d0006d.pdf

8/11/2019 D0006D.pdf

1/14

Current Diabetes Reviews, 2007, 3, 127-140 12

1573-3998/07 $50.00+.00 2007 Bentham Science Publishers Ltd.

Comparative Effectiveness of Pioglitazone and Rosiglitazone in Type 2 Diabetes, Pre-diabetes, and the Metabolic Syndrome: A Meta-Analysis

Susan L. Norris*, Susan Carson and Carol Roberts

Department of Medical Informatics and Clinical Epidemiology, Oregon Health & Science University, Portland, OR, USA

Abstract: Objective -To assess the comparative efficacy and safety of pioglitazone and rosiglitazone.

Research design and methods Multiple electronic databases were searched for randomized, controlled trials (RCTs) of efficacy or ef-fectiveness and for studies of any design which reported adverse events. Pooled estimates were calculated using a random effects model.

Results - Eighty-seven RCTs fulfilled our inclusion criteria for efficacy or effectiveness and 42 studies examined safety or tolerability.Two head-to-head RCTs of type 2 diabetes demonstrated significant improvements in A1c in both groups at follow-up with no significantdifference between groups; a third study found no significant change in A1c in either group. The pooled estimate of effect on A1c for

pioglitazone compared to placebo was -0.99% (95% confidence interval [CI], -1.18, -0.81) and for rosiglitazone was -0.92% (95% CI, -1.2, -0.64). Indirect comparison revealed no significant difference in A1c (between-drug difference -0.07% [95% CI, -0.41, 0.27]).Rosiglitazone increased total cholesterol compared to pioglitazone (net between-drug effect 13.91 mg/dl [95% CI, 1.20 to 26.62]). Bothdrugs increased weight by 2 to 3 kg and rates of adverse events were similar for the two drugs. Data were insufficient to assess compara-tive effects on health outcomes such as cardiovascular events.

Conclusions -Based largely on indirect evidence, the two thiazolidinediones appear to have similar effects on glycemic control and simi-lar side-effect profiles. Rosiglitazone may increase total cholesterol compared to pioglitazone. Studies are needed which provide direct

comparisons between the two drugs, particularly for long-term health outcomes.

Keywords:Type 2 diabetes, Oral hypoglycemic agents, Systematic review, Meta-analyisis.

There are two thiazolidinediones (TZDs) approved for prescrip-tion use in the United States, pioglitazone hydrochloride (Actos)and rosiglitazone maleate (Avandia). A third TZD (Troglita-zone) was removed from the market in 1999 due to adverse he-patic effects [1]. Both pioglitazone and rosiglitazone are approvedby the U.S. Food and Drug Administration for use in adults for the

treatment of type 2 diabetes, either as monotherapy, or in combina-tion with insulin, metformin, or sulfonylurea when diet, exerciseand a single agent does not result in adequate glycemic control

[2,3]. Neither drug is currently approved for use in pre-diabetes orin the metabolic syndrome.

The mechanisms of action of TZDs in lowering plasma glucose

among persons with type 2 diabetes are thought to include the fol-lowing: increase in insulin sensitivity, decrease in endogenous glu-

cose production and postprandial gluconeogenesis, suppression offree fatty acid release from the liver, increase in fasting and post-prandial glucose clearance, and beneficial effects on beta-cell func-

tion [4]. In addition to hypoglycemic effects, TZDs may have car-dioprotective effects that are independent of glucose lowering andmay be due to anti-oxidant, anti-inflammatory, or calcium channel-

blocking properties [5].

The objective of this systematic review was to examine the

comparative effect of pioglitazone and rosiglitazone on blood glu-cose, lipid concentrations, and blood pressure among persons withtype 2 diabetes, pre-diabetes, or the metabolic syndrome. Among

persons with diabetes, we examined the effect of these drugs on themacrovascular and microvascular complications of diabetes; among

persons with the metabolic syndrome or pre-diabetes, we examinedthe prevention and delay of onset of type 2 diabetes.

METHODS

Two independent reviewers identified potentially relevant titlesand abstracts from the Cochrane Central Register of Controlled

*Address correspondence to this author at 3181 SW Sam Jackson Park

Road, Mail Stop B1CC, Portland, OR 97239, USA; Tel: (503) 418-1432;Fax: (503) 494-4551; E-mail: [email protected]

Trials (3rd

Quarter 2005), Cochrane Database of Systematic Reviews (3

rdQuarter 2005), MEDLINE (1966 to July, Week 4 2005)

and EMBASE (3rd

Quarter 2005). Search terms included drugnames and indications and are available from the authors upon request. We reviewed reference lists of included studies and reviewsas well as dossiers submitted by pharmaceutical companies to the

Drug Effectiveness Review Project. Two independent reviewerachieved consensus on all included and excluded articles.

Study participants were adults with type 2 diabetes, prediabetes, or the metabolic syndrome (the latter defined using any othe widely accepted definitions) [6]. Included studies of type 2 dia

betes had to present one or more of the primary outcomes of thi

review: glycemic control either hemoglobin A1c (A1c) or fastingblood sugar; occurrence or progression of microvascular disease

(nephropathy, retinopathy, or neuropathy); occurrence or progression of macrovascular disease (cardiovascular disease, cerebravascular disease, or amputation); other complications of diabetes

mortality; or quality of life.

Published and unpublished English-language reports in any

geographic setting were included if they had a total sample size oten or more participants. For the assessment of efficacy and effectiveness, we included reports of randomized controlled trial

(RCTs) and controlled clinical trials. Head-to-head trials directlycomparing pioglitazone and rosiglitazone and trials comparing either of these drugs to placebo were included. As few placebo

controlled RCTs were available which examined population subgroups, we expanded our inclusion criteria to encompass all study

designs and trials where the comparator was another pharmacotherapeutic agent (active-control trials). For the assessment otolerability and adverse effects, we also examined all study designs

Data were abstracted into a standardized template in a relationadatabase (Microsoft Office, Access 2003).

We assessed the internal validity (quality) of controlled clinica

trials using predefined criteria based on the approach used by theU.S. Preventive Services Task Force [7] and the National HealthService Centre for Reviews and Dissemination [8]: randomization

method; allocation concealment; blinding of participants, investigators, and assessors of outcomes; the similarity of comparison

8/11/2019 D0006D.pdf

2/14

128 Current Diabetes Reviews, 2007, Vol. 3, No. 2 Norris et a

groups at baseline; adequate reporting of attrition; post-allocationexclusions; and the use of intention-to-treat analysis. These criteria

were then used to categorize studies into good, fair, and poor qual-ity studies. Studies were not excluded on the basis of poor quality

as there is a lack of empirical evidence for a relationship betweencriteria thought to measure validity and study outcomes [9]. A sen-sitivity analysis was performed comparing pooled estimates of good

or fair quality studies, to poor quality studies.

When there was a sufficient number of conceptually homoge-

neous studies, we performed a meta-analysis using the mean differ-ence between baseline and follow-up measures for the control andintervention groups and the standard error of each difference. A

pooled estimate of effect for each drug was calculated using therandom effects model proposed by DerSimonian and Laird [10].Review Manager (RevMan) was used for the meta-analysis (version

4.2 for Windows; Copenhagen: The Nordic Cochrane Centre, TheCochrane Collaboration, 2003).

Heterogeneity among trial results was tested using a standard

chi-squared test with a significance level of alpha=0.1, in view ofthe low power of such tests [9]. We also examined heterogeneity

among studies using I2

(the percentage of the variability in effectestimates that is due to heterogeneity rather than sampling error)[11].

An adjusted indirect comparison between pioglitazone androsiglitazone was made for A1c, lipids, and weight by using the

results of the meta-analyses comparing each drug with placebo[12]. Meta-regression was performed to determine whether thestudy-level characteristics of duration of the intervention and study

sponsorship (industry or private) affected the between-groupchange in A1c for placebo-controlled trials. For studies using acombination of a TZD and another hypoglycemic agent, we exam-

ined the effects of insulin, metformin or sulfonylurea on A1c. Forthe meta-regression we used STATA (version 9, StataCorp LP,College Station, Texas, USA).

In order to minimize publication bias [13] we sought unpub-lished data from pharmaceutical companies and consulted experts.

We did not utilize the technique of funnel plots to assess the poten-tial for publication bias as they can be misleading; there are many

causes of funnel plot asymmetry [13,14], and symmetry does notpreclude publication bias [15].

RESULTS

Type 2 Diabetes

A total of 2,173 titles and abstracts were identified by our

searches; of these, 87 RCTs fulfilled our inclusion criteria for effi-cacy or effectiveness and an additional 42 studies examined safetyor tolerability.

A1c improved to a significant degree with no between-groupdifferences in two fair-quality, head-to-head RCTs (in three publi-

cations) (Table 1) [16-18]. A third head-to-head study reported nosignificant change in A1c in either group when the study drugswere prescribed 2 weeks after discontinuation of troglitazone [19].

A total of 8,181 participants were examined in placebo-controlled, efficacy trials of pioglitazone [20-38] and 6,109 partici-pants with rosiglitazone [39-63] (Table 1). For pioglitazone, the

pooled effect estimate for net between-group change in A1c was -0.99% (95% CI, -1.18, -0.81) and for rosiglitazone was -0.92%

(95% confidence interval [CI], -1.2, -0.64). Similar pooled esti-mates were noted when studies were stratified by whether the TZDwas delivered as monotherapy or combined with another hypogly-

cemic agent (Table 2). Significant statistical heterogeneity wasnoted for all groups of studies examined for A1c (p0.002). Stud-ies of fair or good quality had similar results to those of poor qual-

ity for both TZDs. Indirect comparison of the two drugs revealed nosignificant difference in A1c (between-group difference -0.07%

[95% CI, -0.41, 0.27]). Data were insufficient to perform subgroupanalyses by different types of combination therapy.

Follow-up interval (p=0.537) and study quality (p=0.526) werenot associated with change in A1c after controlling for drug (piogli

tazone and rosiglitazone) in meta-regression. A significant interaction was detected between drug and funding source (p=0.0195) andtherefore each drug was examined separately. For pioglitazone

studies, a significant difference in change in A1c was found between studies funded by industry and those where funding sourc

was not reported (the crude mean change in A1c for the industrygroup was -1.17% and for the not-reported group -0.25%; between-group p=0.041). There were only two studies in the noreported group, however, so conclusions may not be robust. No

relationship between funder and A1c was found for rosiglitazonstudies.

Rosiglitazone increased total cholesterol (13.70 mg/dl [95% CI1.06, 26.35]) and pioglitazone decreased triglyceride levels (-1.08mg/dl [95% CI, -2.08, -0.09]) (Table 2). Using indirect compari

sons, rosiglitazone increased total cholesterol compared to pioglitazone (net between-drug effect 13.91 mg/dl [95% CI, 1.20 to26.62]). Rosiglitazone did not have a significant effect on blood

pressure compared to placebo; this outcome was not examined inplacebo-controlled studies of pioglitazone.

Two fair-quality studies demonstrated positive effects for cardiovascular outcomes among patients with type 2 diabetes andknown coronary artery disease [23,60]. Wang and colleagues [60

examined patients who had undergone a percutaneous coronaryintervention and found that coronary events (between-group

p

8/11/2019 D0006D.pdf

3/14

Comparative Effectiveness of Pioglitazone and Rosiglitazone Current Diabetes Reviews, 2007, Vol. 3, No. 2 12

Table 1. Study and Participant Characteristics for Head-to-Head and Placebo-Controlled Trials of Type 2 Diabetes

Study

[Reference]

Dosage

Mono- or Combination

Therapy

Total Sam-

ple Size

Follow-up

(weeks)

Age (years)

% Female

Other Population

Characteristics

Baseline*

Weight (kg)

BMI (kg/m2)

A1c (%)

Quality

Funder

Head-to-head trials

Derosa 2004

2005 [16, 17]

Pio: 15 mg qd

Rosi: 4 mg bid

Both groups received

glimepiride 4 mg qd

87 52 Pio: 53(6); Rosi 54(5)

Pio: 53%; Rosi 48%

Participants also had

metabolic syndrome

Pio: 68.9(3.5); Rosi: 67.8(3.1)

Pio: 24.4(0.8); Rosi 24.3(0.7)

Pio: 8.2(0.7); Rosi: 8.0(0.8)

Fair

NR

Goldberg 2005

[18]

Pio: 30-45 mg qd

Rosi: 4 mg qd-bid

Monotherapy

735 24 Pio: 55.9(10.5); Rosi:

56.3(11.3)

Pio: 54%; Rosi: 55%

Participants had un-

treated dyslipidemia

Pio: 93.7(20.6); Rosi 92.5(21.0)

Pio: 33.7(12.9); Rosi 32.6(6.6)

Pio: 7.6(1.2); Rosi: 7.5(1.2)

Fair

Study jointly funded

by Eli Lilly and Ta-

keda Pharmaceuticals

North America

Kahn 2002 [19] Pio: 15-45 mg qd

Rosi: 2 mg qd to 4 mg bid

Monotherapy, troglitazone

withdrawn

127 16 Pio: 57.8(11.0); Rosi:

57.1(12.1)

Pio: 48%; Rosi 55%

Pio: 101.4( 24.2); Rosi:

103.2(24.8)

Pio 35.2(7.4); 35.6(7.4)

Pio: 8.0(1.7); Rosi: 7.9(1.9)

Fair; open-label study

NR

Pioglitazone versus placebo

Aronoff 2000

[20]

7.5, 15, 30, 45 mg qd

Monotherapy

399 26 53.7(NR)

42%

(for all groups com-

bined)

90.4(13.1)

NR

10.4(2.0)

Poor

Takeda America

Dormandy 2005

[23], Charbonnel

2005 [22]

Titrated up to 45 mg qd

Combined with various

hypoglycemic agents

5238 156 (mean

34.5 months)

61.6(7.8)

34%

Evidence of macrovas-

cular disease

NR

31.0(4.8)

7.9(NR)

Good

Takeda Pharmaceuti-

cal company and Eli

Lilly and Company

Herz 2003 [24] 30, 45 mg qd

Monotherapy

297 16 58.0(10.7)

50.5%

Poorly controlled DM2

on diet only

86.3(17.4)

31.7(4.5)

7.5(NR)

Fair

Eli Lilly

Kipnes 2001 [25] 15, 30 mg qd

Added to SU

560 16 56.9(8.9)

42%

NR

32.0(4.9)

9.9(1.4)

Fair

Takeda Pharmaceuti-

cals

Mattoo 2005 [26] 30 mg qd

Combined with insulin

289 26 58.8(7.4)

57%

Using insulin for 3

months

NR

32.5(4.8)

8.9(1.3)

Fair

Eli Lilly and Takeda

McMahon 2005

[27]

45 mg qd

Used with insulin

16 12 52.5(NR)

11%

Using insulin

NR

32.3(4.1)

7.7(0.6)

Poor

Takeda, American

Heart Association,

NHLBI

Miyazaki 2002

[30]

7.5, 15 mg qd

Monotherapy

58 26 58.0(9.9)

73%

90(13.3)

32.8(5.3)

8.6(1.7)

Fair

Takeda

Miyazaki 2001

[28] 2004 [29]

45 mg qd

Added to SU

23 16 55(13.3)

45%

Generally healthy

82(16.6)

30(3.3)

8.2(1.0)

Data from 2004 (2001 baseline

data slightly different)

Poor

Takeda America (in

part)

8/11/2019 D0006D.pdf

4/14


Table 1. contd.

Study

[Reference]

Dosage


Therapy

Total Sample

Size

Follow-up

(weeks)

Age (years)

% Female

Other Population

Characteristics

Baseline*

Weight (kg)

BMI (kg/m2)

A1c (%)

Quality

Funder

Negro 2004 [31] 45 mg qd

Added to metformin

40 8 61.9(6.0)

NR

NR

26.7(2.4)7.7(0.6)

Poor

NR

Rosenblatt 2001

[32]

30 mg qd

Monotherapy

197 16 55.2(10.01)

43.8%

87.2(18.4)

30.7(5.0)

10.4(1.7)

Fair

Takeda Pharmaceuti-

cals

Rosenstock 2002

[33]

15, 30 mg qd

Added to insulin

566 16 56.7(9.4)

55%

95.4(17)

33.2(5.2)

9.8(0.1)

Fair

Takeda Pharmaceuti-

cals

Saad 2004 [34] 45 mg qd

Monotherapy

147 12 54(NR)

40%

NR

31(NR)

8.1(NR)

Fair

Funding NR; one

author affiliation

Novo-Nordisk Phar-

maceuticals, Prince-

ton, NJ

Scherbaum 2002

[35]

15, 30 mg qd

Monotherapy

235 26 59.1(NR)

44%

84.8(NR)

29.2(NR)

8.8(1.1)

Poor

Takeda Pharmaceuti-

cals, Europe

Smith 2005 [36]

Bogacka 2004

[21]

45 mg qd

Monotherapy

42 24 53.1(9.3)

53%

91.5(14.9)

31.9(5.0)

6.5(0.7)

Poor

Takeda Pharmaceuti-

cals, Inc, USA

Takagi 2003 [37] 30 mg qd

Combined with various

treatments

44 26 65(9)

50%

Known coronary heart

disease

NR

24.5(2.9)

6.7(1.2)

Poor

NR

Wallace 2004

[38]

45 mg qd

Monotherapy

30 12 62.6(10)

27%

Diet-controlled

85.2(4.3)

28.9(2.8)

6.7(0.9)

Fair

Takeda, UK

Mean (range) 7.5 to 45 mg qd 511 (16-5238)

Total: 8181

Median 16

(8 to 156)

Age: 57.6 (52.5-65.0)

range 0-57%

Weight: 88.1(82.0-95.4) kg

BMI: 30.6(24.5-33.2) kg/m2

A1c: 8.4(6.5-10.4)%

Quality:Good 1, Fair

8, Poor 7

Funder:

Industry: 11

Funding NR but 1 or

more authors list their

affiliation as industry:

1

Mixed: 2

NR: 2

Rosiglitazone versus placebo

Agarawal 2003

[39]

4 mg qd, 2 mg bid

Various SU

523 26 57.2(8.0)

33.5%

Normal renal function

(see subgroups for

renal-impaired)

NR

30.7(4.0)

9.2(1.4)

Fair (based on secon-

dary data)

NR

Barnett 2003 [40] 4 mg bid

Various SU

177 26 54.1(NR)

25%

Participants Indian

60%) Pakistani (27%)

NR

26.4

9.1(1.3)

Fair

SmithKlineBeecham

Pharmaceuticals

8/11/2019 D0006D.pdf

5/14


Table 1. contd.

Study

[Reference]

Dosage


Therapy

Total Sample

Size

Follow-up

(weeks)

Age (years)

% Female

Other Population

Characteristics

Baseline*

Weight (kg)

BMI (kg/m2)

A1c (%)

Quality

Funder

Fonseca 2000

[41]

4,8 mg qd

With metformin

349 26 58.8(9.2)

25.7%

NR

30.3(4.4)8.6(1.3)

Fair

SmithKline BeechamPharmaceuticals

Gomez-Perez

2002 [42]

2 mg bid, 4 mg bid

With metformin

105 26 53.4(7.5)

70.6%

NR

28.5(3.9)

9.8(NR)

Fair

NR; 3 authors (in-

cluding corresponding

author) from

GlaxoSmithKline

Hallsten 2002

[43]

4 mg bid

Monotherapy

28 26 57.7(7.1)

29%

Without complications

88.3(9.4)

NR

6.3(0.1)

Fair

Academy of Finland,

Novo Nordisk Foun-

dation, Finnish Diabe-

tes Research Society,

and GlaxoSmithKline

Honisett 2003

[44]

4 mg qd

Monotherapy

31 12 NR

100%

Postmenopausal

women

NR

NR

7.6(3.2) (rosi group)

Poor

NR

Iozzo 2003 [45] 8 mg qd

Monotherapy

19 26 57(6.3)

40%

No prior pharmacother-

apy for DM2

NR

31.5(4.7)

6.1(0.7)

Fair

GlaxoSmithKline

Jones 2003 [46] 4,8 mg qd

With metformin

217 26 60.2(9)

26%

BMI 25-30 (obese

presented in subgroups)

NR

27.7(1.4)

8.8(1.4)

Fair

NR; 3 of 4 authors

from GlaxoSmith-

Kline

Kim 2005 [47] 4 mg qd

Monotherapy

120 12 58.1(9.5)

63%

Taking metformin or

SU

62.3(11.0)

24.5(3.0)

9.3(1.3)

Fair

National R&D pro-

gram, Ministry of

Science Technology,

Republic of Korea

Lebovitz 2001

[48]

4,8 mg qd

Montherapy

493 26 59(10.9)

34%

NR

29.9(4.1)

9.0(1.7)

Poor

NR; 5 of 6 authors

from SmithKline

Beecham Pharmaceu-

ticals

Miyazaki 2001

[49]

8 mg qd

Monotherapy

29 12 56(2)

36%

87(18.7)

30.1(3.7)

8.3(1.5)

Fair

SmithKline Beecham

Natali 2004 [50] 8 mg qd

Monotherapy

46 (rosi and

placebo

groups)

8 58(9)

18%

NR

30.2(3.1)

7.6(0.8)

Fair

GlaxoSmithKline

Nolan 2000 [51] 4,8,12, mg qd

Monotherapy

369 8 62.3(9.5)

39%

81.3(14.5)

29.6(4.4)

NR

Fair

NR; 3 of 4 authors

from SmithKline

Beecham Pharmaceu-

ticals

8/11/2019 D0006D.pdf

6/14


Table 1. contd.

Study

[Reference]

Dosage


Therapy

Total Sample

Size

Follow-up

(weeks)

Age (years)

% Female

Other Population

Characteristics

Baseline*

Weight (kg)

BMI (kg/m2)

A1c (%)

Quality

Funder

Patel 1999 [52] 4 mg qd

Monotherapy

380 12 56.8(11.5)

31%

NR

28.9(4.0)9.1(NR)

Fair

Authors from Smith-Kline Beecham and

VA funding NR

Phillips 2001 [53] 2 bid, 4 qd, 4 bid, 8 qd

Monotherapy

908 26 56.8(9.2)

31%

NR

29.1(4.2)

8.9(1.5)

Fair

NR, author affilia-

tions include Smith-

Kline Beecham

Pharmaceuticals,

USA

Raskin 2001 [55] 2, 4 mg bid

With insulin

313 26 55.6(10.3)

44%

NR

32.7(4.5)

8.9(1.1)

Good

NR; individual

authors have received

support from Smith-

Kline Beecham

Raskin 2000 [54] 2,3,6, bid

With insulin

284 8 60.1(9.4)

40.6%

NR

30.4(4.2)

0.087(0.0163) (reference range

27 kg/m2

108.0(29)

36.3(2,5)

9.8(1.6)

Poor

Health management

Resources and

GlaxosmithKline

Tan 2005(a) [57] 4 mg bid

Monotherapy

24 12 52.3(10.1)

46%

No prior pharmacother-apy for DM2

NR

32.8(4.9)

7.5(1.0)

Fair

GlaxoSmithKline

van Wijk 2005

[58]

4 mg bid

Monotherapy

19 8 60(4.4)

26%

(for entire population at

baseline)

NR

29.2(4.8)

6.2(0.9)

Fair

GlaxoSmithKline

Virtanen 2003

[59]

4 mg bid qd

Monotherapy

28 (rosi and

placebo group

only)

26 58(7.5)

40%

88.3(9.7)

30.7(4.9)

6.3(0.4)

Fair

Academy of Finland,

Novo Nordisk Foun-

dation, Finnish Diabe-

tes Research Society,

and GlaxoSmithKline

Wang 2005 [60] 4 mg qdMonotherapy

70 26 62.2(8.6)20%

Coronary artery disease

after percutaneous

coronary intervention

NR25.6(2.7)

7.33(0.17)

FairMajor National Basic

Research Program of

PR China and Chi-

nese National Natural

Science Foundation

Wolffenbuttel

2000 [61]

1,2 mg bid

With various SU

574 26 61.9(9.1)

43%

Using SU for >6

months

NR

28.1(4.1)

9.2(1.3)

Fair

NR; One of 5 authors

from SmithKline-

Beecham

8/11/2019 D0006D.pdf

7/14


Table 1. contd.

Study

[Reference]

Dosage


Therapy

Total Sample

Size

Follow-up

(weeks)

Age (years)

% Female

Other Population

Characteristics

Baseline*

Weight (kg)

BMI (kg/m2)

A1c (%)

Quality

Funder

Yang 2002 [62] 4 mg qd

With various SU

64 26 57.8(8.9)

62%

65.3(11.2)

25.84(3.5)9.7(1.4)

Fair

Smith-Kline BeechamPharmaceuticals and a

grant from the De-

partment of Education

of the Republic of

China

Zhu 2003 [63] 2,4 mg bid

With various SU

530 24 58.9(7.7)

54%

Chinese, no hepatic

impairment

NR

25.1(2.8)

9.8(1.3)

Fair

SmithKlineBeecham

Research and Devel-

opment

Mean (range) Range: 4 to 12 mg qd Range: 18 to

908

Total 6109

Median 26

(8 to 26)

Age: 58 (52.3 to 62.3)

% Female: Range 20-

100%

Weight: 71.5(8.3 to 108) kg

BMI: 29.3 (24.5 to 36.3) kg/m2

A1c: 8.4 (6.1 to 9.8)%

Quality:Good 1, Fair

20, Poor 3

Funder:

Industry: 6FundingNR but 1 or

more authors list their

affiliation as industry:

10

Public: 2

Mixed: 4

NR: 3

*Participant characteristics are presented for the control group unless otherwise indicate; standard deviation is given in parentheses ( ); bid, two times daily

NR, not reported; qd, once daily; pio, pioglitazone; rosi, rosiglitazone; SU, sulfonylurea; DM2, type 2 diabetes

pioglitazone [69]. Data for pioglitazone, however, were not pre-sented, limiting conclusions that can be drawn.

The majority of RCTs were rated fair quality, however, 12% ofstudies of rosiglitazone and 44% of pioglitazone were rated poor

quality (Table 1). For both drugs combined, adequate randomiza-tion and allocation concealment were only reported in 15% and

12% of trials, respectively. In 40% of studies the baseline character-istics of the treatment groups were either not compared or not com-

parable (and no adjustment made for baseline differences). In 51%of studies attrition was not adequately reported, and, although most

studies were reported as double blind, in only 17% of studies wasit explicitly stated that the outcomes assessor or data analyst was

blinded to study group.

Studies were frequently funded by industry (Table 1): 69% forpioglitazone and 24% for rosiglitazone. In 40% of rosiglitazone

studies funding was not reported, but one or more of the studyauthors were identified as affiliated with a pharmaceutical com-

pany.

Metabolic Syndrome and Pre-Diabetes

Few data were available on the comparative effect of pioglita-zone and rosiglitazone on cardiovascular risk factors among per-sons with pre-diabetes or the metabolic syndrome. Pioglitazonedecreased LDL, total cholesterol, and triglycerides compared torosiglitazone (p

8/11/2019 D0006D.pdf

8/14


participants with type 2 diabetes or the metabolic syndrome [16],there were no significant differences in rates of adverse events or inchange body mass index (pioglitazone 15 mg, 1.2 kg/m

2; rosiglita-

zone 4 mg, 1.5 kg/m2) after 12 months of treatment. In pre-diabetes,

a head-to-head study [71] reported weight gain with pioglitazone(2.5 kg, SD 6.3), rosiglitazone (0.3 kg, SD 5.5), and the controlgroup (2.0 kg, SD 1.6) (p>0.05 for pioglitazone versus rosiglita-zone).

There were generally no differences in rates of adverse eventsbetween the active-treatment and placebo groups (Table 3), and themost frequently reported adverse events were edema, hypoglyce-

mia, and weight gain. Withdrawal rates due to adverse events intreatment arms ranged from 7% to 33% with pioglitazone [2327,34-36] and 0 to 27% with rosiglitazone [42-44,48-51,53,54,5859,62,63,70,72], and were not significantly different from placebo.

We identified no reports of macular edema in either trials or ob

servational studies. Elevations in alanine aminotransferase (>3

times the upper limit of normal) were rare in efficacy trials.

Both drugs increased weight compared to placebo: pioglitazon

2.96 kg (95% CI, 0.73, 5.20) and rosiglitazone 2.12 kg (95% CI

0.89,3.36),withnosignificantdifferencebetweenthetwodrugs

Table 2. Meta-Analysis Results for Type 2 Diabetes

DrugNumber of

StudiesTotal N

Weighted Mean Difference

(95% CI)

Test for Heteroge-

neity

(p-value)

Pioglitazone Versus

Rosiglitazone*

A1c (%) Pioglitazone

All studies 16 7219 -0.99(-1.18, -0.81)

8/11/2019 D0006D.pdf

9/14


Table 3. Adverse Effects of Pioglitazone and Rosiglitazone

Adverse Event Pioglitazone vs Placebo Rosiglitazone vs Placebo

Circulatory system

Cardiac-related events

Monotherapy

3.6% vs 6.3% [20] (cardiac adverse events, not speci-fied)

Combination therapy 6% vs 5% [25] (SU) (cardiac adverse events, not speci-

fied)

7.9% vs 7.0% [33] (Insulin) (cardiovascular adverse

events, not specified)

0.2% vs 0% (MI) [63] (SU)

3.9% (MI, bundle branch block and tachycardia) vs 2.9% (bundle

branch block) [42] (Met)

Congestive heart failure

Monotherapy

11% vs 8%* [23]

Combination therapy 1% vs 0% [33] (Insulin)

12.5% vs 0% [27] (Insulin)

Anemia

Combination therapy 1.9% vs 0.7% [39] (SU)

7.1%* vs 2.2% [46] (Met)

0.9% (4 mg), 0.9% (8 mg) vs 0% [41] (Met)

Peripheral edema

Monotherapy 0% vs 0% (15 mg), 3% vs 0% (30 mg) [35]

3.6% vs 0% [20]

5% vs 1% [32]

14%(30 mg), 16% (45 mg) vs 16% [24]

15.3% vs 7% [33]

22% vs 13% 2005 [23]

5.2% (4 mg), 6.4% (8 mg), 4.1% (2 mg bid), 6.6% (4 mg bid)* vs

1.6% [53]

Combination therapy

7%* vs 2% [25] (SU)

12.5% vs 0% [27] (Insulin)

14.1%* vs 3.4% [26] (Insulin)

Thrombocytopenia

Combination therapy 4.1% (4 mg); 7.7% (8 mg)* vs 3.6% [63] (SU)

Digestive system

Diarrhea, flatulence

Combination therapy 12.7% vs 15.6% [46] (Met)

7% vs 2% [40] (SU)

Liver function test abnormal(alanine aminotransferase >3

times upper limit normal)

Monotherapy 0.77% vs 1.3% [23]

1.3% (7.5 mg), 2.4% (30 mg) vs 1.3% [20]

0.44% vs 0% [52]

0% (2 mg), 0.6% (4 mg) vs 0% [48]

0% vs 0% [51]

0.14% vs 0% [53]

Combination therapy 0% vs 0% [25] (SU)

0% (15 mg), 0.5% (30 mg) vs 0% [33] (Insulin)

0% vs 0% [42] (Met)

0% vs 0.5% [39] (SU)

0% vs 0% [41] (Met)

8/11/2019 D0006D.pdf

10/14


Table 3. contd.

Adverse Event Pioglitazone vs Placebo Rosiglitazone vs Placebo

Endocrine system

Hyperglycemia

Combination therapy

1% vs 9%* [40] (SU)

9.3% (2 mg), 5.3% (4 mg) vs 17.2% [61] (SU)

Hypoglycemia

Monotherapy 0% vs 0% [32]

1.2% vs 0% [20]

10% (30 mg); 11% (45 mg) vs 11% [24]

28%* vs 20% [23]

Combination therapy 0% (15 mg), 3.8% (30 mg) vs

8/11/2019 D0006D.pdf

11/14


Table 3. contd.

Adverse Event

Pioglitazone vs Placebo

Rosiglitazone vs Placebo

Combination therapy 4.9% vs 7.9% [39] (SU)

6% vs 9% [40] (SU)

6.5% vs 8.9% [46] (Met)

Paresthesias


Vision abnormalities

Combination therapy 2.3% vs 0% [63] (SU)

Respiratory tract

Cough


Influenza-like symptoms

Monotherapy 2% (15 mg), 9% (30 mg) vs 8% [35]

Combination therapy

10% vs 14% [40] (SU)

Upper respiritary tract infection, rhinitis, sinusitis, bronchitis

Monotherapy 3% (15 mg), 4% (30 mg) vs 6% [35]

15.2% vs 11.4% [20]


8.6% vs 7.9% [39] (SU)

16%* vs 8.9% [46] (Met)

16.7% (4 mg); 10.0% (8 mg) vs 5.4% [63] (SU)

Urinary tract

Urinary tract infection

Monotherapy

6.7% (15 mg), 3.8% (30 mg) vs 7% [35]

Combination therapy 9.0% (4 mg); 10.9% (8 mg) vs 7.1% [63] (SU)

Other

Injury or accident

Monotherapy 2% vs 2% [23]

Combination therapy 0.9% (4 mg); 1.4% (8 mg) vs 5.4% [63]

6.6% vs 5.7% [39] (SU)

8% vs 7.6% [46] (Met)

*Indicates p-value vs placebo

8/11/2019 D0006D.pdf

12/14


cemia noted with combined therapies and no liver toxicity. Chi-quette et al. [87] reviewed placebo-controlled trials and concluded

that both drugs decreased A1c and increased weight to a similardegree. Pioglitazone lowered triglyceride levels (p

8/11/2019 D0006D.pdf

13/14


[23] Dormandy JA, Charbonnel B, Eckland DJA, et al.Secondary pre-vention of macrovascular events in patients with type 2 diabetes inthe PROactive Study (Prospective Pioglitazone Clinical Trial In

Macrovascular Events): a randomised controlled trial. Lancet 2005;366: 1279-89.

[24] Herz M, Johns D, Reviriego J, et al.A randomized, double-blind,placebo-controlled, clinical trial of the effects of pioglitazone onglycemic control and dyslipidemia in oral antihyperglycemic medi-cation-naive patients with type 2 diabetes mellitus. Clin Ther 2003;

25: 1074-95.

[25]

Kipnes MS, Krosnick A, Rendell MS, Egan JW, Mathisen AL,Schneider RL. Pioglitazone hydrochloride in combination with sul-fonylurea therapy improves glycemic control in patients with type2 diabetes mellitus: a randomized, placebo-controlled study. Am JMed 2001; 111: 10-7.

[26] Mattoo V, Eckland D, Widel M, et al.Metabolic effects of piogli-tazone in combination with insulin in patients with type 2 diabetesmellitus whose disease is not adequately controlled with insulintherapy: Results of a six-month, randomized, double-blind, pro-spective, multicenter, parallel-group study. Clin Ther 2005; 27:554-67.

[27]

McMahon GT, Plutzky J, Daher E, Bhattacharyya T, GrunbergerG, DiCarli MF. Effect of a peroxisome proliferator-activated recep-tor-(gamma) agonist on myocardial blood flow in type 2 diabetes.Diabetes Care 2005; 28: 1145-50.

[28] Miyazaki Y, Mahankali A, Matsuda M, et al. Improved glycemiccontrol and enhanced insulin sensitivity in type 2 diabetic subjects

treated with pioglitazone. Diabetes Care 2001; 24: 710-9.[29] Miyazaki Y, Mahankali A, Wajcberg E, Bajaj M, Mandarino LJ,

DeFronzo RA. Effect of pioglitazone on circulating adipocytokinelevels and insulin sensitivity in type 2 diabetic patients. J Clin En-

docrinol Metab 2004; 89: 4312-9.[30] Miyazaki Y, Matsuda M, DeFronzo RA. Dose-response effect of

pioglitazone on insulin sensitivity and insulin secretion in type 2diabetes. Diabetes Care 2002; 25: 517-23.

[31] Negro R, Dazzi D, Hassan H, Pezzarossa A. Pioglitazone reducesblood pressure in non-dipping diabetic patients. Minerva Endocri-

nol 2004; 29: 11-7.[32] Rosenblatt S, Miskin B, Glazer NB, Prince MJ, Robertson KE. The

impact of pioglitazone on glycemic control and atherogenic dyslip-idemia in patients with type 2 diabetes mellitus. Coron Artery Dis2001; 12: 413-23.

[33] Rosenstock J, Einhorn D, Hershon K, Glazer NB, Yu S. Efficacy

and safety of pioglitazone in type 2 diabetes: a randomised, pla-

cebo-controlled study in patients receiving stable insulin therapy.Int J Clin Pract 2002; 56: 251-7.

[34] Saad MF, Greco S, Osei K, et al.Ragaglitazar improves glycemiccontrol and lipid profile in type 2 diabetic subjects: a 12-week,double-blind, placebo-controlled dose-ranging study with an open

pioglitazone arm. Diabetes Care 2004; 27: 1324-9.[35] Scherbaum WA, Goke B, German Pioglitazone Study G. Metabolic

efficacy and safety of once-daily pioglitazone monotherapy in pa-tients with type 2 diabetes: a double-blind, placebo-controlledstudy. Horm Metab Res 2002; 34: 589-95.

[36] Smith SR, De Jonge L, Volaufova J, Li Y, Xie H, Bray GA. Effectof pioglitazone on body composition and energy expenditure: arandomized controlled trial. Metabolism 2005; 54: 24-32.

[37] Takagi T, Yamamuro A, Tamita K, et al. Pioglitazone reducesneointimal tissue proliferation after coronary stent implantation in

patients with type 2 diabetes mellitus: an intravascular ultrasoundscanning study. Am Heart J 2003; 146: E5.

[38]

Wallace TM, Levy JC, Matthews DR. An increase in insulin sensi-tivity and basal beta-cell function in diabetic subjects treated with

pioglitazone in a placebo-controlled randomized study. Diabet Med2004; 21: 568-76.

[39] Agrawal A, Sautter MC, Jones NP. Effects of rosiglitazone maleatewhen added to a sulfonylurea regimen in patients with type 2 dia-

betes mellitus and mild to moderate renal impairment: a post hocanalysis. Clin Ther 2003; 25: 2754-64.

[40] Barnett AH, Grant PJ, Hitman GA, et al.Rosiglitazone in type 2diabetes mellitus: an evaluation in British Indo-Asian patients.

Diabet Med 2003; 20: 387-93.[41] Fonseca V, Rosenstock J, Patwardhan R, Salzman A. Effect of

metformin and rosiglitazone combination therapy in patients withtype 2 diabetes mellitus: a randomized controlled trial. JAMA2000; 283: 1695-702.

[42] Gomez-Perez FJ, Fanghanel-Salmon G, Antonio Barbosa J, et a

Efficacy and safety of rosiglitazone plus metformin in Mexicanwith type 2 diabetes. Diabetes Metab Res Rev 2002; 18: 127-34.

[43] Hallsten K, Virtanen KA, Lonnqvist F, et al.Rosiglitazone but nometformin enhances insulin- and exercise-stimulated skeletal muscle glucose uptake in patients with newly diagnosed type 2 diabetes. Diabetes 2002; 51: 3479-85.

[44] Honisett SY, Stojanovska L, Sudhir K, Kingwell BA, Dawood TKomesaroff PA. Rosiglitazone lowers blood pressure and increase

arterial compliance in postmenopausal women with type 2 diabetes

Diabetes Care 2003; 26: 3194-5.[45] Iozzo P, Hallsten K, Oikonen V, et al. Effects of metformin an

rosiglitazone monotherapy on insulin-mediated hepatic glucose uptake and their relation to visceral fat in type 2 diabetes. DiabeteCare 2003; 26: 2069-74.

[46]

Jones TA, Sautter M, Van Gaal LF, Jones NP. Addition of rosiglitazone to metformin is most effective in obese, insulin-resistant patients with type 2 diabetes. Diabetes Obes Metab 2003; 5: 163-70.

[47] Kim YM, Cha BS, Kim DJ, et al.Predictive clinical parameters fotherapeutic efficacy of rosiglitazone in Korean type 2 diabetes mellitus. Diabetes Res Clin Pract 2005; 67: 43-52.

[48]

Lebovitz HE, Dole JF, Patwardhan R, Rappaport EB, Freed MIRosiglitazone Clinical Trials Study Group. Rosiglitazone monotherapy is effective in patients with type 2 diabetes. J Clin Endocrinol Metab 2001; 86: 280-8.

[49] Miyazaki Y, Glass L, Triplitt C, et al. Effect of rosiglitazone onglucose and non-esterified fatty acid metabolism in type II diabetic

patients. Diabetologia 2001; 44: 2210-9.[50] Natali A, Baldeweg S, Toschi E, et al.Vascular effects of improv

ing metabolic control with metformin or rosiglitazone in type 2diabetes. Diabetes Care 2004; 27: 1349-57.

[51] Nolan JJ, Jones NP, Patwardhan R, Deacon LF. Rosiglitazontaken once daily provides effective glycaemic control in patientwith Type 2 diabetes mellitus. Diabet Med 2000; 17: 287-94.

[52] Patel J, Anderson RJ, Rappaport EB. Rosiglitazone monotherapyimproves glycaemic control in patients with type 2 diabetes: twelve-week, randomized, placebo-controlled study. Diabetes Obe

Metab 1999; 1: 165-72.[53] Phillips LS, Grunberger G, Miller E, Patwardhan R, Rappaport EB

Salzman A. Once- and twice-daily dosing with rosiglitazone improves glycemic control in patients with type 2 diabetes. DiabeteCare 2001; 24: 308-15.

[54] Raskin P, Rappaport EB, Cole ST, Yan Y, Patwardhan R, Freed

MI. Rosiglitazone short-term monotherapy lowers fasting and post

prandial glucose in patients with type II diabetes. Diabetologi2000; 43: 278-84.

[55] Raskin P, Rendell M, Riddle MC, Dole JF, Freed MI, RosenstockJ. A randomized trial of rosiglitazone therapy in patients with inadequately controlled insulin-treated type 2 diabetes. Diabetes Car2001; 24: 1226-32.

[56] Reynolds LR, Konz EC, Frederich RC, Anderson JW. Rosiglitazone amplifies the benefits of lifestyle intervention measures inlong-standing type 2 diabetes mellitus. Diabetes Obes Metab 20024: 270-5.

[57] Tan GD, Fielding BA, Currie JM, et al.The effects of rosiglitazonon fatty acid and triglyceride metabolism in type 2 diabetes. Diabetologia 2005; 48: 83-95.

[58] van Wijk JP, de Koning EJ, Castro Cabezas M, Rabelink TJRosiglitazone improves postprandial triglyceride and free fatty acidmetabolism in type 2 diabetes. Diabetes Care 2005; 28: 844-9.

[59] Virtanen KA, Hallsten K, Parkkola R, et al.Differential effects o

rosiglitazone and metformin on adipose tissue distribution and glucose uptake in type 2 diabetic subjects. Diabetes 2003; 52: 283-90.

[60] Wang G, Wei J, Guan Y, Jin N, Mao J, Wang X. Peroxisome proliferator-activated receptor-gamma agonist rosiglitazone reduce

clinical inflammatory responses in type 2 diabetes with coronaryartery disease after coronary angioplasty. Metabolism 2005; 54590-7.

[61] Wolffenbuttel BH, Gomis R, Squatrito S, Jones NP, PatwardhanRN. Addition of low-dose rosiglitazone to sulphonylurea therapyimproves glycaemic control in type 2 diabetic patients. Diabet Me

2000; 17: 40-7.[62] Yang WS, Jeng CY, Wu TJ, et al.Synthetic peroxisome prolifera

tor-activated receptor-gamma agonist, rosiglitazone, increaseplasma levels of adiponectin in type 2 diabetic patients. DiabeteCare 2002; 25: 376-80.

8/11/2019 D0006D.pdf

14/14


[63] Zhu XX, Pan CY, Li GW, et al.Addition of rosiglitazone to exist-

ing sulfonylurea treatment in Chinese patients with type 2 diabetesand exposure to hepatitis B or C. Diabetes Technol Ther 2003; 5:33-42.

[64] Choi D, Kim S-K, Choi S-H, et al.Preventative effects of rosiglita-zone on restenosis after coronary stent implantation in patients withtype 2 diabetes. Diabetes Care 2004; 27: 2654-60.

[65] Kreider M, Heise M. Rosiglitazone in the management of older

patients with type 2 diabetes mellitus. Int J Clin Pract 2002; 56:538-41.

[66]

Rajagopalan R, Perez A, Ye Z, Khan M, Murray FT. Pioglitazoneis effective therapy for elderly patients with type 2 diabetes melli-tus. Drug Aging 2004; 21: 259-71.

[67] King AB, Armstrong DU, Chinnapongse S. Comparison of glyce-mic and lipid response to pioglitazone treatment in Mexican-Americans and non-Hispanic Caucasians with type 2 diabetes. Dia-

betes Care 2003; 26: 245-6.

[68] Vongthavaravat V, Wajchenberg BL, Waitman JN, et al.An inter-national study of the effects of rosiglitazone plus sulphonylurea in

patients with type 2 diabetes. Curr Med Res Opin 2002; 18: 456-61.

[69] Manley HJ, Allcock NM. Thiazolidinedione safety and efficacy in

ambulatory patients receiving hemodialysis. Pharmacotherapy2003; 23: 861-5.

[70] Wang TD, Chen WJ, Lin JW, Chen MF, Lee YT. Effects of rosigli-

tazone on endothelial function, C-reactive protein, and componentsof the metabolic syndrome in nondiabetic patients with the meta-

bolic syndrome. Am J Cardiol 2004; 93: 362-5.

[71] Durbin RJ. Thiazolidinedione therapy in the prevention/delay oftype 2 diabetes in patients with impaired glucose tolerance and in-sulin resistance. Diabetes Obes Metab 2004; 6: 280-5.

[72]

Hung YJ, Hsieh CH, Pei D, et al. Rosiglitazone improves insulin

sensitivity and glucose tolerance in subjects with impaired glucosetolerance. Clin Endocrinol 2005; 62: 85-91.

[73] Boyle PJ, King AB, Olansky L, et al. Effects of pioglitazone androsiglitazone on blood lipid levels and glycemic control in patientswith type 2 diabetes mellitus: a retrospective review of randomly

selected medical records. Clin Ther 2002; 24: 378-96.[74]

Delea TE, Edelsberg JS, Hagiwara M, Oster G, Phillips LS. Use of

thiazolidinediones and risk of heart failure in people with type 2

diabetes: a retrospective cohort study. Diabetes Care 2003; 26:2983-9.

[75] Frenchman IB. Treatment and treatment-related adverse events of

type 2 diabetes mellitus in residents of long-term care facilities: Aretrospective study. Curr Ther Res Clin Exp 2003; 64: 1-9.

[76]

Gegick CG, Altheimer MD. Thiazolidinediones: comparison of

long-term effects on glycemic control and cardiovascular risk fac-

tors. Curr Med Res Opin 2004; 20: 919-30.[77] Harmel AP. Treating diabetes: Cardiovascular benefits of antidia-

betes drugs. Am J Manag Care 2002; 8: S219-S28.[78] Hussein Z, Wentworth JM, Nankervis AJ, Proietto J, Colman PG.

Effectiveness and side effects of thiazolidinediones for type 2 dia-

betes: real-life experience from a tertiary hospital. Med J Aust2004; 181: 536-9.

[79] King AB. A comparison in a clinical setting of the efficacy andside effects of three thiazolidinediones. Diabetes Care 2000; 23557.

[80] King KA, Levi VE. Prevalence of edema in patients receivingcombination therapy with insulin and thiazolidinedione. Am Health Syst Pharm 2004; 61: 390-3.

[81] LaCivita KA, Villarreal G. Differences in lipid profiles of patientgiven rosiglitazone followed by pioglitazone. Curr Med Res Opin2002; 18: 363-70.

[82] Lebovitz HE, Kreider M, Freed MI. Evaluation of liver function in

type 2 diabetic patients during clinical trials: evidence that rosiglitazone does not cause hepatic dysfunction. Diabetes Care 2002; 25815-21.

[83] Olansky L, Marchetti A, Lau H. Multicenter retrospective assessment of thiazolidinedione monotherapy and combination therapy in

patients with type 2 diabetes: comparative subgroup analyses oglycemic control and blood lipid levels. Clin Ther 2003; 25 (SuppB): B64-80.

[84] Rajagopalan R, Iyer S, Perez A. Comparison of pioglitazone withother antidiabetic drugs for associated incidence of liver failure: noevidence of increased risk of liver failure with pioglitazone. Diabetes Obes Metab 2005; 7: 161-9.

[85] Tang WH, Francis GS, Hoogwerf BJ, Young JB. Fluid retentionafter initiation of thiazolidinedione therapy in diabetic patients withestablished chronic heart failure. J Am Coll Cardiol 2003; 411394-8.

[86] Boucher M, McAuley L, Brown A, Keely E, Skidmore B. Com

parative clinical and budget evaluations of rosiglitazone and pioglitazone with other anti-diabetic agents. Ottawa Canadian Coordinating Office for Health Technology Assessment Technology overview no 9, 20.

[87] Chiquette E, Ramirez G, Defronzo R. A meta-analysis comparingthe effect of thiazolidinediones on cardiovascular risk factors. ArchIntern Med 2004; 164: 2097-104.

[88] Czoski-Murray C, Warren E, Chilcott J, Beverley C, Psyllaki MACowan J. Clinical effectiveness and cost-effectiveness of pioglitazone and rosiglitazone in the treatment of type 2 diabetes: a sys

tematic review and economic evaluation. Health Technol Asses2004; 8: iii.

[89] Moher D, Fortin P, Jadad A, et al. Completeness of reporting trialpublished in languages other than English: implications for conducand reporting of systematic reviews. Lancet 1996; 347(8998): 3636.

[90] Moher D, Pham B, Klassen T, et al. What contributions do lan

guages other than English make on the results of meta-analyses? Clin Epidemiol 2000; 53(9): 964-72.

[91] Hollis S, Campbell F. What is meant by intention to treat analysis?Survey of published randomised controlled trials. BMJ 1999; 319670-74.

[92]

Chan A, Hrobjartsson A, Haahr M, Gotzsche P, Altman D. Empirical evidence for selective reporting of outcomes in randomized trials: Comparison of protocols to publish articles. JAMA 2004; 2912457-65.

[93] Glenny A, Altman D, Song F, et al. Indirect comparisons of competing interventions. . Health Technol Assess (Winchester, England) 2005; 9 (26): 1-134.

Received: 12 December, 2006 Revised: 19 January, 2007 Accepted: 30 January, 2007

d0006d.pdf

Documents