ИСТИНА

Background/ Introduction The cardiovascular death rate significantly declined last decades. But premature atherosclerosis burden remains an unresolved problem. Purpose The study aimed to assess the association of clinical factors with types of premature atherosclerosis onset. Methods Date of 702 patients (pts) (523 men and 179 women) with premature atherosclerosis (men≤55 (48.6±6.2), women≤60 (52.7±7.0) years of age) were analyzed with decision tree method using SPSS 23.0 program with the Python GUI module. Clinical and instrumental variables (n=109) were used. The test sample was formed by the cross-validation method. Results Myocardial infarction at the onset of atherosclerosis (n=542, 77,2%) was associated with the presence of peripheral atherosclerosis (1st order node, p <0.0001, F = 93,174). The 2nd order node was a uric acid level (p <0.0001, F = 26.493) in pts without peripheral atherosclerosis. In pts with uric acid level, less than 225 mmol/L-the left ventricle posterior wall thickness more than 10 mm was a 3d order node (p <0.0001, F = 30.143). Area under the ROC-curve 0.916, p = 0.011. Multivessel lesion according to coronary angiography data (102 patients) was associated with family history of cardiovascular disease (p = 0.001, F = 13.238), the area under the ROC-curve was 0.667, p = 0.041. For pts. with peripheral atherosclerosis (n=66, 9,4%) the aortic root diameter obtained by an echo was the 1st order node (p <0.0001, F = 36.057). In pts with aortic root diameter over 27 mm, a 2nd order node was creatinine level above 90 mmol/L (p = 0.036, F = 9.945) and in pts with a smaller diameter of aortic root was the history of hypertension emergency (p = 0.001, F = 13.897). Area under the ROC-curve 0.676, p = 0.02. For pts. with ischemic stroke (n=26, 3,7%) as atherosclerosis onset 1st order node was brachiocephalic atherosclerosis lesion (p <0.0001, F = 30.259). Among them, untarget BP level was 2nd order node (p = 0.033, F = 4.958). For pts without atherosclerosis age over 46.7 yrs was a 2nd order node (p <0.0001, F = 24.515), and for pts younger than 46.7 yrs admission glucose higher than 11.06 mmol/L was a 3rd order node (p = 0.026, F = 12.382). This model had high predictive accuracy (area under the ROC-curve 0.963, p <0.0001). Conclusion(s) Thus multiple clinical variants of premature atherosclerotic cardiovascular disease onset appeal to develop an individualized approach to early diagnosis and managemen