Background: Patients with anorexia nervosa (AN) are at high risk of kidney damage. Impairments in kidney function are often underestimated. Due to their severe malnourishment, including lower muscle mass, the estimated glomerular filtration rate (eGFR) based on creatinine (Cr) may be overestimated and delay the diagnosis of renal dysfunction. Cystatin C (CysC) could be a promising renal marker given it is unaltered by body size, composition, and muscle mass.This study aims to compare eGFR calculated with different equations based on Cr and/or CysC in this population and to evaluate their association with body composition.Method: Prospective study of adolescents with AN diagnosis followed in a tertiary pediatric center from January 2024 to August 2025 who had serum creatinine and CysC values, as well as bioelectrical impedance analysis (BIA) measurements. Data on gender, age, body-mass-index (BMI) z-score, body fat mass, fat mass percentage, muscle mass, muscle mass percentage, and fat-to-muscle (FTM) ratio. The eGFR was calculated using creatinine-based equations, CysC based equations, and equations combining creatinine and CysC. Continuous variables were expressed as the median and interquartile range (IQR). Parameters were compared using non-parametric tests (Mann-Whitney). Values of p<0.05 were considered significant.Results: A total of 26 adolescents were included, 96.2% female, with a median (IQR) age at evaluation of 16.58 (15.47-17.65) years and at diagnosis of 14.37 (13.27-15.21) years. The median BMI z-score was -2.22. The median total fat mass was 4.95kg (<2nd percentile for age and sex), the median percentage of fat mass was 12.15% (<2nd percentile for age and sex), and the median FTM ratio was 0.26.When compared with the Schwartz-bedside equation, all the CysC and combined based formulas showed lower eGFR median values (CKiD Under25-CysC (p<0.001), CKiD Under25-Mean-Cr-CysC (p<0.001), EKFC-CysC (p<0.001), EKFC-Mean-Cr-CysC (p<0.001), Zapitelli-CysC (p<0.001), Zapitelli-combined-Cr-CysC (p<0.001)), and all the creatinine-based formulas showed similar values (CKiD Under25-Cr (p=0.228), EKFC-Cr (p=0.849)). With creatinine-based formulas, all patients had normal renal functions. When CKiD Under25-CysC formula was used 69.2% of patients had mild kidney dysfunction (stage 2 chronic kidney disease - CKD), 42.3% with EKFC-CysC, and 53.8% with Zapitelli-CysC.Those who had stage 2 CKD with the CKiD Under25-CysC equation had significantly lower BMI z-scores (p<0.001), total fat mass (p<0.001), percentage of fat mass (p=0.005), total muscle mass (p=0.035), and FTM ratio (p=0.005). With the EKFC-CysC equation, the patients who had stage 2 CKD had significantly lower BMI z-scores (p=0.001), total fat mass (p=0.005), percentage of fat mass (p=0.011), total muscle mass (p=0.013), and FTM ratio (p=0.013). Lastly, with the Zapitelli-CysC formula, the adolescents with stage 2 CKD had lower BMI z-scores (p=0.031) and total fat mass (p=0.023).Conclusion: In this sample, all patients had normal kidney function when using creatinine-based formulas. However, the CysC-based formulas allowed the detection of mild kidney damage in adolescents with AN who had lower BMI z-score, lower fat mass and fat mass percentage, and lower muscle mass. Therefore, these equations might be more suitable to detect impaired kidney function in these patients, which might go undetected otherwise.
