Hemilaryngeal Transplantation in the Canine Model: Technique and Implications

Transcription

1 Hemilaryngeal Transplantation in the Canine Model: Technique and Implications Robert J. Andrews, MD,* Gera]d S. Berke, MD,* Keith E. BIackwe]], MD,* MichaeI Jakobsen, MD, * Marilene B. Wang, MD, * and Joel A. Sercarz, MD, *~- Purpose: There is no ideal method for reconstruction of hemilaryngeal defects because there is no autologous flap or graft that can reproduce the unique structural properties of the larynx. In this article, the technique, potential research, and clinical applications of hemilaryngeal transplantation are addressed. Materials and Methods: In a canine model, transplantation of a hemilarynx was performed. The thyroarytenoid muscle was reinnervated, and an arytenoid adduction was performed to ensure a competent larynx during the early postoperative period. Results: The canine tolerated the procedure well and the transplanted larynx remained healthy and well vascularized during the postoperative period. Electromyography of the transplanted thyroarytenoid muscle verified reinnervation 2 months after the procedure. During induced phonation, vibration was symmetrical with a normal-appearing laryngeal geometry. Conclusions: Preliminary experience indicates that this technique has unique advantages compared with other available techniques for laryngeal reconstruction. Only with additional progress in transplantation medicine could this procedure be considered an option for reconstruction of human partial laryngeal defects. (Am J Otolaryngo12000;21: Copyright 2000 by W.B. Saunders Company) Optimal phonation is possible when both vocal folds have similar muscular and mucosal properties. There are many potential causes of asymmetric vocal fold vibration including paralysis, neoplasms, and pesttraumatic scarring. After reconstruction of hemilaryngeal defects, the remaining vocal cord typically vibrates opposite a reconstructed pseudocord consisting of mucosa, cartilage, or myofascia] flaps. These methods seek to optimize phonation while allowing airway patency and protection. However, most conventional reconstruction techniques provide a suboptimal solution for replacement of the physical properties and 3-dimensional configuration of the excised laryngeal tissue. We have recently developed a technique of From the *Division of Head and Neck Surgery, UCLA Medical Center, Los Angeles, CA, and -I-Division of Head and Neck Surgery, Harbor-UCLA Medical Center, Torrance, CA. Supported by a VA Merit Review Grant, Washington, DC. Address reprint requests to Joel Sercarz, MD, Division of Head and Neck Surgery, UCLA Medical Center, Le Conte Ave, Los Angeles, CA Copyright 2000 by W.B. Saunders Company /00/ /0 hemilaryngeal transplantation (HLT). This is an outgrowth of previous research in the field of whole organ laryngeal transplantation (LT) and experiments conducted to compare vocal function resulting from several hemilaryngeal reconstruction methods. 1-3 Although the procedure was compared with other hemilaryngeal reconstruction approaches in a previous study, the technique has not been previously described in detail. Because no current method of autologous laryngeal repair can precisely replicate the layered structure of the larynx, HLT was included in a hemilaryngeal reconstruction protocol as a theoretical ideal for an anatomically accurate repair method. In this article, HLT is described in a canine subject. The procedure included an arytenoid adduction with reinnervation of the donor thyroarytenoid muscle using the anterior branch of the recurrent laryngeal nerve (RLN). HLT provided an opportunity to investigate the feasibility of partial organ transplantation and compare this method with the performance of other reconstruction techniques. The future role of HLT in the treatment of patients after vertical hemilaryngectomy is discussed. American Journal of Otolaryngology, Vo121, No 2 (March-April), 2000: pp

2 86 ANDREWS ET AL MATERIALS AND METHODS Overview Many aspects of the hemilaryngeal transplant protocol, including canine selection, immunosuppression, surgical preparation, and postoperative care, were identical to those used by Berke et al for canine laryngeal transplantation. 1 One pair of male litter mate beagles, approximately 1-year old and weighing 12 to 15 kg were used for this experiment. Mixed leukocyte cultures (MLC) were performed to assess the major histocompatibility reactivity at the DLA-D region and ensure an appropriate match of donor and recipient. At approximately 2 months after transplantation, electromyography recordings were obtained from the transplanted thyroarytenoid muscle during spontaneous respiration and recurrent laryngeal nerve stimulation. Immunosuppression The immunosuppression protocol used in this study was developed for previous laryngeal transplantation research. 1 During the postoperative period, the recipient subject was treated with azathioprine, cyclosporine, and prednisone. The azathioprine dosage was modified accordingly for neutropania. Serum cyclosporine levels were checked approximately 12 hours postdosage (trough values) by fluorescent polarization immunoassay, using the TDX monitor (Abbott Laboratories, Abbott Park, IL). The cyclosporin dosage was modified to maintain the trough values in the range of 300 to 500 ng/ml. Operative Procedure Both canine subjects were kept fasting for 12 hours before transplantation. Following preoperative sedation with intramuscular acepromazine maleate (0.2 mg/kg), pentobarbital sodium was administered intravenously to achieve corneal anesthesia. The subjects were placed supine on the operating table, intubated with a cuffed endotracheal tube, and a mixture of 0.5% to 2.0% halothane and oxygen was given to maintain anesthesia throughou} the procedure. Ampicillin and gentamicin were administered intravenously as antimicro- bial prophylaxis and continued for 7 days postoperatively in the recipient. The recipient animal was prepared first. Following sterile skin preparation, a tracheotomy was performed at the 5th tracheal ring with an inferiorly based flap sutured to the skin edge. The oral endotracheal tube was removed, and a sterile cuffed endotracheal tube was placed in the stoma for assisted ventilation. A midline vertical skin incision was made which extended from the level of the hyoid bone to the cricoid cartilage. Skin flaps were elevated in the subcutaneous plane. The strap muscles were then carefully transected to provide exposure to the laryngeal framework. The left external jugular vein and common carotid artery were isolated and prepared for vascular anastomoses. The anterior branch of the left RLN was identified, confirmed with a nerve stimulator, and tagged before transection immediately distal to the bifurcation of the RLN. A scalpel was used to make thyroid cartilage cuts in the midline and at the junction of the posterior one-third and anterior two-thirds of the thyroid lamina on the side of resection. A midline thyrotomy was made, and superior and inferior cuts were completed through the thyrohyoid and cricothyroid membranes. A mucosal incision was made in the midline of the posterior commissure and continued through the interarytenoid muscle and the midline raphe of the posterior cricoarytenoid muscle. The excised specimen included the left true vocal cord, ventricle, false vocal cord, intrinsic laryngeal musculature, arytenoid cartilage, and anterior two-thirds of the thyroid cartilage (Fig 1). In the donor animal, skin preparation, tracheostomy, skin incisions, and laryngeal exposure were then performed as in the recipient. The dissection included preservation of the left hyoid venous arch and the external jugular vein providing venous drainage for the graft. The cranial thyroid artery, the blood supply to the canine larynx, was preserved with the attached common carotid artery. The thyroarytenoid branch of the RLN was identified, tagged, and transected. Cartilage, mucosal, and muscular cuts were then made in an identical manner as for the recipient, although a small cuff of additional mucosa was preserved to avoid a tight closure. The left donor hemilarynx was removed (Fig 2) after ligating

3 HEMILARYNGEAL TRANSPLANTATION 87 Fig 1. Axial cross section of the larynx at the level of the glottis. Dashed line indicates structures excised in hemilaryngeal resection. the common carotid artery superior and inferior to the cranial thyroid branch and the external jugular vein proximal and distal to the hyoid venous tributary. The donor hemilarynx was placed in the recipient bed and the arterial supply was reestablished by a common carotid end-to-end anastomosis. An end-to-end external jugular venous anastomosis was then completed. Total ischemia time was under 30 minutes. A nerve anastomosis was performed connecting the recipient anterior branch of the RLN to the donor thyroarytenoid branch. The other intrinsic laryngeal muscles were not reinnervated. The donor arytenoid was fixed to the recipient cricoid with nonabsorbable suture and an arytenoid adduction was performed. The donor posterior cricoarytenoid and interarytenoid muscles were then separately approximated in the midline to their recipient counterparts with interrupted 4-0 braided absorbable sutures. The donor and recipient laryngeal mucosa was then approximated posteriorly, inferiorly, and superiorly with interrupted 5-0 polyglactin sutures. The anterior mucosal edge of the intact recipient hemilarynx (right) was sutured to the ipsilateral anterior edge of the inner thyroid perichondrium and thyroid cartilage using interrupted 5-0 absorbable gut sutures placed superior and inferior to the true vocal cord. The same was done to the donor hemilarynx (left). The larynx was then closed anteriorly with interrupted 5-0 polyglactin sutures, which were placed through the thyroid cartilage and external thyroid perichondrium to close the anterior commissure. These measures reconstruct the anterior commissure while preventing web formation. Four titanium microplates were then placed across junction of the donor and recipient thyroid cartilages to stabilize the laryngeal framework. A cervical esophagostomy was made for placement of a small feeding tube. The strap muscles were reapproximated in the midline. The wound was irrigated with saline and a Jackson-Pratt suction drain was placed deep to the strap muscles. When the canine was capable of spontaneous respiration, a size-11 Hollinger metal laryngectomy tube was inserted into the tracheostomy. Postoperative Care Fig 2. The arterial supply, venous drainage, and external cartilage and tissue incisions for the donor hemilarynx. The canine was closely monitored during the postoperative period. The animal was fed for 1 week through an esophagostomy tube;

4 88 ANDREWS ET AL thereafter, oral feeding was tolerated without aspiration. The canine was not decannulated so that the transplant hemilarynx could be evaluated with a flexible endoscope placed through the tracheotomy stoma. In this manner, the hemilarynx was monitored for gross signs of organ rejection or ischemia. Postoperative Laryngeal Assessment The canine was assessed 2 months after hemilaryngeal transplantation. Laryngoscopy was performed with a zero degree Wolf endoscope (Wolf Medical Instrument Corp, Vernon Hills, IL). Spontaneous electromyography (EMG) recordings were made during light anesthesia to detect reinnervation potentials in the thyroarytenoid muscle. 4 Electromyography was performed using the Nicolet Viking IV system (Nicolet, Madison, WI). Concentric needle electrodes (26 gauge, 1.5 inches long, recording area ram2; Nicolet, Madison, WI) were used, placing the needles into the transplanted thyroarytenoid muscle via a transoral approach. Recordings were obtained from the transplanted thyroarytenoid muscle during spontaneous respiration and during recurrent laryngeal nerve stimulation. After data collection, a laryngectomy was performed, and the subject was painlessly euthanized. The larynx was then prepared for gross and histological evaluation. Gross and Histological Evaluation The larynx was fixed in formaldehyde and whole organ axial sections were made at the levels of the supraglottis, ventricle, glottis, and subglottis for gross and histological examination. The specimens were then embedded in paraffin, and hematoxylin and eosin slides were prepared from each axial section. ible endoscopy via the tracheostomy stoma. Throughout the postoperative period, the transplant tissue was well vascularized and showed no signs of organ rejection. During the same period, graft mucosal edema gradually resolved. An endoscopic evaluation 2 months after the procedure (Fig 3) showed that the transplanted hemilarynx (left) was similar in appearance to the native side (right). The native vocal fold abducted normally with respiration. The transplant vocal fold position was in the midline, consistent with the arytenoid adduction. The transplant mucosa appeared viable and well healed; all suture lines had healed without granulation, excessive scar or web formation. There was a normal anteriorposterior diameter of the glottis. The native vocal fold adducted to the midline and tensed with RLN stimulation. The transplant vocal fold was fixed in the midline and showed no glottic gap during RLN stimulation of the opposite vocal fold. The transplant thyroarytenoid muscle appeared to contract with stimulation of the RLN although evoked EMG was not performed. Reinnervation The electromyographic recording in Fig 4 shows reinnervation of the thyroarytenoid muscle 2 months following HLT. A concentric EMG needle was placed in the thyroarytenoid muscle. Increased EMG activity correlated with the dog's phonation efforts. Gaps corresponded to inspiratory efforts, thus verifying reinnervation of the muscle. RESULTS Clinical Outcome The subject had no postoperative complications. Bowel function returned by postoperative day 3. The esophagostomy tube was removed on postoperative day 7, and the canine was fed by mouth the next day. The heinilaryngeal transplant was examined daily using flex- Fig 3. Two-month postoperative endoscopy of left hemilaryngeal transplant with arytenoid adduction,

5 HEMILARYNGEAL TRANSPLANTATION 89 SPH RECORD transplanted TR 16~10:57 I00 uu FOOT SNITCH STRTUS:I~Olml/ RUN TRIG:NNsI,'IuUe I s HUP a MUP our ms I Fascics : MUP Pattern= Polgphasi s~ Ma E~f0rt : Fig 4. Electromyographic recording from the thyroarytenoid muscle at 2 months posttransplantation. The EMG activity attenuated synchronously with the dogs phonation efforts. Gaps correspond to inspiratory efforts. The lower portion of the figure shows polyphasic reinnervation potentials. Whole Organ Axial Sections Fig 5 shows a whole organ axial histological section through the glottis. The airway appears to be narrow because of the adducted position of the opposite vocal fold, but was adequate for respiration with the native vocal fold (right) abducted in the live subject. The junction of the native and transplant thyroid cartilage is noted at the anterior commissure and posterior 1/3 of the left hemilarynx. The vascular pedicle to the HLT is seen external to the left hemilarynx. The transplant tissue is well vascularized and there is no evidence of graft rejection. Fig 5. Left hemilaryngeal transplant--whole organ axial histological section of glottis. DISCUSSION Laryngeal surgeons are familiar with the challenge of reconstructing the hemilarynx with mucosal, muscle, or cartilage flaps. No method is capable of fully restoring the highly complex anatomy and physiology that accounts for the coordinated functions of speech, swallowing, and airway protection. Blaugrund et al ~ highlighted the problems with phonation after vocal fold reconstruction. Their research indicated that voice production often occurs at the supraglottic level, rather than by the remaining vocal fold vibrating against the psuedocord. 5 This typically suboptimal outcome motivated our search for a better method of hemilaryngeal reconstruction. It is not possible, using conventional reconstruction techniques, to reproduce the same viscoelastic properties and 3-dimensional configuration of the excised larynx. The purpose of this article is to describe our surgical technique and initial experience with hemilaryngeal transplantation in the canine. We attempted this procedure in order to compare a theoretically ideal method of hemilaryngeal reconstruction to other conventional techniques. A previous study documented that, after HLT reconstruction, laryngeal vibration was more symmetric and produced improved objective measures of vocal function when compared with other methods. 2 Although this study was performed in 1 canine subject, the results suggest that HLT may be an ideal method of partial laryngeal reconstruction. After revascularization, no mucosal and muscular tissue ischemia noted and only a brief period of laryngeal edema occurred. Electromyography confirmed reinnervation of the thyroarytenoid muscle. Furthermore, there was a pause in the EMG activity of the TA during inspiratory efforts, suggesting successful reinnervation with adductory fibers from the recipient canine. Previous experience with canine reinnervation indicates that laryngeal function after reinnervation continues to improve for about 9 months. 6 The authors are mindful that it would strengthen this article to include a larger number of animal subjects. No further studies are underway because in canine laryngeal transplantation, few animals survive the early postoperative period. During the decade of the

6 90 ANDREWS ET AL 1990s, there has been only 1 canine laryngeal transplantation survivor outside of our laboratory, an animal studied by Anthony et al. 7 that survived 33 days before having graft rejection. The authors do not anticipate performing further HLT because of the limited information that would be obtained by studying additional animals. Nonetheless, the analysis of only 1 canine subject is a weakness of this study. In the present experiment, arytenoid adduction and thyroarytenoid muscle reinnervation were performed to maintain medialization and provide tension to the transplanted vocal fold. If HLT were to be applied to human hemilaryngeal reconstruction, it is conceivable that reinnervation of each motor and sensory element would be achieved. This would include separate reinnervation of the posterior and anterior branches of the RLN to allow physiological opening and closure of the larynx under central nervous system control. Experience in whole organ LT indicated that physiological motion of the larynx, such as laryngeal dilation upon endotracheal tube occlusion, occurs following RLN reinnervation distal to the bifurcation of the RLN into it's anterior and posterior branches. 1 Finally, end-to-end neurorrhaphy of the internal and external branches of the SLN would be performed to activate cricothyroid function and supraglottic sensation. Because of current risks to the recipient, advances in immunosuppression therapy must occur before the clinical application of HLT in a cancer patient requiring hemilaryngectomy. This requires the development of immunosuppressive agents specifically targeted against rejection of the donor organ, while sparing the infection and cancer surveillance functions of the immune system. If developed, such agents would have a profound impact on the potential for human LT. There are some situations, albeit rare, in which HLT could be offered to motivated patients in the near future. (1) One potential recipient would be a previous transplant recipient (eg, heart, lung, kidney, or liver) already taking immunosuppressive agents who requires hemilaryngectomy. It is possible that such a patient would incur few additional risks from the immunosuppression necessary to maintain the laryngeal graft. This depends upon the relative immunogenicity of the differ- ent organs, an issue that has not yet been studied for the human. Strome et al 8 have studied the effect of varying doses of cyclosporin on laryngeal allograft survival in the rat model. The conclusion was that, with a dosage of 10 mg/kg, there was little evidence of rejection and only mild rejection at 7.5 mg/kg. (2) A less likely indication might be a patient with persistent aspiration or a poor speech outcome who has survived 5 years after a hemilaryngeal resection for cancer. At that time, the risk of recurrent cancer may be low enough that immunosuppression may become an acceptable option. Nerve-sparing approaches to laryngeal resection are currently being used at UCLA in order to make this feasible in the future. (3) Lastly, a patient with devastating laryngeal trauma might accept immunosuppression as an alternative to a poorly functioning laryngeal system. However, using available immunosuppressive regimens, a wide range of side effects including a susceptibility to infection must be weighed carefully. If hemilaryngeal transplantation were to become a realistic clinical option, then appropriate matching of the donor larynx would be necessary. The donor larynx would ideally come from a young individual because of the loss of laryngeal elasticity that occurs with aging. Unlike other organs, there would have to be a strict match in the size and sex of the organ donor and recipient. Finally, an appropriate antigen match would be necessary between donor and recipient. It is currently not clear whether HLA or ABO Rh compatibility, or both, would be required. Currently, the most important barrier to human LT is the need to provide adequate immunosuppression without increasing the risks of cancer recurrence, the development of second primaries, or complications related to infection. As methods of immunosuppression improve, it is possible that LT will become an acceptable medical option for patients with a functionally useless or absent larynx. At that point in time, it is our contention that HLT will be the method of choice to reconstruct hemilaryngeal defects in the future. Transplantation of the hemilarynx might be a possible outgrowth of experience with whole organ LT in humans in the future. In conclusion, HL was developed in the

7 HEMILARYNGEAL TRANSPLANTATION 91 laboratory because of its theoretical advantages over other reconstruction techniques. At the present time, this method may be used as a "gold standard" for comparison to other techniques of laryngeal reconstruction in animal models. The strong interest in laryngeal transplantation has recently culminated in a human trial of the operation. In the future, potential clinical applications of HLT in humans may be envisioned. REFERENCES 1. Berke GS, Ming Y, Block RM, et al: Orthotopic laryngeal transplantation. Is it time? Laryngoscope 103: , Andrews RJ, Sercarz JA, Ye M, et al: Vocal function following vertical hemilaryngectomy: Comparison of four reconstruction techniques in the canine. Ann Otol Rhinol Laryngol 106: , Kevorkian K, Sercarz JA, Ye M, et al: Extended canine laryngeal preservation for transplantation. Laryngoscope 107: , Green DG, Berke GS, Graves MC, et al: A functional study of ansa to recurrent anastomosis: Future directions for laryngeal reinnervation. Otolaryngol Head Neck Surgery 104: , Blangrund SM, Gould wj, Haji T, et al: Voice analysis of the partially ablated larynx. A preliminary report. Ann Otol Rhinol Laryngol 93: , Peterson KL, Sercarz JA, Andrews RA, et al: Objective measures of laryngeal function following reinnervation of the anterior and posterior recurrent laryngeal nerve branches. Laryngoscope 108: , Anthony JP, Allen DB, Trabulsy PP, et al: Canine laryngeal transplantation: Preliminary studies and a new heterotopic allotransplantation model. European Archives of Oto-Rhino-laryngology 252: , Strome M, Strome S, Darrell J, et al: The effects of cyclosporin A on transplanted rat allografls. Laryngoscope 103: , 1993