Thoracoscopic esophageal atresia (EA) repair affords many benefits to the patient; however, intracorporeal suturing of the anastomosis is technically challenging. Esophageal magnetic compression anastomosis (EMCA) is a compelling option for endoluminal EA repair, but available EMCA devices have prohibitive rates of recalcitrant stricture. Connect-EA is a new endoluminal EMCA device system that employs 2 magnetic anchors with a unique mating geometry designed to reliably create a robust anastomosis and decrease rates of leak and stricture. We describe our first-in-human experience with this novel endoluminal device for staged EA repair in 3 patients (Gross type A, B, and C) at high risk for conventional surgical repair. First, the esophageal pouches were approximated thoracoscopically. After acute tension subsided, the device anchors were endoscopically placed in the esophageal pouches and mated. Anchors were spontaneously excreted in 2 cases. Endoscopic repositioning and retrieval of the anchors were required in 1 patient because of narrowed esophageal anatomy. There were no perioperative complications. Patients were managed for 14 to 18 months. The strictures that developed in the patients were membranous and responded well to dilation alone, resolving after 4 to 5 outpatient dilations. Gastrostomies were closed between 6 and 11 months and all patients are tolerating full oral nutrition. Early experience with this new endoluminal EMCA device system is highly favorable. The device offers considerable benefit over conventional handsewn esophageal anastomosis and anastomotic outcomes are superior to available EMCA devices.
Surgical intervention for repair of esophageal atresia (EA)1,2 is highly morbid.3 Establishing esophageal continuity with a handsewn esophagoesophagostomy is the mainstay of repair and can be performed via an open (thoracotomy) or minimally invasive (thoracoscopic) approach. Despite decades of investigation to improve outcomes, anastomotic complications, primarily leak and stricture,3,4 and musculoskeletal sequela after thoracotomy5,6 persist as major sources of postoperative morbidity. The multifactorial benefits of thoracoscopic repair, such as avoiding long-term musculoskeletal morbidities after thoracotomy,7,8 are well established.9–14 However, thoracoscopic EA repair is recognized as one of the most challenging operations in pediatric surgery.15 The technical demands of thoracoscopic repair,5,11,15 particularly the intracorporeal handsewn esophageal anastomosis,16 are reflected by the significant learning curve.12,16 Consequently, adoption of thoracoscopic repair has been limited17 and thoracotomy-related morbidities persist. Further compounding the morbidity of repair, the incidence of anastomotic complications has remained unchanged since the 1990s3,4,18,19 despite >40 years of attempts to optimize operative technique and postoperative management.20
Alternative approaches for EA repair are needed to improve outcomes. An esophageal magnetic compression anastomosis (EMCA) device21 and suture canalization have been described for long-gap esophageal atresia (LGEA).22 However, recalcitrant stricture has been reported in >50% of patients treated using the EMCA device.23 Esophageal suture canalization requires prolonged weekly esophageal dilations postoperatively and is associated with significant risk of mortality.24 Connect-EA is a novel EMCA device system designed to improve outcomes after EA repair. We report our first-in-human experience in 3 patients in accordance with the Surgical Case Report guidelines.25
Device Properties and Supporting Evidence
Connect-EA is part of the Magnamosis (University of California San Francisco Surgical Innovations, San Francisco, CA) family of devices developed on the basis of extensive preclinical evidence supporting the unique mating surface geometry26–29 that was later validated clinically.30,31 The Connect-EA device system employs two 8-mm-diameter magnetic anchors with a neodymium-iron-boron magnetic core encapsulated by biocompatible plating. The anchors are placed via endoluminal approach, with 1 anchor each placed in the upper and lower esophageal pouch. Once in proximity, the anchors auto align and mate, compressing interposed esophageal tissue (Fig 1A). The convex-concave geometry and novel curvature of the mating surface produces a tissue compression profile (Fig 1B)29 that is markedly different from other EMCA devices. The mating surface is designed to induce simultaneous central necrosis and peripheral mucosal bridging to gradually form an anastomosis. The mated anchor pair is designed to detach from surrounding tissue once the anastomosis is physiologically mature and migrate through the intestinal tract to be excreted with stool, without the risks commonly associated with accidentally ingested magnets.
A, Schematic of tissue compression between magnetic anchors. B, Computer simulation revealing the unique profile of compression produced in interposed esophageal tissue. The novel curvature and convex-concave geometry of the mating surface were designed to produce tissue compression with both axial and radial components, with the axial component decreasing as a function of the centerline. The larger amount of pressure applied at the center of the device (solid arrow) induces necrosis, whereas the light pressure applied at the periphery of the device (dashed arrow) is insufficient to induce necrosis, creating a peripheral rim of nonnecrotic tissue that confines necrosis to the center of the device and protects surrounding tissue thereby minimizing perianastomotic inflammation. The version of this work has been used with the permission of Magnamosis, Inc. Max, maximum; Min, minimum.
A, Schematic of tissue compression between magnetic anchors. B, Computer simulation revealing the unique profile of compression produced in interposed esophageal tissue. The novel curvature and convex-concave geometry of the mating surface were designed to produce tissue compression with both axial and radial components, with the axial component decreasing as a function of the centerline. The larger amount of pressure applied at the center of the device (solid arrow) induces necrosis, whereas the light pressure applied at the periphery of the device (dashed arrow) is insufficient to induce necrosis, creating a peripheral rim of nonnecrotic tissue that confines necrosis to the center of the device and protects surrounding tissue thereby minimizing perianastomotic inflammation. The version of this work has been used with the permission of Magnamosis, Inc. Max, maximum; Min, minimum.
The Connect-EA EMCA device was thoroughly tested in preclinical studies to establish feasibility32,33 and short-34 and longer-term outcomes. A mechanically robust anastomosis (supraphysiologic35 burst pressures >100 mm Hg)34 formed in 7 to 10 days, with no complications. Progressive epithelialization (stricture protective) was observed early and likely contributed to observed long-term patency and absence of complications.34
Ethics
Patients with preexisting comorbidities at high risk for surgical EA repair were considered. Consent involved both parents and included extensive discussion of the risks and benefits for all management options, including immediate and delayed surgical repair, and device-specific risks, including experimental status and lack of previous human experience. Cases were reviewed by the ethics commission of the state of Rhineland-Palatinate, which acts as both the governing regulatory body and ethics review board in Germany. Permission for use of the experimental device was granted on a case-by-case basis.
Patient Characteristics and Operative Technique
Patients had the 3 most common EA subtypes (Table 1).36 Gastrostomy was placed in the early postnatal period in all patients.
Patient Characteristics
| Patient | |||
|---|---|---|---|
| 1 | 2 | 3 | |
| Type of EA via Gross classification: description | A: pure EA without fistula | B: EA with proximal fistula | C: EA with distal fistula |
| Birth wt (percentile), g | 1830 (sixth percentile) | 1980 (40th percentile) | 3700 (56th percentile) |
| Gestational age at birth, wk | 35 | 33 | 42 |
| Comorbidities | Prematurity, small for gestational age, persistent ductus arteriosus, and persistent foramen ovale | Prematurity, maternal hepatitis B infection, and trisomy 21 (no cardiac anomalies) | Dextrocardia, ventricular septal defect, and left descending aorta (outside hospital: previous right thoracotomy, open gastrostomy) |
| Interventions before thoracoscopic esophageal approximation (d of life) | Laparoscopic gastrostomy (3) and upper pouch endoscopy (3) | Laparoscopic gastrostomy (2), upper pouch endoscopy (2), and cervical ligation of proximal fistula (2) | Open gastrostomy (1); right thoracotomy, fistula ligation, and EA repair aborted because of unusual anatomy (1) |
| Thoracoscopic approximation | |||
| Age at intervention, d | 25 | 53 | 93 |
| Operative time, mina | 119 | 102 | 130 |
| EA gap distance, vertebral bodies | 4 | 4 | 2 |
| Endoscopic anchor placement | |||
| Age at intervention, d of life | 51 | 58 | 105 (first attempt); 112 (anchor repositioning) |
| Wt at intervention, g (percentile) | 2400 (first percentile) | 2130 (below first percentile) | 5290 (39th percentile) |
| Procedure time, mina | 30 | 20 | 29 (first attempt); 25 (repositioning) |
| Patient | |||
|---|---|---|---|
| 1 | 2 | 3 | |
| Type of EA via Gross classification: description | A: pure EA without fistula | B: EA with proximal fistula | C: EA with distal fistula |
| Birth wt (percentile), g | 1830 (sixth percentile) | 1980 (40th percentile) | 3700 (56th percentile) |
| Gestational age at birth, wk | 35 | 33 | 42 |
| Comorbidities | Prematurity, small for gestational age, persistent ductus arteriosus, and persistent foramen ovale | Prematurity, maternal hepatitis B infection, and trisomy 21 (no cardiac anomalies) | Dextrocardia, ventricular septal defect, and left descending aorta (outside hospital: previous right thoracotomy, open gastrostomy) |
| Interventions before thoracoscopic esophageal approximation (d of life) | Laparoscopic gastrostomy (3) and upper pouch endoscopy (3) | Laparoscopic gastrostomy (2), upper pouch endoscopy (2), and cervical ligation of proximal fistula (2) | Open gastrostomy (1); right thoracotomy, fistula ligation, and EA repair aborted because of unusual anatomy (1) |
| Thoracoscopic approximation | |||
| Age at intervention, d | 25 | 53 | 93 |
| Operative time, mina | 119 | 102 | 130 |
| EA gap distance, vertebral bodies | 4 | 4 | 2 |
| Endoscopic anchor placement | |||
| Age at intervention, d of life | 51 | 58 | 105 (first attempt); 112 (anchor repositioning) |
| Wt at intervention, g (percentile) | 2400 (first percentile) | 2130 (below first percentile) | 5290 (39th percentile) |
| Procedure time, mina | 30 | 20 | 29 (first attempt); 25 (repositioning) |
Reflects duration of general anesthesia.
Patients 1 and 2 (Gross type A and B) were high risk for anastomotic complications (LGEA37 ; birth weight <2 kg)38–40 and complications of general anesthesia with respiratory compromise (prematurity, neonatal lung disease).40,41 Thoracoscopic suture approximation of the pouches was performed (Supplemental Video 1) to induce esophageal tissue growth with traction.42,43 Single lung ventilation with thoracoscopy was tolerated poorly, limiting future thoracoscopic repair. The EMCA device was offered to avoid thoracotomy, prolonged anesthesia exposure, and morbidities associated with delayed repair caused by prolonged hospitalization44 and oral-motor dysphasia.45
Acute tension was allowed to subside for at least 5 days after approximation. The anchors were endoscopically placed in the upper and lower esophageal pouch (via gastrostomy) (5 mm) (Supplemental Video 2). Ferromagnetic attraction to a standard endoscopic grasper was used to position the anchors.
Patient 3 (Gross type C) was transferred from an outside facility after an attempted repair via right thoracotomy was aborted because of anatomic variation and poor surgical exposure. Distal fistula ligation was successful (Table 1).
The device’s endoluminal approach eliminated the risk of operative complications related to atypical anatomy, as well as adhesive scar tissue and reoperation after thoracotomy. The initial endoscopic anchor placement was difficult. The lower esophagus was narrow and decreased maneuverability within the lumen and interfered with the anchors mating. The anchors separated on postoperative day 1 but were endoscopically repositioned and successfully mated a week later (Supplemental Video 3).
Procedures were performed with fluoroscopic guidance. All patients were extubated in the operating room.
Outcomes
EMCA was successful in all patients (Table 2). In the first 2 patients, the anchors detached spontaneously from the anastomosis (days 7 and 12) and were excreted 4 days later (Fig 2). Patient 3 showed signs of anastomotic patency (day 12), but the position of the anchors did not change. On postoperative day 14, endoscopy revealed the mated anchors unattached to tissue immediately superior to the anastomosis and were easily removed. The anastomosis was widely patent and appeared well healed, without inflammation (Supplemental Video 4).
Typical time frame and associated radiographic findings during formation of the esophageal anastomosis with our EMCA device. A, Initially, the anchors are oriented horizontally. B, The anchors partially release from the anastomosis as they detach from healed areas of surrounding tissue causing them to tilt. C, The anchors pass into the stomach. D, The anchors are then eliminated in the stool via the rectum. POD, postoperative day.
Typical time frame and associated radiographic findings during formation of the esophageal anastomosis with our EMCA device. A, Initially, the anchors are oriented horizontally. B, The anchors partially release from the anastomosis as they detach from healed areas of surrounding tissue causing them to tilt. C, The anchors pass into the stomach. D, The anchors are then eliminated in the stool via the rectum. POD, postoperative day.
Patient Outcomes
| Patient (Type) | |||
|---|---|---|---|
| 1 (Type A) | 2 (Type B) | 3 (Type C) | |
| Perioperative outcomes | |||
| Complications, description (POD) | None | None | Anchor separation (1) |
| — | — | Endoscopic device repositioning (7) | |
| — | — | Endoscopic device retrieval (14) | |
| Time to patency of anastomosis: description, d | 7: change in device orientation on radiograph | 12: anchors in situ | 12: saliva from gastrostomy (anchors in situ) |
| — | 14: anchors in stomach | — | |
| Time to device excretion: clinical indicator, d | 11: observed in stool | 16: observed in stool | 14: required endoscopic anchor retrievala |
| Total GA exposure, min | 149 | 122 | 206 |
| Follow-up | |||
| Duration of follow-up, mo | 18 | 14 | 15 |
| Wt at last follow-up (percentile), g | 10200 (26th percentile) | 7995 (10th percentile)b | 8110 (seventh percentile) |
| No. esophageal dilatations | 4 | 4 | 5 |
| Esophageal dilations, POD (balloon caliber in mm) | 30 (9) | 57 (9) | 27 (8) |
| 49 (12) | 113 (10) | 45 (11) | |
| 75 (15) | 120 (10) | 67 (12) | |
| 124 (8, calibration only) | 297 (10) | 122 (10) | |
| — | — | 214 (12) | |
| Gastrostomy closure, POD | 122 | 297 | 273 |
| Endoscopic biopsies at 1 y | Mild esophagitis limited to lower esophagus | Mild esophagitis limited to lower esophagus | Mild esophagitis limited to lower esophagus |
| Duration of PPI (omeprazole) acid suppression, mo postoperative | 0–12 | 0–12 | 0–12 |
| — | — | 15 mo to present | |
| — | — | (14 mo: dysphagia; 15 mo: endoscopy, no macroscopic changes or AS, biopsy with moderate esophagitis; 15 mo: PPI restarted; complete symptom resolution <1 wk) | |
| Clinical status at last follow-up | All PO nutrition and no limitations | All PO nutrition | All PO nutrition and soy formula (cow’s milk allergy) |
| Normal development | Mild hypotonia and mild DD (consistent with trisomy 21) | Normal development | |
| Patient (Type) | |||
|---|---|---|---|
| 1 (Type A) | 2 (Type B) | 3 (Type C) | |
| Perioperative outcomes | |||
| Complications, description (POD) | None | None | Anchor separation (1) |
| — | — | Endoscopic device repositioning (7) | |
| — | — | Endoscopic device retrieval (14) | |
| Time to patency of anastomosis: description, d | 7: change in device orientation on radiograph | 12: anchors in situ | 12: saliva from gastrostomy (anchors in situ) |
| — | 14: anchors in stomach | — | |
| Time to device excretion: clinical indicator, d | 11: observed in stool | 16: observed in stool | 14: required endoscopic anchor retrievala |
| Total GA exposure, min | 149 | 122 | 206 |
| Follow-up | |||
| Duration of follow-up, mo | 18 | 14 | 15 |
| Wt at last follow-up (percentile), g | 10200 (26th percentile) | 7995 (10th percentile)b | 8110 (seventh percentile) |
| No. esophageal dilatations | 4 | 4 | 5 |
| Esophageal dilations, POD (balloon caliber in mm) | 30 (9) | 57 (9) | 27 (8) |
| 49 (12) | 113 (10) | 45 (11) | |
| 75 (15) | 120 (10) | 67 (12) | |
| 124 (8, calibration only) | 297 (10) | 122 (10) | |
| — | — | 214 (12) | |
| Gastrostomy closure, POD | 122 | 297 | 273 |
| Endoscopic biopsies at 1 y | Mild esophagitis limited to lower esophagus | Mild esophagitis limited to lower esophagus | Mild esophagitis limited to lower esophagus |
| Duration of PPI (omeprazole) acid suppression, mo postoperative | 0–12 | 0–12 | 0–12 |
| — | — | 15 mo to present | |
| — | — | (14 mo: dysphagia; 15 mo: endoscopy, no macroscopic changes or AS, biopsy with moderate esophagitis; 15 mo: PPI restarted; complete symptom resolution <1 wk) | |
| Clinical status at last follow-up | All PO nutrition and no limitations | All PO nutrition | All PO nutrition and soy formula (cow’s milk allergy) |
| Normal development | Mild hypotonia and mild DD (consistent with trisomy 21) | Normal development | |
POD indicate days from initial endoscopic anchor placement. AS, anastomotic stricture; DD, developmental delay; GA, general anesthesia; PO, per oral; POD, postoperative day; PPI, proton pump inhibitor; —, not applicable.
Endoscopic procedure time 22 min.
Trisomy 21 growth chart.
Follow-up was 14 to 18 months (Table 2). The introduction of pureed (4 months) and solid foods (6–8 months) was successful. Gastrostomies were closed by 6 to 9 months. Patients were maintained on omeprazole for the first year. Endoscopic biopsies at 1 year revealed mild esophagitis of the lower esophagus and omeprazole was discontinued. At 14 months, patient 3 developed dysphagia. Endoscopic biopsy revealed moderate esophagitis. Omeprazole was restarted with complete symptom resolution.
Endoscopic dilations were performed if feeding difficulty was reported. A limited number of dilations were needed (4 or 5; Table 2) and none required inpatient admissions. Strictures were membranous, dilated easily, and responded well to dilations alone. Esophageal caliber with dilations was not recorded.
Discussion
Staged EA repair with the Connect-EA device system was successful in 3 patients who were high risk for conventional repair and revealed highly favorable outcomes and minimal morbidity. Endoscopic anchor placement is technically feasible with short procedure times (20–30 minutes; Table 1). In patient 3, the need for endoscopic anchor repositioning and retrieval was likely a consequence of narrow esophageal anatomy and does not reflect device failure.
There is significant potential for operative risk mitigation with endoluminal EMCA. Conventional EA repair requires prolonged exposure to anesthesia because of long operative times (open: 202 minutes46 ; thoracoscopic: 315 minutes12 ), with an increased risk for anesthesia-related complications, especially in preterm and low birth weight infants.38,40 Options for surgical approach are limited. The advantages of thoracoscopic repair over thoracotomy are well established.5,6,11 However, thoracoscopic repair requires longer operative times and greater anesthesia exposure and is physiologically demanding because of the cardiorespiratory effects of single lung ventilation.40 Endoluminal EMCA with Connect-EA mitigates the risks related to anesthesia with thoracoscopy and the morbidities associated with thoracotomy. Operative times are highly favorable (20–30 minutes; Table 1) compared with conventional handsewn anastomosis. In surgically complex patients, the endoluminal approach for esophageal anastomosis eliminates the potential risk of major perioperative complications by avoiding reoperation.
Tension, quality, and perfusion of esophageal tissue are key risk factors for complications after handsewn anastomosis.4,47 Extensive pouch mobilization and esophageal reanastomosis compromise perfusion and further impair anastomotic healing.48–50 The risk for anastomotic complications increases 15-fold with anastomotic tension after handsewn anastomosis,47 with anastomotic leak reported in 40.7% patients after LGEA repair.4 Despite multiple risk factors for complications, anastomotic outcomes in all 3 of our patients were highly favorable, with no perioperative complications, including anastomotic leak. Strictures responded well to a limited number of dilations.
On the basis of our early experience, Connect-EA appears to compare favorably to endoluminal EMCA devices described previously. An available EMCA device indicated for LGEA has been reported to result in stenosis requiring dilations in 100% of patients, with 60% going on to develop recalcitrant stricture, with several requiring surgical intervention.23 Furthermore, strictures associated with this earlier EMCA device have been reported to require an average of 13 dilations.51
The available EMCA device indicated for LGEA employs bullet-shaped magnets with a convex-convex mating configuration and small (3.33-mm-diameter), flat central mating surfaces.23 We hypothesize that the convex-concave mating configuration and novel curvature of the Connect-EA’s anchor mating geometry across the full 8-mm diameter reduces the risk of recalcitrant stricture by limiting tissue necrosis to the center of the device to ensure perianastomotic inflammation is minimized. Early neoepithelialization in preclinical studies34 and early clinical experience support our hypothesis. In contrast to the high rates of recalcitrant stricture requiring prolonged serial multimodal interventions reported after the earlier EMCA device, Connect-EA outcomes were comparable to conventional repair for patients with LGEA who require a similar number of dilations.52
Our early experience indicates that Connect-EA is a technically simple and promising endoluminal alternative for EA repair. The Connect-EA device system is only indicated if the esophageal pouches are in approximation; it is not indicated if a gap remains. Although patients with long-gap atresia routinely have a gastrostomy,37 the need for gastrostomy for standard risk repairs requires further consideration. EMCA has the potential to reduce operative time, provide a nonsurgical alternative for staged repair, and improve morbidity-related anesthesia and anastomotic complications. Further investigation in clinical trials is warranted, and we welcome participation of experience centers.
Acknowledgments
We thank the families for their courage and for their contributions in the advancement and innovation of the care of children with EA in the future.
FUNDING: Intradepartmental funding (Mainz, Munich) was provided. This work was also supported by federal grants P50FD003793 (San Francisco, CA) and P50FD006425 (Los Angeles, CA) from the US Food and Drug Administration.
Dr Muensterer conceptualized and performed the surgical procedure, obtained ethical approval for compassionate care, planned and completed data collection, and drafted the manuscript; Dr Evans assisted with product development, planned data collection, and drafted the manuscript; Drs Sterlin, Lindner, and König performed patient care, obtained the informed consent, and helped with the surgical procedures; Drs Sahlabadi and Arbindi assisted with product development; Dr Harrison conceptualized and developed the product, conceptualized the procedure, and planned data collection; and all authors reviewed and revised the manuscript, approved the final manuscript as submitted, and agree to be accountable for all aspects of the work.
All anonymous data concerning the patients and procedures are available from the corresponding author after reasonable request, taking into account data privacy requirements.
- EA
esophageal atresia
- EMCA
esophageal magnetic compression anastomosis
- LGEA
long-gap esophageal atresia
Comments