Construction of a Mortality Prediction Model for Acute Organophosphorus Pesticide Poisoning Patients Undergoing...

Organophosphorus pesticide poisoning (AOPP) leads to high mortality among patients. Hemoperfusion can remove organophosphates in the blood circulation and is the treatment of choice in these patients. However, studies on mortality risk factors in AOPP patients undergoing hemoperfusion are limited.

A retrospective study was conducted on AOPP patients undergoing hemoperfusion between January 2016 and August 2023. Baseline characteristics, laboratory test results, and in-hospital mortality data were collected. The relationships between risk factors and survival status were analyzed. A nomogram predictive model was constructed and evaluated.

Of 216 patients included, 183 and 33 were in the survival and non-survival groups, respectively. Univariate and multivariate analyses revealed that the Glasgow Coma Scale (GCS) at admission, age, white blood cell count, total protein, pH, and cholinesterase change after first hemoperfusion were statistically significantly associated with in-hospital death (odds ratios with 95% CI were 0.71, 0.58-0.85; 1.10, 1.04-1.16; 1.13, 1.03-1.24; 0.88, 0.81-0.96; 0.00, 0.00-0.64; and 0.17, 0.05-0.66, respectively). A nomogram incorporating these six variables demonstrated an area under the curve of 0.966 in the receiver operating characteristic curve. Bootstrap self-sampling internal validation showed a strong correlation between the predicted and actual results of the nomogram (C-index 0.966). The decision curve analysis also indicated satisfactory model performance.

The admission GCS, age, white blood cell count, total protein, pH, and cholinesterase change after the first hemoperfusion could predict in-hospital mortality in AOPP patients undergoing hemoperfusion. A nomogram model based on these six variables could provide accurate mortality prediction in these patients.