Factors Associated With Survival-to-Discharge in Patients With High-Risk Pulmonary Embolism Requiring Venoarterial Extracorporeal Membrane Oxygenation

Highlights

• •A total of 33 patients with high-risk pulmonary embolism were managed with venoarterial extracorporeal membrane oxygenation during the study period.
• •Seventeen patients (52%) survived to discharge.
• •Chronic kidney disease, higher lactate, and lower pH were all associated with in-hospital mortality.
• •Survivors more frequently had surgical or catheter-based treatment after cannulation.
• •Seventy-seven percent of patients treated with anticoagulation alone did not survive to discharge.

High-risk pulmonary embolism (PE) is characterized by hemodynamic instability¹ and is imminently life-threatening, with associated mortality rates as high as 41% for patients with shock and 74% for those with cardiac arrest.2, 3, 4 Despite developments in therapeutic techniques, mortality rates have been largely unchanged over many years.⁵ More recently, venoarterial extracorporeal membrane oxygenation (VA-ECMO) has emerged as a method to provide hemodynamic support and facilitate more definitive surgical or catheter-based therapies, although evidence for this approach is limited⁸. Several retrospective studies have shown improved survival and hemodynamics in the case of circulatory collapse or cardiac arrest with the use of VA-ECMO, often followed by surgical or catheter-based therapy. The goal of this descriptive analysis was to identify predictors of survival-to-discharge in patients with high-risk PE on VA-ECMO.

After institutional review board approval, all cases of high-risk PE (by 2019 European Society of Cardiology guidelines)¹ requiring VA-ECMO between 2015-2022 were reviewed. At the primary author’s institution, all high-risk PE patients are evaluated by a multidisciplinary shock team for systemic fibrinolysis, hemodynamic stabilization using VA-ECMO, or a primary catheter or surgical approach. VA-ECMO exclusion criteria include active bleeding, irreversible neurological injury, known comorbidity significantly limiting life expectancy, or multiorgan failure. Demographics, clinical characteristics, vitals and hemodynamics, laboratory data, procedural data, PE management strategy, and in-hospital outcomes were collected. The primary outcome of interest was survival to discharge.

All statistical analyses were performed using SAS v9.4 (SAS Institute, Cary, NC). Cohorts were hospital “survivors” and “nonsurvivors”. Continuous variables were depicted as mean (SD) or median (interquartile range) and compared using Student’s t-tests or Mann–Whitney U tests. Categorical variables were expressed as frequency (percentage) and compared using chi-square tests or Fisher exact tests when appropriate. p < 0.05 was considered significant.

A total of 33 patients were included in the study population; 17 (52%) survived to discharge. Major causes of death included pulseless electrical activity arrest (7 patients), septic shock (5 patients), and hemorrhagic shock or anoxic brain injury (3 patients). Forty-eight percent were men; the mean age was 59 years (Table 1). Nonsurvivors had a significantly higher proportion of chronic kidney disease (27% vs. 0%, p = 0.04). There were no statistically significant differences in other baseline characteristics in survivors compared to nonsurvivors.

Table 1.

Baseline characteristics stratified by in-hospital survival of patients

Variable	Overall (n = 33)	Survivors (n = 17)	Nonsurvivors (n = 16)	p value
Age (years)	59 ± 16	55 ± 18	64 ± 12	0.11
Male sex	16 (48%)	7 (41%)	9 (56%)	0.39
BMI	32.8 [28.1, 39.2]	32.2 [28.9, 39.4]	33.3 [28.1, 36.5]	0.97
Comorbidities
Coronary artery disease	5 (15%)	1 (5.9%)	4 (25%)	0.17
Congestive heart failure	3 (9.1%)	0 (0%)	3 (19%)	0.09
Hyperlipidemia	10 (30%)	4 (24%)	6 (38%)	0.46
Diabetes	7 (21%)	4 (24%)	3 (19%)	1
Hypertension	24 (73%)	13 (77%)	11 (69%)	0.62
Chronic pulmonary disease	2 (6.1%)	0 (0%)	2 (13%)	0.23
Cerebrovascular accident	2 (6.1%)	0 (0%)	2 (13%)	0.23
Chronic kidney disease	4 (12%)	0 (0%)	4 (27%)	0.04
Tobacco use, current or prior	8 (24%)	4 (25%)	4 (25%)	1
Prior VTE	8 (24.2%)	4 (23.5%)	4 (26.7%)	1
Malignancy	5 (15.2%)	3 (17.6%)	2 (13.3%)	1
Hormone use∗	1 (3.0%)	0 (0%)	1 (6.7%)	0.95
Recent immobilization†	16 (48.5%)	9 (52.9%)	7 (46.7%)	1
Recent surgery†	14 (42.4%)	8 (47.1%)	6 (40.0%)	0.96
Initial lab data
WBC Count (cells/uL)	13.3 ± 6.4	15.5 ± 7.6	11.1 ± 4.1	0.06
Hgb (g/dL)	11.1 ± 2.5	11.2 ± 2.8	11.0 ± 2.2	0.76
Platelet count (cells/uL)	154 ± 61	140 ± 58	166 ± 62	0.29
Creatinine (mg/dL)	1.70 ± 0.91	1.95 ± 1.14	1.50 ± 0.65	0.24
Bicarbonate (mmol/L)	19 ± 4	18 ± 5	20 ± 4	0.14
AST (U/L)	1184 ± 2184	1688 ± 2580	597 ± 1510	0.21
ALT (U/L)	821 ± 1526	1254 ± 1949	315 ± 536	0.12
PTT (sec)	68 ± 56	72 ± 58	64 ± 55	0.73
INR	1.6 ± 0.5	1.7 ± 0.6	1.6 ± 0.5	0.53
pH	7.26 ± 0.14	7.31 ± 0.08	7.21 ± 0.16	0.06
Lactate (mmol/L)	6.8 ± 5.1	4.8 ± 8.4	8.4 ± 5.7	0.06
Lab data - 48 h
WBC Count (cells/uL)	13.0 ± 4.6	12.2 ± 4.3	14.1 ± 5.1	0.34
Hgb (g/dl)	8.8 ± 1.5	8.7 ± 1.5	8.9 ± 1.6	0.77
Platelet count (cells/uL)	93 ± 39	104 ± 38	80 ± 38	0.13
Creatinine (mg/dL)	1.90 ± 0.86	1.54 ± 0.75	2.27 ± 0.83	0.04
Bicarbonate (mmol/L)	23 ± 5	25 ± 4	20 ± 5	0.02
AST (U/L)	438 ± 603	447 ± 648	428 ± 575	0.94
ALT (U/L)	431 ± 664	548 ± 811	293 ± 432	0.34
PTT (sec)	51 ± 16	52 ± 18	50 ± 15	0.74
INR	1.4 ± 0.3	1.4 ± 0.2	1.6 ± 0.3	0.09
pH	7.40 ± 0.07	7.41 ± 0.08	7.40 ± 0.06	0.69
Lactate (mmol/L)	2.0 ± 1.6	1.4 ± 0.8	2.6 ± 2.0	0.05
VA-ECMO data
Duration (days)	6.2 ± 7.3	6.01 ± 3.92	6.49 ± 9.89	0.85
Duration <24 h	3 (9.1%)	0 (0%)	3 (19%)	0.11
Cannulated, cath lab or OR‡	10 (40%)	7 (54%)	3 (33%)	0.09
Cannulated, bedside‡	13 (52%)	5 (38%)	8 (67%)	0.16
Active CPR§	13 (39%)	4 (36%)	9 (69%)	0.22
Timing of PE-directed therapy∗∗				0.03
Pre-ECMO cannulation	7 (21.2%)	3 (17.6%)	4 (26.7%)
Post-ECMO cannulation	14 (42.4%)	11 (64.7%)	3 (20%)
PE-directed therapy on VA-ECMO				0.04
Catheter-based thrombectomy	12 (33%)	7 (41%)	4 (25%)	-
Catheter-directed thrombolysis	4 (15%)	1 (6%)	4 (25%)	-
Combination of catheter-based therapies	2 (3.0%)	1 (6%)	0 (0%)	-
Surgical embolectomy	4 (12%)	4 (24%)	0 (0%)	-
Pulmonary thromboendarterectomy	2 (6.1%)	1 (6%)	1 (6%)	-
No PE specific invasive therapy	9 (27.3%)	3 (18%)	7 (44%)
Outcomes
Hospital length of stay (days)	34 ± 37	42 ± 24	28 ± 46	0.35
ICU length of stay (days)	19 ± 17	26 ± 17	13 ± 15	0.05
Bleeding‖	16 (53%)	8 (50%)	8 (57%)	0.69
Limb ischemia	9 (27%)	5 (29%)	4 (25%)	1
CRRT	11 (33%)	6 (35%)	5 (31%)	0.81
Cerebrovascular accident	5 (16%)	2 (13%)	3 (19%)	1
Infection#	12 (44%)	7 (50%)	5 (39%)	0.55

Data presented are mean ± SD or n (%). p ≤ 0.05 are indicated in bold.

Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CPR, cardiopulmonary resuscitation; CRRT, continuous renal replacement therapy; Hgb, hemoglobin, ICU, intensive care unit; INR, international normalized ratio; IQR, interquartile range; PE, pulmonary embolism; PTT, partial thromboplastin time; VA-ECMO,venoarterial extracorporeal membrane oxygenation; VTE, venous thromboembolism; WBC, white blood cell.

^∗Prothrombotic hormone use such as estrogen-containing medications.

^†Recent was defined as within 7 d of the index hospitalization.

^‡Cannulation location was unable to be obtained for 10/33 patients.

^§Unable to determine if CPR actively occurring during cannulation in 9/33 patients.

^‖Defined as clinically overt bleeding with transfusion of at least 1 unit of packed red blood cell units and an identified source of blood loss.

^#Includes urinary tract infections, bacterial and viral pulmonary infections, cholangitis, bacteremia, and cutaneous infections.

^∗∗Includes surgical embolectomy, pulmonary thromboendarterectomy, catheter-based thrombectomy, and catheter-directed thrombolysis.

On initial laboratory assessment, patients who survived to discharge were substantially less acidotic (pH 7.31 vs. 7.21, p = 0.06) and had a lower serum lactate (4.8 vs. 8.4, p = 0.06); other initial laboratory values were similar. After being on VA-ECMO for 48 hours, patients who survived had maintained a stable renal function (Cr 1.54 vs. 2.27, p = 0.04) and cleared their lactate (lactate 1.4 vs. 2.6, p = 0.05).

In terms of cannulation strategy, a numerically higher proportion of patients who survived to discharge were cannulated in the cath lab or operating room (54% vs. 33%, p = 0.09) as opposed to bedside. A large proportion of the overall cohort was receiving cardiopulmonary resuscitation during cannulation (39%), and this number was numerically higher in cohort that did not survive to discharge (69% vs. 36%, p = 0.22). There were no significant differences in major bleeding or thrombotic events while on VA-ECMO, and rates of continuous renal replacement therapy initiation were similar in both cohorts.

In terms of definitive management strategy, a numerically higher proportion of survivors were treated with surgical or catheter-based therapies in addition to anticoagulation and VA-ECMO (82% vs. 56%, p = 0.1). These definitive treatments more frequently occurred following VA-ECMO cannulation as opposed to before cannulation in those that survived to discharge (65% vs. 20%, p = 0.03).

As demonstrated in this study, high-risk PE carries substantial morbidity and mortality. Proper patient selection and risk-stratification are therefore crucial. Results of this descriptive analysis demonstrate that chronic kidney disease, higher baseline lactate, and lower baseline pH were all associated with in-hospital mortality for patients receiving VA-ECMO for high-risk PE, indicating that at the extreme end of shock state, VA-ECMO support may not be as effective to support survival as opposed to earlier shock stages.

Although this study demonstrated no significant differences in survival to discharge based on PE management strategy, a numerically higher proportion of survivors were treated with surgical or catheter-based therapies in addition to anticoagulation and VA-ECMO. In contrast, 77% of patients receiving anticoagulation alone did not survive to discharge. Given high mortality noted with anticoagulation alone and substantial bleeding risks associated with systemic fibrinolytics, the use of targeted surgical or catheter-based treatments appears reasonable in this patient population, and VA-ECMO provides a useful bridge to hemodynamically support patients before definitive therapy.⁴ Although only a small proportion of patients were on VA-ECMO, the FlowTriever for Acute Massive PE study demonstrated that mechanical thrombectomy was associated with low in-hospital adverse outcomes (17%) and mortality (1.9%).⁵ Additional research is needed to clarify the ideal management strategy after VA-ECMO initiation.

Results of this study suggest that the timing of definitive therapy may also be important. Survivors more frequently received definitive surgical or catheter-based therapies following VA-ECMO cannulation. This supports an “VA-ECMO first, definitive therapy second” approach has been gaining popularity in high-risk PE treatment.

This study had several important limitations. First, despite including several years of data, the sample size was small, which limited the analyses that could be performed and may have left the study underpowered to identify associations between invasive therapy and mortality. The retrospective design also limited the analysis to identification of associations; although randomized data would be optimal, it is difficult to obtain in this population given the clinical acuity. Lastly, data were obtained from a single, quaternary-care institution that is capable of VA-ECMO and may not be generalizable to other institutions.

In conclusion, high-risk PE requiring VA-ECMO is associated with significant in-hospital mortality. This descriptive analysis demonstrated several factors associated with in-hospital mortality for patients on VA-ECMO, particularly those related to worsened shock state on arrival. Careful patient selection is required to identify patients likely to survive to discharge, and further data are needed to validate these risk factors and determine if modifying any of them including offering early definitive therapy can lead to improved survival for these patients.

Ethics Statement

This study was approved by the primary authors' institutional review board. Requirement for individual informed consent was waived.

Funding

The authors have no funding to report.

Disclosure Statement

The authors report no conflict of interest.