Highlights
- •
Doubling in the incidence of ED presentations for CRAO between 2016 and 2021.
- •
A linear increase in utilization for stroke workup imaging modalities in the ED.
- •
Despite AAO/AHA guidelines, >50% of CRAO patients received no stroke workup imaging.
Purpose
To investigate trends in acute central retinal artery occlusion (CRAO) diagnostic assessments and sociodemographic characteristics of patients in the United States (US) emergency department (ED).
Design
Retrospective trend study.
Subjects
Adult patients with CRAO presenting to ED from 2016 through 2021.
Methods
The National Emergency Department Sample was queried with International Classification of Diseases, Tenth Revision (ICD-10) codes with a primary diagnosis of CRAO. Stroke workup modalities assessed included brain imaging (CT/computed tomography angiography and magnetic resonance imaging/magnetic resonance angiography), carotid imaging (US, computed tomography angiography, magnetic resonance angiography), cardiac diagnostics (echocardiogram/electrocardiogram), and laboratory workup (erythrocyte sedimentation rate/C-reactive protein).
Main Outcome Measure
Proportion of studied participants receiving stroke workup modalities.
Results
A total of 3736 patients were identified with mean age of 69 [13.72] years, majority were male (53.02%), and predominantly insured by Medicare (62.77%). Comorbidities included hypertension (75.45%), hyperlipidemia (44.67%), coronary artery disease (20.26%), diabetes (27.38%), and obesity (11%). CRAO incidence significantly increased over the 5-year period from 1698 to 3526 ( P = .024). Utilization rates of all workup modalities including brain imaging, carotid imaging, heart, and laboratory tests showed a linear increase from 2016 through 2021. The proportion of patients receiving no stroke workup decreased from 66.75% to 57.87% ( P = .259).
Conclusions
The US has increased the screening and stroke workup of CRAO in the ED over time, yet greater than 50% of patients continue to not receive any type of stroke workup imaging. Greater awareness of these trends and current guidelines could result in improved screening and patient outcomes.
INTRODUCTION
C entral retinal artery occlusion (CRAO) is a medical emergency that presents as an acute painless vision loss and can be a harbinger of an impending stroke. A prompt stroke workup upon diagnosis and admission to the emergency department (ED) is crucial since the 48 hours following CRAO present the highest risk of stroke; as high as 30% rate of acute ischemic stroke on magnetic resonance imaging (MRI) within the first 7 days of diagnosis. Treatment strategies to reverse acute vision loss have included ocular massage, decreasing intraocular pressure, and anterior chamber paracentesis. These methods all aim to dislodge emboli and restore blood flow, although many of these maneuvers have shown mixed results. TPA has been increasingly utilized over recent years and has shown some success. As such, early detection and management of CRAO is critical to reduce morbidity and mortality in these patients.
In 2019, the American Academy of Ophthalmology (AAO) released the latest Preferred Practiced Pattern (PPP) of acute symptomatic ophthalmic artery occlusion or CRAO that recommends an immediate referral to the nearest stroke center for prompt assessment for consideration of an acute intervention. The recommendations included a systemic evaluation for carotid occlusive and thromboembolic disease as well as suspecting giant cell arteritis (GCA) in patients over 50 years of age. The American Heart Association (AHA) also released a statement in 2020 on the management of acute CRAO, recognizing it as a medical emergency, with recommendations for systems to prioritize early detection and triage to emergency medical attention with a suggested treatment protocol that includes a stroke workup.
This study aims to explore national trends in CRAO presentations to the ED, including the incidence, patient sociodemographics, and medical comorbidities. Furthermore, this study will quantify and analyze a variety of stroke workup imaging modalities and laboratory testing received by patients in the ED.
METHODS
DATA SOURCE
A population-level, retrospective cross-sectional study was performed utilizing deidentified data from the National Emergency Department Sample (NEDS) from the start of 2016 through 2021. The NEDS is the largest all-payer ED database in the United States (US), composed of discharge records data from 993 hospitals and 40 states. The database represents a 20% stratified sample of hospital-owned EDs, encompassing over 100 million visits. National estimates in NEDS are calculated using weights assigned to each record which are adjusted for the sample’s stratified design, hospital characteristics, and nonresponse. The national estimates are obtained by multiplying the sample data by these weights. This study was waived from Institutional Review Board and patient consent requirements from the George Washington University due to the use of deidentified NEDS patient data.
STUDY POPULATION
Patients with a primary diagnosis of CRAO (H34.1) were identified in the NEDS using the International Statistical Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10). Patients with a nonprimary or secondary diagnosis of CRAO were excluded from the study.
DEFINITIONS OF VARIABLES
Current Procedural Terminology codes were used to identify imaging and laboratory tests received by patients. The studied variables of patients included descriptive statistics analyzed for age, sex, primary payer status, household income, and medical comorbidities. Age was stratified into the following cohorts: Less than 25, 25 to 35, 35 to 45, 45 to 55, 55 to 65, 65 to 75, and greater than 85 years. Patient household income was determined based on a quartile classification of the estimated median household income of patients in their residential zip codes. Common patient comorbidities were identified using the Elixhauser comorbidity index and ICD-10 codes.
Stroke workup was categorized as the codes for brain imaging, carotid imaging, cardiac tests, and lab tests. Brain imaging composed of computed tomography (CT) and CT angiography (CTA) of head (70450, 70460, 70470, 70496), MRI and magnetic resonance angiography (MRA) of the head (7055X, 70544, 70545, 70546). Carotid imaging composed of ultrasound imaging (9388X), carotid CTA (70498), and carotid MRA (70547, 70548, 70549). Cardiac tests composed of electrocardiogram (ECG) (93000, 93005, 93010) and transthoracic echocardiography (Echo) (9330X). Lab tests were composed of erythrocyte sedimentation rate (ESR) (85652, 85651) and C-reactive protein (CRP) (86140, 86141). Lab tests (ESR/CRP) were only calculated in patients greater than 50 years of age to screen for GCA as recommended by the AAO PPP guidelines. Patients were assessed for stroke workup by calculating the frequency of imaging and lab modalities. The annual incidence of CRAO presentations to the ED was calculated using NEDS-derived national estimates values and US census bureau population estimates data to calculate an incidence per 100,000 population value. The Mann-Kendall Test for monotonic trend was utilized to calculate significant trend changes in CRAO cases, stroke imaging modalities, and laboratory testing over the specified time frames.
The study was conducted in accordance with the tenets of the Declaration of Helsinki and followed the Strengthening the Reporting of Observation Studies in Epidemiology reporting guidelines. The collection and evaluation of all protected patient health information was performed in a Health Insurance Portability and Accountability Act-compliant manner.
RESULTS
A total of 3736 patients with a primary diagnosis of CRAO were identified over the analyzed period (mean [standard deviation] age, 69 [13.72] years) ( Table 1 ). This cohort was composed of a slight majority of male patients (53.02%) and predominantly insured by Medicare (62.77%). Comorbidities in this cohort included hypertension (75.45%), hyperlipidemia (44.67%), diabetes (27.38%), coronary artery disease (20.26%), atrial fibrillation (14.99%), and obesity (11.00%). Overall, there was a significant increase in the incidence of CRAO ED presentations over time, rising from 0.53 per 100,000 population in 2016 to 1.06 per 100,000 population in 2021 ( Figure 1 ).
| Characteristic | Total n = 3736 (%) |
|---|---|
| Age | |
| <25 | 33 (0.80) |
| 25-35 | 52 (1.39) |
| 35-45 | 109 (2.92) |
| 45-55 | 288 (7.71) |
| 55-65 | 749 (20.05) |
| 65-75 | 1084 (29.02) |
| 75-85 | 974 (26.07) |
| >85 | 447 (11.96) |
| Sex a | |
| Males | 1981 (53.02) |
| Females | 1755 (46.98) |
| Insurance payer | |
| Medicare | 2345 (62.77) |
| Private | 838 (22.43) |
| Medicaid | 277 (7.41) |
| Self-pay | 175 (4.68) |
| No charge | 14 (0.37) |
| Other | 80 (2.14) |
| Income quartile | |
| First | 1003 (26.85) |
| Second | 963 (25.78) |
| Third | 877 (23.47) |
| Fourth | 833 (22.30) |
| Comorbidities | |
| Hypertension | 2819 (75.45) |
| Hyperlipidemia | 1669 (44.67) |
| Diabetes | 1023 (27.38) |
| CAD | 757 (20.26) |
| Atrial fibrillation | 560 (14.99) |
| Obesity | 411 (11.00) |
| Valvular disease | 147 (3.93) |
| Elevated blood glucose | 129 (3.45) |
| Alcohol use | 75 (2.01) |
| Giant cell arteritis | 59 (1.58) |
| Systemic erythematous lupus (SLE) | 23 (0.62) |
Figure 2 illustrates the utilization trend in each stroke imaging modality by year. Overall, utilization rates of all workup modalities including brain imaging, carotid imaging, heart, and laboratory tests showed a linear increase from 2016 to 2021. Brain CT/CTA ( P = .024), carotid CTA ( P = .008), carotid MRA ( P = .024), ECG ( P = .024), and echocardiogram ( P = .024) displayed a significant increasing trend. Screening rates for GCA in patients >50 years of age also saw a significant increase over time. The proportion of patients receiving CRP ( P = .024), ESR ( P = .025), and CRP/ESR ( P = .024) screenings also demonstrated a significant trend increase ( Figure 3 ). Trend significance values were calculated using the Mann-Kendall Test ( α = 0.05).
Table 2 compares the sociodemographic characteristics of patients receiving stroke workup vs those without. Males were significantly more likely to receive stroke imaging ( P = .01). Trauma center status of II and III shows a linear significant difference to receive stroke imaging. Metropolitan nonteaching hospitals and nonmetropolitan hospitals were significantly more likely to receive stroke workup in comparison to metropolitan teaching hospitals ( P < .01). Insurance status and income quartile did not display a significant difference for receiving stroke workup.
| Characteristic | No Stroke Workup b n = 3071 (%) | Stroke Workup b n = 682 (%) | aOR | 95% CI | P Value |
|---|---|---|---|---|---|
| Sex a | |||||
| Female | 1471 (47.9) | 291 (42.67) | Ref | Ref | Ref |
| Male | 1600 (52.1) | 391 (57.33) | 1.29 | 1.06-1.58 | .01 |
| Ethnicity | |||||
| White | 1315 (42.82) | 413 (60.56) | Ref | Ref | Ref |
| Black | 275 (8.95) | 54 (7.92) | 0.82 | 0.58-1.14 | .24 |
| Hispanic | 146 (4.75) | 25 (3.67) | 0.69 | 0.43-1.09 | .11 |
| Asian | 46 (1.5) | 5 (0.73) | 0.43 | 0.17-1.13 | .08 |
| Native American | 9 (0.29) | 2 (0.29) | 0.70 | 0.15-3.34 | .65 |
| Insurance status | |||||
| Medicare | 1908 (62.13) | 445 (65.25) | Ref | Ref | Ref |
| Medicaid | 252 (8.21) | 27 (3.96) | 0.85 | 0.48-1.46 | .55 |
| Private | 677 (22.04) | 165 (24.19) | 1.15 | 0.83-1.59 | .38 |
| Self-pay | 149 (4.85) | 27 (3.96) | 0.85 | 0.46-1.54 | .59 |
| Income quartile | |||||
| First | 832 (27.09) | 175 (25.66) | Ref | Ref | Ref |
| Second | 801 (26.08) | 169 (24.78) | 0.86 | 0.64-1.13 | .28 |
| Third | 706 (22.99) | 177 (25.95) | 1.09 | 0.81-1.46 | .57 |
| Fourth | 684 (22.27) | 149 (21.85) | 0.93 | 0.67-1.38 | .67 |
| Trauma center status | |||||
| Trauma center level I | 1059 (34.48) | 193 (28.3) | Ref | Ref | Ref |
| Trauma center level II | 500 (16.28) | 134 (19.65) | 1.57 | 1.15-2.13 | <.01 |
| Trauma center level III | 235 (7.65) | 91 (13.34) | 1.80 | 1.22-2.64 | <.01 |
| Not a trauma center | 1242 (40.44) | 261 (38.27) | 1.23 | 0.92-1.65 | .15 |
| Hospital teaching status | |||||
| Metropolitan teaching | 2380 (77.5) | 482 (70.67) | Ref | Ref | Ref |
| Metropolitan nonteaching | 491 (15.99) | 120 (17.6) | 1.47 | 1.09-1.97 | .01 |
| Nonmetropolitan hospital | 200 (6.51) | 80 (11.73) | 2.12 | 1.48-3.14 | <.01 |
a Defined as sex assigned at birth.
b Stroke workup defined as having obtained at least one imaging modality in each category (head, neck, cardiac) during the encounter.
Table 3 shows the number of unique stroke imaging modalities received by patients. The proportion of patients receiving no stroke workup decreased from 66.75% to 57.87% during the study period ( P = .259). There was a significant trend increase in administration of one ( P = .024), four ( P = .009), and five ( P = .024) imaging modalities in a single stroke workup.
| Number of Modalities | ||||||
|---|---|---|---|---|---|---|
| Y | No Imaging | 1 a | 2 | 3 | 4 a | 5+ a |
| 2016 | 66.75% | 9.69% | 11.26% | 5.50% | 5.76% | 1.05% |
| 2017 | 60.75% | 11.92% | 11.92% | 7.24% | 6.31% | 1.86% |
| 2018 | 62.64% | 8.81% | 8.43% | 11.30% | 6.51% | 2.30% |
| 2019 | 60.93% | 7.58% | 9.77% | 11.95% | 6.81% | 2.96% |
| 2020 | 61.54% | 4.88% | 6.96% | 14.41% | 9.28% | 2.93% |
| 2021 | 57.87% | 2.97% | 4.34% | 5.70% | 10.16% | 18.90% |
| Trend ( P value) | .259 | .024 | .060 | .259 | .009 | .024 |
a Denotes a significant decreasing monotonic trend over time as calculated using the Mann-Kendall Test ( α = 0.05).
Table 4 depicts the ED ophthalmic imaging composed of optical coherence tomography (OCT) macula, fundus photography, and fluorescein angiography with 24, 19, and 20 patients, respectively, receiving imaging with no significant trend over the study period. Brain MRI/MRA and CT/CTA were performed in 14.10% and 29.04% of patients, respectively ( Table 5 ). Carotid imaging US, CTA, MRA was performed in 8.62%, 16.92%, and 3.82% of patients, respectively. Cardiac ECG and echocardiogram were performed in 27.78% and 12.50% of patients, respectively. ESR and CRP lab tests were performed in 22.70% and 19.58% of patients greater than 50 years old, respectively. At least one type of brain imaging (including MRI, MRA, CT, and CTA) occurred in 43.10% of patients, carotid imaging (including US, CTA, and MRA) in 29.36% of patients, and cardiac imaging (including ECG and Echo) in 40.28% of patients. Ophthalmic imaging was performed in 1.7% of the cohort.
| Y | OCT Macula | Fundus Photography | Fluorescein Angiography (FA) |
|---|---|---|---|
| 2016 | 5 | 3 | 10 |
| 2017 | 3 | 1 | 1 |
| 2018 | 1 | 1 | 1 |
| 2019 | 4 | 5 | 2 |
| 2020 | 6 | 4 | 3 |
| 2021 | 5 | 5 | 3 |
| Total | 24 | 19 | 20 |
| Trend ( P value) | .57 | .24 | .70 |
| Imaging Modality | Frequency (%) |
|---|---|
| Brain | |
| MRI or MRA | 526 (14.10) |
| CT or CTA | 1085 (29.04) |
| Carotid imaging | |
| US | 322 (8.62) |
| CTA | 632 (16.92) |
| MRA | 143 (3.82) |
| Cardiac evaluation | |
| ECG | 1038 (27.78) |
| Echo | 467 (12.50) |
| Ophthalmic imaging a | 63 (1.7) |
| Laboratory test b | |
| ESR | 773 (22.70) |
| CRP | 667 (19.58) |
| ESR & CRP | 624 (18.32) |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
