Abstract: TH-PO1095
Validation of the Kidney Risk Failure Equations (KFRE) in Patients with CKD in Singapore
Session Information
- CKD: Clinical, Outcomes, Trials - I
October 25, 2018 | Location: Exhibit Hall, San Diego Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: CKD (Non-Dialysis)
- 1902 CKD (Non-Dialysis): Clinical, Outcomes, and Trials
Authors
- Liew, Ian tatt, Tan Tock Seng Hospital, Singapore, Singapore
- Weng, Wanting, Tan Tock Seng Hospital, Singapore, Singapore
- Ang, Yee, National Healthcare Group, Singapore, Singapore
- Yap, Chun wei, National Healthcare Group, Singapore, Singapore
- Shen, Peng, Tan Tock Seng Hospital, Singapore, Singapore
- Ooi, Xi yan, Tan Tock Seng Hospital, Singapore, Singapore
- Lim, Regina, Tan Tock Seng Hospital, Singapore, Singapore
- Liew, Adrian, Tan Tock Seng Hospital, Singapore, Singapore
Background
Accurate identification of patients with CKD at risk of progressing to require renal replacement therapy (RRT) would allow optimal care. Tangri et al developed models to predict progression in patients with CKD stages 3-5 (eGFR 10-59ml/min/1.73m2) to kidney failure using data from Canadian cohorts. The aim of our study is to validate the 4-variable and 8-variable KFREs in the multiethnic population in Singapore.
Methods
Demographics, clinical and laboratory data of patients with CKD stages 3-5 referred from primary care physicians to TTSH Renal Department between 1 January 2009 to 31 December 2012 were collected. Kidney failure was defined as the initiation of chronic dialysis or kidney transplant. The 2 year and 5 year kidney failure risks were predicted using the 4-variable(age, sex, albuminuria and eGFR) and 8-variable(4-variable and calcium, phosphate, bicarbonate and albumin) models. This was compared with actual kidney failure rate at 2 and 5 years of follow up. Patients who died before kidney failure were excluded. Model performance was evaluated using the area under the receiver operating characteristic curve (ROC-AUC) and by comparing the observed and predicted risks of kidney failure.
Results
The KFRE models were validated using 2,238 and 1,845 patients for the 2 and 5 year kidney failure risks respectively. 158 (7.1%) and 355 (15.9%) patients developed kidney failure at 2 years and 5 years respectively. ROC-AUC for the 4-variable KFRE at 2 and 5 years are 0.921 (95% confidence interval [CI] 0.899-0.943) and 0.866 (CI 0.844-0.887) while those for the 8-variable KFRE are 0.925 (CI 0.905-0.945) and 0.862 (CI 0.840-0.883) respectively. The mean differences between the observed and predicted kidney failure risk was lower for the 4 variable equation compared to the 8 variable equation at 2 and 5 years.
Conclusion
Discrimination of the 4-variable and 8-variable KFRE were similar and were able to predict the progression to kidney failure in our patients. To further improve the accuracy of the KFRE, we would need to recalibrate the 8-variable equation for our local multiethnic population
Mean difference between observed and predicted risks
| 2 years risk | 5 years risk | |
| 4-variables KFRE | -1.5% | -2.9% |
| 8-variables KFRE | -4.1% | -8.6% |