When someone sustains a serious brain injury and is unresponsive, how soon can doctors say whether he or she has a chance of meaningful recovery? That has always been a difficult question to answer, and it’s being made even tougher by new guidelines from the American Academy of Neurology.
As an intensive care unit pediatrician, I often work with families whose child has had a potentially catastrophic or devastating brain injury. Although the causes of these injuries vary — trauma, stroke, bleeding into the brain, and more — they create a common constellation of problems: the individuals are unconscious or only minimally arousable, and they typically need a ventilator to breathe for them.
For families, the overarching question is whether their loved one will recover consciousness at all and, if so, to what degree will he or she be able to function and interact with others and their environment. To answer this question, standard ICU practice is to support the patient long enough to allow the brain to stabilize after injury, and perform the diagnostic studies and imaging necessary to recommend to the family whether life support should be withdrawn, allowing the patient to die, or whether to go forward with surgically placing a tracheostomy for connection to a ventilator and a feeding tube, in order to facilitate transfer to a rehabilitation hospital for continued treatment in the hope of improvement.
The current practice in most ICUs is to help families make a decision about whether to withdraw life support within the first three to five days of the injury. A recent review stated that 72 hours of observation after the injury is “extensively supported in the literature” as sufficient for physicians to be confident in predicting a poor outcome.
There are several reasons for wanting to make these decisions as soon as possible. First and foremost is the well-being of the family. If there is confidence that the patient will never regain a quality of life that the patient would find acceptable, then early decisions spare the patient and family the suffering of a prolonged and ultimately futile hospital stay.
Second is the desire to avoid what may be the worst outcome: prolonged survival without meaningful recovery. Some patients who never recover consciousness regain the ability to breathe without a ventilator over the first few weeks of treatment. Disconnecting the ventilator before such recovery (somewhat cynically referred to as a “window of opportunity”) always leads to death in a short time. If this window has closed, families often feel compelled to continue treatment, usually with transfer to a long-term care facility. While it is possible to limit future treatments by withdrawing nutrition and fluids, these decisions are often fraught and put the family through even more emotional suffering.
Third is the fact that ICU beds are a limited medical resource: most ICUs run at or near capacity. Early decision-making for patients who are not going to survive with an outcome that is satisfactory to the patient and family makes it possible to treat more of those who can truly be helped by ICU care.
Yet this approach — trying to make a prognosis in three to five days — goes against the recent guidelines from the American Academy of Neurology. They argue that current ICU practice is based on flawed and outdated information, and categorically state that when “discussing prognosis with caregivers of patients with a disorder of consciousness during the first 28 days post-injury, clinicians must avoid statements that suggest these patients have a universally poor prognosis.”
Paradoxically, the more we learn about the prognosis of acute severe brain injury, the less we seem to know. The approach that physicians like me have used now appears to be in direct conflict with the American Academy of Neurology guidelines.
Changing practice in the ICU to conform with the guidelines presents huge challenges. ICUs do not have the capacity to treat all patients with severe brain injury for a month or so. In addition, most ICU physicians do not believe they are incapable of accurately predicting poor outcomes in the most severe cases. And even if the guidelines were to become the rule, there aren’t enough high-quality rehabilitation beds to support the demand.
Given that we lack the tools to more accurately predict which patients will do well, some patients who would have had an acceptable recovery will die following early withdrawal of life support, while others will undergo prolonged, yet ultimately futile, attempts at rehabilitation.
In my career, it is very likely that I’ve erred in withdrawing life support too soon. I’ve also had families who chose to continue life support and rehabilitation against my recommendation, only to face the disappointment of never seeing their child improve. I’ve witnessed firsthand how devoting years to supporting and caring for a loved one who will never wake up takes an enormous toll on a family’s emotional health and financial well-being, not to mention on marriages and career opportunities. Yet decisions need to be made, and the cost of being wrong is a very high price to pay, regardless of whether one errs on the side of too little treatment or too much.
What’s needed now are better ways of predicting which patients in the ICU with severe brain injuries have the greatest potential for meaningful improvement. A recent report in the New England Journal of Medicine represents an important step forward. It showed that the information obtained from electroencephalography (EEG), a technology universally available in hospitals, could help identify a subset of patients with a higher likelihood of good prognoses.
Until we have more information like this, however, families will have no choice but to make life and death decisions for their loved ones in the face of conflicting recommendations from medical experts.
Robert Truog, M.D., is a pediatric intensive care physician at Boston Children’s Hospital; professor of medical ethics, anesthesiology, and pediatrics at Harvard Medical School; and director of the school’s Center for Bioethics.