The use of flash visual evoked potentials in the early diagnosis of suspected optic nerve lesions due to craniofacial trauma

Journal of Cranio-MaxiLlofacial Surgery ( 1996 ) 24, 1- 11 © 1996 European Association of Cranio-Maxillofacial Surgery

The use of flash visual evoked potentials in the early diagnosis of suspected optic nerve lesions due to craniofacial trauma C.-P. Cornelius 1, E. Altenm~iller 2, M. Ehrenfeld1

1Department of Maxillofacial Surgery (Head." Prof. Dr. Dr. N. Schwenzer), 2Department of Neurology (Head." Prof. Dr. J. Dichgans), University of Tfibingen, Germany S U M M A R Y. Craniofacial trauma encroaching on the orbital apex and optic canal can result in direct or indirect optic nerve lesions, leading to visual impairment or blindness. Early diagnosis of a visual loss and immediate therapy are generally considered crucial for a successful restoration of vision in indirect trauma. However, in comatose or sedated patients the assessment of optic nerve function by testing pupillary reactivity may be severely compromised or impossible because of tensely swollen eyelids, conjunctival oedema, concussion of the ciliary muscle or pharmacological effects. In the event that clinical ophthalmic examination, computer tomography or nuclear magnetic resonance scanning fail to clarify the state of the optic nerve, visual evoked potentials (VEPs) to flash stimulation appear to provide reliable information on function within the visual pathway. On this basis, treatment with corticosteroids and/or surgical decompression can be rapidly initiated. Our results in a preliminary patient series confirm the value of acutely monitored VEPs as an objective test of optic nerve function in cases of suspected optic nerve injury immediately after admission to the emergency care unit. The imaging techniques usually applied may be complemented by VEPs to show the functional significance of structural abnormalities found in the vicinity of the optic nerve.

INTRODUCTION Visual impairment or loss caused by optic nerve lesions constitute a major complication in craniofacial trauma. Holt and Holt (1983) reported an incidence of 3% unilateral blindness in facial fractures. Most frequently the intracanalicular portion of the optic nerve is damaged as a consequence of severe or even trivial injury (Walsh, 1979). The latter site in the optic pathway is an area predisposed to trauma, since the nerve is fixed in its position within the bony canal by the dura mater, which itself is rigidly adherent to the bone. Another reason for the peculiar vulnerability of this fixed nerve segment is found in its delicate vascular supply by a network of pial-end branches from the ophthalmic and internal carotid arteries. It has been suggested that optic nerve lesions are caused by a variety of pathogenic mechanisms (Kline et al., 1984), which affect the neural structures, the local blood circulation or both simultaneously (Frenkel and Spoor, 1987). The injury may be direct (i.e. partial or complete transsection or disruption of the nerve) or indirect (i.e. concussion, contusion, compression or laceration of the nerve) due to fractures deforming or displacing the optic canal. External compression of the optic nerve can also occur from impinging bone fragments after orbital wall fractures. The nutrient pial arteries, which are vital for optic

nerve function may suffer spasm, thrombosis or avulsion by compression, torsion, stretching or shearing. Moreover, intraneural and optic nerve sheath hemorrhage may cause compression in the non-yielding intracanalicular passage, followed by demyelination and atrophy of the nerve. Persistent ischemia due to pressure occlusion of the arteries leads primarily to intraneural edema formation, which increases the soft tissue volume in a vicious circle that eventually results in tissue infarction. Depending on the nature of the underlying pathological process, very few conditions (e.g. blowin fractures or compromised microcirculation and its sequelae) will respond to any form of treatment aimed at relieving the pressure within the optic canal. Surgical decompression is considered promising in cases with gradual deterioration of vision, since the continuity of the optic nerve is preserved (Samii and Draf, 1989). From the above considerations there can be little doubt that indirect optic nerve trauma is reversible only if decompression becomes effective shortly after injury. In striking contrast to this, positive results have been reported with surgical optic nerve decompression performed even after several weeks (Fukado, 1981). In general, however, surgical decompression in longstanding cases meets with very limited success and is controversial because of poor and inconsistent results (Brihaye, 1981; Schroeder et al. 1989).

2

Journalof Cranio-MaxillofacialSurgery

Although the relative or absolute deadline for an effective elimination of the primarily reversible causative factors in indirect injuries is virtually unknown, immediate treatment is considered to be essential for a favourable outcome by most authors (Lipkin et al., 1987; Osgutthorpe and Soffermann, 1988; Stoll et al., 1988; Dumbach et al., 1991; Kallela et al., 1994; Hammer, 1995). In recent years, a number of refined surgical techniques and novel approaches under endoscopic control (e.g. Hassler and Eggert, 1985; Wiegand, 1989; Rice and Schaefer, 1993; Perneczky et al., 1993) have been introduced in order to reduce the morbidity of transfacial or transcranial/transfrontal decompression procedures and make them immediately applicable to patients with multiple injuries. The intravenous administration of megadose corticosteroids (methyl-prednisolone: loading dose 30mg/kg body weight, second dose 15 mg/kg 2 h after initial dose, followed by 15 mg/kg every 6 h over 48 h) has been proposed as an alternative or adjunct to surgical treatment (Frenkel and Spoor, 1987; Lipkin et al., 1987; Spoor et al., 1990, 1991). In any case, accurate and reliable evaluation of optic nerve function during the immediate diagnostic work-up is a top priority for a successful subsequent treatment (Beuthner, 1974; Behrens-Baumann and Miehlke, 1979; Clifford-Jones et al., 1980; Schroeder et al., 1989). Psycho-physical testing in a patient who is able to communicate and co-operate is certainly the most appropriate form of clinical visual testing and permits early recognition of visual impairment or loss. However, in patients who are intoxicated with drugs or alcohol, sedated, unconscious or comatose, clinical testing of vision is essentially confined to the assessment of pupillary size and reactivity to light. Under these circumstances a compromised optic nerve can be detected by the presence of a MarcusGunn pupil (afferent pupillary defect), which is elicited with the 'swinging flashlight test' (see Frenkel and Spoor, 1987; Wilhelm, 1991; Hammer, 1995). But this examination in fact is often difficult or impossible to perform, because of tensely swollen eyelids, conjunctival oedema, concussion of the ciliary muscle or pharmacological effects. In such cases, flash visual evoked potentials (VEPs) may provide readily accessible information on the state of the optic nerve and serve as a guideline in the decision to operate or begin corticosteroid therapy (Feinsod and Auerbaeh, 1973; Feinsod, 1976; Shaked et al., 1982; Obertaeke et al., 1986; Mahapatra and Bhatia, 1989). In a pilot study to assess their reliability and value in the detection of clinically unrecognizable visual impairment, VEPs were performed in normal subjects, patients with manifest unilateral blindness after previous optic nerve injury and patients during the

acute phase after craniofacial trauma. Any attempt to substantiate a relationship between the prospects for visual recovery and the timing of surgery was beyond the limits of this investigation, since most of the acute cases demanding decompression, were in such critical condition, that no option for an immediate surgical intervention remained.

METHODS Flash VEP - recording technique

The recording of flash VEPs represents a non-invasive electrophysiological technique assessing for the functional integrity of the optic pathway (Feinsod and Auerbach, 1973; Feinsod, 1976; Altenmiiller et al., 1991). A schematic view of the VEP setup is depicted in Figure 1: stroboscopic light stimuli are applied to each eye individually via goggles containing a set of four crosswise oriented light emitting diodes (LED) inside (Fig. 2 a, b). The cortical responses are obtained with surface electrodes placed occipitally above the visual cortex (Oz) against a fronto-central reference (Fz) (Fig. 1). Isolated cortical responses do not normally exceed a few microvolts and are hidden by higher-voltage spontaneous EEG-activity (Picton, 1987; Altenmiiller et al., 1989). The flash stimulus is regenerated continuously at a specified repetition rate in order to achieve discrimination of the synchronized response. After preamplification, the incoming digitized signals are passed to a so-called 'averager', that is, a microprocessor which sums up signals in phase with the stimulus presented. Meanwhile ambient electro-physiological background activity is eliminated. The averaging technique improves the signal-to-noise ratio in direct relation to the number of acquired sweeps. The findings can be cursor-analysed on the monitor screen and an integrated printer or plotter documents the results. At least two recording runs are repeated to check for consistency of the findings. With a portable recording system, VEPs can be registered on admission to a trauma care unit. Technical details and specifications

A principal handicap for the recording of evoked potentials in emergency or intensive care units is that artifacts and interferences are induced by alternating current from the omnipresent electrical equipment in such rooms. To minimize these disturbances, it is essential to lower the skin impedance to below a level of 3 kilo Ohm by gently scouring and wiping the skin with a gauze pad dipped into a water/pumice powder mixture or into alcohol before attaching the silver/

The use of flash visual evoked potentials in the early diagnosis of suspected optic nerve lesions due to craniofacial trauma

3

Oz

Fig. 1 - VEP setup and electrode position sites. Oz = contralateral visual cortex 3 cm above the inion, Fz = frontocentral reference. Signal triggering generates input to the averager.

(A) Flash VEP gogglesin a patient under general anesthesia. (B) Cross-wiseoriented LED's inside each goggle (polished rivet in the centre).

Fig. 2 -

silver chloride disc electrodes (TOnnies, Freiburg, Germany) to the scalp surface. The active electrode is placed over the occipital pole about 3 cm above the inion against a reference in a midfrontal position. The electrode paste used was Grass EC-2; it has a gum-like consistency and functions as an adhesive for fixation of the metal electrode; a wrap around electrode on one wrist served as ground connection. The plug-in preamplifier (PA 89, Medelec LTD, Old Woking, Surrey, U K ) was located close to the head of the patient to avoid further electrical interferences. The portable recording system (MS 92a, Medelec LTD), which was interfaced with a visual stimulation unit (external trigger, Fig. 1), simultaneously ampli-

fled, averaged, stored and displayed the resulting montage on a single channel. The controls of the recording system were set to a timebase of 500 ms following the stimulus (time/ horizontal division 50 ms) and to a gain of 10/~V-20 #V/vertical division. The bandpass of the filter was set between 0.1 Hz and 100 Hz. The flashes of light were applied unilaterally using the L E D goggles (Medelec LTD) and had a single duration m o d e of 5 ms and a repetition frequency of 1 Hz. For each single eye examination, 128 successive sweeps were averaged. Latencies and amplitudes of the tracings were then cursor-analysed and printed together with the recordings on a strip chart. A full recording session, including the placement

4

Journal of Cranio-Maxillofacial Surgery

of the electrodes and two runs on each eye took about 10 to 15 rain.

RESULTS

0.85 #V (SD: +4.2 #V). The latencies differed by up to 18 ms (mean: 0.17 ms; SD: _+7.7 ms) in recordings repeated at varying time intervals in the same individual, and the amplitudes showed a divergence between 0.3 and 8.8 #V (mean: 0.75 #V; SD:_+3.3 #V).

Reference data

A prerequisite for the application of flash VEPs in patients is the knowledge of normal values and waveforms. Because of minor, but inevitable discrepancies in the characteristics of the stimulation and amplifier components of each recording system, which in turn depend on slight differences in the intensity, duration and frequency of the light flashes, every laboratory is obliged to define its own reference data. For this purpose, a group of 20 normal subjects (age 20-49 years) (= 40 eyes) was recruited. To simulate conditions after craniofacial trauma the eyelids were kept closed during the examination. Certain features of the flash VEP (Fig. 3) proved to be fairly consistent across the control subjects: after an initial series of negative and positive deflections below 80 ms corresponding to an electroretinogram (ERG), a reproducible high-amplitude, positive component within a range of latencies between 96 and 158ms was found (mean :131.2 ms; standard deviation [SD]: _+16.9 ms). In accordance with pattern-shift evoked potentials, in which a similar deflection can be observed, the encountered peak was termed 'P 100'. P stands for positivity, which is an electrophysiological convention for a downward deflection of the registered curve. The added number designates the latency, after which the peak can normally be anticipated (for review see Celsia et al., 1977; Altenmiiller et al., 1989). The amplitudes of 'P 100' had a mean of 13.1 #V with a standard deviation of _+4.9 #V (lower and upper limit: 3.2 #V and 30.7 #V resp.). Similiarly with the latencies, interocular left-right differences of up to 8.8 ms (mean: 0.43ms; SD: _+12.7 ms) were noted. The side differences for the amplitudes lay between 0.1-7.6 #V with a mean of

Evaluation criteria

Based on experience with the control group, criteria had to be predefined, when a flash VEP is in fact accounting for a severe impairment or loss of vision in patients with suspected optic nerve injury. 'P 100' was referred to as the main diagnostic parameter. Traumatic visual loss was considered to be indicated by the complete abolition of this major peak over the occipital cortex. A problem and drawback in VEP evaluation is the dependence of the visual cortical response on the luminance reaching the retinal receptors; this should not fall short of a value of 10 cd (=candela)/m2 (see Altenmiiller et al., 1989). If the light is increasingly filtered on its way to the retina, the latencies of the evoked potentials will be progressively delayed and the amplitudes diminished, until all responses are finally extinguished. Judicious use of the VEP in craniofacial trauma cases is therefore mandatory. One must be aware that even closure, oedema or haematoma of the eyelids result in peak delays and attenuation of the amplitudes. Nevertheless the amount of light entering the eye in periorbital oedema and ecchymosis and/or chemosis is sufficient to produce VEPs and to demonstrate whether the optic nerve is conducting impulses or not. In the presence of bleeding into the anterior or posterior segments of the globe (hyphaema or vitreous hemorrhage) and in severe retinal detachment, light input to the retina and the propagation of impulses may be reduced to such an extent that reliable interpretation of the VEPs becomes questionable.

Left Eye

Right Eye

v

0

~ Latency

Amplitude

5 0 0 ms 122

ms

11,64

pV

0

~ Latency

5 0 0 ms 120

Amplitude 15,00

ms ~tV

Fig. 3 - Flash VEP configuration of a normal subject. L = Latency; A = Amplitude. Arrows m a r k 'P100', the main diagnostic parameter. Actual latencies and amplitudes for 'P100' are indicated.

The use of flash visual evoked potentials in the early diagnosis of suspected optic nerve lesions due to craniofacial trauma

The latter conditions have to be excluded by clinical judgement and/or CT-scanning, as far as possible, in order not to produce erroneous results. Owing to the high individual variance of the latencies of up to 18 ms, unilat~ al delays of °P100' were only accepted as evidence of visual impairment if the prolongation was extreme. Unilateral decline in amplitudes was considered as a clue to optic nerve lesions only in the event that lid haematomas and conjunctival oedema were absent and the amplitude more than 50% below that of the normal side. In addition the lowered amplitude had to be reproducible at least twice. The analysis of the VEPs was principally restricted to the latency range above 90 ms to avoid confusion with the ERG, which is composed of high amplitude potentials with latencies between 30 and 80 ms that are generated by retinal activity and are displayed distinctly when a frontal reference electrode is employed. In cases of optic nerve lesions, the ERG is known to be preserved for several weeks (for a review see Zrenner, 1989) and even to be enhanced in the acute onset injury (Feinsod and Auerbach, 1973).

z= ,-~ 0

~

0

0-~

~

.~ o . ~ . ~ o . ~ o . ~ . ~ >Z>>Z>Z>>

~

0

o

VEPs in patients To date, we have had the opportunity of monitoring flash VEP's in 17 patients (Tables 1 and 2). In three patients (Table 1), the injuries and subsequent monocular blinding dated back a maximum of three years. These three patients volunteered to validate the VEP recording system as controls. Consistent with their defects in pupillary reactivity, all three exhibited no cortical response after stimulation of the amaurotic eye, whereas the regular sequence of potentials was obtained from the normal side (Fig. 4). The remaining 14 patients (Table 2) were investigated in the acute phase ( 1 hour-4 days) after trauma. All patients in this group had an altered level of consciousness and were unable to cooperate; 11 were intubated and under general anaesthesia. With the exception of the first patient (case 1) pupillary reactivity was either suspicious for an optic nerve injury or impossible to test because of periorbital soft tissue alterations (cases 6, 7, 8). In accordance with the normal clinical findings, case 1 showed regular VEPs bilaterally. Five patients (cases 2-5) with abnormal pupillary function and one of the non-examinable patients (case 6) displayed a typical VEP configuration with undisturbed conduction of visual information in the optic nerve. The electrophysiological findings were consonant with normal vision as indicated by psychophysical examinations after the patients had regained normal consciousness and began to cooperate. Thus the VEP recordings reliably recognized the state of

Z~

O0

~

~

ZZ

ZZ

G~

~

z ,.o

o~ 0 o

0

z

g,

t~h

8

8

0

'8 .~

.=

o

._=

¢m

K o

o

o

o m

5

6

Journal of Cranio-Maxillofacial Surgery

o.c~

o . _ ~ . ~

~

oo~

,~,.~

o

o

~

o

o

~0

o

o

~

~

8

~ ~ ~:~ 0

d

~

od ~

e ~

~ "

~

~

~ .

~.~

..~.

~

".~

0

0

¢~

0

~ o

~

0

~i. oo

.

.

~

.

<

o

~z

.

.

~

.

.

~

~

0

~g~zz

oo

o

ZZ

0

zz

0

0

0

0

<

% ,-1

•~ ~

"~ ~

o

.=

.~ z

~

z

0

.~

"~

~ ~

~ z

o

~o

~

~

~

~o ~

o i ~ ~

= ~

. ~o .~~o

. .o. ~. o

~

~

o

i

a

a

o~

~ .o~ o

.

•

8

©

~'~

n~

0'~

~

0

~'~

~

0 ~'~ m

-~'~

'~

~ ~.~

'~

~ ~ ~ ~ .~, .~ "~

o ~ ~ ~.~ ~ .o .~ ~ - ~

~.~

" ~

~o ~ : ~

_~.~ .~ ~ ~ ~ '.~.-N o

~.~

The use of flashvisualevokedpotentialsin the earlydiagnosisof suspected optic nerve lesionsdue to craniofacialtrauma o~

,.~

~-J+a ~

o

o

o

~

.~

~

~.~

°.

": 0

o o

zz

0

O

zz

zo Z0

ee~, o

O

ZO ZO

: ~

• '.~a

~:~'~

Z

~D

-,

0"~

'~

©

0

o

o

":

o

0

0

8

~ 0 0 0

z

" : ©~ ""

o

0

~

~'-"

-'~

Z ~-~

o~

"6

"~ ~-.~ " 7

~.~

.*-e. " ~ "~,

a.~=g

.-z~

d--~ o ~

eq

~-

~'.~

~.~_

¢'q

7

the optic nerve initially in these cases, and turned out to be more sensitive than pure clinical testing. Eight patients (cases 7-14) of the acutely assessed group presented pathological VEPs. Two of these patients (cases 7 and 8) belong to the subset, in which clinical examination was impossible. A missing 'P100' on the right side and a marked abnormality in the latency of the left side, respectively, predicted bilateral blindness in one of these two patients (case 7) after severe panfacial trauma (Fig. 5). In the other patient (case 8), unilateral amaurosis on the right became apparent by the abolished cortical response to flash stimulation. In the remaining six patients (cases 9-14), with pathological VEPs, pupillary testing beforehand had disclosed irregularities indicating unilateral afferent dysfunction. Clinical analysis ultimately revealed a unilateral amaurosis correlated to a unilateral loss of 'P100' in five patients of the latter group (cases 9-13). One conscious, but very inebriated patient (case 14), with a bone fragment antero-medial to the optic foramen in the CT scan (Fig. 6a) after a malar fracture on the right, showed a more than 50% unilateral reduction in the amplitude of 'P100' compared with the contralateral side (Fig. 6b). This finding turned out to be equivalent to a subtotal visual field defect of the right eye during follow-up clinical examination. This patient had been recommended for surgical exposure and decompression, but informed consent could not be obtained from him or his relatives for any type of treatment. Transfacial unroofing of the medial wall of the optic canal was performed in one patient (case 13) on the basis of the extinct VEP after a malar fracture, but no improvement was achieved. Although we think that immediate surgical decompression should be offered to every craniofacial trauma patient with visual loss on the premise that the underlying pathology is deemed reversible, the other patients in the series with pathological VEPs (cases 7-11) had to be left untreated surgically in the immediate period after trauma due to their critical state (e.g. cerebral oedema with intracranial hypertension, abdominal injuries, continuous massive bleeding requiring anterior and posterior nasal packing, etc.). In one patient with a pathological VEP (case 12), a fracture line directly crossed the optic canal, so that megadose corticosteroid therapy was tried on the assumption that it was a probable direct injury. A final patient (not tabulated) developed the obvious features of retrobulbar hemorrhage (severe proptosis, dilated unreactive pupil, tense globe and extreme swelling of the eyelids (Ord, 1981)) after malar fracture reduction and orbital floor repair on the right. Flash VEPs could not be elicited from this eye in accordance with the clinical loss of vision. Immediate surgical re-entry for decompression of the

8

Journal of Cranio-Maxillofacial Surgery

Left Eye

Right Eye

0

5 0 0 ms Latency Amplitude

148 13,08

0

5 0 0 ms

ms ~V

4 - Control patient no. 1. VEPs 2.5 years after an atypical fracture of the right malar. Persistent monocular blindness after drilling down the medial wall of the right optic nerve. Normal VEP configuration on the left. No evoked potentials available from the right eye. Fig.

Right Eye

Left Eye

[-

Lr

0

5 0 0 ms Latency Amplitude

228 3,75

0

5 0 0 ms

ms pV

Case no. 7. VEPs of a multi-traumatized patient with panfacial fractures soon after injury. Marked abnormalities in the VEP from left eye with extremelyincreased latency (228 ms) and low amplitude of 'P100'. No evoked potential available from the right eye. Bilateral amaurosis was assumed and later confirmed. Fig. 5 -

orbital contents by multiple drainage, combined with megadose corticosteroids fortunately resulted in recovery of usable vision (visual acuity 0.3) and reappearance of the VEPs. Postoperative values were recorded immediately (latency 152ms, amplitude 5/~V) and one week after orbital decompression (latency 126 ms, amplitude 3 #V). This case history is mentioned only to illustrate that the VEP recording m a y be useful documentation for medico-legal purposes.

DISCUSSION The possible role of the VEP as a diagnostic aid in suspected optic nerve lesions after craniofacial trauma has long been realised (see Feinsod and Auerbach, 1973; Feinsod, 1976), but former technical equipment, for instance the need for a Faraday cage, made the examination too cumbersome and impractical for frequent application (Kline et al., 1984). The compact, portable recording device and the use of LED-goggles for stimulation now make it easy to monitor flash VEPs immediately after admission to an emergency care unit. Our experience with the patient series described,

shows that flash VEPs can indeed be registered in the acute phase after trauma and m a y be of indisputable value in objectively testing the optic nerve function of unconscious or otherwise uncooperative patients (Mahapatra and Bathia, 1989; Osguthorpe and Sofferrnan, 1989). The VEP recordings in all 14 patients assessed as soon as possible after trauma closely correlated to the later ophthalmological findings. In five cases, however, a discrepancy was noted between the initial pupillary testing and the VEP findings. In contrast to the clinical deficiencies in pupillary function, flash VEPs showed a bilaterally preserved 'P100', which emerged as the main parameter of relevance in the normal and control subjects. In a subset o f three patients, whose pupillary reactivity could not be examined clinically, flash VEPs were enormously helpful in clarifying the functional state of the optic nerve and delivered the information clinically inaccessible. The recordings were indicative of uni- and bilateral visual loss in two of these cases and detected normal electrical responses in the other. The clinical suspicion of a unilateral optic nerve lesion was supported in the remaining six cases. In comparison with clinical examination of pupillary function ('swinging flash light test'), the VEPs

The use of flash visual evoked potentials in the early diagnosis of suspected optic nerve lesions due to craniofacial trauma

Left Eye

9

Right Eye L_ L

0

5 0 0 ms

Latency B

Amplitude

160 8,36

ms jiM

0

5 0 0 ms

Latency 120 ms Amplitude 3,59 /~V

Fig. 6 - Case no. 14. Comminuted malar fracture on the right (A) CT-scan showing involvement of the optic canal by bone fragment presumably impinging on the optic nerve anteromedially. (B) VEP recording to substantiate the functional significance of the CT-findings in the uncooperative and very drunk patient. The lateneies on both sides were within normal limits, but the amplitude of 'P100' on the right had decreased more than 50%. The pathological significance of this attenuation of amplitude was confirmed later by the detection of a subtotal visual field defect. In this case decompression of the optic nerve was indicated, but any form of treatment was refused by the patient and his family.

were more accurate in assessing the situation and predicting the clinical outcome and can help to preclude faulty therapeutic decisions. The success of surgical or 'medical' decompression is particularly dependent on the underlying pathology and the remaining viability of the neural tissues (Lipkin et al., 1987). In potentially reversible pathological conditions, early detection and reliable examination of optic nerve function is of decisive importance for salvaging the patient's vision by subsequent decompression (Beuthner, 1974; Behrens-Baumann and Miehlke, 1979; Osgutthorpe and Sofferrnann, 1988). On the basis of the VEP findings, we performed a transfacial de-roofing of the medial wall of the left optic canal in one patient (case 13). Although the time lapse between injury and surgery was short (4 h), the patient regrettably did not regain visual function. In the majority of patients, the VEP verification of optic nerve damage had no therapeutic consequences in terms of surgical exposure owing to vital problems or an unstable condition not allowing surgery. Since the VEP's are affected by a large number of

factors, such as stimulation and patient parameters, the appropriate recording technique and reliable interpretation of the results in trauma cases requires training and an in-depth knowledge of the caveats. Foremost among methodological problems that limit the use of flash evoked potentials (see Evaluation criteria) are: (1) the filtering effects of intraocular hemorrhage, which diminish the retinal luminous density, (2) the restrained significance of latency delays of the 'P100', and (3) the risk of confusing this peak with ERG waves. To obviate this risk, simultaneous two-channel monitoring of VEP and E R G will be carried out in future trials. Despite adverse recording conditions attributed to closed eyelids, lid haematoma, conjunctival oedema and pharmacological agents inducing mydriasis or miosis during general anaesthesia, flash VEPs yielded meaningful information in our present series and carried weight in the differential diagnosis of an optic nerve lesion. Partial and complete visual dysfunction of the optic nerve were discerned by tentatively assigning clearcut amplitude reductions of more than 50% to a

10

Journal of Cranio-Maxillofacial Surgery

partial blockage of conduction in the anterior visual pathway. If applied judiciously, flash VEPs facilitate prompt recognition of optic nerve lesions after craniofacial trauma. Neuro-ophthalmologic and radiological findings can be supported, and VEPs may have an impact in indicating the need for immediate therapeutic intervention (Obertackeet al., 1986). It is also conceivable that a gradual deterioration or loss of vision (Neubauer, 1987) in patients with prolonged alteration of consciousness necessitating optic nerve decompression can be detected at first onset by regular flash VEP rechecks and thus improve the prospects of therapy.

CONCLUSION

Although flash VEPs are not to be regarded as a first line diagnostic tool in routine use, this non-invasive electrophysiological technique can deliver diagnostic information when conventional methods fail. Before surgical decompression of the optic nerve is undertaken, VEPs may also be used as a safeguard to reaffirm the clinical deficiencies in pupillary function. However, the limitations of the method have to be borne in mind. In short, four conclusions can be drawn from our present observations: 1. Flash VEPs reliably test optic nerve function in uncooperative or comatose patients with suspected injury to the anterior optic pathway 2. The sensitivity and predictive value of VEPs seem superior to those of clinical pupillary testing after craniofacial trauma 3. Flash VEPs can distinguish partial and complete dysfunction of the optic nerve, thus complementing the information gained by modern imaging techniques such as X-ray, CT and N M R scanning 4. Repeated VEPs during follow-up can detect a gradual loss of vision necessitating optic nerve decompression independent of the patient's cooperation. References Altenm~ller, E., H. C Diener, Y. Dichgans: Visuell evozierte Potentiale (VEP). In: St6hr, M., J, Dichgans, It. C. Diener, U. W. Buettner: Evozierte Potentiale SEP-VEP-AEP. Springer, Berlin 1989, 279-381 Altenmaller, E., C. P. Cornelius, 1t. Uhl."Blitz-evozierte visuelle Potentiale in der Frtihdiagnostik yon Optikusschaden nach kranio-fazialen Frakturen. Z. EEG-EMG 22 (1991) 224-229 Behrens-Baumann, W., A. Miehlke: Zur rhinobasalen Dekompression des traumatisch gesch~digten Nervus Opticus. Kiln. Monatsbl. Augenheilkd. 175 (1979) 584-591 Beuthner, D.: Analyse zur Frage der Nervus-opticusDekompression - zugleich eine 1)bersicht t~ber 10 Jahre praventiv-sanierende Versorgung yon Rhinobasisfrakturen (1964-1973). Laryngol. Rhhlol. 53 (1974) 830-835 Brihaye, J.: Transcranial decompression of optic nerve after

trauma. In: Samii, M., Janetta, P. J. (eds.): The cranial nerves. Springer, Berlin 1981, 116-124 Celesia, G. G., R. ~ Daly: Visual electroencephalographic computer analysis (VECA). A new electrophysiologic test for the diagnosis of optic nerve lesions. Neurology (Minneap) 27 (1977) 637-641 Clifford-Jones, R. E., D. N. Landon, W. I. McDonald. Remyelination during optic nerve compression. J. Neurol. Sci. 46 (1980) 239-243 Dumbach, J., W. Y. Spitzer, E. W. Steinhduser: Zur Therapie von L~isionen des Nervus Opticus bei Mittelgesichtsfrakturen. In: Schwenzer, N., G. Pfeifer (eds.): Fortschritte der Kiefer- und Gesichtschirurgie, Vol. XXXVI. Thieme, Stuttgart 1991 162-165 Feinsod, M.: Electrophysiological correlates of traumatic visual damage. In: McLaurin, R. L. (ed.): Head injuries: Proceedings of the second Chicago Symposium on Neural Trauma. ch. 13. Grune and Stratton, New York 1976, 95-100 Feinsod, M., E. Auerbaeh: Electrophysiological examinations of the visual system in the acute phase after head injury. Eur. Neurol. 9 (1973) 56-64 Frenkel, R. E., T. C. Spoor: Diagnosis and management of traumatic optic neuropathies. Adv. Ophthal. Plast. Reconstruct. Surg. 6 (1987) 71-90 Fukado, Y.: Microsurgical transethmoidal optic nerve decompression: experience in 700 cases. In: Samii, M., Janetta, P. J. (eds.): The cranial nerves. Springer, Berlin 1981 125-128 Hammer, B.: Orbital fractures - diagnosis, operative treatment, secondary corrections. Hofgrefe & Huber Publishers, Seattle 1995 Hassler, W., H. R. Eggert: Extradural and intradural microsurgical approaches to lesions of the optic canal and the superior orbital fissure. Acta Neurochirurgica 74 (1985) 87-93 Holt, G. R., J. E. Holt: Incidence of eye injuries in facial fractures: an analysis of 727 cases. Otolaryngol. Head Neck Surg. 91 (1983) 276-279 Kallela, I., T. Hyrkgis, P. Paukku, T. Iizuka, C. Lindqvist: Blindness after blunt maxillofacial trauma - Evaluation of candidates for optic nerve decompression surgery. J. Cranio-Maxillofac, Surg. 22 (1994) 220-225 Kline, B. L., 17. B. Morawetz, S. N. Swaid: Indirect injury of the optic nerve. Neurosurgery 14 (1984) 756-764 Lipkin, A. F , G. E. Woodson, R. H Miller: Visual loss due to orbital fracture. The role of early reduction. Arch. Otolaryngol. Head Neck Surg, 113 (1987) 81-83 Mahapatra, A. K., K Bhatia: Predictive value of visual evoked potentials in unilateral optic nerve injury. Surg. Neurol. 31 (1989) 339-342 Neubauer, H.: Das frontobasale Trauma- Diagnostik und Behandlungs-ablauf. Ophthalmologische Aspekte. In: Schwenzer, N., Pfeifer, G. (eds.): Fortschritte der Kiefer-und Gesichtschirurgie, Bd. XXXII Thieme, Stuttgart 1987, 223-226 Obertacke, U., T. Joka, F. Hdrting, H..-E. Nau, W.. Sauerwein: Amaurose dutch traumatische Sch~tdigung des Nervus opticus. Unfallchirurg 89 (1986) 132-137 Ord, JR. A.: Postoperative retrobulbar haemorrhage and blindness complicating trauma surgery. Brit. J. Oral Surg. 19 (1981) 202-207 Osguthorpe, J. D., R. A. Soffermann. Optic nerve decompression. Otolaryngol. Clin. North Am. 21 (1988) 155-169 Perneczky, A., M. Tsehabitseher, K. D. M. Resche: Endoscopic anatomy for neurosurgery. Thieme, Stuttgart, 1993 Picton, T. IV..: The recording and measurement of evoked potentials. In: Halliday, A. M., Butler, S. R., Paul, R. A. (eds.): A textbook of clinical neurophysiology. John Wiley, Chichester 1987, 23-40 Rice, D. H., S. D. Schaefer: Endoscopic paranasal sinus surgery. Raven Press, New York 1993 Samii, M., W. Draf" Surgery of the skull base - an interdisciplinary approach Springer, Berlin 1989 Schroeder, M., 11. Kolenda, E. Loibnegger, 11. Miihlendyck: Opticusch~idigung nach Sch~tdel-Hirn-Trauma. Eine kritische Analyse der transethmoidalen Dekompression des N. opticus. Laryng-Rhino-Otol. 68 (1989) 534-538

The use of flash visual evoked potentials in the early diagnosis of suspected optic nerve lesions due to craniofacial trauma Shaked, A., M. Hadani, M. Feinsod: Ct and VER follow-up of reversible visual loss with fracture of the optic canal. Acta Neurochir. 62 (1982) 91-94 Spoor, T. C., W C. Hartel, D. B. Lensink, M. J. Wilkinson: Treatment of traumatic optic neuropathy with corticosteroids. Am. J. Ophthalmol. 110 (1990) 665-669 [published erratum appeared in Am. J. Ophthalmol. 111 ( 1991 ) 526 Treatment of traumatic optic neuropathy with corticosteroids - correction (letter)] Stoll, W., H. Busse, P. Kroll: Decompression of the orbit and the optic nerve in different diseases. J. Cranio-Maxillofac. Surg. 16 (1988) 308-311 Walsh, F. B.: Trauma involving the anterior visual pathways. In: Freeman, H. M. (ed.): Ocular trauma, Appleton-CenturyCrafts, New York 1979 Wiegand, M. E.: Endoskopische Chirurgie der Nasennebenh6hlen und der vorderen Schfidelbasis. Thieme, Stuttgart 1989

11

Wilhelm, H.: Pupillenreaktionen - Pupillenst6rungen. Kohlhammer, Stuttgart 1991 Zrenner, E.: The physiological basis of the pattern electroretinogram. In: Osborne, N., Chader, J. (eds.): Progress in retinal research. Vol. 9. Pergamon 1989, 427-464 Carl-Peter Cornelius

Klinik f~ir Kiefer- Gesichtschirurgie der Universit~t Tt~bingen Osianderstr. 2-8 D-72076 TUbingen Germany Tel: 07071/29 61 70 Fax: 07071/29 57 89 Paper received: 18 October 1995 Paper accepted: 30 January 1996