Review of the Refeeding Syndrome
Michael D. Kraft, PharmD*†; Imad F. Btaiche, PharmD, BCNSP*†; and
Gordon S. Sacks, PharmD, BCNSP‡
*Department of Clinical Sciences, College of Pharmacy, University of Michigan, Ann Arbor, Michigan; †Department of
Pharmacy Services, University of Michigan Health System, Ann Arbor, Michigan; and the ‡Pharmacy Practice
Division, School of Pharmacy, University of Wisconsin–Madison, Madison, Wisconsin
ABSTRACT: Refeeding syndrome describes a constella-
tion of metabolic disturbances that occur as a result of
reinstitution of nutrition to patients who are starved or
severely malnourished. Patients can develop fluid and
electrolyte disorders, especially hypophosphatemia, along
with neurologic, pulmonary, cardiac, neuromuscular, and
hematologic complications. We reviewed literature on
refeeding syndrome and the associated electrolyte abnor-
malities, fluid disturbances, and associated complications.
In addition to assessing scientific literature, we also con-
sidered clinical experience and judgment in developing
recommendations for prevention and treatment of refeed-
ing syndrome. The most important steps are to identify
patients at risk for developing refeeding syndrome, insti-
tute nutrition support cautiously, and correct and supple-
ment electrolyte and vitamin deficiencies to avoid refeed-
ing syndrome. We provide suggestions for the prevention
of refeeding syndrome and suggestions for treatment of
electrolyte disturbances and complications in patients
who develop refeeding syndrome, according to evidence in
the literature, the pathophysiology of refeeding syndrome,
and clinical experience and judgment.
The term refeeding syndrome (RS) is generally
reserved to describe the metabolic alterations that
occur during nutrition repletion of underweight,
severely malnourished, or starved individuals. The
hallmark sign of RS is severe hypophosphatemia
and its associated complications. However, RS actu-
ally encompasses a constellation of fluid and electro-
lyte abnormalities affecting multiple organ systems,
including neurologic, cardiac, hematologic, neuro-
muscular, and pulmonary function. This article will
review the pathophysiology of RS, its physiologic
complications, the treatment of associated metabolic
disturbances, and provide guidelines for its recogni-
tion and prevention.
The classic study describing RS was conducted by
Keys and colleagues1in 1944 on male conscientious
objectors of World War II. The participants had
undergone semistarvation for 6 months and upon
nutrition replenishment, some subjects developed
cardiac failure. With the advent of modern-day par-
enteral nutrition (PN) and enteral nutrition (EN),
reports of similar complications were noted in
severely undernourished patients who received
aggressive nutrition supplementation. Weinsier and
resulting in death of 2 chronically undernourished
women who received aggressive PN. Both patients
were well below ideal body weight (IBW; 40% and
70%, respectively) and exhibited low serum concen-
trations of potassium and phosphorus before PN
initiation. Large amounts of carbohydrate and pro-
tein were delivered (approximately 75 kcal/kg from
dextrose and 3.5 g/kg of protein) at PN initiation,
rather than gradually increasing PN calories to goal
over the following days. Within 48 hours, both
patients experienced cardiac abnormalities and pul-
monary failure requiring mechanical ventilation.
Severe hypophosphatemia, hypokalemia, and hypo-
magnesemia occurred despite the presence of sup-
plemental electrolytes in the PN formulations. One
patient died on hospital day 6 and the other died
during the third week of hospitalization. These out-
comes represent the most severe responses to
refeeding but underscore the importance of under-
standing this syndrome, recognizing patients at
risk, and providing appropriate treatment in the
event of its occurrence.
Overview of Refeeding Syndrome
Understanding the physiology of starvation pro-
vides insight into the morbid sequelae associated
with refeeding a severely undernourished individ-
Correspondence: Michael D. Kraft, PharmD, Clinical Assistant
Professor and Clinical Pharmacist, University of Michigan
Health System, Department of Pharmacy Services, UH/B2 D301,
Box 0008, 1500 East Medical Center Drive, Ann Arbor, MI
48109-0008. Electronic mail may be sent to firstname.lastname@example.org.
Nutrition in Clinical Practice 20:625–633, December 2005
Copyright © 2005 American Society for Parenteral and Enteral Nutrition
ual. During the initial period of starvation (24–72
hours), the liver uses glycogen stores for energy and
skeletal muscle to provide amino acids as a source
for new glucose production (ie, gluconeogenesis) for
glucose-dependent tissues, such as the brain, renal
medulla, and red blood cells. After 72 hours of
starvation, metabolic pathways shift to derive
energy from ketone production as a result of free
fatty acid oxidation while sparing protein mobiliza-
tion from skeletal muscle.3Other adaptive mecha-
nisms include an overall decrease in liver gluconeo-
genesis, a decline in basal metabolic rate, reduction
in the secretion of insulin, and an increased use of
free fatty acids by the brain as the primary energy
source in place of glucose.
With the reintroduction of carbohydrate via oral
feeding, EN, or PN, there is a sudden shift back to
glucose as the predominant fuel source, creating a
high demand for the production of phosphorylated
intermediates of glycolysis (ie, red blood cell adeno-
sine triphosphate [ATP] and 2,3-diphosphoglycerate
[DPG]) with inhibition of fat metabolism. This
results in hypophosphatemia, the hallmark sign of
RS. Additional mechanisms identified as contribut-
ing to low serum phosphorus concentrations include
preexisting low total body stores of phosphorus dur-
ing starvation and enhanced cellular uptake of phos-
phorus during anabolic refeeding. Phosphate is nec-
essary for accrual of lean tissue mass and is a vital
component of metabolic pathways involving the pro-
duction of ATP and 2,3-DPG. Potassium and mag-
nesium also shift intracellularly in response to
anabolism and increased insulin release. Magne-
sium is a cofactor for the Na-K?ATPase pump, so
uncorrected hypomagnesemia can complicate potas-
sium repletion. Other metabolic alterations that
may occur include fluid imbalance and vitamin defi-
ciencies. An expansion of the extracellular water
compartment occurs during refeeding of the mal-
nourished individual. Although the exact mecha-
nism of fluid imbalance in RS is unknown, sodium
and water retention may be due to an antinatri-
uretic effect from hyperinsulinemia4or a possible
interaction between the water, sodium, and carbo-
hydrate homeostasis.5Although it is difficult to
determine whether thiamine deficiency is a result of
RS or is a preexisting deficiency due to starvation, it
is reasonable to presume that an undernourished
individual is at risk for thiamine deficiency. Thia-
mine is an essential cofactor involved in the metab-
olism of carbohydrates.6The phosphorylated form of
glucose is converted to pyruvate, which undergoes
decarboxylation in the presence of pyruvate dehy-
drogenase and thiamine. Acetyl coenzyme A is pro-
duced for entrance into the Krebs cycle and genera-
tion of ATP as an energy source for all living cells.
High doses of carbohydrate can increase the demand
for thiamine use in undernourished subjects with
decreased baseline thiamine stores, thus precipitat-
ing thiamine deficiency and its associated complica-
tions.6As a result, thiamine administration prior to
and during carbohydrate intake is recommended in
patients at risk for RS.
Clinical manifestations of RS are related to the
electrolyte and vitamin deficiencies that are present
and the subsequent abnormalities that develop with
the initiation of nutrition support. Clinical manifes-
tations of RS are summarized in Table 1.
Phosphorus is the major intracellular anion, and
it is important for many metabolic processes involv-
ing ATP and 2,3-DPG as described previously.
Severe hypophosphatemia (eg, serum phosphorus
concentration ?1–1.5 mg/dL) can lead to severe
neurologic, cardiac, respiratory, and hematologic
abnormalities and possibly death. Several reports
describe hypophosphatemia associated with the ini-
tiation of nutrition support (oral, enteral, and par-
enteral).2,7–15Severe hypophosphatemia in these
patients has also led to neurologic symptoms,
including paresthesias, weakness, confusion, disori-
entation, encephalopathy, areflexic paralysis, sei-
zures, coma, and death.8–10,12,13,16,17Silvas and
Paragas8described RS in 3 patients with severe
malnutrition who received aggressive PN for reple-
tion. All 3 adult patients were significantly under-
weight (?50%–60% of usual body weight, or
reported weight loss of approximately 22–30 kg
[50–65 pounds]); PN was initiated at a high rate
(?37–40 kcal/kg/day on day 1), and advanced rap-
idly over the course of 2–5 days (up to ?65–100
weakness, somnolence, lethargy, restlessness, and
muscle aches around day 5 of PN. Two patients
became unresponsive, developed seizures (on days 8
and 16 of PN) and coma; 1 patient expired 5 days
after the initial seizure (4 days after PN was discon-
decreased over the first 5–8 days and reached a
nadir of 0.1–0.5 mg/dL. Of note, 2 of the patients
gained approximately 1.8 kg (4 pounds) relatively
quickly after initiation of PN, likely reflecting fluid
Severe hypophosphatemia has also been show to
impair cardiac function18and respiratory func-
tion.14,19,20O’Connor et al18described diminished
cardiac function in patients with serum phosphorus
concentrations of 0.7–1.4 mg/dL. Stroke volume
(SV), mean arterial pressure (MAP), and left ven-
tricular stroke work (LVSW) were all decreased and
pulmonary artery wedge pressure (PAWP) was
increased. SV, MAP, and LVSW increased signifi-
cantly, with a significant decrease in PAWP, after
phosphate repletion over 8 hours (serum levels ?
Vol. 20, No. 6
KRAFT ET AL
1.6–4.7 mg/dL). Severe hypophosphatemia has also
been shown to impair diaphragmatic contractility20
and lead to acute respiratory failure requiring intu-
bation and mechanical ventilation.14,19Youssef14
described a case of a woman with multiple intestinal
fistulae who underwent a laparotomy and then
began PN. She developed respiratory failure on
postoperative day 2 (her second day of PN) and went
on to develop generalized convulsions, coma, and
required intubation and mechanical ventilation.
Hypophosphatemia can lead to decreases in ATP
and 2,3-DPG as described previously. This may lead
to further abnormalities in oxygen transport and
delivery,7,11,21–23and impaired glucose metabolism.7
Hypophosphatemia and a subsequent decrease in
2,3-DPG increase the affinity of hemoglobin for
oxygen and shifts the oxygen dissociation curve
to the left.21–23Sheldon and Grzyb11described
hypophosphatemia and associated abnormalities in
a series of 19 trauma patients, 8 of whom inadver-
tently were given PN without phosphate supplemen-
tation. Patients who developed hypophosphatemia
also had decreased levels of ATP and 2,3-DPG. The
authors further found a significant correlation
between total calories administered and the fall in
serum phosphorus concentration, and a significant
correlation between the amount of phosphate
administered and the increase in serum phosphorus
concentration. Travis et al7found that within 5–7
days of PN initiation (3–4 L/day) that did not
contain phosphate, 5 of 8 patients developed severe
hypophosphatemia (serum phosphorus concentra-
tion ?1 mg/dL, mean ? 0.5 mg/dL). Hypophos-
phatemia also led to reductions in erythrocyte ATP
and 2,3-DPG, with an associated increase of hemo-
globin affinity for oxygen (P50? 19.5 mm Hg, nor-
mal ?27 ? 1.1 mm Hg). Furthermore, hypophos-
phatemia led to significant decreases in erythrocyte
glucose-6-phosphate and fructose-6-phosphate, and
a significant increase in total triose phosphates
(eg, glyceraldehyde-3-phosphate, dihydroxyacetone
phosphate), suggesting a decrease in erythrocyte
glycolysis. These decreases in oxygenation and glu-
cose metabolism may also lead to central nervous
system and respiratory symptoms, as discussed.
Potassium is the major intracellular cation, with
approximately 98% of total body potassium residing
in the intracellular space but also in bone and
physiologic functions, including regulation of electri-
cal cellular membrane potential, cellular metabo-
lism, glycogen synthesis, and protein synthesis.
Hypokalemia alters the electrical action potential
across cell membranes and leads to membrane hy-
perpolarization and impaired muscular contrac-
tion.24–27Mild to moderate hypokalemia (eg, serum
potassium concentration ? 2.5–3.5 mEq/L) can
cause nausea, vomiting, constipation, and weak-
ness. If left untreated, severe hypokalemia (eg,
serum potassium concentration ?2.5 mEq/L) can
lead to paralysis, respiratory compromise, rhabdo-
myolysis, muscle necrosis, and changes in myocar-
dial contraction and signal conduction.26–29Patients
Clinical manifestations of refeeding syndrome
Impaired oxygen transport
and delivery, hypoxia
Impaired cardiac function
Alterations in myocardial
Presence of U-waves
Altered mental status
Refractory hypokalemia and
Torsade de pointes
REVIEW OF THE REFEEDING SYNDROME
with severe hypokalemia may develop electrocardio-
graph changes such as ST-segment depression,
T-wave flattening, T-wave inversion, or the presence
of U-waves.26,27,29Patients may also develop cardiac
arrhythmias, including atrial tachycardia, brady-
cardia, atrioventricular block, premature ventricu-
lar contractions, ventricular tachycardia, ventricu-
lar fibrillation, and possibly sudden death.25–28
Magnesium is the second most abundant intracel-
lular cation, with most of the total body magnesium
found in bone, muscle, and soft tissue.30–32Approx-
imately 1% of the total body magnesium resides in
the extracellular fluid.30–32Magnesium is an impor-
tant cofactor for many enzymes and in many bio-
chemical reactions, including reactions during oxi-
Hypomagnesemia (serum magnesium concentra-
tion ?1.5 mg/dL) is frequently observed in critically ill
patients34–37and has been associated with increased
of hypomagnesemia can resemble those of hypokale-
mia or hypophosphatemia. Patients with mild to mod-
erate hypomagnesemia can experience weakness,
muscle twitching, tremor, altered mental status,
anorexia, nausea, vomiting, and diarrhea.30–32,35,40,41
Moderate to severe hypomagnesemia (eg, serum mag-
nesium concentration ?1.0 mg/dL) can manifest
such signs and symptoms as electrocardiographic
changes (eg, prolonged PR, widened QRS, prolonged
QT, ST depression, peaked T-wave, or T-wave
flattening),30–32,42cardiac arrhythmias (eg, atrial
fibrillation, torsade de pointes, ventricular arrhyth-
mias, ventricular tachycardia),32,35,39tetany, convul-
sions, seizures, coma, and even death.30–32,35,41Hypo-
magnesemia, if left untreated, can also complicate the
treatment of coexisting hypokalemia and hypocalce-
mia. Hypomagnesemia-induced hypokalemia is likely
due to impaired Na?/K?-ATPase activity.43Hypo-
magnesemia-induced hypocalcemia is likely a result of
impaired parathyroid hormone release and/or activi-
Thiamine is an important cofactor in carbohy-
drate metabolism.6Thiamine is a water-soluble
vitamin, and total body stores can quickly become
depleted with weight loss and malnutrition. With
carbohydrate intake, there is an increased demand
for thiamine, a cofactor in glycolysis. Thiamine defi-
ciency in malnourished patients has led to Wer-
nicke’s encephalopathy in patients who were given
PN with high carbohydrate loads.12,17,47With thia-
mine deficiency, pyruvate is then converted to lac-
tate.48Excessive lactate formation leading to lactic
acidosis and death was reported in patients who
received PN without thiamine supplementation.49–52
The role of other vitamin deficiencies (especially
water-soluble vitamins) in RS is less clear.
Sodium Retention/Fluid Overload
Sodium retention and expansion of extracellular
water that may occur in the early phases of RS can
lead to fluid overload, pulmonary edema, and car-
diac decompensation.53,54This may be especially
devastating to patients at risk for RS (eg, patients
with severe malnutrition) because they may have
reduced cardiac mass and function.53,55Fluid and
sodium restriction are indicated when initiating
nutrition support in patients at risk for RS. Patients
should be monitored closely for signs of fluid accu-
mulation and overload.
Clearly, preventing RS is the primary goal when
initiating nutrition support in severely malnour-
ished and cachectic patients. There are several key
steps that clinicians should take to avoid RS and the
morbidity and mortality associated with RS. It is
essential to first identify patients who are at risk for
RS before initiating nutrition support (Table 2).
Regardless of the method used to estimate caloric
goals (eg, Harris-Benedict equation, kcal/kg, etc), it
is essential to avoid overfeeding. The minimum
glucose requirement for a 70-kg adult to suppress
gluconeogenesis, spare proteins, and supply fuel to
the central nervous system is approximately 100–
150 g/day.3A reasonable goal for protein intake in
adults is approximately 1.5 g/kg/day, although some
patients may have increased (eg, severe trauma,
severe burns, continuous renal replacement ther-
apy, hepatic dysfunction or cirrhosis with encepha-
lopathy [CRRT]) or decreased requirements (eg,
renal failure with uremia).
When initiating nutrition support in patients at
risk for RS, the rule of thumb is to “start low and go
slow.” Nutrition support should be initiated cau-
tiously (eg, approximately 25% of estimated goal
needs on day 1), and gradually increased to goal over
the course of 3–5 days. Any electrolyte abnormali-
ties (especially hypophosphatemia, hypokalemia,
and hypomagnesemia) should be corrected before
nutrition support is initiated. Providing empiric
Identification of patients at risk for refeeding syndrome
Residents admitted from skilled nursing facilities
Unfed for 7–10 days with evidence of stress/depletion
Chronic diseases causing undernutrition (eg, cancer or
cardiac cachexia, chronic obstructive pulmonary disease,
History of excessive alcohol intake
Morbid obesity with massive weight loss
Vol. 20, No. 6
KRAFT ET AL
electrolyte supplementation (in patients with nor-
mal renal function) before and during nutrition
support is advisable. Increasing total caloric load
may decrease serum phosphorus concentration, and
it is necessary to provide a minimum of approxi-
mately 10–15 mmol of phosphate per 1000 kcal to
maintain normal serum concentrations (in patients
with normal renal function).11Patients with severe
malnutrition, critical illness, severe trauma, and
burns will also likely have a depletion of total body
phosphorus (even if serum concentrations are nor-
mal), and their phosphate requirements will be
higher. The same may be true for potassium and
magnesium in these patients as well. After nutrition
support is initiated and titrated upward, electro-
lytes should be supplemented according to serum
electrolyte concentrations and response to therapy.
Because patients at risk for RS may also have
diminished cardiac reserve and can develop fluid
overload, fluid and sodium should be minimized
during the first few days of nutrition support (eg,
mL/day).56Patients should gain no more than 1 kg
per week during repletion. Any weight gain ?1
kg/week would likely be attributed to fluid reten-
Vitamin supplementation should also be pro-
vided. Parenteral multivitamin preparations pro-
vide daily requirements as recommended by the
American Medical Association.57These prepara-
tions contain 3 mg or 6 mg of thiamine daily. How-
cachectic patients, and additional supplementation
has been suggested.58Supplemental thiamine at
50–100 mg/day IV, or 100 mg PO for 5–7 days
should be provided to patients at risk for thiamine
deficiency or RS. Most reports have focused on
thiamine deficiency, but other vitamins may also be
deficient in the malnourished patient. Although the
importance of other vitamin deficiencies in RS is less
clear, administering supplemental vitamins (espe-
cially folic acid) to patients at risk for RS is a
reasonable approach. In addition to thiamine, 1
mg/day folic acid may also be provided for 5–7 days.
Alternatively, providing a supplemental multivita-
min PO daily in addition to EN for 5–7 days is
reasonable. These steps can be done safely and inex-
pensively and may prevent patient morbidity.
Patients should be monitored closely for signs and
symptoms of RS. Vital signs, including heart rate,
blood pressure, respiratory rate, mental status, and
neurologic function, should be monitored routinely,
especially during the first several days of nutrition
support until goal is reached. Finger pulse oximetry
should be used if available, and patients should also
be monitored for any electrocardiographic changes if
possible. In addition, patients should be evaluated
for any neuromuscular signs and symptoms during
daily physical examinations. Patients should also be
assessed for fluid balance, signs of edema, fluid
overload, and weighed on a regular basis.
Treatment of RS includes supportive care and
treatment of any electrolyte disorders. If the patient
exhibits any signs or symptoms consistent with RS,
nutrition support should be interrupted immedi-
ately. Dextrose 10% in water can be initiated
instead at the same rate to avoid rebound hypogly-
cemia if desired. A “stat” laboratory assessment
should be made to evaluate serum electrolyte and
glucose levels. If the patient exhibits any neurologic
changes (eg, mental status changes, encephalopa-
thy), a single dose of IV thiamine 100 mg should be
given. If respiratory distress or other respiratory
symptoms are present, supplemental oxygen should
be provided, and an arterial blood gas obtained.
Cardiovascular changes should be addressed and
treated immediately. Any evidence of fluid overload
should also be treated appropriately (eg, diuretic
The following sections provide suggestions for
treatment of specific electrolyte abnormalities. We
would recommend administering ?50% of the initial
empiric doses of electrolytes (phosphate, potassium,
and magnesium) in patients with impaired renal
function (eg, creatinine clearance ?50 mL/min,
serum creatinine ?2 mg/dL, patients who are oligu-
ric [urine output ?400 mL/day] or anuric [urine
output ?100 mL/day]) who are not treated with
CRRT. In addition, when using weight-based dosing,
there are no definitive data or recommendations for
“adjusting” weight in patients who are significantly
obese. There is also debate on when clinicians
should use an “adjusted” body weight (eg, using a
percentage above IBW or according to body mass
index [BMI]). Adipose tissue is estimated to be
composed of approximately 10%–30% water,59–63
and total body water in men is slightly higher that
that in women. Often in clinical practice, an “adjust-
ment” of 25%–40% of the difference between actual
weight and IBW is added to the IBW to determine
the “adjusted” body weight or dosing weight.
Even though this practice is controversial, adjust-
ing body weight in obese patients may minimize the
risk of overdosing and complications.
Treatment of Hypophosphatemia
Treatment of hypophosphatemia depends on the
magnitude of hypophosphatemia, whether or not the
patient is symptomatic, and the route of administra-
tion that is available (ie, enteral or parenteral).
Patients with mild hypophosphatemia who are
asymptomatic and have a functioning gastrointesti-
nal tract may be treated with oral phosphates.
However, oral absorption can be unreliable, and oral
phosphate products may cause diarrhea. Asymptom-
atic patients with moderate to severe hypophos-
phatemia who cannot receive oral medications and
patients who are symptomatic should receive IV
phosphate supplementation. Phosphate dosing is
largely empiric because serum concentrations may
REVIEW OF THE REFEEDING SYNDROME
not correlate with total body phosphorus stores.
Suggested IV phosphate dosing is provided in Table
3.64–70We recommend providing ?50% of the initial
empiric phosphate dose in patients with impaired
renal function who are not treated with CRRT.
Patients treated with CRRT have continuous phos-
phorus clearance and may require higher initial
doses, depending on the degree of hypophos-
phatemia and whether or not phosphate is used in
the dialysate/replacement fluid. Further phosphate
response to the initial dose.
IV phosphate formulations are available as potas-
sium or sodium salts. One mmol of potassium phos-
phate contains 1.47 mEq of potassium, and 1 mmol
of sodium phosphate contains 1.33 mEq of sodium.
Potassium phosphate can be used in patients with
simultaneous hypokalemia; otherwise sodium phos-
phate should be used. Total phosphate dose should
be infused over 4–6 hours to minimize adverse
effects (eg, thrombophlebitis from potassium phos-
phate) and to reduce the risk of calcium-phosphate
precipitation. Doses can be infused up to a rate of 7
be guidedby clinical
mmol of phosphate per hour (or about 10 mEq of
potassium per hour).69,70Serum phosphorus concen-
tration should be checked 2–4 hours after a dose and
additional phosphate supplementation
until the patient is asymptomatic or the serum
phosphorus concentration is in the normal range.
Serum phosphorus concentration should be moni-
tored at least daily for the first week of nutrition
support. More frequent monitoring may be indicated
in the first several days of nutrition support, espe-
cially in patients who manifest signs or symptoms of
Treatment of Hypokalemia
Hypokalemia can be treated with potassium sup-
plementation via the oral or IV route. The IV route
should be used when treating patients with symp-
tomatic or severe hypokalemia (eg, serum potassium
concentration ?2.5 mEq/L), or when the gastroin-
testinal tract cannot be used. Dosing of potassium is
largely empiric and based on clinical response and
serum concentrations. Suggestions for potassium
dosing are provided in Table 4.71–73We would also
recommend that patients with impaired renal func-
tion who are not being treated with CRRT receive
?50% of the recommended initial dose. Patients
receiving CRRT may have higher clearance of potas-
sium and require higher initial doses. Serum potas-
sium concentration should be checked within 1–4
hours after a dose, and multiple doses of potassium
may be required for full repletion. Potassium can be
safely administered in adult patients at rates of
10–20 mEq/h. Rates ?20 mEq/h are rarely needed,
except in emergent situations. Patients should
receive potassium via a central venous catheter and
should have continuous cardiac monitoring for infu-
sion rates ?10 mEq/h. Potassium should never be
given as a rapid infusion to avoid serious or fatal
consequences. Potassium concentration in solutions
for continuous infusion via a peripheral vein should
be limited to 80 mEq/L, and up to 120 mEq/L can be
used for infusion via a central vein. These standard
recommendations are provided for safety, although
Treatment of hypophosphatemia64–70*
Degree of hypophosphatemia IV phosphate replacement dosage*†
2.3–2.7 mg/dL (mild
1.5–2.2 mg/dL (moderate
?1.5 mg/dL (Severe
*In patients with normal renal function; patients with renal
insufficiency should receive ?50% of the initial empiric dose.
Maximum infusion rate ? 7 mmol phosphate/h.
†We suggest using adjusted body weight (AdjBW) in patients who
are significantly obese (weight ?130% of IBW or BMI ?30 kg/m2):
AdjBW (men) ? (?wt (kg) ? IBW(kg)? ? 0.3) ? IBW; AdjBW
(women) ? (?wt (kg) ? IBW(kg)? ? 0.25) ? IBW.
Treatment of hypokalemia71–73*
Degree of hypokalemia IV potassium
Rate of IV infusion†Maximum concentration
Serum potassium concentration ? 2.5–3.4 mEq/L
(mild to moderate hypokalemia, asymptomatic)
20–40 mEq10–20 mEq potassium/h;
maximum of 40 mEq
80 mEq/L via a peripheral vein;
up to 120 mEq/L via a central
vein (admixed in 0.9% sodium
chloride in water, or 0.45%
sodium chloride in water)
Serum potassium concentration ?2.5 mEq/L, or if
symptomatic (severe symptomatic hypokalemia)
*In patients with normal renal function; patients with renal insufficiency should receive ?50% of the initial empiric dose.
†Continuous cardiac monitoring and infusion via a central venous catheter are recommended for infusion rates ?10 mEq potassium per hour.
Vol. 20, No. 6
KRAFT ET AL
individual recommendations and practices may vary
Oral potassium supplementation can be provided,
but oral supplements can cause gastrointestinal side
effects (eg, cramping, diarrhea), and oral liquid
formulations have an unpleasant taste. We recom-
mend an oral potassium dose of 20–40 mEq, or a
total dose of 40–100 mEq/day as an initial regimen
to correct hypokalemia. Oral doses should be divided
into 2–4 doses to minimize gastrointestinal side
Serum potassium concentration should be moni-
tored at least daily during the first several days of
nutrition support. Because hypomagnesemia may
cause refractory hypokalemia, magnesium defi-
ciency should be corrected, along with potassium
supplementation, in order to facilitate the correction
Treatment of Hypomagnesemia
Magnesium deficiency has been associated with a
total body magnesium deficiency of 1–2 mEq/kg.75
IV treatment of hypomagnesemia should be the
preferred route in patients at risk for RS if symp-
tomatic and when the gastrointestinal tract cannot
be used. Oral magnesium supplements have a slow
onset and are associated with diarrhea and gastro-
intestinal intolerance. Suggestions for empiric IV
dosing of magnesium (for patients with normal renal
function) are listed in Table 5.39,41,75–83Because
magnesium distribution and equilibration between
serum and intracellular spaces and tissues are
of an IV dose of magnesium excreted in the
urine),31,32,75–77,80,82infusion time of IV magnesium is
important. In nonemergent situations, we recommend
infusing doses of ?6 g of magnesium sulfate over 6–12
hours and infusing higher doses over 12–24 hours,
with a maximum of 1 g magnesium sulfate (?8.1 mEq
istration rates may simply increase urinary loss of
magnesium. Additional supplementation may be
required, and total repletion of magnesium may take
several days. Severe symptomatic hypomagnesemia
may require more aggressive dosing in the acute
setting (eg, 4 g magnesium sulfate [?32 mEq elemen-
tal magnesium] over 4–5 minutes has been used in
patients with preeclampsia or eclampsia).75,80
For patients with impaired renal function, we
recommend using ?50% of the suggested empiric
magnesium dose. The patient must be monitored
approach the maximum recommendations (approxi-
mately 12 g magnesium sulfate [?97 mEq elemental
magnesium] over 12 hours).75Serum magnesium
concentration should be checked approximately
12–24 hours after magnesium repletion. Serum
magnesium concentrations can be monitored more
frequently in the acute setting; however, because of
the slow magnesium equilibrium,32,84serum magne-
sium concentration can seem artificially high if
measured too soon after a dose.79Serum concentra-
tions should be monitored at least once daily during
the first several days of nutrition support in patients
at risk for RS.
Restarting Nutrition Support
If a patient manifests signs and symptoms of RS,
nutrition support should be restarted with great
caution. All electrolyte abnormalities should be ade-
quately treated and supplemental electrolytes pro-
vided in the nutrition formulation above what was
previously provided when RS symptoms developed.
Multivitamins should also be supplemented as
described earlier. The patient should be free of
symptoms and stable before restarting nutrition
support. We suggest initiating nutrition support at
?50% of the previous rate when symptoms develop,
and advance nutrition to goal cautiously over at
least 4–5 days. The patient should be monitored
closely for further signs and symptoms of RS.
Treatment of hypomagnesemia39,41,75–83*
Degree of hypomagnesemia IV magnesium replacement dosage*†
Serum magnesium concentration ? 1–1.5 mg/dL
(mild to moderate hypomagnesemia, asymptomatic)
Serum magnesium concentration ?1 mg/dL (severe
Rate of IV infusion
1–4 g magnesium sulfate (8–32 mEq magnesium), up to 1
4–8 g magnesium sulfate (32–64 mEq magnesium), up to 1.5
Maximum of 1 g magnesium sulfate/h (8 mEq magnesium/h),
up to 12 g magnesium sulfate
(97 mEq magnesium) over 12 h if asymptomatic; up to 32
mEq magnesium over 4–5 min in severe symptomatic
*In patients with normal renal function; patients with renal insufficiency should receive ?50% of the initial empiric dose.
†We suggest using adjusted body weight (AdjBW) in patients who are significantly obese (weight ?130% of IBW or BMI ?30 kg/m2):
AdjBW (men) ? (?wt (kg) ? IBW(kg)? ? 0.3) ? IBW; AdjBW (women) ? (?wt (kg) ? IBW(kg)? ? 0.25) ? IBW.
‡One gram magnesium sulfate ? 8.1 mEq magnesium.
REVIEW OF THE REFEEDING SYNDROME
RS is a serious condition that can develop in
underweight, severely malnourished, or starved
individuals during nutrition repletion. RS involves
significant electrolyte, fluid, and vitamin abnormal-
ities that can lead to significant morbidity and
mortality. Clinicians should be aware of RS, identify
patients at risk of developing RS, and most impor-
tantly take steps to prevent RS. Patients who
develop signs and symptoms of RS require aggres-
sive electrolyte supplementation, vitamin supple-
mentation, and supportive care, and nutrition sup-
port should be restarted with great caution.
1. Keys A, Brozek J, Henschel A, Mickelson O, Taylor HD, eds. The
Biology of Human Starvation, Vols. 1, 2. Minneapolis, MN:
University of Minnesota Press; 1950.
2. Weinsier RL, Krumdieck CL. Death resulting from overzealous
total parenteral nutrition: the refeeding syndrome revisited. Am J
Clin Nutr. 1980;34:393–399.
3. Cahill GF. Starvation in man. N Engl J Med. 1970;282:668–675.
4. DeFronzo RA, Cooke CR, Andres R, Faloona GR, Davis PJ. The
effect of insulin on renal handling of sodium, potassium, calcium,
and phosphate in man. J Clin Invest. 1975;55:845–855.
5. Guirao X, Franch G, Gil MJ, Garcia-Domingo MI, Girvent M,
Sitges-Serra A. Extracellular volume, nutritional status, and
refeeding changes. Nutrition. 1994;10:558–561.
6. Van Way CW, Longoria M, Sacks GS. Do surgeons need to worry
about vitamin deficiencies? Nutr Clin Pract. 2001;16(suppl):S5–
7. Travis SF, Sugerman HJ, Ruberg RL, et al. Alterations of red-cell
glycolytic intermediates and oxygen transport as a consequence of
hypophosphatemia in patients receiving intravenous hyperali-
mentation. N Engl J Med. 1971;285:763–768.
8. Silvis SE, Paragas PD. Paresthesias, weakness, seizures, and
hypophosphatemia in patients receiving hyperalimentation. Gas-
9. Sand DW, Pastore RA. Paresthesias and hypophosphatemia
occurring with parenteral alimentation. Am J Dig Dis. 1973;18:
10. Furlan AJ, Hanson M, Cooperman A, Farmer RG. Acute areflexic
paralysis: association with hyperalimentation and hypophos-
phatemia. Arch Neurol. 1975;32:706–707.
11. Sheldon GF, Grzyb S. Phosphate depletion and repletion: relation
to parenteral nutrition and oxygen transport. Ann Surg. 1975;
12. Baughman FA Jr, Papp JP. Wernicke’s encephalopathy with
intravenous hyperalimentation: remarks on similarities between
Wernicke’s encephalopathy and the phosphate depletion syn-
drome. Mt Sinai J Med. 1976;43:48–52.
13. Silvis SE, DiBartolomeo AG, Aaker HM. Hypophosphatemia and
neurological changes secondary to oral caloric intake. Am J
14. Youssef HAE. Hypophosphatemic respiratory failure complicat-
ing PN: an iatrogenic potentially lethal hazard [letter]. Anesthe-
15. Hayek ME, Eisenberg PG. Severe hypophosphatemia following
the institution of enteral feedings. Arch Surg. 1989;124:1325–
16. Vanneste J, Hage J. Acute severe hypophosphatemia mimicking
Wernicke’s encephalopathy [letter]. Lancet. 1986;1:44.
17. Mattioli S, Miglioli M, Montagna P, Lerro MF, Pilotti V, Gozzetti
G. Wernicke’s encephalopathy during PN: observation in one
case. JPEN J Parenter Enteral Nutr. 1988;12:626–627.
18. O’Connor LR, Wheeler WS, Bethune JE. Effect of hypophos-
phatemia on myocardial performance in man. N Engl J Med.
19. Newman JH, Neff TA, Ziporin P. Acute respiratory failure asso-
ciated with hypophosphatemia. N Engl J Med. 1977;296:1101–
20. Aubier M, Murciano D, Lecocguic Y, et al. Effect of hypophos-
phatemia on diaphragmatic contractility in patients with acute
respiratory failure. N Engl J Med. 1985;313:420–424.
21. Benesch R, Benesch RE. The effect of organic phosphates from the
human erythrocyte on the allosteric properties of hemoglobin.
Biochem Biophys Res Commun. 1967;26:162–167.
22. Lichtman MA, Miller DR, Cohen J, Waterhouse C. Reduced red
cell glycolysis, 2,3-diphosphoglycerate and adenosine triphos-
phate concentration, and increased hemoglobin-oxygen affinity
caused by hypophosphatemia. Ann Intern Med. 1971;74:562–568.
23. Thomas HM, Lefrak SS, Irwin RS, Fritts Jr HW, Caldwell PRB.
The oxyhemoglobin dissociation curve in health and disease: role
of 2,3-diphosphoglycerate. Am J Med. 1974;57:331–348.
24. Mandal AK. Hypokalemia and hyperkalemia. Med Clin North
25. Halperin ML, Kamel KS. Potassium. Lancet. 1998;352:135–140.
26. Freedman BI, Burkart JM. Hypokalemia. Crit Care Clin. 1991;7:
27. Brophy DF, Gehr TWB. Disorders of potassium and magnesium
homeostasis. In: Dipiro JT, Talbert RL, Yee GC, et al, eds.
Pharmacotherapy: A Pathophysiologic Approach. 5thed. New
York, NY: McGraw-Hill; 2002:981–993.
28. Gennari FJ. Hypokalemia. N Engl J Med. 1998;339:451–458.
29. Agarwal A, Wingo CS. Treatment of hypokalemia. N Engl J Med.
30. Wacker WEC, Parisi AF. Magnesium metabolism. N Engl J Med.
1968;278:658–663, 712–717, 772–776.
31. Graber TW, Yee AS, Baker FJ. Magnesium: physiology, clinical
disorders, and therapy. Ann Emerg Med. 1981;10:49–57.
32. Reinhart RA. Magnesium metabolism: a review with special
reference to the relationship between intracellular content and
serum levels. Arch Intern Med. 1988;148:2415–2420.
33. Wacker WE. The biochemistry of magnesium. Ann N Y Acad Sci.
34. Reinhart RA, Desbiens NA. Hypomagnesemia in patients enter-
ing the ICU. Crit Care Med. 1985;13:506–507.
35. Ryzen E, Wagers PW, Singer FR, Rude RK. Magnesium deficiency
in a medical ICU population. Crit Care Med. 1985;13:19–21.
36. Chernow B, Bamberger S, Stoiko M, et al. Hypomagnesemia in
patients in postoperative intensive care. Chest. 1989;95:391–397.
37. Frankel H, Haskell R, Lee SY, et al. Hypomagnesemia in trauma
patients. World J Surg. 1999;23:966–969.
38. Rubiez GJ, Thill-Baharozian M, Hardie D, Carlson RW. Associa-
tion of hypomagnesemia and mortality in acutely ill medical
patients. Crit Care Med. 1993;21:203–209.
39. Ceremuzynski L, Hao NV. Ventricular arrhythmias late after
myocardial infarction are related to hypomagnesemia and mag-
nesium loss: preliminary trial of corrective therapy. Clin Cardiol.
40. Kingston ME, Al-Siba’I MB, Skooge WC. Clinical manifestations
of hypomagnesemia. Crit Care Med. 1986;14:950–954.
41. Salem M, Munoz R, Chernow B. Hypomagnesemia in critical
illness: a common and clinically important problem. Crit Care
42. Seelig MS. Electrographic patterns of magnesium depletion
appearing in alcoholic heart disease. Ann N Y Acad Sci. 1969;162:
43. Skou JC. Further investigations on a Mg??? Na?-activated
adenosine triphosphate, possibly related to the active, linked
transport of Na?and K?across the nerve membrane. Biochem
Biophys Acta. 1960;42:6–23.
44. Weisinger JR, Bellorin-Font E. Magnesium and phosphorous.
45. Anast CS, Winnacker JL, Forte LR, Burns TW. Impaired release
of parathyroid hormone in magnesium deficiency. J Clin Endo-
crinol Metab. 1976;42:707–717.
46. Fatemi G, Ryzen E, Flores J, Endres DB, Rude RK. Effect of
experimental human magnesium depletion on parathyroid hor-
mone secretion and 1,25-dihydroxyvitamin D metabolism. J Clin
Endocrinol Metab. 1991;73:1067–1072.
Vol. 20, No. 6
KRAFT ET AL
47. Reuler JB, Girard DE, Cooney TG. Wernicke’s encephalopathy. Download full-text
N Engl J Med. 1985;312:1035–1039.
48. Romanski SA, McMahon MM. Metabolic acidosis and thiamine
deficiency. Mayo Clin Proc. 1999;74:259–263.
49. Kitamura K, Takahashi T, Tanaka H, et al. Two cases of thiamine
deficiency-induced lactic acidosis during PN. Tohoku J Exp Med.
50. Barrett TG, Forsyth JM, Nathavitharana KA, Booth IW. Poten-
tially lethal thiamine deficiency complicating parenteral nutrition
in children. Lancet. 1993;341:901–902.
51. Centers for Disease Control and Prevention. Deaths associated
with thiamine-deficient PN. MMWR Morb Mortal Wkly Rep.
52. Centers for Disease Control and Prevention. Lactic acidosis
traced to thiamine deficiency related to nationwide shortage of
multivitamins for PN. MMWR Morb Mortal Wkly Rep. 1997;46:
53. Heymsfield SB, Bethel RA, Ansley JD, Gibbs DM, Felner JM,
Nutter DO. Cardiac abnormalities in cachectic patients before
and during nutritional repletion. Am Heart J. 1978;95:584–594.
54. Huang YL, Fang CT, Tseng MC, Lee YJ, Lee MB. Life-threaten-
ing refeeding syndrome in a severely malnourished anorexia
nervosa patient. J Formos Med Assoc. 2001;100:343–346.
55. Gottdiener JS, Gross HA, Henry WL, Borer JS, Ebert MH. Effects
of self-induced starvation on cardiac size and function in anorexia
nervosa. Circulation. 1978;58:425–433.
56. Apovian CM, McMahon MM, Bistran BR. Guidelines for refeeding
the marasmic patient. Crit Care Med. 1990;18:1030–1033.
57. Nutrition Advisory Group, AMA Department of Foods and Nutri-
tion. Multivitamin preparations for parenteral use, 1975. JPEN J
Parenter Enteral Nutr. 1979;3:258–262.
58. Brooks MJ, Melnik G. The refeeding syndrome: an approach to
understanding its complications and preventing its occurrence.
59. Hankin ME, Munz K, Steinbeck AW. Total body water content in
normal and grossly obese women. Med J Aust. 1976;2:533–537.
60. Watson PE, Watson ID, Batt RD. Total body water volumes for
adult males and females estimated from simple anthropometric
measurements. Am J Clin Nutr. 1980;33:27–39.
61. Schoeller DA, van Santen E, Peterson DW, et al. Total body water
measurement in humans with18O and2H labeled water. Am J
Clin Nutr. 1980;33:2686–2693.
62. Webster JD, Hesp R, Garrow JS. The composition of excess
weight in obese women estimated by body density, total body
water and total body potassium. Hum Nutr Clin Nutr. 1984;38:
63. Foster GD, Wadden TA, Mullen JL, et al. Resting energy expen-
diture, body composition, and excess weight in the obese. Metab-
64. Lentz RD, Brown DM, Kjellstrand CM. Treatment of severe
hypophosphatemia. Ann Intern Med. 1978;89:941–944.
65. Vannatta JB, Whang R, Papper S. Efficacy of intravenous phos-
phorous therapy in the severely hypophosphatemic patient. Arch
Intern Med. 1981;141:885–887.
66. Andress DL, Vannatta JB, Whang R. Treatment of refractory
hypophosphatemia. South Med J. 1982;75:766–767.
67. Vannatta JB, Andress DL, Whang R, Papper S. High-dose intra-
venous phosphorus therapy for severe complicated hypophos-
phatemia. South Med J. 1983;76:1424–1426.
68. Kingston M, Al-Siba’i MB. Treatment of severe hypophos-
phatemia. Crit Care Med. 1985;13:16–18.
69. Rosen GH, Boullata JI, O’Rangers EA, Enow NB, Shin B. Intra-
venous phosphate repletion regimen for critically ill patients with
moderate hypophosphatemia. Crit Care Med. 1995;23:1204–1210.
70. Clark CL, Sacks GS, Dickerson RN, Kudsk KA, Brown RO.
Treatment of hypophosphatemia in patients receiving specialized
nutrition support using a graduated dosing scheme: results from
a prospective clinical trial. Crit Care Med. 1995;23:1504–1511.
71. Kruse JA, Carlson RW. Rapid correction of hypokalemia using
concentrated intravenous potassium chloride infusions. Arch
Intern Med. 1990;150:613–617.
72. Hamill RJ, Robinson LM, Wexler HR, Moote C. Efficacy and
safety of potassium infusion therapy in hypokalemic critically ill
patients. Crit Care Med. 1991;19:694–699.
73. Kruse JA, Clark VL, Carlson RW, Geheb MA. Concentrated
potassium chloride infusions in critically ill patients with hypo-
kalemia. J Clin Pharmacol. 1994;34:1077–1082.
74. Ryan MP. Interrelationships of magnesium and potassium
homeostasis. Mineral Electrolyte Metab. 1993;19:290–295.
75. Flink EB. Therapy of magnesium deficiency. Ann N Y Acad Sci.
76. Heaton FW. The kidney and magnesium homeostasis. Ann N Y
Acad Sci. 1969;162:775–785.
77. Martin HE. Clinical magnesium deficiency. Ann N Y Acad Sci.
78. Dickerson RN, Brown RO. Hypomagnesemia in hospitalized
patients receiving nutritional support. Heart Lung. 1985;14:561–
79. Rasmussen HS, McNair P, Norregard P, Backer V, Lindeneg O,
Balslev S. Intravenous magnesium in acute myocardial infarc-
tion. Lancet. 1986;1:234–236.
80. Oster JR, Epstein M. Management of magnesium depletion. Am J
81. Sacks GS, Brown RO, Dickerson RN, et al. Mononuclear blood cell
magnesium content and serum magnesium concentration in crit-
ically ill hypomagnesemic patients after replacement therapy.
82. Hebert P, Mehta N, Wang J, Hindmarsh T, Jones G, Cardinal P.
Functional magnesium deficiency in critically ill patients identi-
fied using a magnesium-loading test. Crit Care Med. 1997;25:
83. Huycke MM, Naguib MT, Stroemmel MM, et al. A double-blind
placebo-controlled crossover trial of intravenous magnesium sul-
fate for foscarnet-induced ionized hypocalcemia and hypomag-
nesemia in patients with AIDS and cytomegalovirus infection.
Antimicrob Agents Chemother. 2000;44:2143–2148.
84. Barnes BA. Magnesium conservation: a study of surgical patients.
Ann N Y Acad Sci. 1969;162:786–801.
REVIEW OF THE REFEEDING SYNDROME