Treatment of the complications of gastro-oesophageal reflux disease and failed gastro-oesophageal surgery

Published on 16/04/2015 by admin

Filed under Surgery

Last modified 22/04/2025

Print this page

This article have been viewed 1941 times

Treatment of the complications of gastro-oesophageal reflux disease and failed gastro-oesophageal surgery

Robert Mason

Introduction

Gastro-oesophageal reflux affects up to a quarter of the UK population on a regular basis. Most will not seek medical help and self-medicate. Of those who do see a doctor the vast majority will be well controlled by medical therapy – invariably proton-pump inhibitors. It has been calculated that only 1% of patients with gastro-oesophageal reflux disease (GORD) will develop complications of the reflux. These complications are consequent on damage to the mucosa resulting in erosive oesophagitis with, rarely, a peptic oesophageal ulcer, a peptic stricture or Barrett’s oesophagus (see Chapter 15). The vast majority of sufferers of GORD have non-erosive reflux disease (NERD) and never develop either erosive oesophagitis or Barrett’s oesophagus.¹

There are two main treatments for GORD. These are medical therapy with proton-pump inhibitors or surgical therapy based on either a total or partial fundoplication. Comparative studies suggest there is little to choose between the two therapeutic arms in either the control of symptoms or prevention of complications. This is well documented in the results of the multicentre LOTUS trial, comparing esomeprazole and laparoscopic antireflux surgery, with good control of symptoms and microscopic oesophagitis in over 85% of cases in both arms out to 5 years.^2,³

In recent years there has been a significant increase in the use of fundoplication based on the laparoscopic approach. This is claimed to be a one-off long-term treatment that removes the need for expensive long-term ingestion of proton-pump inhibitors. However, it is recognised that with time the use of these drugs does increase in patients who have had a successful initial surgical outcome, reaching 25% at 10 years.^4,⁵

In the USA, the number of antireflux operations increased by 260% between 1993 and 2000 but decreased by 40% between 2000 and 2006.⁶ Similar trends have been seen in Europe, particularly in Sweden. The reasons are unclear but may reflect disappointing long-term results for surgery, with a significant proportion of patients back on medical therapy as described above and recognition of the complications, which although rare can have a serious impact on a young, previously fit patient.

Gastro-oesophageal reflux results from failure of the lower oesophageal sphincter to either provide a mechanical barrier (volume/supine refluxers) or which relaxes inappropriately, usually secondary to gaseous distension of the gastric fundus (upright refluxers). These differences can have an effect on the outcome of antireflux surgery whereas defective oesophageal motility, other than undiagnosed achalasia, does not.^7,⁸

Complications of GORD

Complications arise from damage caused by the refluxate to the squamous mucosa of the oesophagus. The contents of the refluxate may include any combination of acid and/or ‘bile’ (duodenogastric contents). Chronic exposure of the oesophageal mucosa to this fluid in patients with an intact stomach causes inflammation, can result in erosive oesophagitis with ulceration and subsequent peptic stricture, and in some results in metaplastic change and the development of Barrett’s oesophagus. It appears that acid alone is the principal factor in determining the severity of oesophagitis, whereas in patients experiencing symptoms who are on long-term proton-pump inhibitors weak acid or bile is implicated.^9,¹⁰

Short oesophagus

A complication of chronic reflux that promotes much debate among upper gastrointestinal (GI) surgeons is the short oesophagus. This is claimed to be a complication of prolonged reflux damage and fibrosis, and is manifested by an inability to mobilise the oesophagogastric junction at surgery to allow 2–3 cm of oesophagus below the hiatus. Whether this exists in reality is contentious. In this author’s experience it is very rare, although it may have been more prevalent 30–40 years ago in the pre-antisecretory era. By opening the hiatus and fully mobilising the oesophagus, the oesophagus can always be brought down without tension.

Those who recognise the complication advocate a Collis gastroplasty to lengthen the oesophagus through either an open abdominal or laparoscopic approach. It may provide a better outcome in such patients when compared with fundoplication alone.¹¹ In this procedure a 40 F bougie is passed into the stomach and kept on the lesser curve. A neo-oesophagus is created using a circular and linear cutting stapler and a loose wrap performed using the mobilised stomach (Fig. 14.1). This has the potential to leave a tube of acid-secreting stomach above the wrap, with consequent complications of bleeding, ulceration and stenosis. It has been used in revisional surgery, where mobilisation of the gastro-oesophageal junction below the diaphragm has proven difficult. In experienced hands the results are reported to be good.¹²

Figure 14.1 A Collis gastroplasty. Stapled lines shown with Z line above the wrap.

Gastrointestinal haemorrhage

The treatment of erosive oesophagitis is medical with proton-pump inhibitors in the first instance, with excellent healing rates, although double-dose treatment may be required with maintenance therapy after healing has been achieved.¹³ To attribute gastrointestinal haemorrhage to erosive oesophagitis is a matter of exclusion after full endoscopic examination as it is not a common cause of significant bleeding except in those with a tendency to bleeding, usually as a result of other medication they are taking.

In rare cases a true peptic ulcer can arise in an area of oesophagitis or Barrett’s oesophagus. Such an ulcer can erode into blood vessels and bleed. This can be controlled endoscopically by injection of the ulcer base with adrenaline and application of clips or a heater probe. In rare cases where the bleeding cannot be controlled, embolisation can help and possibly the use of a Sengstaken tube. Surgery will be difficult as it requires oversewing of the bleeding vessel or possible resection through a left thoracoabdominal approach. In all cases of ulceration the surrounding mucosa and ulcer margins should be biopsied to exclude Barrett’s oesophagus, dysplasia and malignancy when the acute episode has passed.

Peptic oesophageal stricture

The incidence of symptomatic stricture is extremely low in the pantheon of reflux disease. The management involves endoscopy and biopsy to exclude serious pathology followed by optimum acid inhibition with proton-pump inhibitors and H₂-receptor antagonists and gentle dilatation with an endoscopic balloon (Fig. 14.2). The diameter is dependent on the size of the stricture but it is better to undertake sequential dilatation starting with a 10-mm balloon rather than use too big a balloon and risk splitting the oesophagus. The risk of perforation is of the order of 2–3%. This should be recognised at the time and treated conservatively with nil by mouth, intravenous proton-pump inhibitors and antibiotics plus placement of a nasojejunal tube under screening for feeding. Success can be achieved in over 90% of cases with conservative management. If the perforation is not recognised and the patient develops sepsis then a more aggressive approach is required, including drainage and surgical repair. Such patients are best referred to specialist centres with full intensive care and interventional radiological support.^14,¹⁵ The use of self-expanding plastic stents in such perforations shows promise but can cause problems and is best avoided. Any stent used in this context must be removable.

Figure 14.2 Endoscopic balloon dilatation of peptic stricture.

Patients with symptomatic peptic strictures will often require repeat dilatation to maintain swallowing. In such cases injection of steroids can be beneficial and division of a tight band-type stricture with a laser can reduce the frequency of dilatation.¹⁶ The use of self-expanding plastic stents is controversial and in resistant strictures only removable stents (plastic) or the new biodegradable stents should be used.¹⁷ The patient with a stricture resistant to these treatments may require surgical resection. Such cases may be technically difficult as there has been transmural inflammation or perforation and an open approach is recommended. An algorithm is shown in Box 14.1.

Box 14.1 Algorithm for the management of recalcitrant peptic oesophageal strictures

1. Repeat graded dilatation

2. Injection of steroid (laser short stricture – Schatski ring)

3. Removable or biodegradable stent

4. Resect

Failed antireflux surgery

Failure may be due to persistence or recurrence of reflux symptoms, development of new symptoms or complications of the surgical procedure. The incidence of complications 17–19 is listed in Table 14.1. Complications and failures tend to occur soon after laparoscopic surgery and later after open surgery.

Table 14.1

Complications of antireflux surgery

Pneumothorax		2%
Paraoesophageal hernia		7%
Dysphagia – Slipped wrap, tight wrap, tight hiatus	early late	34% 6%
Perforated oesophagus		1%
Bloating/diarrhoea		30%

As described above, the success rate for fundoplication in controlling symptoms is in the region of 85% using accepted criteria for surgical success – volume reflux, failed medical treatment, etc.2–5 This implies that 15% fail and with time even successful initial surgery results in 25% of patients subsequently requiring regular medication.

It is now recognised that the type of wrap (partial or total) does not have a long-term effect on dysphagia, although total fundoplication may give better control of reflux symptoms whereas partial fundoplication has less dysphagia in the short term, although this difference disappears with time.^20,²¹

The results of 24-hour pH monitoring post-fundoplication are very interesting.²² The usual pattern is that there is no measurable reflux in spite of symptoms that seem classical of reflux. This relates to the observation that many patients undergoing revisional surgery for recurrent symptoms have a wrap that is in good position and is intact. The cause of the symptoms is therefore unclear and caution must be expressed on the successful outcome of re-operation. When the pH studies are positive and there is good symptom correlation and positive DeMeester score, this will usually be associated with wrap disruption and a better outcome can be expected with revision.

Management of failure after antireflux surgery

Failure will be discussed on the basis of the clinical problem that arises from ‘failure’. It should be recognised that some patients will actually have a persistence of symptoms that they had prior to surgery and that these symptoms were not actually related to their identified reflux. The importance of a careful history that takes into account the patient’s preoperative history, the symptoms with which they have re-presented and the time course of development of these symptoms cannot be overemphasised.

Recurrence of reflux symptoms

Recurrence of symptoms as described earlier is more common than generally recognised and a significant proportion of patients (25%) are back on medication post-surgery, although their symptoms are better controlled. It is wise to repeat the physiology studies in such patients to identify those who have true reflux, especially if considering revisional surgery.²³

The reasons for this recurrence of symptoms include wrap disruption, wrap slippage and migration into the chest. The latter is more frequent when at the first operation there was a large hiatal defect, which was not closed properly with sutures, and if the sac was not excised from the chest or the repair was associated with tension (Fig. 14.7). Closure of a large hiatus with removal of the sac can be difficult, especially when undertaken laparoscopically, and if problems arise this should lead to conversion to open surgery. An alternative is to close the hiatus with a mesh, preferably from behind the oesophagus. Good results have been reported using this technique but caution is needed as the mesh can erode the oesophagus and cause dense strictures at the hiatus.^23,²⁴ Removal will invariably require open surgery, usually via a thoracoabdominal incision with an oesophagogastrectomy. The use of man-made products to cover the hiatus adjacent to the oesophagus should be avoided if at all possible.

Figure 14.7 Laparoscopic view of wrap migrated into the chest.

Failure of the symptoms to improve after surgery questions the original diagnosis and may be associated with a sensitive oesophagus. It can be related to a ‘bad’ operation when the wrong part of the stomach is brought up (Fig. 14.8). This can produce the equivalent of a gastric band as used in bariatric surgery. Such patients often experience significant weight loss in the postoperative period and require revisional surgery.

Figure 14.8 Diagram of a ‘bad wrap’ where the greater curve is used rather than fundus.

Persistence of preoperative symptoms

Persistence of symptoms suggests that the original diagnosis was at fault. Review of preoperative 24-hour pH and oesophageal manometry together with a repeat study may reveal underlying motility disorders such as achalasia, lack of symptom correlation and low DeMeester scores. Such patients tend to be upright refluxers or if the original studies were negative may represent ‘functional heartburn’.²⁵ Taking down the wrap in the absence of dysphagia is unlikely to benefit the patient and is only a last resort. Other causes of pain must be sought and 24-hour pH study on and off full-dose medical treatment can be revealing.

Dysphagia

Dysphagia tends to occur early following laparoscopic surgery and late following open surgery. It can be an immediate complication in the first 24 hours and, if total, requires immediate return to theatre as the wrap may have slipped or twisted. Redoing the wrap or converting to a partial wrap will lead to resolution.

Dysphagia persisting beyond 2 months may be due to a pre-existing motility disorder, an over-tight wrap, scarring in the hiatus or over-tight repair of the hiatus. If in doubt, especially in cases where fluids only can be taken, early barium studies can be very helpful,²⁶ revealing migrated or slipped wraps or strictures at the hiatus (Figs 14.3 and 14.4). If the stenosis is thought to be due to stricturing of the hiatus then care should be taken with dilatation as the oesophageal wall can be compressed against a hard fibrous band and the result is invariably poor (Fig. 14.9). The solution to this is re-operation and incision of the fibrous hiatal ring to release the oesophagus. It is important to visualise and avoid injury to the inferior phrenic vein when undertaking this procedure. This procedure can be achieved laparoscopically in most cases.

Figure 14.9 Balloon dilatation of a tight hiatal ring demonstrating the resistant short stricture characteristic of this complication.

When dysphagia arises after a period of normal swallowing this may be due to wrap slippage or migration but may indicate the development of malignancy, especially if the operation was performed for Barrett’s oesophagus. The first investigation is endoscopy (Fig. 14.10), biopsy of any suspicious lesion and balloon dilatation of the wrap up to 20 mm with an endoscopic balloon dilatator. Barium studies and CT are also indicated as wrap migration into the chest may not be obvious on endoscopy (Fig. 14.11).

Figure 14.10 Retroflection and view of partial fundoplication on endoscopy.

Figure 14.11 Wrap migration into the chest.

The incidence of dysphagia is initially less when a partial wrap is performed, in contrast to a full 360° Nissen fundoplication. The use of a partial fundoplication is increasing, with no apparent reduction in efficacy in the control of reflux.^20,²¹ Certainly, in patients with pre-existing motility disorders such as scleroderma, or as part of surgery for achalasia, a partial wrap should be undertaken or dysphagia is likely.

Other symptoms

Gas bloat and change in bowel habit often indicate a functional gastrointestinal disorder that has either been exacerbated by the surgery or has arisen de novo. In some series the incidence of these complications can be as high as 60%, with half arising de novo. Such patients tend to be air swallowers and re-operation should be avoided at all costs and medical management undertaken. A not infrequent complication of surgery at the gastro-oesophageal junction is damage to the vagus nerve, especially the anterior vagus, which may influence gastric emptying. In such cases, poor emptying will precipitate GORD and will resolve with pyloric dilatation. Functional solid and liquid emptying studies will confirm this and should be undertaken if formal surgery on the stomach is considered. Trial of dilatation of the pylorus with a 20-mm balloon can be very instructive. Surgery in such cases would involve an antrectomy and Roux-en-Y anastomosis.

Revisional surgery following failed antireflux surgery

Before embarking on revisional surgery it is crucial to gather all the above information and have detailed consultations with the patient on what can be expected to be achieved. There are situations such as a tight hiatus, bad wrap or slippage into the chest (whether the whole stomach as a volvulus or a rolling component) in which there is either a significant risk of complications or a good chance of a successful outcome where surgery can be recommended. This includes missed motility disorders in which conversion to a partial wrap and myotomy, if appropriate, can improve symptoms.

Care should be taken in recommending re-operation in patients with persistent or recurrent symptoms in whom the investigations reveal an intact wrap in a good position and normal physiological studies. However, if re-operation is undertaken for the correct reasons then the outcome is generally good, although the outcome is worse if re-operation is for dysphagia.^12,²⁷

This should not be undertaken by an occasional surgeon or by one who is unfamiliar with entering the chest. The most frequently used procedure is revision of the fundoplication, either redoing a 360° wrap or converting it to a partial wrap, together with appropriate hiatal repair (Figs 14.12 and 14.13). This can be undertaken laparoscopically, especially if the previous procedure was laparoscopic, but should be converted to open surgery if any difficulty is encountered.

Figure 14.12 Revision with 360° wrap.

Figure 14.13 Revision with 270° wrap. Sutures between wrap and right crus are clearly visualised.

The principles are to take the whole wrap down and return the anatomy to normal prior to undertaking a re-fundoplication. This enables good visualisation of the crura and hiatus, and must be associated with mobilisation of the oesophagus while preserving the vagi. Great care must be undertaken to avoid opening the oesophagus and stomach, and placement of a small-diameter endoscope into the stomach enables accurate identification of the structures and also enables leak testing to identify small defects. A guide to the wrap is the previous sutures and lifting these with forceps will often reveal the wrap.

The author’s preferred approach is an open operation through a bilateral subcostal incision using an Omnitract retractor. If there are dense adhesions the oesophagus can be approached over the top of the left lobe of the liver, which can then be mobilised.

This approach makes reduction of a stomach that has slipped into the chest easier. The sac can be reduced by blunt dissection from the front of the hiatus, dividing the peritoneum with a diathermy point and extending this to the right and left, thus exposing the crura. The sac should be fully excised (Fig 14.14).

Figure 14.14 Dissection of sac and exposure of crura in revision of large hiatal defect.

Posterior repair of the hiatus can be achieved by placing non-absorbable sutures in the left crus and bringing them behind the oesophagus to place in the right crus. Usually, three well-placed sutures are enough to close the hiatus without tension. This invariably enables avoidance of the use of mesh. Full mobilisation of the fundus and short gastric vessels enables a loose 360° or 270° wrap to be performed. On completion of the wrap, it is sutured to the right crus with two or three non-absorbable sutures (Fig. 14.13).

In cases associated with gastric volvulus, the author adds a tube gastrostomy on the greater curve to fix the stomach in three places – gastrostomy, hiatus and duodenum. This avoids any need for a nasogastric tube and by leaving it in place for 3 weeks fixes the stomach and prevents any recurrence.

It is reasonable to have one attempt at repairing and revising a wrap. Further operations may well result in a devitalised gastro-oesophageal zone necessitating oesophagogastric resection via a left thoracoabdominal approach with a 25-mm stapled anastomosis at the aortic arch.²⁸ The gastric tube is based on the right gastroepiploic artery and should be no more than 3 cm wide. Unless there is previous peptic ulceration, a pyloroplasty is not needed and the addition of a wrap at the anastomosis does not protect from reflux. If the stomach is not suitable then a short-segment colonic interposition is an alternative approach. The functional result of such a procedure will never be ‘normal’.

In cases with a failed second operation, further surgery should only be undertaken if there are major nutritional or quality-of-life issues and a second specialist opinion should be sought. In such cases the options will rest between a subtotal gastrectomy and Roux-en-Y anastomosis if the problem is reflux and resection if a there is a persistent stricture that is leading to inability to eat.

Complex revisional surgery

Such patients fall into three distinct categories: those with oesophageal occlusion; those with a cervical oesophagostomy and gastrostomy following salvage surgery for failed resection for cancer or Boerhaave’s syndrome; and the late complications following previous surgery for oesophageal atresia. These cases must be managed in specialist units with full access to advanced imaging, thoracic, microvascular and upper GI surgeons, and full ITU back-up.

Oesophageal occlusion usually results from ingestion of caustic fluids with either complete occlusion of the oesophagus from the pharynx or a long tight complex stricture.²⁹ Rare causes include epidermolysis bullosa.³⁰ In such cases the pharynx is also involved with inflammation and fibrosis and patients are unable to swallow their own saliva and have permanent gastrostomy feeding. A colonic interposition between the pharynx and stomach or a jejunal loop can in many cases restore swallowing, although pharyngeal fibrosis inhibits the initiation of swallowing. If the stricture is high with a relatively normal oesophagus below, then a pharyngoplasty can help.

It is now recognised that in cases of failed oesophageal resection with tube necrosis of major anastomotic breakdown, resection of the conduit with cervical oesophagostomy and stapling off the distal end can be life saving. Similar oesophageal resection or exclusion can have benefits in Boerhaave’s syndrome with oesophageal necrosis and mediastinitis.³¹ Such patients require reconstruction to restore continuity when they have recovered. This can usually be achieved by the use of either a colonic interposition or a long ‘supercharged’ jejunal Roux-en-Y loop.^32,³³

If possible, the author prefers the colon in the first instance, using the transverse and descending colon based on the ascending branch of the left colic artery. An angiogram CT scan should be performed prior to surgery to ensure a good anastomosis at the splenic flexure. The colon can be brought up either substernally or subcutaneously in unfit patients. A sternal split is preferred together with clearing of the anterior mediastinum of fat and thymus, and mobilising the left brachiocephalic vein. This removes any compression on the conduit at the thoracic inlet, which is the major reason for failure. If a ‘blind’ retrosternal tunnel is used then the manubrium and sternoclavicular joint must be removed or compression and venous infarction will result. The author prefers not to use the right colon as it is more bulky and more difficult to straighten out at the hepatic flexure. A short length of terminal ileum should be used to anastomose to the oesophagus if the right colon has to be used.

In cases where the colon is unavailable, a supercharged jejunal loop can be used. In order to reach the neck, the proximal part of the loop will be ischaemic and requires ‘supercharging’ by a microvascular anastomosis between a mesenteric artery and usually the left internal mammary artery and a venous anastomosis to any suitable vein (Figs 14.15 and 14.16).

Figure 14.15 A long supercharged Roux loop showing the long mesentery and dusky appearance at the top end.

Figure 14.16 The same loop following microvascular anastomosis in the neck and the well-vascularised bowel.

In a series of 25 consecutive patients (seven supercharged jejunum and 18 colonic interpositions), 75% are able to maintain oral nutrition and 20% require nocturnal jejunostomy nutritional supplementation. Mortality was 4% but morbidity was 57%, emphasising the need to centralise such cases.

There are now an increasing number of patients in their thirties who had successful surgery for oesophageal atresia in infancy. This was invariably a colonic interposition via the left chest with multiple neck and thoracic operations. In such cases presenting with difficulty in swallowing, it is crucial to investigate thoroughly prior to considering any reconstruction to determine accurately the site of obstruction. This can be at the root of the neck, which bulges on swallowing, in redundant colonic loops in the chest, or at the level of the diaphragm where a fibrotic stricture causes a mechanical obstruction and proximal dilatation. Investigation includes endoscopy, barium studies and contrast-enhanced CT scanning (Fig. 14.17).

Figure 14.17 A contrast swallow demonstrating redundant colonic loops in the left chest.

The most common causes for difficulty in swallowing are redundant colonic loops and strictures at the diaphragm.^33,^34,³⁵ A long left thoracoabdominal approach gives the best access, and it is crucial to identify the vascular pedicle at an early stage and preserve it at all costs. The diaphragm is split to the hiatus, the cologastric anastomosis taken down, redundant colon excised with a harmonic scalpel keeping to the bowel wall, preserving the vascular arcade and re-anastomosis of the colon to stomach. The diaphragm is closed without any compression at the new pseudo-hiatus. Such patients will require jejunostomy feeding for some time.

If the problem occurs in the neck, excision of the manubrium and left costoclavicular joint and pexy of the redundant colon may help. The colon can be resected in the left chest with primary anastomosis, again taking extra care to preserve the vascular arcade.

It cannot be overstressed that a conservative approach is best as such patients have a life expectancy of many decades, and a realistic discussion regarding outcome should take place at the first consultation. Taking such an approach can lead to good results and quality of life.