Abstract
Background:
Magnetic compression anastomosis (MCA) is an alternative technique for patients with long-gap esophageal atresia (EA). It allows for preservation of the native esophagus. We aimed to systematically summarize the current literature on MCA in EA.
Methods:
Studies where neonates with EA were treated with MCA devices were included, while studies on esophageal stenosis were excluded. All clinical studies, including comparative studies, case series, and case reports, were eligible for inclusion. Methodological quality assessment was performed using a validated tool.
Results:
Twelve studies with a total of 42 patients were included in this review. There was a wide variation among these studies with regard to the time of initiation of MCA (1 day to 7 months), procedure time (13–320 minutes), and magnet characteristics (strength, size, and shape of the magnets used). The time to achieve anastomosis ranged from 1 to 12 days. Stricture at the anastomotic site was reported in almost all the patients, which required multiple endoscopic dilatations (median no. of dilatations/patient = 9.8). Stent placement for refractory stricture was required in 9 (21%) patients, and surgery for stricture was required in 6 (14%) patients. Long-term outcomes included esophageal dysmotility (n = 3) and recurrent pulmonary infections (n = 3) were reported in only four studies.
Conclusion:
As per the findings of this review, neonates with long-gap EA undergoing MCA would invariably require multiple sittings of endoscopic dilatations (median no. of dilatations/patient = 9.8). Also, there is a wide variation among the included studies in terms of the procedure of MCA. Future studies with a standardized procedure for achieving MCA are needed to determine additional outcomes in this fragile patient population.
Introduction
Esophageal atresia (EA) is a rare malformation affecting infants with a birth incidence of ∼1 in 3000 live births. 1 EA is characterized by a discontinuous esophagus with or without a tracheoesophageal fistula. The spectrum of EA is varied with many different classifications. The surgical treatment landscape for EA has undergone considerable advancements over the years, yet the morbidity associated with long-term outcomes remains high, indicating the demand for improved interventions. 2 The thoracoscopic approach was introduced as a minimally invasive intervention that offered promising prospects, but technical complexities have hindered its widespread acceptance. 3 Nevertheless, outcomes seem at least equal between thoracoscopy and thoracotomy.4,5 However, the treatment of long-gap EA continues to pose significant difficulties.
The technical challenges associated with long-gap EA often derive from the tension generated while trying to approximate the esophageal segments. In response, different surgical strategies have been conceived, particularly for patients unsuitable for initial primary repair of EA.6,7 While the thoracoscopic approach entails less morbidity for fistula ligation without reconstruction, it has also proven helpful in evaluating the length between the esophageal segments. 8 Nonoperative maneuvers for shortening the gap include spontaneous growth and different methods of bougienage. Operative measures are many, with evidence supporting best practices being low and no single method superior to the others. 7 Although esophageal replacement techniques can create continuity without tension, they fail to maintain the functionality of the native esophagus.9,10
Evidently, longer distances between esophageal segments intuitively pose greater challenges and the standardization of classifications to steer effective treatment still needs refinement. Thus, while substantial progress has been made in managing EA, the search for more effective, minimally invasive, and functional-preserving strategies, particularly for long-gap EA, continues.
To circumvent tension and avoid multiple surgeries but at the same time still maintain the native esophagus, magnet-assisted bougienage of the proximal and distal segments was introduced as a case report by Hendren and Hale 11 and as a case series in 2009 by Zaritzky et al. 12 The initial technique was not without limitations and frequently yielded anastomotic strictures demanding additional interventions. 13 Magnet-assisted approximation techniques have evolved since then, both in terms of the magnet characteristics and application. 14 The magnetic approach involves magnets to apply a constant approximating force between the proximal and distal esophageal segments, through which an iatrogenic perforation generates a full-thickness compression anastomosis. Once the magnets couple, the intervening tissue, which is compressed by the magnetic force, becomes ischemic, and then sloughs off centrally, while the outer anastomotic rim heals. 15 Magnetic compression anastomosis (MCA) has been suggested as a minimally invasive alternative method for patients with long-gap EA.
We sought to systematically characterize the current literature for all MCA techniques in EA. Our main outcome measures were to assess the utility, safety, complications, and long-term outcomes of the MCA approach in these children. Our secondary outcome measures were to identify the different sizes, shapes, and strengths of the magnets used in different MCA techniques.
Materials and Methods
Search strategy
The literature search was conducted as per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Two investigators (N.K. and S.A.) independently conducted the searches on PubMed, EMBASE, Scopus, and Web of Science databases on June 3, 2022. The search keywords used were (“magnetic compression anastomosis” OR “magnamosis”) AND (“esophageal atresia” OR “tracheoesophageal fistula”). The total search records were analyzed and duplicates were removed. Subsequently, the eligibility criteria were applied to screen the studies.
Eligibility criteria
The inclusion criteria were as follows: all neonates with EA in which magnetic compression devices were used to achieve anastomosis. Exclusion criteria were studies in which such devices were used for esophageal stenosis. In view of paucity of literature on this technique, all clinical studies, including comparative studies, case series, and case reports, were included. Literature reviews, commentaries, conference abstracts, and opinion articles were excluded.
Data extraction
Search results were obtained by two independent researchers (A.V. and J.D.). Extracted information was as follows: first author's name, publication year, sample size, gap between the upper and lower pouch before initiation of magnetic compression, age (in days) at the time of initiation of magnetic compression, specifications (shape, size, strength) of the magnet used, length of time to achieve anastomosis by magnetic compression, time to initiation of nasogastric and oral feeds, failure rate, length of the hospital stay, incidence of complications such as stricture or leak, and the requirement of secondary procedures to treat complications such as dilatations. Disagreements were settled by discussion and consensus with another author (N.P.). Data synthesis was independently performed by two investigators (A.V. and J.D.) using Microsoft Excel spreadsheets.
Quality assessment
Methodological quality assessment of the included studies (case reports and case series) was performed using a tool proposed by Murad et al. 16 The tool was modified in this present study to exclude question nos. Four through six from the causality domain as these were specific for adverse drug events. The quality of the included studies was assessed independently by two authors (D.K. and N.K.).
Results
A total of 214 articles were identified through database searching and an additional 3 articles using snowballing of the references. After removal of duplicates, 201 abstracts were screened. The full text of 13 articles was screened and all these articles satisfied the inclusion and exclusion criteria. One study was excluded from the final analysis 12 since an update on this study, which included a greater number of cases, was published later by the same authors. 17 A total of 12 articles were finally included in the systematic review. A detailed PRISMA flow diagram is included in Figure 1.

PRISMA flow diagram for selection of relevant studies. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.
Study characteristics
Selected articles consisted of case series (n = 8) and case reports (n = 4). Majority of the published studies were from the United States (n = 7). Table 1 provides the summary of selected studies.
Summary of the Characteristics of the Included Studies
EA, esophageal atresia; MCA, magnetic compression anastomosis.
Details of the included subjects
Anastomosis using magnetic compression devices was attempted in a total of 44 patients. In the study by Wolfe et al., 18 the anastomosis using magnetic compression failed in one patient and the patient was excluded from the study by the authors themselves and not reported. In another patient, MCA was used for esophageal stricture. Therefore, this patient was excluded from our analysis. Finally, 42 patients were included in our analysis. The birth weight of the included babies ranged from 890 to 3700 g (median 1900 g).
Initial approximation of the pouches was tried in a few patients using thoracoscopic suture approximation (n = 5), thoracoscopic mobilization (n = 1), Foker's technique (n = 1), thoracotomy and suture approximation (n = 1), Bakes dilator (n = 2), and primary repair (n = 2). The initial vertebral gap between the ends ranged from 1.2 to 7 vertebral bodies (median 4). Table 2 provides a summary of the included subjects.
Summary of the Included Subjects
In the study by Shieh et al., 19 magnetic compression failed to create an anastomosis in all three patients as the scar tissue (created due to prior surgery to close the tracheoesophageal fistulae) prevented magnet-induced esophageal approximation, thereby leading to either magnets not attracting enough or erosion into the surrounding structures.
Details of MCA
There was a wide variation among the studies in the time of initiation of magnetic compression ranging from 1 day to 7 months (median 4 months). The procedure time was reported by only a few studies and ranged between 13 and 85 minutes (mean 36 minutes). The strength of the magnets used was 2500G in one study, 12000G in two studies, and 12800G in two studies. The size of the magnets used was 5 mm in four studies, 8 mm in one study, and 10 F in two studies. The shape of the magnets used was cylindrical in two studies, convex–concave in two studies, and bullet shaped in two studies. The time to achieve anastomosis ranged from 1 to 12 days (mean = 7.6 days).
Anastomosis could not be achieved in 3 patients in the study by Shieh et al. 19 Time of initiation of NG/oral feeds was reported by only four studies. Table 3 provides a summary of the details of MCA.
Summary of Details of Magnetic Compression Anastomosis
Complications and long-term outcomes
Stricture at the anastomotic site was reported in 36/42 patients (85%). Anastomotic leak occurred in 1 patient, which subsided with conservative management. Multiple esophageal dilatations were required for the treatment of anastomotic stricture (median no. of dilatations/patient = 9.8). Stent placement for refractory stricture after MCA was performed in 9 patients (21%). Surgery for stricture was required in 6 (14%) cases. Long-term outcomes were reported in four studies, in which complications, including esophageal dysmotility (n = 3), recurrent pulmonary infections (n = 3), tracheomalacia (n = 3), gastroesophageal reflux disease (GERD) (n = 2), esophagitis (n = 3), and oral aversion (n = 1), were reported.
Quality assessment
The detailed quality assessment by two independent observers for each study is depicted in Table 4. The domain of selection was adequately addressed by the included studies. However, the weaknesses included the ascertainment of outcomes and causality domains as the long-term outcomes and complications apart from stricture were reported by only a few studies. Similarly, reporting domain was not adequately addressed by half of the included studies as the details of the magnetic anastomosis was not described sufficiently. Kappa statistics showed a value of 0.964 (P < .001), highlighting a substantial agreement among the two observers.
Methodological Quality Assessment by Two Independent Observers
Discussion
MCA for EA is a compelling concept as a minimally invasive technique to achieve the continuity of the esophagus. The main advantage of this technique is the preservation of native esophagus. This technique has so far been used in a select group of centers across North America, Europe, and China and is still not standardized. Postprocedural stricture is common as the anastomosis created by magnetic compression is narrow. Therefore, multiple endoscopic dilatations are required for anastomotic strictures in all patients.
Recently, Holler et al. 14 published a systematic review of MCA, focusing on the comparison of their novel approach (4 patients) with previous reports to compare their outcomes. In contrast to the high rates of recalcitrant stricture requiring prolonged serial multimodal interventions after the esophageal magnetic compression anastomosis (EMCA) device, the outcomes of their novel Connect-EA device were comparable with conventional repair for patients with long-gap EA. The anchor pair in this novel device was designed to detach from the surrounding tissue once the anastomosis is physiologically mature and migrates through the intestinal tract to be excreted with stool, without the risks commonly associated with accidentally ingested magnets. The authors concluded there was no difference in the overall postoperative stricture. Their new approach required lesser number of postoperative dilatations and did not require any surgical procedure for the stricture.
The study concluded that MCA is a safe and technically simple alternative to treat complex or long-gap EA. Compared with the study by Holler et al., 14 we identified three additional studies (Shieh et al., 19 Evans et al., 20 and Conforti et al. 21 ) that were not included by them. We have included these three studies in the present systematic review. In addition, despite utilizing similar techniques, it was quite interesting to witness contrasting results between the studies by Holler et al. 14 and Evans et al. 20 The latter used the same “novel application technique” and the patients required postoperative dilatations comparable with the previous approaches. Moreover, one patient required a secondary procedure (stent placement) for the management of esophageal stricture.
This review highlights wide variability in the initiation and procedure times of magnetic compression among the studies. In addition, the strength, size, and shape of the magnets used also varied among the included studies. The time of initiation of magnetic compression ranged from 1 day to 7 months, procedure time ranged from 13 to 85 minutes, strength of the magnets used varied between 2500G and 12800G in two studies, the size of the magnets used ranged from 5 to 8 mm, and the shape of the magnets used was cylindrical, convex–concave, or bullet shaped.
Owing to the limited data, the above variables could not be linked to the postoperative outcomes. The time to achieve anastomosis ranged from 1 to 12 days. Anastomosis could not be achieved in four patients. The time of initiation of oral feeds after MCA ranged from 10 to 73 days. Stricture at the anastomotic site was reported in 36/42 patients (85%). Multiple esophageal dilatations were required for the treatment of anastomotic strictures in these patients (median no. dilatations = 9.8/patient). Long-term outcomes reported include complications, such as esophageal dysmotility, recurrent pulmonary infections, tracheomalacia, GERD, and esophagitis.
Complications related to the standard operative correction of EA consist mainly of anastomotic strictures, anastomotic leakage, and anastomotic dehiscence. These complications each present differently. The complication rates of operative treatment vary depending on many factors, for example, the type of EA, operative method, other medical conditions, and birth weight. The overall complication rate has been estimated to be 14%, with strictures developing in 6.1%, leakage in 2.6%, anastomotic dehiscence in 1.3%, and perforation in 1.3% of patients. 22 For long-gap EA, the reported complication rates are higher. The incidence of anastomotic leak is 28.7%–35.2%, and the incidence of anastomotic stricture 57%–67.6%, depending on the method of operative intervention.9,23 In addition to the complications presented here, patients are susceptible to long-term complications such as GERD, dysphagia, and respiratory problems.24,25
Deterioration of postoperative esophageal motility is attributed to many of the late complications. 26 The etiology of esophageal dysmotility remains unsolved, but theories of congenital defects secondary to malformations or iatrogenic causes have been proposed. 27 Esophageal dysmotility and recurrent pulmonary infections were reported in three patients each after MCA by Zaritzky et al. 17 However, with the current limited data, it remains unclear whether esophageal dysmotility is related to long-term complications of MCA.
Several aspects of the current management of EA are without consensus. Members of the European Reference Network for Rare Inherited Congenital Anomalies (ERNICA) conducted a conference to facilitate the standardization of guidelines, 28 where consensus was achieved on several technical aspects, but controversial issues remain. While MCA might emerge as a promising technique in the future, the current evidence is very limited. Therefore, it was not included in the options for operative management at the ERNICA consensus conference.
MCA is not a first-line alternative but may be considered in complex cases with a higher risk of complications. Given the current complication rates detected from the collective evidence, MCA has limited utility and patients require follow-up interventions. Stricture is uniform after MCA, and all patients require additional endoscopic dilatations of anastomotic strictures. Also, the long-term results still need to be improved. One disadvantage of the MCA technique is the usual waiting period of 2–3 months for the natural growth of the esophageal pouches, which may necessitate a lengthy hospital stay and a constant risk of aspiration pneumonia. On the contrary, in the study by Wolfe et al., 18 MCA was done successfully in a 1-day-old neonate, thereby demonstrating the feasibility of the procedure in the newborn period.
Therefore, these factors must be balanced against the merits and demerits of esophageal replacement before any definite conclusions in this regard are drawn.
There are some limitations of the present review. First, only case reports and case series are available for this technique. Further studies with a larger sample size could provide more data about the utility of MCA. Second, the characteristics of the included subjects such as age at the time of initiation of magnetic compression, initial gap between esophageal pouches varied among the included studies. Third, there is a lack of uniformity in the technique of magnetic compression used among these studies. The included studies differ with respect to the magnet characteristics (type, size, strength), the time of initiation of magnet compression, and the duration of magnetic compression used. There is also variable reporting of data on the details of the technique, initiation of feeds after successful anastomosis, etc.
Finally, due to a limited sample size, useful comparison between the various techniques of magnetic compression could not be made in this review.
Conclusion
MCA is a feasible minimally invasive technique in long-gap EA, which allows for preservation of the native esophagus. However, the patients would invariably require multiple endoscopic dilatations due to the small size of the resultant anastomosis. Long-term studies are needed to determine additional outcomes in this fragile patient population.
Footnotes
Authors' Contributions
S.A., N.K., and N.P. contributed to the study conception and design. Data acquisition was done by D.K., J.D., and A.V. Analysis and interpretation of data were done by S.A., D.K.Y., and P.G. Drafting of the article was performed by N.K., D.K., J.D., A.V., D.K.Y., and P.G. The article was revised by S.A. and N.P. The final draft of the article was read and approved by all the authors. S.A. will act as the guarantor of this article.
Disclosure Statement
No competing financial interests exist.
Funding Information
No funding was received for this article.
