A prospective, randomized trial of esophageal submucosal tunnel closure with a stent versus no closure to secure a transesophageal natural orifice transluminal endoscopic surgery access site

ORIGINAL ARTICLE: Experimental Endoscopy

A prospective, randomized trial of esophageal submucosal tunnel closure with a stent versus no closure to secure a transesophageal natural orifice transluminal endoscopic surgery access site Brian G. Turner, MD, Min-Chan Kim, MD, Denise W. Gee, MD, Abdulmetin Dursun, MD, Mari Mino-Kenudson, MD, Edward S. Huang, MD, Patricia Sylla, MD, David W. Rattner, MD, William R. Brugge, MD, FASGE Boston, Massachusetts, USA

Background: Secure esophagotomy closure methods are a critical element in the advancement of transesophageal natural orifice transluminal endoscopic surgery (NOTES) procedures. Objective: To compare the clinical outcomes in swine receiving an esophageal stent or no stent after a submucosal tunnel NOTES access procedure. Design: Prospective, randomized, controlled trial in 10 Yorkshire swine. Setting: Academic center. Intervention: An endoscopic mucosectomy device was used to create an esophageal mucosal defect. An endoscope was advanced through a submucosal tunnel into the mediastinum and thorax, and diagnostic mediastinoscopy and thoracoscopy were performed. Ten animals were randomized to no stenting (n ⫽ 5) or stenting (n ⫽ 5) with a prototype small-intestine submucosa– covered stent. Main Outcome Measurements: Gross and histologic appearance of the mucosectomy and esophagotomy sites as well as clinical outcomes. Results: There was a significant difference in the overall procedure time between the animals that received a stent (35.0 min, range 27-46.0 min) and those with no closure (19.0 min, range 17-32 min) (P value ⫽ .018). The unstented group achieved endoscopic and histologic evidence of complete re-epithelialization and healing (100%) at the mucosectomy site compared with the stented group (20%, P ⫽ .048). Stent migration into the stomach occurred in two swine. Both groups had complete closure of the submucosal tunnel and well-healed esophagotomy sites. Limitations: Animal study, small number of subjects. Conclusion: The placement of a covered esophageal stent significantly interferes with mucosectomy site healing. (Gastrointest Endosc 2011;73:785-90.)

Transesophageal natural orifice transluminal endoscopic surgery (NOTES) access to the mediastinum and thorax is promising for certain populations of patients, but

identifying a reliable, secure esophagotomy closure method is paramount. Previous animal studies report transesophageal approaches to mediastinoscopy, thora-

Abbreviations: NOTES, natural orifice transluminal endoscopic surgery.

doi:10.1016/j.gie.2010.11.025

DISCLOSURE: Funding for this project was provided by a grant from the Center for Integrative Medicine and Technology. D. Rattner disclosed a speaker relationship with Olympus and receipt of an honorarium. W. Brugge is a consultant for Boston Scientific. Cook Medical, Inc donated the prototype stent and additional endoscopic supplies essential for completing this trial. No other financial relationships relevant to this publication were disclosed.

Received May 6, 2010; Accepted November 11, 2010.

Copyright © 2011 by the American Society for Gastrointestinal Endoscopy 0016-5107/$36.00

Reprint requests: William R. Brugge, MD, Gastroenterology Unit, Massachusetts General Hospital, 55 Fruit Street, Boston, MA 02114.

www.giejournal.org

Current affiliations: Gastrointestinal Unit (B.G.T., E.S.H., W.R.B.), Department of Surgery (M.-C.K., D.W.G., A.D., P.S., D.W.R.), Department of Pathology (M.M.-K.), Massachusetts General Hospital, Boston, Massachusetts, USA; Department of Surgery, Minimally Invasive and Robot Center (M.-C.K.), Dong-A University College of Medicine, Busan, Korea.

Volume 73, No. 4 : 2011 GASTROINTESTINAL ENDOSCOPY 785

Esophageal stenting for NOTES access site closure

coscopy, pleural biopsy, sympathectomy, and mediastinal lymphadenectomy.1-7 Esophagotomy closure of the esophageal wall surgical site has been accomplished by a variety of techniques, including endoscopic clips,7 submucosal tunneling techniques,1,5 and suturing devices.7 These techniques offer adequate closure but can be challenging because of the narrow working space of the esophagus and/or blind deployment of the device. Ideally, the closure device also should aid in diverting esophageal contents from the NOTES surgical site to prevent leakage into the mediastinal space. A covered esophageal stent would be a suitable alternative or adjunct device to the aforementioned methods. Recently, our group performed a preliminary study examining the use of a prototype, biomaterial-covered esophageal stent as an adjunct device to be placed after the creation of a submucosal tunnel for NOTES thoracic procedures. The results of the study demonstrated that placement of an esophageal stent provided an effective method of closure with complete healing of the submucosal tunnel.8 Interestingly, we also found that delayed healing of the mucosectomy site occurred in each animal. Necropsy after a 14-day survival period revealed that 67% of animals had poor healing of the esophageal mucosa at the mucosectomy site. This finding may have important implications in the choice of the NOTES access technique or NOTES closure, or in the future this finding may play a role in the use of biomaterial-covered stents in humans after more substantial procedures such as endoscopic submucosal dissection for early esophageal cancers. The goal of this study was to directly compare the clinical outcomes of two separate closure methods (stent versus no stent) at a NOTES access site created with a submucosal tunnel technique. To our knowledge, this study represents the first randomized, controlled trial comparing closure methods at a transesophageal NOTES access site.

MATERIALS AND METHODS This study was approved by the subcommittee for research animal care at the Massachusetts General Hospital, Boston, Massachusetts. The measured outcomes included endoscopic and histologic healing of the esophageal surgical sites, gross and histopathologic examination, and necropsy findings.

Anesthesia and NOTES procedure Ten Yorkshire swine underwent thoracic NOTES procedures. Swine were randomized to the unstented or stented group by using random number tables before the start of the procedure. The 10 fasted swine underwent general anesthesia with induction by Telazol (A.H. Robins, Richmond, Va) 4.4 mg/kg intramuscular plus xylazine 2.2 mg/kg intramuscular and endotracheal intubation. Anesthesia was maintained with 1.5% to 3.0% isoflurane 786 GASTROINTESTINAL ENDOSCOPY Volume 73, No. 4 : 2011

Turner et al

Take-home Message ●

●

In this randomized trial, the creation of a submucosal tunnel alone or a tunnel and the adjunct use of an esophageal stent after natural orifice transluminal endoscopic surgery procedures proved safe and effective. In animals that underwent placement of a biomaterialcovered esophageal stent, mucosectomy site healing was significantly delayed, compared with animals with no stent.

and oxygen at 3.0 L/minute. Intravenous clindamycin at 300 mg was given preoperatively. With the animal supine, a double-channel gastroscope (Pentax Medical Inc, Montvale, NJ) was introduced into the esophageal lumen and used to direct the placement of an overtube in the proximal esophagus. A Duette Multiband mucosectomy device (Cook Medical, Inc, Winston-Salem, NC) was positioned 10 cm proximal to the gastroesophageal junction, and an esophageal mucosal defect was created with the snare resection of banded esophageal tissue (Fig. 1). Closed biopsy forceps and air insufflation were used to create a short, 5-cm, submucosal tunnel at the mucosectomy site. Within the tunnel, a prototype rotating hook knife (Cook Medical) was used to make an incision through the muscular layer of the esophageal wall and enter the mediastinum. The pleura was then incised with endoscopic flexible scissors (Ethicon Endo-Surgery, Cincinnati, Ohio) or a hook knife and the endoscope was passed into the right thoracic cavity. Mediastinoscopy and thoracoscopy were performed to identify structures including lung, pericardium, lymph nodes, diaphragmatic surface, and chest wall.

Stent placement On completion of thoracoscopy, swine randomized to the stent group underwent placement of a prototype, fully covered, plastic esophageal stent (28 mm diameter, 65 mm length, 37 mm proximal flare) lined with small-intestine submucosa (Surgisis; Cook Biotech, West Lafayette, In) material around the exterior of the stent body. Stent size was chosen based on our previous preliminary work with esophageal stenting in transesophageal NOTES access sites.8 First, a guidewire was inserted through the endoscope channel and advanced to the stomach. The endoscope was withdrawn and the stent sheath inserted over the guidewire. Under direct endoscopic visualization, the stent was advanced to the mucosectomy site and deployed over both the submucosal tunnel and mucosal defect. A circumferential loop of string around the proximal cage of the stent permitted any necessary adjustments. After placement, two 18-mm Eclipse Wire Guided Balloon Dilators (Cook Medical) were inserted simultaneously into the stent lumen. The first balloon was fully inflated to 18 mm, www.giejournal.org

Turner et al

Figure 1. Closed biopsy forceps indicate the esophageal mucosectomy site immediately following application of electrocautery to the snare surrounding the banded tissue.

and the second dilator balloon was sufficiently inflated to result in maximal expansion of the stent body.

Postoperative events After completion of the procedures, the swine were monitored for weight gain, ambulation, oral intake, clinical signs of infection, and interactive appearance. On postoperative day 0, the animals were given an additional 300 mg of oral clindamycin. The antibiotic regimen on postoperative days 1 to 3 was 300 mg of oral clindamycin twice daily. In the first 48 hours postoperatively, the swine diet was restricted to liquids and soft foods (ie, yogurt and Ensure (Abbott Laboratories, Abbott Park, Illinois, U.S.A.)), and then the diet was advanced to a regular diet for the remainder of the survival period. Necropsy was performed on day 14. On the day of necropsy, endoscopy was used to assess the esophagotomy sites and/or placement of the stent. Animals with a stent did not undergo removal of the stent at the time of endoscopy to prevent any potential damage to the exit site, which could alter histopathologic interpretation. The animals were then euthanized and examined for evidence of pleural fluid and signs of infection such as thoracic or mediastinal abscess. The esophagi were explanted, grossly examined, and sent for histopathologic review by an expert GI pathologist.

Statistical analysis The sample size calculation was based on a previous pilot study of 9 survival swine that underwent placement of a prototype Surgisis-covered stent to secure the mucosectomy and transesophageal exit sites created during a NOTES procedure. In the pilot study, we concluded that esophageal stents appear to provide safe closure of the surgical site but may impair healing at the mucosectomy site. Incomplete healing was observed at a frequency of 100%. Based on these data, we performed a sample size calculation for a 2-sample comparison of proportions, assuming that incomplete healing of the esophageal mucosectomy site would occur in all (100%) swine receiving an www.giejournal.org

Esophageal stenting for NOTES access site closure

esophageal stent. The calculation resulted in a required sample size of 4 swine in each group (a total of 8 swine) to attain a study power of 80% and an alpha of 0.05. Procedural times were reported as a median with ranges and defined as the time from overtube placement to completion of the NOTES procedure in unstented animals or completion of stent deployment. Analysis of the outcome and histologic data was performed by using SAS (version 9.2; SAS Institute, Inc, Cary, NC). Continuous variables were compared by using a 2-tailed independent samples t test for normally distributed variables and Wilcoxonranked sum for non-normal distribution, whereas categorical variables were compared by using the Fisher exact test of significance. A P value ⬍ .05 was considered statistically significant.

RESULTS Transesophageal NOTES access to the mediastinum and chest cavity was successfully performed in all but one swine (median weight 41.5 kg, range 36.0-47.0 kg). In one swine, mediastinoscopy alone was performed because of extensive mediastinal webbing and preexisting adhesions that prohibited safe access to the pleural space. There was a significant difference in the overall procedure time between the group that received a stent (35.0 min, range 27-46.0 min) and the group with no stent (19.0 min, range 17-32 min), P value ⫽ .018. There was no significant difference in the time required to perform mediastinoscopy or thoracoscopy in either group, P ⫽ .45. There was one procedural complication in the esophageal stent group because of a tiny, 1-mm, full-thickness puncture from the use of biopsy forceps. Necropsy results demonstrated 3 swine with adhesions in the stented group, but this did not reach statistical significance. A small amount of pleural fluid was seen in two swine in the stented group, whereas one swine in the unstented group had a moderate pleural effusion (P ⬎ .99). No abscesses or postoperative complications were noted in either group. Gross and histologic assessment revealed a significant difference in healing at the mucosectomy sites (Figs. 2 and 3). The unstented group achieved endoscopic and histologic evidence of complete re-epithelialization and healing (100%) at the mucosectomy site compared with the stented group (20%), P ⫽ .048. The stent migrated completely into the stomach in the two swine with complete re-epithelialization. The time of migration is unknown because neither serial endoscopy nor imaging was performed. Submucosal tunnels were well-healed in each group of animals. The transesophageal exit site through the muscular layer had fibrosis and chronic inflammation in the majority of animals but was not significantly different in either group. A comparative analysis of all clinicopathologic outcomes of each group is presented in Table 1. Volume 73, No. 4 : 2011 GASTROINTESTINAL ENDOSCOPY 787

Esophageal stenting for NOTES access site closure

Turner et al

Figure 2. Gross pathologic specimens of explanted esophagi after 14 day survival period. The unstented swine’s esophagus showed a well-healed, nonvisible mucosectomy site (left), whereas the stented swine’s esophagus (right) exhibited poor healing and minimal re-epithelialization at the mucosectomy site (white arrow).

DISCUSSION One of the major challenges in natural orifice transluminal procedures is secure closure. Full-thickness healing and secure tissue apposition are critical to preventing infectious complications. Some groups, including ours, have used a submucosal tunneling technique alone for closure, with good outcomes.1,5 We prefer to create the NOTES access site with a mucosectomy and tunnel as opposed to a transverse esophageal incision for two reasons. First, the submucosal tunnel acts as a secure closure and does not require the blind firing of suture devices (which might damage mediastinal structures) to achieve closure. Second, on manipulation of the access site during the procedure, transverse incisions may enlarge or tear, resulting in bleeding and making repair more challenging. Overall, the defect size may be equivalent to that of a mucosectomy site. Apart from the submucosal technique, other groups have successfully used endoscopic clips and prototype T-tag anchor devices to close the esophagotomy sites after NOTES procedures.7 Endoscopic sutures appear to provide full-thickness healing of the mucosal and muscular layers. There are currently no published data on the use of esophageal stents to aid in NOTES access site closure. Based on nonrandomized, observational human studies, stenting resulted in better outcomes when stents were placed immediately after iatrogenic perforations.9 Another series of patients with nonmalignant GI perforations or anastomotic leaks/disunion of the esophagus and colon reported the potential of covered stents for providing a “stent-guided regeneration and re-epithelialization” of the GI lumen.10 788 GASTROINTESTINAL ENDOSCOPY Volume 73, No. 4 : 2011

Figure 3. Histology slides of representative mucosectomy sites. Hematoxylin and eosin (H&E) stain at 20x magnification. The esophageal lumen is oriented at the top of each panel. A, The low-power view of a stented animal’s esophagus demonstrates lack of re-epithelialization at the mucosectomy site, with granulation tissue and no overlying squamous epithelium (black arrow) after 14 day survival. B, The unstented animal demonstrates complete healing with squamous restitution at the mucosectomy site (black arrow).

Although each of these techniques offers promise, there are no published studies comparing any of these methods of closure in a randomized, prospective fashion. The aim of our study was to assess outcomes of two closure methods and to determine the frequency of mucosal healing at an esophageal mucosectomy site in swine undergoing transesophageal NOTES procedures with submucosal tunnel closure alone compared with those receiving an adjunctive esophageal stent. In terms of www.giejournal.org

Turner et al

Esophageal stenting for NOTES access site closure

TABLE 1. Clinicopathologic results of transesophageal NOTES procedures Unstented group (n ⴝ 5)

Stented group (n ⴝ 5)

P value*

19.0 (17.0-32.0)

35.0 (27.0-46.0)

.018

Mediastinoscopy time, median (range), min

4.0 (3.0-16.0)

4.0 (3.0-6.0)

.46

Thoracoscopy time, median (range), min

6.0 (5.0-6.0)†

6.0 (5.0-8.0)

.75

Stent placement time, median (range), min

n/a

6.0 (6.0-13.0)

n/a

Mortality, no. (%)

0 (0)

0 (0)

n/a

3.0 (2.0-4.0)

3.0 (0.0-9.0)

.77

Ambulating postoperatively, no. (%)

5 (100)

5 (100)

n/a

Bright and interactive postoperatively, no. (%)

5 (100)

5 (100)

n/a

Oral intake postoperatively, no. (%)

5 (100)

5 (100)

n/a

Procedural complications, no. (%)§

0 (0)

1 (20)

.048

Adhesions

0 (0)

3 (60)‡

.18

Pleural effusion

1 (20)

2 (40)

⬎ .99

Abscess

0 (0)

0 (0)

n/a

Postoperative complications

0 (0)

0 (0)

n/a

0 (0)

4 (80)

.048

Complete re-epithelialization of mucosectomy site

5 (100)

1 (20)

.048

Serosal esophageal inflammation at transesophageal exit site

3 (60)

4 (80)

⬎ .99

Serosal esophageal fibrosis at transesophageal exit site

3 (60)

5 (5)

.44

Microscopic esophageal abscess

0 (0)

0 (0)

n/a

Clinical and procedural findings Total procedure duration, median (range), min

Weight gain, median, (range), pounds

Necropsy, no. (%)

Histopathologic examination, no. (%) Esophageal erosions

NOTES, Natural orifice transluminal endoscopic surgery; n/a, not applicable. *Continuous variables were compared by using a 2-tailed independent samples t test, whereas categorical variables were compared by using the Fisher exact test of significance. †Thoracoscopy was not performed in one animal because of extensive mediastinal webbing inhibiting safe entry. ‡In these 3 animals, a small periesophageal adhesion was found as well as adhesions located between the esophagus and aorta and between the lung and chest wall. §A tiny, 1-mm hole occurred in the muscular wall secondary to use of biopsy forceps in the tunnel.

mucosal healing, the stented group demonstrated significantly impaired healing compared with the unstented group (P ⫽ .048). Although this just reaches statistical significance, stent migration occurred in the one swine achieving complete re-epithelialization. The precise reason for the delayed healing in the stented group is unknown. Possibilities include tissue ischemia secondary to increased radial forces from the stent or direct contact of the Surgisis covering in the prototype stent. Although the Surgisis mesh has proven to be effective in human hernia repairs,11-13 there is recent evidence that 4 commercial porcine biologic meshes induce cytokine expression that adversely affects tissue remodeling.14 In the study, the Surgisis product induced significantly www.giejournal.org

fewer inflammatory mediators in comparison with the other products. A trial directly comparing the Surgisiscovered stent with a conventional covered stent would be required to help clarify the etiology of delayed healing. Finally, several growth factors in human saliva, including basic fibroblast growth factor, have wound healing and tissue regeneration properties.15,16 The diversion of particulate food matter and saliva away from the esophagotomy site and submucosal tunnel is critical; however, it also may prevent necessary growth factors from reaching the mucosectomy site and facilitating healing. Although the primary goal of the stent is to create tissue apposition in the submucosal tunnel and to divert esophageal contents, delayed mucosal healing remains a concern. Delayed Volume 73, No. 4 : 2011 GASTROINTESTINAL ENDOSCOPY 789

Esophageal stenting for NOTES access site closure

healing at the site may result in infection or an esophageal ulcer with bleeding, or it may jeopardize healing of the submucosal tunnel. These potential complications were not seen in our study of only 5 stented animals. As mentioned, the size of the stent may be influencing the rate of healing because of its diameter and the resulting radial forces. The stent size was based on a previous trial from our laboratory. In that trial, the necessary diameter of the proximal flare needed to help prevent stent migration was assessed. Additionally, the appropriate diameter of the stent body necessary to allow sufficient contact of the small-intestine submucosa– covered portion of the stent with the esophageal wall was determined.8 Because the NOTES access mucosectomy site is about 15 mm, a significant length of stent is not required. However, it is possible that the proximal flare is too large and disrupts mucosal blood flow. Further studies will need to be carried out to reveal optimal stent characteristics. When the clinical application of the stent is considered, it may not be necessary for the entire mucosectomy site to heal. Rather, it may be more important that the submucosal tunnel completely heals (as in this study) and that the stent helps to divert particulate food matter from entering the tunnel and ultimately the mediastinum. In a clinical setting, it might be practical to remove the stent after 2 weeks of placement. Leaving the stent for longer than 2 weeks might result in the stent eroding through the esophagus, and the likelihood of stent migration would be higher. There was one complication in the stented group that resulted from a tiny, full-thickness tear caused by biopsy forceps. This complication is likely related to the swine model, given the relatively thin nature of the esophagus. In humans, the esophageal wall is thicker, and such a complication would not be anticipated. Our study did not demonstrate a significant difference in the two groups of swine with regard to mortality, postoperative outcomes, or the development of abscesses, adhesions, or pleural effusions; however, the current study is not powered to fully assess the outcomes of these secondary endpoints. Large, randomized studies will be necessary to determine whether there are truly significant differences in clinical outcomes of the animals, because the swine overall did very well postoperatively. There are no current large animal trials examining infectious outcomes in transesophageal NOTES thoracic procedures. Other limitations to this study include small sample size and relatively short-term survival. In this randomized trial, creation of a submucosal tunnel alone or a tunnel and the adjunct use of an esophageal stent after NOTES procedures was successful, safe, and effective. The use of an esophageal stent successfully diverts intraluminal contents away from the tunnel opening and may prevent contamination of the mediastinum and thorax. However, the results of this prospective, randomized study demonstrate that stent placement significantly interferes with mucosectomy site healing and that animals with a submucosal tunnel alone have complete healing with similar out790 GASTROINTESTINAL ENDOSCOPY Volume 73, No. 4 : 2011

Turner et al

comes. The clinical impact of the esophageal stent on mucosectomy site healing may not outweigh the benefits of stent placement, but this remains to be seen in larger, randomized trials. ACKNOWLEDGMENTS We would like to thank Cook Medical, Inc for donation of the prototype stent and additional endoscopic supplies essential for completing this trial. In particular, we would like to thank Richard DuCharme and Vihar Surti for their contributions to the project. We also express our deep gratitude to Katherine Briggs, who is our dedicated animal research laboratory coordinator and veterinary technician. REFERENCES 1. Gee DW, Willingham FF, Lauwers GY, Brugge WR, Rattner DW. Natural orifice transesophageal mediastinoscopy and thoracoscopy: a survival series in swine. Surg Endosc 2008;22:2117-22. 2. Turner BG, Gee DW, Cizginer S, et al. Feasibility of endoscopic transesophageal thoracic sympathectomy (with video). Gastrointest Endosc 2010;71:171-5. 3. Fritscher-Ravens A, Olagbaiye OA, Holland C, et al. Randomized comparative trial of transesophageal endoscopic versus thoracoscopic removal of marked mediastinal lymphnodes in a long-term animal survival study. Gastrointest Endosc 2009;69:AB121. 4. Woodward T, McCluskey D, 3rd, Wallace MB, et al. Pilot study of transesophageal endoscopic surgery: NOTES esophagomyotomy, vagotomy, lymphadenectomy. J Laparoendosc Adv Surg Tech A 2008;18: 743-5. 5. Sumiyama K, Gostout CJ, Rajan E, et al. Transesophageal mediastinoscopy by submucosal endoscopy with mucosal flap safety valve technique. Gastrointest Endosc 2007;65:679-83. 6. Willingham FF, Gee DW, Lauwers GY, et al. Natural orifice transesophageal mediastinoscopy and thoracoscopy. Surg Endosc 2008;22:1042-7. 7. Fritscher-Ravens A, Patel K, Ghanbari A, et al. Natural orifice transluminal endoscopic surgery (NOTES) in the mediastinum: long-term survival animal experiments in transesophageal access, including minor surgical procedures. Endoscopy 2007;39:870-5. 8. Turner BG, Cizginer S, Kim MC, et al. Stent placement provides safe esophageal closure in thoracic NOTES (TM) procedures. Surg Endosc. E-pub 2010 Sept 4. 9. Fischer A, Thomusch O, Benz S, et al. Nonoperative treatment of 15 benign esophageal perforations with self-expandable covered metal stents. Ann Thorac Surg 2006;81:467-72. 10. Amrani L, Menard C, Berdah S, et al. From iatrogenic digestive perforation to complete anastomotic disunion: endoscopic stenting as a new concept of ”stent-guided regeneration and re-epithelialization.” Gastrointest Endosc 2009;69:1282-7. 11. Agresta F, Bedin N. Transabdominal laparoscopic inguinal hernia repair: is there a place for biological mesh? Hernia 2008;12:609-12. 12. Ansaloni L, Catena F, Gagliardi S, et al. Hernia repair with porcine smallintestinal submucosa. Hernia 2007;11:321-6. 13. Oelschlager BK, Pellegrini CA, Hunter J, et al. Biologic prosthesis reduces recurrence after laparoscopic paraesophageal hernia repair: a multicenter, prospective, randomized trial. Ann Surg 2006;244:481-90. 14. Orenstein SB, Qiao Y, Klueh U, et al. Activation of human mononuclear cells by porcine biologic meshes in vitro. Hernia 2010;14:401-7. 15. Ishizaki H, Westermark A, van Setten G, et al. Basic fibroblast growth factor (bFGF) in saliva—physiological and clinical implications. Acta Otolaryngol Suppl 2000;543:193-5. 16. Zelles T, Purushotham KR, Macauley SP, et al. Saliva and growth factors: the fountain of youth resides in us all. J Dent Res 1995;74:1826-32.

www.giejournal.org