TABLE 2.

Associations between lifestyle factors and fasting plasma total homocysteine concentration from a univariate and multivariate linear regression model in men and women from the training set of the pooled cohort (n = 1619)¹

	Univariate result2		Multivariate result3
Lifestyle factor	β in log(homocysteine), μmol/L (per category or 1 SD)4	P	β in log(homocysteine), μmol/L (per category or 1 SD)4	P
Smoking		<0.001		0.003
Never	0 (referent)		0 (referent)
Past	−0.020		−0.018
Current	0.068		0.062
Multivitamin use		<0.001		0.002
No	0 (referent)		0 (referent)
Yes	−0.084		−0.091
Caffeine intake,5 mg/d	0.027	<0.001	0.020	0.01
Alcohol intake, g/d	0.025	<0.001	0.024	<0.001
Dietary folate intake,5,6 μg/d	−0.026	<0.001	−0.016	0.01
Supplementary folate intake,5,6 μg/d	−0.046	<0.001	−0.019	<0.001
Total riboflavin intake,5 mg/d	−0.020	0.004	0.003	0.84
Total pyridoxine intake,5 mg/d	−0.020	0.003	−0.008	0.35
Total cyanocobalamin intake,5 μg/d	−0.024	<0.001	0.003	0.97
Betaine/choline intake,5 mg/d	−0.015	0.02	−0.008	0.75

¹Fasting plasma homocysteine concentrations were ln-transformed, corrected for batch variation, and adjusted for age, gender, and total calorie intake by the residual method. Median log(homocysteine) is 2.46 μmol/L (mean = 2.46, SD = 0.27, min = 1.57, and max = 3.63).

²Separate univariate models were run for each lifestyle factor.

³Intercept (SE) = 2.530342146 (0.03599243); model R² = 0.07.

⁴One increment is the SD; SD is 211mg/d for caffeine intake, 12 g/d for alcohol intake, 120 μg/d for dietary folate intake, 221 μg/d for supplementary folate intake, 12 mg/d for total riboflavin intake, 29 mg/d for total pyridoxin intake, 14 μg/d for total cyanocobalamin intake, and 83 mg/d for betaine/choline intake.

⁵Nutrient values are energy-adjusted intake.

⁶We included dietary folate and supplementary folate separately in the model because of differential bioavailability between the 2 sources of folate.