Estimation formulas for ionized magnesium in critically ill children

Introduction

Clinically, ionized magnesium (iMg) measurement requires specialized equipment, leading many healthcare providers to rely on total magnesium (tMg) as a surrogate marker. However, accurately assessing magnesium homeostasis necessitates precise iMg measurement. Given this challenge, we conducted a study to develop predictive formulas for estimating iMg levels based on molecules involved in acid-base regulation.

Materials and methods

This exploratory study was conducted prospectively with pediatric patients under 18 years of age admitted to the pediatric intensive care unit in a tertiary children’s hospital in Japan between July and August 2024. We measured serum iMg levels using the Stat Profile Prime analyzer from Nova Biomedical. Concurrent blood samples were collected for ion analysis following the Stewart approach. Multiple regression analysis was utilized to develop an equation for estimating iMg levels. The accuracy of the developed equations was further assessed using Bland-Altman analysis and weighted κ coefficients.

Results

A total of 200 samples were analyzed. Spearman’s rank correlation coefficient between iMg and tMg concentrations was 0.915, indicating a strong and statistically significant association in the simple regression analysis (β = 0.275, 95 % confidence interval 0.271–0.278). Three predictive equations were developed using multivariate analysis. These models were tested on another 86 additional samples. Bland-Altman analysis demonstrated minimal error between measured and estimated values, with κ coefficients indicating high concordance.

Conclusions

We successfully developed an iMg estimation model using routinely measured clinical parameters.