Evidence Table 4

Quality assessment of randomized controlled trials (New for Update 1)

AuthorYearRandomization method adequate/Allocation concealment adequate?Groups similar at baseline?CommentInclusion criteria specified?Exclusion criteria specified?Outcome assessors masked?Care provider masked?Patients masked?Attrition reported?Withdrawal Rate differential or high?CommentLoss to followup differential or high?CommentITT analysis?CommentPost- randomization exclusions?CommentFundingOther
Agarwal2005Method not describedYesYesYesYesUnable to determineNoNoYesNo1/22 (4.5%) pio vs 3/22 (13.6%)
glipizide
NoUnable to determinenumber analyzed not clear: 40/44?Yes2/44 discontinued from the study drug for unclear reason; 1 discontinued by physicianTakedaEnd point was blinded to personnel performing the technical and data analysis
Bakris2006Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindYesYes71/389 (18.3%) overall; 19.6% rosi vs 16.8% glyNoYes374/389 (96.1%) analyzedYes12/389 excluded for protocol deviation; 20/389 excluded for "other"GlaxoSmit hKline
Basu2006Method not describedMethod not describedYesYesYesNRNRNRYesNoNoNo2/21 subjects (9.5%) withdrew and were not analyzedNoTakeda
Belfort2006YesYesYesYesYesNRYesNRYesNoNoNo47/55 analyzed (85.5%)Yes2 withdrawan for CAD
Bhatt2007Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindNoUnable to determine85 of 200 did not have CIMT data, but # withdrawn not clear for other outcomes.Unable to determineUnable to determineall results have same sample size as baseline group; unclear if no attrition or if only completers analyzedUnable to determineGlaxoSmit hKline
Cao2006Method not describedMethod not describedYesOnly completers examined; more LAD in rosi group and more circumflex and RA lesions in control groupYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindYesYes63/360 (17.5%) did not complete followup.Unable to determineNoUnable to determineNot reported
Chappuis2007Method not describedMethod not describedYesYesYesNRNRNRNoUnable to determineUnable to determineUnable to determineUnable to determineSwiss Diabetes Foundatio
DREAM Trial Investigators2006YesYesYesYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindYesYes713+612 stopped drug or withdrew out of 5269 (23%); no differentialNoYesNo
Dailey2004Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindYesYes36/181 (19.9%) in rosi group vs 68/184 (37.0%) placeboNoYesLOCF, but high loss to followupYes5 in placebo group withdrawn for reason 'other'Bristol-Myers Squibb
Dargie2007YesMethod not describedNoHigher rate of NYHA class II and prior MI in rosi groupYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindYesYesNoYesITT population for placebo
110/114; rosi 108/110
Unable to determineGlaxoSmit hKline
DeRosa (Pharmacotherap y), 2006 (Diab, Obes, Metab)2005YesYesYesYesYesUnclear, reported as double blindUnclear, reported as double blindYesYesNoNoYes95/99 analyzed (96%)YesWithdrawal for protocol violation (1 or 2 patients only)Not reported
Derosa (J Int Med Research) 2006, 2006, 2007, 20072006aYesYesYesYesYesUnclear, reported as double blindUnclear, reported as double blindYesYesNoNoNoYesSome exclusions for protocol violationsNot reported
Garber2006Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindYesYesYesNoYesYes11 patients withdrawn for reason "other"Bristol- Myers Squibb
Gastialdelli2007Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindYesUnable to determineUnable to determineUnable to determineUnable to determine
Goldstein2006Method not describedMethod not describedYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindYesUnable to determineUnable to determineUnable to determineGlaxoSmit hKline
Hanefeld2007Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindYesYesYes114/598 (19.1%); higher for rosi 4 mg (23.5%)NoYes587/598 analyzed (98.2%) LOCFNoSmithKline Beecham
Hanefield, Betteridge 20052004Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindYesYesYesYesSmithKline Beecham
Heliovaara2007Method not describedMethod not describedYesHOMA and HDL cholesterol differedYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindNoUnable to determineUnable to determineUnable to determineUnable to determineEli LillyNo information on attrition, number analyzed not explicit.
Home, Home 20052007YesYesNoBackground met stratum had lower age, higher BMI, and shorter duration since diagnosis than background SU stratumYesYesYesNoNoYesYesYes10%YesYesGlaxoSmit hKline
Jain2006Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindYesYesYes240/502 (52.1%)Yes43/502 (8.6%)YesUsed LOCF but very high withdrawal rateYes61 excluded for patient noncompliance, protocol violaton, or investigator discretionNot reported
Kahn2006Method not describedMethod not describedYesYesYesYesUnclear, reported as double blindYesYesYes1733/4360 withdrew (39.7%)Yes234/4360 (5.4%)Yes4127/4360 analyzed (95%)YesGlaxoSmit hKline
Kulenovic2006Method not describedMethod not describedNoOlder in Gliben groupYesYesNRNRNRNoUnable to determineUnable to determineUnable to determineStates ITTUnable to determineNot reported
Lautamaki2005Method not describedMethod not describedNRReported for analyzed patients only (54/62)YesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindYesNoNoNoCompleters only analyzed (54/62)Yes3 excluded due to protocol violation, 1 due to unstable anginaFoundatio ns and GlaxoSmit hKline
Mazzone2006Method not describedMethod not describedNoMore hx MI in glimep than pioYesYesUnclear, reported as double blindUnclear, reported as double blindYesYesNoYes458/462 analyzed (99.1%)Yes1 protocol violation, 3 investigator decisionTakeda
Negro2005Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindNoUnable to determineUnable to determineUnable to determineUnable to determineNot reported
Nishio2006Method not describedMethod not describedYesYesYesNoNRNRNoUnable to determineUnable to determineYesNumber analyzed equals total NUnable to determineNot reported
Osman2004Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindYesUnable to determineUnclear for A1c, 11/16 did not consent to f/u angio for CIMTUnable to determineNoUnable to determineNIH
Perriello2006Method not describedMethod not describedYesYesNoUnclear, reported as double blindUnclear, reported as double blindYesYesNoNoYes275/283 analyzed (97.2%)Yes275/283 analyzed (97.2%)Not reported
Pfutzner2006Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindNoUnable to determineUnable to determineUnable to determineUnable to determine
Pfutzner, Forst 2005, Forst 20052005Method not describedMethod not describedNRReported for per-protocol population only (173/192 enrolled)YesYesNoNoNoYesNo90.5% completedUnable to determineYes173/179 (96.6%) analyzedUnable to determineTakeda
Raskin2004Method not describedMethod not describedYesYesYesNoNoNoYesYes71/252 (28.2%). Lower d/c rate for combination vs montx groups; 40% in monotx groupsUnable to determineYesYes8 excluded for non- complianceNovo Nordisk
Roden (Trial 1)2005Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindNoUnable to determineUnable to determineNo1098/1194 analyzed (92%)Unable to determineEli Lilly and Takeda
Roden (Trial 2)2005Method not describedMethod not describedYesYesYesNRNRNRNoUnable to determineUnable to determineYes606/639 analyzed (94.8%)Unable to determineEli Lilly and Takeda
Rosenstock2006Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindYesYesCompleters: Intervention 90/116 (78%), placebo 57/111 (51%)NoYesYes11/227 for protocol violationGlaxoSmit hKline
Sharma2006YesYesYesYesYesUnclear, reported as double blindUnclear, reported as double blindYesYesNoYes8.6% overall; 1/18 (5.5%) met vs 2/17 pio (11.8%)No30/35 analyzed (85.7%)No
Sourij2006Method not describedMethod not describedYesYesYesUnclear, reported as double blindUnclear, reported as double blindYesNoUnable to determineUnable to determineUnable to determineUnable to determineTakeda AustriaUnable to determine number analyzed
Stocker2007YesMethod not describedYesYesYesNoNoNoYesYesNoNo75/92 analyzed (81.5%)NoGlaxoSmit
Teramoto2007Method not describedMethod not describedYesYesYesNoNoNoYesNoUnable to determineExact loss to F/U NRYes91/92 analyzed (98.9%)Unable to determineJapan Pioglitazo ne Study Group (Type unclear)
Tseng2005Method not describedMethod not describedYesNoNoUnclear, reported as double blindUnclear, reported as double blindUnclear, reported as double blindNoUnable to determineUnable to determineUnable to determineResults reported for 48/48 randomized; unclear if any withdrawalsUnable to determineNot reported
Umpierrez2006Method not describedMethod not describedNoYesYesNoNoNoYesNo29/210 (13.8%)Yes4/101 glimepiride (4%) vs 15/109 pio (13.8%)Yes203/210 analyzed (96.7%)YesSanofi-Aventis
Yamanouchi2005YesYesYesYesYesNRNRNRYesNo8/114 (7%)No3/114 (2.6%)Noonly completers analyzed=93%NoNot reported

From: Evidence Tables

Cover of Drug Class Review: Thiazolidinediones
Drug Class Review: Thiazolidinediones: Final Report Update 1 [Internet].
Norris SL, Carson S, Thakurta S, et al.
Portland (OR): Oregon Health & Science University; 2008 Aug.
Copyright © 2008, Oregon Health & Science University, Portland, Oregon.

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