Case
A 30 year old man was a cyclist involved in collision. He “flipped over the handle bars” after hitting a rock and was thrown forward, striking his forehead on the concrete, without loss of consciousness. After the accident, there was no amnesia or focal neurologic symptoms except for visual loss in his right eye (OD). On examination, visual acuity was counting fingers OD. He had a right relative afferent pupillary defect, normal extraocular motility and no evidence of globe injury. The fundoscopy demonstrated normal optic disc and retinal appearances OU. A non-contrast CT scan of the head showed no acute abnormalities.
Introduction
Traumatic optic neuropathy may be caused by direct or indirect injury to the optic nerve. The most common demographic are young males. Traumatic optic neuropathy (TON) often results in profound visual loss in one or both eyes with dramatic consequences for future employment, social functioning and quality of life.
Direct, Indirect and Blast Injury
Direct optic nerve injury may occurs after penetrating trauma or when bony fragments or orbital hemorrhage impinge on the optic nerve.
Indirect injury is more common and may occur after head injury or globe injury. During globe injury, rotational forces may shear and avulse the optic nerve [1], rapid changes in intraocular pressure may transmit energy at the optic disc and RGC (Retinal Ganglion Cells) may die in association with retinal injury [1–3].
During head injury, energy transmitted through the skull may damage the optic nerve.
Blast injury may also damage the optic nerve and visual pathways. The different components of blast injury are characterized as: primary blast injury, caused by the direct effects of the blast wave; secondary blast injury caused by projectiles energized in the explosion; and tertiary blast injury caused by the blast wind moving the victim. Secondary and tertiary blast injury are penetrating and blunt injury, although the severity is often higher than in non-blast injury, but primary blast injury represents a distinct mechanism.
Mechanisms of Indirect Injury
Epidemiologic studies show that optic neuropathy may occur after relatively minor head injury [4]. The optic nerve is vulnerable to injury in the optic canal, where it is tethered by dural and vascular attachments [5]. Frontal forces applied to the skull cause deformation of and focusing of energy at the orbital apex [6] and optic canal [7, 8]. Post-mortem optic nerve pathology after fatal closed head injuries revealed intra-sheath hemorrhage in 83%, interstitial hemorrhage in 36% and necrotic and shearing lesions, defined as areas of axonal disruption and demyelination, in 44%, strongly suggesting vascular injury as a mechanism of optic nerve damage after severe head injury [9].
Direct mechanical disruption of axons is termed primary axonal injury. Delayed impairment of axonal function and structure is termed secondary axonal injury and is the predominant mechanism of axonal loss after traumatic brain injury [10]. Mechanical perturbation of the axonal surface activates enzymatic processes, which degrade the cytoskeleton. In a mouse model of indirect axonal injury (the fluid percussion model), secondary degeneration of axons occurred 1–3 h after injury [11]. In a guinea pig model of stretch-induced axonal injury, visual evoked potentials maximally detected structural injury at 28 min after injury, with a proportion of injured nerves having normal early VEP [12]. Delayed axonal degeneration (weeks after injury) occurs in animal models after ocular and brain primary blast injury [13, 14].
Epidemiology
The British Ophthalmic Surveillance Unit (BOSU) found an annual incidence of 1/1,000,000 [15], although it has also been reported in 2% of head injuries and 6% of facial fractures requiring surgical repair [16], which suggests that the true incidence may be much higher, given that 160,000 people suffer a traumatic brain injury in the UK every year (headway.org).
Most affected patients are young adult males [15]. In a series of 326 patients reviewed in a neuro-ophthalmic clinic for head trauma, 26 had indirect optic nerve injury no cases of direct optic nerve injury or optic neuropathy secondary to globe injury were reported [17], suggesting that indirect injury is the most common mechanism after head injury.
Diagnosis
The diagnosis of TON is made clinically when there is a history of trauma and evidence of an optic neuropathy, including a RAPD (Relative Afferent Pupillary Defect ) in unilateral TON. At presentation, the patient may be alert with no obvious evidence of injury, or the patient may be unconscious in cases of severe head injury.
Presentation
History
The initial step in the clinical evaluation is to obtain a complete history with chronological details. The history may need to be taken from family members, friends or witnesses at the scene of the trauma. It is important to ascertain the patient’s visual status prior to the injury, not only for clinical management but also for medico-legal purposes. Conscious patients complain of reduced vision, although in up to 10% of patients, awareness of visual loss may be delayed for several hours (which may or may not reflect secondary injury) [18].
Examination
Patients with head injury and traumatic optic neuropathy often have polytrauma and are unconscious on ITU when first seen by an ophthalmologist. In this case absence of a RAPD does not rule out TON, as injury may be bilateral and a RAPD may be difficult to assess in the brightly-lit ITU environment where the pupils are miosed by opiods. It is important here to exclude or treat globe injuries and orbital compartment syndrome. Resistance to retropulsion of the globe and elevated intraocular pressure suggest an orbital cause.
Presenting visual acuity ranges widely from 20/20 to no perception of light, the latter seen invariably in cases of optic nerve avulsion. Color vision is usually reduced in proportion to visual acuity.A RAPD should be present on the side of injury unless there is a contralateral optic neuropathy or gross retinal pathology from prior disease or concurrent trauma.
Extensive retinal pathology may also cause a RAPD, and must be excluded by dilated fundus examination. Before dilating the pupils, the primary team should be consulted in neurologically unstable cases.
Formal visual field testing should be attempted in alert and cooperative patients for documentation and monitoring purposes. Few conclusions can be drawn regarding topographic diagnosis of visual fields although altitudinal field defects in the context of TON tend to suggest the location of injury to be within the optic canal where superior pial vessels surrounding the optic nerve are thought to be the most susceptible to shearing forces, leading to an inferior altitudinal field defect.
Fundoscopy
Vitreous hemorrhage may indicate optic nerve avulsion or be associated with a posterior vitreous detachment and so should raise a suspicion for retinal detachment. However, Terson syndrome, associated with subarachnoid hemorrhage, should also be considered. If the vitreous hemorrhage is so extensive preclude fundoscopy, B-scan ultrasound examination will identify any retinal detachment. When fundoscopy is possible, the fundus should be carefully examined for evidence of retinal detachment and dialysis, vascular occlusion and hemorrhage, commotio and sclopetaria retinae, Purtscher’s retinopathy, traumatic macular hole and choroidal rupture.
The appearance of the optic disc depends on the site of the injury. If the optic nerve is avulsed as it enters the globe, there may be vitreous hemorrhage seen overlying the region of the optic disc and peripapillary retinal and subretinal hemorrhage with an excavated disc. Injury that occurs anterior to the entry of the central retinal vessels to the optic nerve may cause a central or branch retinal artery occlusion or a central retinal vein occlusion. Indirect injury that occurs posterior to the entry of the central retinal artery – most commonly within the optic canal - does usually not produce acute fundus changes. The optic disc may appear normal for up to 5 weeks, before pallor develops, dependent on the severity of optic nerve injury. Intraorbital optic nerve compression by intrasheath hemorrhage may cause the optic disc to swell. Bilateral optic disc swelling suggests papilledema caused by raised intracranial pressure (usually associated with an underlying head injury).
Ipsilateral retinal pathology needs to be assessed in combination with other examination findings, and a final determination made clinically whether the main visual problem is TON, a retinal problem or both.
OCT Features
OCT allows a more precise determination of RGC and axonal loss than the assessment of pallor, and OCT changes correspond topographically to visual field loss [19]. OCT evidence of RGC degeneration develops as early as 1.5 weeks after injury [20]. Segmentation of the macular cube on OCT to analyse the ganglion cell complex (including the ganglion cell layer and inner plexiform layer) shows thinning before peripapillary retinal nerve fiber layer analysis [20], probably because initial axonal edema obscures thinning.
Imaging Findings
CT scanning is routine in patients with head injury and fine (2 mm) orbital cuts with bony windows will assist with the assessment of orbital fractures (up to 2/3 of patients [21]) and globe injury. MR imaging is contra-indicated if there is any suspicion of a metallic foreign body but provides better contrast for brain, nerve and orbital soft tissues.
Neuroimaging findings after indirect optic nerve trauma are quite variable; typical signs that may help in the diagnosis include optic canal fracture, intrasheath hemorrhage, and possible alteration of MRI signal within the optic nerve itself. Severe visual loss is more often seen in cases with neuroimaging evidence of a fracture involving the optic canal, indicating likelihood of direct injury to the optic nerve. Contrast-enhancement may be seen after trauma on both CT and MR imaging, but is not usually diagnostically helpful [22], although it may indicate ongoing hemorrhage [23]. However, in many cases, no signs are present on routine imaging and the diagnosis of indirect trauma must be inferred from the clinical presentation. Diffusion weighted and diffusion tensor imaging may provide an alternative means to assess axonal injury. These measures may be normal initially, developing reduced fractional anisotropy (FA) by 1–7 days and then restricted diffusion as a later change [24–26], although reduced FA and restricted diffusion are not reported by all authors [25, 26].
Prognostic Features
With conservative management, 10–60% spontaneous improvement may be expected [18, 27]. Initial visual acuity of no light perception (NLP) carries a worse prognosis than light perception or better, but even those with NLP vision may experience significant recovery [18, 28, 29]. Facial and orbital fractures and facial penetrating injuries may be associated with a worse prognosis, particularly where blood is present in the posterior ethmoidal air cells, although this finding is not reproduced in all series [21, 27, 28, 30]. Age over 40 and loss of consciousness at time of injury may also confer a worse prognosis [30].
Management: A Review of the Evidence
Management includes observation, medical therapy, and surgery. There is no class I evidence to guide the clinician in the care of patients with this disorder, and the decision to treat or not are therefore often made on an individual basis.
Up to 60% of patients have significant visual recovery with conservative management [18], and any treatment must be evaluated in this context.
Corticosteroid therapy has been proposed for use in TON, although evidence to support their use is largely anecdotal, with reports of visual improvement in up to 2/3 of treated patients [REF Spoor 1990]. In the past 10 years, studies have emerged that have called into question the safety and efficacy of steroid therapy for this condition (see below). Novel medical therapies for TON such as erythropoietin infusion [31] showed early promise pre-clinical and early phase studies, but failed in a phase III trial [32]. Transcranial electrical stimulation also has been used in small cohorts of patients with greater improvement than conservative therapy on a non-standard visual field testing algorithm, but not other measures of visual function [33]. Numerous animal studies have tested multiple neuroprotective agents after optic nerve crush injury and other models including blast and fluid percussion, but rarely use functional outcome measures that test optic nerve integrity [2, 3, 34].
Similarly, surgical intervention for TON, typically optic canal decompression via endoscopic or transcranial surgery, has been correlated with visual improvement in several small case series [28, 35–37]. However, neither retrospective [28, 38] nor prospective [18] studies of medical therapy and/or surgery for TON have been able to demonstrate superiority of intervention to observation, leading to ongoing controversy in the management of this condition.
Corticosteroids in CNS Injury
Animal Studies
In rat studies, in which megadose corticosteroids were administered 30 min after optic nerve crush, treatment increased axonal loss [39], and meta-analysis of animal studies found low levels of randomization, masking to treatment allocation, or masked assessment of outcome in studies of corticosteroids in head injury models [40].
Historic animal studies of corticosteroid therapy after spinal cord injury suggested a functional benefit of megadose (15–30 mg/kg loading dose followed by infusion) [41].
Spinal Cord Injury
The NASCIS studies and others randomized 1580 patients with spinal cord injury to a variety of steroid doses from 1 g/day to 30 mg/kg loading dose, followed by 5.4 mg/kg/h (>11 g/day for a 70 kg adult). All studies failed their primary endpoints. The lower dose had no effect, but the higher doses were associated with some improvement in functional outcomes in post-hoc subgroup analyses in patients treated between 3 and 8 h after injury, but also an increase in wound infections with 1 g/day methylprednisolone and a trend towards an increased risk of septic complications in all other studies (http://www.trauma.org/archive/spine/steroids.html accessed 17/2/18).
Traumatic Optic Neuropathy
Meta-analysis of corticosteroid case series suggested that they improved visual outcome [27]. The International Optic Nerve Trauma Study was a prospective, non-randomised cohort study of 127 patients with TON, which found the greatest rate of visual improvement in the untreated group, but no significant differences in visual outcome between conservative management, medical management or surgery [18]. One small, prospective, randomized, double-masked, placebo-controlled trial, which recruited patients within 7 days of injury, found no benefit of 1 g methylprednisolone daily for 3 days followed by an oral taper when compared to placebo [42].
Head Injury
The CRASH trial was an international multicenter study that randomized 10,008 patients to megadose (maximum NASCIS dose) steroids <8 h after injury. There was no benefit in terms of functional outcome but there was an increase in all cause mortality in the steroid treated group [43].
The CRASH and NASCIS results are very relevant to patients with TON because most have head injuries, and so are vulnerable to the increase in mortality and the only (questionable) benefits seen were when treatment was started between 5 and 8 h after injury and the 1 g/day dose of methylprednisolone has never been shown to be beneficial.
International Variation in Practice
Management in the United States
There is no standard approach or established standard of care for the management of TON, and therapeutic decisions are guided by other factors, such as the patient’s or clinician’s desire to “do something” other than observation [38], or by other medical considerations (i.e. infection) that may limit medical and/or surgical treatment options. Until about 10 years ago, intravenous methylprednisolone 1 g/d (either single dose or divided four times daily) was given to many TON patients, especially those who were diagnosed within 72 h of their trauma, based on anecdotal and retrospective data suggesting efficacy [27] and the clinical belief that there was minimal risk in using this treatment in an otherwise healthy patient population. Based on the results of the CRASH trial and more recent interpretations of the NASCIS data, researchers have published reviews and opinions that corticosteroids are of no proven benefit but are of proven harm and should not be used for TON treatment ([44–46], and their use is falling out of favor [47].
Surgical management of TON patients is centered around identification of fracture and bony impingement upon the optic nerve within the optic canal and its potential for surgical reversal. Almost all facial trauma patients in the US undergo high-resolution CT scanning upon or shortly after hospital admission, and these studies should be adequate to identify a canal fracture. Radiographic findings must be correlated with the clinical exam; that is, a patient who presents with a canal fracture and no light perception vision may be much less likely to have a good response to surgery than one who has better preoperative vision. A documented decline in visual acuity during the first 48 h after injury also may indicate ongoing compression of the nerve that might be reversed by surgical canal decompression. If optic nerve sheath hemorrhage is present either in isolation or with a canal fracture, then a documented visual decline again may be considered an indication for surgery. If the patient will be undergoing cranial and/or endoscopic surgery for other reasons (repair of traumatic CSF leak, aneurysm clipping, drainage of hematoma), then canal decompression may carry less additional risk and could be considered. Overall, the trend in the US is toward observation of patients with TON [47], giving them appropriate counseling on the lack of evidence for the efficacy of corticosteroid and the potential harm that might result.
Global Perspective (Additional Commentary by Richard J. Blanch MD)
Management in Europe
Management is variable, with groups in Italy reporting the routine use of steroids and endoscopic decompression [30, 48], in Spain, steroid use and observation [49], in France observation and endoscopic decompression [50], groups in Germany describe the use of corticosteroids as historical and surgery as having no benefit [51], and in Austria decompression with or without steroids is reported [52]. Management in the UK is largely conservative [53, 54], although variation exists with a significant minority of clinicians administering corticosteroids [54].
Management in Singapore/Asia
The current management of TON across Asia differs somewhat between countries with the most numerous reports coming from China, where endoscopic optic nerve decompression is commonly performed, with or without steroids and with emphasis on treating without delay [21, 55–57]. A center in Taiwan suggests that megadose pulse steroid therapy as per the NASCIS II protocol should be used on patients who do not have contraindications to receiving steroids and have incurred injury less than 24 h previously [58]. In India, a combined modality protocol including mega-dose steroid therapy and endoscopic optic nerve decompression is considered a reasonable treatment option [59]. In Singapore, surgery is reserved for patients with radiological evidence of compression [60]. Intravenous methylprednisolone therapy (250 mg Q6h) may be given occasionally following discussion with the patient and acceptance that benefit of the therapy has not been proven. Physicians in Malaysia have advocated giving intravenous followed by oral corticosteroids [61].
Long–term Prognosis
The outcome of TON, regardless of management, may be most dependent upon the severity of the trauma as indicated by presenting visual acuity, although there is significant variation not explained by presenting visual function. Unfortunately, the behaviors that can lead to initial TON may put patients at risk for a second TON event, and recovery in such instances is uniformly poor irrespective of treatment approach [36]. Patients should be encouraged to wear eye protection if they are functionally monocular, and ongoing ophthalmic surveillance for other late complications of globe injury (e.g. glaucoma) should be conducted.
Current treatment of TON is often largely by patient and physician preference and institutional norms, with limited reference to the evidence-base. Future treatment of TON should be guided by an improved understanding of the response of retinal ganglion cells (RGC) to injury, and in particular, the time window in which the RGC remain viable and potentially responsive to methods that might preserve and/or restore optic nerve function before irreversible axonal disconnection and RGC death occur. Mechanistic pre-clinical studies and randomized prospective trials are needed to evaluate emerging therapies for this challenging disorder.
Conclusions
Traumatic optic neuropathy may result in association with a variety of ocular, orbital, and intracranial injuries. In both the United States and more broadly in the world, there is no consensus on medical or surgical treatment, although in USA and Europe, clinicians are moving towards observational management as the norm. The current evidence base would not support treatment of most cases because of potential side effects including systemic morbidity, as well as a lack of proof of superiority to observation. Early recognition of the condition may be useful in counseling the patient and family regarding prognosis, and in specific cases, to guide treatment aimed at preventing progressive visual deterioration. Because patients often have other life-threatening injuries, diagnosis of the visual loss may be delayed. The ophthalmologist should seek to determine the visual status even of unconscious and uncooperative patients as early as possible after injury with the goals both of establishing baseline status and to track potential deterioration versus stability that will guide any treatment decisions. Additional research is needed to understand the rate at and mechanisms by which permanent optic nerve and retinal ganglion cell damage occurs in order to help to define a therapeutic window for future interventions.
Key Points
Traumatic optic neuropathy typically occurs from deceleration injury but may happen with seemingly slight blows to the forehead
Indirect traumatic optic neuropathy is much more common than direct (i.e. penetrating or perforating) optic nerve injury.
The injury may go unrecognized because of co-existing CNS insult that makes visual acuity assessment difficult.
No standard treatment regimen exists, and no intervention has been shown to be superior to conservative management with supportive therapy.