Most patients admitted to hospital today with symptomatic gastric outlet obstruction have cancer. A palliative gastric drainage procedure may be indicated if the tumor is unresectable, or if the patient’s condition prohibits a more substantial operation. Dependent loop gastrojejunostomy is the surgical procedure of choice but in certain situations endoscopic stenting may be better (see below). When “double bypass” is done for palliation of pancreatic cancer, the gastric anastomosis is placed distal to the biliary anastomosis. When bypassing malignant gastric outlet obstruction, it is important to choose a site on the stomach that is likely to remain tumor free for at least several months. If this cannot be done without placing the anastomosis high on the stomach, alternative methods of palliation (e.g., gastrostomy and jejunostomy; and/or endoscopic stenting) should be considered. Anterior gastrojejunostomy is acceptable if placed in the distal third of the stomach. All patients with palliative gastrojejunostomy should be maintained on acid-suppressive medication unless shown to be achlorhydric since the operation is inherently ulcerogenic. Admittedly, most of these patients will not live long enough to experience the complications of marginal ulcer.

When the obstruction is in the proximal duodenum and/or distal stomach, gastrojejunostomy usually helps the troublesome symptoms of abdominal pain and vomiting. But when the duodenal obstruction is distal to the ampulla of Vater, gastrojejunostomy may not help, since with this procedure the biliary and pancreatic secretions must decompress retrograde through a functional pylorus. In this case consideration should be given to adding a pyloroplasty or pyloromyotomy to the dependent gastrojejunostomy. An alternative may be to anastomose a Roux limb or jejunal loop first to the proximal duodenum, and then to the dependent portion of the greater gastric curvature. We have used both
alternatives with good clinical results. Extending the gastrostomy through the pylorus into the duodenal bulb and doing a long gastroduodeno-jejunostomy anastomosis, is awkward gastroduoders-jejunostomy.

Meta-analysis suggests that endoscopic stenting compares favorably with surgical bypass in patients with malignant gastric outlet obstruction. While this may be true for “all comers,” in this author’s opinion, surgical bypass of malignant gastric outlet obstruction should not be performed in patients deemed to have a survival measured in weeks. Palliative endoscopic stenting of malignant gastric outlet obstruction should be considered in these patients. In many patients this palliates the vomiting and may allow them to take some liquids orally.

The causes of benign gastric outlet obstruction include peptic ulcer (by far the most common), duodenal webs, hypertrophic pyloric stenosis, and Crohn’s disease. Although distal gastric resection is the preferred operation in the good risk patient with ulcer-related gastric outlet obstruction, the lesser operation of highly selective vagotomy and gastrojejunostomy has been shown in a prospective randomized study to give comparable clinical results. Perhaps lifelong oral PPIs and gastrojejunostomy would be just as good or better. One attraction of gastrojejunostomy is that it can be performed laparoscopically or with a laparoscopic-assisted technique. Furthermore, if this procedure fails to improve the obstruction and/or ulcer symptoms, it does not preclude distal gastrectomy. Finally if gastrojejunostomy were to be associated with intolerable dumping, it is reversible. One disadvantage of gastrojejunostomy to treat obstructing peptic ulcer is that pyloric or duodenal cancer can be missed. In general, pyloroplasty (particularly the Heineke–Mikulicz and Finney type) requires pliability of the pyloric channel, and is therefore not a good option to relieve obstruction from peptic ulcer. Occasionally the chronic cicatrix is disk-like (rather than long and cylindrical) and near the pylorus, and amenable to pyloroplasty. It is worth remembering the Jaboulay pyloroplasty technique (essentially a side to side duodenogastrostomy) in the occasional patient in whom gastrojejunostomy is difficult due to adhesions and/or previous operations.

Gastrojejunostomy is a good operation to relieve the obstructive symptoms caused by proximal duodenal webs or Crohn’s disease. Obstruction distal to the ampulla of Vater is best treated with a proximal side-to-side duodenojejunostomy (Roux or loop). Duodenal webs or peptic strictures are often amenable to stricturoplasty using the Heineke–Mikulicz technique.

Pyloroplasty can be used as a closure technique for a longitudinal pyloroduodenal incision made, for example, to oversew a posterior bleeding duodenal ulcer or to excise a benign proximal duodenal lesion (polyp or small carcinoid). Alternatively, the longitudinal enterotomy may be closed longitudinally, and in the case of ulcer disease a highly selective vagotomy, or truncal vagotomy and gastrojejunostomy, added if desired.

It is generally accepted that a drainage operation should be added to truncal or total gastric (i.e., selective) vagotomy since the antrum and pylorus are denervated with these procedures. Denervation results in a spastic pylorus and a sluggish antrum, the net effect of which is symptomatic gastric stasis in some (but not all) patients. Gastric drainage is unnecessary with highly selective vagotomy, which preserves the vagal innervation of the antrum and pylorus. Most surgeons perform a pyloromyotomy or pyloroplasty as part of a transhiatal or transthoracic esophagogastrectomy unless a vagal sparing approach is used. In the unusual circumstance when a transabdominal proximal subtotal gastrectomy is performed, severing the vagal trunks and anastomosing the distal stomach to the distal thoracic esophagus, we do not disrupt the pylorus since we feel that preserving the pylorus may spare the patient (usually a poor risk patient with a large tumor) the unenviable sequelae of bile esophagitis.

Patients with postfundoplicaton gastroparesis may benefit from a gastric drainage procedure (usually pyloroplasty or pyloromyotomy), since the etiology is generally thought to be related to vagal trauma. However, most patients can and should be managed medically. The best way to make the diagnosis of total gastric vagotomy post fundoplication is (a) compare the 6-month postoperative gastric emptying scan to the preop scan and (b) obtain a sham feeding pancreatic polypeptide test to assess vagal integrity (a flat curve suggests complete vagotomy, but a normal curve—indicating the preservation of vagal input to the pancreas—may not rule out total gastric vagal denervation). If it is recognized during fundoplication or hiatal hernia repair that both the anterior and posterior vagal trunks have been severed, we would not add a gastric drainage operation at that time but the patient should be informed, and studied if symptomatic.

Increasingly, symptomatic patients without previous operation are being identified with quantifiable delayed gastric emptying. Some have other demonstrable abnormalities in gastrointestinal (GI) motility, and/or irritable bowel syndrome, and/or diabetes. These patients are a challenging group and remind us that there is much about GI pathophysiology that is unknown or poorly understood. The managing physician for these complex patients should usually be a gastroenterologist with motility expertise, not the general surgeon. Rarely, carefully selected patients may benefit from a gastric drainage procedure, but this surgical decision should never be made without the participation of the managing gastroenterologist, patient, and family. While the T½ gastric emptying time measured by scintigraphy may be improved by pyloroplasty, this can easily be outweighed by the pain and bilious vomiting associated with the attendant increased duodenogastric reflux sometimes associated with this procedure. Remember, we cannot use the large body of surgical literature on peptic ulcer surgery to predict the functional complications of pyloroplasty in this unusual group with primary gastroparesis. Also, remember that pyloroplasty is not reversible. Prior to considering pyloroplasty to facilitate gastric emptying for gastroparesis, it may be prudent to render the sphincter temporarily incompetent with balloon dilation and/or botox. A favorable response to these maneuvers may increase the likelihood of a favorable response to pyloroplasty. Alternatively, loop gastrojejunostomy, a reversible operation, may be considered and PPIs prescribed. Finally many of these patients may benefit from gastric pacing.

Not infrequently patients who present with gastroesophageal reflux disease (GERD) symptoms and objective criteria indicating fundoplication are found to have delayed gastric emptying on scintigraphy. While these patients may have a somewhat higher incidence of gas bloat symptoms post fundoplication, the large majority do well without the addition of pyloroplasty. It has been shown that fundoplication increases gastric emptying in most of these patients, and only rarely will subsequent pyloroplasty be necessary.

Most surgeons feel that pyloromyotomy and Heineke–Mikulicz pyloroplasty are comparable gastric drainage procedures. This was the conclusion of at least one prospective randomized trial in adults undergoing esophagectomy. The advantage of the pyloromyotomy is presumed to be a lower leakage rate. The disadvantage is there may be a tendency among nonpediatric surgeons to leave undivided pyloric muscle fibers, which could produce a suboptimal clinical result.

By far the easiest and most commonly performed pyloroplasty is the Heineke–Mikulicz type, but general surgeons should be familiar with the other types as well. For instance, if a 10-cm gastroduodenotomy is required to deal with a large posterior bleeding ulcer, closure with an Heineke-Mikulicg pyloroplasty may be difficult or impossible, but closure with the Finney technique may be straightforward. Or if the original plan was gastrojejunostomy but the proximal jejunum is involved with dense adhesions (e.g., s/p left colectomy) or Crohn’s disease, Jaboulay pyloroplasty would be a straightforward alternative.

These procedures are equally effective as gastric drainage operations. The major advantage of the pyloroplasty is that it does not require entry into the inframesocolic abdomen, and perhaps a leak might be more easily managed with drainage and gastric suction, since it is proximal to the entry of bile and pancreatic juice. The major advantage of the gastrojejunostomy is that it is easily reversible (unlike pyloroplasty). Furthermore, if subsequent distal gastrectomy is required, previous gastrojejunostomy rarely presents a technical challenge. But previous pyloroplasty, particularly the Finney or Jaboulay, may make the handling of the duodenal stump quite problematic. The major disadvantage of gastrojejunostomy is that it is ulcerogenic since the jejunal mucosa is less resistant to acid peptic injury than the proximal duodenum with its higher concentration of buffering and cytoprotective factors. Long-term acid suppression may mitigate this disadvantage. Another potential disadvantage of gastrojejunostomy is dysfunction, associated with the “circus movement” of duodenal contents, or proximal small bowel obstruction associated with kinking or adhesions.

Loop gastrojejunostomy involves one anastomosis, while Roux gastrojejunostomy involves two anastomoses and is more ulcerogenic (due to the complete lack of duodenal contents in the vicinity of the gastrojejunostomy). Therefore Roux gastrojejunostomy should be used only rarely as a drainage procedure with an intact stomach. On occasion we have used a Roux limb to drain the distal stomach (along with transection of the duodenal bulb and preservation of the right gastroepiploic artery) in patients with a history of severe bile esophagitis following proximal subtotal gastrectomy.