Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33)
Randomised trial in 3,867 adults with newly diagnosed type 2 diabetes: a policy of intensive glucose lowering with sulphonylurea or insulin (FPG target <6 mmol/L) versus conventional diet-first care achieved lower median HbA1c (7.0% vs 7.9%) and a 12% relative reduction in any diabetes-related aggregate endpoint over 10 years, driven by microvascular benefit, without significant differences in diabetes-related death or all-cause mortality in the primary summary; hypoglycaemia and weight gain favoured conventional care.
Design
- Population: 3,867 adults with newly diagnosed type 2 diabetes after 3 months diet treatment; median age 54 y (IQR 48–60)
- Arms: intensive policy with sulphonylurea (chlorpropamide, glibenclamide, or glipizide) or insulin vs conventional policy starting with diet alone; drugs in the conventional arm added only for hyperglycaemic symptoms or FPG >15 mmol/L
- Intensive glycaemic aim: FPG <6 mmol/L
Glycaemic separation (10-year medians; abstract)
- HbA1c: 7.0% (6.2–8.2) intensive vs 7.9% (6.9–8.8) conventional (~11% relative reduction in HbA1c)
- No difference in achieved HbA1c among the three intensive pharmacologic options
Aggregate clinical endpoints vs conventional (abstract)
- Any diabetes-related endpoint: 12% lower risk in intensive group (95% CI 1–21%; P = 0.029)
- Microvascular endpoints (major driver): 25% lower risk (95% CI 7–40%; P = 0.0099), including need for retinal photocoagulation
- Diabetes-related death: 10% lower (95% CI −11 to 27%; P = 0.34)
- All-cause mortality: 6% lower (95% CI −10 to 20%; P = 0.44)
Safety / tolerability
- Major hypoglycaemia per year: 0.7% conventional vs 1.0% chlorpropamide, 1.4% glibenclamide, 1.8% insulin (P < 0.0001 on intensive vs conventional for hypoglycaemia burden)
- Weight gain higher on intensive therapy (mean +2.9 kg vs conventional; P < 0.001), largest with insulin (+4.0 kg vs +2.6 kg chlorpropamide vs +1.7 kg glibenclamide)
Evidence hygiene
This is a 1990s pharmacologic glycaemic-policy trial—not a DASH/Mediterranean feeding study and not modern GLP-1 outcome evidence; pair with incident-diabetes prevention trials (DPP) and contemporary CV-outcome glucose-lowering trials when counselling individuals.
Publication
UK Prospective Diabetes Study (UKPDS) Group. Lancet. 1998 Sep 12;352(9131):837-853. PMID 9742976.
Outcomes
- HbA1c (Glycated Hemoglobin)-0.9% (Absolute Change)
- Any diabetes-related aggregate endpoint: 12% lower risk (95% CI 1–21%; P=0.029) for intensive vs conventional policy over 10 years (UKPDS 33).
- Microvascular endpoints: 25% lower risk (95% CI 7–40%; P=0.0099) with intensive vs conventional policy, including retinal photocoagulation need (UKPDS 33 abstract).