Prediction of Chemotoxicity, Unplanned Hospitalizations and Early Death in Older Patients with Colorectal Cancer Treated with Chemotherapy

Abstract

Simple Summary Chemotoxicity, unplanned hospitalizations (Uhs) and early death (ED) are common among older patients with cancer who receive chemotherapy. Our objective was to determine factors predicting these complications. A predictive score for these three complications based on geriatric, tumor and laboratory variables was developed in a series of 215 older patients with colorectal carcinoma receiving chemotherapy. The use of this score may reliably identify patients at risk to have excessive toxicity with chemotherapy, UH or ED, thus helping to plan treatment, implement adaptive measures, and intensify follow-up. Purpose: To identify risk factors for toxicity, unplanned hospitalization (UH) and early death (ED) in older patients with colorectal carcinoma (CRC) initiating chemotherapy. Methods: 215 patients over 70 years were prospectively included. Geriatric assessment was performed before treatment, and tumor and treatment variables were collected. The association between these factors and grade 3-5 toxicity, UH and ED (<6 months) was examined by using multivariable logistic regression. Score points were assigned to each risk factor. Results: During the first 6 months of treatment, 33% of patients developed grade 3-5 toxicity, 31% had UH and 23% died. Risk factors were, for toxicity, instrumental activities of daily living, creatinine clearance, weight loss and MAX2 index; for UH, Charlson Comorbidity Score, creatinine clearance, weight loss, serum albumin, and metastatic disease; and for ED, basic activities in daily living, weight loss, metastatic disease, and hemoglobin levels. Predictive scores were built with these variables. The areas under receiver operation characteristic (ROC) curves for toxicity, UH and ED were 0.70 (95% CI: 0.64-0.766), 0.726 (95% IC: 0.661-0.799) and 0.74 (95% IC: 0.678-0.809), respectively. Conclusion: Simple scores based on geriatric, tumor and laboratory characteristics predict severe toxicity, UH and ED, and may help in treatment planning.