Background It is a controversial issue whether pancreas divisum (PD) induces pancreatitis. All previous studies have investigated this issue based on endoscopic procedures, which inevitably involve a selection bias.

Objectives To determine the unbiased prevalence rate of PD in a community population and to investigate the effect of PD on idiopathic pancreatitis using a non-invasive magnetic resonance (MR) technique.

Design Cross-sectional study.

Patients The study enrolled 504 subjects from the community who participated in the medical check-up programme and 46 patients with idiopathic pancreatitis (8 acute, 23 chronic, 15 recurrent) extracted from 70 122 consecutive MR studies performed at an academic tertiary care hospital.

Interventions All subjects underwent magnetic resonance (MR) scanning and medical examination.

Main outcome measures Statistical comparison between subjects from the community and patients with idiopathic pancreatitis was made for the rate of PD (and its subtypes: classical PD, PD with absent ventral duct, and incomplete PD), MR findings, and clinical features.

Results Multiple logistic regression analysis revealed PD as a significant factor that induces pancreatitis (OR 23.4; p<0.0001). The PD rate was significantly higher for all/chronic/recurrent idiopathic pancreatitis patients (35%/43%/33%; p<0.001 for all) than for subjects in the community group (2.6%), but was not higher for acute pancreatitis (13%; p=0.357). All PD subtypes were indicated to induce idiopathic pancreatitis but showed different associations with each onset type of pancreatitis.

Conclusions This is the first study to describe the prevalence of PD and PD subtypes in a community population and their association with idiopathic pancreatitis in vivo based on the findings of non-invasive MR and with minimal selection bias. It is concluded that PD should be considered a predisposing factor for chronic and recurrent pancreatitis.