Weight loss over time and survival: a landmark analysis of 1000+ prospectively treated and monitored lung cancer patients.

BACKGROUND: Eligibility criteria and endpoints for cancer cachexia trials-and whether weight loss should be included-remain controversial. Although most cachexia trials enrol patients after initial cancer diagnosis, few studies have addressed whether weight loss well after a cancer diagnosis is prognostic. METHODS: We pooled data from non-small cell lung cancer patients from prospectively conducted trials within the Alliance for Clinical Trials in Oncology (1998-2008), a nationally funded infrastructure. We examined (i) weight data availability and weight changes and (ii) survival. RESULTS: A total of 822 patients were examined. Of these, 659 (80%) were on treatment at the beginning of Cycle 2 of chemotherapy; weight was available for 656 (80%). By Cycles 3 and 4, weight was available for 448 (55%) and 384 (47%), respectively. From baseline to immediately prior to Cycle 2, 208 (32%) gained weight; 225 (34%) lost <2% of baseline weight; and 223 (34% of 656) lost 2% or more. Median survival from the beginning of Cycle 2 was 13.0, 10.9, and 6.9 months for patients with weight gain, weight loss of <2%, and weight loss of 2% or more, respectively. In multivariate analyses, adjusted for age, sex, performance score, type of treatment, and body mass index, weight loss of 2% or more was associated with poor overall survival compared with weight gain [hazard ratio (HR) = 1.66; 95% confidence interval (CI): 1.33-2.07; P < 0.001] and compared with weight loss of <2% (HR = 1.57; 95% CI: 1.27-1.95; P < 0.001). Although weight loss of <2% was not associated with poorer overall survival compared with weight gain, it was associated with poorer progression-free survival (HR = 1.24; 95% CI: 1.01-1.51; P = 0.036). Similar findings were observed in a separate 255-patient validation cohort. CONCLUSIONS: Weight should be integrated into cancer cachexia trials because of its ease of frequent measurement and sustained prognostic association.

Duke Scholars

Author Xiaofei Wang Biostatistics & Bioinformatics, Division of Biostatistics