Temporal distribution of deaths in cancer patients during the day in different settings

All living organisms perform their functions normally according to circadian rhythms. Certain diseases, such as ischemic heart disease and asthma, produce symptoms that are distributed during the day in a nonrandom fashion. Chronomodulated therapy with some regimens of chemotherapy and other drugs produce better results than traditional schedules. Even death is not evenly distributed during the day. Significant differences in the time of death through the day could influence the work planning and care activities.

To determine whether timing of death from a population of cancer patients admitted at our Oncology Institute varied during the day and according to different settings: at home (H), at the palliative care unit (PCU), and at other services (OS) of the hospital. Comparing the timing of deaths from different settings can give some clues about the possible existence of a circadian rhythm and the influence of external circumstances in the time of death of cancer patients.

We conducted a retrospective study of the records of time of death at the different settings.

The study involved 772 patients from the PCU and 997 from OS who died between May 25, 1996, and May 24, 2000, and 347 patients who died at H between April 1, 1999, and December 31, 2001. A statistically significant difference was found in the distribution of time of death in patients at the PCU (p <.001), but not at OS or at H. There were two peaks between 08:00 and 10:00 and between 00:00 and 02:00, and one trough between 04:00 and 08:00. This suggests that a temporal variation occurs in the time of death of cancer patients dying in the PCU, but not in other settings.

The clinical relevance of the results obtained in this study would depend on the amplitude of the eventual variation detected in the number of deaths during the day. Therefore, although there was a statistically significant variation at the time of death during the day, its amplitude is not high enough to make it clinically significant. The differences observed among the various settings suggest that, even if circadian rhythms exist at the time of death in cancer patients, external factors can overcome these rhythms.