Background <p>Systemic microvascular dysfunction is proposed as a key pathophysiological process in heart failure with preserved ejection fraction (HFpEF). This study compared microvasculature across vascular beds in HFpEF patients and controls.</p> Methods <p>This prospective, case-control study included subjects ≥ 60years. HFpEF patients were diagnosed in an expert centre. Controls without HF were selected from the Maastricht Study, a population cohort enriched with diabetes mellitus. Microvascular assessments included central retinal venular/arteriolar calibres (CRVE/CRAE), flicker-light-induced retinal dilation, skin microvascular flowmotion and heat-induced hyperemia, and urinary albumin-to-creatinine ratio (UACR). Group differences were evaluated with confounder-adjustments (age, sex, blood pressure, body mass index, diabetes, haemoglobin, smoking). Interactions with sex and diabetes mellitus were tested, and stratified analyses were performed when significant interactions were present.</p> Results <p>Microvascular assessments were performed in 138 HFpEF patients and 2140 controls. Microvascular differences were present between groups in all vascular beds. However, confounder-adjusted analyses attenuated differences. Confounder-adjusted analyses indicated that HFpEF patients versus controls still had retinal differences: narrower CRVE (− 8.1&#xa0;μm, <i>p</i> = 0.008) and narrower CRAE trend (− 3.5&#xa0;μm, <i>p</i> = 0.073), but similar flicker-light-induced retinal venular/arteriolar dilation (− 0.23%, <i>p</i> = 0.392; − 0.18%, <i>p</i> = 0.593, respectively). Confounder-adjusted analyses showed similar skin flowmotion measures (i.e. endothelial power − 0.09log(PU²), <i>p</i> = 0.181), and heat-induced hyperemia (0.02log(%), <i>p</i> = 0.605) between groups. UACR remained higher in HFpEF after confounder adjustments (0.56log(g/mol), <i>p</i> = &lt; 0.001). Interaction analyses revealed that female patients had narrower CRVE versus controls (p_int=0.023; females − 13.8&#xa0;μm, <i>p</i> &lt; 0.001; males 1.2&#xa0;μm, <i>p</i> = 0.812). Patients had lower skin endothelial flowmotion power only when diabetes was co-occurring (p_int=0.048; − 0.36 log(PU² ), <i>p</i> = 0.014). UACR was higher in male and female patients versus controls, but was more pronounced in males (p_int=0.002).</p> Conclusions <p>HFpEF patients showed microvascular differences versus controls across all vascular beds studied. However, confounder-adjusted differences remained significant in eyes and kidneys. The findings across multiple organs support that MVD is likely a more systemic process than only local MVD in HFpEF, and possible sex-specific underlying pathophysiology.</p> Registration <p>URL: <a href="https://onderzoekmetmensen.nl">https://onderzoekmetmensen.nl</a>; Unique identifier: NL7655.</p> Graphical abstract <p></p>