Prognostication based on the change in the palliative prognostic index for patients with terminal cancer.

CONTEXT: The use of the Palliative Prognostic Index (PPI) in relation to the clinical time course has not yet been established. OBJECTIVES: To investigate the association between the changes in the PPI over time and the survival of terminal cancer patients in a palliative care unit (PCU). METHODS: This retrospective cohort study analyzed data from 374 terminal cancer patients who were admitted to the PCU of a university hospital in Japan. Clinical data, such as age, gender, body mass index, vital signs, initial PPI, and subsequent PPI, were collected from the medical records. The PPI change per day (ΔPPI) was calculated using the initial PPI at admission and the one after five to seven days. The factors associated with death within three weeks were identified using Cox proportional hazards model analysis. RESULTS: After their admission to the PCU, 147 (39.3%) patients were deceased within three weeks. The multivariate-adjusted analysis showed that body temperature (hazard ratio [HR] 0.7; 95% CI 0.5, 1.0), initial PPI (HR 1.3; 95% CI 1.2, 1.4), and ΔPPI (HR 6.6; 95% CI 4.9, 9.0) were significantly and independently associated with death within three weeks. In the subanalysis, the ΔPPI was significantly associated with death within three weeks in the group with initial PPI ≤ 4 (HR 9.3; 95% CI 5.8, 15.0), 4 < initial PPI ≤ 6 (HR 14.4; 95% CI 5.7, 36.2), and initial PPI > 6 (HR 9.0; 95% CI 4.1, 20.0). CONCLUSION: Our data suggest that the ΔPPI may be useful for predicting the survival of terminally ill cancer patients.