The early-onset pancreatic cancer paradox: a systematic review and meta-analysis.

Abstract

Background: Globally, the rate of early-onset pancreatic cancer (EOPC) is increasing, but its clinical features and outcomes relative to average-onset pancreatic cancer (AOPC) remain uncertain. We compared clinicopathological features, treatment patterns, and survival outcomes.

Methods: PubMed, Embase, Web of Science, the Cochrane Library, and Scopus were searched through January 2026 (PROSPERO: CRD420261346604). Observational studies comparing EOPC with AOPC using age-based thresholds were included. Study quality was assessed with the Newcastle-Ottawa Scale and ROBINS-E. Hazard and odds ratios were pooled using random-effects models for 39 outcomes across 9 predefined PROSPERO domains. All tests were two-sided. Certainty was assessed using GRADE.

Results: Forty-eight studies from 15 countries were included, encompassing 1,659,432 patients (non-overlapping estimate: 695,936). EOPC patients had a lower prevalence of diabetes (odds ratio [OR], 0.39; 95% confidence interval [CI], 0.21 to 0.72), more advanced disease (stage IV: OR, 1.18; 95% CI, 1.02 to 1.36), and higher chemotherapy utilization (OR, 2.04; 95% CI, 1.69 to 2.45; P < .001). Overall survival was comparable (hazard ratio, 1.03; 95% CI, 0.86 to 1.23; P = .76; prediction interval, 0.50 to 2.12). Exploratory dose-response meta-regression suggested a nonlinear (U-shaped) association between age threshold and survival (P = .008). GRADE certainty was very low for all outcomes.

Conclusions: Patients with EOPC present with more advanced disease, receive more aggressive treatment, and achieve comparable survival, although certainty is very low. These findings support equitable access to intensive multimodal therapy and germline testing before age 50; however, stage-specific survival data are required to confirm these observations.