Optic Canal Decompression in Indirect Optic Nerve Trauma

Optic Canal Decompression in Indirect Optic Nerve Trauma Leonard A. Levin, MD, PhD,1 Michael P. Joseph, MD,2 Joseph F. Rizzo III, MD,1 Simmons Lessell, MD l Background: The proper management of neurogenic visual loss after blunt head trauma is controversial. Non-treatment, corticosteroids, and surgical decompression of the optic canal are all currently considered to be reasonable alternatives. The goal of this study was to identify factors affecting improvement in patients treated with canal decompression. Methods: A retrospective analysis of 31 cases in which transethmoidal decompression of the optic canal had been performed for neurogenic visual loss after closed head trauma was conducted. Each patient was alert and free of injury to the globe when evaluated before surgery. Surgery was performed within 6 days of injury, and all were given perioperative steroids. Results: Visual acuity improved in 22 (71%) patients, with 6 (19%) regaining visual acuity of 20/40 or better. The mean improvement from preoperative visual deficit was 42.0% ± 6.6%, with a median improvement of 45.2%. Both univariate and multivariate analysis suggested that vision improved more in patients who were younger than 40 years of age than in patients who were 40 years of age or older. Interval between injury and surgery, preoperative visual acuity, and the presence of optic canal fracture did not affect outcome. Conclusion: Any future randomized trials of therapy should stratify patients based on age. Enrollment of patients with no light perception or who experienced delay between injury and treatment may be reasonably considered. Ophthalmology 1994;101:566-569

The proper management of neurogenic visual loss after blunt head trauma is controversial.I'f Non-treatment, corticosteroids, and surgical decompression of the optic canal are all considered to be reasonable alternatives, but the issue is not apt to be resolved until a controlled, randomized trial is conducted. Pending such a trial, useful Originally received: July 16, 1993. Manuscript accepted: September 30, 1993. J Neuro-Ophthalmology Unit, Department of Ophthalmology, Massachusetts Eye and Ear Infirmary, Boston, Massachusetts. 2 Department of Otolaryngology, Massachusetts Eye and Ear Infirmary, Boston, Massachusetts. Presented in part as a poster at the American Academy of Ophthalmology Meeting, Dallas, November 1992. Supported in part by a Heed Foundation Fellowship (Dr. Levin). The authors have no proprietary interest in any research or materials presented in this article. Reprint requests to Leonard A, Levin, MD, PhD, F4j340 CSC, University of Wisconsin Hospital and Clinics, 600 Highland Avenue, Madison WI 53792-3220.

566

information may nevertheless be obtained from the outcome oftreatment in uncontrolled case series. We reported the use oftransethmoidal decompression of the intracanalicular segment of the nerve combined with perioperative corticosteroid treatment in 14 cases of indirect optic nerve trauma.? Thirty-one patients have now been treated in this fashion, and the expanded series allows analysis of factors within the group that influence improvement in visual acuity.

Patients and Methods The study group included 31 consecutive alert patients seen at the Massachusetts Eye and Ear Infirmary within 6 days of closed head trauma that resulted in unilateral neurogenic visual loss without evidence of injury to the globe. Neuro-ophthalmic examination and computed tomography of the head and orbit in the axial and coronal planes with soft tissue and bone windows were completed shortly after presentation. Clean, well-illuminated plastic

Levin et al

Optic Canal Decompression

Table 1. Acuity Conversion Chart

Table 2. Entry Characteristics of Patients

Visual Acuity

logMAR Units

Characteristics

No light perception Light perception Hand motions Counting fingers 20/800 20/400 20/200 20/100 20/80 20/60 20/50 20/40 20/30 20/25 20/20 20/15 20/13

-4.70 -3.70 -2.70 -1.40 -1.60 -1.30 -1.00 -0.70 -0.60 -0.48 -0.40 -0.30 -0.18 -0.10 0.00 0.12 0.19

Age (yrs) Mean ± standard deviation Range M:F Visual acuity No light perception Light perception to 20/300 20/200 or better Optic canal fracture Hours until surgery 0-24 25-72 73-144

wall charts were used for acuity testing before and after surgery. The physicians conducting the examinations were fully acquainted with the history of every patient, including visual acuity before surgery, at the time of the final examination. Patients received 8 to 10 mg of dexamethasone intravenously at the time of admission, during surgery, and every 8 hours during the first 24 to 48 hours after surgery. Four of the patients had received 1 gm of intravenous methylprednisolone a day for 1 to 3 days at other hospitals before referral. All surgery was performed by one surgeon (MPJ) under general anesthesia as previously described.' For the purpose ofanalysis, visual acuity was converted into 10gMAR (logarithm of the minimum angle of resolution) units, with values for no light perception, light perception, hand motions, and count fingers extended as previously described (Table 1).6.7 To eliminate a bias in favor of patients with worse acuity before surgery resulting from their greater potential for improvement, change in visual acuity was calculated as acuity units regained as a proportion of the number of units initially lost, assuming 16/13 (20/15) as an arbitrary baseline. We analyzed the effect of the following variables on visual acuity improvement with surgery: age, gender, hours until operation, use of high-dose steroids, presence of optic canal fracture (demonstrated radiologically or intraoperatively), and absence of light perception before surgery. The arcsin transformation was used to normalize the proportion of visual acuity regained after surgery." Univariate analysis of dichotomous factors was performed using Student's t test and the nonparametric Mann-Whitney U statistic. Two methods were used to study the contribution of confounding variables. The contribution of each of the clinical variables to improvement in visual acuity was studied with multiple linear regression. In ad-

No.

%

32.4 ± 18.4 13-74 25:6

(81%:19%)

15 10 6 10

48.4 32.2 19.3 32.3

10 9 12

32.3 29.0 38.7

dition, univariate analysis was performed after stratification of each group based on possible confounding variables, using two-way analysis of variance.

Results Table 2 displays the entry characteristics of the study patients. Except for the expected preponderance of male patients," there was a broad distribution of ages, preoperative visual acuity, and intervals between injury and surgery. Visual acuity improved in 22 (71 %) of the patients, with 6 (19%) regaining visual acuity of 20/40 or better. The mean improvement was 1.4 ± 0.3 10gMAR units. When expressed as percentage improvement from visual deficit before surgery, the mean improvement was 42.0% ± 6.6%, with a median improvement of 45.2%. Figure 1 shows the effect of certain variables on visual outcome. Vision improved significantly more in patients younger than 40 years of age (n = 20) than in those 40 years of

Fi------l

Overall ll11JrOYement • • • • • •

HOU"S:524 Hours>24

,====r~--t

j

Male Female

NoFractlnl Fractlnl

Ugh! Perceptioo or Better

'=====I=~

j

No Light Perception

20

40 60 % Improvement

80

Figure 1. Comparison of outcome (in percentage improvement) between various groupings of patients. See text for significance levels of cornparisons.

567

Volume 101, Number 3, March 1994

Ophthalmology

Table 3. Stratified Analysis of Age-related Differences in Improvement Older than 40 Yrs

Younger than 40Yrs

Unadjusted Canal fracture Yes No Adjusted Preoperative visual acuity No light perception Light perception or better Adjusted Interval to surgery ~24 hrs >24 hours Adjusted

No.

%±SEM

No.

%±SEM

20

52.6 ± 7.6

11

22.7 ± 10.3

6 14

63.1 ± 11.9 48.1 ± 9.6

4 7

7.3 ± 14.5 31.4 ± 13.6

10 10

43.1 ± 13.2 62.1 ± 7.8

5 6

13.5 ± 18.6 30.3 ± 10.1

9 11

54.8 ± 12.2 50.8 ± 10.3

1 10

0 24.9 ± 10.8

P

0.025

0.013

0.022

0.058

SEM = standard error of the mean.

age or older (n = 11; 52.6% ± 7.6% vs. 22.7% ± 10.3%; P = 0.025 by Mann-Whitney U test). Univariate analysis of other variables, specifically gender, presence of canal fracture, lack oflight perception before surgery, treatment with high-dose steroids, and interval before surgery of longer than 24 hours, did not show significant differences. The small number of female patients and of patients treated with high-dose steroids, however, gave the study insufficient power to detect clinically important differences. To analyze the possible confounding effect of certain variables on age-dependent response to surgery, a stratified analysis was performed (Table 3). Even when stratified by presence or absence ofcanal fracture and visual acuity before surgery, age was still significantly associated with outcome. Stratification by length of interval before surgery resulted in a less apparent effect of age on outcome. The interaction term in the analysis of variance, however, was not significant (P = 0.56), which is consistent with a lack of association between length of interval before surgery and the effect of age on outcome. Table

4.

To further examine the effects of covarying factors, a multiple linear regression was performed. All variables for which there were sufficient (n > 6) patients in each group were used to model percentage improvement in visual acuity (Table 4). Only age was a significant covariant in predicting worse visual outcome after surgery (P = 0.038). The multiple correlation coefficient of the model was 0.459, suggesting that factors other than those studied contributed to outcome. A similar regression using all of the variables yielded similar results (data not shown).

Discussion The data obtained from this retrospective analysis of 31 patients bears only peripherally on the issue of how posterior indirect optic nerve trauma should be managed. There was only one form of treatment, neither the patients nor the physicians performing the examinations were masked, and the techniques were not standardized. The physicians who performed the examinations, aware that

... Linear Regression Model of Visual Acuity Improvement"

Variable

Coefficient (sin" improvement)

Standard Error

Student's t

P

Age younger than 40 yrs No light perception preoperatively Optic canal fracture Surgery within 24 hrs Intercept

-0.200 0.096 -0.018 -0.007 -0.441

0.092 0.082 0.091 0.098

2.19 1.17 0.19 0.08

0.038 0.251 0.847 0.940

• The following transformed linear model was used: sin-1(improvement)

=

a, (age < 40) + a2 (NLP) + a, (fracture) + a, (hours < 24) + b

where the coefficients (a.) are from the table and the dichotomous variableshave value 1 if absent, -1 if present.

568

Levin et al . Optic Canal Decompression the patients had been treated, may have been biased toward obtaining better visual acuity after surgery, and patients may have benefited from the practice effect. There is no reason to suspect, however, that the design or execution of our study would not allow valid intragroup comparisons. In the univariate analysis, age proved to be a significant variable; on average, younger patients had a better visual outcome than older patients. This was true when results were stratified by visual acuity before surgery, length of time before surgery, and presence of optic canal fracture, suggesting that these variables were not significantly confounding. Similarly, multivariate analysis demonstrated a statistically significant effect of age. The model was not highly predictive of outcome overall, however, implying that other factors not studied may be relevant. If younger patients have a greater potential for recovery, the reason is inapparent. Interestingly, in the Optic Neuritis Treatment Trial, older patients with optic neuritis were less likely than younger patients to regain visual acuity of 20/20 (Beck RW, personal communication, March 1993). Some studies found that recovery from aphasia after stroke and mechanical head trauma is faster in younger patients than in older patients, although other studies did not. 10,1 I Results of animal experiments also suggest that remyelinization might not be as effective in the central nervous system of older animals.'? Univariate and multivariate analysis demonstrated that patients who were totally blind before surgery were not significantly less likely to improve than those with better vision before surgery. This suggests that patients with complete loss of vision at impact may have lesions that are merely more severe, rather than qualitatively different, than those in patients who retain some vision. Unfortunately, the absence of tissue for examination precludes testing ofthis hypothesis. Furthermore, the small number of patients studied makes the possibility of a type II error likely. Power calculations reveal that at least 52 patients would be necessary to detect a 10%difference in improvement, given the variance of the population outcome. Similarly, the effect of gender and treatment with steroids could not be analyzed because of the small number of female patients and ofpatients treated with high-dose steroids. Although the data from this study are not sufficient to formulate guidelines for the management of indirect optic nerve trauma, information is provided that may help determine whether optic nerve canal decompression, high-

dose corticosteroid treatment, or observation is most appropriate. The fact that the visual acuity of many of the patients who had no light perception before surgery did improve after treatment, although not as much as in patients with visual acuity of light perception or better, suggests that surgical intervention should not be denied solely because the patient has no light perception, as has been suggested." Patients in this study who underwent decompression could not be shown to have a statistically worse outcome, suggesting that this method of treatment could be reasonably considered within several days of trauma. Furthermore, decompression was performed in our patients within 6 days of injury, and therefore the data do not preclude the possibility that surgery might be of benefit after even longer intervals.

References I. Lessel1 S. Indirect optic nerve trauma. Arch Ophthalmol 1989; I07:382-6. 2. Joseph MP, Lessell S, Rizzo J, Momose KJ. Extracranial optic nerve decompression for traumatic optic neuropathy. Arch Ophthalmol 1990; 108: 1091-3. 3. Spoor TC, Hartel WC, Lensink DB, Wilkinson MJ. Treatment of traumatic optic neuropathy with corticosteroids. Am J Ophthalmol 1990;110:665-9. 4. Hughes B. Indirect injury of the optic nerves and chiasm. Bul1 Hopkins Hosp 1962;111:98-126. 5. Anderson RL, Panje WR, Gross CEo Optic nerve blindness fol1owing blunt forehead trauma. Ophthalmology 1982;89: 445-55. 6. Gottlob I, Fendick MG, Guo S, et al. Visual acuity measurements by swept spatial frequency visual-evoked-cortical potentials (VECPs): clinical application in children with various visual disorders. J Pediatr Ophthalmol Strabismus 1990;27:40-7. 7. Costa VP, Smith M, Spaeth GL, et al. Loss of visual acuity after trabeculectomy. Ophthalmology 1993;100:599-612. 8. Winer BJ. Statistical Principles in Experimental Design, 2nd ed. New York: McGraw-Hil1, 1971. 9. Kline LB, Morawetz RB, Swaid SN. Indirect injury of the optic nerve. Neurosurgery 1984;14:756-64. 10. Kertesz A. Recovery from aphasia. Adv Neurol 1984;42: 23-39. II. Reinvang I. The natural history of aphasia. Adv Neurol 1984;42: 13-22. 12. Ludwin SK. Chronic demyelination inhibits remyelination in the central nervous system. An analysis of contributing factors. Lab Invest 1980;43:382-7. 13. Walsh FB, Hoyt WF. Clinical Neuro-Ophthalmology, 3rd ed. Baltimore: Wil1iams & Wilkins, 1969.

569