Back
Currently some calculators estimate the risk of progression of chronic kidney disease (CKD). The 4- and 8-variable KFRE estimates the probability of requiring renal replacement therapy (RRT) within 2 to 5 years in patients with CKD stages 3 to 5. The KFRE has been validated in more than 30 countries, but not yet individually in Latin America. The KFRE allows the definition of diagnostic and therapeutic strategies, the timely planning of the initiation of RRT, and the standardized analysis of the quality of the programs, adapted to the case mix and to an external comparator.
Uruguay's National Renal Health Program (PSRU) covers approximately 80% of the country's population. Patients are enrolled in a national CKD registry with baseline and follow-up data. The objective is to validate the KFRE equation for a Latin American country population in its 4- and 8-variable versions.
Observational, longitudinal and retrospective cohort study of the population admitted to the PSRU and followed up for at least 2 years. The ACR was calculated according to the formula of Weaver et al. when these data were not available. The 4-variable and 8-variable KFRE equations were used to estimate the probability of KRT at 2 and 5 years of follow-up. The analysis of ROC curves (AUC) and the ratio of estimated to observed events were used to assess discrimination and calibration ability.
The project was approved by the ethics committee. IBM SPSS v29 software was used.
14,833 patients were included in the analysis. 56.6% men, 69.1 ± 12.2 years, GFR: 37.6 ± 12.1 ml/min, 28% proteinuric (Pru/cru: 1.41 ± 2.1 g/g). KFRE of 4 variables showed a high discriminatory ability (specificity and sensitivity) with AUC of 0.893 at 2 years and 0.876 at 5 years. The observed KRT was higher than the expected probability (41.7% vs. 29.1%, p < 0.01) only in the higher risk quintile (> 5.1%) with AUC: 0.688. KFRE showed good calibration with AUC greater than 0.85 in all stages of CKD and albuminuria. In 668 patients, KFRE was calculated for 8 variables with an AUC of 0.903 at 2 years and 0.898 at 5 years. The frequency of expected (20.2%) and observed (23.2%) events was similar. In the highest risk quintile (>38%), the expected (74.5%) and observed (71.9%) event rates were similar (p=NS), in contrast to the 4-variable KFRE estimate where the observed event rate was 43% higher.
The use of KFRE allows estimation of the risk of chronic kidney disease progression and the need for KRT. The 4-variable KFRE showed excellent discrimination and calibration capacity, except in patients with a very high risk of KRT, where it is underestimated, so that the use of the 8-variable KFRE is suggested in this group. KFRE could be used in Uruguay to plan interventions, establish control strategies and guide individual decision making, as well as in risk-adjusted evaluation of the results of health interventions and quality control of programs and institutions.
This abstract was partially presented at the Uruguayan Society of Nephrology Congress, 2023.