NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Shyangdan D, Clar C, Ghouri N, et al. Insulin Sensitisers in the Treatment of Non-Alcoholic Fatty Liver Disease: A Systematic Review. Southampton (UK): NIHR Journals Library; 2011 Nov. (Health Technology Assessment, No. 15.38.)

Cover of Insulin Sensitisers in the Treatment of Non-Alcoholic Fatty Liver Disease: A Systematic Review

Insulin Sensitisers in the Treatment of Non-Alcoholic Fatty Liver Disease: A Systematic Review.

Show details

Appendix 3Quality of included studies

TABLE 14Quality of included trials

StudyMethod of randomisationAllocation concealmentBlindingITT data analysisPercentage who completed trialPower calculationSimilarity of groups at baselineSponsorship/author affiliationTotal score
Aithal 2008101Computer randomisation in blocks of fourNot givenDouble blindNot givenI: 84%

C: 81%
Yes (requiring 33 in each group)Most parameters; lower alkaline phosphatase and fasting insulin in pioglitazone group at baselineUnclear, investigator-initiated study, but Takeda UK provided pioglitazone and placebo tablets; one author received funding by Takeda to attend meetings4
Belfort 2006102Computer generation by research pharmacyNot givenDouble blindNot givenPio: 23/26 (88%)

Pbo: 17/21 (81%)
NoYesGrants from the National Centre for Research Resources, Takeda Pharmaceuticals, and the Veterans Affairs Medical Research Fund

One of the authors is a member of the speakers bureau of Eli Lilly, one a consultant to Eli Lilly and the third a member of the advisory board and speakers bureau of Takeda Pharmaceuticals. No other conflict of interest
Sanyal 2004103By an independent statistician in random blocks of two, four and sixNoPathologist blindedNoPio: 8/10 (80%)

Vit E: 10/10 (100%)
NoYesSupported in part by a grant from the National Institutes of Health to the General Clinical Research Centre at Virginia Commonwealth University and by an award to AJS4
Sanyal 2010104Administered centrally by the DCC, web-based application; assigned participants in permuted blocks of treatments stratified by clinical centreRandomisation centrally, request made by a web based application and participants randomised in permuted blocksDouble blindYesPio: 70/80 (88%)

Vit E: 80/84 (95%)

Pbo: 72/83 (87%)
YesYesSupported by National Institute of Diabetes and Digestive and Kidney Disease

Additional funding by Takeda Pharmaceuticals North America Inc.
Bugianesi 200587Randomisation procedure was centralised in Bologna, and based on a random sequenceSealed envelopes were used to conceal randomisationBoth patients and investigator were not blind to the treatmentITT, LOCF analysisI: 100%

C: 100%
NoMost parameters; significant difference in sex between groups (100% men in Vit E group) and in cholesterol levels (significantly lower in diet group)Not given4
Garinis 2010105Not givenNot givenOpen labelNot givenI: 15/20(75%)

C: 25/25 (100%)

Five participants dropped out because of non-compliance to the treatment
Not givenTwo treatment groups showed similar clinical and biochemical variablesNot clear2
Haukeland 2009106Computer-assisted process of minimalisation (to minimalise baseline difference in serum ALT and liver histology)Allocation code was blinded to participants and investigators; two pathologists blindedDouble blindPer-protocol analyses (44 participants that completed the trial)I: 83%

C: 100%

Four participants treated with metformin did not complete the study due to gastrointestinal side effects (n=1), development of exanthema (n=1), withdrawal of consent (n=1) and erroneous use of study medication (n=1)
Yes, based on liver steatosis, at least 36 participants in each group to have 80% power to detect a significant difference between Met and placebo and 10% mean reduction; only 44 participants completed the trial, the power to detect a significant difference between treatment groups was reduced to 60%Metformin-treated participant were younger, less often treated for hypertension and slightly more obese (significance not reported)Eastern Norway Regional Health Authority (grant) and Merck Santé (delivery of study medication); no conflicts of interest7
Nadeau 2009107Research pharmacist. Stratified by ethnic group as well as insulin level >40 or <40 to assure similar distributions of drug and placebo groupsNoDouble blind; study radiologist blindedNoI: 28/37 (76%)

C: 10/13 (77%)
NoYesGrant by General Clinical Research Centers Program, National Centers for Research Resources, Insulin Control of Fat Regulation and Exercise in Teens and the Kettering Family Foundation4
Nar 2009108Not givenNot givenOperator performing ultrasound blind to treatmentNot given but presumably all patients completed the studyNot given, presumably 100%Not givenYesNot given3
Shields 2009109Randomised by the pharmacy using a computer-generated program into group A or BNoTwo study pathologists were blindedYesI: 7/10 (70%)

C: 9/9 (100%)
NoYes; two groups similar regarding their laboratory and anthropometric data; treatment group was older and predominantly male; (but difference not significant)No conflict of interest; funding source not reported5
Torres 2009100Not given; randomisedNot givenOpen label;

pathologist blinded
Not givenTotal number of participants allocated in each group is reported in the abstract as 108 till dateNot givenNot givenNot given1
Uygun 2004110Random sampling numbersNoUnblindedNoI: 15/17 (88%) completed in 6 months and 11/17 (65%) in 12 months; liver biopsies at study end by 13 only

C: 17/17 (100%); liver biopsies at study end by 10 only
NoYesNot given3
Ratziu 2008111Randomisation (pre-sealed envelopes) conducted by blocks of four and stratified on metformin usePre-sealed envelopesDouble blind

Pathologist blinded
NoI: 32/32 (100%)

C: 31/31 (100%)
The inclusion of 29 participants per arm was deemed necessary for rejecting the null hypothesis with 80% power and a type I error rate of 0.05. The total number of included participants was set at 32 per arm to account for 20% lost to follow-upYesGlaxoSmithKline partly funded the trial, investigator-initiated trial. One of the authors is a consultant to Astellas, Gilead, Pfizer, Sanofi-Aventis, and Trophos. Another author is a consultant for, and owns 15% of, BioPredictive, a company that markets FibroTest and SteatoTest. None of the authors has a personal conflict of interest with the manufacturer of any of the marketed TZDs7
Idilman 2008112Random assignment 1 : 2, method not reportedNot givenPathologist assessing liver biopsies blinded to patient and treatmentNo (but only one patient did not complete the study)Met: 100%

Rosi: 96%

C: 100%
Not givenYesStated that there was no conflict of interest with respect to funding3
Omer 2010113Not givenNot givenOpen labelNot givenMet: 86%

Rosi: 100%

Met + Rosi: 100%
Not givenAll three groups were similar at baseline except baseline serum insulin level (significantly higher in the metformin group and metformin-plus-rosiglitazone group compared with the rosiglitazone group)Committee of Ege University Medical School, Department of Gastroenterology and Endocrinology2

C, control; I, intervention; LOCF, last observation carried forward; Met, metformin; Pbo, placebo; Pio, pioglitazone; Rosi, rosiglitazone; Vit E, vitamin E.

© 2011, Crown Copyright.

Included under terms of UK Non-commercial Government License.

Bookshelf ID: NBK98345


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (995K)

Other titles in this collection

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...