Highlights
• In a 20-year linked-records cohort of 328,049 adults in England with newly diagnosed prediabetes, most younger people (<55 years) remained in the prediabetes state over 10 years, while older adults had higher probabilities of progression to type 2 diabetes, cancer, and death.
• Cancer incidence after prediabetes increased with age and was only modestly higher among those who developed type 2 diabetes versus those who remained with prediabetes (maximum ≈4–5 additional cases per 1,000 person-years over 10 years).
• Sociodemographic and lifestyle factors (BMI, smoking, socioeconomic status, ethnicity) influenced state occupancy probabilities and sojourn times, supporting targeted prevention and screening strategies keyed to age and risk profile.
Background
Prediabetes—defined by blood glucose above the normal range but below thresholds for diabetes—affects a large and growing proportion of adults worldwide. It is clinically important as a risk state for progression to type 2 diabetes and for cardiovascular disease. Observational data have also linked higher glucose concentrations and diabetes with an elevated risk of several cancers, but the temporal and competing-risk relationships between prediabetes, conversion to type 2 diabetes, incident cancer, and death have been incompletely characterized at a population level.
The Lancet Diabetes & Endocrinology study by Zaccardi and colleagues reports long-term, multistate trajectories in a large English cohort with newly recorded prediabetes, providing a richer, time-updated view of how individuals move between prediabetes, type 2 diabetes, cancer, and death over two decades. This analysis can help clinicians and policymakers better calibrate prevention, surveillance, and resource planning according to age and modifiable risk factors.
Study design and methods
The authors used the Clinical Practice Research Datalink (CPRD) primary care records linked to hospital and mortality data to identify adults aged 18–100 years with a new-record diagnosis of prediabetes between Jan 1, 1998 and Nov 30, 2018 in England. Individuals were followed from prediabetes diagnosis until death or end of follow-up (Nov 30, 2018). Two intermediate outcomes were modelled: (1) diagnosis of type 2 diabetes and (2) diagnosis of cancer. A multistate modelling framework estimated state occupancy probabilities and sojourn (length of stay) times across eight combined states and seven possible transitions (for example: prediabetes → type 2 diabetes; prediabetes → cancer; prediabetes → type 2 diabetes → cancer; and transitions to death from each state).
Analyses were stratified by sex and age at prediabetes diagnosis (<55, 55–64, 65–74, ≥75 years) and assessed the influence of BMI, smoking, socioeconomic status, and ethnicity on trajectory metrics. Median follow-up was 7.7 years, and the cohort included 328,049 persons with 163,782 transitions observed.
Key findings
Population-level occupancy and transition probabilities over 10 years after a prediabetes diagnosis varied markedly by age.
Primary quantitative outcomes reported:
• Probability of remaining in the prediabetes state at 10 years: ranged from 23.2% (men aged ≥75 at diagnosis) to 72.1% (men aged <55 at diagnosis).
• Probability of death following prediabetes at 10 years: 1.2% (women aged <55) up to 38.7% (women aged ≥75).
• Probability of developing type 2 diabetes and remaining in that state at 10 years: 7.9% (men aged ≥75) to 24.0% (women aged <55).
• Probability of developing cancer and remaining in that state at 10 years: ranged in men and women from 1.9% (men aged <55) to 7.8% (men aged ≥65 to <75 years).
Cancer incidence rates were higher in older age groups and only marginally higher among those who developed type 2 diabetes compared with those who remained in prediabetes. Over a 10-year horizon, the largest observed differences were 4.1 more cancer cases per 1,000 person-years in women and 4.8 more per 1,000 person-years in men for those who developed type 2 diabetes versus those who did not.
Sojourn (expected duration spent in each state during the 10-year window) also varied by age. For example, men aged <55 spent a mean of 8.34 years in the prediabetes state during the decade following diagnosis, while men aged ≥75 spent about 5.34 years.
Sociodemographic and lifestyle factors meaningfully affected trajectories: higher BMI, current smoking, greater socioeconomic deprivation, and some ethnic groups were associated with higher probabilities of progression to type 2 diabetes or cancer and altered sojourn times.
Clinical interpretation of the magnitude
The observed absolute differences in cancer incidence associated with conversion to type 2 diabetes were relatively small on an annual-person-years basis but can be clinically meaningful when aggregated across populations. For an individual patient, age and competing risks (notably mortality) dominate the trajectory. Younger people with prediabetes are likely to remain in that state for years, creating a window for primary prevention focused on lifestyle and metabolic risk modification. Older adults have higher absolute risks of cancer and death, and competing mortality reduces time at risk for progression to diabetes.
Strengths of the study
• Large, population-based sample with linkage across primary care, hospital, and mortality records and long follow-up (up to 20 years), increasing the external validity for the English population.
• Use of multistate models to explicitly account for competing risks and sequential transitions (eg, prediabetes → diabetes → cancer), yielding clinically intuitive metrics such as occupancy probabilities and sojourn times.
• Stratification by age and assessment of modifiable covariates helps inform targeted prevention strategies.
Limitations and sources of potential bias
• Observational design cannot establish causality. Residual confounding (including by unmeasured variables such as diet, physical activity, and screening behaviour) may influence associations.
• Prediabetes ascertainment relied on routine clinical records; misclassification is possible (some individuals with dysglycaemia may have been undiagnosed or recorded inconsistently over time). Practices and diagnostic thresholds changed over the study period (1998–2018), which could affect incidence and transition estimates.
• Detection bias is possible: individuals developing diabetes may have greater clinical contact and diagnostic work-up, potentially increasing cancer detection rates compared with those remaining in prediabetes.
• The analysis reports aggregate cancer incidence; cancer is heterogeneous—risk trajectories likely differ by cancer site (eg, liver, pancreatic, colorectal, breast) and by stage at diagnosis. Cause-specific analyses are needed to tailor screening recommendations.
• Generalizability outside England and similar healthcare contexts may be limited.
Expert commentary and clinical implications
This study refines our understanding of the natural history after a prediabetes diagnosis by quantifying competing trajectories to diabetes, cancer, and death. For clinicians, the principal, actionable insight is the heterogeneity by age and modifiable risk factors:
• Younger adults with prediabetes (<55 years) typically have a prolonged period in the prediabetes state, representing an opportunity for intensive lifestyle interventions, weight management, and pharmacologic prevention (where indicated) to prevent progression and reduce long-term cardiometabolic and possibly cancer risk.
• Older patients carry higher immediate absolute risks of cancer and mortality; discussions should prioritize comorbidities, appropriate cancer screening per age-based guidelines, and careful consideration of the benefit-risk balance for diabetes prevention interventions.
• The modest incremental cancer risk associated with progression to type 2 diabetes suggests that prevention strategies should remain focused on established benefits (reducing diabetes incidence, cardiovascular risk, and microvascular complications), while recognizing potential downstream benefits for cancer prevention via weight loss, reduced hyperinsulinaemia, and improved metabolic health.
• Sociodemographic and lifestyle determinants influenced trajectories substantially; integrated approaches (addressing obesity, smoking cessation, and social determinants) are likely to be most effective. From a policy perspective, linking diabetes prevention programmes to cancer screening and risk-reduction services could improve efficiency.
Research and policy implications
Key next steps include:
• Cause-specific cancer analyses to identify which malignancies are most affected by transitions from prediabetes to diabetes, and whether stage at diagnosis differs.
• Interventional trials and implementation studies that evaluate whether diabetes prevention interventions (lifestyle programs, GLP-1 receptor agonists or other pharmacotherapies where evidence supports use) reduce cancer incidence or improve cancer-related outcomes.
• Health-economic evaluations to guide age- and risk-stratified allocation of prevention resources, and to assess the value of integrated prevention-screening pathways.
• Greater attention to equity and access—programmes should address barriers for socioeconomically deprived and ethnically diverse groups who bear disproportionate risks.
Conclusion
Zaccardi et al. provide an important, population-level picture of post-prediabetes trajectories over up to 20 years. Age at prediabetes diagnosis is the dominant determinant of subsequent occupancy in prediabetes, progression to type 2 diabetes, incident cancer, and death. Younger individuals frequently remain in prediabetes for many years, presenting a key prevention window, whereas older adults have higher competing risks from cancer and mortality. Clinical strategies should be age-stratified and address BMI, smoking, and social determinants to reduce progression to diabetes and to optimize early identification and prevention of cancer in people with dysglycaemia.
Funding and clinicaltrials.gov
Funding reported by the study: National Institute for Health and Care Research (NIHR) Applied Research Collaboration, East Midlands; NIHR Leicester; British Retail Consortium; Hope Against Cancer. ClinicalTrials.gov: Not applicable (observational cohort study).
References
1. Zaccardi F, Ling S, Gillies C, Brown K, Davies MJ, Khunti K. Trajectories of type 2 diabetes and cancer in 330 000 individuals with prediabetes: 20-year observational study in England. Lancet Diabetes Endocrinol. 2025 Nov 19:S2213-8587(25)00285-2. doi: 10.1016/S2213-8587(25)00285-2. Epub ahead of print. PMID: 41274303.
2. American Diabetes Association. Standards of Medical Care in Diabetes—2024. Diabetes Care. 2024;47(Suppl 1):S1–S200.
Note: Readers should consult the full Lancet paper for detailed methods, supplementary analyses, and site-specific cancer results when available.
