What is the pathophysiology of reflux patients developing Barrett's esophagus?
A. Watson (London)
Around 10% of patients with gastroesophageal reflux disease (GERD) develop Barrett's columnar-lined esophagus (CLE) and in these, some 10% will develop adenocarcinoma  and around 40% will develop other complications, including stricture and ulcer . Although Barrett's CLE is a complication of GERD in which squamous epithelium becomes destroyed and replaced by metaplastic columnar epithelium, only a small number of patients with Barrett's CLE have a documented history of GERD These factors suggest that patients with Barrett's CLE have specific pathophysiological features, and several studies have been performed to elicit differences in the various components of the pathophysiological profile between patients with erosive esophagitis than those with Barrett's CLE.
Lower esophageal sphincter function
Patients with Barrett's CLE have low values of resting lower esophageal sphincter (LES) pressure, shorter LES length and intra-abdominal length and a higher proportion of a mechanically defective LES than patients with erosive esophagitis. Mean resting LES pressure. was 4.5 mm/Hg in patients with Barrett's CLE compared to 8.2 mm/Hg in Stein's series  and 4.0 mm/Hg compared to 9.6 mm/Hg in our laboratory [4,5]. Stein's study also showed a shorter LES length and intra-abdominal length in Barrett's patients compared to those with erosive esophagitis, and the proportion of Barrett's patients with a mechanically defective LES, on the basis of resting pressure, length and intra-abdominal length , was 93% in Stein's series , 90% in DeMeester's series  and 72% in our laboratory, compared to a range of 29%-46% in the series of patients with erosive esophagitis. No difference in the proportion of patients with a mechanically defective LES has been demonstrated in Barrett's patients between those with and without complications .
Esophageal body function
Patients with Barrett's CLE have a lower mean amplitude of contractions in the lower third of the esophagus, a higher incidence of contractions less than 30 mm/Hg and a higher proportion have contractions less than 30 mm/Hg than patients with erosive esophagitis. In Stein's study, the mean amplitude of contractions in the lower third of the esophagus was 38 mm/Hg compared to 75 mm/Hg in patients with erosive esophagitis. At 24-hour ambulatory manometry, they found that the frequency of low amplitude contractions (less than 30 mm/Hg) was 32% of all contractions in Barrett's patients compared to 18.9% in patients with erosive esophagitis . In our laboratory, the proportion of patients with mean amplitude of contractions less than 30 mm/Hg in the lower third of the esophagus was 65% in Barrett's patients compared to 35% of patients with erosive esophagitis.
In addition to peristaltic impairment due to low amplitude contractions, patients with Barrett's CLE also have a higher incidence of simultaneous, non-propagated peristaltic waves than patients with erosive esophagitis. In our laboratory, 31% of patients with Barrett's CLE had an incidence of greater than 50% of simultaneous contractions compared to 7% of patients with erosive esophagitis. DeMeester's study showed that 83% of patients with Barrett's CLE had "peristaltic failure" in the distal esophagus manifest by inadequate contractility and/or abnormalities of wave progression, including simultaneous contractions, dropped waves and interrupted waves . They also showed that there was no difference in the incidence of manometric defects in the esophageal body between patients with uncomplicated Barrett's CLE and those who developed complications.
Esophageal acid exposure
There appears to be a greater frequency of higher levels of pathological acid exposure across the pH ranges in patients with Barrett's CLE compared to those with erosive esophagitis. DeMeester's study showed that 93% of patients with Barrett's CLE have pathological acid exposure . In Stein's study , the mean percentage total time the pH was less than 4 was 24% in Barrett's patients compared to 10.3% in patients with erosive esophagitis, and in our laboratory, the corresponding figures were 19.2% compared to 8.2%, both studies showing increased numbers of reflux episodes and an increased number of episodes greater than 5 minutes duration in Barrett's patients. Stein's group showed that the increased acid exposure in Barrett's patients occurred whether pH of less than 4, less than 3 or less than 2 were measured . In both DeMeester's and our series, there was no significant difference in the degree of acid exposure between Barrett's patients with and without complications [8, 9].
Gastric acid secretion
Several workers have documented higher basal and pentagastrin stimulated gastric acid secretion in Barrett's CLE [3, 10-12]. In Stein's series, mean basal acid output was 6.3 mmol/hr in Barrett's patients compared to 2.7 in patients with erosive esophagitis, corresponding figures for stimulated acid output being 23.4 and 13.2 mmol/hr respectively [3, 10]. Gastric hypersecretion was present in 48% of patients with Barrett's CLE compared with 21% of patients with erosive esophagitis. DeMeester's study showed no difference in gastric secretory status between patients with uncomplicated Barrett's CLE and those developing complications .
Several studies have shown increased esophageal alkaline exposure in patients with Barrett's CLE, in association with duodenogastric alkaline reflux, and more recently the availability of a portable spectrophotometer system has confirmed the presence of increased bilirubin in the esophagus of patients with Barrett's CLE. In Stein's study, the mean percentage total time the pH was greater than 7 was 9.5% in Barrett's patients compared to 5.5% in patients with erosive esophagitis and 3.5% in normal volunteers [3, 10). In two groups of Barrett's patients studied in our laboratory, the mean percentage total time pH was greater than 7 was 11.5 and 14.4% respectively, with a significantly higher level (16% vs 8%) in patients not responding to medical treatment and in patients developing complications of Barrett's CLE compared to uncomplicated cases (24.2% vs 8.4% ) [4, 11]. DeMeester's study also showed increased levels of alkaline exposure in Barrett's patients developing complications, 93% of patients with increased alkaline exposure developing complications compared to 44% of those with normal alkaline exposure .
The association between increased esophageal alkaline exposure and duodenogastric alkaline reflux has been demonstrated by 24-hour ambulatory combined esophageal and gastric pH monitoring in several studies [3, 8, 10, 13]. Stein et al. confirmed the presence of duodenogastric alkaline reflux in a higher proportion of Barrett's patients compared to those with erosive esophagitis by cholescintigraphy and by ambulatory intragastric spectrophotometric (Bilitec) bilirubin monitoring. Using the latter technique, duodeno-gastric reflux of bile was detected in 40% of patients with Barrett's CLE compared to 25% with erosive esophagitis and 0% in normal volunteers, and was found to occur predominantly in the postprandial and supine periods . Gillen et al. confirmed these findings using gastric aspiration techniques  and the same group subsequently performed ambulatory intraesophageal Bilitec monitoring, showing a significantly higher duration of bilirubin exposure in Barrett's patients compared to those with erosive esophagitis, particularly in the supine and interdigestive periods . Kauer et al. performed intraesophageal Bilitec monitoring and correlated simultaneous pH and bilirubin absorbence readings . They demonstrated abnormal bilirubin exposure in 80% of patients with Barrett's CLE compared to 40% of patients with erosive esophagitis, the
mean duration of bilirubin exposure being 14.5% and 1% of total time respectively. Importantly, they found that the abnormal bilirubin exposure occurred predominantly in the pH range 4-7, with a mixed refluxate of gastric and duodenal contents. Within this pH range, bile salts are capable of damaging esophageal mucosa and reflux events within this pH range will be undetected by standard ambulatory pH monitoring. Studies to differentiate the degree of intraesophageal bilirubin exposure between uncomplicated Barrett's patients and those developing complications are awaited.
Stein et al. found no difference in gastric emptying of a radio-labelled semi-solid meal between patients with Barrett's CLE and erosive esophagitis, although in both groups emptying was delayed compared to normal volunteers [3, 10]. In DeMeester's study, there was no difference in the rate of gastric emptying between Barrett's patients with and without complications . The occurrence of duodenogastric alkaline reflux is likely to be associated with more subtle alterations in antro-duodenal dysmotility.
Patients with Barrett's CLE have reduced sensitivity to perfusion of the esophageal mucosa with acid. This is suggested by the repeated observation that patients who develop major complications of GERD, such as reflux stricture and Barrett's CLE complain of less heartburn than patients with erosive esophagitis and indeed often have no symptoms prior to the onset or detection of the complication [3, 17]. Furthermore, in an epidemiological study of the prevalence of Barrett's CLE in Olmsted County, USA, Cameron found the prevalence to be 22.6 per 100,000 in patients undergoing endoscopy, whereas autopsy studies revealed the true prevalence in the community to be 376 per 100,000 population .
We studied 20 patients with Barrett's CLE and reflux stricture by perfusing the esophageal mucosa with N/10 Hydrochloric acid using a modified Berstein test. A scoring system was developed based on the degree of symptomatic awareness and the volume of HCl needed to elicit symptomatic awareness, and the scores were compared with those of 20 patients with erosive esophagus . All patients with erosive esophagitis exhibited symptomatic awareness of acid perfusion and 72% recorded the maximum symptomatic score. In the patients with reflux stricture and Barrett's CLE, 40% experienced no symptoms during this perfusion and only 15% exhibited the maximum score achieved by the majority of patients with erosive esophagitis.
It is evident, therefore, that patients with Barrett's CLE are at the extreme end of the pathophysiological spectrum of gastroesophageal disease. They have a higher incidence and degree of lower esophageal sphincter dysfunction and peristaltic failure in the esophageal body than patients with erosive esophagitis, together with greater levels of acid and alkaline exposure, the latter associated with esophageal exposure to a mixed refluxate of gastric and duodenal contents. It is unfortunate that this constellation is compounded by severe impairment and frequently absence of mucosal sensitivity to acid perfusion.
Whilst the management of patients with Barrett's CLE uncomplicated by dysplasia is controversial, it is apposite to draw appropriate therapeutic inferences from our knowledge of the pathophysiology of the condition. It is generally believed that the aims of therapy in gastroesophageal reflux disease are to alleviate symptoms and to prevent the onset of complications, the latter being more prevalent in the more severe forms of the disease. As many patients with Barrett's CLE do not have symptoms consequent on impaired mucosal sensitivity, the wide body of opinion which advocates no treatment in the absence of symptoms is open to question, particularly as this philosophy results in those patients with GERD with the most severe pathophysiological disturbances being treated less aggressively than those with milder forms of the disease. In the prevention of complications, namely ulcer, stricture and carcinoma, which perhaps is a more worthwhile objective in patients with Barrett's CLE, it should be recognized from the various studies performed that the only pathophysiological difference between those patients who go on to develop complications from those who do not, is an increased frequency and level of alkaline exposure [8, 9], which has been shown to be associated with a mixed refluxate of gastric and duodenal juice, occurring mostly within the pH range 4-7 at which bile salts are most likely to be damaging to the esophageal mucosa . Experimentally, the exposure of esophageal mucosa to duodenal contents results in adenocarcinoma, and there would appear therefore to be some merit in preventing alkaline exposure . Acid suppression therapy does not achieve this, which is likely to account for the observation that Barrett's CLE progresses, and reflux stricture and ulcer can develop on acid-suppression therapy [2, 4, 20, 21]. Antireflux surgery is the only modality which can effectively eliminate esophageal alkaline exposure, which is likely to explain the superior response over acid suppression in prevention of stricture and ulcer [4, 20] and the observations of arrest of progression of Barrett's CLE [1, 2, 20], and indeed partial or complete regression in a small proportion of cases [4, 20, 22-24]. Since the incidence of adenocarcinoma is proportional to the length of columnarised segment [21, 25], it seems reasonable to suggest that antireflux surgery may influence the incidence of adenocarcinoma, although this is currently unknown. It is, however, documented that adenocarcinoma can occur after successful antireflux surgery , but this may reflect the stage in the cycle of genomic instability at which intervention occurs, rather than implying an overall absence of effect of antireflux surgery.
It is on account of the controversy regarding management of Barrett's CLE in the absence of dysplasia, and particularly the influence of the various therapeutic modalities on the incidence of adenocarcinoma, which is increasing dramatically, that the author has proposed a multi-centre prospective randomized study to compare the effect of maximal conservative therapy and antireflux surgery in the management of Barrett's CLE with particular reference to the incidence of adenocarcinoma, which it is hoped will be mounted in conjunction with OESO.
1. Naef AP, Savary M, Ozello I. Columnar lined lower esophagus: an acquired lesion with malignant predisposition. Report on 140 cases of Barrett's esophagus with 12 adenocarcinomas. J Thorac Cardiovasc Surg 1975;70:826-834.
9. Attwood SEA, Ball CS, Barlow AP, Jenkinson L, Norris TL, Watson A. Role of intragastric and intraesophageal alkalinisation in the genesis of complication in Barrett's columnar lined esophagus. Gut 1993;34:11-15.
13. Ball CS, Jenkinson LR, Norris TL, Watson A. The value of simultaneous 24-hour ambulatory esophageal and gastric pH monitoring: an analysis of 100 investigations. In: Little AG, Ferguson MK, Skinner DB, eds. Diseases of the esophagus. Mount Kisco, New York, Futura Publishing, 1990:55-61.
16. Kauer WKH, Peters JH, DeMeester TR, et al. Mixed reflux of gastric and duodenal juices is more harmful to the esophagus than gastric juice alone: the need for surgical therapy re-emphasized. Ann Surg 1995;222:525-533.
18. Cameron AJ, Zinsmeister AR, Ballard DJ, Carney JA. Prevalence of columnar-lined (Barrett's) esophagus: comparison of population-based clinical and autopsy findings. Gastroenterology 1990;99:918-922.