- Email: [email protected]
Traumatic optic neuropathy in children and adolescents
Major Articles Traumatic Optic Neuropathy in Children and Adolescents Nitza Goldenberg-Cohen, MD,d Neil R. Miller, MD,a,b,d and Michael X. Repka, MDc,d Purpose: Retrospective review to describe the features and visual outcome of traumatic optic neuropathy (TON) in children and adolescents. Methods: Interventional and observational case series of children ⬍18 years with TON. Results: Forty children, 24 boys and 16 girls, were identified with 43 affected eyes. The mean age was 11.6 years (range, 2 to 18). The most common causes were motor vehicle accidents (62%) and sports injuries (22%). The trauma was blunt in 78% of cases and penetrating in 22%. Visual acuity at presentation for 27 patients whose visual acuity could be assessed ranged from no light perception to 20/80. Interventions included corticosteroids (n ⫽ 18), decompression of the optic canal (n ⫽ 3), and optic sheath nerve fenestration (n ⫽ 1). Visual outcome for 22 children with at least 1 month of follow-up was no light perception in 10 (45%), ⬍ 20/200 in 7(32%), 20/200 to ⬍ 20/80 in 1(4%), and ⱖ 20/80 in 4 (18%). There was no difference in the rate or degree of improvement between treated and untreated patients. Conclusion: TON in children is caused by mechanisms similar to those that cause TON in adults. The severity of visual loss as well as the rate and degree of improvement are also similar. Our data do not indicate that treatment improves visual outcome. (J AAPOS 2004;8:20-27) raumatic optic neuropathy (TON) is an uncommon, serious insult that often leads to permanent visual loss. TON occurs in 0.5% to 5% of patients after sustaining head trauma.1,2 Typically, the loss of vision occurs at the time of the injury; however, a minority of patients experience delayed visual loss, usually occurring hours to a few days later. TON may be of the anterior variety; however, in most cases the optic disc is normal, suggesting damage to the posterior orbital, canalicular, or intracranial portions of the nerve. Approximately 50% of patients experience no improvement in vision regardless of treatment.1 Traumatic optic neuropathy can result from direct or indirect trauma to the nerve in the orbit, optic canal, or intracranial space, although the optic canal appears to be the site of injury in most cases. The canalicular portion of the nerve is fixed and thus subject to force transmitted by the bones of the skull to the optic canal. Thus, even slight longitudinal deformation of the optic canal can stretch or compress the nerve and result in trauma to the axons or vascular supply.3 Hematoma, thrombosis, and edema
T
within and around the nerve then develop. In addition, trauma may fracture the optic canal or the anterior clinoid processes and produce bone fragments that impinge on the optic nerve. The management of TON is controversial.1,4,5 Various options include high-dose6-8 or low-dose9 corticosteroids, immediate decompression of the canalicular portion of the optic nerve,10 decompression of the canalicular optic nerve after a course of systemic corticosteroids,11,12 observation, and optic nerve sheath fenestration in cases of anterior TON.3 Unfortunately, findings in published reports have been inadequate to determine which, if any, of these treatments improves visual outcome after TON, mainly because of retrospective design, small numbers of matched cases, treatment bias, and unmasked outcomes. The only prospective, controlled, randomized trial—the International Optic Nerve Trauma Study—was abandoned because of insufficient recruitment.4 Only a few studies of TON have focused on children and adolescents.11,13 The purpose of this report is to describe our experience with the presentation and outcome of TON in children and adolescents.
PATIENTS AND METHODS a
b
c
From the Departments of Neurology, Neurosurgery, and Pediatrics, Wilmer Ophthalmological Institute, Johns Hopkins University School of Medicine,d and Johns Hopkins University School of Medicine, Baltimore, MD. Supported by the Isabel and Zanvyl Krieger Fund (NGC), Baltimore, MD; and in part by the Richard Baks Fellowship Fund, Baltimore, MD; the Stewart Wolff Fellowship Fund, Baltimore, MD; and the Judith and Paul Romano Binocular Vision and Strabismus Fellowship Endowment Fund, Baltimore, MD. Submitted April 2, 2003. Revision accepted August 22, 2003. Reprint requests: Michael X. Repka, MD, 233 Wilmer, Johns Hopkins Hospital, 600 N Wolfe St, Baltimore, MD, 21287-9028. Copyright © 2004 by the American Association for Pediatric Ophthalmology and Strabismus. 1091-8531/2004/$35.00 ⫹ 0 doi:10.1016/j.jaapos.2003.08.009
20
February 2004
A retrospective chart review of the investigators’ diagnostic databases was used to identify all patients with TON seen between 1973 and 2003. The diagnosis of TON in the alert child was based on signs of optic neuropathy: reduced acuity and color vision, relative afferent pupillary defect (RAPD), visual field defect, and light or red color desaturation. The diagnosis in an uncooperative child was based on the presence of an RAPD along with supportive findings on neuroimaging scan and fundus examination. Eligible patients had to have been ⬍18 years old at the time of their injury and examined by one of us on at least Journal of AAPOS
Journal of AAPOS Volume 8 Number 1 February 2004
Goldenberg-Cohen, Miller, and Repka
21
Table 1 Demographic and clinical characteristics
Characteristic No. of patients Mean age in years (SD) Median age (yr) Male (%) Female (%) Type of injury (n) Vehicle/bicycle accident Sports* Fall from bridge Gunshot Other/unknown State of consciousness Coma (%) Altered conscious state (%) Normal (%) Unknown (%) Time to medical care Immediate (%) Within 1 day (%) Within 2 weeks (%) Uncertain (%) Site of initial care (%) Johns Hopkins Hospital Other hospital in United States Outside United States Baseline visual acuity (PD) Unable to test† Unknown NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Canal fracture on imaging Missed on initial scan
Optic Nerve Decompression Surgery 4 8.8 (5.0) 10.0 3 (75) 1 (25)
Total
Observation
High-Dose Steroids
40 11.6 (4.2) 12.0 24 (60) 16 (40)
18 11.3 (4.4) 12.0 9 (50) 9 (50)
18 12.5 (3.7) 13.5 12 (67) 6 (33)
25 9 1 2 3
14 1 0 0 3
8 8 0 2 0
3 0 1 0 0
10 (25) 4 (10) 18 (45) 8 (20)
6 3 6 3
4 0 12 2
0 1 0 3
29 (73) 3 (8) 1 (2) 7 (18)
11 0 1 6
14 3 0 1
4 0 0 0
13 (32.5) 21 (52.5) 6 (15)
6 9 3
3 12 3
4 0 0
6 7 12 3 10 2 0 11 4
5 5 3 1 4 0 0 5 2
1 2 5 2 6 2 0 3 2
0 0 4 0 0 0 0 3 0
CF ⫽ counting fingers; NLP ⫽ no light perception. *Baseball (n ⫽ 2), soccer (n ⫽ 11), softball (n ⫽ 1), golf (n ⫽ 1), hockey (n ⫽ 1), rollerblade (n ⫽ 1), running (n ⫽ 1), and sledding (n ⫽ 1). †Unconscious (n ⫽ 5) and swollen eyelids (n ⫽ 1).
one occasion. Patients treated before our examination were included only if pretreatment examination findings were available. Patients with a ruptured globe (n ⫽ 1), shaken-baby syndrome (n ⫽ 3), or no recorded visual acuity at outcome (n ⫽ 2) were excluded. Data tabulated included the age, gender, mechanism of injury, visual acuity at diagnosis and at outcome, presence of an RAPD, treatment (type and timing), presence of fracture or intrasheath hemorrhage on neuroimaging scan, presence and location of facial and orbital fractures, reconstruction surgeries if performed, visual acuity at most recent visit, and length of follow-up.
RESULTS We identified 40 patients (24 boys and 16 girls) with 43 affected eyes (mean ⫾ SD ⫽ 11.6 ⫾ 4.2 years; median ⫽ 12.0; range, 2 to 18 years). The important demographic and clinical characteristics are listed in Table 1. Initial
medical care was provided at John Hopkins Hospital for 13 children (32.5%), elsewhere in the United States for 21 (52.5%), and outside the United States for 6 (15%). The diagnosis was made within 24 hours of the injury for 32 of 33 patients for whom such data were available. The right eye was affected in 15, the left in 22, and both in 3 children. The 3 children with bilateral TON sustained their injuries in motor vehicle accidents (MVAs). Two of these children had no light perception (NLP) in one eye at presentation. The mechanism of TON was blunt facial trauma in 31 (78%) and penetrating orbital trauma in 9 patients (22%) (Table 1). The most common etiologies were an MVA in 23 (57.5%) and a sports injury in 9 (22.5%) patients. All of the MVA victims experienced blunt trauma. Two were drivers, 13 were passengers (11 in the back seat and 2 in the front seat), 1 was riding a motorcycle, 2 were riding bicycles, and 5 were pedestrians. Cases of penetrating
22
Journal of AAPOS Volume 8 Number 1 February 2004
Goldenberg-Cohen, Miller, and Repka
Table 2 Final visual acuity and optic nerve atrophy measurements for all patients
Measurement Final visual acuity NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Optic nerve atrophy
Total (n ⴝ 40)
Observation (n ⴝ 18)
High-Dose Steroids (n ⴝ 18)
Optic Nerve Decompression Surgery (n ⴝ 4)
12 1 13 3 11 40
6 0 4 0 8 18
5 1 7 2 3 18
1 0 2 1 0 4
CF ⫽ counting fingers; NLP ⫽ no light perception.
orbital injury were caused by an antenna, stick (n ⫽ 2), tree branch, screwdriver, pencil, finger, and gunshot (n ⫽ 2). Visual acuity within 3 days of the trauma was known for 27 patients (29 eyes) (68%) (Table 1). Twelve (44%) had NLP, 3 (11%) had LP, 10 (37%) had counting fingers (CF) to ⬍ 20/200, and 2 (7.4%) had visual acuity 20/200 to 20/80. Baseline acuity after injury was not known for 13 children, 5 of whom were in a coma, 1 of whom had severely swollen eyelids, and 7 of whom had been referred more than 3 days after the injury with incomplete data regarding vision at presentation. An RAPD was detected in 14 of the 20 children for whom the presence of an RAPD was specifically recorded at the initial examination. Among the other 6 patients, an RAPD developed within 3 days after the injury. The 3 other children had evidence of bilateral optic neuropathy. Twenty-two children had evidence of a retrobulbar optic neuropathy, with normal-appearing optic discs, whereas 2 had swollen discs. Twelve had evidence of intraocular trauma as well as an optic neuropathy. Retinal hemorrhages were seen in 6 patients, 3 with orbital fracture, 1 with choroidal rupture, and 1 with optic disc swelling. Other intraocular findings included single cases of hyphema, commotio retinae, retinal tear, and macular hole. Fundoscopy could not be performed in 12 of the patients at the time of the injury, mainly because of severe eyelid swelling from associated facial fractures. Computed axial tomography (CAT) scanning was performed at the time of the trauma on 34 patients. Plain radiographs only were taken of 2 patients. Findings from 7 of the CAT scans showed normal results, whereas findings from 27 showed abnormalities related to the trauma. Fourteen patients had evidence of an orbital fracture on neuroimaging. The orbital fractures involved the floor (n ⫽ 6), posterior orbit (n ⫽ 4), and orbital rim (n ⫽ 4). Optic nerve sheath hemorrhage was seen in 1, optic nerve swelling in 6, and optic canal fractures in 11 patients. Other imaging findings included skull fractures in 4, orbital foreign bodies in 3, and nonorbital facial bone fractures in 7 patients. Associated brain injuries seen on CAT scan findings were subdural hematoma (n ⫽ 3), temporal lobe infarct (n ⫽ 1), and intraventricular hemorrhage (n ⫽ 1). Two patients were also studied using magnetic reso-
nance imaging. The results of one study were normal, whereas the other study showed changes consistent with a hemorrhage within the sheath of the orbital portion of the optic nerve. Neurologic Status Ten (25%) children were in a coma after their trauma, 4 (10%) had altered consciousness, and 18 (45%) were alert and oriented. No data were available for 8 patients (20%) (Table 1). Additional cranial neuropathies were found in 9 patients; most had multiple cranial nerve pareses. Six children had an oculomotor nerve paresis, 4 had an abducens nerve paresis, 2 had a trigeminal neuropathy, 2 had a facial nerve paresis, and 1 had a trochlear nerve paresis. Treatment and Visual Outcome for Entire Cohort Twenty-two patients were treated for their TON. Treatment included high-dose corticosteroids in doses ranging from 1 mg/kg/d to 1,000 mg/d prednisone equivalent (n ⫽ 17), retrobulbar steroid injection (n ⫽ 1), optic canal decompression (n ⫽ 3), and optic nerve sheath fenestration (n ⫽ 1). The remaining 18 patients were not treated. Table 2 lists the visual acuity outcomes for the entire cohort of patients. Of the 40 patients, 11 (28%) had vision ⱖ 20/80 in the affected eye at the most recent follow-up. Visual Outcome for Subgroup of Children With Both Baseline and Follow-Up Visual Acuities An outcome analysis was performed on a subgroup of patients followed up for at least 1 month who had quantitative baseline and final visual acuity measurements. Twenty-two patients met these criteria (Table 3). Their mean time to follow-up was 8.6 months (median ⫽ 5 months; range, 1 month to 2.5 years). Nine of the 22 patients (41%) who had baseline visual acuity measurements had improved visual acuity at follow-up; 4 of the 9 had intraorbital penetrating trauma. Table 3 lists the outcomes for this group of patients. The outcomes are stratified by initial acuity as well as treatment. In brief, among the 7 patients who were not treated, 2 experienced improvement in visual acuity, 1 from counting fingers to 20/50 and another from 20/200 to 20/20. Four untreated patients had no change in vision, and 1 experienced dete-
Journal of AAPOS Volume 8 Number 1 February 2004
Goldenberg-Cohen, Miller, and Repka
23
Table 3 Visual acuity measurement at final visit by treatment group for patients with documented baseline acuity and at least 1 month of follow-up Optic Nerve Decompression Surgery Total Observation High-Dose Steroids (n ⴝ 4) Visual Acuity (n ⴝ 22) (n ⴝ 7) (n ⴝ 11) Baseline visual acuity NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Outcome visual acuity by baseline acuity Baseline: NLP NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Baseline: LP NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Baseline: ⬍ 20/200 to CF NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Baseline: ⬍ 20/80 to 20/200 NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80
11 2 7 2 0
3 0 3 1 0
4 2 4 1 0
4 0 0 0 0
11 7
3 3
4 3
4 1
3 1
0 0
1
2 1
2 1
0 0
2 1
0 0
1
0
1
0
0 7 2 1 2
0 3 1
0 4 1 1 2
0 0 0
2 2
2 1
0 1
0 0
2
1
1
0
0
CF ⫽ counting fingers; NLP ⫽ no light perception.
rioration to NLP. Of 15 patients in this subgroup who were treated, 11 were treated with systemic corticosteroids, and four were treated with immediate optic canal decompression. Among the 11 patients treated with corticosteroids, 4 experienced visual improvement, 4 had no change in vision, and 3 experienced deterioration in vision. All 4 of the patients treated surgically (3 with optic canal decompression and 1 with optic sheath fenestration) had NLP vision at presentation. Among these patients, 3 improved to 20/200, whereas one remained status NLP. Secondary analyses were performed including only those patients with follow-up intervals of ⬎ 1 week and ⬎ 3 months. Thirteen of 27 patients (48%) had improved visual acuity after ⱖ 1 week of follow-up, whereas 8 of 18 patients (44%) had improved visual acuity after ⱖ 3 months of follow-up. Outcome by Subgroups Based on Potential Prognostic Factors Blunt Versus Penetrating Trauma. Initial and final visual acuity measurements subdivided by type of injury are listed in Table 4. Improvement was noted in 4 of 8
children (50%) after penetrating injury, whereas 2 had deteriorated of acuity despite receiving high-dose steroid therapy. Improvement was noted in 9 of 19 (48%) patients after blunt trauma. Optic Canal Fracture. Eleven patients had neuroimaging evidence of an optic canal fracture (Table 5). Three of these patients were treated with optic canal decompression and 3 with corticosteroids. The remaining 5 patients were not treated (Table 4). The acuities at presentation were NLP in 7, LP in 2, and unknown in 2 patients. Acuity improved from NLP in 3 of 3 patients in the optic canal decompression surgery group to 4/200 and 5/200, respectively. One untreated patient spontaneously improved from NLP to 20/50. One of the 2 patients with visual acuity of NLP improved after steroid therapy to a final acuity of 20/200 PD. Four patients (36%) had a delay in diagnosis (2 patients were scanned only on referral, and 2 scans were reinterpreted on referral). Two of these children were treated with corticosteroids, and 2 children were not treated. Optic Nerve Swelling. Swelling of the intraorbital portion of the optic nerve was noted in 5 patients on CAT
24
Journal of AAPOS Volume 8 Number 1 February 2004
Goldenberg-Cohen, Miller, and Repka
Table 4 Type of injury and visual outcome* Penetrating Trauma (n ⴝ 8) Initial Acuity
Outcome Visual outcome NLP LP ⬎ 20/200 to CF ⬎ 20/80 to ⱖ 20/200 ⱖ 20/80 Status (%) Improved Stable Deteriorated
Blunt Trauma (n ⴝ 19) Final Acuity
2 1 3 2 0
Initial Acuity
3 0 3 0 2
Final Acuity
10 2 6 1 0
7 1 6 2 3
4 (50) 2 (25) 2 (25)
9 (47) 9 (47) 1 (5)
CF ⫽ counting fingers; NLP ⫽ no light perception. *Table includes only patients with initial and final visual acuity measurements.
Table 5 Optic nerve findings on imaging and initial/final acuity measurements Optic Canal Fracture (n ⴝ 11)
Optic Nerve Edema (n ⴝ 6) Visual Acuity NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Unknown
Severed Optic Nerve (n ⴝ 1)
Orbital Foreign Body (n ⴝ 3)
Initial
Final
Initial
Final
Initial
Final
Initial
Final
2 1 3 0 0
3 0 2 1 0
7 0 2 0 0 2
5 0 3 2 1
0 0 1 0 0
1 0 0 0 0
1 0 2 0 0
1 0 2 0 0
CF ⫽ counting fingers; NLP ⫽ no light perception.
scan and in one patient on quantitative ultrasound (Table 4). Two of these patients had NLP in the affected eye at presentation, whereas one had LP vision in the affected eye, and three had acuity of counting fingers. Four of these five patients were treated with corticosteroids, after which two improved from CF to 3/200 and 20/400, respectively; two patients deteriorated from CF to and LP to NLP, respectively. The one patient who underwent optic nerve sheath fenestration improved from NLP to 20/200. The patient with evidence of optic nerve swelling who was not treated remained NLP at follow-up.
DISCUSSION This study identified 40 children with TON, the largest series of children reported to our knowledge. Most of these patients experienced immediate and severe visual loss at the time of injury. For patients with baseline acuities and follow-up ⱖ 1 month, acuity improved in 41%, but only 18% eventually regained ⱖ 20/80 vision. Only five of 40 patients had visual acuity ⱖ 20/50 in the affected eye at follow-up. Improvement appeared to be more likely when vision was ⱖ 20/200 at presentation, regardless of treatment. Conversely, patients with NLP acuity at presentation rarely experienced significant visual improvement, regardless of treatment. Two other case series6,11 included large numbers of children (Table 6). These stud-
Table 6 Pediatric traumatic optic neuropathy Findings Number of patients ⬍ 18 years Most common etiology (n)
Percent with NLP at presentation Treatment modalities Observation (%) High-dose steroids (%) Surgical decompression (%) Timing of decompression surgery Percent with VA improvement Visual outcome NLP ⱖ 20/80
Present Study
Lessell6
Mahapatra et al11
40
16
35
Motor vehicle Motor vehicle Fall (20), (23), bicycle (7), bicycle (6) motor vehicle (2), falls (1) falls (0) (12) 12 71
18 (45) 18 (45) 4 (10) ⬍9d
9 (56) 6 (38) 3 (19) Within 48 h
0 35 (100) 5 (14) After 3 wk
41
44
34
12 11
1 8
NA NA
NLP ⫽ no light perception.
ies found rates of improvement of 34 and 44% of children, respectively. For children who were not treated in this study, the rate of improvement was 29% compared with 36% for those treated with corticosteroids. These rates are lower than the 50% probability of improvement reported
Journal of AAPOS Volume 8 Number 1 February 2004
for adults.4 This lower rate of improvement in children with TON may be related to the more severe nature of the damage in children or possibly to our inability to identify milder cases. MVAs were the most common cause of TON in our series, accounting for 57.5% of the injuries. The patient was most often injured as a back seat passenger, likely because most children ride in the back seat. The injury was noted less often when the patient had been a driver, bicyclist, or pedestrian. Sports injuries (22.5%) were the second most common etiology in our patients. MVAs were also the most common cause in children in a report from Japan.14 Studies from India found falls to be the most common cause,11,13 whereas bicycle injuries were most common in an earlier report from the United States.6 TON in adults is also most often caused by MVAs.1,4 Diagnosing TON in children can be a challenge. Children may not notice the visual deficit, may not report it, or may not be able to perform acuity measurements. Three children in our series initially did not report their monocular blindness. In addition, a limited visual acuity examination by counting fingers performed at near in an accident ward failed to identify two patients who had partial loss of vision. These children were not correctly diagnosed until another evaluation was performed, in one case almost 6 months after initial presentation. Last, many children could not perform afferent system testing because of age, fear, or loss of consciousness. Because of the difficulties inherent in examining children after severe head trauma, we believe that the single most important clinical test of optic nerve function is the swinging flashlight test to identify a relative afferent pupillary defect. The presence of a RAPD in the absence of ophthalmoscopic evidence of retinal damage almost invariably indicates optic nerve dysfunction. Thus, pupillary light reflex testing and a swinging flashlight test, which can be performed whether or not a patient is conscious, should be performed for every child who experiences head trauma. In addition, a recent report found poorer visual outcomes in patients with TON who had a RAPD, making the test useful not only for diagnosis but also as a predictor of prognosis.15 In this study, children were treated with systemic corticosteroids (n ⫽ 18), surgery (n ⫽ 4), or observation (n ⫽ 18) based on best clinical judgment. Steroids were given systemically in high doses (⬎ 1mg/kg/d orally or intravenously) to 17 patients, whereas 1 patient received a retrobulbar corticosteroid injection. No significant differences in improvement or final acuity measurements between this treatment and no treatment were observed. Surgery was not performed in all of the patients with optic nerve swelling in this study because of a time delay in reaching our hospital or a delayed diagnosis during the initial evaluation. We examined several features of the condition for an association with prognosis. These included type of trauma,
Goldenberg-Cohen, Miller, and Repka
25
presence of an optic canal fracture, and poor vision at presentation. Although Wang et al8 noted a better visual outcome after blunt trauma than after penetrating trauma in their series of patients with TON (most of whom were adults), we found no difference in outcome between blunt and penetrating trauma in our series. Eleven of our patients (28%) had evidence of a fracture of the optic canal on imaging. This rate is between the 14% rate reported among children in India11 and the 44 % rate identified in adults by CAT imaging.16 All of these studies likely underestimate the actual incidence of optic canal fractures in patients with TON. For example, Wohlrab et al10 found that 75% of their patients with TON had fractures at surgical exploration. The importance of CAT scanning with respect to treatment and prognosis in patients with TON is controversial.1 Although it is reasonable to treat patients in whom CAT scanning shows a bone fragment impinging on the optic nerve, the presence or absence of a simple fracture of the optic canal does not seem to be predictive of prognosis, regardless of treatment. Nevertheless, the presence of a fracture of the optic canal on CAT scan was a poor prognostic sign in this series. Only 1 of 11 patients (9%) with evidence of an optic canal fracture on imaging had a final acuity ⱖ 20/80 compared with 10 of 24 patients (42%) without optic canal fractures who eventually achieved ⱖ 20/80. Severe initial visual loss also was a predictor of a poor visual outcome among our patients. Patients with NLP were unlikely to improve irrespective of management. Such an observation has been made by others.8,17,18 However, Levin at al,19 in a study of 31 surgically decompressed adults, found no difference in outcome based on preoperative visual acuity. Mahapatra and Tandon11 noted that the presence of a normal or even abnormal but detectable VEP was associated with a high chance of visual recovery, whereas an absent VEP at presentation was almost never associated with substantial improvement. Optic nerve swelling after blunt trauma has been associated with a favorable prognosis for visual recovery. Brodsky et al20 reported three patients who had partial recovery of vision. In this study, two patients had optic disc swelling. One was status NLP and improved to 20/200 after sheath fenestration, and the other improved from counting fingers to 20/400 after parenteral administration of steroids. As noted above, there is no consensus as to the optimum treatment for TON. The results of several retrospective case series8,21,22 suggest a beneficial effect of systemic corticosteroids, whereas others3,8,10-12 found improved visual outcome after surgical therapy. Still other series1,6,9,19,23 have failed to demonstrate any clear-cut benefit of treatment compared with observation alone. The largest prospective comparison study of patients of all ages with TON found no difference in outcome among patients treated with surgical decompression, high-dose steroids, or observation.4 In that study, approximately
26
Goldenberg-Cohen, Miller, and Repka
50% of the patients in the steroid and observation groups improved by three or more lines of acuity, whereas 38% improved in a similar fashion in the surgically treated group. These investigators concluded that the medical evidence did not support one particular approach and that clinicians should continue “to decide to treat or not treat on an individual patient basis.”19 The results of our study support this recommendation because we were unable to identify any treatment that was more likely to result in visual improvement than observation alone. Recently, Steinsapir et al5,24 found that methylprednisolone exacerbated axonal loss after optic nerve crush injury in a rodent model. Whether the findings of these studies are applicable to humans is unclear. Human studies have not shown a deleterious effect of high-dose steroids,4 nor is it certain that a rodent optic nerve crush model is representative of human injury. However, this work raises an additional consideration for the clinician considering high-dose steroid therapy for an adult or child with TON. The findings of our study are subject to a number of important biases. Visual acuity before the injury was not known for any of the patients, so it is possible that some patients with evidence of an optic neuropathy after trauma had pre-existing optic nerve damage. A second bias is the possibility that cases of minor optic nerve injury were not identified in children because of the failure of such children to notice or report mildly decreased vision. Furthermore, patients with milder cases of TON might have been less likely to be referred for evaluation and treatment by us. These sampling biases would increase the proportion of NLP cases we have reported. Because NLP may be a predictor of a poor prognosis as identified in this report and in the report by Wang et al,8 these ascertainment biases would make the outcome for childhood TON appear to be worse than it actually is. Other biases include inability to obtain initial visual acuities in all patients and a lack of standardized method of visual acuity measurement because of the retrospective nature of the study. In addition, although we believe our study to include the largest number of children with TON to date, the number of patients in this study nevertheless is relatively small, making it impossible to detect small differences in treatment outcomes. The retrospective nature of the study means that treatments were selected at the discretion of the treating physician and that the timing of treatment was not part of any defined protocol. In addition, there likely was a bias toward treatment, especially in patients who experienced progressive loss of vision after injury, thus potentially making interventions appear to be less successful. The lack of uniform follow-up examination beyond 1 month might have underestimated the visual acuity outcomes for those patients with shorter follow-up because of late recovery, thus making the outcome for TON appear to be worse than it actually was. However, in secondary analyses we found no difference in the rates of improve-
Journal of AAPOS Volume 8 Number 1 February 2004
ment with minimum follow-up intervals of 1 week, 1 month, or 3 months. In conclusion, TON in childhood occurs most often from blunt trauma under circumstances similar to those in adults and typically results in visual loss that rarely improves significantly, regardless of treatment. The diagnosis may be difficult to establish but may be aided by a careful evaluation of pupillary responses. Factors suggestive of a worse visual outcome include NLP acuity at presentation, bone fracture(s) in the optic canal or posterior orbit, and evidence of intraorbital optic nerve swelling. As in the case of adults, clinicians should individualize treatment for children with TON by recognizing that spontaneous recovery may occur and that all treatments have potential risks.
References 1. Steinsapir KD, Goldberg RA. Traumatic optic neuropathy. Surv Ophthalmol 1994;38:487-518. 2. Steinsapir KD, Goldberg R. Traumatic optic neuropathies. In: Miller NR, Newman NJ, editors. Walsh and Hoyt’s Clinical NeuroOphthalmology. 5th ed. Baltimore (MD): Williams and Wilkins; 1998. p. v. 1. 3. Lu¨bben B, Stoll W, Grenzebach U. Optic nerve decompression in the comatose and conscious patients after trauma. Laryngoscope 2001;111:320-8. 4. Levin LA, Beck RW, Joseph MP, Seiff S, Kraker R, International Optic Nerve Study Group. The treatment of traumatic optic neuropathy: the International Optic Nerve Trauma Study. Ophthalmology 1999;106:1268-77. 5. Steinsapir KD, Seiff SR, Goldberg RA. Traumatic optic neuropathy: where do we stand? Ophthal Plast Reconstr Surg 2002;18:232-4. 6. Lessell S. Indirect optic nerve trauma. Arch Ophthalmol 1989;107: 382-6. 7. Mauriello JA, DeLuca J, Krieger A, Schulder M, Frohman L. Management of traumatic optic neuropathy—a study of 23 patients. Br J Ophthalmol 1992;76:349-52. 8. Wang BH, Robertson BC, Girotto JA, Liem A, Miller NR, Ilift N, Manson PN. Traumatic optic neuropathy: a review of 61 patients. Plast Reconstr Surg 2001;107:1655-64. 9. Yip CC, Chng NW, Au Eong KG, Heng WJ, Lim TH, Lim WK. Low-dose intravenous methylprednisolone or conservative treatment in the management of traumatic optic neuropathy. Eur J Ophthalmol 2002;12:309-14. 10. Wohlrab TM, Maas S, de Carpentier JP. Surgical decompression in traumatic optic neuropathy. Acta Ophthalmol Scand 2002;80:28793. 11. Mahapatra AK, Tandon DA. Traumatic optic neuropathy in children: a prospective study. Pediatr Neurosurg 1993;19:34-9. 12. Thakar A, Mahapatra AK, Tandon DA. Delayed optic nerve decompression for indirect optic nerve injury. Laryngoscope 2003;113: 112-9. 13. Mahapatra AK. Optic nerve injury in children: a prospective study of 35 patients. J Neurosurg Sci 1992;36:79-84. 14. Shokunbi T, Agbeja A. Ocular complications of head injury in children. Childs Nerv Syst 1991;7:147-9. 15. Alford MA, Nerad JA, Carter KD. Predictive value of the initial quantified relative afferent pupillary defect in 19 consecutive patients with traumatic optic neuropathy. Ophthal Plast Reconstr Surg 2001; 17:323-7. 16. Seiff S. High dose corticosteroids for treatment of vision loss due to indirect injury to the optic nerve. Ophthalmic Surg 1990;21: 389-95.
Journal of AAPOS Volume 8 Number 1 February 2004
Goldenberg-Cohen, Miller, and Repka
17. Waga S, Kubo Y, Sakakura M. Transfrontal intradural microsurgical decompression for traumatic optic nerve injury. Acta Neurochir (Wien) 1988;91:42-6. 18. Mine S, Yamakami I, Yamaura A, Hanawa K, Ikejiri M, Mizota A, et al. Outcome of traumatic optic neuropathy. Comparison between surgical and nonsurgical treatment. Acta Neurochir 1999;141:27-30. 19. Levin LA, Joseph MP, Rizzo JF III, Lessell S. Optic canal decompression in indirect optic nerve trauma. Ophthalmology 1994;101: 566-9. 20. Brodsky MC, Wald KJ, Chen S, Weiter JJ. Protracted posttraumatic optic disc swelling. Ophthalmology 1995;102:1628-31.
27
21. Spoor TC, Hartel WC, Lensink DB, Wilkinson MJ. Treatment of traumatic optic neuropathy with corticosteroids. Am J Ophthalmol 1990;110:665-9. 22. Cook MW, Levin LA, Joseph MP, Pinczower EF. Traumatic optic neuropathy. A meta-analysis. Arch Otolaryngol Head Neck Surg 1996;122:389-92. 23. McNab AA. Orbital and optic nerve trauma. World J Surg 2001;25: 1084-8. 24. Steinsapir KD, Goldberg RA, Sinha S, Hovda DA. Methylprednisolone exacerbates axonal loss following optic nerve trauma in rats. Restor Neurol Neurosci 2000;17:157-63.
An Eye on the Arts – The Arts on the Eye
Dr. Graveline saved himself for the richest patients. The regulars got cut on every winter, and Rudy counted on their business. He reassured his surgical hypochondriacs that there was nothing abnormal about having a fifth, sixth, or seventh blepharoplasty in as many years. Does it make you feel better about yourself? Rudy would ask them. Then it’s worth it, isn’t it? Of course it is. Such a patient was Madeleine Margaret Wilhoit, age sixty-nine, of North Palm Beach. In the course of their acquaintance, there was scarcely a square inch of Madeleine’s substantial physique that Dr. Rudy Graveline had not altered. Whatever he did and whatever he charged, Madeleine was always delighted. And she always came back the next year for more. Though Madeleine’s face reminded Dr. Graveline in many ways of a camel, he was fond of her. She was the kind of steady patient that offshore trust funds are made of. On January fourth, buoyed by the warm sunny drive to Whispering Palms, Rudy Graveline set about the task of repairing for the fifth, sixth, or seventh time (he couldn’t remember exactly) the upper eyelids of Madeleine Margaret Wilhoit. Given the dromedarian texture of the woman’s skin, the mission was doomed and Rudy knew it. Any cosmetic improvement would have to take place exclusively in Madeleine’s imagination, but Rudy (knowing she would be ecstatic) pressed on. —Carl Hiaasen (from Skin Tight)
T
within and around the nerve then develop. In addition, trauma may fracture the optic canal or the anterior clinoid processes and produce bone fragments that impinge on the optic nerve. The management of TON is controversial.1,4,5 Various options include high-dose6-8 or low-dose9 corticosteroids, immediate decompression of the canalicular portion of the optic nerve,10 decompression of the canalicular optic nerve after a course of systemic corticosteroids,11,12 observation, and optic nerve sheath fenestration in cases of anterior TON.3 Unfortunately, findings in published reports have been inadequate to determine which, if any, of these treatments improves visual outcome after TON, mainly because of retrospective design, small numbers of matched cases, treatment bias, and unmasked outcomes. The only prospective, controlled, randomized trial—the International Optic Nerve Trauma Study—was abandoned because of insufficient recruitment.4 Only a few studies of TON have focused on children and adolescents.11,13 The purpose of this report is to describe our experience with the presentation and outcome of TON in children and adolescents.
PATIENTS AND METHODS a
b
c
From the Departments of Neurology, Neurosurgery, and Pediatrics, Wilmer Ophthalmological Institute, Johns Hopkins University School of Medicine,d and Johns Hopkins University School of Medicine, Baltimore, MD. Supported by the Isabel and Zanvyl Krieger Fund (NGC), Baltimore, MD; and in part by the Richard Baks Fellowship Fund, Baltimore, MD; the Stewart Wolff Fellowship Fund, Baltimore, MD; and the Judith and Paul Romano Binocular Vision and Strabismus Fellowship Endowment Fund, Baltimore, MD. Submitted April 2, 2003. Revision accepted August 22, 2003. Reprint requests: Michael X. Repka, MD, 233 Wilmer, Johns Hopkins Hospital, 600 N Wolfe St, Baltimore, MD, 21287-9028. Copyright © 2004 by the American Association for Pediatric Ophthalmology and Strabismus. 1091-8531/2004/$35.00 ⫹ 0 doi:10.1016/j.jaapos.2003.08.009
20
February 2004
A retrospective chart review of the investigators’ diagnostic databases was used to identify all patients with TON seen between 1973 and 2003. The diagnosis of TON in the alert child was based on signs of optic neuropathy: reduced acuity and color vision, relative afferent pupillary defect (RAPD), visual field defect, and light or red color desaturation. The diagnosis in an uncooperative child was based on the presence of an RAPD along with supportive findings on neuroimaging scan and fundus examination. Eligible patients had to have been ⬍18 years old at the time of their injury and examined by one of us on at least Journal of AAPOS
Journal of AAPOS Volume 8 Number 1 February 2004
Goldenberg-Cohen, Miller, and Repka
21
Table 1 Demographic and clinical characteristics
Characteristic No. of patients Mean age in years (SD) Median age (yr) Male (%) Female (%) Type of injury (n) Vehicle/bicycle accident Sports* Fall from bridge Gunshot Other/unknown State of consciousness Coma (%) Altered conscious state (%) Normal (%) Unknown (%) Time to medical care Immediate (%) Within 1 day (%) Within 2 weeks (%) Uncertain (%) Site of initial care (%) Johns Hopkins Hospital Other hospital in United States Outside United States Baseline visual acuity (PD) Unable to test† Unknown NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Canal fracture on imaging Missed on initial scan
Optic Nerve Decompression Surgery 4 8.8 (5.0) 10.0 3 (75) 1 (25)
Total
Observation
High-Dose Steroids
40 11.6 (4.2) 12.0 24 (60) 16 (40)
18 11.3 (4.4) 12.0 9 (50) 9 (50)
18 12.5 (3.7) 13.5 12 (67) 6 (33)
25 9 1 2 3
14 1 0 0 3
8 8 0 2 0
3 0 1 0 0
10 (25) 4 (10) 18 (45) 8 (20)
6 3 6 3
4 0 12 2
0 1 0 3
29 (73) 3 (8) 1 (2) 7 (18)
11 0 1 6
14 3 0 1
4 0 0 0
13 (32.5) 21 (52.5) 6 (15)
6 9 3
3 12 3
4 0 0
6 7 12 3 10 2 0 11 4
5 5 3 1 4 0 0 5 2
1 2 5 2 6 2 0 3 2
0 0 4 0 0 0 0 3 0
CF ⫽ counting fingers; NLP ⫽ no light perception. *Baseball (n ⫽ 2), soccer (n ⫽ 11), softball (n ⫽ 1), golf (n ⫽ 1), hockey (n ⫽ 1), rollerblade (n ⫽ 1), running (n ⫽ 1), and sledding (n ⫽ 1). †Unconscious (n ⫽ 5) and swollen eyelids (n ⫽ 1).
one occasion. Patients treated before our examination were included only if pretreatment examination findings were available. Patients with a ruptured globe (n ⫽ 1), shaken-baby syndrome (n ⫽ 3), or no recorded visual acuity at outcome (n ⫽ 2) were excluded. Data tabulated included the age, gender, mechanism of injury, visual acuity at diagnosis and at outcome, presence of an RAPD, treatment (type and timing), presence of fracture or intrasheath hemorrhage on neuroimaging scan, presence and location of facial and orbital fractures, reconstruction surgeries if performed, visual acuity at most recent visit, and length of follow-up.
RESULTS We identified 40 patients (24 boys and 16 girls) with 43 affected eyes (mean ⫾ SD ⫽ 11.6 ⫾ 4.2 years; median ⫽ 12.0; range, 2 to 18 years). The important demographic and clinical characteristics are listed in Table 1. Initial
medical care was provided at John Hopkins Hospital for 13 children (32.5%), elsewhere in the United States for 21 (52.5%), and outside the United States for 6 (15%). The diagnosis was made within 24 hours of the injury for 32 of 33 patients for whom such data were available. The right eye was affected in 15, the left in 22, and both in 3 children. The 3 children with bilateral TON sustained their injuries in motor vehicle accidents (MVAs). Two of these children had no light perception (NLP) in one eye at presentation. The mechanism of TON was blunt facial trauma in 31 (78%) and penetrating orbital trauma in 9 patients (22%) (Table 1). The most common etiologies were an MVA in 23 (57.5%) and a sports injury in 9 (22.5%) patients. All of the MVA victims experienced blunt trauma. Two were drivers, 13 were passengers (11 in the back seat and 2 in the front seat), 1 was riding a motorcycle, 2 were riding bicycles, and 5 were pedestrians. Cases of penetrating
22
Journal of AAPOS Volume 8 Number 1 February 2004
Goldenberg-Cohen, Miller, and Repka
Table 2 Final visual acuity and optic nerve atrophy measurements for all patients
Measurement Final visual acuity NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Optic nerve atrophy
Total (n ⴝ 40)
Observation (n ⴝ 18)
High-Dose Steroids (n ⴝ 18)
Optic Nerve Decompression Surgery (n ⴝ 4)
12 1 13 3 11 40
6 0 4 0 8 18
5 1 7 2 3 18
1 0 2 1 0 4
CF ⫽ counting fingers; NLP ⫽ no light perception.
orbital injury were caused by an antenna, stick (n ⫽ 2), tree branch, screwdriver, pencil, finger, and gunshot (n ⫽ 2). Visual acuity within 3 days of the trauma was known for 27 patients (29 eyes) (68%) (Table 1). Twelve (44%) had NLP, 3 (11%) had LP, 10 (37%) had counting fingers (CF) to ⬍ 20/200, and 2 (7.4%) had visual acuity 20/200 to 20/80. Baseline acuity after injury was not known for 13 children, 5 of whom were in a coma, 1 of whom had severely swollen eyelids, and 7 of whom had been referred more than 3 days after the injury with incomplete data regarding vision at presentation. An RAPD was detected in 14 of the 20 children for whom the presence of an RAPD was specifically recorded at the initial examination. Among the other 6 patients, an RAPD developed within 3 days after the injury. The 3 other children had evidence of bilateral optic neuropathy. Twenty-two children had evidence of a retrobulbar optic neuropathy, with normal-appearing optic discs, whereas 2 had swollen discs. Twelve had evidence of intraocular trauma as well as an optic neuropathy. Retinal hemorrhages were seen in 6 patients, 3 with orbital fracture, 1 with choroidal rupture, and 1 with optic disc swelling. Other intraocular findings included single cases of hyphema, commotio retinae, retinal tear, and macular hole. Fundoscopy could not be performed in 12 of the patients at the time of the injury, mainly because of severe eyelid swelling from associated facial fractures. Computed axial tomography (CAT) scanning was performed at the time of the trauma on 34 patients. Plain radiographs only were taken of 2 patients. Findings from 7 of the CAT scans showed normal results, whereas findings from 27 showed abnormalities related to the trauma. Fourteen patients had evidence of an orbital fracture on neuroimaging. The orbital fractures involved the floor (n ⫽ 6), posterior orbit (n ⫽ 4), and orbital rim (n ⫽ 4). Optic nerve sheath hemorrhage was seen in 1, optic nerve swelling in 6, and optic canal fractures in 11 patients. Other imaging findings included skull fractures in 4, orbital foreign bodies in 3, and nonorbital facial bone fractures in 7 patients. Associated brain injuries seen on CAT scan findings were subdural hematoma (n ⫽ 3), temporal lobe infarct (n ⫽ 1), and intraventricular hemorrhage (n ⫽ 1). Two patients were also studied using magnetic reso-
nance imaging. The results of one study were normal, whereas the other study showed changes consistent with a hemorrhage within the sheath of the orbital portion of the optic nerve. Neurologic Status Ten (25%) children were in a coma after their trauma, 4 (10%) had altered consciousness, and 18 (45%) were alert and oriented. No data were available for 8 patients (20%) (Table 1). Additional cranial neuropathies were found in 9 patients; most had multiple cranial nerve pareses. Six children had an oculomotor nerve paresis, 4 had an abducens nerve paresis, 2 had a trigeminal neuropathy, 2 had a facial nerve paresis, and 1 had a trochlear nerve paresis. Treatment and Visual Outcome for Entire Cohort Twenty-two patients were treated for their TON. Treatment included high-dose corticosteroids in doses ranging from 1 mg/kg/d to 1,000 mg/d prednisone equivalent (n ⫽ 17), retrobulbar steroid injection (n ⫽ 1), optic canal decompression (n ⫽ 3), and optic nerve sheath fenestration (n ⫽ 1). The remaining 18 patients were not treated. Table 2 lists the visual acuity outcomes for the entire cohort of patients. Of the 40 patients, 11 (28%) had vision ⱖ 20/80 in the affected eye at the most recent follow-up. Visual Outcome for Subgroup of Children With Both Baseline and Follow-Up Visual Acuities An outcome analysis was performed on a subgroup of patients followed up for at least 1 month who had quantitative baseline and final visual acuity measurements. Twenty-two patients met these criteria (Table 3). Their mean time to follow-up was 8.6 months (median ⫽ 5 months; range, 1 month to 2.5 years). Nine of the 22 patients (41%) who had baseline visual acuity measurements had improved visual acuity at follow-up; 4 of the 9 had intraorbital penetrating trauma. Table 3 lists the outcomes for this group of patients. The outcomes are stratified by initial acuity as well as treatment. In brief, among the 7 patients who were not treated, 2 experienced improvement in visual acuity, 1 from counting fingers to 20/50 and another from 20/200 to 20/20. Four untreated patients had no change in vision, and 1 experienced dete-
Journal of AAPOS Volume 8 Number 1 February 2004
Goldenberg-Cohen, Miller, and Repka
23
Table 3 Visual acuity measurement at final visit by treatment group for patients with documented baseline acuity and at least 1 month of follow-up Optic Nerve Decompression Surgery Total Observation High-Dose Steroids (n ⴝ 4) Visual Acuity (n ⴝ 22) (n ⴝ 7) (n ⴝ 11) Baseline visual acuity NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Outcome visual acuity by baseline acuity Baseline: NLP NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Baseline: LP NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Baseline: ⬍ 20/200 to CF NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Baseline: ⬍ 20/80 to 20/200 NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80
11 2 7 2 0
3 0 3 1 0
4 2 4 1 0
4 0 0 0 0
11 7
3 3
4 3
4 1
3 1
0 0
1
2 1
2 1
0 0
2 1
0 0
1
0
1
0
0 7 2 1 2
0 3 1
0 4 1 1 2
0 0 0
2 2
2 1
0 1
0 0
2
1
1
0
0
CF ⫽ counting fingers; NLP ⫽ no light perception.
rioration to NLP. Of 15 patients in this subgroup who were treated, 11 were treated with systemic corticosteroids, and four were treated with immediate optic canal decompression. Among the 11 patients treated with corticosteroids, 4 experienced visual improvement, 4 had no change in vision, and 3 experienced deterioration in vision. All 4 of the patients treated surgically (3 with optic canal decompression and 1 with optic sheath fenestration) had NLP vision at presentation. Among these patients, 3 improved to 20/200, whereas one remained status NLP. Secondary analyses were performed including only those patients with follow-up intervals of ⬎ 1 week and ⬎ 3 months. Thirteen of 27 patients (48%) had improved visual acuity after ⱖ 1 week of follow-up, whereas 8 of 18 patients (44%) had improved visual acuity after ⱖ 3 months of follow-up. Outcome by Subgroups Based on Potential Prognostic Factors Blunt Versus Penetrating Trauma. Initial and final visual acuity measurements subdivided by type of injury are listed in Table 4. Improvement was noted in 4 of 8
children (50%) after penetrating injury, whereas 2 had deteriorated of acuity despite receiving high-dose steroid therapy. Improvement was noted in 9 of 19 (48%) patients after blunt trauma. Optic Canal Fracture. Eleven patients had neuroimaging evidence of an optic canal fracture (Table 5). Three of these patients were treated with optic canal decompression and 3 with corticosteroids. The remaining 5 patients were not treated (Table 4). The acuities at presentation were NLP in 7, LP in 2, and unknown in 2 patients. Acuity improved from NLP in 3 of 3 patients in the optic canal decompression surgery group to 4/200 and 5/200, respectively. One untreated patient spontaneously improved from NLP to 20/50. One of the 2 patients with visual acuity of NLP improved after steroid therapy to a final acuity of 20/200 PD. Four patients (36%) had a delay in diagnosis (2 patients were scanned only on referral, and 2 scans were reinterpreted on referral). Two of these children were treated with corticosteroids, and 2 children were not treated. Optic Nerve Swelling. Swelling of the intraorbital portion of the optic nerve was noted in 5 patients on CAT
24
Journal of AAPOS Volume 8 Number 1 February 2004
Goldenberg-Cohen, Miller, and Repka
Table 4 Type of injury and visual outcome* Penetrating Trauma (n ⴝ 8) Initial Acuity
Outcome Visual outcome NLP LP ⬎ 20/200 to CF ⬎ 20/80 to ⱖ 20/200 ⱖ 20/80 Status (%) Improved Stable Deteriorated
Blunt Trauma (n ⴝ 19) Final Acuity
2 1 3 2 0
Initial Acuity
3 0 3 0 2
Final Acuity
10 2 6 1 0
7 1 6 2 3
4 (50) 2 (25) 2 (25)
9 (47) 9 (47) 1 (5)
CF ⫽ counting fingers; NLP ⫽ no light perception. *Table includes only patients with initial and final visual acuity measurements.
Table 5 Optic nerve findings on imaging and initial/final acuity measurements Optic Canal Fracture (n ⴝ 11)
Optic Nerve Edema (n ⴝ 6) Visual Acuity NLP LP ⬍ 20/200 to CF ⬍ 20/80 to ⱖ 20/200 ⱖ 20/80 Unknown
Severed Optic Nerve (n ⴝ 1)
Orbital Foreign Body (n ⴝ 3)
Initial
Final
Initial
Final
Initial
Final
Initial
Final
2 1 3 0 0
3 0 2 1 0
7 0 2 0 0 2
5 0 3 2 1
0 0 1 0 0
1 0 0 0 0
1 0 2 0 0
1 0 2 0 0
CF ⫽ counting fingers; NLP ⫽ no light perception.
scan and in one patient on quantitative ultrasound (Table 4). Two of these patients had NLP in the affected eye at presentation, whereas one had LP vision in the affected eye, and three had acuity of counting fingers. Four of these five patients were treated with corticosteroids, after which two improved from CF to 3/200 and 20/400, respectively; two patients deteriorated from CF to and LP to NLP, respectively. The one patient who underwent optic nerve sheath fenestration improved from NLP to 20/200. The patient with evidence of optic nerve swelling who was not treated remained NLP at follow-up.
DISCUSSION This study identified 40 children with TON, the largest series of children reported to our knowledge. Most of these patients experienced immediate and severe visual loss at the time of injury. For patients with baseline acuities and follow-up ⱖ 1 month, acuity improved in 41%, but only 18% eventually regained ⱖ 20/80 vision. Only five of 40 patients had visual acuity ⱖ 20/50 in the affected eye at follow-up. Improvement appeared to be more likely when vision was ⱖ 20/200 at presentation, regardless of treatment. Conversely, patients with NLP acuity at presentation rarely experienced significant visual improvement, regardless of treatment. Two other case series6,11 included large numbers of children (Table 6). These stud-
Table 6 Pediatric traumatic optic neuropathy Findings Number of patients ⬍ 18 years Most common etiology (n)
Percent with NLP at presentation Treatment modalities Observation (%) High-dose steroids (%) Surgical decompression (%) Timing of decompression surgery Percent with VA improvement Visual outcome NLP ⱖ 20/80
Present Study
Lessell6
Mahapatra et al11
40
16
35
Motor vehicle Motor vehicle Fall (20), (23), bicycle (7), bicycle (6) motor vehicle (2), falls (1) falls (0) (12) 12 71
18 (45) 18 (45) 4 (10) ⬍9d
9 (56) 6 (38) 3 (19) Within 48 h
0 35 (100) 5 (14) After 3 wk
41
44
34
12 11
1 8
NA NA
NLP ⫽ no light perception.
ies found rates of improvement of 34 and 44% of children, respectively. For children who were not treated in this study, the rate of improvement was 29% compared with 36% for those treated with corticosteroids. These rates are lower than the 50% probability of improvement reported
Journal of AAPOS Volume 8 Number 1 February 2004
for adults.4 This lower rate of improvement in children with TON may be related to the more severe nature of the damage in children or possibly to our inability to identify milder cases. MVAs were the most common cause of TON in our series, accounting for 57.5% of the injuries. The patient was most often injured as a back seat passenger, likely because most children ride in the back seat. The injury was noted less often when the patient had been a driver, bicyclist, or pedestrian. Sports injuries (22.5%) were the second most common etiology in our patients. MVAs were also the most common cause in children in a report from Japan.14 Studies from India found falls to be the most common cause,11,13 whereas bicycle injuries were most common in an earlier report from the United States.6 TON in adults is also most often caused by MVAs.1,4 Diagnosing TON in children can be a challenge. Children may not notice the visual deficit, may not report it, or may not be able to perform acuity measurements. Three children in our series initially did not report their monocular blindness. In addition, a limited visual acuity examination by counting fingers performed at near in an accident ward failed to identify two patients who had partial loss of vision. These children were not correctly diagnosed until another evaluation was performed, in one case almost 6 months after initial presentation. Last, many children could not perform afferent system testing because of age, fear, or loss of consciousness. Because of the difficulties inherent in examining children after severe head trauma, we believe that the single most important clinical test of optic nerve function is the swinging flashlight test to identify a relative afferent pupillary defect. The presence of a RAPD in the absence of ophthalmoscopic evidence of retinal damage almost invariably indicates optic nerve dysfunction. Thus, pupillary light reflex testing and a swinging flashlight test, which can be performed whether or not a patient is conscious, should be performed for every child who experiences head trauma. In addition, a recent report found poorer visual outcomes in patients with TON who had a RAPD, making the test useful not only for diagnosis but also as a predictor of prognosis.15 In this study, children were treated with systemic corticosteroids (n ⫽ 18), surgery (n ⫽ 4), or observation (n ⫽ 18) based on best clinical judgment. Steroids were given systemically in high doses (⬎ 1mg/kg/d orally or intravenously) to 17 patients, whereas 1 patient received a retrobulbar corticosteroid injection. No significant differences in improvement or final acuity measurements between this treatment and no treatment were observed. Surgery was not performed in all of the patients with optic nerve swelling in this study because of a time delay in reaching our hospital or a delayed diagnosis during the initial evaluation. We examined several features of the condition for an association with prognosis. These included type of trauma,
Goldenberg-Cohen, Miller, and Repka
25
presence of an optic canal fracture, and poor vision at presentation. Although Wang et al8 noted a better visual outcome after blunt trauma than after penetrating trauma in their series of patients with TON (most of whom were adults), we found no difference in outcome between blunt and penetrating trauma in our series. Eleven of our patients (28%) had evidence of a fracture of the optic canal on imaging. This rate is between the 14% rate reported among children in India11 and the 44 % rate identified in adults by CAT imaging.16 All of these studies likely underestimate the actual incidence of optic canal fractures in patients with TON. For example, Wohlrab et al10 found that 75% of their patients with TON had fractures at surgical exploration. The importance of CAT scanning with respect to treatment and prognosis in patients with TON is controversial.1 Although it is reasonable to treat patients in whom CAT scanning shows a bone fragment impinging on the optic nerve, the presence or absence of a simple fracture of the optic canal does not seem to be predictive of prognosis, regardless of treatment. Nevertheless, the presence of a fracture of the optic canal on CAT scan was a poor prognostic sign in this series. Only 1 of 11 patients (9%) with evidence of an optic canal fracture on imaging had a final acuity ⱖ 20/80 compared with 10 of 24 patients (42%) without optic canal fractures who eventually achieved ⱖ 20/80. Severe initial visual loss also was a predictor of a poor visual outcome among our patients. Patients with NLP were unlikely to improve irrespective of management. Such an observation has been made by others.8,17,18 However, Levin at al,19 in a study of 31 surgically decompressed adults, found no difference in outcome based on preoperative visual acuity. Mahapatra and Tandon11 noted that the presence of a normal or even abnormal but detectable VEP was associated with a high chance of visual recovery, whereas an absent VEP at presentation was almost never associated with substantial improvement. Optic nerve swelling after blunt trauma has been associated with a favorable prognosis for visual recovery. Brodsky et al20 reported three patients who had partial recovery of vision. In this study, two patients had optic disc swelling. One was status NLP and improved to 20/200 after sheath fenestration, and the other improved from counting fingers to 20/400 after parenteral administration of steroids. As noted above, there is no consensus as to the optimum treatment for TON. The results of several retrospective case series8,21,22 suggest a beneficial effect of systemic corticosteroids, whereas others3,8,10-12 found improved visual outcome after surgical therapy. Still other series1,6,9,19,23 have failed to demonstrate any clear-cut benefit of treatment compared with observation alone. The largest prospective comparison study of patients of all ages with TON found no difference in outcome among patients treated with surgical decompression, high-dose steroids, or observation.4 In that study, approximately
26
Goldenberg-Cohen, Miller, and Repka
50% of the patients in the steroid and observation groups improved by three or more lines of acuity, whereas 38% improved in a similar fashion in the surgically treated group. These investigators concluded that the medical evidence did not support one particular approach and that clinicians should continue “to decide to treat or not treat on an individual patient basis.”19 The results of our study support this recommendation because we were unable to identify any treatment that was more likely to result in visual improvement than observation alone. Recently, Steinsapir et al5,24 found that methylprednisolone exacerbated axonal loss after optic nerve crush injury in a rodent model. Whether the findings of these studies are applicable to humans is unclear. Human studies have not shown a deleterious effect of high-dose steroids,4 nor is it certain that a rodent optic nerve crush model is representative of human injury. However, this work raises an additional consideration for the clinician considering high-dose steroid therapy for an adult or child with TON. The findings of our study are subject to a number of important biases. Visual acuity before the injury was not known for any of the patients, so it is possible that some patients with evidence of an optic neuropathy after trauma had pre-existing optic nerve damage. A second bias is the possibility that cases of minor optic nerve injury were not identified in children because of the failure of such children to notice or report mildly decreased vision. Furthermore, patients with milder cases of TON might have been less likely to be referred for evaluation and treatment by us. These sampling biases would increase the proportion of NLP cases we have reported. Because NLP may be a predictor of a poor prognosis as identified in this report and in the report by Wang et al,8 these ascertainment biases would make the outcome for childhood TON appear to be worse than it actually is. Other biases include inability to obtain initial visual acuities in all patients and a lack of standardized method of visual acuity measurement because of the retrospective nature of the study. In addition, although we believe our study to include the largest number of children with TON to date, the number of patients in this study nevertheless is relatively small, making it impossible to detect small differences in treatment outcomes. The retrospective nature of the study means that treatments were selected at the discretion of the treating physician and that the timing of treatment was not part of any defined protocol. In addition, there likely was a bias toward treatment, especially in patients who experienced progressive loss of vision after injury, thus potentially making interventions appear to be less successful. The lack of uniform follow-up examination beyond 1 month might have underestimated the visual acuity outcomes for those patients with shorter follow-up because of late recovery, thus making the outcome for TON appear to be worse than it actually was. However, in secondary analyses we found no difference in the rates of improve-
Journal of AAPOS Volume 8 Number 1 February 2004
ment with minimum follow-up intervals of 1 week, 1 month, or 3 months. In conclusion, TON in childhood occurs most often from blunt trauma under circumstances similar to those in adults and typically results in visual loss that rarely improves significantly, regardless of treatment. The diagnosis may be difficult to establish but may be aided by a careful evaluation of pupillary responses. Factors suggestive of a worse visual outcome include NLP acuity at presentation, bone fracture(s) in the optic canal or posterior orbit, and evidence of intraorbital optic nerve swelling. As in the case of adults, clinicians should individualize treatment for children with TON by recognizing that spontaneous recovery may occur and that all treatments have potential risks.
References 1. Steinsapir KD, Goldberg RA. Traumatic optic neuropathy. Surv Ophthalmol 1994;38:487-518. 2. Steinsapir KD, Goldberg R. Traumatic optic neuropathies. In: Miller NR, Newman NJ, editors. Walsh and Hoyt’s Clinical NeuroOphthalmology. 5th ed. Baltimore (MD): Williams and Wilkins; 1998. p. v. 1. 3. Lu¨bben B, Stoll W, Grenzebach U. Optic nerve decompression in the comatose and conscious patients after trauma. Laryngoscope 2001;111:320-8. 4. Levin LA, Beck RW, Joseph MP, Seiff S, Kraker R, International Optic Nerve Study Group. The treatment of traumatic optic neuropathy: the International Optic Nerve Trauma Study. Ophthalmology 1999;106:1268-77. 5. Steinsapir KD, Seiff SR, Goldberg RA. Traumatic optic neuropathy: where do we stand? Ophthal Plast Reconstr Surg 2002;18:232-4. 6. Lessell S. Indirect optic nerve trauma. Arch Ophthalmol 1989;107: 382-6. 7. Mauriello JA, DeLuca J, Krieger A, Schulder M, Frohman L. Management of traumatic optic neuropathy—a study of 23 patients. Br J Ophthalmol 1992;76:349-52. 8. Wang BH, Robertson BC, Girotto JA, Liem A, Miller NR, Ilift N, Manson PN. Traumatic optic neuropathy: a review of 61 patients. Plast Reconstr Surg 2001;107:1655-64. 9. Yip CC, Chng NW, Au Eong KG, Heng WJ, Lim TH, Lim WK. Low-dose intravenous methylprednisolone or conservative treatment in the management of traumatic optic neuropathy. Eur J Ophthalmol 2002;12:309-14. 10. Wohlrab TM, Maas S, de Carpentier JP. Surgical decompression in traumatic optic neuropathy. Acta Ophthalmol Scand 2002;80:28793. 11. Mahapatra AK, Tandon DA. Traumatic optic neuropathy in children: a prospective study. Pediatr Neurosurg 1993;19:34-9. 12. Thakar A, Mahapatra AK, Tandon DA. Delayed optic nerve decompression for indirect optic nerve injury. Laryngoscope 2003;113: 112-9. 13. Mahapatra AK. Optic nerve injury in children: a prospective study of 35 patients. J Neurosurg Sci 1992;36:79-84. 14. Shokunbi T, Agbeja A. Ocular complications of head injury in children. Childs Nerv Syst 1991;7:147-9. 15. Alford MA, Nerad JA, Carter KD. Predictive value of the initial quantified relative afferent pupillary defect in 19 consecutive patients with traumatic optic neuropathy. Ophthal Plast Reconstr Surg 2001; 17:323-7. 16. Seiff S. High dose corticosteroids for treatment of vision loss due to indirect injury to the optic nerve. Ophthalmic Surg 1990;21: 389-95.
Journal of AAPOS Volume 8 Number 1 February 2004
Goldenberg-Cohen, Miller, and Repka
17. Waga S, Kubo Y, Sakakura M. Transfrontal intradural microsurgical decompression for traumatic optic nerve injury. Acta Neurochir (Wien) 1988;91:42-6. 18. Mine S, Yamakami I, Yamaura A, Hanawa K, Ikejiri M, Mizota A, et al. Outcome of traumatic optic neuropathy. Comparison between surgical and nonsurgical treatment. Acta Neurochir 1999;141:27-30. 19. Levin LA, Joseph MP, Rizzo JF III, Lessell S. Optic canal decompression in indirect optic nerve trauma. Ophthalmology 1994;101: 566-9. 20. Brodsky MC, Wald KJ, Chen S, Weiter JJ. Protracted posttraumatic optic disc swelling. Ophthalmology 1995;102:1628-31.
27
21. Spoor TC, Hartel WC, Lensink DB, Wilkinson MJ. Treatment of traumatic optic neuropathy with corticosteroids. Am J Ophthalmol 1990;110:665-9. 22. Cook MW, Levin LA, Joseph MP, Pinczower EF. Traumatic optic neuropathy. A meta-analysis. Arch Otolaryngol Head Neck Surg 1996;122:389-92. 23. McNab AA. Orbital and optic nerve trauma. World J Surg 2001;25: 1084-8. 24. Steinsapir KD, Goldberg RA, Sinha S, Hovda DA. Methylprednisolone exacerbates axonal loss following optic nerve trauma in rats. Restor Neurol Neurosci 2000;17:157-63.
An Eye on the Arts – The Arts on the Eye
Dr. Graveline saved himself for the richest patients. The regulars got cut on every winter, and Rudy counted on their business. He reassured his surgical hypochondriacs that there was nothing abnormal about having a fifth, sixth, or seventh blepharoplasty in as many years. Does it make you feel better about yourself? Rudy would ask them. Then it’s worth it, isn’t it? Of course it is. Such a patient was Madeleine Margaret Wilhoit, age sixty-nine, of North Palm Beach. In the course of their acquaintance, there was scarcely a square inch of Madeleine’s substantial physique that Dr. Rudy Graveline had not altered. Whatever he did and whatever he charged, Madeleine was always delighted. And she always came back the next year for more. Though Madeleine’s face reminded Dr. Graveline in many ways of a camel, he was fond of her. She was the kind of steady patient that offshore trust funds are made of. On January fourth, buoyed by the warm sunny drive to Whispering Palms, Rudy Graveline set about the task of repairing for the fifth, sixth, or seventh time (he couldn’t remember exactly) the upper eyelids of Madeleine Margaret Wilhoit. Given the dromedarian texture of the woman’s skin, the mission was doomed and Rudy knew it. Any cosmetic improvement would have to take place exclusively in Madeleine’s imagination, but Rudy (knowing she would be ecstatic) pressed on. —Carl Hiaasen (from Skin Tight)