Table 3.

Linear regression analysis of the predictors of the nicotine metabolite ratio (NMR) in the intent-to-treat group (N=1155)

	Nicotine Metabolite Ratio (NMR) R2 = 0.065a; P<0.001
Predictor	B	Standard Error	β	P value	% Variationb
Ethnicityc	0.048 (0.053)	0.015 (0.012)	0.118 (0.130)	0.002 (<0.001)	0.81 (1.5)
Genderd,e	0.058 (0.058)	0.012 (0.012)	0.144 (0.144)	<0.001 (<0.001)	1.9 (1.9)
Birth control pill use	0.031 (0.030)	0.057 (0.057)	0.016 (0.015)	0.58 (0.60)	0.03 (0.02)
Hormone replacement therapy use	0.158 (0.158)	0.062 (0.062)	0.074 (0.073)	0.010 (0.011)	0.53 (0.53)
Body mass index	−0.003 (−0.003)	0.001 (0.001)	−0.087 (−0.088)	0.004 (0.003)	0.69 (0.71)
Alcohol use (# drinks/week)	0.002 (0.002)	0.001 (0.001)	0.056 (0.055)	0.062 (0.062)	0.29 (0.28)
Cigarettes	0.002 (0.002)	0.001 (0.001)	0.077 (0.077)	0.010 (0.011)	0.55 (0.53)
Menthol	0.007	0.015	0.018	0.627	0.02

Numbers in parentheses indicate values when the mentholated cigarette use variable is removed from the model to facilitate comparison to the total group in Table 2 where this variable is missing

^aTogether the predictors account for 6.5% of the variation in NMR with and without mentholated cigarette use in the model

^bCalculated by squaring the part correlation coefficient (not shown), and multiplying by 100

^cAfrican Americans and Caucasians were coded as ‘0’ and ‘1’, respectively, in the model

^dMales and females were coded as ‘0’ and ‘1’, respectively, in the model

^eWhen we restricted the model to females only (N=516), to further examine the effect of birth control pill and hormone replacement therapy use, the predictors (gender and menthol are excluded) together explained 5.3% of the variation in NMR. The standardized beta values for birth control pill and hormone replacement therapy use were 0.021 (P=0.63) and 0.10 (P=0.021), respectively, in the female-only group. They uniquely contributed 0.04% and 1% of the variation in NMR, respectively, in females.