Skip to main content
. 2010 Sep 24;119(1):144–149. doi: 10.1289/ehp.0901878

Table 2.

Effect of lead and hemoglobin on IQ by DRD2 Taq IA genotype: results from stratified and interaction multivariate GEE models.a

Lead [ln (μg/dL)]
Hemoglobin (g/dL)
Model Genotype n β-Coefficient (95% CI) p-Value β-Coefficient (95% CI) p-Value
−4.22 (−7.10 to −1.36) 0.004 2.30 (1.26 to 3.08) < 0.0001
2: Stratified model A1/A1 73 −9.12 (−18.08 to −0.16) 0.047 −1.82 (−5.28 to 1.64) 0.306
A1/A2 290 −4.35 (−8.24 to −0.44) 0.023 2.24 (0.57 to 3.91) 0.001
A2/A2 354 −3.87 (−8.28 to 0.54) 0.009 3.03 (1.60 to 4.45) < 0.0001
3: Stratified model A1/A1 73 −9.12 (−18.08 to −0.16) 0.046 −1.82 (−5.28 to 1.64) 0.302
A1/A2 + A2/A2 644 −4.00 (−7.21 to −0.69) 0.017 2.72 (1.58 to 3.86) < 0.0001
4: Interaction modelb A1/A1 717 −7.27 (−17.53 to −0.20) > 0.28c −1.32 (−4.66 to 2.02) 0.022c
A1/A2 + A2/A2 −3.94 (−7.10 to −0.79) 2.72 (1.60 to 3.84)
a

Controlling for midarm circumference, age, sex, family income, parental education, and family size and accounting for school and classroom levels.

b

The model assessing the interaction of lead and DRD2 Taq IA controlled for confounding by hemoglobin and vice versa for the model assessing hemoglobin and DRD2 Taq IA interaction.

c

p-Values for interaction term coefficients.