Introduction

Many injectable materials such as teflon, collagen, silicone and gelfoam 1-3 have been used to medialize the vocal fold in cases of laryngeal paralysis with commitment of glottic closure. More recently, fat has been used as a viable alternative, since it has a number of qualities: it is autologous and easy to collect by liposuction, presents minimum risk of forming granuloma and of foreign body reaction or hypersensitivity 4, 5.

However, the use of fat in intracordal injections also has its limitations. The main one is the inability to predict the final result of medialization, since fat can be reabsorbed at variable percentages (30 to 50%), be it injected to compensate future losses by reabsorption 4, 5, 7.

Another factor that hinders an estimate of the final outcome is the lack of knowledge of the space the injected fat is going to occupy in the vocal fold and what is the effect of its presence over glottic configuration, the contour of the free margin and viscoelastic properties of each vocal fold layer.

Generally speaking, it is believed that deep injections in the thyroarytenoid muscle (TA) allow the appropriate medialization of the vocal fold with little interference over the viscoelastic properties of the more superficial layers, and therefore, the vibration phenomenon. However, there is no local standardization in the vocal fold injections: if it should be conducted in one or more points, in the TA muscle or the paraglottic space, how deep it should go. Mikaelian et al. (1992)7 and Shaw et al. (1997)4 suggested that the injection should be made in the TA muscle in two or three points. Mikus et al. (1995)8 conducted the injection only on the medium third of the vocal fold, at 2mm deep. Brandenburg et al. (1996)5 preferred an injection immediately anterior and posterior to the medium third of the vocal fold. Shindo et al. (1996)9 made injections at two sites: medium third of the vocal fold and parallel to the vocal process of the arytenoid. Most of the authors do not specify depth and distance concerning the free margin in which they make the injection, only referring that it should be made in the TA muscle.

As to glottic configuration, the literature shows that medialization promoted by intracordal injection of fat is limited to anterior portions of the vocal folds and does not allow the correction of posterior chinks 10.

The objective of our study was to assess the effects of injected fat in the TA muscle at the medium site of the vocal folds and laterally to the vocal process of the arytenoid in relation to the spatial distribution of fat and glottic configuration after the injections.
Material and method

The study was conducted in six human adult larynges excised from cadavers from the Death Checking Service, Medical School, University of Sao Paulo. Four larynges were from male and two from female subjects. The mean age of donating cadavers was 45.33 years (25 to 60 years).

Larynges were removed in an interval below 24 hours after the death. The fat for intracordal injection was obtained at the same time by liposuction from the abdominal region of one of the cadavers, according to the technique described by Brandenburg et al. (1992). No larynges showed morphological abnormalities of the structures and the vocal folds and arytenoids were symmetrically moving.

Supraglottic structures were removed for adequate exposure of the glottis and the larynges were fixed in the appropriate support. The midline of the glottis was delimited through needle or thread of Nylon 2-0, through the anterior commissure up to the middle point of the cricoid cartilage lamina.

Fat collected by liposuction was decanted in gauze to remove its serous component and then introduced in the pressure syringe used for fat injections. The syringe was coupled to Gelco gauge 18.

In three larynges, the injection was conducted in the medium point of the vocal fold, in the vestibular surface, in order to reach the TA muscle. The needle was introduced about 3-4mm laterally to the free margin, and 2-4mm deep. In the other larynges, the injection was posterior, in a lateral point to the vocal process of the arytenoid, at the same depth. The needle lamina was directed laterally in all injections. In each larynx, we injected the fat slowly and in amounts that were enough for the free margin to reach the glottic midline. The injection was made in one single fold, maintaining the other one as control.

Cranial view photos of the glottis were made before and after the injection using a digital camera (Sony Mavica), sustained by a tripod.

Next, larynges were fixed in formol at 10% for 10 days and later decalcified in nitric acid at 7% for three more days. They were then cut at the frontal site in the anterior, middle and posterior aspect of the vocal fold and at the level of the arytenoid cartilage. Sections were made manually, using a disposable microtome lamina.

The cut surface was placed next to a millimeter scale and photographed with the same digital camera. The digital images were then processed by the software Corel PhotoPaint v.10, to illustrate the regions occupied by the block of fat (shown in yellow), TA muscle (in red) and the internal surface of the thyroid cartilage (in blue).

Results

Glottic configuration
The free margin of the vocal fold tended to be bulged (with a convexity) when the fat was injected in the middle third of the fold (Figure 1c). In none cases there was adduction of the vocal process of the arytenoid and closure of the cartilaginous portion of the glottis (Figure 1). This fact was observed, however, when the fat was injected laterally to the vocal process of the arytenoid (Figure 2). Conversely, injections at this site not always allowed the appropriate closure of the anterior portions of the glottis (Figure 2a). The analysis of frontal sections of the larynx showed that in cases in which there was medialization of the vocal process, the fat occupied the lateral region of the arytenoid cartilage body (Figure 3).

Distribution of injected fat
We observed that the injected fat bolus occupied a cylindrical or ovoid space, throughout the longest axis of the vocal fold. Thus, in sequential frontal sections, fat seemed to be circular or ovoid, occupying a specific area inside the TA muscle, which tended to be smaller in the ends and larger at the site in which the injection was made (Figure 4). In many larynges, the injected fat that was too deep in the TA muscle disseminated and came close to the superficial layers of the vocal folds (Figure 5). In one larynx, the fat bolus occupied the corresponding portions to the vocal fold lamina propria (Figure 5c). In posterior portions of the vocal fold, fat tended to limit to the injected muscle (TA) and not to overflow to other muscles, such as the lateral cricoarytenoid (LCA) (Figure 6). Even when injected in the lateral region of the vocal process of the arytenoid, fat tended to be anteriorly distributed, occupying the referred space (Figure 7).

Depth of vocal fold injection
In two larynges in which the superficial injection of the TA was made, the fat entered the muscle fibers and reached the peripheral area, promoting bulging of the vestibular surface of the vocal fold (Figure 8). In addition to not promoting appropriate medialization of the vocal folds, fat deposition in the region favored its extrusion through the puncture orifice (Figure 8a).






Discussion

Since fat was proposed as the material to be used for intracordal injections 11, many studies have approached the problem of reabsorption of fat in the postoperative period and, consequently, deterioration of vocal quality 4-6, 8-12. Literature, however, is poor in addressing spatial distribution of injected fat and its effects over the contour of the free margin, glottic configuration and affections of the viscoelastic properties of each layer of the vocal folds. In this study, we tried to understand these issues. The limited number of larynges used and the experimental nature of the study suggested that a transposition of results to the clinical practice should be carefully made. We can infer, however, about what can happen when we inject fat in the vocal folds. We believe that this knowledge can provide foundations to future clinical trials.

As to glottic configuration, we observed differences concerning the injection made in the medium point of the vocal fold and laterally to the vocal process of the arytenoid. In none larynges of the first group we observed medialization of the vocal process, with closure of the posterior glottis. Conversely, this fact was observed at higher or lower degree in all larynges of the second group. Therefore, we considered the fat injection to promote posterior glottic closure by medialization of the vocal process of the arytenoid, differently from what the literature says10.

In larynges in which the fat is injected in the medium site of the vocal fold, the medialization was obtained by bulging of the membranous portion of the glottis, with formation of a convexity of the free margin. In clinical practice, the more lateralized the paralyzed vocal fold is, the more fat is necessary to promote its medialization and, consequently, the higher is the convexity of the free margin. This non-flat configuration of the free margin of the vocal fold can however compromise the appropriate closure of the glottis. Studying the use of fat for the treatment of unilateral laryngeal paralysis, Shaw et al. (1997)4 reported that vocal quality in many patients improved as a result of time, attributing the improvement to reabsorption of excess fat injected and reduction of the convexity of the free margin, with consequent improvement of the glottic closure. Therefore, injections only of the medium portion of the vocal fold, when it is lateralized, lead to a paradox: in order to promote good glottic closure it can be necessary that the free margin be a little convex. On the other hand, convexity can impair glottic closure.

We considered, therefore, that the fat injection at a lateral site of the arytenoid vocal process, and in smaller amount, of the middle third of the vocal fold, similarly to that described by Shindo et al. (1996)9, can result in complete glottic closure, without being necessary to cause a marked deformation of the free margin. At least in experimental situations, this technique seems to have advantages compared to exclusive injection of the middle third of the vocal fold.

We also observed that injection at only one single site can allow complete glottic closure by tendency to distribute the fat in the anterior-posterior axis of the TA muscle. As illustrated by Figure 7, fat when injected laterally to the vocal process, can disseminate to the anterior third of the glottis, without exceeding the limits of the TA muscle in the frontal plan. TA muscle, therefore, can provide more resistance to diffusion of fat in the frontal plan than the longitudinal axis of the vocal fold. This finding is theoretically expected since the fibers of the muscle run anterior-posteriorly along the vocal folds. Each muscle fiber is attached to adjacent fibers by supported connective tissue and a proteins complex (dystrofin-glycoprotein complex) that keep them together 13. Conversely, the spaces between the muscle fibers, occupied by endomise, create corridors that provide little mechanical resistance to diffusion of substances. Thus, the restriction imposed by muscle fibers to fat diffusion is smaller in the longest axis of the muscle than perpendicular to it.

This tendency of anterior-posterior diffusion, however, does not promote migration to the injected fat in the middle third of the vocal fold to the lateral region of the vocal process of the arytenoid. As said before, no case in which the injection was made in the middle third of the vocal fold resulted in medialization of the vocal process of the arytenoid. We considered that in the extremities of the TA muscle, the muscle fibers are very close one to the other, reducing the inter-fiber space that can increase tissue resistance to diffusion of fat. Even though the fat that is injected close to the posterior extremity of the TA muscle can be diffused to the middle and anterior thirds of the vocal fold, the opposite seem to be less likely.

According to the findings of our study, we consider it is not necessary to make injections in many sites of the vocal fold as advocated by some authors 4, 5, 7, 9. Since injected fat can overflow through the injection site 8, the greater the number of injections sites, the greater the risk of overflow and the worse the control over the final volume of fat.

As to the possibility of overflow, we observed that fat occasionally reached the periphery of the TA muscle and promoted the undesirable bulging on the vestibular surface of the vocal fold. Once the fat reaches this surface, the possibility of extrusion is higher. It occurs when the fat injection is very superficial or very close to the free margin of the vocal fold and the diffusion radius of injected fat overflows the limits of the muscle. To avoid this type of situation, it is necessary that the injected fat be deep in the TA muscle.

Even following these instructions, in some cases, injected fat was diffused quite considerably, coming close to the superficial layers of the vocal fold. These findings allowed us to make theoretical considerations concerning the effects of injected fat over the structural organization in vocal fold layers and viscoelastic properties of each layer.

In Type I thyroplasty, medialization of the folds is obtained by introducing a block of silicone that pushes medially the paraglottic structures. In this case, the structural organization in vocal fold layers is maintained with progressive reduction of rigidity, as it moves towards the more superficial layers. According to Hirano (1975)14, the difference in viscoelastic properties of each layer plays an important role in the vibration characteristics of the vocal fold. This fact is observed in cases of laryngeal paralysis with marked atrophy of the TA muscle. Even if there is glottic complete closure, there is no formation of mucosa vibration wave owing to loss of support of the mucosa promoted by the TA muscle. Both the mucosa and the TA muscle are flaccid with the situation. In the case of fat, when the injected bolus is close to the superficial layers of the vocal folds, the difference in rigidity of the layers can be changed and reduced, becoming homogenous. Since fat is a soft and malleable tissue, it does not provide enough support to the mucosa during cord vibration. Thus, it is possible that the vibration characteristics of the fold that received fat injection be modified, with reduction of the formation of mucosa waves, even in the presence of complete glottic closure. In fact, upon studying the patients with unilateral paralysis of the vocal folds that were submitted to fat intracordal injection, Shaw et al. (1997)4 demonstrated that despite the significant improvement of glottic closure, patients did not present significant improvement considering the formation of mucosa waves after 1, 6 and 12 months postoperative. However, perceptual analysis (GRBAS) and subjective analyses conducted with the patients in the study showed significant improvement in postoperative vocal quality. The impairment of the mucous-undulation movement of the mucosa in such patients is a theoretical limitation whose clinical relevance should be better determined.

Conclusions

This study allowed us to conclude that in experimental conditions:

 Fat injection, when made laterally to the vocal process of the arytenoid, allowed the closure of the posterior glottis by medialization of the vocal process of the arytenoid. When made only in the medium third of the vocal fold it promoted bulging of the membranous portion, without closure of the cartilaginous portion of the vocal fold.
 The fat injected in the TA muscle tends to occupy a cylindrical space along the longest axis of the vocal fold, so that the injection in only one site can promote bulging of the whole membranous glottis. This fact could be observed when fat was injected laterally to the vocal process of the arytenoid.

References

1. Schmidt PJ, Wagenfeld D, Bridger MW, van Nostrand AW, Briant TD. Teflon injection of the vocal cord: a clinical and histopathologic study. J Otolaryngol 1980; 9(4):297-302.
2. Stein J, Eliachar I, Myles J, Munoz-Ramirez H, Strome M. Histopathologic study of alternative substances for vocal fold medialization. Ann Otol Rhinol Laryngol 2000; 109(2):221-6.
3. Ford CN, Bless DM, Campbell D. Studies of injectable soluble collagen for vocal fold augmentation. Rev Laryngol Otol Rhinol (Bord) 1987; 108(1):33-6.
4. Shaw GY, Szewczyk MA, Searle J, Woodroof J. Autologous fat injection into the vocal folds: technical considerations and long-term follow-up. Laryngoscope 1997; 107(2):177-86.
5. Brandenburg JH, Unger JM, Koschkee D. Vocal cord injection with autogenous fat: a long-term magnetic resonance imaging evaluation. Laryngoscope 1996; 106(2):174-80.
6. Horl HW, Feller AM, Biemer E. Technique for liposuction fat reimplantation and long-term volume evaluation by magnetic resonance imaging. Ann Plast Surg 1991; 26:248-58.
7. Mikaelian DO, Lowry LD, Sataloff RT. Lipoinjection for unilateral vocal cord paralysis. Laryngoscope 1991; 101(5):465-8.
8. Mikus JL, Koufman JA, Kilpatrick SE. Fate of liposuctioned and purified autologous fat injections in the canine vocal fold. Laryngoscope 1995; 105(1):17-22.
9. Shindo ML, Zaretsky LS, Rice DH. Autologous fat injection for unilateral vocal fold paralysis. Ann Otol Rhinol Laryngol 1996; 105(8):602-06.
10. Chang HP, Chang SY. Autogenous fat intracordal injection as treatment for unilateral vocal palsy. Chung Hua I Hsueh Tsa Chih (Taipei) 1996; 58(2):114-20.
11. Brandenburg JH, Kirkham W, Koschkee D. Vocal cord augmentation with autogenous fat. Laryngoscope 1992; 102(5):495-500.
12. Kononas TC, Bucky LP, Hurley C, May Jr JW. The fate of suctioned and surgically removed fat after reimplantation for soft-tissue augmentation: A volumetric and histologic study in the rabbit. Plast Reconstr Surg 1993; 91(5):763-8.
13. Barohn RJ. In: Goldman L, Bennett, JC. Cecil - Textbook of Medicine, 21st. ed. Philadelphia: W.B. Saunders; 2000. p.2201-2.
14. Hirano, M. Phonosurgery. Basic and Clinical investigations. Otologia (Fukuoka) 1975; 21:239-440.

---

1Ph.D., Assistant physician, Division of Clinical Otorhinolaryngology, Hospital das Clínicas, Medical School, University of Sao Paulo.
2Full Professor, Discipline of Otorhinolaryngology, Medical School, University of Sao Paulo.
3Ph.D. studies under course, Discipline of Otorhinolaryngology, Medical School, University of Sao Paulo.
4Ph.D., Otorhinolaryngologist, Medical School, University of Sao Paulo.
5Full Professor, Discipline of Otorhinolaryngology, Medical School, University of Sao Paulo.
Study conducted at the Division of Clinical Otorhinolaryngology, Hospital das Clínicas, Medical School, University of Sao Paulo.
Address correspondence to: Dr. Rui Imamura - Av. Liberdade, 91 conj. 91/3
CEP: 01503-000 Sao Paulo, SP, Brazil
Study presented at 36o. Congresso Brasileiro de Otorrinolaringologia, held on November 18 - 23, 2002, in Florianópolis - SC.
Article submitted on May 30, 2003. Article accepted on July 17, 2003.