Table 4.

Multivariable regression analyses of the associations between U-As and telomere length stratified for arsenic metabolism efficiency (above and below median of fraction of each metabolite)^a.

Study group	TLb	% iAs	n	β1 (95% CI)c	p-valuec	p-valued
Andes	0.41	< 11	94	0.0024 (−0.0089 to 0.014)	0.67	0.35
	0.45	≥11	114	0.0089 (0.00036 to 0.017)	0.041
Chaco	1.01	< 17	80	0.011 (−0.0059 to 0.028)	0.20	0.35
	1.19	>17	79	0.018 (0.0019 to 0.034)	0.029
		% DMA
Andes	0.45	< 80.9	114	0.0069 (−0.0012 to 0.015)	0.093	0.73
	0.41	≥80.9	94	0.0051 (−0.0076 to 0.018)	0.43
Chaco	1.23	< 71.1	79	0.012 (−0.0028 to 0.026)	0.11	0.61
	1.001	≥71.1	80	0.024 (−0.0053 to 0.054)	0.11

^aU-As referred to as sum of urinary arsenic metabolites. The number of individuals in Andes in each < and ≥ median split groups varied for telomere length as telomere length measurements were only available for individuals sampled in 2008.

^bPresented as median relative values of telomere length.

^cTelomere length = α + β1 × U-As (per 100 μg/L) + β2 × age + β3 × gender.

^dp-value for interaction (β4) from the equation: telomere length = α + β1 × U-As (per 100 μg/L) + β2 × age + β3 × gender + β4 × (U-As × < and > median %iAs/%MMA/%DMA).