Outcome assessment of acute methanol poisoning: A risk-prediction nomogram approach for in-hospital mortality

Abstract

Acute methanol poisoning could be associated with high morbidities and fatalities. Stratifying high-risk patients is crucial in improving their prognosis. Hence, this study aimed to identify patients with methanol poisoning at high risk of in-hospital mortality. Also, the risk factors for blindness were assessed. The study included 180 acutely methanol-poisoned patients who received standard medical care. Out of 180 patients, 52 (28.9 %) patients presented with blindness, and 43 (23.9 %) patients died. The predictive model was based on four significant variables, including blindness, mean arterial pressure, serum bicarbonate, and serum creatinine. The presence of blindness and elevated serum creatinine significantly increased the likelihood of mortality by 14.274 and 5.670 times, respectively. Likewise, decreases in mean arterial pressure and serum bicarbonate significantly increased mortality risk by 0.908 and 0.407 times, respectively. The proposed nomogram exhibited excellent discriminatory power (area under the curve (AUC)=0.978, accuracy=93.3 %), which outperforms the AUCs of individual predictors. The provided nomogram is easily applicable with outstanding discrimination, making it clinically helpful in predicting in-hospital mortality in acutely methanol-poisoned patients. Regarding the risk factors for blindness, multivariable regression analysis revealed that delayed time for admission (OR=1.039; 95 % CI=1.010-1.069; p= 0.009) and elevated anion gap (OR=1.053; 95 % CI=1.007-1.101; p= 0.023) were significant risk factors. The current study assists physicians in identifying methanol-poisoned patients with a high probability of mortality or blindness on admission. Future studies are recommended for external validation of the created nomogram, in addition to follow-up for patients with visual impairment.

Keywords: Methanol Poisoning; Mortality; Nomogram; Outcome; Severe Visual Impairment, Blindness.

Graphical abstract

Fig. 1

Risk-prediction nomogram for prediction of in-hospital mortality among methanol-poisoned patients.

Fig. 2

Applying the nomogram to predict the probability of in-hospital mortality of a 40-year-old male patient with acute methanol poisoning. He was presented with blindness. On-admission, parameters included MAP = 60 mmHg, HCO₃ = 4.5 mmol/L, and serum creatinine = 1.9 mg/dl. The probability was calculated: blindness, MAP, HCO₃, and serum creatinine correspond to 1.3, 3, 8.1, and 1.6 points, respectively. The result of the summation of these points is 14, which means > 99 % risk of mortality.

Fig. 3

(A) ROC curves for the quantitative predictors from the logistic regression model (MAP, HCO₃, and serum creatinine) predicting in-hospital mortality among methanol-poisoned patients.; (B) ROC curves testing the performance of the developed nomogram.

Fig. 4

Calibration curve testing the accuracy of the developed nomogram in predicting in-hospital mortality among methanol-poisoned patients.