Evidence Based Review Article

The Journal of Informed Pharmacotherapy 2001;4:204-207.

The addition of rosiglitazone to metformin improves glycemic control in patients with type 2 diabetes: Are things looking "ROSI'ER"?

Download a PDF version of this article

Reviewer: Beena Kuriakose B.Sc. (Pharm) and Lisa Dolovich, Pharm.D.
Reviewer's email address:
beena.kuriakose@utoronto.ca and ldolovic@mcmaster.ca
Reviewer's profession/specialty:
Ambulatory care

Original Citation

Fonseca V, Rosenstock J, Patwardhan R, Salzman A.  Effect of metformin and rosiglitazone combination therapy in patients with type 2 diabetes mellitus: a randomized controlled trial.  JAMA 2000;283:1695-702. PubMed Cit

Overall Study Question

This study was conducted to determine the effectiveness and safety of adding rosiglitazone to maximal doses of metformin in patients with inadequately controlled diabetes?

Three hundred and forty-eight patients (mean age 58 years, mean HbA1C 8.8%) who had type 2 diabetes with a fasting plasma glucose (FPG) between 7.7 and 16.7 mmol/L (mean approximately 12 mmol/L) while taking metformin 2.5 g/day, had an insulin secretory capacity (fasting C-peptide concentration) of at least 0.27 nmol/L at screening, a body mass index between 22 to 38 (mean 30.1), and a weight change of no more than 10% between screening and baseline were included in this study.  Exclusion criteria included clinically significant renal or hepatic disease, angina, cardiac insufficiency (NYHA class III or IV), symptomatic diabetic neuropathy, clinically significant ECG abnormalities, abnormal laboratory results, chronic insulin use, participation in any rosiglitazone-related study, or the use of any investigational drug (with the exception of metformin) within 30 days of the study.

Patients were allocated to metformin + placebo (n=116), metformin + rosiglitazone 4 mg/day (n=119), or metformin + rosiglitazone 8 mg/day (n=113) for 26 weeks.  The main outcome measures were the change in HbA1C, insulin sensitivity (HOMA-S), Beta-cell function (HOMA-B), and lipid levels. Safety was also assessed.  

Are the Results of the Study Valid?

1. Was assignment of patients randomized?

Yes.  Randomization was computer generated in a 1:1:1 ratio.

2. Were all patients who entered the trial properly accounted for and attributed at its conclusion?

Yes.  348 patients were randomized and 290 completed the study. 16 patients withdrew due to lack of efficacy, 18 withdrew due to an adverse effect, and 17 withdrew for other reasons.  Seven patients were lost to follow-up.

3. Were patients, their clinicians, and study personnel 'blind' to treatment?

Yes.  Patients, investigators, and sponsors were blind to treatment product and allocation.

4. Were the groups similar at the start of the trial?

Yes.  Baseline characteristics were similar among the treatment groups. There were slightly more men in the control group (metformin and placebo).

5. Aside from the experimental intervention, were the groups treated equally?


6. Overall, are the results of the study valid?


What were the Results?

1. How large was the treatment effect? 

There was a dose dependent statistically significant decrease in HbA1C and FPG at week 26 in both the rosiglitazone 4 mg/day (-0.56% and -1.8 mmol/L, respectively) and rosiglitazone 8 mg/day (-0.78%, and -2.7 mmol/L, respectively) groups when compared with an increase in the control group (0.45% and 0.33 mmol/L, respectively).  Thus, when compared to placebo, the HbA1C after 26 weeks of therapy was 1% and 1.2% lower in the rosiglitazone 4 mg and 8 mg groups, respectively. 

There was a statistically significant increase (Absolute Risk Difference (ARD)) in the total cholesterol (4mg, 0.53 mmol/L; 8mg, 0.60 mmol/L), HDL (4mg, 0.08 mmol/L; 8mg,0.10mmol/L), and LDL levels (4mg, 0.36 mmol/L; 8mg 0.40 mmol/L) when both doses of rosiglitazone were compared to the control group.  HOMA-S and HOMA-B values also increased in a dose-dependent manner with the addition of rosiglitazone to metformin.  

Metformin plus rosiglitazone 4 mg/day, 8 mg/day or placebo 
for Type 2 Diabetes after 26 weeks of therapy 

% of patients achieving outcome 

  1% decrease-HgA1C  HbA1C < 8% FPG-7.7mmol/L
Metformin + Placebo 7.0% 35.9%  8.9% 
Metformin + Rosiglitazone 4mg/d 32.8% - 19.8%
Metformin + Rosiglitazone 8mg/d 37.2% 57.3% 32.6%
NNT (Metformin + Rosiglitazone 4mg/d) 3.9 - 9.2
NNT (Metformin + Rosiglitazone 8mg/d) 3.3 4.7 4.2

The incidence of side effects was not different among all three treatment groups with approximately 76% of patients experiencing at least one adverse event.  However, both rosiglitazone groups experienced small decreases in hemoglobin (4mg, -5.0g/L; 8mg, -8.0g/L) and hematocrit, more edema (4 mg, 2.5%; 8 mg, 3.5%; control, 0.9%) and more weight gain (control, -1.2kg; 4mg, 0.7kg; 8mg, 1.9kg) as compared to the placebo arm.  No one in the rosiglitazone arms experienced a clinically significant increase in ALT (i.e. greater than 3 times the upper limit of normal) during the study. Mean changes in AST, total bilirubin, and alkaline phosphatase were not different among groups.

2. How precise was the estimate of the treatment effect?

The results are statistically significant in favour of the treatment groups when comparing treatment to control groups.  Confidence intervals (CI) and standard deviations are not provided in the text.  However results provided in the figures show that compared to control, the addition of rosiglitazone to metformin produces a statistically different in HgA1c (p<0.001)and FPG (p<0.001).  The 95% confidence intervals provided on the graph examining changes in HgA1c for both rosiglitazone groups compared to control after 26 weeks of therapy demonstrate that even the lower ends of the 95% CI show that a minimal clinically important difference in HgA1c has occurred for both rosiglitazone groups compared with the control group (lower ends of the 95% CI shows approximately a 0.65% reduction (4 mg) and 0.8% reduction (8 mg) in HgA1c).

Will the Results Help Me in Caring for My Patients?

1. Can the results be applied to my patient care?

Yes.  The patients in this study were obese patients with Type 2 diabetes who had high FPG despite monotherapy with metformin. This patient population is very representative of patients with type 2 diabetes.

2. Were all clinically important outcomes considered?

No.  Surrogate endpoints were measured in this study. There were no long term macrovascular or microvascular outcomes considered.

3. Are the likely treatment benefits worth the potential harms and costs?

Yes.  Improving blood glucose control (with drug monotherapy) even by a reduction in HgA1c of 0.9% can reduce the risk of any diabetes related endpoint (including microvascular and macrovascular endpoints) by 12% relative to diet therapy alone (NNT =20). (1)  The adverse effects associated with rosiglitazone (edema, weight gain, headache, reduction in hemoglobin, upper respiratory tract infections, and increases in LDL and cholesterol) can be detected with careful monitoring and therefore dealt with if they prove to be bothersome or clinically detrimental to the patient.


With the introduction of the new thiazolidinediones to the market, there are more options available for the treatment of type 2 diabetes.  In the UKPDS trial the investigators demonstrated that, despite intensive treatment, monotherapy often failed to maintain optimal glycemic control. (1)  Many patients (39%) required the addition of insulin or metformin to maintain control.  The results of the current study and the study by Wolffenbuttal and colleagues (involving rosiglitazone in combination with sulphonylureas) suggests that rosiglitazone can improve glycemic control when used in combination therapy in patients who have experienced secondary failure to one oral hypoglycemic agent. (2)  However, the question remains as to whether these agents are equally or more effective than the current standard therapy (metformin plus a sulfonylurea). (3)  

It is interesting to examine the effects of the combination of a sulfonylurea plus metformin versus metformin alone.  In one trial, 22% of patients on combination therapy achieved a FPG of 7.7 mmol/L or less, compared to 3% in the metformin monotherapy group after 29 weeks of treatment. (4)  Although direct comparisons cannot be made between these results and the results from Fonseca et al, it may be hypothesized that the effects would be similar between the two combinations (i.e. metformin + sulfonylurea vs. metformin + rosiglitazone). To date, there have been no trials comparing the effects of the combination of metformin and a sulfonylurea versus the combination of a thiazolidinedione and either metformin or a sulfonylurea.  Another question that remains is whether patients who have failed therapy with a combination of metformin and a sulfonylurea would benefit from the addition of a thiazolidinedione.  Further investigations need to be performed to assess whether rosiglitazone (and its counterpart pioglitazone) have a more extensive place in therapy.


1. UKPDS Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33).  Lancet 1998;352: 837-53.
2. Wolffenbuttel B, Gomis R, Squatrito S, et al. Addition of low-dose rosiglitazone to sulphonylurea therapy improves glycaemic control in Type 2 diabetic patients. Diab Med 2000;17:40-7.
3. Meltzer S, Leiter L, Daneman D, et al. 1998 clinical practice guidelines > for the management of diabetes in Canada. CMAJ 1998;159(8 Suppl):S1-29.
4. DeFronzo R, Goodman A, and the Multicenter Metformin Study Group. Efficacy of metformin in patients with non-insulin-dependent diabetes mellitus. N Engl J Med 1995;333:541-9.

Copyright © 2000 by the Journal of Informed Pharmacotherapy. All rights reserved.