Benign biliary strictures (BBSs) are usually a consequence of other conditions such as chronic pancreatitis, chronic cholangitis, and surgical procedures. BBS as a complication of surgery is commonly associated with procedures such as cholecystectomy and liver transplantation, especially live-donor liver transplantation (LDLT). BBS is much more frequent after LDLT than after orthotopic liver transplantation (OLT), with reported incidence rates of 8.3–31.5% [1]. The conventional treatment for BBS is balloon dilation followed by the insertion of multiple plastic stents via the endoscopic retrograde cholangiopancreatography (ERCP) or percutaneous transhepatic biliary drainage (PTBD) approach. These methods are effective in most patients [2], but may fail in patients with narrow or even completely obstructed strictures. Unfortunately, in such patients, open choledochojejunostomy or long-term external drainage in the form of PTBD may be inevitable.

Magnetic compression anastomosis (MCA) is a novel method for the recanalization of disconnected tracts. For MCA, a daughter magnet is placed percutaneously at the proximal end of the stricture, and a parent magnet is placed endoscopically at the distal end of the stricture. The attraction force between the two magnets leads to necrosis of the fibrotic stricture tissue and the creation of a new transmural anastomosis. This method has been successfully used for gastrointestinal anastomosis, enteroenterostomy, and choledochojejunostomy [3,4,5]. For patients with completely obstructed biliary strictures, MCA is a suitable alternative treatment that can improve the long-term prognosis. However, few reports [6,7,8,9,10,11,12,13,14,15,16,17,18] of the use of MCA for the treatment of bilio-biliary strictures are available; less than 100 such cases have been reported thus far, and most of these were from Korea, Japan, and Turkey. In China, our team is the first and only one to conduct systematic research on MCA. After thorough experiments, we have used magnets for complete biliary-enteric anastomosis [19], pancreatic intestinal anastomosis [20], and inferior vena cava reconstruction in a clinical setting. Recently, using our newly designed magnets, we were able to successfully treat several patients with biliary MCA. The aim of the present study was to determine the efficacy of this novel technique for the recanalization of complicated BBSs.

Patients and methods

Patients

We enrolled 9 patients who underwent MCA for BBSs at the First Affiliated Hospital of Xi’an Jiaotong University, between 2012 and 2018. The inclusion criteria were as follows: (1) complete BBS measuring ≤ 10 mm long, and (2) incomplete BBS measuring ≤ 10 mm long plus a history of more than 3 attempts at standard ERCP treatment (termed as refractory biliary stricture). All procedures performed in this study involving human participants were in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki Declaration and its later amendments. Informed consent was obtained from all participants prior to enrollment in the study. All patients underwent magnetic resonance cholangiography (MRCP) to localize the stricture and clarify its anatomical relationships.

MCA device

The MCA device consisted of two identical nickel-plated NdFeB magnets (rating, N45), termed the parent magnet and the daughter magnet. Each magnet was shaped like a cylinder with a tail at one end for connecting with a silk thread. We developed magnets of seven different sizes (diameters, 2–5 mm; height, 10 mm) for different clinical scenarios. For a given patient, a suitable magnet was selected according to the diameter of the duct and the features of the stricture (Fig. 1A).

Fig. 1
figure 1

A The parent and daughter magnets. B The daughter magnet is grasped with a 14-F soft pusher. C A long thread is introduced through the endoscopic channel, with the parent magnet on the tip. D The parent magnet can be grasped with an 8.5-F pusher or a snare

Full size image

Preparation for MCA

To relieve obstructive jaundice and/or biliary infection and prepare the patient for MCA, we first performed PTBD under ultrasound or fluoroscopic guidance. The bile duct in the right anterior section or segment III was targeted. MRCP is a sensitive, non-invasive test for the diagnosis of biliary stricture [21]. However, we found that MRCP was not suitable to evaluate the feasibility of MCA. Owing to the obstruction, the recipient duct was not full of bile, and therefore, MRCP could not display all the relevant details around the stricture. In contrast, cholangiography through PTBD and/or ERCP from a different direction clearly showed the stricture details, such as the length and shape of the stricture, and whether the two biliary stumps were parallel. As magnets in different planes cannot be mated together, patients with biliary stumps located in different planes were excluded.

The PTBD approach was used for percutaneous delivery. An 8.5-F biliary drainage catheter was placed at first. Every week, the catheter was replaced with a larger catheter to gradually enlarge the tract, from 8.5 to 16 F. The largest size catheter (16 F) was retained for 2 weeks to enhance the strength of the route.

MCA procedure

A 16-F sheath was inserted into the PTBD tract. Then, the daughter magnet connected to a 15-cm-long silk thread was advanced into the upper end of the stricture with a 14-F pusher (Fig. 1B). ERCP was performed using a duodenal videoendoscope (TJF 160, Olympus Optical, Tokyo, Japan). After successful papillary cannulation, a mid-endoscopic sphincterotomy (mid-EST) was performed, followed by balloon dilatation of the papilla. The balloon for this purpose was selected according to the diameter of the lower common bile duct, and usually measured 8–10 mm. After the endoscope was withdrawn, the parent magnet with a small tail was connected to the tip of the endoscope with a long thread in the channel (Fig. 1C). The endoscope was again advanced up to the papilla, and the parent magnet was carefully inserted into the common bile duct with an 8.5-F pusher (Fig. 1D). The use of our newly developed smaller magnets made this step simple. The parent magnet grasped by a snare could be directly inserted into the channel and then introduced into the common bile duct. During the insertion of the parent magnet into the lower end of the stricture, the parent and daughter magnets would attract each other once their positions were properly aligned (Fig. 2A, B). Finally, the long sheath was removed, and an indwelling 12–14-F PTBD was inserted.

Fig. 2
figure 2

Magnetic compression anastomosis (MCA) for benign biliary stricture after orthotopic liver transplantation (patient 4). A Percutaneous transhepatic biliary drainage and endoscopic retrograde cholangiopancreatography confirm complete biliary obstruction. B The daughter magnet is inserted through the sinus, and the parent magnet is advanced endoscopically. The MCA device is coupled. C Recanalization is confirmed using cholangiography. D The MCA device is retrieved, and the bile duct is continuous. E A fully covered metallic stent is deployed across the new stoma (arrow). F The stent was removed 7 months later, and endoscopic retrograde cholangiopancreatography showed a smooth and continuous duct

Full size image

Procedure after MCA

An abdominal X-ray film was routinely obtained on day 1 and day 3 or 4 after the operation to check the coupling of the magnets; further X-ray examinations were carried out if the patient’s condition necessitated this. The distance between and position of the magnets were carefully observed. When the volume of bile drained via the PTBD had markedly decreased, cholangiography was performed to confirm recanalization (Fig. 2C). If the biliary tract was recanalized, ERCP or cholangioscopy was performed to remove the MCA device (Fig. 2D) and insert multiple plastic stents or a fully covered metallic stent into the biliary tract for more than 6 months (Fig. 2E). An alternative to the placement of internal stents was the placement of a 12–16-F PTBD catheter.

Follow-up

After the successful treatment of BBS with MCA, all patients were placed under regular follow-up. Internal stents were removed more than 6 months after MCA. We evaluated their clinical symptoms, laboratory results (including total bilirubin, aspartate aminotransferase, alanine aminotransferase, and alkaline phosphatase), and imaging findings (abdominal ultrasonography). MRCP was performed if restenosis was suspected.

Results

General information

The general characteristics of the patients are shown in Table 1. The mean age of the 9 patients was 49 ± 12.9 years (range 29–74 years). There were seven male patients and two female patients. Most patients had complete BBSs. Only two patients had refractory BBSs; one of these patients had previously undergone four unsuccessful ERCP treatments, while the other had undergone three unsuccessful ERCP treatments. The BBS developed after OLT in 6 patients and after other surgical procedures in 3 patients. The patients were diagnosed with BBS at 6 ± 3.8 months after the operation, and the strictures had been present for 3 ± 5.2 months before MCA. Two patients had a history of having undergone three MCA attempts, all of which failed because of insufficient magnetic power or too great a distance between the magnets.

Table 1 General characteristics of the patients
Full size table

Outcomes of the MCA procedure

The MCA procedure was successfully performed in all 9 patients (Table 2). The diameters of the magnets used were 2–5 mm. The mean stricture length was 3 ± 1.7 mm (range 2–7 mm). Recanalization was confirmed by cholangiography in all patients after 16.3 ± 13.2 days (range 7–50 days). The abdominal X-ray examination before recanalization showed that the MCA device was well coupled in the bile duct in 6 patients. However, in 3 patients (patients 6, 7, and 8), the approximated magnets had separated. For patient 7, we immediately added one more daughter magnet, and the three magnets then closely attracted each other (Fig. 3A–C). For the other 2 patients, no additional operation was performed. However, they both experienced severe abdominal pain on the following day, and a subsequent X-ray film showed that the two magnets were again attracting each other closely (Fig. 3D–F). The pain may have been caused by the strong compressive force, and was relieved the next day.

Table 2 Intraoperative and postoperative details
Full size table
Fig. 3
figure 3

A–C Patient 7. A A plain X-ray film showing that the magnetic compression anastomosis (MCA) device is closely coupled on the first day after MCA. B On day four, the MCA device separated unexpectedly. C Another daughter magnet (4 mm) was added, and the MCA device was again tightly approximated. D–F Patient 6. D An MCA device is weakly coupled during the procedure. E The MCA device separated 2 days after the procedure. F The MCA device is closely coupled on day 4 after the patient suffered severe abdominal pain. Arrowheads, parent and daughter magnets; arrow, third magnet

Full size image

ERCP was performed in all patients to remove the magnets and insert stents through the new anastomosis. Multiple plastic stents were used in four patients, and a fully covered, self-expandable metallic stent was used in four patients. In one patient, a 16-F PTBD catheter was used instead of a stent. All patients recovered uneventfully after the last ERCP procedure. Furthermore, we measured the changes in the magnetic force in the last four patients, and found that the force was almost the same before and after MCA (Fig. 4).

Fig. 4
figure 4

Magnetic force before and after the magnetic compression anastomosis procedure is almost identical (Patient 5)

Full size image

The internal stents were removed in 7 patients at 7–9 months after recanalization. Three of these patients had multiple plastic stents (patients 1–3), and four patients had metallic stents (patients 4–7). For patient 8, the PTBD catheter was replaced with a new PTBD catheter at 6 months after MCA, and patient 9 still had internal stents at the time of this writing. After 2–66 months of stent-free follow-up, no stenosis recurrence was found. The patient with the longest follow-up (66 months) has almost normal liver function.

Adverse events

Two patients developed mild complications: patient 4 developed cholangitis, and patient 6 developed biliary bleeding. The biliary bleeding was caused by injury to the PTBD tract by the sheath, and ceased spontaneously 1 day later without requiring blood transfusion or any other treatment. Cholangitis was caused by the injection of too much contrast medium, and recovered with conservative therapy within 5 days.

Discussion

BBS is a common biliary disease, especially after iatrogenic biliary injury and liver transplantation. Since Yamanouchi et al. first reported a clinical trial using the modern MCA method [22], various animal and clinical studies have reported the feasibility and safety of MCA for cases of BBS that cannot be treated by conventional methods [6,7,8,9,10,11,12,13,14,15,16,17,18], such as ERCP and PTBD. Animal studies have shown that the MCA technique induces the development of anastomoses as evidenced by the continuity of the serosal, submucosal, and mucosal layers [23], and the patency of the anastomotic site after MCA is similar to that achieved after conventional treatments [24]. Our previous experiment also showed that vascular anastomoses formed by the MCA method were better than those formed using manual sutures in terms of continuity of the vascular intima and inflammation [25]. Thus far, 15 clinical studies have reported biliary MCA in literature (Table 3) [6,7,8,9,10,11,12,13,14,15,16,17,18]. Although most of these studies were case reports, the long-term results seem promising.

Table 3 Literature review of magnetic compression anastomosis (MCA) for the recanalization of bilio-biliary strictures
Full size table

Several different magnet-delivery routes have been reported, including a surgically formed percutaneous–jejunum tract [16, 26], a percutaneous–percutaneous tract [26], and a percutaneous–peroral tract; the last of these three routes is most commonly used for bilio-biliary strictures. Technical difficulties in magnet delivery during the MCA procedure are mainly related to the insertion of the parent magnet through the duodenal papilla. In this study, we used mid-EST and balloon dilation to expand the papilla, making it easier to advance the magnet into the biliary tract. Some researchers have inserted a fully covered metallic stent across the papilla 1 day before the MCA operation, and withdrawn the stent after the MCA device was approximated [8, 10,11,12]. However, the use of metallic stents is expensive, and the above techniques lead to an increase in the number of ERCP sessions. In our opinion, mid-EST with 8–10-mm balloon dilation is sufficient to allow the insertion of a magnet across the papilla. Using a smaller magnet is another method to overcome the above problem. Park et al. [17] developed a Φ2.4-mm magnet for MCA; this magnet can be passed through an endoscope channel and be delivered into the bile duct with a 7-F pusher, provided that only mid-EST is performed.

The stable coupling of the MCA device is a key point for the success of the MCA technique. Studies have reported success rates of 77.7–89.7% for MCA [12, 16, 17]. The failure of MCA is mainly attributable to the following factors: (1) Adequate magnetic force is the most important factor for successful MCA, but the optimal magnetic power for human subjects is unknown. One study has applied a magnetic force determination algorithm (MAGDA) [27] to quantify the magnetic strength. Magnetic force can be increased by the use of larger magnets or by the addition of multiple magnets (as in the case of patient 7 in this study). (2) Stricture length also affects MCA-device approximation. Most studies report that the stricture length should be below 15 mm [6,7,8,9,10, 13, 17]. In the present study, the stricture length was strictly limited to be within 10 mm, and the median stricture length was 3 mm. (3) A slim or tortuous duct could impede the advancement of the magnet and result in a greater inter-magnet distance and lower magnetic force. The unsuccessful MCA attempts in patients 4 and 7 were caused by this reason; we therefore selected smaller magnets that could be approximated more closely, and successfully performed MCA in these patients. (4) Inconsistencies in the axial directions of the two biliary stumps could mean that the magnetic poles are not in the same plane and result in failure of the technique [12, 17]. We excluded patients with such inconsistencies via preoperative PTBD and ERCP evaluation.

Recanalization time was difficult to estimate. Several methods were used to detect recanalization in this study, including changes in the bile volume obtained via PTBD, stool color, and magnet position within the duct. The final confirmation was made using cholangiography. Recanalization time is closely related to stricture length, magnetic force, and stricture characteristics. Matsuno et al. [9] recanalized a 2-mm stricture with an MCA device in 10 days. Okajima et al. [7] also performed MCA to treat a 2-mm stricture, and were able to retrieve the magnet after 42 days. Among the 22 cases of MCA reported before 2014, most strictures had developed after LDLT; the average stricture length was 6.4 mm (2–15 mm), and the average recanalization time was 53.3 days (9–181 days) [28]. Recently, Parlak et al. performed MCA for the treatment of BBS after LDLT [17]; they reported that the mean stricture length was 4 ± 1.2 mm (2.5–6 mm), and recanalization occurred 8.1 ± 4.7 days (2–14 days) after MCA. In our study, the median recanalization time was 16.3 days, much shorter than that reported by Jang et al. [28] This may be attributable to the use of NdFeB magnets (which are strong and stable) and the short stricture length in our study. Furthermore, it should be pointed out that since X-ray examination could not be performed every day, the recanalization times in all of the above reports may have been longer than the actual times.

Unlike conventional ERCP treatment, MCA causes ischemic necrosis of the stricture scar by magnetic compression. This is similar to resection, and may help reduce the restenosis rate. However, most authors still believe that the new anastomosis should be supported by stents for at least 6 months after recanalization [17, 18]. The type of device used to support the new anastomosis has varied among studies. Jang et al. [18] used PTBD catheters, and this method was successful in 33 of 35 patients after 175.3 days (48–381 days). In the remaining 2 patients, the stricture recurred after an average follow-up of 41.9 months (7.1–73.4 months). Parlak et al. [17] used 6- or 8-mm balloon dilation and multiple biliary plastic stents after recanalization. They reported no stenosis recurrence after 4.8 ± 3.8 months (1–11 months) of follow-up. In addition to the above two techniques, we also used fully covered metallic stents in some of our patients. Our outcomes were consistent with those of the above studies. A review of the literature showed that the stricture-recurrence rate after MCA is reported to be 3.9%, which is lower than the rates reported for ERCP (3–37.5%) [1] and PTBD (5–17.1%) [29]. Of course, it is still necessary to conduct larger clinical trials with longer follow-up periods.

The complication rate of MCA for BBS is very low. In this study, only two patients developed mild adverse events (one patient developed bleeding from the PTBD tract, and one patient had cholangitis). None of the patients developed bile leakage. We also reviewed a total of 57 cases reported in literature [6,7,8,9,10,11,12,13,14,15,16,17,18], and found that only 1 patient developed mild cholangitis, which healed with conservative treatment in 30 days [12]. In our experience, sufficient biliary drainage should be provided before and after the procedure to minimize the risk of cholangitis. No instances of bleeding from the new anastomosis or bile leakage were reported, which is consistent with the results of animal studies [30].

Although the present study demonstrates the potential advantages of MCA in the treatment of biliary strictures, this study is only a single-center clinical experience with a small number of patients. Multicenter studies with a large sample size are needed to evaluate the long-term effects of this technique. In addition, more stringent selection criteria, operational specifications, and accessories are needed to further optimize the safety and effectiveness of MCA and reduce the incidence of adverse events. The development of novel materials to strengthen the magnetic force will also improve MCA outcomes.

The clinical success rate of BBS treatment is improving because of developments in endoscopic and other interventional techniques, but some strictures cannot be treated using conventional methods. As a treatment modality for such BBS cases, MCA is safe and feasible, with a high success rate and a low stricture-recurrence rate. For effective MCA, doctors must perform accurate pre-assessment, secure the routes of magnet delivery, maintain a < 1-cm distance between the magnets, and be aware of the magnet characteristics.