Weight loss prior to cancer treatment carry a negative impact on clinical outcomes. However, few studies have addressed whether weight assessment over time is plagued by high dropout rates and whether weight change carries a prognostic association in the same manner as it does at baseline.
A pooled analysis of individual patient data was undertaken among non-small cell lung cancer patients who participated in prospective cancer treatment trials from the Alliance for Clinical Trials in Oncology from 1998-2008. This study examined 1) rates of missing weight data over time and 2) the prognostic association of weight beyond baseline assessment.
822 chemotherapy-treated patients were examined. 659 (80%) were still on treatment at the beginning of cycle 2. Weight was available for 656 (80%) patients. However, by cycles 3 and 4, weight was available for only 448 (55%) and 384 (47%) patients, respectively. From baseline to immediately prior to cycle 2, 224 patients (34% of 656) lost more than 2% baseline weight, and 226 (34%) lost between 0 and 2%. With respect to prognostic associations, the median survival time from the beginning of cycle 2 was 6.9, 10.9, and 13.0 months for patients with weight loss of 2% or more, loss of < 2%, and those with weight gain, respectively. In multivariate analyses, after adjustment for age, gender, performance score, type of treatment, and body mass index, weight loss of 2% or more was associated with poor survival compared to weight loss of < 2% (hazard ratio (HR) = 1.57; 95% CI [1.27 to 1.95]; P <.001). Although weight gain was not associated with improved overall survival, it was associated with better progression-free survival outcomes (HR = 0.81; 95% CI [0.66 to 0.99]; P=.04).
Weight is a clinically useful endpoint and should be integrated into cancer cachexia trials because of its ease of frequent measurement and sustained prognostic association.
Clinical trial identification
Legal entity responsible for the study
Fred C and Katherine B Andersen Foundation and the United States National Cancer Institute.
All authors have declared no conflicts of interest.