Improving outcomes of extracorporeal membrane oxygenation: predicting risk, understanding complications

Extracorporeal membrane oxygenation (ECMO) is a support modality that temporarily takes over the function of the heart and/or lungs in patients who are critically ill. Blood is drained outside the body through an artificial lung, given oxygen, and returned to the patient. ECMO does not cure the underlying illness, but it can buy time: to recover, or to bridge to a next step, such as a mechanical heart pump or an organ transplant.
Although the use of ECMO has greatly increased in recent decades, a large proportion of patients does not survive. This thesis looks at two related themes: the complications that occur during and after ECMO support, and the way doctors try to estimate how likely a patient is to recover.
The first part shows that complications such as bleeding and a sudden deterioration after weaning from ECMO occur regularly and are linked to a higher mortality. Reassuringly, patients who survive ECMO are not more often anxious or depressed than other patients who have been critically ill.
The second part focuses on predicting chances of survival. Existing prediction models often turn out to be unreliable and do not take changes in a patient's condition during treatment into account. Therefore, a model was developed that continuously updates its prediction using the most recent data, similar to the way doctors themselves asses the course of illness.
Together, these insights contribute to a better understanding of ECMO-related complications and course of disease, and form a step toward better supporting doctors, patients, and families in making difficult decisions in ECMO care.