Endoscopic ultrasound-guided hepaticogastrostomy (EUS-HGS) is becoming a well-established method for treating obstructive jaundice. However, it has been associated with severe adverse events,^1–3 of which stent migration is the most severe and may be fatal.^3–5 According to previous reports, stent migration usually occurs within 7 days after EUS-HGS, although it may appear later. Thus, frequent follow-up by computed tomography (CT) may predict its occurrence.⁶ Herein, we present a case in which repeated monitoring using CT prevented stent migration 11 days after EUS-HGS.

A 61-year-old man was admitted to our hospital for acute cholangitis two years after undergoing subtotal stomach-preserving pancreaticoduodenectomy for distal bile duct cancer (Table 1). CT revealed dilation of the intrahepatic bile duct (IHBD) and hilar biliary obstruction, suspected to be a recurrence of the bile duct cancer (Fig. 1A). Bilateral drainage was necessary as the right and left IHBDs were separated. We decided to perform EUS-HGS. We inserted the scope (GF-UCT260; Olympus) into the stomach and punctured B3 with a 19-gauge needle (SonoTip TopGain; Medi-Globe). Cholangiography revealed severe hilar biliary obstruction (Fig. 1B). After we advanced a 0.025-inch guidewire (M-through^TM; ASAHI INTECC Co.) into the right IHBD, a double-lumen cannula (UNEVEN DOUBLE LUMEN CANNULA; Piolax Medical Devices Inc.) was inserted to assist the placement of another guidewire (VisiGlide 2^TM; Olympus) from the second lumen. We placed an uncovered self-expanding metal stent (SEMS) (Zeostent V^TM, 8 mm diameter, 6 cm length; Zeon Medical Inc.) between the right and left IHBD. The fistula dilation was performed using a slim mechanical dilator (ES DILATOR^®; Zeon Medical Inc.); subsequently, a partially covered SEMS (bare-end type, Niti-S biliary S-type^®, 8 mm diameter, 12 cm length; TaeWoong Medical) was placed from the liver to the stomach. We released the stent using intra-conduit release method to prevent stent migration.⁷ The metallic stents were successfully placed without immediate postoperative complications (Fig. 1C, 1D).

Figure 1. (A) Abdominal computed tomography revealed hilar biliary obstruction with intrahepatic bile duct (IHBD) dilation. (B) The left and right IHBDs were separated by the fluoroscopic image. (C, D) We confirmed that the length of the metallic stent on the gastric side was sufficient with fluoroscopic and endoscopic images.

Table 1 . Laboratory Data on Admission.

Parameters	Levels
Hematology
WBC (/μL)	9,270
RBC (/μL)	3.22×106
Hemoglobin (g/dL)	11.0
Hematocrit (%)	33.7
Platelets (/μL)	250×103
Neutrophils (%)	84.0
Eosinophils (%)	0.6
Lymphocytes (%)	13.4
Coagulation
Prothrombin activity (%)	83
Partial thromboplastin time (s)	26.8
Tumor marker
CEA (ng/mL)	8.9
CA 19-9 (U/mL)	7,568
Serum chemistry
Total protein (g/dL)	7.1
Albumin (g/dL)	3.3
Blood urea nitrogen (mg/dL)	14
Creatinine (mg/dL)	0.97
Total bilirubin (mg/dL)	1.2
Direct bilirubin (mg/dL)	0.4
AST (U/L)	34
ALT (U/L)	30
Lactate dehydrogenase (U/L)	255
ALP (U/L)	2,464
γ-GTP (U/L)	714
Amylase (U/L)	80
Na (mmol/L)	138
K (mmol/L)	3.8
C-reactive protein (mg/dL)	6.11

WBC, white blood cells; RBC, red blood cells; CEA, carcinoembryonic antigen; CA 19-9, carbohydrate antigen 19-9; AST, aspartate transaminase; ALT, alanine transaminase; ALP, alkaline phosphatase; γ-GTP, gamma-glutamyl transpeptidase; Na, sodium; K, potassium.

We confirmed the stent alignment by performing CT immediately after the EUS-HGS (Fig. 2A). Although CT performed the next day revealed that the distance between the liver and stomach had increased because the stomach wall had returned to its original position (Fig. 2B), the length of the stent to the gastric side was still sufficient. However, CT performed 11 days after EUS-HGS revealed that the stent on the gastric side was significantly shortened (Fig. 2C). To avoid stent migration, we performed an endoscopic procedure using gastroscopy (GIF-Q260J; Olympus) and confirmed that the stent had nearly migrated out of the gastric lumen (Fig. 3A). First, a guidewire was inserted into the IHBD through a metallic stent. Second, the SEMS was pulled back as much as possible by grasping it with forceps. Lastly, we inserted a fully covered SEMS (X-suit NIR^®, 8 mm diameter, 6 cm length; Olympus) into the previous stent using the stent-in-stent technique (Fig. 3B).

Figure 2. (A) A computed tomography (CT) image obtained immediately after endoscopic ultrasonography-guided hepaticogastrostomy (EUS-HGS). The double yellow arrow denotes the length between the stomach wall and liver; we confirmed that the stent was properly positioned. (B) According to this CT performed on the next day following the EUS-HGS, the distance between the left liver and the outer gastric wall became longer. (C) CT performed 11 days after the procedure showed that the proximal end of the stent had nearly migrated.

Figure 3. (A) The metallic stent remained slightly in the luminal cavity. (B) We inserted another stent into the previous stent to avoid stent migration. (C) Computed tomography performed 40 days after the first procedure showed that the gastric side of the previous stent (arrow) had migrated into the gastric wall. (D) The length of the second stent on the gastric portion (arrowhead) remained appropriate.

The patient was discharged 3 days after the second procedure with no complications. A CT scan 40 days after the initial procedure revealed that the patient had increased ascites and the gastric side of the initial stent had migrated; however, the second stent was in an appropriate position (Fig. 3C, 3D). The patient remained free of obstructive jaundice until he died of cancer progression 101 days after the EUS-HGS.

The study was approved by the ethical committee of Kyushu University Hospital (approval no. 23059). The informed consent was waived because of the retrospective nature of the study.

Stent migration, a severe complication of EUS-HGS, has an occurrence rate of 1.7%–2.7%.^1,3 Although some preventive methods such as the “ClipFlap” anchoring method and Crisscross anchor-stents have been proposed,^8,9 implementing them is impractical due to high costs and invasiveness. It is difficult to prevent stent migration completely, even when we employ the intra-scope channel stent release technique, which has been reported to be effective in preventing stent migration.⁷ Martins et al⁵ described a case of a patient who died from stent migration 5 days after successful metallic stent placement using EUS-HGS. Here, we describe a case in which frequent CT follow-ups helped to prevent stent migration after EUS-HGS.

CT is a costly but minimally invasive examination, and reports have demonstrated its efficiency in preventing stent migration after EUS-HGS.^6,10 Ochiai et al¹⁰ utilized CT to measure the distance from the stomach to the liver before EUS-HGS and compared migration and non-migration groups. In the migration group, the distance before EUS-HGS was significantly longer than that in the non-migration group (47.0 mm ± 11.3 mm vs. 20.1 mm ± 12.8 mm, P < 0.001). Thus, a long distance from the stomach to the liver on preoperative CT indicated a high risk of stent migration. Yamamura et al⁶ compared the length of the stent from the gastric wall to the liver using CT at the early (day 1) and late phases (after 7 days). The stent length in the abdominal cavity in the late phase (8.55 mm ± 2.4 mm) was longer than that in the early phase (7.13 mm ± 2.1 mm).⁶ This is because the gastric wall returns to its original position after EUS-HGS and this usually occurs within several days after stent placement. Therefore, CT performed days after EUS-HGS is as effective in preventing stent migration as preoperative CT.

Recently, the usefulness of a partially covered SEMS with anchoring properties (Niti-S Spring Stopper Stent [SSS]; Taewoong Medical) for preventing stent migration after EUS-HGS has been reported. Ishii et al¹¹ reported a retrospective pilot study of about 37 patients who were performed EUS-HGS using SSS. The rates of technical (97.3%) and clinical successes (89.2%) were high and no stent migration occurred. The strong anchoring force of SSS also has a possibility to be effective for preventing stent migration.¹² However, there is a report of a case in which the hepatic side of the SSS migrated into the abdominal cavity. Further investigation is required to determine the best stent for EUS-HGS.

In this case, immediately after the EUS-HGS procedure, CT confirmed that the distance from the stomach wall to the liver was short (1.6 cm), while the metallic stent on the gastric side was long (4.9 cm) (Fig. 2A). After monitoring with CT, we observed that the stent length in the abdominal cavity increased and the length on the gastric side decreased, indicating the risk of stent migration. Considering this, we conducted CT 1 and 11 days after the EUS-HGS procedure. The stent length between the stomach and liver increased to 3.2 and 5.8 cm 1 day after EUS-HGS, respectively. Additionally, the stent length on the gastric side shortened to 2.9 and 1.1 cm after 1 and 11 days, respectively (Fig. 2B, 2C). As the latter CT indicated a risk of stent migration, we placed an additional stent.

Although follow-up CT was conducted 11 days after EUS-HGS in our case, previous reports have shown that stent migration can occur within 7 days after EUS-HGS.^5,13 Yamamura et al⁶ performed follow-up CT at least one week after EUS-HGS (median: 14 days, range: 7–25 days) and no cases of stent migration were observed. However, here, the risk of stent migration persisted 7 days after the procedure. A rapid increase in the volume of ascites due to disease progression may also be a risk factor for stent migration. Although it is difficult to decide a specific interval or timing for follow-up CT, closely repeated CT scans may be appropriate in cases where the distance from the stomach to the liver is large on pretreatment CT or when uncontrollable ascites is expected to be increased rapidly. Additional studies must be conducted to determine the optimal timing of follow-up CT after EUS-HGS.

In conclusion, this case demonstrated that stent migration may occur even 7 days after EUS-HGS and repeated CT monitoring may help prevent it.

None.

The data that support the findings of this study are available from the corresponding author upon reasonable request.

No potential conflict of interest relevant to this article was reported.