Abstract Background Heart failure with preserved ejection fraction (HFpEF) represents nowadays the most frequent presentation of heart failure (HF) in patients aged > 65 years. The underlying reason for the failure of the vast majority of randomized controlled trials (RCTs) in HFpEF patients has been identified in the heterogeneity of pathophysiology and clinical phenotype of this clinical syndrome. Coronary microvascular dysfunction (CMD) represents one among various pathophysiological mechanisms, together with pulmonary vascular disease, pericardial restraint, impaired chronotropic reserve and abnormal autonomic tone. Whether CMD-HFpEF endotype differs from the others in terms of echocardiographic parameters and clinical outcomes is still a matter of debate. Therefore, a systematic review and meta-analysis were performed in order to compare HFpEF populations with or without CMD in terms of echocardiographic features and clinical outcomes. Methods We searched for articles published in PubMed, Scopus and Wiley comparing HFpEF population with or without CMD up to 1st September 2022. Observational studies, reporting echocardiographic parameters mentioned in HFA-PEFF score and/or clinical time to event data, were included. E/e’ ratio, left atrial volume index (LAVi) and left ventricular mass index (LVMi) constituted our three parameters of choice and Hedge's g was the summary effect size. The composite of HF hospitalization and all-cause death represented our clinical endpoint. Meta-regressions according to follow-up time were performed in order to explore potential heterogeneity sources across studies. Results We identified 9 prospective observational studies, enrolling 797 patients with HFpEF. On pooled analysis, patients with CMD present a more severe echocardiographic phenotype, determined by a higher LAVi effect size (ES) 0.40; Confidence Interval (CI) 0.11, 1.69, E/e’ ratio (ES 0.65; CI 0.28, 1.02) and LVMi (ES 0.27; CI 0.01, 0.53) compared to no-CMD patients. Furthermore, CMD patients showed a significant higher rate of the composite endpoint of all-cause-death and hospitalization for HF (HR 3.22, CI 1.2-8.5, p 0.02). At meta-regression, a significant correlation was found between logarithmic hazard ratios and follow-up time reported (z=2.03, p value 0.04), suggesting that long-term follow-up of CMD-HFpEF patients is required to track the natural trajectory of the disease. Conclusions Aside from being considered a pathophysiological hallmark in the development of HFpEF, CMD seems to play an aggravating role in the progression of the disease, leading both to more severe echocardiographic parameters and worse clinical outcomes compared to other endotypes. Thus, different echocardiographic thresholds could allow for a better prognostic stratification and for identifying the subset of patients who deserve a CMD assessment and who are eventually more likely to benefit from drugs targeting CMD.