Reversal of vision loss after traumatic optic neuropathy

Abstract

Traumatic optic neuropathy is sinister sequelae of craniofacial trauma leading to vision loss. The decision between early medical or surgical intervention is usually individualised. Visual evoked potentials may guide the treatment plan. We describe a young male presenting 5 days after a road traffic accident with no perception of light vision in the right eye. He was managed medically with high dose of intravenous steroids. At the 3-month follow-up, he reported a reversal of vision loss with return of visual acuity to 3/60, which improved to 6/36 at 5 months and remained stable at 8 months.

Background

Vision loss after craniofacial trauma without globe injury results from traumatic optic neuropathy (TON). Hippocrates had noted that blows to the eyebrows may in fact cause blindness.¹ Two variants—direct and indirect are described; indirect TON is thought to be more common.

In direct TON, there is evidence of trauma to the optic nerve from a projectile or sharp object including bony fragments from fractures of the optic canal leading to strain, compression or transection of axons of the optic nerve or its vascular supply.² Sarkies described five mechanisms of direct TON: optic nerve avulsion, transection, optic nerve sheath haemorrhage, orbital haemorrhage and orbital emphysema.¹

Indirect TON is suspected when after a head injury, the patient develops reduced vision and an afferent pupillary defect despite a normal slit-lamp examination (in the acute setting) and normal MRI and CT of the optic nerve and canal.^{2 3} The mechanism results from the transmission of compressive forces to orbital apex and optic canal after blunt trauma. Within the tight optic canal, contusion of pial vasculature leads to inflammatory oedema of the axons and development of a compartment syndrome with a vicious cycle of hypoxia, ischaemia and cell death.¹

With advances in imaging technologies, high-resolution CT of the orbit and paranasal sinuses with thin sections reconstructed digitally can help detect subtle injuries and fractures of the optic canal. This may reclassify a seemingly indirect TON to a direct variant. Anatomical relationship of the optic nerve and posterior nasal sinuses can be studied in these high-resolution images that may detect variations predisposing the patient to optic nerve injury.⁴ Haemorrhage in sphenoid sinus and posterior ethmoid cells, especially the Onodi cell variant, may point towards an underlying fracture of the optic canal with an impinging bony fragment leading to a direct injury to the optic nerve.⁵

Clinical practice world over is divided between surgical and medical management for optic nerve decompression in the acute setting. The International Optic Nerve Trauma Study (IONTS) remained inconclusive while comparing high-dose corticosteroids or surgical optic nerve decompression.⁶ A Cochrane review by Yu-Wai-Man et al recommended that surgery in TON is controversial and should be individualised.⁷ Combined management with initial high-dose steroids followed by surgical optic nerve decompression can also be offered.

Delayed presentation with profound vision loss and absence of visual evoked potential (VEP) at presentation are predictors for a poor recovery.^{6 8 9} Literature does demonstrate anecdotally recovery of vision from no perception of light (no PL) in patients of both direct and indirect TON.^10–13

Case presentation

A 17-year-old male individual presented to the ophthalmology department with vision loss in the right eye following a roadside accident 5 days prior. He was hit by a speeding bike at a crosswalk and fell on his face. He developed swelling on the right cheek and periocular region. There was no history of loss of consciousness or ear or nose bleed. He did not seek immediate medical intervention. Five days later, when the periocular swelling reduced, he noted the inability to see with the right eye. This prompted him to seek medical attention and he presented to the ophthalmology department. At presentation, his visual acuity was no PL in the right eye and 6/6 in the left eye. Pupillary reactions revealed an afferent pupillary defect in the right eye and the presence of direct but absence of consensual response in the left eye. An abrasion with scabbing was noted in the superolateral sub-brow region. The rest of the anterior segment evaluation including ocular motility function was normal. Fundus evaluation was unremarkable with no evidence of disc pallor or retinal haemorrhage. The red–green colour vision on Ishihara colour plates and contrast sensitivity on Pelli-Robson of the uninvolved left eye were normal (1.80 at 1 m). The systemic evaluation including a comprehensive general physical examination was unremarkable.

Investigations

A CT of the orbit and paranasal sinuses was requested. There was evidence of fracture of the posteromedial wall of the orbit, fracture along the right optic canal, haemorrhage in the sphenoid sinus and presence of an Onodi cell. The right optic canal was dehiscent with the optic nerve coursing within the sphenoid sinus, type 3 DeLano and a corresponding deep carotico-optic recess. The sphenoid septum was noted to attach directly onto the optic nerve. On the left side, the optic nerve caused an impression on the lateral sphenoid sinus wall, type 2 DeLano. The rest of the orbital walls and globe were intact (figure 1).

Figure 1.

(A) Axial (mid orbit) section of CT demonstrates haemorrhage in the right sphenoid sinus and an Onodi cell on the right side (asterix). In addition, the optic canal is dehiscent as the optic nerve courses within the sphenoid sinus, type 3 DeLano, with a deep carotico-optic recess. The green arrow marks the fracture along optic canal and the red arrow shows the attachment of the intersphenoid septum. (B) Coronal section corresponding to the axial cut in (A shows the intersphenoid septum attaching directly onto the dehiscent optic nerve (red arrow). (C) Axial cut (superior orbit) shows a fracture of the posterior medial wall of the right orbit (yellow arrow).

Differential diagnosis

In view of the fractures along the right posteromedial wall of orbit and optic canal, presence of Onodi cell and a dehiscent right optic canal within the sphenoid sinus, a diagnosis of right TON, possibly a direct variant, was established.

Treatment

The treating team of ophthalmologists and otolaryngologists discussed the options of surgical decompression or intravenous methylprednisolone (IVMP). In view of vision in the right eye of no PL and delay in presentation of 5 days, an initial trial of IVMP (1 g/day) was instituted. After two doses, the option of surgical decompression was reconsidered. A pattern VEP was requested to guide the decision. The right P100 was non-recordable while the left P100 potential was normal (figure 2). The prognosis for vision recovery was limited and the same was discussed with the patient’s parents. They opted to continue with medical management and surgical decompression of optic nerve was declined.

Figure 2.

Visual evoked potential (VEP) (pattern) at time of injury. The right eye waveform and P100 potential were non-recordable while the left eye was normal. The visual acuity was no perception of light in the right eye and 6/6 in the left eye.

Outcome and follow-up

The patient received five doses of IVMP followed by oral steroids (1 mg/kg/day) for 11 days. These were then tapered to 10 mg/week over the next 4 weeks. There was no visual recovery at the 4-week follow-up. The patient then returned for his next follow-up after a gap of 3 months after the initial injury. He reported ability to perceive nearby objects with the right eye over the past few days. He was receiving bi-weekly injection of methyl cobalamin and daily oral folic acid supplements, as per advice from a neurologist, for the past 4 weeks. Ophthalmic evaluation revealed a visual acuity of 3/60 in the right eye with a relative afferent pupillary defect. Right optic disc demonstrated temporal pallor. The superolateral sub-brow wound had healed with residual scarring. Colour vision assessment with Ishihara charts showed the inability to read demo plate with the right eye and a normal perception in the left eye. Contrast sensitivity on Pelli-Robson testing was 0.00 in the right eye and 1.80 in the left eye (at 1 m). A flash VEP was repeated and a waveform was recordable from the right eye, though it was inconsistent and with a reduced amplitude (figure 3). The patient was counselled regarding limited improvement in the right eye, development of right eye exotropia and asked to maintain a regular follow-up. At a follow-up at 5 months, he regained a visual acuity of 6/36 in the right eye with a sensory right exotropia (figure 4). The visual acuity remained stable at the 8-month follow-up. Optical coherence tomography (OCT) scans were performed using Spectralis (Heidelberg Engineering, Heidelberg, Germany). The central macular thickness and macular contour and configuration were normal in both eyes. OCT of the peripapillary retinal nerve fibre layer demonstrated atrophy on the right side (figure 5).

Figure 3.

Visual evoked potential (VEP) (flash) 3 months later showing normal response in the left eye and reduced amplitude and prolonged latency on the right eye. The visual acuity was 3/60 in the right eye and 6/6 in the left eye.

Figure 4.

Follow-up at 5 months: external photograph demonstrating right eye exotropia and scar in the right superolateral sub-brow region (red arrow). The final visual acuity remained stable: 6/36 in the right eye and 6/6 in the left eye at 8 months.

Figure 5.

Optical coherence tomography (OCT) of the macula (A, B) and optic nerve head (C, D) of right and left eyes. The macular OCT did not reveal any structural abnormality with a central macular thickness of 203 μm in the right eye and 205 μm in the left eye. The peripapillary retinal nerve fibre layer demonstrated atrophy on the right side, consistent with the clinical finding of optic nerve pallor due to atrophy secondary to trauma.

Discussion

Vision loss due to TON is a rare but devastating sequela of craniofacial trauma. TON is classically described as direct or indirect, with the latter being reported as a more common variant.^1–3 Indirect TON results from transmitted forces leading to inflammatory oedema of the axons within the narrow bony optic canal and development of a compartment syndrome. The IONTS compared corticosteroid treatment or optic canal decompression versus observation in patients with indirect optic nerve injury. Patients with direct trauma to the optic nerve were not recruited in this study. The results remained inconclusive and the study could not establish a standard of care.⁶

Poor presenting visual acuity, absence of visual evoked potential (VEP) and delay in intervention predict a graver prognosis.^{6 8 9} Goldenberg-Cohen et al reported that improvement in TON is likely if vision at presentation is equal to or better than 6/60, regardless of intervention.¹⁴ Mahapatra and Tandon demonstrated that the presence of a detectable VEP at presentation (even with an abnormal waveform) was associated with a high chance of visual recovery as compared with an absent VEP. In their study of 43 patients, at the final follow-up, 7/7 with normal VEP, 10/12 with abnormal (but present) VEP and only 2/24 with absent VEP had a recovery of vision. The remaining 22 with a persistently absent VEP waveform did not show any visual improvement.¹⁵ Correlation of VEP and its use as a predictor for visual outcome after TON has also been demonstrated by Agarwal and Mahapatra.¹⁶

Anecdotal reports have shown that patients of TON with no PL vision and absent waveform may experience full recovery without intervention.^{11 12 17} Wolin and Lavin described one patient of indirect TON who regained vision from no PL to 20/50 at 5 months, without any intervention and two cases who demonstrated an initial visual recovery before initiation of corticosteroids that improved further after therapy.¹¹ Rosenberg et al demonstrated recovery of visual acuity from no PL and return of VEP waveform after 3 months in a patient of significant head trauma with depressed cortical functions after a skiing accident. The eventual visual acuity was 20/70 at 1 year. They concluded that despite no evidence of visual function on repetitive ophthalmic examinations including electrophysiological studies, there remained potential for long-term spontaneous recovery of visual function. Vision assessment was complicated in their case due to the severe brain injury and absence of initial communication and feedback from the patient.¹⁷

In contrast, a study by Yu et al evaluated the role of intervention with methylprednisolone followed by surgical decompression in 96 patients of indirect TON with no PL vision. The improvement in visual acuity was 63.6% if treated within 3 days of injury, compared with 35.7% if delayed by more than 7 days. The same study also documented 26 patients who regained vision after the initial steroid therapy and did not require surgical decompression.¹³

In the direct variant of TON, as there is ‘direct’ physical trauma to the optic nerve, the neuronal damage and vision loss are profound and visual recovery is scant. There is limited literature describing visual recovery after direct TON. In the presence of direct projectile or sharp fragment impinging on the optic nerve, an urgent institution of corticosteroids followed by surgical decompression to relieve the impacting trauma has shown to aid visual recovery, as described by Nazir et al. They reported a case who developed direct TON 1 day after endoscopic endonasal orbital decompression for Grave’s ophthalmopathy. The patient underwent immediate surgical removal of the bony fragment compressing the optic nerve, followed by intravenous and oral steroids. She regained vision from no PL to 6/6 over a course of 4 months.¹⁰ Table 1 summarises relevant literature demonstrating visual recovery from no PL in patients with TON.

Table 1.

Summary of studies demonstrating vision recovery from no perception of light in patients of TON

References	Cases	Clinical presentation	Outcome
Wolin and Lavin11	4	Indirect TON with no demonstrable optic nerve compression or impingement Rx: corticosteroid therapy in 3; observation in 1	All 4 regained useful vision; 1 recovered spontaneously to 20/50
Aggarwal and Mahapatra16	100	71 had abnormal VEP and 29 at absent VEP waveform at presentation Rx: patients were prescribed corticosteroids if they presented within a month of injury. Those who remained static after steroid therapy were undertaken for surgical decompression This study demonstrated the role of both positive and absent VEP waves in predicting the outcome	23 patients demonstrated improvement in visual acuity; out of these 22 had presented within 3 weeks after injury 15 patients who had persistently negative VEP did not show any vision improvement
Nazir et al10	1	Vision loss of 1 day duration due to direct TON with bony fragment of lamina papyracea impinging the optic nerve, after undergoing orbital decompression for Grave’s ophthalmopathy Rx: decompression surgery followed by intravenous and oral steroid therapy	6/6 at 4 months
Rosenberg et al17	1	Severe brain injury, multiple skull fractures, subdural and subarachnoid haemorrhages, and bifrontal intraparenchymal haemorrhagic contusions. Clinically and electrophysiologically demonstrable loss of vision	Return of VEP waveform Vision improved to 20/200 at 3 months and 20/70 at 1 year
Weed et al12	1	Fracture of posterolateral wall of left sphenoid sinus in area of left optic canal Rx: observation only	Spontaneous improvement to 20/40 at 3 weeks
Yu et al13	96 + 26	All patients of indirect TON with no PL vision Rx: initial management with intravenous methylprednisolone for 3 days; surgery offered to those who did not improve (n=96)	Visual acuity improved in 45/96 patients (46.9%) after surgery 26 patients improved with steroids alone
Our case	1	Optic canal fracture with anatomical predispositions of deLano type 3 and Onodi cell (suggestive of direct TON) Rx: methylprednisolone started at 5 days after injury followed by oral steroids	6/36 at 8 months

no PL, no perception of light; TON, traumatic optic neuropathy; VEP, visual evoked potential.

Our case presented with the absence of a detectable waveform on VEP. Thin slices and digital image reconstruction of the CT scan demonstrated the following features—anatomical predispositions to TON due to dehiscent optic canal within the sphenoid sinus (deLano type 3), posterior-most pneumatised ethmoid sinus cell lying in close proximity to the optic nerve (Onodi cell), fracture along the optic canal and haemorrhage in the sphenoid sinus. The presence of these factors prompted a diagnostic possibility of ‘direct’ TON. Decompression surgery was offered but the patient declined. High-dose intravenous steroids followed by slow oral taper was instituted. The patient had a moderate recovery of vision of 6/36 and a detectable (but diminished) waveform on VEP at 5 months which remained stable at 8 months. Though the presentation was delayed (5 days), the intervention with steroids was beneficial.

There is no consensus regarding the appropriate cut-off time period for the effectiveness of steroid therapy in TON. Thakar et al have reported that a delay of greater than 2 weeks from the onset of TON to active intervention results in a reduced potential for vision recovery.¹⁸ Another study on patients of indirect TON managed with methylprednisolone±surgery reported an improvement of vision to >3/60 in up to 70% if treatment was initiated within 7 days of trauma, compared with only 24% if treatment was delayed.⁹ Valerie et al reported statistically comparable improvement in vision after high-dose steroids in patients who presented within or beyond 72 hours of indirect optic nerve injury, advocating a role of steroid therapy despite a delayed presentation.¹⁹

Our report thus highlights the recovery of vision and return of VEP waveform in a patient of TON. We demonstrate the role of radiology to help detect anatomical variations predisposing to TON. In addition, we discuss the relevance of induction of high-dose intravenous corticosteroid therapy despite a delay in presentation.

Patient’s perspective.

After my accident, I was unable to open my right eye. Four days later when I opened it, I was unable to see even light with my right eye. Doctors at PGI informed me regarding trauma to nerve of eye and started injections. When they performed test with flashes of light, they told the nerve is not responding. They suggested surgery but informed that result may be equivocal. I opted to continue with medicines only and refused surgery. Today, after 4 months, I can see better with right eye but not as good as my left. Doctors have told me that my eye is deviating outwards and I have been advised exercises for that. I am grateful to the almighty for my return in vision and I am hopeful that it will improve further. I was counselled very patiently at every step during my treatment and I understand that this sort of return of vision is seen rarely. I am very thankful for the doctors who have treated me and wish them the best.

Learning points.

• In patients with craniofacial trauma without globe injury, visual acuity, pupillary reactions and visual evoked potentials aid the diagnoses of traumatic optic neuropathy (TON).

• Results with medical or surgical management in TON are equivocal and the decision must be individualised from case to case.

• CT with thin sections or digital reconstruction is vital in detecting pre-existing bony variations that may predispose a patient to TON.

• The concept of indirect versus direct TON needs to be revisited. High-resolution CT with digital image reconstruction may reveal haemorrhage in sphenoid sinus and small fractures of the optic canal with underlying optic nerve bruising, making the diagnosis of a direct nerve injury more appropriate.

• Late presentation with no perception of light, traditionally thought to be irreversible, may show limited recovery with intervention, thus, highlighting the beneficial role of delayed therapy with intravenous methylprednisolone.