Hypophosphatemia and its clinical implications in critically ill children: a retrospective study.

PURPOSE: The aims of this study were to determine the prevalence of hypophosphatemia and to discuss the clinical implications of hypophosphatemia in critically ill children.

MATERIALS AND METHODS: A retrospective review of the medical records of children admitted to the pediatric intensive care unit from December 2006 to December 2007 was conducted.

RESULTS: In 60.2% (n = 71) of the patients, any serum phosphorous level at admission and at the third day or seventh day after admission to pediatric intensive care unit was in hypophosphatemic range. Sepsis was present in 22.9% (n = 27) of the children studied and was associated with hypophosphatemia (P = .02). Hypophosphatemia was also associated with use of furosemide (P = .04), use of steroid (P = .04), use of β(2) agonist (P = .026), and use of an H(2) blocker (P = .004). There was a significant association between hypophosphatemia and the rate to attain target caloric requirements by enteral route (P = .007). The median time to attain target caloric requirements by enteral route was 2.9 ± 1.9 (0.2-10) days in the normophosphatemic group and 4.4 ± 2.8 (0.3-12) days in the hypophosphatemic group. In the multiple regression model, solely the rate to attain the target caloric requirements by enteral route demonstrated independent association with hypophosphatemia (P = .006; β = .27; 95% confidence interval, 0.02-0.09). Significant association was found between hypophosphatemia and the duration of mechanical ventilation and between hypophosphatemia and pediatric intensive care unit length of stay (P = .02 and P = .001, respectively).

CONCLUSIONS: Critically ill pediatric patients are prone to hypophosphatemia, especially if they cannot be fed early by enteral route. Hypophosphatemia is associated with an increased duration of mechanical ventilation and increased length of stay in the pediatric intensive care unit, suggesting that active repletion might improve these parameters.