Significant weight loss, cachexia, and being bedbound signal that a cancer patient is dying. However, identifying the specific signs that give physicians the ability to predict death is not well described in the literature. To better understand why predicting death is an important part of the care continuum, The ASCO Post spoke with palliative care specialist and medical oncologist, David Hui, MD, MSc, Assistant Professor in the Departments of Palliative Care/Rehabilitation Medicine and General Oncology at The University of Texas MD Anderson Cancer Center, Houston. Dr. Hui and his colleagues recently completed a study in this underreported area of palliative care.
Study Details
Please describe the trial design.
It was a prospective longitudinal observational cohort study designed to identify the physical signs of impending death. It involved the MD Anderson Cancer Center in Houston and the Barretos Cancer Hospital in Brazil. We basically followed the cancer patients from the day of admission to an acute palliative care unit.
We documented clinical signs of interest twice a day, every day until they were discharged from the unit or they died. We were able to tell how often certain systems occurred prior to death. After aggregating and analyzing the data, we identified a number of physical signs that allow us to predict death within 3 days.
Difficult Clinical Situation
Why is it important to be able to identify the clinical signs of impending death?
For a number of reasons. Clinical decisions at the end of life, such as discharging patients and ending and beginning medication schedules, are dependent on a patient’s prognosis. In patients who are actively dying, we need to focus all our efforts on comfort measures that prevent suffering. The key question becomes: how can we as oncologists tell when our patients are actively dying?
This is not a subject without controversy. In many countries, there is an intense debate about the medical propriety in the use of integrated care pathways for patients who are imminently dying, such as the United Kingdom’s Liverpool Care Pathway for the Dying Patient.
Part of this controversy stems from the fact that, to date, we still do not have a lot of clinical confidence to determine when a patient has entered this irreversible phase and will die within a few days. That lack of knowledge creates unease about frank discussions about care in this very difficult clinical setting. Plus, if we cannot predict with certainty that the patient is imminently dying, then it opens the door for critics of the pathway to say that option was taken away from patients still eligible for life-sustaining therapy.
Family Considerations
How does having a better way to predict death affect the patient’s family?
Families of terminally ill cancer patients often want to know when to expect death so they can appreciate the process and understand what their loved one is going through. This information can also help their efforts as caregivers, and they can plan their lives around the dying process. For instance, if a father is admitted to the hospital and his son lives in another state, it’s important to be able to know if the patient is in the final days of life so his son can make appropriate travel plans for a peaceful closing and to say goodbye.
Overestimating Survival
In your work, you indicate that physicians tend to overestimate survival time in their terminally ill cancer patients. Why is this?
A body of literature from our group at MD Anderson, as well as from other institutions, shows that doctors and nurses routinely overestimate survival, and this is largely due to our lack of accurate prognostic tools. Moreover, we often lean on the side of overcaution when breaking bad news to patients and their families; there are so many implications to a diagnosis of imminent death that doctors often avoid the conversation altogether. Caution is understandable, but honesty based on accurate information is needed in this difficult setting.
Role of Hospice
Is this an area of inquiry that might help hospice workers?
We looked at patients in MD Anderson’s acute palliative care unit, because they are being carefully monitored, in the inpatient setting. We suspect what we learned might be applicable to other settings such as hospice; however, this would need to be further tested. Our next step is to conduct a larger validation study to further confirm our findings. We could then develop a diagnostic tool to help clinicians and caregivers identify the specific signs of impending death.
Key Symptoms
What were the symptoms that best predicted impending death?
There were 10 clinical signs leading to the death of patients we followed in the palliative care unit. Built on our findings, we were able to divide the signs into two groups based on the frequency of occurrence, their onset relative to death, and their predictive value for impending death.
The early signs are decreased level of consciousness, decreased performance status, and dysphagia. Most patients who die have these signs, which occur about a week before death. These signs have only moderate predictive value in telling us that the patient has 3 days left.
The later signs that we documented were pulselessness of the radial artery, apnea, Cheyne-Stokes breathing, death rattle, peripheral cyanosis, respiration with mandibular movement, and decreased urine output. These signs occur due to bodily changes seen with very decreased level of consciousness—perhaps activity in the brainstem or decreased cardiovascular profusion. These signs usually occur in the last 3 days of life. They are not always present in patients who die within 3 days, but when they are present, they are strongly predictive of impending death.
So with simple bedside observations, a clinician can predict what is going to happen to the patient by detecting the signs, such as respiration with mandibular movement, which essentially means the jaw drops during breathing. This is a telltale sign that the patient will die in the next few days.
Having tools that help predict impending death will make the stressful process easier for the patient, the doctor, and the family and caregivers who need to prepare for the end. As mentioned, our next step is more research from which we can design a tool to help clinicians predict when death is imminent. ■
Disclosure: Dr. Hui reported no potential conflicts of interest.
