Keywords
vestibular schwannoma - facial reconstruction - hypoglossal - nerve transfers - facial
schwannoma
Introduction
Facial paralysis caused by facial nerve (N7) lesion due to trauma or surgery is a
devastating condition that may result in lifelong loss of function of the muscles
in the frontal, orbital, and oral segments, affecting the ability to frown and close
the eye and mouth. In addition, the quality of life is diminished by the loss of the
ability to express emotion through smiling.[1]
[2]
[3]
[4]
In the past decades, different techniques for facial nerve reconstruction have been
proposed.[3]
[5]
[6]
[7]
[8]
[9]
[10] One of these is the hypoglossal–facial nerve (N7–N12) transfer, of which numerous
technical modifications have been described.[10] One of them is the partial (hemi) use of the hypoglossal nerve with direct end to
side coaptation to reduce hemiatrophy of the tongue and diminish recovery time.[2]
[9]
[11]
[12]
[13]
Over the years, many systems to grade the facial nerve function have been developed.[14]
[15] Historically, the House–Brackmann (H-B) score is the most well-known and widely
used grading system to score facial nerve function, using both characteristics at
rest and in motion.[16] Although originally not developed to score the facial function after reconstruction
and despite its shortcomings, the H-B grading system is also frequently used in studies
reporting on the outcome of the N7–N12 transfer, namely, in around 70%.[17] The H-B grading does not clearly differentiate between the function of different
segments of the face at rest and in an active phase. Therefore, detailed information
about the potential differences between the function in a static or dynamic phase,
for instance smiling, is limited.[14] We know that a good smile suggests increased intelligence, happiness, and social
status. Therefore, smiling is fundamental in facial reanimation.[4]
In this study, we evaluate our results of facial reconstruction using the N7–N12 transfer
and specifically focus on the ability to smile. Five medical doctors blinded for the
side of the N7–N12 transfer independently assessed photographs of the whole face at
rest and while smiling. Additionally, the photographs were divided in three segments
(frontal, orbital, and oral) to determine to what extent it is possible to generate
a smile following an N7–N12 transfer.
Material and Methods
Patients
In this retrospective cohort study, patients who underwent an N7–N12 nerve transfer
between 2001 and 2019 were included. Sixteen patients had an N7 lesion following skull
base surgery and one patient after cholesteatoma surgery (mastoidectomy). We consider
the N7–N12 nerve transfer as the best first step in facial reanimation in this situation
as it potentially reinnervates all muscles of the face. Clinical data were collected
from medical records, including the cause of the loss of facial nerve function, interval
between facial paralysis and surgical reconstruction, outcome of H-B grading, and
complications during reconstruction.[18]
Patients were excluded when (1) the follow-up was less than 1 year; (2) the facial
nerve deficit occurred following resections of malignant tumors; (3) postoperative
photographs were unavailable; and (4) major static procedures were additionally performed
(e.g., forehead lift).
Digital photographs of the entire face were made by a clinical photographer, and if
not present, they were provided by the patients following instructions. The patients
were asked to keep the face at rest and to smile to the best of their ability as they
would normally do. The photographs of the entire face were digitally divided into
three segments: frontal, orbital, and oral ([Fig. 1]). The boundaries of the orbital segment were just cranial and caudal to the supra-
and infraorbital margins covering the area of the orbicularis oculi muscle. The frontal
segment was the part cranial to the orbital segment, and the oral segment was the
part caudal to the orbital segment.
Fig. 1 of photographs and segments (a,b) at rest and (c,d) in the active phase.
All the segmented photographs at rest and while smiling were mixed at random and separated
from those of the entire face.
The photographs were assessed by five medical doctors individually (2 neurosurgeons,
2 ear, nose, and throat [ENT] surgeons, and 1 ENT resident) who were blinded to the
side of the N12–N7 transfer. First, the segmented photographs were assessed, and 1
week later those of the entire face. The assessors were asked to indicate whether
the face was symmetrical or asymmetrical and if asymmetrical to identify the affected
side. If the face was asymmetrical, they had to score whether it was mildly or severely
disfiguring. The identification of the affected side was compared with the clinical
data, and defined as correct or incorrect.
This study was evaluated by the medical ethics committee of the Leiden University
Medical Center (LUMC). The committee judged that medical ethical review was not required
because of the retrospective nature, and patients were not subjected to any procedures
and/or behavioral restriction.
Surgical Technique of N12–N7 Transfer
The surgical technique applied was extensively described previously.[2] In short, the extratemporal portion of the facial nerve was identified via a parotid
incision, using the posterior belly of the digastric muscle and tragal pointer. The
vertical part of the facial canal in the mastoid bone was unroofed. The intratemporal
part of the facial nerve was mobilized and transected at the external facial nerve
(second) genu. The hypoglossal nerve was identified at the level of the carotid bifurcation
and neurolyzed as proximally as possible. The hypoglossal nerve was partially cut
such that the exposed area corresponded to the cross-sectional area of the facial
nerve. A tensionless end-to-side coaptation between the two nerves was made using
10–0 sutures and glue.
Statistical Analysis
Statistical analysis was performed using GraphPad Prism (version 9). The assessments
in which the affected side was correctly identified were tested using Fisher's exact
test for a comparison between the assessments of the total face at rest and in the
active phase (smiling). For the segmented assessments, this was done with the chi-squared
test. The hypothesis was that the observer can identify the affected side more accurately
in a smiling patient. The interobserver variability was scored using the Fleiss kappa
analysis in IBM SPSS Statistics version 28.0.1.0. The following strength of agreement
was used: less than 0.20 poor, 0.21 to 0.40 fair, 0.41 to 0.60 moderate, 0.61 to 0.80
good, and 0.81 to 1.00 very good.[19] Pre- and postoperative H-B grading was tested for significance. A p-value of less than 0.05 was considered significant.
Results
The study comprised 11 female and 6 male patients with a mean age of 43.5 years at
the time of facial paralysis (SD ± 17.7; range: 8–68; median: 44; [Table 1]). All the patients had a facial paralysis following surgery for vestibular schwannoma
(VS; n = 12), facial schwannoma (n = 2), hemangioma (n = 1), epidermoid cyst (n = 1), and cholesteatoma (n = 1; [Fig. 2a]). The average interval between facial paralysis and reconstructive surgery was 5.2
months (SD ± 4.6; range: 0–15; median: 4). Four patients with preserved facial nerve
continuity during VS resection had a longer interval (average: 10 months) as compared
with the overall average, reflecting the time that passed to assess whether potential
spontaneous recovery would occur. Two patients had surgery and/or facial nerve reconstructions
elsewhere (patient nos. 3 and 5) and were reconstructed late (>12 months). In one
of these patients (no. 5) initially an end-to-end coaptation of the facial nerve was
performed and in the second instance the N7–N12 transfer was performed. All patients
scored H-B VI prior to the N12–N7 transfer. No complications occurred following the
N12–N7 surgery. The H-B grading was performed during outpatient visits with a mean
postnerve transfer interval of 62.5 months (SD ± 49.8; range: 17–172; median: 40).
Postoperatively, 13 patients improved to H-B grade III (76%), 1 patient to grade IV
(6%), 1 to grade V (6%), and in 2 patients facial function did not recover, with a
persisting H-B grade VI (12%; [Table 1], [Fig. 2b]). Patients 3 and 5, who underwent a reconstruction late had postoperative H-B grade
VI (no. 3) and IV (no. 5). Of the two facial schwannoma patients, one had postoperative
H-B grade V and one had a postoperative H-B grade VI. A gold weight was inserted in
the upper eyelid of the affected side to improve closure in three patients. Tarsorrhaphies
were performed in three patients. Four patients had synkinesis, which was treated
with botulinum toxin. None of the patients perceived loss of function of the tongue.
Table 1
Results and patient characteristics of the patients who underwent a hypoglossal to
facial nerve transfer
|
Patient no.
|
Age (y)
|
Gender
|
Pathology
|
Preoperative HB
|
Postoperative HB
|
Interval
lesion-surgery (mo)
|
FU (mo)
|
|
1
|
43
|
F
|
VS[a]
|
6
|
3
|
4
|
38
|
|
2
|
63
|
F
|
VS
|
6
|
3[b]
|
1
|
117
|
|
3
|
18
|
F
|
VS[a]
|
6
|
6[c]
[d]
|
15
|
172
|
|
4
|
37
|
F
|
VS
|
6
|
3[b]
[c]
|
4
|
128
|
|
5
|
8
|
M
|
Cholesteatoma
|
6
|
4[c]
|
13
|
93
|
|
6
|
21
|
F
|
VS
|
5
|
3
|
2
|
20
|
|
7
|
51
|
F
|
VS
|
6
|
3
|
2
|
21
|
|
8
|
68
|
M
|
VS
|
6
|
3[d]
|
1
|
59
|
|
9
|
37
|
M
|
FS
|
6
|
5[d]
|
5
|
22
|
|
10
|
57
|
F
|
VS
|
6
|
3
|
11
|
26
|
|
11
|
44
|
M
|
VS[a]
|
6
|
3
|
7
|
80
|
|
12
|
55
|
M
|
Hemangioma
|
6
|
3[b]
|
2
|
40
|
|
13
|
43
|
F
|
VS
|
6
|
3
|
2
|
117
|
|
14
|
63
|
M
|
FS
|
6
|
6
|
4
|
18
|
|
15
|
20
|
F
|
Epidermoid cyst
|
6
|
3
|
7
|
17
|
|
16
|
55
|
F
|
VS[a]
|
6
|
3
|
9
|
19
|
|
17
|
55
|
F
|
VS
|
6
|
3[b]
|
2
|
94
|
Abbreviations: F, female; FS, facial schwannoma; FU, routine follow-up in months;
HB, House–Brackmann; M, male; VS, vestibular schwannoma.
a Neuropraxia.
b Botulinum toxin injections.
c Gold weight eyelid.
d Tarsorrhaphy.
Fig. 2 (a) Overview of different pathologies per number of patients resulting in facial nerve
deficit. VS, vestibular schwannoma; FS, facial schwannoma. (b) Pre- and postoperative House–Brackmann classification.
Postoperative photographs were made with a mean interval of 93 months after the N12–N7
reconstruction (SD ± 67.3; range: 12–212; median: 108). In total, 85 assessments of
photographs at rest and during smile were performed. For the frontal, orbital, and
oral segments, 255 assessments (17 patients, 3 segments, and 5 accessors) both at
rest and during smile were performed.
The results of the photographic analyses are shown in [Fig. 3] and [Tables 2],[3],[4]. The total face at rest was symmetrical in 22 of 85 (26 %) patients. The affected
side was significantly less well identified at rest as compared with during smiling
(48/63 vs. 75/80; p = 0.003; [Fig. 3a]). Asymmetry (n = 63) was judged as mildly disfiguring in 48/63 (76%) of the patients. When smiling,
asymmetry was scored as severely disfiguring in 30/80 (38%) patients, which was 15/63
(24%) patients at rest.
Fig. 3 (a) Results of photographical analysis of the total face, showing the proportion of
correct and incorrect scores by the five medical observers. A total of 17 patients
were scored by five observers. Fisher's exact test shows a significant difference
between the two phases. (b) Results of the photographical analysis of three segments, showing the proportional
correct versus incorrect per segment at rest versus the active phase. The chi-squared
test shows a significant difference in all three segments.
Table 2
Symmetry assessment of the entire face at rest and during smile (active) following
hypoglossal to facial nerve transfer (n = 85). Correct identification of the affected side was only scored with asymmetry
|
Symmetry
|
Asymmetry
|
Correct identification
|
|
Rest
|
22 (26%)
|
63 (74%)
|
48 (76%)
|
|
Active
|
5 (6%)
|
80 (94%)
|
75 (94%)
|
Table 3
Symmetry and level of disfigurement in photographical analysis of the entire face
following a hypoglossal to facial nerve transfer, at rest and during smile (active),
n = 85
|
Symmetry
|
Disfigurement
|
|
Mild
|
Severe
|
|
Rest
|
22 (26%)
|
48 (56%)
|
15 (18%)
|
|
Active
|
5 (6%)
|
50 (59%)
|
30 (35%)
|
Table 4
Assessment of the frontal, orbital, and oral segment photographs
|
|
Symmetry, n (%)
|
Asymmetry, n (%)
|
Correct, n (%)
|
|
Frontal
|
Rest
|
54 (64)
|
31 (36)
|
27 (87)
|
|
Active
|
57 (67)
|
28(33)
|
25 (89)
|
|
Orbital[a]
|
Rest
|
17 (20)[a]
|
68 (80)
|
47 (69)
|
|
Active
|
13 (15)[a]
|
72 (85)
|
64 (89)
|
|
Oral
|
Rest
|
30 (35)[a]
|
55 (65)
|
40 (73)
|
|
active
|
5 (6)[a]
|
80(94)
|
77 (96)
|
a Significant difference between the orbital and oral symmetrical segments (p = 0,012) using Fisher's exact test. Correct identification of the affected segment
was only scored in asymmetric faces.
In the analysis of the segments, the oral segment during smiling was asymmetrical
in 80 of 85 (94%) patients. At rest, the oral segment was symmetrical in 30 of 85
(35%) patients. The affected side in the oral segment during smiling was correctly
identified in 77 (96%) patients. The frontal segment was symmetrical at rest in 54
(64%) patients and during smiling in 57 (67%) patients. The orbital segment during
smiling was asymmetrical in 72/85 (85%) patients. In the rest phase, the orbital segment
was symmetrical in 17/85 (20%) patients. There was a significant difference (p = 0.012) between the orbital and oral segments regarding symmetry at rest and during
smiling ([Table 4]). The identification of the affected side in the active phase in all three segments
differed significantly from the rest phase (p < 0.0001; [Fig. 3b]).
The results of the interobserver variability using the Fleiss kappa analysis showed
that the level of agreement during smile was good in the total face (0.771) and in
the oral segment (0.641) and moderate in the orbital segment (0.420; [Table 5]).
Table 5
Interobserver variability scores using Fleiss' kappa analysis
|
Assessment
|
Intraclass correlation (kappa)
|
95% confidence interval
|
|
Lower bound
|
Upper bound
|
|
Rest total face
|
0.167
|
0.057
|
0.277
|
|
Active total face
|
0.771
|
0.636
|
0.906
|
|
Score of total face, rest
|
0.215
|
0.105
|
0.326
|
|
Score of total face, active
|
0.363
|
0.234
|
0.491
|
|
Rest frontal segment
|
0.313
|
0.219
|
0.406
|
|
Rest orbital segment
|
0.134
|
0.031
|
0.238
|
|
Rest oral segment
|
0.326
|
0.217
|
0.436
|
|
Active frontal segment
|
0.206
|
0.110
|
0.302
|
|
Active orbital segment
|
0.420
|
0.307
|
0.533
|
|
Active oral segment
|
0.641
|
0.513
|
0.769
|
Discussion
The N12–N7 transfer is widely used to treat the sequelae of a facial nerve lesion.
However, not much is known about the functional recovery of facial muscles. This is
due to the fact that the grading systems that have been applied in the reports on
outcome of the N12–N7 transfer do not clearly differentiate between the face at rest
and during an active phase, such as smiling. In this study, we found that during smiling,
professionals could correctly assess the affected side in more than 90% of cases.
There was a 20% increase of asymmetry between rest and smile which was mainly caused
by the oral segment of the face and to a lesser extent by the orbital region. These
findings suggest that application of additional dynamic procedures to improve the
oral segment may be a logical first step to improve smiling after the N12–N7 transfer.
The affected side with the face at rest was correctly identified in only 76%. Apparently,
it was not evident to distinguish which side of the face was normal and which was
reinnervated by the N12–N7 transfer. This might indicate that the appearance of what
was mistakenly perceived as the unaffected side, but actually was the N12–N7 reinnervated
side, cannot be grossly abnormal. However, the N12–N7 transfer results in a combination
of flaccid paralysis components mixed with synkinetic activity, superimposed on faces
that may also demonstrate normal aging phenomena.
In this study, we asked the patients to smile as they would normally do to the best
of their ability. Providing these instructions did not lead to an active smile. It
is worth noting that patients can generate a smile following an N12–N7 transfer. To
do so, they have to consciously and forcefully push the tongue against the hard palate.
Thereby, the original motor program of the tongue is used to activate the facial muscles.
Apparently, the central program to activate a spontaneous smile does not activate
hypoglossal motoneurons, which would require central plastic changes to occur.[20]
The insufficient activity of the oral segment after the N12–N7 transfer may be improved
by additional static or dynamic techniques.[4]
[7]
[8]
[10] One option that we currently use is to combine the N12–N7 transfer with a transfer
of the masseteric nerve branch (N5, trigeminal nerve) to the oral branch of the facial
nerve. An N5–N7 transfer alone does not provide symmetry at rest as good as the N12–N7
transfer. Therefore, a combination might prove optimal.[17] To create a smile after an N5–N7 transfer, however, one has to close the jaw. Although
this is also different from spontaneous smiling, this action comes closer to a natural
smile. Additionally, clenching the teeth to smile is easier to perform than pushing
the tongue against the palate.
Cross-facial nerve grafting is another option to reanimate the facial musculature.
If the facial musculature is withered, it is one of the very few options, but should
be accompanied with a gracilis muscle transfer to regain dynamic function. This technique
provides a positive trend in disease-specific quality of life.[21] However, the cross-facial technique is complex and requires multiple surgeries,
each of which has failure rates. These factors have to be taken into consideration
when choosing the best type of treatment, and these should be discussed with the patients
to achieve optimal shared decision-making.
In a previous study of our group, we reported the outcome of the N7–12 transfer procedures
using the H-B grading system. In that study, 86% of the patients had a H-B grade III
in contrast to 76% in the present study.[2] The difference can be explained by the fact that in our earlier report, patients
with facial schwannomas were excluded. Facial schwannoma causes a slowly progressing
paralysis, which usually takes years to develop. Irreversible atrophy of a part of
the facial musculature occurs over time, excluding muscle fibers for reinnervation
by a nerve transfer that causes a negative impact on the outcome.[22]
[23]
[24] If we exclude the patients with a facial schwannoma, the overall H-B grade III score
in the remaining series increases to 87%, which is comparable with earlier reports.[6]
[11] Optimally, facial reinnervation following a complete injury right from the start
is performed within 6 months after onset.[5]
[8] Since the process of facial nerve function deterioration is a gradual process in
case of a facial nerve schwannoma, patient counseling with regard to the timing of
nerve transfer is key for good outcome.
The weaknesses of this study are the relatively small number of patients, the fact
that intraobserver variability was not assessed, and the retrospective nature of the
study. The strength is that the outcome was independently performed by five assessors.
In addition, the method of segmental analysis of the reconstructed (N12–N7) face provides
deeper insight into the contribution of the frontal, orbital, and oral parts of the
face to obtain symmetry and the generation of a smile. In our opinion, this study,
using (segmented) photographs and observer assessments, is unique. Other studies concerning
facial reanimation and postoperative results use different scoring systems, which
are categorized in observational, mathematical, and computer-graphical measurements.[14] Nevertheless, this study addresses the question of whether an N7–N12 transfer generates
a good smile, which was observed by five medical assessors.
Conclusion
Following an N7–N12 transfer, the majority of patients achieve a good symmetry of
the face at rest, but they are unable to generate a natural smile. Static and dynamic
analysis of the facial nerve–innervated muscle function not only is essential to adequately
evaluate the outcome of facial nerve reconstructions but also provides clues that
additional dynamic procedures may be required to improve the overall outcome.
