Predictors of early death risk in older patients treated with first-line chemotherapy for cancer.

PURPOSE: Objective factors for making choices about the treatment of elderly patients with cancer are lacking. This investigation aimed to help physicians select appropriate treatments through the identification of factors that predict early death (< 6 months) after initiation of chemotherapy treatment. PATIENTS AND METHODS: Previously untreated patients greater than 70 years of age who were scheduled for first-line chemotherapy for various types of cancer were included. Baseline abbreviated comprehensive geriatric assessment (aCGA), including the Mini-Mental State Exam, Timed Get Up and Go (GUG), Activities of Daily Living (ADL), Instrumental Activities in Daily Living (IADL), Mini Nutritional Assessment (MNA), Geriatric Depression Scale (GDS15), and comorbidities index (Cumulative Index Rating Scale-Geriatric), was carried out. Prognostic factors of early death were sought from aCGA results and traditional oncology measures. RESULTS: A total of 348 patients were included across 12 centers in Southwest France (median age, 77.45 years; ratio of men to women, 1.47; advanced disease, 65%). Abnormal aCGA scores were observed for 18.1% of patients on the ADL, 73.0% of patients on the IADL, 24.1% of patients on the GUG, 19.0% of patients on the MMS, 44.0% of patients on the GDS15, and 64.9% of patients on the MNA. Advanced disease (odds ratio [OR], 3.9; 95% CI, [1.58 to 9.73]), a low MNA score (OR 2.77; 95% CI, [1.24 to 6.18]), male sex (OR, 2.40; 95% CI, [1.2 to 4.82]), and long GUG (OR, 2.55; 95% CI, [1.32 to 4.94] were associated with higher risk of early death. CONCLUSION: In patients greater than 70 years of age with cancer, advanced disease, a low MNA score, and poor mobility predicted early death. We recommend that the MNA and GUG, performed by a trained nurse, be maintained as part of routine pretreatment workup in these patients to identify at-risk patients and to inform the decision-making process for chemotherapy.