Bowel perforation after liver transplantation for biliary atresia: a retrospective study of care in the transition from children to adulthood

The timing of LT for BA has remained controversial. In a retrospective cohort study of 347 pediatric patients, it was reported that the patients who were underwent KPE and required LT after the first year of age showed better patient and graft survival than those who required LT within the first year of age [11]. Although some BA patients reached adulthood with their native livers, there has been limited evidence of the impact of LT on the outcome of patients who reach adulthood after KPE. Uchida et al. [15] reported that the outcome of LT in adult BA patients was significantly poorer in comparison to pediatric patients. They noted that the rates of post-transplant intestinal perforation, intra-abdominal bleeding requiring relaparotomy and biliary leakage were significantly higher in adult patients. The cumulative 5- and 10-year survival rates were 70 and 56% in adult patients, respectively. In contrast, those in pediatric patients were 87 and 81%, respectively. On the other hand, Sampedro et al. [14] and Kyoden et al. [16] reported that the outcomes of LT were satisfactory in adult BA patients. They concluded that LT can be performed safely in adult patients. In this study, with respect to the preoperative statuses of the patients, the number of previous laparotomies and the coexistence of pulmonary precomplications were higher in Group C. The PELD/MELD scores in Group A were significantly higher in comparison to Groups B and C. This is probably because the indication for LT was portal hypertension or repeated cholangitis with relatively mild liver damage. Since almost all LT patients who were treated in our department underwent living donor LT, the GV/SLV decreased as the age at LT increased. Regarding the intraoperative outcomes, the operative time for LT and the duration of hepatectomy were significantly longer in Group C. This result indicates that prolonged adhesiotomy was required due to severe intra-abdominal adhesion in Group C; thus, the incidence of BPLT was higher in Group C. However, satisfactory patient survival was achieved in all three groups.

Complications after LT are relatively common in BA patients and result in significant mortality [28, 29]. Bowel perforation is a noteworthy complication that occurs after LT in BA patients [8‐11]. The incidence of this complication after pediatric LT is reported to be 6.4–20% [17, 19‐23]. In the present study, bowel perforation was the most common surgical complication after LT. The incidence in the present study was relatively high (15.9%)—probably because the study population included greater numbers of adolescent and adult patients than children. Various etiologies of BPLT were reported [17‐26]. One possible cause is thermal injury to the bowel during LT. In most BA patients, KPE was performed before LT and intra-abdominal adhesion was severe. Thus, we need to perform adhesiotomy carefully and gently when performing LT for BA patients. Previous studies reported that previous laparotomy and a prolonged operative time for LT were risk factors; this suggests that difficult adhesiotomy and thermal injury due to electrocautery were etiologies of BPLT. Although the number of previous laparotomies was not identified as a risk factor for BPLT in the present study, most sites of bowel perforation were observed where adhesiotomy had been performed during LT. Interestingly, the sites of perforation were observed in different locations in the patients in Group A (around the liver) and those in Groups B and C (throughout the abdominal cavity). This result indicates that KPE caused adhesion around the liver in Group A, while repeated cholangitis and multiple laparotomy caused adhesions throughout the abdominal cavity in groups B and C. Bowel injuries in the lower abdomen occurred during adhesiotomy, not only for hepatectomy but also for the construction of Roux-en Y hepaticojejunostomy in Group C. Ruptured sutures at the fixation of the Roux-en Y limb to the peritoneum due to strong traction occurred in two cases. Following these results, we usually make a Roux-en Y limb of 30 cm through the antecolic route in KPE to prevent difficult adhesiotomy and a short limb in LT. Moreover, we place adhesion barriers (Seprafilm^®, Kaken Pharmaceutical Co., Ltd, Japan) around the liver at the end of KPE. Further studies will be necessary to establish the utility of these procedures in preventing this complication. The results of the present study suggested that Seprafilm^® did not have a significant effect in the prevention of BPLT (p = 1.0, data not shown). This result indicated that the cause of adhesion was not only KPE but also the cholangitis that occurred after the Seprafilm^® was resorbed. Recent studies have shown laparoscopic portoenterostomy for BA to have equivalent outcomes to open portoenterostomy [30‐33]. Because laparoscopic portoenterostomy can decrease adhesion, further studies are necessary to evaluate its benefit in LT. Because laparoscopic portoenterostomy can decrease adhesion, further studies to evaluate its benefit to LT are required. Development of new surgical devices is also required to perform adhesiotomy safely and quickly.

Pulmonary complications are well recognized in chronic liver disease. The development of portal hypertension is fundamental in the pathogenesis [34]. The incidence of pulmonary complications in BA patients increased as the age of patients increased in the present study. It was reported that pulmonary complications were a risk factor for surgical complications (including infection, biliary complications, portal vein thrombosis and bowel perforation) after LT [35‐37]. Multiple perforation was only observed in the patients with pulmonary complications in this study. In addition to these two cases, a patient in Group C suffered from multiple bowel perforations following splenectomy before LT. Although pulmonary complications did not show strong statistical power because of the small sample size of the present study, pulmonary complications remain a risk factor for BPLT.

In our cases, fever or the elevation of inflammatory marker levels was observed in most patients, despite the patients being immunosuppressed. Bowel perforation was suggested to have triggered the severe immune response and distressed the patients. Because the sites of perforation in Group C were localized in the lower tract, peritonitis was relatively severe, even if the diagnosis was immediate. Aggressive operations, such as segmental enteral resection and primary anastomosis or enterostomy were, therefore, indicated for most of the patients in Group C. Xiong et al. [38] reported that among adults with BPLT, the patients who had severe abdominal cavity contamination tended to die despite undergoing enterostomy. Thus, an early diagnosis may ensure better survival. While the rate of mortality due to bowel perforation after pediatric LT is reported to be 30–50% [17], we did not encounter any cases of BPLT-associated mortality in the present study. Timely laparotomy and aggressive operations seemed to prevent deaths due to this life-threatening complication in our department.

As in previous reports [17, 19], a prolonged operative time for LT was found to be an independent risk factor for BPLT in a multivariate analysis with logistic regression in the present study. This result also suggests that severe intra-abdominal adhesion is a risk factor for BPLT because the duration of adhesiotomy had a strong impact on the operative time for LT. While a prolonged operative time for LT was identified as an independent risk factor, the duration of hepatectomy did not differ to a statistically significant extent. This is also depended on the duration of adhesiotomy. A possible reason for this difference is that adhesiotomy was performed not only during hepatectomy but also during Roux-en Y hepaticojejunostomy construction. In two cases of bowel perforation, the perforation occurred due to the rapture of sutures at the site of fixation of the Roux-en Y limb to the peritoneum. Although thermal injury due to electrocautery was not the etiology, adhesiotomy for the construction of Roux-en Y hepaticojejunostomy took a long time in these two cases. Thus, a prolonged operative time for LT showed strong statistical power. In this point, we must be careful until the end of the operation, even if a longer operative time is required.

In summary, the timing of LT did not affect patient survival after LT for BA. However, the incidence of BPLT was significantly higher in patients who were older than 12 years of age. Furthermore, these patients suffered from more severe peritonitis, which required aggressive surgery. The patients who required LTt after 12 years of age showed a potentially higher risk of BPLT because they required longer operations due to severe adhesion throughout the abdominal cavity and the coexistence of pulmonary complications. Since the clinical course after transplantation was complicated after adolescence, LT should be performed, as early as possible for patients who are diagnosed with progressive liver disease after KPE. Patients who undergo LT after adolescence, should be carefully observed to allow for an immediate diagnosis of BPLT and timely laparotomy should be performed to treat this lethal complication. The population of the present study was small. Thus, further studies are necessary to clarify the optimal timing of LT for BA patients.