Predictive value of estimated ionized magnesium for mechanical ventilation in children with asthma: a single-centre...

Asthma exacerbations are a frequent cause of paediatric intensive care admissions, and early identification of patients at risk of respiratory failure is essential. Because direct ionized magnesium measurement is not routinely available, we evaluated whether estimated ionized magnesium derived from routinely available biochemical data was associated with subsequent invasive mechanical ventilation. We conducted a retrospective observational study of children admitted to a paediatric intensive care unit in Japan between 2016 and 2024 with asthma. Primary exposure was considered as the admission-phase estimated ionized magnesium value, defined as the earliest available value before initiation of invasive ventilation, and applied consistently to ventilated and non-ventilated children. The primary outcome was the need for mechanical ventilation. Receiver operating characteristic analysis and multivariable logistic regression were used to evaluate the predictive performance of estimated ionized magnesium while adjusting for arterial pH and arterial carbon dioxide tension. Among 115 eligible children, 51 (44.3%) required invasive mechanical ventilation. Estimated ionized magnesium was lower in ventilated than in non-ventilated children (median 0.467 vs 0.632 mmol/L; p < 0.001). The area under the receiver operating characteristic curve was 0.90, with a 95% confidence interval of 0.83 to 0.96. Each 0.1-mmol/L decrease in estimated ionized magnesium remained associated with higher odds of mechanical ventilation after adjustment for pH and arterial carbon dioxide tension (adjusted odds ratio: 9.46; 95% confidence interval: 3.94 to 22.71; p < 0.001). Admission-phase estimated ionized magnesium was strongly associated with invasive mechanical ventilation in children with asthma exacerbations and may provide a practical early risk marker.