Background: Medication errors due to inaccurate measures of kidney function are common among elderly patients. We investigated differences between estimated glomerular filtration rate (eGFR) based on creatinine and cystatin C and how these differences would affect prescribing recommendations among acutely hospitalized elderly patients. We also identified factors associated with discrepancies between estimates. Methods: Estimated glomerular filtration rate and chronic kidney disease (CKD) classifications were determined for 338 acutely hospitalized elderly patients using equations from Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), Berlin Initiative Study (BIS) and Cockcroft-Gault (CG). Prescribed renal risk medications were compared with dosing guidelines in Renbase ® . Linear regression models were used to identify explanatory variables for eGFR discrepancies between equations. Muscle weakness was assessed by handgrip strength; inflammation was assessed by smoking status, serum C-reactive protein (CRP), soluble urokinase plasminogen activator receptor (suPAR) and neutrophil gelatinase-associated lipocalin (NGAL); and organ dysfunction was assessed by thyroid-stimulating hormone (TSH) and FI-OutRef. Results: Median eGFR values were 65.5, 60.7, 54.1, 57.1, 55.1 and 57.6 mL/ min/1.73m 2 according to CKD-EPI Cr , CKD-EPI Comb , CKD-EPI Cys , BIS Cr , BIS Comb and CG Cr , respectively. Depending on choice of equation, renal risk medications were prescribed at higher than recommended dose in 13.6% to 22.5% of patients using normalized GFR units and 9.9% to 19.1% of patients using absolute units. Age, handgrip strength, CRP, suPAR, NGAL and smoking status had significant association with eGFR discrepancies between creatinine-and cystatin C-based equations. Conclusions: Significant discrepancies in eGFR and CKD classification were observed when switching between eGFR equations in acutely hospitalized elderly patients. Switching from a creatinine-based equation to its corresponding cystatin C-based equation resulted in lower GFR estimates, and these differences were M. B. Houlind and J. Petersen are sharing last authorship.