Nutrition meets nephrology in head and neck cancer: the predictive impact of body composition on the onset of acute kidney disease

Background – Patients with head and neck cancer undergoing high-dose cisplatin chemotherapy combined with radiotherapy are particularly at risk of developing nephrotoxicity. While Acute Kidney Injury risk is well established, the role of subclinical renal impairment and its potential interplay with nutritional status remains poorly defined. We aim to evaluate how body composition pre-treatment can predict the onset of Acute Kidney Disease. Materials and methods – We conducted a prospective, monocentric, observational study involving 110 patients with locally advanced HNC treated with concurrent cisplatin (≥200 mg/m²) and radiotherapy. Patients were treated either in a definitive or postoperative adjuvant setting, according to institutional clinical practice. Baseline body composition was assessed by bioelectrical impedance analysis (BIA). Renal function and biochemical data were collected at three time points across 21 days. Variables associated with AKD were identified via univariate analysis (Mann–Whitney U test), followed by multivariate logistic regression for independent predictors. Results – AKD developed in 20% of patients during treatment. Those who developed AKD exhibited a significantly higher median increase in serum creatinine compared to non-AKD patients (Δcreatinine: 0.48 vs. 0.04 mg/dL, p<0.001). Low body cell mass normalized for height (BCM/kg/m) emerged as a significant independent predictor of AKD (OR=1.29, p=0.04). A trend toward association was also observed for leukocyte count and platelet levels. Other nutritional indicators (BMI, phase angle, SMI) did not show a significant association. Importantly, all patients completed radiotherapy and reached the target cisplatin dose. Conclusions – A trend toward an association between low lymphocyte count and reduced muscle mass with AKD development suggests that pre-treatment nutritional and immunological status may influence nephrotoxicity risk. Bioimpedance-derived metrics may provide a non-invasive, reproducible method to identify high-risk patients.