Middle hepatic vein (MHV) reconstruction with an interposition vessel graft has been established as a standard procedure for living donor liver transplantation (LDLT) using a modified right lobe graft. MHV reconstruction resulted in a new need for vascular allografts in the field of LDLT. Increasing demand for LDLT volume leads to significant shortage in vessel allografts supply. Regarding availability, synthetic vascular grafts have a definite merit of unlimited supply, and their patency rates were acceptably high [1-3]. However, unwanted migration of the synthetic vascular graft into the hollow viscus has been sporadically reported worldwide [3-10]. We previously used ringed expanded polytetrafluoroethylene (PTFE) grafts for a long period, but that were replaced with collagen-impregnated woven double-velour polyester grafts (Hemashield Platinum, Maquet; Rastatt, Germany; Hemashield grafts) because of its limited supply and accidental PTFE graft migration [11]. Even after conversion to use Hemashield grafts, we encountered recurrence of incidental graft migration in a LDLT recipient. Thus, we present a case of Hemashield graft migration into the duodenum following LDLT.

This patient was a 64-year-old man with hepatitis B virus–associated liver cirrhosis and hepatocellular carcinoma, with an end-stage liver disease score of 25. This patient had undergone transarterial chemoembolization nine times before transplantation. Because of deterioration of liver function and altered mentality (Fig. 1A, B), we decided to perform LDLT. The donor was his 35-year-old daughter. An LDLT operation using a modified right lobe graft was performed. The V and VIII vein openings in the right hepatic graft were reconstructed with a 10 mm-sized Hemashield graft, which was anastomosed to the common opening of the recipient’s middle-left hepatic vein trunk (Fig. 1C, D). The patient recovered uneventfully after LDLT operation and was discharged 13 days after transplantation.

Figure 1. Computerized tomography (CT) findings before and after liver transplantation. Pretransplant CT images showed over liver cirrhosis (A, B). CT scan taken 1 week after transplantation showed no abnormal findings with good patency of the Hemashield graft conduit (C, D) (arrows).

This patient underwent dynamic computerized tomography (CT) at each follow-up (Fig. 2). CT scans taken at one year (Fig. 2D), and two years (Fig. 3A, B), revealed no abnormal abdomen findings. However, a gastroduodenoscopic examination at two years revealed accidental migration of the Hemashield graft into the duodenal bulb (Fig. 4). The patient had no signs or symptoms and no problem with diet. The Hemashield graft migration was identified by retrospective review of the 1-year CT (Fig. 2D) and 2-year CT scan (Fig. 3A, B).

Figure 2. Follow-up computerized tomography (CT) scans. The Hemashield conduit showed good patency at the 1-month CT (A), partial occlusion at the 2-month CT (B), and near-complete occlusion at the 3-month CT (C). One-year CT scan showed complete luminal occlusion of the Hemashield graft conduit, but its location within the duodenum was not identified (D). Arrows indicate the location of a Hemashield graft conduit.

Figure 3. Follow-up computerized tomography (CT) scans with overt findings of Hemashield graft migration (arrows). A Hemashield graft was located within the duodenum in the 2-year CT scan, but this finding was missed due to lack of radio-opacity (A, B). Follow-up CT scan taken after two years and six months showed no significant change in the location of the migrated Hemashield graft and its surrounding tissue reaction (C, D). Arrows indicate the location of the Hemashield graft.

Figure 4. The gastroduodenal endoscopic finding of Hemashield graft migration taken years after transplantation. The distal stump of the Hemashield graft is located within the duodenum (A). The Hemashield graft penetrates the duodenal wall without severe inflammation at the duodenal wall (B).

Because of the potential risk of Hemashield graft migration-associated complications, surgical removal was recommended, but the patient wished to observe more with taking CT scans more frequently. A follow-up CT scan taken at two years and six months showed no significant changes compared with the 2-year CT scan (Fig. 3C, D). The patient is doing well two years and six months after LDLT.

MHV reconstruction with vascular graft interposition is widely accepted as an essential procedure for LDLT using a modified right liver graft. The increased frequency of LDLTs has led to relative lake of vessel allograft supply. Cryopreserved iliac vein allografts have traditionally been considered the most suitable conduit for MHV reconstruction, but its supply shortage has been a critical problem. Synthetic vascular grafts have a definite advantage of unlimited supply regarding availability as well as high patency rates [1-5].

However, we experienced accidental displacement of a PTFE graft into the hollow viscus [3,5]. Migration of such foreign material into the stomach or duodenum can trigger life-threatening complications, necessitating removal. Hsu et al. [4] reported that 1.5% of the patients developed complications related to the PTFE graft. One patient developed fulminant thrombosis with sepsis at 24 months and died of multiple-organ failure; the other three patients experienced graft migration to the second portion of the duodenum, which required surgical removal.

The physical features of Hemashield grafts are similar to PTFE grafts, but the former has some merits. Hemashield grafts have the circular pleats (Concentricrimp, Maquet) and a thin-walled flexible structure formed using a woven double velour polyester material, making it easier to handle and sew. Its longitudinal colored lines (Guideline Strip, Maquet) are also helpful in correcting the alignment during anastomosis. In addition, its bovine collagen-impregnated and the woven double velour structure facilitates prevention of early luminal thrombus formation and tissue reaction at the anastomosis sites [11]. However, Hemashield grafts are not radio-opaque, unlike our expectations.

Because of these physical features of Hemashield grafts, we expected the risk of accidental graft migration to be lower than that associated with PTFE. We did not experience accidental migration in the first two years after using Hemashield grafts [11]. However, we encountered an accidental Hemashield graft during 5-year experience with more than 600 cases of LDLT using Hemashield graft.

Until now, accidental migration of synthetic vascular grafts into the hollow viscus could be easily detected on CT images because PTFE graft is radio-opaque, thus suggesting contrast-enhanced CT as the best early detection of synthetic vascular graft-associated complications [3,10]. However, two sessions of contrast-enhanced CT scan missed Hemashield migration in the present case, which was retrospectively identifies after the performance of endoscopic gastroduodenoscopy. This experience makes two points. First, the Hemashield graft is not radio-opaque, thus special care should be paid to identify its location. We strongly recommend adding radio-opaque markers to Hemashield grafts during manufacturing. Second, Hemashield grafts also have a penetrative force comparable to that of PTFE grafts. Final closure of synthetic vascular grafts with a sizeable surgical polymer clip (Hem-o-lok Weck; Teleflex, Wayne, PA, USA) may increase the risk of graft migration, as this clip may be associated with strong tissue reaction and adhesion. Thus, we recommend sewing the graft end manually rather than end clipping.

So far, we have treated surgically displaced synthetic vascular grafts through surgical removal after laparotomy [3,5]. However, to date, there is no definite consensus on the treatment of such graft migration into the hollow viscus. Most patients underwent definitive treatment including surgical removal, but some remained asymptomatic and under close observation at the time of writing [3-10].

In conclusion, every synthetic vascular graft can penetrate into the adjacent organs and soft tissues and its incidence may not be negligibly low. Thus, lifelong surveillance is necessary to detect unexpected rare complications in LDLT recipients who had MHV reconstruction using synthetic vascular grafts.

There was no funding related to this study.

All authors have no conflicts of interest to declare.

Conceptualization: RKO, SH. Data curation: RKO, CSA, DBM, GWS, DHJ, GCP, WHK, SH. Formal analysis: CSA, DBM, TYH, GWS, DHJ, GCP, WHK, YIY, HL, SH. Funding acquisition: TYH. Investigation: RKO, CSA, TYH, GWS, DHJ, GCP, WHK, YIY, HL, SH. Methodology: All. Supervision: SH. Visualization: SH. Writing – original draft: All. Writing – review & editing: RKO, SH.