Key points are not available for this paper at this time.
Abstract Background and Aims Determining prognosis in patients with autosomal dominant polycystic kidney disease (ADPKD) heavily relies on the height-adjusted total kidney volume (HtTKV) Mayo classification. This classification predicts kidney outcomes on a group level but lacks precision in individual patients. Recently, several urine and blood biomarkers have demonstrated prognostic value in ADPKD, independent of HtTKV and baseline eGFR. This study compares the prognostic value of various biomarkers to aid selection when developing novel risk stratification tools. Method Included were participants of the DIPAK studies with a follow-up duration of 1 year and ≥3 eGFR measurements. We excluded patients with diabetes mellitus and patients using tolvaptan or somatostatin analogues. At baseline, urine samples were analyzed for albumin, urea, creatinine, β2-microglobulin (β2-MG), monocyte chemotactic protein-1 (MCP-1), TmP/GFR (tubular maximum reabsorption rate of phosphate to GFR), and blood was drawn to measure creatinine, urea (for assessment of urinary-to-plasma urea ratio; UP urea), glucagon, fibroblast growth factor 23 (FGF-23) and copeptin. Albumin, β2-MG and MCP-1 were divided by urinary creatinine concentration to account for concentration effects. The cross-sectional and longitudinal association with baseline eGFR and eGFR decline over time was analyzed using multivariate linear regression and repeated measures linear mixed model regression. Furthermore, logistic regression models were tested to identify patients with rapidly-progressive disease (eGFR slope ≤ −3.0 ml/min/1.73 m2/year). Lastly, multivariate Cox regression analysis was conducted to evaluate the time to the combined endpoint of 30% decline in eGFR, reaching an eGFR ≤ 15 ml/min/1.73 m2, initiation of dialysis or receiving a kidney transplant. Results 565 patients were included with a mean age of 47 ± 11 years, eGFR of 64 ± 27 ml/min/1.73 m2 and HtTKV of 858 IQR: 540, 1305 ml/m. During a mean follow-up of 5.2 ± 2.2 years, the mean eGFR decline was −3.38 ± 2.37 mL/min/1.73 m2/year. Of these patients, 283 (50.1%) had rapidly progressive disease and 267 (47.3%) reached the combined kidney endpoint. All biomarkers except glucagon were associated with baseline eGFR and eGFR decline over time after adjustment for established clinical risk factors for disease progression (sex, PKD mutation and baseline age and eGFR). After adjustment for the same covariates, albumin/creatinine, MCP-1/creatinine, FGF-23, UP-urea and copeptin were predictive of rapid disease progression. The area under the curve (AUC) value for predicting rapidly progressive disease for the Mayo HtTKV classification alone was 0.694, which could be improved to 0.766 (P .001) by a risk prediction model that also included the aforementioned predictive clinical risk factors. The introduction of several biomarkers further improved this clinical model, including albumin/creatinine (AUC 0.794, p = 0.03), MCP-1/creatinine (AUC 0.804, p = 0.04) and copeptin (AUC 0.789, p = 0.05). Similar results were seen in multivariate Cox regression models after adjustment for covariates, with independent predictors of the combined endpoint being again albumin/creatinine (P .001), MCP-1 creatinine (P .001), copeptin (p = 0.012), but now also UP-urea (p = 0.008) and FGF-23 (P .001). These findings appeared robust in subgroup analyses of 200 patients with Mayo HtTKV class 1C and 122 patients with early disease (age ≤ 40 years and eGFR ≥ 60 ml/min/1.73 m2 at baseline), in whom predicting prognosis is most challenging. Conclusion In patients with ADPKD, albumin/creatinine, MCP-1/creatinine, FGF-23 and copeptin are consistent predictors of kidney outcomes after adjustment for established clinical risk factors. Our data suggest that these biomarkers should be prioritized when developing novel multimarker models to enhance precision of risk prediction in ADPKD.
Bais et al. (Wed,) studied this question.