Evaluating the frequency rate of hypomagnesemia in critically ill pediatric patients hy using multiple regression analysis and a computer-based neural network

Objectives: To determine the frequency rate of hypomagnesemia in patients a dmitted to the pediatric intensive care unit (ICU), and to identify subsets of patients (grouped by disease) who are at greatest risk of hypomagnesemi a. We also compared a neural network model with multiple regression analysi s to identify independent variables that would correlate with hypomagnesemi a and to predict serum magnesium values in critically ill pediatric patient s overall. Design: Prospective, multicenter study. Setting: Tertiary level medical/surgical pediatric ICUs. Patients: Data were obtained at admission to the pediatric ICU for 463 pati ents from newborn to 18 yrs old who were admitted with a variety of surgica l and nonsurgical conditions. Interventions: None. Measurements and Main Results: Total serum magnesium values were obtained w ithin the first 24 hrs after admission in 463 pediatric patients admitted t o four pediatric ICUs. Hypomagnesemia (defined as total serum magnesium <0. 75 mmol/L) was found in 51 (11%) of the 463 patients, with the highest freq uency rate (72%) and lowest mean serum magnesium level (0.66 +/- 0.17 mmol/ L) in patients admitted after surgery with extensive osseous involvement (s pinal fusion and craniofacial reconstruction). To determine whether hypomag nesemia could be predicted on the basis of other laboratory and clinical cr iteria, multiple regression analysis was performed and showed age, weight, and albumin levels weakly associated (r(2) = .14, p < .001) with magnesium levels within the different diagnostic groups. These data were used to prod uce a mathematical model able to predict magnesium levels within 5% of the actual values in 23% of patients. A neural network was also created to comp are its predictive capabilities to those of the multiple regression model. Once trained on a random subset (85%) of the patient population, the neural network was able to predict magnesium revels to within 5% of actual values for 88% of the remaining 15% of patients, comparing favorably with the pre dictions derived from the multiple regression model. Conclusions: Hypomagnesemia is not uncommon (11%) in critically ill pediatr ic patients, but is very common (72%) in patients admitted after surgery fa r spinal fusion or craniofacial reconstruction. Patients who undergo surger y for correction of scoliosis and craniofacial anomalies should have serum magnesium levels monitored closely after surgery. In other patients, a neur al network or multiple regression model could help predict which patients w ould be at risk of developing hypomagnesemia, thereby focusing testing on p atients likely to benefit from such testing.