Long-Term Outcomes After Central Retinal Artery Occlusion Treated Acutely With Hyperbaric Oxygen Therapy: A Case Series

Abstract

Purpose:

This study assesses the long-term outcomes, including neovascular complications, of central retinal artery occlusion (CRAO) treated acutely with hyperbaric oxygen therapy (HBOT).

Methods:

Four cases of CRAO treated acutely with HBOT were reviewed. Visual and structural outcomes were reviewed. Ocular complications including neovascularization were assessed and risk factors determined.

Results:

Two patients with a history of non-insulin dependent diabetes mellitus (NIDDM) developed early-onset ocular neovascularization within 1 month following treatment, with final vision of light perception over 1 year after injury. One patient with NIDDM and 1 patient without NIDDM did not develop ocular neovascularization; both had improvement in final visual acuity to 20/400 and 20/250, respectively.

Conclusions:

Patients treated acutely with HBOT for CRAO may require more frequent and earlier monitoring for complications, especially in patients with diabetes. Further research is needed to determine the long-term safety and efficacy of HBOT for CRAO, especially in the setting of systemic disease such as diabetes.

Introduction

Central retinal artery occlusion (CRAO) is a vision-threatening ophthalmic emergency with an estimated incidence of approximately 1 in 100 000. When the central retinal artery is acutely occluded, resultant ischemia and infarction ultimately lead to visual loss. Approximately 80% of patients following CRAO will have a final visual acuity (VA) of 20/400 or worse in the affected eye. ¹ Traditional treatment recommendations for CRAO attempt to lower intraocular pressure (IOP), increase retinal perfusion pressure, or displace the embolus. These include ocular massage, glaucoma medications, diuretics, vasodilators, anterior chamber paracentesis, and hypercapnic vasodilation (ie, breathing into a paper bag). Novel interventions have been described, such as revascularization through laser or surgical arteriotomy, and intra-arterial fibrinolysis. ²

Despite the various interventions described, including in several case reports and series demonstrating successful improvement in VA with the described therapies, there is no generally accepted definitive treatment regimen for CRAO. Of the few randomized clinical trials completed, none provide definitive evidence for improved visual outcomes with any of the treatments listed, and some, such as fibrinolysis, may be associated with significantly higher adverse events. ³ More recently, hyperbaric oxygen treatment (HBOT) has been used as a novel therapy for CRAO. The proposed mechanism for HBOT in the acute treatment of CRAO is that resultant hyperoxygenation of the choroidal circulation allows for diffusion of oxygen into the inner retinal layers and potentially decreases ischemic damage to the retina. There are few case reports and series as well as no randomized controlled trials that demonstrate the efficacy of HBOT in the acute treatment of CRAO. This is likely related to the rarity of the disease and the necessity for acute intervention in a relatively small treatment window. The authors present 4 cases of patients who received HBOT for CRAO, specifically focusing on the long-term visual outcomes and complications associated with this therapy.

Methods

This study is a case series conducted at a single-center, tertiary care academic facility. Four cases of individuals diagnosed with CRAO and subsequently treated acutely with HBOT were reviewed and the cases described. HBOT varied in treatment time and number of treatments, following the hyperbaric medicine’s guidelines and the patient’s response to treatment. In most instances, protocol included continued treatment until VA returned to baseline or plateaued over a 2- to 3-day period. Visual and structural outcomes were evaluated and compared. Ocular complications including neovascularization were assessed and risk factors were determined.

Results

Findings including visual outcomes and ocular complications are demonstrated in Table 1.

Table 1.

Summary of Findings.

Case	VAi	VAf	Onset to presentation, h	Onset to HBOT, h	Diabetes (HbA1c)	CRS	Time to ONV, d
1	HM	LP	5	12	+ (5.8)	+	13
2	LP	LP	11	19	+ (5.9)	+	31
3	CF at 2 feet	20/400	8	22	+ (6.0)	+	NA
4	HM	20/250	3	6	–	+	NA

Abbreviations: CF, counting fingers; CRS, cherry-red spot present on presentation; HbA_1c, hemoglobin A_1c at presentation; HBOT, hyperbaric oxygen therapy; HM, hand motion; LP, light perception; NA, not available; ONV, ocular neovascularization; VAf, final visual acuity (at last follow-up > 1 year from initial presentation); VAi, initial visual acuity (at presentation).

Case 1

The patient was a 74-year-old man with a history of hypertension (HTN), noninsulin-dependent diabetes (NIDD), and hyperlipidemia who presented to the emergency department 5 hours after sudden-onset vision loss in his right eye. He had a history of rhegmatogenous retinal detachment in both eyes after pars plana vitrectomy with endolaser in the right eye and scleral buckle in the left eye, and he had 20/20 VA in both eyes prior to presentation. There was no history of diabetic retinopathy (DR) documented. At the time of presentation, he was found to have hand motion vision in the right eye with a pale retina, arteriolar attenuation, and a cherry-red spot consistent with a diagnosis of CRAO. Fluorescein angiography (FA) demonstrated almost complete retinal arteriolar hypoperfusion, late peripapillary choroidal staining, and minimal to no venous filling more than 2 minutes into the study. A complete workup was negative for giant cell arteritis and systemic hypercoagulability.

The patient underwent 2 subsequent treatments in a hyperbaric oxygen chamber (2.8 atmosphere absolute [ATA] of 100% oxygen for 90-minute sessions), which started approximately 12 hours after the onset of initial symptoms, with no improvement in his final VA. The patient reported decreased vision in the right eye 13 days from diagnosis. At that time he was found to have neovascularization of the angle (NVA), neovascularization of the iris (NVI), and neovascularization of the cornea. Panretinal photocoagulation was then performed, and the neovascularization was regressed at his subsequent visits. The patient ultimately developed neovascular glaucoma in his right eye less than 1 month after presentation, and 2 years after the initial presentation his VA was light perception (LP) only.

Case 2

The patient was an 80-year-old woman with a history of primary open-angle glaucoma and pseudophakia in both eyes, as well as HTN, NIDD, coronary artery disease, 3 myocardial infarctions, and hemispheric transient ischemic attack, who presented to the emergency department with sudden vision loss in the right eye that occurred 11 hours prior to presentation. On examination, she was found to have LP vision with retinal whitening and a cherry-red spot without visualization of retinal emboli, consistent with a diagnosis of CRAO (Figure 1). FA demonstrated diminished retinal arteriolar filling, macular nonperfusion, and delayed arteriovenous transit time (Figure 2). She had no evidence of DR on examination or prior documentation of DR. The patient underwent an extensive vascular, cardiac, and hypercoagulability workup. She received ocular massage for 10 minutes in addition to alternating topical brimonidine and dorzolamide/timolol ophthalmic drops for 4 cycles. The patient then proceeded to the hyperbaric oxygen chamber approximately 19 hours after her initial symptom onset and received 2 treatments per day for 3 days (90-minute periods of 100% oxygen at 2.0 to 2.8 ATA). Her VA did not improve after the HBOT.

Figure 1.

Right fundus photograph of the patient in case 2, on initial presentation, demonstrating retinal whitening and a cherry-red spot without embolus, consistent with a diagnosis of central retinal artery occlusion.

Figure 2.

Fluorescein angiography of the patient presented in case 2, revealing (left) remarkable slowing of arteriolar filling in the right eye (early phase) and (right) delayed arteriovenous transit (with minimal venous filling more than 2 minutes into the study).

The patient subsequently presented to the emergency department 31 days after her initial presentation with complaints of a right-sided headache, red eye, and eye pain in the right eye. She was found to have no light perception vision, 360° of NVI, a layered hyphema, and IOP elevated to 31 mm Hg. After unsuccessful lowering of her IOP on maximum medical therapy, she underwent a right anterior chamber paracentesis followed by intravitreal antivascular endothelial growth factor (anti-VEGF) injection. Ahmed tube shunt placement in the right eye was performed 2 days later. The patient subsequently underwent panretinal photocoagulation and 4 additional treatments with intravitreal anti-VEGF for management of her neovascular glaucoma. Her final VA 1 year after her initial presentation was LP OD.

Case 3

The patient was a 63-year-old woman with a history of pseudophakia in both eyes, HTN, NIDD, and hyperlipidemia who presented to the optometry clinic with sudden vision loss in the left eye that began approximately 8 hours prior to presentation. On examination by the ophthalmology department, she could count fingers (CF) at 2 feet and was found to have retinal whitening with a cherry-red spot in the left eye and moderate nonproliferative DR, consistent with a diagnosis of CRAO. FA demonstrated hypofluorescence of the macula and delayed arteriovenous transit time (∼40 seconds).

On presentation, the patient was hypertensive to 214/128 mm Hg, so HBOT was delayed until her blood pressure was successfully treated. She received her first HBOT 22 hours following initial symptoms, and her VA improved to 20/800. She continued to receive twice-daily treatment for a total of 7 treatments (90- to 278-minute sessions of 100% oxygen at 2 to 2.8 ATA), with a final best-corrected VA of 20/400 at 1 year after her initial presentation. At that time, she had no evidence of neovascular changes.

Case 4

The patient was a 66-year-old man with a history of HTN and hypothyroidism who presented with 3 hours of sudden onset, painless vision loss in the left eye. He was found to have hand-motion vision OS with a pale retina and cherry-red spot. Additionally, multiple Hollenhorst plaques were visualized throughout the superior and inferior arterial arcades in the left eye. The patient received ocular massage, topical timolol, brimonidine, and dorzolamide in the clinic. FA was performed and demonstrated delayed arteriovenous transit time (∼308 seconds). He was admitted to the hospital for a stroke workup to include cardiac echocardiogram, carotid ultrasound imaging, computed tomography and computed tomography angiogram of the head and neck, and magnetic resonance imaging of the brain. While an inpatient, the patient received HBOT twice daily for 9 total treatments (2.0 to 2.8 ATA for 90 to 122 minutes per session) . His first treatment was performed approximately 6 hours after the onset of symptoms. His vision improved to CF at 3 feet after receiving therapy.

During his hospital stay, the patient was found to have complete occlusion of his left common carotid artery as well as critical narrowing of his right internal carotid artery. He was treated with dual antiplatelet therapy, and he subsequently underwent a right transcarotid artery revascularization with vascular surgery 10 days after his initial presentation. A repeat FA on postoperative day 1 showed drastically improved macular perfusion. One year after his initial presentation, his VA improved to 20/250. At that time, there was no evidence of ocular neovascularization (ONV).

Conclusions

HBOT is an American Heart Association–level IIB recommendation and one of the few treatments available for CRAO. Early treatment may improve VA; the current recommendation for emergent hyperbaric medicine referral is within 12 to 24 hours of presentation of symptoms, ^4,5 although the best outcomes have been reported to occur when HBOT is started between 8 to 12 hours after vision loss. ⁶ There are various suggested treatment protocols, but the most common includes starting at 2.0 ATA and increasing up to 2.8 ATA for 3 to 10 treatments, discontinuing treatment after the VA plateaus. ^4,7

Despite early enthusiasm with HBOT, few studies have examined the long-term potential risks related to therapy. The cases presented here demonstrate the potentially increased risk of ONV in patients with diabetes who present with CRAO and receive HBOT. Patients 1 and 2 both had diabetes and subsequently developed ONV within 1 month after the initial onset of their symptoms, with a final VA of LP in both patients. This is in contrast to the 1 patient with moderate nonproliferative DR and 1 patient with no history of diabetes who had no evidence of ONV at follow-up, both of whom had improvement in final VA (20/400 from CF at 2 feet and 20/250 from CF at 3 feet).

Even in the absence of DR, diabetes may increase the risk of ONV. Mason et al ⁸ reported that NIDD was a risk factor for ONV, with an adjusted odds ratio of 5.2. Jung and colleagues ⁹ similarly reported diabetes in 50% of patients with CRAO and NVI vs 17.3% in patients with CRAO without NVI; the strongest predictor of NVI was found to be the lack of retinal reperfusion, leading to chronic ischemia. Neither group of authors included patients treated with HBOT. ^8,9 A proposed mechanism for higher rates of neovascularization in patients with diabetes who receive HBOT may be that high levels of choroidal and retinal arteriolar oxygen keep the already ischemic retina alive longer, consequently causing a longer release of anti-VEGF in response to the CRAO, leading to neovascularization.

Hadanny et al published the largest cohort to date of patients with nonarteritic CRAO who received HBOT within 20 hours of symptom onset. The authors concluded that the most important predictor of successful treatment was the absence of a cherry-red spot at the time of funduscopic examination, so a cherry-red spot may be a marker for irreversible anoxic retinal damage. However, the reported results included only immediate adverse outcomes following HBOT (such as middle ear or sinus barotraumas) and immediate post-HBOT VAs, without the assessment of long-term complications such as ONV. ¹⁰

The development of ONV may be due to chronic retinal ischemia, as seen in other conditions such as proliferative DR, central retinal vein occlusion, and ocular ischemic syndrome. In the study by Mason et al, in which patients were not treated with HBOT or other prophylactic therapy, the average time from diagnosis of CRAO to development of ONV was 30.7 days, ranging from the date of presentation to 137 days. Of the 83 patients in that case series, 14.5% developed ONV, most frequently NVI (83.3%) and neovascularization of the disc (16.7%). ⁸ Patients with diabetes may be at higher risk because of chronic ischemia with lower incidence for reperfusion. ^8,9 Each of the patients in our study presented within the past 2 years, and 2 of the 3 patients with diabetes subsequently went on to develop ONV within 1 month of presentation, with 1 patient having developed NVI, NVA, and neovascularization of the cornea.

Similar to the cases presented in this study, Tang and colleagues described a case of a patient with diabetes and a CRAO who received 5 days of HBOT, as per their facility’s protocol. The patient presented 24 days after initial onset of symptoms with elevated IOP and NVI and NVA on anterior segment examination with gonioscopy. The authors concluded patients undergoing HBOT for CRAO may need closer monitoring for ONV. ¹¹ The findings presented in the previously described case series concur with this conclusion and also suggest patients with diabetes may especially require closer observation for ONV.

Limitations to this study include relatively small sample size, multiple HBOT regimens, and varied medical histories, as well as limitiations inherent to a retrospective study design. Future studies may include prospective randomized trials of patients with vs without diabetes undergoing HBOT for CRAO.