Abstract Background Autosomal dominant polycystic kidney disease (ADPKD) causes about 10% of cases of end stage renal disease. Disease progression rate is heterogeneous. Tolvaptan is presently the only specific therapeutic option to slow kidney function decline in adults at risk of rapidly progressing ADPKD with chronic kidney disease (CKD) stages 1–4. Thus, a reliable evaluation of kidney function in patients with ADPKD is needed. Methods We evaluated the agreement between measured (mGFR) and estimated glomerular filtration rate (eGFR) by 61 formulas based on creatinine and/or cystatin-C (eGFR) in 226 ADPKD patients with diverse GFR values, from predialysis to glomerular hyperfiltration. Also, we evaluated whether incorrect categorization of CKD using eGFR may interfere with the indication and/or reimbursement of Tolvaptan treatment. Results No formula showed acceptable agreement with mGFR. Total Deviation Index averaged about 50% for eGFR based on creatinine and/or cystatin-C, indicating that 90% of the estimations of GFR showed bounds of error of 50% when compared with mGFR. In 1 out of 4 cases with mGFR < 30 ml/min, eGFR provided estimations above this threshold. Also, in half of the cases with mGFR between 30 and 40 ml/min, formulas estimated values < 30 ml/min. Conclusions The evaluation of renal function with formulas in ADPKD patients is unreliable. Extreme deviation from real renal function is quite frequent. The consequences of this error deserve attention, especially in rapid progressors who may benefit from starting treatment with tolvaptan and in whom specific GFR thresholds are needed for the indication or reimbursement. Whenever possible, mGFR is recommended. Graphic abstract