Life after extracorporeal membrane oxygenation : long-term survival and quality of life

Abstract

Background: The use of Extracorporeal membrane oxygenation (ECMO) has steadily increased in the last decades and has evolved in several ways. From originally being a means to support the neonatal patient with respiratory or circulatory failure for days to weeks, ECMO is now mainly used in paediatric and adult patients. Survival after ECMO varies depending on the underlying condition, and survival to discharge and 6-12 months has been readily reported in the literature. Likewise, quality-of-life and health status are well-investigated in the 3-24-month period after discharge. However, there is a paucity of data concerning long-term outcomes several years after ECMO treatment.

Aim: To identify long-term survival and causes of death in ECMO treated patients (study I and II), and to investigate the long-term health and mental status after treatment, including cognitive functions and brain radiographic findings (study III), pulmonary function, pulmonary morphology, mood disorders and quality of life (study IV).

Overview of methods: Using the Swedish national causes of death registry, study I and II attained survival status and causes of death in all commonly treated patient groups at the ECMO Centre of the Karolinska University Hospital. Survival was depicted using the Kaplan-Meier technique. For study III and IV, a retrospective cohort was created by contacting consecutive long-term adult survivors, starting with the first adult survivor treated at the centre. Thirty-eight patients treated with ECMO for respiratory failure were investigated. This included magnetic resonance imaging of the brain and extensive neurocognitive tests (study III), followed by computed tomography of the lungs, spirometry, a six-minute walk test and self-reported forms of quality of life and mood symptoms (study IV, including Short form 36, St George’s respiratory questionnaire, Hospital anxiety and depression scale and Trauma screening questionnaire).

Summary of research results: Survival status in 255 adults was investigated in median 4.4 years after treatment (study I). The mortality was high in the first three months after treatment (17% of the ECMO survivors died in the first 90 days). This time point served as a cut-off to define late survival. In patients who were alive at 90 days, 87% were alive five years later. Long-term survival differed between groups and was highest in patients treated for a known or suspected infectious disease. In study II, 400 children were investigated in median 7.2 years after treatment. Similar to the results in adults, there was high 90-day mortality, and 93% of neonates and 89% of paediatric patients were alive 10 years later in the group who survived to this time point. Patients who died generally had severe comorbidities or an underlying disease which caused deterioration later in life. Brain lesions were seen in 37% of the long-term survivors (14/38, study III). In the group treated with venoarterial ECMO, 64% had signs of brain lesions. General intelligence depicted as the full-scale intelligence quotient (normal mean 100, SD 15) was 97 in median (IQR 86-104). In patients with brain lesions, the median full-scale intelligence quotient was 88, compared to 102 in patients with normal brain imaging (p=0.28). Memory functions and executive functions, also reported as indices with a normal mean of 100 and a SD of 15, were significantly reduced in patients with brain lesions (p=0.03 and 0.02, respectively). Patients with hypoxaemia during ECMO treatment, defined as <93% pulse oximetry haemoglobin saturation in median during ECMO treatment (or the first 10 days if treated for a long time) had similar intelligence as patients with normoxaemia. Quality of life was reduced in the present cohort, but the results were similar to previously published data on patients with acute respiratory distress syndrome not treated with ECMO. A reduction in diffusion capacity was seen in 47% of the patients, and lung function varied greatly between patients. Lung parenchymal damage was common, in mean 7% of the parenchyma was damaged (range 0-44%). In 50% of the patients, this damage was predominantly localised anteriorly, possibly indicating ventilator-induced lung injury. Parenchymal damage correlated with time on ECMO and time with mechanical ventilation, and with reductions in quality of life and diffusion capacity.