Shibuya TY, Feinberg SM, Mathog RH, Li K, Kim SA, Stachler RJ, Yoo GH. Visual Risks of Facial Fracture Repair in the Setting of Traumatic Optic Neuropathy. Arch Otolaryngol Head Neck Surg. 2006;132(3):258-264. doi:10.1001/archotol.132.3.258
To identify whether facial fracture repair in patients with traumatic optic neuropathy results in visual deterioration.
A retrospective analysis was performed of all patients admitted from 1992 through 1997 with the diagnosis of facial fracture and traumatic optic neuropathy. Vision was recorded before and after fracture repair using logarithm of the minimum angle of resolution measurements. Visual outcome was compared with a nonsurgically treated group of patients with a similar diagnosis.
University trauma hospital.
A total of 700 medical charts were reviewed, and 54 patients met study criteria. All patients received megadose corticosteroid treatment and were divided into 3 groups: (1) facial fracture repair alone, (2) optic nerve decompression (OND) + facial fracture repair, or (3) nonsurgical treatment.
For the 16 patients in the fracture repair alone group, 12 (75%) had improved vision and 4 (25%) had no change postoperatively. For the 10 patients in the OND + fracture repair group, 3 (30%) had improved vision, 5 (50%) had no change, and 2 (20%) had worsened vision postoperatively. For the 28 patients in the nonsurgical group, 18 (64%) had improved vision, 9 (32%) had no change, and 1 (4%) had worsened vision by discharge. Facial fracture repair alone and the nonsurgical groups both demonstrated significant visual improvement by discharge. The amount of improvement was not significantly different between all 3 groups (facial fracture repair, 0.38 ± 0.40; OND + facial fracture repair; 0.32 ± 1.38; and nonsurgical, 0.69 ± 1.07).
Facial fracture repair in the setting of traumatic optic neuropathy had no adverse effect on vision. Patients requiring OND + fracture repair had a significantly worse visual prognosis.
Loss of vision after facial trauma is an uncommon complication, with a reported incidence of 2%-5%.1- 4 Factors affecting vision are vascular occlusion, orbital compartment syndrome, retrobulbar hemorrhage, and retinal detachment. Direct injuries to the optic nerve, optic tract, or central visual centers can result in loss of vision as well. Traumatic optic neuropathy (TON) as defined by Anton Nuhn in 1845 is a lesion of the optic nerve resulting in posttraumatic amaurosis.5
The optic nerve lies within the subarachnoid space covered by a dural sheath. It then enters the optic canal, a completely osseous channel within the lesser wing of the sphenoid bone. Within the optic canal, the nerve is adherent to bone and vulnerable to traumatic injury from either shearing or compression.6 Within the orbit, the nerve runs from the orbital apex in an S-shaped course through orbital fat to the posterior pole of the globe to innervate the retina.
The management of TON is controversial,5 with no standardized approach to the management of facial fractures in an individual with TON or blindness. Weymuller7 has previously conducted a poll in which he found wide variations in patient management. Responses ranged from absolutely no surgical repair or manipulation of the facial skeleton to willingness to repair fractures on only seeing eyes.
For the purpose of examining whether surgery of the facial skeleton in patients with TON resulted in visual deterioration, we retrospectively studied all patients admitted to Detroit Receiving Hospital, Detroit, Mich, between 1992 and 1997, with the diagnosis of both facial fracture and TON. Using the logarithm of the minimum angle of resolution (log MAR) (Table 1), we graded each patient's degree of TON using a standard ophthalmologic conversion measurement from the values of no light perception, light perception, hand motion, finger counting, and visual acuities from 20/800 down to 20/15.8,9 A comparison of visual outcomes was made between one group of patients undergoing facial fracture repair and another group of patients with similar injury but not undergoing repair. The purpose of this comparison was to determine whether surgery of facial fractures had any significant impact on visual outcome. In addition, results of individuals undergoing optic nerve decompression (OND) with facial fracture repair were compared with the groups of facial fracture repair alone group or those treated without surgery.
All individuals admitted to Detroit Receiving Hospital from January 1, 1992, to December 31, 1997, with the diagnosis of TON and facial fracture were included in this study. All data were obtained and handled in accordance with the policies of the Wayne State University School of Medicine (Detroit) on confidentiality and approved by the institutional review board. Seven hundred cases were reviewed and 54 patients met the study criteria. Patient demographics were recorded for age, sex, mechanism of injury, types of fractures, surgical repair, and days from injury to repair (Table 2 and Table 3). The management of TON for all individuals included megadose treatment with corticosteroids or megadose treatment with corticosteroids plus OND (Table 4). Megadose corticosteroid treatment consisted of a loading dose of methylprednisolone acetate at 30 mg/kg, followed by a 15-mg/kg dose 2 hours later and then repeated dosing every 6 hours.10
Using a standard ophthalmologic conversion method, we used the log MAR for comparison of data between groups.8,9 Visual acuities as measured from the Snellen chart, ranging from no light perception, light perception, hand motion, finger counting, and visual acuities from 20/800 down to 20/15, were converted to MAR using the conversion logarithm (Table 1).8,9 In our series, worsening of vision despite megadose corticosteroid treatment prompted OND by the Ophthalmology service. Patients with multiple trauma, an unstable medical condition, minimally displaced fracture, or refusal of surgery received megadose corticosteroid therapy without surgical intervention.
One group composed of patients with the diagnosis of TON and facial fracture (Table 2 and Table 4 [patients 1-16]) underwent surgical repair. Vision was recorded preoperatively and postoperatively (Table 4 [patients 1-16]). A second group composed of patients with deteriorating vision (Table 2 and Table 4 [patients 17-26]) underwent OND followed by facial fracture repair at the same setting. Visual acuity for these patients was recorded preoperatively and postoperatively (Table 4 [patients 17-26]). Visual outcome for both groups was then graded as improved (increase in log MAR value), no change (unchanged log MAR value), or worsened (decrease in log MAR value) (Table 4 [patients 1-26]) and compared between groups (Table 4 and Table 5).
The third group was composed of patients with the diagnosis of TON and facial fracture who were treated conservatively without surgery. All individuals had their visual acuity measured at the time of admission and at discharge (Table 3 and Table 4 [patients 27-54]). Visual outcomes again were graded as improved (increase in log MAR value), no change (no change in log MAR value), or worsened (decrease in log MAR value) (Table 4 [patients 27-54]). This group functioned as a control population of patients for comparison with patients who had surgery.
Individual groups were compared for differences preoperatively and postoperatively or at the time of admission and at the time of discharge using the Wilcoxon signed rank test. Groups were compared for significant differences using the Kruskal-Wallis test performed by SAS statistical software (SAS Institute Inc, Cary, NC). The Bonferroni method was used to adjust the P values for multiple comparisons.
Of the 54 patients who met the study criteria, 16 underwent surgical repair of facial fractures (Table 2, patients 1-16), 10 underwent facial fracture repair and OND (Table 2, patients 17-26), and 28 did not have surgery (Table 3). There were 7 women and 47 men (Table 2 and Table 3). The average patient age was 40.5 years (range, 25-73 years) (Table 2 and Table 3). The mechanisms of injury were assault in 26 patients (48%), motor vehicle accident in 19 patients (35%), gunshot wound in 8 patients (15%), and fall in 1 patient (2%) (Table 2 and Table 3).
Among the 26 patients in the surgically treated group, fractures included the orbit in 26 (100%), the nose in 17 (65%), the maxilla in 16 (62%), the zygomaticomaxillary complex in 13 (50%), the frontal sinus in 8 (31%), the naso-orbital-ethmoid in 5 (19%), and the mandible in 3 (11%), and were further classified as Le Fort III in 3 (11%), Le Fort I in 2 (8%), Le Fort II in 2 (8%), and basal skull in 16 (4%) (Table 2). There were 4 women and 22 men. The average patient age was 39.8 years (range, 25-57 years) (Table 2). As a group, the mechanisms of injury were assault in 11 patients (42%), motor vehicle accident in 11 patients (42%), and gunshot wound in 4 patients (15%) (Table 2).
The fractures in the 28 nonsurgically treated patients included the orbit in 27 (96%), the maxilla in 14 (50%), the zygomaticomaxillary complex in 12 (43%), the nose in 4 (14%), the naso-orbital-ethmoid in 4 (14%), the frontal sinus in 3 (11%), and the mandible in 1 (4%) (Table 3). There were no Le Fort or skull base injuries. There were 3 women and 25 men. The average patient age was 40.1 years (range, 25-73 years) (Table 3). As a group, the mechanisms of injury were assault in 15 patients (54%), motor vehicle accident in 8 (29%), gunshot wound in 4 patients (14%), and fall in 1 patient (Table 3).
All patients in this study were treated with megadose corticosteroids as a standard protocol recommended by the ophthalmology service. Only selected cases with worsening vision underwent OND as deemed necessary by the ophthalmology service (Table 4 [patients 17-26]). For clarity, all facial fractures were repaired after OND.
Of the 16 patients in the fracture repair only group, 12 (75%) demonstrated visual improvement, 4 (25%) had no change in vision, and none demonstrated worsening of vision. The mean preoperative vision score was −1.44 ± 1.67 log MAR units, the mean postoperative vision score was −1.07 ± 1.83 log MAR units, and the mean overall improvement was 0.38 ± 0.40 log MAR units (Table 4 [patients 1-16]). Vision did significantly improve postoperatively in this group compared with preoperative vision scores (P<.001, Wilcoxon singed rank test).
Of the 10 patients in the OND plus facial fracture repair group, 3 (30%) demonstrated visual improvement, 5 (50%) had no change, and 2 (20%) demonstrated worsening of vision. The mean preoperative vision score was −3.57 ± 1.62 log MAR units, the mean postoperative vision score was −3.25 ± 1.99 log MAR units, and the mean overall improvement was 0.32 ± 1.38 log MAR units (Table 4 [patients 17-26]). Vision did not significantly improve postoperatively in this group compared with preoperative vision scores (P = .88, Wilcoxon singed rank test).
In total, of the 26 patients in the surgically treated group with either facial fracture repair or facial fracture repair plus OND, 15 (58%) demonstrated improvement, 9 (35%) had no change, and 2 (8%) demonstrated worsening of vision overall. There was no significant difference in visual outcomes between this whole group and the nonsurgically treated group.
Of the 28 patients in the nonsurgically treated group, 18 (64%) demonstrated visual improvement, 9 (32%) had no change, and 1 (4%) demonstrated worsening of vision. The mean admission vision score was −1.76 ± 1.67 log MAR units, the mean discharge vision score was −1.08 ± 1.49 log MAR units, and the mean overall improvement was 0.69 ± 1.07 log MAR units (Table 4 [patients 27-54]). Interestingly, vision did significantly improve at the time of discharge in this group compared with the vision score at time of admission (P<.001, Wilcoxon singed rank test). There was no statistical difference in visual improvement among these 3 treatment groups (P = .30, Kruskal-Wallis test) (Table 5).
Management of TON is controversial. In a meta-analysis of 244 cases of TON from 46 journal articles, Cook et al5 found insufficient data to conclude whether treatment with corticosteroids, extracranial decompression, or both was the best treatment. They noted, however, that either treatment or both treatments was better than no treatment at all. The management of the facial fractures frequently associated with TON is also not clear. A retrospective review of 61 patients by Wang et al8 compared treatment with high-dose steroids, surgical intervention, and observation alone in patients with blunt and penetrating trauma. Patients with penetrating trauma were found to have worse outcomes compared with patients with blunt trauma, and no difference was found between patients who underwent surgical vs nonsurgical intervention. In the present study, patients who presented with no light perception fared considerably worse than all others. Patient age and timing of surgery were not found to have an effect on outcome.8 In another retrospective study of 45 patients with TON and facial fractures, Hsieh et al9 found no difference in visual outcome between patients who underwent surgical repair of facial fractures and those receiving corticosteroid treatment alone. In addition, patients in whom surgery occurred within 48 hours following injury did not have significantly different outcomes compared with those who had surgery later than 48 hours.9 Similar to all 3 studies mentioned,5,8,9 we found that surgery that included OND did not significantly improve visual outcome. In addition, patients with severely injured vision had a worse prognosis with little evidence of visual improvement.
Weymuller7 polled 13 experts from the fields of ophthalmology, oral surgery, otolaryngology, and plastic surgery and asked individuals whether they would repair a LeFort III fracture in the setting of blindness. Responses varied from no surgical manipulation of the facial skeleton to willingness to repair the fracture on an only seeing eye. This illustrates the controversy that exists in the management of fractures in the setting of ocular injury. Our study found that manipulation of the facial skeleton for the repair of facial fractures in the setting of TON had no negative effects on vision and when combined with OND had an overall 8% incidence in visual deterioration. This percentage of visual deterioration was not significantly different from the 4% deterioration seen in a group of individuals with a similar diagnosis of facial fracture and TON who did not undergo fracture repair (Table 5).
Visual loss after facial fracture repair has been well studied. In their 11-year study of 1240 cases, Girotto et al11 identified that the incidence of visual loss after facial fracture repair was 0.242%. This review suggested that increased intraocular pressure was the main cause in 67% of cases resulting in blindness. In examining postmortem cases of traumatic blindness, Hughes12 and Walsh13 found hemorrhage into the optic nerve sheath or contusion of the nerve with edema and compression as the main causes of blindness. Direct osseous compression, laceration of the nerve, or retinal artery hemorrhage was not commonly identified. Other suggested causes of visual deterioration have been attributed to perioperative ischemic optic neuropathy secondary to intraoperative hypotension, acute blood loss, and anemia.11,14 The ischemia is believed to be due to the diminished perfusion of the nutrient vessels that supply the optic nerve.11 A study by Stanley et al15 examined the management of displaced lateral orbital wall fractures associated with visual disturbances in 11 patients. The mechanism of vision loss in this study was postulated to be direct injury to the optic nerve caused by displaced bony fragments; no patients had worsening of vision from surgical manipulation.15
Holographic interferometry has been performed by Anderson et al16 to identify the stress forces acting on the orbit resulting in optic nerve injury. Anderson et al16 identified that a surface load placed on the skull near the supraorbital rim leads to stress concentrated in the orbital roof 5 mm from the optic foramen. Similarly, load placed on the malar eminence inferiorly leads to stress concentrated on the orbital floor 5 mm from the optic foramen as well. Injuries in these regions could have a significant impact on the delicate perineural blood supply to the optic nerve. This is especially true in the intracanalicular portion of the nerve, which is tightly encased in bone.11 The intracanalicular segment is the most frequently injured portion of the nerve and is highly sensitive to hemorrhage or traumatic edema, which can lead to compromise of the perineural blood supply.17 Delayed visual loss after facial trauma has been reported by a number of authors.18- 21 The mechanisms of injury have only been speculative in nature.
At our institution, TON is managed primarily by the ophthalmology service. Patients are treated with megadose corticosteroids unless contraindicated by their medical condition. The standard administration of megadose corticosteroids consists of a loading dose of methylprednisolone at 30 mg/kg, followed by a 15-mg/kg dose 2 hours later and then repeated dosing every 6 hours.10 It is also common practice for all patients to have the head of the bed elevated to reduce facial edema. All patients in this study received megadose corticosteroids. Only a subgroup (Table 4 [patients 17-26]) in whom vision deteriorated while receiving megadose corticosteroid treatment underwent OND.
Our review of 16 patients with the diagnosis of TON undergoing facial fracture repair without OND during a 6-year period found significant improvement (P = .001) in vision postoperatively (Table 4 [patients 1-16]). We do not believe that facial fracture correction was the cause for improved vision. Rather, we believe improvement in vision was the natural progression of TON in a group who sustained a milder injury (−1.44 ± 1.67 log MAR unit; Table 4 [patients 1-16]) and also received megadose corticosteroid treatment. This is suggested by the fact that significant visual improvement (P<.001) was also noted in the nonsurgically treated group of patients who also sustained a milder injury (−1.76 ± 1.67 log MAR unit; Table 4 [patients 27-54]) and also received megadose corticosteroid treatment. Only the severely injured group of patients (−3.57 ± 1.62 log MAR unit; Table 4 [patients 17-26]) who underwent OND, facial fracture repair, and megadose corticosteroid treatment did not demonstrate any significant improvement. Our findings are similar to those previously reported by other studies that suggest that a more severe visual injury has worse visual prognosis.8,9
More importantly, there was no increased rate of visual loss in surgically treated patients when compared with the nonsurgically treated patients (Table 5). It would appear that manipulation of the facial skeleton for facial fracture repair did not provide any significant risk to vision if all patients prior to any surgical intervention were cleared by the ophthalmology service and fracture repair was delayed until vision had stabilized. It would be prudent to delay facture repair until the patient and family comprehend the severity of the injury and realize the operative reduction of the facial fracture is a separate issue from the optic nerve lesion. On the other hand, because this was a retrospective study with the usual limitations of not being able to control multiple variables, a prospective clinical trial should be performed to confirm these results. We also cannot appreciate the benefits of OND in the setting of facial fractures, and again, a controlled prospective study would be helpful. It would appear from our experience that there is no contraindication to facial fracture repair (with or without OND) in the setting of TON, but clearly, future prospective trials would be helpful in confirming these findings.
In conclusion, surgical manipulation of the facial skeleton for facial fracture repair in the setting of TON had no adverse effect on vision. Three groups of patients with the diagnosis of facial fracture and TON who all received megadose corticosteroid treatment were managed with either (1) facial fracture repair alone, (2) OND plus facial fracture repair, or (3) nonsurgical treatment. In comparing treatment groups, there was no significant difference in visual improvement. Patients who had severe visual injury requiring megadose corticosteroid treatment, OND, and facial facture repair had a significantly worse visual prognosis and a reduced chance for improvement compared with the other groups.
Correspondence: Terry Y. Shibuya, MD, Department of Otolaryngology/Head & Neck Surgery, University of California, Irvine School of Medicine, 101 The City Drive South, Bldg 56, Suite 500, Orange, CA 92868 (TShibuya@uci.edu).
Submitted for Publication: May 25, 2005; accepted September 7, 2005.
Financial Disclosure: None.