The effects of rosiglitazone on inflammatory biomarkers and adipokines in diabetic, hypertensive patients

Abstract

OBJECTIVE:

To compare the effects of a 12-week treatment course of a rosiglitazone-based versus a metformin- or glyburide-based strategy on inflammatory biomarkers and adipokine levels in hypertensive, type 2 diabetes patients.

METHODS:

One hundred three treatment-naive patients or patients on monotherapy with either metformin or glyburide, and a hemoglobin A1C (A1C) ≥7.5%, were randomly assigned to either rosiglitazone add-on (4 mg/day ± titration to 8 mg/day) or a combination of metformin (250 mg twice per day [BID] titrated to 500 BID if A1C ≥7.5% and ≤8.0%; 500 mg BID titrated to 1 g BID if A1C >8.0%) and glyburide (2.5 mg BID titrated to 5 mg BID if A1C ≥7.5% and ≤8.0%; 5 mg BID titrated to 10 mg BID if A1C >8.0%).

RESULTS:

Rosiglitazone add-on produced significantly greater reductions in high-sensitivity C-reactive protein (2.1 mg/L to 0.9 mg/L) and increases in adiponectin (8.7 mg/mL to 14.8 mg/mL) levels compared with metformin/glyburide (both P<0.005). At close-out, all patients had improved fasting plasma glucose and A1C levels (8.5% to 7.4% and 8.8% to 7.1% for rosiglitazone add-on and metformin-glyburide, respectively [P<0.001 for both arms]) relative to the corresponding baseline values.

CONCLUSIONS:

The present study demonstrated that in hypertensive, diabetic subjects, a rosiglitazone-based treatment strategy results in favourable changes in inflammatory biomarkers compared with metformin/glyburide.

Type 2 diabetes, a growing global health burden (1,2), is often associated with the presence of hypertension (3,4), which alone accounts for approximately 75% of the cardiovascular risk in patients with diabetes (5). The recent report of the Canadian Chronic Disease Surveillance System (6), noting that more than one million Canadians have concomitant type 2 diabetes and hypertension, has highlighted the need to define appropriate management for such patients.

Although current guidelines encourage a multifactorial approach to treat type 2 diabetes (7,8), often the focus is on lowering glucose levels. Oral glucose-lowering agents are commonly prescribed if pancreatic beta-cell function is believed to be preserved. The biguanide metformin and the thiazolidinedione rosiglitazone are insulin sensitizers that work via different mechanisms. Metformin is a recommended first-line antihyperglycemic agent (9) that decreases hepatic gluconeogenesis while augmenting peripheral glucose uptake (10–12). Rosiglitazone acts via the peroxisome proliferator-activated receptor-gamma (PPAR-γ) and, until recently, was widely prescribed for the management of type 2 diabetes because of its apparent ability to improve pancreatic beta-cell function, improve whole-body insulin sensitivity and afford durable glycemic control (13).

Adipose tissue is recognized as an important source of molecular mediators, collectively known as adipokines, which are active regulators of the inflammatory milieu involved in mediating vascular injury, insulin resistance and atherogenesis (14). Among the proinflammatory mediators are C-reactive protein (CRP), interleukin (IL)-6, leptin and plasminogen activator inhibitor (PAI)-1. Adipose tissue also secretes adiponectin, an adipokine believed to be anti-inflammatory and to protect against obesity-linked insulin resistance (15). Accordingly, strategies directed at elevating adiponectin levels may retard atherosclerosis and improve whole-body insulin sensitivity.

The present study sought to determine whether a rosiglitazone-based regimen was superior to a metformin-based approach in altering the levels of selected markers of inflammation, endothelial dysfunction and procoagulant imbalance, all of which contribute to global vascular risk in diabetic, hypertensive patients (16). Specifically, patients with type 2 diabetes mellitus and hypertension were enrolled and followed over a 12-week period because previous work indicated that this duration was sufficient to reliably detect fluctuations in inflammatory markers in patients with coronary artery disease (17,18).

METHODS

Study design and population

The design of the present study has previously been described (19). Briefly, it was a single-centre, randomized, open-label, comparative parallel-group study that was approved by a central ethics review committee. Enrollment occurred between April 2006 and April 2009, and written informed consent was provided by all subjects.

Patients who met the inclusion criteria were >40 and <80 years of age with type 2 diabetes and suboptimal glycemic control (hemoglobin A1C [A1C] ≥7.5%), were treatment-naive for glucose lowering, or receiving metformin or glyburide monotherapy, and were being treated with an angiotensin-receptor blocker (ARB) for hypertension and/or diabetic nephropathy, or had an indication for ARB therapy (systolic blood pressure [BP] >129 mmHg and/or diastolic BP >79 mmHg, or an albumin:creatinine ratio of >2.0 mg/mmol for men and >2.8 mg/mmol for women). ARB therapy was included as part of the protocol to ensure that in all study subjects, BP was not only well controlled but also ideally with an agent that was deemed to be metabolically ‘neutral’. Individuals were excluded if they were receiving insulin therapy or multiple oral glucose-lowering agents, were pregnant, breastfeeding or HIV positive, had a history of significant cardiac, hepatic, renal or systemic inflammatory disease, were taking a PPAR-γ agonist, potassium-sparing diuretics, steroids, chemotherapy drugs or nonsteroidal anti-inflammatory drugs other than acetylsalicylic acid, or had other significant and potentially compromising laboratory abnormalities.

Figure 1 provides an outline of the study algorithm. Participants were asked to continue with their prerandomization antihyperglycemic regimen. Study-initiated glucose-lowering treatment was stratified and allocated at visit 2 (weeks 1 to 2). Treatment-naive subjects were randomly assigned to either rosiglitazone (4 mg/day force titrated to 8 mg/day) or metformin (250 mg twice per day [BID] titrated to 500 mg BID if baseline A1C ≥7.5% and ≤8.0%, or 500 mg BID titrated to 1 g BID if baseline A1C >8.0%). Patients taking metformin before randomization were randomly assigned to the addition of either rosiglitazone (4 mg/day force titrated to 8 mg/day) or glyburide (2.5 mg BID titrated to 5 mg BID if baseline A1C ≥7.5% and ≤8.0%, or 5 mg BID titrated to 10 mg BID if baseline A1C >8.0%). Individuals taking glyburide before randomization were randomly assigned to the addition of either rosiglitazone (4 mg/day) or metformin (250 mg BID titrated to 500 mg BID if baseline A1C ≥7.5% and ≤8.0% or 500 mg BID titrated to 1 g BID if baseline A1C >8.0%). Additional medication, if required, was dispensed at week 4 (visit 4). Compliance and adverse events were assessed, both in clinic and by telephone, at all visits.

Figure 1).

Study algorithm. A1C Hemoglobin A1C; aPAI Active plasminogen activator inhibitor; hs-CRP High-sensitivity C-reactive protein; IL Interleukin; MMP Matrix metalloproteinase; sVCAM Soluble vascular cell adhesion molecule

Outcome measures

The primary end point of the study was the change in adiponectin level (from baseline to 12 weeks) in the rosiglitazone-treated versus the metformin/glyburide-treated subjects. Fasting blood samples collected at the time of enrollment (visit 1, day 0), visit 2 and close-out (visit 6, week 12) were centrifuged to obtain serum or plasma, which were stored at −80°C. Samples were run simultaneously to minimize interassay variability. Glycemic and lipid profiles as per serum samples collected at visits 1 and 6 were determined at diagnostic laboratories routinely used by the recruitment clinic. Serum levels of adiponectin, interleukin (IL)-6, leptin, matrix metalloproteinase (MMP)-9, active PAI (aPAI)-1 and soluble vascular cell adhesion molecule (sVCAM)-1 in samples collected at visit 2 and visit 6 were quantified using commercially available kits (LINCOplex, Millipore, USA) while separate sample aliquots collected at the same visits were analyzed for high-sensitivity CRP (hs-CRP) at an accredited diagnostic laboratory. Intra-assay and interassay coefficient of variation values, respectively, were: adiponectin (9.2%, 15.9%), IL-6 (7.8%, 18.0%), leptin (7.9%, 15.0%), MMP-9 (6.8%, 11.7%), aPAI-1 (6.6%, 10.0%) and sVCAM-1 (4.5%, 8.5%).

Statistical analysis

Prestudy power calculations indicated that a sample size of 50 patients per group was required to detect differences in adiponectin changes between the two study arms. Categorical variables are described with absolute frequencies while continuous variables are summarized as median (25th percentile, 75th percentile). Adiponectin, hs-CRP, IL-6, leptin, MMP-9, aPAI-1 and sVCAM-1 values underwent natural logarithmic transformation before assessment with ANOVA for repeated measures. Differences were considered to be statistically significant at a two-sided P<0.05. Statistical analyses were performed using SPSS version 16.0 (IBM Corporation, USA).

RESULTS

Participants

Of the 115 subjects enrolled, nine withdrew before randomization and three were lost to follow-up. Thus, the final study cohort consisted of 103 patients, of whom 33 were treatment-naive for glucose-lowering and 70 were receiving either metformin or glyburide monotherapy before study enrollment. Of these 103 patients, 55 were randomly assigned to the addition of rosiglitazone, and 48 to the addition of either metformin or glyburide. Baseline characteristics are provided in Table 1. Subjects in the two subgroups were well matched.

TABLE 1.

Baseline characteristics of the study population

	All (n=103)	Rosiglitazone add-on (n=55)	Metformin/glyburide (n=48)
Demographic, median (percentiles 25th, 75th)
Male sex, n (%)	60 (58)	30 (55)	30 (63)
South Asian, n (%)	91 (88)	50 (91)	41 (85)
Age, years	59 (51, 65)	61 (51, 64)	56 (51, 65)
Physical, physiological and biochemical measurements, median (percentiles 25th, 75th)
Weight, kg	76 (67, 84)	74 (65, 81)	77 (70, 85)
Body mass index, kg/m2	27.7 (25.1, 29.8)	27.7 (23.7, 29.4)	28 (26, 30)
Fasting plasma glucose*
mmol/L	10.1 (8.1, 12.8)	10.0 (7.8, 12.5)	10.3 (8.4, 12.9)
mg/dL	182 (146, 230)	180 (140, 225)	185 (151, 232)
A1C, %	8.6 (7.8, 9.9)	8.5 (7.9, 10.0)	8.8 (7.8, 9.8)
Blood pressure, mmHg
Systolic	132 (119, 146)	131 (120, 149)	134 (119, 142)
Diastolic	80 (77, 86)	80 (77, 84)	81 (77, 89)
Serum creatinine
μmol/L	74 (67, 84)	76 (68, 85)	74 (66, 84)
mg/dL	0.84 (0.76, 0.95)	0.86 (0.77, 0.96)	0.83 (0.75, 0.95)
Cardiovascular risk factors, n (%)
Current smoker	4 (4)	1 (2)	3 (6)
Family history of premature coronary artery disease	10 (10)	2 (4)	8 (17)
Diabetes	103 (100)	55 (100)	48 (100)
Hypertension	63 (62)	31 (56)	32 (67)
Hyperlipidemia	67 (65)	33 (60)	34 (71)
Cardiovascular history, n (%)
Myocardial infarction	15 (15)	6 (11)	9 (19)
Angiographic stenosis	14 (14)	7 (13)	7 (15)
>49% luminal diameter
Stroke/TIA	2 (2)	0 (0)	2 (4)
Previous revascularization, n (%)
PTCA	9 (9)	3 (5.5)	6 (12.5)
CABG	6 (6)	5 (9)	1 (2)

^*n=100 because three patients did not complete two glycemic workups. A1C Hemoglobin A1C; CABG Coronary artery bypass graft; PTCA Percutaneous transluminal coronary angioplasty; TIA Transient ischemic attack

Anthropometric changes

Neither the rosiglitazone-based nor metformin/glyburide combination treatment had any appreciable impact on weight or body mass index at the end of the 12-week study period (Table 2).

TABLE 2.

Baseline and final anthropometric profiles of the study population

Parameter	Rosiglitazone add-on (n=55)		Metformin/glyburide (n=48)		P
	Baseline	Close-out (week 12)	Baseline	Close-out (week 12)
Weight, kg	74 (65, 81)	76 (65, 82)	77 (70, 85)	77 (69, 86)	0.257
Body mass index, kg/m2	27.7 (23.7, 29.4)	27.9 (24.1, 29.7)	27.8 (25.6, 30.0)	27.3 (25.6, 31.2)	0.213

Data presented as median (25th percentile, 75th percentile). P values represent intergroup comparisons of delta change for baseline and close-out values

Changes in glycemic, BP and lipid profiles

As shown in Table 3, improvements in A1C (P<0.0001) and fasting plasma glucose (P<0.005) levels were observed independent of the randomized treatment strategy. Rosiglitazone-based treatment was associated with a lower BP at 12 weeks compared with baseline. No appreciable change in lipid parameters was observed in either group.

TABLE 3.

Baseline and final glycemic, blood pressure and lipid profiles of study population

Parameter	Rosiglitazone add-on*		P†	Metformin/glyburide*		P†	P‡
	Baseline	Close-out (week 12)		Baseline	Close-out (week 12)
Fasting plasma glucose
mmol/L	10.0 (7.8, 12.5)	7.1 (6.4, 8.5)	<0.0001	10.3 (8.4, 12.9)	7.7 (6.9, 10.5)	<0.005	0.248
mg/dL	180 (140, 225)	128 (115, 153)		185 (151, 232)	139 (124, 189)
A1C, %	8.5 (7.9, 10.0)	7.4 (6.9, 8.3)	<0.0001	8.8 (7.8, 9.8)	7.1 (6.6, 8.2)	<0.0001	0.935
Blood pressure, mmHg
Systolic	131 (120, 149)	124 (113, 138)	<0.05	134 (119, 142)	125 (116, 140)	NS	0.281
Diastolic	80 (77, 84)	78 (69, 83)	<0.05	81 (77, 89)	79 (73,84)	NS	0.731
Total cholesterol
mmol/L	4.9 (3.9, 5.6)	4.9 (4.1, 5.6)	NS	4.8 (3.9, 5.3)	4.5 (4.0, 5.3)	NS	0.662
mg/dL	191 (152, 218)	191 (160, 218)		187 (152, 207)	177 (156, 207)
Low-density lipoprotein-cholesterol
mmol/L	2.9 (2.1, 3.5)	2.8 (2.2, 3.4)	NS	2.5 (2.0, 3.1)	2.4 (2.0, 3.2)	NS	0.819
mg/dL	111 (82, 137)	109 (86, 133)		98 (78, 121)	94 (78, 125)
HDL-cholesterol
mmol/L	1.2 (1.0, 1.4)	1.2 (1.1, 1.5)	NS	1.1 (0.9, 1.3)	1.1 (1.0, 1.3)	NS	0.127
mg/dL	45 (39, 55)	49 (43, 59)		45 (35, 51)	44 (39, 51)
Total cholesterol:HDL cholesterol ratio	4.0 (3.3, 4.5)	3.8 (3.0, 4.7)	NS	4.1 (3.3, 5.3)	4.1 (3.2, 5.2)	NS	0.304
Triglycerides
mmol/L	1.5 (1.2, 2.2)	1.4 (1.1, 1.9)	NS	1.7 (1.3, 2.9)	1.7 (1.4, 2.4)	NS	0.345
mg/dL	136 (107, 196)	122 (98, 169)		151 (116, 258)	154 (125, 214)

Data presented a median (25th percentile, 75th percentile).

^*n is variable because three patients did not complete two glycemic work-ups, one did not provide a sample for lipid analysis at the final visit and four subjects had high triglyceride readings that precluded the calculation of low-density lipoprotein levels;

^†P for intragroup comparisons between baseline and close-out values;

^‡P for intergroup comparisons of delta change for baseline and close-out values. HDL High-density lipoprotein; NS Not statistically significant

Changes in adipokine and inflammatory biomarker profiles

Table 4 details the levels of adiponectin, hs-CRP, IL-6, leptin, MMP-9, aPAI-1 and sVCAM-1 detected in fasting serum samples collected at visits 2 and 6. At the end of the 12-week study period, patients randomly assigned to rosiglitazone exhibited higher levels of the anti-inflammatory adipokine adiponectin (8.7 mg/mL versus 14.8 mg/mL; P<0.001), and significantly lower serum hs-CRP (2.1 mg/L versus 0.9 mg/L; P= 0.002) and MMP-9 concentrations (75.3 ng/mL versus 50.7 ng/mL; P=0.004). Conversely, patients randomly assigned to the combination of metformin and glyburide demonstrated no significant difference in adiponectin, hs-CRP and MMP-9 levels between baseline and the end of the study. While sVCAM-1 levels were reduced by a similar degree in both study groups, no statistically significant differences were observed with respect to IL-6, aPAI-1 or leptin levels between groups.

TABLE 4.

Baseline and final adipokine, inflammatory and matrix remodelling biomarker profiles of the study population

Marker	Rosiglitazone add-on				Metformin/glyburide				P‡
	n*	Baseline	Close-out (week 12)	P†	n*	Baseline	Close-out (week 12)	P†
Adiponectin, mg/mL	33	8.7 (6.2, 10.0)	14.8 (11.2, 19.6)	<0.001	46	9.2 (6.1, 13.2)	8.5 (6.4, 11.9)	0.605	<0.001
hs-CRP, mg/L	55	2.1 (1.2, 3.6)	0.9 (0.7, 1.5)	0.002	47	2.7 (1.4, 4.3)	2.0 (1.3, 4.3)	0.625	0.01
IL-6, pg/mL	47	2.7 (1.6, 3.6)	2.0 (1.2, 3.2)	0.077	45	2.9 (1.8, 3.7)	2.0 (1.4, 3.2)	0.211	0.569
Leptin, ng/mL	55	10.0 (5.6, 17.4)	12.9 (6.2, 20.5)	0.29	47	13.3 (6.7, 19.0)	13.8 (8.9, 24.1)	0.325	0.624
MMP-9, ng/mL	55	75.3 (51.4, 114.4)	50.7 (38.6, 79.4)	0.004	48	91.1 (60.1, 140.5)	81.4 (60.7, 108.2)	0.163	0.664
aPAI-1, ng/mL	38	14.2 (11.2, 24.8)	12.9 (6.5, 19.4)	0.061	41	22.2 (15.3, 34.2)	19.4 (10.2, 30.7)	0.161	0.569
sVCAM-1, ng/mL	55	1626 (1310, 1860)	1178 (971, 1452)	<0.001	48	1670 (1211, 2066)	1390 (1054, 1595)	0.024	0.235

Data presented as median (25th percentile, 75th percentile).

^*n is variable because levels of markers in some samples were outside of the detection window.

^†P for intragroup comparisons between baseline and close-out values;

^‡P for intergroup comparisons of delta change for baseline and close-out values. aPAI Active plasminogen activator inhibitor; hs-CRP High-sensitivity C-reactive protein; IL Interleukin; MMP Matrix metalloproteinase; sVCAM Soluble vascular cell adhesion molecule

DISCUSSION

The findings of the present study indicate that in patients with type 2 diabetes and hypertension, 12 weeks of rosiglitazone-based treatment was associated with a more favourable anti-inflammatory profile compared with a metformin/glyburide combination strategy. In patients assigned to rosiglitazone, we observed greater reductions in hs-CRP levels, with concomitant increases in the protective anti-inflammatory adipokine adiponectin. Notably, however, rosiglitazone treatment did not result in any appreciable changes in IL-6, leptin, MMP-9, aPAI-1 and sVCAM-1 levels relative to metformin/glyburide pharmacotherapy.

In recent years, it has become increasingly apparent that a complex interplay exists among systemic inflammation, activation of vascular cells, procoagulant imbalances and arterial structural modifications. All of these changes have, individually and in various permutations, been reported to contribute to diabetogenesis and atherosclerotic risk, and markers of these disorders include adiponectin, CRP, IL-6, leptin, PAI-1, VCAM-1 and MMP-9.

Hypoadiponectinemia is a strong predictor of diabetes (20,21) while higher adiponectin levels are associated with more favourable glycemic, lipid and inflammatory profiles in type 2 diabetes (22). For every 1 g/mL decrease in adiponectin level, common carotid artery intima-media thickness increased by 3.48 μm (95% CI 1.23 μm to 5.73 μm) in men and 2.39 μm in women (95% CI 10.50 μm to 4.27 μm) (23), suggesting that hypoadiponectinemia is associated with the development of early atherosclerosis. Along similar lines, subjects with type 2 diabetes and macrovascular disease were found to have lower circulating adiponectin levels than those without macrovascular disease (24), and an inverse relationship was noted between circulating adiponectin and cardiovascular events (RR 0.71 [95% CI 0.53 to 0.95]) (25).

Evidence suggests that hs-CRP, aside from being an independent predictor of cardiovascular events (26), may also provide a means of identifying and providing prognostic information regarding the state of diabetogenesis (27,28). Of note, hs-CRP has demonstrated strong correlations with nonfatal (29) and fatal (30–32) cardiovascular events in patients with type 2 diabetes. In a seven-year follow-up study completed by Soinio et al (32), patients with type 2 diabetes and hs-CRP concentrations of >3.0 mg/L exhibited a greater risk of coronary artery disease (RR 1.72) than did those with hs-CRP levels ≤3.0 mg/L. Identification of a strategy that mitigates rises in hs-CRP levels while improving glycemic control and lipid profiles may be of particular benefit to patients with type 2 diabetes.

Following the observation of heightened MMP-9 activity in the aortas and blood samples of diabetic rats (33), similar increases in MMP-9 activity were reported in the plasma (34) and atherosclerotic plaques (35,36) of patients with type 2 diabetes mellitus and peripheral or coronary arterial disease. Because dysregulated MMP production can cause net degradation of the extracellular matrix (37), it is reasonable to postulate that excessive MMP activity may aggravate atherosclerosis in type 2 diabetes, thereby contributing to type 2 diabetes-associated plaque vulnerability.

Concordant with earlier reports in type 2 diabetic patients (38–40), we noted that rosiglitazone elicited a significant increase in adiponectin levels. Robust amelioration of hs-CRP concentrations in the rosiglitazone study arm similarly mirrored those previously discussed (7,17,18,38,41,42, 43). The fact that leptin and IL-6 levels were unchanged concurs with the reports of Betrand et al (38) and Marx et al (17). In accord with the study by Reynolds et al (42), but not two others (38,44), the median aPAI-1 level was nonsignificantly lower in rosiglitazone-treated patients. The differences among studies for the IL-6, leptin and aPAI-1 outcomes are not surprising, and may be attributed, in part, to heterogeneous patient characteristics at baseline, the varying durations of therapy and the different single time points selected for evaluating these biomarkers.

Under immunostimulatory situations, VCAM-1 functions as both a scaffold for leukocyte migration and a trigger of endothelial signalling. Unlike the other biomarkers examined, however, sVCAM-1 levels were significantly reduced in both study arms, suggesting that the two antihyperglycemic strategies are similarly effective at modulating endothelial cell activation. Interestingly, a causal relationship between rosiglitazone and decreased sVCAM-1 was reported by Dolezalova et al (44) but was not observed by other groups (17,38).

Significant lowering of serum MMP-9 levels has been documented in patients with type 2 diabetes and coronary artery disease as early as two weeks into rosiglitazone therapy (17) and sustained for up to 12 months postinitiation of rosiglitazone treatment (17,38,41). The current findings extend those previously reported by demonstrating a corresponding inhibitory and protective effect of rosiglitazone on MMP-9 expression in type 2 diabetes patients with hypertension.

Weight gain is a common concern despite the simultaneous improvements in insulin sensitivity and BP afforded by rosiglitazone. No significant weight fluctuations were observed in this study population, which contradicts earlier reports showing significant weight gain with rosiglitazone usage (38,40,42,45), weight loss following metformin monotherapy (45,46) and less weight gain/no weight gain in those undergoing rosiglitazone-metformin cotherapy (45). Because these studies were conducted over longer periods of time (≥24 weeks), it is plausible that the duration of our study was too short to reveal any appreciable changes in weight and body mass index.

The Multiple Risk Factor Intervention Trial (MRFIT) (47) was the first to recognize that for any given systolic pressure, diabetes is associated with a significant (>2-fold) increase in age-adjusted cardiovascular death rate. Subsequent meta-analysis by the Blood Pressure Lowering Treatment Trialists’ Collaboration (48) revealed that lowering systolic BP by 6/4 mmHg in diabetic subjects was associated with a 25% decrease in major cardiovascular events and a 24% reduction in total mortality.

In our study, rosiglitazone not only improved glycemic status but also favourably modified the proinflammatory profile and lowered BP. The enthusiasm for the pleiotropic advantages of rosiglitazone has, however, been significantly dampened by the increased risk of heart failure and/or myocardial infarction in patients with type 2 diabetes (44).

Several limitations of our study warrant consideration. Our study cohort was small, derived from a single site and predominantly of South Asian descent, factors that may limit the generalizability of our findings to other populations. Participation in the study may have itself partially accounted for the improved glycemic and biomarker profiles. It is unknown whether patients were comanaged by another specialist or if they had integrated changes in diet and lifestyle during the study window (we did, however, advise maintenance of diet and exercise patterns over the duration of the study, plus other drugs were not changed during the study). All of these factors could have contributed to fluctuations in the final outcome measures. Although our power calculations indicated that a sample size of 50 patients per group was required to detect differences in adiponectin changes between the two study arms, it is possible that this sample size may not be similarly powered for the other biomarkers assessed; thus, we cannot exclude the possibility our study may be underpowered to accurately evaluate all biomarkers. Finally, because the present analysis was a short-term study, our findings do not shed further insight regarding the safety of rosliglitazone. It is also important to note that the changes observed in circulating biomarkers with rosiglitazone add-on must be balanced with the available clinical data that suggest that rosiglitazone therapy is associated with an untoward cardiovascular profile in clinical trials. Specifically, rosiglitazone has been linked with a significant increase in the risk of myocardial infarction (OR 1.43 [95% CI 1.03 to 1.98]; P=0.03) and with an increase in the risk of death from cardiovascular causes (OR 1.64 [95% CI 0.98 to 2.74]; P=0.06) (50).

CONCLUSION

The present study demonstrated that 12 weeks of rosiglitazone-based therapy, in subjects with type 2 diabetes and hypertension, is associated with evidence of reduced hs-CRP levels and elevated concentrations of adiponectin, indicative of a shift in the balance toward an improved inflammatory profile. The long-term benefits of rosiglitazone’s effects on inflammatory markers and adiponectin with respect to vascular risk reduction remain to be determined.