Intended for healthcare professionals


Refeeding syndrome

BMJ 2004; 328 doi: (Published 15 April 2004) Cite this as: BMJ 2004;328:908
  1. Stephen D Hearing, consultant gastroenterologist (stephen.hearing{at}
  1. >Department of Gastroenterology, Staffordshire General Hospital, Weston Road, Stafford ST16 3SA

    Is underdiagnosed and undertreated, but treatable

    Refeeding syndrome was first described in Far East prisoners of war after the second world war.1 Starting to eat again after a period of prolonged starvation seemed to precipitate cardiac failure. The pathophysiology of refeeding syndrome has now been established.2 In starvation the secretion of insulin is decreased in response to a reduced intake of carbohydrates. Instead fat and protein stores are catabolised to produce energy. This results in an intracellular loss of electrolytes, in particular phosphate. Malnourished patients' intracellular phosphate stores can be depleted despite normal serum phosphate concentrations. When they start to feed a sudden shift from fat to carbohydrate metabolism occurs and secretion of insulin increases. This stimulates cellular uptake of phosphate, which can lead to profound hypophosphataemia.3 This phenomenon usually occurs within four days of starting to feed again.

    Phosphate is necessary for the generation of adenosine triphosphate from adenosine diphosphate and adenosine monophosphate and other crucial phosphorylation reactions. Serum phosphate concentrations of less than 0.50 mmol/l (normal range 0.85-1.40 mmol/l) can produce the clinical features of refeeding syndrome, which include rhabdomyolysis, leucocyte dysfunction, respiratory failure, cardiac failure, hypotension, arrhythmias, seizures, coma, and sudden death.4 5Importantly, the early clinical features of refeeding syndrome are non-specific and may go unrecognised.

    Refeeding syndrome can occur with parenteral as well as enteral feeding. In the United Kingdom patients with anorexia nervosa, cancer, alcoholism, and some patients after operations are known to be at risk of refeeding syndrome.6 However, other groups, such as patients with neurological dysphagia who are being fed through nasogastric or percutaneous endoscopic gastrostomy tubes, may also be at risk.

    Understanding of refeeding syndrome and its treatment is limited among general physicians and surgeons. Many patients at risk of refeeding syndrome are not treated on specialist nutrition units. Measurement of serum phosphate may not be done in patients at risk, and when phosphate is measured the importance of grossly abnormal results may not be recognised. The other barrier is a lack of consensus on treatment. Intravenous phosphate is required as oral supplementation is inadequate. Previously recommended regimens for treatment of severe hypophosphataemia have been developed mainly from experience in treating small numbers of patients in intensive care settings.7 Typically regimens recommend multiple infusions based on weight, with frequent monitoring of serum phosphate, but these complex regimens may be impractical on general wards. Ideally a treatment regimen for refeeding syndrome would not require monitoring blood tests more often than daily, would not require adjustment for weight, and would provide an adequate dose of phosphate.

    Results of such a regimen have recently been published.8 Thirty patients with refeeding syndrome, normal renal function, and a phosphate concentration of less than 0.50 mmol/l were treated with 50 mmol intravenous phosphate over 24 hours (500 mlPhosphates Polyfusor, Fresenius Kabi, Warrington, United Kingdom). This treatment was effective; 93% (28/30) achieved a serum phosphate concentration of 0.50 mmol/l or more after four days. Importantly, five patients required further phosphate as severe hypophosphataemia recurred after initial correction. The treatment seemed safe; no patient developed renal failure, although three patients developed mild transient hyperphosphataemia and four asymptomatic hypocalcaemia. Importantly, all patients were managed on general wards. Although this study is uncontrolled, it is the largest published series of the treatment of severe hypophosphataemia due to refeeding syndrome.

    Treatment of refeeding syndrome can be helped by the input of hospitals' nutrition teams. Dieticians and nutrition nurses can help in identifying malnourished patients at risk of developing refeeding syndrome. When these patients require artificial feeding (enteral or parenteral), this should be started at a reduced calorific rate (25-50% of estimated requirements) to reduce the risk of refeeding syndrome developing. Serum phosphate, magnesium, calcium, potassium, urea, and creatinine concentrations should be measured before feeding and repeated daily for four days after feeding is started. When hypophosphataemia occurs it should be corrected in addition to other electrolyte abnormalities, such as hypokalaemia and hypomagnesaemia. Feeding can be continued. If after 24 hours the serum phosphate concentration remains low or falls subsequently then further phosphate should be administered.

    No randomised controlled trials of treatment for refeeding syndrome have been performed, and the optimal regime therefore remains to be determined. In addition the exact degree of hypophosphataemia requiring treatment remains to be determined, although most experts on nutrition would recommend treatment of hypophosphataemia with measurements of 0.32-0.50 mmol. Randomised controlled trials of the treatment of refeeding syndrome, with clinical end points such as survival, are therefore needed. The current priority is to improve awareness of refeeding syndrome among general physicians and surgeons and to convince them of its importance. Many doctors remain unconvinced of the importance of treating hypophosphataemia, while in contrast accepting the need to treat other electrolyte deficiencies such as hypokalaemia and hypomagnesaemia. With further liaison between hospital nutrition teams and ward staff and the use of an adequate treatment regimen, it may be possible to improve the prognosis of this currently underecognised and undertreated group substantially.


    • Competing interests None declared.


    View Abstract

    Log in

    Log in through your institution


    * For online subscription