Long-term patterns of fasting blood glucose levels and pancreatic cancer incidence

Abstract

Background Whether type 2 diabetes is cause or consequence, or both, of pancreatic cancer (PaC) remains unresolved. Leveraging repeated measurements of fasting blood glucose (FBG), we examined the temporal relationship between hyperglycemia and PaC incidence. Methods We conducted a nested case-control study of 278 cases and 826 matched-controls from the Korean National Health Insurance Service-Health Screening Cohort. Over 11 years before index date (date of PaC diagnosis for cases), all participants had at least one FBG measurement in each of the three time windows: -11 to -8, -7 to -4, and -3 to 0 years. Using conditional logistic regression, we estimated odds ratios(ORs) of PaC and 95% confidence intervals (CIs) for hyperglycemia in the overall period and at each interval; for major long-term patterns of FBG across the three intervals (recent-onset, medium-term, and long-standing hyperglycemia). Results Higher FBG over the past 11 years was associated with an increased odds of PaC (p(trend) < .0001), with recent FBG more predictive of PaC than distant FBG. By FBG assessed in the -3 to 0 interval, OR was 1.97 (95% CI 1.32-2.93) for 110-125 mg/dL and 3.17 (95% CI 2.09-4.80) for >= 126 mg/dL. By long-term patterns of FBG, compared to consistent normoglycemia, OR was 2.02 (95% CI 1.24-3.31) for long-standing hyperglycemia and 3.38 (95% CI 1.87-6.13) for recentonset hyperglycemia. These associations were more pronounced among never-smokers than ever-smokers (p(interaction) = .06). Conclusion Recent-onset hyperglycemia may be an early manifestation of undetected PaC, while long-lasting hyperglycemia may serve as a moderate etiologic factor for PaC.