Factors that limit the use of flash visual evoked potentials for surgical monitoring

Electroencephalography and clinical Neurophysiology , 1988, 71: 142-145

142

Elsevier Scientific Publishers Ireland, Ltd. EEG02053

Short communication

Factors that limit the use of flash visual evoked potentials for surgical monitoring C. Cedzich, J. Schramm1, C.F. Mengedoht and R. Fahlbusch Department of Neurosurgery, Unwersity of Erlangen-Niirnberg, D-8520 Erlangen (F. R. G.) (Accepted for publication: 26 November 1987)

Summary

A study was conducted comparing the incidence with which the N 2 / P 2 / N 3 was obtained after flash VEP in 3 groups: anterior visual pathway lesions, non-tumor craniotomies and non-cranial surgery. These groups allowed evaluation of the effects of anesthesia, visual pathway lesions and craniotomy on the stability of the flash VEP. It was found that the latency was not significantly affected in the 3 groups, whereas the incidence of obtainable peaks and the amplitudes were adversely affected by anesthesia, cranial surgical manipulation and especially by the presence of a visual pathway lesion. These adverse effects were so marked that the application of flash VEP for intraoperative monitoring seems of little use.

Key words: Visual

evoked potentials; Intraoperative monitoring; Visual pathway

Although it is well known that the inter- and intraindividual variability of flash visual evoked potentials (VEPs), even in normal awake subjects, is very high (Shearer and Dustman 1980; Lesser et al. 1985; Jones 1986; Oken et al. 1987), intraoperative monitoring of visual function by VEPs has been described as helpful for intraoperative guidance during surgery close to the visual pathways (Feinsod and Auerbach 1971; Wright et al. 1973; Feinsod et al. 1976; Wilson et al. 1976; Koshino et al. 1978; Albright and Sclabassi 1985; Costa e Silva et al. 1985). Our experience and that of others has been inconsistent (Allen et al. 1981; Raudzens 1982; Schramm et al. 1986; Cedzich et al. 1987). When monitoring patients with perisellar tumors we found large fluctuations of amplitudes and latencies and also potential losses which were not associated with significant clinical loss. Furthermore, we had the impression that certain manipulations were more disturbing than others. This study was conducted to elucidate these disturbing facts and to quantify factors influencing intraoperative variability of flash VEPs.

1 This study was supported by a grant of the Deutsche Forschungsgemeinschaft. Parts of these results were presented at the International Evoked Potential Symposium, Sept. 1986, Berlin.

Correspondence to: C. Cedzich, Department of Neurosurgery, University of Erlangen-Niirnberg, Schwabachanlage 6, 8520 Erlangen (F.R.G.).

Material Three groups were formed: Group 1:25 patients with perisellar tumors, i.e., patients with impairment of visual function and with manipulation of the anterior visual pathway. Group 2:10 patients without surgery close to the visual pathway, but craniotomy. Normal visual acuity and fields. Group 3: As a control group 10 patients with neither visual pathway surgery nor craniotomy, but with non-cranial surgery (lumbar disks). In all cases permission by the patients for intraoperative recording was obtained.

Methods All VEPs were recorded with a CA 1000 Signal Averager (Nicolet Biomedical). Averages were obtained after stimulation of both eyes separately. As stimulus we used red light emitting diodes (Nicolet Biomedical) through closed eyes. Stimulus frequency was 1.9 flashes/sec; 100 flashes per average were analyzed. Recordings were performed via silver cup electrodes attached to the scalp in a 1-channel montage, occipital (Oz) against frontal (Fz). Bandpass 5-100 Hz, analysis time 250 msec; electrode impedance was always under 2 kIL Anesthesia was a standardized combination of the fentanyl plus nitrous oxide and muscle-relaxant type, with additional benzodiazepine in all 3 groups. In our initial cases of the tumor group (4 of 25) we also used halothane or enflurane, but this

0168-5597/88/$03.50 © 1988 Elsevier Scientific Publishers Ireland, Ltd.

INTRAOPERATIVE M O N I T O R I N G OF FLASH VEP was soon abandoned when we realized that VEPs were profoundly changed by these agents alone (Schramm et al. 1986). In the other 2 groups, using the same anesthesia, in some cases additional enflurane in a dose of 0.2-0.6 vol% was necessary to keep blood pressure down. We chose the N 2 / P 2 / N 3 wave complex as the point of analysis as this complex was the most prominent and consistent. Incidences (i.e., the frequency of obtainabihty) of these peaks pre- and intraoperatively, the ranges, mean values and standard deviatons of latencies and amplitudes were evaluated separately for both eyes. In order to evaluate statistically a difference in the values of amplitudes and latencies occurring between the particular groups preoperatively and intraoperatively, the t value was calculated by the group samples test for all patients. Similarly, the differences between the preoperative and the first intraoperative measured values for all 3 groups were tested by the paired samples t test (significance: P < 0.05). Statistical comparison of latencies and amplitudes during the whole surgery did not seem useful, because the duration of surgery varied considerably and, especially, because we saw a high rate of potential losses in the tumor and craniotomy groups, with values not applicable for statistical evaluation.

143 NON-

CRANIAL

SURGERY

(LUMBAR

DiSC )

P r e - OP 104

~

-~ -~

i

I NORMAL

Positioning - start

surgery

//

2 5 0 ms VISUAL FUNCTION

Results

Fig. 1. The pre- and intraoperative records in lumbar disk surgery. There is no potential loss intraoperatively and the patient has normal visual acuity and normal visual fields. Positivity downwards.

The latencies and amplitudes of all patients in the 3 groups were statistically normally distributed, as determined by the Kolmogoroff-Smirnov test.

Preoperative versus intraoperative amplitudes

Preoperative latency data The preoperative latency data did not show significant differences in variability between the 3 groups (group samples t test: P > 0.05).

By within-group comparison of the preoperative and the early intraoperative amplitudes we found significantly smaller amplitudes intraoperatively in all groups (paired samples t test, P < 0.05). Only for P 2 / N 3 in the tumor group we did not find a statistical difference between the pre- and intraoperative values ( P = 0.072).

Preoperative versus intraoperative latency The within-group comparison of the preoperative latencies and the first measured intraoperative latencies did not show significant differences for all peaks (paired samples t test: P > 0.05), except that for N2 in the tumor group we observed a significant difference between the preoperative and intraoperative values ( P = 0.026); that means, the N2 values were significantly higher intraoperatively.

Intraoperative amplitude data The between-group comparison of the early intraoperative amplitudes showed no significant differences in amplitude for all groups (group samples t test: P > 0.05). Only for N 2 / P 2 of the tumor group we obtained significantly lower values than in the other 2 groups ( P = 0.01).

Potential loss lntraoperative latency The between-group comparison for the early intraoperative latencies showed no difference for all peaks and all groups (group samples t test: P > 0.05).

Preoperative amplitude data For the preoperative amplitudes we observed significant differences between the values of the tumor group and the other 2 groups (group samples t test: P < 0.05). That is, the amplitudes of the tumor group were significantly smaller than in the other two. No differences in amplitudes were found between the craniotomy and the non-cranial surgery groups.

Furthermore we saw a complete potential loss in the tumor group in 21 cases (in 4 cases only during anesthesia alone, in 9 cases during trepanation or the transnasal approach and in 8 cases during tumor dissection). The potential loss during trepanation was more often found on the trephined side than on the other side, as we described elsewhere (Cedzich et al. 1987). In the craniotomy group a complete potential loss was found in 3 cases. Here we also found flat records earlier on the trephined side than on the other. In the non-cranial surgery group, the control group, we never had potential loss during surgery (Fig. 1).

144

C. CEDZICH ET AL.

Discussion

INACTIVE

Our experience of 35 monitored patients with flash VEP (Cedzich et al. 1987) was discouraging because of so many false positive and false negative results, i.e., there was no agreement between the intraoperative recorded potentials and the postoperative clinical situation. In other words, like Allen et al. (1981) and Raudzens (1982), we had a high proportion of potential losses without clinical sequelae. The groups reporting fewer problems and advocating flash VEP monitoring (Feinsod and Auerbach 1971; Wright et al. 1973; Feinsod et al. 1976; Wilson et al. 1976; Koshino et al. 1978; Costa e Silva et al. 1985) used similar technical parameters, i.e., stimulus rates, frequency bandpass, stimulus mode. Raudzens showed in 1982 that the variability of flash VEPs could not be improved by changing the color of the LED flashes or the interstimulus intervals; alteration of the stimulus rate from 1.5 to 1.0 Hz had no effect on the variability, only on the amplitude attenuation at higher frequencies (Raudzens 1982). We think a stimulus frequency of 1.9 Hz is not too high

Stim. Rt. Eye

P2

• Pre-OP d125

S

RE

FRONTAL

-,, _

/~

/~

154

. %,',

""

~,./

.=

tumor f\//

"

.

' _ ~.-~T-'~,'" -....

",-~'~- . j

~

- v I

~

~

~

trepanation

duraopening tumor removal tumor removal tumor removal

~

duraclosure

]

R" 2

skin closure 5 ~V

I

"

., _j"

"

tumor

'x

remov@i

/-b ~ ~'~/'/"

" Op.-

end

/ \\

incision of skin

incision of muscle

~

cyst

pre-op

/

~

sphenoid sinus prep•

.

'

~-J-~,#, / ~

mucosal Incision

~

_

R. Eye

Positioning

127

',/~

ASTROCYTOMA

ADENOMA

250 ms Normal visual acuity and fields

Fig. 2. The pre- and intraoperative records of a 20-year-old woman with a frontal astrocytoma on the right side. There is a reversible potential loss during trepanation and tumor dissection.

3~o/~W/

250 ms Vis. Acuity 1.0 Rt.Eye

[6J~v-P°st-OP I

Fig. 3. The pre- and intraoperative records of a 57-year-old man with an inactive pituitary adenoma. There is a reversible potential loss during sphenoid sinus dissection and tumor removal although the patient postoperatively has an improvement of his visual acuity from 0.6 to 1.0 and no temporal hemianopia. as we obtained good cortical records in patients without visual pathway lesions. Other authors used stimulus frequencies of 2 Hz without problems (Costa e Silva et al. 1985) or even 3 Hz (Allen et al. 1981). Flash VEPs are also sensitive to other factors, for example, body temperature or anesthetics. Markand et al. (1984) showed that the flash VEP was abolished by reducing the body temperature to 25 o C, but the AEPs and SEPs could be obtained at this temperature. Russ et al. (1984) described latency prolongation of P2 at temperatures between 35 and 25°C. In long neurosurgical procedures temperature reduction down to 35.5 ° C may occur, but that is certainly insignificant for the

INTRAOPERATIVE MONITORING OF FLASH VEP VEP. The influence of anesthesia can be determined in the non-cranial surgery group. In our series the mean latencies increased (but not significantly) and the means of amplitudes decreased using nitrous oxide, fentanyl and benzodiazepine. These results agree with those of other studies (Bergamasco 1967; Russ et al. 1984; Thurner et al. 1987). The effects of halothane and enflurane were less marked in the patients with normal visual systems (Thurner et al. 1987) and could therefore be used in them in concentrations up to 0.6 vol%, but the effects were prominent in visual pathway lesions and we learned after a few cases that halogenated agents could not be used in these patients. In such cases, we saw early complete potential losses with low doses of these agents (0.4 vol%), as observed by others (Costa e Silva et al. 1985). The use of an ultrasonic aspirator or bipolar coagulation are not possible causes for high variability of the intraoperative VEP, as these instruments were rarely or never used during the transnasal approach and anyway the monitoring was interrupted during the application of both instruments. A further increase of the range of variability and in the incidence of flat records was observed during trepanation. Furthermore, we found that the VEP after stimulation on the operated side disappeared more often during trepanation than after contralateral stimulation. Thus, the trepanation itself and the side of trepanation seem to have an adverse effect for monitoring flash VEPs, although no definite explanation, such as vibration and acceleration of the skull, is available (Fig. 2). The possible influence of direct brain manipulation (pressure, mobilization) is no adequate explanation, as it does not happen during the transnasal approach, and it is similar in the craniotomy group. In patients with compressed visual pathway structures this kind of variability was even more pronounced and the incidence of potential losses even higher, especially in the stage of tumor dissection, but also during trepanation or transnasal dissection (Fig. 3). Therefore, the high variability of intraoperative VEPs in the tumor group is due to the combined effects of anesthesia, surgical manipulation and, mainly, compression of the visual pathways. As the major adverse effects on flash VEP monitoring are caused by the lesion plus the operative approach, it follows that only changes due to anesthesia might be controlled. As long as the high proportion of potential losses and large variability of flash VEPs due to the lesion plus craniotomy cannot be reduced, one conclusion of this study is that flash VEPs are not helpful for intraoperative guidance during surgery close to the visual system. The authors thank B. Yenprapine for secretarial and E. Seebach for graphical assistance. Special acknowledgment for Dr. J. Kimura's helpful criticism.

References Albright, L.A. and Sclabassi, R.J. Cavitron ultrasonic surgical aspirator and visual evoked potential monitoring for chiasmal gliomas in children. J. Neurosurg., 1985, 63: 138-140. Allen, A., Starr, A. and Nudleman, K. Assessment of sensory

145 function in the operating room utilizing cerebral evoked potentials: a study of fifty-six surgically anesthetized patients. Clin. Neurosurg., 1981, 28: 457-482. Bergamasco, B. Modification of cortical responsiveness in humans induced by drugs acting on the central nervous system. Electroenceph. clin. Neurophysiol., 1967, 23: 186. Cedzich, C., Schramm, J. and Fahlbusch, R. Are flash visual evoked potentials useful for intraoperative monitoring of visual pathway function? Neurosurgery, 1987, in press. Costa e Silva, J., Wang, A.D. and Symon, L. The application of flash visual evoked potentials during operations on the anterior visual pathways. Neurol. Res., 1985, 7: 11-16. Feinsod, M. and Auerbach, E. The electroretinogram and the visual evoked potentials in two patients with tuberculum sellae meningioma before and after decompression of the optic nerve. Ophthalmologica (Basel), 1971, 163: 360-368. Feinsod, M., Selhorst, J.B., Hoyt, W.F. and Wilson, Ch.B. Monitoring optic nerve function during craniotomy. J. Neurosurg., 1976, 44: 29-31. Jones, S.J. The value of evoked potentials in surgical monitoring. In: R.Q. Cracco and J. Bodis-Wollner (Eds.), Evoked potentials. ARL Press, New York, 1986: 421-427. Koshino, K., Kuroda, R., Mogami, H. and Takimoto, H. Hashing diode evoked responses for detecting optic nerve function during surgery. Med. J. Osaka Univ., 1978, 29: 39-47. Lesser, R.P., Li~ders, H., Klein, G. and Dinner, D.S. Visual potentials evoked by light-emitting diodes mounted in goggles. Cleve. Clin. J. Med., 1985, 52: 223-228. Markand, O.N., Warren, C.H., Moorthy, S.S., Stoelting, R.K. and King, R.D. Monitoring of multimodality evoked potentials during open heart surgery under hypothermia. Electroenceph, clin. Neurophysiol., 1984, 59: 432-440. Oken, B.S., Chiappa, K.H. and Gill, E. Normal temporal variability of P100. Electroenceph. clin. Neurophysiol., 1987, 68: 153-156. Raudzens, P.A. Intraoperative monitoring of evoked potentials. Ann. NY Acad. Sci., 1982, 388: 308-326. Russ, W., Krurnholz, W. and Hempelmann, G. Visuell evoziert~ Potentiale (VEP) in An~isthesie und Intensivmedizin. An~isthesist, 1984, 33: 154-160. Schramm, J., Cedzich, C., Fahlbusch, R., Mokrusch, Th. and Hochstetter, H. Intraoperative monitoring of evoked potentials for surgery in and around the third ventricle and the brain-stem. In: M. Samii (Ed.), Surgery in and around the Brain Stem and Third Ventricle. Springer, Berlin, 1986: 153-160. Shearer, D.E. and Dustman, R.E. The pattern reversal evoked potentials: the need for laboratory norms. Am. J. EEG Technol., 1980, 20: 185-200. Thurner, F., Schramm, J. und Pasch, Th. Wirkung von Fentanyl und Enfluran auf sensorisch evozierte Potentiale des Menschen in F l u n i t r a z e p a m / N 2 0 - - Basisnarkose. An~isthesist, 1987, in press. Wilson, W.B., Kirsch, W.M., Neville, H., Stears, J., Feinsod, M. and Lehmann, R.A.W. Monitoring of visual function during parasellar surgery. Surg. Neurol., 1976, 5: 323-329. Wright, J.E., Arden, G. and Jones, B.R. Continuous monitoring of the visually evoked response during intra-orbital surgery. J. Ophthalmol. Soc. U.K., 1973, 93: 311-314.