Esophageal Mucosa in Health and Disease

Defining a Criterion of Irreversibility: Visible Columnar Lined Esophagus

The first step in preventing adenocarcinoma is to recognize the point of irreversibility that signals the inability to prevent progression to adenocarcinoma. In GERD, at this point in time, that point of irreversibility is the occurrence of vCLE. In the United Kingdom, vCLE defines Barrett esophagus. In the United States and Europe, intestinal metaplasia is required for the diagnosis of Barrett esophagus.

Medical treatment does not reverse vCLE or prevent its progression to intestinal metaplasia, increasing dysplasia, and adenocarcinoma. Present medical treatment of GERD therefore commits 10% of all patients to irreversibility every 5 years.

The advantage with defining irreversibility in GERD by the presence of vCLE is that there is no evidence that any patient who does not have vCLE progresses to adenocarcinoma. If we prevent vCLE, we will prevent adenocarcinoma.

It can be reasonably argued that the person who is endoscopically normal with intestinal metaplasia at the normal SCJ is at risk for adenocarcinoma of the “gastric cardia.” However, present management guidelines recommend that such patients with GERD should not undergo biopsies because the risk of cancer in patients who have intestinal metaplasia is unknown. The argument, therefore, has no practical merit at this time. It may change in the future if an increased cancer risk is defined in this group. If and when that happens, preventing intestinal metaplasia at the SCJ in the endoscopically normal person will become necessary.

The detection of vCLE requires endoscopy. The present management guidelines delay endoscopy to the point of treatment failure. At this point a significant number of patients will already have vCLE. If endoscopy is performed proactively without waiting for treatment failure, as was done in the Pro-GERD study, 240 of 2721 (8.8%) patients would already have vCLE. In addition, the following endoscopic findings were predictive of progression to vCLE in the next 5 years: (1) presence of erosive esophagitis with risk increasing to 19.7% in patients with severe erosive esophagitis ; and (2) presence of intestinal metaplasia in a biopsy taken from the SCJ of an endoscopically normal patient, with such patients having a 25% risk of progression to vCLE within 5 years. The patients in the Pro-GERD study had well-established GERD, often with severe symptoms and a long duration. It is probable that endoscopy performed at the onset of GERD would have a lower prevalence of vCLE.

In the Pro-GERD study, the nonendoscopic findings that were significantly associated with progression to vCLE in GERD patients under medical therapy were male gender, alcohol use, and regular PPI use.

None of the nonendoscopic criteria that are predictive for development of vCLE within 5 years listed previously are indications for endoscopy in the GERD patient. The indication remains the occurrence of treatment failure. The main reason for this is the lack of any desire to prevent vCLE in the minds of the medical community. To them, vCLE is simply another inevitable event in the course of GERD that occurs in a minority of GERD patients. The fact that it is a cellular change whose end point is a lethal malignancy is ignored.

This is a nihilistic attitude that permits conversion of the patient without risk to one whose progression to adenocarcinoma becomes inevitable. The only excuse for this attitude is that cancer is rare in GERD patients. With the sevenfold increase in the incidence of GERD-induced adenocarcinoma over the past 4 decades, this excuse has become increasingly lame and unacceptable.

If the presence of vCLE is recognized as the point of irreversibility in GERD, there can be a new objective of management of the GERD patient: the prevention of progression to vCLE.

This would then provide an incentive and demand for earlier endoscopy before failure with empiric treatment with PPI in the patient with GERD. Early endoscopy presently has the ability only to recognize the presence of vCLE and predict its occurrence within the next 5 years by the presence of severe erosive esophagitis (19.7%) and intestinal metaplasia at the normal SCJ (25%). Successful repair of the damaged LES in the patient with a high risk of vCLE in 5 years has a high probability of preventing vCLE.

These reasons for early endoscopy are presently not justified because of the cost associated with increasing the number of endoscopies. However, it emphasizes the fact that any push to prevent adenocarcinoma must change the indications for endoscopy to an earlier stage in the progression of GERD. This will only happen if a new and more accurate method of predicting progression of GERD to vCLE becomes available. The new histologic measure of LES damage that we propose can be that test.

Cause of Visible Columnar Lined Esophagus

To be effective in preventing vCLE, we must identify its cause. It is certain that vCLE is the result of exposure of the esophagus above the endoscopic GEJ to gastric contents as a result of reflux. As such, it is also certain that if reflux can be prevented, vCLE will not occur.

There is strong evidence that the risk of vCLE increases with increasing severity of reflux (demonstrated by objective evidence of acid exposure by a pH test), increasing duration of reflux, male gender, regular PPI therapy, and possibly alcoholism and smoking. Nason et al. showed that the prevalence of Barrett esophagus was higher in patients whose symptoms were controlled with PPI therapy. They suggested that the present practice of waiting for treatment failure was irrational if the objective for endoscopy was the detection of Barrett esophagus.

The most dominant factors in the etiology of vCLE are the severity and duration of reflux. Patients with Barrett esophagus are known to have a higher prevalence of an abnormal pH test than any other category of GERD. There is no specifically defined level of abnormality in the pH test or a specific number of years of reflux that correlates with the occurrence of vCLE. If the objective is preventing vCLE, success will demand intervention at the earliest practical time after the onset of reflux. For prevention of vCLE to be certain, intervention must occur before any significant reflux occurs into the thoracic esophagus.

From a practical standpoint, it is necessary to identify criteria that separate very low and high risk of impending and future vCLE by some defined severity and/or duration of reflux. This is not possible by presently available tests. We will propose that the new test of LES damage provides accurate criteria for predicting future vCLE.

It is certain that the severity of reflux into the thoracic esophagus correlates with the frequency of LES failure, which in turn correlates with the severity of LES damage. In relation to our objective of preventing vCLE, this recognition establishes a new more practical objective: prevention of reflux into the thoracic esophagus that is severe enough to cause vCLE.

Defining the Normal and Defective Lower Esophageal Sphincter by Manometry

The functional state of the LES can be defined manometrically by three separate components : its mean pressure, its total length, and the length of its abdominal segment. Manometric studies of asymptomatic subjects indicate that the LES pressure is greater than 15 mm Hg, the total LES length is 40 to 50 mm, and the length of the abdominal segment (a-LES) is 30 to 35 mm.

The criteria that define a defective LES that correlates with the presence of sufficient reflux into the esophagus to produce clinical GERD are (1) a decrease in the mean LES pressure to less than 6 mm Hg, (2) a decrease in total LES length to less than 20 mm, and (3) a decrease in a-LES length to less than 10 mm. At these levels of LES damage, sphincter failure occurs so frequently that it results in an abnormal pH test and significant exposure of the squamous epithelium in the body of the esophagus to reflux. LES damage defined by these criteria correlate with an increased probability of symptoms of GERD, severe grades of erosive esophagitis, and vCLE.

There is a significant gap between the previously listed criteria that define a normal LES and a defective LES that is associated with abnormal reflux into the esophagus, as defined by an abnormal pH test and the presence of clinical GERD. The mean LES pressure must decrease from a normal of greater than 15 mm Hg to less than 6 mm Hg; the total LES length must decrease from a normal of 40 to 50 mm to less than 20 mm; and the a-LES length must decrease from 30 to 35 mm to less than 10 mm before it becomes a criterion of LES failure.

At least part of this gap between a normal and defective LES represents the reserve capacity of the LES. As LES damage increases, its reserve capacity is progressively reduced. However, as long as it is not exhausted, the LES maintains its competence (green zone in Table 3.1 ).

TABLE 3.1

Length of Abdominal Segment of Lower Esophageal Sphincter (a-LES) Damage (Measured by the New Test), Length of the Residual Functional a-LES, and Their Correlation With Lower Esophageal Sphincter Failure and Severity of Reflux *

Open full size image

GERD, Gastroesophageal reflux disease; LES, lower esophageal sphincter. Green areas, The LES is competent with damage that is within its reserve capacity. Orange areas , Clinical GERD from onset of symptoms to point of transition from postprandial reflux to incessant reflux and an increasing prevalence of visible columnar-lined esophagus (vCLE). Red areas, The LES is incompetent with severe reflux and a high prevalence of vCLE.

* We assume that the patient has an initial a-LES length of 35 mm, that a heavy meal causes 10 mm of dynamic shortening of the a-LES in the postprandial phase, and that LES failure occurs at an a-LES length of less than 10 mm.

This early LES damage cannot be recognized by any present criterion for the diagnosis of GERD: the patient has no symptoms, no endoscopic abnormality, no manometric criteria of a defective LES, and no abnormal pH test. This state where the LES is damaged within its reserve functional capacity can be called the phase of compensated LES damage . We will show that histologic examination with new criteria can define and measure this early LES damage.

Before the onset of LES damage, all persons have an initial a-LES length that is equal to the length of the abdominal esophagus. Zaninotto et al. reported that the manometric length of the a-LES in 49 asymptomatic volunteers had the following distribution (I have taken the liberty of removing one outlier that had an a-LES length of >50 mm): less than 10 mm in 1; 10 to 15 mm in 6; 15 to 20 mm in 10; 20 to 25 mm in 17; 25 to 30 mm in 11; and 30 to 35 mm in 5 persons.

The manometric measurement of the a-LES at any given point in a person's life after the LES has developed completely can be expressed by the following formula:

This only assumes that the anatomic part of the abdominal esophagus that contains the a-LES does not disappear into thin air when LES pressure is lost.

Manometrically, LES damage is equivalent to loss of pressure. When this occurs at the distal end, it results in shortening of the manometric a-LES. The damaged LES is distal to the end of the residual LES at manometry and therefore identical in its pressure characteristics to the proximal stomach.

In a patient with LES damage, the distal limit of the manometric LES is not the end of the esophagus (see Fig. 3.2 ). The true end of the esophagus includes the damaged a-LES. Any manometric interpretation that makes the assumption that the esophagus ends at the distal end of the manometric LES is potentially wrong by as much as 35 mm (the entire initial a-LES length). For example, if the distal limit of the manometric LES is above the diaphragmatic pressure impression, this is not necessarily a hiatal hernia because the true end of the esophagus cannot be defined by manometry.

At present, the previous formula that defines the LES cannot be applied because two elements, LES damage and the initial LES length, are unknown. As a result, manometry has no practical value in the diagnosis of GERD. However, it illustrates the critically important and misunderstood concept that the manometric definition of the distal end of the a-LES is not the end of the esophagus. The true end of the esophagus must include the damaged a-LES that is present distal to the manometric end of the LES in virtually all people This cannot be measured at present.

In Zaninotto et al. , therefore, the measured manometric a-LES does not necessarily represent individual variation of the length of the normal a-LES, simply because the subjects had no symptoms of GERD. It could be the result of shortening of the a-LES by progressive a-LES damage. The data in the study can be explained by assuming that the initial a-LES length was 35 mm (the highest length) in all patients, and the distribution represents different degrees of a-LES damage. For example, an asymptomatic person with a measured manometric a-LES length of 22 mm (the median a-LES length in the study) could have an initial length of 35 mm with 13 mm of a-LES damage (see Table 3.1 ). That person is asymptomatic because the LES, though damaged, is still sufficiently competent to prevent reflux.

The distribution of acid exposure in these volunteers in Zaninotto et al. showed a pH less than 4 for a mean of 1.57%, a median of 1.1%, and a range of 0% to 6% of the 24-hour period. This shows that these asymptomatic persons had evidence of mild reflux with 5% reaching the pH test definition for abnormal reflux. This was objective evidence of LES failure, despite the fact that they did not have symptoms.

The data in Zaninotto et al. raise the intriguing but obvious probability that the LES has a reserve capacity. It can shorten significantly from its initial length while remaining competent—that is, there is a phase of compensated LES damage where patients have LES damage within its reserve capacity without significant LES failure and reflux into the thoracic esophagus. A person without significant reflux into the thoracic esophagus will be at zero risk for developing vCLE.

Based on this understanding, we can divide the severity of LES damage into (1) compensated, (i.e., LES damage is such that it does not produce LES failure, where the pH test is zero; green zone in Table 3.1 ); (2) LES damage that causes infrequent LES failure and mild reflux (i.e., pH test is greater than zero but pH test normal; <4.5% of time pH < 4 or DeMeester score < 14), noting that vCLE is extremely unlikely in such patients (orange zone in Table 3.1 ); and (3) severe LES damage with LES failure sufficient to produce an abnormal pH test and a high prevalence of vCLE (red zone in Table 3.1 ).

This further refines our objective into an LES-based objective to prevent vCLE: prevention of a-LES damage beyond the point where reflux is sufficiently severe to cause vCLE. In Table 3.1 , this corresponds to preventing a-LES damage from reaching 25 mm. When there is a measure of a-LES damage, there is a range of zero to 25 mm of LES damage that is available for intervention to prevent progression of LES damage. Prevention of vCLE becomes theoretically very feasible in this method.

This new objective clearly shows the futility of present management of GERD. The presently accepted criteria that define GERD (troublesome symptoms, erosive esophagitis, an abnormal pH test, and a defective LES on manometry where the a-LES is <10 mm) are the very things that must be prevented if we hope to prevent esophageal adenocarcinoma.

Result of Abdominal Lower Esophageal Sphincter Damage: the Dilated Distal Esophagus

A largely unappreciated normal function of the a-LES is to maintain the tubular shape of the abdominal esophagus. The high resting pressure of the a-LES continually opposes the dilatory tendency of the positive (around +5 mm Hg) intraluminal pressure of the abdominal esophagus.

When the a-LES is damaged, the protection provided by the tonic contraction of the LES is lost. The dilatory positive intraluminal pressure will be accentuated during meals when the stomach distends and the intragastric pressure increases. The distal abdominal esophagus that has lost LES pressure will therefore dilate to form the dilated distal esophagus (see Fig. 3.3 ).

With LES damage, the tubular abdominal esophagus shortens, the damaged esophagus dilates and takes up the gastric contour and becomes part of the reservoir, and the angle of His becomes more obtuse. Mucosal rugal folds, which are a feature of all reservoir organs, develop in this dilated distal esophagus that results from loss of abdominal LES function (as discussed later).

The dilated distal esophagus has a variable length that is equal to the amount of shortening of the a-LES due to damage. The equation that defines the a-LES now resolves as follows:

The end of the tubular esophagus, which has been used by pathologists to define the GEJ since Hayward in 1961 , is proximal to the true GEJ by the length of the dilated distal esophagus.

This “gastricization” of the abdominal esophagus that has lost LES tone occurs at a manometric, endoscopic, and gross anatomic level. This has led to confusion that has created error in this region from the beginning of time and continues to the present.

We will show that it is only the correct interpretation of the histology of this region that can resolve this error.

Mechanism of Abdominal Lower Esophageal Sphincter Damage

LES damage is the result of pressure exerted from below, as a result of a heavy meal that causes gastric overdistention. Ayazi et al. and Robertson et al. showed elegantly that gastric overdistention causes “effacement” of the distal part of the LES, resulting in a temporary decrease in LES length. The squamous epithelium lining the effaced LES is exposed to gastric juice because the pH transition point has moved proximally (see Fig. 3.4 ).

The phenomenon of effacement of the distal end of the LES can be demonstrated at endoscopy. In a person with normal endoscopy, the SCJ is the GEJ. In retroflex view, when the stomach is insufflated with air, the SCJ moves downward and becomes visible. When the same thing happens during a heavy meal, the squamous epithelium is in the stomach, below the pH transition point.

There is a pocket of strong acid at the height of the food column during a meal. Repeated and frequent exposure of the squamous epithelium to this acid pocket during gastric overdistention during heavy meals results first in reversible injury to the distal esophageal squamous epithelium, followed by permanent columnar metaplasia of the squamous epithelium.

If LES damage occurs because of pressure from below, it must follow that LES damage begins at its distal end and progresses upward. Loss of length therefore begins in the distal a-LES. Robertson et al. showed that early LES shortening produced by a heavy meal in asymptomatic volunteers was entirely in the abdominal segment and did not affect the thoracic LES.

LES damage can therefore be considered to be basically the result of an eating disorder. Viewed in this light, each person can be regarded as having a unique relationship between his/her eating habit, the response of the LES to this overeating, and the damage caused to the esophageal squamous epithelium by exposure to gastric juice.

At one extreme, the patient's LES is not damaged by the effect of his/her eating habit on the LES. This patient never has LES failure and reflux, the pH test is zero, and this person never gets GERD. At the other extreme, the patient's LES is damaged early in life by an excessive eating habit and/or an LES susceptible to damage and progresses rapidly to LES incompetence and severe reflux into the thoracic esophagus at a relatively young age. This damage includes erosive esophagitis and becomes irreversible when vCLE occurs.

Between these two extremes is the entire clinicopathologic spectrum of GERD. Progression of GERD can therefore be defined by the rate of progression of LES damage resulting from a person's eating habit ( Table 3.2 ).

TABLE 3.2

Changes With Age of the Functional Residual Length of the Abdominal Lower Esophageal Sphincter *

Open full size image

* Assuming that the original length at maturity is 35 mm, that lower esophageal sphincter (LES) damage begins at age 15 years, and that LES damage has a linear progression over the long term. Note : The abdominal LES lengths in green represent lengths at which the LES is likely to be competent. The lengths in orange represent an LES that is susceptible to failure with gastric distention (i.e., at risk of postprandial reflux). The lengths in red represent an LES that is below the length at which LES failure occurs at rest.