Unbuffered Highly Acidic Gastric Juice Exists at the Gastroesophageal Junction After a Meal

Transcription

1 GASTROENTEROLOGY 2001;121: ALIMENTARY TRACT Unbuffered Highly Acidic Gastric Juice Exists at the Gastroesophageal Junction After a Meal JONATHAN FLETCHER,* ANGELA WIRZ,* JOANNE YOUNG,* RAMSEY VALLANCE, and KENNETH E. L. McCOLL* *Department of Medicine and Therapeutics, Gardiner Institute, Western Infirmary, and Department of Radiology, Gartnavel General Hospital, Glasgow, Scotland Background & Aims: Gastroesophageal reflux typically occurs after meals. During dual gastric and esophageal ph monitoring, we observed that postprandial refluxate was often more acidic than the gastric contents. This study aimed to investigate this phenomenon. Methods: Dual gastric and esophageal ph tracings were analyzed from 40 dyspeptic patients. Dual ph electrode pullthrough studies were performed in healthy volunteers to document regional variation in intragastric ph under both fasting and postprandial conditions. The squamocolumnar junction was identified using radio-opaque endoscopic clips. We also examined in vitro partitioning of gastric juice added to a homogenized fatty meal. Results: The dual ph traces confirmed that esophageal refluxate was frequently more acidic than the body of the stomach after meals but not during fasting. The pull-through studies showed a pocket of acid at the gastroesophageal junction that escaped the buffering effect of meals, remaining highly acidic (median ph 1.6) compared with the body of the stomach (ph 4.7; P < 0.001). This proximal acid pocket extended from the cardia across the squamocolumnar junction 1.8 cm into the distal esophagus. The in vitro studies showed that acidic gastric juice could partition on top of a homogenized fatty meal. Conclusions: After eating, highly acidic unbuffered gastric juice is present at the gastroesophageal squamocolumnar junction and is likely to contribute to the high prevalence of disease at this site. Exposure of the distal esophagus to acidic gastric juice and the accompanying symptoms of heartburn are most common in the postprandial period. 1,2 This relationship of gastroesophageal reflux symptoms to eating contrasts with that of duodenal ulcer disease, in which symptoms are relieved rather than induced by eating. 3 Relief of duodenal ulcer pain by eating is explained by the increase in intragastric ph that occurs after ingestion of food. Although food stimulates the stomach to secrete acid, the buffering effect of the food decreases the acidity of the gastric juice, causing its ph to increase from fasting values of to It is generally assumed that the acidity of the gastric refluxate reaching the distal esophagus is similar to or slightly less than that of the stomach from which it originated. In view of the marked reduction in the acidity of gastric contents after eating, it is surprising therefore that symptoms of heartburn are prominent during the postprandial period. Esophageal acid-perfusion tests using acid of differing strengths indicate a threshold of approximately ph ,6 Above this ph, esophageal perfusion rarely illicits symptoms and fails to trigger acid-induced salivation. In the postprandial period, the ph of the contents of the body of the stomach will be above this threshold of ph 2.5, and thus it is surprising that reflux symptoms are common at this time. While performing dual gastric and esophageal ph monitoring, we observed that the ph of the esophageal refluxate after meals was frequently more acidic than the ph in the body of the stomach. In this study, we formally analyzed dual ph recordings and confirmed this observation. To further investigate this phenomenon, we studied the ph in different regions of the stomach and at the gastroesophageal junction (GEJ) under both fasting and postprandial conditions. These studies in healthy volunteers show that after a meal there is an unbuffered pocket of acid at and immediately above the gastroesophageal squamocolumnar junction. The aims of this study were (1) to determine the relationship of the ph of the esophageal refluxate to that of the stomach both under fasting conditions and after a meal; (2) to determine the ph in the most proximal Abbreviations used in this paper: GEJ, gastroesophageal junction; LES, lower esophageal sphincter by the American Gastroenterological Association /01/$35.00 doi: /gast

2 776 FLETCHER ET AL. GASTROENTEROLOGY Vol. 121, No. 4 region of the stomach relative to that more distally both under fasting conditions and after a meal (a region of unbuffered high acidity was identified in the most proximal stomach after eating); (3) to identify the location of this unbuffered acid pocket in relation to the gastroesophageal squamocolumnar junction; and (4) to determine whether acidic gastric juice can partition on top of a homogenized meal, forming an unbuffered acid layer. Subjects and Methods Studies of the Relationship of Esophageal Refluxate ph to Intragastric ph Under Fasting Conditions and After a Meal Dual gastric and esophageal 24-hour ph recordings of 28 dyspeptic patients were analyzed retrospectively. These patients all had negative results of tests for Helicobacter pylori and normal findings on upper gastrointestinal endoscopy. Their mean age was 45 (range, 22 68) years; 12 were men and 16 women. Esophageal ph monitoring revealed a mean total percent of time with ph 4 of 5.7%, with half the group having abnormal esophageal acid exposure. Another 12 patients were recruited prospectively, and their 24-hour ph recordings were analyzed in an identical manner. These patients also had negative H. pylori results and normal endoscopic findings. Their mean age was 52 (range, 37 73) years, and half were men. Esophageal ph monitoring revealed a mean total percent of time with ph 9.3%, with two thirds of this group having abnormal esophageal acid exposure. ph recordings were performed using a dual-channel ph catheter with 2 antimony electrodes (Synectics Medical, Enfield, UK) spaced 15 cm apart and connected to a digital data logger (Diggitrapper MK III; Synectics). The electrodes were calibrated in buffer before and after the procedure, and electrode drift was considered acceptable if it was 0.5 ph units. No significant electrode drift was seen. The data was downloaded onto a personal computer and analyzed by the Esophogram 5.4 software (Synectics). The procedure for ph monitoring was identical in the 28 patients analyzed retrospectively and the 12 patients analyzed prospectively. Both groups fasted to have the lower esophageal sphincter (LES) identified by station pull-through using a water-perfused manometry catheter (Mui Scientific, Mississauga, Ontario, Canada). A dual-channel ph catheter was then placed with the proximal electrode 5 cm above the upper border of the LES in the esophagus and the distal electrode 15 cm below this point recording from the gastric body. Esophageal and intragastric ph was recorded for 24 hours. Times of meals and supine periods were noted. Antisecretory medication was withheld for 3 weeks before each examination. Analyses of dual ph recording. The upright fasting and postprandial buffered periods were analyzed separately to allow comparison. The supine period was excluded because of the tendency to sudden increases in intragastric ph. This tendency is unrelated to food buffering and remains a poorly explained phenomenon. 7 The postprandial buffered period was defined as starting when the gastric ph increased by more than 0.5 ph units above the baseline fasting gastric ph and finishing when it fell below this threshold ph. An episode of gastroesophageal reflux was defined as starting when the esophageal ph fell below ph 4 and finishing when it rose above this threshold ph. Each upright reflux episode was examined to determine the minimum esophageal ph during that reflux episode and the corresponding minimum gastric ph over the 60 seconds leading up to this esophageal ph minimum. For each patient, the median esophageal ph minimum (refluxate minimum ph) and median gastric ph minimum (gastric minimum ph) were calculated for fasting reflux episodes and for reflux episodes occurring during meal-related gastric buffering. Studies of the ph of the Most Proximal Stomach Compared With That More Distally Under Fasting Conditions and After a Meal Ten consecutive healthy volunteers with no history of dyspepsia were recruited. Their mean age was 34 (range, 18 51) years, and 5 were men. They fasted before the procedure and had the dual-channel ph catheter passed via the nose into the stomach so that both electrodes were recording gastric acid ph. At this point the proximal electrode was 50 cm from the nostril and the distal electrode was 65 cm from the nostril. An abdominal x-ray was taken to verify that the catheter was in the correct position. With the subject semirecumbent, the ph catheter was gently withdrawn by 1-cm increments every minute. The Diggitrapper event marker was used to mark each 1-cm change in catheter position. Catheter withdrawal continued in this manner until both electrodes were recording esophageal ph. At this point the ph catheter was advanced to restore the original position with both electrodes in the stomach, and a second abdominal radiograph was taken to confirm the catheter position. The subjects were then given a meal of french fried potatoes and either fish (n 5), smoked sausage (n 4), or chicken (n 1). Fifteen minutes after completion of the meal, a second ph catheter pull-through was carried out, again pausing at 1-cm intervals. The distance from the nostril to the point at which the ph reading increased from gastric ph to esophageal ph (the ph step-up) was identified for the 2 electrodes on both the fasting and postprandial pull-through. Fifteen separate ph data points were recorded over 1 minute at each electrode position during the pull-through. For each electrode position, these data points were summarized as the minimum ph (ph min) and mean ph (ph mean). Studies to Identify the Location of the Unbuffered Acid Pocket Relative to the Esophageal Squamocolumnar Junction Ten healthy volunteers (4 of whom had participated in the previous part of the study), mean age 32 (range, 19 51) years, with no history of dyspepsia were recruited. They fasted

3 October 2001 UNBUFFERED ACID AT THE GASTROESOPHAGEAL JUNCTION 777 before undergoing upper gastrointestinal endoscopy, which was carried out under conscious sedation with intravenous midazolam. Two metal clips, 180 apart, were attached to the mucosa at the squamocolumnar junction. We used standard hemostatic clips (MD-59; Keymed/Olympus Southend-on-Sea, UK) that close taking a firm grip of mucosa. The clips are deployed using an endoscopic clip-fixing device (Olympus). After each subject had recovered from sedation, a 2-channel ph catheter was passed via the nose into the stomach so that both electrodes were reading gastric ph. With the subject semirecumbent, the ph catheter was withdrawn using the pull-through technique described in part B. The catheter position was fixed with tape when the distal electrode was judged to be at the ph step-up. This was defined as the catheter position at which a distinct increase was seen from gastric to esophageal ph. A supine abdominal radiograph was taken, and the distance between the ph electrode and the endoscopic clips was measured. The ph catheter was then repositioned with both electrodes in the stomach and the subjects were given a meal of french fried potatoes and battered fish. Fifteen minutes after completion of the meal, the ph pull-through was repeated and another abdominal radiograph was taken with the ph catheter fixed at the ph step-up. The abdominal radiograph was again examined to determine the location of the ph electrode relative to the clips. Abdominal radiographs were also examined to assess the position of the ph electrode relative to the gastric air bubble. As in the previous procedure, the ph data recorded at each position were summarized as ph min and ph mean. In Vitro Studies to Determine Whether Acidic Gastric Juice Can Partition on Top of a Homogenized Meal Forming an Unbuffered Acid Layer An in vitro study was performed to determine whether acidic gastric juice could remain on top of a homogenized meal and result in an unbuffered layer. For this, a meal of french fried potatoes and fish was blended and mixed with water to form the consistency of porridge. The meal was then placed in a transparent container. Freshly aspirated pentagastrin-stimulated human gastric juice was applied with a dropper onto the food mixture, and a 2-cm layer of acid formed on its surface. A ph electrode was used to measure ph in the container. The electrode was positioned in the surface layer of gastric juice and then pushed deeper into the meal until it reached the bottom of the container. The beaker was then gently agitated for 60 seconds until it had a homogenous appearance. The ph electrode was once again positioned in the surface layer and again pushed through the mixture until it reached the bottom of the container. Statistics The ph data for reflux episodes were summarized as median values with SEM. Reflux episodes were compared using the Mann Whitney U test. The ph data for each electrode position were compared with the ph data 1 cm below the GEJ using the Mann Whitney U test because these data also were not distributed normally. The change in ph step-up distance was compared using the Wilcoxon signed rank test. Ethics The study was approved by the North Glasgow University NHS Trust Ethics Committee, and each subject gave written informed consent. Results Relationship of Esophageal Refluxate ph to Intragastric ph Under Fasting Conditions and After a Meal The results of the retrospective and prospective studies were similar and are therefore presented together. (The results can be identified separately in Figure 1.) The intragastric ph electrodes showed the characteristic pattern with a median fasting ph of 1.3 (range, ) and an increase in ph after each meal. The mean duration of meal-related buffering was 5 hours (range, 1:00 9:10 hours). The mean peak ph during this time was 5.2 (range, ). The esophageal electrodes mainly recorded a ph above 4 for the 24-hour period. The mean total percent of time ph was 4 was 6.8%. During the fasting period, a mean of 29 (range, 0 66) gastroesophageal reflux episodes were recorded per patient. The median gastric minimum ph during these fasting reflux episodes was 1.3 (range, ). The median refluxate minimum ph during the fasting reflux Figure 1. Refluxate minimum ph and corresponding gastric minimum ph under fasting conditions and during meal buffering. Gastric minimum ph is the median value of the lowest gastric ph measured during reflux episodes per patient; refluxate minimum ph is the median value of the lowest esophageal ph measured during reflux episodes per patient. F, Analyzed retrospectively; E, analyzed prospectively.

4 778 FLETCHER ET AL. GASTROENTEROLOGY Vol. 121, No. 4 episodes was 2.5 (range, ). The esophageal refluxate was significantly less acidic than the intragastric contents during the fasting period (P 0.01; Figure 1). During the postprandial gastric buffered period, a mean of 22 (range, 1 78) gastroesophageal reflux episodes were recorded per patient. The median gastric minimum ph over the 1 minute before these postprandial reflux episodes was 3.2 (range, ). The median refluxate minimum ph during these postprandial reflux episodes was 3.0 (range, ; Figure 1). The ph recorded from the gastroesophageal refluxate was more acidic than that recorded from the stomach over this postprandial period, although the difference was not statistically significant. Most of these reflux episodes occurred after the meals had finished, so the lower esophageal ph cannot be explained by ingestion of acidic drinks. The buffering effect of the meal caused a marked increase in the intragastric ph during reflux episodes. The median gastric minimum ph during fasting reflux episodes was 1.3, compared with 3.2 during postprandial reflux episodes. In contrast, the buffering effect on the acidity of the gastroesophageal refluxate was much less apparent. The median reflux minimum ph during fasting was 2.5 versus 3.0 after a meal (Figure 1). ph of the Most Proximal Stomach Compared With That More Distally Under Fasting Conditions and After a Meal From 10 subjects, 20 fasting ph pull-throughs and 19 postprandial pull-throughs could be analyzed (the proximal electrode could not be readvanced into the stomach postprandially because of discomfort in 1 subject). Figure 3. Median ph min for each electrode position in the stomach and esophagus after a meal. The electrode positions were set 1 cm apart. ph min is the lowest ph recorded over 1 minute at each electrode position. *ph greater than at the electrode position marked (P 0.01). In the fasting state, stable gastric ph was recorded throughout the stomach (median, 1.4; range, ). A clear step-up in ph corresponding to the GEJ could be seen in all pull-throughs and occurred at a median distance of 44 cm from the nostril (range, cm). The ph step-up of the GEJ remained apparent after a meal, but its position had changed. After the meal, the ph step-up occurred at a median distance of 42 cm from the nostril (range, 37 46) which was 2 cm more proximal than that observed under fasting conditions (P 0.05). This proximal migration of the ph step-up was observed in each of the 10 subjects (Figure 2). After the meals, gastric buffering was not uniform throughout the stomach. The maximum increase in ph was seen 6 cm distal to the postprandial ph step-up (median ph 4.7). Beyond this point, the extent of buffering decreased progressively (Figure 3). In addition, a region of high acidity was identified immediately distal to the postprandial ph step-up (Figures 3 and 4). The median ph min for electrode positions 1 and 2 cm below the postprandial ph step-up was significantly more Figure 2. The distance from the nostril at which the ph probe changed from reading a low gastric ph to a neutral esophageal ph under fasting and postprandial conditions. The distances recorded in each individual are correct to 1 cm. Median values are indicated by horizontal bars. Postprandial distances are less than fasting distances (P 0.05). Figure 4. Example of a ph tracing recorded from an electrode during the catheter pull-through technique in 1 subject while fasting and again after a meal. The postprandial recording shows a region of high acidity corresponding to the location of the ph step-up point observed under fasting conditions.

5 October 2001 UNBUFFERED ACID AT THE GASTROESOPHAGEAL JUNCTION 779 acidic than any of the more distal electrode positions (Figure 3). The median ph min 1 cm below the postprandial step-up point was 1.6 (range, 0.9 5), compared with a median ph min 4 cm below the postprandial step-up point of 4.4 (range, ; P 0.01). The same pattern was seen when the median ph mean for each electrode position was analyzed. The ph just distal to the ph step-up point after a meal was similar to the gastric fasting ph, suggesting that this region was not exposed to the buffering influence of the meal. This region of unbuffered acid (ph 2) was observed on a median of 2 (range, 1 4) consecutive 1-cm increment electrode positions. This indicated that it extended for a median distance of 2 cm. This pocket of unbuffered acid was identified in 9 of the 10 healthy volunteers studied. In the remaining volunteer, food-related ph buffering was not seen after the meal; the intragastric ph remained below 2 right up to the ph step-up point. In each subject, the ph profiles recorded by the proximal and distal ph electrode were similar (Figure 5). This finding confirmed that the changes in ph were related to the position of the electrode rather than changes in intragastric ph over time. These studies indicated that the ph step-up at the GEJ migrated 2 cm proximally after the meal. The mechanism behind this migration was unclear. It could be caused by the GEJ being displaced upward and thus both the ph step-up and the squamocolumnar junction migrating proximally together. Alternatively, it is possible that the distention of the stomach was pulling apart the cardia and/or the distal LES and thereby resulting in the ph step-up moving proximally relative to the squamocolumnar junction. The above studies also clearly demonstrate an unbuffered acid pocket in the region of the GEJ postprandially. However, the exact location of this pocket relative to the squamocolumnar junction is unclear. To clarify the mechanism of this proximal migration of the ph step-up postprandially and the location Figure 5. Mean ph at each electrode position during the postprandial pull-through of both the proximal and distal electrode in 1 individual. A pocket of unbuffered acid is seen in the region of the GEJ consistently with both electrodes. of the postprandial acid pocket, we undertook further pull-through studies in 10 subjects whose squamocolumnar junction was marked with metal clips at endoscopy. Studies to Identify the Location of the Unbuffered Acid Pocket Relative to the Esophageal Squamocolumnar Junction From 10 subjects whose squamocolumnar junction had been marked with metal clips, 19 fasting and 19 postprandial ph pull-throughs could be analyzed (the proximal electrode could not be advanced into the stomach because of discomfort in 1 subject). A distinct pocket of unbuffered acid was seen just distal to the ph step-up in 6 of 10 subjects after a meal. In these 6 subjects, a pocket of acid was apparent on both the proximal and distal electrodes on the ph catheter. An electrode position was deemed to reflect an acid pocket if it recorded a ph 2. In 2 subjects, no acid pocket was seen with a high intragastric ph stepping up to intraesophageal ph at the GEJ. In another 2 subjects, there was no buffering effect of the meal, and the intragastric ph remained 2 right up to the ph step-up point. Analysis of ph data from all 10 subjects showed a region of unbuffered acid (ph 2) starting at the postprandial ph step-up point and extending over a median of 2 (range, 1 4) consecutive electrode positions (1-cm increments), indicating that the acid pocket extended for a median distance of 2 cm. The median ph min 1 cm below the postprandial step-up point was 1.6 (range, ) compared with a median ph min 3 cm below the postprandial step-up point of 4.4 (range, ; P 0.01). The location of the ph step-up distal to the nostril under fasting conditions and after a meal was measured using the 1-cm distance markers on the ph catheter. Under fasting conditions, a clear step-up in ph could be seen in all subjects when the ph catheter was moved from gastric ph to esophageal ph. This occurred a mean of 45.5 (range, 43 50) cm from the nostril. After the meal, the ph step-up occurred a mean of 43.9 (range, 41 47) cm from the nostril and had therefore moved proximally by a mean of 2 cm (range, 0 4 cm; P 0.05). This pull through technique using 10-mm increments has an inherent inaccuracy in determining the exact distance to the ph step-up point. If the ph electrode starts at any position from 1 to 9 mm below the ph step-up point, it will register a step-up in ph if withdrawn by 10 mm. The measured distance from the nostril on the ph catheter will then represent a 1 9-mm underestimate of the actual distance to the ph step-up point. Because the electrode starting position below the ph step-up point is random, we expect the average underesti-

6 780 FLETCHER ET AL. GASTROENTEROLOGY Vol. 121, No. 4 mate to be 5 mm. Correcting for this gives the ph step-up occurring at 46 cm fasting and 44.4 cm after the meal. The distance of the squamocolumnar junction from the nostril was derived by measuring the distance on radiography between the metal clips (marking the squamocolumnar junction) and the ph electrode and adding this distance to the position of the ph electrode, which was known from the distance markers on the ph catheter. The radiography was taken immediately after the ph step-up occurred both under fasting conditions and after the meal. This measurement was possible in 9 of the 10 subjects (1 subject did not have radiographs). Under fasting conditions, this was a mean of 46.3 (range, ) cm from the nostril. After the meal, the squamocolumnar junction lay a mean of 46.2 (range, ) cm from the nostril, indicating no change in its position. These measurements indicate that under fasting conditions, the location of the ph step-up corresponded to that of the squamocolumnar junction. However, after the meal the ph step-up moved proximally (by 1.6 cm), whereas the position of the squamocolumnar junction remained the same, resulting in the ph step-up point occurring on average 1.8 cm proximal to the squamocolumnar junction (Figure 6). The proximal movement of the ph step-up indicated that there was opening of the distal esophagus after a meal, allowing acid to extend up onto squamous mucosa. The unbuffered acid pocket clearly demonstrated in 6 of the 10 subjects included in this part of the study occurred immediately distal to the location of the ph Figure 6. Distance from the nostril to the squamocolumnar junction and ph step-up before and after a meal in the 10 patients who had their squamocolumnar junctions marked with clips. The squamocolumnar junction was measured using the radiograph of the distal electrode position in 9 subjects (the ph step-up point of the proximal electrode was not identified by radiography). The ph step-up was measured using both electrodes in 9 subjects and 1 electrode in 1 subject. *Distance from nostril to postprandial ph step-up less than under fasting conditions (P 0.01). step-up. The ph of the acid pocket remained 2 for a mean of 2 consecutive 1-cm increment electrode positions. This suggests that it actually extended a mean distance of 2 cm. The mean proximal migration of the ph step-up in these 6 subjects was 1.7 cm, implying that the unbuffered acid pocket was located at the squamocolumnar junction and extended approximately 1.7 cm proximally and 0.3 cm distally to it. The postprandial radiographs did not show any signs of a gastric air bubble in the region of the squamocolumnar junction or proximal acid pocket. In Vitro Studies of ph Partitioning Within the Stomach After a Meal When gastric juice was poured onto the beaker containing the homogenized meal, it formed a separate layer floating on the surface of the meal. As the ph electrode was pushed down through the beaker contents before agitation, it recorded a ph of 0.6 in the layer of surface acid, increasing to a ph of 6.2 deeper into the food mixture. After 60 seconds of gentle agitation, the mixture of food and acid looked homogeneous. Despite this appearance, layering of gastric acid was shown to persist. The second push through of the ph electrode revealed a ph of 1.7 on the surface, increasing to 3.7 deeper into the food mixture. This shows the existence and persistence of an unbuffered layer of acid on top of the food mixture (Figure 7). Discussion Gastroesophageal reflux is probably the most common chronic gastrointestinal disorder. 2 Reflux of gastric contents into the esophagus is most frequent after meals because of an increase in the number of transient LES relaxations during this period. 1,8 Meals would be expected to reduce the noxious effects of gastric juice by increasing its ph through their buffering effect. However, our current studies indicate that the gastric juice closest to the squamocolumnar junction largely escapes the buffering effect of food. Our studies also indicate that this pocket of unbuffered gastric acid actually traverses the squamocolumnar junction after a meal, exposing the distal esophageal mucosa to highly acidic gastric juice. Our interest in intragastric ph near the squamocolumnar junction was stimulated by the observation that after a meal the ph of esophageal refluxate was frequently more acidic than that of the body of the stomach. We first formally analyzed dual gastric and esophageal ph traces. Under fasting conditions, the combined intragastric and intraesophageal monitoring studies showed that the acidity of the esophageal refluxate was

7 October 2001 UNBUFFERED ACID AT THE GASTROESOPHAGEAL JUNCTION 781 Figure 7. ph recorded from an electrode at the top and bottom of a beaker containing a homogenized meal onto which acid gastric juice has been poured. (A) Before and (B) after agitation of the mixture for 1 minute. considerably less than that of the gastric juice. The median ph of gastroesophageal refluxate under fasting conditions was 2.5, whereas the simultaneous intragastric ph was 1.4, representing a 10-fold higher hydrogen ion concentration. To reach the esophageal electrode, gastric refluxate will have had to travel 5 cm up the esophagus. In doing so it comes into contact with bicarbonate derived from swallowed saliva, leading to a gradual increase in refluxate ph. Simultaneous ph measurement at 5, 11, and 17 cm above the LES has confirmed that during episodes of acid reflux the ph nadir progressively increases as the electrode position moves further away from the squamocolumnar junction. 9 Analyses of the dual gastric and esophageal ph traces after eating showed a different pattern. After eating, the median intragastric ph increased and remained elevated for 1 2 hours. During the period of postprandial buffering, the esophageal refluxate ph was similar to the intragastric ph rather than being 10 times less acidic as it was under fasting conditions. We also found that the ph of the esophageal refluxate was similar under fasting and postprandial conditions and was unlike the intragastric ph, which was markedly increased during the postprandial buffered period. Some of this difference in the relationship of esophageal refluxate ph to intragastric ph under fasting versus postprandial conditions could be artifactual. Reflux episodes were defined on the basis of ph being 4. As a consequence, we will have excluded reflux episodes with a refluxate ph 4, which may have been more likely to occur during the postprandial period when intragastric ph was higher. However, examination of individual traces showed numerous examples of the refluxate being much more acidic than the intragastric ph. This raised the possibility that there might be a region in the proximal stomach that escaped the buffering effects of ingested food and was able to reflux into the esophagus during the postprandial period. Further studies were undertaken to directly investigate this possibility. In 10 healthy volunteers we measured the ph throughout the stomach and distal esophagus by slowly withdrawing ph electrodes. Under fasting conditions, the intragastric ph was low and showed little change on passing through the stomach before showing a steep increase on entering the esophagus. A different ph pattern was seen after the meal. The buffering effect of the meal was maximal in the body region of the stomach and became less on moving distally into the antral region. In addition, in each subject there was a region of low ph just before the ph step-up on entering the esophagus. Previous studies have observed differences in the ph of the body and antral regions of the stomach during the postprandial period. 4,10,11 These studies have shown that the increase in postprandial ph is most marked in the body region compared with the antrum. The more pronounced buffering effect seen in the body region may be explained by the retention of solid food in this part of the stomach. 12 Our current observation confirms and extends the findings of previous studies by showing that the most proximal cardia region of the stomach largely escapes the buffering effect of meals. The presence of a local pocket of unbuffered acid in the most proximal stomach is consistent with recent observations with esophageal impedence catheters. These have shown that after a meal there can be episodes of nonacid (i.e., buffered gastric juice) reflux closely interspersed with episodes of acid reflux. 13,14 An additional finding from our studies of the intragastric ph profile was that the location of the esophageal ph step-up was approximately 2 cm more proximal after a meal than under fasting conditions. This area in the region of the squamocolumnar junction that registered a neutral ph under fasting conditions but an acid ph postprandially corresponded to the location of the proximal pocket of unbuffered acid observed after a meal. Further studies were therefore undertaken to identify the location of the fasting and postprandial ph step-up points and postprandial unbuffered acid pocket relative to the gastroesophageal squamocolumnar junction. These studies were performed by endoscopically attaching radioopaque metal clips to the squamocolumnar junction to indicate its position. Under fasting conditions, the ph step-up corresponded almost exactly to the esophageal squamocolum-

8 782 FLETCHER ET AL. GASTROENTEROLOGY Vol. 121, No. 4 nar junction. After a meal, the ph step-up migrated proximally by a mean of 1.6 cm. In contrast, the location of the squamocolumnar junction did not change after a meal. This indicated that after the meal, acidic gastric juice was extending above the squamocolumnar junction on to the esophageal mucosa. Therefore, these further studies indicated that the proximal unbuffered pocket of acid observed after the meal was actually traversing the squamocolumnar junction and extended from the cardia region of the stomach to the distal esophagus. The further studies allowed us to identify the location of the postprandial unbuffered acid pocket not just in subjects with the metal clips in situ but also in the 10 subjects who had been studied without the clips. This was because the clip study showed that the squamocolumnar junction corresponded to the fasting ph step-up point and did not change its distance from the nostril after a meal. Using the fasting step-up to indicate the squamocolumnar junction allowed us to deduce the position of the unbuffered acid pocket in the 10 subjects without squamocolumnar junction clips and showed that the acid pocket in this group also extended from the cardia up into the distal esophagus (Figure 4). The pattern of ph seen with the 2 groups of subjects undergoing the ph catheter pull-through technique was similar. This suggests that the endoscopy procedure and the application of metal clips did not influence this phenomenon in the second group of subjects. These new findings raise a number of questions. The first is the mechanism explaining our observation of the exposure of the most distal esophageal mucosa to acid after a meal. The exposure of this squamous mucosa to acid was not a transient phenomenon because it was clearly observed on each ph electrode traversing the region 10 minutes apart. This feature and the fact that it extended a mean of only 1.8 cm above the squamocolumnar junction indicate that it cannot be caused by transient relaxations of the LES. It seems more likely to be caused by the postprandial distention of the proximal stomach opening the LES and thereby exposing the distal esophagus to gastric acid. Oberg et al. 15 proposed such a hypothesis in 1997 when reporting evidence linking intestinal metaplasia of cardiac mucosa to reflux disease. Manometric studies of the LES during distention of the stomach with air show shortening of the length of the high-pressure zone and reveal that the sphincter opening starts in the distal portion of the sphincter before spreading proximally to involve the remainder of the LES. 16 These observations would again suggest that distending the stomach will tend to pry open the distal LES and allow gastric acid to reach the squamous mucosa of the distal esophagus. The mechanism by which the region around the squamocolumnar junction escapes the buffering effect of food also requires consideration. Recent magnetic resonance imaging studies have shown marked partitioning of food and its constituents in the stomach. 17 The lipid component forms a layer floating on top of the more aqueous phase. We wondered whether acid secreted by the proximal stomach might be trapped above this lipid layer and thus be separated from the buffering effect of the food. To investigate this possibility, we performed in vitro studies using a homogenized meal. These confirmed that acidic human gastric juice added on top of the homogenized meal escaped the buffering effect of the food and that the ph gradient persisted even after agitation of the mixture. The most proximal part of the stomach is relatively quiescent after a meal, with the motor contractions arising distal to this point and then moving away toward the antrum. 18 This would facilitate persistence of an acid pocket in this site. Our study has made 2 new observations that are likely to be important in the etiology of the high prevalence of disease at the squamocolumnar junction. The first is that gastric juice in the most proximal stomach escapes the buffering effect of food and remains highly acidic during the postprandial period. The second is that this proximal pocket of unbuffered acid actually traverses the esophageal squamocolumnar junction and overlies both the gastric cardia mucosa of the stomach and the squamous mucosa of the distal esophagus. It is now recognized that both the cardia mucosa and the distal esophageal mucosa are common sites of mucosal inflammation and intestinal metaplasia as well as of increasing incidence of carcinoma. 19,20 None of our healthy volunteers complained of heartburn at the time of clear evidence of acid extending above their squamocolumnar junction. Instillation of acid onto esophageal mucosa induces symptoms in patients with reflux disease but in only approximately 15% of healthy volunteers. 21,22 However, there is a delay of approximately 6 minutes between the beginning of the infusion and the development of heartburn. In addition, induction of symptoms requires the infusion of highly acidic HCl of ph 1. Instillation of acid with ph 2.5 rarely induces symptoms. 5 The induction of symptoms might also depend on the area of esophagus exposed to acid, and exposure of the most distal esophagus may not be sufficient to induce symptoms. This could explain why significant disease in the region of the squamocolumnar junction, including intestinal metaplasia and cancer, frequently presents without a preceding history of heartburn or other symptoms. 19,23 Further understanding of the mechanism of the development of the proximal unbuffered acid pocket and its extension onto the distal esophagus may open up new

9 October 2001 UNBUFFERED ACID AT THE GASTROESOPHAGEAL JUNCTION 783 approaches to the management of diseases at the squamocolumnar junction. The possible role of the fat content of the meal as well as postprandial posture needs to be investigated. After most meals there is probably sufficient buffer present to increase the ph of the entire gastric contents to nondamaging levels but it needs to be more evenly distributed throughout the stomach. Considerable attention has been given to the role of transient LES relaxations in the etiology of esophageal disease, and therapies are being developed to reduce their frequency. Our current studies indicate that opening of the distal esophageal sphincter, allowing short-segment reflux after meals, may also be important and may require different therapeutic strategies. The presence of an unbuffered acid pocket at the squamocolumnar junction also has implications with respect to monitoring of adequacy of ph control by antisecretory medication in reflux disease. Intragastric ph studies are often performed in such patients to determine that the ph is 4. However, our present studies show that the ph may remain low near the squamocolumnar junction despite being 4 in the body of the stomach, where ph electrodes are normally positioned. The contents of the stomach are often regarded as a homogenous entity. However, our current studies indicate that the contents of the most proximal stomach may escape the buffering effect of food, providing a source of highly acidic gastric juice available for reflux into the esophagus during episodes of LES relaxation. In addition, our studies indicate that this pocket of unbuffered acid may itself extend from the cardia across the squamocolumnar junction. These observations are likely to be relevant to the high prevalence of mucosal pathology recognized to occur at, just above, and just below the squamocolumnar junction. References 1. Dodds WJ, Dent J, Hogan WJ, Helm JF, Hauser R, Patel GK, Egide MS. Mechanisms of gastroesophageal reflux in patients with reflux esophagitis. N Engl J Med 1982;307: Nebel O, Fornes M, Castell DO. Symptomatic gastroesophageal reflux: Incidence and precipitating factors. Dig Dis Sci 1976;21: Knill-Jones RP. A formal approach to symptoms in dyspepsia. Clin Gastroenterol 1985;14: McLaughlan G, Fullarton GM, Crean GP, McColl KE. Comparison of gastric body and antral ph: a 24 hour ambulatory study in healthy voluteers. Gut 1989;30: Smith JL, Opekun AR, Larkai E, Graham DY. Sensitivity of the esophageal mucosa to ph in gastroesophageal reflux disease. Gastroenterology 1989;96: Sudhir K, Dutta HBM, Meirowitz RF, Vaeth J. Modulation of salivary secretion by acid infusion in the distal oesophagus in humans. Gastroenterology 1992;103: Verdu EF, Fraser R, Murphy G, Blum AL, Armstrong D. The origin of nocturnal intragastric ph rises in healthy subjects. Scand J Gastroenterol 1995;30: Holloway RH, Kocyan P, Dent J. Provocation of transient lower esophageal sphincter relaxations by meals in patients with symptomatic gastroesophageal reflux. Dig Dis Sci 1991;36: Shaker R, Dodds WJ, Helm JF, Kern MK, Hogan WJ. Regional esophageal distribution and clearance of refluxed gastric acid. Gastroenterology 1991;101: Barlow AP, Hinder RA, DeMeester TR, Fuchs K. Twenty-four-hour gastric luminal ph in normal subjects: influence of probe position, food, posture, and duodenogastric reflux. Am J Gastroenterol 1994;89: Fisher RS, Sher DJ, Donahue D, Knight LC, Maurer A, Urbain JL, et al. Regional differences in gastric acidity and antacid distribution: is a single ph electrode sufficient? Am J Gastroenterol 1997;92: Collins PJ, Houghton LA, Read NW, Horowitz M, Chatterton BE, Heddle R, Dent J. Role of the proximal and distal stomach in mixed solid and liquid meal emptying. Gut 1991;32: Sifrim D, Holloway RH, Tack J, Silny J, Lerut T, Janssens J. Transient Les relaxations: are they more frequent in patients with gastroesophageal reflux disease? Gastroenterology 1999;116:A Vela M, Camacho-Lobato L, Hatlebakk J, Katz PO, Castell DO. Effect of omeprazole (PPI) on ratio of acid to nonacid gastroesophageal reflux. Studies using simultaneous intraesophageal impedance and ph measurement. Gastroenterology 1999;116:A Oberg S, Peters JH, DeMeester TR, Chandrasoma P, Hagen JA, Ireland AP, Ritter MP, Mason RJ, Crookes P, Bremner CG. Inflammation and specialized intestinal metaplasia of cardiac mucosa is a manifestation of gastroesophageal reflux disease. Ann Surg 1997;226: Kahrilas PJ, Shi G, Manka M, Joehl RJ. Increased frequency of transient lower esophageal sphincter relaxation induced by gastric distention in reflux patients with hiatal hernia. Gastroenterology 2000;118: Boulby P, Gowland P, Adams V, Spiller R. Use of echo planar imaging to demonstrate the effect of posture on the intragastric distribution and emptying of an oil/water meal. Neurogastroenterol Motil 1997;9: Kelly KA. Motility of the stomach and gastroduodenal junction. In: Johnson LR, ed. Physiology of the gastrointestinal tract. New York, NY: Raven, 1981: Trudgill NJ, Suvarna SK, Kapur KC, Riley SA. Intestinal metaplasia at the squamocolumnar juntion in patients attending for diagnostic gastroscopy. Gut 1997;41: Blot WJ, Devesa SS, Kneller RW. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. JAMA 1991;265: Behar J, Biancani P, Sheahan DG. Evaluation of esophageal tests in the diagnosis of reflux esophagitis. Gastroenterology 1976; 71: Hewson EG, Sinclair JW, Dalton CB, Wu WC, Castell DO, Richter JE. Acid perfusion test: does it have a role in the assessment of non cardiac chest pain? Gut 1989;30: Cameron AJ, Lomboy CT, Pera M, Carpenter HA. Adenocarcinoma of the esophagogastric junction and Barrett s esophagus. Gastroenterology 1995;109: Received October 16, Accepted June 13, Address requests for reprints to: Kenneth E. L. McColl, M.D., Department of Medicine and Therapeutics, Gardiner Institute, Western Infirmary, Glasgow, G11 6NT Scotland. k.e.l.mccoll@clinmed.gla. ac.uk; fax: (44)