Table 3 Integrated discrimination improvement and net reclassification improvement for pulse pressure amplification

From: An outcome-driven threshold for pulse pressure amplification

Cohort Endpoint Risk indicator

Integrated discrimination improvement

Net reclassification improvement

IDI (%)

P

NRI (%)

P

Discovery (n = 3945)

 Cardiovascular

   PPA < 1.3 vs ≥1.3

0.12 (−0.25, 0.50)

0.52

16.1 (1.73, 30.5)

0.028

   Continuous

0.27 (−0.15, 0.69)

0.21

9.26 (−5.29, 23.8)

0.21

 Coronary

   PPA < 1.3 vs ≥1.3

0.41 (−0.41,1.22)

0.33

33.0 (11.9, 54.2)

0.002

   Continuous

0.40 (−0.47, 1.27)

0.36

21.6 (−0.41, 43.7)

0.054

Replication (n = 1663)

 Cardiovascular

   PPA < 1.3 vs ≥1.3

0.50 (−0.72, 1.71)

0.42

33.7 (10.0, 57.4)

0.005

   Continuous

0.76 (−0.35, 1.88)

0.18

22.7 (−1.74, 47.1)

0.069

 Coronary

   PPA < 1.3 vs ≥1.3

0.89 (−2.08, 3.85)

0.56

44.9 (11.7, 78.2)

0.008

   Continuous

1.08 (−0.60, 2.76)

0.21

30.2 (−4.45, 64.8)

0.088

All (n = 5608)

 Cardiovascular

   PPA < 1.3 vs ≥1.3

0.16 (−0.20, 0.53)

0.38

15.2 (2.76, 27.6)

0.017

   Continuous

0.29 (−0.10, 0.68)

0.15

11.8 (−0.70, 24.3)

0.064

 Coronary

   PPA < 1.3 vs ≥1.3

0.49 (−0.37, 1.34)

0.26

30.8 (12.7,48.9)

0.001

   Continuous

0.46 (−0.28, 1.21)

0.22

21.5 (2.85, 40.2)

0.024

  1. The base model included cohort (random effect), sex, age, body mass index, heart rate, smoking and drinking, total-to-HDL serum cholesterol ratio, estimated glomerular filtration rate, antihypertensive drug intake, history of cardiovascular disease, and diabetes. The integrated discrimination improvement (IDI) is the difference between the discrimination slopes of the base model and the base model extended by PPA. The discrimination slope is the difference in predicted probabilities (%) between participants without and with an endpoint. The net reclassification index (NRI) is the sum of the percentages of participants reclassified correctly in individuals without and with an endpoint. IDI and NRI estimates are given with 95% confidence interval