Table 4.
Linear regression models to evaluate effect modification by age in tertiles on associations of blood and tibia lead with uric acid in all lead workers, with outliers removed (method 1), and with additional control for systolic blood pressure (method 2) and serum creatinine (model 3) (n = 803).
| Method 1
|
Method 2
|
Method 3
|
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Variable | β-coefficient | SE β | p-Value | β-coefficient | SE β | p-Value | β-coefficient | SE β | p-Value |
| Blood lead model | |||||||||
| Intercept | 4.9217 | 0.0757 | < 0.01 | 4.9350 | 0.0759 | < 0.01 | 4.8528 | 0.0736 | < 0.01 |
| Age (years) | −0.0182 | 0.0039 | < 0.01 | −0.0199 | 0.0040 | < 0.01 | −0.0210 | 0.0039 | < 0.01 |
| Systolic blood pressure (mm Hg) | — | — | — | 0.0047 | 0.0023 | 0.04 | 0.0046 | 0.0022 | 0.04 |
| Serum creatinine (mg/dL) | — | — | — | — | — | — | 2.1830 | 0.2666 | < 0.01 |
| Blood lead (μg/dL) | 0.0111 | 0.0041 | < 0.01 | 0.0105 | 0.0041 | 0.01 | 0.0071 | 0.0039 | 0.07 |
| Blood lead × age category 2 | −0.0109 | 0.0057 | 0.05 | −0.0107 | 0.0056 | 0.06 | −0.0063 | 0.0054 | 0.25 |
| Blood lead × age category 1 | −0.0150 | 0.0058 | 0.01 | −0.0148 | 0.0058 | 0.01 | −0.0107 | 0.0056 | 0.06 |
| Tibia lead model | |||||||||
| Intercept | 4.8932 | 0.0749 | < 0.01 | 4.9087 | 0.0750 | < 0.01 | 4.8430 | 0.0735 | < 0.01 |
| Age (years) | −0.0155 | 0.0039 | < 0.01 | −0.0174 | 0.0040 | < 0.01 | −0.0184 | 0.0038 | < 0.01 |
| Systolic blood pressure (mm Hg) | — | — | — | 0.0052 | 0.0022 | 0.02 | 0.0048 | 0.0022 | 0.03 |
| Serum creatinine (mg/dL) | — | — | — | — | — | — | 2.1808 | 0.3189 | < 0.01 |
| Tibia lead (μg Pb/g bone mineral) | 0.0036 | 0.0018 | 0.04 | 0.0031 | 0.0018 | 0.08 | 0.0019 | 0.0017 | 0.28 |
| Tibia lead × age category 2 | −0.0057 | 0.0028 | 0.04 | −0.0053 | 0.0028 | 0.06 | −0.0019 | 0.0028 | 0.49 |
| Tibia lead × age category 1 | −0.0071 | 0.0029 | 0.02 | −0.0067 | 0.0029 | 0.02 | −0.0044 | 0.0029 | 0.13 |
—, Variable not included in method. Models were also adjusted for sex, BMI, and alcohol use. The oldest age tertile is the reference category. Slopes in the middle (age category 2) and youngest (age category 1) age categories are obtained by adding their respective β-coefficients (of the cross-product term for age × lead) to the β-coefficient of the reference category (oldest age group). p-Values for the cross-product terms reflect the statistical significance of the difference between the slopes of the regression line in that age category and the regression line for the oldest age group.