Heart-eye-fingertip axis in ischaemia with non-obstructive coronary arteries: retinal neuroaxonal thinning and nailfold capillary rarefaction in coronary microvascular dysfunction

Background Coronary microvascular dysfunction (CMD) is increasingly recognised as an important component of coronary artery disease. Ischaemia with non-obstructive coronary arteries (INOCA) due to CMD is increasingly viewed as a systemic microangiopathy. However, whether coronary microvascular impairment aligns with concordant abnormalities in extra-coronary microvascular beds remains incompletely defined. Purpose To evaluate whether CMD in INOCA is associated with (i) reduced nailfold capillary density and (ii) retinal neuroaxonal thinning assessed by optical coherence tomography (OCT), compared with individuals without CMD. Methods In this prospective study, 82 participants underwent invasive coronary function testing by bolus thermodilution technique to measure coronary flow reserve (CFR) and index of microvascular resistance (IMR). CMD was defined as CFR <2.5 and/or IMR ≥25, with classification into structural and functional CMD endotypes. Nailfold videocapillaroscopy quantified capillary density (capillaries/mm) in the full cohort. Comprehensive ophthalmic screening and retinal OCT (optic nerve head and macula) were performed in a predefined subgroup with healthy eyes (87 eyes from 44 CMD participants; 47 eyes from 26 controls). Peripapillary retinal nerve fibre layer (pRNFL) and macular ganglion cell complex (GCC) thickness were extracted using standard segmentation. All participants underwent 24-hour ambulatory blood pressure monitoring (ABPM). Mediation analysis explored ABPM-derived pulse pressure index (PPi) metrics in the CMD–pRNFL relationship. Results Participants with CMD were younger than controls (mean age 51.9 vs 55.7 years) and more frequently women (69% vs 44%). Nailfold capillary density was lower in CMD versus controls (9.3 ± 2.9 vs 11.0 ± 1.8 capillaries/mm; P=0.04) and remained significant after multiplicity adjustment (adjusted P<0.05). In the OCT subgroup, inferior pRNFL was thinner in CMD than controls (median 124 [101–145] vs 134 [109–166] μm; P=0.009). The day–night PPi difference significantly mediated the association between CMD and inferior pRNFL thickness (P=0.002). Higher PPi and a greater day–night PPi difference independently predicted thinner total, inferior, and superior pRNFL and correlated with thinner GCC (all P<0.05). Conclusion INOCA attributable to CMD demonstrates a multisystem microvascular phenotype, characterised by peripheral capillary rarefaction and retinal neuroaxonal thinning alongside coronary microvascular impairment. Nailfold videocapillaroscopy and retinal optical coherence tomography may provide accessible, non-invasive markers of coronary microvascular function, supporting earlier identification and refined risk stratification in patients with angina pectoris and non-obstructive epicardial coronary artery disease.