Series: Landmark Trials in Endocrinology & Metabolism | Study #19
Category: Type 2 Diabetes · Glycaemic Control | Design: Multicentre, randomised controlled trial | n: 3,867 | Follow-up: 10 years (median)
📋 Summary
Authors: UK Prospective Diabetes Study (UKPDS) Group
Journal: Lancet 1998;352:837–853 | PMID: 9742976
UKPDS 33 was the glycaemic control arm of the United Kingdom Prospective Diabetes Study, the largest long-term trial of glycaemic therapy in newly diagnosed type 2 diabetes conducted in its era. A total of 3,867 patients with newly diagnosed type 2 diabetes and a mean fasting plasma glucose of 6.1 to 15.0 mmol/L after three months of dietary treatment were randomised to intensive blood-glucose control with either sulphonylurea (chlorpropamide, glibenclamide, or glipizide) or insulin, targeting a fasting plasma glucose below 6 mmol/L, or to conventional therapy with diet alone, with drug added only for symptoms or fasting glucose above 15 mmol/L. Over 10 years, the mean HbA1c was 7.0% in the intensive group versus 7.9% in the conventional group. Three aggregate endpoints were pre-specified. Intensive therapy reduced the risk of any diabetes-related endpoint by 12% (95% CI 1 to 21%; p=0.029), driven predominantly by a 25% reduction in microvascular endpoints including retinal photocoagulation (7 to 40%; p=0.0099). There was no significant reduction in diabetes-related deaths (10% reduction; −11 to 27%; p=0.34) or all-cause mortality (6% reduction; −10 to 20%; p=0.44). No significant differences in outcomes were observed between sulphonylurea and insulin within the intensive arm. Intensive therapy was associated with more weight gain (mean +2.9 kg) and more hypoglycaemia.
📊 Key Findings
| Endpoint | Intensive | Conventional | Effect Size |
|---|---|---|---|
| Any diabetes-related endpoint | — | — | 12% risk reduction (1–21%) · p=0.029 |
| Microvascular endpoints | — | — | 25% reduction (7–40%) · p=0.0099 |
| Diabetes-related death | — | — | 10% reduction (−11 to 27%) · NS |
| All-cause mortality | — | — | 6% reduction (−10 to 20%) · NS |
| Any MI | — | — | 16% reduction (−2 to 31%) · p=0.052 (trend) |
| Mean HbA1c achieved | 7.0% | 7.9% | Difference −0.9% |
| Major hypoglycaemia (insulin group) | 1.8%/year | 0.7%/year | Increased with intensive therapy |
💬 Expert Commentary
UKPDS 33 was the foundational trial establishing that glycaemic control in type 2 diabetes reduces microvascular complications, and it did so in a newly diagnosed population followed for a decade — a design that captured the full natural history of early type 2 diabetes and the accumulating effect of modest glycaemic improvements over time. The 25% reduction in microvascular endpoints at a mean HbA1c difference of just 0.9 percentage points is consistent with the epidemiological modelling from the Stratton et al. analysis of the UKPDS cohort, which showed a continuous log-linear relationship between HbA1c and microvascular risk throughout the range studied. The absence of statistically significant macrovascular benefit in UKPDS 33 has been extensively debated. Several interpretations are consistent with the data: the macrovascular event rate in a newly diagnosed cohort over 10 years may be insufficient to power the trial for cardiovascular endpoints; the HbA1c difference achieved (0.9%) may be too modest to produce detectable macrovascular benefit within the trial timeframe; and, as the UKPDS post-trial follow-up would later demonstrate, macrovascular benefits may require longer observation to become statistically apparent.
The finding that sulphonylureas and insulin produced equivalent glycaemic and clinical outcomes was reassuring in the context of concerns about sulphonylurea-related cardiovascular harm that had persisted since the University Group Diabetes Programme trial of the 1960s. UKPDS 33 did not support a class-specific cardiovascular hazard with sulphonylureas in newly diagnosed type 2 diabetes, a conclusion that was later reinforced by multiple observational studies and other randomised data. The post-trial UKPDS legacy analysis published 10 years later demonstrated that the microvascular benefit persisted for at least a decade after trial closure with near-complete glycaemic convergence between groups, mirroring the metabolic memory concept established in DCCT/EDIC for type 1 diabetes and establishing that the same phenomenon operates in type 2 disease.
Limitations: The trial was not powered for macrovascular endpoints as a primary outcome. The glycaemic separation achieved (0.9%) was modest by modern standards. Multiple secondary comparisons within the study require cautious interpretation. The study was funded by the UK Medical Research Council, British Diabetic Association, and industry partners.
🔑 BOTTOM LINE
UKPDS 33 established that intensive blood-glucose control with sulphonylurea or insulin reduces microvascular complications by 25% in newly diagnosed type 2 diabetes over 10 years, with a 0.9% HbA1c difference, confirming the glucose hypothesis for type 2 diabetes microvascular disease and motivating intensive glycaemic management from diagnosis in this population.
⭐ Clinical Impact Rating: ●●●●● Practice-defining
