Skip to main content

Table 5 Continuous net reclassification improvement (NRI) and integrated discrimination improvement (IDI) for the identification of individuals who experienced death or cardiovascular event during 5-year follow-up by cardiovascular risk factor models including either KCD20, eGFRstep, or eGFR60, in comparison with cardiovascular risk factor model without kidney function parameter

From: Kidney age - chronological age difference (KCD) score provides an age-adapted measure of kidney function

 

Estimate (95% CI)

P value

Cardiovascular risk factor model including KCD20, in comparison with cardiovascular risk factor model without kidney function parameter.

 NRI (events & non-events)

0.126 (0.042, 0.201)

0.0020

 NRI (events)

−0.504 (−0.575, − 0.424)

< 0.0001

 NRI (non-events)

0.630 (0.580, 0.662)

< 0.0001

 IDI

0.00359 (0.00098, 0.00620)

0.0069

Cardiovascular risk factor model including eGFRstep, in comparison with cardiovascular risk factor model without kidney function parameter.

 NRI (events & non-events)

−0.010 (− 0.163, 0.122)

NS

 NRI (events)

−0.607 (− 0.705, 0.599)

NS

 NRI (non-events)

0.597 (−0.485, 0.755)

NS

 IDI

0.00174 (−0.00024, 0.00372)

NS

Cardiovascular risk factor model including eGFR60, in comparison with cardiovascular risk factor model without kidney function parameter.

 NRI (events & non-events)

0.206 (0.105, 0.295)

< 0.0001

 NRI (events)

−0.329 (− 0.401, − 0.235)

< 0.0001

 NRI (non-events)

0.534 (0.447, 0.576)

< 0.0001

 IDI

0.00603 (0.00259, 0.00947)

0.0006

  1. CI confidence interval, eGFR60 eGFR < 60 ml/min/1.73 m2, eGFRstep age-dependent stepped eGFR criteria of Delanaye et al. [10], KCD20 Kidney age-Chronological age Difference score ≥ 20 years
  2. Continuous NRI for time-to-event data with inverse probability weighting calculated as described by Pencina et al. [33,34,35] and CI estimated with bootstrapping. IDI calculation based on sex-specific 5-year absolute risk, calculated as described by Pencina et al. [33]