Pancreas divisum has been claimed to be a harmless congenital variant or to occasionally cause acute relapsing pancreatitis (ARP), chronic pancreatitis (CP), or a chronic abdominal pain (CAP) syndrome. Both surgical and endoscopic approaches to accessory papilla decompression have been promulgated and widely disparate results reported in the literature. We retrospectively reviewed a five-year experience with dorsal pancreatic duct decompression at our institution utilizing a variety of endotherapeutic techniques. Data collected included procedural complications; patient interpretation of pre- and posttherapy pain, frequency, and intensity graded on an analog pain scale; frequency of hospitalization; and patient perception of "global" improvement to endotherapy. At a mean follow-up of 20 months, there was a statistically significant decrease in pancreatitis incidence in 15 patients with ARP (P = 0.016) and 19 patients with CP (P = 0.025). The frequency and intensity of chronic pain was also significantly improved (P < 0.001) in the latter group. In contrast, only one of five patients with CAP and normal dorsal pancreatography and secretin tests experienced global improvement, and there was no improvement utilizing an analog pain scale (P = 0.262) in the group as a whole. There was a 20% incidence of mild procedure or subsequent stent-related pancreatitis and an 11.5% accessory papilla restenosis rate. It is concluded that a subset of carefully selected patients with pancreas divisum may respond to endotherapy but that long-term follow-up will be required to define its ultimate place in the management of symptomatic patients with this anomaly.