Table 2.
Variable (Increment) | Adjusted Difference in PCSK9 Concentration, ng/ml [95% Confidence Interval]a | Proportional Marginal Variance Decompositionb [95% Confidence Interval]a | P Value |
---|---|---|---|
Statin treatment | 70.0 [74.9 to 65.1] | 0.1480 [0.1313 to 0.1662] | <0.001 |
LDL cholesterol, 10 mg/dl | 2.6 [2.0 to 3.1] | 0.0167 [0.0111 to 0.0235] | <0.001 |
Women | 21.5 [16.6 to 26.4] | 0.0162 [0.0099 to 0.0228] | <0.001 |
Diabetes mellitus | 13.6 [8.6 to 18.7] | 0.0054 [0.0020 to 0.0099] | <0.001 |
HDL cholesterol, 10 mg/dl | 3.1 [1.8 to 4.5] | 0.0037 [0.0010 to 0.0081] | <0.001 |
Body mass index, 1 kg/m2 | 0.92 [0.52 to 1.33] | 0.0036 [0.0012 to 0.0073] | <0.001 |
Cardiovascular disease | 11.3 [5.8 to 16.7] | 0.0028 [0.0007 to 0.0061] | <0.001 |
Age, 10 yr | −3.7 [–5.8 to –1.6] | 0.0026 [0.0006 to 0.0055] | <0.001 |
Current smoking | 5.3 [2.2 to 8.4] | 0.0022 [0.0004 to 0.0050] | <0.001 |
Lipoprotein(a),a 10 mg/dl | 1.09 [0.35 to 1.82] | 0.0006 [0.0000 to 0.0030] | 0.07 |
UACR,b 100 mg/g | 0.30 [0.06 to 0.54] | 0.0001 [0.0000 to 0.0018] | 0.42 |
eGFR, 10 ml/min per 1.73 m2 | 0.41 [–0.89 to 1.70] | 4.16E-05 [0.0000 to 0.0016] | 0.65 |
hs-CRP,b 1 mg/L | 0.23 [–5.32 to 0.52] | 8.69E-06 [0.0000 to 0.0009] | 0.83 |
Σ = 0.2019 (total R2) |
A multivariable regression analysis was performed to identify variables that are significantly associated with PCSK9 concentrations in the entire population at the baseline investigation. All variables were included in one single model and are ordered by their relative importance to the model given by the proportional marginal variance decomposition (pmvd) metric. Estimates for the continuous variables are provided for a clinically relevant increment. UACR, urine albumin-creatinine ratio; hs-CRP, high-sensitivity C-reactive protein.
For variables that were skewed, β-estimates were taken from the untransformed model, whereas their relative importance (pmvd) and P values were calculated from the log-transformed model.
pmvd is defined as the relative importance of all included variables on proprotein convertase subtilisin/kexin type 9 concentrations.