Introduction. Heart failure with preserved left ventricular ejection fraction (HFPEF) is an epidemiologically serious disease. Noninvasive diagnosis of HFPEF remains challenging. The current diagnosis is based on evidence of diastolic dysfunction, albeit systolic dysfunction is also present but not included in the diagnostic algorithm. The aim of this study was to analyze the longitudinal (long), circumferential (circ) and radial (rad) component of systolic left ventricular (LV) function in patients with exertional dyspnea of unexplained etiology and normal left ventricular ejection fraction (LVEF). Methods. One hundred and twenty-two patients with exertional dyspnea of unexplained etiology and normal LVEF and 21 healthy controls, underwent echocardiography examination at rest and at the end of symptom-limited exercise. We analysed the longitudinal, circumferential and radial deformation of myocardium using two dimensional speckle tracking echocardiography in all subjects. Results. Patients with exertional dyspnea and preserved LVEF were divided into group A1 (46 patients meeting the criteria for the diagnosis HFPEF) and group A2 (76 patients without HFPEF). Group A1 had significantly worse longitudinal and circumferential systolic LV function than group A2. Subjects in group A1 compared to group A2 showed significantly different strain rates during atrial contraction (SR A), circ and ratio of peak early trans-mitral flow velocity (E) and strain rate E wave (E / SR E) circ. Multivariate logistic regression analysis showed that the SR parameter A circ is an independent predictor of HFPEF (odds ratio 0.550, 95% confidence interval: 0.370 -0.817, P value 0.003). Conclusion. Longitudinal and circumferential LV deformation was significantly more impaired in patients with HFPEF than in patients with exertional dyspnea without HFPEF. In patients with exertional dyspnea and normal LVEF, the value of SRA circ appears to be a significant and independent predictor of HFPEF. This parameter may, in the future complement the diagnostic algorithm for HFPEF.