OBJECTIVES: Autosomal dominant polycystic kidney disease (ADPKD) is a genetic disorder characterised by enlarged kidneys and declining renal function. ADPKD progression rates are heterogeneous, influenced by age, gender, renal size and genotype. Disease models often utilise progression rates derived from published studies. This study aimed to compare ADPKD progression, in terms of changes in total kidney volume (TKV) and renal function, modelled from summary versus patient-level data (PLD), and assess the consistency of predictions with trial observations. METHODS: Regression equations were derived from the TEMPO 3:4 trial placebo arm (natural history) to predict annual changes in TKV and estimated glomerular filtration rate (eGFR). Candidate covariates included age, gender, ethnicity, region/country, TKV and eGFR. Predictions were compared using the PLD regression equations or linear interpolation of summary rates of change in four patient categories. Finally, the model was initiated with published baseline patient profiles representing early and late disease from the HALT-PKD trials, and predicted progression compared to trial observations. RESULTS: For patients initiated with the average TEMPO 3:4 placebo profile, predicted eGFR trajectories based on PLD or summary data were similar (average decline: -5.3 and -5.1ml/min/1.73m/year, respectively); however, TKV predictions deviated as TKV exceeded 2,500ml, with increasingly rapid growth predicted based on summary data. The model closely replicated ADPKD progression among patients with early disease; all predicted values within the 95% confidence interval of HALT-PKD observations. In patients with late disease, modelled baseline TKV of 1,000-1,500ml led to closest replication of eGFR observations (average decline: -3.2 to -4.4, versus -3.9ml/min/1.73m/year during trial). CONCLUSIONS: Though predictions based on summary and PLD were consistent, the PLD regression equations produced more realistic results at extreme values. The availability of relevant PLD to describe the natural history of ADPKD progression provides a more robust foundation for disease and economic modelling than summary data alone.