Obstruction of the Biliary Limb of the Roux Loop After Successful Kasai Portoenterostomy for Biliary Atresia: Report of Two Cases

Introduction

Kasai portoenterostomy (KPE) may achieve jaundice clearance in children with biliary atresia (BA), resulting in long-term survival with the native liver [1–3]. However, cholangitis can occur despite adequate bile drainage and jaundice clearance after KPE [4]. Post-KPE cholangitis is multifactorial [4–7]. The overall incidence of cholangitis after KPE for BA ranges from 40% to 93%; up to 76% of patients experience recurrent episodes, most commonly within the first year after KPE [5–8]. Our overall jaundice clearance rate is 56.3%, and 75% during the period of the cases reported below [1,9]. A previous study showed that the incidence of cholangitis at our institution is 64.3%, with a 74% recurrence rate, a mean of 3.6 (range, 1–15) episodes per child, a peak incidence within 180 days of KPE, and a mean hospital stay of 14.8 (range, 2–64) days [10].

Recurrent cholangitis can result in substantial morbidity, increased health care resource utilization, and additional costs [10]. It may lead to progressive hepatic fibrosis, decompensation, and an increased need for liver transplantation [11]. Accordingly, strategies have been developed to prevent recurrent cholangitis; these primarily comprise pharmacologic approaches, including prophylactic antibiotics and steroids [8]. Surgical interventions have generally been reserved for drainage of bile lakes or Roux loop modification [12]. Roux loop obstruction has been rarely reported in association with recurrent cholangitis [13,14]. Here, we highlight the importance of the Roux-en-Y jejunum by presenting 2 cases of BA after KPE with recurrent cholangitis refractory to pharmacologic therapy. Dense adhesions of the Roux loop resulted in dilatation and bile stasis within the biliary limb. Surgical relief of Roux loop stasis was associated with either resolution or a pronounced reduction in cholangitis frequency.

Case Reports

Two children of Southeast Asian descent (aged 47 and 42 months) underwent KPE for BA at 33 and 54 days of life, respectively, with total bilirubin levels of 15 and 6 μmol/L at 3 months postoperatively (Table 1).

KPE was performed using a 50 cm Roux-en-Y jejunal limb in a retrocolic configuration without modification. The jejunum was transected with a linear cutting stapler 10 to 15 cm from the duodenojejunal flexure, depending on the availability of a suitable mesenteric window. The biliary limb of the Roux loop was constructed to approximately 50 cm in length. The staple line at the hepatic end was oversewn with 5-0 polydioxanone suture. Enteral continuity was restored with a 2-layer end-to-side jejunojejunostomy (inner layer: 4-0 Vicryl continuous; outer layer: 5-0 polydioxanone interrupted). KPE was completed using a single layer of interrupted 5-0 polydioxanone sutures at the liver hilum after transection of the fibrous cord of the extrahepatic bile ducts. Care was taken to avoid redundancy of the supracolic Roux limb by keeping it as short as possible; the remaining loop was positioned in the infracolic compartment. Mesenteric windows were carefully closed to prevent kinking of the bowel. A thorough saline lavage was performed to remove residual contrast from the operative cholangiogram and any extravasated intestinal secretions. Anti-adhesion barriers were not used. The abdomen was closed using a mass closure technique. Histopathologic examination of liver biopsies and the fibrous cord revealed findings consistent with BA. Post-KPE steroids are not routinely administered in our unit.

The children experienced recurrent cholangitis beginning 20 and 570 days after KPE, with cumulative hospital stays of 129 and 80 days, respectively. In case 1, cholangitis developed at 1, 2, 3, and 4 months postoperatively; in case 2, episodes occurred at 7, 8, 11, 15, and 18 months postoperatively. Cholangitis was diagnosed based on fever (>37.5°C), clinically worsening jaundice, transaminitis, or acholic stools, with or without positive blood cultures. Physical examination findings were unremarkable. Recurrent cholangitis was evaluated through abdominal radiographs (which were unremarkable; Figure 1) and hepatobiliary ultrasound, which did not show biliary lakes. Cholangitis was treated with parenteral antibiotics at therapeutic doses (either culture-directed or empiric regimens including ampicillin, gentamicin, metronidazole, or ampicillin-sulbactam in the absence of positive cultures) for at least 2 weeks, with duration guided by clinical and biochemical responses (eg, liver function tests [LFTs]).

Hepatobiliary scintigraphy using technetium-99m mebrofenin was performed to assess biliary drainage in the context of frequent cholangitis despite antibiotic therapy. This analysis demonstrated prompt hepatic uptake and excretion, with subsequent stasis in the biliary limb of the Roux loop at 6 hours in both children (Figure 2A demonstrates case 2). Laparotomy (at 4 and 19 months of age, respectively) revealed dense obstructive adhesions at the biliary limb in case 1 and at the T-junction of the enteroenterostomy of the Roux loop in case 2, resulting in dilatation of the biliary limb with intraluminal bile stasis (Figure 3). Adhesiolysis, resection of the obstructed segment, and reanastomosis were performed. LFTs showed normalization, with 0 and 3 episodes of cholangitis during follow-up periods of 33 and 23 months, respectively. Postoperative growth parameters improved (weight percentiles increased from 6 to 13 in case 1 and from 82 to 85 in case 2; height percentiles increased from 2.4 to 11 in case 1 and from 95 to 97 in case 2 at 3 and 6 months post-KPE, respectively). Scintigraphy was repeated in case 2 at 11 months post-revision due to 3 additional episodes of cholangitis; it demonstrated unimpeded drainage from the liver into the distal small bowel (Figure 2B). These episodes resolved with short courses of antibiotics, and LFTs again showed normalization. The mean length of hospital stay for post-reoperation cholangitis in case 2 was 5.6 days, whereas the overall mean length of stay for cholangitis at our institution is 14.8 days.

Discussion

Post-KPE cholangitis has been defined as recurrent (>1 episode), relapsing (>3 episodes within 6 months), or intractable (cholangitis requiring >4 weeks of antibiotic therapy or ≥3 episodes spaced less than 1 month apart) [15,16]. The proposed mechanisms – including intestinal bacterial translocation and colonization, alterations in the intestinal microbiota, hematogenous spread via the portal vein, and immune-inflammatory responses – are based on the hypothesis that enteric bacteria play a central role in the development of cholangitis. This hypothesis is supported by porcine BA models demonstrating bacterial colonization of the Roux-en-Y loop as early as 1 week postoperatively [17]. Furthermore, Levy et al recognized that mechanical obstruction of the Roux loop at the mesocolic window may occur as a late complication after KPE for BA [13]; Altman et al reported successful surgical management of Roux loop obstruction as a treatment for refractory cholangitis [18]. Additionally, De Moor et al documented cholangitis secondary to Roux loop obstruction by a biliary stone after liver transplantation for BA [19]. Finally, Houben et al described 3 children with Roux loop obstruction whose cholangitis was relieved by surgical widening of the mesocolic window in 2 cases and a Finney-type enteroplasty in the third case [14].

The clinical presentation of cholangitis due to mechanical obstruction of the Roux loop is similar to that of typical post-Kasai cholangitis without obstruction. Both early and late presentations have been reported [13,14,20]. Diagnosis is based on imaging findings of a dilated extrahepatic bowel loop, identified via hepatobiliary scintigraphy or direct percutaneous cannulation with antegrade contrast visualization of the dilated biliary limb of the Roux loop [13,14]. The index of suspicion and timing of hepatobiliary scintigraphy have been tailored to clinical circumstances, including early evaluation in cases of persistent LFT derangement during quiescent periods, regardless of the number of episodes, and in cases of intractable cholangitis [14,20]. In cases of Roux limb obstruction, operative management is required to prevent further episodes of recurrent cholangitis. Various procedures have been proposed, including adhesiolysis, revision of the jejunal anastomosis, and exteriorization of the biliary limb as a Mikulicz stoma [13,14,20,21].

Both of our patients were examined via hepatobiliary scintigraphy due to the intractable nature of the cholangitis, rather than persistent LFT abnormalities during quiescent periods. After unremarkable ultrasonography findings were noted, hepatobiliary scintigraphy was selected as the diagnostic modality of choice, instead of more invasive methods (eg, transhepatic cholangiography), to evaluate the portoenterostomy after the Kasai procedure. Hepatobiliary scintigraphy is regarded as a sensitive modality for the evaluation of neonatal jaundice in BA [22]. It has also been described as a sensitive tool for assessing biliary-enteric anastomoses [23]; its use in the evaluation of post-KPE cholangitis has been reported as a noninvasive and sensitive method for diagnosing Roux loop obstruction.

In our cases, surgical intervention was undertaken to address persistent subhepatic stasis; we aimed to exclude hilar cysts while inspecting, confirming, and relieving Roux loop obstruction as indicated by intraoperative findings. During laparotomy in case 1, adhesions were identified around the jejunojejunostomy, along with ectatic dilatation and redundancy in the supracolic portion of the Roux loop (biliary limb). In case 2, dense adhesions were noted around the jejunojejunostomy. However, enteric continuity appeared to be preserved in both cases, as demonstrated by the ability to transfer bowel contents across the jejunojejunostomy from the duodenojejunal segment to the distal small bowel. It is possible that a minor leak or a kink during the initial operation contributed to dense adhesion formation, with obstruction confined to the biliary limb. In both cases, laparotomy with adhesiolysis, resection of the jejunojejunostomy, and redo jejunojejunostomy improved bile flow, as demonstrated by postoperative hepatobiliary scintigraphy findings in the second child. The first child experienced complete resolution of cholangitis, whereas the second child showed a reduced frequency (from 5 episodes over 1 month before Roux loop surgery to 3 episodes over 23 months after surgery) and a shorter mean length of hospital stay for cholangitis (from 16 days before surgery to 5.6 days after surgery). We acknowledge that cholangitis recurrence in case 2 suggests that Roux limb obstruction was not the sole cause. Continued episodes after Roux loop surgery likely reflect the multifactorial pathogenesis of post-KPE cholangitis.

The recognition of a surgically correctable cause of recurrent cholangitis after successful KPE for BA warrants consideration of preventive strategies that may be adopted during KPE. A recent meta-analysis identified the anti-reflux valve as a protective factor against post-KPE cholangitis [24]. However, factors other than anatomy may be involved, given that neither valve creation nor Roux loop lengthening has been consistently shown to affect the incidence of post-KPE cholangitis [25]. In a prospective randomized controlled trial comparing a 30- to 40-cm Roux loop with a 13- to 20-cm Roux loop, no significant difference was observed in the incidence of post-KPE cholangitis [25, 26]. Conversely, the reconstruction technique appears to influence myoelectric activity within the Roux loop in non-BA models [27,28]. The occurrence of 2 cases within our series of 86 patients with BA may reflect either a low incidence of Roux loop obstruction or a technical factor [10].

This report does not establish a causal association between recurrent cholangitis and bile stasis within the biliary limb of the Roux loop. The presence of dense adhesions around the jejunojejunostomy in both cases underscores the importance of meticulous surgical technique, avoidance of kinking, and minimization of redundancy in the supracolic Roux limb. After surgical correction, radiologic findings demonstrated free and unimpeded flow with clearance from the biliary limb, without evidence of stasis. Strategies to reduce postoperative adhesions include minimally invasive approaches, reduced bowel handling during surgery, and the implementation of anti-adhesion barriers and solutions [29]. Although routine use of anti-adhesion agents may decrease abdominal adhesions, such agents may not eliminate obstruction and could increase the risk of anastomotic leakage [30]. Furthermore, studies comparing open and laparoscopic KPE have shown improvements in operative time, blood loss, and time to return to a normal diet, but they have not identified significant differences in complications, cholangitis, or native liver survival [31].

Conclusions

Mechanical obstruction may be associated with bile stasis in the biliary limb without overt obstruction on the enteric side of the Roux loop among patients investigated for recurrent cholangitis, even after successful KPE. Strategies to minimize Roux-loop-related complications and the need for further surgery should be considered following jaundice clearance after KPE for BA.