Non-invasive score identifies ultrasonography-diagnosed non-alcoholic fatty liver disease and predicts mortality in the USA

Table 5 Association between LFS and mortality

Mortality from:	Simple model^a			Full model^b			Full model + NFS adjustment^b
Mortality from:	HR	95% CI	Pvalue	HR	95% CI	Pvalue	HR	95% CI	Pvalue
All causes
Low (ref) versus high LFS	1.51	0.99 to 2.3	0.056	1.59	0.93 to 2.72	0.087	1.65	0.96 to 2.85	0.07
Int (ref) versus high LFS	1.52	1.03 to 2.23	0.034	1.60	1.01 to 2.54	0.048	1.61	1 to 2.58	0.049
LFS (continuous)	1.07	0.99 to 1.14	0.082	1.09	1.01 to 1.19	0.039	1.1	1.024 to 1.19	0.011
CVD
Low (ref) versus high LFS	2.29	1.24 to 4.26	0.009	2.24	1.03 to 4.88	0.042	2.30	1.02 to 5.19	0.046
Int (ref) versus high LFS	2.20	1.18 to 4.12	0.015	2.30	1.19 to 4.48	0.015	2.32	1.18 to 4.55	0.015
LFS (continuous)	1.11	1.04 to 1.19	0.002	1.11	1.03 to 1.19	0.006	1.13	1.05 to 1.21	<0.001
Liver disease
Low (ref) versus high LFS	9.80	2.01 to 47.62	0.006	31.25	3.13 to 333.33	0.004	46.45	5.13 to 420.51	<0.001
Int (ref) versus high LFS	17.24	2.6 to 111.11	0.004	30.30	4-250	0.001	32.99	4.57 to 237.95	<0.001
LFS (continuous)	1.25	1.14 to 1.36	<0.001	1.32	1.12 to 1.55	0.001	1.3	1.12 to 1.51	<0.001
Diabetes
Low (ref) versus high LFS	15.87	6.02 to 41.67	<0.001	2.87	0.65 to 12.66	0.161	3.13	0.67 to 14.56	0.143
Int (ref) versus high LFS	5.24	2.24 to 12.2	<0.001	1.98	0.71 to 5.52	0.185	2.04	0.72 to 5.79	0.176
LFS (continuous)	1.37	1.27 to 1.48	<0.001	1.21	1.02 to 1.44	0.034	1.21	1.03 to 1.42	0.024
Malignancy
Low (ref) versus high LFS	0.55	0.25 to 1.24	0.147	0.75	0.27 to 2.11	0.578	0.75	0.27 to 2.09	0.576
Int (ref) versus high LFS	0.65	0.29 to 1.44	0.281	0.76	0.32 to 1.79	0.519	0.76	0.32 to 1.79	0.518
LFS (continuous)	0.92	0.86 to 0.99	0.026	0.97	0.87 to 1.07	0.521	0.95	0.8 to 1.12	0.495

CVD, cardiovascular disease; Int, intermediate; LFS, liver fat score.
^aSimple model: adjusted for age and sex.
^bFull model: further adjusted for race/ethnicity, education, income, diabetes, hypertension, history of CVD, lipid-lowering medication, smoking status, waist circumference, alcohol consumption, caffeine consumption, total cholesterol, high-density lipoprotein cholesterol, transferrin saturation, and C-reactive protein; Because of smaller numbers of liver- and diabetes-related deaths, we did not adjust for education, income, history of CVD, lipid-lowering medication, and C-reactive protein in the full model.