Table 2.
Novel CKD ASCVD risk prediction models developed in the CRIC study
| Model Parameters | Model 1: ACC/AHA PCEs Variables | Model 2: CRIC Clinical Model | Model 3: CRIC Enriched Model | |||
|---|---|---|---|---|---|---|
| Variables | HR (95% CI) | β, per 1-unit | HR (95% CI) | β, per 1-unit | HR (95% CI) | β, per 1-unit |
| Age, per 5 yr | 1.15 (1.09 to 1.21) | 0.0275 | 1.23 (1.16 to 1.31) | 0.0420 | 1.19 (1.12 to 1.26) | 0.0347 |
| Male sex | 1.09 (0.87 to 1.38) | 0.0900 | ||||
| Black race | 1.21 (0.96 to 1.51) | 0.1885 | ||||
| Total cholesterol, per 10 mg/dl | 1.02 (1.00 to 1.04) | 0.0020 | 1.08 (1.01 to 1.15) | 0.0073 | ||
| HDL cholesterol, per 10 mg/dl | 0.87 (0.80 to 0.95) | −0.0134 | 0.88 (0.81 to 0.95) | −0.0127 | 0.85 (0.77 to 0.94) | −0.0164 |
| Systolic BP, per 10 mm Hg | 1.13 (1.08 to 1.19) | 0.0126 | 1.08 (1.02 to 1.14) | 0.0077 | ||
| Use of BP-lowering medications | 1.54 (0.97 to 2.46) | 0.4342 | ||||
| History of diabetes | 1.88 (1.49 to 2.37) | 0.6300 | ||||
| Current smoking | 2.09 (1.57 to 2.78) | 0.7382 | 2.03 (1.53 to 2.69) | 0.7086 | 1.94 (1.46 to 2.57) | 0.6615 |
| Log (urine ACR), mg/g | 1.17 (1.11 to 1.23) | 0.1553 | 1.12 (1.05 to 1.19) | 0.1127 | ||
| HbA1c, per 0.5% | 1.09 (1.05 to 1.12) | 0.1634 | 1.08 (1.05 to 1.12) | 0.1616 | ||
| Hemoglobin, per 5 g/dl | 0.64 (0.46 to 0.88) | −0.0903 | ||||
| Apolipoprotein B, per 10 mg/dl | 0.89 (0.79 to 0.99) | −0.0121 | ||||
| Log (hsCRP), mg/L | 1.11 (1.02 to 1.21) | 0.1045 | ||||
| Log (troponin-T), pg/ml | 1.23 (1.09 to 1.39) | 0.2090 | ||||
| Log (NT-proBNP), pg/ml | 1.22 (1.11 to 1.35) | 0.2011 | ||||
| Score calculations | ||||||
| Baseline survival at 10 years, S(10) | 0.9045 | 0.9075 | 0.9100 | |||
| Predicted probability of ASCVD at 10 yearsa | 1– S(10)exp(ΣbX–3.6963) | 1– S(10)exp(ΣbX–3.3283) | 1– S(10)exp(ΣbX–4.5580) | |||
a
The predicted 10-year probability for each model can be calculated as 1– S(10)exp(ΣbX–betaavg) where b is the regression coefficient (β), X is the individual patient’s level for each risk factor, and betaavg is the sample mean of the β×value sum.