Characteristic	Garber 2006	Goldberg 2005	Hanefeld 2007	Hällsten 2002	Jung 2005
Intervention 1 (I1) / intervention 2 (I2) / control 1 (C1)	I1: rosiglitazone + metformin C1: glibenclamide + metformin	I1: pioglitazone C1: rosiglitazone	I1: rosiglitazone 4 mg + placebo I2: rosiglitazone 8 mg + placebo C1: glibenclamide + placebo	I1: rosiglitazone I2: metformin C1: placebo	I1: rosiglitazone + glimipiride C1: metformin + glimipiride
Randomised controlled clinical trial (RCT)	Y	Y	Y	Y	Y
Non‐inferiority / equivalence trial	N	N	Y	N	N
Controlled clinical trial	N	N	N	N	N
Design: parallel study	Y	Y	Y	Y	Y
Design: crossover study	N	N	N	N	N
Design: factorial study	N	N	N	N	N
Crossover study: wash‐out phase	NA	NA	NA	NA	NA
Crossover study: carryover effect tested	NA	NA	NA	NA	NA
Crossover study: period effect tested	NA	NA	NA	NA	NA
Method of randomisation	?	stratified for being previously treated with oral antidiabetic drugs and according to sex	?	?	?
Unit of randomisation (individuals, cluster ‐ specify)	individuals	individuals	individuals	individuals	individuals
Randomisation stratified for centres	?	N	?	?	?
Randomisation ratio	1:1	1:1	1:1:1	1:1	1:1
Concealment of allocation	?	?	?	?	?
Stated blinding (open; single, double, triple blind)	double‐blind	double‐blind	double‐blind, double‐dummy	double‐blind	?
Actual blinding: participant	Y	?	Y	Y	NA
Actual blinding: caregiver / treatment administrator	?	?	?	?	NA
Actual blinding: outcome assessor	?	?	?	?	NA
Actual blinding: others	?	?	?	?	NA
Blinding checked: participant	N	N	N	N	NA
Blinding checked: caregiver / treatment administrator	N	N	N	N	NA
Primary endpoint defined (power calculation)	Y	Y	Y	N	N
[n] of primary endpoint(s)	1	1	1	?	?
[n] of secondary endpoints	7	16	13	?	?
Total [n] of endpoints	8	17	14	13	10
Prior publication of study design	N	N	N	N	N
Outcomes of prior/current publication identical	NA	NA	NA	NA	NA
Power calculation	Y	N	Y	N	N
[n] participants per group calculated	150	NA	?	NA	NA
Non‐inferiority trial: interval for equivalence specified	NA	NA	Y	NA	NA
Intention‐to‐treat analysis (ITT)	Y	Y	Y	?	?
Per‐protocol‐analysis	N	Y	?	?	?
ITT defined	Y	Y	Y	NA	NA
Missing data: last observation carried forward (LOCF)	?	Y	Y	N	N
Missing data: Other methods	N	N	N	N	N
LOCF defined	NA	N	N	NA	NA
Analysis stratified for centres	N	N	?	N	N
[n] of screened patients (I1 / I2/ C1/ total)	total: 356	total: 4410	total: 662	?	?
[n] of randomised participants (I1/ I2 / C1 / total) ‐ primary endpoint	I1: 158 C1: 160 total: 318	I1: 369; C1: 366; total: 735	I1: 200 I2: 191 C1: 207 total: 598	I1: 15 I2: 16 C1: 14 total: 45	I1: 15 C1: 15 total: 30
[n] of participants finishing the study (I1/ I2 / C1 / total)	I1: 133 C1: 131 total: 264	I1: 299 C1: 286 total: 585	I1: 153 I2: 158 C1: 173 total: 484	I1: 14 I2: 13 C1: 14 total: 41	I1: 14 C1: 13 total: 27
[n] of participants analysed (I1/ I2 / C1 / total) ‐ primary endpoint	I1: C1: total: 305	I1: 363 C1: 356 total: 719	I1: 195 I2: 189 C1: 202 total: 586 (ITT population)	I1: 14 I2: 13 C1: 14 total: 41	I1: 14 C1: 13 total: 27
Description of discontinuing participants	N	Y	Y	N	N
Drop‐outs (reasons explained)	Y	Y	?	N	Y
Withdrawals (reasons explained)	Y	Y	Y	Y	N
Losses‐to‐follow‐up (reasons explained)	?	Y	?	N	N
[n] of participants who discontinued (I1/ I2 / C1 / total)	I1: 25 C1: 29 total: 54	I1: 70 C1: 80 total: 150	I1: 47 I2: 33 C1: 34 total: 114	I1: 1 I2: 3 C1: 0 total: 4	I1: 1 C1: 2 total: 3
[%] discontinuation rate (I1/ I2 / C1 / total)	I1: 16 C1: 18 total: 17	I1: 19 C1: 22 total: 20	I1: 23.5 I2: 17.3 C1: 16.4 total: 19	I1: 7 I2: 19 C1: 0 total: 9	I1: 7 C1: 13 total: 10%
Discontinuation rate similar between groups	N	Y	N	?	Y
[%] crossover between groups	?	?	?	?	?
Differences [n] calculated to analysed patients	N	NA	?	NA	NA
Adjustment for multiple outcomes / repeated measurements	N	N	N	N	N
Baseline characteristics: Clinically relevant differences	Y 9% more men in I1 than C	N	N rosiglitazone 8 mg less male participants)	? HbA1c not included in baseline characteristics	? HbA1c and resistin not included in baseline characteristics
Treatment identical (apart from intervention)	Y	Y	Y	Y	Y
Compliance measured	N	N	N	Y	N
Other important covariates measured (specify)	N	N	N	N	N
Co‐morbidities measured	N	Y	N	N	N
Co‐medications measured	N	N	N	N	N
Specific doubts about study quality	N	N	N	N	N
Funding: commercial	Y	Y	Y	Y	N
Funding: non‐commercial	?	?	N	Y	Y
Publication status: peer review journal	N	Y	Y	Y	N
Publication status: journal supplement	N	N	N	N	N
Publication status: abstract	N	N	N	N	N
Publication status: other	N	N	N	N	N
Notes	commercial funding not explicitly stated but three of five authors from pharmaceutical company	no quantitative data on adverse events	sample size not reported, LOCF parameter unclear, baseline no of participants 587	patients received written diet instructions	poor reporting on quality criteria
Footnotes Y = yes; N = no; ? = unclear I = intervention; C = control; (baseline) = if numbers for certain features could ne be derived from the text, numbers from baseline characteristics were used