Cluster Analysis as a Method of Assessing the Severity of Non-Alcoholic Fatty Liver Disease in Patients with Hypertension

Abstract

The purpose of the study was to determine the possibilities of cluster analysis as a method for assessing the severity of Non-alcoholic fatty liver disease in its comorbidity with hypertension. Materials and methods. In the study, we examined 63 patients with non-alcoholic steatohepatitis and hypertension, 62 patients with isolated non-alcoholic steatohepatitis and 20 healthy individuals as a control group. All patients underwent anthropometric examinations, including determination of waist circumference, measuring systolic and diastolic blood pressure, performing routine laboratory tests, revealing plasma levels of kallistatin, interleukin-1β and interleukin-10 by enzyme-linked immunosorbent assay. Abdominal ultrasonography was performed to determine the thickness of the right, left and caudal lobes of the liver, and 2-dimensional shear wave elastography with calculating the liver parenchyma stiffness (E, kPa) was performed to determine the fibrotic changes of the liver. Cluster analysis was performed in the program "STATISTICA 8.0." using a hierarchical method and k-means algorithm with pre-standardized data. Results and discussion. According to the results of clustering, 3 subgroups of patients with non-alcoholic fatty liver disease and hypertension were identified. Cluster 1 was characterized by low systolic and diastolic blood pressure, waist circumference, low levels of alanine aminotransferase and aspartate aminotransferase, moderate interleukin-1β levels, low activity of kallistatin and interleukin-10, ultrasound signs of hepatomegaly and increased liver parenchyma stiffness. Cluster 2 included patients with the highest systolic and diastolic blood pressure, moderate waist circumference, highest alanine aminotransferase, aspartate aminotransferase and interleukin-1β levels, low activity of kallistatin and interleukin-10, the most pronounced ultrasound hepatomegaly signs and the worst liver elastography results. Cluster 3 was characterized by moderate systolic and diastolic blood, minimum alanine aminotransferase, aspartate aminotransferase and interleukin-1β content, highest kallistatin and interleukin-10 levels, minimum thickness of liver lobes and lowest liver parenchymal stiffness. Clusters differed significantly from the control group in all parameters, and from the isolated non-alcoholic fatty liver disease group, significant differences were found in all characteristics except waist circumference and interleukin-1β levels. The clusters varied in at least 4 parameters, and the maximum dissimilarity was found between clusters 2 and 3, which differed by 9 indicators. Conclusion. Thus, the clustering of examination results of the patients with non-alcoholic fatty liver disease and hypertension showed a statistically significant distribution of groups according to the severity of liver fibrosis. The elevated systolic and diastolic blood pressure, increased alanine aminotransferase and aspartate aminotransferase, decreased activity of kallistatin and interleukin-10 and signs of hepatomegaly according to ultrasound data could be considered as markers of severe liver fibrosis. The cluster distribution of patients with non-alcoholic fatty liver disease and hypertension opens prospects for the development of new integrated methods in assessment of liver parenchyma changes severity in patients with this comorbidity