Table 2.
Cross-sectional stepwise multivariable linear regression modelling for homoarginine (hArg) (n = 2057).
| Explanatory variable | β | 95% CI | P-value |
|---|---|---|---|
| Male sex | 0.14 | [0.10, 0.17] | 4.1 × 10−13 |
| Body mass index (kg/m2) | 0.014 | [0.0093, 0.018] | 1.2 × 10−9 |
| Daily smoking | −0.089 | [−0.12, −0.056] | 1.3 × 10−7 |
| Age (years) | −0.0067 | [−0.0097, −0.0038] | 7.8 × 10−6 |
| lnSHBG (nmol/L) | 0.061 | [0.032, 0.089] | 3.2 × 10−5 |
| lnTriglycerides (mmol/L) | 0.045 | [0.019, 0.072] | 8.2 × 10−4 |
| LDL cholesterol (mmol/L) | 0.022 | [0.0045, 0.040] | 0.014 |
| lnCRP (mg/L) | 0.016 | [0.0023, 0.029] | 0.021 |
Statistics: In the bi-directional stepwise regression modelling applied, serum hArg was used as a dependent variable and all the variables shown in Table 1 and lnSHBG as explanatory variables. Those variables that were selected by Akaike’s information criterion (AIC) using the stepAIC R function with the default settings and had a p-value < 0.05 are shown above. HArg, SHBG, triglycerides and CRP were natural log-transformed. For the continuous variables, β (95% CI) are shown for each 1‐unit change in the variable. Abbreviations: SHBG, sex hormone-binding globulin; LDL, low-density lipoprotein; CRP, C-reactive protein.