Introduction
Diabetes is associated with increased pancreatic cancer (PC) risk. Beta-blockers may mitigate PC risk and progression by inhibiting β-adrenergic receptor-mediated signaling pathways involved in tumor cell proliferation, invasion, and angiogenesis. However, the potential oncopreventive effect of beta-blockers on PC risk in patients with type 2 diabetes (T2D) remains unclear.
Aims & Methods
We aimed to assess the association between beta-blocker use and PC risk in newly diagnosed T2D patients after adjusting for risk factors. This territory-wide retrospective cohort study included adults (≥18 years) newly diagnosed with T2D between 2001 and 2020. Patients with type 1 diabetes, prior PC, immunoglobulin G4-related disease, pancreatectomy, or fewer than two HbA1c results were excluded. To minimize reverse causality between PC and T2D, those diagnosed with PC within 1 year of T2D diagnosis or with <1 year of follow-up were excluded. Primary exposure was beta-blocker use (e.g., acebutotlol, atenolol, bisoprolol, carvedilol, celiprolol, esmolol, labetalol, metoprolol, nadolol, nebivolol, pindolol, propranolol, and sotalol). In the primary (time-fixed) analysis, exposure was defined as ≥90 days of use after T2D diagnosis (i.e. index date). The outcome was incident PC. Patients were followed until PC, pancreatectomy, death, or study end (December 31, 2020). Adjusted hazard ratios (aHRs) were derived from multivariable Cox models adjusting for 31 covariates: age, sex, smoking status, alcohol use disorder, body index mass, glycemic control (in terms of time-weighted mean HbA1c), pancreatitis, cholangitis, colectomy, Charlson’s comorbidity index, metabolic and cardiovascular comorbidities, and concomitant medications (aspirin, non-steroidal anti-inflammatory drugs, statins, and other antidiabetics). Sensitivity analyses included time-dependent analysis (drug use was defined as ≥180 days/year; to address immortal time bias), and multiple propensity score (PS)-based approaches (matching, regression adjustment, weighting, and matching weights). Dose- and duration-response relationships were evaluated.
Results
Among 431,261 T2D patients (mean age: 59.1±11.4 years; male: 52.5%; 61,917 beta-blocker users and 369,344 beta-blocker non-users), 1,303 (0.3%) developed PC during a mean follow-up of 9.4±5.0 years. The 18-year cumulative incidence was 0.68% (95% CI: 0.57%-0.80%) in beta-blocker users and 0.76% (95% CI: 0.70%-0.82%) in non-users (P = 0.02; Figure 1). Beta-blocker use was associated with lowered PC risk in time-fixed analysis (aHR: 0.81; 95% CI: 0.70-0.95), time-dependent analysis (aHR: 0.75; 95% CI: 0.62-0.91), and all PS-based methods (aHR range: 0.83-0.86). A dose- and duration-dependent inverse relationship was observed (Ptrend <0.001). Higher cumulative defined daily dose (cDDD) reduced the risk (aHR: 0.84, 95% CI: 0.72-0.98, for <4,000 cDDD; aHR: 0.31, 95% CI: 0.12-0.82, for ≥4,000 cDDD). Longer use also lowered the risk (aHR: 1.12, 95% CI: 0.92-1.37, for <5 years; aHR: 0.70, 95% CI: 0.57-0.87, for 5 to <15 years; aHR: 0.11, 95% CI: 0.04-0.35, for ≥15 years). Each additional 1,000 cDDD and 1-year increase in use was associated with a 21% (aHR: 0.79, 95% CI: 0.72-0.88) and 6% (aHR: 0.94, 95% CI: 0.92-0.96) risk reduction, respectively.
Conclusion
Beta-blocker use was associated with a lowered risk of PC among patients with newly diagnosed T2D, in a dose- and duration-dependent manner.