**3. Selection of animal models**

118 Otolaryngology

then healing can occur by secondary intention. In this case granulation tissue formation and epithelial migration occur (Woo 1995), and there is correspondingly more scar tissue formation Voice rest is usually prescribed after surgery (Ishikawa and Thibeault 2010), but even with a totally compliant patient, apposition of epithelial flaps edges can be difficult to maintain. Thus various methods like micro-sutures and fibrin glue (Bleach, Milford et al. 1997; Flock 2005; Kitahara, Masuda et al. 2005; Finck, Harmegnies et al. 2010; Skodacek, Arnold et al. 2011) have been used to improve wound closure and minimize scar tissue

The use of microsutures in vocal fold wound closure was proposed by Woo et al in 1995, hypothesizing that microsutures would allow precise positioning of wound edges and maintenance of the approximation (Woo 1995). This would reduce exposure of the wound site and permit primary healing to occur. They carried out the procedure in 18 patients, finding improved voice results after surgery. As there was no control group and basis for comparison in Woo et al's study, Fleming et al attempted to compare the amount of scar formation with and without microsutures in a canine model (Fleming, McGuff et al. 2001). A small sample group of 4 dogs were used, with bilateral microflap defects created in each dog. 6-0 fast absorbing gut sutures were used to close the microflap on only one side, leaving the contralateral side unclosed. The amount of scar was evaluated between 39 and 49 days post surgery. Un-sutured vocal folds were found to have at around 75% larger scar formation than sutured vocal folds, concurring with Woo et al's hypothesis that the use of

Fleming et al also identified the length of time required for suture placement as the main disadvantage of this technique, suggesting that practice and familiarization with the technique using larger sutures before actual surgery could help mitigate the learning curve. Tsuji et al recently proposed an improvement to the microsuture technique (Tsuji, Nita et al. 2009) by pre-tying a small length of 4-0 non-absorbable nylon suture to the free end of a 7-0 absorbable suture. The nylon acted as an anchor at the epithelial surface, preventing the thread from escaping and removing the need for an assistant surgeon to maintain tension on the free end of the suture. This improved the ease of performing the technique. Their new technique was tested on human cadaveric larynges for a total of 10 sutures and they

Despite good wound healing results demonstrated by micro-sutures, many surgeons prefer using adhesives to hold down epithelial flaps to achieve wound closure. Tissue adhesives such as cyanoacrylates and fibrin glue have been used (Flock 2005) and may be easier to apply than that of sutures. A potential limitation of tissue adhesives includes increased scar tissue formation if glue accumulates between the epithelial edges preventing proper approximation, or by adhering the epithelium to the underlying connective tissue without proper reformation of the intervening layered structure. Rapid curing can also restrict the surgeon from re-apposing malpositioned flaps. Lack of tensile strength of the adhesive is another concern. Fibrin glue can take several minutes for initiation of curing and several hours to develop its full strength. Especially during the curing phase it may not possess

formation.

**2.3.1 Microsutures** 

microsutures improves postoperative wound healing.

reported a placement time of 5 to 7 minutes per suture.

**2.3.2 Tissue adhesives** 

Experimentation on live humans is not possible or ethical in most situations. Cadaveric human larynges are also difficult or expensive to obtain. Hence, when studying a new technique or device, an animal model can provide a systematic platform for the experimentation and validation. However due to differences in vocal fold size and structure, one animal may not suit all research requirements.

Depending on the research question to be addressed and the methodological approach, these differences can limit applicability of experimental results. Selection of an appropriate animal model needs careful consideration. Practical issues like size, availability of animal, availability of the facilities to house or carry out the procedures, procurement cost and maintenance of the animal for the duration of the study can restrict researchers from acquiring their ideal animal model.

Characteristics of particular interest when considering operative techniques include the size, shape and position of the larynx and other upper airway structures to simulate surgical access. Similarity of vocal fold shape and location is essential for testing microsurgical techniques, while similar tissue composition is necessary when assessing in-vivo behaviour of implanted materials and tissue responses.

Various animal models have been used extensively in vocal fold studies with their results compared across models (Garrett, Coleman et al. 2000; Jiang, Raviv et al. 2001; Titze and Alipour 2006; Alipour and Jaiswal 2007; Alipour and Jaiswal 2008; Bless and Welham 2010; Alipour, Jaiswal et al. 2011). Three of the more commonly used models in operative studies are rabbits (Thibeault, Gray et al. 2002; Thibeault, Bless et al. 2003; Branski, Rosen et al. 2005; Carneiro and Scapini 2009; Campagnolo, Tsuji et al. 2010), dogs (Garrett, Coleman et al. 2000; Fleming, McGuff et al. 2001; Rousseau, Hirano et al. 2003; Karajanagi, Lopez-Guerra et al. 2011) and pigs (Blakeslee, Banks et al. 1995; Garrett, Coleman et al. 2000; Jiang, Raviv et al. 2001; Alipour and Jaiswal 2008; Fonseca, Malafaia et al. 2010).

#### **3.1 Rabbit models**

Due to their docile nature, relatively abundant numbers, ease of housing and management, rabbits are popular animal models. Rabbits are often used in immunological studies and exhibit similar vocal fold histology to humans. However access to rabbit vocal folds by standard suspension laryngoscopy is limited due to the smaller size of the rabbit larynx. Carneiro et al (Carneiro and Scapini 2009) used rabbits to study vocal fold grafts by exposing their vocal folds via a neck incision and laryngofissure. Branski et al (Branski, Rosen et al. 2005) studied the healing process of rabbit vocal fold after injury, using a neonatal laryngoscope to access the vocal fold. Campagnolo et al (Campagnolo, Tsuji et al. 2010) studied the healing effects of injectable corticosteroids after vocal fold surgery using a custom made laryngoscope for access.

Investigation of Experimental Wound Closure Techniques in Voice Microsurgery 121

However, from our experience with intubated animals, neither of these features created a

Regner used high-speed digital imaging to compare the vocal fold vibratory characteristics of ex-vivo bovine, canine, ovine, and porcine larynges with human vocal folds. By measuring amplitude, oscillation frequency, and phase difference of vocal fold vibration, it was concluded that canine and porcine larynges are the most appropriate models for vibratory or kinetic studies on phonation (Regner, Robitaille et al. 2010). Alipour also studied vibratory characteristics of excised pig, cow and sheep larynges, and concluded that the porcine larynx had the highest range of phonation frequencies, making it a good

In a similar study, Jiang et al. (Jiang, Raviv et al. 2001) concluded that pigs models provided the most similarity in vocal fold stiffness and was a reasonable alternative for phonation studies. As pigs are a common livestock, the high availability of pig larynges from local abattoirs poses less of an ethical concern for sacrificing animals for research purposes.

Extensive ex-vivo experiments have been carried out for phonation studies (Regner and Jiang ; Jiang, Zhang et al. 2003; Skodacek, Arnold et al. 2011), for modeling the vibratory dynamics of the vocal folds. These experiments allow precise and independent control of various parameters affecting phonation, enabling systematic investigation and

A typical setup of such experimental systems consists of a mounting assembly, a pseudo lung, humidifiers, thermometers, flow and pressure meters. The mounting assembly where the excised larynx is housed consists of one lateral pronged micromanipulator sutured to the anterior tip of the thyroid cartilage and two other micromanipulators attached bilaterally to the arytenoid cartilages. This allows the elongation of the vocal folds to be controlled precisely. Airflow is generated by either an internal building source or a conventional compressor and is conditioned by heaters/humidifiers in order to prevent the larynges from drying out. The excised larynx is clamped directly to a tube from the pseudo lung and flowmeters and pressure meters are used to measure subglottal airflow and pressure before entry to the larynx. This experimental system can be easily adapted for use in ex-vivo surgical experimentation and can provide a platform to assess the effects of surgical

Alternatively, a mechanical model was proposed by Choo et al (Choo, Lau et al. 2010) specifically for the simulation of experiments on the vocal fold. In their design, they proposed the use of agarose as a material substitute for human vocal folds, mapping the mechanical properties of agar concentrations to that of vocal fold cover and ligament. By repeated casting of different concentrations of agarose into a mould, the phantom vocal folds were designed to mimic the layered structure of the vocal fold. In addition, vocal fold vibration was actuated externally with the use of vibrators, allowing the control of the vocal

fold vibration frequency. Glottal gap and airflow could also be customized.

significant hindrance to exposure or access to the vocal folds.

candidate for animal studies (Alipour and Jaiswal 2008).

**4. Ex-vivo experimental setups** 

measurements of vocal fold vibrations.

procedures on vocal fold vibration.

**4.2 Mechanical models** 

**4.1 Using animal models** 

#### **3.2 Canine models**

Canine models are used extensively in phonation studies. Comparing vocal fold structure across dogs, monkeys, pigs and human models using histology and laryngeal videostroboscopy, Garrett et al (Garrett, Coleman et al. 2000) found that unlike the human vocal fold, which has a higher elastin concentration in the deeper layers of the lamina propria, both pig and dog had a thin band of elastin concentrated just deep to the epithelial basement membrane zone. Just deep to this thin band, collagen and the elastin were less concentrated as in humans. The mucosal wave on stroboscopy was most similar between humans and canines and it was concluded that dog vocal folds were the most ideal for use in surgical studies due to its similarity in size, histology and mucosal wave. However, Fleming et al (Fleming, McGuff et al. 2001) noted that the slight differences in vocal fold structure like the thicker lamina propria and the lack of a well defined vocal ligament would have implications on its vibratory characteristics. Also, the higher cost and ethical considerations of using a companion animal for experimental studies are practical issues that need to be decided upon. Nevertheless, Fleming et al argued that as canine vocal fold healing was found to be similar to humans and similar human pathological conditions have been found to occur in canine models, they are still suitable for use in vocal fold microsurgery. Hahn et al (Hahn, Kobler et al. 2005; Hahn, Kobler et al. 2006; Hahn, Kobler et al. 2006) also compared collagen and elastin distribution in human, dog, pig and ferret larynges. They found that canine lamina propria collagen levels were most similar to those of humans, but on quantitative histology, elastin and collagen distribution in the human lamina propria was best matched by the porcine vocal fold.

#### **3.3 Porcine models**

Pigs are also common models for vocal fold studies. Based on our experience with pig models, the dimensions of the larynx in a 30 to 40 kg pig are similar to that of the adult human (Garrett, Coleman et al. 2000; Jiang, Raviv et al. 2001). The vocal folds have a similar configuration, and the intrinsic muscles and distribution of the recurrent laryngeal nerve is similar as demonstrated by detailed dissections of cadaveric porcine laryngeal neuromuscular anatomy (Knight, McDonald et al. 2005). Other phonatory characteristics such as rotational mobility of the cricothryoid joint, and relative size and innervation of the cricothyroid muscle have also been studied and found to be similar to that of humans (Jiang, Raviv et al. 2001); although these features are not of direct relevance to endoscopic laryngeal microsurgery.

An important difference between the pig and human vocal folds is that the pig has an additional fold in the vertical plane separated by a ventricle. The presence of a superior and inferior fold could relate to the thyroarytenoid muscle having two separate bellies (Knight, McDonald et al. 2005). It has been suggested that the inferior fold is the true vocal fold and the superior fold is akin to the ventricular fold in humans. However this remains a subject for debate as there is a further ventricle above the superior fold. It is suggested that vibration occurs at both folds as well as in the supraglottic structures during phonation (Kurita, Nagata et al. 1983; Alipour and Jaiswal 2008).

The pig larynx also differs in the structure of the arytenoid complex. The arytenoid cartilages have been described as fused across the posterior commissure, making laryngoscopic exposure more difficult (Garrett, Coleman et al. 2000). In addition to this, the arytenoids are also positioned more superiorly resulting in a steeper angle to the vocal folds. However, from our experience with intubated animals, neither of these features created a significant hindrance to exposure or access to the vocal folds.

Regner used high-speed digital imaging to compare the vocal fold vibratory characteristics of ex-vivo bovine, canine, ovine, and porcine larynges with human vocal folds. By measuring amplitude, oscillation frequency, and phase difference of vocal fold vibration, it was concluded that canine and porcine larynges are the most appropriate models for vibratory or kinetic studies on phonation (Regner, Robitaille et al. 2010). Alipour also studied vibratory characteristics of excised pig, cow and sheep larynges, and concluded that the porcine larynx had the highest range of phonation frequencies, making it a good candidate for animal studies (Alipour and Jaiswal 2008).

In a similar study, Jiang et al. (Jiang, Raviv et al. 2001) concluded that pigs models provided the most similarity in vocal fold stiffness and was a reasonable alternative for phonation studies. As pigs are a common livestock, the high availability of pig larynges from local abattoirs poses less of an ethical concern for sacrificing animals for research purposes.
