Ongoing assessment of nutritional status in children with malignant disease

The nutritional status of a child on cancer therapy influences both tolerance of and response to treatment. However, it is difficult to assess nutritional status on a daily basis because an accurate quantitation of the calorie intake is difficult. Anthropometric and biochemical parameters are prone to error and often reflect past rather than current nutritional status. In practice, a subjective clinical assessment is usually relied upon. This study objectively appraises the value of such an assessment. Based on clinical symptoms that alter oral intake and absorption of food, a scoring system was designed to assess nutritional status on a day to day basis. A symptom score (SS) of 10 implied "normality"; 0 indicated maximum debility. Over a 2-year period 511 daily scores were recorded in 30 patients aged 0.7-17.5 years. Patients were studied at presentation and during treatment for acute lymphoblastic leukemia (ALL, n = 14; solid tumors receiving megatherapy with autologous bone marrow rescue (ABMR, n = 8), and chemotherapy for different tumors (miscellaneous, n = 8). The SS was compared with other nutritional parameters, including sequential anthropometric indices, serum albumin, insulin-like growth factor-I (IGF-I), IGF binding protein-3 (IGFBP-3), and whole-body protein turnover (WBPT) using [1-(13) C]leucine. The mean SS was reduced at diagnosis for all leukemic patients (median score = 8), improved during first remission (p < 0.002), fell to a minimum during febrile neutropenia (p = 0.0009), and improved with clinical and hematological recovery (p = 0.0009). A reduction in SS was related to fever (p < 0.001) and a fall in neutrophil count (p < 0.001). There was no correlation with anthropometric indices or IGF-I and IGFBP-3 levels. Paired WBPT studies in 9 patients showed that SS correlated well with protein breakdown (p = 0.026). The results suggest that the ongoing nutritional status of children with malignancy undergoing chemotherapy is best assessed using simple clinical parameters.