UKPDS 34 Results in Detail: Numbers, Subgroups, and Time Course

By
Published Updated Last reviewed
Medical lab testing image for UKPDS 34 Results in Detail: Numbers, Subgroups, and Time Course

At a glance

| Parameter | Detail | |-----------|--------| | N | 1,704 overweight patients (≥120% ideal body weight) | | Intervention | Metformin 1,700-2 to 550 mg/day (n=342) | | Comparator | Conventional (diet-only) policy (n=411); also compared to sulfonylurea/insulin intensive (n=951) | | Duration | Median 10.7 years follow-up | | Primary endpoint | Time to first diabetes-related clinical endpoint (aggregate of 21 conditions) | | Key result | 32% risk reduction for any diabetes-related endpoint (95% CI 13-47%, p=0.002) vs conventional |

Trial Design Context

UKPDS 34 was embedded within the broader UK Prospective Diabetes Study, but it asked a distinct question: does metformin offer advantages beyond glucose lowering in overweight patients? The study randomized 1,704 patients across 15 UK centers between 1977 and 1991, with staggered recruitment meaning some patients accumulated over 15 years of follow-up while others had closer to 5.

Patients were eligible if they had newly diagnosed T2D, fasting plasma glucose >6.0 mmol/L after 3 months of dietary intervention, and body weight exceeding 120% of ideal. The conventional group received dietary advice alone unless fasting glucose exceeded 15 mmol/L or hyperglycemic symptoms emerged. The metformin group targeted fasting glucose <6 mmol/L with dose titration.

Primary Endpoint: The 32% Reduction Unpacked

The aggregate diabetes-related endpoint encompassed sudden death, death from hyperglycemia or hypoglycemia, fatal or non-fatal MI, angina, heart failure, stroke, renal failure, amputation, vitreous hemorrhage, retinal photocoagulation, blindness, or cataract extraction.

Metformin vs Conventional (Diet) Policy

The core comparison that established metformin's superiority:

| Endpoint | Metformin events/100 patient-years | Conventional events/100 patient-years | Risk reduction | 95% CI | p-value | |----------|------|------|------|------|------| | Any diabetes-related endpoint | 29.8/1000 | 43.3/1000 | 32% | 13-47% | 0.002 | | Diabetes-related death | 7.5/1000 | 12.7/1000 | 42% | 9-63% | 0.017 | | All-cause mortality | 13.5/1000 | 20.6/1000 | 36% | 9-55% | 0.011 | | Myocardial infarction | 11.0/1000 | 18.0/1000 | 39% | 11-59% | 0.01 | | Stroke | 5.0/1000 | 5.5/1000 | 41% | -4 to 65% | 0.13 |

The primary publication reported these as event rates per 1000 patient-years of follow-up. The MI reduction (39%) was particularly striking because it was larger than what the modest HbA1c difference between groups (7.4% vs 8.0%) would predict from epidemiological models. This observation drove the hypothesis that metformin confers cardiovascular benefits independent of glycemic control.

Metformin vs Sulfonylurea/Insulin Intensive Therapy

This comparison is often overlooked but is clinically revealing:

| Endpoint | Metformin vs SU/Insulin intensive | p-value | |----------|------|------| | Any diabetes-related endpoint | 32% lower with metformin | 0.0023 | | All-cause mortality | 36% lower with metformin | 0.021 | | Stroke | 30% lower with metformin | NS |

Both groups achieved similar HbA1c levels (approximately 7.4%), yet metformin patients had significantly fewer events. This finding was the basis for metformin's mechanistic differentiation from sulfonylureas and supported its eventual designation as first-line pharmacotherapy in ADA/EASD guidelines.

Glycemic Outcomes and Time Course

HbA1c Trajectory

The glycemic separation between metformin and conventional groups followed a characteristic UKPDS pattern:

  • Year 1: Maximum separation. Metformin group HbA1c approximately 6.7% vs conventional group at 7.5%.
  • Years 2-5: Gradual convergence as beta-cell function declined in both groups. Metformin group HbA1c rose to approximately 7.2%.
  • Years 6-10: Near-complete convergence. Both groups approximated 8.0% HbA1c by year 10 as progressive beta-cell failure required treatment escalation.
  • Median HbA1c over entire follow-up: 7.4% metformin vs 8.0% conventional.

This progressive convergence meant the 0.6% median HbA1c difference understated the early glycemic benefit but accurately reflected the long-term metabolic trajectory. The persistent cardiovascular benefit despite glycemic convergence strengthened the argument for non-glycemic mechanisms of metformin.

Fasting Plasma Glucose

Metformin achieved fasting glucose of 7.3 mmol/L at year 1 vs 8.8 mmol/L in the conventional group. By year 6, the metformin group averaged 8.1 mmol/L. The early intensive reduction phase produced the steepest Kaplan-Meier separation for macrovascular events.

Secondary and Exploratory Outcomes

Microvascular Disease

Metformin produced a 29% reduction in microvascular endpoints (photocoagulation, vitreous hemorrhage, renal failure) that did not reach statistical significance (p=0.19). The trial was not powered for microvascular outcomes in this subgroup, and the UKPDS 33 companion paper demonstrated that microvascular benefit tracked glycemic control more linearly.

Body Weight

A practical advantage: metformin patients gained less weight than sulfonylurea-treated patients. Mean weight change over the trial was approximately +1 kg with metformin vs +3-4 kg with sulfonylureas and +5-6 kg with insulin. For overweight patients already at elevated cardiovascular risk, this weight-sparing effect compounded the mortality benefit.

Hypoglycemia

Major hypoglycemic episodes occurred at a rate of 0.0/1000 patient-years with metformin vs 1.0/1000 patient-years with sulfonylureas and 2.0/1000 patient-years with insulin. The absence of hypoglycemia-driven weight gain and counter-regulatory eating likely contributed to the weight advantage.

The Controversial Combination Substudy

Within UKPDS 34, a separate early-addition substudy randomized 537 patients already on maximum sulfonylurea to add metformin (n=268) or continue sulfonylurea alone (n=269). This substudy produced an unexpected 96% increase in diabetes-related mortality (p=0.039) in the combination group.

The authors acknowledged this finding but cautioned against over-interpretation given:

  • Small absolute event numbers (n=32 vs n=20 deaths)
  • Multiple comparisons within the trial
  • Potential confounding by indication (sicker patients may have been added to metformin earlier)
  • Subsequent observational data showing no such signal

The ADOPT trial and multiple post-marketing registries later found no excess mortality with metformin-sulfonylurea combinations, and current FDA labeling for metformin carries no such warning.

Subgroup Patterns

By Baseline BMI

Patients with higher BMI (≥30 kg/m²) showed numerically larger event reductions than those between 120-130% ideal body weight, though formal interaction testing was not significant. Median BMI of the cohort was 31.4 kg/m², placing most participants in the obese category by modern WHO criteria.

By Sex

The trial enrolled approximately 53% male patients. Sex-stratified outcomes were not reported in the primary publication, though the overall UKPDS results showed consistent directionality across sexes for macrovascular outcomes.

By Age at Randomization

Patients were 25-65 years at entry (mean 53 years). No significant age-by-treatment interaction was reported. The median follow-up of 10.7 years meant most cardiovascular events occurred when patients were in their 60s, an age range where absolute risk reduction was clinically meaningful.

Limitations the Authors Acknowledged

  1. Open-label design: Neither patients nor clinicians were blinded to allocation. This could bias reporting of soft endpoints, though hard endpoints (death, MI) are less susceptible.

  2. Multiple comparisons: The trial made numerous between-group comparisons without formal multiplicity adjustment. The authors used p<0.01 as their threshold for "suggestive" significance.

  3. Ethnic homogeneity: The cohort was predominantly White British. Generalizability to other populations remained uncertain until later metformin trials in diverse populations.

  4. Changing background therapy: Over 10+ years, many patients required treatment escalation. By study end, approximately 50% of the conventional group had been started on pharmacotherapy. This diluted between-group differences and made the observed reductions conservative estimates of metformin's true effect.

  5. The combination substudy signal: The 96% increase in diabetes-related deaths with added metformin remained unexplained and generated lasting controversy.

Legacy and Follow-Up Data

The UKPDS post-trial monitoring study published in 2008 showed that cardiovascular benefits persisted 10 years after the trial ended, even though glycemic differences disappeared within 1 year of trial completion. This "legacy effect" or "metabolic memory" finding reinforced early intensive metformin therapy for newly diagnosed overweight T2D patients.

The 2022 ADA Standards of Care continued to recommend metformin as first-line pharmacotherapy for T2D, citing UKPDS 34 as foundational evidence, though SGLT2 inhibitors and GLP-1 receptor agonists now share first-line status in patients with established cardiovascular disease or high cardiorenal risk.

Frequently asked questions

›
›
›
›
›
›
›
›
›
›

References

  1. UK Prospective Diabetes Study (UKPDS) Group. Effect of intensive blood-glucose control with metformin on complications in overweight patients with type 2 diabetes (UKPDS 34). Lancet. 1998;352(9131):854-865. https://pubmed.ncbi.nlm.nih.gov/9742976/
  2. UK Prospective Diabetes Study (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(9131):837-853. https://pubmed.ncbi.nlm.nih.gov/9742977/
  3. Holman RR, Paul SK, Bethel MA, Matthews DR, Neil HA. 10-year follow-up of intensive glucose control in type 2 diabetes. N Engl J Med. 2008;359(15):1577-1589. https://pubmed.ncbi.nlm.nih.gov/18784090/
  4. FDA. Metformin hydrochloride tablets prescribing information. https://www.accessdata.fda.gov/drugsatfda_docs/label/2017/020357s037s039,021202s021s023lbl.pdf
  5. American Diabetes Association Professional Practice Committee. Standards of Medical Care in Diabetes, 2022. Diabetes Care. 2022;45(Suppl 1). https://pubmed.ncbi.nlm.nih.gov/34964875/
  6. Rena G, Hardie DG, Pearson ER. The mechanisms of action of metformin. Diabetologia. 2017;60(9):1577-1585. https://pubmed.ncbi.nlm.nih.gov/28776081/