Table 5.

Assessment of confounding by population stratification for the association between CYP1A1/2 haplotypes and lung cancer among African Americans participating in the San Francisco Bay Area Lung Cancer Study, 1998-2003

Block	Haplotype	OR*	(95% CI)*	OR†	(95% CI)†	CRR‡
1	A	1.00	Ref	1.00	Ref
	B	1.10	(0.79-1.52)	1.11	(0.80-1.55)	0.98
	C	0.97	(0.62-1.52)	1.06	(0.67-1.70)	0.91
	D	1.10	(0.67-1.80)	1.09	(0.67-1.79)	1.01
	E	0.87	(0.52-1.43)	0.83	(0.50-1.38)	1.04
	F	0.72	(0.41-1.28)	0.74	(0.41-1.31)	0.98
	G	1.36	(0.74-2.52)	1.57	(0.83-2.98)	0.87§
			0.70		0.53
2	A	1.00	Ref	1.00	Ref
	B	1.02	(0.72-1.45)	0.99	(0.69-1.40)	1.03
	C	1.01	(0.72-1.42)	1.04	(0.73-1.48)	0.97
	D	1.40	(0.85-2.30)	1.38	(0.84-2.26)	1.02
	E	0.83	(0.48-1.44)	0.83	(0.47-1.44)	1.01
			0.65		0.66
3	A	1.00	Ref	1.00	Ref
	B	1.06	(0.81-1.39)	1.07	(0.81-1.41)	0.99
	C	1.21	(0.72-2.02)	1.20	(0.72-2.01)	1.00
			0.75		0.74
4	A	1.00	Ref	1.00	Ref
	B	0.90	(0.66-1.24)	0.87	(0.63-1.20)	1.04
	C	1.14	(0.82-1.59)	1.18	(0.84-1.65)	0.97
	D	0.70	(0.39-1.26)	0.70	(0.39-1.25)	1.01
			0.34		0.23

^*Adjusted for age, sex, and number of pack-years

^†Adjusted for age, sex, number of pack-years, and admixture (Amerindian and European ancestry)

^‡CRR = confounding risk ratio (pooled OR/ adjusted OR) (39)

Haplotypes correspond to haplotypes shown in Supplemental Figure 4 and in Supplemental Table 3. Rare haplotypes having frequencies < 5% are not included in the regression models.

Global tests for haplotype association are indicated in italics (comparing model with and without haplotypes).

^§Indicates CRR had a change of 10% or greater

Odds ratios and 95% confidence intervals are derived using 530 subjects (three individuals are missing number of pack-years and two individuals were unable to have their haplotypes estimated).