Traumatic Optic Neuropathy Management: A Survey Assessment of Current Practice Patterns.

INTRODUCTION: The treatment of traumatic optic neuropathy (TON) is highly controversial with a lack of substantiated evidence to support the use of corticosteroids or surgical decompression of the optic nerve. The aim of the study was to determine if there was a general consensus in the management of TON despite controversy in the literature.
METHODS: An anonymous survey of members of the American Society of Ophthalmic Plastic and Reconstructive Surgery and the North American Neuro-Ophthalmology Society regarding their practice patterns in the management of patients with TON was performed.
RESULTS: The majority of 165 respondents indicated that they treated TON with corticosteroids (60%) while a significant minority (23%) performed surgical interventions (P < 0.0001). Subgroup analysis comparing rates of treatment with steroids among oculoplastic surgeons and neuro-ophthalmologists (67% vs. 47%) was not significant (Fisher's Exact test [FET], P =0.11) while results did suggest that a higher proportion of oculoplastic surgeons (33%) than neuro-ophthalmologists (11%) recommended surgical intervention (FET, P =0.004). In cases where visual acuity exhibited a downward trend treatment with steroids was the most commonly employed management. In general, neuro-ophthalmologists trended toward observation over treatment in TON patients with stable visual acuity while oculoplastic surgeons favored treatment with corticosteroids.
CONCLUSIONS: In spite of the lack of class I evidence supporting intervention of TON, the majority of respondents were inclined to offer corticosteroid treatment to patients whose visual acuity showed progressive decline following injury. Copyright:

INTRODUCTION

Traumatic optic neuropathy (TON) is rare in the general population with a reported incidence of approximately 1 per million.[1] However, young males are known to be at increased risk and patients who suffer craniofacial trauma with midfacial fractures have up to a 3% incidence of TON in historical case series.[23] Among recent combat veterans with orbital fractures the rate of TON may be as high as 50%.[4] The visual morbidity of TON ranges from mild to permanent and severe visual disability; estimates of the spontaneous recovery rate are variable, ranging from 20% to as high as 57% of patients.[1]

The treatment of TON remains controversial. The basic treatment options for TON include observation, high or mega-dose corticosteroids, or surgical decompression with or without concomitant steroids. The use of “megadose” systemic corticosteroids was reported by small case series in the early 1980s and the practice was lent support by the publication of the National Acute Spinal Cord Injury Studies which showed a benefit from early steroid treatment in those with spinal cord injury.[5678] Positive results from surgical decompression of the optic canal through transfrontal and transethmoidal approaches were also reported in several early studies.[91011]

Contrary to these earlier trials the landmark International Optic Nerve Trauma Study reported no significant benefit for treatment with either corticosteroids or surgical decompression when compared to observation in an observational study of 133 TON patients.[12] Compounding this lack of clarity, a recent randomized controlled trial comparing corticosteroid treatment with surgical decompression failed to demonstrate a significant difference in recovery of visual acuity between the two intervention groups.[13]

Concern regarding the potential for harm with these treatment options has also been raised. Most notably, the Corticosteroid Randomization After Significant Head injury (CRASH) trial reported an increase in mortality and serious disability in patients with traumatic brain injury treated with corticosteroids.[1415] Investigators from the Bascom Palmer Eye Institute performed transphenoidal medial optic canal decompression on cadavers and microscopic examination of the optic nerve afterwards demonstrated significant iatrogenic damage to the dura and optic nerve.[16] In a Cochrane Database review of randomized controlled trials of TON patients, very few studies met inclusion criteria, and the authors found no clear evidence of benefit from surgery or the use of corticosteroids compared to observation in the treatment for TON.[1718] Some authors have continued to recommended treatment with either corticosteroids or decompression based on the results of meta- or sub-group analyses of retrospective reviews.[1920] Other have investigated alternative therapies and recently a role for intravenous erythropoietin has been suggested.[21]

Objectives

Conflicting expert opinion and a lack of consistent evidence of benefit for specific TON treatments leaves physicians without guidance resulting in variation in clinical practice among practitioners and institutions. We report the results of a survey of practitioners involved in the management of TON patients to elucidate whether there is a consensus on treatment in spite of conflicting and sparse class I evidence.

METHODS

Study design

Institutional review board approval was obtained to anonymously survey providers involved in the management of TON regarding the clinical features used to determine patient management and the specific treatments employed. Participants did not receive compensation for their participation in the survey. Survey questions were developed by a multidisciplinary committee comprised of otolaryngologists and ophthalmologists [Supplemental 1]. Demographic data regarding location, years in practice as well as specialty/sub-specialty training were collected from respondents. The usage of common imaging modalities (computed tomography [CT] and magnetic resonance imaging [MRI]) and the relative importance of clinical features, such as the patient's initial visual acuity or pupillary exam, were queried. Respondents were asked to state their management strategies at varying levels of visual acuity as well as in the setting of specific imaging findings. Respondents were also asked to specify surgical approaches and steroid dosages utilized because heterogeneity in treatment strategy has been suggested as a factor leading to contradictory results in the literature.[19] The range of steroid doses (in equivalents of milligrams of methylprednisolone) used to query respondents was selected to conform to those used by Levin et al.[12]

Study population

Surveys were submitted to the mailing lists of ophthalmologic specialty societies whose members could reasonably be expected to be involved in the management and evaluation of TON (American Society of Ophthalmic Plastic and Reconstructive Surgery [ASOPRS] and the North American Neuro-Ophthalmology Society [NANOS]). Responses were collected from 4/15/19 to 6/20/19 before the survey was closed and analysis began. Respondents that indicated they were not involved in the evaluation and treatment of TON were not included in the final analysis.

Analysis/statistical methods

The analysis of responses was carried out utilizing an omnibus test (Friedman test for ranked responses, Chi-square for categorical responses) followed by post-hoc analyses for significant results. The Fisher's Exact test (FET) was used in place of Chi-square testing when the number of expected results was <5 in 20% of cases as proposed by Cochran.[22] A generalization of FET was employed as implemented by Kirkman for survey questions involving more than two independent or dependent variables (i.e., analyses requiring a >2 × 2 contingency table).[2324] Post hoc analyses comparing raw number of responses were performed with FET. In the case of Friedman testing, Nemenyi testing was used to perform post-hoc analysis. Post hoc analysis comparing proportions of responses from each subspecialty in a pair-wise manner were compared using the z-test for proportions of two populations. A standard significance level of 5% was used for all hypothesis testing and in cases of multiple post-hoc tests with FET on the same family of data, the Bonferroni correction was applied to this significance level to minimize Type I error.[25] The statistical treatment was reviewed by UC Davis' Clinical and Translational Science Center's biostatistics department.

RESULTS

Participants

Survey responses were received from 165 unique participants. The response rate was 10.8% (105/971) from the ASOPRS membership mailing list and 7.6% (60/788) for NANOS. Survey questions and data from these respondents are summarized in Supplemental 1.

Descriptive data

Participants were asked to provide demographic information in order assess for correlation between practice patterns and factors such as geography, years in practice [Table 1]. Respondents overwhelmingly practiced at academic centers and level I trauma centers. On average respondents had been in practice for 13.1 years (weighted mean of binned responses). Participants were primarily ophthalmologists with a minority (8 of 165 respondents) identifying as nonophthalmologic specialties (otolaryngology, neurology or not specified) and the majority having undergone sub-specialty fellowship training [Table 2]. Individual responses demonstrated that CT, followed by MRI, were the single most routinely ordered imaging modalities for the evaluation of patients suspected of having TON [Figure 1]; orbital ultrasonography was routinely ordered by fewer than 5% respondents. Differences between specialties' imaging modality preferences (neuro-ophthalmology and oculoplastic surgeons) did not achieve statistical significance once the Bonferroni correction was applied (P >.003).

Table 1

Demographic features of survey respondents

Demographic Variable	Respondents
Society membership
ASOPRS	105
NANOS	60
Years in practice
0-5	43
6-10	24
11-15	12
16-20	23
>21	63
Practice setting
Academic	90
Semi-academic	49
Private	26
Trauma level
I	99
II	20
III	6
IV	5
V	5
NA	29

NA: Not applicable, ASOPRS: American Society of Ophthalmic Plastic and Reconstructive Surgery, NANOS: North American Neuro-Ophthalmology Society

Table 2

Sub-specialty training of survey respondents

Specialty Type	Respondents
Ophthalmology	157
Neuro-ophthalmology	53
Plastics	101
Other	1
None	1
Not specified	1
Neurosurgery	0
Neurology	4
Otolaryngology	1
Not specified	3
Total	165

Figure 1

Percent of specialty routinely ordering indicated combination of imaging modalities. Imaging combinations selected by zero respondents are omitted for clarity. No imaging modality demonstrated statistically significant inter-specialty differences

Main results

One hundred and twenty-nine of 165 respondents ranked all choices in survey question #8 regarding their ranking of clinical factors in their assessment of TON patients [Figure 2]. Presenting visual acuity (mean rank 2.04), pupillary examination (2.29) and the trend in visual acuity (2.53) were all significantly more important than imaging features (3.15) [Table 3].

Figure 2

Participant's ranked patients' initial visual acuity as the most important clinical factor influencing their decision making; differences between the trend in acuity and pupillary exam were not statistically significant but imaging factors were considered least important. 1– most important, 4– least important

Table 3

Post hoc analysis of respondent’s ranking of the importance of clinical features in traumatic optic neuropathy management

	Initial acuity	Acuity trend	Imaging feature	Pupillary exam
Initial acuity	-
Acuity trend	0.018	-
Imaging feature	<0.001	<0.001	-
Pupillary exam	0.44	0.47	<0.001	-

Values of P<0.05 are considered statistically significant

As a group, respondents were more likely to treat TON with corticosteroids rather than surgery (60% vs. 23%; P <.0001). Respondents practicing for fewer than 5 years were more likely to observe patients with TON than those practicing for >5 years (55.8% vs. 25.4%) while variations in management across practice setting were minimal [Figure 3].

Figure 3

Variations in management of traumatic optic neuropathy patients between practitioners by clinical experience (left) and practice setting (right)

Subgroup analysis comparing rates of treatment with steroids amongst oculoplastic surgeons and neuro-ophthalmologists (67% vs. 47%) was not significant (FET, P =0.11) while results did suggest that a higher proportion of oculoplastic surgeons (33%) than neuro-ophthalmologists (11%) recommended surgical intervention (FET, P =0.004). In cases with stable visual acuity surgical decompression was significantly less common than treatment with steroids or observation at all initial visual acuities [Figure 4]. In cases where visual acuity exhibited a downward trend treatment with steroids was the most commonly employed management and observation was only more common than surgery in patients with better than 20/200 vision. The proportion of each subspecialty preferring to observe patients with stable visual acuity is exhibited in Figure 5; detailed subspecialty results of this type for the four scenarios queried in the survey are reported in Supplemental 2. In general neuro-ophthalmologists trended toward observation over medical or surgical treatment in patients with stable visual acuity while oculoplastic surgeons favored treatment with corticosteroids. In cases of declining visual acuity, the proportion of neuro-ophthalmologists who preferred to observe was reduced but remained significantly increased compared to oculoplastic surgeons; surgical intervention remained rare in this scenario. Surgical decompression was the most common management in cases where imaging was significant for a bony spicule or optic nerve hematoma [Figure 4]. However, in both of these scenarios, surgery was much less likely to be considered in patients whose vision was better than 20/200. In cases where visual acuity was >20/200 neuro-ophthalmologists were more likely to recommend surgical intervention than oculoplastic surgeons on subgroup analysis [Supplementals 2-4].

Figure 4

Management of traumatic optic neuropathy patients based on initial acuity (upper left), trend in visual acuity (upper right) and in the setting of an imaging finding of a bony spicule (bottom left) or hematoma (bottom right). Asterisks indicate visual acuities with overall statistically significant differences and points within these groups that do not meet significance with post-hoc testing are circled. NLP-No light perception, LP– Light perception, HM– Hand motion, CF– Count fingers

Figure 5

Sub-specialty specific proportion of respondents who reported observing patients with stable visual acuity grouped by sub-specialty and patient presenting visual acuity. *Indicate significant difference (P < 0.05) on two population proportion z-test. Neuro-Neuro-ophthalmology, Plastics– Oculoplastic surgeons

Respondents who indicated that they treated TON surgically were further queried regarding specific surgical approaches (endoscopic, craniofacial, and craniotomy) and overall the three approaches were equally likely to be offered (χ2 [df 2, n = 39] =1.11, P = 0.57). The type of approach selected was not significantly affected by respondent sub-specialty (FET, P = 0.61). The majority of respondents indicated that the latest period for surgical intervention is within 72 h of injury and this did not vary with sub-specialty (FET, P = 1.0). 85% of the surgery group treated with perioperative corticosteroids and this finding did not vary significantly between subspecialties (P = 0.60).

Respondents treating TON with systemic corticosteroids were asked what doses and duration of steroid therapy were employed [Figure 6a]. The majority (73%) of responses indicated that the initial dose of steroid fell in the range between 500 mg and 1,999 mg; 64% indicated this response for maintenance dosing. Neuro-ophthalmologists and oculoplastic surgeons reported similar initial steroid dosages (FET, P = 0.90) but differed in maintenance dosage (FET, P = 0.03) with oculoplastic surgeons prescribing higher maintenance dosages on average. Most respondents indicated that they treated for courses of 14 days or less (sub-specialty comparison, P = 0.07) and that they would not initiate steroids more than 3 days following the initial injury and this did not differ based on specialty (FET, P = 0.33) [Figure 6b-c].

Figure 6

(a) Initial and maintenance dose ranges. (b) Corticosteroid treatment duration. (c) Number of days post-injury after which steroids are no longer initiated

DISCUSSION

Key results/interpretation

The aim of the study was to determine if there was a general consensus in the management of TON despite at least three decades of controversy in the literature regarding the relative risk-benefit ratio of treatment.[2627] The collected responses add to the current understanding of the management of TON by highlighting the diversity of approaches and lack of strong consensus. There are a number of potential explanations for discordance between observed practice patterns and the evidence reported in the literature. While the potential morbidity of surgical intervention is fairly apparent, the risks of systemic corticosteroids may be less apparent. Participants may be unaware of studies suggesting a potential for mortality with corticosteroid treatment, such as the CRASH study, or animal studies showing a lack of effect and potential for optic nerve damage.[14282930] Alternatively, respondents may have assumed that patients did not have traumatic brain injury, and thus the findings of the CRASH study were not applicable, when answering survey questions as this factor was not specified in the question stem. This is not to say that there are no publications supporting treatment in the literature;[193132] a third explanation is that respondents are relying on these positive studies or applying anecdotal evidence from their clinical experience to select patients they believe will benefit from treatment.

Given the predilection for treatment exhibited by respondents these results may offer those PR actioners inclined to treat additional guidance with regard to practices considered reasonable by consensus opinion. Specifically, most respondents were inclined to offer treatment to patients whose visual acuity showed progressive decline following injury. Most often this treatment consisted of corticosteroid with dosages of 500–2000 mg methylprednisolone and durations of 7–14 days. Corticosteroid treatment was generally not offered for those patients presenting >3 days after their initial injury. Finally, surgical intervention was generally reserved for patients with specific indications, such as a culprit lesion on imaging, and in those patients with severely reduced vision (count fingers and worse). These guidelines are not based on patient outcomes as these could not be assessed in this anonymous and self-reported survey format but, as truly randomized studies have been historically difficult to perform due to low incidence and poor follow-up, they are worthy of consideration from clinical and medicolegal points of view.[31213]

Limitations

A limitation of the study is the representativeness of the results. Response rate was limited which is typical of most electronic surveys without incentives. The response rate is also limited due to the low incidence of TON cases. Reported practice patterns were based on the respondent's experience encountering patients with TON and confidence in its management and thus lends itself to selection and recall bias. Finally, the study was unable to correlate practice patterns with patient outcomes.

CONCLUSIONS

Controversy persists regarding the proper management of patients afflicted by TON; the results of this survey indicate that these differences in management are not an academic matter confined to the literature but are manifest in clinical practice. While the responses illustrate significant heterogeneity in the practice patterns of physicians of different specialties a sentiment in favor of offering treatment to patients with declining vision or unfavorable imaging features is clear. While intuitive, this practice may not take into account the potential for harm to patients treated with corticosteroids or decompression. These results do not demonstrate which practice patterns are most beneficial to patients but warrant further discussion, and formal investigation with randomized studies, to better define safe and effective evidence-based clinical practice.

Research quality and ethics statement

The authors of this manuscript declare that this scientific work complies with reporting quality, formatting and reproducibility guidelines set forth by the EQUATOR Network. The authors also attest that this clinical investigation was determined to require the Institutional Review Board/Ethics Committee review, and the corresponding protocol/approval number is IRB #1333824. We also certify that we have not plagiarized the contents in this submission and have done a Plagiarism Check.

Financial support and sponsorship

The project described was supported by the National Center for Advancing Translational Sciences (NCATS), National Institutes of Health (NIH), through grant #UL1 TR001860.

Conflicts of interest

There are no conflicts of interest.

Subspecialty trends in the management of traumatic optic neuropathy patients based on initial, stable, visual acuity. *Indicate significance of the two population proportion ztest between indicated groups at individual visual acuities. (a) Observation, (b) corticosteroids, (c) decompression

Subspecialty trends in the management of traumatic optic neuropathy patients based on initial, downtrending, visual acuity. *Indicate significance of the two population proportion ztest between indicated groups at individual visual acuities.(a) Observation, (b) Corticosteroids, (c) Decompression

Subspecialty trends in the management of traumatic optic neuropathy patients based on initial visual acuity in the setting of a bony spicule. * indicate significance of the two population proportion ztest between indicated groups at individual visual acuities. (a) Observation, (b) Corticosteroids, (c) Decompression

Subspecialty trends in the management of traumatic optic neuropathy patients based on initial visual acuity in the setting of a nerve sheath hematoma. *indicate significance of the two population proportion ztest between indicated groups at individual visual acuities. (a) Observation, (b) Corticosteroids, (c) Decompression