A Prospective, Randomized Trial of Clear Liquids vs Low-fat Solid Diet as the Initial Meal in Mild Acute Pancreatitis

Abstract & Commentary

By Malcolm Robinson MD, FACP, FACG, Emeritus Clinical Professor of Medicine, University of Oklahoma College of Medicine, Oklahoma City. Dr. Robinson reports no financial relationship to this field of study.

Synopsis: Patients with mild acute pancreatitis probably can be fed solid food with results as good as those seen with initial feeding with the traditional clear liquid diet.

Source: Jacobson Brian C, et al. Clin Gastroenterol and Hepatol. 2007;5:946-951.

There are almost 250,000 cases of pancreatitis annually in the United States. Most are classified as mild disease and resolve within a few days. Management of pancreatitis has traditionally entailed IV hydration, pain control, and elimination of oral feeding (NPO). The NPO status is presumed to avert early and potentially deleterious pancreatic stimulation. Once re-feeding is deemed appropriate, clear liquid diet has been the usual starting point. Diet is then advanced slowly and as tolerated. These authors hypothesized that re-feeding with a low fat solid diet (LFSD) would be as well tolerated as a clear liquid diet (CLD). It was hoped that the more advanced initial diet might lead to a shorter hospital stay. The study was well powered to detect a difference in length of hospital stay between groups receiving initial CLD vs LFSD. Out of a total of 1,355 patients at Brigham and Women's Hospital admitted with pancreatitis between 1999 and 2005, 121 patients with clear-cut mild pancreatitis were ultimately randomized into this non-blinded study. Those excluded were mostly ineligible on the basis of pancreatic enzymes that were not > 3 times normal or > twice normal with clear CT evidence of pancreatitis or eliminations for comorbidities. The primary prospective endpoint of this study was to have been a reduction in length of hospitalization (LOH). Analysis of data included the elimination of patients who had been fed less than 6 hours prior to eating. Even in this fine hospital, some study patients received the wrong initial meal. In any case, there was no difference in LOH between the CLD and LFSD groups. There were 113 patients in the per protocol evaluation group. Despite the failure to demonstrate a change in LOH, first meal and first day caloric intake and fat intake were higher in the LFSD group than in CLD patients. The authors thought that the lengths of stay in study patients might have been impacted by factors that were not related to diet. For example, a number of patients had extended hospitalizations for additional diagnostic studies. Previous studies elsewhere also have shown solid diet toleration in pancreatitis patients although they too failed to demonstrate decreased LOH. The authors conclude that early initiation of solid food intake is certainly worth considering in patients who are hospitalized with mild acute pancreatitis.


As was mentioned in an accompanying editorial, this paper raises more questions than answers.1 There is little justification in the concept that restriction of oral feeding is beneficial in many cases of pancreatitis. The inflamed pancreas probably doesn't respond normally to stimulation, and some patients, even those with relatively severe pancreatitis, might have no problems with initiation of feeding early in their course. We know little or nothing about the relative effects of any particular dietary constituents on the inflamed pancreas. The editorial writer mentions the potentially parallel notion that nasogastric tubes were deemed indispensable after many forms of gastrointestinal surgery, especially surgery with fresh intestinal anastomoses. This concept has been disproved, and most modern surgeons now dispense with the routine use of nasogastric suction in such cases. The whole concept of graduated diets in most hospitalized patients is either invalid or very poorly validated. In particular, the classical progression from clear liquid diet to the so-called full liquid diet seems especially ridiculous. Full liquids are mostly dairy-based, and large numbers of people in the general population are dairy intolerant for any of several reasons (especially lactose intolerance). Such individuals would almost certainly do better with initiation of a general diet. It has always seemed to me that the diets of hospitalized patients should be individualized. Patients who are doing well and who really want to try a particular food or beverage often should be accommodated. I would urge that patients with pancreatitis who have been able to reduce their requirements for opiatebased analgesic therapy could have pro-active dietary advancement. If a larger study were to be performed, I am quite confident that this hypothesis would be upheld and that shortened hospital stay would be demonstrable. Meanwhile, we already have evidence that dietary advancement in patients with mild pancreatitis is generally safe and tolerable.


1. Reber HA. Clin Gastroenterol and Hepatol. 2007;5:915-916.