TABLE 7.

Relation between dietary choline and levels of steatosis and fibrosis¹

	Steatosis2		Fibrosis3
	Values	P	Values	P
Children 9–13 y old4	1.20 (0.37, 3.90)	0.76	0.64 (0.20, 2.04)	0.45
Males ≥14 y old5	0.68 (0.33, 1.38)	0.28	1.89 (0.94, 3.79)	0.07
Premenopausal women ≥19 y old5	1.57 (0.61, 4.06)	0.35	2.55 (1.00, 6.48)	0.05
Postmenopausal women5	0.88 (0.42, 1.86)	0.74	3.37 (1.58, 7.19)	0.002

¹All values are cumulative ORs; 95% CIs in parentheses. The cumulative OR of worse steatosis or fibrosis associated with a deficient daily choline intake (less than one-half the defined ADI) was assessed by using an ordinal logistic regression model with known contributors to NAFLD and NASH controlled for. ORs and P values were derived from multiple ordinal logistic regression models. Dietary choline was analyzed as a dichotomous variable, whereby choline values less than the deficient intake concentration were compared with values greater than or equal to deficient intake. Values of deficient choline intake are specific to each group. ADI, adequate dietary intake; NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis.

²Defined as <34%, 34–66%, and >66%.

³Defined as stage 0; stages 1a, 1b, and 1c combined; stage 2; and stages 3 and 4 combined.

⁴Age, race (white, Hispanic, or other), BMI z score, triglyceride concentration, hemoglobin A_1c, and daily caloric intake split at the median, and HOMA-IR >3.5 (mg · dL⁻¹ · μU · mL⁻¹ · 405⁻¹) (yes compared with no) were controlled for.

⁵Age, race (white, Hispanic, or other), BMI (kg/m²), waist circumference (cm), triglyceride concentration, hemoglobin A_1c and daily caloric intake split at the median, HOMA-IR >3.5 (mg · dL⁻¹ · μU · mL⁻¹ · 405⁻¹) (yes compared with no), alcohol use (yes compared with no), and steroid use (yes compared with no) were controlled for.