Acute Pancreatitis: Feeding, ERCP, and the Gallbladder

0:00Welcome to Board Pearls. This is episode two of three of the Acute Pancreatitis chapter, in the Pancreatic and Biliary Disease module. This episode is the first days of management: when to feed, when to do an ERCP, and when to take out the gallbladder, and in all three the intuitive older instinct turns out to be net harmful.

0:20Start with nutrition, because the old answer and the new answer point in opposite directions. The old answer was pancreatic rest: keep the patient NPO, run TPN, wait for the gland to quiet, and the mechanism made intuitive sense, since food stimulates the pancreas, a stimulated pancreas keeps secreting enzymes, and enzymes keep the autodigestion going, so starve the stimulus and you starve the disease. The data refused to cooperate, because patients on TPN had more infections, not fewer, and by a wide margin.

0:55The reason sits in the gut, not the pancreas. The gut barrier is an active structure that depends on nutrients in the lumen to maintain its tight junctions, mucus, and antimicrobial activity, so when the lumen is empty the villi atrophy within days and the wall becomes permeable, and gut bacteria translocate across that permeable wall into peripancreatic tissue, where they reach devitalized tissue and seed infected necrosis. So pancreatic rest, the intuitive answer, is actually the mechanism by which mild pancreatitis becomes infected severe pancreatitis.

1:27Enteral feeding does the opposite: nutrients in the lumen keep the villi tall and the junctions intact and the bacteria where they belong, and the inflammation in the pancreas doesn't care that food is going past in the duodenum, because the cytokine cascade is already running and what determines whether the patient infects the necrotic tissue is the integrity of the gut wall, not the absence of stimulation.

1:49So the rule inverts. In mild disease, feed early, within twenty-four to forty-eight hours, and the bar is low: if the patient is hungry, has bowel sounds, and isn't vomiting, give them food. The old clear-liquids-then-advance ritual is unnecessarily restrictive, because a low-fat solid diet is safe from the start and delivers more calories, so the patient who can eat a regular tray should eat one. Take the young woman with mild gallstone pancreatitis, lipase falling, pain resolving on day one, and hungry: the temptation is to keep her NPO until the lipase normalizes because the number is reassuring and waiting feels cautious, but it isn't cautious, because holding her NPO prolongs her stay, raises her infection risk, and does nothing for the gland, so the right answer is breakfast.

2:34Predicted severe disease is the harder case, and this is where practice was reshaped by asking whether these patients need an early nasojejunal tube within twenty-four hours to bypass the stomach and protect the gut barrier, compared with simply feeding by mouth on demand once the nausea and ileus resolve. The answer was no difference in major infection or mortality, so early nasojejunal placement is not mandatory, and a predicted severe patient who can tolerate oral intake can advance orally, with the tube reserved for the patient who genuinely can't eat rather than forced on everyone with a high severity score.

3:09And when a tube is needed, the level doesn't much matter, because nasogastric, nasoduodenal, and nasojejunal feeding come out equivalent, which changes what you do at three in the morning when a patient can't eat: place the nasogastric tube and start feeding rather than waiting twelve hours for interventional radiology to position a post-pyloric tube under fluoroscopy, because the mechanism that justified post-pyloric feeding, bypassing pancreatic stimulation in the duodenum, was the wrong mechanism, and gut-barrier preservation works at any level of the small bowel. Formula choice is the same story, with standard and semi-elemental formulas performing equivalently, so there's no need for the expensive elemental product.

3:49TPN is the failure path, used only when the patient truly can't tolerate enteral feeding for a prolonged period from severe ileus, hemodynamic intolerance, or surgical anatomy, because otherwise enteral feeding wins on every outcome that matters.

4:04That brings us to the biliary tree, because the second big decision is what to do about the gallstones. The instinct from the older literature was an urgent ERCP on every patient with gallstone pancreatitis, on the theory that clearing the duct early prevents ongoing obstruction, and that instinct was wrong for an anatomic reason: most stones that trigger gallstone pancreatitis have already passed by the time the patient presents, because the stone hit the ampulla, raised ductal pressure briefly, set off the cascade, and then dropped into the duodenum, so the patient arrives with a quiet duct and an inflamed pancreas, and an ERCP on that patient finds nothing to retrieve while still exposing them to the well-known post-ERCP pancreatitis risk for no benefit.

4:41So the decision splits into three patients. The first has cholangitis, with fever, jaundice, leukocytosis, and biliary dilation, and since antibiotics alone can't sterilize an obstructed biliary tree and persistent obstruction drives sepsis and organ failure, this patient needs ERCP within twenty-four hours, the same urgent rule as cholangitis from any cause. The second has persistent obstruction without cholangitis, with a bilirubin and ALT that are rising rather than falling, a dilated duct, and a stone still visible, and failure to clear the duct invites recurrent cholangitis and ongoing pancreatic injury, so this patient gets ERCP within twenty-four to seventy-two hours. The third has mild gallstone pancreatitis without cholangitis or persistent obstruction, where the ALT was elevated on admission but is falling, the bilirubin is normal or trending down, and the duct isn't dilated, and this patient gets no benefit from ERCP, which is exactly the population the definitive trial studied: predicted severe gallstone pancreatitis without cholangitis showed no difference in death or major complications between urgent and on-demand ERCP, retiring the old universal-ERCP practice and confirming that cholangitis or persistent obstruction is the lever, not severity prediction.

6:02So the operational rule is clean: cholangitis or persistent obstruction means ERCP, mild disease without obstruction means no ERCP, and the trap is the sick-looking patient with a high severity score and gallstones, because the indication is the duct, not the pancreas. The patient with a positive Murphy sign, a rising bilirubin, and a dilated duct needs ERCP within twenty-four hours regardless of how the pancreas looks, and the patient with a falling ALT and a non-dilated duct doesn't need it even if the severity score is high.

6:29That handles the duct, and the remaining decision is the gallbladder. The older practice was interval cholecystectomy at four to six weeks, on the reasoning that the inflamed gland needed time to settle and the gallbladder could wait, but the problem was the wait, because during those weeks the patient still had stones in the gallbladder and the same physiology that put a stone at the ampulla once could do it again.

6:51When that risk was measured directly, comparing same-admission cholecystectomy within a few days of pain resolution against interval surgery about a month later, the interval group had far more recurrent biliary events without any reduction in surgical complications, so a six-week wait costs roughly one in six patients a recurrent biliary event before they even reach their planned operation, whether that's biliary colic, cholecystitis, a duct stone, or another bout of pancreatitis. So same-admission cholecystectomy is now standard for mild gallstone pancreatitis, with the target within three days of pain resolution.

7:23There are two exceptions. In severe or necrotizing disease, the peripancreatic inflammation distorts the anatomy around the porta hepatis, the dissection planes are obscured, and the bile-duct-injury rate during cholecystectomy rises sharply, so you defer surgery four to eight weeks until the inflammation subsides, accepting the recurrent-event risk because the operative-injury risk is worse. And in the patient who can't have surgery at all, the elderly patient with severe lung disease, advanced heart failure, and chronic kidney disease who is too high-risk for laparoscopic cholecystectomy under any conditions, the durable alternative is ERCP with a biliary sphincterotomy, because a wide-open ampulla lets any stone that later passes from the gallbladder slip through without raising ductal pressure or triggering pancreatitis, so the gallbladder stays in place but the source of pancreatitis is neutralized.

8:15So put the three decisions together. Feed early because the gut barrier protects against infected necrosis. Reserve ERCP for cholangitis or persistent obstruction because the typical patient has already passed the stone and an empty ERCP only adds risk. And operate same-admission for mild disease because the interval wait costs about one in six patients a recurrent biliary event. The unifying logic is that each intervention earns its place against its own complication profile, and the older instincts toward NPO, urgent ERCP, and interval cholecystectomy each turn out to be net harmful.

8:51Episode three picks up where this stops. We work through necrotizing disease and the step-up logic of delay, drain, and debride, the vascular complications of splanchnic venous thrombosis and the pseudoaneurysm that mandates angiography before any drainage, the prevention of post-ERCP pancreatitis with indomethacin, lactated Ringer's, and a pancreatic duct stent, and the workup of recurrent or idiopathic disease, including endoscopic ultrasound for microlithiasis and the pancreas divisum trap.

9:23For the full chapter, the practice vignettes, and the topic-tagged question bank, head to board pearls dot com. You'll find the rest of the series on Apple Podcasts, Spotify, or wherever you listen to podcasts. That brings us to the end of episode two of three of chapter twenty-five, and I'll see you in the next one.