Predictors of Apnea and Prediction of Time to Death in Donation After Cardiac Death

Donation after cardiac death (DCD) is a method by which patients have the opportunity to donate organs following elective withdrawal of life-sustaining therapies (WLST). Often these patients have suffered a catastrophic neurological injury although not progressed to brain death or have a severe medical condition for which ongoing medical care is considered non-beneficial. DCD has become an increasingly significant source of organs for transplantation in a time of growing wait lists and organ shortfall.

In order to donate organs through DCD, the potential donor must progress to death within a certain time window after withdrawal of life sustaining therapies. This timeframe varied but is usually less than 120 minutes following WLST. Up to 40% of potentially eligible DCD donors in Ontario do not proceed to organ procurement for these reasons (internal data from Trillium Gift of Life).

The uncertainty and variability in the potential for successful organ procurement has an impact on families, health care teams and organ retrieval teams. It is important that this emotional and resource burden only occurs in candidates with a reasonable likelihood of being eligible to donate organs. The amount of time elapsing between WLST and circulatory arrest (and organ procurement) has important implications for the quality of the procured organs, and some organs are unable to be transplanted after death. Finally, maintaining organ procurement teams and an operating room on standby consumes valuable hospital resources and removes these human and physical resources from other clinical duties.

Several different prediction tools have been proposed to predict time to death following WLST in potential DCD donors, but none has been proven useful. To derive a more generalizable prediction tool it is necessary to identify valid predictors that are common to a wide variety of patients undergoing WLST. We propose the addition of features that predict apnea or respiratory insufficiency in the development of a new predictive model.

While previous studies propose important variables for prediction of time to death, we hypothesize that features that are focused on apnea (neurologic and non-neurologic) will be independently predictive of time of death following WLST.

Neurological predictors of apnea: The previous studies have consistently identified one or two neurologic risk factors for apnea associated with time to death. These risk factors have been evaluated in isolation and have never been rigorously studied in a broader population or in combination. We wish to evaluate the following neurological risk factors for apnea which have been previously found to be associated with time to death in certain studies: Glasgow Coma Scale, absence of brain stem reflexes (corneal, pupil, cough, gag), and controlled mode of mechanical ventilation.

Non-Neurological predictors of apnea: No studies to date have evaluated features focused on non-neurologic causes of apnea. Given that neurologic predictors of apnea have been the one consistent feature that has been identified across most studies, we hypothesize that evaluating additional predictors of apnea beyond neurologic causes could similarly have an association with time to death as they both result in profound hypoxia. While some researchers have evaluated the contribution of BMI and weight, we hypothesize that the addition of more definitive features of upper airway obstruction will strengthen the performance of our prediction tool. Neck circumference, absence of endotracheal tube cuff leak, fluid balance, and history of obstructive sleep apnea (OSA) are novel and could improve the operating characteristics of prediction tools.

Neck circumference has been found to be associated with airway obstruction and has been incorporated into numerous prediction tools for OSA. In a study evaluating prevalence and predictors of upper airway obstruction following stroke, neck circumference was independently associated with any upper airway obstruction that occurred 24 hours following acute stroke. In a pooled analysis, the absence of cuff leak has been shown to be predictive of post extubation stridor with 92% specificity, thus suggesting impending loss of airway patency. Fluid balance has also been found to be associated with post extubation failure.

We will conduct a prospective, observational cohort study of all patients being considered for DCD in whom consent for donation is obtained to evaluate the association between neurologic and non-neurologic risk factors for apnea, other clinically important variables and time to death after WLST, and use these data to derive a generalizable predictive model for the prediction of the time to death following WLST in potential DCD donors.