Table 3.
Associations of cardiac biomarkers and incident HF in the CRIC Study
| Cardiac Biomarker | Model 1a | Model 2b | Model 3c | |||
|---|---|---|---|---|---|---|
| Hazard Ratio (95% Confidence Interval) | P Value | Hazard Ratio (95% Confidence Interval) | P Value | Hazard Ratio (95% Confidence Interval) | P Value | |
| Continuous predictors | ||||||
| Log(hsTnT) per 1 SD (1.12) increase | 1.60 (1.36 to 1.88) | <0.001 | 1.50 (1.27 to 1.78) | <0.001 | 1.27 (1.07 to 1.52) | 0.008 |
| Log(NT-proBNP) per 1 SD (1.39) increase | 2.30 (1.99 to 2.66) | <0.001 | 2.11 (1.80 to 2.47) | <0.001 | 1.99 (1.69 to 2.35) | <0.001 |
| Categorical predictorsd | ||||||
| hsTnT, pg/ml (reference: undetectable) | ||||||
| >5–9.3 | 1.75 (0.96 to 3.22) | <0.001 | 1.96 (1.02 to 3.79) | <0.001 | 1.98 (1.03 to 3.83) | 0.001 |
| >9.3–15.0 | 2.47 (1.38 to 4.39) | 2.70 (1.44 to 5.05) | 2.51 (1.34 to 4.71) | |||
| >15.0–26.5 | 2.72 (1.52 to 4.88) | 2.96 (1.57 to 5.59) | 2.47 (1.31 to 4.66) | |||
| >26.5 | 4.85 (2.68 to 8.80) | 4.77 (2.49 to 9.14) | 3.71 (1.93 to 7.13) | |||
| NT-proBNP, pg/ml (reference: ≤47.6) | ||||||
| >47.6–95.8 | 2.71 (1.28 to 5.75) | <0.001 | 2.71 (1.23 to 5.98) | <0.001 | 2.57 (1.17 to 5.68) | <0.001 |
| >95.8–188.8 | 4.11 (1.99 to 8.50) | 3.72 (1.72 to 8.04) | 3.41 (1.58 to 7.39) | |||
| >188.8–433.0 | 5.73 (2.80 to 11.74) | 5.47 (2.55 to 11.74) | 4.76 (2.21 to 10.23) | |||
| >433.0 | 11.83 (5.75 to 24.34) | 9.57 (4.40 to 20.83) | 8.07 (3.70 to 17.64) | |||
Model 1 adjusted for age, sex, race/ethnicity, clinical center, diabetes status, self-reported cardiovascular disease at baseline, current smoking, alcohol use, log(24-hour urine total protein excretion), eGFR, systolic BP, body mass index, LDL, and HDL.
Model 2 adjusted for model 1 variables plus hemoglobin, angiotensin-converting enzyme/angiotensin receptor blocker use, diuretic use, β-blocker use, serum phosphorus, log(PTH), and log(FGF23).
Model 3 adjusted for model 2 variables plus the alternative cardiac biomarker (e.g., adjusted for NT-proBNP and hsTnT in the models and vice versa).
Detectable levels of hsTnT were divided into quartiles; NT-proBNP categories were defined by quintiles.