13 (10.9)a | 62 (8.6)a | 19 (27.5)b | <.001 |
| Cardiac arrest | 8 (6.7)a | 22 (3.1)a | 0 (0.0)a | .036d |
| LOS, days | 9.3 ± 24.9a | 6.4 ± 10.0a | 21.5 ± 21.1b | <.001 |
| Non-routine dischargee | 33 (30.8) | 160 (23.6) | 19 (30.2) | .169 |
| In-hospital mortality | 7 (5.9) | 38 (5.3) | 5 (7.2) | .780 |
| Hospital charge, USD | 56,756 ± 59,152a | 55,352 ± 55,092a | 182,491 ± 167,473b | <.001 |
AE, Adverse event; CDT, catheter-directed thrombosis; CPR, cardiopulmonary resuscitation; ECMO, extracorporeal membrane oxygenation; ICU, intensive care unit; LOS, length of stay; PMT, percutaneous mechanical thrombectomy; USD, United States dollars.
Data presented as number (%) or mean ± standard deviation.
a,b,cSuperscripts denote post-hoc between group differences after Bonferroni adjustment. Different letters represent a significant difference between groups.
Denotes Fisher’s exact test.
Excludes patients deceased in hospital; non-routine discharge includes any discharge disposition other than home or home with home health.
Discussion
This study analyzed 6 years of statewide data to examine how hospital and patient characteristics influence the selection of advanced interventions for acute PE. We found that both patient and hospital characteristics significantly impacted the type of intervention used. The adoption of advanced PE therapies has increased steadily over the past decade, with rising utilization of catheter-based interventions (PMT and CDT), systemic thrombolytics, and ECMO.3 In addition to the overall proliferation of these treatments, there is a significant interest in PMT given ongoing industry-sponsored trials examining outcomes in intermediate and high-risk PE.18, 19, 20, 21 Our results help provide a better understanding of current trends and factors that influence PE intervention as ongoing trials and new data influencing standards of care emerge.
A key finding of this study was the substantial shift from CDT to PMT during the study period, with this trend accelerating markedly in the post-pandemic era. The increased adoption of PMT likely reflects the availability of large-bore mechanical thrombectomy devices, which offer single-session treatment and potentially reduce ICU LOS. Although our results found no significant differences in ICU LOS among patients treated with CDT and PMT, PMT patients were significantly less likely to be admitted to the ICU (45.8 vs 85.7%). As the time frame of this study included the height of the COVID-19 pandemic, it is possible that the decreased requirement for ICU care with PMT made this treatment option particularly attractive during this resource constrained period and led to its continued adoption post-pandemic.
PMT was more common in larger hospitals with higher PE volumes, whereas CDT utilization was associated with smaller hospital size, lower PE volume, and—unexpectedly—trauma center designation. The association with trauma centers may reflect variability in specialist availability and procedural logistics. Initial catheter placement for CDT is typically faster than PMT, and the competing demands for operating room, hybrid room, or interventional radiology time in trauma centers could influence intervention choice toward the more expedient option. Other studies have demonstrated significant variability in other vascular procedures at trauma centers and variability in specialties involved. It is reasonable to conclude PE treatment is similarly variable in trauma centers.22 We did not specifically evaluate type of specialist performing the advanced PE intervention or level of trauma designation in detail. Interestingly, the absence of cardiac surgery did not appear to limit large-bore mechanical thrombectomy use. This is despite potential risk of decompensation of patients requiring ECMO or intraprocedural technical complications requiring cardiac intervention. There has been a significant number of studies demonstrating reasonable safety profiles of some of the more commonly utilized devices.21,23, 24, 25, 26 Between both industry-sponsored trials looking at more user-friendly mechanical devices and significant interest by a variety of specialists in adoption of these technologies, it is unsurprising that hospitals without cardiac surgery are engaged in their use. There are no “best practices” guidelines for adoption of these technologies within the context of PERTs. The PERT Consortium and various other professional organizations such as the American College of Chest Physicians do not clearly proscribe that cardiac surgery be readily available for use of these technologies.
As anticipated, surgical embolectomy occurred predominantly at large trauma centers with cardiac surgery and ECMO capabilities and higher PE volumes. These tertiary centers were likely the destination for the most complex patients. Surgical embolectomy patients were more frequently transferred from external facilities, likely reflecting the need for specialized capabilities unavailable at referring institutions. The PERT Consortium does provide a framework for interhospital transfer that specifically recommends transfer for high-risk patients with PE who may need ECMO support. Notably only two hospitals within our analysis have the ability to provide ECMO support. This may have influenced the number and type of procedures performed at certain hospitals; however, our database does not capture hospital-specific transfer patterns for high-risk patients with PE, making it difficult to determine whether such patients receive systemic thrombolysis or transfer to ECMO-capable centers rather than undergoing catheter-based intervention at the initial hospital.
Our analysis revealed concerning disparities in advanced PE intervention utilization. African American patients and individuals from socially vulnerable backgrounds—characterized by minority status and language barriers—were more likely to receive CDT rather than PMT or surgical intervention. The impact of SDOH on patient outcomes, treatment allocation, and systemic inequities in health care access on vascular surgery patients are well-documented in established literature.9,27,28 Additionally, previous literature documents the impact of SDOH on PE outcomes.29,30 Our findings raise critical questions about how SDOH influence clinical decision-making in PE management. Comorbidity burden increased progressively across intervention types, from CDT to PMT to surgical embolectomy. However, patients undergoing surgical embolectomy were notably younger and more often insured by Medicaid, whereas CDT and PMT recipients were predominantly Medicare beneficiaries. This pattern suggests that insurance type—serving as a proxy for age and socioeconomic status—may shape treatment pathways. The observed variation in intervention use reflects the multifactorial complexity of PE management decisions and is consistent with previously documented literature documenting the influence of socioeconomic status on patient care.9,27 Hospital infrastructure, availability of advanced technologies, clinician expertise, and institutional protocols all contribute to these patterns. Patient-specific factors—including age, comorbidities, and insurance status—further influence therapeutic selection, creating a complex interplay of clinical, logistical, and systemic determinants.
Economic factors may also influence intervention selection, though we did not directly examine costs. CDT incurs expenses related to thrombolytic drugs, prolonged ICU stays, and multiple angiography suite visits for treatment assessment. Although large bore thrombectomy devices carry substantial upfront costs, these may be offset by reduced procedural complexity and less frequent ICU utilization. Maryland’s unique total cost of care model adds another layer of complexity, as hospitals may select interventions based on available resources within their global budgets. Both CDT and PMT may be particularly attractive within this capitated model as they incurred average charges approximately one-third of that of surgical embolectomy in our data. This represents an important target for future investigations into resource allocation in advanced PE management.
This analysis is limited by its reliance on a single-state database from Maryland, which operates under a unique total cost of care model that may not generalize to other health care systems. We lacked granular data regarding the specialist teams providing PE interventions and the presence of formalized PERT programs, both of which likely influenced treatment decisions. Additionally, we could not assess patient transfer patterns or participation in clinical trials, which may affect intervention selection at individual institutions. Like all database studies, our analysis was limited by dependence on the accuracy of coding data.
Conclusions
Understanding the hospital and patient factors that determine PE intervention type is essential for addressing inequities in advanced PE care delivery. Our findings highlight significant disparities based on race, social vulnerability, and hospital characteristics that warrant systemic attention. Efforts to reduce these disparities should focus on improving equitable access to all treatment modalities across diverse populations, standardizing clinical decision pathways through PERT implementation, and addressing systemic barriers within health care infrastructure. Future research should prospectively evaluate intervention efficacy across demographic groups and examine the impact of targeted policy interventions aimed at reducing treatment inequities in PE management.
Author Contributions
Conception and design: ES, JT, TD, GJ
Analysis and interpretation: ES, JT, TD, GJ
Data collection: ES, JT, GJ
Writing the article: ES, JT, GJ
Critical revision of the article: ES, JT, TD, GJ
Final approval of the article: ES, JT, TD, GJ
Statistical analysis: JT
Obtained funding: Not applicable
Overall responsibility: GJ
Funding
None.
Disclosures
None.
From the Eastern Vascular Society
Footnotes
Additional material for this article may be found online at www.jvsvenous.org.
The editors and reviewers of this article have no relevant financial relationships to disclose per the Journal policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest.
Appendix
Additional material for this article may be found online at www.jvsvenous.org.
Appendix (online only)
References
- 1.Prabhu W., Soukas P.A. Pulmonary embolism in 2017: increasing options for increasing incidence. R I Med J (2013) 2017;100:27–32. [PubMed] [Google Scholar]
- 2.Bell E.J., Lutsey P.L., Basu S., et al. Lifetime risk of venous thromboembolism in two cohort studies. Am J Med. 2016;129:339.e19. doi: 10.1016/j.amjmed.2015.10.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Abumoawad A., Shatla I., Behrooz L., et al. Temporal trends in the utilization of advanced therapies among patients with acute pulmonary embolism: insights from a national database. Eur Heart J Acute Cardiovasc Care. 2023;12:711–713. doi: 10.1093/ehjacc/zuad092. [DOI] [PubMed] [Google Scholar]
- 4.Abou Ali A.N., Cherfan P., Zaghloul M.S., et al. Institutional trends over a decade in catheter-directed interventions for pulmonary embolism. J Vasc Surg Venous Lymphat Disord. 2022;10:287–292. doi: 10.1016/j.jvsv.2021.06.024. [DOI] [PubMed] [Google Scholar]
- 5.Silver M.J., Gibson C.M., Giri J., et al. Outcomes in high-risk pulmonary embolism patients undergoing FlowTriever mechanical thrombectomy or other contemporary therapies: results from the FLAME study. Circ Cardiovasc Interv. 2023;16 doi: 10.1161/CIRCINTERVENTIONS.123.013406. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Toma C., Jaber W.A., Weinberg M.D., et al. Acute outcomes for the full US cohort of the FLASH mechanical thrombectomy registry in pulmonary embolism. EuroIntervention. 2023;18:1201–1212. doi: 10.4244/EIJ-D-22-00732. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Russell N., Sayfo S., George T., Gable D. Effect of a pulmonary embolism response team on the management and outcomes of patients with acute pulmonary embolism. J Vasc Surg Venous Lymphat Disord. 2023;11:1139–1148. doi: 10.1016/j.jvsv.2023.05.016. [DOI] [PubMed] [Google Scholar]
- 8.Hobohm L., Farmakis I.T., Keller K., et al. Pulmonary embolism response team (PERT) implementation and its clinical value across countries: a scoping review and meta-analysis. Clin Res Cardiol. 2023;112:1351–1361. doi: 10.1007/s00392-022-02077-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Farmakis I.T., Valerio L., Giannakoulas G., et al. Social determinants of health in pulmonary embolism management and outcome in hospitals: insights from the United States nationwide inpatient sample. Res Pract Thromb Haemost. 2023;7 doi: 10.1016/j.rpth.2023.100147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.HSCR C HSCRC overview. https://hscrc.maryland.gov/Pages/About-Us.aspx
- 11.Narayan M., Tesoriero R., Bruns B.R., Klyushnenkova E.N., Chen H., Diaz J.J. Acute care surgery: defining mortality in emergency general surgery in the state of Maryland. J Am Coll Surg. 2015;220:762–770. doi: 10.1016/j.jamcollsurg.2014.12.051. [DOI] [PubMed] [Google Scholar]
- 12.Murray R. Setting hospital rates to control costs and boost quality: the Maryland experience. Health Aff (Millwood) 2009;28:1395–1405. doi: 10.1377/hlthaff.28.5.1395. [DOI] [PubMed] [Google Scholar]
- 13.HSCR C Clinical public use data requests. https://hscrc.maryland.gov/Pages/hsp-data-request.aspx
- 14.Horn S.D., Horn R.A., Sharkey P.D. The Severity of Illness Index as a severity adjustment to diagnosis-related groups. Health Care Financ Rev. 1984;Suppl(Suppl):33–45. [PMC free article] [PubMed] [Google Scholar]
- 15.CDC CDC/ATSDR Social Vulnerability Index. 2023. https://www.atsdr.cdc.gov/placeandhealth/svi/index.html
- 16.Austin J.M., D'Andrea G., Birkmeyer J.D., et al. Safety in numbers: the development of Leapfrog's composite patient safety score for U.S. hospitals. J Patient Saf. 2014;10:64–71. doi: 10.1097/PTS.0b013e3182952644. [DOI] [PubMed] [Google Scholar]
- 17.Group TL Leapfrog Hospital Survey scoring algorithms. 2025. https://www.leapfroggroup.org/survey-materials/scoring-and-results
- 18.Jaber W.A., Gonsalves C.F., Stortecky S., et al. Large-bore mechanical thrombectomy versus catheter-directed thrombolysis in the management of intermediate-risk pulmonary embolism: primary results of the PEERLESS randomized controlled trial. Circulation. 2025;151:260–273. doi: 10.1161/CIRCULATIONAHA.124.072364. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Rosovsky R.P., Konstantinides S.V., Moriarty J.M., et al. A prospective, multicenter, randomized controlled trial evaluating anticoagulation alone vs anticoagulation plus computer assisted vacuum thrombectomy for the treatment of intermediate-high-risk acute pulmonary embolism: rationale and design of the STORM-PE study. Am Heart J. 2025;288:1–14. doi: 10.1016/j.ahj.2025.03.018. [DOI] [PubMed] [Google Scholar]
- 20.Klok F.A., Piazza G., Sharp A.S.P., et al. Ultrasound-facilitated, catheter-directed thrombolysis vs anticoagulation alone for acute intermediate-high-risk pulmonary embolism: rationale and design of the HI-PEITHO study. Am Heart J. 2022;251:43–53. doi: 10.1016/j.ahj.2022.05.011. [DOI] [PubMed] [Google Scholar]
- 21.Giri J., Mahfoud F., Gebauer B., et al. Peerless II: a randomized controlled trial of large-bore thrombectomy versus anticoagulation in intermediate-risk pulmonary embolism. J Soc Cardiovasc Angiogr Interv. 2024;3 doi: 10.1016/j.jscai.2024.101982. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Kerwin A.J., Lombardo S., Haut E.R., Knudson M.M. Variability of prophylactic inferior vena cava filter use in young trauma patients: analysis of the consortium of leaders in the study of traumatic thromboembolism database. J Am Coll Surg. 2025;241:601–609. doi: 10.1097/XCS.0000000000001417. [DOI] [PubMed] [Google Scholar]
- 23.Sista A.K., Horowitz J.M., Tapson V.F., et al. Indigo aspiration system for treatment of pulmonary embolism: results of the EXTRACT-PE trial. JACC Cardiovasc Interv. 2021;14:319–329. doi: 10.1016/j.jcin.2020.09.053. [DOI] [PubMed] [Google Scholar]
- 24.Tu T., Toma C., Tapson V.F., et al. A prospective, single-arm, multicenter trial of catheter-directed mechanical thrombectomy for intermediate-risk acute pulmonary embolism: the FLARE study. JACC Cardiovasc Interv. 2019;12:859–869. doi: 10.1016/j.jcin.2018.12.022. [DOI] [PubMed] [Google Scholar]
- 25.Piazza G., Hohlfelder B., Jaff M.R., et al. A prospective, single-arm, multicenter trial of ultrasound-facilitated, catheter-directed, low-dose fibrinolysis for acute massive and submassive pulmonary embolism: the SEATTLE II study. JACC Cardiovasc Interv. 2015;8:1382–1392. doi: 10.1016/j.jcin.2015.04.020. [DOI] [PubMed] [Google Scholar]
- 26.Tapson V.F., Sterling K., Jones N., et al. A randomized trial of the optimum duration of acoustic pulse thrombolysis procedure in acute intermediate-risk pulmonary embolism: the OPTALYSE PE trial. JACC Cardiovasc Interv. 2018;11:1401–1410. doi: 10.1016/j.jcin.2018.04.008. [DOI] [PubMed] [Google Scholar]
- 27.Arya S., Binney Z., Khakharia A., et al. Race and socioeconomic status independently affect risk of major amputation in peripheral artery disease. J Am Heart Assoc. 2018;7 doi: 10.1161/JAHA.117.007425. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Barshes N.R., Minc S.D. Healthcare disparities in vascular surgery: a critical review. J Vasc Surg. 2021;74:6S–14S.e1. doi: 10.1016/j.jvs.2021.03.055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Zakai N.A., McClure L.A., Judd S.E., et al. Racial and regional differences in venous thromboembolism in the United States in 3 cohorts. Circulation. 2014;129:1502–1509. doi: 10.1161/CIRCULATIONAHA.113.006472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Barco S., Valerio L., Ageno W., et al. Age-sex specific pulmonary embolism-related mortality in the USA and Canada, 2000-18: an analysis of the WHO Mortality Database and of the CDC Multiple Cause of Death database. Lancet Respir Med. 2021;9:33–42. doi: 10.1016/S2213-2600(20)30417-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.