Dying from Dementia

Abstract & Commentary

By Mary Elina Ferris, MD, Clinical Assistant Professor, University of Southern California. Dr. Ferris reports no financial relationship to this field of study.

Synopsis: Advanced dementia mortality rates rise after episodes of fever, pneumonia, and eating problems. Treatment decisions often lead to burdensome interventions and distressing symptoms that might be avoided if caregivers were better informed about the expected complications and prognosis of this condition.

Source: Mitchell SL, et.al. The clinical course of advanced dementia. N Engl J Med 2009;361:1529-1538.

Nursing home residents in 22 facilities were tracked prospectively over 18 months to better describe their clinical prognosis, and to determine how health care decisions made by their proxies influenced medical outcomes. Of 1763 nursing home residents, 572 met the advanced dementia eligibility criteria and 323 were recruited for the study (mean age, 85 years; 85% female, and 90% white). Median nursing home stay was 3 years, and median interval since dementia diagnosis was 6 years. Dementia diagnosis was established using scores on standard data sets and clinical description. Eligibility required the highest score on the Global Deterioration Scale, which reflects profound cognitive deficits such as inability to recognize family members, minimal verbal communication, total functional dependence, incontinence of urine and stool, and inability to ambulate independently. The study also required the availability of an appointed health care proxy who could communicate in English; the proxies were interviewed quarterly.

Within 18 months, 55% had died, with a median survival of 478 days. The overall probability of death within 6 months was 25%; this increased to nearly 50% if an episode of fever, pneumonia, or eating problems had occurred within the prior 6 months. Of those who died, 90% had documented eating problems (weight loss, swallowing or chewing problems, refusal to eat or drink) in the last 3 months of life. Sentinel events such as strokes, hip fracture, seizures, and gastrointestinal bleeding occurred in 10% but rarely were the cause of death. Distressing symptoms such as dyspnea, pain, pressure ulcers, and agitation increased as the end of life approached, ranging from 39% to 54%.

Health care proxies overwhelmingly supported comfort as the primary goal, but only 26% of the proxies for those who died expected death within 6 months. For the residents who died, 12% had been hospitalized, 30% received intravenous therapy, and 7% had tube feeding in the last 3 months of their life. The most common reason for hospitalization was pneumonia and other infections. If their health care proxies believed that they had less than 6 months to live, those residents were less likely to undergo one of these interventions.


Understanding advanced dementia as a terminal illness requiring palliative care will help improve the suboptimal treatment that many currently receive. Burdensome interventions lead to little comfort for the demented elderly in their last days of life, yet medical progress has allowed us to extend those lives, often beyond any rational reason. Patients with dementia who receive hospice care, in the model used for terminal cancer, have fewer hospitalizations, better pain control, and their families report greater satisfaction than when hospice is not used.1

If families and health care proxies understand the limited prognosis, they are more likely to avoid interventions, as this study shows. It describes the "clinical trajectory of end-stage dementia" that can help both families and nursing home staff understand its terminal nature, and adapt care to provide comfort rather than cure. Although it does not provide information about the duration of dementia after an initial diagnosis, it does help us establish when the end of life is near. After the occurrence of pneumonia, fever, or eating problems, survival is often 6 months or less; this is an opportunity to initiate discussions about hospice care or other palliative care programs.


1. Sachs GA. Dying from dementia. N Engl J Med 2009; 361:1595-1596.