Chemotherapy near the end of life

BMJ 2014; 348 doi: http://dx.doi.org/10.1136/bmj.g1529 (Published 04 March 2014) Cite this as: BMJ 2014;348:g1529
  1. Michael W Rabow, professor of clinical medicine
  1. 1Department of Medicine, University of California, 1545 Divisadero Street, San Francisco, CA 94143-0320, USA
  1. mrabow{at}medicine.ucsf.edu

A difficult decision with potentially unexpected implications

Even as cancer treatments become more effective, we can still wonder about the symbolic meaning behind decisions to pursue chemotherapy near the end of life. Although most patients with metastatic cancer choose to receive palliative chemotherapy, evidence suggests that most do not clearly understand its intent.1 In decision making about chemotherapy, doctors are supposed to describe, and patients are supposed to understand, the direct outcomes of the proposed treatment (for example, clinical response rates and side effects). However, the broader implications of such decisions can be just as important. In the linked paper by Wright and colleagues (doi:10.1136/bmj.g1219),2 choosing palliative chemotherapy was associated with a whole set of outcomes that may not have been known, expected, or discussed by patients, their family caregivers, and their oncologists.

Wright and colleagues studied the outcomes of a cohort of 386 cancer patients who died during the Coping with Cancer Study, a federally funded cohort study of terminally ill cancer patients and their informal caregivers.3 Wright and colleagues’ participants were patients at eight US outpatient oncology clinics who had advanced cancer refractory to one or more chemotherapy regimens and were identified by their oncologists as terminally ill at study enrollment. They were categorized as receiving (56%) or not receiving chemotherapy at enrollment, and the outcomes were compared using propensity scoring to balance the groups for important confounding factors. Study participants died a median of 4.0 months after enrollment.

Wright and colleagues found that patients who, months before death, had decided to receive palliative chemotherapy were significantly more likely than patients not receiving chemotherapy to undergo mechanical ventilation or cardiopulmonary resuscitation in the last week of life. They were more likely to be referred to a hospice late (one week or less before death). All these outcomes have been associated with poorer quality of life for patients, worse distress for caregivers, and increased costs. Patients who received chemotherapy were also less likely than others to die in their preferred location, more likely to die in an intensive care unit, and more likely to be tube fed in the last week of life. Notably, receiving chemotherapy was not associated with longer survival.

Late chemotherapy has already been associated with decreased use of hospice care.4 Wright and colleagues are the first to report an association between palliative chemotherapy and place of death. Previous research has focused on chemotherapy received days or weeks before death, but Wright and colleagues studied decisions made months before patients’ deaths. Whereas only 6.2% of cancer patients use chemotherapy two weeks before death,5 and 20-50% use it within 30 days of death,6 7 fully 62% use it within two months before death.8 Given the frequency of patients choosing to have chemotherapy months rather than days before death, Wright and colleagues argue reasonably that this earlier timeframe is actually a much more important time during which to understand decisions about treatment and their implications.

Most cancer patients are likely to choose chemotherapy for the promise of improved survival or quality of life. Most patients in the new study said they would be willing to accept chemotherapy if it gave them just one extra week of life. A key implication of this research is the need to better identify patients who are likely to benefit from chemotherapy near the end of life. Many of Wright and colleagues’ participants may have chosen chemotherapy for the promise of even a single extra week of life, but whether they received even that is not clear.

An important and perhaps more achievable improvement that should be pursued as a result of this research is to encourage oncologists to discuss with patients the broader implications of palliative chemotherapy when making decisions about treatment. For all patients, it is time to be clear about the full extent of possible harms.

Clearly, the choice to start chemotherapy should not be the end of decision making. Patients nearing the end of life should be able to refine their decisions as they get more information and understand their ongoing likelihood of benefiting from various treatments. We learnt long ago that patients do not like the either/or requirement of hospice care in the United States (where patients often have to forgo cancer treatments to receive the benefits of a hospice). Nor should we accept that a course of chemotherapy effectively commits patients to a host of outcomes that they do not want. Especially perhaps near the end of life, the choice to accept chemotherapy should not lead oncologists and patients to abandon key care assessments and planning, including regular evaluation of the efficacy of chemotherapy, advance care planning, exploration of the benefits of hospice care, and pursuit of one’s preferred place of death.

Future work should explore how a decision to start chemotherapy, even many months before death, comes to be associated with how and where someone dies. Perhaps chemotherapy and the associated outcomes in Wright and colleagues’ study are primarily the consequence of the zeal not to “give up.” Perhaps the associations are due, at least in part, to inadequate involvement of expert palliative care. Patients may well want, and we might help them to achieve, care near the end of life that is individualized, nuanced, and consistent with their wishes. Working with oncology teams, early palliative care can help patients in making such decisions and in pursuing treatments, or avoiding them, to live as long as possible, with the best possible quality of life.9


Cite this as: BMJ 2014;348:g1529


  • Research, doi:10.1136/bmj.g1219
  • Competing interests: I have read and understood the BMJ Group policy on declaration of interests and declare the following interests: none.

  • Provenance and peer review: Commissioned; not externally peer reviewed.