Mathematical Modelling in the Differential Diagnosis of Anemia of Chronic Diseases and Iron Deficiency Anemia in Patients With Inflammatory Joint Diseases

Abstract

Aim. To develop based on the indicators of clinical blood tests, some biochemical markers, iron metabolism and cytokines a mathematical model that allows with high sensitivity and specificity to conduct differential diagnosis of anemia of chronic diseases (AHD) and iron deficiency anemia (IDA) in patients with inflammatory joint diseases (rheumatoid arthritis (RA), psoriatic arthritis (PSA), ankylosing spondylitis (AS)). Material and methods. The study included 104 patients with inflammatory joint diseases and anemic syndrome (37 males/67 females, age 48.4±5.42 years old), of which 54 patients with RA, 27 patients with PSA and 23 patients with AS. The control group consisted of 22 patients with inflammatory diseases of the joints without anemia (13 males/9 females, age 47.8±3.55 years old), of which 7 patients with RA, 10 with PSA, 5 with AS. For all patients, in addition to hemogram parameters, the concentrations of interleukin-6, interleukin-10, interleukin-1β, interferon-gamma, tumor necrosis factor alpha, ferritin, C-reactive protein, hepcidin, transferrin, soluble transferrin receptor were determined. For quantitative indicators, the median, standard error of the mean, and interquartile range was calculated. The significance of differences between several unrelated groups was determined using the Kruskal-Wallis test. The calculation of a mathematical model for the differential diagnosis of ACD and IDA was carried out using discriminant analysis. Results. As a result of the study, a canonical linear discriminant function (CLDF) was obtained: CLDF = 1.612171-0.002725× Hepcidin -0.005429× Ferritin The coordinates of the centroids are calculated, for ACD it is -1.44222008, and for IDA it is 1.52705656. A patient whose CLDF value is determined based on ferritin and hepcidin concentrations should be classified into the ACD or IDA group based on the minimum distance to the corresponding centroid. The resulting mathematical model has 100% sensitivity and 80% specificity. Conclusions. The equation based on hepcidin and ferritin (obtained from the results of discriminant analysis) allows, with high sensitivity, specificity and information ability, to carry out differential diagnosis of ACD and IDA in patients with inflammatory joints diseases instantly at the stage of initial contact with the doctor. Using of the equation will make it possible to more effectively diagnose and correct these types of anemia in this pathology, including for the purpose of more effective treatment.