Do retinal microvascular abnormalities shed light on the pathophysiology of lacunar stroke?

Background. Lacunar strokes account for 25% of all ischaemic stroke but the exact nature of the causative cerebral small vessel abnormality remains unknown. Pathological studies are technically difficult and brain imaging cannot adequately characterise the cerebral small vessels. The retinal blood vessels are of similar size and physiology to the cerebral small vessels and may act as a surrogate marker for these cerebral small vessels. We therefore investigated retinal microvascular abnormalities in lacunar stroke. Methods. We performed a systematic review of retinal microvascular abnormalities in lacunar stroke to clarify associations and identify where further research was required. We then established a cohort of patients presenting with lacunar stroke with cortical stroke controls to investigate differences in retinal microvascular abnormalities between stroke subtypes. All patients had MRI brain at presentation and digital retinal photography of both eyes. We investigated the prevalence of retinopathy (hard and soft exudates or haemorrhages/microaneurysms), focal arteriolar narrowing and arteriovenous nicking . We developed, validated and used novel semi-automated techniques for measuring retinal arteriolar and venular widths, retinal arteriolar geometry (branching co-efficients (change in arteriolar cross sectional area across a bifurcation) and branching angles) and fractal dimensions (reflecting branching complexity) of the vasculature. We also assessed MRI parameters in lacunar stroke. We used multivariable analysis to correct for baseline imbalances in vascular risk factors. Results. From the systematic review we demonstrated that retinal microvascular abnormalities are associated with incident and prevalent stroke but that in general, strokes were inadequately characterised and there were no data regarding retinal microvascular abnormalities in ischaemic stroke subtypes. We recruited 253 patients, 129 lacunar strokes and 124 cortical strokes, mean age 68 years. We found no difference in the prevalence of retinopathy, arteriovenous nicking, focal arteriolar narrowing or arteriolar widths between lacunar and cortical stroke subtypes. We found that venules were wider in lacunar stroke. We found no differences in arteriolar branching co-efficients or arteriolar branching angles between lacunar and cortical strokes but found that deep white matter white matter hyperintensities on MRI were associated with increased branching co-efficients and periventricular white matter hyperintensities associated with decreased branching co-efficients. We found that the fractal dimension of the vascular tree was decreased in lacunar stroke. Furthermore we found that enlarged perivascular spaces on MRI are associated with lacunar stroke and white matter disease. Conclusions. We have clearly demonstrated that retinal microvascular abnormalities differ between lacunar and cortical stroke suggesting that a distinct small vessel vasculopathy may cause lacunar stroke. We have also identified MR markers of lacunar stroke. These results suggest that venular disease (a hitherto underresearched area) may play a role in the pathophysiology of lacunar stroke. Retinal microvascular abnormalities can act as markers for cerebral small vessel disease. We plan collaborative analyses with colleagues who have performed similar studies to further assess retinal abnormalities in lacunar stroke.