In a prospective study, 85 patients with chronic pancreatitis have been subjected to evaluation by morphologic analysis (endoscopic retrograde cholangiopancreatography), by exocrine function tests (bentiromide PABA and 72-hour fecal fat testing), and by endocrine function tests (oral glucose tolerance test and fat-stimulated release of pancreatic polypeptide). All patients were graded on a five-point system, with 1 point assessed for an abnormal result in each of the five tests performed. Zero score denoted mild disease; 1-2 points signaled moderate disease; and 3-5 points indicated severe disease. In 68 patients, both an initial and late (mean follow-up period of 14 months) evaluation were performed. Forty-one patients underwent modified Puestow side-to-side Roux-en-Y pancreaticojejunostomy. The Puestow procedure alone was performed in 18 patients. Eight patients also had drainage of pseudocysts, seven also had a biliary bypass, and eight had pseudocyst drainage plus bypass, in addition to the Puestow. There were no deaths. Of the 68 patients who were studied twice, 30 had operations and 38 did not. None of the patients with severe disease improved their grade during follow-up. Of 24 patients who did not undergo operation, 17 (71%) who were graded mild/moderate progressed to a severe grade at follow-up. By contrast, only three of the 19 patients operated on (16%) and who were initially graded as mild/moderate progressed to severe disease at follow-up testing. More than 75% of all of the patients had a history of weight loss. Twenty-six of 30 patients operated on (87%) (all of whom had lost weight before surgery) gained a mean 4.2 kg (range 1.4-2.7 kg) after surgery, compared with no significant weight change (range -3.6-2.7 kg) among patients not operated on. These findings support a policy of early operation for chronic pancreatitis, perhaps even in the absence of disabling abdominal pain.