Although originally described as a pancreatic endocrine tumour, the definition has also come to include extrapancreatic gastrin-secreting tumours. The majority of tumours lie within an area defined by the junction of the cystic and common bile ducts superiorly, the junction of the second and third portions of the duodenum inferiorly, and the junction of the neck and body of the pancreas medially: the ‘gastrinoma triangle of Stabile’⁵ (Fig. 18.1). Where the condition is due to a pancreatic tumour, in two-thirds of cases the tumour will be multifocal within the pancreas.⁶ At least two-thirds will be histologically malignant. One-third will already have demonstrable metastases by the time of diagnosis.⁷ The most common extrapancreatic site is in the wall of the duodenum. Less frequently (6–11% of cases) ectopic gastrinoma tissue has been identified in the liver, common bile duct, jejunum, omentum, pylorus, ovary and heart.⁸ These extrapancreatic tumours rarely metastasise to the liver and, even though they do metastasise just as frequently to regional lymph nodes, they tend to have a better prognosis than primary pancreatic tumours.

Figure 18.1 Gastrinoma triangle of Stabile et al.⁵

One-quarter of patients with ZES have other endocrine tumours as part of a familial multiple endocrine neoplasia (MEN-1) syndrome, particularly hyperparathyroidism.⁷ This group of patients has a much worse prognosis than sporadic ZES, in part due to the multifocal nature of the disease.

Diagnosis

Diagnosis may be confirmed by paradoxical fasting hypergastrinaemia associated with gastric acid hypersecretion. Hypergastrinaemia may be expected to occur in cases of achlorhydria such as ingestion of antisecretory drugs, postvagotomy, pernicious anaemia, atrophic gastritis, antral G-cell hyperplasia or gastric outlet obstruction. Hypergastrinaemia is also associated with a retained antrum after a Billroth II/Polya-type gastrectomy where a small cuff of antrum has been included in the ‘duodenal’ closure (if a retained antrum is suspected, technetium pertechnetate scan may be useful in identifying the antral mucosa). If there is diagnostic uncertainty or the basal serum gastrin level is marginal, dynamic assay of serum gastrin following secretin (or alternatively calcium or glucagon) provocation may be required. Gastrin response to a standard meal helps to differentiate hypergastrinaemia due to antral G-cell hyperplasia, which will result in an increase in serum gastrin levels, while no response would be expected in cases of gastrinoma. Serum chromogranin A, a non-specific marker for neuroendocrine tumours, should also be measured.⁹

Tumour localisation

Tumours may be localised initially by computed tomography (CT). This may also identify metastatic disease. Endoscopic ultrasound (EUS) is highly accurate in the localisation of pancreatic tumours and gastrinomas in the duodenal wall as small as 4 mm. Octreotide scan and selective arterial secretagogue injection (SASI) testing are the most reliable approaches to localising gastrinomas. Liver metastases can frequently be detected by conventional imaging, but octreotide scan has proved a more sensitive investigation that may prevent unnecessary surgical exploration. SASI involves selective catheterisation of the feeding arteries of the duodenum and pancreas and the hepatic veins. Secretin is injected in turn into the splenic, gastroduodenal (GDA) and superior mesenteric (SMA) arteries. Corresponding hepatic venous gastrin levels are measured and allow identification of the main feeding vessel. More precise localisation can be achieved by more peripheral cannulation of the SMA and GDA or different points along the splenic artery. The test has greater than 90% sensitivity and specificity for preoperative tumour localisation.¹⁰

Surgery for ZES

The surgical management of ZES is characterised by controversy and little evidence. Historically, the debate centred around the radicality of surgical approaches to eliminate end organ acid production such as total gastrectomy. This is generally accepted, as unnecessary given that adequate acid suppression can usually be achieved with proton-pump inhibitors (PPIs), albeit at much higher doses than those usually recommended. How aggressively surgery should be pursued for the gastrinoma itself became the next area of controversy. With adequate acid suppression, patients may be rendered asymptomatic and the natural history of the gastrinoma tended to one of only very slow progression. Nevertheless, 60–90% of gastrinomas are reported to be malignant and some do have a more aggressive course. It became more acceptable to consider resection, as 30–50% with sporadic disease may be cured or at least have a reduced rate of development of liver metastases. Whether that should be local enucleation or a wider resection remains controversial. In the past many would say that patients with MEN-1 and those with liver metastases should not be treated surgically. Nevertheless, impressive results have been reported even in the former group and there is evidence that surgical resection of metastatic liver disease does offer long-term survival, even where resection may be incomplete. One of the largest series has shown that surgical exploration and resection resulted in excellent long-term results, with a 15-year disease-related survival rate of 98% compared to 74% for non-operated cases.¹¹

Surgical exploration when preoperative investigations have failed to precisely localise a tumour is now a less frequent problem, particularly in specialist centres with access to SASI and octreotide scan. Nevertheless, a laparotomy will detect a third more gastrinomas than even octreotide scan. If surgical exploration is performed then the pancreas must be mobilised along its entire length, inspected, palpated and if the facilities are available re-scanned intraoperatively by endoluminal or standard ultrasound. Palpation of the duodenal wall will identify 61% of duodenal gastrinomas. Duodenal transillumination by endoscopy will improve detection to 84% and duodenotomy identifies the remaining cases.¹² If no gastrinoma is found in the usual locations, other ectopic sites should be examined carefully. Resection of these primary ectopic tumours can sometimes lead to durable biochemical cures. Gastrinomas may be identified in 96% of surgical explorations if these approaches are adopted.¹¹ With the use of SASI in particular, though a tumour cannot be precisely localised, it may be sufficiently ‘narrowed down’ to allow a limited pancreatic and/or duodenal resection.¹⁰ The intraoperative secretin test in which gastrin levels in response to secretin are measured before and after resection can be useful in assessing the effectiveness of resection.

Emergency management of complicated peptic ulcer disease

Although very few patients now require elective surgery, the number who require surgery for the complications of peptic ulcer disease has remained constant for many years.

Perforation

A number of factors associated with poor outcome in perforated peptic ulcer have been identified: delay in diagnosis, coexistent medical illness, shock on admission, leucocytosis and age over 75. A delay in treatment of greater than 24 hours is associated with a sevenfold increase in mortality, threefold risk of morbidity and a twofold increase in hospital stay. The elderly are particularly vulnerable and often more difficult to diagnose because of poorly localised symptoms and signs and fewer preceding symptoms. The principles of treatment of peptic ulcer perforation involve resuscitation, control of contamination and prevention of recurrence.

Careful resuscitation and perioperative optimisation play a significant role in reducing morbidity and mortality associated with perforated peptic ulcer disease.¹³

Conservative management

Study of the natural history of perforated peptic ulcers suggests that they frequently seal spontaneously by omentum or adjacent organs and that, particularly when this occurs rapidly, contamination can be minimal. Taylor showed that the mortality in his series of patients with peptic ulcer disease was half that of the contemporary (pre-1946) reported mortality for perforation treated surgically.¹⁴ Conservative treatment today consists of parenteral broad-spectrum antibiotics, intravenous acid antisecretories, intravenous fluid resuscitation and nasogastric aspiration. In addition, water-soluble meal and limited follow-through is recommended to confirm that the leak has sealed. CT is becoming increasingly used in the diagnosis of acute abdominal pain and the degree of fluid contamination may also serve as a useful guide as to whether peritoneal lavage or drainage is necessary.

The conservative approach may be particularly appropriate in patients under 70, whose comorbidity (recent myocardial infarction, severe chronic obstructive pulmonary disease, etc.) place them at high operative risk. Sometimes, where there is some delay in presentation and patients exhibit an already improving clinical picture with minimal abdominal signs, conservative treatment is likely to be successful. Such an approach, however, requires careful interval assessment by an experienced surgeon with a low threshold for performing surgery if clinical improvement is not apparent, both to confirm the diagnosis and oversew an unsealed perforation. Small series have demonstrated low, comparable mortality for the conservative approach with conversion to operation of around 20–30%. Those that failed conservative treatment were more often over 70 years of age.15–17

Surgery

In most cases the treatment of choice for patients with perforation of the duodenum is still laparotomy, peritoneal lavage and simple closure of perforation, usually by pedicled omental patch repair (Fig. 18.2). The routine use of drains is unnecessary and may in fact increase morbidity. This simple treatment is safe and effective in the long term, when combined with pharmacological acid suppression. Ninety per cent of perforations are associated with HP infection,¹⁸ and HP eradication further significantly reduces the risk of ulcer recurrence.¹⁹

Figure 18.2 Closure of duodenal perforation with pedicled omental patch.

In cases of ‘giant’ perforation, where the defect measures 2.5 cm or more, partial gastrectomy with closure of the duodenal stump should be considered (see also management of bleeding from giant duodenal ulcer below). Alternatively, in situations where the clinical situation or expertise dictates more expeditious surgery, the duodenal perforation should be closed as well as possible around a large Foley or T-tube catheter to create a controlled fistula. This can be combined with venting gastrototomy and feeding jejunostomy.²⁰ The advances in understanding of the medical treatment of peptic ulcer disease together with the decrease in experience of elective antiulcer surgery have made the argument for definitive ulcer surgery in the emergency setting almost untenable.

Although laparoscopic treatment of peptic ulcer perforation was first reported in 1990 ²¹ and many excellent series have been reported since, a European population study demonstrated that the proportion of cases performed laparoscopically is as low as 6%,²² and even in centres with a specialist interest in laparoscopic surgery the proportion of cases completed laparoscopically is less than 50%.²³

There are several series that show favourable results with the laparoscopic approach,²⁴ but there are very few prospective studies.²⁵ Shock, delayed presentation, confounding medical condition, age greater than 70 years, poor laparoscopic expertise and ASA grade are risk factors for open conversion and for poorer outcome.²⁶ In ‘low-risk’ patients the laparoscopic approach may have significant advantages. For higher-risk patients (prolonged perforation for > 24 h, shock on admission and confounding medical conditions) there is no evidence that the laparoscopic approach is advantageous.

Whatever the approach, basic surgical tenets must be observed: careful preparation for theatre with timely expeditious operation involves thorough peritoneal lavage and secure closure of the defect. Any marginal benefit of the minimally invasive approach is lost if any of these tenets are compromised.

Bleeding

Management of acute haemorrhage from peptic ulceration of the stomach and duodenum has been revolutionised by rapidly developing endoscopic technology and expertise. The principle of successful management is by prompt resuscitation, accurate endoscopic diagnosis and the timely application of appropriate therapy.

Medical therapy

Although there is evidence that proton-pump inhibitors given pre-endoscopy reduce the incidence of endoscopic findings of stigmata of recent haemorrhage and the need for endoscopic intervention, there is no evidence that this or any other specific pre-endoscopy medical intervention has any effect on overall morbidity, mortality, or specifically the risk of re-bleeding or need for surgery.²⁷

There is compelling evidence that proton-pump inhibitors given after endoscopic control of bleeding are beneficial. A randomised controlled study (n = 240) from Hong Kong has demonstrated a significant reduction in re-bleeding following endoscopic treatment with a protocol of intravenous omeprazole (omeprazole 80 mg i.v. bolus followed by 8 mg/h infusion for 72 hours).²⁸ This is supported by meta-analysis of subsequent studies.²⁹

Meta-analysis of randomised, double-blinded trials with tranexamic acid reveals no significant difference in the incidence of re-bleeding but an increase in complications related to therapy such as stroke, myocardial infarction, deep vein thrombosis and pulmonary embolism.³⁰

Somatostatin decreases gastric acid and pepsin secretion. Nevertheless, there is no proven benefit of somatostatin or its analogue (octreotide) in the management of active non-variceal upper gastrointestinal bleeding. Prostaglandin E₂ and its analogue (misoprostol) inhibit gastric acid production, stimulate mucosal perfusion, and promote bicarbonate and mucus secretion. Small studies to date have demonstrated no benefit of stopping acute bleeding or preventing re-bleeding.

Endoscopic therapy

The various techniques of endoscopic haemostasis have dramatically reduced the need for emergency surgery for bleeding due to peptic ulceration.

Meta-analysis suggests that endoscopic therapy reduces the mortality of acute upper gastrointestinal bleed in patients’ active bleeding or non-bleeding visible vessel by avoiding the often considerable morbidity or mortality of emergency surgery.³¹

Ulcers with a clean base or non-protuberant pigmented dot in an ulcer bed, which are at low risk of re-bleeding, do not require endoscopic treatment. For all others, including those who have active bleeding or non-bleeding visible vessels or have adherent blood clot, endoscopic treatment should be given.³²

Injection with 4–16 mL 1:10 000 adrenaline around the bleeding point and then into the bleeding vessel achieves haemostasis in up to 95% of cases. Additional injection with sclerosants (sodium tetradecyl sulphate, polidoconal, ethanolamine) or absolute alcohol does not confer additional benefit and may cause perforation. Fibrin glue and thrombin may be more effective, but they are not widely available.

Techniques used commonly are the heater probe, multipolar coagulation (BICAP) and argon plasma coagulation. There is no strong evidence to recommend one thermal haemostasis technique over another.

Mechanical clips have had variable success reported when compared with other techniques. This may reflect the technical difficulties with their placement. In certain situations, such as active bleeding from a large vessel, they may be particularly useful.

There is strong evidence that, for patients at higher risk of re-bleeding, treatment by a combination of two different modalities is more beneficial than relying on one modality alone.³³ The commonest combination is likely to be adrenaline injection and heater probe application.

There is no evidence to support a repeat endoscopy unless there is a suggestion of further active bleeding or it is felt that the initial endoscopic treatment was suboptimal. Nevertheless, some clinicians do choose to re-evaluate higher risk cases after 24–48 hours and consider further endoscopic treatment.

If there is evidence of re-bleeding following endoscopic treatment, unless there is evidence of haemodynamic instability or the ulcer size was greater than 2 cm, a further attempt at endoscopic control is recommended.³⁴

Surgery

Operative or radiological intervention is mandatory if initial control of bleeding is not possible endoscopically. Further intervention should also still be considered if re-bleeding occurs following initially successful endoscopic treatment. Re-bleeding may be observed directly endoscopically or indirectly by continuing haematemesis, or the continuing need for transfusion. If there is doubt as to whether re-bleeding has occurred, a check endoscopy should be performed before subjecting a patient to surgery.

Surgical intervention should be anticipated where there is a significant risk of re-bleeding. Various scoring criteria have been suggested to predict risk of significant re-bleeding and death; one commonly used is the Rockall system (Table 18.1). In addition, the size of the ulcer (particularly > 2 cm) and its proximity to major vessels, such as the gastroduodenal ulcer on the posterior inferior wall of the duodenal bulb and the left gastric artery high on the lesser curve of the stomach, suggest a high risk of massive bleeding.

Table 18.1

Rockall scoring system for risk of re-bleeding and death after admission to hospital for acute gastrointestinal bleeding

A total score of > 3 is associated with good prognosis; < 8 is associated with high risk of death.

BP, blood pressure; GI, gastrointestinal; IHD, ischaemic heart disease; SRH, stigmata of recent haemorrhage.

Bleeding duodenal ulcer: The first step is to make a longitudinal duodenotomy immediately distal to the pyloric ring. Haemostasis can be initially achieved by digital pressure. While it may be necessary to extend the duodenotomy through the pyloric ring, the pylorus should be preserved if at all possible. Older texts frequently assume that vagotomy is an integral part of ulcer surgery and recommend a larger pyloroduodenotomy, but this is usually not necessary. The stomach and duodenum should be cleared of blood and clots using suction to obtain optimal view of the bleeding site. If access is still difficult, kocherisation of the duodenum may help, along with drawing up of the posterior duodenal mucosa using Babcock’s forceps.

The actively bleeding or exposed vessel should be secured. Points of note in securing the vessel are the limited access, the proximity of underlying structures such as the common bile duct and the tough fibrous nature of the base of a chronic ulcer. In view of these problems, a small, heavy, round-bodied or taper-cut semicircular needle with 0 or No. 1 suture material should be used. The argument of absorbable versus non-absorbable sutures is irrelevant: the sutures probably slough off as the ulcer heals. Securing bleeding from the gastroduodenal artery may involve a horizontal mattress ‘U-stitch’ to incorporate posterior and medial perforating vessels (Fig. 18.3).

Figure 18.3 Suture control of bleeding from gastroduodenal artery illustrating the ‘U-stitch’ incorporating any perforating vessels.

The duodenotomy may be closed longitudinally. If vagotomy has been performed the pyloric ring should be divided and the duodenotomy closed transversely to create a Heineke–Mikulicz pyloroplasty (Fig. 18.4a). If transverse closure is difficult because of the length of the duodenotomy, longitudinal closure may be performed and a gastrojejunostomy considered. Alternatively, a Finney pyloroplasty may be fashioned (Fig. 18.4b).

Figure 18.4 (a) Heineke–Mikulicz pyloroplasty. (b) Finney pyloroplasty.

In a giant ulcer, the first part of the duodenum may be virtually destroyed and, once opened, impossible to close. In this situation it is necessary to proceed to partial gastrectomy. The right gastric and right gastroepiploic arteries are divided. The stomach is disconnected from the duodenum by a combination of blunt and sharp dissection. Antrectomy is perfomed and continuity restored by a gastrojejunostomy. The duodenal stump can then be closed. Although this can be achieved by pinching the second part of the duodenum away from the ulcer to allow conventional closure, this is probably more safely achieved by the technique of Nissen (Fig. 18.5). The duodenal stump is drained by either a tube or Foley catheter either through the duodenal suture line or more securely though the healthy sidewall of the second part of the duodenum (Fig. 18.6).

Figure 18.5 Nissen technique.

Figure 18.6 Duodenal drainage following partial gastrectomy for duodenal ulcer.

Long-term acid suppression is required postoperatively. With the advent of proton-pump inhibitors and the recognition of the role of HP, vagotomy should have no part of surgery for bleeding duodenal ulceration.

Bleeding gastric ulcer: The precise site of bleeding should already have been identified endoscopically. If not, intraoperative endoscopy and careful palpation of the stomach for induration should identify the site of the bleeding ulcer. If there is still doubt a generous incision should be made across the pylorus and duodenum, followed by a more proximal gastrotomy if the source of bleeding is still not clear. Most chronic gastric ulcers are at the incisura or in the antrum. The traditional treatment for such ulcers that fail endoscopic therapy is partial gastrectomy. Some groups have advocated simple under-running of bleeding gastric ulcers. While this may be appropriate in selected cases with small bleeding gastric ulcers such as the Dieulafoy lesion, the only randomised trial to date (n = 129) suggests that this ‘conservative’ approach has a higher mortality and is more likely to result in re-bleeding if used unselectively.³⁵

For proximal gastric ulcers, typically those high on the lesser curve eroding through into the left gastric artery, the choice of operation lies between total gastrectomy or local excision of the lesser curve (Pauchet’s manoeuvre). Frequently such limited procedures involve as much mobilisation of the stomach as total gastrectomy. There is no evidence to recommend one approach over another, though the experience of the surgeon is a major factor in the decision-making process.

Interventional radiology

There are no randomised controlled trials comparing surgery with transcatheter arterial embolisation as salvage treatment following endoscopy. Despite this, in some centres, interventional radiology has become the ‘gold standard’ intervention following failed endoscopy. Loffroy et al. looked at 15 studies that included 819 patients who had had failed endoscopic control of bleeding; 93% of cases were reported to have ‘technical success’.³⁶ Of these, 67% had immediate cessation of bleeding. Of those that continued to bleed after initial treatment, half responded to a second embolisation. Overall, 20% of patients underwent salvage surgery. The overall mortality in these series was 28%. This seems disappointing and no better than one might expect from a series of bleeders salvaged by surgery. In fact, there was a wide variation between reported mortality in the series, which may indicate different levels of expertise, and case selection. The series were also collected over a 17-year period, during which time there has been considerable improvement in radiological technique. There have been two small retrospective comparisons involving a total of 161 patients.^37,³⁸ They showed that although the radiologically treated patients tended to be older and less fit, they had a comparable (26% vs. 21%) or lower (3% vs. 14%) mortality compared with the surgically treated group.

With a limited evidence base it seems that angiography is a reasonable option where the expertise and facilities are readily available, particularly if the surgery itself is deemed high risk.^36,³⁷

Pyloric stenosis

Gastric outlet obstruction can result from peptic ulcer disease of the duodenum or pre-pyloric region. It is a condition usually associated with chronic relapsing ulceration and is now fairly uncommon in the Western world.

Resuscitation and medical therapy

Initial management should consist of aggressive parenteral fluid and biochemical restoration with nutritional and vitamin supplementation as necessary. Nasogastric intubation with a wide-bore tube allows gastric washout of undigested food and so reduces antral stimulation. Aggressive parenteral antisecretory therapy and Helicobacter pylori eradication, if appropriate, are used. In cases where the obstruction has been due to oedema and spasm, the situation can be expected to resolve once medical treatment has healed the ulcer.³⁹ Dietary changes to decrease the fibre content while providing a high calorie and protein intake are important until ulcer healing has occurred. In cases where the obstruction is due to fibrosis and cicatrisation of a pyloric ulcer, some form of intervention will be necessary.

Endoscopic treatment

The group of patients who develop gastric outflow obstruction are generally elderly with established comorbidities and tolerate major surgery poorly. As a result, minimally invasive approaches such as endoscopic management are often more appropriate in the first instance. Initial reports of successful resolution of pyloric stenosis following endoscopic balloon dilatation were challenged due to the relatively high number of cases that ultimately required open surgery. Nevertheless, this remains a useful first-line endoscopic procedure that can be repeated on several occasions with good long-term results in up to 80% of patients.⁴⁰

Only if a combination of intensive medical treatment and repeated dilatation fails to reopen the gastric outlet is surgery indicated.⁴⁰

Surgery

There are no published series that prove which procedure achieves the best results in this situation. Initial fears about the capacity of a large atonic stomach to resume function have not been realised. The operation with least complications is simple pyloroplasty (or gastroenterostomy where the inflammation around the pylorus is particularly intense), with the use of long-term medical acid suppression. Antrectomy and selective vagotomy or subtotal gastrectomy are more aggressive alternatives less likely to result in re-stenosis, but with a higher mortality and incidence of both short- and long-term side-effects.

Laparoscopic highly selective vagotomy with balloon dilatation has been attempted with some success in cases of pyloric stenosis. This has not been proven to be superior to dilatation and long-term acid suppression. Laparoscopic truncal vagotomy and gastroenterostomy has proven to be a technically feasible solution with good symptomatic, sustained response.⁴¹

Complications of previous ulcer surgery

Although elective surgery for benign ulcer disease is now rare, there remains a large cohort of patients operated on prior to the mid-1980s with a variety of surgical procedures, of whom a small percentage will develop further symptoms, some of which may be severely disabling. Although numerous clinical syndromes have been well described (Box 18.1), patients presenting with pure syndromes are uncommon. The majority presents with a mixed picture, but usually have a dominant symptom complex suggesting one main problem. This needs to be elucidated by a careful and detailed history of the clinical events occurring during a bad attack.

Preoperative evaluation

Endoscopy

Endoscopic examination is essential, and as with patients after previous antireflux surgery, it should be carried out by the surgeon considering any revisional procedure. The exact anatomy, size of the gastric remnant, size and position of any drainage procedure, the presence of enterogastric reflux of bile, recurrent ulceration, the general state of the gastric mucosa, and the presence of a hiatus hernia and/or reflux oesophagitis can be assessed. All abnormalities should be biopsied. All patients should be assessed for the presence of HP.

Radiology

Barium meal examination of the stomach is a useful adjunct where the anatomy remains unclear.

Gastric-emptying studies

Gastric-emptying studies may occasionally be useful. Barium meal examination may show rapid emptying of the contrast from the stomach and may demonstrate gross intestinal hurry with the meal reaching the caecum within a short time of leaving the stomach. Gastric emptying is, however, best studied using a radioactively labelled meal, either liquid or solid. In general, the radioactive liquid meals are easier to interpret than solid meals. The normal measured indices such as 10-minute emptying, the T_1/2 and the percentage retention after 60 minutes are often used in assessment. However, after gastric surgery these indices can be misleading as the patients often show a fast initial emptying component followed by a slower component.

Other tests

Congo red for the evaluation of the completeness of vagotomy and dumping provocation tests are now seldom performed. Oesophageal function tests will be required in those patients suspected of having gastro-oesophageal reflux. Enterogastric reflux can be assessed using the hepatobiliary dimethylacetanilide iminodiacetic acid (HIDA) scan. Bacterial overgrowth can be diagnosed by aspiration and culture of jejunal contents or by the [¹⁴C]glycocholate breath test.

Various nutritional indices, including weight, serum albumin, transferase and corrected serum calcium concentration, should be measured in all patients. In selected patients full assessment for metabolic bone disease should be undertaken, especially in postmenopausal women. A full haematological survey should be carried out including measurement of serum iron, iron-binding capacity, folate and vitamin B₁₂ levels.

Enterogastric reflux

Reflux of alkaline duodenal content into the stomach occurs following surgery that damages, bypasses or removes the pylorus. Enterogastric reflux is more common after gastrectomy where reconstruction as a Billroth II gastrojejunostomy has been carried out.

The symptoms consist of persistent epigastric discomfort, sometimes made worse by eating and frequently associated with intermittent vomiting of bile-stained fluid or food mixed with bile, usually occurring within 90 minutes of a meal. Some patients become malnourished because of inadequate food intake, and anaemia develops in about a quarter of the patients as a result of chronic blood loss from the associated gastritis. Gastro-oesophageal reflux disease may also develop.

Endoscopy shows a diffuse gastritis with an oedematous hyperaemic friable mucosa and frequently superficial erosions. Endoscopic biopsy shows typical histological features including foveolar hyperplasia, glandular cystification, oedema of the lamina propria and vasocongestion of the mucosal capillaries, all in association with inflammatory cell infiltration.

Medical treatment

Cholestyramine has been shown to be an effective bile-acid-binding agent in vitro, although the results of several therapeutic trials have been disappointing. Antacids containing aluminium hydroxide have also been studied because of their bile-acid-binding capacity but the results have been equally unimpressive. In clinical trials sucralfate has been shown to reduce the inflammation within the gastric mucosa but this has not been associated with any improvement in symptoms. Prokinetic agents have also been used to improve clearance of the refluxate from the stomach, and the occasional patient may respond. These agents may, however, worsen dumping and diarrhoea. Ursodeoxycholic acid has been shown in one study to almost abolish the nausea and vomiting associated with enterogastric reflux and to significantly decrease the intensity and frequency of pain.

Surgical treatment

In patients with a previous truncal vagotomy and drainage, reversal of the drainage procedure can be undertaken provided at least 1 year has elapsed from the original operation. This is based on the premise that the stomach will regain some of its lost motility during this time. In fact, more than half of the patients with truncal vagotomy probably did not require a drainage procedure in the first place. Improvement or complete relief usually follows closure of gastrojejunostomy for enterogastric reflux and bile vomiting in the vast majority of patients. The risks of gastric stasis are minimal and conversion to a pyloroplasty should be avoided.

Reconstruction of the pylorus after pyloroplasty is a relatively straightforward operation. Having cleared the anteropyloroduodenal segment of all adhesions, the scar of the previous pyloroplasty is accurately opened. The pyloric ring is palpated and the scarred ends freshened if necessary. One approach is to make a small antral gastrotomy to allow the insertion of a size 12 or 14 Hegar dilator through the area of the pyloric reconstruction into the duodenum. Using a double-ended monofilament suture the pyloric ring is accurately opposed around the Hegar dilator before reapproximating the duodenum and antrum using a continuous serosubmucosal technique. Withdrawal of the Hegar dilator allows fingertip palpation of the reconstructed pylorus prior to closure of the antral gastrotomy. The overall results of pyloric reconstruction show that 80% of patients gain a satisfactory or good result,⁴² although in one study only half of the patients with enterogastric reflux had a satisfactory or good response.⁴³

If enterogastric reflux is not relieved, then the duodenal switch operation would seem an appropriate further remedial procedure for patients whose symptoms necessitate further surgery ⁴⁴ (Fig. 18.7). Recent experience with this has shown good results, although acid suppression is needed to prevent jejunal ulceration.⁴⁵

Figure 18.7 Pylorus-preserving Roux loop – the duodenal switch operation.

In patients who have had a gastric resection or in those with a gastrojejunostomy with pyloric stenosis, a Roux limb (approximately 45 cm in length) would seem an appropriate revisional procedure (with antrectomy in patients with pyloric stenosis). The procedure, however, does carry risks, as it is ulcerogenic because it diverts the buffering effect of upper gastrointestinal contents away from the gastroenteric anastomosis. The second problem is the development of delayed gastric emptying of solid food, producing a symptom complex of satiety, epigastric pain and non-bilious vomiting that has been termed the ‘Roux syndrome’. Although many patients will demonstrate objective evidence of delayed gastric emptying of solids, this is usually of little or no clinical consequence except in a minority. The Roux syndrome is more likely to develop in patients who demonstrate delay in gastric emptying of solids prior to construction of the Roux limb and those who have a large residual gastric pouch. A completion vagotomy at the time of revisional surgery may make these symptoms more likely. Where these conditions exist, the operative procedure required is a more extensive gastric resection. The entire anastomosis should be resected to leave a small gastric pouch, and the Roux limb should be anastomosed to the stomach as an end-to-side Polya-type gastrojejunostomy. In those patients who develop severe symptoms from the Roux syndrome postoperatively, then the treatment is near-total resection of the gastric remnant with a Polya-type gastrojejunostomy.

Roux diversion will control enterogastric reflux in over 70% of patients. Recurrent jejunal ulcers can be avoided by checking and if necessary completing the truncal vagotomy as part of the operative procedure. More commonly, today, one would consider relying on long-term treatment with proton-pump inhibitors rather than perhaps risking further side-effects of vagotomy.

Chronic afferent loop syndrome

The afferent loop syndrome can only occur after gastrojejunostomy or a Billroth II-type reconstruction after partial gastrectomy. The condition is caused by intermittent postprandial obstruction of the afferent limb of the gastrojejunostomy. The clinical picture is very similar to that produced by enterogastric reflux (Table 18.2). The problem is rarely encountered if surgeons use a short afferent jejunal loop. The obstruction may be due to anastomotic kinking, adhesions, internal herniation, volvulus of the afferent limb or obstruction of the gastrojejunal stoma itself (Fig. 18.8). Once diagnosed, the treatment is always surgical. Conversion to a Billroth I anastomosis or a Roux-en-Y reconstruction of the afferent limb both produce good results.

Table 18.2

Differentiation between the chronic afferent loop syndrome and enterogastric reflux

Chronic afferent loop syndrome	Enterogastric reflux
Meal-related pain – relieved by vomiting	Constant pain (worsened by eating) – not relieved by vomiting
Vomitus contains bile	Vomitus contains bile and food
Vomiting projectile	Vomiting non-projectile
Rarely associated with bleeding/anaemia	Bleeding/anaemia found in 25% of patients

Figure 18.8 Causes of afferent limb syndrome.

Dumping

The literature shows a considerable variability in the incidence of dumping after each procedure due at least partly to variations in definitions of the syndrome. A significant number of patients will develop dumping-type symptoms in the early period after their initial gastric operation but the majority have sufficient reserve to adjust to the changes without developing severe sequelae.

The symptoms of early dumping can be divided into vasomotor and gastrointestinal, as shown in Box 18.2. In a severe attack, the vasomotor symptoms are usually experienced by the patient towards the end of a meal or within 15 minutes of finishing, and the gastrointestinal symptoms develop a little later, but usually within 30 minutes after eating.

Early dumping is associated with rapid gastric emptying leading to hyperosmolar jejunal content causing massive fluid shifts from the extracellular space into the lumen. This is associated with a significant fall in plasma volume. It is also known that plasma concentrations of several gut regulatory peptides are elevated in patients with the dumping syndrome, but it is not clear whether this is coincidental or causative. Late dumping symptoms are the result of reactive hypoglycaemia. Taking a careful history, delineating the vasomotor and gastrointestinal components, usually makes the diagnosis of the dumping syndrome. Where there is any doubt, the patient should be encouraged to keep a diary card recording the foods eaten and the symptoms that develop thereafter. A provocative test for assessing dumping syndrome can be used to confirm clinical suspicion. This test is a modification of the oral glucose tolerance test and involves the ingestion of 50 or 75 g glucose in solution after an overnight fast. Immediately before and up to 180 min after ingestion of this solution, the blood glucose concentration, haematocrit, pulse rate and blood pressure are measured at 30-min intervals. The provocative test is considered positive if late (120–180 min) hypoglycaemia occurs, or if an early (30 min) increase in haematocrit of more than 3% occurs. The best predictor of dumping syndrome seems to be a rise in the pulse rate of more than 10 b.p.m. after 30 min.⁴⁶

Medical treatment

The majority of patients displaying the dumping syndrome can be managed satisfactorily by dietary manipulation. Reducing the carbohydrate content and restricting fluid intake with meals will help many of these patients. Avoiding extra salt and eating more frequent small meals may also be required. Lying down after eating helps to slow gastric emptying and may minimise symptoms. Guar gum, a vegetable fibre, is known to reduce postprandial hyperglycaemia in both normal and diabetic patients. In a small study of postgastric surgery patients it has been shown to prevent the dumping syndrome and increase food tolerance in the majority of patients.⁴⁷ Pectin also delays gastric emptying but may precipitate diarrhoea. The use of acarbose, an alpha-glycoside hydrolase inhibitor, interferes with carbohydrate absorption and has been shown to help in patients with late dumping. Octreotide, given subcutaneously prior to eating, has been shown to significantly reduce or abolish the symptoms of dumping.⁴⁸ The use of short-acting octreotide was quite troublesome for patients and only around half saw long-term benefit.⁴⁹ Encouraging results have been seen with a longer-acting repeatable (LAR) formulation of octreotide, which only needs administration once a month, rather than with each meal.⁵⁰

An algorithm for the treatment of postoperative dumping is shown in Fig. 18.9.⁵¹

Figure 18.9 Algorithm for treating postoperative dumping. Adapted from Tack J, Arts J, Caenepeel P et al. Pathophysiology, diagnosis and management of postoperative dumping syndrome. Nat Rev Gastroenterol Hepatol 2009; 6(10):583–90.

Surgical treatment

For patients with truncal vagotomy and drainage procedures, taking down the gastrojejunostomy ⁵² should cure or improve dumping in over 80% of patients. Reconstruction of the pylorus produces similar results.⁵³ After gastrectomy, a number of procedures have been advocated for dumping. The simplest and probably the best is to convert the drainage procedure to a 45-cm Roux-en-Y gastrojejunostomy. The delay in liquid emptying after this procedure is thought to be due to myoelectrical abnormalities within the Roux limb itself, causing a degree of retrograde contraction. The delay in emptying of solids is probably a result of the vagotomy leading to a degree of gastric atony and loss of the antral propulsive force to propel solid food into the small intestine. Reversal of the proximal 10 cm of the jejunal limb to create an antiperistaltic interposition is unnecessary and may lead to further stasis and dilatation of the interposed segment. This will worsen any symptoms of gastric retention. The interposition of a segment of upper jejunum between the gastric remnant and the duodenum has been advocated. Both isoperistaltic and antiperistaltic interpositions have been used, but these procedures can be associated with serious complications, and the long-term success rate is variable.⁵⁴