A primary aim of the study was to investigate whether retinal microvascular findings in patients with suspected TIA helped to predict their risk of stroke or CVE within 90 days.

We undertook a cohort study of adult patients with a NIH Stroke Scale ≤ 3 admitted to three emergency departments’ (EDs’) observation units for an accelerated diagnostic protocol for suspected TIA or stroke. Clinical and imaging results were recorded prospectively. Non-mydriatic fundus photographs were obtained as part of the accelerated diagnostic protocol and reviewed for retinal microvascular findings, defined as retinal hemorrhages, cotton wool spots, retinal emboli or occlusions, hard exudates, or microaneurysms. Stroke and CVE were adjudicated by a stroke neurologist.

Among the 395 patients enrolled (median age: 57 years (interquartile range [IQR]: 50–66); 219 (55%) women; 253 (64%) black; 34 (9%) retinal microvascular findings present), 154 (39%) completed 90-day follow-up visits.  Within 90 days, 3/395 (0.8%) had a stroke, 3 additional subjects had a recurrent TIA, and none had a CVE.  All patients with a stroke or recurrent TIA had an ABCD² score >3 vs. 70% (273/295; p=0.25) of other patients, and none had any microvascular findings vs. 9% (34/392; p=0.59) of other patients.

Although we have found that retinal microvascular findings among patients with suspected TIA in the ED are an independent factor differentiating TIA and stroke from mimics, our study is unable to provide evidence that they improve short-term predictions of future ischemic events.  However, our study was underpowered due to a much lower occurrence of short-term stroke than expected at the time of its design, likely due in large part to improved secondary prevention, in particular, from the type of accelerated diagnostic protocol our patients underwent.