Technically, using the gallbladder as the biliary conduit is undoubtedly the fastest and the easiest approach to bilioenteric bypass. Distal CBD obstruction usually leads to a markedly dilated gallbladder (Courvoissier’s sign), making access to the gallbladder quite easy. When choosing the gallbladder as the conduit, the surgeon must consider the proximity of the neoplasm to the entrance of the cystic duct into the extrahepatic biliary tree—this consideration will be discussed fully below in the section Selection of Bilioenteric Bypass.

Historically, cholecystogastrostomy and/or cholecystoduodenostomy were the early operations for bilioenteric bypass because of the proximity of the gallbladder to these organs and because of the ease (and safety) of the anastomosis. Mobilization of the fundus of the dilated gallbladder without disruption of the cystic artery will allow the gallbladder to be “folded over” for anastomosis to the stomach or duodenum. Enthusiasm for this procedure waned in the 1960’s as more evidence developed showing problems with bile gastritis/esophagitis, bilious vomiting, and ongoing problems when the duodenum became obstructed from local progression of the pancreatic neoplasm. This approach to bilioenteric bypass is largely of historic interest even in benign conditions because of the reasons described above, but also because the anastomosis often strictures over time (not really a problem when used in the palliation of pancreatic cancer).

Cholecystojejunostomy: Due to the mobility of the jejunum, this anastomosis is the quickest, easiest, and safest form of bilioenteric drainage and does not require any mobilization of the gallbladder. A loop of proximal jejunum is brought antecolic and sewn to the anterior aspect of the gallbladder fundus; in rare cases, a retrocolic approach through the right transverse mesocolon may be easier, but usually not. The anastomosis can be created in a hand sewn, usually one layer fashion due to the thin-walled gallbladder (Fig. 5.1A), or in stapled fashion (Fig. 5.1B); if a gastrojejunostomy is also planned, an intraluminal circular stapler can be introduced in the enterotomy to be used for the gastrojejunostomy and passed distally to a site where a stapled circular anastomosis can be created (usually a 21 or 25 mm stapler is sufficient) (Fig. 5.1C). Some surgeons then create a Braun-type, side-to-side enteroenterotomy between the afferent and efferent limbs of the loop cholecystojejunostomy but most do not believe this diverting enteroenterostomy to be necessary. Overall, this is a very safe anastomosis with little morbidity.

One very important point needs to be stressed. Strong consideration needs to be directed at the possibility of future obstruction of the cystic duct where it enters the CBD as the pancreatic neoplasm enlarges; obstruction of the cystic duct will lead to recurrent biliary obstruction. If contemplating a cholecystojejunostomy, the gallbladder should be dilated or, if chronically inflamed, should contain bile. Some advocates of this easy anastomosis suggest a cholecystogram to confirm both patency of the cystic duct outflow as well as the site of entrance of the cystic duct into the CBD and its distance from tumor. If the cystic duct enters the CBD low near the duodenum, or if the neoplasm is in close proximity to the entrance of the cystic duct, using the gallbladder as the biliary conduit is contraindicated.

Most HPB surgeons currently use the common hepatic duct as the biliary conduit for a palliative bilioenteric bypass. The rationale behind this somewhat more involved approach than using the gallbladder or the CBD is because of worry related to local progression of the pancreatic neoplasm potentially involving the entrance of the cystic duct into the CBD and/or involving the CBD and/or any anastomosis to the CBD. Being several centimeters rostral, the common hepatic duct is involved only rarely with direct spread from a pancreatic neoplasm or only very late in the disease.

Choledochoduodenostomy/hepaticoduodenostomy: When possible, some groups have advocated a choledocho- or hepaticoduodenostomy. The extrahepatic biliary tree
and the duodenum are quite close anatomically and, when the extrahepatic biliary tree is dilated, the anastomosis becomes quite easy. After a Kocher maneuver to gain duodenal mobility, this anastomosis can be done as a side-to-side anastomosis or less commonly by transecting the common hepatic duct at its junction with the CBD, oversewing the distal CBD, and carrying out an end-to-side hepaticoduodenostomy (Fig. 5.2); the latter requires circumferential dissection around the common hepatic duct and oversewing of the CBD, which seems unnecessary if a more simple, straightforward, side-to-side hepaticoduodenostomy is feasible technically. This anastomosis is straightforward, facile, and safe. Usually a cholecystectomy is also performed.

Arguments against this approach cite the potential for anastomotic obstruction in conjunction with duodenal obstruction from local extension of the pancreatic neoplasm. An increasing number of proponents with good, long-term data maintain that (1) this approach requires only one, easy, safe anastomosis, (2) late obstruction of the anastomosis is rare, and (3) when the more distal duodenum becomes obstructed, biliary drainage can still occur retrograde through the stomach and a gastrojejunostomy. The more limited reported experience appears quite good, and the arguments against hepaticoduodenostomy are largely theoretic and not actually supported by objective data.

Hepaticojejunostomy: Most HPB surgeons prefer enteric biliary drainage via the common hepatic duct based on the theoretic argument that the chance of future malignant obstruction by local extension of the pancreatobiliary neoplasm is much less likely—see “Selection of Bilioenteric Bypass”. Hepaticojejunostomy can be carried out either to a loop of jejunum or more commonly to a Roux limb (Fig. 5.3). The latter is usually preferred for technical ease. Bringing the intestinal conduit (loop of jejunum or Roux limb) up to the hepatic duct is much easier technically and a much shorter distance via a retrocolic approach through the right transverse mesocolon rather than up and over the omentum and right transverse colon; bringing up a jejunal loop through the mesocolon is more bulky, and worry always remains about obstruction of the loop at the mesocolon; therefore, a Braun-type enteroenterostomy (a second anastomosis) is also usually created. A Roux limb is less bulky, more mobile, and easier to work with than the loop. The argument that a Roux drainage requires the potential morbidity of a second anastomosis (end-to-side jejunojejunostomy) is offset by the ease of construction of the hepaticojejunostomy and the usual construction of a Braun enteroenterostomy with the loop construction anyway. Any worry about reflux of enteric chyme into the
biliary tree should not be of concern, because this procedure is palliative, and prolonged survival greater than 2 years is unexpected. As with hepaticoduodenostomy, if the gallbladder is still in situ, most surgeons will add a cholecystectomy for several reasons: (1) the dilated gallbladder can be “in the way,” and (2) by bypassing the gallbladder, there can be stasis in the gallbladder lumen, which can lead to cholecystitis requiring intervention, a condition that no one would want to develop in a patient with unresectable pancreatic cancer. Cholecystectomy is easy and adds minimal morbidity.

The hepaticojejunostomy can be constructed either in a side-to-side or end-to-side fashion (Fig. 5.3). These authors see no benefit of transecting the common hepatic duct to carry out an end-to-side hepaticojejunostomy, which requires its circumferential mobilization and oversewing of the CBD distally, although some surgeons prefer this technique. A side-to-side anastomosis is equally easy to construct. The anastomosis is usually a one-layer technique; if the biliary tree is adequately dilated, a running technique is faster. Again, because this is a palliative procedure, either absorbable or permanent 3- or 4-0 suture material is appropriate. A second serosal layer to the periductal tissues can be added if so desired. Use of any form of decompressive biliary “T-tube” or retrograde bilio-external drainage tube is unnecessary unless the surgeon is very worried about the integrity of the anastomosis for some reason. Anastomotic leak should be rare (<2%). Presence of an endobiliary stent placed preoperatively can either be removed retrograde through the hepatic ductotomy at the time of hepaticojejunostomy or, if there is any concern about the operative anastomosis, the stent can be left to provide another pathway of enteric drainage/decompression but will necessitate endoscopic removal postoperatively. Overall, this procedure has proven to be a very safe, durable anastomosis.

The need for any type of intrahepatic enteric bypass would be extremely unusual as a technique of operative palliation. Situations that might in very rare cases justify such
an approach would involve extensive nodal disease extending up the hepatoduodenal ligament. Possible operative procedures in this situation might involve a left hepatic duct-jejunostomy after taking down the hilar plate with a Hepp-Couinaud type of anastomosis, a segment 3 approach (ligamentum teres approach), or the classic Longmire, left-sided cholangiojejunostomy after excising the lateral most aspect of segment 2 of the liver. These procedures are largely of historic interest, because currently, such extensive involvement of the hepatoduodenal ligament would not have escaped notice on imaging preoperatively (i.e., intrahepatic biliary dilation with dilated extrahepatic ducts) which indicates unresectable disease; should the patient require operative palliation for some other complication (e.g., duodenal obstruction), biliary decompression would be attained by a percutaneous biliary drain without the increased morbidity of one of these intrahepatic bilioenteric bypasses.