INTRODUCTION: Parenteral nutrition (PN) is widely used in palliative care (PC), but there is limited evidence to support its use at the end of life (EOL). This aim of this was to investigate the relationship between routine laboratory parameters and survival in patients receiving PN, and to develop a decision tree model to support clinicians decide whether to start or forgo PN.
METHODS: The laboratory parameters of 113 patients with advanced diseases who were admitted to a specialized palliative care unit (PCU) were analyzed at two points in time: T0 = before PN, T1 = two weeks after initiation of PN. Univariate Mann-Whitney U-tests and multivariate linear regression models, as well as a decision tree analysis were computed; all in relation to survival time.
RESULTS: The final regression model was significant with p = 0.001 (adjusted R2 = 0.15) and included two predictors for survival time after PN initiation: the CRP/albumin ratio and urea at T1 (ps = 0.019). Decision tree analysis revealed three important predictors for classification of survival time after PN initiation: CRP, urea, and LDH (all at T0).
DISCUSSION: The decision tree model may help to identify patients likely to benefit from PN, thus supporting the clinical decision whether or not to start PN.