Nota : Para ver los artículos completos es necesario suscribirse aquíLos hipervínculos no están operativos |
![[Feedback]](feedbackACT.gif){kind=small}
![[For Subscribers]](subscriberACT.gif){kind=small}
![[Archive]](archiveACT.gif){kind=small}
![[Search]](searchACT.gif){kind=small}
![[Contents]](tocACT.gif){kind=small}
|
Annu. Rev. Med. 1999. 50:469-506.
Tom R. DeMeester, MD,1, 2 Jeffrey H. Peters, MD,1 Cedric G.
Bremner, MD,1 and Parakh Chandrasoma, MD3
1Department of Surgery, 2Department of
Cardiothoracic Surgery, and 3Department of Pathology, University of Southern
California School of Medicine, Los Angeles, California 90033-4612; e-mail: demeester@surgery.hsc.usc.edu
KEY WORDS: barrier, reflux composition, cardiac-type mucosa, carditis, intestinal metaplasia
 |
ABSTRACT |
|---|
 Top NextLiterature Cited |
|---|
This chapter reviews the biology of gastroesophageal reflux disease, relating pathophysiology to medical and surgical therapy. Various definitions of the disease are presented and workable criteria are developed to identify patients with the disease. The central importance of the lower esophageal high-pressure zone as a barrier to reflux is emphasized, along with an analysis of its biomechanical alteration in disease. The composition of the refluxed gastric juice is characterized in regard to its potential for mucosal injury. Evidence is provided that cardiac-type mucosa is an acquired sequel to acid-induced squamous mucosal injury in the terminal esophagus. A hypothesis regarding the process of intestinalization of cardiac-type mucosa to form Barrett's esophagus is presented. An integrated concept of the pathophysiology of gastroesophageal reflux disease is constructed. Practical concepts regarding the treatment of gastroesophageal reflux disease are developed, based on a review of studies on the natural history of the disease and the long-term outcome of therapy.
|
DEFINITION OF GASTROESOPHAGEAL REFLUX DISEASE |
|---|
| TopPreviousNextLiterature Cited |
|---|
Gastroesophageal reflux disease (GERD) is typically manifested by heartburn and regurgitation and commonly by atypical symptoms of chronic cough, wheezing, hoarseness, and chest pain. Diagnosis can be a challenge because there is no universally accepted definition for the disease. Over the years, the definition has varied according to the available therapy. For example, when symptomatic therapy was the only treatment available, the presence of symptoms defined the presence of the disease, and a cure was declared when the symptoms were abated by antacid or medication to suppress acid secretion. Once endoscopy became available, the presence of endoscopic esophagitis defined the disease, and a cure occurred when the esophagitis was healed. After antireflux surgery was perfected, the measurement of increased esophageal acid exposure defined the disease, and when the exposure was reduced to normal by surgery the disease was considered cured.
With progressive understanding of the pathophysiology of GERD, clinical investigators have learned that a symptomatic definition presents problems. First, ascribing typical symptoms such as heartburn and regurgitation to gastroesophageal reflux in the absence of some objective findings can be deceiving. Other disorders can cause similar symptoms, such as achalasia, diffuse spasm, pyloric stenosis, cholelithiasis, gastritis, gastric or duodenal ulcer, and coronary artery disease. Second, a symptomatic definition that includes atypical symptoms loses so much specificity that its usefulness is limited. Furthermore, when both atypical and typical symptoms are present, atypical symptoms—such as choking or shortness of breath—may be so devastating that they override the typical symptoms. Third, the imprecision of a symptomatic definition of the disease encourages the use of "shotgun therapy" rather than a more detailed evaluation of poorly articulated patient complaints.
There are also problems with defining GERD by the presence of endoscopic esophagitis. First, this definition assumes that asymptomatic patients with esophagitis have excessive regurgitation of gastric juice into the esophagus. In the presence of typical symptoms, this assumption is almost always correct, but when symptoms are absent or atypical the esophagitis can result from other etiologies, the most common being unrecognized chemical injury from drug ingestion. A second problem is that this definition leaves undiagnosed a large number of patients who have typical symptoms but do not have endoscopic esophagitis. Third, esophagitis is a mucosal injury that can result from GERD, but is not synonymous with the presence of the disease. Defining a disease by its complications is not a workable solution.
A third approach to defining GERD is to measure the amount of esophageal exposure to gastric juice. This measurement requires prolonged monitoring of the esophagus body for acid, a tag indicating the presence of gastric juice . Most normal individuals experience short episodes of esophageal acid exposure, usually after meals; the disease occurs when esophageal acid exposure exceeds normal. The unit of measurement is either the percent time the esophageal pH is below 4 over a 24-h recording period or a composite number called the 24-h pH score, which incorporates not only the percent time the pH is below 4 but also how often this occurs during a 24-h period, the body position in which it occurs (upright or supine), and duration of the episodes . Most importantly, 24-h esophageal pH monitoring provided a method to quantitate the disease and made available an objective measurement to assess the effect of treatment .
Currently, for the clinician, the most workable requirement to identify patients with GERD is the presence of typical symptoms and at least one piece of objective evidence of reflux, such as an abnormal esophageal acid exposure on 24-h pH monitoring, visible esophagitis (i.e. linear erosion or cobble-stoning of the lower esophagus on endoscopy), or a biopsy showing inflammatory cells infiltrating the mucosa on histology. For patients with atypical symptoms, the clinician should be more conservative and require two pieces of objective evidence. In the patient with atypical respiratory symptoms, endoscopic laryngitis can be counted as one piece of evidence .
|
THE BARRIER TO REFLUX |
|---|
|
TopPreviousNextLiterature Cited |
|---|
The gastrointestinal tract is a continuous, hollow tube. Its functions are the ingestion and digestion of food, the absorption of chemical energy, and the elimination of residue. These functions are performed in separate compartments whose boundaries differ from the customary anatomic divisions of the gastrointestinal tract. Each compartment has a pumping mechanism to propel contents into a receptacle or reservoir portion; a sphincter to separate the pump from the reservoir; and the ability to maintain within the reservoir a distinct chemical, enzymatic, and pH environment appropriate to its function. In the most proximal compartment, the tongue and pharynx function as a pump; the upper esophageal sphincter, soft palate, and epiglottis as valves; and the striated muscle portion of the upper esophagus as a receptacle. In the second compartment, the smooth muscle portion of distal esophagus, characterized by peristaltic contractions of high amplitude, pumps food through a valve (the lower esophageal sphincter) into the gastric fundus, which acts as a reservoir. In the third compartment, the antrum behaves as a pump, propelling chyme through a valve (the pyloris) into a receptacle, the duodenum. Similarly, the small intestine pumps its contents through the ileocecal valve into a capaciance organ, the cecum. An important principle is that the breakdown of function in one compartment of the gastrointestinal tract tends to produce secondary effects in the proximal compartments rather than in the distal compartments. Thus, problems originating in the stomach commonly cause symptoms in the esophagus or symptoms referable to the pharyngeal and laryngeal area. This concept is important in understanding the pathophysiology of GERD and structuring a rational approach to its therapy.
In humans, the barrier that confines the gastric environment to the stomach is the lower esophageal "sphincter." It has no anatomical landmarks but can be identified by a rise in pressure over gastric baseline pressure as a pressure transducer is pulled from the stomach into the esophagus (Figure 1). This high-pressure zone is normally present except in two situations: (a) After a swallow, it is momentarily dissipated to allow passage of food into the stomach; (b) when the fundus is distended with gas, it is eliminated to allow venting of the gas. The common denominator for virtually all episodes of gastroesophageal reflux, whether physiologic or pathologic, is the loss of the normal resistance to the flow of gastric juice from an environment of higher pressure, the stomach, to an environment of lower pressure, the esophagus. In severe disease, this is usually caused by a permanently nonexistent or reduced high-pressure zone. In early disease or in normal subjects, the loss is transient .
|
Three characteristics of this high-pressure zone, commonly called the lower esophageal sphincter, maintain its function as a barrier to intragastric and intra-abdominal pressure challenges (Figure 1). Two of these characteristics—the overall length of the zone and its pressure measured at the respiratory inversion point—work together and depend on each other for proper sphincter function. The resistance of the lower esophageal sphincter to the flow of gastric juice into the esophagus is a function of both its pressure and the length over which the pressure is exerted . The shorter the overall length of the high-pressure zone, the higher the pressure must be to maintain sufficient resistance to remain competent (Figure 2). Consequently, a normal sphincter pressure can be nullified by a short overall sphincter length. Furthermore, as the stomach fills, the length of the sphincter decreases, rather like the neck of a balloon shortening as the balloon is inflated. If the overall length of the sphincter is abnormally short when the stomach is empty, then with minimal gastric distention there will be insufficient sphincter length for the existing pressure to maintain sphincter competency, and reflux will occur. It is now possible to measure the integrated effects of radial pressure exerted over the entire length of the high-pressure zone, forming a three-dimensional computerized image of the sphincter. The volume of this image reflects the sphincter's resistance and is called the sphincter pressure vector volume (Figure 3). A calculated volume less than the fifth percentile of normal subjects is an indication of a permanently defective sphincter.
|
|
The third characteristic of the lower esophageal sphincter is its position, in that a portion of the overall length of the high-pressure zone should be exposed to positive intra-abdominal pressure. This portion of the high-pressure zone is commonly called the abdominal length of the sphincter (Figure 1). During periods of increased intra-abdominal pressure, the resistance of the lower esophageal sphincter would be overcome if the abdominal pressure were not applied equally to the high-pressure zone and stomach . Think of sucking on a soft soda straw immersed in a bottle of coke; the hydrostatic pressure of the fluid and the negative pressure inside the straw due to sucking cause the straw to collapse instead of allowing the liquid to flow up the straw in the direction of the negative pressure. If the abdominal length is inadequate, the sphincter cannot collapse in response to applied intra-abdominal pressure, and reflux is more liable to result.
If the high-pressure zone has an abnormally low pressure, a short overall length, or minimal exposure to the abdominal pressure environment in the fasting state, the result is a permanent loss of lower esophageal sphincter resistance and the unhampered reflux of gastric contents into the esophagus throughout the circadian cycle. This is known as a permanently defective sphincter and is identified by one or more of the following characteristics: a high-pressure zone with an average pressure of less than 6 mm Hg, an average overall length of 2 cm or less, and/or an average length exposed to the positive pressure environment of the abdomen of 1 cm or less . Compared with normal subjects, these values are below the 2.5 percentile for each parameter. The most common cause of a permanently defective sphincter is an inadequate pressure but the efficiency of a sphincter with a normal pressure can be nullified by an inadequate abdominal length or an abnormally short overall length .
For the clinician, the finding of a permanently defective sphincter has several implications. Foremost, it is almost always associated with esophageal mucosal injury and predicts that the patient's symptoms will be difficult to control medically. Surgery is likely to be needed for consistent, long-term symptom control. It is now accepted that a permanently defective sphincter is irreversible, even when the associated esophagitis is healed. A permanently defective sphincter is commonly associated with reduced esophageal body function, and if the disease is not brought under control, the progressive loss of effective esophageal clearance can lead to severe mucosal injury, repetitive regurgitation, aspiration, and pulmonary failure .
Transient loss of the high-pressure zone is usually caused by a functional problem of the gastric reservoir. Ingestion of excessive air or food can result in gastric dilatation and, if the active relaxation reflex has been lost, increased intragastric pressure. When the stomach is distended, the vectors produced by gastric wall tension pull on the gastroesophageal junction with a force that varies according to the geometry of the cardia; that is, the forces are applied more directly when a hiatal hernia exists than when a proper angle of His is present . These forces pull on the terminal esophagus, causing it to be "taken up" into the stretched fundus, thereby reducing the length of the high-pressure zone ). This process continues until a critical length is reached, usually about 1 to 2 cm, when the pressure drops precipitously and reflux occurs (Figure 4). If only the pressure and not the length of the high-pressure zone is measured, as with a Dent sleeve , this event will appear as a spontaneous dissipation or "relaxation" of the high-pressure zone.
|
The mechanism by which gastric distention contributes to shortening of the high-pressure zone, so that its pressure drops and reflux occurs, provides a mechanical explanation for transient relaxations of the lower esophageal sphincter without invoking a neuromuscular reflex. Rather than a transient muscular relaxation, there is a mechanical shortening of the high-pressure zone, secondary to progressive gastric distention, to the point where competence is lost. Consequently, non–swallow-induced relaxations of a normal high-pressure zone are inappropriately termed transient lower esophageal sphincter relaxations; rather they should be called transient lower esophageal sphincter shortenings. These transient sphincter shortenings occur in the initial stages of GERD and are the mechanisms for the early complaint of excessive postprandial reflux.
After gastric venting, the length of the high-pressure zone is restored and competence returns until distention again shortens it and encourages further venting and reflux. This sequence results in the common complaints of repetitive belching and bloating in patients with GERD. The increased swallowing frequency seen in patients with GERD contributes to gastric distention and is due to their repetitive ingestion of saliva in an effort to neutralize the acid refluxed into their esophagus . Thus, GERD may begin in the stomach, secondary to gastric distention caused by overeating and the increased ingestion of fried foods, which delay gastric emptying . Both characteristics are common in Western society and may explain the high prevalence of the disease in the Western world.
If mechanical forces set in play by gastric distention are important in pulling on the terminal esophagus and shortening the length of the high-pressure zone, then the geometry of the cardia, i.e. the presence of a normal acute angle of His or the abnormal dome architecture of a sliding hiatus hernia, should influence the ease with which the sphincter is pulled open. There is a close relationship between the degree of gastric distention necessary to overcome the high-pressure zone and the morphology of the cardia . Greater gastric dilatation, as reflected by a higher intragastric pressure, was necessary to "open" the sphincter in patients with an intact angle of His compared with those who have a hiatus hernia (Figure 5). This is what would be expected if the high-pressure zone were shortened by mechanical forces and explains why a hiatal hernia is often associated with GERD.
|
In normal subjects, almost all reflux episodes are precipitated by belching. In patients, belching remains an important but decreasing cause of reflux as the severity of the disease, reflected by the grade of esophagitis, worsens. Activities that produce a pressure gradient across the diaphragm, like coughing, sniffing, or straining, become increasingly important in precipitating reflux as the disease progresses. In patients with severe grades of esophagitis, episodes of acid reflux occurred spontaneously, reflecting that the sphincter is permanently defective in its resting state and that there is a persistent loss of the barrier.
Reflux episodes associated with belching are by inference caused by gastric distention and are responsible for increased esophageal acid exposure in patients with early or less mucosal disease. In this situation, the loss of the barrier is transient. Mucosal damage caused by repetitive exposure to gastric juice results in an inflammatory injury of underlying muscle, which leads to a permanently defective high-pressure zone, initially due to the loss of abdominal length and eventually due to the loss of pressure and overall length . Subsequent inflammation of the esophageal body results in the loss of its clearance ability, leading to prolonged esophageal exposure to gastric juice . This signals the presence of advanced disease and places the patient at risk for Barrett's metaplasia, stricture formation, and aspiration.
|
COMPOSITION OF THE REFLUXED GASTRIC JUICE AND MUCOSAL INJURY |
|---|
| TopPreviousNextLiterature Cited |
|---|
The failure of the antireflux barrier leads to increased esophageal exposure to gastric juice, which can cause injury to the esophageal mucosa and/or respiratory epithelium along with the loss of esophageal and lung function owing to subsequent repair and fibrosis. The injury and repair in the esophagus can be visualized endoscopically; the former appears as a linear or interlacing ulceration called cobble-stoning of the mucosa, and the latter appears as a stricture and/or metaplasia of the mucosa. If a biopsy shows columnarization of the squamous epithelium and intestinalization of the columnar epithelium, the condition is called Barrett's esophagus. The prevalence and severity of these complications are related to the presence of a permanently defective high-pressure zone (Figure 6) and the composition of the material refluxed (Figure 7) . The observation that a permanently defective high-pressure zone occurs in 42% of patients without mucosa injury suggests that the amount of esophageal exposure to gastric juice can be reduced by the effective clearance of a vigorously contracting esophageal body. Eventual failure of the esophageal body allows prolonged exposure of the esophageal mucosa to gastric juice, resulting in further mucosal injury and deterioration of esophageal contractility, as is commonly seen in patients with strictures and Barrett's metaplasia . The loss of esophageal contractility also increases the potential for regurgitation and aspiration.
|
|
The composition of the refluxed gastric juice is particularly important in the development of complications. Acid alone is not very damaging in physiologic concentrations, but in high concentrations, as seen in patients with hypersecretion and particularly in the presence of pepsin, the incidence of mucosal damage is increased. Similarly, the reflux of duodenal juice alone does little damage but is particularly noxious when combined with gastric acid (Figure 7) .
Bile acids are considered the most damaging component of duodenal juice. When monkeys underwent a Roux-en Y antrojejunostomy without a vagotomy and resection of their lower esophageal sphincter with esophagogastrostomy, only 1 of 7 developed esophagitis, even though most incurred a perforated jejunal ulcer from the high acid secretion stimulated by the construction of a Mann Williams preparation. In contrast, a simple cholecystogastrostomy and resection of the lower esophageal sphincter with esophagogastrostomy resulted in all seven monkeys developing severe esophagitis and stricture, and one developed a perforated esophageal ulcer . Several other studies with different animals have established the principle that severe mucosal injury results from exposure to a combination of acid and bile , and this principle appears applicable to humans . When patients with symptoms of heartburn and regurgitation, increased esophageal acid exposure on 24-h pH monitoring, and endoscopic esophagitis were treated with either a Nissen fundoplication or a Roux-en Y diversion of duodenal juice, their symptoms improved and their esophagitis healed. The only difference between the two forms of therapy was that following the Roux-en Y, esophageal acid exposure remained increased, whereas after the Nissen fundoplication it returned to normal. In fact, after the duodenal diversion, the time the esophagus was exposed to the pH range of 3–4 was greater than before surgery. Despite this increased esophageal acid exposure, the patients' symptoms improved and the esophagitis surprisingly healed. These studies indicate that, in GERD, duodenal juice is an important component in the reflux that causes damage to the esophageal mucosa.
Continuous esophageal aspiration studies with analysis of non-pooled samples have confirmed the presence of bile acid at toxic levels in the esophagus of patients with GERD compared with normal subjects (Figure 8). Patients with erosive esophagitis showed a tenfold increase in bile acid concentration compared with those who had no injury. Of importance is that esophageal acid exposure was similar in both groups. Patients with strictures or Barrett's esophagus had more esophageal acid exposure than the other groups but had bile acid concentrations significantly greater than those in patients with erosive esophagitis. These findings reinforce the concept of an acid and bile synergism in causing mucosal injury.
|
The development of the Bilitec probeTM to monitor bilirubin as a marker for duodenal juice has greatly simplified the clinical study of the prevalence of duodenogastroesophageal reflux . Furthermore, experience with the device has shown that the exposure time above the absorbance of 0.2 correlates, as well, to bile acid exposure. In large population studies of patients with GERD, 33% refluxed only gastric juice and 57% gastroduodenal juice; 10% had sufficient duodenogastric reflux to neutralize all gastric acid so that only duodenal juice was detected in the esophagus. Eighty percent of the patients who refluxed gastroduodenal juice had mucosal injury, compared to 54% of those who refluxed only acid and 10% who refluxed only duodenal juice (Figure 9) .
|
The effect of duodenogastric reflux is to elevate the pH of gastric juice (Figure 10). The height of the elevation depends on the baseline level of gastric pH, which can vary depending on whether the patient is taking acid-suppression therapy. Elevating the pH of gastric juice has four known effects. First, when the pH exceeds 4, heartburn and regurgitation diminish . Second, when the pH enters the 3–5 range, it stimulates phenotypic differentiation of the cardiac-type mucosa toward intestinalization, with proliferation of the mucosal glandular cells . Third, when the pH reaches 4.5, bacteria normally present in the mouth begin to grow in the stomach, and bile acids can be deconjugated to release the more noxious free bile acid. Fourth, when the pH enters the 3–6 range, bile acids become soluble, and a portion dissociate into their ionized salt and free H+ while the remainder persist as a lipophilic, non-ionized acid. As the pH approximates 7, over 90% of bile acids become soluble and completely ionized. Acidification of bile to a pH below 2 results in an irreversible bile-acid precipitation. Consequently, under normal physiologic conditions, bile acids in the stomach precipitate and have minimal effect in an acid gastric environment. On the other hand, in a more alkaline gastric environment—such as occurs with excessive duodenogastric reflux or acid-suppression therapy, or after vagotomy—bile acids remain in solution and are only partially dissociated. When nondissociated, nonpolar bile-acid molecules reflux into the esophagus, they can enter the mucosal cell. Once in the cell, where the pH is 7, they become completely dissociated into polar ions and are trapped intracellularly in concentrations of up to seven times the luminal concentration . In the cell, bile acid salts can at low concentrations impair mitochondrial function; at high concentrations they become cytotoxic and function as co-mutagens or likely direct mutagens .
|
If soluble bile acids are to remain innocuous in a patient with chronic reflux managed by acid-suppression therapy, they must remain completely ionized. This requires that a gastric pH of 6–7 be maintained 24 hours a day, 7 days a week, for the patient's lifetime. This is not only impractical but probably impossible without very high doses of medication. Insufficient medication allows the pH to drift down to 4–5 and cause cellular mucosal damage while the patient remains relatively asymptomatic (Figure 11) (48A, 48B).
|
The injury can result in mild to erosive esophagitis, ulceration, stricture, or the development of a columnar lined esophagus with intestinal metaplasia, i.e. Barrett's esophagus , the incidence of which has increased progressively since 1986 (Figure 12).
|
Barrett's esophagus initiates a sequence of mucosal changes that can ultimately lead to esophageal adenocarcinoma. It is unknown whether the movement toward cancer is caused by mitogenesis secondary to chronic mucosal injury or by mutagenesis from exposure to a direct mutagen.
Throughout the 1950s and 1960s, complications of Barrett's esophagus were mainly acid-related, i.e. inflammation, ulceration, and stricture formation. There were few reports of adenocarcinoma of the esophagus and most authors believed that when adenocarcinoma did occur it was a gastric cancer that had crept up into the esophagus. Since the 1970s, when potent acid-suppression therapy became available, the acid complication of Barrett's esophagus became less common and the malignant complication more common . This has raised the suspicion that altering the gastric pH by acid-suppression therapy has encouraged the emergence of Barrett's esophagus and its malignant complications . In a cohort of 9928 patients prescribed cimetidine and followed for 8 years, the ratio of observed to expected death from adenocarcinoma of the cardia and esophagus was 3.7 to 1 (p < 0.05) . In the 1990s, convincing evidence emerged of an explosion of Barrett's esophagus (Figure 12) and adenocarcinoma of the esophagus (Figure 13) that has yet to be explained . Several studies are under way to determine whether bile acids, now solubilized in gastric juice owing to the widespread use of acid-suppression therapy, are mutagens. Recent cell culture studies indicate that repetitive, short exposure of cells to bile salts at a pH of 5 increases the mutation frequency without altering the growth curve of the cultured cells . If bile salts are demonstrated to contribute to the development of malignancy, then early surgical intervention to reestablish a lost antireflux barrier or prevent its transient shortening with gastric dilatation should be encouraged .
|
|
CARDIAC-TYPE MUCOSA IS AN ACQUIRED SEQUEL OF SQUAMOUS MUCOSAL INJURY |
|---|
| TopPreviousNextLiterature Cited |
|---|
Manometric recordings of the gastroesophageal junction show three distinct physiological regions. The first region, located proximal to the gastroesophageal junction, has a subatmospheric resting baseline pressure with swallow-induced peristaltic contractions. It corresponds to the tubular esophagus and is normally not exposed to gastric juice. The second region, located in the vicinity of the gastroesophageal junction, is a high-pressure zone 3–4 cm long. It has an average pressure of 15 mm Hg above gastric baseline pressure, which relaxes on swallowing. It is normally not exposed to gastric juice and is referred to as the lower esophageal sphincter. The third region is distal to the gastroesophageal junction and has a resting baseline pressure equal to the supra-atmospheric intra-abdominal pressure. It corresponds to the cardiac zone of the stomach and is normally intermittently exposed to gastric juice .
The absence of dependable anatomical, endoscopic, or histological landmarks that correlate with the lower esophageal high-pressure zone, and the fact that a sphincter cannot be recognized either during endoscopy, at surgery, or on gross examination of surgical specimens, have contributed to the mystery regarding pathologic changes within the region of the gastroesophageal junction. To add further to the mystery, the squamocolumnar junction normally lies within the high-pressure zone but does not correlate with the gastroesophageal junction, and it migrates cephalad in patients with GERD . The extent of the cephalad migration correlates with the severity of disease; Barrett's esophagus represents the extreme of this migration . As a consequence of these ill-defined and moving landmarks, the pathology and pathophysiology of the gastroesophageal junctional zone have been an enigma for many years.
Hayward stated that the lower 1–2 cm of the esophagus is normally lined by metaplastic mucus secreting columnar epithelium that has the ability to resist acid-peptic digestion. He argued that this mucosa prevents squamous epithelial digestion by providing a buffer between it and the acid-pepsin–producing fundic mucosa. Although Hayward's description placed this cardiac mucosa in the lower esophagus, it has over time (without any logical or scientific basis) come to be regarded as the cardiac zone of the stomach. Recently, Chandrasoma reviewed the histology of the gastroesophageal junction in a large number of autopsies whose medical records did not mention GERD. He reported that in the vast majority of children and adults under age 20, the squamous epithelium transitioned directly with the oxyntic mucosa of the gastric fundus. Cardiac mucosa appeared in specimens from patients older than 20, but its length was almost always less than 1 cm. However, in a significant number of individuals in this older group, the squamous epithelium transitioned directly to oxyntic fundic mucosa without intervening cardiac mucosa .
Several authors have evaluated this finding further . Öberg et al obtained endoscopic biopsies above, at, and below the gastroesophageal junction identified by the proximal extent of the rugal folds in 334 consecutive patients with symptoms of foregut disease, who had no prior history of gastric or esophageal surgery or endoscopic evidence of a columnar-lined esophagus. In 88 of the 334 patients (26%), biopsy revealed no cardiac-type mucosa. The absence of cardiac-type mucosa was strongly associated with the absence of the hallmarks of GERD (Table 1) . In a similar study by Speckler's group, multiple biopsies of an endoscopically normal-appearing gastroesophageal junction also showed cardiac-type mucosa in only a minority of patients.
|
Based on these investigations, the traditional view that the anatomical gastric cardia is lined by cardiac-type mucosa may not be correct; rather, cardiac-type mucosa may be a normal sequel to squamous mucosal injury and an early sign of GERD.
Our interpretation of these findings is that in the normal state, the squamous epitheleal lining of the lower esophagus and high-pressure zone changes abruptly to oxyntic mucosa of the gastric fundus without a "gastric cardia" mucosa . We hypothesize that since oxyntic fundic mucosa is not affected by gastric juice, cardiac-type glandular mucosa must come from exposure of the esophageal squamous mucosa to gastric juice. This process probably represents a change in the direction of differentiation of the germinative cells of the squamous epithelium towards columnar secreting cells .
Exposure of squamous mucosa to gastric juice is most likely to occur postprandially, when the stomach is distended and the squamous-lined distal esophagus is taken up by the expanding fundus . When limited to physiologic exposure of gastric juice, the cardiac mucosa has a chance to heal, with progressive oxyntic transformation of the mucous glands within the cardiactype mucosa. This produces a glandular mucosa that is similar to fundic mucosa in that it contains both parietal and mucous cells and, judging by the degree of inflammation compared with cardiac-type mucosa, is more resistant to damage by refluxed gastric acid. Of importance is that intestinal metaplasia rarely occurs within fundic mucosa as a result of exposure to gastric juice. Oxyntic transformation of cardiac-type mucosa protects against the development of intestinal metaplasia and thus is probably a beneficial change .
In nearly all circumstances, cardiac-type mucosa when present shows inflammatory and reactive changes referred to as carditis . This inflamed cardiac-type mucosa is the only mucosal type that progresses to intestinal metaplasia. When this occurs, the hypertrophic cardiac mucous cells develop acid rather than neutral mucin and take on a positive stain with Alcian blue. Subsequently, well-formed goblet cells appear. Once intestinalized, the cardiac-type mucosa seems to have an increased ability to withstand damage by refluxed gastric juice, as biopsy commonly shows little or no inflammation. As opposed to the fundic-like mucosal transformation, the development of intestinal metaplasia within cardiac-type mucosa is considered a detrimental change, since this mucosa can progress to dysplasia and adenocarcinoma ).
|
THE PROCESS OF INTESTINALIZATION OF CARDIAC-TYPE MUCOSA |
|---|
| TopPreviousNextLiterature Cited |
|---|
The early literature regarded Barrett's esophagus as a congenital disorder, and not until 1959 had sufficient clinical evidence accumulated to suggest that it was an acquired abnormality caused by GERD . The acquired etiology was confirmed experimentally in the landmark report of Bremner et al in 1970 . In 1976, it was recognized that three different types of glandular epithelium could line the lower esophagus in patients with Barrett's esophagus : gastric-fundic–type mucosa containing chief and parietal cells; junctional-cardiac-type mucosa containing mucous cells and no chief and parietal cells; and an intestinal type of columnar epithelium with a villiform surface, mucous glands, Alcian-blue-staining goblet cells, and no parietal or chief cells. When present, the intestinalized columnar mucosa was always most proximal and gastric-fundic–type mucosa most distal, with junctional cardiac-type mucosa interposed between them. To understand the development of Barrett's esophagus, it is critical to realize that attempts to differentiate cardiac-type mucosa of the gastric cardia from non-intestinalized columnar mucosa that can occur higher in the esophagus have led to confusion. Today, most pathologists agree that the two are indistinguishable and that in both areas intestinal metaplasia can occur.
In the 1980s, it became evident that the malignant transformation that complicated Barrett's esophagus occurred in the intestinalized cardiac-type mucosa . Once this association was recognized, the term Barrett's esophagus was applied only to an esophagus lined by intestinalized cardiac-type mucosa. The length of intestinalized mucosa could be as short as a few millimeters . In other words, a patient is considered to have Barrett's esophagus if any amount of cardiac-type mucosa in the lower esophagus has histologic evidence of goblet cells. The length of the intestinalized mucosa could be limited to the gastroesophageal junction or the lower 3 cm of the esophagus, or it could involve the esophageal body.
The observation that some patients have cardiac-type mucosa without intestinalization suggests that intestinalization requires a specific condition or stimulus. In other words, the development of the mucosal change typical of Barrett's esophagus is a stepwise process. In the first step, cardiac-type mucosa forms from the squamous epithelium as a normal response to acid exposure . Under the proper luminal conditions and stimuli, intestinalization of the cardiac-type mucosa follows . This process may take a period of time. Indeed, in children under age five who acquire a columnar-lined esophagus, intestinal metaplasia is rare ; goblet cells begin to appear only later. Clinical studies have shown a time lag of five to seven years after the onset of reflux symptoms for intestinalized cardiac-type mucosa to appear in adults . In support of the two-step concept, Hamilton & Yardley reported the development of Barrett's mucosa in the esophagus above the anastomosis in three patients after esophagastrectomy. In two, squamous epithelium progressed to cardiac-type mucosa with subsequently intestinal metaplasia of the cardiac-type mucosa over 6.3 and 10 years, respectively .
Öberg et al have shown that the prevalence of intestinalization of cardia-type mucosa is dependent on the status of the lower esophageal sphincter and the severity of esophageal acid exposure. In a small percentage of patients with a normal high-pressure zone, intestinal metaplasia develops in an endoscopically indistinguishable segment of cardiac-type mucosa at a normal-appearing gastroesophageal junction. As the high-pressure zone weakens, cardiac-type mucosa develops within the distal 3 cm of the esophagus, i.e. within the vicinity of the high-pressure zone, and a higher percentage of patients become intestinalized. With time and deterioration of the high-pressure zone, cardiac-type mucosa develops in the esophageal body, and almost all patients show intestinalization (Figure 14). At each step, the patients exhibit more profound hallmarks of GERD, including decreasing length and pressure of the lower esophageal high-pressure zone, increasing esophageal acid exposure, and the eventual loss of esophageal contraction amplitude (Table 2). The compromised sphincter probably results from inflammatory changes in the muscularis propria of the high-pressure zone secondary to increased esophageal acid exposure. The same process causes the loss of esophageal body contractility classically seen in patients with long segments of Barrett's esophagus .
|
|
Recent studies have hypothesized that intestinalization of the cardiac-type mucosa is initiated by possible exposure to pH 3–5 resulting from the interaction of the refluxed gastric juice, duodenal juice, and saliva at the squamocolumnar interface. Exposure to this pH range has been shown to encourage the phenotypic expression of intestinalization by columnar epithelial cells. It is hypothesized that as the length of intestinalized cardiac-type mucosa increases, the more distal portion is exposed to more acid and less saliva (and consequently less exposure to pH 3–5); it then reverts to cardiac-type mucosa by losing its intestinal characteristics and may eventually become fundic mucosa ). This may explain the observation that intestinal metaplasia is found consistently in the proximal portion of a long segment of Barrett's metaplasia, with gastric-fundic–type mucosa most distal and cardiac-type mucosa interposed between them .
In contrast to this stepwise extension theory is the common belief that Barrett's esophagus develops rapidly to its full extent with no subsequent change. This concept is based largely on a study in which all patients had esophageal columnar lining of 3 cm or more . Twenty-one patients with Barrett's esophagus were followed for a mean of 7.3 years. The average initial length of columnar epithelium was 8.29 cm and the final length was 8.33 cm, not significantly different. The strongest evidence that the extent of Barrett's esophagus did not change significantly over time was that the mean length of columnar epithelium was similar in all age groups. This finding would be expected, since patients with long segments of columnar lining uniformly have profound defects in the antireflux mechanism, resulting in very high esophageal exposure to gastric contents. At this stage of disease, further deterioration of the antireflux mechanism or greater esophageal exposure to gastric juice would not be expected. Rather than implicating the sudden occurrence and unchanging extent of Barrett's metaplasia, the findings may be the consequence of the normal gradient between the positive pressure environment in the stomach and the negative pressure environment in the thoracic esophagus. When the antireflux barrier completely deteriorates, gastric juice flows to the point of lowest esophageal pressure, which is in the mid-thoracic esophagus . Under these conditions, intestinalized cardiac-type mucosa could appear within the lower third to lower half of the esophageal body without any apparent subsequent change in length over time. This is the third step in the stepwise extension theory. The second step is short-segment Barrett's, i.e. less than 3 cm in length, and the first is microscopic intestinalization of cardiac-type mucosa below an endoscopically normal-appearing gastroesophageal junction.
A controversial issue is whether intestinalization of cardiac-type mucosa, particularly at the gastroesophageal junction, is due to GERD or is secondary to Helicobacter pylori infection. In part, this controversy exists because some authors report the type of underlying mucosa in which intestinalization occurs, e.g. cardiac-type mucosa ), and others report the anatomical area in which it occurs, e.g. the cardia of the stomach . Confusion occurs in the latter because it is difficult to differentiate the anatomic cardia from the fundus. When the underlying mucosa is identified, intestinal metaplasia arising in fundic or antral mucosa has been associated with H. pylori infection, but this relationship does not persist for cardiac-type mucosa. In contrast, intestinal metaplasia in cardiac-type mucosa is not associated with H. pylori or other gastric pathology; rather, it is associated with GERD . This observation is in keeping with the rapid and parallel increase of adenocarcinoma of the esophagus and the gastroesophageal junction .
|
INTEGRATED HYPOTHESIS OF THE PATHOPHYSIOLOGY OF GASTROESOPHAGEAL REFLUX DISEASE |
|---|
| TopPreviousNextLiterature Cited |
|---|
We hypothesize that GERD begins in the stomach. Fundic distention occurs because of overeating and delayed gastric emptying secondary to the high-fat Western diet . The distention causes the sphincter to be "taken up" by the expanding fundus, exposing the squamous epithelium of the terminal esophagus to gastric juice . Repeated exposure causes inflammation of the squamous epithelium, the formation of cardiac-type mucosa, and carditis . This initial step explains why, in early disease, the esophagitis is mild and commonly limited to the very distal esophagus. The patient compensates by increased swallowing , which bathes the injured mucosa in saliva and alleviates the discomfort induced by exposure to gastric acid. Increased swallowing results in aerophagia, bloating, and repetitive belching . The distention induced by aerophagia leads to further exposure and repetitive injury to additional squamous epithelium and to further development of cardiac-type mucosa. This repetitive process results in inflammation of the cardiac-type mucosa, commonly called carditis, and explains the complaint of epigastric pain so often registered by patients with early disease. The process can lead to a fibrotic mucosal ring at the squamocolumnar junction or short tongues of cardiac-type mucosa resulting in an irregular squamocolumnar junction . Extension of the inflammatory process into the muscularis propria causes a progressive loss in the length and pressure of the distal esophageal high-pressure zone and is associated with increased esophageal exposure to gastric juice and the symptoms of heartburn and regurgitation . The loss of the barrier occurs in a distal-to-proximal direction and eventually results in the permanent loss of lower esophageal sphincter resistance and the explosion of gastric juice into the esophagus, with all the clinical manifestations of severe esophagitis. This accounts for the observation that severe esophageal mucosal injury is almost always associated with a permanently defective sphincter . If there is sufficiently high esophageal acid exposure, cardiac-type mucosa can form within the esophageal body. At any time during this process and under specific luminal conditions or stimuli, such as exposure to a specific pH range, intestinalization of the cardiac-type mucosa can occur and set the stage for malignant degeneration (Figure 15). This almost always occurs in patients with long segments of cardiac-type mucosa.
|
|
THERAPY OF GASTROESOPHAGEAL REFLUX DISEASE |
|---|
| TopPreviousLiterature Cited |
|---|
Judging from the high prevalence of heartburn in the general population, GERD is very common . Most patients with mild symptoms self-medicate, whereas those with more severe and persistent symptoms seek medical attention. Consequently, studies on the natural history of GERD are rare, and the few that do exist usually involve patients receiving some form of therapy. This is unlikely to change, since few patients would endure the symptoms of the disease without obtaining therapy.
In one of the most detailed studies on the natural history of the disease, an intensive endoscopic follow-up of a defined population of 959 patients was performed over a 30-year period. The study involved only patients with endoscopic esophagitis and did not include those who had symptoms without mucosal injury. In about 45% of patients, esophagitis developed as an isolated episode and did not return while the patient was on acid-suppression therapy. In the remaining patients, esophagitis intermittently recurred on acid-suppression therapy, and in 42% it progressed on therapy to more severe mucosal injury. This latter group makes up about 23% of the initial population of patients with esophagitis. While on therapy, and within as short a period as six weeks, 18% of the initial population acquired a columnar-lined lower esophagus with intestinal metaplasia.
Comparative studies on the long-term outcome of medical or surgical therapy on the natural history of GERD are even more rare. One of the best is a 17–22-year follow-up study of 60 patients treated at one institution in Finland, a country with a relatively nonmobile, stable population . Of the 50 patients who were conservatively managed without H-2 antagonist or omeprazole, 72% had milder symptoms, 10% were unchanged, and 18% became worse over the course of their disease. Only 12% became symptom-free. Ten additional patients, with uncontrollable symptoms, had a surgical antireflux procedure. Nine were symptom-free over the same follow-up period, and in one, the symptoms were milder. Improvement in 24-h esophageal acid exposure occurred in both groups but was more marked after surgery. Endoscopic esophagitis had similar healing rates in both groups but 15% of the conservatively treated patients developed Barrett's esophagus, which did not occur in the surgically treated group.
Long-term studies on the outcome of either medical or surgical therapy on the natural history of GERD are also uncommon, and the measurements used to show the effect are variable. The following three outcome measures have been used to evaluate the effectiveness of therapy for GERD: (a) the therapy's effect on the symptoms, (b) its ability to heal and prevent further mucosal injury, and (c) its ability to correct the condition that causes the disease, i.e. the reflux of gastric juice into the esophagus. In general, most reports focus on the first of these measures and some incorporate the second but few include all three.
The mainstay of medical therapy is acid suppression, and several reports have established that chronic, life-long therapy is necessary to avoid recurrence of symptoms and mucosal injury . One long-term study showed that acid suppression is effective in healing esophagitis but dose escalation is often necessary to maintain the effect over five years . Twenty-four-hour esophageal pH monitoring of patients on acid suppression showed that despite high doses of proton pump inhibitors, nocturnal gastric acid breakthrough occurs and commonly results in esophageal acid exposure . Consequently, the consistency of medical therapy, even over a 24-h period, is problematic. Failure to fully inhibit gastric acid secretion, particularly in a patient who refluxes a combination of gastric and duodenal juice, can result in esophageal exposure to gastric juice with a pH of 3–5, causing persistent mucosal damage in an asymptomatic patient. This is a practical issue, since studies indicate that even short-term medication is taken inconsistently . Asymptomatic patients are unlikely to take medication regularly for 20–50 years, and even if they do, some reflux of gastric juice into the esophagus is still likely.
The reality of ongoing mucosal injury during medical therapy has been shown . A population of 138 patients with GERD were studied retrospectively with a focus on prior medical treatment, duration of reflux symptoms, former endoscopic findings, the grade of esophagitis before treatment began, and whether Barrett's metaplasia developed while on effective intermittent or continuous medical treatment if no evidence of Barrett's metaplasia was present before medical therapy began. Thirty-four percent of the 74 patients on effective intermittent acid-suppression medication developed Barrett's metaplasia, although they did not have it before medical treatment started. Further, Barrett's metaplasia developed in 58% of the 12 patients treated by effective continuous omeprazole administration. The median duration of medical treatment in those who developed Barrett's metaplasia was 7 years, ranging from 5 months to 20 years. The median number of endoscopic procedures done before the Barrett's metaplasia was diagnosed was 9, ranging from 3 to 17, which suggests that the finding of Barrett's metaplasia was not simply overlooked on previous examinations. Patients with Barrett's esophagus who were on continuous therapy with omeprazole or H-2 antagonist had no recurrence of acute esophagitis on endoscopy and were relieved of heartburn and retrosternal pain; however, 69% still complained of frequent belching, bilious vomiting, nausea, regurgitation, or dysphagia. These symptoms were also present in 80% of the patients who developed Barrett's esophagus while under treatment with intermittent omeprazole or H-2 antagonist therapy, even though they did not complain of heartburn and had no signs of acute esophagitis on endoscopy.
In contrast to medical therapy, which centers on acid suppression, surgical therapy focuses on reestablishing an effective barrier between the stomach and esophagus. Like medical therapy, surgical therapy has been the subject of few long-term outcome studies. Those that exist have shown that for early and late disease, a properly selected and performed surgical antireflux procedure is durable. An actuarial analysis of 100 patients with normal esophageal length showed that a Nissen fundoplication enjoyed a 91% success rate in controlling reflux symptoms over a 10-year period. This was associated with healing of esophagitis in all patients who had had this symptom before surgery and with normalization of esophageal acid exposure on 24-h pH monitoring in 88% of the 34 patients who volunteered for postoperative testing.
Pearson et al reported the long-term outcome following the combination of gastroplasty and partial fundoplication in 382 patients with obvious acquired shortening of the esophagus owing to severe esophagitis and stricture, or with subtler degrees of acquired shortening frequently encountered in patients who require reoperation or who have a massive sliding hiatal hernia . Of these 382 patients, 250 were followed for 5 or more years and 101 for more than 10 years. Two hundred and fifteen patients had an acquired short esophagus caused by peptic stricture (138 patients) or severe ulcerative esophagitis without significant stricture (77 patients). One hundred and twelve of these were followed for more than 5 years and 46 for more than 10 years. They had a 93% good to excellent outcome sustained over the long follow-up period. Fifty-four patients had a large sliding hiatal hernia or intrathoracic stomach. Thirty-eight of these were followed for more than 5 years and 4 for more than 10 years. They had a 91% good to excellent outcome, and none required a further operation. Unfortunately, endoscopy and 24-h esophageal pH monitoring were not used routinely during the follow-up period and were only done in symptomatic patients.
Because operative therapy is invasive, mortality from the procedure is always a possibility. Collective experience with the open transabdominal Nissen fundoplication shows an operative mortality of 0.1%. The transthoracic gastroplasty and partial fundoplication procedure has an associated operative mortality of 0.5% .
Very few authors have reported on the development of Barrett's metaplasia after a surgical antireflux procedure. A review of commonly cited studies with endoscopic follow-up of patients treated by antireflux surgery found no study that specifically addressed postoperative development of Barrett's esophagus . Most concentrated on the healing of esophagitis or commented on the postoperative length of Barrett's metaplasia that was known to be present before surgery. In a collected review of 216 patients followed for 3–60 months, there was not a single reported case of Barrett's esophagus that was not present preoperatively. Thus, despite the limitations of the surgical approach, the de novo development of Barrett's esophagus appears exceedingly rare in patients who have had an effective antireflux repair.
Some practical concepts regarding the treatment of the GERD can be drawn from the few studies on its natural history and those on the long-term outcome of medical and surgical therapy. First, not all patients with symptoms of GERD develop the complication of esophagitis, and in about half of those who do, it is an isolated event. Second, both medical and surgical therapy are effective in controlling the symptoms of the disease and in healing esophagitis. Third, dose escalation of acid-suppression therapy is often necessary to prevent recurrence of esophagitis. Fourth, relapse following cessation of medical therapy is common and indicates the need for life-long maintenance therapy. Fifth, if patients are properly screened for surgical therapy and the appropriate operation is selected and correctly performed, freedom from the symptoms and complications of GERD for up to 10 years or more can be expected. Sixth, the development of Barrett's esophagus is apparently rare after effective surgical therapy but is not uncommon during medical therapy.
Based on these concepts, it seems reasonable that patients presenting with symptoms of heartburn and regurgitation without obvious complications can be placed on simple antacids for 8–12 weeks before extensive investigations are performed. Given the over-the-counter availability of H-2 blockers, many patients will have already self-medicated their symptoms before seeing a physician. Patients should be advised to elevate the head of the bed; avoid tight clothing; eat small, frequent meals; avoid eating the night-time meal shortly before retiring; lose weight; and avoid alcohol, coffee, chocolate, peppermint, and fatty foods, all of which may aggravate the symptoms.
Alginic acid in combination with simple antacids may augment symptomatic relief by creating a floating physical barrier to reflux as well as neutralizing gastric acid secretions. Alginic acid reacts with sodium bicarbonate in the presence of saliva to form a highly viscous solution that floats on the surface of the gastric juice. When reflux occurs, this protective layer is refluxed into the esophagus and acts as a barrier against the noxious gastric contents. Medications to promote gastric emptying, e.g. metoclopramide, domperidone, or cisapride, are beneficial in early disease but of little value in more advanced disease.
After three to four weeks, simple antacid therapy should be discontinued. If symptoms persisted during therapy or return after discontinuation, the patient should be endoscoped. Failure of antacids or self-administered H-2 antagonist to relieve the symptoms suggest that the patient may have another disease. The immediate return of symptoms after stopping treatment suggests relatively severe disease and the necessity of maintenance therapy. An endoscopy at this time provides the opportunity to identify evidence of another disease, or, in those with GERD, to assess the severity of mucosal injury and the presence of Barrett's esophagus.
The administration of hydrogen potassium proton pump inhibitors such as omeprazole in doses of 20–60 mg/day can cause an 80–90% reduction in gastric acidity. This is advocated for patients whose symptoms persist despite antacid therapy. Proton pump inhibitors usually heal mild and moderate esophagitis but severe esophagitis may heal in only half of the patients unless high doses are used. It is important to realize that in patients who reflux a combination of gastric and duodenal juice, acid-suppression therapy may give symptomatic improvement while still allowing the reflux of gastroduodenal juice with a pH of 4–5 . This can produce an environment that allows persistent mucosal damage in an asymptomatic patient.
The traditional stepwise therapeutic approach to GERD, i.e. from antacids to acid-suppression drugs to antireflux surgery, should be reexamined in view of a more complete understanding of the pathophysiology and natural history of the disease. The rising incidence of Barrett's esophagus, a known risk factor for the current epidemic of adenocarcinoma, should be considered as well. The approach, outlined in Figure 16, should be to identify risk factors for persistent and progressive disease and encourage proper antireflux or resectional surgery when appropriate risk factors are present. Patients who show mucosal injury on endoscopy should have 24-h pH and bilirubin monitoring to determine the degree and pattern of esophageal exposure to gastric and duodenal juice. The status of the lower esophageal high-pressure zone and the function of the esophageal body should also be measured. These studies can help identify features that predict a poor response to long-term acid-suppression therapy, the likelihood of frequent relapses, and the potential for developing complications. These features include a permanently defective lower esophageal high-pressure zone, poor contractility of the esophageal body, supine reflux, duodenogastroesophageal reflux, severe erosive esophagitis, or a columnar-lined esophagus. Patients with these risk factors should be encouraged to have antireflux surgery, with the expectation of long-term control of symptoms and complications .
|
In the past, poorly controlled symptoms or persistence of mucosal injury and a permanently defective lower esophageal high-pressure zone were the primary indications for surgery, and many internists and surgeons were reluctant to advise surgery in their absence. Recently, the advent of laparoscopic technology has catalyzed a renewed interest in antireflux surgery. Successful procedures can be performed by laparoscopic access, with minimal disruption of the patient's life (the hospital stay is typically less than 48 h) and markedly less pain than is associated with the open procedure . Consequently a physician should not be deterred from considering the option of antireflux surgery in a patient with typical symptoms, with or without a permanently defective sphincter . The laparoscopic approach is particularly applicable to the patients with early disease who need chronic, life-long therapy with proton pump inhibitors or who require increasing doses to control their symptoms. As noted above, patients generally fail to take their medication consistently, which can establish an environment that promotes persistent mucosal injury. Significantly for these patients, a good response to medical therapy predicts an excellent outcome following antireflux surgery.
The requirements for consideration of antireflux surgery are: (a) the objective demonstration of increased esophageal acid exposure or 24-h pH monitoring; (b) symptoms that are typical of GERD; and (c) symptoms that improve on acid-suppression therapy. Whether the patient has a permanently defective sphincter is not critical, since recent studies have shown that the Nissen fundoplication prevents the shortening of a normal sphincter during gastric distention. Patients who fulfill the above criteria have a 97.4% chance of having an excellent to good long-term outcome from laparoscopic Nissen fundoplication . When the criteria are met, this procedure can be performed with minimal morbidity, making it a very competitive alternative to chronic acid-suppression therapy .
Approximately 25% of patients who develop severe esophagitis with their first episode of GERD will develop persistent progressive disease while on medical therapy . They usually have a structurally defective lower esophageal sphincter and reflux both gastric and duodenal juice into their esophagus . Clinically, they have supine reflux, causing position-dependent regurgitation and repetitive pulmonary aspiration, or persistent esophagitis of the same severity commonly associated with Barrett's metaplasia. These patients should be encouraged to have early surgical therapy; otherwise, the repetitive mucosal injury can lead to "esophageal shortening" , a term surgeons use when they cannot bring the gastroesophageal junction down into the abdominal cavity without tension .
Shortening of the esophageal body occurs as complications of the disease become more severe. It is caused by scarring throughout the esophageal wall and is manifested by hiatal herniation and periesophageal inflammation. Radiologically, it is associated with fixation of the hiatal hernia, i.e. the hernia does not reduce in the upright position. Any hernia that measures on endoscopy greater than 5 cm between the crura and the gastroesophageal junction, identified as the proximal extent of the gastric rugal folds, is likely to be associated with esophageal shortening. Manometrically, the peristaltic amplitude in the distal esophagus is often subnormal. If a short esophagus is detected at the time of an abdominal fundoplication, the surgeon's options are severely limited. It is much better to suspect it ahead of time and plan the operative strategy accordingly.
Patients with normal esophageal length and normal esophageal body motility are best served by a transabdominal Nissen fundoplication, now normally done laparoscopically. Usually this situation is found in early disease, before severe complications have developed. The presence of a short esophagus is an indication for a transthoracic approach. This allows the surgeon to mobilize the esophagus to a much greater extent than is possible through the abdomen. If after mobilization there is sufficient esophageal length, a transthoracic Nissen fundoplication can be done, provided esophageal contractility is adequate (>20 mm Hg over the distal half of the esophagus). If after extensive mobilization the esophagus is too short for a tension-free repair, a Collis gastroplasty can be performed to provide an extra 4–5 cm of "neoesophagus" around which a Belsey partial fundoplication can be added . The partial fundoplication is done because esophageal contractility is usually reduced when extensive esophageal shortening is present. In this situation, a complete fundoplication would create too much resistance and lead to postoperative dysphagia . In the relatively rare case where low distal peristaltic amplitude is associated with normal length, a transabdominal partial fundoplication is a reasonable option.
End-stage GERD occurs when the patient has persistent dysphagia associated with poor motility, a stricture that needs repetitive and frequent dilatations, a history of multiple unsuccessful antireflux operations, or Barrett's esophagus with high-grade dysplasia. These conditions are best served by esophageal replacement. The most durable substitute is the colon, and the functional results are especially good if the vagus nerves can be left intact .
Perhaps the most important issue in the treatment of GERD is the development of Barrett's esophagus and its progression to cancer during therapy. The development of Barrett's esophagus is almost always associated with symptoms or evidence of GERD; it is unusual in normal individuals who do not reflux. Consequently, the best protection against Barrett's esophagus is an effective antireflux barrier. Patients with GERD who undergo a surgical antireflux procedure prior to the development of Barrett's esophagus rarely develop it after surgery, provided the repair remains competent . The best deterrent to the development of Barrett's esophagus and subsequent adenocarcinoma, therefore, is early referral of patients with risk factors of progressive disease for an antireflux procedure. These risk factors are those known to make the disease difficult to control medically, such as a permanently defective sphincter, supine reflux, esophageal bile exposure, young age, and persistent symptoms .
There is evidence that once Barrett's esophagus develops, surgical therapy is associated with a reduced incidence of dysplasia and adenocarcinoma compared with medical therapy. McCallum et al followed 181 patients with Barrett's esophagus free of dysplasia; 29 had antireflux surgery and the other 152 were treated medically. After a mean follow-up of 62 months in the surgical group and 49 months in the medical group, there was a significant difference in the incidence of dysplasia and adenocarcinoma. Dysplasia was found in 3.4% of the surgical group compared with 19.7% in the medically treated group. No patient in the surgically treated group developed adenocarcinoma of the esophagus, compared with 1.3% of the medically treated patients. The authors concluded that compared with medical therapy an antireflux operation was significantly associated with the prevention of dysplasia and cancer in patients with Barrett's esophagus. Similarly, Ortiz et al studied 59 patients with Barrett's esophagus who were randomized to medical or surgical therapy and followed for a mean of four years. One of the 30 surgical patients had a failed fundoplication and developed high-grade dysplasia and adenocarcinoma postoperatively. In comparison, 6 of 27 medically treated patients developed dysplasia (5 low-grade, 1 high-grade), and adenocarcinoma was ultimately identified in the patient with high-grade dysplasia. The difference did not reach statistical significance, given the small number of patients.
More recently, Katz et al compared the development of dysplasia and adenocarcinoma during endoscopic surveillance of patients with Barrett's esophagus in 15 patients who had an antireflux procedure and 82 who received acid-suppression therapy only. Their analysis showed a lower risk of developing dysplasia and adenocarcinoma after antireflux surgery than with medical treatment, p < 0.05. This suggests a protective effect of surgery and supports the hypothesis that progression to dysplasia is less likely in patients who are consistently protected from epithelial injury.
One factor complicating any analysis of progression of Barrett's esophagus after antireflux surgery is that the cellular and genetic alterations leading to dysplasia and adenocarcinoma may have already occurred prior to the antireflux procedure. It has been estimated to take up to six years for adenocarcinoma to develop within Barrett's esophagus with low-grade dysplasia, and thus some cancers, particularly those that present during the first few postoperative years, may not necessarily represent progression of disease after surgery. McDonald et al found invasive adenocarcinoma in two patients and carcinoma in situ in one patient during surveillance after antireflux surgery, but they noted that no patient developed carcinoma after 39 months despite a follow-up for a median of 6.5 years and a maximum of 18.2 years.
A review of the English-language literature since 1975 revealed 11 series that followed a total of 346 patients with Barrett's esophagus after fundoplication . Fifteen patients developed esophageal adenocarcinoma after antireflux surgery. In addition, four isolated reports described adenocarcinoma developing in Barrett's esophagus after an antireflux operation. Although the length of follow-up was not always available, 11 of the 19 cancers (58%) developed within 3 years of fundoplication, and 15 (79%) developed within 5 years of fundoplication. The remaining 4 cancers developed 5–10 years after fundoplication but each of these patients had recurrent reflux determined by symptoms or positive 24-h pH monitoring. Thus, a functioning fundoplication seems to provide protection from progression of Barrett's esophagus to adenocarcinoma. The first 3–5 years after antireflux surgery remain critical, though, since most of the cancers that will occur will be discovered during this period. Consequently, the best management of a patient with Barrett's esophagus seems to be an antireflux procedure followed by yearly surveillance.
Annu. Rev. Med. 1999. 50:469-506
Copyright © 1999 by Annual Reviews. All rights reserved
0066-4219/99/0201-0469
|
Reprint
(PDF) Version of this Article 
= p <0.05 vs normals, no injury or erosive
esophagitis. (Printed with permission from Reference ) 
