Journal of Pediatric Surgery 49 (2014) 1131–1137

Contents lists available at ScienceDirect

Journal of Pediatric Surgery

j ourna l homepage: www.e lsev ie r .com/ locate / jpedsurg

Magnetic gastrointestinal anastomosis in pediatric patients

Mario Zaritzky a,⁎, Ricardo Ben b, Krystal Johnston c

a Department of Radiology, The University of Chicago Medicine, Comer Children’s Hospital, 5721 S. Maryland Avenue, Chicago, IL 60637, USAb Department of Gastroenterology, Hospital de Niños de La Plata, Calle 14 Nro 1631, La Plata, Buenos Aires, Argentinac MED Institute, Inc., 1 Geddes Way, West Lafayette, IN 47906, USA

a b s t r a c ta r t i c l e i n f o

⁎ Corresponding author at: Department of RadiologMedicine, Comer Children’s Hospital, 5721 S. MarylanTel.: +1 773 702 2075; fax: +1 773 834 7452.

E-mail address: mzaritzky@radiology.bsd.uchicago.ed

http://dx.doi.org/10.1016/j.jpedsurg.2013.11.0020022-3468/© 2014 Elsevier Inc. All rights reserved.

Article history:

Received 11 July 2013Received in revised form 29 October 2013Accepted 2 November 2013

Key words:Surgical procedures,minimally invasive/instrumentationAnastomosis, surgical/instrumentationElectromagnetic phenomena/therapeutic useMagnetics/instrumentationEsophageal atresia/therapy

Background/Purpose: To describe 17 patients who underwent magnetic, non-surgical gastrointestinal(GI) anastomoses.Methods: Patients with GI obstruction, stenosis, or atresia were treated with image-guided and/orendoscopically placed discoid magnet pairs or catheter-based bullet-shaped magnet pairs.Results: Anastomosiswas achieved in 7 days in an 11-year-oldwith gastric outlet obstruction due tometastaticcolon cancer. Anastomosis was achieved in 8 and 10 days in 2 patients (age 2.0 years and 3.4 years) who hadrectocolonic stenosis. Re-anastomosis was achieved in an average of 6 days (range 3 to 7 days) in 5 patients(age 6 months to 5.9 years) with severe recurrent postsurgical esophageal stenosis refractory to dilatation.Primary esophageal anastomosis was achieved in an average of 4.2 days (range 3 to 6 days) in 9 patients withesophageal atresia (Type A or Type C surgically converted to Type A) with a gap length of 4 cm or less. Theaverage age of these esophageal atresia patients was 3 months (range 23 days to 5 months).

Conclusion: Minimally invasive magnet placement was feasible and achieved anastomosis in all patients.

© 2014 Elsevier Inc. All rights reserved.

Compression devices to facilitate intestinal anastomoses haveachieved intermittent popularity in the medical literature. The firstdescription of compression anastomosis in an animal model was putforth by Denans in 1826; his device was composed of metal rings. TheMurphy button, introduced in 1892, also consisted of metal rings heldin place by a purse-string suture [1–3]. In the mid-1980s, additionalcompression devices were introduced, which were composed ofbiodegradable rings and non-biodegradable rings with or withoutmagnets included in the design [1,2]. These devices were primarilyused to create small or large bowel anastomoses.

In the 1970s, Hendren and Hale published clinical outcomes ofmagnet-induced growth in the treatment of esophageal atresia andhigh-pouch imperforate anus; magnetic bougienage allowed primaryesophageal or rectal anastomosis without interposition of additionaltissue [4–6]. From1995 to 2001, Cope published four papers describinggastroenteric and cholecystogastric/jejunal anastomoses in animals[7–10]. In 2005, Chopita et al. published results of 15 patients who hadmalignant upper GI obstruction and who underwent successfulmagnetic compression gastroenteric anastomosis [11]. Additionally,Takao et al. published a case report of a patient with benign commonbile duct obstruction who underwent choledochoduodenostomy bymagnetic compression anastomosis [12]. More recently, magnetic

y, The University of Chicagod Avenue, Chicago, IL 60637.

u (M. Zaritzky).

compression anastomosis treatment for biliary obstruction has beenreported in case reports and small case series [13–16].

Gastrointestinal (GI) anastomoses are traditionally created bysuture or staple methods. The goal of any anastomosis is to connectthe two structures and allow sufficient healing to discourageanastomotic leak. If a safe, minimally invasive treatment techniquewere developed, it could become the preferred anastomotic treatmentmethod. Therefore, creation of a functional magnetic compressionanastomosis through minimally invasive techniques warrants appro-priate investigation. The purpose of this analysis is to describe patientswith GI obstruction, stenosis, or atresia who underwent magnetic,non-surgical GI anastomoses.

1. Materials and methods

Between June 2001 and December 2012, 17 pediatric patientswere treated by magnetic compression anastomosis at a single centerin Argentina (Table 1). Approval was obtained from the site’s EthicsCommittee and informed consent was obtained from the parents orlegal guardians.

1.1. Discoid magnet pairs

A circular discoid magnet pair (Cook Medical, Winston-Salem, NC;device not currently marketed) was used to treat GI obstruction in8 patients. One patient had gastric outlet obstruction due tometastatic colon adenocarcinoma, 2 patients had rectocolonic

Table 1Underlying conditions and magnets used for each patient.

Patientnumber

Indication for magnetic compression anastomosis Type of magnet used

1 Gastric outlet obstruction (metastatic colonadenocarcinoma)

Discoid paira

2 Rectocolonic stenosis (post-Hirschprung’s repair) Discoid pair3 Rectocolonic stenosis (post-Hirschprung’s repair) Discoid pair4 Esophageal stenosis (post-surgical repair of

esophageal atresia)Discoid pair

5 Esophageal stenosis (post-surgical repair ofesophageal atresia)

Discoid pair

6 Esophageal stenosis (post-surgical repair ofesophageal atresia)

Discoid pair

7 Esophageal stenosis (post-surgical repair ofesophageal atresia)

Discoid pair

8 Esophageal stenosis (post-surgical repair ofesophageal atresia)

Discoid pair

9 Type A Esophageal atresia, unrepaired Catheter-based pairb

10 Type A Esophageal atresia, unrepaired Catheter-based pair11 Type A Esophageal atresia, unrepaired Catheter-based pair12 Type A Esophageal atresia, unrepaired Catheter-based pair13 Type A Esophageal atresia, unrepaired Catheter-based pair14 Type A Esophageal atresia, unrepaired Catheter-based pair15 Type C Esophageal atresia, esophagus unrepaired

(tracheoesophageal fistula had been repaired)Catheter-based pair

16 Type C Esophageal atresia, esophagus unrepaired(tracheoesophageal fistula had been repaired)

Catheter-based pair

17 Type C Esophageal atresia, esophagus unrepaired(tracheoesophageal fistula had been repaired)

Catheter-based pair

a Discoid magnet pairs contained a 12 mm and a 14 mm diameter magnet with amagnetic power of 12,550 G.

b Bullet-shaped catheter-based magnets were 5 mm in diameter with a magneticpower of 12,800 G.

Fig. 1. A circular discoid magnet pair (12 mm and 14 mm diameter, respectively).Magnets are composed of neodymium, iron, and boron. The discoid magnets have amagnetic power of 12,550 G.

1132 M. Zaritzky et al. / Journal of Pediatric Surgery 49 (2014) 1131–1137

stenosis after surgical repair of Hirschprung’s disease, and 5 patientshad severe esophageal stenosis after primary surgical repair ofesophageal atresia. The discoid magnets (neodymium, iron, andboron, or NdFeB) were 12 mm and 14 mm in diameter and had amagnetic power of 12,550 G (Fig. 1). The distal 12 mm magnet wasplaced under fluoroscopic guidance via a catheter advanced over awire guide. The proximal 14 mm magnet was placed with the sametechnique, using endoscopic forceps to move the magnet intoposition, if necessary. Magnetic attraction was monitored by repeatradiographs. Upon successful anastomosis, the coupled magnet pairwas removed by endoscopic introduction of a third magnet held byforceps. Once attracted, all three magnets were removed underendoscopic and fluoroscopic guidance.

1.2. Catheter-based bullet-shaped magnet pairs

Nine patients with previously untreated esophageal atresia under-went placement of catheter-based bullet-shaped magnet pairs (CookMedical, Winston-Salem, NC; device not currently marketed). Mid-term outcomes for 5 of these patients have been previously reported[17]. Six patients had Type A esophageal atresia (an esophageal pouchand a gastric pouch with no tracheoesophageal fistula, or TEF) and 3patients had Type C esophageal atresia (an esophageal pouch and agastric pouch with a TEF). Patients with Type C atresia underwentsurgical repair of the TEF prior to magnetic compression anastomosistreatment. The gap between the upper and lower pouches wasevaluated by placement of metal probes viewed on anterioposterior(AP) and lateral chest x-rays. Only children with a gap of 4 cm or lessbetween the esophageal and gastric pouches were treated with thecatheter-based device. Additionally, all 9 patients previously under-went surgical gastrostomy and had a mature gastrostomy tract at thetime of catheter-based treatment.

The esophageal catheter used for anastomosis of primary esoph-ageal atresia contained 2 lumens. One lumen had a 5 mm bullet-shaped NdFeB magnet at the tip and the second lumen facilitated

suction of excess saliva from the upper pouch. The gastric cathetercontained 3 lumens. One lumen had a 5 mm bullet-shaped NdFeBmagnet at the tip, another lumen had a retention balloon to keep thedevice in place within the stomach, and one lumen permitted gastricfeeding. Each catheter-based magnet contained a central opening forpassage of a 0.038 inch wire guide and had a magnetic power of12,800 G (Fig. 2A, B, and C).

AP and lateral chest radiographs were obtained immediately afterplacement of the catheter-based magnets to verify appropriatecatheter position and magnet alignment. Radiographs were obtaineddaily thereafter until magnetic coupling was demonstrated. Whenanastomosis was achieved, the proximal end of the inner esophagealguiding catheter was cut, the retention balloon of the gastric catheterwas deflated, and awire guidewas introduced and passed through theesophageal catheter, through the anastomosis, and through thegastrostomy port. The esophageal catheter was pushed distallytoward the stomach while the gastric catheter was pulled awayfrom the patient, thus removing the esophageal catheter, coupledmagnets, and gastric catheter as a unit. The outer esophageal catheterwas left in place for use as an orogastric tube.

2. Results

2.1. Discoid magnet pairs

Gastroenteric anastomosis was achieved in 7 days in an 11-year-old patient with gastric outlet obstruction due to metastatic coloncancer (Fig. 3A to F). Immediately, a 25 mm diameter partiallycovered duodenal stent was placed. One week after magnetremoval, the patient required balloon dilatation for recurrentanastomotic obstruction caused by tumor ingrowth at an uncoveredend of the stent. Importantly, the patient was able to ingest anormal diet until his death, which occurred 6 months afterplacement of the discoid magnets.

Colorectal anastomosis was achieved in 8 and 10 days in 2 patientswho developed rectocolonic stenosis after surgical correction ofHirschprung’s disease. These patients were 2 years old and 3.4 yearsold, respectively, at the time of magnet therapy. In both patients,magnets were introduced through the existing colostomy. Bothpatients required anastomotic balloon dilatation 15 days after magnetremoval and 1 patient also underwent prophylactic placement of a28 mm diameter fully covered colonic stent. The colostomies wereeventually taken down in both patients and neither patientexperienced difficulty passing stool.

Esophageal re-anastomosis was achieved in an average of 6 days(range 3 days to 7 days) in 5 patients who previously underwentsurgical correction of esophageal atresia but experienced severerecurrent postsurgical esophageal stenoses that were refractory todilatation. These patients were aged 6 months to 5.9 years. Repre-sentative images from one of these patients are shown in Fig. 4A to E.

Fig. 2. (A) Catheter-based bullet-shaped magnet pair (2-lumen esophageal catheter at top, 3- lumen gastric catheter at bottom). Magnets are composed of neodymium, iron, andboron. The catheter-based magnets have a magnetic power of 12,800 G. (B) AP representation of the catheter-based magnets in situ. (C) Lateral representation of the catheter-basedmagnets in situ.

1133M. Zaritzky et al. / Journal of Pediatric Surgery 49 (2014) 1131–1137

Three of these patients also underwent prophylactic covered stentplacement 2 weeks after placement of the discoid magnets. Stentswere removed 2 months after placement.

2.2. Catheter-based bullet-shaped magnet pairs

Primary esophageal anastomosis was achieved in an average of4.2 days (range 3 to 6 days) in 9 patients with previously unrepairedesophageal atresia. The average patient age in this group was3 months (range 23 days to 5 months). Representative images fromone of these patients are shown in Fig. 5A to D. No patient experiencedan anastomotic leak. One patient developed sepsis (defined as feverand elevated white blood cell count) 48 h after magnet placement. Inthis case, the catheter-based magnets were removed, the patient wassuccessfully treated with antibiotics, and the catheter-based magnetswere replaced to complete the magnetic treatment. Eight patientsdeveloped anastomotic strictures that required dilatation and 2 ofthese patients with intractable esophageal stenosis also underwentplacement of 10 mm diameter fully covered biliary stents afterdilatation. One patient (who underwent several dilatations and stentplacement) ultimately required surgical re-anastomosis.

Three esophageal atresia patients were lost to long-term follow-up. All 6 patients with long-term follow-up data are ingesting normalresidue diets for their age. At the time of this writing, the oldest

patient is over 12 years old. Several patients in this group demon-strate common comorbidities associated with esophageal atresia. Twopatients have gastroesophageal reflux disease (GERD) and tracheo-malacia. Three patients have esophageal dysmotility requiringtreatment and 3 patients have asthma or recurrent pulmonaryinfections. None of the patients have scoliosis or rib deformities. Ofnote, 1 patient (17 months old at the time of writing) carriesconcurrent diagnoses of GERD, tracheomalacia, esophageal dysmoti-lity, and asthma/recurrent pulmonary infection. This patient is at the15th growth percentile for age at the time of publication.

3. Discussion

Temporary, minimally invasive magnet placement achieved GIanastomosis in all 17 patients in this series. Discoidmagnet pairs wereplaced under fluoroscopic and endoscopic guidance. Catheter-basedbullet-shaped magnet pairs were placed and monitored underradiographic guidance. Anastomoses were created in the esophagus,the gastric outlet, and the colon/rectum and all were accomplishedbetween 3 and 10 days after magnet placement.

In animals, the integrity of small intestinal or gastroentericmagnetic anastomoses have been shown to be equivalent to or betterthan sutured [18] and stapled [18,19] anastomoses. An animal studydemonstrated a slightly higher average burst pressure for magnetic

Fig. 3. An 11-year-old patient with gastric outlet obstruction due tometastatic colon cancer. (A) Pyloric obstruction is demonstrated endoscopically. (B) Endoscopic placement of thefirst discoid magnet distal to the pylorus. (C) Second discoid magnet in position in the stomach. (D) Initial radiologic image after magnets placed. (E) Radiologic image showingcoupling of the magnets. (F) Radiologic image after prophylactic stent placement.

1134 M. Zaritzky et al. / Journal of Pediatric Surgery 49 (2014) 1131–1137

compression duodenocolonic anastomoses when compared to stapledanastomoses [20], while another animal study demonstrated asignificantly higher burst pressure for magnetic compression jejuno-jejunal anastomoses compared to sutured anastomoses [21]. It isunlikely that a randomized controlled trial will be devised in humansto compare magnetic and sutured or stapled anastomoses, but it isimportant to recognize that the suture and staple techniques mostcommonly used in GI surgeries may not be the optimal choice.Further, magnetic compression anastomoses do not leave behindpermanent foreign bodies like sutured and stapled anastomoses.

An animal study of magnetic colorectal anastomoses describedmagnet placement through a hybrid natural orifice translumenalendoscopic surgery (NOTES) procedure [22]. Both side-to-side andend-to-side anastomoses were created in these animals. Two animalsdeveloped asymptomatic anastomotic stenosis, but without furtherintervention, at 30-day follow-up, both anastomoses were easilycrossed with the endoscope.

Eleven patients in this series required anastomotic dilatation,including eight patients who had esophageal atresia and were treatedwith catheter-based magnets. However, stenosis following magneticcompression anastomosis is not necessarily related to the catheter-based device. For example, Takamizawa et al. described a patient whounderwent surgical repair of esophageal atresia, had recurrentpostsurgical esophageal stenosis refractory to dilatation, and wastreated with one magnet suspended on a nylon thread and onemagnet introduced through the mouth [23]. Because these magnetsdid not attract, larger magnets were introduced in the same fashiontwo months later; reanastomosis was achieved but the patient stillrequired several balloon dilatations over the subsequent threemonths. In contrast, the catheter-based magnets in the presentstudy all functioned appropriately and achieved magnetic coupling inall patients.

Postoperative esophageal stenosis is also common after surgicalintervention. Outcomes of surgical esophageal atresia repair can vary

Fig. 4. A patient with recurrent postsurgical esophageal stenosis refractory to dilatation. (A) Endoscopic view of the distal discoid magnet placed through the gastrostomy.(B) Radiographic image of the distal discoid magnet in position. (C) Addition of the proximal discoid magnet placed via the mouth. (D) Radiologic image showing coupling of themagnets. (E) Endoscopic image showing the final anastomosis.

1135M. Zaritzky et al. / Journal of Pediatric Surgery 49 (2014) 1131–1137

by center (i.e., surgeon experience with the chosen surgicaltechnique) and can be influenced by the presence and degree ofunderlying comorbidities. Postoperative esophageal stenoses occur in27% to 44% of patients who undergo Type A esophageal atresia repair[24–26]. Nevertheless, the overall rate of anastomotic stenosis in thepresent study was particularly high (64.7%, 11/17); this rate can beaddressed by revision ofmagnet design and/ormagnet power. Furtherinvestigations are needed to refine the technique and to develop asuitable alternative to open surgical or laparoscopic repair.

Seven patients in this series underwent stent placement aftermagnetic compression anastomosis, including five patients who hadstents placed in the esophagus. Since there are no currently approvedpediatric esophageal stents, tracheobronchial and fully covered biliary

stents have been used in the treatment of pediatric esophagealstenosis. One group recommends consideration of tracheobronchialstent placement in the treatment of all recurrent severe pediatricesophageal strictures [27]. In adults, a study investigating magneticgastroenteric anastomosis followed by routine stent placement wasterminated prematurely due to a serious adverse event (i.e., stent-related perforation and sepsis leading to death) [28]. These in-vestigators noted that magnetic compression itself was feasible andsafe, but concluded that the requirement for subsequent stentplacement caused unnecessary risk. However, the adult patients inthat study had a primary malignant diagnosis, and thus, represent adifferent patient population than the pediatric patients with benignpathologywho are described in the present study. Even so, it would be

1136 M. Zaritzky et al. / Journal of Pediatric Surgery 49 (2014) 1131–1137

most beneficial if magnetic compression anastomoses did not requirestent placement to maintain the anastomotic tract; however, since astent is often necessary, the best type of stent (e.g., biodegradable,self-expanding, covered, or uncovered) and the appropriate duration

Fig. 5. Radiologic images from a patient with esophageal atresia (Type A or Type C converted tthe esophageal gap. (B) Image verifying initial catheter-based magnet placement. (C) Imag

of stent placement, two currently unknown factors, require furtherinvestigation.

In conclusion, minimally invasive magnet placement was feasibleusing either discoid or catheter-based bullet-shaped magnet pairs. GI

o Type A) and a gap length of 4 cm or less. (A) AP image demonstratingmeasurement ofe showing coupling of the magnets. (D) First esophagram after anastomosis.

1137M. Zaritzky et al. / Journal of Pediatric Surgery 49 (2014) 1131–1137

anastomosis was achieved in all patients. Further design refinementsare necessary to reduce the rate of post-anastomotic stenosis, butmagnetic compression anastomosis appears to be a promising newtherapeutic option.

Acknowledgments

The authors thank Andres Aguirre, Michael Brecht, and Aerie Thuofor serving as study engineers and thank Tim Anderson for his artisticexpertise.

References

[1] Aggarwal R, Darzi A. Compression anastomoses revisited. J Am Coll Surg 2005;201:965–71.

[2] Kopelman D, Hatoum OA, Kimmel B, et al. Compression gastrointestinalanastomosis. Expert Rev Med Devices 2007;4:821–8.

[3] Kaidar-Person O, Rosenthal RJ, Wexner SD, et al. Compression anastomosis:history and clinical considerations. Am J Surg 2008;195:818–26.

[4] Hendren WH, Hale JR. Electromagnetic bougienage to lengthen esophagealsegments in congenital esophageal atresia. N Engl J Med 1975;293:428–32.

[5] Hendren WH, Hale JR. Esophageal atresia treated by electromagnetic bougienageand subsequent repair. J Pediatr Surg 1976;11:713–22.

[6] Hendren WH, Hale JR. High-pouch imperforate anus treated by electromagneticbougienage and subsequent perineal repair. J Pediatr Surg 1976;11:723–33.

[7] Cope C. Creation of compression gastroenterostomy by means of the oral,percutaneous, or surgical introduction of magnets: feasibility study in swine.J Vasc Interv Radiol 1995;6:539–45.

[8] Cope C. Evaluation of compression cholecystogastric and cholecystojejunalanastomoses in swine after peroral and surgical introduction of magnets. J VascInterv Radiol 1995;6:546–52.

[9] Cope C, Clark TW, Ginsberg G, et al. Stent placement of gastroenteric anastomosesformed by magnetic compression. J Vasc Interv Radiol 1999;10:1379–86.

[10] Cope C, Ginsberg GG. Long-term patency of experimental magnetic compressiongastroenteric anastomoses achieved with covered stents. Gastrointest Endosc2001;53:780–4.

[11] Chopita N, Vaillaverde A, Cope C, et al. Endoscopic gastroenteric anastomosis usingmagnets. Endoscopy 2005;37:313–7.

[12] Takao S, Matsuo Y, Shinchi H, et al. Magnetic compression anastomosis for benignobstruction of the common bile duct. Endoscopy 2001;33:988–90.

[13] Jang SI, Kim JH, Won JY, et al. Magnetic compression anastomosis is useful inbiliary anastomotic strictures after living donor liver transplantation. GastrointestEndosc 2011;74:1040–8.

[14] Itoi T, Kasuya K, Sofuni A, et al. Magnetic compression anastomosis for biliaryobstruction: review and experience at Tokyo Medical University Hospital.J Hepatobiliary Pancreat Sci 2011;18:357–65.

[15] Itoi T, Yamanouchi E, Ikeuchi N, et al. Magnetic compression duct-to-ductanastomosis for biliary obstruction in a patient with living donor livertransplantation. Gut Liver 2010;4(Suppl 1):S96–8.

[16] Avaliani M, Chigogidze N, Nechipai A, et al. Magnetic compression biliary–entericanastomosis for palliation of obstructive jaundice: initial clinical results. J VascInterv Radiol 2009;20:614–23.

[17] Zaritzky M, Ben R, Zylberg GI, et al. Magnetic compression anastomosis as anonsurgical treatment for esophageal atresia. Pediatr Radiol 2009;39:945–9.

[18] Jamshidi R, Stephenson JT, Clay JG, et al. Magnamosis: magnetic compressionanastomosis with comparison to suture and staple techniques. J Pediatr Surg2009;44:222–8.

[19] Pichakron KO, Jelin EB, Hirose S, et al. Magnamosis II: magnetic compressionanastomosis for minimally invasive gastrojejunostomy and jejunojejunostomy.J Am Coll Surg 2011;212:42–9.

[20] Gonzales KD, Douglas G, Pichakron KO, et al. Magnamosis III: delivery of amagnetic compression anastomosis device using minimally invasive endoscopictechniques. J Pediatr Surg 2012;47:1291–5.

[21] Fan C, Ma J, Zhang HK, et al. Sutureless intestinal anastomosis with a novel deviceof magnetic compression anastomosis. Chin Med Sci J 2011;26:182–9.

[22] Wall J, Diana M, Leroy J, et al. MAGNAMOSIS IV: magnetic compressionanastomosis for minimally invasive colorectal surgery. Endoscopy 2013;45:643–8 [Epub ahead of print].

[23] Takamizawa S, Yamanouchi E, Muraji T, et al. MCRA of an anastomotic stenosisafter esophagoesophagostomy for long gap esophageal atresia: a case report.J Pediatr Surg 2007;42:769–72.

[24] Burjonrappa S, Thiboutot E, Castilloux J, et al. Type A esophageal atresia: a criticalreview of management strategies at a single center. J Pediatr Surg 2010;45:865–71.

[25] Holland AJ, Ron O, Pierro A, et al. Surgical outcomes of esophageal atresia withoutfistula for 24 years at a single institution. J Pediatr Surg 2009;44:1928–32.

[26] Antoniou D, Soutis M, Christopoulos-Geroulanos G. Anastomotic stricturesfollowing esophageal atresia repair: a 20-year experience with endoscopicballoon dilatation. J Pediatr Gastroenterol Nutr 2010;51:464–7.

[27] Best C, Sudel B, Foker JE, et al. Esophageal stenting in children: indications,application, effectiveness, and complications. Gastrointest Endosc 2009;70:1248–53.

[28] van Hooft JE, Vleggaar FP, Le Moine O, et al. Endoscopic magnetic gastroentericanastomosis for palliation of malignant gastric outlet obstruction: a prospectivemulticenter study. Gastrointest Endosc 2010;72:530–5.