Abstract
OBJECTIVES
Haemorrhagic and thrombotic complications are known obstacles in extracorporeal life support (ECLS), and patients requiring post-cardiotomy (PC)-ECLS are particularly prone. The objective of this study was to characterize the incidence, type and clinical relevance of bleeding and thrombotic events in patients on PC-ECLS.
METHODS
A total of 504 patients receiving PC-ECLS between 2000 and 2021 at a single centre were included in a retrospective analysis. Incidence and type of haemorrhagic and thrombotic complications in patients on PC-ECLS were assessed. Overall survival was compared, and perioperative risk factors for bleeding and thrombotic events were assessed by binary logistic regression.
RESULTS
Of the 504 patients requiring PC-ECLS, 196 patients (38.9%) had 235 bleeding events [surgical site: n = 135 (26.8%); cannulation site: n = 68(13.4%); requiring surgical revision: n = 39 (7.7%); cannulation site change: n = 17 (3.4%); fatal cannulation site bleeding: n = 4(0.8%); intracranial haemorrhage: n = 11 (2.1%); gastrointestinal haemorrhage: n = 8 (1.6%); pulmonary haemorrhage: n = 8 (1.6%); and intra-abdominal/retroperitoneal haemorrhage: n = 5 (1%)]. Overall mortality was higher in patients with major bleeding complications than in patients without bleeding complications (P < 0.0001).
A total of 74 patients (14.7%) had 84 thrombotic events [ischaemic stroke, n = 39 (7.7%); cannula/circuit thrombosis, n = 26 (5.2%); peripheral embolism, n = 11 (2.2%); device exchange for haemolysis, n = 8 (1.6%)]. Another 246 patients (48.8%) had at least 1 haemocompatibility-related adverse event. Preoperative dual antiplatelet therapy [adjusted odds ratio (OR): 1.83, 95% confidence interval (CI): 1.063–3.137] and ECLS duration (adjusted OR: 1.14, 95% CI: 1.086–1.197) were identified as independent risk factors for haemorrhage. Prior stroke/transient ischaemic attack (adjusted OR: 1.91, 95% CI: 1.08–3.83) and ECLS duration (adjusted OR: 1.09, 95% CI: 1.04–1.15) were identified as risk factors for thrombotic events.
CONCLUSIONS
Bleeding complications in patients on ECLS are common and significantly impair survival. Nearly half of the patients were affected by any haemocompatibility-related event.
Keywords: post-cardiotomy extracorporeal life support, bleeding in ECLS, extracorporeal membrane oxygenation
Extracorporeal life support (ECLS) is an effective and valuable treatment option for patients in cardiogenic shock requiring temporary mechanical circulatory support.
INTRODUCTION
Extracorporeal life support (ECLS) is an effective and valuable treatment option for patients in cardiogenic shock requiring temporary mechanical circulatory support. ELCS-related complications are common and yield a high morbidity and mortality in these critically ill patients [1–3]. Adverse events related to haemocompatibility, i.e. bleeding and thrombotic events, account for a large proportion of all ECLS-related complications and represent a major burden in patients on ECLS in general [1, 4–6]. Bleeding events have been identified as the complications most affecting survival in patients with cardiogenic shock requiring temporary mechanical circulatory support [7]; thus, awareness and preventive measures to avoid these events are clearly necessary to improve outcome.
Although the ECLS circuit and related coagulopathy may increase the probability for both bleeding and thrombosis, the risk for developing haemorrhagic complications is greatly increased by the impact of recent major surgery and related tissue trauma as well as by coagulopathy caused by prior cardiopulmonary bypass (CPB). Consequently, a rate of bleeding events as high as 45% has been described in the PC setting [1, 8]. Ongoing haemorrhage during ECLS adversely impacts ECLS flow by constant volume loss and subsequent substitution, leading to an unstable intravascular volume status. The latter leads to suboptimal support conditions and limited efficiency of ECLS with impaired end organ perfusion and increased risk of secondary renal and liver failure and potential transfusion-related complications. Thromboembolic complications such as ischaemic stroke and circuit thrombosis, although less frequent than haemorrhage, are associated with high morbidity and mortality [4, 9].
Available data on the incidence and consequences of haemorrhagic and thrombotic events in the post-cardiotomy (PC) setting are limited by small sample sizes or the reporting of mixed cohorts of PC and non-PC patients [10].
The primary goal of this investigation was a detailed characterization of the incidence, location, clinical relevance and risk factors of haemorrhagic complications in a large, single-centre PC-ECLS cohort. The secondary goal was to assess the incidence and types of thrombotic events that occur in patients on PC-ECLS in order to provide comprehensive data on haemocompatibility-related PC-ECLS complications.
METHODS
Ethics statement
The Medical University of Vienna PC-ECLS registry (institutional review board no. 1086/2019, date: 16/05/2019) and the present study (institutional review board no. 1632/2021, date: 06/08/2021) were approved by the institutional ethics committee of the Medical University of Vienna. Informed consent was waived due to the retrospective study design.
Study population
All consecutive patients receiving PC-ECLS at a single centre from February 2000 to April 2021 were included in an institutional registry described previously [11].
Inclusion criteria
All patients ≥18 years of age who received ECLS after cardiac surgery either intraoperatively for CPB weaning failure or within 72 h from the end of the operation and who did not have any exclusion criteria, were included in the study (Fig. 1).
Figure 1:
Study flow chart.
Exclusion criteria
Exclusion criteria were age <18 years, ECLS initiated prior to the operation, ECLS initiated >72 h after the end of the operation, duration of the ECLS run < 6 h, ECLS initiated at the time of a left ventricular assist device (LVAD) implant for temporary right ventricular support, thoraco-abdominal aortic aneurysm repair, patients on ECLS transferred from other hospitals and patients with missing outcome or procedural information.
Clinical definitions and end points
All patients were assessed for the occurrence of haemorrhagic and thrombotic events during the ECLS run as defined below.
The primary end point was overall survival in patients with compared to patients without haemorrhagic complications during PC-ECLS.
Secondary end points were type and location of haemorrhagic events in patients on PC-ECLS. Additionally, the occurrence and types of thrombotic events during PC-ECLS were assessed as secondary end points.
Definition of a bleeding event
Haemorrhagic complications were defined on the basis of the registry definitions of the Extracorporeal Life Support Organization (ELSO) as any haemorrhage requiring surgical intervention or a red-blood-cell transfusion > 20 ml/kg/24 h of packed red blood cells or >3 units of packed red blood cells/24 h per calendar day. (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).
In addition, any haemorrhagic event leading to haemodynamic instability due to volume depletion, or limiting vital organ function (e.g. intracranial haemorrhage or cardiac tamponade), was considered a major bleeding event irrespective of transfusion requirements or intervention (e.g. fatal intracranial haemorrhage). In case of surgical site haemorrhage, the decision for bleeding revision was based on a case-to-case evaluation considering the following criteria: ongoing chest tube output ≥ 200 cc/h for ≥3 h, evidence of haematopericardium, cardiac tamponade or haematothorax. Additionally, criteria such as coagulation disorders that need correction, the presence of hypothermia, the presence or absence of haemodynamic instability and suspicion of a potentially active, surgical source of bleeding were considered.
Definition of thrombotic events
Adverse events related to thromboembolism included the following: (i) Circuit or cannula thrombosis that was treated with exchange of circuit components or led to termination of ECLS; (ii) ischaemic stroke confirmed by cerebral computed tomography, with a modified Rankin scale (MRS) at the time of hospital discharge ≥2 according to the “Proposed standardized neurological end-points for cardiovascular clinical trials” [12]; and (iii) peripheral thromboembolism other than cerebral thromboembolism (e.g. clinically apparent embolism to limb, gastrointestinal or renal arteries).
MANAGEMENT OF PATIENTS ON EXTRACORPOREAL LIFE SUPPORT
Indications for extracorporeal life support
The interdisciplinary decision-making process for PC-ECLS implementation followed at our institution has been described elsewhere [11]. Briefly, ECLS was initiated either in the operating room for failure of CPB weaning or within 72 h from the end of the operation for post-cardiotomy low cardiac output syndrome, haemodynamic or respiratory instability necessitating ECLS or postoperative cardiac arrest.
Extracorporeal life support cannulation
Cannulation for PC-ECLS is performed either via indirect axillary artery cannulation via a side graft or femoral arterial cannulation (percutaneous or surgical). An antegrade leg perfusion cannula is placed in case of femoral arterial cannulation at the time of ECLS initiation at the discretion of the implanting surgeon and has become the standard approach in recent years. In the case of mediastinal (central) cannulation, cannulas are placed via the ascending aorta and right atrium.
Anticoagulation on extracorporeal life support
The institutional protocol for anticoagulation on PC-ECLS has been described previously [11, 13]. When ELCS is implemented in the operating room for failure to wean from CPB, full-dose protamine is administered for complete heparin reversal, and anticoagulation for ECLS is started 24–48 h postoperatively, given that no bleeding tendency is present. Unfractionated heparin is used at a continuous intravenous infusion rate of 7.5–20 U/kg/h and monitored by activated partial thromboplastin time with a target therapeutic range of 1.5–2.5 x baseline. If heparin-induced thrombocytopaenia is suspected, anticoagulation is switched to argatroban and continued in case heparin-induced thrombocytopaenia is confirmed by diagnostic testing [14]. Anticoagulation is reduced at the discretion of the treating intensivists and surgeons in case of a tendency to bleed and discontinued in case of severe mediastinal haemorrhage [13].
Data collection and follow-up
Baseline characteristics, medications, laboratory data, in-hospital clinical course including the occurrence of bleeding and thrombotic events and outcome data were retrieved from hospital records. Survival upon follow-up was assessed by patient records and the federal Statistics Austria database (Statistics Austria, Vienna, Austria).
Statistical analyses
All statistical analyses were performed using Prism for Mac OS Version 9.3.1 (GraphPad Software Inc., San Diego, CA, USA). Data are presented as absolute numbers and percentages or medians and interquartile ranges. Group comparisons of categorical variables were made using the χ2 test for all variables with n > 5 in each group, and the Fisher exact test for variables with n ≤ 5 in 1 or more of the groups. The Mann–Whitney U test was used for group comparisons of continuous variables between 2 groups, and the Kruskal-Wallis test was used for group comparisons of continuous variables between more than 2 groups. Survival was estimated using Kaplan–Meier curves, and groups were compared by the log-rank test. Binary logistic regression was used to identify risk factors for haemorrhagic and thrombotic complications in patients on ECLS, including all clinically relevant variables of interest with P < 0.1 in the univariable analysis, and no stepwise methods were applied. Two-sided P < 0.05 was considered statistically significant.
RESULTS
Patients
Between February 2000 and April 2021, a total of 607 consecutive patients required PC-ECLS at our centre. After application of the inclusion and exclusion criteria listed in Fig. 1, we included 504 patients in the analysis.
In the primary analysis, patients were grouped according to the occurrence of haemorrhagic events during the ECLS run: group 1: no bleeding event (n=308); group 2: 1 or more bleeding events (n=196).
The secondary analysis included haemorrhagic and thrombotic events in patients on ECLS, and patients were divided into 4 groups: group 1: neither a haemorrhagic or a thrombotic event while on ECLS (n=258); group 2: one or more haemorrhagic event while on ECLS (n=172); group 3: One or more thrombotic events while on ECLS (n=50); group 4: both haemorrhagic and thrombotic events while on ECLS (n=24).
Incidence of haemorrhagic and thrombotic events
A total of 196 patients (38.9%) developed 1 or more major bleeding event during the ECLS run, and 308 patients (61.1%) had no bleeding event. In total, 235 bleeding events occurred in 196 patients while on ECLS.
Regarding all haemocompatibility-related adverse events (HRAEs), 258 patients (51.2%) had no event, 172 patients (34.1%) had haemorrhagic events only, 50 patients (9.9%) had thrombotic events only and 24 patients (4.8%) had both haemorrhagic and thrombotic events (Fig. 1). An overview of all haemorrhagic and thrombotic events in the cohort is given in Fig. 2.
Figure 2:
Overview of haemocompatibility-related adverse events.
Types of haemorrhagic events
Surgical site bleeding was most frequent, occurring in 135 patients (26.8%), followed by cannulation site-related bleeding in 68 patients (13.4%). Details are given in Table 1. Eleven patients (2.1%) developed intracranial haemorrhage while on ECLS. Severe gastrointestinal bleeding occurred in 8 cases (1.6%), pulmonary haemorrhage in 8 cases (1.6%), intra-abdominal haemorrhage in 3 (0.6%) and retroperitoneal bleeding in 2 cases (0.4%) (Fig. 1; Table 1).
Table 1:
Haemorrhagic and thrombotic events in patients on extracorporeal life support: Overview of incidence and type
| Total study cohort (n = 504) (n = 436) |
|
|---|---|
| Haemorrhagic events (235 events occurred in 196 patients) | |
| Surgical site bleeding | 135 (26.8) |
| Re-exploration | 131 (26) |
| Surgical source of bleeding identified | 56 (11.1) |
| Mediastinal/intrapericardial | 119 (23.6) |
| Haematothorax | 42 (8.3) |
| Cannulation site bleeding | 68 (13.4) |
| Surgical revision of cannulation site | 39 (7.7) |
| Change of cannulation site for bleeding | 17 (3.4) |
| Fatal cannulation site haemorrhage | 4 (0.8) |
| Central nervous system/intracranial haemorrhage | 11 (2.1) |
| Gastrointestinal | 8 (1.6) |
| Pulmonary/bronchial | 8 (1.6) |
| Intraperitoneal | 3 (0.6) |
| Retroperitoneal | 2 (0.4) |
| Thrombotic events (84 events occurred in 74 patients) | |
| Ischaemic stroke (MRS ≥ 2) | 39 (7.7) |
| Ischaemic stroke (MRS ≥ 4) | 25 (5) |
| Circuit/cannula thrombosis | 26 (5.2) |
| Device/cannula exchange | 12 (2.4) |
| ECLS explanted | 14 (2.8) |
| Oxygenator/circuit change for haemolysis | 8 (1.6) |
| Peripheral embolism | 11 (2.2) |
ECLS: extracorporeal life support; MRS: modified Rankin scale.
Baseline characteristics and preoperative anticoagulant and antiplatelet medication
There were no differences in baseline characteristics and laboratory parameters between patients with and without haemorrhagic events (Table 2).
Table 2:
Baseline characteristics, preoperative laboratory parameters and preoperative anticoagulant and antiplatelet medications of patients with and without haemorrhagic events on post-cardiotomy extracorporeal life support
| Total study population (n = 504) | No bleeding event (n = 308) | Bleeding event (n = 196) | P-value | |
|---|---|---|---|---|
| Baseline data and medical history | ||||
| Age (years) | 66.8 (57.1-73.7) | 66.7 (56.7-73.7) | 66.9 (57.4-73.6) | 0.70 |
| Male | 333 (66.1) | 197 (64) | 136 (69.4) | 0.21 |
| EuroSCORE II | 14.2 (5.9-30.9) | 13.8 (5-28.2) | 14.8 (6.6-32) | 0.13 |
| BMI (kg/m2) | 26.8 (23.9-30.1) | 26.8 (23.9-30.6) | 26.6 (24.2-29.4) | 0.26 |
| Insulin-dependent diabetes mellitus | 40 (7.9) | 26 (8.4) | 14 (7.1) | 0.6 |
| Coronary artery disease | 274 (54.4) | 173 (56.2) | 101 (51.5) | 0.31 |
| Peripheral arterial disease | 74 (14.7) | 48 (15.6) | 26 (13.2) | 0.47 |
| Cerebrovascular disease | 71 (14.1) | 43 (14) | 28 (14.3) | 0.92 |
| Previous stroke/TIA | 78 (15.5) | 42 (13.6) | 36 (18.4) | 0.15 |
| Previous cardiac surgery | 153 (30.4) | 88 (28.6) | 65 (33.2) | 0.27 |
| Active endocarditis | 52 (10.3) | 28 (9.1) | 24 (12.2) | 0.26 |
| LVEF * | 0.58 | |||
| <15% | 57 (11.3) | 39 (12.7) | 18 (9.2) | |
| LVEF 16-30% | 100 (19.8) | 62 (20.1) | 38 (19.4) | |
| 31-50% | 120 (23.8) | 69 (22.4) | 51 (26) | |
| ≥51% | 167 (33.1) | 103 (33.4) | 64 (32.7) | |
| LVEF unknown | 36 (7.1) | 20 (6.5) | 16 (8.2) | |
| Previous LVAD | 24 (4.8) | 15 (4.9) | 9 (4.6) | 0.89 |
| Preoperative laboratory parameters | ||||
| Creatinine (mg/dl) | 1.3 (1-1.8) | 1.3 (1-1.7) | 1.3 (1-1.8) | 0.67 |
| Blood urea nitrogen (mg/dl) | 23.6 (16.8-36.5) | 23 (16.9-35.4) | 24.3 (16.3-37.8) | 0.44 |
| Total bilirubin (mg/dl) | 1 (0.5-1.3) | 1 (0.5-1.2) | 0.9 (0.5-1.4) | 0.36 |
| ASAT (U/l) | 29 (22-49) | 30 (21-49) | 29 (22-49) | 0.77 |
| ALAT (U/l) | 27 (18-43) | 27 (18-44) | 25.5 (17-41) | 0.45 |
| Gamma-GT (U/l) | 59 (33-111) | 56 (34-100) | 61 (33-127) | 0.54 |
| Total protein (g/l) | 69 (62-74) | 69 (62-74) | 69 (62-74) | 0.66 |
| Albumin (g/l) | 39 (33-42) | 39 (34-43) | 38 (32-42) | 0.13 |
| Haemoglobin (mg/dl) | 12.1 (10.3-13.8) | 12.2 (10.4-13.8) | 12 (10.2-13.8) | 0.45 |
| Haematocrit (%) | 36.6 (32.1-41) | 37.1 (32.1-41.1) | 36.2 (32.2-40.6) | 0.65 |
| C-reactive protein (mg/dl) | 1 (0.3-3.4) | 1 (0.3-4) | 0.7 (0.3-2.8) | 0.44 |
| Preoperative coagulation parameters | ||||
| aPTT (s) | 40.1 (35.8-47.1) | 40.5 (35.5-47.3) | 39.4 (35.9-46.1) | 0.54 |
| Platelet count (G/L) | 198 (158-259) | 200 (159-260) | 192 (154-255) | 0.25 |
| Thromboplastin time (%)** | 77 (58-94) | 76 (57-94) | 77 (59-94) | 0.69 |
| Thrombin time (s) | 15.5 (14.1-17.7) | 15.6 (14.1-18) | 15.4 (14.2-17.2) | 0.64 |
| Clauss fibrinogen assay (mg/dl) | 397 (327-479) | 399 (328-484) | 394 (323-473) | 0.49 |
| Antithrombin III activity (%) | 93 (81-103) | 92 (80-103) | 94 (84-104) | 0.23 |
| Preoperative medication | ||||
| ASA monotherapy (previous 7 days) | 125 (24.8) | 74 (24) | 51 (26) | 0.61 |
| Clopidogrel monotherapy | 8 (1.6) | 7 (2.3) | 1 (0.5) | 0.16 1 |
| Prasugrel monotherapy | 1 (0.2) | 0 (0) | 1 (0.5) | 0.39 1 |
| Dual antiplatelet therapy (previous 7 days) | 73 (14.5) | 37 (12) | 36 (18.4) | 0.048* |
| ASA+clopidogrel | 53 (10.5) | 28 (9.1) | 25 (12.8) | 0.23 |
| ASA+prasugrel | 8 (1.6) | 2 (0.6) | 6 (3.1) | 0.06 1 |
| ASA+ticagrelor | 11 (2.2) | 6 (1.9) | 5 (2.6) | 0.76 1 |
| ASA+cangrelor | 1 (0.2) | 1 (0.3) | 0 (0) | >0.99 1 |
| Phenprocoumon/warfarin | 82 (16.3) | 49 (15.9) | 33 (16.8) | 0.78 |
| DOAC/NOAC (previous 72 h) | 13 (2.6) | 9 (2.9) | 4 (2) | 0.77 1 |
| Low-molecular-weight heparin (12 hrs)(previous24hrs) | 163 (32.3) | 96 (31.2) | 67 (34.2) | 0.48 |
| Unfractionated heparin (6 h) | 47 (9.3) | 25 (8.1) | 22 (11.2) | 0.24 |
| Systemic thrombolysis (24 h) | 2 (0.4) | 1 (0.3) | 1 (0.5) | >0.99 1 |
LVEF not applicable in patients with LVAD. Patients with LVAD are listed separately.
Laboratory assay for thromboplastin time changed over the study period.
Indicated P-values were calculated using the Fisher exact test instead of the χ2 test because n was ≤ 5 in 1 or more groups based on prespecified criteria.
ALAT: alanine aminotransferase; aPTT: activated partial thromboplastin time; ASA: acetylsalicylic acid; ASAT: aspartate aminotransferase; BMI: body mass index; DOAC: direct oral anticoagulants; ECLS: extracorporeal life support; GT: glutamyl transferase; LVAD: left ventricular assist device; LVEF: left ventricular ejection fraction; NOAC: novel oral anticoagulants; TIA: transient ischaemic attack.
Preoperative dual antiplatelet therapy (DAPT) was more prevalent in patients who developed haemorrhagic complications (P = 0.048) (Table 2).
Perioperative and extracorporeal life support data
Procedural and ECLS run details are shown in Table 3. The duration of the ECLS run was longer in patients with haemorrhagic complications (P < 0.0001).
Table 3:
Procedural details and extracorporeal life support run details for patients on post-cardiotomy extracorporeal life support with and without haemorrhagic complications
| Total study population (n = 504) | No bleeding event (n = 308) | Bleeding event (n = 196) | P-value | |
|---|---|---|---|---|
| Procedure | ||||
| Isolated CABG | 68 (13.5) | 45 (14.6) | 23 (11.7) | 0.44 |
| Valve repair/replacement | 153 (30.4) | 87 (28.2) | 66 (33.7) | |
| Combined CABG/valve surgery | 121 (24) | 78 (25.3) | 43 (21.9) | |
| Aortic aneurysm surgery | 14 (2.8) | 5 (1.6) | 9 (4.6) | |
| Acute type A aortic dissection | 35 (6.9) | 21 (6.8) | 14 (7.1) | |
| HTX | 85 (16.9) | 55 (17.9) | 30 (15.3) | |
| Congenital heart disease | 6 (1.2) | 4 (1.3) | 2 (1) | |
| Other | 22 (4.4) | 13 (4.2) | 9 (4.6) | |
| CPB (min) | 242 (170-322) | 222 (165-316) | 262 (187-327) | 0.092 |
| Aortic cross-clamp (min) | 103 (72-156) | 102 (70-156) | 108 (75-156) | 0.51 |
| ECMO indication | ||||
| CPB weaning failure (initiated during initial surgery) | 375 (74.4) | 227 (73.7) | 148 (75.5) | 0.65 |
| Haemodynamic decline/CPR/respiratory failure after cardiac surgery (implanted within 72 h after end of surgery) | 129 (25.6) | 81 (26.3) | 48 (24.5) | |
| CPR before implant | 112 (21.8) | 63 (20.5) | 49 (25) | 0.23 |
| Implant during ongoing CPR | 30 (6) | 15 (4.8) | 15 (7.7) | 0.2 |
| Duration of support (days) | 4.3 (2.6-6.8) | 3.6 (2.3-5.8) | 5.6 (3.4-8.6) | <0.0001 |
| Cannulation site | ||||
| Peripheral | 461 (91.5) | 287 (93.2) | 174 (88.8) | 0.084 |
| Central | 43 (8.5) | 21 (6.8) | 22 (11.2) | |
CABG: coronary artery bypass graft; CPB: cardiopulmonary bypass; CPR: cardiopulmonary resuscitation; ECMO: extracorporeal membrane oxygenation; HTX: heart transplant.
Laboratory parameters at the time of admission to the intensive care unit are shown in Table S1.
Survival
Overall mortality in patients with major bleeding events during ECLS was higher compared to that of patients without major bleeding complications (P < 0.0001; Fig. 3a). Survival of patients with no haemocompatibility-related adverse events, with bleeding events only, with thrombotic events only and with both is visualized in Fig. 3b.
Figure 3:
(A) Kaplan–Meier estimates of survival of patients on post-cardiotomy extracorporeal life support with and without haemorrhagic events. Overall mortality was significantly higher in patients on extracorporeal life support who had one or more bleeding events compared to patients on extracorporeal life support without bleeding events (shaded areas = 95% confidence interval). (B) Kaplan-Meier estimates of survival of patients with no haemocompatibility-related adverse events (n=258), bleeding events only (n=172), thrombotic events only (n=50) or both (n=24).
Types of thrombotic events
A total of 84 thrombotic events were observed in 74 patients. Ischaemic stroke was the most common type and occurred in 39 cases (cranial computed tomography-confirmed ischaemic stroke with a MRS ≥ 2 at discharge). Twenty-five patients had an MRS ≥ 4 at the time of discharge (Fig. 1).
Circuit and/or cannula thrombosis occurred in 26 patients (5.2%), necessitating circuit/cannula exchange in 12 patients (2.4%) and premature termination of ECLS in 14 patients (2.8%). Pump/oxygenator exchange due to haemolysis was conducted in 8 patients (1.6%).
Peripheral arterial thromboembolism apart from cerebral arterial embolism (e.g. mesenterial, renal or limb artery thromboembolism not related to cannulation) occurred in 11 patients (2.2%) (Figs. 1, 2).
Multivariable analysis of risk factors for haemorrhagic and thrombotic events in patients on extracorporeal life support
A binary logistic regression analysis conducted to identify risk factors for bleeding while on ECLS included the following variables: preoperative DAPT (univariable P = 0.048), ECLS duration (univariable P < 0.0001), CPB duration (univariable P = 0.092) and fibrinogen level at the time of admission to the intensive care unit (univariable P = 0.06).
Preoperative DAPT (adjusted OR: 1.83, 95% CI: 1.063–3.137, P = 0.03) and duration of ECLS (adjusted OR: 1.14, 95% CI: 1.086–1.197, P < 0.0001) were identified as independent risk factors for haemorrhagic complications in patients on PC-ECLS (Table 4).
Table 4:
Multivariable logistic regression analysis of perioperative risk factors for haemorrhagic complications in patients on post-cardiotomy extracorporeal life support
| Adjusted odds ratio | P-value | 95% CI | |
|---|---|---|---|
| Preoperative DAPT | 1.83 | 0.03 | 1.063–3.137 |
| CPB duration (min) | 1.001 | 0.13 | 0.9996–1.003 |
| Fibrinogen (ICU admission) | 0.99 | 0.16 | 0.996–1 |
| ECLS duration (d) | 1.14 | <0.0001 | 1.086–1.197 |
CI: confidence interval; CPB: cardiopulmonary bypass; DAPT: dual antiplatelet therapy; ECLS: extracorporeal life support; ICU, intensive care unit.
The multivariable analysis used to identify risk factors for thrombotic events in patients on ECLS included the following variables: procedure (univariable P = 0.026); previous stroke/transient ischaemic attack (univariable P = 0.07), ECLS duration (univariable P < 0.0001) and aortic cross-clamp duration (univariable P = 0.043) (Tables S2-S5). ECLS duration (adjusted OR: 1.09, 95% CI: 1.04–1.15, P = 0.0002) and previous stroke/transient ischaemic attack (adjusted OR: 1.91, 95% CI: 1.08–3.86, P = 0.049) were identified as risk factors for thrombotic events in patients on PC-ECLS (Table S6).
DISCUSSION
This large, single-centre analysis including exclusively patients on ECLS post-cardiotomy shows that haemorrhagic complications during PC-ECLS are frequent, affecting 38.9% of patients. A total of 14.7% of patients had thrombotic events, and nearly half of the patients (48.8%) were affected by at least 1 HRAE. The study confirmed the distinct negative impact of bleeding on survival, and we observed a cumulative adverse effect of bleeding and thrombotic events on outcome, with the most deaths in patients who experienced both types of HRAE.
Uncontrollable bleeding has been suggested by recently published guidelines as the sole absolute contraindication for PC-ECLS [15]; these guidelines also stress the prognostic importance of haemorrhage in these patients. Ongoing haemorrhage with constant volume loss and the subsequent need for volume and blood product substitution leads to unstable ELCS flow and suboptimal support conditions. The actual advantage of ECLS in post-cardiotomy shock—the ability of ECLS to provide full cardiac output and adequate end organ perfusion—is severely impaired by haemorrhagic complications, which promote the development of secondary complications such as end organ failure (kidney, liver) and potential transfusion-related complications. In contrast to veno-venous extracorporeal membrane oxygenation, where thrombotic events have been reported as the predominant type of HRAE [16], bleeding prevails in ECLS. Post-cardiotomy patients are at particular risk to develop haemorrhagic complications, and high rates of haemorrhagic complications in these patients have been described in prior studies [1, 4, 8].
The aetiology of post-cardiotomy shock likely has an impact on the occurrence of bleeding events, and we hypothesize that patients with right ventricular failure are likely to be more prone to bleeding events because of systemic and hepatic congestion leading to elevated central venous pressure and coagulopathy. On the other hand, one advantage of ECLS is the relief of systemic congestion by effective drainage of the venous system. In this study, it was not possible to clearly differentiate right ventricular failure from other aetiologies of post-cardiotomy shock because of the retrospective design. This question remains to be answered by future prospective studies. The low rate of surgical sources of bleeding described in the operative reports (distinct surgical bleeding source identified in 56 of 131 patients who underwent revision for bleeding) suggests a high incidence of coagulopathy-related surgical site haemorrhage in this cohort.
This study identified the ECLS run duration as a risk factor for both haemorrhagic and thrombotic complications. In line with this finding, ECLS duration has been identified as a risk factor for HRAE in a large Extracorporeal Life Support Organization Registry analysis (mixed cohort including 17.7% post-cardiotomy patients) [4]. The duration of ECLS as a risk factor for haemorrhage has also been a ubiquitous finding in prior studies with non-PC [17] and PC patients [8].
Additionally, ECLS duration is consistently reported as a risk factor for death in PC-ECLS [18, 19], emphasizing once again the importance of a comprehensive weaning evaluation, of keeping the duration ECLS to the necessary minimum, and of avoiding delaying ECLS explantation for non-medical (i.e. organizational) reasons.
The second risk factor for haemorrhage in patients on ECLS in our study was preoperative dual antiplatelet therapy. Extracorporeal circulation is well known for affecting platelet count as well as platelet aggregation and activation [20–22] and for shear stress-induced acquired von-Willebrand syndrome [23], which may result in both impaired haemostasis and increased risk of thrombosis [22, 23].
Platelet deficiency and dysfunction in patients on ECLS has been linked to bleeding complications as well as death [22]; thus it is not surprising that the additional effect of irreversible platelet inhibition by P2Y12 inhibitors leads to a further increase of bleeding complications [24]. We suggest that patients who receive DAPT preoperatively be closely monitored for bleeding and that the anticoagulation regimen for patients on ECLS should be adapted to this situation.
The ambitious question about how best to manage anticoagulation in patients on PC-ECLS has still not been answered by the present study, and it is not possible to define a strategy suitable for every case. A previous study showed no increase in thromboembolic events with a lower target activated clotting time range in patients on ECLS [25], but there is still insufficient evidence on whether and when it may be safe in certain situations to lower anticoagulation targets. It appears to be reasonable and necessary to adapt and lower the anticoagulation target range in high-risk situations for haemorrhage (e.g. early postoperative phase) while the patient is still on full ECLS flow, because in this situation, the risk of bleeding may outweigh the risk for thrombotic events, which may change later during the ECLS run [26].
Raffa et al. described a decrease in the incidence of bleeding events in patients on PC-ECLS after implementation of changes in anticoagulation/patient management including full intraoperative heparin reversal and delayed postoperative initiation of anticoagulation by not starting heparin before 24 h post-surgery and after careful evaluation of the bleeding tendency [1]. Our centre has followed a similar approach, and we tend to delay heparinization even more in order to avoid haemorrhagic complications in patients on PC-ECLS. Von Stumm et al. reported no increase in thromboembolic events and a potential reduction of bleeding events with delayed heparinization in neonatal patients on PC-ECLS [27].
Steady re-evaluation is then necessary during the ECLS run because anticoagulation should reach the defined therapeutic range before entering the weaning phase. By that time, the patient should be in a stabilized condition, bleeding should be controlled and the target activated partial thromboplastin time should be reached before reduction of ECLS flow to avoid thrombus formation in the circuit and subsequent thromboembolic complications.
In this study, the most common sites of bleeding were the surgical site and the cannulation site, stressing the importance of diligent surgical haemostasis, a meticulous cannulation technique and the proactive avoidance of haemorrhagic complications.
Although direct left ventricular unloading strategies were only infrequently applied in this study (∼1%), the utilization of direct left ventricular unloading is likely going to increase in the future, and this change is expected to have an impact on the occurrence of haemorrhagic events. An increase of bleeding events has been demonstrated with the concomitant use of ECLS and the Impella Support System (Abiomed, Danvers, MA, USA) (in non-post-cardiotomy patients) [28].
A prior analysis by Mariscalco et al. has shown significantly increased rates of bleeding complications with central cannulation [29]. Central cannulation has also been associated with adverse survival and higher complication rates [30].
We observed a non-significant trend towards a higher rate of central cannulation among patients with bleeding complications. However, at our centre, peripheral access for ECLS (axillary arterial or femoral arterial with antegrade perfusion cannula) is the preferred approach, and the rate of centrally cannulated patients in the present study was low (8.5%). For this reason, a comparison of bleeding and thrombotic events between central and peripheral cannulation in our cohort may not be representative, and the analysis might also be underpowered in this regard because of the low number of centrally cannulated patients and biased by the centre-specific preference of peripheral cannulation.
LIMITATIONS
Several limitations apply to the present study. The study period covers more than 21 years, during which there have been changes in the management of patients on ECLS as well as in the management of all patients undergoing cardiac surgery over this long time period. The single-centre design, although enabling the provision of good data granularity, limits generalizability of the results. Comparability of patient groups is not granted in a retrospective analysis, and although great care was taken to provide a detailed description of baseline and procedural characteristics, patient characteristics may vary in parameters that are not captured by the present analysis, and bleeding may correlate with the occurrence of other events not captured by this study.
A clear differentiation of heart failure aetiology between left-, right- and biventricular failure was not possible in the retrospective cohort but likely plays a role in the occurrence of bleeding events. ECLS flow has an impact on haemorrhagic and thrombotic events, but detailed information on ECLS flow was not included in the retrospective registry.
CONCLUSION
Bleeding complications in patients on ECLS are common and significantly impair survival. Nearly half of the patients were affected by a haemocompatibility-related event. Thus, reducing bleeding and thrombotic complications is key to improving the outcome of patients on post-cardiotomy ECLS. This issue may be approached by the triad of accurate surgical haemostasis, optimizing anticoagulation management and limiting the duration on ECLS.
Supplementary Material
Glossary
ABBREVIATIONS
- ECLS
extracorporeal life support
- PC
post-cardiotomy
- OR
odds ratio
- CI
confidence interval
- CPB
cardiopulmonary bypass
- MRS
modified Rankin scale
- aPTT
activated partial thromboplastin time
- HRAE
haemocompatibility-related adverse events
- DAPT
dual antiplatelet therapy
Contributor Information
Anne-Kristin Schaefer, Department of Cardiac Surgery, Medical University of Vienna, Vienna, Austria.
Michaela Latus, Department of Cardiac Surgery, Medical University of Vienna, Vienna, Austria.
Julia Riebandt, Department of Cardiac Surgery, Medical University of Vienna, Vienna, Austria.
Georg Goliasch, Department of Internal Medicine II, Medical University of Vienna, Vienna, Austria.
Martin H Bernardi, Division of Cardiac Thoracic Vascular Anaesthesia and Intensive Care Medicine, Medical University of Vienna, Vienna, Austria.
Günther Laufer, Department of Cardiac Surgery, Medical University of Vienna, Vienna, Austria.
Daniel Zimpfer, Department of Cardiac Surgery, Medical University of Vienna, Vienna, Austria.
Dominik Wiedemann, Department of Cardiac Surgery, Medical University of Vienna, Vienna, Austria.
SUPPLEMENTARY MATERIAL
Supplementary material is available at EJCTS online.
Funding
This study received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Conflict of interest: Dominik Wiedemann is a proctor for Abbott and a scientific advisor for Xenios/Fresenius. Daniel Zimpfer is a proctor for Abbott, Medtronic and Berlin Heart. He also has research grants from Abbott, Medtronic, Edwards and Berlin Heart. He has received speaker fees from Abbott, Medtronic, Berlin Heart and Edwards. He has received travel support from Abbott, Medtronic, Berlin Heart and Edwards. The remaining authors have no conflicts of interest to report.
DATA AVAILABILITY
The data are available on request to the corresponding author.
Author contributions
Anne-Kristin Schaefer: Writing—original draft; Conceptualization; Michaela Latus: Data curation, Writing—review & editing; Julia Riebandt: Writing—review & editing; Georg Goliasch: Writing—review & editing; Martin H. Bernardi: Writing—review & editing; Günther Laufer: Supervision; Resources; Daniel Zimpfer: Writing—review and editing; Dominik Wiedemann: Conceptualization; Supervision; Resources; Writing—review and editing.
Presented at the 2022 Annual Meeting of the American Association for Thoracic Surgery, Session: Controversies and Challenges in ECMO Management.
REFERENCES
- 1. Raffa GM, Gelsomino S, Sluijpers N, Meani P, Alenizy K, Natour E. et al. In-hospital outcome of post-cardiotomy extracorporeal life support in adult patients: the 2007-2017 Maastricht experience. Crit Care Resusc 2017;19(Suppl 1):53–61. [PubMed] [Google Scholar]
- 2. Meani P, Matteucci M, Jiritano F, Fina D, Panzeri F, Raffa GM. et al. Long-term survival and major outcomes in post-cardiotomy extracorporeal membrane oxygenation for adult patients in cardiogenic shock. Ann Cardiothorac Surg 2019;8:116–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Lorusso R, Raffa GM, Alenizy K, Sluijpers N, Makhoul M, Brodie D. et al. Structured review of post-cardiotomy extracorporeal membrane oxygenation: part 1-Adult patients. J Heart Lung Transplant 2019;38:1125–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Chung M, Cabezas FR, Nunez JI, Kennedy KF, Rick K, Rycus P. et al. Hemocompatibility-Related Adverse Events and Survival on Venoarterial Extracorporeal Life Support: an ELSO Registry Analysis. JACC Heart Fail 2020;8:892–902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Willers A, Swol J, Buscher H, McQuilten Z, van Kuijk SMJ, Ten Cate H. et al. Longitudinal Trends in Bleeding Complications on Extracorporeal Life Support Over the Past Two Decades-Extracorporeal Life Support Organization Registry Analysis. Crit Care Med 2022;50:e569–e80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Nguyen TP, Phan XT, Nguyen TH, Huynh DQ, Tran LT, Pham HM. et al. Major Bleeding in Adults Undergoing Peripheral Extracorporeal Membrane Oxygenation (ECMO): prognosis and Predictors. Crit Care Res Pract 2022;2022:5348835. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Takahashi K, Kubo S, Ikuta A, Osakada K, Takamatsu M, Taguchi Y. et al. Incidence, predictors, and clinical outcomes of mechanical circulatory support-related complications in patients with cardiogenic shock. J Cardiol 2022;79:163–9. [DOI] [PubMed] [Google Scholar]
- 8. Melehy A, Ning Y, Kurlansky P, Kaku Y, Fried J, Hastie J. et al. Bleeding and Thrombotic Events During Extracorporeal Membrane Oxygenation for Postcardiotomy Shock. Ann Thorac Surg 2022;113:131–7. [DOI] [PubMed] [Google Scholar]
- 9. Toivonen F, Biancari F, Dalen M, Dell'Aquila AM, Jonsson K, Fiore A. et al. Neurologic Injury in Patients Treated With Extracorporeal Membrane Oxygenation for Postcardiotomy Cardiogenic Shock. J Cardiothorac Vasc Anesth 2021;35:2669–80. [DOI] [PubMed] [Google Scholar]
- 10. Ellouze O, Abbad X, Constandache T, Missaoui A, Berthoud V, Daily T. et al. Risk Factors of Bleeding in Patients Undergoing Venoarterial Extracorporeal Membrane Oxygenation. Ann Thorac Surg 2021;111:623–8. [DOI] [PubMed] [Google Scholar]
- 11. Schaefer AK, Distelmaier K, Riebandt J, Goliasch G, Bernardi MH, Zimpfer D. et al. Access site complications of postcardiotomy extracorporeal life support. J Thorac Cardiovasc Surg 2021; [DOI] [PubMed] [Google Scholar]
- 12. Lansky AJ, Messe SR, Brickman AM, Dwyer M, Bart van der Worp H, Lazar RM. et al. Proposed Standardized Neurological Endpoints for Cardiovascular Clinical Trials: an Academic Research Consortium Initiative. Eur Heart J 2018;39:1687–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Buchtele N, Staudinger T, Schafer AK, Bogl MS, Schoergenhofer C, Schwameis M.. Anticoagulation in Critically Ill Adults during Extracorporeal Circulation. Hamostaseologie 2021;41:294–306. [DOI] [PubMed] [Google Scholar]
- 14. Schaefer AK, Donhauser B, Kroeckel I, Fureder L, Holaubek C, Braunschmid T. et al. Heparin-induced thrombocytopaenia diagnostic testing after adult cardiac surgery: a single-centre experience and development. Eur J Cardiothorac Surg 2019;55:722–8. [DOI] [PubMed] [Google Scholar]
- 15. Lorusso R, Whitman G, Milojevic M, Raffa G, McMullan DM, Boeken U. et al. 2020 EACTS/ELSO/STS/AATS expert consensus on post-cardiotomy extracorporeal life support in adult patients. J Thorac Cardiovasc Surg 2021;161:1287–331. [DOI] [PubMed] [Google Scholar]
- 16. Nunez JI, Gosling AF, O'Gara B, Kennedy KF, Rycus P, Abrams D. et al. Bleeding and thrombotic events in adults supported with venovenous extracorporeal membrane oxygenation: an ELSO registry analysis. Intensive Care Med 2022;48:213–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Hu W, Zhang J, Wang M, Chen W, Chai L, Leung EL. et al. Clinical Features and Risk Factors Analysis for Hemorrhage in Adults on ECMO. Front Med (Lausanne) 2021;8:731106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Schaefer AK, Riebandt J, Bernardi MH, Distelmaier K, Goliasch G, Zimpfer D. et al. Fate of patients weaned from post-cardiotomy extracorporeal life support. Eur J Cardiothorac Surg 2022;61:1178–85. [DOI] [PubMed] [Google Scholar]
- 19. Distelmaier K, Wiedemann D, Binder C, Haberl T, Zimpfer D, Heinz G. et al. Duration of extracorporeal membrane oxygenation support and survival in cardiovascular surgery patients. J Thorac Cardiovasc Surg 2018;155:2471–6. [DOI] [PubMed] [Google Scholar]
- 20. Chen Z, Mondal NK, Zheng S, Koenig SC, Slaughter MS, Griffith BP. et al. High shear induces platelet dysfunction leading to enhanced thrombotic propensity and diminished hemostatic capacity. Platelets 2019;30:112–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Jiritano F, Serraino GF, Ten Cate H, Fina D, Matteucci M, Mastroroberto P. et al. Platelets and extra-corporeal membrane oxygenation in adult patients: a systematic review and meta-analysis. Intensive Care Med 2020;46:1154–69. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Siegel PM, Chalupsky J, Olivier CB, Bojti I, Pooth JS, Trummer G. et al. Early platelet dysfunction in patients receiving extracorporeal membrane oxygenation is associated with mortality. J Thromb Thrombolysis 2022;53:712–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Heilmann C, Geisen U, Beyersdorf F, Nakamura L, Benk C, Trummer G. et al. Acquired von Willebrand syndrome in patients with extracorporeal life support (ECLS). Intensive Care Med 2012;38:62–8. [DOI] [PubMed] [Google Scholar]
- 24. Gelbenegger G, Jilma B.. Clinical pharmacology of antiplatelet drugs. Expert Rev Clin Pharmacol 2022;15:1177–97. [DOI] [PubMed] [Google Scholar]
- 25. Hong JI, Hwang J, Shin HJ.. Satisfactory outcome with low activated clotting time in extracorporeal membrane oxygenation. Rev Cardiovasc Med 2021;22:1589–94. [DOI] [PubMed] [Google Scholar]
- 26. Patel B, Diaz-Gomez JL, Ghanta RK, Bracey AW, Chatterjee S.. Management of Extracorporeal Membrane Oxygenation for Postcardiotomy Cardiogenic Shock. Anesthesiology 2021;135:497–507. [DOI] [PubMed] [Google Scholar]
- 27. von Stumm M, Subbotina I, Biermann D, Gottschalk U, Mueller G, Kozlik-Feldmann R. et al. Impact of delayed systemic heparinization on postoperative bleeding and thromboembolism during post-cardiotomy extracorporeal membrane oxygenation in neonates. Perfusion 2020;35:626–32. [DOI] [PubMed] [Google Scholar]
- 28. Schrage B, Becher PM, Bernhardt A, Bezerra H, Blankenberg S, Brunner S. et al. Left Ventricular Unloading Is Associated With Lower Mortality in Patients With Cardiogenic Shock Treated With Venoarterial Extracorporeal Membrane Oxygenation: results From an International. Multicenter Cohort Study. Circulation 2020;142:2095–106. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Mariscalco G, Salsano A, Fiore A, Dalen M, Ruggieri VG, Saeed D. et al. Peripheral versus central extracorporeal membrane oxygenation for postcardiotomy shock: multicenter registry, systematic review, and meta-analysis. J Thorac Cardiovasc Surg 2020;160:1207–16 e44. [DOI] [PubMed] [Google Scholar]
- 30. Kowalewski M, Zielinski K, Brodie D, MacLaren G, Whitman G, Raffa GM. et al. Venoarterial Extracorporeal Membrane Oxygenation for Postcardiotomy Shock-Analysis of the Extracorporeal Life Support Organization Registry. Crit Care Med 2021;49:1107–17. [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.
Supplementary Materials
Data Availability Statement
The data are available on request to the corresponding author.