Introduction
Diabetes is a known risk factor for pancreatic cancer (PC), potentially mediated by mechanisms such as hyperinsulinemia, insulin resistance, chronic inflammation, and hyperglycemia. Prediabetes shared many of these pathophysiological features; however, its association with PC remains inconclusive.
Aims & Methods
We aimed to evaluate the association between prediabetes and risk of PC and to determine whether changes in glycemic status modify the subsequent PC risk. This retrospective cohort study used data from a territory-wide electronic healthcare database. Adults (≥18 years) with at least one hemoglobin A1c (HbA1c) or fasting glucose (FG) measurement between January 2007 and December 2022 were included. Baseline glycemic status was classified as: (1) normoglycemia (HbA1c <5.7% or FG <5.6 mmol/L), (2) prediabetes (HbA1c 5.7-6.4% or FG 5.6-6.9 mmol/L), or (3) diabetes (HbA1c ≥6.5% or FG ≥7 mmol/L). The index date was the date of first recorded HbA1c or FG. Patients were followed until PC, pancreatectomy, death, or study end date (December 31, 2022). Exclusion criteria included prior PC, immunoglobulin G4-related disease, and type 1 diabetes. To minimize reverse causality between diabetes and PC, individuals with PC diagnosed within two years of the index date or with less than two years of follow-up were excluded. The primary outcome was incident PC. Covariates included demographic factors, body mass index, smoking status, alcohol use disease, metabolic and cardiovascular comorbidities, diabetes-related complications, and concomitant medications (including aspirin, non-steroidal anti-inflammatory drugs [NSAIDs] statins, and antidiabetic agents). Adjusted hazard ratios (aHRs) were estimated using multivariable Cox regression. Among individuals with normoglycemia or prediabetes at baseline, glycemic trajectories over the first 2 years were categorized as: (1) persistent normoglycemia, (2) reversion from prediabetes to normoglycemia, (3) persistent prediabetes, and (4) progression from prediabetes to diabetes, with PC risk evaluated accordingly.
Results
A total of 1,977,149 individuals were included (median age:55.3 [IQR:45.0-65.0] years; 44.5% male), with 1,295,290 (65.5%) classified as normoglycemia, 469,236 (23.7%) as prediabetes, and 212,623 (10.8%) as diabetes at baseline. Over a median follow-up of 8.8 (IQR:5.4-11.6) years, PC developed in 1,419 (1.26/10,000 person-years) normoglycemic, 859 (2.19/10,000 person-years) prediabetic, and 551 (2.92/10,000 person-years) diabetic individuals. The 14-year cumulative PC incidence among three groups was shown in Figure 1 (log-rank P <0.001). Compared with normoglycemia, prediabetes and diabetes were associated with 27% (aHR:1.27; 95%CI:1.17-1.39) and 56% (aHR:1.56; 95%CI:1.40-1.75) higher PC risk, respectively. Each 1% increase in HbA1c was associated with an 11% increased risk (aHR:1.11; 95%CI:1.07–1.15). At year 2, of those with prediabetes at baseline, 68,268 (14.5%) progressed to diabetes, 81,122 (17.3%) reverted to normoglycemia, and 319,846 (68.2%) remained prediabetic. Among normoglycemics, 81,122 (6.3%) remained persistently normoglycemic. Compared with persistent normoglycemia, PC risk was higher in persistent prediabetes (aHR:1.37; 95%CI:1.22-1.53) and prediabetes progressing to diabetes (aHR:2.22; 95%CI:1.91-2.59), but not in those reverting to normoglycemia (aHR:1.17; 95%CI:0.96-1.42).
Conclusion
Prediabetes was independently associated with an increased risk of PC compared with normoglycemia. Early intervention to revert prediabetes to normoglycemia may be a modifiable oncopreventive strategy to reduce PC risk.