Refining Long-Term Prediction of Cardiovascular Risk in Diabetes – The VILDIA Score

Figures

• 

  Figure 1. Variable selection by a bootstrap resampling procedure based on Cox regression analysis (cut off level for selection: 80%). To avoid collinearity the following variables were selected for the bootstrap procedure due to their highest univariate predictive value (defined by the hazard ratio for 1-SD change derived from Cox regression) from clusters with close correlations (r > 0.4): Cystatin (NOT selected: BUN, creatinine, ADMA, SDMA, eGFR), free fatty acids, LDL-triglycerides and Lp-PLA2 (NOT selected: total cholesterol, HDL-triglycerides, lipoprotein(a), free glycerol, apolipoprotein A-I, HDL-cholesterol ester, HDL-phospholipid, HDL-cholesterol, cholesterol ester, LDL-cholesterol, apolipoprotein E, LDL-apolipoprotein B, apolipoprotein B, LDL phospholipid, triglycerides, HDL free cholesterol, apolipoprotein C-III, a-Tocopherol, phospholipid, VLDL apolipoprotein B, VLDL-cholesterol ester, VLDL-cholesterol, LDL free cholesterol, VLDL free cholesterol, apolipoprotein A-II), body mass index (NOT selected: waist-to-hip ratio), antithrombin III (NOT selected: retinol), lycopin (NOT selected: alpha-carotin, beta-carotin, lutein, zeaxanthin, all-trans beta-carotin, cis-beta-carotin,…

• 

  Figure 2. Kaplan-Meier estimates of mortality. Kaplan-Meier estimates for tertiles of VILDIA score for all-cause mortality (A) and cardiovascular mortality (B) (all P < 0.001 between all tertiles, log rank test).

• 

  Figure 3. Risk prediction of simplified risk score and association with established cardiovascular medication. The bars show 10-year all-cause (3A) and cardiovascular (3B) mortality stratified by the simplified biomarker score. We further found an inverse associations between the simplified risk and aspirin use (for trend, p < 0.007, Fig. 3C), beta blocker use (for trend, p < 0.001, Fig. 3D) and statin use (for trend, p < 0.001, Fig. 3E) at discharge. The simplified score was based on the following cut-off values: [male = 1] + [age ≥ 75years (86th percentile) = 1] + [NT-proBNP ≥ 400 ng/L (51th percentile) = 1, ≥2000 ng/L (85th percentile) = 2] + [diabetes duration ≥ 5 years (81th percentile) = 1] + [renin ≥ 50 pg/ml (74th percentile) = 1)] + [25-OH vitamin D3 < 10 ng/L (29th percentile) = 1] + [Lp-PLA2 > 450 U/l (43rd percentile) = 1].