Table 4. Linear regression models evaluating effect of HFE genotypes on the association between lead exposure and iron metabolism.

	β	SE	Bcoff	t value	P-value1	R 2	F value	P-value2
Variable				BPb(µg/L)
Content	340.588	65.512		5.199	0.000
BIC	−5.507	1.953	−0.119	−2.280	0.005
BIC×HH or HD DD	−1.149	0.432	−0.099	−2.661	0.008	0.594	135.187	0.000
				BIC(mg/kg)
Content	40.661	0.532		76.374	0.000
BPb×HH or HD DD	−0.002	0.000	−0.284	−9.254	0.000	0.444	135.408	0.000
				Tf(g/L)
Content	3.649	0.109		33.474	0.000
BPb	−0.002	0.001	−0.538	−2.877	0.004
BPb×HH or HD DD	0.000	0.000	0.414	2.237	0.026	0.141	26.626	0.000

Abbreviations: BPb, blood lead; Tf, transferrin; BIC, body iron content; HH, wild-type; HD, H63D heterozygous variant; DD, H63D homozygous variant.

Linear models were adjusted stepwise for age (year), gender (male vs. female), education (lower than high school vs. higher than high school), marriage (yes vs. no), tobacco use (yes vs. no), alcohol consumption (yes vs. no), occupational lead exposure (unexposed, dissolved lead operations or electrolytic lead operations) and work years. H63D genotype (HH vs. HD or DD), iron metabolic index/BPb and the cross-product with the genotype and each iron metabolic index/BPb. While BPb was independent factor, each iron metabolic index and the cross-product with the genotype were put separately into the model.

¹ P value for each statistic;

² P value for each regression model.