Predictors of mortality in patients with acute methanol intoxication: a retrospective analysis of clinical, laboratory, and radiological findings

This study evaluated clinical, laboratory, and radiological predictors of in-hospital mortality in ICU patients with acute methanol intoxication. Fifty-seven adult patients admitted to a tertiary ICU between December 2024 and March 2025 were included. Normality was assessed using the Shapiro-Wilk test. Normally distributed variables are presented as mean ± SD (independent t-test) and non-normally distributed variables as median (IQR) (Mann-Whitney U test). Categorical variables were compared with Fisher’s exact test. ROC analysis with bootstrap 95% confidence intervals (1000 resamples) was performed for all significant continuous predictors; optimal cutoffs were identified using the Youden index. DeLong’s method was used to formally compare AUC values between ROC curves. Multivariable binary logistic regression was limited to a maximum of two predictors in accordance with the events-per-variable (EPV ≥ 10) principle. Twenty-nine patients (50.9%) died. Composite physiological severity scores demonstrated the highest discriminatory performance: APACHE II score achieved AUC = 0.950 (95%CI 0.890–0.991; an exploratory cutoff of > 23 was identified by the Youden index in this cohort) and SOFA score AUC = 0.895, without a statistically significant difference between them (DeLong p = 0.201). The two primary DeLong comparisons confirmed that composite scores significantly outperformed arterial pH (ΔAUC = 0.155, p = 0.011) and serum methanol (ΔAUC = 0.281, p < 0.001). In multivariable logistic regression including APACHE II and arterial pH, APACHE II was the only significant predictor (OR = 1.346, 95%CI 1.152–1.572, p < 0.001; pH: OR = 0.038, p = 0.134). All non-survivors required mechanical ventilation. Diffuse/severe brain involvement was associated with 90.9% mortality in a small subgroup ( n = 11, p = 0.005); this finding is exploratory. In acute methanol intoxication, composite physiological severity scores outperformed individual toxicological and metabolic biomarkers in mortality discrimination. These findings suggest that comprehensive physiological assessment may offer superior risk stratification compared with individual toxicological or metabolic parameters in this population. These results pertain to mortality prediction and do not address the therapeutic roles of arterial pH and serum methanol concentration, which remain essential for guiding antidotal therapy and dialysis decisions.