Table 2.

Associations between whole grain intake with risk factors for nonalcoholic fatty liver disease. Data are from meta-analyses only.

	Relative risk ratio/weighted mean difference compared to controls*	P value (after adjustment for potential confounders unless otherwise stated)	Median or average whole grain intake (high versus low; g whole grain/d)	Whole grains consumed	Study type	Number of cohorts/studies included in the meta-analysis	Reference
Cardiovascular disease (incidence)	0.79 (0.74, 0.85)	<0.001	44 versus 0	Mixed (mainly US studies)	Prospective cohort	9	[20]
	0.79 (0.73, 0.85)	<0.001	40 versus 3.2			7	[21]
Type 2 diabetes (incidence)	0.74 (0.69, 0.80)	<0.001	44 versus 0	Mixed (mainly US studies)	Prospective cohort	6	[20]
	0.79 (0.72, 0.87)	<0.001	32 versus 0+	Mixed (mainly US studies)	Prospective cohort	6	[22]
Fasting insulin (pmol/L)	−0.29 (−0.59, 0.01)	<0.001	>50 versus <20	Mixed	Intervention	10	[20]
	−0.011 (−0.015, −0.007)	<0.001	16 versus 0+	Mixed	Prospective cohort	14	[23]
Fasting glucose (mmol/L)	−0.93 (−1.65, −0.21)	<0.001	>50 versus <20	Mixed	Intervention	11	[20]
	−0.009 (−0.013, −0.005)	<0.001	16 versus 0+	Mixed	Prospective cohort	14	[23]
Total cholesterol (mmol/L)	−0.83 (−1.24, −0.42)	<0.001	>50 versus <20	Mixed	Intervention	16	[20]
LDL-cholesterol (mmol/L)	−0.72 (−1.34, −0.11)	<0.001	>50 versus <20	Mixed	Intervention	15	[20]
Weight gain (kg)	−0.18 (−0.54, 0.18)	ns	>50 versus <20	Mixed	Intervention	12	[20]
	−0.17 (−0.22, −0.11)	<0.001	16 versus 0+	Mixed (US cohorts)	Prospective cohort	3	[24]

*Highest versus lowest categories of whole grain intake in prospective cohort studies and weighted mean difference compared to controls in intervention studies.

**These values are not actual intake but are the difference intake estimated to lead to the corresponding change in biomarker concentration.