Abstract
AIM: To investigate development of refeeding hypophosphatemia during the refeeding period and the extent of the decrease in the serum phosphorus level among anorexia nervosa patients with severe malnutrition. OBJECTIVE: The accurate prediction of the severity of refeeding hypophosphatemia in patients with anorexia nervosa during acute treatment is of great importance. Although some predictors were found in previous reports, these studies used binominal data-the presence or absence of hypophosphatemia-as an outcome indicator but not the extent of serum phosphorus level decrease. It is crucial in clinical settings to predict the extent of the serum phosphorus level decrease as well as development of refeeding hypophosphatemia, in particular, for patients with severe malnutrition, who has a higher risk of death. METHODS: We investigated 63 admissions from 37 patients with anorexia nervosa who had severe malnutrition (admission body mass index 11.5 ± 1.6) and carried out a linear discriminant regression analysis for the development of refeeding hypophosphatemia. The extent of the decrease in the serum phosphorus level were investigated using multiple linear regression analysis. Explanatory variables included data upon admission (age, sex, body mass index, blood urea nitrogen to creatinine ratio, albumin, initial serum phosphorus level, anorexia nervosa type, i.e., restrictive or binge-purge) as well as treatment-related indicators (calorie intake, amount of phosphate administered, and rate of weight gain). RESULTS: Development of refeeding hypophosphatemia and a change in serum phosphorus levels were predicted by body mass index and elevated blood urea nitrogen to creatinine ratio. CONCLUSIONS: Our study found that refeeding hypophosphatemia among patients with severe malnutrition was predicted by a lower body mass index and elevated blood urea nitrogen to creatinine ratio.