Nailfold capillaroscopy in diabetic retinopathy: microvascular insights, clinical value, and translational challenges - a narrative review.

摘要

Diabetic retinopathy (DR) is a leading microvascular complication of diabetes mellitus and a major cause of preventable vision loss. DR is now viewed as a manifestation of systemic diabetic microangiopathy rather than an isolated ocular disorder. Retina-centered assessment may not fully capture systemic microvascular burden, particularly where access to specialized retinal imaging is constrained. Nailfold capillaroscopy (NFC) enables direct, non-invasive visualization of peripheral capillary morphology and has been investigated as a marker of systemic microvascular abnormalities in diabetes. This narrative review examines shared mechanisms of retinal and peripheral microvascular injury and summarizes evidence linking NFC abnormalities to DR presence and severity. The available cross-sectional evidence was used to develop an exploratory three-tiered, pattern-based framework. Background features include capillary tortuosity, mild dilation, and microhemorrhages, which have been associated with prevalent DR but show limited consistency in discriminating current severity. Auxiliary features include neoangiogenic patterns, megacapillaries, and crossing capillaries, while the proposed core pattern comprises reduced capillary density and avascular areas. Among these parameters, avascular areas showed the most consistent pooled association with current DR severity. Selected single-cohort analyses reported AUC values above 0.80 for cross-sectional discrimination of prevalent DR or PDR; however, these estimates have not been prospectively or externally validated. In one cross-sectional cohort, a capillary density cutoff of <7.5 capillaries/mm yielded 95% sensitivity but only 65% specificity for prevalent DR, while a multi-parameter composite score showed an AUC of 0.745 in another cohort. The cumulative layering model and three-tiered pattern-based framework synthesize these findings for exploratory analysis of cross-sectional NFC-DR associations and discrimination of DR severity. However, the available evidence is predominantly cross-sectional and heterogeneous, with variability in imaging equipment, morphological definitions, sample sizes, and confounder adjustment. The framework is therefore hypothesis-generating and is not suitable at present for guiding routine clinical management or ophthalmology referral. Its longitudinal relevance requires prospective validation. NFC remains investigational. Further evaluation requires standardized imaging protocols, automated quantitative analysis, and large prospective studies.