Prediction model for time until death

Organ donation after circulatory death (DCD donation) is only possible if the potential donor dies within 2 hours after the withdrawal of medical treatment. After that, the quality of the organs decreases too much. How can you better estimate when someone will die?

Why was this research necessary?

Angela Kotsopoulos: 'Usually, DCD donors are patients who are in the intensive care unit (ICU) after, for example, a brain hemorrhage, stroke, or accident. If recovery is ruled out, the attending physician, in consultation with the family, stops life-sustaining treatment. The patient then dies naturally after some time.'

When consent for organ donation is given, the deceased is taken to the operating room for organ retrieval after the doctor has declared death. This is only possible if there is no more than 2 hours between the withdrawal of life-sustaining treatment and death. It would be nice to be able to estimate whether someone will die within those 2 hours after the withdrawal of life-sustaining treatment.'

'This applies not only to the patient themselves, but also to the family, to the potential recipients, and to the deployment of staff and resources. Because starting a donation procedure requires so much from everyone involved, you would ideally like to estimate in advance whether it will actually happen.'

Better prediction with a model

Together with Radboudumc and the NTS, Kotsopoulos therefore developed a prediction model. She collected data from over 400 patients for this purpose. 'The model looks at a number of patient characteristics. The worse the patient scores on these, the greater the chance that he or she will die within one or two hours after the withdrawal of life-sustaining treatment.'

Model not yet suitable for clinical practice

The model predicts quite well which patients will die within 1 or 2 hours. Nevertheless, Kotsopoulos advises against basing decisions in clinical practice on this.

'There will always be patients who do die within two hours, even though the model predicted otherwise. If you do not start a donation procedure for them, you lose a potential organ donor. You really want to prevent that.'

'We want to fulfill the wish of the patient and their family to donate. The family is often disappointed if the donation does not go ahead. And of course, with the major shortage of donor organs, we do not want to lose any donors. In case of doubt, I would therefore always start the donation procedure.'

'The model can contribute to communication with the patient's family. 'For them, it is nice to have an estimate.'

Other application: treating longer

Kotsopoulos' co-supervisor, donation intensivist Farid Abdo, sees another application. 'In some cases, the model predicts that the potential donor will likely not die within 2 hours. In these patients, in consultation with the family, you could wait a little longer before withdrawing life-sustaining treatment.'

'The patient does not suffer from this, because comfort care is paramount for all these patients. But in the meantime, the body progresses a bit further in the dying process. That increases the chance that the patient will still die within 2 hours, and thus the wish for organ donation can still be fulfilled.'

Advice for donation intensivists

What is the best decision in practice now? Abdo: ‘If you estimate with great certainty that someone will not die within 2 hours of withdrawing life-sustaining treatment, then discuss with the family the possibility of waiting a little longer until the dying process is in a further stage.’

Kotsopoulos: ‘And if you are in doubt whether someone will die within 2 hours, always start the donation procedure.’