High prevalence of hypophosphataemia at PICU admission in non-malnourished children

Dear Editor, Hypophosphataemia, which is frequent in critically ill adults [1], may induce myocardial dysfunction [2]. Hypophosphataemia was recently reported in two-thirds of children in a Brazilian paediatric intensive care unit (PICU) [3], with malnutrition as the main risk factor [4]. No studies have assessed prevalence of hypophosphataemia at PICU admission in countries where malnutrition is exceedingly rare. We reviewed the medical records of all patients admitted to our PICU in Paris, France, from January 2007 to April 2009. We collected causes of admission; sex; age; weight; PRISM III score and predicted risk of mortality; calcaemia; phosphataemia; C-reactive protein (CRP); use of steroids, diuretics, and catecholamines; times on mechanical ventilation and in the PICU; and deaths. Hypophosphataemia was defined as phosphate B1.15 mmol/L [5]. The Z-score for weight was used as a marker of overall nutritional status, because stature and brachial circumference were rarely collected at admission. Data were described as median (inter-quartile range) or percentage and compared using Wilcoxon, Mann–Whitney, or v test. Factors associated with hypophosphataemia (p \ 0.10) on univariate analysis were entered into a multivariable model using a multipleimputation method to include missing data, and odds ratios with 95% confidence intervals were computed. Mortality was compared using Wilcoxon–Mann–Whitney test. All calculations were performed using SAS v9.1 software (SAS Institute Inc., Cary, NC, USA). Of the 1,537 PICU stays, 613 stays had phosphataemia measured at admission. These patients were older [24 (4;95) months versus 14 (2;57) months; p = 0.001] and had higher risk of mortality [2.2% (1.0%;5.9%) versus 1.9% (0.8%;3.7%); p \ 0.0001] than children without phosphataemia measurements. The results showed hypophosphataemia for 128 stays, including severe hypophosphataemia (\0.6 mmol/L) for 6 stays (Table 1). On univariate analysis, factors significantly associated with hypophosphataemia at admission were CRP [100 mg/L [2.74 (1.74;4.31), p \ 0.0001], PRISM [4 [1.55 (1.00;2.38), p = 0.05], steroids [2.32 (1.38;3.92), p = 0.002], female gender [1.41 (0.95;2.08), p = 0.09], postoperative period [0.60 (0.34;1.07), p = 0.08] and hypocalcaemia [1.56 (1.00;2.43), p = 0.05]. On multivariable analysis, only CRP [100 mg/L [2.86 (1.79;4.59), p \ 0.0001] and PRISM [4 [1.61 (1.03;2.51), p = 0.03] remained significant. Hypophosphataemia at admission had no effect on mechanical ventilation duration (p = 0.40), PICU length of stay (p = 0.91) or mortality (p = 0.82). In this large sample, one-fifth of patients had hypophosphataemia at PICU admission. We employed a similar threshold to define hypophosphataemia as has been used in recent studies [3, 4]. Weight Z-score