Extracorporeal membrane oxygenation (ECMO) for acute heart failure in adults

5 Safety

This section describes safety outcomes from the published literature that the Committee considered as part of the evidence about this procedure. For more detailed information on the evidence, see the interventional procedure overview.

5.1 Death during extracorporeal mechanical oxygenation (ECMO) support was reported in 40% (52/131) of patients in a case series of 131 patients with cardiogenic shock; in 'most cases' this was because of multi-organ failure or sepsis.

5.2 Death within 30 days was reported in 76% (167/219) of patients in the case series of 219 patients with refractory postcardiotomy cardiogenic shock; the main cause of death was low cardiac output syndrome secondary to refractory myocardial failure.

5.3 Intracranial haemorrhage was reported in 2% of patients in the register reporting on 2312 cardiac failure patients (absolute numbers not reported; timing unclear).

5.4 Major or significant bleeding was reported to have a pooled incidence estimate of 41% (95% confidence interval [CI] 14.8% to 63.6%) in a systematic review of case series of 260 patients (5 studies).

5.5 Rethoracotomy for bleeding or tamponade was reported to have a pooled incidence estimate of 42% (95% CI 16.1% to 83.7%) in a systematic review of case series of 828 postcardiotomy patients (6 studies).

5.6 Stroke was reported in 9% (7/81) of patients in the case series of 81 patients (timing unclear). A pooled incidence estimate of 6% (95% CI 4.2% to 8.3%) was reported for stroke in a systematic review of case series of 630 patients (3 studies).

5.7 Femoral artery perforation (leading to uncontrollable bleeding and subsequent death) was reported in 2 patients in the case series of 131 patients. Inferior vena cava tear was reported in 2% (2/92) of patients in a systematic review of case series of 92 patients (2 studies).

5.8 Lower extremity amputation was reported to have a pooled incidence estimate of 5% (95% CI 2.3% to 9.3%) in a systematic review of case series of 192 patients (5 studies).

5.9 Fasciotomy or compartment syndrome was reported to have a pooled estimate of 10% (95% CI 7.35% to 14.5%) in a systematic review of case series of 335 patients (5 studies).

5.10 Lower extremity ischaemia was reported to have a pooled estimate of 17% (95% CI 12.5% to 22.6%) in a systematic review of case series of 677 patients (13 studies). Lower limb ischaemia was reported in 13% (28/219) of patients in the case series of 219 patients; fasciotomy for severe leg ischaemia was needed in 6% (13/219) of patients (timing unclear).

5.11 Venous thrombus in either the femoral vein or the inferior vena cava was reported in 8% (18/217) of patients (4 studies) and arterial thrombus in 7% (13/192) of patients (3 studies) in a systematic review of case series.

5.12 Significant post-ECMO infection had a pooled incidence estimate of 30% (95% CI 13.7% to 64.5%) in a systematic review of case series of 992 patients (10 studies). Significant infection was defined as sepsis or suspected sepsis needing antibiotics, which occurred in 8 of the 10 studies. Ventilator-associated pneumonia (1 or more episode) was reported in 49% (40/81) of patients, surgical wound infections in 17% (14/81) of patients, bacteraemia in 14% (11/81) of patients, and catheter-related infections in 6% (5/81) of patients in the case series of 81 patients (timing unclear).

5.13 Mechanical complications, including oxygenator failure (15%; 36% of these patients survived), cannula problems (4%; 27% survived), tubing rupture (less than 1%; none survived) and pump malfunction (less than 1%; 28% survived), were reported in cardiac failure patients (n=2312) included in the register of patients treated by ECMO (absolute numbers not reported). Clots in the ECMO circuit were reported in 19% (56/295) of patients and air embolus in 2% (5/295) of patients in the case series of 295 patients.

5.14 Brachial plexus injury was reported in 2 case reports of patients treated with ECMO. One of these patients died 55 days later from multiple systemic complications. Symptoms improved at 4‑month follow-up in the other patient.

5.15 The specialist advisers said that anoxic neurological complications were additional adverse events reported in the literature. They listed the following anecdotal adverse events: intrathoracic bleeding, left ventricular thrombus formation and acute pulmonary injury.