The New England Journal of Medicine February 07, 2002

Reduction in the Incidence of Type 2 Diabetes with Lifestyle Intervention or Metformin

Diabetes Prevention Program Research Group (W.C. Knowler, E. Barrett-Connor, et al.)

Bottom Line

An intensive lifestyle modification program or metformin therapy significantly reduced the incidence of type 2 diabetes in high-risk adults with impaired glucose tolerance compared to placebo, with the lifestyle intervention proving superior to metformin.

Key Findings

1. Over an average follow-up of 2.8 years, the incidence of diabetes was 11.0, 7.8, and 4.8 cases per 100 person-years in the placebo, metformin, and lifestyle groups, respectively [1.1.3].
2. The intensive lifestyle intervention reduced the incidence of type 2 diabetes by 58% (95% CI, 48 to 66%) compared with placebo.
3. Metformin reduced the incidence of type 2 diabetes by 31% (95% CI, 17 to 43%) compared with placebo.
4. The lifestyle intervention was significantly more effective than metformin.
5. To prevent one case of diabetes during a period of three years, 6.9 persons would have to participate in the lifestyle-intervention program.

Study Design

Design
RCT
Double-Blind
Sample
3,234
Patients
Duration
2.8 yr
Median
Setting
Multicenter, US
Population Adults without diabetes who had elevated fasting (5.3-6.9 mmol/L) and post-load plasma glucose concentrations (impaired glucose tolerance) and a BMI of 24 or higher.
Intervention Intensive lifestyle modification program (goal of ≥7% weight loss and ≥150 minutes of physical activity per week) OR Metformin (850 mg twice daily).
Comparator Placebo plus standard lifestyle recommendations.
Outcome Incidence of type 2 diabetes.

Study Limitations

The intensive lifestyle intervention was resource-intensive (involving a 16-lesson curriculum and frequent one-on-one counseling), making it difficult to replicate and sustain in standard real-world clinical practice.
The lifestyle modification arm was inherently unblinded, which could introduce behavioral and assessment biases, although hard biochemical endpoints for diabetes diagnosis mitigate this risk.
Gastrointestinal symptoms were significantly more common in the metformin group, which may have affected adherence and partially unblinded participants in that arm.

Clinical Significance

The Diabetes Prevention Program (DPP) was a landmark trial that definitively established type 2 diabetes as a preventable condition in high-risk individuals. By demonstrating that an intensive lifestyle intervention was almost twice as effective as metformin (58% vs. 31% risk reduction), it cemented structured diet and exercise programs as the gold standard, first-line management strategy for prediabetes. These findings catalyzed the creation of widespread, community-based diabetes prevention programs worldwide.

Historical Context

Prior to 2002, the escalating global prevalence of type 2 diabetes was recognized as a major public health crisis, but evidence for prevention was largely observational or derived from smaller, homogenous populations (such as the Finnish Diabetes Prevention Study). The role of pharmacological interventions, particularly insulin sensitizers like metformin, in preventing the progression from impaired glucose tolerance to overt diabetes was also theoretical. The DPP provided robust, multicenter, diverse-population data that fundamentally shifted the paradigm from solely treating diabetes to actively preventing it through structured behavioral modifications.

Guided Discussion

High-yield insights from every perspective

Med Student
Medical Student

How does the mechanism of action of metformin compare to the physiological changes induced by weight loss and exercise in preventing the progression from impaired glucose tolerance to type 2 diabetes?

Key Response

This tests foundational knowledge of pathophysiology. Metformin primarily works by reducing hepatic glucose production and improving peripheral insulin sensitivity via AMPK activation. In contrast, lifestyle modifications (weight loss and exercise) directly reduce visceral adiposity, decrease systemic inflammation, and acutely upregulate GLUT4 translocation in skeletal muscle independent of insulin, which profoundly preserves beta-cell function and insulin sensitivity.

Resident
Resident

Based on the DPP subgroup analyses, which specific patient populations with prediabetes are most likely to benefit from starting metformin, and how does age or BMI influence your choice between metformin and lifestyle modifications?

Key Response

This focuses on clinical application and management. While lifestyle intervention was universally effective across all subgroups in the DPP, metformin was significantly more effective in younger patients (under age 60) and heavier patients (BMI 35 or greater), and relatively ineffective in older or leaner patients. This directly informs targeted pharmacological therapy.

Fellow
Fellow

The DPP established the superiority of lifestyle interventions over metformin, but the subsequent DPPOS (Outcomes Study) revealed a gradual decline in the preventative effect over 15 years. How should we integrate this evidence regarding the long-term durability of beta-cell preservation with the emergence of highly effective incretin therapies (GLP-1 RAs/GIPs) for obesity and prediabetes?

Key Response

Fellows must integrate historical landmark trials with modern advances. The rationale addresses the fundamental progressive nature of beta-cell failure in prediabetes. While DPP interventions delayed onset, they did not halt beta-cell decline permanently. Comparing this to the potential disease-modifying effects of modern incretin therapies encourages nuanced interpretation of treatment durability.

Attending
Attending

Given the overwhelmingly positive results of the 16-lesson intensive lifestyle intervention in the DPP, what are the systemic barriers to implementing such resource-heavy programs in routine primary care, and how can we adapt these findings into practical, sustainable models for our patients?

Key Response

Attendings must consider practice-changing implications and healthcare delivery. The DPP's lifestyle arm involved intense, individualized coaching which is costly and hard to scale. Translating clinical trial efficacy to real-world effectiveness involves discussing group classes, telehealth, insurance reimbursement (like the Medicare Diabetes Prevention Program), and systemic practice redesign.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD
PhD

The DPP trial was terminated one year early by the Data and Safety Monitoring Board due to the overwhelming efficacy of the lifestyle intervention. How does early trial termination for benefit potentially introduce 'random high' bias in the estimation of treatment effect size, and what statistical penalties or Bayesian approaches should be utilized to adjust for this?

Key Response

PhD-level researchers must critique study design and statistical methodology. Truncated trials often overestimate the magnitude of treatment effects because they stop at a 'peak' of statistical significance. Exploring the methods to adjust for alpha spending and the implications for the reported 58% risk reduction highlights advanced epidemiological concepts.

Journal Editor
Journal Editor

In evaluating the internal validity of the DPP, how might the lack of blinding in the lifestyle intervention arm have differentially influenced participant attrition, self-reported physical activity outcomes, and healthcare-seeking behaviors compared to the double-blinded metformin and placebo arms?

Key Response

A journal editor focuses on critical appraisal and threats to validity. Behavioral interventions cannot be blinded, which introduces performance and detection bias. The 'Hawthorne effect' in the lifestyle group could drive differential adherence or reporting, requiring the reviewer to critically assess how the authors mitigated and reported these inherent biases.

Guideline Committee
Guideline Committee

The DPP demonstrated a Number Needed to Treat (NNT) of 7 for lifestyle intervention and 14 for metformin over 3 years. How do these findings directly shape current ADA Standards of Care regarding prediabetes, and should current guidelines be updated to mandate structured lifestyle programs as a prerequisite before pharmacological reimbursement is authorized?

Key Response

Guideline committees translate evidence into formal recommendations. Current ADA guidelines (Level A evidence) strongly recommend intensive lifestyle behavior change programs for all patients with prediabetes, while recommending metformin specifically for those with BMI >=35, age <60, or prior GDM (directly mirroring DPP findings). The question pushes the committee to evaluate the strength of evidence against healthcare economic policies.

Clinical Landscape

Noteworthy Related Trials

1997

Da Qing IGT and Diabetes Study

n = 577 · Diabetes Care

Tested

Clinic-based diet, exercise, or diet plus exercise interventions

Population

Adults with impaired glucose tolerance

Comparator

Standard clinic care

Endpoint

Incidence of type 2 diabetes

Key result: Over 6 years, diet, exercise, and diet plus exercise interventions reduced the risk of developing diabetes by 31 percent, 46 percent, and 42 percent, respectively.
2001

Finnish Diabetes Prevention Study

n = 522 · NEJM

Tested

Intensive lifestyle intervention with individualized diet and exercise counseling

Population

Overweight middle-aged subjects with impaired glucose tolerance

Comparator

Standard care control group

Endpoint

Incidence of type 2 diabetes

Key result: The cumulative incidence of diabetes was 11 percent in the intervention group and 23 percent in the control group, representing a 58 percent reduction in risk.
2002

STOP-NIDDM Trial

n = 1429 · Lancet

Tested

Acarbose 100 mg three times daily

Population

Patients with impaired glucose tolerance

Comparator

Placebo

Endpoint

Development of type 2 diabetes

Key result: Acarbose significantly reduced the risk of progressing to type 2 diabetes by 25 percent compared to placebo over an average of 3.3 years.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis