Study | Medication studied | No. of patients | Study period | Mean length of follow-up, yr | Covariates | ||
---|---|---|---|---|---|---|---|
Exposed group | Comparison group | Exposed group | Comparison group | ||||
RCT | |||||||
Dormandy et al., 2005 (PROactive study, multicentre)7 | Pioglitazone | No TZD use | 2 605 | 2 633 | 2001–2004 | 2.9 | NA |
Kahn et al., 2006 (ADOPT, multicentre)16 | Rosiglitazone (monotherapy) | No TZD use (metformin or glibenclamide monotherapy) | 1 456 | 2 895 | 2000–2006 | 3.4 | NA |
Home et al., 2009 (RECORD, multicentre)17 | Rosiglitazone (+ sulfonylurea or metformin) | No TZD use (sulfonylurea and metformin) | 2 220 | 2 227 | 2001–2008 | 5.5 | NA |
Sanofi-Aventis 2009 (multicentre, United States)21 | TZD (unspecified) (+ insulin glargine and sulfonylurea or metformin) | No TZD use (insulin glargine, metformin and sulfonylurea) | 256 | 130 | 2006–2008 | NR§ | NA |
Cohort | |||||||
Oliveria et al., 2008 (United States)18* | TZD (unspecified) | No TZD use | NR | NR | 2000–2004 | 3.9 | Yes¶ |
Lewis et al., 2011 (Kaiser Permanente, California, United States)10 | Pioglitazone | No pioglitazone use | 30 173 | 162 926 | 1997–2008 | 3.3 (exposed group); 6.2 (comparison group) | Yes** |
Tseng, 2011 (National Health Insurance, Taiwan)11 | Pioglitazone and/or rosiglitazone | No TZD use | 1 028 | 112 520 | 2003–2005 | 3.0 | Yes†† |
Tseng, 2012 (National Health Insurance, Taiwan)20 | Pioglitazone | No pioglitazone use | 2 545 | 52 383 | 2006–2009 | NR | Yes‡‡ |
Neumann et al., 2012 (National Health Insurance, France)19† | Pioglitazone | No pioglitazone use | 155 535 | 1 335 525 | 2006–2009 | 2.4 (pioglitazone exposed group) | Yes§§ |
Rosiglitazone | No rosiglitazone use | 153 334 | 1 337 726 | ||||
Case–control | |||||||
Piccinni et al., 2011 (FDA Adverse Event Reporting System, United States)9‡ | Pioglitazone | No pioglitazone use | 37 841 | 561 244 | 2004–2009 | NA | NR |
Note: ADOPT = A Diabetes Outcome Progression Trial, FDA = US Food and Drug Administration, NA = not applicable, NR = not reported, PROactive trial = PROspective pioglitAzone Clinical Trial in macroVascular Events, RCT = randomized controlled trial, RECORD = Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycemia in Diabetes, TZD = thiazolidinedione.
↵* The total number of patients in this study was 191 223.
↵† The total number of patients in this study was 1 491 060.
↵‡ Case/noncase study design.
↵§ The duration of the trial was 12 weeks, and 10% of patients were lost to follow-up.
↵¶ Age, sex, risk factors for bladder cancer (pelvic radiation, schistosomiasis). concentration, smoking status, history of bladder conditions (urinary tract
↵** Age, sex, race/ethnicity, incident diabetes at baseline, baseline hemoglobin A1C conditions, urolithiasis, incontinence and “other bladder or urethral conditions”), comorbidities (congestive heart failure, renal function, other cancer before baseline), use of other glucose-lowering therapies (metformin, sulfonylureas, other TZDs, other oral glucose-lowering drugs, insulin), income.
↵†† Age, sex, duration of diabetes, comorbidities (urinary tract disease, nephropathy, hypertension, chronic obstructive pulmonary disease, stroke, ischemic heart disease, peripheral arterial disease, eye disease, dyslipidemia), use of glucose-lowering agents (metformin, sulfonylureas, acarabose, TZDs, insulin), other medications (statins, fibrates, angiotensin-converting-enzyme inhibitors and/or angiotensin-receptor blockers, calcium-channel blockers), living region, occupation.
↵‡‡ Same covariates as in Tseng 2011 study, as well as heart failure, use of meglinitide, use of rosiglitazone (instead of “TZDs”) and other cancers before baseline (excluding bladder cancer).
↵§§ Age, sex, use of other glucose-lowering agents.