Abstract:

This retrospective study included 139 critically ill patients who received micafungin therapy at the First Affiliated Hospital of Nanchang University between January 2020 and December 2023. Patients were classified into a hypomagnesemia group (serum magnesium < 0.75 mmol/L) and a non-hypomagnesemia group. Demographic characteristics, laboratory data, and concomitant medications were collected. Univariate and multivariate logistic regression analyses were performed to identify independent risk factors. A predictive nomogram was subsequently developed using the rms package in R software and underwent internal validation. The incidence of hypomagnesemia in this cohort was 33.8% (47/139). Univariate analysis revealed significant between-group differences in SOFA score, duration of micafungin therapy, baseline serum creatinine, baseline serum magnesium, and post-treatment serum creatinine (all P < 0.05). Multivariate analysis further identified micafungin treatment duration (OR = 1.117, 95% CI: 1.056–1.181; P < 0.001), post-treatment albumin (OR = 0.889, 95% CI: 0.807–0.979; P = 0.017), and baseline serum creatinine (OR = 0.992, 95% CI: 0.987–0.997; P = 0.001) as independent predictors of hypomagnesemia. The nomogram demonstrated strong discriminatory ability (AUC = 0.807, 95% CI: 0.728–0.886), with corresponding sensitivity and specificity of 66.0% and 87.0%, respectively. The calibration curve showed a mean absolute error of 0.027, and the Hosmer-Lemeshow test confirmed satisfactory model calibration (χ² = 13.12, P = 0.11). Overall, this study revealed a notably high incidence of micafungin-induced hypomagnesemia among critically ill patients. Treatment duration, post-treatment albumin level, and baseline serum creatinine emerged as independent risk factors. The proposed nomogram offered a practical and reliable tool for predicting the risk of micafungin-related hypomagnesemia in the critically ill population.

Key words: Critically ill patients, Micafungin, Hypomagnesemia, Risk factors, Prediction model