UKPDS
(1998)Objective
Metformin - To evaluate whether intensive blood glucose control with metformin reduces diabetes-related complications, including stroke, in overweight patients with type 2 diabetes.
Study Summary
• Metformin reduced stroke risk more than sulfonylurea, insulin, or diet control.
Intervention
Randomized controlled trial involving 1704 overweight patients with newly diagnosed type 2 diabetes. Compared intensive blood-glucose control using metformin versus conventional therapy (mainly diet). A secondary analysis compared metformin with intensive therapies (chlorpropamide, glibenclamide, or insulin). Median follow-up was 10.7 years.
Inclusion Criteria
Overweight (>120% ideal body weight), age ~53, newly diagnosed type 2 diabetes, fasting plasma glucose 6.1–15.0 mmol/L after 3 months of dietary run-in.
Study Design
Arms: Metformin vs. Conventional (Diet); also Metformin vs. Sulfonylurea/Insulin (secondary analysis)
Patients per Arm: Metformin: 342; Conventional: 411
Outcome
• All-cause mortality: 36% lower with metformin; RR 0.64 (95% CI 0.45–0.91), p=0.011.
• Stroke: Metformin had lower risk vs. both conventional and other intensive therapies; p=0.032 for comparison vs. sulfonylureas/insulin.
• Non-fatal stroke: RR 0.42 (95% CI 0.12–1.45) vs. conventional; trend toward benefit.
• Fatal/non-fatal stroke: 12 events in metformin group vs. 23 in conventional group (3.3 vs. 5.5 per 1000 patient-years).
Bottom Line
Intensive blood-glucose control with metformin in overweight patients with newly diagnosed type 2 diabetes significantly reduced the risk of any diabetes-related endpoint, diabetes-related death, and all-cause mortality, with less weight gain and fewer hypoglycaemic attacks compared to sulphonylureas or insulin, suggesting it may be the preferred first-line pharmacological therapy in these patients. However, early addition of metformin to sulphonylurea in patients already on maximum dose was associated with an increased risk of diabetes-related death.
Major Points
- 753 overweight patients with newly diagnosed type 2 diabetes were randomized to conventional diet policy (n=411) or intensive control with metformin (n=342) for a median of 10.7 years.
- Median glycated haemoglobin (HbA1c) was 7.4% in the metformin group vs 8.0% in the conventional group.
- Metformin, compared to conventional group, showed risk reductions of 32% (95% CI 13-47, p=0.002) for any diabetes-related endpoint, 42% (9-63, p=0.017) for diabetes-related death, and 36% (9-55, p=0.011) for all-cause mortality.
- Among intensive blood-glucose control patients, metformin showed a greater effect than chlorpropamide, glibenclamide, or insulin for any diabetes-related endpoint (p=0.0034), all-cause mortality (p=0.021), and stroke (p=0.032).
- A supplementary trial showed early addition of metformin in sulphonylurea-treated patients was associated with an increased risk of diabetes-related death (96% increased risk [95% CI 2-275], p=0.039) compared to continued sulphonylurea alone.
- A combined analysis of diet/metformin and sulphonylurea/metformin studies showed fewer metformin-allocated patients having diabetes-related endpoints (risk reduction 19% [2-33], p=0.033).
- Metformin was associated with less weight gain and fewer hypoglycaemic attacks than insulin and sulphonylureas.
Study Design
- Study Type
- Randomised controlled trial (main trial) and supplementary randomised controlled trial
- Randomization
- Yes
- Blinding
- No placebo was given, implying open-label for participants and investigators. Outcomes were aggregated to minimize statistical tests.
- Sample Size
- 1704
- Follow-up
- Median 10.7 years (main trial); median 6.6 years (supplementary trial)
- Centers
- 23
- Countries
- UK
Primary Outcome
Definition: Aggregates of any diabetes-related clinical endpoint (sudden death, death from hyperglycaemia or hypoglycaemia, fatal or non-fatal myocardial infarction, angina, heart failure, stroke, renal failure, amputation [of at least one digit], vitreous haemorrhage, retinopathy requiring photocoagulation, blindness in one eye, or cataract extraction), diabetes-related death, and all-cause mortality.
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| Conventional: 43.3 events per 1000 patient-years for any diabetes-related endpoint; 12.7 for diabetes-related death; 20.6 for all-cause mortality. | Metformin: 29.8 events per 1000 patient-years for any diabetes-related endpoint; 7.5 for diabetes-related death; 13.5 for all-cause mortality. | 0.68 (0.53-0.87 (any diabetes-related endpoint); 0.37-0.91 (diabetes-related death); 0.45-0.91 (all-cause mortality)) | 0.0023 (any diabetes-related endpoint); 0.017 (diabetes-related death); 0.011 (all-cause mortality) |
Limitations & Criticisms
- The study design did not allow a direct comparison of phenformin with sulphonylurea, a limitation inherited from the UGDP study.
- The supplementary trial's finding of increased diabetes-related death with metformin added to sulphonylurea might be due to patient differences (older, more hyperglycaemic, less overweight, shorter follow-up) or chance, rather than a direct negative effect of metformin in this context.
- The epidemiological assessment, while not showing increased mortality with combined sulphonylurea and metformin, had wide confidence intervals, limiting its conclusiveness.
- The primary endpoints were aggregates, which, while minimizing statistical tests, can obscure individual event differences. For single endpoints, 99% CIs were used to account for potential Type 1 errors, which is a very conservative approach and may have missed some effects.
- The study was open-label (no placebo), which could introduce bias in patient management and reporting of non-blinded outcomes, although outcomes were aggregated for analysis.
Citation
Lancet 1998; 352: 854-65