Systematic reviews identified for the update review

Susan L Norris; Susan Carson; Sujata Thakurta; Benjamin KS Chan

This publication is provided for historical reference only and the information may be out of date.

Appendix ESystematic reviews identified for the update review

Publication Details

Table

Publication Details

Copyright

Publisher

Oregon Health & Science University, Portland (OR)

NLM Citation

Norris SL, Carson S, Thakurta S, et al. Drug Class Review: Thiazolidinediones: Final Report Update 1 [Internet]. Portland (OR): Oregon Health & Science University; 2008 Aug. Appendix E, Systematic reviews identified for the update review.

Author Year Quality	Aims	Databases searched; Literature search dates; Other data sources	Eligibility criteria	Number of trials/Number of patients	Characteristics of identified articles: Study designs	Characteristics of identified articles: populations	Characteristics of identified articles: Interventions	Efficacy and effectiveness results	Subgroups	Adverse events	Comments
Berlie 2007	To obtain a precise estimate of the odds for developing TZD-induced edema and to compare rates between pio and rosi and with various combinations of oral agents	Medline (1966-5/2006), CINHAL (1982-5/2006), Cochrane Control Trials Register (to 1st quarter 2006), EMBASE (1996–2005) “manual search for review articles and original manuscript” (no details) Abstracts; ADA, AHA, ACC (2003-present) Takeda and GlaxoSmithKline were contacted for studies in their new Drug Applications	Prospective, RCTS, active- or placebo-controlled; monotherapy or combined therapy; data on edema	26 RCTs 15,332 subjects with DM2	7/26 were open label trials	All subjects with DM2 Average age range across studies: 53.7 to 61.9y Average duration diabetes across studies: 5.6m to 13.6y Mean baseline A1c across studies 7.5 to 10.2% with pio and 7.9 to 9.1% with rosi	Total daily dosage (mg) Monotherapy-placebo trials: pio 7.5–45, rosi 4–8 Combination trials: pio 15–30; rosi 4–8	A1c mean reduction; 0.56 – 2.3%	None	Weight gain (kg): pio −0.59 3.86; rosi 1.2 to 5.0 Pooled OR: All included studies (pio and rosi, all comparators): 2.26(95% CI, 2.02 – 2.53), P <0.000001 Placebo-controlled pio compared with rosi, indirect comparisons: 3.03(95% CI, 2.15 – 3.91) TZD monotherapy placebo-controlled studies: 2.35 (95% CI, 1.40 – 3.91) TZD combination therapy placebo- or no-treatment control: 2.14 (95% CI, 1.88 – 2.43)
Bolen 2007 (and AHRQ 2007 Review)	To summarize the literature on the benefits and harms of oral agents in the treatment of adults with type 2 diabetes	Medline, EMBASE, Cochrane Central Register of Controlled Trials (inception to 1/2006); industry data and FDA web-site; hand search 15 journals, reviewed reference lists	English-language, assessed benefits and/or harms of oral diabetes drugs (excluding 1st generation SU) compared to other oral agents (not insulin)	RCTs: 216 Systematic reviews: 2	Intermediate outcomes: 135 RCTs and 1 systematic review Final health outcomes: Adverse events: 167 studies, 1/3 of which were RCTs; 2 systematic reviews	Adults with type 2 diabetes	Various dosages and combinations	A1c: decreased 1% which was similar to SU and metformin HDL: pio increased more than rosi (1–2 mg/dL); pio increased compared with metformin or SU (3–5 mg/dL) LDL: rosi increased LDL more than pio (10–15 mg/dL) Triglycerides: pio decreased (15–52 mg/dL) compared with rosi (increase 6–13 mg/dL); pio decreased more than metformin	Age, sex, comorbidities	Mortality: insufficient data CV D mortality: insufficient data CHF: risk is increased with TZDs compared with other oral agents Microvascular outcomes: insufficient data Weight: TZDs increased weight similar to SU (3kg) as monotherapy or in combination with other oral agents; TZDs increased weight compared with metformin, acarbose, and repaglinide Hypoglycemia: less frequent with TZDs than SU (risk difference 4–9%) Edema: TZDs higher risk than SU (absolute risk difference 2 to 21%) CHF: TZDs higher risk than with metformin or SU, absolute risks 0.8 to 3.6%; absolute risk difference 0.7 to 2.2%) Mild anemia: TZDs higher risk than other drugs (absolute risk difference 1 to 5%) Elevated ALT: low rates (<1%) with TZDs Hospitalizations for acute cholecystitis: pio 12 patients compared with placebo 1 patient (pooled, unpublished analysis of 1526 patients)
Eurich 2007	To review literature on the association between antidiabetic agents and morbidity and mortality in people with heart failure and diabetes	Medline, Health-STAR, EMBASE, CINAHL, international Pharmaceutical Abstracts, Allied and Complementary Medicine, cochran Controlled Trials Register date of inception to 7/07 Manual search reference lists; contacted experts	Contemporaneous comparison group; examined association between antidiabetic agents and hospital admission or mortality	4 TZD trials 22,476 patients	Unclear from review; reviewing primary studies, 1 RCT and 3 cohort with comparison	Patients with diabetes and heart failure	TZD compared with a variety of comparators and placebo	NR	NR	Pooled OR for TZDs compared with other treatments for all cause mortality: 0.83 (95% CI, 0.71 – 0.97) Pooled OR for TZDs compared with other treatments on hospital admissions for heart failure 1.13 (95% CI, 0.1.04 – 1.22)	Pooled estimates included observational studies and RCTs AEs
Lagu 2007	To examine the risk of heart failure and of cardiac death in patients given TZDs	Medline, Database of Abstracts of Reviews of Effects, Cochrane Library; up to 3/2007; start date NR; databases of European Society of Cardiology, AHA, ACC, ADA by hand; reference lists	RCTs, double-blind studies with risk estimates or frequency data for congestive heart failure and cardiovascular death	7 trials 20191 patients	RCTs only	Diabetes and prediabetes	TZD compared with a variety of comparators and placebo	NR	NR	Risk of CHF compared to controls (placebo- and active-controlled trials): RR 1.72 (95% CI, 1.21 – 2.42), P =0.002; placebo-controlled trials only: RR 1.97 (95% CI, 0.94 – 4.13); pio only: RR 1.32 (95% CI, 1.04 – 1.68); rosi only: RR 2.18 (95% CI, 1.44 – 3.32), P =0.0003 Risk of cardiovascular death compared to controls: RR 0.93 (95% CI, 0.67 – 1.29), P =0.68; placebo-controlled trials only: RR 1.08 (95% CI, 0.66 – 1.76): pio only: RR 1.01 (95% CI, 0.51 – 2.09); rosi only: RR 0.91 (95% CI, 0.63 – 1.3) Pio: Overall event rate for CHF: TZD 2.3%; comparator group 1.4% NNH for CHF 107 over 29.7m F/U; range 35 to 491
Phatak, 2006	To examine factors affecting the size of A1c response to TZDs	PubMed, EBSCO, Sci-lit; dates NR GlaxoSmithKline public web-site	RCTs, English-language, placebo- and active-controlled	42 8322 subjects	RCTs only	Diabetes; mean age 57.5y; 42.3% female subjects; baseline A1c 8.9% (SD 0.8)	TZDs compared with a variety of comparators and placebo; 50% of studies were monotherapy; mean baseline A1c 9.1%(SD 1.0)	Weighted between-group change in A1c (all comparators) TZDs: −0.82% (SD 0.13) Pio: −1.04% (SD 0.07) Rosi: −0.67% (SD 0.10) Weighted between-roup change in A1c for placebo-controlled trials: Pio: −1.03 (SD 0.19) Rosi: −0.98 (SD 0.18) Change in A1c greater with higher baseline A1c (>9.0%) (no statistics) Duration of study treatment correlated with decrease in A1c (P =0.003)	Study duration, age, sex duration therapy examined with meta-regression	NR
Riche, 2007	To evaluate the impact of TZDs on repeat TVR after PCI	Medline, EMBASE, CINAHL, Cochrane database; through 7/2006 (start date NR); English only; abstracts from AHA, ACC, ADA searched 2001 – 2006	RCTS evaluating TZDs compared with standards of care; ≥ 6m follow-up; data provided on repeat TVR with number of patients receiving repeat TVR reported	7 608 subjects	RCTs only; all placebo-controlled with comparator standard drug therapy	1/7 studies non-diabetic; 1/7 metabolic syndrome; 5/7 type 2 diabetes	Pio or rosi given at various dosages either 1-day pre-operatively or 1–2 weeks post-operatively	Repeat TVR RR Overall: range 0.13 to 0.67; pooled estimate 0.35 (95% CI, 0.22 – 0.57) Pio: 0.24 (95% CI, 0.11 –0.51) Rosi: 0.45 (95% CI, 0.25 – 0.83) Diabetes: 0.34 (95% CI, 0.19 – 0.63) No diabetes: 0.37 (95% CI, 0.18 – 0.77)	Pio, rosi, diabetes, no diabetes	NR
Richter, 2007 (Rosi cochran)	To assess the effects of rosi in the treatment of type 2 diabetes	Medline, EMBASE, Cochrane database; last search 8/2006; reference lists searched	RCTs in adults with type 2 diabetes; study duration ≥ 24w	18 3888	RCTs only; various comparators	Type 2 diabetes, largely Caucasian populations; mostly persons on other oral hypoglycemic agents; mean age patients 47 – 61y; diabetes duration 4 – 9y; baseline A1c 6.8 – 9.5%, mean 8.8%	Rosi mono- or combined therapy at various dosages	Mortality: only reported by 1 study (ADOPT): 2.3% with rosi, 2.1% metformin, 2.2% glyburide A1c: similar reductions with rosi as metformin, glibenclamide, or glimepiride	NR	Edema: OR rosi compared with comparators, random effects model: 4.62 (95% CI, 2.28 – 9.38) Severe hypoglycemic episodes: somewhat lower with rosi than active monotherapy, particularly SU; no pooled data and no statistics From ADOPT trial only: CVD events: % serious/% total events: rosi 3.4/4.3; metformin 3.2/4.0; glyburide 1.8/2.8 CHF, total events (%): rosi 1.5, metformin 1.3, glyburide 0.6 Fracture rates: higher with rosi than; no statistics reported
Richter, 2006 (Pio cochran)	To assess the effects of pio in the treatment of type 2 diabetes	Medline, EMBASE, Cochrane database; last search 8/2006; reference lists searched	RCTs in adults with type 2 diabetes; study duration ≥ 24w	22 6200	RCTs only; various comparators	Type 2 diabetes, largely Caucasian populations; mostly persons on other oral hypoglycemic agents; mean age patients 53 – 63y; diabetes duration 3 – 14y; baseline A1c 7.4 – 10.3%	Pio mono- or combined therapy at various dosages	Mortality: only reported by 1 study (Proactive) as part of a composite endpoint (mortality, stroke, nonfatal MI, surgical vascular intervention: placebo-controlled, HR 0.90 (95% CI, 0.80 – 1.02) A1c: reductions similar to other oral agents	NR	7/22 trials reported AEs Overall and serious AEs comparable between intervention groups Hb: decrease noted in 6 studies examining this outcome: range 0.5 – 0.75 g/dL Weight: increased in 15 studies examining this outcome: up to 3.9kg Edema: RR 2.86 (95% CI, 1.14 – 3.18) Hypoglycemia episodes: Pio rates < SU rates; pio + insulin increased rates No pooled data and no statistics
Rosmarakis, 2007	To review RCT evidence on the effect of TZDs on in- stent restenosis after PCI	PubMed, last search 6/2006; reference lists reviewed	RCTs examining TZDs compared with various comparators and effect on in-stent restenosis after coronary stent implantation; in English	5 235 (text states 259)	RCTs only; various comparators	4/5 studies diabetes; 1/5 non diabetes	Pio or rosi; dosages and use of other oral agents NR; first dose of TZD given between 1d prior to procedure up to 2w after; 3/5 studies compared pio to standard treatment; 2/5 studies compared pio to rosi; all studies 6 month duration	Restenosis rate measured with quantitative coronary angiography at 6m: pio or rosi compared with standard therapy: OR 0.29 (95% CI, 0.15 – 0.56)	NR	Mortality: 2/259: 1 in control arm; 1 with TZD “No drug-related side effects”
Singh 2007 (Diabetes Care)	To evaluate the risk of CHF with TZDs in type 2 diabetes and to classify this AE under the dose-time-susceptibility system	RCTS: existing reviews; PubMed (1/2003 to 9/2006); manufacturer’s web-site Controlled observational studies and case reports: PubMed (to 9/2006, start date NR); Web of Knowledge Cited References and PubMed related articles Case reports also: EMBASE, Google Scholar (to 9/2006, start date NR)	Controlled observational studies with data to calculate OR of new onset CHF in patients receiving TZDs compared with other oral agents Case reports with CHF and TZD	RCTs: 3 10,731 Observational studies: 4 67,382 Case reports: 162 case subjects	RCTs, observational studies, case reports	Prediabetes or diabetes	TZD compared with placebo at various dosages	NR	NR	New onset CHF: RCTs: (3): OR 2.1 (95% CI, 1.08 – 4.08) Observational studies: (4): OR 1.55 (95% CI, 1.33 – 1.80) Case reports: 162 case subjects with 99 analyzable cases; median duration of onset of CHF 24w; CHF occurred in subjects <60y (26% of cases) and with low and high dosage
Singh 2007 (JAMA)	To review the long-term cardiovascular risks of rosi	Medline, GlaxoSmithKline clinical trials register, FDA website; product information sheets; last search 5/2007	RCTs in DM2 or IGT and trial duration ≥ 12m; provided numerical data on all AEs and monitored CVD AEs	4 14,291 Also 3 systematic reviews	RCTs, systematic reviews, meta- analyses; rosi compared with placebo or active oral agent	Prediabetes or type 2 diabetes	Rosi compared with active drug or placebo	NR	NR	Relative risk (95% CI) rosi compared with comparator: MI: 1.42 (1.06 – 1.91) Heart failure: 2.09 (1.52 – 2.88) CV mortality: 0.90 (0.63 – 1.26) Number needed to harm: MI: 822 per year with rosi if baseline risk 0.29% (low risk, ADOPT) CHF; 383 per year with rosi if baseline risk 0.24 (low risk, ADOPT)

: Abbreviations: ACC, American College of Cardiology; AEs, adverse events; ADA, American Diabetes Association; AHA, American Heart Association; CHF, congestive heart failure; CVD, cardiovascular disease; DM2, type 2 diabetes mellitus; m, month(s); MI, myocardial infarction; NR, not reported; NSD, no significant difference; OR, odds ratio; PCI, percutaneous coronary intervention; pio, pioglitazone; RCTs, randomized controlled trials; rosi, rosiglitazone; RR, relative risk; SU, sulfonylurea; TVR, target vessel revascularization; TZD, thiazolidinedione; w, week(s); y, year(s).