Hembd, Austin M.D.; Nagarkar, Purushottam M.D.; Perez, Justin M.D.; Gassman, Andrew M.D.; Tolley, Philip B.S., M.S.; Reisch, Joan Ph.D.; White, Charles L. III M.D.; Rozen, Shai M. M.D.

A two-stage facial reanimation procedure with a cross-facial nerve graft followed by a free muscle transfer is an approach to treat longstanding facial palsy that offers the advantages of a synchronous, spontaneous, emotional smile.¹ However, unreliable or suboptimal results have been noted to occur in the older patient when using this approach.²,³ The cause of this pattern is likely multifactorial: research in rodent models and humans has shown that axonal regeneration and total axon counts in peripheral and central nerves can decrease with age.^4–14 The latter is an imperative focus in the midst of growing facial reanimation literature that links improved function of free muscle transfers with increased axon input.³,^15–20

We hypothesize that a decrease in axonal load at the seventh cranial nerve donor branches for cross-facial nerve grafts is one cause of the decreased reliability of two-stage facial reanimation in the older patient. However, many prior histomorphometric analyses of the human facial nerve in relation to age have used small sample sizes, and have been near the brainstem, thus including more proximally bound motor axons, or both.⁷,⁹,¹²,¹³,¹⁷ This study used 126 hemifaces to characterize the relationship between age and facial nerve motor axon counts at a site with direct implications for cross-face nerve graft coaptation.

MATERIALS AND METHODS

Sixty-three fresh, cadaveric heads were dissected to expose the facial nerve from the stylomastoid foramen, through the parotid, into the entry point of the zygomaticus major muscle. For each hemiface, two cross-sectional facial nerve samples were taken: one proximal as the nerve exits the stylomastoid foramen, and one distal at a point 1 cm posterior to the anterior parotid border along the buccal and zygomatic branches. We traced these branches to ensure that they eventually inserted into the zygomaticus major muscle. Nerve samples were preserved in 10% neutral buffered formalin and embedded in paraffin, and the proximal end of each sample was sectioned onto a glass slide and stained with SMI-31 antineurofilament stain (Fig. 1).

Fig. 1

Stained specimens were then scanned using an Aperio ScanScope CS (Leica Biosystems Inc., Buffalo Grove, Ill.) whole slide scanner and axonal counts were measured using ImageScope software (Leica Biosystems). Only axons greater than 1 μm were counted.²¹ Scanned images were reviewed to verify adequate and uniform staining and for appropriate, transverse sections through the nerve. Quality control was strict to ensure the most accurate axon quantification: specimens that had any oblique or longitudinal portions of nerve tissue, or specimens that had any nondiscrete immunohistochemical staining, were identified and the process was repeated. If nerve tissue had failed to be adequately sectioned or stained after the second round of sampling, we elected to omit the specimen, as each sectioning process requires another 5 mm of nerve tissue for the paraffin-embedding process, and we wanted to maintain consistent anatomical sampling points. Manual verification of counts was carried out on a randomly selected sample of sections.

The University of Texas Southwestern Division of Biostatistics completed the statistical analysis. Correlation analysis between age and axon counts was completed using a Pearson correlation coefficient at both proximal and distal points on each hemiface. Unpaired t tests were used to analyze whether there was a difference between axon counts at the proximal or distal sampling point in male versus female cadavers or left versus right laterality. Unpaired t tests were also used to analyze the difference of means at the proximal or distal sampling points between these different age cohorts.

RESULTS

Thirty-six female and 27 male cadavers were dissected, with an average age of 71 years (range, 22 to 97 years). No cause of death directly or indirectly affected the nervous system and thus the facial nerves were considered to be normal. At the distal sampling point (zygomatic or buccal branch, 1 cm proximal to the anterior parotid border), there was a significant negative correlation between age and axonal load (r = −0.45; p < 0.0001; n = 114) (Fig. 2). At the proximal sampling point (the nerve at the styloid foramen), there was also a significant negative correlation between age and axonal load (r = −0.26; p < 0.01; n = 104) (Fig. 3).

Fig. 2

Fig. 3

There was no significant statistical difference of means between left and right axon counts at the proximal (p = 0.77) or distal (p = 0.22) sampling point. In addition, there was no significant statistical difference of means between female and male axon counts at the proximal (p = 0.07) or distal (p = 0.20) sampling point. Overall, we omitted 12 specimens from the distal sampling point and 22 specimens from the proximal sampling point that did not provide adequate nerve tissue after repeated sectioning and analysis according to our algorithm described above. Average axonal loads at the proximal and distal sampling points from cadavers aged 22 to 59 (n = 26), 61 to 79 (n = 53), and 80 to 97 (n = 33) years are depicted in Figure 4.

Fig. 4

We divided our cadaveric specimens into these three chronologic age groups to ascertain whether there was a true difference in axonal counts between successively aging decades. Exact divisions were based on attempts to create roughly equivalent age spans and attempting to maintain adequate specimen numbers in each cohort to allow statistically significant comparisons.

The difference of mean axon counts at the distal sampling point was statistically significant between both the youngest and middle-aged cohorts (p = 0.02), and the middle-aged and oldest cohorts (p = 0.002). The difference of mean axon counts at the proximal sampling point was statistically significant between the youngest and oldest cohorts only (p = 0.03).

DISCUSSION

The observation that free muscle transfers innervated by masseteric nerve transfer have better excursion at the oral commissure than those powered by cross-facial nerve grafts has been confirmed.^22–24 In addition, it has been shown that the masseteric nerve has greater axonal load than facial nerve donor branches and distal ends of cross-facial nerve grafts.²⁵,²⁶ Indeed, increasing donor branch axonal load or adding sensory input has been shown to improve functional outcomes and motoneuron regeneration with cross-face nerve grafts, respectively.³,¹⁵,²⁷ With the understanding that donor axon counts have a directly proportional relationship with the outcomes of one- or two-stage procedures with a free muscle transfer, it is prudent to consider insufficient axonal load as a possible explanation for why facial reanimation in older patients is less successful.

Kullman et al.¹⁴ and Fujii and Goto⁷ correlated facial nerve axon/fiber counts with increasing age, the latter of which found a negative correlation. Both, however, used sample sizes of fewer than 11 cadavers and were analyzing portions of the facial nerve at or proximal to the stylomastoid foramen. More recently, a larger study of 20 cadavers analyzed facial nerves taken near the brainstem and showed a marked decrease in motoneuron counts with increasing age, and this was postulated as a cause of worsening recovery from Bell palsy in the elderly.⁹ In addition, age was associated with decreasing nerve regeneration rates in facial reanimation patients measured by time until the Tinel sign at the distal cross-face nerve graft.⁴ In contrast, Jacobs et al.²⁰ did not see a decrease in facial nerve regeneration or in axons with increasing age. The latter study used a larger sample size (30 patients), but the only correlation of age and axon counts was from distal ends of nerve grafts after coaptation and not the donor zygomatic/buccal branch axon counts. Therefore, that analysis was focused more on regenerative changes with age and not donor branch axon count changes with age.

Terzis et al.³ used a large sample size (69 patients) and conducted a histomorphometric analysis of donor branch specimens, and correlated age and donor axon counts each with functional outcomes separately, but not together. They showed that although some older patients, in addition to patients with fewer than 900 axons, had good functional results, overall, age was significantly different between patients with poor to moderate versus good to excellent functional outcomes in those with greater than 900 axons.

Using animal models, Streppel et al.¹² showed delayed regrowth of facial axons after transection in aged rats, and Sturrock¹³ found decreased neuronal counts in the facial nerve motor nucleus of aged mice. Interestingly, the latter study showed that instead of a gradual decrease of neurons with age, neuronal counts in mice remained constant until a threshold of age, after which the counts fell precipitously. This has been shown in human sural nerves as well, with degeneration and demyelination of myelinated fibers increasing after the age of 60 years.⁸ Sturrock also stated, in contrast to studies showing a relatively constant neuron number of the trigeminal motor nucleus, that the decrease was perhaps attributable to mimetic musculature and platysma being more prone to aging changes, which might be reflected in fewer motoneurons, compared with muscles of mastication.

We have previously described that an approximate 1-cm retrograde intraparotid dissection would provide a greater than 90 percent chance of selecting a donor buccal or zygomatic branch with over 900 axons, thus possibly increasing the chances for an improved functional outcome.³,²⁸ This is our preferred site of coaptation of cross-facial nerve grafts and thus how we chose the location of our distal sampling point. The negative linear correlation between age and axon count that our data shows (Fig. 2) is therefore at a relevant and anatomically reproducible site for cross-facial nerve coaptation.

We analyzed samples from the proximal sampling point at the stylomastoid foramen as a quality control measure to support that there is a real decrease in axonal load occurring with increasing age. If just the distal sampling point showed a decrease in axonal counts with age, but not the proximal sampling point, it could potentially call into question the quality of our methodology. This does not appear to be the case.

We found an average of 1670 ± 973 total axons at the distal sampling site along the buccal branch. This is consistent with mean axon counts at donor buccal branches published by Placheta et al. (1826 axons) and Jacobs et al. (1736 axons), despite using different staining modalities.²⁰,²⁹ At our proximal sampling point at the stylomastoid foramen, we found an average of 5329 ± 1376 axons. Van Buskirk found the mean number of axons in the facial nerve at the brainstem to be 6999,³⁰ and Fujii and Goto⁷ similarly found the mean number of axons to be 6254 when using a Luxol, periodic acid–Schiff, hematoxylin stain. Similarly, results by Kondo et al. showed an average of 6245 facial nerve axons near the brainstem.⁹ Our sampling point was more distal to the brainstem, and thus we conclude that our average of 5329 axons is consistent with previous studies. In addition, there was no difference in axonal counts between female and male cadavers or between left and right facial nerves, which is consistent with previously published data not finding a gender or laterality difference with motoneuron counts near the facial nerve nucleus.⁹

Although our study included samples from cadavers aged 22 to 97 years, the median age was 72 years, and only 11 percent (28 hemifaces) of our cadaveric specimens were younger than 60 years. Ideally, we would use equal numbers of cadavers in every decade of life to more completely determine the trend of the changes of axonal load with age over time. However, as we were accessing these cadavers through the Dallas Willed Body Program, most donors are of older age. Despite this limitation, our data displayed a significant down-trending of axonal load from the third decade and beyond at a surgically relevant location for facial reanimation.

Because of the current unreliability of cross-facial nerve grafting in the elderly population, particularly in the seventh and eighth decades of life, we currently opt to use the one-stage reanimation approach with a functional free muscle transfer of the gracilis to the masseteric nerve in these patients. We choose this donor nerve for the reasons of improved axonal load and excursion as discussed above.

However, the masseteric nerve has drawbacks, namely, an animation deformity while chewing and less reliability in providing the emotional, spontaneous smile that the facial nerve does to the functional free muscle flap. Although there are a number of reasons why the masseteric nerve could provide a more robust result (e.g., less distance required for nerve regeneration, only one suture line compared with the two-stage cross-facial nerve graft reanimation technique), the exact causes have not been elucidated. Our main goal with this study was to characterize with high-quality data whether decreasing axonal load could be a plausible cause of the lack of efficacy of cross-facial nerve grafts in the elderly. We conclude that it could very well be a cause.

We do not recommend the proximal nerve point we took samples from in this study as a plausible cross-facial nerve graft coaptation site, as going more proximal than the pes anserinus would come at the cost of decreasing specificity of the neurotized free muscle graft. We suggest that a coaptation site on the donor buccal branch near the area of our distal sampling point (1 cm proximal to the anterior parotid border within the substance of the parotid gland) can increase axonal load²⁸ which, by virtue of the trend shown in this study, is particularly essential with aging facial palsy patients.

Whether an intraparotid coaptation in a two-stage reanimation technique may be plausible in an older age group (seventh or eighth decade of life) is difficult to answer. In our opinion, this is not the only solution, as nerve regeneration in this group may be so poor that it is insurmountable by increasing the donor axonal load alone. Nevertheless, this information could possibly be used in patients in the fifth or sixth decade of life and younger to improve outcomes.

CONCLUSIONS

As age increases, the axonal load of the facial nerve at the zygomatic and buccal branches decreases. These results propose that decreasing axonal load can be a factor in less reliable outcomes of cross-facial grafting in the aging population. Moreover, this underscores the importance of recruiting more donor axons in attempting to improve facial reanimation in the older patient.

ACKNOWLEDGMENTS

The authors would like to acknowledge Ping Shang for performing the microtomy and immunostaining, and Kathryn Bartush, Keri Phillips, and Jeff Harris for performing the histologic processing of nerve tissue. They also thank the David M. Crowley Foundation for their continued support.

REFERENCES