Table. Randomized Clinical Trials (RCTs) With Posttreatment Histological Features of the Liver Assessing Thiazolidinedione Therapy in Nonalcoholic Steatohepatitis (NASH) Included in the Meta-analysis.

Source	No. of Patients	Mean Age, y	Male, %	Mean BMI	Diabetes, %	Agent (Daily Dosage)	Trial Duration, mo	Comparator	BMI Change From Baseline, %	Quality Scorea
Ratziu et al,15 2008	63	54	59	31	31	Rosiglitazone maleate (8 mg)	12	Placebo	+1	7 (H)
Sanyal et al,16 2004	20	46	50	32	0	Pioglitazone hydrochloride (30 mg)	6	Vitamin E	0	7 (E)
Belfort et al,17 2006	55	51	45	34	48	Pioglitazone hydrochloride (45 mg)	6	Placebo	+2.7	7 (E)
Aithal et al,18 2008	74	54	61	31	0	Pioglitazone hydrochloride (30 mg)	12	Placebo	+3	7 (E)
Idilman et al,19 2008	74	47	59	32	0	Rosiglitazone maleate (8 mg)	12	Metformin hydrochloride, placebo	−2.6	4 (B, C, D, E)
Omer et al,20 2010	64	49	55	31	70	Rosiglitazone maleate (4 mg)	12	Metformin hydrochloride, metformin hydrochloride plus rosiglitazone maleate	0	4 (B, C, D, E)
Sanyal et al,21 2010	247	46	40	34	0	Pioglitazone hydrochloride (30 mg)	24	Vitamin E, placebo	+4.8	8
Cusi et al,12 2016	101	51	70	34	51	Pioglitazone hydrochloride (45 mg)	18	Placebo	+1	8

Abbreviation: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared).

^aThe Cochrane Risk of Bias Tool (score range, 0-8) score for RCTs is reported, with failing items in parentheses. Quality items of RCTs according to the Cochrane Risk of Bias Tool are as follows: A (adequate method of sequence generation), B (masking of participants performed), C (masking of personnel performed), D (masking of assessors performed), E (randomization concealment adequate), F (adequate assessment of each outcome), G (selective outcome reporting avoided), and H (intent-to-treat analysis of the results).