Figure 1. Roux-en Y hepaticojejunostomy
anatomy. Reprinted with permission from Elsevier.6
Figure 2. Endoscope navigating jejunojejunostomy;
afferent limb tattooed. Reprinted with permission from Elsevier.6
Figure 3. Short Roux limb with tattoo
on afferent limb.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Felder SI, Menon VG, Nissen NN, Margulies DR, Lo S, Colquhoun SD. Hepaticojejunostomy Using Short-Limb Roux-en-Y Reconstruction. JAMA Surg. 2013;148(3):253–257. doi:10.1001/jamasurg.2013.601
Importance When performing biliary reconstruction, one of the long-standing
tenets of surgery is that Roux-en-Y (RY) reconstruction should use
a long hepatic limb to decrease the risk for postoperative cholangitis.
However, this practice is not well supported and may also make postoperative
biliary endoscopy difficult. While some authors recommend Roux limbs
of up to 75 cm, we have routinely used a Roux length of 20 cm to facilitate
possible postoperative endoscopic access.
Objective To review our experience with short-limb RY hepaticojejunostomy
(HJ) and examine the short-term and long-term outcomes following this
procedure, as well as the success of future biliary interventions.
Design Retrospective medical record review of all patients who underwent
short-limb RYHJ by 2 surgeons (N.N.N. and S.D.C.).
Setting Tertiary care, university-affiliated teaching hospital.
Participants One hundred patients who underwent RYHJ were identified, with
30 of those patients being excluded owing to creation of an RYHJ to
intrahepatic bile ducts with concomitant liver resection.
Main Outcomes and Measures Patient records were reviewed to determine the incidence of
postoperative cholangitis and biliary stricture. Secondary outcomes
were the need for postoperative biliary endoscopy and success rates
for endoscopic biliary interventions.
Results Seventy patients underwent short-limb RYHJ over an 11-year period
(2001-2012). Indications included benign stricture (n = 18),
malignant stricture (n = 12), choledochal cyst (n = 5),
choledocholithiasis (n = 3), idiopathic cholangitis (n = 2),
and deceased donor or live donor liver transplant (n = 30).
Seven patients, including 4 liver transplant patients, developed clinical
or radiographic evidence of postoperative biliary stricture, and all
patients underwent successful endoscopic cholangiography. Four of
these patients required dilation and/or stone extraction, which were
accomplished endoscopically in all cases.
Conclusions and Relevance Short-limb RYHJ is safe and associated with a low incidence
of postoperative complications. In addition, biliary intervention,
when indicated, can be performed endoscopically with a high degree
of success. In the absence of any evidence demonstrating longer limbs
to be superior, we recommend using short-limb RY reconstruction for
The creation of a Roux-en-Y (RY) hepaticojejunostomy (HJ) is
a critical component of many types of hepatobiliary operations. Indisputable
tenets of this procedure include the creation of a durable jejunojejunostomy,
followed by the creation of a tension-free anastomosis between the
hepatic duct and the defunctionalized jejunal limb. However, other
technical considerations, including the ideal length of the defunctionalized
jejunal limb and the distance from the ligament of Treitz to the jejunojejunostomy,
are open to question. Surgical textbooks describe creating a long
jejunal hepatic limb of up to 75 cm, purportedly to decrease the risk
for enterobiliary reflux–causing cholangitis.1-4 Because of variability in training programs and regional dogma,
the practicing surgeon is left without clear guidelines in how best
to create the RYHJ.
Quiz Ref IDWith the recent advent of increasingly sophisticated
and facile endoscopic techniques,the RYHJ reconstruction has taken
on added relevance.5 The development
of single-balloon and double-balloon enteroscopy to treat biliary
pathology after HJ highlights the significance of limb length in gastrointestinal
reconstruction. The length of an RY-afferent limb has been found to
increase the difficulty of biliary endoscopy after RYHJ.5 Because of this, our practice has been to
create short-limb RYHJ for biliary reconstruction. In this technique,
both the distance of the jejunojejunostomy from the ligament of Treitz
and the length of the defunctionalized hepatic jejunal limb are deliberately
kept as short as possible (Figure 1). The current study was undertaken to review our
experience with short-limb RYHJ and to specifically examine the short-term
and long-term outcomes after this procedure, as well as the success
of future biliary interventions.
This study is a retrospective institutional review board–approved
review of consecutive patients who underwent RYHJ by 2 specialized
hepatobiliary surgeons (N.N.N. and S.D.C.) between 2001 and 2012.
Patient medical records were reviewed to determine the indication
for operation; details of the operative procedure; occurrence of postoperative
biliary complications including cholangitis, biliary stricture, and
Roux limb revision; and the need for future biliary intervention.
Complications were classified as immediate if they occurred within
90 days of operation and were deemed delayed if they occurred after
90 days. In patients requiring biliary intervention, the type and
number of treatments, as well as the success of treatment, were reviewed.
Endoscopic success was defined as the ability to reach the HJ anastomosis
on first attempt without excessive technical difficulty, as reported
in the procedure note.
To examine the most uniform population possible, we excluded
patients who underwent HJ with concomitant liver resection and anastomosis
to intrahepatic ducts (except in the case of live donor liver transplantation)
and patients without specifically documented limb lengths. The complexity
of transected liver parenchyma introduces the possibility for additional
biliary complications not specific to Roux limb length, potentially
obscuring outcomes of the area of interest—the length of Roux
Roux-en-Y HJ was performed by first creating a hand-sewn or
stapled jejunojejunostomy approximately 20 cm from the ligament of
Treitz in a manner that created a defunctionalized jejunal segment
of approximately 20 cm (Figures 1,2, and 3). The general approach was to keep
both the distance from the ligament of Treitz and the length of the
Roux limb as short as possible but as long as necessary to avoid tension.
The Roux limb was typically brought retrocolic to keep the length
short unless the anatomy was prohibitive and, in these cases, it was
brought antecolic. The HJ anastomosis was performed with interrupted,
absorbable monofilament suture using Loupe magnification and selective
silastic internal stenting (4 French). Stents were left in place until
spontaneous migration or clinical need to endoscopically remove them.
In our recent experience, the afferent limb was tattooed submucosally
with an endoscopic India ink dye at the level of the jejunojejunostomy
to provide the endoscopist a clear roadmap to navigate toward the
bilioenteric anastomosis should endoscopy be required at a later date
(Figure 2 and Figure 3).
One hundred patients who underwent RYHJ reconstruction were
identified; however, 30 patients were excluded owing to the creation
of an RYHJ to intrahepatic bile ducts with concomitant liver resection,
leaving 70 patients for analysis. No patient was lost to follow-up.
Patient demographics and the indication for RYHJ are presented in Table 1. The mean and median follow-up
times for the group were 56 months and 49 months, respectively.
Quiz Ref IDImmediate biliary complication occurred in
2 of the 70 patients reviewed (3%), with both patients requiring surgical
revision in the early postoperative period. One patient underwent
orthotopic liver transplantation and required revision of the HJ anastomosis
for biliary leak on postoperative day 14. The other patient underwent
short-limb RYHJ following common bile duct injury during cholecystectomy
and returned on postoperative day 2 for HJ revision secondary to a
bile leak. Neither of these patients developed clinical evidence of
biliary stricture in long-term follow up.
Quiz Ref IDDelayed biliary complications occurred in
7 patients (10%), including cholangitis in the absence of biliary
stricture in 1 patient (1%), anastomotic biliary strictures in 2 patients
(3%), and intrahepatic (nonanastomotic) biliary strictures in 4 patients
(6%) (Table 2). The
2 patients with anastomotic strictures included 1 patient who had
undergone live donor liver transplantation for primary sclerosing
cholangitis (PSC) and 1 patient who had undergone 2 prior bile duct
resections for papillary cholangiocarcinoma. Of the 4 patients with
nonanastomotic strictures, 3 patients developed intrahepatic biliary
strictures after undergoing liver transplantation. Two of these 3
patients had undergone liver transplantation for a diagnosis of PSC
and, in these cases, the intrahepatic strictures were attributed to
recurrent PSC. The third liver transplant patient with an intrahepatic
stricture experienced postliver transplantation biliary caste syndrome.
The fourth patient with an intrahepatic stricture after short-limb
RYHJ had a history of bile duct injury during laparoscopic cholecystectomy
and had undergone an attempt at operative repair prior to biliary
reconstruction at our center. No patient with antecolic reconstruction
experienced a complication requiring reoperation or endoscopy.
Quiz Ref IDEndoscopic retrograde cholangiography (ERC)
was attempted in all 7 patients with delayed biliary complications
after short-limb RYHJ. The techniques and instrumentation used in
these patients included balloon enteroscopy in 4 patients, a variable
stiffness colonoscope in 1 patient, and push enteroscopy in the remaining
2 patients. In all patients, the ERC successfully reached the biliary
anastomosis on the first attempt and was considered an adequate diagnostic
study (100% diagnostic success rate). Four patients required therapeutic
endoscopic interventions including anastomotic needle-knife stricturoplasty
and dilatation in 1, biliary stenting for intrahepatic stricture,
and anastomotic balloon dilation. One of the 4 patients who underwent
endoscopic intervention experienced clinical cholangitis; however,
no abnormalities were found on repeated ERC. In the 2 patients with
recurrent PSC, no intervention was performed given that there was
no therapeutically applicable stricture on ERC. Two patients required
repeated ERC for diagnosis and treatment of recurrent cholangitis;
this was again successful in both cases (Table 3).
The first patient (patient 2) underwent endoscopy with placement
of biliary stents for clinical cholangitis. This procedure documented
no anastomotic stricture. The patient returned 14 days later for removal
of the biliary stents, as the patient's liver enzymes remained elevated.
It was felt that the stents were not beneficial and could potentially
become problematic. The second patient (patient 7) underwent endoscopy
without intervention for clinical cholangitis on 2 separate occasions,
with a time interval of 38 months. No evidence of stricture, biliary
debris, or other pathology was identified and no endoscopic intervention
Of the 12 patients with malignant strictures included, 10 were
diagnosed as having cholangiocarcinoma and the remainder as having
hepatocellular carcinoma. One patient with cholangiocarcinoma developed
a benign anastomotic stricture 17 months postoperatively and was treated
successfully with endoscopy.
The creation of an RYHJ is vital to the success of many hepatobiliary
surgical endeavors. We have adopted the approach of short-limb RYHJ
in which a deliberate effort is made to keep the distance from the
ligament of Treitz to the HJ anastomosis as short as possible. In
our practice, we routinely keep this aggregate distance at or less
than 40 cm. These results show that this approach is associated with
a very low incidence of both early and late biliary complications.
In light of this, short-limb RYHJ appears to be a safe method for
biliary reconstruction. Our overall rate of biliary complications
of 13%, which includes both early and late complications, compares
favorably with the literature in which biliary complications following
standard RYHJ ranges from 7% to 38%.7-11 Equally important, this short-limb construction has allowed for
100% endoscopic success in patients who have required postoperative
biliary intervention. Although the number of cases requiring ERC was
low in our series, the high success rate was significant. In comparison,
several small series cite endoscopic success rates varying from 60%
to 90% for reaching the biliopancreatic limb in patients with standard
RYHJ, with successful ERC ranging from 46% to 80%.12-19
The short-limb RYHJ described has been part of our standard
practice for more than 10 years, but we have found no prior description
of this approach. In fact, surgical textbooks continue to espouse
a long-limb reconstruction, suggesting that the defunctionalized jejunum
should range from 40 to 75 cm in length. We have similarly found no
reports comparing short-limb and long-limb reconstruction. To our
knowledge, our report is the first to describe this technique and
its associated short-term and intermediate-term results.
The historical descriptions of the RY configuration warrant
some consideration. Cesar Roux's 1893 report of RY reconstruction
to divert biliopancreatic fluids following gastric surgery described
his experience with 50 patients, in which the jejunum was divided
15 to 30 cm below the ligament of Treitz and a jejunojejunostomy was
created about 12 cm below the gastrojejunostomy.20,21 This RY configuration
was later applied to biliary reconstruction. Later, contemporaries
of Whipple in the 1940s attempted to determine the optimal length
of a Roux limb using an animal model. The report concluded 30 cm to
be the optimal length for a Roux limb because at this length, enteric
reflux was avoided.22 Several surgical
texts1,2 and atlases
depict a long jejunal limb ranging from 40 to 75 cm in length; however,
the evidence to justify such a length is omitted.
Quiz Ref IDThe rationale for the longer length of defunctionalized
jejunum relates to concern of cholangitis secondary to reflux of food
particles into the hepatic ducts. The assumption is that a longer
limb will prevent this phenomenon. However, this logic breaks down
at 2 levels. First, there is little evidence that reflux of food or
enteric contents causes cholangitis.23 Second, there is no evidence that a longer limb prevents enterobiliary
and hepaticoduodenostomy are examples of procedures that should allow
easy reflux of proximal enteric contents into the biliary tree and
yet the risk for cholangitis with these procedures is quite low.26-28
A very important consideration in the construction of RYHJ is
forethought to address future complications and potential interventions.5 For example, we routinely tattoo the biliary
limb of the reconstruction to help guide the endoscopist should biliary
intervention become necessary (Figure
2 and Figure 3).
Our endoscopists have reported the landmark to be useful when navigating
up the Roux limb, minimizing inadvertent intubation of the blind or
distal jejunal limb. Similarly, we believe short-limb RYHJ facilitates
successful endoscopic biliary intervention in the few patients who
have biliary complications. In our series, all patients who required
biliary intervention underwent successful ERC with no need for percutaneous
intervention. Standard RYHJ with a longer limb of jejunum may make
ERC more complex and pose greater risk. While the need for biliary
endoscopy will only occur in a minority of patients after RYHJ, the
short-limb reconstruction described here makes this eventuality less
worrisome. Although the number of cases requiring ERC was low in our
series, the high success rate of this procedure was significant.
In conclusion, short-limb RYHJ can be performed in a wide variety
of biliary conditions with excellent short-term and long-term outcomes
and with a rate of biliary complications that appears similar to or
better than standard RYHJ. In addition, endoscopic biliary interventions
have a high rate of success in patients with short-limb RYHJ. These
findings suggest that long-limb RYHJ, which has long been part of
the surgical canon, is obsolete and that short-limb RYHJ should be
the preferred method of biliary reconstruction when an RY configuration
Correspondence: Seth I. Felder,
MD, Department of Surgery, Cedars Sinai Medical Center, 8700 Beverly
Blvd, Ste 8215, Los Angeles, CA 90048 (firstname.lastname@example.org).
Accepted for Publication: August 14,
Author Contributions:Study concept and design: Felder, Menon, Nissen, Margulies,
and Colquhoun. Acquisition of data: Felder,
Menon, and Lo. Analysis and interpretation of data: Felder, Menon, Nissen, and Colquhoun. Drafting
of the manuscript: Felder, Menon, and Nissen. Critical revision of the manuscript for important intellectual content: Felder, Nissen, Margulies, Lo, and Colquhoun. Statistical analysis: Felder, Menon, and Nissen. Study supervision: Nissen, Margulies, Lo, and Colquhoun.
Conflict of Interest Disclosures: None
Previous Presentation: This study was
presented as a poster at the Pacific Coast Surgical Association's
82nd Annual Meeting; February 18-21, 2011; Scottsdale, Arizona.