A prospective study examining cachexia predictors in patients with incurable cancer

Abstract

Background: A predictive model of cachexia development would help identify those at greatest risk for early therapeutic intervention. The aims of this study were to identify predictors of cachexia development and to use these to create and evaluate the accuracy of a predictive model. Methods: A secondary analysis of an international, prospective, observational multicentre study was conducted. Patients with incurable cancer and without cachexia, who attended a palliative care unit were amenable to the analysis. Cachexia was defined as weight loss (WL)>5% (6 months) or WL>2% and body mass index<20kg/m2. Disease related, demographic and patient reported markers were evaluated as possible predictors with Cox analysis. The predictor “cancer type” was divided into high risk (pancreatic or gastric cancer), low risk (sarcoma, breast or hematologic cancer) or neutral risk (all others). A classification and regression tree analysis was used to create a model based on optimal combinations and cut-offs of significant predictors for cachexia development, and accuracy was evaluated with receiver operating characteristic curve analysis. Results: 628 patients, 57% female, were included in the analysis. Median (IQR) age was 65 (17) years and Karnofsky performance status 70 (10). Median (IQR) follow-up was 109 (108) days, and 159 (25%) patients developed cachexia. Initial WL, cancer type, appetite and chronic obstructive pulmonary disease were significant predictors (p≤0.04). A four-level model was created with each level carrying an increasing risk of cachexia development. Level 1-patients (WL<3%, low risk cancer type and no or little appetite loss) had a relative hazard ratio (RHR) of 0.27 of cachexia development, while level 4-patients (WL 3-5%) had a RHR of 4.9. Accuracy of cachexia predictions at 3 months was 76%, and a risk level ≥3 (1-3% WL and neutral risk cancer type OR<3% weight loss and high risk cancer type) yielded a sensitivity of 46% and a specificity of 90% of cachexia development. Conclusions: Important predictors of cachexia have been identified and used to construct a predictive model of cancer cachexia. Patients with a risk level ≥3 have a high risk of cachexia development, and should be targeted for therapeutic intervention.