Abstract
Postoperative visual loss (POVL) is a rare but devastating complication after nonocular surgery. POVL can be caused by central retinal artery occlusion, ischemic optic neuropathy (ION), and cortical blindness. The most common cause of POVL related to prone spine surgery in adults is ION. Risk factors for the development of ION associated with spine surgery include male sex, Wilson frame use, obesity, a lower percentage of colloid administration, longer anesthetic duration, and a greater amount of estimated blood loss. One recent study using national data demonstrates that the incidence of POVL related to ION after spine surgery has decreased over the past decade and a half. This decreased incidence may be related to prior studies identifying risk factors and large-scale efforts to educate practitioners caring for these high-risk patients.
Keywords
American Society of Anesthesiologists Postoperative Visual Loss Registry, Anesthesia, Central retinal artery occlusion, Cortical blindness, Ischemic optic neuropathy, Perioperative visual loss, Postoperative visual loss, POVL, Spinal fusion
Introduction
Postoperative visual loss (POVL) after nonocular surgery is an uncommon but devastating complication. POVL is most common after cardiac surgery followed by prone spine and head and neck procedures, with a national incidence ranging from 0.03 to 0.28% in these higher risk surgeries. Specific institutions may see lower or higher rates, which may be related to the types of procedures performed, the patient population, or specific practice patterns. The most common diagnoses associated with POVL are central retinal artery occlusion (CRAO), ischemic optic neuropathy (ION) [anterior ION (AION) and posterior ION (PION)], and cortical blindness (also known as cerebral visual loss). This chapter will discuss the major POVL diagnoses including the predisposing or associated factors and procedures, presentation, evaluation, and management.
Central Retinal Artery Occlusion
CRAO occurs as the result of decreased blood supply to the retina. The mechanism of decreased blood flow can be secondary to emboli or external pressure on the eye. In the perioperative period, the usual cause is from improper head positioning and direct pressure on the eye. Pressure on the eye leads to increased intraocular pressure that compromises flow in the central retinal artery, the major vascular supply to the retina, leading to retinal ischemia. Some of the first reported cases of POVL were related to CRAO and started appearing in the literature around 1950. All these early reports were thought to be caused by direct pressure on the eye from a face mask and attachments or from a horseshoe headrest while in the prone position, and possibly exacerbated by hypotension. These findings were even replicated in animal models by decreasing blood volume and blood pressure while applying direct pressure on the eye while under general anesthesia.
CRAO almost always presents as unilateral visual loss upon awakening from anesthesia. Patients can have little or no light perception and an afferent pupillary defect with slowed or absent pupillary light reflex. There can also be signs of trauma to or around the affected eye including periorbital edema, ptosis, proptosis, corneal abrasion, ecchymosis, and paralysis of extraocular eye muscles. Fundoscopic examination reveals a cherry-red spot in the macular region and diffuse retinal ischemia. Prevention of CRAO is key, as prognosis is poor and treatment options such as hyperbaric oxygen therapy, inhaled 5% carbon dioxide in oxygen, and directed intra-arterial thrombolysis have inconsistent results and are frequently not possible in the postoperative period.
Every attempt should be made during positioning to ensure that the head is in the proper position and that there is no pressure on the eyes. Head movement can occur intraoperatively, and frequent eye checks should be done to confirm that the eyes remain free of pressure throughout the entire procedure. Horseshoe headrests have been associated with numerous case reports and case series of CRAO thought related to the narrow margin of safety as the firm portion of the headrest comes very close to the eyes. However, sporadic case reports of CRAO have also been associated with newer headrests, thus highlighting the importance of frequent eye checks. Although well-intentioned, the use of eye goggles may actually increase the risk of pressure on the eyes in the prone position because of the narrow margin of clearance around the eye, making it necessary to avoid two obstacles, instead of one, while positioning and throughout the case.
Ischemic Optic Neuropathy
ION can be associated with a variety of procedures including prostatectomy, liposuction, major vascular surgery, and more commonly with cardiac surgery, head and neck surgery, and spine surgery. ION can be classified into two types depending on the location of the injury on the optic nerve. The central retinal vessels and the optic nerve pass through a connective tissue layer called the lamina cribrosa on their way to the globe. Injury to the optic nerve anterior to the lamina cribrosa is classified as AION, while injury posterior to the lamina cribrosa is classified as PION.
AION can be further subdivided into arteritic and nonarteritic. Arteritic AION is almost never seen perioperatively and is caused by temporal arteritis. Nonarteritic AION is the most common cause of acute optic neuropathy in the community setting in patients older than 50 years. It also occurs perioperatively and is usually associated with cardiopulmonary bypass procedures, major vascular operations, head and neck surgery, and spine surgery. The initial presentation of perioperative nonarteritic AION can occur immediately upon awaking from anesthesia or may occur several days after surgery. It is more often bilateral. Poorly reactive pupils are seen on pupillary examination, and a relative afferent pupillary defect occurs if the injury is unilateral or asymmetric. Fundoscopic examination will initially show optic disc edema, which is often accompanied by splinter or peripapillary flame-shaped hemorrhages. Disc edema may last for several weeks to months after which the optic disc becomes pale and atrophied and the heme is absorbed. In these later stages, it is difficult to distinguish AION from PION as they will both demonstrate optic nerve pallor with an otherwise normal fundus. The prognosis of postoperative AION is poor, and no treatment has shown benefit in controlled trials. Optimizing oxygen delivery with adequate hemoglobin and return to baseline perfusion pressure are frequently recommended by ophthalmologists, although exact transfusion thresholds and blood pressure ranges have not been methodically identified. Published reports of attempted treatments for spontaneously occurring AION outside the operative arena include therapy with diphenylhydantoin, aspirin, systemic and intravitreal corticosteroids, anti–vascular endothelial growth factor agents, erythropoietin (EPO) or EPO receptor agonists, optic nerve sheath decompression, optic neurotomy, and hyperbaric oxygen. None of the treatments that have been tested in randomized controlled trials have shown a benefit.
PION, like AION, can be arteritic or nonarteritic. Nonarteritic postoperative PION is a more common cause of POVL involving head and neck procedures and prolonged prone spine surgeries when compared to AION. PION presents on awakening from anesthesia and usually affects both eyes. It has been suggested that PION presents with more severe clinical impairment than AION because the posterior portion of the optic nerve travels though rigid, inflexible structures and optic nerve swelling leads to more intense axonal compression and optic nerve damage. Pupillary examination and visual field defects are similar to those found in perioperative AION. Initial funduscopic examination is normal with PION but progresses to show optic disc pallor after several weeks to months. As with AION, the recovery is poor and no treatments have been shown to be efficacious.
Cortical Blindness
Cortical blindness is the least common cause of POVL following nonocular surgery in adults and is thought to be caused primarily by emboli and less commonly by cerebral hypoperfusion. It has a very high association with pediatric surgical procedures in national databases, but it is unclear if this association is related to the high rates of cerebral microemboli that have been detected in some patients in pediatric studies, or to the common practice of the combined use of deliberate hypotension and anemia in high-blood-loss procedures or to preexisting diagnoses of cortical blindness in patients with coexisting congenital diseases such as cerebral palsy. Approximately one quarter of adults have a patent foramen ovale providing the opportunity for paradoxical emboli. Cortical blindness is most often associated with cases with high embolic loads such as cardiopulmonary bypass, spine and bone and joint surgeries. In fact, a study by Takahashi et al. used transesophageal echocardiography to show that 80% of patients undergoing instrumented spine procedures had moderate to severe embolic events. Pedicle screw insertion was associated with the highest numbers of pulmonary emboli. Despite these findings, the incidence of cortical blindness in national databases in these instrumented spine operations in adults is very low.
Cortical blindness from global cerebral hypoperfusion can be caused by severe intraoperative hypotension or less severe hypotension with tight stenoses or atretic vessels in the cerebral vasculature. It is more often reported in procedures such as cardiac, hip, and spine surgery where there can be large deviations from baseline cerebral perfusion related to blood loss, cardiac disease, or surgeon requests for deliberate hypotension to reduce blood loss.
Cortical blindness usually presents on awakening from anesthesia. Pupillary light reflexes and funduscopic examination are usually normal, and these are distinguishing features of this diagnosis from CRAO or AION. Homonymous hemianopia can be seen in unilateral lesions, while bilateral injury to the occipital cortex can lead to complete blindness. There may also be other focal neurologic deficits depending on the location of the lesion. Computed tomography or magnetic resonance imaging can help identify the location of the lesion. No definitive treatment for cortical blindness has been identified; however, some ophthalmologic consultants recommend correcting anemia and maintaining normotension. Visual fields may return in a few days, but spatial perception and issues with distinguishing relationships between size and distance may be long lasting. Recovery from cortical blindness in children related to hypoperfusion perioperatively is reasonably good with 11 of 15 children in one study having complete recovery.