Abstract: FR-PO648

Incremental Value of Renal Pathological Score to Kidney Failure Risk Equation in Advanced Diabetic Nephropathy

Session Information

Category: Diabetes

  • 502 Diabetes Mellitus and Obesity: Clinical

Authors

  • Yamanouchi, Masayuki, Graduate School of Medical Sciences, Kanazawa University, Kanazawa, Japan
  • Hoshino, Junichi, Toranomon Hospital, Tokyo, Japan
  • Ubara, Yoshifumi, Toranomon Hospital, Tokyo, Japan
  • Toyama, Tadashi, Kanazawa University Hospital, Kanazawa, Japan
  • Kitajima, Shinji, Department of Nephrology, Kanazawa University Hospital, Kanazawa, Japan
  • Hara, Akinori, Kanazawa University, Kanazawa, Japan
  • Shimizu, Miho, Division of Nephrology, Kanazawa University Hospital, Kanazawa, ishikawa, Japan
  • Furuichi, Kengo, Kanazawa University, Kanazawa, Japan
  • Wada, Takashi, Kanazawa University, Kanazawa, Japan
Background

The Kidney Failure Risk Equation (KFRE) that enables to accurately predict end-stage renal disease (ESRD) in patients with chronic kidney disease (CKD) in stages 3 to 5 was developed and validated worldwide. We aimed to first assess the performance of KFRE on individuals with biopsy proven diabetic nephropathy in advanced CKD stages at the time of biopsy and then evaluate the incremental value of pathologic information of renal biopsy to KFRE on them.

Methods

296 individuals with biopsy proven diabetic nephropathy in CKD stages 3 to 5 at the time of biopsy was identified at four nephrology centers in Japan. Pathological classification was performed by three pathologists based on the Pathologic Classification of Diabetic Nephropathy. Individuals were randomly assigned to two cohorts (2:1 cohorts). The development cohort of 198 was used to validate the KFRE and to assess the incremental value of pathological score on diabetic nephropathy (D-score). Model performance was assessed by chi-squared test, Akaike information criterion, and Harrell’s c-statistics. Incremental value of D-score to KFRE was evaluated with net reclassification improvement (NRI), integrated discrimination improvement (IDI), Integrated sensitivity (IS) and integrated specificity (IP). Validation of the models was performed in the validation cohort of 98.

Results

Median follow-up durations (25th, 75th percentiles) were 1.8 (1.0, 5.0) years and 2.0 (1.0, 3.8) years, respectively (p=0.89). Both KFRE and KFRE+D-score were significant predictors of ESRD in both the development cohort and validation cohort (Hazard Ratios of KFRE and KFRE+D-score in the development cohort were 2.44 (1.89-3.15) and 1.08 (1.04-1.13), and those in the validation cohort were 2.30 (1.57-3.38) and 1.08 (1.01-1.16), respectively). Incremental value of D-score using free cut-points showed positive overall NRI (0.5%; CI, 0.1-0.5%) but the IDI showed no significant change (0.0002; CI, -0.00007-0.0005). The KFRE+D-score model improved the both IS and IP but showed little differences.

Conclusion

KFRE worked well to identify individuals at high risk of ESRD in both the development and validation cohort. Adding pathological information to the KFRE improved the risk prediction of ESRD but did not statistically outperform the KFRE.