The aim of study: clinical validation of a model for early detection of patients at high risk of developing complicated forms of alcohol withdrawal syndrome (hereinafter - AWS), based on laboratory diagnostics.
 Subjects and methods: The study was a prospective, cohort, observational. The total number of included patients was 200, 9 of them were excluded, and the included patients were distributed as follows: uncomplicated AWS developed in 98 patients (51.3%), alcohol delirium in 67 patients (35.1%), and alcohol-induced psychotic disorder in 26 patients (13.6%). Potassium, sodium, calcium and platelet count levels were compared between the groups, and the relationships between gender, age and potassium levels in patients with alcohol delirium were studied.
 Results: levels of potassium, sodium and platelet count were significantly different in groups of patients with alcohol delirium, uncomplicated AWS and alcohol-induced psychotic disorder. Blood calcium levels were significantly different between patients with alcohol delirium and uncomplicated AWS. Women were found to be less sensitive to metabolic disorders that developed as a result of alcohol consumption, as evidenced by their lower incidence of delirium. The high prevalence of alcohol delirium among elderly patients is a consequence of decompensation of metabolic regulation mechanisms in them. Predictive model of alcohol delirium, based on the results of blood potassium level study, proved effectiveness in clinical application.