Table 2.

Primary Outcome: Sustained Reduction in ALT Level by Treatment Group

	Vitamin E (n = 58)	Metformin (n = 57)	Placebo (n = 58)	P Valuea
				Vitamin E vs Placebo	Metformin vs Placebo
Sustained reduction in ALT level, No. (%) [95% CI]b	15 (26) [15 to 39]	9 (16) [7 to 28]	10 (17) [9 to 29]	.26	.83
Relative efficacy vs placebo, % (95% CI)c	50 (−36 to 206)	−9 (−149 to 109)
Change in ALT level from baseline, mean (95% CI), U/Ld
Week 24	−49.2 (−64.4 to −33.9)	−3.0 (−21.1 to 15.0)	−24.5 (−43.0 to −5.9)	.005	.09
Week 48	−44.5 (−60.3 to −28.7)	−11.7 (−45.3 to 22.0)	−25.0 (−43.7 to −6.4)	.04	.52
Week 72	−44.2 (−65.9 to −22.5)	−20.5 (−59.8 to 18.8)	−36.4 (−57.1 to −15.8)	.29	.51
Week 96	−48.3 (−66.8 to −29.8)	−41.7 (−62.9 to −20.5)	−35.2 (−56.9 to −13.5)	.07	.40

Abbreviations: ALT, alanine aminotransferase; CI, confidence interval.

^aBased on χ² test for binary outcomes or analysis-of-covariance model regressing change from baseline on treatment group and baseline value of the outcome for continuous outcomes.

^bSustained reduction defined as ALT ≤40 U/L or ≤0.5×baseline ALT level at 48-, 60-, 72-, 84-, and 96-week visits. Primary outcome was imputed as no sustained reduction if data were missing either at week 96 or at all 4 visits from week 48 to week 84; numbers imputed were 8, 6, and 9 for the vitamin E, metformin, and placebo groups, respectively.

^cRelative efficacy vs placebo=(1 − relative risk)× 100.

^dNumber of patients with complete data was similar across treatment groups at each follow-up visit and ranged from 45 to 52.