A multimarker approach to the diagnosis of heart failure with preserved ejection fraction

Background. Against the background of a steady increase in the incidence of heart failure with preserved ejection fraction (HFpEF), the highly laborious diagnosis of the disease hinders the timely implementation of therapeutic and preventive measures, which determines the relevance of our study. Aim: To evaluate the significance of biomarkers, indices of myocardial strain and diastolic function in the diagnosis of HFpEF. Material and Methods. According to a diastolic stress test (DST), 88 patients (65.9±5.4 years; 47.7% men) with left ventricular ejection fraction ≥50% were divided into 2 groups: 1 (n = 49) – with HFpEF, 2 (n = 39) – without HFpEF. All patients underwent echocardiography (ECHOCG) at rest and DST, evaluation of left atrial reservoir strain (LASr), left atrial stiffness index (LASI), diastolic function index E/e΄; biomarkers of immune inflammation; neurohumoral (NT-proBNP) and sympathoadrenal systems; fibrosis; growth factors; insulin resistance were studied: triglycerides/glucose (TyG) index; renal function; anemia, iron (Fe) status with assessment of percentage of transferrin iron saturation (%TIS); cortisol; progesterone (PGN). Factors associated with HFpEF were identified using simple or multivariate logistic regression, and 2 mathematical models of the disease were created. ROC analysis was used to identify cut-off points for indicators and HFpEF models were estimated. Results. In group 1, women predominated (63.3 vs 38.5%; p = 0.021), mainly with functional class II according to the New York classification, with a higher body mass index (BMI) (32.4 ± 4.0 vs 29.9 ± 3.9 kg/m²; p = 0.003), the incidence of obesity (67.3 vs 43.6%; p = 0.025), iron deficiency (43.9 vs 34.6%; p = 0.033) and the use of β-blockers (73.5 vs 51.3%; p = 0.032). In multivariate analysis, HFpEF was associated with LASr: OR (95% CI) 0.722 (0.540–0.965) (p = 0.028) and E/e΄ at DST: OR (95% CI) 9.263 (1.912–44.885) (p = 0.006). Due to the unavailability of DST in practical healthcare, the following were included in 1 model without DST: gender, BMI - 30.4 kg/m², NT-proBNP - 360 pg/ml, LA volume index - 30 ml, E/e΄ - 9.8 at rest, LASr < 26%, TyG - 8.7 units, %TIS interaction < 27.7%, PGN < 1.7 nmol/l. The area under the ROC curve (AUC) was 0.879 (95% CI 0.806–0.951) and p < 0.001, specificity was 75.7%, sensitivity – 86.2% and the overall predictive value was 79.5%. Due to the secondary nature of the fibrotic process with the development of diastolic dysfunction, the following were included in model 2 without ECHOCG: gender, BMI - 30.4 kg/m², NT-proBNP - 360 pg/ml, cardiotrophin 1 - 587.0 pg/ml, TyG - 8.7 units, %TIS < 27.7%, PGN < 1.7 nmol/l. The AUC of the model was 0.830 (95% CI 0.734–0.927) and p < 0.001, specificity – 75.7%, sensitivity – 80.4%, overall predictive value – 78.3%. Conclusion. Multimarker analysis revealed the diagnostic significance in HFpEF of a decrease in PGN level less than 1.7 nmol/l due to age-related hormonal remodeling, an increase in NT-proBNP level more than 360 pg/ml with neurohumoral activation, a decrease in transferrin iron saturation percentage less than 27.7% as a marker of iron deficiency, an increase in TyG insulin resistance index more than 8.7 units associated with obesity, LASr less than 26%, LASI more than 0.38 units, E/e΄ ratio more than 9.8 units at rest and more than 12.1 units during DST. Using a panel of laboratory markers including the previously mentioned levels of sST2, NT-proBNP, %LVAT, and hsTnT allows for a more precise assessment of the disease or its high risk of development, even without performing resting echocardiography or DST.