Abstract Background Estimation of GFR (eGFR) using formulae based on serum creatinine concentrations are commonly used to assess kidney function. Physical exercise can increase creatinine turnover and lean mass; therefore, this method may not be suitable for use in exercising individuals. Cystatin-C based eGFR formulae may be a more accurate measure of kidney function when examining the impact of exercise on kidney function. The aim of this study was to assess the agreement of four creatinine and cystatin-C based estimates of GFR before and after a 12-month exercise intervention. Methods One hundred forty-two participants with stage 3–4 chronic kidney disease (CKD) (eGFR 25–60 mL/min/1.73 m2) were included. Subjects were randomised to either a Control group (standard nephrological care [n = 68]) or a Lifestyle Intervention group (12 months of primarily aerobic based exercise training [n = 74]). Four eGFR formulae were compared at baseline and after 12 months: 1) MDRDcr, 2) CKD-EPIcr, 3) CKD-EPIcys and 4) CKD-EPIcr-cys. Results Control participants were aged 63.5[9.4] years, 60.3% were male, 42.2% had diabetes, and had an eGFR of 40.5 ± 8.9 ml/min/1.73m2. Lifestyle Intervention participants were aged 60.5[14.2] years, 59.5% were male, 43.8% had diabetes, and had an eGFR of 38.9 ± 8.5 ml/min/1.73m2. There were no significant baseline differences between the two groups. Lean mass (r = 0.319, p