- Email: [email protected]
Barrett's esophagus: Too much acid, alkali, or both?
798 SELECTED SUMMARIES
clinical disease states, such as chronic hepatitis B. In this report nomenclature also was problematic, and the terms used in the “clinical” diagnoses did not match those used for “final” diagnosis. For example, the patients were diagnosed clinically as mild or severe chronic hepatitis B, mild or severe chronic hepatitis (not further specified), or chronic non-A, non-B hepatitis. The same group after histology was obtained included the diagnoses of chronic persistent hepatitis B, chronic active or chronic persistent non-A, non-B hepatitis, or delta hepatitis. Some rather vague clinical diagnoses were used, for example cholestatic liver disease. The final diagnoses were more precise, as might be expected. This study of liver biopsy should not be taken as a guide to be applied to standard work-up in a nonstudy setting. For example, not every patient needs an antimitochondrial antibody or HIV titer, as was done here. Also, recent work would suggest that cryptogenic liver disease may be associated with the heterozygous state for a,-antitrypsin (Stand J Gastroenterol1985;20:835-842). Therefore, if this diagnosis is under consideration, and it probably should be in most patients with asymptomatic elevations in transaminase, a quantitative level of a,-antitrypsin is not sufficient, or even necessary, and a Pi phenotype must be determined. One test that was used here, the alkaline phosphatase of hepatic origin, is unclear to me. Does this mean only elevated alkaline phosphatases associated with an elevated gamma glutamyl transpeptidase (GGTP)? Oddly, there are some lapses in editing in this paper. Tables 2 and 3 show the number of patients with fatty liver as 27, giving a total patient number of 100, whereas the text lists the number of 17, presumably the correct number to add up to the total number of 96 patients studied. Hemochromatosis is spelled in two different ways. The major difficulty with this report is that the authors do not tell us how they actually calculated the data on the accuracy of the clinical diagnosis as determined by the biopsy diagnosis. What we really want to know is how many of the patients changed group after the results of their biopsy were known. For example, we know that 25 carried a diagnosis of alcoholic liver disease and that 26 had it shown on biopsy. Was it the same 25 patients plus one extra? The miscellaneous group grew by 4 patients when biopsy confirmation was in hand. From which group did these subjects come? Attempts to back calculate to these numbers from their table giving percentages for sensitivity and predictive value were unsuccessful. Unfortunately, the authors have missed the opportunity to explore the real problems for specialists in liver disease remaining with liver biopsy as a diagnostic tool. The weak link in many settings is the interpretation of the pathology. There are several causes for these difficulties. First, many pathologists do not routinely do the multiple stains done by these authors, and considered standard at some centers. In interpreting liver biopsy material, an H&E stain only may leave many diagnostic clues hidden. Second, interpretation of liver biopsy material does require expertise and experience, which may be sorely lacking. Finally, the good working relationship between pathologist and clinician described here often does not exist. The pathologist cannot provide optimal interpretation with only limited clinical input. Another real problem with liver biopsy not addressed here is the question of diagnostic accuracy. It is known that biopsy is relatively good at providing a correct morphologic diagnosis, given a specimen of adequate size, a problem not addressed in this report. Unfortunately, etiology often remains obscure. For example, some of the cases of fatty liver described here may be indeed alcohol induced, as patients often minimize their alcohol intake. There are settings in which the biopsy may be wrong. For example, in macronodular cirrhosis, the needle may pass through a large regenerative nodule, yielding a diagnosis of normal tissue and misleading the clinician. Diagnostic accuracy is much higher in diffuse diseases such as alcoholic liver disease (Lancet 1986:1:523-525). Like this paper, liver biopsy remains imperfect. Nevertheless, we are gratified to be reassured that it is useful in assessing patients
GASTROENTEROLOGY
Vol. 98, No. 3
with chronically abnormal liver function tests. Also, we are particularly grateful that this important message has again been brought to the generalists who make up the audience of the Annals and are frequently faced with the problem of what to do with chronically abnormal liver tests. They should know that, when used with knowledge of its limitations, liver biopsy remains the gold standard for diagnosis in liver disease. C. A. RIELY, M.D.
BARRETT’S ESOPHAGUS: TOO MUCH ACID, ALKALI, OR BOTH? Mulholland MW, Reid BI, Levine DS, et al. [Departments of Surgery and Medicine, University of Washington, Seattle, Washington]. Elevated gastric acid secretion in patients with Barrett’s metaplastic epithelium. Digest Dis Sci 1989;34:13291335. Twenty consecutive patients with chronic, symptomatic gastroesophageal reflux, many of whom were referred to the University of Washington for known or suspected Barrett’s esophagus, were recruited for a study assessing gastric acid secretion in Barrett’s patients. Of the 20 patients, 17 agreed to undergo protocol investigations. Ten of these patients (9 men, 1 woman: mean age 59 + 4 years] were subsequently proven by endoscopic biopsy to have Barrett’s esophagus. The other 7 patients [2 men, 5 women; mean age 50 + 5 years) had gastroesophageal reflux without Barrett’s specialized metaplasia. Normal controls, matched with regard to age and sex for the first 10 reflux patients to enter the study, were chosen from a group of volunteers. The control group consisted of 5 men and 5 women, with a mean age of 50 k 4 years. Gastric acid secretion and release of endogenous gastrin were stimulated with a liquid peptone meal using the technique of intragastric titration in all subjects. Response to exogenous gastrin was assessed using dose-response studies with i.v. synthetic human gastrin 17-I. In addition to baseline studies performed in the unmedicated state, gastric acid secretion in response to exogenous human gastrin 17-I was measured while patients were receiving either 300 mg of cimetidine every 6 hours or 2.5 mg of cimetidine plus methscopolamine every 6 hours. Continuous 24-h esophageal pH monitoring was performed in all gastroesophageal reflux patients in the unmedicated state. Number of reflux episodes was counted and percentage of time with esophageal pH <4 was calculated. Basal acid output was significantly greater in the Barrett’s patients (8.56 + 1.82 mEq/h; x + SEM) than in the agematched controls (2.78 A 1.08 mEq/h), but not different from the patients with uncomplicated gastroesophageal reflux (5.58 * 2.90 mEq/h). Gastric acid secretion also was higher for the Barrett’s patients after each dose level of gastrin 17-I than either the reflux esophagitis group or normal control group. However, these latter differences were only statistically significant when the Barrett’s esophagitis group was compared with the controls. One-hour peak acid output stimulated by the peptone meal, serum gastrin release, both basally and with meal stimulation, and sensitivity to gastrin, expressed as the concentration causing half-maximal acid secretion, were similar among the study groups. Both reflux
March 1990
groups showed similar decreases in basal and gastrinstimulated acid secretion after the administration of cimetidine and methscopolamine. No significant differences were noted when patients with uncomplicated reflux were compared with the Barrett’s patients in terms of frequency or total time of reflux episodes during 34-h esophageal pH monitoring. The authors concluded that differences in acid secretion, both basally and in response to gastrin 17-1, may characterize patients with Barrett’s esophagus. They speculated that elevated basal acid production may contribute to the pathogenesis of this specialized metaplastic epithelium. However, the mechanisms responsible for their observations were not elucidated by the current study and remain an area for future investigation. Comments. Today it is well recognized that the Barrett’s esophagus is an acquired condition that occurs as a sequela of gastroesophageal reflux disease (GERD). What is not so obvious is why some patients develop esophagitis while others develop specialized columnar metaplasia of the distal esophagus. As a group, patients with Barrett’s esophagus have weaker lower esophageal sphincter (LES) pressure and more prolonged acid exposure than do patients with severe esophagitis [Arch Surg 1983;18:543-549; Br J Surg 1987;74:774776). However, there is much overlap among individuals within these groups such that neither manometry nor pH monitoring can be used as a diagnostic test for Barrett’s esophagus. Gastric emptying rates are also similar in Barrett’s patients when compared with healthy volunteers and esophagitispatients [Dig Dls Sci 1986;32:12171220). Therefore, we must begin to look past the underlying esophageal motility disorder at other potential causative factors for the development of Barrett’s epithelium. Another potential culprit might be the gastric contents bathing the distal esophagus, with acid being the most likely candidate. In fact, it is somewhat surprising that gastric secretion has only recently been studied in these patients. The Seattle group found basal acid output and the gastric acid secretion response to synthetic gastrin to be significantly greater in Barrett’s patients when compared with healthy volunteers, with uncomplicated esophagi& patients falling between these 2 groups. Although the numbers were small, their findings are similar to a preliminary report from Collen et al at Georgetown University(Gastroenterologyl987392:1350A].Thisgroup reported significantly higher basal acid output in 13 Barrett’s patients (11.9 + 4.6 mEq/h, *SD) compared with 59 patients with GERD (5.8 + 5.4 mEq/h) and 40 healthy volunteers (3.2 f 2.7 mEq/ h). Basal acid output was greater than IO mEq/h in 69% (9 of 13) of patients with Barrett’s esophagus compared with 24% (14 of 59) of GERD patients. Furthermore, these findings may have therapeutic importance as the same group has observed that 15 of 36 Barrett’s patients required more than the usual doses of ranitidine to effect healing of esophagitis (Gastroenterology 1988;94:73A]. Neither group reported the proportion of patients with underlying duodenal ulcer disease. Although intriguing, these studies may be flawed by important methodological problems. Patient groups tended to be age-matched, but the Barrett’s patients were nearly all men, while the esophagitis and control groups had a more equal sex distribution. This male preponderance may be important as the basal acid output is higher in men than in women (Gastroenterology 1984$6:843-8511. More importantly, the previous studies did not exclude associated duodenal ulcer disease as a confounding factor. Dr. Basil Hirschowitz has recently studied 14 Barrett’s patients comparing basal and maximal gastric acid and pepsin secretion with 114 esophagitis patients stratified for the presence of duodenal ulcer disease (39 with duodenal ulcer, 75 without duodenal ulcers] (Gastroenterology 1989;96:212A). Three patients with Barrett’s esophagus and duode-
SELECTED SUMMARIES
799
nal ulcer disease had basal acid outputs ranging from 8.7-11.5 mEq/h, comparable to the values reported in the 2 previous studies, but not different from control duodenal ulcer patients with or without esophagitis. However, the 11 Barrett’s patients without duodenal ulcer disease, all of whom were men, had a mean basal acid output of 2.9 + 0.7 mEq/h or slightly, but not significantly, below the mean values for the comparable nonulcer patients with esophagitis (3.5 + 0.5 mEq/h) or without esophagitis (4.4 * 0.5 mEq/ h). Pepsin secretion was also not different between these groups. Duodenogastroesophageal reflux may be another factor contributing to the genesis and complications in Barrett’s esophagus. In a dog model, prolonged acid or acid and bile, but not bile alone, produces columnar epithelialization of the distal esophagus (Br J Surg 1988;75: 113-115). Gillen et al. found no significant difference in 24-h esophageal acid exposure between Barrett’s patients with or without complications (strictures, ulcerations] (Br J Surg 1987;74:774-776). However, the same group found abnormally high postprandial bile acid concentrations in the stomachs of patients with complicated Barrett’s esophagus (Br J Surg 1988;75:540-543). Bile reflux has been implicated in the causation of intestinal metaplasia and stomach cancer after gastric surgery (World J Surg 1979;3:731-7361, suggesting the possibilities of similar sequelae in Barrett’s esophagus. Carcinomas in Barrett’s patients have been reported to develop more readily after gastric surgery (Gut 1989:30:14-18). Others have noted regression of Barrett’s epithelium after vagotomy, antrectomy, and Roux-en-Y reconstruction incorporating a 70-cm loop (Gastroenter01 Clin Bioll988;12:709-712). Therefore, we must conclude that the pathogenic role for gastric contents (acid, bile, and/or proteolytic enzymes) in Barrett’s esophagus is unknown at this time. Future long-term studies with omeprazole may begin to unravel this mystery (New Engl J Med 1989;320: 1497-1498). However, some concern has been raised that marked acid suppression may promote the progression and malignant degeneration of Barrett’s epithelium via the mechanisms of alkaline reflux (T. DeMeester. personal communication). Acid, alkali, or both-maybe time will tell? 1. E. RICHTER, M.D.
GASTRIC CANCER AND ADENOVIRUS Koike K, Hinrichs SH, Isselbacher K], et al. (Department of Pathology University of California, Davis, California and Cancer Center, Harvard Medical School, Boston, Massachusetts]. Transgenic mouse model for human gastric carcinoma. Proc Nat1 Acad Sci USA 1989;86:5615-5619. This study reports that mice transfected with adenovirus developed gastric cancers. A segment of the adenovirus type 12 genome, known as early region 1 [El], was modified by replacing the normal promoter sequence of the genome with a regulatory gene segment known as mouse mammary tumor virus long terminal repeat (MMTV 5’ LTR). The MMTV LTR segment is not expressed during early embryonic development and thus, the El gene with the MMTV LTR segment would not be expressed until after birth. This change in the regulation of the gene was made to prevent early gene expression that might lead to developmental abnormalities that could disrupt embryogenesis. The 6.7-kb fragment containing the 2 components of the El gene, ElA and EIB, along with the MMTV LTR enhancer (promoter) were microinjected into fertilized eggs. Transgenic mice comprised 36% of the offspring. Twenty-three of the transgenie mice developed to maturity (11 males and 12 females]. All the males died within 3-4 months and 2 of the females
clinical disease states, such as chronic hepatitis B. In this report nomenclature also was problematic, and the terms used in the “clinical” diagnoses did not match those used for “final” diagnosis. For example, the patients were diagnosed clinically as mild or severe chronic hepatitis B, mild or severe chronic hepatitis (not further specified), or chronic non-A, non-B hepatitis. The same group after histology was obtained included the diagnoses of chronic persistent hepatitis B, chronic active or chronic persistent non-A, non-B hepatitis, or delta hepatitis. Some rather vague clinical diagnoses were used, for example cholestatic liver disease. The final diagnoses were more precise, as might be expected. This study of liver biopsy should not be taken as a guide to be applied to standard work-up in a nonstudy setting. For example, not every patient needs an antimitochondrial antibody or HIV titer, as was done here. Also, recent work would suggest that cryptogenic liver disease may be associated with the heterozygous state for a,-antitrypsin (Stand J Gastroenterol1985;20:835-842). Therefore, if this diagnosis is under consideration, and it probably should be in most patients with asymptomatic elevations in transaminase, a quantitative level of a,-antitrypsin is not sufficient, or even necessary, and a Pi phenotype must be determined. One test that was used here, the alkaline phosphatase of hepatic origin, is unclear to me. Does this mean only elevated alkaline phosphatases associated with an elevated gamma glutamyl transpeptidase (GGTP)? Oddly, there are some lapses in editing in this paper. Tables 2 and 3 show the number of patients with fatty liver as 27, giving a total patient number of 100, whereas the text lists the number of 17, presumably the correct number to add up to the total number of 96 patients studied. Hemochromatosis is spelled in two different ways. The major difficulty with this report is that the authors do not tell us how they actually calculated the data on the accuracy of the clinical diagnosis as determined by the biopsy diagnosis. What we really want to know is how many of the patients changed group after the results of their biopsy were known. For example, we know that 25 carried a diagnosis of alcoholic liver disease and that 26 had it shown on biopsy. Was it the same 25 patients plus one extra? The miscellaneous group grew by 4 patients when biopsy confirmation was in hand. From which group did these subjects come? Attempts to back calculate to these numbers from their table giving percentages for sensitivity and predictive value were unsuccessful. Unfortunately, the authors have missed the opportunity to explore the real problems for specialists in liver disease remaining with liver biopsy as a diagnostic tool. The weak link in many settings is the interpretation of the pathology. There are several causes for these difficulties. First, many pathologists do not routinely do the multiple stains done by these authors, and considered standard at some centers. In interpreting liver biopsy material, an H&E stain only may leave many diagnostic clues hidden. Second, interpretation of liver biopsy material does require expertise and experience, which may be sorely lacking. Finally, the good working relationship between pathologist and clinician described here often does not exist. The pathologist cannot provide optimal interpretation with only limited clinical input. Another real problem with liver biopsy not addressed here is the question of diagnostic accuracy. It is known that biopsy is relatively good at providing a correct morphologic diagnosis, given a specimen of adequate size, a problem not addressed in this report. Unfortunately, etiology often remains obscure. For example, some of the cases of fatty liver described here may be indeed alcohol induced, as patients often minimize their alcohol intake. There are settings in which the biopsy may be wrong. For example, in macronodular cirrhosis, the needle may pass through a large regenerative nodule, yielding a diagnosis of normal tissue and misleading the clinician. Diagnostic accuracy is much higher in diffuse diseases such as alcoholic liver disease (Lancet 1986:1:523-525). Like this paper, liver biopsy remains imperfect. Nevertheless, we are gratified to be reassured that it is useful in assessing patients
GASTROENTEROLOGY
Vol. 98, No. 3
with chronically abnormal liver function tests. Also, we are particularly grateful that this important message has again been brought to the generalists who make up the audience of the Annals and are frequently faced with the problem of what to do with chronically abnormal liver tests. They should know that, when used with knowledge of its limitations, liver biopsy remains the gold standard for diagnosis in liver disease. C. A. RIELY, M.D.
BARRETT’S ESOPHAGUS: TOO MUCH ACID, ALKALI, OR BOTH? Mulholland MW, Reid BI, Levine DS, et al. [Departments of Surgery and Medicine, University of Washington, Seattle, Washington]. Elevated gastric acid secretion in patients with Barrett’s metaplastic epithelium. Digest Dis Sci 1989;34:13291335. Twenty consecutive patients with chronic, symptomatic gastroesophageal reflux, many of whom were referred to the University of Washington for known or suspected Barrett’s esophagus, were recruited for a study assessing gastric acid secretion in Barrett’s patients. Of the 20 patients, 17 agreed to undergo protocol investigations. Ten of these patients (9 men, 1 woman: mean age 59 + 4 years] were subsequently proven by endoscopic biopsy to have Barrett’s esophagus. The other 7 patients [2 men, 5 women; mean age 50 + 5 years) had gastroesophageal reflux without Barrett’s specialized metaplasia. Normal controls, matched with regard to age and sex for the first 10 reflux patients to enter the study, were chosen from a group of volunteers. The control group consisted of 5 men and 5 women, with a mean age of 50 k 4 years. Gastric acid secretion and release of endogenous gastrin were stimulated with a liquid peptone meal using the technique of intragastric titration in all subjects. Response to exogenous gastrin was assessed using dose-response studies with i.v. synthetic human gastrin 17-I. In addition to baseline studies performed in the unmedicated state, gastric acid secretion in response to exogenous human gastrin 17-I was measured while patients were receiving either 300 mg of cimetidine every 6 hours or 2.5 mg of cimetidine plus methscopolamine every 6 hours. Continuous 24-h esophageal pH monitoring was performed in all gastroesophageal reflux patients in the unmedicated state. Number of reflux episodes was counted and percentage of time with esophageal pH <4 was calculated. Basal acid output was significantly greater in the Barrett’s patients (8.56 + 1.82 mEq/h; x + SEM) than in the agematched controls (2.78 A 1.08 mEq/h), but not different from the patients with uncomplicated gastroesophageal reflux (5.58 * 2.90 mEq/h). Gastric acid secretion also was higher for the Barrett’s patients after each dose level of gastrin 17-I than either the reflux esophagitis group or normal control group. However, these latter differences were only statistically significant when the Barrett’s esophagitis group was compared with the controls. One-hour peak acid output stimulated by the peptone meal, serum gastrin release, both basally and with meal stimulation, and sensitivity to gastrin, expressed as the concentration causing half-maximal acid secretion, were similar among the study groups. Both reflux
March 1990
groups showed similar decreases in basal and gastrinstimulated acid secretion after the administration of cimetidine and methscopolamine. No significant differences were noted when patients with uncomplicated reflux were compared with the Barrett’s patients in terms of frequency or total time of reflux episodes during 34-h esophageal pH monitoring. The authors concluded that differences in acid secretion, both basally and in response to gastrin 17-1, may characterize patients with Barrett’s esophagus. They speculated that elevated basal acid production may contribute to the pathogenesis of this specialized metaplastic epithelium. However, the mechanisms responsible for their observations were not elucidated by the current study and remain an area for future investigation. Comments. Today it is well recognized that the Barrett’s esophagus is an acquired condition that occurs as a sequela of gastroesophageal reflux disease (GERD). What is not so obvious is why some patients develop esophagitis while others develop specialized columnar metaplasia of the distal esophagus. As a group, patients with Barrett’s esophagus have weaker lower esophageal sphincter (LES) pressure and more prolonged acid exposure than do patients with severe esophagitis [Arch Surg 1983;18:543-549; Br J Surg 1987;74:774776). However, there is much overlap among individuals within these groups such that neither manometry nor pH monitoring can be used as a diagnostic test for Barrett’s esophagus. Gastric emptying rates are also similar in Barrett’s patients when compared with healthy volunteers and esophagitispatients [Dig Dls Sci 1986;32:12171220). Therefore, we must begin to look past the underlying esophageal motility disorder at other potential causative factors for the development of Barrett’s epithelium. Another potential culprit might be the gastric contents bathing the distal esophagus, with acid being the most likely candidate. In fact, it is somewhat surprising that gastric secretion has only recently been studied in these patients. The Seattle group found basal acid output and the gastric acid secretion response to synthetic gastrin to be significantly greater in Barrett’s patients when compared with healthy volunteers, with uncomplicated esophagi& patients falling between these 2 groups. Although the numbers were small, their findings are similar to a preliminary report from Collen et al at Georgetown University(Gastroenterologyl987392:1350A].Thisgroup reported significantly higher basal acid output in 13 Barrett’s patients (11.9 + 4.6 mEq/h, *SD) compared with 59 patients with GERD (5.8 + 5.4 mEq/h) and 40 healthy volunteers (3.2 f 2.7 mEq/ h). Basal acid output was greater than IO mEq/h in 69% (9 of 13) of patients with Barrett’s esophagus compared with 24% (14 of 59) of GERD patients. Furthermore, these findings may have therapeutic importance as the same group has observed that 15 of 36 Barrett’s patients required more than the usual doses of ranitidine to effect healing of esophagitis (Gastroenterology 1988;94:73A]. Neither group reported the proportion of patients with underlying duodenal ulcer disease. Although intriguing, these studies may be flawed by important methodological problems. Patient groups tended to be age-matched, but the Barrett’s patients were nearly all men, while the esophagitis and control groups had a more equal sex distribution. This male preponderance may be important as the basal acid output is higher in men than in women (Gastroenterology 1984$6:843-8511. More importantly, the previous studies did not exclude associated duodenal ulcer disease as a confounding factor. Dr. Basil Hirschowitz has recently studied 14 Barrett’s patients comparing basal and maximal gastric acid and pepsin secretion with 114 esophagitis patients stratified for the presence of duodenal ulcer disease (39 with duodenal ulcer, 75 without duodenal ulcers] (Gastroenterology 1989;96:212A). Three patients with Barrett’s esophagus and duode-
SELECTED SUMMARIES
799
nal ulcer disease had basal acid outputs ranging from 8.7-11.5 mEq/h, comparable to the values reported in the 2 previous studies, but not different from control duodenal ulcer patients with or without esophagitis. However, the 11 Barrett’s patients without duodenal ulcer disease, all of whom were men, had a mean basal acid output of 2.9 + 0.7 mEq/h or slightly, but not significantly, below the mean values for the comparable nonulcer patients with esophagitis (3.5 + 0.5 mEq/h) or without esophagitis (4.4 * 0.5 mEq/ h). Pepsin secretion was also not different between these groups. Duodenogastroesophageal reflux may be another factor contributing to the genesis and complications in Barrett’s esophagus. In a dog model, prolonged acid or acid and bile, but not bile alone, produces columnar epithelialization of the distal esophagus (Br J Surg 1988;75: 113-115). Gillen et al. found no significant difference in 24-h esophageal acid exposure between Barrett’s patients with or without complications (strictures, ulcerations] (Br J Surg 1987;74:774-776). However, the same group found abnormally high postprandial bile acid concentrations in the stomachs of patients with complicated Barrett’s esophagus (Br J Surg 1988;75:540-543). Bile reflux has been implicated in the causation of intestinal metaplasia and stomach cancer after gastric surgery (World J Surg 1979;3:731-7361, suggesting the possibilities of similar sequelae in Barrett’s esophagus. Carcinomas in Barrett’s patients have been reported to develop more readily after gastric surgery (Gut 1989:30:14-18). Others have noted regression of Barrett’s epithelium after vagotomy, antrectomy, and Roux-en-Y reconstruction incorporating a 70-cm loop (Gastroenter01 Clin Bioll988;12:709-712). Therefore, we must conclude that the pathogenic role for gastric contents (acid, bile, and/or proteolytic enzymes) in Barrett’s esophagus is unknown at this time. Future long-term studies with omeprazole may begin to unravel this mystery (New Engl J Med 1989;320: 1497-1498). However, some concern has been raised that marked acid suppression may promote the progression and malignant degeneration of Barrett’s epithelium via the mechanisms of alkaline reflux (T. DeMeester. personal communication). Acid, alkali, or both-maybe time will tell? 1. E. RICHTER, M.D.
GASTRIC CANCER AND ADENOVIRUS Koike K, Hinrichs SH, Isselbacher K], et al. (Department of Pathology University of California, Davis, California and Cancer Center, Harvard Medical School, Boston, Massachusetts]. Transgenic mouse model for human gastric carcinoma. Proc Nat1 Acad Sci USA 1989;86:5615-5619. This study reports that mice transfected with adenovirus developed gastric cancers. A segment of the adenovirus type 12 genome, known as early region 1 [El], was modified by replacing the normal promoter sequence of the genome with a regulatory gene segment known as mouse mammary tumor virus long terminal repeat (MMTV 5’ LTR). The MMTV LTR segment is not expressed during early embryonic development and thus, the El gene with the MMTV LTR segment would not be expressed until after birth. This change in the regulation of the gene was made to prevent early gene expression that might lead to developmental abnormalities that could disrupt embryogenesis. The 6.7-kb fragment containing the 2 components of the El gene, ElA and EIB, along with the MMTV LTR enhancer (promoter) were microinjected into fertilized eggs. Transgenic mice comprised 36% of the offspring. Twenty-three of the transgenie mice developed to maturity (11 males and 12 females]. All the males died within 3-4 months and 2 of the females