Table 19

Studies examining subgroups based on demographic characteristics or comorbidities

Author, Year QualityCountry SettingStudy designRace/ethnicityConcurrent hypoglycemic treatmentInclusion criteria Exclusion criteriaMean age(SD) GenderBaseline A1c (SD) Weight (SD) or BMI (SD)A1c outcomesAdverse events and tolerability
Jun JK 2003

Fair, for case series
Single center
Time series retrospective chart reviewHispanic 100%SU 50%
Insulin 52%
Metformin 70%
Hispanic, >18 y, DM2, uncontrolled hyperglycemia with A1c≥8.0%; have taken Pio for at least 6m; A1c within 1m before start of Pio; have at least 2 A1c measures at 3-m intervals during the 6-m period; lipid panel within 1m before start
Exclusion criteria noncompliant with Pio as noted in chart:
54.6 (8.5) yr
83% female
10.4% (1.7%)

78.9 (21.4) kg

32.0 (8.1) kg/m2
6-month follow-up A1c: −2.0% (P<0.0001)8 patients (5.6%) withdrew secondary to significant peripheral edema; 1 patient had exacerbation of congestive heart failure, 1 reported myalgias.
King AB 2003

Fair(for cohor study)
Single center
Cohort with comparison group
Retrospective chart review
98 non-Hispanic
Caucasians and 81 Mexican-Americans
SU 55%
Insulin 0%
Metformin 21%
Clinic patients with DM2, treated with Pio 45 mg/d for 6m or more without interruption; A1c and lipids available on the chart within 4w of starting treatment and approximately 4m into treatment
Exclusion criteria: patients whose lipid-lowering medication was changed during study period
Hispanics: 52.7 (15.2) yr

Non-Hispanics: 61.2 (12.8) yr

% female NR
Hispanics: 8.2% (1.9%) non-Hispanics: 8.0% (1.9%)

Hispanics: 89.2 (NR) kg
Non-Hispanics: 99.6 (NR) kg
A1c at 3-m follow-up Hispanic: -1.2(1.8)
Non-Hispanic: 1.1(1.4)
No AEs presented Weight gain: Hispanics 1.41 kg, Caucasians 1.64 kg (P=0.54)
Rajagopalan R., 2004

NA (based on 5 other studies, 1 of fair quality; data not available in 4)
Multicenter trials
5 RCTs, 1 published (Rosenblatt 2001), others unpublished by Takeda PharmaceuticalsNR2 placebo-controlled Pio monotherapy trials; 1trial each of Pio combined with metformin, sulfonylurea, or insulinInclusion: Patients 30–75 years, BMI 25–40 mg/m2, fasting c-peptide>0.331 nmol/L, normal thyroid function
Exclusion: NYHA class III or IV status, significant renal or hepatic disease, uncontrolled hypertension, coronary artery disease or stroke in last 6m
Two subgroups examined: <65 and ≥65 years; mean age and % female NR< and >65 years reported as ranges for the 5 studies combined A1c: 9.8% to 10.9%; 8.9% to 10.3% BMI, weight NRMean decrease from baseline in A1c 0.53 to 1.94%; older group had similar response to younger group; both groups also benefits to a comparable degree for lipid levelsAdverse cardiovascular events and hypoglycemia were similar in the younger and older age groups treated with Pioglitazone monotherapy and with Pioglitazone combined with metformin.
Hypoglycemia was2-fold higher in the older-aged group using Pioglitazone combined with a sulfonylurea or insulin.
Tan M 2004 (glimepiride study)

RCT, AC, DBHispanic 99%, white 1%NonePatients with DM2 and A1c >7.5% and ≤11.0%in patients who were not receiving oral hypoglycemic agents, and >7.5% and ≤ 9.5% in patients who were receiving oral agents.
Patients must have had a trial of diet and lifestyle interventions before study enrollment
Exclusion criteria: significant functional limitation (NYHA class III or IV; triglycerides>400 mg/dl; serum creatinine >2.0 mg/dl; renal transplantation or current renal dialysis; ALT or AST > 2.5 times upper limit of normal; clinical signs or symptoms of liver disease; Hg<115 g/l for women and <115g/l for men; BMI <25 or >35 kg/m2; signs or symptoms of substance abuse
55.3 (NR) yr

51% female

74.4 (NR) kg
A1c at 1-year follow-up Pio: -0.8%
Glimepiride: −0.7% Between-group Pvalue = 0.64
Incidence of treatment-emergent and severe AEs was similar in the 2 groups
Agrawal, A 2003

Fair, based on secondary data
RCT, PC, DB, secondary data from 3 RCTs examined subgroup with decreased renal function (creatinine clearance 30–80 ml/min)NRAdded to various SUPatients currently treated with SU
Exclusion criteria: patients of child-bearing potential, serum creatinine level >1.8 mg/dl
61.6 (NR) yr

38% female
9.15% (NR)

28.8 (NR) kg/m2
A1c at 6m: Between-group change −1.1% for both renal impaired and nonimpaired patients% AEs was similar for patients in both treatment groups when comparing those with renal impairment and those without, including incidence of hypoglycemia; edema more common in patients with normal renal function in both treatment groups (no statistics)
Barnett, A 2003

RCT, PC, DBIndian: 60%; Pakistani: 27%; Bangladeshi: 9.5%; Sri Lankan: 3%; Mauritian: less than 1%Added to SUPatients with DM2 taking SU for at least 4 months with dose unchanged within 2 months before start of study, those taking medications that affect glucose or lipids were eligible if doses remained constant at screening and during study period
Exclusion criteria: patients of child-bearing potential, severe hypertension, anemia or blood disorders, congestive heart failure, significant liver disease, a weight variance of >5% between screening baseline
54.2 (NR) yr

22% female
9.13% (NR) 26.6 (NR) kg/m2A1c at 26 weeks
Rosi: −1.16 Placebo 0.26 (P<0.001)
Treatment-emergent AEs in 70% Rosi and 75% with placebo; withdrawals for AEs: Rosi 5%, placebo 10% Weight (kg): Rosi 3.9, placebo −0.1(P<0.001)
Chan NN 2004 (Observational study)USA
Single cente
Cohort, single groupChineseMonotherapyTwelve insulin-treated DM2 patients with nephropathy who were started on ROSI due to suboptimal glycemic control and progressive weight gain
All patients had diabetic nephropathy, with urinary albumin-creatinine ratio >25 mg/mmol; mean serum creatinine 223.1 (68.1)
Exclusion criteria: none reported
65 (8.3) yr

58% female
8.6% (NR)

71.7 (NR) kg
A1c at 15.5m: −1.1 (P=0.01)LFT: no significant increase in ALT
Hematocrit: NSD weight gain 2.2 kg (P=0.08)
Choi D 2004 (Observational study)Korea
Single center
RCTKoreanCombined therapy with a variety of hypoglycemic agents used by both groups (SU, metformin, α-glucosidase inhibitor, or insulin); % son each drug not specified95 previously-treated diabetics who had recent acute MI or stable or unstable angina and underwent coronary stent implantation at a Korean university hospital
Exclusion criteria: prior treatment with TZDs, ejection fraction <35%, liver or renal disease, pregnancy, reference vessel diameter <2.75mm
59.9 (9.3) yr

7.72% (1.13%)
68.1 (11.0) kg
24.8 (3.35) kg/m2
6 months: Intervention change: −0.61 (1.15)
Control change: −0.75 (1.07)
“No patient had significant side effects, such as an elevation in the liver enzyme levels.”
Honisett, S 2003

RCT, PC, DBNR80% continued their use of metformin, SU, or bothWomen, diagnosed with DM2 1–12 years prior to study; all postmenopausal
Exclusion criteria: none reported

100% female
A1c change at 12 weeks: −1.2%, P=0.001No AEs were reported to the investigators
Jones, T 2003

RCT, PC, open-labelNRAdded to metforminPatients of non-child-bearing potential, aged 40–80 years, diagnosed with DM2, fasting C-peptide >0.8 ng/ml at screening, maintaining a FPG level (between >140 mg/dL- <300 mg/dL) prior to randomization
Exclusion criteria patients with clinically: significant renal or hepatic disease, angina, cardiac insufficiency, symptomatic diabetic neuropathy, significant clinical abnormality on ECG, history of chronic insulin therapy, participation in any previous rosi-related study
59.9 (NR) yr

32% female
8.83% (NR) 28.2
BMI<25: Rosi 8 mg+metformin −0.3; metformin alone 0.3
BMI 25–30: Rosi 8 mg+ metformin: − 0.7; metformin alone 0.1
BMI >30: Rosi: 8 mg+ metformin −1.0; metformin alone 0.2
Data from graphs, exact values NR rosi vs. metformin P<0.025 for all 3 groups
AE profile not different between normal weight, overweight, and obese
Kreider M 2002

NA (based on 8 other studies, primary data not available)
Secondary data: 8 studies, either PC or AC, DB% White: <70years: 79%
>70years: 91%
Monotherapy, elderlyDM2, FPG varied among studies, range 7.8–16.9 mmol/l; age varied, range 30–80y; BMI 22–38 kg/m2
Patients stratified by < or >=70y
Efficacy data pooled from 3 monotherapy studies of 26w duration
Significant renal disease; angina or cardiac insufficiency, symptomatic diabetic neuropathy, hepatic disease, history of diabetic ketoacidosis, history of chronic insulin use, other serious major illness
<70 years: 56 >70 years: 73

37% female
<70 years: Rosi: 8.8% (1.5%); placebo 9.0% (1.7%) >70 years: rosi: 8.6% (1.4%); placebo 8.9% (1.5%)
BMI: <70 years: Rosi: 29.8 (4.1) kg/m2; placebo 29.8 (4.2) kg/m2 >70 years: Rosi: 28.3 (3.9) kg/m2; placebo 28.4 (4.1) kg/m2
A1c at 26 weeks
<70 years: Rosi 4 mg daily: −0.2; 8 mg daily −0.5; placebo 0.8
>70 years: Rosi 4 mg daily: −0.1; 8 mg daily: −0.4; placebo 1.0 NSD between the 2 age groups
Hypoglycemic episodes occurred in <1% on ROSI in either age group; 2 patients <70y in Rosi group discontinued treatment because of hypoglycemia
Vongthavara vat V., 2002
Various Asia and South AmericaMulticenterRCT, no-treatment control, open-labelWhite (38.3%); Black (3.0%); Asian (57.5%); Other (1.2%)Added to SUPatients with DM2 (as defined by the National Diabetes Data group criteria) who had been receiving SU therapy (glibenclamide, glipizide, gliclazide, chlorpropamide, tolbutamide, or glimepiride) for at least 6 months and if SU dose had been constant for at least 2 months before the screening visit; between 40 and 80 years of age and FPG 126 to 270 mg/dl at screening. Exclusion criteria: Significant renal or hepatic impairment, hypertension, anemia, abnormal blood cell counts or hypertension severe angina, coronary insufficiency, heart failure, EKG evidence of left ventricular hypertrophy; patients requiring insulin or who had taken investigational drugs within 30 days of screening.;56.0 (NR) yr

56% female
NR 68.9 kg 27.1 kg/m2A1c change at 26
weeks:Rosi+SU: −1.1(95% CI−1.37, −0.89); SU control: 0.1(−0.1−0.2)
Hypoglycemia (%)Rosi+SU: 11.6; SU control: 1.2 (P<0.001) Serious AE (%): Rosi+SU: 2.4; SU control: 5.3
Wang G., 2005
Single center
RCT, no-treatment control, open-labelChinese (assumed)MonotherapyAged 50 to 73, with a diagnosis of coronary artery disease (>50% stenosis as proven on angiography) and established DM2
Exclusion criteria:
Acute MI during the preceding 12 weeks, cardiac insufficiency, renal function impairment, liver function impairment, systemic inflammatory disease, infectious disease, cancer, or a serious illness that would affect participation; insulin treatment.
61.2 (8.6)yr

18% female
7.33% (0.17%)

25.6 (2.7) kg/m2
Change in A1c reported graphically only (difficult to interpret)
Rosi: decreased at 6m compared to control group (P<0.05)
Weight gain: NSD from baseline level and from control group (data not provided)
Wang, T 2004 (Metabolic syndrome only)

RCT, PC, open-labelChinese (assumed)MonotherapyPresence of metabolic syndrome and meet at least of the following 3 criteria: waist circumference of >90 cm in men and >80 cm in women, serum TG > 150 mg/dl, HDL <40 mg/dl in men and <50 mg/dl in women, IFG 110–125 mg/dl, BP >130/85 mm Hg or treated hypertension.
Exclusion criteria:
Patients with acute coronary events, stroke or coronary revascularization within the preceding 3 months; diabetes mellitus according to the criteria of the American Diabetes Association, overt liver disease, chronic renal failure, hypothyroidism, myopathy alcohol/drug abuse, several other significant diseases, use of other lipid-lowering therapy, immunosuppressants, erythromycin, hormone replacement therapy.
59.5 (NR)y

42% female

25.4 (NR) kg/m2
A1c NR
FPG: NSD within or between groups (P>0.05)
AEs reported as none
Pioglitazone and rosiglitazone
Manley HJ 2003


(Cohort study)
Single Center
Retrospective cohortNRCombined therapy, variousChart review of patients receiving hemodialysis at a US clinic who were prescribed either rosi or pio from 4/2001 to 5/2002
Diabetes was the cause of ESRD in 92.5%
Exclusion criteria: none reported
64.8 (11.5) yr
Range: 46–85 yr

35% female
8.6% (2.2%)

Comparison of rosi to pio: interdialytic weight change Rosi: 3.6 kg at baseline and 3.97 at 3m follow-up ( P=0.0032); hematocrit:
Rosi 34.89 at baseline and 34.0 at follow-up; data not provided for pio, but difference between pio and rosi for these 2 variables was reported as significant, but NR direction of pio effects compared to rosi
No data provided on AEs.

Abbreviations: AE, adverse event; DB, double blind; MI, myocardial infarction; NR, not recorded; NSD, no significant difference; NYHA, New York Heart Association; PC, placebo-controlled; pio, pioglitazone; RCT, randomized controlled trial; rosi, rosiglitazone; SU, sulfonylurea.

From: Results

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.