GI Nutrition: Refeeding, Malnutrition, and Enteral Access

0:00Welcome to Board Pearls. This is episode one of four of the GI Nutrition chapter, in the Special Populations and Acute or Supportive Care module. In this episode we cover the nutrition framework and the enteral access decisions: refeeding syndrome with the NICE high-risk criteria and the cautious calorie reintroduction with prior thiamine, hospital malnutrition diagnosed by GLIM phenotypic plus etiologic criteria with albumin called out as not a nutrition marker, and enteral access selection with its tube-feed complications.

0:30Start with the patient who has been starving. The single most consequential metabolic event in inpatient nutrition is the moment carbohydrate goes back into a patient who has been adapted to starvation. Refeeding syndrome is what happens when that reintroduction outpaces the body's capacity to handle the electrolyte shifts that follow. The reasoning starts with what starvation does to intracellular stores. Glycogen empties in the first twenty-four to forty-eight hours. Fatty acids and ketones take over as the dominant fuels. Intracellular phosphorus, potassium, and magnesium drift downward as renal conservation lags behind ongoing losses. Total body stores fall while serum levels still look fine, because the serum compartment is small and gets buffered by ongoing release from the much larger intracellular pool. The labs lie. The patient is depleted in the cells; the serum is the last place to show it.

1:23Now reintroduce carbohydrate. Insulin surges. Glucose, potassium, magnesium, and phosphorus all get driven into cells together, and the small serum compartment crashes within hours because the previously masked total-body deficit becomes unmasked all at once. The most consequential of those electrolytes is phosphorus, because phosphorus depletion produces diaphragmatic ATP failure and respiratory collapse. ATP regeneration in muscle requires phosphate, and the diaphragm is muscle. Hypophosphatemia also produces arrhythmia through disrupted cardiac myocyte action potentials, and acute encephalopathy from cerebral energy failure. The mortality of severe untreated refeeding syndrome approaches thirty percent. The intervention is not heroic; it is restrained.

2:13Thiamine sits inside the same crisis through a parallel mechanism that is worth pulling apart on its own. Thiamine is a cofactor for transketolase and pyruvate dehydrogenase, the enzymes that move pyruvate into the TCA cycle. In a thiamine-depleted patient, glucose drives glycolysis but pyruvate cannot get into the citric acid cycle. ATP fails in vulnerable thalamic and mammillary neurons. Wernicke encephalopathy emerges acutely, and if not corrected it progresses to permanent Korsakoff syndrome with anterograde amnesia and confabulation. That is why thiamine always precedes glucose. The rule is not a ritual. It is the direct consequence of glycolytic dependence on a cofactor that the starving patient no longer has.

3:00The NICE high-risk criteria identify the patients in whom this combination of starvation adaptation and thiamine depletion is predictable. The single-criterion list is the strong one. Any one of these alone makes a patient high risk. BMI below sixteen. Weight loss above fifteen percent in the prior three to six months. More than ten days of negligible nutritional intake. Or pre-feeding hypokalemia, hypophosphatemia, or hypomagnesemia. The two-criterion list is the weaker one, where any two together qualify. BMI below eighteen and a half. Weight loss above ten percent. More than five days of low intake. Or a history of alcohol misuse, insulin, chemotherapy, or diuretic use. The high-risk populations follow from the mechanism. Anorexia nervosa. Prolonged NPO beyond seven days. Chronic alcohol use disorder. Post-bariatric surgery. Oncology cachexia. Prolonged hyperemesis gravidarum. If you can tell yourself a coherent story about why the patient is starvation-adapted and thiamine-depleted, the patient is high risk.

4:05Management is restraint as the active intervention. Caloric initiation is ten to twenty kilocalories per kilogram per day in the first twenty-four to forty-eight hours, advancing slowly over four to seven days to full target. The restraint itself is the treatment, because the magnitude of the insulin surge tracks with the carbohydrate load, and a smaller surge produces a smaller intracellular shift. Thiamine two hundred to three hundred milligrams orally or intravenously is given before any glucose load and continued daily for the first five to seven days. Phosphorus, potassium, and magnesium are checked twice daily during the first seventy-two hours and aggressively repleted to keep ahead of the intracellular shift. Caloric escalation does not begin until those electrolytes are stable on appropriate replacement. The temptation in a malnourished patient is to feed faster. That instinct is wrong here. The slow start is the active treatment.

5:00That brings us to the second decision the inpatient nutrition team makes. The patient is either malnourished in a way that nutrition support can fix, or catabolic from acute illness in a way that no amount of feeding will reverse. This is the place where most nutrition consults go wrong, and the trap usually runs through albumin. So pause on albumin first. Albumin is a negative acute-phase reactant. It is suppressed by IL-six and TNF-alpha during inflammation. The liver gets reprogrammed under cytokine drive to produce positive acute-phase reactants like CRP, fibrinogen, and ferritin, and hepatic albumin synthesis is downregulated to make room. A low albumin in a hospitalized patient reflects inflammatory burden, not caloric or protein insufficiency. Prealbumin behaves the same way for the same reason, with a shorter half-life of two to three days that gives it a false reputation as a more responsive nutrition marker. Both are suppressed by the same cytokines. Both fail as nutrition markers in the inflamed patient. The corollary is the testable one. Aggressive caloric or protein supplementation does not normalize albumin in inflamed patients and does not improve survival. The early weight gain seen in tube-fed or parenterally-fed catabolic patients in randomized trials is fluid, not muscle. The GLIM framework explicitly excludes both albumin and prealbumin from its diagnostic criteria precisely because they do not respond to nutrition support in the inflamed state.

6:29GLIM stands for the Global Leadership Initiative on Malnutrition, and it formalizes the catabolic-versus-true-deficit distinction by requiring one phenotypic criterion plus one etiologic criterion for the diagnosis. The two-criterion structure is intentional. The phenotypic component identifies the metabolic deficit. The etiologic component identifies the cause that nutrition support is meant to address. Without the etiologic side, you are treating a phenotype with no defined target. The phenotypic criteria come in three forms. Unintentional weight loss, defined as above five percent within six months or above ten percent beyond six months. Low body mass index, defined as below twenty in adults under seventy, or below twenty-two in those over seventy. Or reduced muscle mass on validated measurement such as mid-upper-arm circumference, calf circumference, or imaging. The etiologic criteria are reduced intake or assimilation (less than half of estimated requirement for over a week, any degree for over two weeks, or chronic malabsorption) or inflammation from acute illness or chronic disease. Screening uses the Nutritional Risk Screening two thousand and two tool in the inpatient setting and the Malnutrition Universal Screening Tool in the community. Both flag candidates for full GLIM-based evaluation rather than serving as the diagnosis themselves.

7:53Calorie and protein needs in critical illness run roughly twenty-five to thirty kilocalories per kilogram and one and two-tenths to one and a half grams of protein per kilogram. Disease-specific patterns layer on top of the framework. Cirrhotic patients carry a high prevalence of sarcopenia and benefit from protein at one and two-tenths to one and a half grams per kilogram with a late-evening snack to blunt overnight gluconeogenesis. Inflammatory bowel disease produces both inflammatory cachexia and specific micronutrient deficits from ileal disease and steroid use. Celiac disease produces a specific malabsorption phenotype with iron, folate, calcium, and vitamin D deficiencies that resolve with the gluten-free diet. Chronic pancreatitis produces fat malabsorption and fat-soluble vitamin deficiency that respond to pancreatic enzyme replacement.

8:43The clinical position that follows from all of this is more restrained than the consult-service instinct suggests. Nutrition support is indicated when expected oral intake will be inadequate for more than seven days, when malabsorption produces ongoing weight loss, or when hypermetabolism in burns, polytrauma, or severe pancreatitis outpaces oral capacity. Nutrition support is not indicated to chase albumin. It is not indicated to cure the cachexia of metastatic malignancy. Both of those impulses are common, and both produce harm without benefit.

9:19Once the decision is made to feed, the route follows a single operating principle. A functional gut wins. Enteral nutrition preserves gut mucosal integrity, maintains the microbiome, supports gut-associated lymphoid tissue, and avoids the central catheter complications of parenteral nutrition. If the gut works at all, you use it. The decision after that is short-term versus long-term access. Short-term feeds, meaning under four to six weeks, use nasogastric or nasojejunal tubes because the durability requirement is brief and the procedural cost of a percutaneous tube is not justified. Longer-term feeds over four to six weeks use percutaneous endoscopic gastrostomy, with surgical or radiologic alternatives when endoscopy is not feasible.

10:04PEG has two principal placement techniques worth knowing by name. The pull technique, sometimes called the Ponsky technique, brings the gastrostomy tube through the mouth and out through the abdominal wall using an endoscopic snare. The introducer technique, sometimes called the Russell technique, places the tube directly through the abdominal wall under endoscopic visualization. The pull technique is the default. The introducer technique is preferred when oropharyngeal cancer raises concern about tumor seeding of the gastrostomy tract, because pulling the tube through the tumor field risks implanting tumor cells along the tract. PEG with jejunal extension, written as PEG-J, is reserved for patients who need post-pyloric feeding because of severe gastroparesis, recurrent aspiration, or proximal anatomic obstruction. The jejunal extension delivers nutrient distal to the gastric outlet so the stomach is bypassed.

10:57PEG candidates need a functional stomach without prohibitive gastric residuals, mental capacity to participate in care or a designated caregiver, and an expected need exceeding four to six weeks. PEG in advanced dementia is a board-tested indication where the data point in one direction. PEG does not prevent aspiration in advanced dementia, because aspiration in dementia is largely from oropharyngeal dysphagia rather than gastric content reflux. PEG does not extend survival in that population. PEG is not standard of care for advanced dementia. Counseling families against PEG in this context is the testable position, and the reasoning is the part to carry. The mechanism of aspiration matters; the tube does not solve the mechanism present in this disease.

11:47Two complications of established PEGs come up repeatedly. Buried bumper syndrome presents as a triad in a long-dwelling PEG: pain with feeding, resistance to flushing, and inability to rotate the tube. The mechanism is that excessive external traction pulls the internal bumper into the gastric wall, where mucosal overgrowth eventually buries it within the wall and obstructs the internal lumen. Prevention is rotating the tube weekly to break any incipient mucosal seal and avoiding excessive external traction. Treatment is endoscopic or surgical bumper removal and tube replacement. The other complications are predictable. Peristomal cellulitis is managed with topical or oral antibiotics and local wound care. Peristomal leakage from a stoma that has enlarged or a bumper that is malpositioned is managed by tube downsizing, repositioning, or temporary removal to allow the tract to contract. Tube blockage is cleared with warm water and pancreatic enzyme solution rather than carbonated beverages, which do not work and can corrode the lumen. Inadvertent dislodgement is replaced through the mature tract within twenty-four hours; longer delay risks tract closure and forces radiologic or repeat endoscopic placement.

12:59Tube-feed diarrhea is the third trap, and the failure mode the boards punish is switching to a peptide-based or concentrated formula without first reviewing the medication list. Tube-feed diarrhea is overwhelmingly medication-related, not formula-related. Sorbitol-containing liquid medications deliver an osmotic load directly through the tube. Metronidazole oral suspension, acetaminophen elixir, and theophylline syrup are common offenders, and there are many others. The colon cannot absorb sorbitol, which draws water into the lumen and produces osmotic diarrhea. Broad-spectrum antibiotics produce dysbiosis-driven diarrhea or unmask Clostridioides difficile infection. Contaminated formula or open delivery systems can introduce bacteria that produce direct enteritis. Concentrated hypertonic formulas are designed for fluid-restricted patients like heart failure or cirrhosis with ascites; using them as a diarrhea remedy can paradoxically worsen the diarrhea by raising the intraluminal osmotic load. Peptide-based formulas are reserved for documented digestive insufficiency like severe pancreatic exocrine failure or short bowel where intact protein cannot be hydrolyzed adequately. The workflow is therefore to review the medication list for sorbitol vehicles, check for recent antibiotic exposure, send a stool C. diff test, and only after these steps consider changing the formula.

14:19Pull the framework so far together. Refeeding syndrome is what happens when a starvation-adapted patient meets a carbohydrate load, and the active intervention is restraint with thiamine before glucose, because the phosphorus that fails the diaphragm and the thiamine that fails the mammillary neurons are the two deficits the carbohydrate load exposes. Hospital malnutrition is not what a low albumin says it is; GLIM separates the phenotypic deficit from the etiologic cause precisely because the inflamed liver lies about both albumin and prealbumin. And the route of nutrition follows the gut's functional state, with a functional gut winning every time, enteral access matched to expected duration, and the PEG complications driven by tube mechanics while tube-feed diarrhea is driven by the medication list before the formula.

15:07The next episode takes the patient whose gut cannot do the work: parenteral nutrition, its composition rules that follow from chemistry, its long-term complications from liver disease to catheter infection, and short bowel syndrome with its anatomic phenotypes and teduglutide.

15:24For 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 one of four of chapter thirty two, and I'll see you in the next one.