Refeeding Syndrome Risk: How Cautiously to Advance Calories, and How Fast to Load Thiamine
Everyone in the room agrees thiamine goes in before or with the first calorie, no exceptions. What they don't agree on is whether the traditional 'start low, go slow' caloric ramp is actually the safer path, or whether it's the more conservative-sounding choice that leaves a starving teenager malnourished for longer than a newer, faster protocol would.
S.V., a 16-year-old girl, has run competitive distance for her high school team since seventh grade and made varsity as a freshman, a trajectory her coaches describe as remarkable and her mother now describes, in hindsight, as the beginning of something none of them named early enough. Training mileage crept up gradually over the past year alongside a parallel, quieter restriction in what she allowed herself to eat — the two reinforcing each other in a way that read, for months, as dedication rather than illness, until she collapsed at practice ten days ago and was found, at the emergency department that day, to weigh substantially less than her team physical had recorded just three months earlier.
She is being admitted today specifically for medical stabilization and refeeding, and her intake over the past ten days — by her own accounting and her mother's independent estimate, which broadly agree — has been under 500 kilocalories most days, sometimes closer to 200. That combination of profound recent intake restriction, rapid weight loss, and a starting weight well below expected for her age and height places her squarely in the high-risk category for refeeding syndrome by every standard clinical criterion the team uses. The physiology is well characterized: prolonged starvation depletes total-body phosphate even while serum phosphate still reads normal; the moment real caloric intake resumes, an insulin surge drives glucose, phosphate, potassium, and magnesium sharply into cells, and thiamine — a cofactor multiple carbohydrate-metabolism enzymes depend on directly — gets consumed by that same sudden metabolic demand, precipitating deficiency if it isn't already replete. Everyone on the team agrees thiamine goes in before or with the very first calorie, without exception. What they don't agree on is the caloric ramp itself. The traditional, still widely taught approach starts conservatively low, historically around 5-10 kilocalories per kilogram per day, out of concern that faster refeeding directly causes the syndrome it's named for. A newer body of evidence — a multi-site randomized trial led by Garber and colleagues, comparing higher- and lower-calorie inpatient refeeding protocols in hospitalized adolescents with anorexia nervosa — found that higher-calorie refeeding, done under the same close phosphate and electrolyte monitoring either approach requires, restored weight faster and shortened hospital stay without an increase in refeeding syndrome. Her phosphate is 3.8 and her potassium 3.9 — both unremarkable on paper, and both measured before any calories have been given, which is exactly when starvation physiology hides depletion behind a normal serum value. The trial's reassurance was earned under monitoring, not instead of it, and S.V.'s heart rate of 38 sits at the edge of the population that reassurance was drawn from.
Admission day 1, caloric protocol decision
I'd start conservatively here — 5 to 10 kilocalories per kilogram per day, advancing slowly. Her heart rate is 38. She's lost eighteen percent of her body weight in three months and was eating under 500 calories a day for the last ten. This is exactly the acute, high-risk presentation the traditional low-and-slow approach exists to protect. I'd rather move cautiously with a patient this compromised than lean on a newer protocol before I'm confident it applies to someone this acute.
I want to be clear that the traditional caution isn't untested reasoning we're choosing to override — Garber and colleagues directly tested it. Their multi-site randomized trial compared higher- and lower-calorie refeeding protocols in hospitalized adolescents with anorexia nervosa, the exact population S.V. belongs to, under the same close phosphate and electrolyte monitoring either approach requires. The higher-calorie arm restored weight faster and shortened hospital stay without an increase in refeeding syndrome. "Start low and go slow" sounds safer, but the actual randomized comparison found it mostly just meant staying malnourished for longer, not meaningfully safer.
I understand the instinct to default to caution given how sick she looks right now, but that instinct is exactly what this trial was built to test directly, rather than just assumed.
I don't think either of you is wrong about the evidence — I think the open question is where S.V. actually falls within it. The Garber trial's population was broadly "hospitalized adolescents with anorexia nervosa," but her specific combination — heart rate in the 30s, eighteen percent loss in three months, ten days under 500 calories — may sit toward the more severe tail even within that trial's own enrolled patients. I'd propose starting at the higher end of what's still reasonably described as a cautious range rather than either extreme: not the most conservative 5 kcal/kg historically taught, and not necessarily the trial's fastest advancement on day one, with a plan to accelerate toward the trial's protocol within 24 to 48 hours if her phosphate, potassium, and magnesium stay stable on close monitoring. That gives us a real answer specific to her numbers rather than a blanket pick between the two named approaches.
Agreed unanimously and without debate: thiamine 200mg IV before the first caloric intake, with proactive phosphate, potassium, and magnesium monitoring every 6-12 hours regardless of which caloric strategy is chosen. Caloric advancement will start at an intermediate level — above the most conservative historically taught range, below the higher-calorie trial's fastest protocol — with a plan to advance toward the trial's demonstrated-safe rate within 24-48 hours if electrolytes remain stable on close monitoring.
Not agreed: the exact trigger for accelerating from the intermediate starting point to the faster protocol. The Adolescent Medicine Physician wanted a defined, time-based trigger — 24 hours of stable electrolytes, advance regardless of any other single data point — to avoid caution drifting into an indefinitely slow ramp by default. The Attending Psychiatrist preferred a clinical-judgment- based trigger tied to S.V.'s overall trajectory (vitals, engagement, not just labs), worried that a fixed time trigger could advance calories before her hemodynamic instability has genuinely resolved. Both plans will be reassessed together at morning rounds tomorrow with actual overnight data in hand.
Educational content only — a composite teaching case, not a real patient encounter or a substitute for clinical guidance. About These Cases →