Is quality of life predictive of the survival of patients with advanced nonsmall cell lung carcinoma?

BACKGROUND. Several studies have recently reported on the importance of qua lity of life (QOL) in predicting the survival of patients with lung carcino ma. To confirm these reports, the relationship between survival and QOL, as measured by the European Organization for Research and Treatment of Cancer (EORTC) Quality of Life Questionnaire and Duke-UNC Social Support Scale, w as examined within a group of 206 patients with advanced nonsmall cell lung carcinoma treated in a randomized clinical trial conducted by the Cancer a nd Leukemia Group B (CALGB 8931). METHODS. Patients completed the EORTC instrument and the Duke-UNC Social Su pport Scale at baseline in the clinic. The Cox proportional hazards model w as used to determine the incremental contribution QOL provided in predictin g survival beyond the effect of known clinical prognostic variables. RESULTS. Clinical factors that were jointly predictive of poorer survival i ncluded law performance status, nonadenocarcinoma histology, presence of dy spnea, weight loss greater than 5%, albumin level less than 3.5 mg/dL, and adrenal metastases. Univariate analyses showed that patient-reported EORTC subscales describing increased pain, appetite loss, fatigue, lung carcinoma symptoms, poorer overall QOL, and poorer physical functioning predicted si gnificantly poorer survival. Multivariate analyses showed that, after adjus tment for clinical factors, overall QOL was not a significant predictor of survival. Rather, the only EORTC subscale of prognostic importance was the pain subscale, in which a 40-point increase was associated with a 27% incre ase in the hazard rate. CONCLUSIONS. This study did not confirm the prognostic importance of overal l QOL. Rather, after adjustment for significant clinical factors, a patient -provided pain report had the greatest prognostic importance. (C) 1999 Amer ican Cancer Society.