Discussion

CRAO is a medical emergency, and the standard of care for this condition is evolving. The Canadian experience with patient presentations and practice patterns for CRAO has not been extensively described in the literature to date. In our 4-year study of CRAO in a large Canadian academic health sciences center, we found that patients seek care urgently following vision loss from CRAO and that generally patients receive timely workup; however, these patients experience poor visual outcomes and commonly have high-risk underlying conditions. Opportunities to strengthen care for this population exist.

Patients generally seek care urgently after CRAO: in our series, more than 1 in 4 first contacted the medical system within 4.5 hours, and nearly 2 out of 3 presented within 24 hours. Similar times to presentation have been reported in American and European centers. ^{Reference Hoyer, Kahlert, Güney, Schlichtenbrede, Platten and Szabo8,Reference Lee, Tschoe and Coffman11,Reference Shah, Gilbert and Melles12} In Smith et al.’s series of patients with CRAO in Edmonton, Canada, only 3 of 27 (11.1%) presented within 4.5 hours. ^{Reference Smith, Benson, Tennant and Jivraj13} The most common site of initial presentation was an emergency department (ED) in more than half (61.8%) of cases, similar to other series, followed by optometry and ophthalmology practices at similar rates (14.5% and 13.2%, respectively). By contrast, American studies consistently show that the second most common initial point of contact was an ophthalmologist for nearly one-third of cases. ^{Reference Flowers, Chan, Meyer, Bruce, Newman and Biousse10–Reference Shah, Gilbert and Melles12} The comparatively lower rate of initial presentation to ophthalmology observed in our study presumably reflects the inability to self-refer to an ophthalmologist in Ontario. This is a reminder that non-ophthalmologists must be aware of this dangerous etiology of acute vision loss as they may be the first to encounter patients with CRAO.

Confirmation of diagnosis is a critical step that requires examination by a vision-care provider, typically an ophthalmologist. Among patients arriving at our tertiary care institution without a confirmed diagnosis of CRAO, the door-to-ophthalmological consultation time exceeded 12 hours in half of patients. This delay may reflect low suspicion for CRAO among providers who first encounter patients or delayed triage/assessment by ophthalmology after consultation has been requested. Delay to definitive ophthalmic diagnosis has been recognized as a major barrier to optimal care in other centers. In a recent US-based study, Shah et al. reported that only 172 (35.5%) of 484 patients with recent CRAO saw an ophthalmologist within 24 hours of vision loss. ^{Reference Shah, Gilbert and Melles12} Fundus photography and/or optical coherence tomography platforms can be used to accurately diagnose CRAO, and data can be reviewed remotely. ^{Reference Vasseneix, Bruce, Bidot, Newman and Biousse15,Reference Lema, De Leacy and Fara16} Implementation of a telemedicine pathway between the ED and ophthalmology that leverages ED-based ophthalmic imaging may be an optimal strategy to expedite the diagnosis (and treatment, if indicated) of CRAO.

The American Heart Association (AHA) Scientific Statement for CRAO emphasizes urgent etiologic workup including assessment for ipsilateral carotid stenosis. ^{Reference Mac Grory, Schrag and Biousse1} Although not explicitly written to address acute retinal ischemia, CSBPR advise brain and vascular imaging be completed as soon as possible following acute stroke or TIA, ideally in an ED for those presenting within 48 hours. ^{Reference Lindsay, Mountain, Gubitz, Dowlatshahi, Casaubon and Smith6} Nearly all patients with CRAO at our institution received acute brain and vascular imaging, and CT/CTA was the favored modality in most cases. By contrast, in the Edmonton series spanning 2018–2020, only 3 of 27 patients had any brain imaging. ^{Reference Smith, Benson, Tennant and Jivraj13} In a large American series of 2802 patients with retinal artery occlusion presenting to EDs between 2006 and 2014, brain and carotid imaging were performed in only 20.3% and 7.1%, respectively. ^{Reference Yousuf, Guiseppi, Katz, Nnorom and Akinyemi17} Compared to that study, a more recent series from an American center of 91 patients with CRAO between 2017 and 2020 reported increased rates of carotid imaging: CTA or MRA was obtained in 51.6% and carotid dopplers in 68.1%. ^{Reference Lee, Tschoe and Coffman11} The increasing rates of brain and vascular imaging likely reflect an evolving standard of care with similarities between CRAO and cerebral stroke being emphasized in recent years. ^{Reference Biousse, Nahab and Newman18}

We examined rates of GCA screening in patients with recent CRAO. The American Academy of Ophthalmology Preferred Practice Pattern for CRAO advises GCA be suspected in patients over age 50. ^{Reference Flaxel, Adelman and Bailey14} Although this is an uncommon cause of CRAO, recognition of GCA is critical given the high risk of fellow eye involvement and therefore of bilateral blindness in patients presenting with vision loss from GCA if treatment is delayed. ^{Reference Chen, Leavitt, Fang, Crowson, Matteson and Warrington19} In our series, GCA screening was frequent but not uniform: clinical and serological assessments were each omitted in approximately 1 in 10 patients. Although this is a low proportion, this is suboptimal given the relatively high regional incidence of GCA. ^{Reference Barra, Pope and Pequeno20} Indeed, the prevalence of GCA in our series of patients with CRAO (8.1% in those aged 50 and above) is higher than reported by other authors. ^{Reference Mac Grory, Schrag and Biousse1}

Recent clinical trials have failed to demonstrate the benefit of intravenous thrombolysis in minor nondisabling acute ischemic cerebral stroke over APT. ^{Reference Khatri, Kleindorfer and Devlin21,Reference Chen, Cui and Zhou22} In contrast, meta-analyses of relatively small observational studies have suggested that thrombolysis may improve rates of functional recovery in selected patients with non-arteritic CRAO. ^{Reference Huang, Wang and Zhang23,Reference MacGrory, Nackenoff and Poli24} At our center, no patient was offered thrombolytic therapy for CRAO. Smith and colleagues similarly reported no patient in their series of 27 in Edmonton, Canada, received thrombolysis. ^{Reference Smith, Benson, Tennant and Jivraj13} Relatively low rates of thrombolysis, despite the willingness to offer this therapy by providers, have been reported in American ^{Reference Chan, Flowers, Meyer, Bruce, Newman and Biousse9} and European ^{Reference Hoyer, Kahlert, Güney, Schlichtenbrede, Platten and Szabo8} centers and reflect the difficulty in operationalizing this treatment for these patients in the hyperacute period after vision loss. A pathway for stroke team activation for suspected acute CRAO does not exist in our center; such pathways have been proposed and implemented in other centers as a strategy to optimize access to thrombolysis for acute CRAO. ^{Reference Mac Grory, Schrag and Biousse1}

We examined the use of antiplatelet agents after CRAO. CSBPR advise DAPT, including loading doses of both agents, for patients with acute high-risk TIA or minor ischemic stroke who are not at high bleeding risk. ^{Reference Lindsay, Mountain, Gubitz, Dowlatshahi, Casaubon and Smith25} The AHA statement suggests a similar practice is reasonable for CRAO. ^{Reference Mac Grory, Schrag and Biousse1} At our center, we observed significant variability in the decision to add APT and to provide a loading dose: among patients who were eligible for APT initiation or escalation, approximately one-third did not receive any antiplatelet agent, and more than half did not receive a loading dose. In Lee et al.’s series of 91 patients with acute CRAO, inconsistent use of APT was observed, with 27.5% of patients not starting any antiplatelet medication. ^{Reference Lee, Tschoe and Coffman11} In another US study of 109 patients with CRAO between 2010 and 2020 referred to a tertiary hospital, APT was started by external providers in fewer than 10% of cases. ^{Reference Flowers, Chan, Meyer, Bruce, Newman and Biousse10} The heterogeneity in practice patterns in our center and others highlights the importance of early involvement after CRAO of stroke experts to guide the nuanced decision around antiplatelet use.

We observed a relatively high rate of concordance between practice patterns at our institution with guidelines that emphasize the importance of managing vascular risk factors after CRAO: ^{Reference Mac Grory, Schrag and Biousse1} neurovascular secondary prevention referrals were omitted for approximately 10% of patients with non-arteritic CRAO. This is a low proportion overall; however, it is important to note that those who experience non-arteritic CRAO are at increased long-term risk of stroke, myocardial infarction and death. ^{Reference Wai, Knapp and Ludwig4} Presumably, some of this risk can be mitigated with appropriate workup and optimization after CRAO. We observed a higher rate of non-referral for ocular secondary prevention of closer to 20%. After CRAO, regardless of etiology, ocular follow-up serves important functions, which include optimizing residual vision, monitoring for neovascular-related complications and preserving the health of the contralateral eye. ^{Reference Mac Grory, Schrag and Biousse1} Other institutional series have not examined rates of secondary prevention referrals to our knowledge.

With respect to the results of etiologic workup, underlying conditions included GCA in 8.1% of patients over 50, and among patients with non-arteritic CRAO, carotid stenosis was found in 25.7% and cardioembolic etiology in 10.5%. This reinforces the urgency of prompt diagnosis and workup after CRAO as these conditions are amenable to time-sensitive intervention to reduce the risk of additional morbidity.

Rates of functional visual recovery in our series (10.5%) are comparable to what has been reported in meta-analysis and demonstrate the poor natural history of vision loss from CRAO. ^{Reference Schrag, Youn, Schindler, Kirshner and Greer26} Fortunately, we observed infrequent additional neurovascular events in our patients. Although follow-up data were not uniformly available for all patients, we interpret this as evidence of relatively effective secondary stroke prevention in the current system.

Our study is the largest Canadian series to date reporting institutional experience with CRAO; and, considering an evolving body of literature regarding the optimal management of this condition, is timely. The relatively high number of patients (for a rare condition) included over a 4-year period is also a strength as this allows us to draw meaningful conclusions. There are limitations. First, as a single-center study, there are inherent issues with generalizability. Nevertheless, The Ottawa Hospital is among the largest academic health sciences centers in Canada and serves a diverse population exceeding 1.4 million; as such, we believe our conclusions may have applicability in other Canadian centers. Supporting this, our findings overall do not deviate significantly to those of the only other Canadian series on the topic, notwithstanding the issues discussed above. ^{Reference Smith, Benson, Tennant and Jivraj13} Second, the retrospective nature and reliance on electronic records have obligatory limitations including missing data and potential for error with time stamps; however, this is not expected to significantly alter our conclusions. Third, while we used a theoretically robust case-finding strategy that relied on pre-specified diagnosis codes that are entered in multiple areas in the chart, it is possible that some cases of CRAO during the study window were not identified. Importantly, the inadvertent exclusion of a small number of cases would not alter our findings in a meaningful way. Finally, our study window was inclusive of the onset and peak of the Covid-19 pandemic. This may have caused more patients to present to the ED as opposed to community-based providers such as optometrists; nevertheless, the distribution of sites of presentation is comparable to previous studies.

Patients with CRAO generally present within a day of vision loss, and typically present to the ED. Expediting diagnosis of CRAO should be a priority and may require telemedicine strategies to decrease the delay in ophthalmological assessment. ^{Reference Vasseneix, Bruce, Bidot, Newman and Biousse15,Reference Lema, De Leacy and Fara16} At our institution, there are generally favorable practice patterns with respect to acute investigations and secondary prevention referrals after CRAO; these should continue to be emphasized in guidelines for CRAO, including the importance of GCA screening given Canada’s high GCA prevalence. ^{Reference Barra, Pope and Pequeno20} Despite growing evidence for the use of thrombolytic therapy in CRAO, ^{Reference Mac Grory, Schrag and Biousse1} our hospital system, one of the largest in Canada, has not yet adopted this within its institutional scope. The role of and strategies for implementation of thrombolysis in the management of CRAO in the Canadian context should be clarified. Patients with CRAO commonly have a high-risk underlying pathology that requires specific treatment to reduce additional morbidity and experience low rates of visual recovery. There is an unmet need for multidisciplinary Canadian guidelines for the management of this condition, especially in the acute care setting.