MATHEMATICAL MODEL FOR PREDICTING THE SEVERITY OF THE COURSE OF ACUTE PANCREATITIS

Abstract

Background. A feature of acute pancreatitis is the high risk of developing complications (occurring in 50% of patients), the mortality rate of which reaches 15%, and with a severe course varies within 40-70%. It has been proven that timely determination of the severity of the disease, selection of appropriate treatment tactics, early diagnosis of complications and their prevention significantly improve treatment results. Aim: to develop a modern mathematical model for predicting the severity of the course of acute pancreatitis, taking into account indicators of calcium-phosphorus metabolism, hemocoagulation factors and serological examination data for Helicobacter pylori. Materials and methods. The study was based on the results of an examination of 280 patients with acute pancreatitis, who were divided into two groups: the main group (n=187) – patients with a severe course and a comparison group (n=93) – patients with a mild and moderate course of the disease. To develop a mathematical model, the following indicators were analyzed and compared: duration of the disease before hospitalization, body mass index, number of leukocytes, C-reactive protein, blood glucose, procalcitonin, interleukin-6, immunoglobulin M to Helicobacter pylori, thrombin-antithrombin III complex, activity of tissue plasminogen activator, serum calcium, albumin corrected calcium, vitamin D. Results. Based on the obtained results, we developed a mathematical model for predicting the severity of the course of acute pancreatitis and revealed a correlation between the calculated scores (according to the mathematical model) and the APACHE II scale (severe course of 8 points and more). The step-by-step creation of a model by the method of multiple regression analysis with a gradual decrease in the number of indicators from 12 to 8 and to 6 allowed us to propose a mathematical model that has high accuracy for predicting the severe course of acute pancreatitis (R=0.82; R2=0.66; p< 0.0001). The obtained data demonstrate the dependence of the "severe course" factor on the content of vitamin D, immunoglobulin M to Helicobacter pylori and the activity of tissue plasminogen activator and substantiate the need for their early determination in patients with acute pancreatitis. Therefore, the developed mathematical model is highly informative and can be used in medical practice for early prediction of the severe course of acute pancreatitis. Conclusions. The dependence of the «severe course» factor in patients with acute pancreatitis on the content of vitamin D, immunoglobulin M to Helicobacter pylori and the activity of tissue plasminogen activator has been proven, and the need for their determination in the early period of the disease is substantiated. Using the method of multiple regression analysis, a mathematical model was developed that has high accuracy for predicting the severe course of acute pancreatitis (R=0.82; R2=0.66; p<0.0001).