Abstract
Background
Extracorporeal membrane oxygenation (ECMO) is increasingly used in patients after cardiac surgery; however, anticoagulation management has consistently been challenging. This study aimed to explore the feasibility of a low-anticoagulation strategy for post-cardiotomy ECMO (PC-ECMO).
Methods
A retrospective comparison was performed between two anticoagulation targets in adult patients undergoing veno-arterial ECMO after cardiac surgery at the Beijing Anzhen Hospital (Beijing, China) between January 2018 and November 2023. The low-anticoagulation (LAC) strategy group consisted of patients with an activated partial thromboplastin time (APTT) ratio of 1–1.5, whereas the conventional anticoagulation (CAC) strategy group included those with an APTT ratio of 1.5–2.5. The primary outcome was thrombotic complications associated with ECMO. Secondary outcomes included bleeding events, pulmonary infection, need for renal replacement therapy, in-hospital mortality rate, ECMO support duration, hours of mechanical ventilation, anticoagulation fraction, length of hospitalization, and transfusion volume(s).
Results
The study included data from 203 patients, who were divided into two groups: LAC (n = 108 [53.2%]) and CAC (n = 95 [46.8%]). Propensity score matching was used to balance confounding variables. A total of 43 patient pairs were successfully matched, and no significant difference was observed in thrombotic complications between the LAC and CAC groups (30.2% versus [vs.] 25.3%, respectively; p = 0.810). Meanwhile, no significant differences were observed in secondary outcomes and subgroups within the matched cohort, except for ECMO support time, which was shorter in the LAC group (119.6 h vs. 146.0 h; p = 0.015).
Conclusion
The low-anticoagulation strategy was feasible for PC-ECMO support.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12871-025-03153-4.
Keywords: Anticoagulation, Post-cardiotomy extracorporeal membrane oxygenation, Feasibility
Introduction
Severe cardiorespiratory shock not responding to conventional treatment can occur as a complication of cardiac surgery, resulting in elevated mortality rates. Extracorporeal membrane oxygenation (ECMO) provides temporary circulatory and respiratory support, serving as a bridge to organ recovery, transplantation, or permanent assistance, thereby facilitating recovery [1–5]. Despite the increased use of post-cardiotomy ECMO (PC-ECMO), no substantial improvement has been observed in short-term survival rates [6–7].
ECMO-related complications are inevitable. The most frequent complication is renal failure [8], followed with bleeding and thrombotic events [9–13]. Studies have shown that anticoagulation decreases thrombotic complications; nevertheless, the degree of anticoagulation is closely related to the risk of bleeding [10, 12, 14]. Previous studies have shown that hemorrhagic complications substantially affect survival rates [13, 15–17]. However, the current researches are insufficient to demonstrate the most effective monitoring method and threshold for anticoagulant therapy in ECMO patients [18]. Protti et al. found that among 273 centers across 50 countries, partial thromboplastin time (APTT) was the most common monitoring method (42%), followed by activated clotting time (ACT; 30%) and anti-factor Xa (23%) [19]. The Extracorporeal Life Support Organization (ELSO) guidelines recommend targeting an APTT of 1.5–2.5 times the patient’s baseline [20]. Given that patients undergoing PC-ECMO are at increased risk of hemorrhagic complications, an efficacious and secure anticoagulation protocol is paramount for successful outcome of PC-ECMO therapy.
Previous retrospective studies have suggested that reduced or no anticoagulation approaches in patients undergoing VA-ECMO are safe [21–26]. Presently, there is a research gap regarding the optimal anticoagulation strategy specifically for PC-ECMO support. Thus, this study aimed to assess the feasibility and safety of a reduced anticoagulation strategy in patients receiving PC-ECMO.
Methods
Extracorporeal life support cannulation and anticoagulation
Details regarding VA-ECMO initiation and management have been described previously [27–28]. A heparin bolus (100 µ/kg) was injected before cannulation in patients undergoing PC-ECMO. After initiation of VA-ECMO, anticoagulation levels were adjusted according to the patient’s bleeding risk. The assessment of bleeding tendency is determined by clinicians based on the patient’s chest tube drainage [29], bleeding at the ECMO cannulation site, coagulation function, and platelet count and function [30]. For those with a bleeding tendency, a lower target APTT ratio of 1.0–1.5 is aimed for, while in patients without bleeding concerns, the target range is 1.5–2.5. APTT is the primary monitoring tool, with ACT used supplementarily. APTT testing occurs twice daily, while ACT is checked every 4–6 h. After ECMO initiation, clinicians determine the appropriate timing to commence anticoagulation based on the anticoagulation target and the trend of APTT ratio/ACT changes.
Study design and population
Data from adult patients, who underwent PC-ECMO support at Beijing Anzhen Hospital between January 2018 and December 2023, were retrospectively collected and reviewed. Patients with ECMO duration < 24 h, as well as those who did not undergo any anticoagulation therapy were excluded. Additionally, patients requiring secondary surgeries for hemorrhage control within the first 24 h and those treated with direct thrombin inhibitors (DTIs) were excluded from the analysis. Additional exclusion criteria included patients with multiple VA-ECMO episodes, those undergoing hybrid ECMO configurations, and those with incomplete key data, specifically those lacking APTT and antithrombin III (ATIII) data. Patients were divided into two groups based on the average APTT ratio during ECMO support: the LAC (low-anticoagulation) group with an average APTT ratio of 1–1.5 and the CAC (conventional anticoagulation) group with an average APTT ratio of 1.5–2.5.
Data collection
The following patient data were collected: demographic variables; comorbidities; pre-ECMO mean arterial pressure (MAP) and lactate levels; intra-aortic balloon pump (IABP); cardiac arrest (CA); ECMO installation locations; cardiac surgery type(s); cardiopulmonary bypass (CPB); time to start anticoagulation and time to end anticoagulation; the worst levels of key blood markers (hemoglobin [HB], platelet [PLT] count, albumin [ALB], D-dimer, and alanine aminotransferase [ALT]) during ECMO; average APTT and ATIII values while on ECMO; compliance rate of the APTT ratio during ECMO support; ECMO blood flow at 24 h of support; Sequential Organ Failure Assessment (SOFA) score after ECMO support; total amount of blood products transfused (red blood cells [RBC], fresh frozen plasma [FFP], and Platelet during ECMO; In-hospital mortality and ECMO-associated morbidities, including thrombotic complications, bleeding complications, pulmonary infections, and need for renal replacement therapy (RRT). Time 1, Time 2 and Time 3 were calculated using the following equations: Time1 = time to start anticoagulation – time to start ECMO; Time 2 = time to end anticoagulation – time to start anticoagulation; and Time 3 = Time 2/ECMO duration.
Outcomes and definitions
The primary outcome was the occurrence of thrombotic complications associated with the ECMO run or resulting from ECMO. Diagnostic imaging is performed when patients demonstrate clinical signs suggestive of thrombosis. To screen for post-ECMO thrombotic events, routine vascular ultrasounds are conducted. Patients receive an ultrasound on the day of weaning. If results are normal, they are monitored clinically; if symptoms arise, another ultrasound is obtained. Adverse events related to thromboembolism were categorized as follows: circuit or cannula blockages leading to component replacements or ECMO termination; ischemic stroke, as confirmed by cerebral computed tomography; and peripheral thromboembolism excluding cerebral events, such as arterial thrombosis, venous thrombosis, pulmonary embolism, or intracardiac thrombosis.
Secondary outcomes included bleeding complications, pulmonary infection, RRT use, in-hospital mortality, duration of ECMO and mechanical ventilation, length of hospital stay, and blood product transfusion volumes. Bleeding complications were categorized following ELSO registry guidelines, counting events requiring surgery, daily RBC transfusions over 20 ml/kg/day, or more than 3 RBC unit transfusions in 24 h. Severe issues included fatal brain bleeding and cardiac compression (pericardial tamponade). (See ELSO Registry Data Definitions, 05/17/2022, available online at https://www.elso.org/portals/0/files/pdf/elso%20registry%20data%20definitions%2005_17_22.pdf).
Statistical analysis
Continuous data are expressed as median and corresponding interquartile range (IQR), and statistical comparisons were performed using the non-parametric Mann–Whitney U test. Categorical data are tabulated as frequency and percentage, and comparisons were performed using the chi-squared test. When expected cell counts were < 5, the more conservative Fisher’s exact test was used. Propensity score matching analysis was performed between the LAC and CAC groups based on estimated propensity scores. Age, gender, body mass index (BMI), comorbidities, smoking, anticoagulant drugs, IABP, CA, ATIII, HB, platelet count, ALT, ALB, lactate (LAC2), Time 1, SOFA, compliance rate and ECMO blood flow were included as matching indicators to address confounding factors. A one-to-one propensity score-matched analysis was performed using nearest-neighbor matching within a caliper width of 0.02 standardized difference (SD) of the pooled propensity scores. Covariates with standardized mean difference (SMD) > 0.20 after matching were unbalanced. To demonstrate the stability of the results, multivariate logistic regressions were performed by entering LAC (vs. CAC) with the following available variables: age, BMI, IABP, D-dimer, ATIII, HB, platelet count, ALT, RBC, FFP, platelet, lactate, compliance rate, ECMO blood flow and VA-ECMO duration. Differences with p < 0.05 were considered to be statistically significant. All analyses were performed using SPSS version 27.0 and R version 4.4.0.
Results
Patient characteristics
Among the 470 adult patients who underwent PC-ECMO from 1 February 2018 to 30 November 2023, 203 fulfilled the inclusion criteria (Fig. 1), and were divided into two groups according to anticoagulation strategy: LAC (n = 108 [53.2%]); and CAC (n = 95 [46.8%]). Before matching, the IQR of the APTT ratio for LAC group and CAC group were 1.26 (1.14,1.36) and 1.79 (1.59,2.15), respectively. The median values of the average APTT were 38.3 s for the LAC group and 54.9 s for the CAC group, as shown in Fig. 2A. Baseline patient characteristics are summarized in Table 1. The median age was 62 years (IQR 54–68 years) and the majority of patients were male (69.5%). IABP was used in 68.0% of patients. The proportion of patients who underwent cannulation in the ICU was the highest (53.5%). Of these patients, 82.8% underwent coronary artery bypass grafting (CABG) or valve repair/replacement surgery. The average ATIII in the study population was lower than the normal value, and the median values of the average ATIII was 45% (IQR 31%–62%). The nadir levels of ALB and PLT in patients after surgery were significantly lower than normal values: median ALB, 25 g/L (IQR 20–28 g/L); and median platelet count, 29 × 109/L (IQR 17–50 × 109/L). The median time to initiate anticoagulation after ECMO support was 39.6 h (IQR 16.1–59.3 h). A significant difference was observed between the LAC and CAC groups in terms of age, gender, ATIII, ALB, platelet count and compliance rate (p < 0.05). Accordingly, propensity score matching was used to address baseline imbalances between the groups. After matching, baseline characteristics were well-balanced between the two groups, with no statistically significant differences observed, as shown in Table 1. Because all SMDs were < 0.2, it was reasonable to assume a rough balance in the distribution of matched covariates between the two groups (Supplemental Fig. S1). After matching, the IQR of the APTT ratio for LAC group and CAC group were 1.24 (1.14,1.38) and 1.77 (1.57,2.15), respectively. The median values of the average APTT were 40.5 s for the LAC group and 54.4 s for the CAC group, as shown in Fig. 2B.
Fig. 1.
Flow chart of the selection process for patients. PC-ECMO post-cardiotomy Extracorporeal Membrane Oxygenation, LAC low anticoagulation strategy, CAC conventional anticoagulation strategy group
Fig. 2.
The distribution of mean APTT values in both groups. A represents mean APTT prior to matching, whereas B exhibits the mean APTT following the matching process
Table 1.
Baseline characteristics (before and after matching)
| Parameter [median (IQR)/%] |
Total | LAC | CAC | p | total | LAC | CAC | P |
|---|---|---|---|---|---|---|---|---|
| n | 203 | 108 | 95 | 86 | 43 | 43 | ||
| Age (years) | 62 (54,68) | 59 (52,68) | 63 (58,68) | 0.035 | 63 (57,68) | 65 (55,68) | 62 (57,67) | 0.406 |
| Gender | 62 (30.5) | 25 (23.1) | 37 (38.9) | 0.022 | 30 (34.9) | 14 (32.6) | 16 (37.2) | 0.821 |
| BMI (kg/m2) | 25 (23,27) | 25.0 (23,28) | 24 (23,27) | 0.138 | 24 (22,27) | 24 (22,27) | 24 (23,26) | 0.598 |
| Myocardial infarction | 34 (16.7) | 16 (14.8) | 18 (18.9) | 0.550 | 12 (34.9) | 7 (16.3) | 5 (11.6) | 0.757 |
| Hypertension | 103 (50.7) | 49 (45.4) | 54 (56.8) | 0.136 | 43 (50.0) | 22 (51.2) | 21 (48.8) | 1.000 |
| Hyperlipidemia | 60 (29.7) | 36 (33.3) | 24 (25.5) | 0.291 | 26 (30.2) | 14 (32.6) | 12 (27.9) | 0.815 |
| Diabetes | 48 (23.6) | 24 (22.2) | 24 (25.3) | 0.548 | 22 (25.6) | 9 (20.9) | 13 (30.2) | 0.459 |
| Heart failure | 22 (10.8) | 9 (8.3) | 13 (13.7) | 0.319 | 11 (12.8) | 5 (11.6) | 6 (14.0) | 1.000 |
| Smoker | 82 (40.4) | 42 (38.9) | 40 (42.1) | 0.747 | 54 (62.8) | 26 (60.5) | 28 (65.1) | 0.824 |
| anticoagulant drugs | 23 (11.4) | 13 (12.1) | 10 (10.5) | 0.888 | 11 (12.8) | 7 (16.3) | 4 (9.3) | 0.520 |
| IABP | 138 (68.0) | 69 (63.9) | 69 (72.6) | 0.237 | 58 (67.4) | 30 (69.8) | 28 (65.1) | 0.818 |
| Cardiac arrest | 35 (17.2) | 21 (19.4) | 14 (14.7) | 0.484 | 13 (15.1) | 8 (18.6) | 5 (11.6) | 0.549 |
| D-dimer (ng/mL) | 3280 (1808,8011) | 3217 (1769,7736) | 3410 (1808,8916) | 0.759 | 4217 (2281,9506) | 5806 (2335,10197) | 3435 (2116,8989) | 0.371 |
| ATIII (%) | 45 (31,62) | 54 (39,71) | 38 (52,79) | < 0.001 | 43 (30,60) | 42 (32,61) | 45 (29,60) | 0.743 |
| ALB (g/L) | 25 (20,28) | 26 (23,30) | 22 (19,27) | 0.008 | 25 (21,30) | 26 (24,30) | 25 (20,28) | 0.300 |
| ALT (IU/L) | 154 (47,670) | 129 (45,443) | 198 (52,800) | 0.175 | 179 (54,746) | 156 (33,784) | 183 (56,733) | 0.911 |
| HB (g/L) | 67 (61,73) | 68 (62,75) | 66 (61,72) | 0.158 | 68 (62,74) | 68 (60,76) | 67 (62,74) | 0.832 |
| Platelet count (109/L) | 29 (17,50) | 39 (23,60) | 23 (12,34) | < 0.001 | 29 (16,50) | 30 (22,51) | 27 (14,42) | 0.231 |
| MAP (mmHg) | 50 (42,59) | 51 (40,59) | 50 (42,60) | 0.326 | 50 (42,58) | 50 (40,59) | 50 (43,58) | 0.678 |
| Lactate (mmol/L) | 8.9 (4.9,14.4) | 8.4 (4.5,13.0) | 9.5 (5.5,15.0) | 0.089 | 9.3 (4.5,14.6) | 8.5 (4.5,15.1) | 9.4 (4.5,14.5) | 0.990 |
| CPB | 150 (73.9) | 83 (76.9) | 67 (70.5) | 0.306 | 71 (82.6) | 36 (83.7) | 35 (81.4) | 1.000 |
| Time1 (h) | 35.9 (16.1,59.3) | 36.2 (18.7,47.3) | 35.4 (14.0,66.4) | 0.731 | 36.6 (20.3,47.6) | 37.0 (23.7,45.6) | 35.7 (15.4,65.8) | 0.997 |
| Compliance rate (%) | 50 (32,75) | 63 (43,80) | 39 (26,60) | <0.001 | 50 (32,71) | 50 (29,80) | 50 (36,67) | 0.962 |
| ECMO blood flow(L/min) | 3.4 (3.1,3.7) | 3.4 (3.2,3.7) | 3.4 (3.0,3.7) | 0.554 | 3.4 (3.2,3.7) | 3.4 (3.2,3.7) | 3.4 (3.1,3.9) | 0.710 |
| SOFA | 11 (9,12) | 11 (9,12) | 11 (10,12) | 0.068 | 11 (9,12) | 11 (9,12) | 11 (10,12) | 0.560 |
| Insatallation location | 0.581 | 0.305 | ||||||
| Cannulation in OR | 76 (38.0) | 42 (38.9) | 34 (37.0) | 35 (40.7) | 21 (48.8) | 14 (32.6) | ||
| Cannulation in ICU | 107 (53.5) | 54 (50.0) | 53 (57.6) | 41 (47.7) | 18 (41.9) | 23 (60.5) | ||
| Cannulation in ER | 4 (2.0) | 3 (2.8) | 1 (1.1) | 1 (1.2) | 1 (2.3) | 0 (0.0) | ||
| Cannulation in CCL | 2 (1.0) | 1 (0.9) | 1 (1.1) | 1 (1.2) | 0 (0.0) | 1 (2.3) | ||
| Cannulation in Ward | 11 (5.5) | 8 (7.4) | 3 (3.3) | 6 (7.0) | 3 (7.0) | 3 (7.0) | ||
| Operation (%) | 0.487 | 0.638 | ||||||
| Valve repair/replacement | 39 (19.2) | 17 (15.7) | 22 (23.2) | 18 (20.9) | 7 (16.3) | 11 (25.6) | ||
| Isolated CABG | 81 (39.9) | 44 (40.7) | 37 (38.9) | 30 (34.9) | 17 (39.5) | 13 (30.2) | ||
| Combined CABG/valve surgery | 37 (18.2) | 19 (17.6) | 18 (18.9) | 17 (19.8) | 8 (18.6) | 9 (20.9) | ||
| TAVI | 11 (5.4) | 8 (7.4) | 3 (3.2) | 4 (4.7) | 3 (7.0) | 1 (2.3) | ||
| Other | 35 (17.2) | 20 (18.5) | 15 (15.8) | 17 (19.8) | 8 (18.6) | 9 (20.9) |
Data are presented as medians (25th–75th percentile) or n (%). LAC low anticoagulation strategy, CAC conventional anticoagulation strategy, IQR interquartile range, BMI body mass index, IABP intra-aortic balloon pump, ATIII anti-thrombin III, ALB albumin, ALT alanine aminotransferase, HB hemoglobin, PLT platelets, MAP mean arterial pressure, LAC lactates, CPB cardiopulmonary bypass, ER emergency room, OR operating room, ICU intensive care unit, CCL cardiac cath lab, CABG coronary artery bypass grafting, TAVI transcatheter aortic valve implantation. SOFA Sequential Organ Failure Assessment
Primary outcome
Fifty-one (25.1%) patients experienced ≥ 1 thrombotic event(s) during ECMO. Ischemic stroke was the most prevalent event, occurring in 26 patients (12.8%), followed by arterial thrombosis in 16 patients (8.4%). The study recorded a comprehensive list of all thrombotic and hemorrhagic events, as summarized in Table 2.The proportion of thrombotic complications among coagulation-related events in subgroups with different anticoagulation levels is shown in Fig. S2. No substantial difference was observed in the occurrence of major thrombotic events between the two groups. The overall prevalence of thrombotic complications was 25.1% before matching and 27.9% after matching (Table 3). After matching, no significant difference was observed in the incidence of thrombotic complications between the LAC and CAC groups (30.2% vs. 25.3%; p = 0.810). The cumulative incidence of thrombotic events in the LAC and CAC groups within 15 days following the initiation of ECMO support makes no difference (Fig. 3). Patients were stratified into two groups according to thrombotic complication status: those with thrombotic events and those without. The baseline characteristics of these two groups are summarized in Table S1 (Supplemental Table S1). In multivariate logistic regression analysis, anticoagulation strategy did not emerge as an independent risk factor for thrombotic complications (adjusted odds ratio [OR] 1.08 [95% confidence interval (CI) 0.36–3.26]; p = 0.886) (Supplemental Table S2 for details). The effect of the LAC on thrombotic complications remained consistent across most evaluated subgroups (Fig. 4). There were no significant interactions between the LAC strategy and the variables that defined the subgroups.
Table 2.
Thrombotic and hemorrhagic events in patients on post-cardiotomy extracorporeal membrane oxygenation
| Thrombotic events (59 events were observed in 51 patients) | |
|---|---|
| Ischemic stroke | 26(12.8) |
| Arterial thrombosis | 16(8.4) |
| Venous thrombosis | 5(2.5) |
| circuit change | 5(2.5) |
| Oxygenator change | 2(1.0) |
| Intracardiac Thrombosis | 3(1.5) |
| Pulmonary embolism | 2(1.0) |
| Hemorrhagic events (39 events were observed in 35 patients) | |
| Surgical site bleeding | 3(1.5) |
| Cannulation site bleeding | 11(5.4) |
| Gastrointestinal bleeding | 8(3.9) |
| Oronasal hemorrhage | 6(3.0) |
| Intracranial hemorrhage | 10(4.9) |
| Other | 1(0.5) |
Table 3.
Outcomes
| Parameter [median (IQR)/%] | Before matching | After matching | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Total(203) | LAC(108) | CAC(95) | P | Total(86) | LAC(43) | CAC(43) | P | ||
| Primary outcome | |||||||||
| Thrombotic Complications | 51(25.1) | 27(25.2) | 24(25.3) | 1.000 | 24(27.9) | 13(30.2) | 11(25.3) | 0.810 | |
| Ischemic stroke | 26(12.8) | 11 (10.2) | 15(15.8) | 0.226 | 14(16.3) | 7(16.3) | 7(16.3) | 0.614 | |
| Arterial thrombosis | 16(7.9) | 9(8.3) | 7(7.4) | 0.799 | 8(9.3) | 5(11.6) | 3(7.0) | 0.356 | |
| Secondary outcomes | |||||||||
| Platelet (U) | 2.0(0.0,4.0) | 1.0(0.0,3.0) | 2.0(1.0,5.0) | 0.003 | 2.0(0.8,4.0) | 2.0(0.0,3.0) | 2.0(1.0,5.0) | 0.260 | |
| RBC (U) | 8.0(4.0,14.0) | 8.0(4.0,14.0) | 10.0(6.0,15.0) | 0.017 | 8.0(5.0,14.0) | 8.0(4.0,14.0) | 10.0(6.0,15.5) | 0.115 | |
| FFP (ml) | 400(0,800) | 400(0,800) | 600(400,1000) | 0.001 | 400(200,800) | 400(0,800) | 400(200,1000) | 0.820 | |
| VA-ECMO duration (hours) | 141.8(95.2,202.0) | 123.2(89.4,177.5) | 159.0(119.4,215.3) | 0.001 | 137.7(90.3,203.4) | 119.6(81.7,149.5) | 146.0(111.0,215.0) | 0.015 | |
| Mechanical ventilation time (hours) | 180.3(96.6,269.8) | 154.0(87.8,267.4) | 207.3(101.7,284.5) | 0.053 | 190.0(99.9,270.5) | 169.8(103.8,270.4) | 207.3(98.3,270.8) | 0.607 | |
| Hospital-stay (days) | 22.0(16.0,30.0) | 22.0(16.0,30.0) | 23.0(17.0,31.0) | 0.739 | 21.0(14.8,28.0) | 22.0(16.0,30.5) | 23.0(15.0,31.0) | 0.156 | |
| Time3 | 0.7(0.4,0.8) | 0.7(0.4,0.8) | 0.6(0.3,0.9) | 0.521 | 0.6(0.5,0.8) | 0.6(0.4,0.8) | 0.7(0.5,0.9) | 0.143 | |
| Pulmonary infection | 177(87.2) | 93 (86.1) | 84 (88.4) | 0.779 | 79(91.9) | 36(83.7) | 37(86.0) | 0.738 | |
| Bleeding Complications | 35 (17.2) | 17 (15.7) | 18 (18.9) | 0.676 | 14(16.3) | 5(11.6) | 9(20.9) | 0.382 | |
| Cannulation site bleeding | 11(5.4) | 4(3.7) | 7(7.4) | 0.354 | 3(3.5) | 2(4.7) | 1(2.3) | 0.500 | |
| Intracranial hemorrhage | 10(4.9) | 5(4.6) | 5(5.3) | 0.835 | 5(5.8) | 4(9.3) | 1(2.3) | 0.180 | |
| RRT | 113(55.7) | 52(48.1) | 61(64.2) | 0.022 | 48(55.8) | 22(51.2) | 26(60.5) | 0.515 | |
| In-hospital mortality | 97 (47.8) | 51 (47.2) | 46 (48.4) | 0.976 | 47(54.7) | 28(65.1) | 19(44.2) | 0.083 | |
LAC low anticoagulation strategy, CAC conventional anticoagulation strategy, IQR interquartile range, RBC red blood cells, FFP fresh frozen plasma, VA-ECMO. Veno-arterial extracorporeal membrane oxygenation, Time1 = time to start anticoagulation - time to start ECMO; Time2 = time to end anticoagulation - time to start anticoagulation; Time3 = Time2/ECMO duration, RRT renal replacement therapy
Fig. 3.
The cumulative incidence of thrombotic events in the LAC and CAC groups within 15 days following the initiation of ECMO support. 4A, before matching, LAC (23.7%) vs. CAC(25.8%) P = 0.73; 4B, after matching, LAC(31.3%) vs. CAC(24.1%) P = 0.49
Fig. 4.
Association between low anticoagulation strategy and thrombotic complication across prespecified subgroups. ATIII anti-thrombin III
Secondary outcomes
The overall in-hospital mortality rate was 47.8% before matching and 54.7% after matching. After propensity score matching, no statistically significant difference was observed in in-hospital mortality between the LAC and CAC groups (65.1% vs. 44.2%, p = 0.083) (Table 3). After incorporating the anticoagulation strategy into the multivariate logistic analysis of mortality-related factors, the results indicate that the level of anticoagulation is not an independent risk factor for mortality, with p = 0.09 and OR 95% CI 1.84 (0.91, 3.74). For details, please refer to Table S3.
Complications, including bleeding events, pulmonary infection, and RRT, were common in both the LAC and CAC groups. Thirty-five (17.2%) patients developed ≥ 1 bleeding event(s) during ECMO. Cannulation site bleeding was the most frequent complication, occurring in 11 (5.4%) patients, followed by intracranial hemorrhage in 10 (4.9%). The details are summarized in Table 2. The proportion of bleeding complications among coagulation-related events in subgroups with different anticoagulation levels is shown in Fig. S2. There was no significant difference in the incidence of major types of bleeding events between the two groups (Table 3). After matching, LAC was associated with similar rates of bleeding complications (11.6% vs. 20.9%; p = 0.382), pulmonary infection (83.7% vs. 86.0%; p = 0.738), and RRT (51.2% vs. 60.5%; p = 0.515) (Table 3).
Before matching, requirements for blood product transfusion were significantly lower in the LAC group compared with the CAC group. Specifically, the LAC group required fewer units or millilitres of RBCs (8.0 vs. 10.0; p = 0.017), FFP (400 vs. 600; p = 0.001), and platelet (1.0 vs. 2.0; p = 0.003). After matching, between-group comparisons of transfusion volumes showed no significant differences: RBCs (8.0 vs. 10.0; p = 0.115); FFP (400 vs. 400; p = 0.820); and PLT (2.0 vs. 2.0; p = 0.260).
The proportion of anticoagulation time during the ECMO procedure, denoted as Time 3, is depicted in Supplemental Fig. S3A and Fig. S3B for before and after matching periods, respectively. The distribution at Time 3 did not exhibit a statistically significant difference between the two groups before and after matching (0.7 vs. 0.6 [p = 0.521]; and 0.6 vs. 0.7 [p = 0.143]) (Table 3).
Discussion
ECMO advancements now result gentler anticoagulation strategies to balance effective circulatory support with reduced complications, especially life-threatening bleeding risks. This article offers insights into the safety and efficacy of a low anticoagulation strategy in PC-ECMO patients.
In summary, the low anticoagulation strategy shows no significant difference compared to conventional anticoagulation in terms of thrombotic complications, other ECMO-related complications, and mortality rates. Notably, the association between anticoagulation strategy and thrombotic complication risk remained consistent across all ATIII levels. There was no significant correlation between the time of initiation of anticoagulation therapy and the occurrence of thrombotic complications.
Existing studies fail to establish the most effective monitoring approaches and thresholds for anticoagulation in ECMO patients, leading to inconsistent definitions of low anticoagulation. Previous researches on anticoagulation strategy have defined different low anticoagulation criteria based on APTT [23], ACT [25], and anti-factor Xa levels [26], respectively. Nevertheless, despite this variability, a consistent finding across these studies is that using a LAC strategy did not lead to a significant increase in thrombotic events. However, recent meta-analyses have not found a correlation between APTT levels and the occurrence of thrombosis and bleeding [31–32]. Given its general availability and low cost, APTT continues to be a widely utilized monitoring method.
The duration of ECMO support is a significant factor influencing thrombus formation [33]. The observed between-group difference in support time (CAC > LAC) may confound outcome interpretation and should be accounted for in sensitivity analyses. When duration of ECMO support was included in the multivariate analysis of thrombotic complications (Table S2), the conclusion remained that anticoagulation strategy is not an independent risk factor for thrombotic complications. The results of previous meta-analyses indicate that the incidence of thrombotic complications in ECMO patients ranges from 26 to 36% [31–32]. Literature suggests that up to 90% of patients undergoing PC-ECMO experience thrombosis (autopsy findings) [34]. Although the incidence of thrombotic complications in current studies is comparable to that reported in meta-analyses, there remains a consistent underestimation of thrombotic complications due to the lack of regular radiological studies and postmortem examinations.
It is noteworthy that after matching, the mortality rate in the low anticoagulation group was 65.1%, while that in the conventional anticoagulation group was 44.2%. Although statistically non-significant (P = 0.083), the 20.9% absolute increase in mortality with low anticoagulation may warrant clinical consideration. After matching, only a small number of patients (43 pairs) were included in the statistical analysis, which raises the possibility of a false negative due to the small sample size. We performed multivariable regression analysis for mortality predictors in the full cohort (n = 203). The multivariate regression analysis showed that anticoagulation intensity was not an independent risk factor for mortality (p = 0.09, OR 95%CI = 1.84 (0.91–3.74). Further studies involving larger cohorts are required to determine the effects of a low anticoagulation approach on mortality rates.
Approximately 48.5% of patients undergoing VA-ECMO treatment experience bleeding complications, as reported in the literature [8]. In this study, patients requiring re-exploration for hemorrhage control within 24 h after ECMO initiation were excluded, which likely contributed to the lower observed incidence of bleeding complications. Consequently, in contrast to previous research, this study failed to reveal a correlation between a LAC strategy and a reduction in bleeding complications [23]. Before matching, a preliminary data analysis revealed that the CAC group exhibited a higher number of RBC, platelet, and FFP transfusions. This finding is consistent with previous investigation [25]. Analysis of baseline data before matching demonstrated that individuals in the CAC group tended to be older, predominantly female, and presented with lower ALB, ATIII, and PLT levels, indicating a more advanced disease state [35–36]. The lower worst platelet counts in the CAC group compared to the LAC group may be attributed to the higher dose of heparin, which could potentially induce DIC or bleeding, thereby leading to increased platelet consumption [37]. Heparin exerts its effect by consuming ATIII, and the higher dose of heparin in the CAC group compared to the LAC group results in lower ATIII levels [38–39]. After matching, patient characteristics were well balanced, and the statistical outcomes became more robust. There were no statistically significant differences in blood product transfusion volumes between the two groups.
Current studies have not determined the lower limit of endogenous AT concentration that permits complete heparin activity at any given dose [38–39]. According to the ELSO guidelines, if a patient on maximum dose of unfractionated heparin (UFH) fails to achieve the desired anticoagulation level, the guidelines recommend considering the augmentation of AT, aiming for an AT activity of 50–80% or higher [20]. Therefore, this study explored the effect of a LAC strategy on thrombotic complications among individuals with ATIII > 50% and ATIII ≤ 50%. Our findings revealed a consistent association between low-anticoagulation strategy and the likelihood of thrombotic complications that persisted uniformly across varying ATIII levels. A recent prospective observational study reported that ATIII levels were not correlated with the body’s response to heparin in patients supported by VA-ECMO [36]. Our study corroborates these earlier findings to a certain extent. More research is needed to explore if monitoring ATIII levels and adding extra ATIII during V-A ECMO is beneficial for enhancing heparin response.
Another notable finding of the study was the lack of significant correlation between the timing of anticoagulation initiation and the incidence of thrombotic complications (Table 2). The findings demonstrate that the primary goal of anticoagulation therapy during ECMO support is to maintain monitoring parameters within the therapeutic range. This suggests that ensuring effective anticoagulation is more critical than the precise timing of therapy initiation.
Limitations
The inherent complexities of anticoagulation management and monitoring in ECMO patients, combined with the retrospective nature of this study, introduce several limitations to the research. Firstly, the use of the APTT ratio, which is related to APTT, as a method for anticoagulation monitoring lacks evidence-based medical support. Although anti-factor Xa is increasingly playing a significant role in anticoagulation monitoring, this test has only recently been implemented at our institution, resulting in a lack of relevant data. Secondly, the retrospective identification of thrombosis, which is only possible when clinical examinations are performed, may lead to underestimation. And, the absence of routine imaging examinations and postmortem analyses may also contribute to the underreporting of thrombotic events. While propensity score matching and adjusted logistic regression models were used to minimize confounding, the retrospective study design remains susceptible to residual bias from unmeasured variables. For example, the absence of data regarding factors such as the dosage of UFH administered and fibrinogen levels may have influenced the observed outcomes. Given the retrospective nature of the study, anticoagulation regimens for some patients may have been determined based on their specific health conditions, potentially influencing the research outcomes. Fourthly, the relatively small sample size of only 86 matched patients and the single-center design may introduce bias and limit the generalizability of the subgroup analyses, necessitating further validation and long-term follow-up for more definitive conclusions.
Future perspectives and outlook
Current monitoring indicators (APTT, ACT, anti-factor Xa) have limitations, and no single parameter fully captures ECMO patients’ coagulation status. Future efforts should focus on advancing multimodal monitoring in clinical practice. The majority of current research consists of single-center, small-scale retrospective studies. Future efforts should focus on conducting higher-quality prospective randomized controlled trials to evaluate the efficacy of various anticoagulation strategies. Machine learning techniques can also be used to analyze the anticoagulation data of patients with ECMO to provide support for clinical decision-making.
Conclusion
Adopting a less aggressive anticoagulation regimen appeared to be feasible in patients who undergo VA-ECMO after cardiac surgery. Despite confirming the safety of a low anticoagulation strategy, this study underscores the need for more extensive randomized controlled trials to fully assess the security and practicality of applying this approach in VA-ECMO treatment.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Acknowledgements
We thank Xiaomeng Wang and Tingting Wu for their assistance in the guidance on statistical methodology.
Abbreviations
- PC-ECMO
Post-cardiotomy extracorporeal membrane oxygenation
- LAC
Low anticoagulation
- CAC
Conventional anticoagulation
- APTT
Activated partial thromboplastin time
- DTIs
Direct thrombin inhibitors
- ELSO
Extracorporeal Life Support Organisation
- VV-ECMO
Veno-venous ECMO
- VA-ECMO
Veno-arterial ECMO
- ATIII
Antithrombin III
- ALT
Alanine aminotransferase
- FFP
Fresh frozen plasma
- RBC
Red blood cells
- RRT
Renal replacement therapy
- SD
Standardized difference
- SMD
Standardized mean difference
- ACT
Activated clotting time
- UFH
Unfractionated heparin
- CABG
Coronary artery bypass grafting
- ALB
Albumin
- CPB
Cardiopulmonary bypass
- IABP
Intra-aortic balloon pump
- HB
Hemoglobin
Author contributions
ZD and TXH were responsible for conception and design of the study. YW, HF, JL collected the patient data. YW, LW performed the statistical analysis. CL, XH, SZ and HW interpreted the outcomes of data analysis. YW and LW contributed to the writing of the main manuscript, HF, JL, HX and SZ contributed the editing, and formatting of the article. ZD, TXH, CL, LW and HW contributed to the supervision of the study and revising of the manuscript. All authors contributed substantially to the final approval of the version to be published, and agreed to be accountable for all aspects of the work.
Funding
This work was supported by the National Key Clinical College Construction Program (Critical Care Medicine), National Natural Science Foundation of China (No. 82100408, to X Hao), Beijing Nova Program (No. 20220484043, to C Li), Young Elite Scientists Sponsorship Program by CAST (2022QNRC001, to L Wang), National Natural Science Foundation of China (82200433, to L Wang), and Beijing Hospitals Authority Youth Programme (QML20230602, to L Wang). The National Natural Science Foundation of China (No. 82170400), Beijing Municipal Administration of Hospitals Incubating Program (PX2024025, to Z Du), The National Key Research and Development Program of China (Grant Nos. 2021YFC2701700 and 2021YFC2701703, to Z Du).
Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
The study was carried out in accordance with the guidelines of the Declaration of Helsinki. This study is a retrospective analysis using fully anonymized electronic medical record data from Beijing Anzhen Hospital (2018–2023), where contacting the participants is not feasible. The study was reviewed and approved by the Ethics Committee of Beijing Anzhen Hospital (Approval No.2025004x), and the requirement for informed consent was waived in accordance with Article 39 of the Ethical Review Measures for Biomedical Research Involving Humans.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Yan Wang and Liangshan Wang contributed equally to this work and share co-first authorship.
Zhongtao Du and Xiaotong Hou contributed equally to this work and share co-corresponding authorship.
Contributor Information
Zhongtao Du, Email: zhongtaodu@126.com.
Xiaotong Hou, Email: xt.hou@ccmu.edu.cn.
References
- 1.Elsharkawy HA, Li L, Esa WAS, Sessler DI, Bashour CA. Outcome in patients who require venoarterial extracorporeal membrane oxygenation support after cardiac surgery. J Cardio-Thorac Vasc Anesth. 2010;24:946–51. [DOI] [PubMed] [Google Scholar]
- 2.Ariyaratnam P, McLean LA, Cale ARJ, Loubani M. Extra-corporeal membrane oxygenation for the post-cardiotomy patient. Heart Fail Rev. 2014;19:717–25. [DOI] [PubMed]
- 3.Rajsic S, Breitkopf R, Jadzic D, Popovic Krneta M, Tauber H, Treml B. Anticoagulation strategies during extracorporeal membrane oxygenation: A narrative review. J Clin Med. 2022;11(17):5147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kowalewski M, Zieliński K, Brodie D, et al. Venoarterial extracorporeal membrane oxygenation for postcardiotomy Shock-Analysis of the extracorporeal life support organization registry. Crit Care Med. 2021;49(7):1107–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Vallabhajosyula S, Arora S, Sakhuja A, et al. Trends, predictors, and outcomes of temporary mechanical circulatory support for postcardiac surgery cardiogenic shock. Am J Cardiol. 2019;123(3):489–97. [DOI] [PubMed] [Google Scholar]
- 6.Mariani S, Heuts S, van Bussel BCT et al. Patient and management variables associated with survival after postcardiotomy extracorporeal membrane oxygenation in adults: the PELS-1 multicenter cohort study. J Am Heart Assoc. 2023;12(14):e029609. [DOI] [PMC free article] [PubMed]
- 7.Lorusso R, Raffa GM, Alenizy K, et al. Structured review of post-cardiotomy extracorporeal membrane oxygenation: part 1-Adult patients. J Heart Lung Transpl. 2019;38(11):1125–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Rajsic S, Treml B, Jadzic D, et al. Extracorporeal membrane oxygenation for cardiogenic shock: a meta-analysis of mortality and complications. Ann Intensive Care. 2022;12(1):93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Thomas J, Kostousov V, Teruya J. Bleeding and thrombotic complications in the use of extracorporeal membrane oxygenation. Semin Thromb Hemost. 2018;44(1):20–9. [DOI] [PubMed] [Google Scholar]
- 10.Martucci G, Giani M, Schmidt M, et al. Anticoagulation and bleeding during Veno-Venous extracorporeal membrane oxygenation: insights from the PROTECMO study. Am J Respir Crit Care Med. 2024;209(4):417–26. [DOI] [PubMed] [Google Scholar]
- 11.Schaefer AK, Latus M, Riebandt J, et al. Bleeding and thrombotic events in post-cardiotomy extracorporeal life support. Eur J Cardiothorac Surg. 2023;63(4):ezad072. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Melehy A, Ning Y, Kurlansky P, et al. Bleeding and thrombotic events during extracorporeal membrane oxygenation for postcardiotomy shock. Ann Thorac Surg. 2022;113(1):131–7. [DOI] [PubMed] [Google Scholar]
- 13.Nunez JI, Gosling AF, O’Gara B, et al. Bleeding and thrombotic events in adults supported with venovenous extracorporeal membrane oxygenation: an ELSO registry analysis [published correction appears in. Intensive Care Med. 2022;48(5):644–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Descamps R, Moussa MD, Besnier E, et al. Anti-Xa activity and hemorrhagic events under extracorporeal membrane oxygenation (ECMO): a multicenter cohort study. Crit Care. 2021;25(1):127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Li C, Cai T, Xie H, et al. Risk factors and outcomes for patients with bleeding complications receiving extracorporeal membrane oxygenation: an analysis of the Chinese extracorporeal life support registry. Artif Organs. 2022;46(12):2432–41. [DOI] [PubMed] [Google Scholar]
- 16.Aubron C, DePuydt J, Belon F et al. Predictive factors of bleeding events in adults undergoing extracorporeal membrane oxygenation. Ann Intensive Care.2016;6(1):97. [DOI] [PMC free article] [PubMed]
- 17.Kalbhenn J, Zieger B, Bleeding During Veno-Venous ECMO. Prevention and treatment. Front Med (Lausanne). 2022;9:879579. Published 2022 May 23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Rajsic S, Treml B. Anticoagulation monitoring during ECMO support: monitor or flip a coin?? Clin Cardiol. 2024;47(12):e70061. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Protti A, Iapichino GE, Di Nardo M, Panigada M, Gattinoni L. Anticoagulation Management and Antithrombin Supplementation Practice during Veno-venous Extracorporeal Membrane Oxygenation: A Worldwide Survey. Anesthesiology. 2020;132(3):562-570. [DOI] [PubMed] [Google Scholar]
- 20.McMichael ABV, Ryerson LM, Ratano D, Fan E, Faraoni D, Annich GM. 2021 ELSO adult and pediatric anticoagulation guidelines. ASAIO J. 2022;68(3):303–10. [DOI] [PubMed] [Google Scholar]
- 21.Lv X, Deng M, Wang L, Dong Y, Chen L, Dai X. Low vs standardized dose anticoagulation regimens for extracorporeal membrane oxygenation: A meta-analysis. PLoS ONE. 2021;16(4):e0249854. Published 2021 Apr 8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Wood KL, Ayers B, Gosev I, et al. Venoarterial-Extracorporeal membrane oxygenation without routine systemic anticoagulation decreases adverse events. Ann Thorac Surg. 2020;109(5):1458–66. [DOI] [PubMed] [Google Scholar]
- 23.Song K, Kim JB. Safety of low-dose anticoagulation in extracorporeal membrane oxygenation using the permanent life support system: a retrospective observational study. J Yeungnam Med Sci. 2023;40(3):276–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Raman J, Alimohamed M, Dobrilovic N, Lateef O, Aziz S. A comparison of low and standard anti-coagulation regimens in extracorporeal membrane oxygenation. J Heart Lung Transpl. 2019;38(4):433–9. [DOI] [PubMed] [Google Scholar]
- 25.Yeo HJ, Kim DH, Jeon D, Kim YS, Cho WH. Low-dose heparin during extracorporeal membrane oxygenation treatment in adults. Intensive Care Med. 2015;41(11):2020–1. [DOI] [PubMed] [Google Scholar]
- 26.Aubron C, McQuilten Z, Bailey M, et al. Low-Dose versus therapeutic anticoagulation in patients on extracorporeal membrane oxygenation: A pilot randomized trial. Crit Care Med. 2019;47(7):e563–71. [DOI] [PubMed] [Google Scholar]
- 27.Wang L, Yang F, Wang X, et al. Predicting mortality in patients undergoing VA-ECMO after coronary artery bypass grafting: the REMEMBER score. Crit Care. 2019;23:11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Li CL, Wang H, Jia M, Ma N, Meng X, Hou XT. The early dynamic behavior of lactate is linked to mortality in postcardiotomy patients with extracorporeal membrane oxygenation support: A retrospective observational study. J Thorac Cardiovasc Surg. 2015;149:1445–50. [DOI] [PubMed] [Google Scholar]
- 29.Lorusso R, Whitman G, Milojevic M, et al. 2020 EACTS/ELSO/STS/AATS expert consensus on post-cardiotomy extracorporeal life support in adult patients. J Thorac Cardiovasc Surg. 2021;161(4):1287–331. [DOI] [PubMed] [Google Scholar]
- 30.Granja T, Hohenstein K, Schüssel P, et al. Multi-Modal characterization of the coagulopathy associated with extracorporeal membrane oxygenation. Crit Care Med. 2020;48(5):e400–8. [DOI] [PubMed] [Google Scholar]
- 31.Rajsic S, Breitkopf R, Treml B, et al. Association of APTT-Guided anticoagulation monitoring with thromboembolic events in patients receiving V-A ECMO support: A systematic review and Meta-Analysis. J Clin Med. 2023;12(9):3224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Rajsic S, Treml B, Jadzic D, et al. APTT-guided anticoagulation monitoring during ECMO support: A systematic review and meta-analysis. J Crit Care. 2023;77:154332. [DOI] [PubMed] [Google Scholar]
- 33.Mansour A, Flecher E, Schmidt M, et al. Bleeding and thrombotic events in patients with severe COVID-19 supported with extracorporeal membrane oxygenation: a nationwide cohort study. Intensive Care Med. 2022;48(8):1039–52. [DOI] [PubMed] [Google Scholar]
- 34.Rastan AJ, Lachmann N, Walther T, et al. Autopsy findings in patients on postcardiotomy extracorporeal membrane oxygenation (ECMO). Int J Artif Organs. 2006;29(12):1121–31. [DOI] [PubMed] [Google Scholar]
- 35.Wang F, Zhang G, Lu Z, et al. Antithrombin III/ SerpinC1 insufficiency exacerbates renal ischemia/ reperfusion injury. Kidney Int. 2015;88(4):796–803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Mansour A, Berahou M, Odot J, et al. Antithrombin levels and heparin responsiveness during venoarterial extracorporeal membrane oxygenation: A prospective Single-center cohort study. Anesthesiology. 2024;140(6):1153–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Oude Lansink-Hartgring A, de Vries AJ, Droogh JM, van den Bergh WM. Hemorrhagic complications during extracorporeal membrane oxygenation - The role of anticoagulation and platelets. J Crit Care. 2019;54:239–43. [DOI] [PubMed] [Google Scholar]
- 38.Helms J, Frere C, Thiele T, et al. Anticoagulation in adult patients supported with extracorporeal membrane oxygenation: guidance from the scientific and standardization committees on perioperative and critical care haemostasis and thrombosis of the international society on thrombosis and haemostasis. J Thromb Haemost. 2023;21(2):373–96. [DOI] [PubMed] [Google Scholar]
- 39.Panigada M, Cucino A, Spinelli E, et al. A randomized controlled trial of antithrombin supplementation during extracorporeal membrane oxygenation. Crit Care Med. 2020;48(11):1636–44. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.