Billroth II Gastrectomy

 

Frank James Branicki and Leslie Karl Nathanson

 

Duodenal Ulcer: Etiology and Management Strategies

Timing of Surgical Intervention

Billroth II Gastrectomy

Difficult Duodenal Stump Closure

Stapled Anastomosis

Laparoscopic Billroth II Gastrectomy

Specific Postoperative Complications

Duodenal Stump Leakage

Summary

Suggested Reading

EDITOR'S COMMENT

 

Table of Contents

Chapter 74 Billroth II Gastrectomy

Duodenal Ulcer: Etiology and Management Strategies

 

A 1989 National Institutes of Health (NIH) consensus statement documented that bleeding was estimated to occur in 100,000 of four million patients with peptic ulceration in the United States annually. Globally, peptic ulceration reaches almost epidemic proportions in the Chinese populations of Hong Kong, the People's Republic of China, and Taiwan. Genetic influences are believed to play a part and it is not uncommon to encounter the disorder in children. Nevertheless, there is now overwhelming evidence for a role for Helicobacter pylori in the etiology of peptic ulcer and the intestinal type of gastric adenocarcinoma.

Therapy of duodenal ulceration has evolved in the past decade with a greater understanding of therapeutic options and more recently with the advent of minimal access surgery. At least 90 percent of duodenal ulcers have an association with H. pylori, which may be diagnosed by histologic examination of endoscopic mucosal biopsy or by the rapid urease test. Occasionally, culture of H. pylori is employed particularly in patients who have antibiotic allergies or failed previous therapies, or in countries where H. pylori has a high level of background antibiotic resistance. There is no justification for treatment of patients for periods longer than 14 days. Breath tests to detect urease can indicate cure of H. pylori 4 weeks after antibiotic therapy, at a time when serologic antibody tests will still give a positive result. Cure has been defined as failure to demonstrate H. pylori by a sensitive technique 4 weeks after cessation of all antimicrobial therapy and proton pump inhibitors, the latter having been shown to interfere with diagnosis. The 1994 NIH consensus conference guidelines for duodenal ulcer include assessment of H. pylori status and eradication of the organism. Recurrence will develop in 10 percent of patients with duodenal ulceration, usually as a result of persistent basal acid hypersecretion or an underlying mucosal defect. Basal acid secretion is increased sixfold in patients with duodenal ulcer but returns toward normal within 6 months after H. pylori eradication.

The efficacy of various drug regimens for eradication of H. pylori has led physicians to adopt a conservative approach to referral for consideration of surgical intervention for peptic ulcer. This policy has been supported by the finding that reinfectivity rates for H. pylori are very low, for example, less than 1 percent in Western societies. It has now become inappropriate to recommend elective surgical intervention for complicated peptic ulcer unless there is evidence that ulceration persists in the absence of Helicobacter infection. Indeed, it has been documented in two reports that subsequent eradication of the organism in patients undergoing nonsurgical management of bleeding peptic ulcer was associated not only with ulcer healing but with no subsequent hemorrhage during follow-up studies.

Thus, it is now considered that the indications for elective surgical intervention in complicated ulcer disease require better definition. Surgery in patients with recurrent duodenal ulceration is warranted for 1) poor patient compliance with medical therapy, 2) patient preference when ulcer relapse occurs frequently despite intermittent medication, and 3) definitive alternative treatment to long-term maintenance therapy in younger patients. The introduction of a variety of laparoscopic techniques, including posterior truncal vagotomy and anterior seromyotomy, highly selective vagotomy (HSV), vagotomy and antrectomy, and partial gastrectomy, now provide opportunities for treatment of duodenal ulceration without the short-term morbidity related to open upper abdominal incisions. It is essential, however, that the procedure of choice at open surgery for the majority of patients with duodenal ulcer (HSV) is also adopted as optimal laparoscopic treatment when a gastric resection is deemed unnecessary. Truncal vagotomy and antrectomy retain a place in the management of some patients with pyloric/duodenal stenosis, prepyloric ulceration, and recurrence despite previous HSV or truncal vagotomy and a drainage procedure. Billroth II gastrectomy is an alternative in some circumstances; even in the recent past, its usage has been widespread in China but it is occasionally associated with side effects of diarrhea and dumping. The role of truncal vagotomy and antrectomy or Billroth II gastrectomy is largely confined to the management of patients with primary ulceration with bleeding or perforation arising from a large ulcer, for example, 2 cm or greater in size, or recurrent ulceration following previous truncal vagotomy. Although truncal vagotomy and antrectomy are associated with a recurrence rate of less than 3 percent, on long-term follow-up studies they have the distinct disadvantage of delayed gastric emptying, which occurs in about 6 percent of patients, regardless of whether continuity has been restored with gastroduodenal or gastrojejunal anastomosis. The combination of vagotomy, which interferes with gastric motor function, and excision of the antral pacemaker mechanism occasionally gives rise to gastric remnant stasis. This may take between 2 and 6 weeks to resolve but always does so spontaneously in our experience. Obviously, nutritional support needs to be provided during this period. Enteral fine-bore tube feeding is sometimes problematic with proximal migration of the distal tube to the stomach with gastric distention; parenteral nutrition is preferred. As truncal vagotomy and antrectomy may require protracted hospitalization in the postoperative period, there is a reluctance to undertake this procedure in the elderly, who often have concomitant illness and run the risk of atelectasis or deep vein thrombosis. Billroth II gastrectomy is therefore favored for older patients, whereas truncal vagotomy and antrectomy offer the benefits of a larger gastric remnant in the younger age group.

The first successful gastric resection was performed by Theodor Billroth on January 29, 1881, in a 43-year-old woman with gastric cancer, the gastric remnant being anastomosed to the duodenum (Billroth I). On January 15, 1885, Billroth performed the first gastrectomy (Billroth II) with closure of both the duodenal stump and the gastric remnant, continuity being restored by fashioning an anterior gastrojejunostomy. Following gastric resection for duodenal ulcer, although gastroduodenal anastomosis has perhaps the greatest physiologic appeal for restoration of continuity, it is well established that lower rates of recurrent ulceration are achieved by a two-thirds gastric resection, duodenal stump closure, and gastrojejunal anastomosis. In addition, gastroduodenostomy is often technically difficult in the presence of a scarred or fibrotic duodenum, necessitating extensive mobilization of both duodenum and remnant to ensure a tension-free anastomosis. Doubts may also be harbored as to whether the reconstructed gastroduodenal outlet will have sufficient caliber for adequate gastric emptying to take place. The Billroth II procedure can be performed with the confidence that a safe adequate stoma is always possible. Gastrojejunal reconstruction is also favored in the management of patients with gastric cancer, as it obviates potential obstructive problems should recurrent disease arise in the vicinity of the pancreatic head. Early gastric cancer may require only a 2-cm resection margin and, depending on the site of the lesion, gastroduodenal anastomosis may be appropriate. In contrast, all advanced cancers are best dealt with by gastrojejunal reconstruction following gastric resection, which does not compromise the extent of lymphadenectomy undertaken.

An awareness of technical variations is important to the gastric surgeon, who may select which is most appropriate according to anatomic and pathologic findings. The variation promoted by Franz Hofmeister (1867–1929) with closure of much of the lesser curvature side of the cut gastric remnant proved popular, as it is said to ensure that the stoma is placed at the most dependent part of the stomach and avoids acute angulation of the jejunum at the lesser curvature, which is claimed to be a disadvantage of the Polya technique (Fig. 74-1). Nevertheless, our own practice of the Polya operation described by Jeno (Eugene) Polya (1876–1944) in which the whole of the cut end of the gastric remnant is sutured to the jejunum has provided satisfactory functional results in a large majority (Fig. 74-2). There appears to be little or no advantage for the crosscut technique with a transverse jejunal incision, as described by Hoerr and Turnbull (Fig. 74-3). Indeed, theoretical considerations concerning uncut Roux en Y loops suggest that a crosscut jejunum may have a greater possibility of impaired motility.

More than two dozen variations of the Billroth II procedure have been described since its introduction in 1885. These encompass variations that involve siting of the loop gastroenterostomy with regard to the gastric remnant and transverse colon and Braun enteroenterostomy or Roux en Y modifications. Anastomosis of the afferent limb of the loop to the greater (Moynihan; Fig. 74-4) or lesser (Balfour; Fig. 74-5) curvatures is an option, and much controversy has centered on whether a loop should be sited in an antecolic or retrocolic position. Tanner held the view that antecolic placement was advantageous, as it rendered revision gastric surgery much simpler should the need ever arise and that carcinoma of the colon occurred frequently enough in England to raise concerns regarding the hazards of radical colonic resection in patients in whom a retrocolic gastrojejunostomy had been performed. These technical considerations were considered sufficient argument for advocating antecolic anastomosis, as this does not appear to compromise function. It has, however, been voiced that antecolic anastomosis is associated with a greater likelihood of colonic or small-bowel obstruction. This risk is believed to be reduced by bringing the colon with gentle traction to the right side of the anastomosis at the conclusion of the procedure. Our preference for Billroth II gastrectomy involves mobilization of the distal stomach, closure of the duodenal stump, and gastric resection followed by antecolic (Balfour) gastrojejunostomy.

 

Table of Contents

Chapter 74 Billroth II Gastrectomy

Timing of Surgical Intervention

 

The timing of surgery in the management of complicated duodenal ulceration is crucial to outcome measurements. Emergency surgical intervention for gastrointestinal bleeding is associated with a 10-fold increase in mortality compared with elective surgery. If at all possible the patient presenting with bleeding duodenal ulcer should undergo urgent therapeutic endoscopy for immediate management, followed in selected patients by elective surgical intervention when operation is deemed essential. Experience with such a policy in the management of over a thousand patients with bleeding peptic ulcer has recorded no operative mortality for elective or early operative intervention during the same admission. In the selection of management strategies for patients who present with acute hemorrhage, it is necessary to identify the "high-risk" patient with coexisting medical illness or hemodynamic instability, or both, and the "high risk" lesion. Therapeutic endoscopy for bleeding peptic ulcer is believed to be of value for active spurting/oozing hemorrhage, or when a nonbleeding visible vessel with or without sentinel clot has been visualized at endoscopy on initial presentation. Endoscopic stigmata of recent hemorrhage are more likely to be seen in large ulcers, greater than 1 cm, in diameter, large ulcers being more frequently encountered in patients over 60 years of age. Ulcer size of greater than 1 cm is associated with an increased risk of rebleeding and mortality, with rebleeding carrying a 10-fold increase in mortality. Postoperative gastrointestinal bleeding from bleeding peptic ulcer has a mortality of 40 percent if reoperation is required. Injection sclerotherapy is considered the procedure of choice for control of postoperative hemorrhage that continues or recurs. Obvious reservations exist regarding distention of the stomach in the immediate postoperative period, but using as little insufflation as possible we have successfully managed suture line bleeding in this way. The portable nature and minimal cost of equipment required have been responsible for the increasing popularity of injection sclerotherapy (with alcohol, epinephrine, polidocanol, or saline) in preference to electrocoagulation, heater probe application, or laser therapy. Sclerotherapy has been shown to be equally efficacious and may be repeated if so required. Gastric wall necrosis and jaundice due to bile duct fibrosis have been reported with the use of epinephrine and polidocanol, respectively. Thus, overzealous injection treatment with excessive injections of aliquots is to be avoided.

Bleeding

When surgical intervention for gastrointestinal hemorrhage is required, adequate control of the bleeding site in an ulcer that is to remain in situ is essential. Figure-of-eight plication of a bleeding vessel may suffice but, occasionally, control may only be obtained with the aid of prior ligation of the gastroduodenal artery adjacent to its origin from the hepatic artery. Nonabsorbable sutures, for example, 2–0 linen, are preferred for transfixion of the bleeding site and suture ligation of feeding vessels required to control oozing hemorrhage from the ulcer base. If hemorrhage is such that there is hemodynamic instability at induction of anesthesia, then pyloroduodenotomy with direct suture of the bleeding point is essential before one embarks on gastric mobilization. Difficulty is sometimes encountered in securing hemorrhage in the fibrotic base of an ulcer. It is important to use a wide-nosed needle holder that has a firm grip on the needle without any free rotational movement of the needle in its holder. When bleeding from extensive ulceration proves difficult to control by direct plication following pyloroduodenotomy, it may be necessary to identify the origin of the gastroduodenal artery and ligate the vessel medially and lateral to the duodenum as it courses posteriorly on the head of the pancreas. Extensive ulceration of the medial duodenal wall may encroach on the ampulla of Vater or the accessory pancreatic duct of Santorini. Encroachment on the ampulla is usually recognized endoscopically or at open operation without too much difficulty, whereas the presence of an accessory pancreatic duct opening may easily be overlooked. Particularly when hemorrhage is brisk, it is necessary to be wary of the possible inadvertent suture ligation of the ampulla of Vater. If ulceration extends to the second part of the duodenum, the location of the ampulla must be identified by palpation or visually, once hemostasis has been achieved. The risk of rebleeding in the postoperative period is believed to be greater in patients with large peptic ulcers greater than 2 cm in size. Accordingly, we hold the view that all large ulcers that have given rise to life-threatening hemorrhage be resected if surgery is required. Penetrating ulcers of the head or body of the pancreas are generally left in situ with exclusion from the gastrointestinal tract when continuity is restored. Plication of the bleeding point in small posteromedial duodenal ulcers may be followed by exclusion of the ulcer base from access to the duodenal lumen by suture closure of the mucosa at the ulcer margins.

Perforation

Large duodenal ulcers, greater than 5 cm in size, are most commonly found in a posteromedial position in the first part of the duodenum. Anterior ulcers generally perforate into the peritoneal cavity before attaining such a size, whereas posteromedial ulceration into the pancreas may not give rise to complication unless bleeding supervenes. In the low-risk patient with a free intraperitoneal perforation from duodenal ulceration, definitive anti-ulcer surgery may have a place. Definitive intervention, some form of vagotomy or excisional surgery, is not recommended 1) if the patient has a history of coexistent medical illness, for example, diabetes; 2) the duration of intraperitoneal contamination exceeds 24 hours; or 3) hemodynamic instability is evident on first presentation, for example, systolic blood pressure less than 100 mm Hg. In such circumstances simple omental patch repair is advisable with the exception of the large ulcer defect, greater than 2 cm in diameter, for which excisional surgery is the preferred option. Omental patch repair of ulcers greater than 2 cm in size is believed to be associated with an increased risk of peripatch leakage of duodenal contents postoperatively. In addition, closure in such circumstances may give rise to obstruction in the duodenum, which may already have some degree of stenosis from chronic scarring. There is, therefore, support for excisional ulcer surgery, duodenal stump closure followed by antrectomy with truncal vagotomy, or Billroth II gastric resection and anastomosis.

 

Table of Contents

Chapter 74 Billroth II Gastrectomy

Billroth II Gastrectomy

 

Prior usage of H2 receptor antagonists or proton pump inhibitors may have resulted in hypochlorhydria with endogenous bacterial overgrowth and a greater likelihood of infective postoperative complications. Gastroduodenal resections are classified as "clean contaminated" in nature and prophylactic antibiotic therapy is indicated; a single cephalosporin injection an hour before operation is usually sufficient. Prior upper endoscopy will have provided in most instances, except perforation, information crucial to decision making as regards the nature of the procedure to be performed. Patients with pyloric/duodenal stenosis need to be fasted and undergo repeated gastric lavage, with nutrition being maintained intravenously for up to a week preoperatively. This helps to ensure adequate motility in the remnant stomach should resection be performed. Truncal vagotomy and pyloroplasty or gastroenterostomy may, however, be ineffective for the treatment of vomiting when there has been chronic gross distention of the stomach with resultant atony. Alternatively, the Billroth II procedure may be considered in patients over 60 years of age with pyloric/duodenal stenosis.

Our preference at open surgery is an upper midline incision that extends to the left side of the xiphisternum to provide adequate exposure of the proximal lesser curvature of the stomach. Access is greatly enhanced with the use of a self-retaining retractor (Hepco, Kansas City, KS), which comprises two vertical poles that are attached by fixators to the operating tube and a central horizontal bar that is then used to bridge the poles. A variety of self-retaining Hepco retractors for xiphisternum, costal margin, and liver are available to facilitate excellent abdominal exposure without the need for manual retraction of wound or liver. Any adhesions to the lower pole of the spleen are divided before gastric mobilization.

Duodenal Transection and Closure

At open operation, the duodenum, stomach, gallbladder, and bile duct are inspected and palpated. Although large duodenal ulcers are generally palpable, small lesions may not be so, especially when situated posteromedially. If a large, active duodenal ulcer is present, an inflammatory mass is occasionally encountered; in the absence of bleeding, truncal vagotomy and gastroenterostomy represent a more prudent choice for therapy than gastric resection. The mass may be adherent to the extrahepatic biliary tree or the inferior surface of the liver; when overt bleeding has been the presenting feature, the ulcer will require direct suture, excision, or exclusion. Rarely, ulceration may abut the extrapancreatic common bile duct, requiring meticulous dissection of the duodenum to facilitate separation of the edges of the ulcer from its bed, which is retained in situ contiguous with the bile duct. If previous acute gastrointestinal hemorrhage has occurred, it is important not to overlook the presence of any second ulcer or lesion in the duodenum or stomach.

The right and left gastroepiploic vessels are identified, and small arterial and venous branches extending from the pylorus to the proximal stomach are then ligated and divided between hemostats. This maneuver, which opens wide the lesser sac, permits inspection of the posterior wall of the duodenal bulb with preservation of the main trunks of the right gastroepiploic vessels as the bulb is separated from the head of the pancreas. The duodenum can be mobilized as described by Kocher by division of the peritoneum lateral to its second and third parts. Mobilization of the first part of the duodenum proceeds with division of the branches of the right gastric artery, which arises from the common hepatic. Without gross scarring stump closure is accomplished with no difficulty when 2 cm of proximal duodenum can be freed on duodenal mobilization. This can then be transected with the aid of two Lang Stevenson clamps (Downs, London, England) placed transversally just distal to the pylorus (Fig. 74-6A). Gauze swabs soaked in chlorhexidine hydrochloride (Hibitane blue) are placed behind the duodenum and scalpel duodenal transection is performed between the clamps. The cut ends are swabbed with Hibitane-soaked gauze and a tie is then placed in figure-of-eight fashion to secure an additional small Hibitane-soaked swab to the clamp, which is left in place on the distal portion of the gastric specimen to be resected. This prevents intraperitoneal contamination from the proximal end of the divided duodenum. A 4–0 Maxon suture (Davis and Geck, Manati, PR) is then placed in a continuous fashion beneath the clamp on the duodenal stump (Fig. 74-6B). This suture is run back and forth to close the stump, the clamp is removed, and excess duodenum that has been clamped is trimmed. This first row of sutures is not hemostatic and to ensure that no bleeding will occur from the cut mucosa a second row of sutures is placed in the same plane using the same 4–0 Maxon suture in an over-and-over fashion (Fig. 74-6C). These suture layers are then inverted with the aid of a continuous 4–0 Maxon seromuscular suture placed so as to invert both corners of the cut edge of duodenum. This layered closure, although it takes a few minutes longer then single-layer or staple closure, has been found to be efficacious, with no leakage observed when it is performed electively for benign disease. Unfortunately, closure is sometimes more problematic without the option of such straightforward suture techniques in healthy tissues. If coincidental gallstones are present in the gallbladder, it is recommended that cholecystotomy and stone removal be performed in preference to cholecystectomy, which is regarded as being more likely to be associated with postoperative digestive sequelae if gastric resection is performed.

 

Table of Contents

Chapter 74 Billroth II Gastrectomy

Difficult Duodenal Stump Closure

 

During duodenal mobilization the extrapancreatic common bile duct must be identified and preserved. Inflammation may shorten the duodenum so that the ampulla of Vater is less than 5 cm from the pylorus. If the duodenum is grossly scarred or edematous, it is necessary to assess the feasibility of stump closure before embarking on gastric resection. Occasionally, a penetrating ulcer is encountered that requires scissors dissection to free the edges of the ulcer from the posteromedial duodenal wall, giving rise to a defect that may be several centimeters in size. The ulcer that remains in situ must be plicated adequately to control hemorrhage if it is regarded as the source of previous bleeding. The defect may be very large when ulceration is almost circumferential and the duodenum is divided at the proximal end of the mucosal bridge to facilitate stump closure. When ulceration is so extensive, it is imperative that the ampulla of Vater or any obvious accessory pancreatic duct opening be located by inspection or palpation. Failure to do so can give rise to suture closure of duct(s) with impeded biliary and pancreatic drainage and/or consequent pancreatitis or formation of a pancreatic fistula.

Rarely, if serious doubt exists as to the location of the ampulla it is justifiable to perform a choledochotomy and insert a pediatric infant feeding tube distally into the common bile duct to emerge from the ampulla. This maneuver, described by Longmire, has also been found to be useful in patients with duodenal involvement with antral carcinoma. A small-caliber T tube will subsequently need to be placed for safe duct closure. When a large ulcer has been left in situ, the cut edge of duodenum may be irregular, making it necessary to suture the anterolateral free cut edge to the inferior margin of the ulcer in situ with continuous or interrupted 4–0 Maxon. This excludes the ulcer from the duodenal lumen and a second row of interrupted 4–0 Maxon sutures is placed between the seromuscular layer of the inrolled duodenum and the ulcer base or superior edge depending on its size. When pyloric/duodenal stenosis results in only a pinhole orifice, simple separation of the stomach and duodenum at this site will lessen the difficulties encountered by fibrosis, with the orifice occasionally being more easily identified internally by a small distal gastrotomy.

Closure of the difficult stump may be accomplished by methods described by Nissen, Graham, or Bseth. Nissen and Graham maneuvers involve covering an ulcer left in situ utilizing the anterosuperior duodenal wall, whereas Bseth advocates posteromedial duodenal wall suture for coverage of the ulcer bed. When compared with the Nissen procedure Graham's technique has the disadvantage of requiring a more extensive dissection of the posterior duodenal wall. Nissen effected closure with sutures that incorporated the inferior margin of the ulcer, whereas mobilization of the duodenum distal to a large ulcer may damage or occlude the accessory pancreatic duct. Graham advocated ulcer exclusion with the duodenum mobilized for 1 to 2 cm distal to the ulcer and subsequent closure with seromuscular suture of the duodenum to the pancreas to cover the ulcer bed. It is not necessary and can be hazardous to attempt to dissect the posterior duodenal wall distal to the inferior border of an ulcer. The secret of successful stump closure is to limit the distal extent of the duodenal dissection. Strauss and others have been proponents of an intramural duodenal dissection with development of a cuff of posterior duodenal wall. A jejunal patch as utilized for repair of traumatic duodenal injuries can also be employed when stump closure appears difficult. Roux en Y duodenojejunostomy can also be considered for the reduction of intraduodenal pressure. An alternative is the Bancroft closure, which suffers the disadvantage of possible retained antral mucosa in the stump and bleeding during separation of the submucosa from the muscular layer. Following subtotal gastrectomy and end-to-side gastrojejunostomy (Moynihan), Appel has performed duodenojejunostomy with anastomosis of the duodenal stump to the side of efferent loop followed by Braun jejunojejunostomy. It is claimed that this utilizes less duodenal tissue than conventional closure and gives an anastomosis that is tension free, as the duodenum is not mobilized.

Resection of no more than 2 cm of duodenum also obviates excessive devascularization. In circumstances in which one has some reservations regarding the integrity of a proposed or completed closure, other options merit consideration. If the appearance of the transected duodenum and ulceration suggests that the risk of subsequent suture line leakage is probable, then catheter tube duodenostomy is advisable. This may entail the use of a Foley or Malecot catheter in the transected end of the duodenum, secured by means of one or two pursestring sutures and brought out percutaneously in a short straight track. A Malecot catheter is preferred as insufflation of a Foley catheter balloon for retention in situ may cause ampullary or proximal duodenal obstruction if sited distal to the ampulla. End duodenostomy is, however, often accompanied by peritubal leakage of duodenal contents, no matter how secure the nonabsorbable suture pursestrings appear. Some surgeons hold the view that lateral tube duodenostomy through noninflamed tissue in the second part of the duodenum will allow for earlier healing than end duodenostomy in a scarred stump. For this reason many surgeons opt for a side duodenostomy with latex T-tube drainage via a stab wound in the second part of the duodenum. Most T tubes available today are siliconized and need to remain in situ for 3 to 6 weeks, 10 days being insufficient to enable formation of a satisfactory track. If stump closure can be accomplished, with leakage considered unlikely, then lateral T-tube duodenostomy may suffice (Fig. 74-7). When closure is regarded as unsafe and highly likely to lead to leakage then end tube/catheter is recommended. Personal preference is for end catheter duodenostomy in the most difficult cases.

It is also imperative in the prevention of stump leakage to avoid afferent limb obstruction at the site of subsequent gastrojejunal anastomosis. If duodenal stump closure is difficult, it is necessary to 1) railroad the nasogastric tube well into the afferent limb with or without suturing to the duodenal mucosa with absorbable sutures; 2) meticulously cover the stump with omentum sutured to duodenum or surrounding structures, or both; and 3) insert a drain into the right upper quadrant space for 5 days. Afferent limb intubation in this way stabilizes the loop and is said to prevent kinking of the limb as well as obviating raised intraduodenal pressure, which is detrimental to stump integrity. If stump leakage occurs it is usually evident on the fourth or fifth postoperative day but may occur between the second and seventh days. Any drain placed at operation to deal with stump leakage should this occur must therefore remain in situ for at least 5 days, even if early postoperative drainage is minimal. It is desirable to remove the drain after 5 days, as occasionally small bowel or omentum may become adherent to the drain if it is left for longer, and this may not only make removal difficult even with sedation/analgesia but can result in damage, giving rise to small-bowel fistula or hemorrhage.

Gastric Transection and Reconstruction

Following duodenal stump closure the lesser curvature of the stomach is mobilized with small anterior and posterior division of branches of the left gastric artery and vein. The dissection is carried superiorly on the lesser curve as the gastrohepatic omentum is divided with electrocoagulation. Branches of the left gastric artery and vein supplying the stomach are divided between hemostats, the dissection proceeding proximally to within 3 cm of the distal esophagus. The stomach is bared on greater and lesser curvatures to such an extent that a two-thirds gastric resection is possible. At least 2.5 cm of bared stomach can be retained on the greater curvature and utilized for anastomosis in the knowledge that intragastric blood supply will be sufficient to maintain viability if short gastric arteries remain intact. Often a natural break in the gastroepiploic arcade will dictate the exact site of gastric transection.

The duodenojejunal flexure is identified and a loop of proximal jejunum is lifted free to lie in juxtaposition to the gastric remnant, being held by two small Babcock clamps at the site proposed for jejunal anastomosis. It has been emphasized that too long an afferent limb is to be avoided, as this can produce bacterial overgrowth in a blind loop syndrome. Nevertheless, antecolic anastomosis is usually desirable and it is possible to do this without tension on the loop provided a suitable site is chosen to anastomose. It is obviously preferable to make the loop 2 to 5 cm longer than to risk excessive tension. The limb with the screw attachment of a curved Lane twin clamp is applied to the jejunum stretched upward by appropriate placement of two Babcock clamps, the jejunum being grasped across its lumen to within 3 cm of the tip of the instrument. The second limb of the twin clamp is applied to the stomach, which is held caudally by a Babcock clamp attached to part of the bared greater curve that is to be resected. The clamp is applied and closed, taking care not to ensnare the nasogastric tube, which has been passed previously and now withdrawn high in the proximal stomach before application of the clamp, which is again positioned so that the lesser curvature of the stomach extends to within 3 cm of its tip. The stomach remnant is then lifted anteriorly, permitting the two limbs of the clamps to be apposed behind it utilizing the eye of the inferior clamp, the screw arrangement being tightened to approximate the stomach and jejunum.

A continuous 4–0 Maxon posterior layer seromuscular suture is placed from lesser to greater curvature of the stomach and the suture needle is attached, laid taut to the left side. The jejunum is divided with electrocoagulation for a length that is judged to match the proposed gastric division. This jejunal incision is slightly shorter than the gastric division and can be lengthened if need be during suturing. The stomach to be excised is then clamped inferiorly with a soft bowel clamp 5 cm from the twin clamp attachment. Diathermy is employed to open the gastric wall adjacent to the twin clamp, with Hibitane swabs used to clean the exposed interior of the stomach and jejunum. A continuous 4–0 Maxon suture that includes all coats anastomoses the jejunum and stomach from lesser to greater curvatures. When the greater gastric curvature is reached, the portion of stomach to be excised is lifted anteriorly and division is completed with electrocoagulation in a semicircular fashion to provide a flap; this will enable inner wall gastrojejunal anastomosis without the need to release the twin clamp. The inner layer of the anastomosis is completed and the twin clamp removed from stomach and jejunum. Should there be a small excess of gastric lumen for approximation, this can be directly sutured to itself on the lesser curvature aspect, with subsequent inversion by seromuscular suture application. The anterior outer layer is fashioned by a continuous seromuscular suture employing the previous 4–0 Maxon, which has been laid to the left side. Reinforcement by additional seromuscular sutures between the stomach and jejunum at the vulnerable lateral angles of the stoma has been advocated. The nasogastric tube is manipulated well into the afferent limb. The transverse colon is pulled with gentle traction to the right of the anastomosis, which is inspected, especially posteriorly, to ensure that there is no bleeding from transected extragastric vessels or wall. Patency can be tested by "finger and thumb" opposition between gastric remnant and jejunal loop. The duodenal stump is inspected for leakage or bleeding and then covered with omentum sutured in place to provide a prophylactic patch.

Nonabsorbable mucosal sutures for gastric anastomoses are to be avoided. While absorbable catgut, Vicryl, polyglycolic acid (Dexon) sutures have proved satisfactory, 4–0 Maxon has the ability to slide more easily through tissues, a greater knot pull strength, and a longer absorption time of 180 days. Continuous sutures are favored, as these secure hemostasis. A majority of surgeons continue to use a two-layer continuous anastomosis with a posterior inner layer that incorporates the full thickness of gastric and jejunal walls and an outer seromuscular layer that inverts the inner layer. Recent trends suggest that a one-layer inverting continuous suture that incorporates serosa, muscular, and submucosal coats is gaining popularity despite the risks of gastric suture line hemorrhage.

 

Table of Contents

Chapter 74 Billroth II Gastrectomy

Stapled Anastomosis

 

An alternative method of duodenal closure involves placement of a TA30 or TA50 horizontal stapling device positioned across the duodenum just distal to the pylorus (Fig. 74-8). A Lang Stevenson clamp is placed more proximally to avoid leakage on division of the duodenum with a scalpel after the TA stapler has been fired; stay sutures in the duodenum are usually unnecessary. The stomach can be closed with the use of a TA90 transverse stapling device employing 4.5-mm rather than 3.8-mm staples, the former being considered to be more secure and hemostatic. After firing, the stomach can be divided with a scalpel, a soft bowel clamp distally preventing soiling with gastric contents. Following removal of the transverse staplers, on firing the cut ends are inspected for bleeding and integrity; placement of interrupted 4–0 Maxon sutures may be required (Fig. 74-9). The use of electrocoagulation is acceptable for staple lines but is not advisable in the vicinity of a continuous suture line, which may be inadvertently damaged. Many surgeons do not invert transverse staple lines with sutures but continuous suture coverage avoids free leakage should the staple line fail.

The GIA side-to-side stapling device can be used to fashion a gastrojejunostomy after approximation of the posterior stomach and jejunum with stay sutures (Fig. 74-10). A small jejunal enterotomy permits placement of one limb of the GIA; the second is introduced via a small, posterior gastric incision appropriately sited to enable a widely patent anastomosis to be fashioned on firing and located 4 cm proximal to the cut end of the stomach so as to avoid ischemic necrosis of the stomach between the GIA and TA staple lines (Fig. 74-11). The GIA device must provide an anastomosis at least 5 cm in diameter to obviate delayed gastric emptying. The limbs of the GIA device are approximated, and the instrument is fired and withdrawn from the lumen. Active bleeding from the staple line may be difficult to detect through the small stab incisions, which is a distinct disadvantage in gastric surgery, during which continuous spurting hemorrhage may go undetected only to be manifest with persistent heavily blood-stained nasogastric aspirates postoperatively. The conjoined defect at the site of stab incisions is closed with a continuous 4–0 Maxon suture. The GIA can be introduced from the left or the right side, with a left-sided entry often being performed if the remnant is small; right-sided placement is technically more difficult.

A variation in technique incorporates excision of a portion of the TA-stapled greater curvature to provide a gastric opening, which is hand sewn to the jejunal limb in Hofmeister fashion (Fig. 74-12). Stapling devices are widely employed for duodenal stump closure but are not so favored for gastric resection by most surgeons, with usage being confined to the creation of gastric tubes for esophageal replacement.

 

Table of Contents

Chapter 74 Billroth II Gastrectomy

Laparoscopic Billroth II Gastrectomy

 

Laparoscopic exposure of the upper abdomen has been facilitated by the advent of videolaparoscopy, simple retraction techniques of the left lobe of the liver, and patient positioning. Complete expeditious mobilization of the vascular pedicles required the development of the automatic clipping devices and more recently the linear cutter/stapler. This same device has greatly enhanced the surgeon's ability to divide bowel and fashion the anastomosis, essential technical steps performed during gastrectomy. These instruments have only been available since 1991. Along with mechanical devices, suturing techniques have evolved and with current high-quality needle holders standard curved needles can be used. A recently launched suturing device (Endostitch, Auto Suture, Norwalk, CT) has potential to dramatically ease the technical demands on the surgeon during suturing. A variety of procedures to fashion the gastroenterostomy are illustrated to allow surgeons with varying access to expensive technology to choose the most appropriate technique for reconstruction.

Patient Positioning and Port Sites

The patient is positioned supine with varying degrees of head uptilt depending on body habitus, to encourage downward displacement of the transverse colon and omentum. A nasogastric drainage tube decompresses the stomach and antibiotics are routinely administered. The position and sizes of the ports are illustrated in (Fig. 74-13). The epigastric port site, just to the right of the xiphisternum, is used to insert the liver retractor (Wilson Cook, Brisbane, Australia), which is then attached to an external holder ("Iron Intern" or "Omnitract") clamped to the operating table rails. Once positioned and anchored, this system atraumatically elevates the left lobe of the liver for the entire operation, freeing the assistant for other tasks.

Gastric Resection

Dissection begins along the greater curvature of the stomach toward the left, outside the gastroepiploic vascular arcades using sharp scissors dissection of the windows, supplemented with electrocautery. Isolated epiploic vessels are clipped using the 9-mm clip of the automatic clip applier and divided (Fig. 74-14). Attention is then turned to the level of division of the stomach and the posterior aspect is cleared of any filmy adhesions above the pancreas. A point along the greater curve is chosen for transection of the antrum and sequential applications of the linear cutter are applied until the lesser curvature is reached. The duodenum is next transected in a similar manner (Fig. 74-15). The lesser omentum is now divided using scissors and cautery for small vessels and clips for larger ones. Once this is completed the mobilized antrum is extracted via the 12-mm port site or, if bulky, it can be temporarily placed above the right lobe of liver to await extraction at the end of the procedure. In our view this is best achieved by placement of the specimen in a retrieval bag to minimize bacterial contamination of the abdominal wall during extraction.

Gastric Reconstruction: Stapled/Sutured Anastomoses

The transverse colon is grasped and elevated cephalad with attached greater omentum. The duodenojejunal flexure is located and the jejunum is followed to allow a loop of sufficient length to curve around transverse colon to reach for antecolic anastomosis without kinking. A 2–0 Prolene suture on a curved needle is introduced and passed through the stomach and jejunum, and both ends are removed via the right upper quadrant port site and held with a hemostat. The port at that site is then re-introduced alongside the stay suture and an enterotomy for the stapling device is created in the jejunum and stomach (Fig. 74-16). Care must be taken to gain access to the thick walled stomach; air insufflation via the nasogastric tube distends and thins out the stomach facilitating this step. For a stapled anastomosis the linear stapler/cutter is introduced and a 60-mm-long anastomosis is created. The residual defect is closed with a single layer of continuous 3–0 PDS, as illustrated in Fig. 74-16C and D. A jamming loop knot enables the commencement of this suture line (Fig. 74-17).

Data for determining an adequate size of stoma are illustrated in Table 74-1. These experimental gastrojejunostomies were created in a piglet model. Those of 40-mm length had a high incidence of stenosis and obstruction at 12 weeks compared with a stoma of 60 mm in length. Of relevance is the thickness of the gastric wall, which produces more scarring than other intestinal anastomoses and subsequent scar contracture during healing, leading to stomal stenosis.

Closure of the residual defect by stapling with the linear cutter is possible. This may, however, very easily cause narrowing of the efferent limb of the jejunum and in our view suturing is safer and more accurate. An aid to simplify technique is available with the Endostitch suturing device; this instrument has only recently become available for clinical practice and shows great promise. With all suture techniques great care must be taken to follow the suture to ensure that tension is maintained and that the edges of the anastomosis are approximated correctly.

An alternative to the stapled anastomosis is to create an entirely sutured one. This is achieved by placement of the stay suture as illustrated above. The single continuous posterior layer of sutures is then placed as illustrated in (Fig. 74-18). Next, the stomach and jejunum are opened and the anastomosis is completed with the anterior layer of sutures as indicated. Spillage of gastric contents is minimized by aspiration on the nasogastric tube before the stomach is opened. It is important to lock the suture every eight stitches to prevent loosening of the running suture line during peristalsis. The area is lavaged clean with warm Ringer's lactate solution after extraction of the resected antrum if this has not already been performed. The fascia of all ports greater than 5 mm are sutured to prevent port site herniation. A retrocolic anastomosis can be fashioned using the same technique as for antecolic placement. The creation of the window in the transverse mesocolon is sited to the left of the middle colic vessels; the anastomosis is created superiorly and on completion is brought inferiorly below the transverse mesocolon and held in position with a few interrupted 3–0 PDS sutures. A drain is placed to the area of the duodenal stump.

Patients are usually able to tolerate fluids by the third day and solids by the fourth postoperative day. Ambulation is encouraged as soon as possible to minimize the cardiorespiratory complications related to recumbency. Reported series of small numbers of patients demonstrate the feasibility of the technique with the discharge from the hospital between 5 and 12 days after the procedure and return to normal activities in 2 to 3 weeks. Assessment of long-term outcome requires evaluation at further follow-up reviews.

 

Table of Contents

Chapter 74 Billroth II Gastrectomy

Specific Postoperative Complications

 

Early postoperative complications include hemorrhage, obstruction, and leakage. Postoperative hemorrhage may be intraperitoneal and the result of splenic injury, which may have been recognized and managed by suture or application of hemostatic agents at operation. Hemorrhage can also occur from small vessels in the omentum or the gastric wall itself. Intraperitoneal hemorrhage is usually manifest within 6 hours after operation and, despite the temptations of conservative management, often requires relaparotomy for expeditious control. It is important to inspect the bed of the pancreas and celiac axis for any evidence of bleeding before proceeding with restoration of gastrointestinal continuity. In addition, the whole anastomosis must be inspected externally on completion, particularly by elevating and rotating the anastomoses if necessary to identify any bleeding point on the posterior wall consequent to gastric transection. Intraluminal hemorrhage can arise from a suture or staple line and can occur early or a few days later in association with anastomotic dehiscence. A second mucosal lesion overlooked at initial endoscopy and/or laparotomy can also give rise to postoperative hemorrhage. It may be possible to control early suture/staple line or peptic ulcer hemorrhage with injection sclerotherapy but if this is unsuccessful relaparotomy will be necessary. If gentle postoperative endoscopy with minimal insufflation reveals tumor, for example, a small leiomyoma, to be the source of the bleeding, then further surgical intervention is advisable as soon as possible.

Obstruction of the efferent limb of a gastrojejunal anastomosis is unusual but delayed gastric emptying can occur if truncal vagotomy has been previously performed for peptic ulcer disease. This is best managed with conservative measures, with endoscopy serving both to verify that the anastomosis is patent and that no mechanical obstruction is evident in the proximal 30 cm of the efferent limb. Afferent limb obstruction is suspected if duodenal stump leakage occurs. Not all stump leakage is attributable to afferent limb holdup, and reoperation can often be avoided with successful conservative treatment.

 

Table of Contents

Chapter 74 Billroth II Gastrectomy

Duodenal Stump Leakage

 

A gastrojejunal anastomosis is far less likely to be the site of leakage than is the duodenal stump, but if leakage is found at reoperation it can usually be dealt with by omental or jejunal seromuscular patch repair or resection. If upper abdominal pain accompanied by features of tachycardia, fever, or hypotension occurs in the postoperative period, leakage can be confirmed, if it is not overt, with the use of a water-soluble (Gastrografin) study that may identify duodenal stump extravasation. Complete afferent limb obstruction will result in failure to image the duodenum. If afferent limb obstruction is confirmed by the presence of a dilated afferent limb evident on ultrasound or computed tomography, then reoperation for its correction is essential if the fistula is to heal. In these circumstances the afferent limb can be transected with the GIA stapler close to its junction with the stomach. The gastric staple line is oversewn and the afferent limb reimplanted into the efferent limb in an end-to-side Roux en Y fashion.

Duodenal stump leakage may be associated with 1) inadequate closure of an extensively scarred and edematous duodenum, 2) afferent loop obstruction, or 3) local pancreatitis, this occurring secondary to injury of the main pancreatic duct or substance of the gland, division or ligation of the accessory pancreatic duct, or, less likely, simply excessive handling of the pancreas. The most important cause is inadvertent damage or division of the accessory pancreatic duct, which is vulnerable in a foreshortened scarred duodenum.

Duodenal stump blowout is potentially a major catastrophe with free intraperitonal leakage and peritonism; the reported mortality for stump leakage is 12 to 50 percent. The metabolic effects of external duodenal fistulization can be devastating. End catheter duodenostomy with or without T-tube drainage of the common bile duct or duodenum may salvage the situation. Treatment options for duodenal stump leakage also include reoperation with 1) immediate resuture or covering of the defect with omentum or a seromuscular jejunal patch, 2) resection, 3) additional drainage, or 4) conservative therapy. Management of overt stump leakage includes 1) nasogastric aspiration, 2) stump/right upper quadrant drainage with or without suction, 3) total parenteral nutrition or feeding jejunostomy, 4) H2 receptor antagonist or proton pump inhibitor, 5) somatostatin infusion, and 6) antibiotic therapy. Strict fluid balance measurements are required with an ongoing evaluation of the volume and nature of the drainage fluid. Blood cultures are necessary in the febrile patient. It is salutary to recall that patients with ongoing leakage who have received antibiotic therapy may often develop systemic infection with Candida. Attention to detail is required with regard to skin protection from pancreatic enzymes and avoidance of excoriation. Culture and sensitivity tests of the fistulous drainage are required at least twice a week. In the presence of ongoing fever despite external drainage, abdominal imaging with ultrasound or preferably computed tomography is indicated. Any localized abscess may require percutaneous drainage under image guidance. Unfortunately, duodenal fistulas are not generally amenable to direct closure but an omental plug can be sutured in place for closure. Once leakage occurs a protracted period of hospitalization is often required before resolution on conservative therapy. Catheter duodenostomy has been recommended if 1) leakage exceeds 200 ml per day when skin excoriation is problematic, 2) drainage is insufficient and bile-stained fluid leaks from the abdominal wound, or 3) features of generalized peritonitis are evident. Subsequent healing can be monitored by contrast studies via the catheter duodenostomy. Extravasation of contrast material will necessitate retention of the catheter for at least 5 days, at which time the study can be repeated. Free passage of contrast into the afferent limb and stomach without extravasation must be observed before clamping of the duodenostomy tube is considered safe. If no features of leakage such as abdominal pain or wound drainage arise overnight, the duodenostomy tube can be removed. Bile leakage via the catheter track will usually subside completely within 1 or 2 days; persistent leakage suggests some degree of afferent limb obstruction. When this occurs bile will be conspicuously absent from a nasogastric aspirate and the loop itself will appear dilated on ultrasound or computed tomography. Subphrenic or subhepatic abscess collections are suspected with the onset of pain and fever within 10 days of operation and are usually secondary to suture/staple line leakage or, less commonly, to infected hematoma.

Late sequelae of Billroth II gastrectomy include recurrent duodenal ulceration, chronic reflux gastritis, dumping, gastroesophageal reflux, and remnant cancer. Recurrent ulceration occurs in 5 percent of patients and may be associated with retained gastric antrum with gastrin production in the "duodenal" stump or hypersecretory states such as gastrinoma in the Zollinger-Ellison syndrome, or a mucosal defect. Long-term medical therapy may be advised or, occasionally, if surgical intervention is deemed necessary, a reresection of the stump with excision of the pylorus is performed if retained gastric antrum is confirmed on investigation. Stomal ulceration that fails to heal and merits surgery requires truncal vagotomy for satisfactory control, although the combination of vagotomy and previous resection may give rise to delayed gastric emptying for several weeks. It may well be that thoracoscopic truncal vagotomy alone may yet find a role in the management of stomal ulceration. Reflux gastritis in the remnant is attributable to bile and pancreatic enzyme damage and is perhaps best dealt with by treatment with mucosal cytoprotective agents such as sucralfate. Prostaglandin therapy is an alternative mode of treatment but is often associated with diarrhea and is unsuitable for prolonged usage. Bilious vomiting occurs when some degree of afferent limb obstruction is present and when bile under pressure in a distended intact afferent limb floods into the stomach, overcoming relative gastrojejunal afferent limb holdup. This may be accompanied by marked reflux esophagitis at endoscopy.

Cytoprotective agents may be helpful but if symptoms are troublesome revision gastric surgery with Roux en Y reconstruction is occasionally recommended. The etiology of postgastrectomy symptoms may be difficult to elucidate and may require endoscopy, 24-hour ambulatory intragastric and esophageal pH monitoring, and perhaps nuclear medicine studies. Early dumping following Billroth II gastrectomy is treated conservatively with small frequent meals and avoidance of liquid intake at mealtimes. Treatment is often unsatisfactory, with drug therapy and antiperistaltic jejunal interposition both proving disappointing; revisional surgery with restoration of gastroduodenal continuity, if technically feasible, is worthy of consideration in the patient with benign disease. Nutritional sequelae relating to iron, B12, and folic acid metabolism require careful follow-up study and monitoring biannually. Supplemental medication and regular B12 injections every 3 months are important aspects of care. There is good evidence to suggest that the postgastrectomy patient is at increased risk of gastric cancer, and annual gastroscopy, even in asymptomatic individuals, has been recommended, the risk being greater after a decade.

 

Table of Contents

Chapter 74 Billroth II Gastrectomy

Summary

 

A number of technical advances have been introduced to simplify and/or expedite the performance of Billroth II gastrectomy. It is imperative, however, to be aware of the perils of life-threatening surgical complications and to take no risks that are otherwise avoidable. The ideal procedure is one that will provide a successful symptom-free outcome that is lifelong. While various techniques have been described to restore gastrointestinal continuity, there is no strong evidence to suggest that simple Polya reconstruction is inferior for the vast majority of patients. A case for Roux en Y reconstruction may, however, be made in the patient with preexisting troublesome reflux esophagitis. The risk of stomal ulceration is believed to be increased if biliary and pancreatic diversion is performed and many surgeons routinely undertake truncal vagotomy concomitantly. The gastric surgeon must be familiar not only with anatomic choices for restoring continuity but also with the latest developments in minimal access techniques, which are likely to revolutionize surgical management of both benign disease and early gastric cancer.