Evidence Table 17Key Question 2: Systematic reviews

Study Characteristics
Author, Year
Quality
Aim(s) of ReviewEligibility Criteria
Characterstics of Included Studies
Number studies
Number of Patients
Characteristics of included populationsCharacteristics of InterventionsMain harms outcomes and results
Lago, 2007
Good
Systematic review and meta-analysis of pooled data from randomized trials of TZDs in subjects with prediabetes or type 2 diabetes to assess the risk of development of heart failure and death from cardiovascular causes in patients given TZDsEligibility: Randomized, double-blind, controlled trials of TZDs; Report of risk estimates, frequency data for congestive heart failure and cardiovascular death; trials with male human patients; written in English Excluded non-randomized clinical trials;No data for cardiovascular outcomes or death

Included Studies: All RCTs published since 2005; Follow-up with pts btwn 12 – 48 mos (mean 29.7 mos); Trial population range: 200–5269 participants, median 4351
7 trials; includes one trial with two control groups

20,191 Patients
Age: 59.2
Male %: 66.9%
BMI: 31 (5.0)
Baseline HbA1c: 7.72 (1.1)

Baseline Medical History
Hypertension: 50.4%
Hyperlipidaemia: 47.9%
Coronary artery disease: 18.6%
CHF: 20%
Chronic kidney disease or neuropathy: 2.4%
Daily TZD dosage:
Rosiglitazone vs. Placebo: 8mg
Rosiglitazone vs. Metformin and Sulf: 4 – 8mg
Rosiglitazone vs. Placebo: 4–8mg
Rosiglitazone vs.
Metformin/Rosiglitazone vs.
Glibeclamide: 4–8mg
Rosiglitazone vs. Placebo: 4–8mg
Pioglitzaone vs. glimepiride: 15–45mg
Pioglitazone vs. placebo: 15–45 mg
CHF Overall Risk, Risk Ratio (95% CI):
TZDs: 1.72 (1.21, 2.41), p=0.002
Rosi: 2.18 (1.44, 3.32)
Pioglitazone: 1.32 (1.04, 1.68)

CHF events
Rosi: 69
Pioglitazone: 145
Controls (Rosi trials): 35
Controls (Pioglitazone trials): 111

Comparison of risk of CHF and CV death: RR (95% CI)
Rosi Trials: 2.41 (1.61, 3.61)
Pioglitazone Trials: 1.32 (1.04, 1.68)
Total: 1.74 (0.97, 3.14)

Comparison of risk of CHF for Rosi and Pioglitazone:
RR (95%CI)
Rosi Trials: 1.01 (0.70, 1.45)
Pioglitazone Trials: 1.01 (0.49, 2.06)
Total: 1.01 (0.73, 1.40)
Mannucci, 2008
Fair
Meta-analysis of RCTs to assess whether pioglitzaone is also associated with increased cardiovascular riskEligibility: RCT; Pioglitazone vs. any other tx; Duration ≥ 4 weeks; Ongoing studies excluded.

Included Studies: RCT; 24 placebo; 19 insulin secretagogues; 8 metformin; 8 PPAR agonists; 1 insulin; 4 α-Glucosidase inhibitors; 4 DPP-IV inhibitors
Studies: 94 – 68 on pts w/ type 2 diabetes; 26 on pts w/ different conditions

Patients: 19, 803 type 2 diabetes pts (excluding PROACTIVE study) 25,041 type 2 diabetes pts (including PROACTIVE study)
Weighted mean age: 51.7 years; Mean diabetes duration: 6.6 yrs; mean HbA1c 8.1% (excluding PROACTIVE)Some combined therapy;
Some monotherapy
Non-fatal CV Events:
Type 2 Diabetes pts: (k=40)
Pioglitazone (n=4259): 44 events
Comparator ( n=3989): 50 events
RR 0.82 ( 0.55–1.23)

CHF:
PROACTIVE:
Pioglitazone associated with increased risk for CHF
Other Trials (k=40)
Pioglitazone (n = 5380): 58 cases
Comparator ( n= 4791): 39 cases
RR 1.32 (0.88–1.98)
Monami, 2008
Poor
Identification of moderators of the effects of rsoiglitazone on the risk of MI an dsevere heart failureEligibility: RCTs; Efficacy of rosiglitazone on type 2 diabetes; Comparator = any other ttreatment; Duration ≥ 4 wks; Type 2 diabetic pts; MI or CHF outcomes; Exclusion of ongoing trials

Included Studies: Duration range: 6–312 wks; Comparators: 52 Placebo; 8
Metformin; 15 Insulin secretagogues; 2
Pioglitazone; 3 Insulin; 12 None (?); 1
Multiple comparators
86 studies

30,003 Patients
Rosiglitazone: N = 16284
Comparators: N = 13719
NRSerious CHF (No. of Cases)
Rosi: 78
Comp: 47
MH-OR 1.59 (1.11–2.28)
< 6 mos: MH-OR 1.18 (0.89–1.63)
≥ 6 mos: MH-OR 1.21 (0.92–1.61)
Pinelli, 2008
Good
To provide a relative compariosn of the efficacy and safey of addint TZDs or exenatide to oral agents for the management of type 2 DM by performing meta- analyses of relevant published studies.Eligibility: Inclusion criteria: English; prospective, RCT; Placebo/Active comparator; ≥ 24 week duration; nonpregnant adults w/type 2 DM; full-text, peer reviewe articles examining efficacy of TZDs or exenatide in combination with other oral agents; reported HbA1c outcomes.
Exclusion criteria: assessed these agents as mnotherapy or adjunctive therapy to insulin-based regimens; open-label extension trials; interim analysis of Phase 3 clinical trials.

Included studies: 5 TZD open-label trials 2 exenatide open-label trials
Randomized; double blind (13) and triple- blind (2) trials
22 studies

9,325 patients
Mean age range: 53–61 Years; Mean baseline
HbA1c range: 7.5–9.9
8 TZD + metformin, sulfonylurea, or combined metformin/sulfonylurea vs. placebo-control trials 3 exenatide + metformin, sulfonylurea, or combined metformin/sulfonylurea vs. placebo-control trials 9 TZD vs. other glucose- lowering agents and open- label subcutaneous insulin 2 exenatide vs. other glucose- lowering agents and open- label subcutaneous insulinAdverse Events:
Severe Hypoglycemia:
Exenatide: 1 study
TZD: 4 studies
Nonsevere hypoglycemic events, TZD vs. other tx arms:
OR: 1.59 (0.76, 3.32)
Exenatide vs. placebo:
OR: 3.53 (0.92, 13.61)

Body Weight
TZD vs. comparator (n=6):
WMD: 1.51 kg (−0.12, 3.15)
Exenatide vs. comparator (n=5):
WMD: −2.74 kg (−4.85, −0.64)
Exenatide vs. placebo (n=3):
WMD: −1.29 kg (−2.22, −0.36)
Exenatide vs. insulin comparator (n=2):
WMD: −4.79 kg (−6.06, −3.52)

Nausea:
Exenatide:
OR: 9.02 (3.66, 22.23)
Vomiting:
Exenatide:
OR: 4.56 (3.13, 6.65)
Diarrhea:
Exenatide:
OR: 2.96 (2.05, 4.26)
Monami, 2009
Fair
Offer a comprehensive and updated synthesis of all available clinical data on safety and efficacy of GLP-1 receptor agonists.Eligibility: RCTs; Cross-ver of parallel series design; pts w/type 2 DM; Duration ≥ 12 weeks; Comparing GLP-1 receptor agonists w/placebo or active drugs; English

Included studies: 16 peer-reviewed publications; 5 unpublished trials; 6 open- label trials; duration of trial range (wks): 12–52
21 studies

8,482 patients
Mean age range: 53–61 years; Mean baseline HbA1c range: 7.0–8.9, Mean baseline BMI range: 23.9–36.09 Liraglutide studies; 12 exenatide studies; 12 placebo- controlled; 6 active comparator studies; 3 two comparator arms vs. placebo and active drugs;Hypoglycaemia (n= 15 trials):
Exenatide bid: 325 patients reported
Comparator: 109 patients reported
MH-OR 2.92 (1.49, 5.75), p = 0.002
Exenatide +Sulfonylurea: MH-OR 4.62 (1.89, 11.21), p = 0.001
Exenatide with or without sulfonylurea: MH-OR 1.37 (0.72, 2.63), p = 0.34
Exenatide vs. insulin: MH-OR 0.61 (0.33, 1.14), p = 0.125
Liraglutide: 78 patients reported
Comparator: 109 patients reported
Severe Hypoglycaemia :
Exenatide vs. Insulin:
MH-OR 0.74 (0.23, 2.39), p = 0.61
Nausea:
GLP-1 receptor agonists (17 trials): No. of cases
Interventional Drug (ID): 1354
Comparator (C): 230
MH-OR 3.88 (2.79, 5.42), p <0.001
Exenatide bid (10 trials): No. of cases
ID: 818
C: 133
MH-OR 8.38 (4.27, 16.48), p <0.001
Liraglutide (6 trials): No. of cases
ID: 522
C: 69
MH-OR 3.48 (2.29, 5.28), p <0.001

Vomiting:
GLP-1 receptor agonists (14 trials): No. of cases
Interventional Drug (ID): 365
Comparator (C): 56
MH-OR 4.23 (2.67, 6.13), p <0.001
Exenatide bid (9 trials): No. of cases
ID: 253
C: 42
MH-OR 4.54 (3.24, 6.38), p <0.001
Liraglutide (5 trials): No. of cases
ID: 108
C: 11
MH-OR 4.26 (1.01, 18.07), p = 0.049

Diarrhea:
GLP-1 receptor agonists (14 trials): No. of cases
Interventional Drug (ID): 396
Comparator (C): 88
MH-OR 2.36 (1.67, 3.33), p <0.001
Exenatide bid (9 trials): No. of cases
ID: 192
C: 49
MH-OR 2.56 (1.85, 3.54), p <0.001
Liraglutide (5 trials): No. of cases
ID: 204
C: 35
MH-OR 2.36 (1.67, 3.33), p <0.001
Richter, 2008
Good
To assess the effects of dipeptydyl peptidase-4 (DPP- 4) inhibitors for type 2 DMEligibilty: RCTs; Adults w/type 2 DM; tx for a minimum of 12 wks w/DPP-4 inhibitors alone or in combination; outcomes measuring HbA1c, adverse events, health- related quality of life

Included studies: Duration range: 12–52 wks; Most trials lasted 24 wks
25;
11 sitagliptin trials

6,743 patients
Most pts were inadequately controlled, either on diet, exercise or both or on metformin, glimepiride with or without metformin or pioglitazone treatment. Sex ration ws rougly balanced between intervention vs. control. Pts mostly white, obese, approx 55 yrs; duration of diabetes 3–5 yrs.6 Sitagliptin monotherapy vs. placebo;
2 Sitagliptin monotherapy vs. hypoglycaemic agent monotherapy;
sitagliptin combination vs. other combination therapies
Adverse Events: 11 trials, n = 12416
RR, (95%CI) 1.15 (1.02, 1.31)
Discontinuation due to AEs : 11 trials, n = 4414
RR, (95%CI) 1.05 (0.77, 1.43)
Serious AEs : 11 trials, n = 4413
RR, (95%CI) 0.97 (0.75, 1.27)
All-cause infections : 8 trials, n = 3589 RR, (95%CI) 1.29 (1.09, 1.52)

Change in body weight: 4 trials, n = 1259
Mean Difference [MD] 0.66 (0.37, 0.94)
Sitagliptin vs. placebo: 3 trials, n = 1109
MD: 0.69 (0.32, 1.06)
Sitagliptin vs. another agent: 1 trials, n = 150
MD: 0.6 (0.13, 1.07)
Nagajothi, 2008
Fair
Meta-analysis of RCTs comparing pioglitazone with either placebo or other oral hypoglycemic agentsEligibility: Randomized, drug-controlled or placebo-controlled trials evaluating pioglitazone; Reports MI as primary, secondary, or adverse outcome; Published data; English

Included Studies: Radomized double- blinded controlled trial;
Comparators: 2 Placebo; 1 metformin or gliclazide; 1 glyburide
1 glimepiride; Duration range: 6 months – 34.5 months
5 Studies

9,755 Patients
NRNRMI (no. of events/total no. of pts)
Pioglitazone: 143/4969
Control: 168/4996
RR (95%CI): 0.86 (0.69–1.07), p = 0.17

Stroke
Pioglitazone: 98/4692
Control: 126/4717
RR (95% CI): 0.79 (0.61–1.02), p = 0.07

Revascularization
Pioglitazone: 200/2861
Control: 264/2889
RR(95% CI): 0.40 (0.13–1.23), p = 0.11

Mortality
Pioglitazone: 185/4969
Control: 198/4996
RR (95%CI):0.94 (0.78–1.15), p =0.56

CV Mortality:
Pioglitazone: 130/4969
Control: 143/4996
RR (95%CI): 0.92 (0.73–1.16), p = 0.47
Phung, 2010
Good
To determine the comparative efficacy, risk of weight gain, and hypoglycemia associated with noninsuling antideiabetic drugs in patients with T2DM not controlled by metformin alonenclusion Criteria: Parallel-design RCTs; Compared noninsulin antidiabetic drugs with either placebo or another noninsulin antidiabetic drug in addition tometformin in all treatment groups; treated patients for ≥ 12 wks but ≤ 52 wks after randomization; only patients who showed inadequate response to stable metformin monotherapy; reported outcomes of HbA1c
Exclusion Criteria: evaluated the addition of more than 1 drug to metformin; participants were not considered to have inadequate response to a stable metformin monotherapy, participants were taking background threapies other than metformin; evaluated insulin.

27 RCTs; mean trial duration [range] 32 [12–52] wks;
N = 27

N = 11, 198
53–62 yrs; 23% – 75% mean; baseline HbA1c range 6.4% – 9.3%Compares classes of drugs: Sulfonylureas, glinides, TZDs, alpha-glucosidase inhibitors; dipeptidyl peptidase 4 inhibitors; GLP-1 Agonists
All studies had to have a mean metformin dose of enrolled patienta of at least 1500 mg/d during the study.
Comparison of noninsulin antidiabetic drugs with placebo, Traditional Meta-analysis:
Change in Body Weight, Kg, N = # of Trials; Weighted
Mean Difference (95% CI):
All drugs: N = 12; 0.14 (−1.37, 1.65)
Sulfonylurea: N = 2; 1.99 (0.86, 3.12)
Glinides: N = 2; 0.91 (0.35, 1.46)
TZDs: N = 1; 2.30 (1.70, 2.90)
AGIs: N = 1; −1.80 (−2.83, −0.77)

Comparison of noninsulin antidiabetic drugs with placebo, Mixed-treatment Meta-analysis:
Change in Body Weight, Kg, Weighted Mean Difference (95% CI):
Sulfonylurea: +2.06 (1.15, 2.96)
Glinides: +1.77 (0.46, 3.28)
TZDs: +2.08 (0.98, 3.17)
AGIs: −1.80 (−3.79, 0.21)

Comparison of noninsulin antidiabetic drugs with placebo, Traditional Meta-analysis:
Overall Hypoglycemia, N = # of Trials; Relative Risk (95% CI):
All drugs: N = 19; 1.43 (0.89, 2.30)
Sulfonylurea: N = 3; 2.63 (0.76, 9.13)
Glinides: N = 2; 7.92 (1.45, 43.21)
TZDs: N = 2; 2.04 (0.50, 8.23)
AGIs: N = 2; 0.60 (0.08, 4.55)
DPP-4 inhibitors: N = 8; 0.67 (0.30, 1.50)
GLP-1 analogs: N = 2; 0.94 (0.42, 2.12)

Comparison of noninsulin antidiabetic drugs with placebo, Mixed treatment Meta-analysis:
Overall Hypoglycemia, Relative Risk (95% CI):
Sulfonylurea: 4.57 (2.11, 11.45)
Glinides: 7.50 (2.12, 41.52)
TZDs: 0.56 (0.19, 1.69)
AGIs: 0.42 (0.01, 9.00)
DPP-4 inhibitors: 0.63 (0.26, 1.71)
GLP-1 analogs: 0.89 (0.22, 3.96)
Loke, 2009
Good
To determine systematically the reisk of fractures associated with thiazolidinedione therapy and to evaluate the effect of the therapy on bone densityRCT; Controlled observational studies; Comparsion of risk of fracture among pts with type 2 diabetes taking TZDs and those not taking TZDs; parallel-design trial; ≥ year duration; pts had impaired glucose tolerance or type 2 DM; use of either placebo or oral therapy as active comparator; Fracture outcomes reported

All trials were double blinded, Included particiapnts with impaired glucose and type 2 diabetes; duration range 1 – 4 years. Data on fractures by sex available for 5 trials.
N = 10

N = 13, 715
Pts in tx groups similar to control pts re: ethnic background, disease duration, HbA1c, BMITZD or controlFractures Overall, No of fractures:
TZD: 185/6122
Control: 186/793
OR (95% CI): 1.45 (1.18–1.79), p < 0.001

Fractures in Women, No. of Fractures:
TZD: 111/1903
Control: 76/2497
OR (95% CI): 2.23 (1.6–3.01), p<0.001

Fractures in men, No. of fractures:
TZD: 64/3064
Control: 95/3937
OR: 1.00 (0.73–1.39), p = 0.98

Difference between male and female subgroups: χ2 12.01, p <0.001

From: Evidence Tables

Cover of Drug Class Review: Newer Diabetes Medications, TZDs, and Combinations
Drug Class Review: Newer Diabetes Medications, TZDs, and Combinations: Final Original Report [Internet].
Jonas D, Van Scoyoc E, Gerrald K, et al.
Portland (OR): Oregon Health & Science University; 2011 Feb.
Copyright © 2011, Oregon Health & Science University.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.