Share this article on:

Study Identifies 8 Signs of Impending Death in Cancer Patients

Samson, Kurt

doi: 10.1097/01.COT.0000464347.26690.00


A prospective review of end-stage cancer patients has identified eight specific physical signs that closely correlate with death within three days.

The researchers say the changes—predominantly neurocognitive in nature—should help clinicians better determine when blood tests and other efforts to prolong life can be halted, when patients should be discharged home or to hospice, and when loved ones should be notified that death will occur soon.

The study, published in the March 15 issue of Cancer (2015;121:960-967), is the first to closely monitor such patients and identify the most telling signs of death, the researchers note, explaining that while other studies have looked at the subject, none has used such strict daily monitoring protocols in a large group of dying patients.

The lead author, David Hui, MD, Assistant Professor in the Department of Palliative Care and Rehabilitation Medicine at MD Anderson Cancer Center, said that although the findings might also apply to the last days of patients with other terminal medical conditions, more research is needed to determine that, as are studies in larger groups of cancer patients.

Back to Top | Article Outline

Study Specifics

The team, along with colleagues at Barretos Cancer Hospital in Brazil, studied the frequency and onset of 52 bedside physical signs and their diagnostic performance for impending death in 357 consecutive cancer patients admitted for palliative care. The changes found were systematically documented every 12 hours from admission until patient death or discharge. The frequency and median onset time for each sign were analyzed from death backwards and “likelihood ratios” (LRs) of dying within three days were calculated.

“We found that simple bedside observations can potentially help us recognize if a patient has entered the final days of life,” Hui said. “If these findings are confirmed, they will help doctors, nurses, and families better recognize the dying process, while providing more compassionate care for such patients.”

A total of 203 patients, or 57 percent, died during the study, and eight physical signs were found to be highly predictable of impending death. The predictive utility of changes ranged from 35 to 78 percent in patients within three days of death. Each change had a late onset and taken together, there was a greater than 95 percent specificity. Chief among them were:

  • Nonreactive pupils (positive LR, 16.7);
  • Decreased response to verbal stimuli (positive LR, 8.3);
  • Decreased response to visual stimuli (positive LR, 6.7);
  • Inability to close the eyelids (positive LR, 13.6);
  • Drooping of the nasolabial fold or relaxation of the facial muscles (positive LR, 8.3);
  • Hyperextension of the neck (positive LR, 7.3);
  • Grunting (positive LR, 11.8); and
  • Bleeding in the upper gastrointestinal tract (positive LR, 10.3).

“Many of these signs are well recognized in dying patients, but ours is the first to categorize them in terms of a likelihood of impending death using strict monitoring. In an earlier study we identified five physical signs that were highly predictive of impending death, but this new study systematically examined a more comprehensive list of signs representing multiple organ systems,” Hui said.

“Although these changes indicated a high likelihood of death within three days, their absence cannot rule out that a patient will die within that time frame.”

He noted that many other signs—including delirium, dysphagia, and incontinence—common at the end of life were found not to be diagnostic of impending death, probably because they had an earlier onset. In addition, while many others, such as epistaxis, myoclonus, and thrush, were not associated with imminent death.

Back to Top | Article Outline

Possible Limitations

There were a number of limitations with the study, as well, he noted: “We included only cancer patients admitted to acute palliative care units, who often have more severe symptoms, and the frequency, onset, and utility of these signs may differ in other health care settings and among patient populations—something that should be further examined.”



Mortality rates at the two hospitals also differed significantly, most likely due to differences in health care systems, he said. In Brazil, patients can stay in the hospital longer than in the United States. Nonetheless, physicians at both units used similar palliative care practices and met with each other over the course of the study.

“Despite different mortality rates, the predictive rates were similar, which we believe strengthens our results and their generalizability,” Hui said.

However, patient signs were documented only every 12 hours, which limited the researchers' ability to identify the median onset, he added. Moreover, the researchers examined a large number of clinical signs in a relatively small number of patients, another limitation.

“Our findings should be considered preliminary until validated in future studies,” he said. “I consider this approach as personalized and prognostic-based decision making. Our doctors tend to rely on their own judgment with such patients, but there has been very little clinical research data. We found, for example, that the prognostic value of nonreactive pupils was very high in terms of dying within three days—with a ratio of 17. Anything over 10 is considered excellent in prognostic terms. This and the other signs tell us that the body is shutting down, and doctors can try making the patient more comfortable rather than trying to prolong their life.”

Back to Top | Article Outline

Symptoms Often Mistaken

Asked for his perspective, Mellar Davis, MD, Professor of Medicine at the Cleveland Clinic Lerner College of Medicine and Chairman of the Palliative Section of the Multinational Association of Supportive Care in Cancer, said: “This well-done study puts everything together that has so far been piecemeal in the literature. Two groups will benefit from the findings—physicians and families.”

He noted that impending death is often not recognized by attending physicians, especially for patients who stop eating or drinking and have difficulty breathing. Nurses, though, who spend more time with such patients, are generally more aware.

“For families, they often ask when the end is near, so these findings will help us in educating them on what to expect so they are not taken by surprise and can be present for their loved one. They also want to know what it is like at the end.”

Timing is also crucial with such patients because some might die while being transported to hospice. “I use these signs all the time, and even though there are some outliers, I do not think that any other paper has put it all together so well. Additional research is probably necessary.”



Davis noted that physicians can mistake these symptoms as being due to other factors, especially opioid or other drug toxicity, so many clinicians may miss these signs of imminent death. Families too can mistake such symptoms as opioid toxicity, and when opioid levels are reduced, patients can unnecessarily experience pain in their final days and hours. Patients with dementia often stop eating or drinking, so these other changes can be very helpful, he said.

These clinical signs are by no means restricted to cancer patients, he added: “In my experience, patients dying from heart failure, cirrhosis, emphysema, and many other end-stage conditions also show these changes before dying.”

Back to Top | Article Outline

Why So Little Previous Research?

So why has there been so little research on common signs of imminent death? “For one thing this is not a pleasant subject,” Davis said. “Researchers in general do not get grants to study this, so funding must come from other sources.”

In this case, MD Anderson funded the study, although the National Institutes of Health also supported the research with a small grant.

If this “checklist” of changes is to become widely distributed, funding will likely come from palliative care organizations and hospice groups, he speculated. “I doubt if there will be a larger study. This one is robust enough that it need not be repeated. In my personal experience, these last changes are the same across all diseases.”

Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Home  Clinical Resource Center
Current Issue       Search OT
Archives Get OT Enews
Blogs Email us!