GASTROENTEROLOGY
1991;100:328-332
Helicobacfer pylori Infection in Pernicious Anemia: A Prospective Controlled Study TSE-LING FONG, CORNELIUS P. DOOLEY, MARGARITA DEHESA, HARTLEY COHEN, RALPH CARMEL, PATRICK L. FITZGIBBONS, GUILLERMO I. PEREZ-PEREZ, and MARTIN J. BLASER Departments of Medicine and Pathology, University of Southern California School of Medicine, Los Angeles, California, and the Division of Infectious Diseases, Vanderbilt University School of Medicine, Nashville, Tennessee
Although some authors believe that Helicobacter pylori is the etiologic agent in chronic nonspecific gastritis, it has also been suggested that the bacterium colonizes inflamed mucosa as a secondary event. This study documents the prevalence of Zf. pylori in 28 patients with pernicious anemia and compares the findings with those of a group of 28 age-, race-, and sex-matched asymptomatic control subjects. All subjects underwent endoscopy with biopsy of the gastric antrum and corpus. A sample of serum was obtained before endoscopy for determination of antibodies (immunoglobulin A and immunoglobulin G) to H. pylori. The prevalence of H. pylori (by biopsy) in patients with pernicious anemia was significantly less than that in controls (11% vs.71%, P < 0.0001).All patients with pernicious anemia had abnormalities of corpus histology (inflammation and/or atrophy). In addition, 50% of patients with pernicious anemia had a lymphocytic infiltration of the antrum. All controls with H. pyZori had gastritis, 50% having active chronic gastritis. Atrophic changes of the corpus were more commonly found in patients with pernicious anemia (75% vs. 7%, P < 0.0001). Serology and biopsy results correlated poorly in the patients with pernicious anemia: all 5 patients with positive serology results had negative biopsy results, whereas all 3 patients with positive cultures on biopsy had negative serological studies. In conclusion, patients with pernicious anemia are protected from infection with H. pylori, and H. pylori does not passively colonize mucosa inflamed by an unrelated process.
M
any investigators believe (l-4) that Helicobacter pylori is the etiologic agent in chronic, nonspe-
cific gastritis
(often termed
type B gastritis).
Con-
versely, a number of investigators (5,6)suggest that the bacterium may incidentally colonize gastric epithelium in the setting of nonspecific gastritis induced by other causes. Attention has been focused on certain specific forms of gastritis (including pernicious anemia and Menetrier’s disease) in an effort to resolve this controversy. It is argued that H. pylori should be found with equal frequency in specific and nonspecific forms of gastritis if the inflammation is merely permissive for H. pylori colonization. At the least, one might anticipate that the prevalence of H. pylori infection in specific forms of gastritis matches that of the background population, and indeed such infection is common in the asymptomatic population (7). Results of preliminary studies (8-11) suggest that H. pylori is uncommon in specific forms of gastritis. However, these studies are flawed because they lack appropriate population controls. The purposes of this study were to estimate the prevalence of H. pylori infection in patients with pernicious anemia and to compare the findings with those of a group of matched asymptomatic subjects. Methods Subjects Thirty-eight patients with suspected pernicious anemia [low-serum vitamin B,, level, megaloblastic anemia) attending the hematology clinic at the Los Angeles CountyUniversity of Southern California Medical Center were asked to participate in this study. Of these, 32 agreed to
o 1991 by the American Gastroenterological oom5005/91/$3.00
Association
February 1991
H. PYLORI
participate but one subsequently failed to keep all appointments. The final diagnosis of pernicious anemia was based on the presence of a low serum B,, level and at least one of the following: (a) an abnormal Schilling test that corrected with administration of intrinsic factor; (b) absence of intrinsic factor in gastric secretions; and (c) presence of antibodies directed against intrinsic factor in the patient’s serum. Based on these criteria, 3 of the study patients were found not to have pernicious anemia (2 were subsequently shown to have intestinal malabsorption). Of the remaining 28 patients, 12 were male, the mean age (*SD) was 59 2 14 years; 20 were Hispanic, 5 were black, and 3 were white. The mean pH (?SD) in gastric aspirates available from 25 of these patients was 7.3 2 0.4 (range, 6.3-8.0). For each case, one age-, sex-, and race-matched asymptomatic volunteer was selected from a large group of subjects studied previously at this institution (7,12). A control subject was randomly selected from a group with the age (2 3 years), sex, and race characteristics of the case. All participants in the study gave written informed consent under protocols approved by the Research Committee of the Los Angeles County-University of Southern California Medical Center.
INFECTION IN PERNICIOUS ANEMIA 329
The identification of H. pylori in all specimens by reviewing the H&E stains (7~3).
was made
Serological Techniques Serum samples were examined for H. pylori by an enzyme-linked immunosorbent assay described by PerezPerez et al. (14). In brief, the antigen used was a pool of the sonicated isolates of H. pylori from five patients with gastritis. Serum samples were diluted 1:50 for the immunoglobulin (1g)A assay and 1:800 for the IgG assay. The second antibody was peroxidase-conjugated goat anti-human IgA or IgG. Analysis of Data A subject was considered to be infected with H. pylori only if the bacterium was cultured from, or observed in, biopsy specimens. Differences between groups were assessed by the x2 test; P < 0.05 was considered significant. Results
Gastrointestinal Endoscopy All subjects underwent endoscopy and biopsy of the upper gastrointestinal tract. Before the endoscopy, 5 to 8 mL of venous blood was withdrawn and the resultant serum was stored at -4°C. During the endoscopy, paired biopsy specimens were obtained routinely from the prepyloric gastric antrum and the gastric corpus. Separate sterile biopsy forceps were used for each site. One biopsy specimen from each pair was placed in transport medium for subsequent microbiological assessment. The remaining biopsy specimen was fixed in neutral buffered formalin and retained for subsequent histological study. Incidental lesions found at endoscopy also underwent biopsy for histological evaluation. Microbiological
Techniques
Immediately after endoscopy, the biopsy specimens were removed from the transport medium, plated onto both chocolate and Campylobacter-selective agars (Scott Laboratories, Los Angeles, CA), and cultured at 37°C under microaerobic conditions (Campypak, BBL Microbiology Systems, Becton, Dickinson and Co., Cockeysville, MD). The plates were checked every 2 to 3 days, and identification of H. pylori was made on the basis of the morphological features of the colony and Gram’s stain, and positive reactions to oxidase, catalase, and urease. Histological Studies Gastric biopsy specimens were stained with H&E and graded for gastritis based on the inflammatory cell infiltrate according to previously published criteria (7). Chronic gastritis denoted a lymphocytic infiltrate and active chronic gastritis described a neutrophilic infiltrate in addition to chronic gastritis. In addition to inflammation, atrophy and intestinal metaplasia were also noted when present.
Endoscopy Endoscopy showed unsuspected gastric lesions in six of the study participants. A gastric polypoid lesion was discovered at endoscopy in one of the patients with pernicious anemia. This was subsequently removed by endoscopic snare and was found to contain a focus of carcinoma in situ. In the control group, four subjects had gastric erosions and one had subepithelial hemorrhages. Prevalence
of Helicobacter
pylori
H. pylori was detected in only three (11%) of the patients with pernicious anemia compared with 20 (71%) of the control group (P < 0.0001). H. pylori was identified by culture only in the three infected pernicious anemia patients. Conversely, H. pylori was identified by culture and stain in 13, stain only in 6, and culture only in 1 of the control group. Two of the 3 pernicious anemia patients with H. pylori also had autoimmune cytopenias (15), a finding not noted in the remaining 25 patients. Gastric Histology Gastric histology was abnormal in all patients with pernicious anemia. Twenty-five of the 28 patients with pernicious anemia were found to have a chronic inflammatory infiltrate in gastric biopsy specimens. Of these, the inflammation was confined to the antrum in 6 cases, to the corpus in 11 cases, and involved both antrum and corpus in 8 cases (Table 1). Inflammation was noted in 2 (antral in 1, corpus in 1) of the 3 patients infected with H. pylori, the third
330
GASTROENTEROLOGY Vol. 100, No. 2
FONG ET AL.
Table I. Summary of Biopsy Findings N
Antrum Inflammation
Pernicious anemia H. pylori (+) HpyIori (-) Controls H. pylori (+)” H. pylon’(- )
corpus
Atrophy
Inflammation
3 25
1 13
0 0
1 18
20 8
18 (7) 1
0 2
20 (7) 2
Atrophy 3 18 1 1
“Numbers in par&theses refer to number with active chronic gastritis at each site. A total of 10 subjects had this lesion (4 in both antrum and corpus).
having atrophic changes only (Table 1). H. pylori was cultured from both antrum and corpus in these 3 patients. Atrophic changes (glandular atrophy and/or intestinal metaplasia) were noted in corpus biopsies from 21 (75%) of the patients with pernicious anemia compared with 2 (7%) of the control group (P < 0.0001, Table 1). Eighteen of these 21 had intestinal metaplasia of the corpus. Intestinal metaplasia of the antrum was found in 2 patients with pernicious anemia (both with atrophic changes in the corpus). One of these 2 also had antral gastritis. H. pylon’ was found in biopsy specimens of both the antrum and corpus from 17 of the 20 infected controls. The remaining 3 had H. pylori in corpus biopsy specimens only. A chronic inflammatory infiltrate was noted in both antrum and corpus in 18 of the 20 controls with H. pylori. The remaining 2 had inflammation of the corpus only (Table 1) and also had H. pylori in corpus biopsy specimens only. Inflammation was observed in 2 additional control subjects who were not found to have H. pylori on culture and stain. One of these had inflammation of the corpus only, the other having inflammation of both antrum and corpus (Table 1). Active chronic gastritis was found only in control subjects infected with H. pylori: 10 of the 20 with H. pylori had this lesion (Table 1). Intestinal metaplasia was not noted in biopsy samples from the control group. Serology
Disparate serological results were noted in the pernicious anemia patients: none of the three patients with H. pylori on gastric biopsy specimens had H. pylori antibodies; four of the remaining patients had IgA antibodies; one of these patients also had IgG antibody directed against H. pylori; one additional patient had IgG antibody directed against the bacterium. Both gastric histology and the clinical features of these five patients were indistinguishable from those of the other 23 patients with pernicious anemia.
Serum was available from 18 of the 20 controls with, and all 8 controls without, H. pylori. In those infected, IgG and IgA antibodies directed against H. pylori were noted in 17 and 15 of the 18 samples, respectively. Of the 8 subjects who did not have H. pylori on gastric biopsy samples, 3 had IgG and IgA antibodies directed against the bacterium. Interestingly, 2 of these 3 were noted to have gastritis on biopsy (Table 1). Discussion This study demonstrates that H. pylori is an infrequent finding in patients with pernicious anemia and confirms the data of previous studies (8-10). The previous studies were retrospective and therefore used the histological technique only. The prevalence of H. pylori in these studies ranged from 3.5% to 21%, but tissue samples were available from the corpus only in many of the patients (8-10). This study is an advance on previous studies in that it was prospective, used both microbiological and histological techniques to diagnose infection with H. pylori, examined tissue from both antrum and corpus in all patients, and used matched controls. The finding of H. pylori in 11% of patients with pernicious anemia contrasted starkly with a prevalence of 71% in the age-, race-, and sex-matched controls. This high prevalence in the controls is accounted for by the advanced age of the study group and the high proportion of Hispanics (7,12). Therefore, H. pylori is found much less commonly in patients with pernicious anemia than in the background population. This suggests that the pathological process in pernicious anemia protects the gastric mucosa from colonization by H. pylori. It is not known why this might occur. Intestinal metaplasia, which is a common sequela in pernicious anemia, would mitigate against colonization by H. pylon’. However, intestinal metaplasia was an uncommon finding in the antra of our patients with pernicious anemia. Achlorhydria, a universal finding in pernicious anemia, might also prevent colonization by H. pylori, perhaps by allowing overgrowth of competing organisms. Conversely, the immunological processes involved in the pathogenesis of pernicious anemia might cause alterations in the glycoprotein calyx of the gastric epithelial cell and interfere with specific bacterial attachment mechanisms (16). There are no data to confirm or refute this hypothesis. The infrequent finding of H. pylori in patients with pernicious anemia adds further support to the contention that the bacterium is the causative agent of chronic nonspecific gastritis. We found chronic non-
February
1991
specific gastritis in 22 of the 28 control subjects. Twenty of these 22 (91%) were found to harbor H. pylori, confirming a consistent relationship between the bacterium and the histological lesion found by numerous centers throughout the world (l-3). Conversely, pernicious anemia is the consequence of a specific form of gastritis and H. pylori is negatively associated with this lesion. In addition, Ormand et al. (11) and Drumm et al. (17) were unable to find H. pylori in any patients with other types of specific gastritis, including M&&trier’s disease, eosinophilic gastritis, and Crohn’s gastritis. Therefore, it would appear that the relationship between H. pylori and the so-called chronic, nonspecific gastritis is unique. Our findings make it extremely unlikely that Zf. pylori is a sequela to gastritis, as has been suggested (6). As expected, all patients with pernicious anemia had evidence of the pathological process in the proximal stomach (18). Loss of gastric secretory function was demonstrated by the findings of achlorhydria and absence of intrinsic factor in gastric aspirates. Inflammation of the antrum (lymphocytic infiltration only) was noted in 50% of the patients with pernicious anemia. Antral gastritis has been noted in previous studies (l&19-21) and varies in frequency from 21% to 100%. The significance of antral inflammation in pernicious anemia is unknown. It may be an extension of the pathological processes at work in the corpus. Alternatively, it may be chronic, nonspecific gastritis superimposed on the pathological process of pernicious anemia (22). The lack of association of this antral gastritis with H, pylori and the lack of a neutrophilic infiltrate in these cases provide strong evidence against this hypothesis. The serological results also deserve comment. In the control group, there was a close correlation between antibody status and results of gastric biopsies, as has been previously validated (7,14,17). Conversely, the serological results in the patients with pernicious anemia were often in conflict with biopsy data. The serological response to H. pylori observed in our patients was predominantly IgA. In a separate serological study, demonstrating a negative association between H. pylori infection and pernicious anemia, 7 (17.5%) of 40 patients had IgA antibodies to H. pylori only (23). Seropositivity in IgA, but not IgG, is uncommon in 23.pylori-infected individuals (7,14,17). We are unable to provide an explanation for this finding in pernicious anemia. In summary, our study confirms previous retrospective reports that H. pylori is an infrequent finding in patients with pernicious anemia. Our controlled data show that the prevalence of the infection in pernicious anemia is significantly less than that of the background population. We conclude that the associ-
H. PYLORI INFECTION
IN PERNICIOUS
ANEMIA
331
ation between H. pylori and chronic, nonspecific gastritis is unique and that the pathological process in pernicious anemia protects against H. pylori infection.
References 1. Dooley CP, Cohen H. The clinical significance of Campylobocterpylori. Ann Intern Med 1988;108:70-79. 2. Marshall BJ, McGechie DB, Rogers PA, Glancy RJ. Pyloric Compylobacter infection and gastroduodenal disease. Med J Aust 1985;142:439-444. 3. Graham DY, Klein PD. Campylobocter pyloridis gastritis: The past, the present, and speculations about the future. Am J Gastroenterol 1987;82:283-286. 4. Blaser MJ. Helicobacterpylori and the pathogenesis of gastroduodenal inflammation. J Infect Dis 1990;161:1237-1241. 5. Varis K. Gastritis-a misused term in clinical gastroenterology. Stand J Gastroenterol 1988;23(Suppl 155):53-58. 6. Peterson WL, Lee E, Feldman M. Relationship between Campylobacterpylori and gastritis in healthy humans after administration of placebo or indomethacin. Gastroenterology 1988;95: 1185-1197. PL, Bauer M, Appleman MD, 7. Dooley CP, Cohen H, Fitzgibbons Perez-Perez GI, Blaser MJ. Prevalence of Helicobacter pylori and histologic gastritis in asymptomatic persons. N Engl J Med 1989;321:1562-1566. 8. O’Connor HJ, Axon ATR, Dixon MF. Campylobacter-like organisms unusual in type A (pernicious anemia) gastritis (letter). Lancet 1984;2:1092. 9. Gonzalez JD, Sancho FJ, Sainz S, Such J, Fernandez M, Mones anemia (letter). Xiol J. Campylobacfer pylori and pernicious Lancet 1988;1:57. 10. Flejou J-F, Bahame P, Smith AC, Stockbrugger RW, Rode J, Price AB. Pernicious anemia and Campylobacfer-like organisms: is the gastric antrum resistant to colonization? Gut 1989;30:60-64. JE, Talley NJ, Shorter CR, Conley CR, Wilson WR, 11. Ormand Phillips SF. Campylobacter pylori prevalence in specific forms of gastritis: further evidence supporting a pathogenic role for C. pylori in chronic antral gastritis (abstr). Gastroenterology 1989; 96:A378. P, Perez-Perez GI, 12. Dehesa M, Dooley CP, Cohen H, Fitzgibbons Blaser MJ. High prevalence of Campylobacter pylori in an asymptomatic Hispanic population (abstr). Gastroenterology 1989;96:A115. PL, Dooley CP, Cohen H, Appleman MD. Preva13. Fitzgibbons lence of gastric metaplasia, inflammation and Campylobacter pylori in the duodenum of a normal population. Am J Clin Path01 1988;90:711-714. GI, Dworkin BM, Chodos JE, Blaser MJ. Campylo14. Perez-Perez bacterpylori antibodies in humans. Ann Intern Med 1988;109: 11-17. AP, Sacks Y, Carmel R. Autoimmune cytopenias in 15. Rabinowitz pernicious anemia: a report of four cases and review of the literature. Eur J Haematol 1990;44:18-23. 16. Evans DG, Evans DJ Jr, Moulds JJ, Graham DY. N-acetylneuraminyllactose-binding fibrillar hemagglutinin of Campylobacter pylori: a putative colonization factor antigen. Infect Immun 1988;56:2896-906. GI, Blaser MJ, Sherman PM. Intrafamil17. Drumm B, Perez-Perez ial clusteringofHeZicobacterpylori. N Engl J Med 1990:322:359363. RG, Mackay IR. A reappraisal of the nature and 18. Strickland
332
19.
20.
21.
22.
FONG ET AL.
significance of chronic atrophic gastritis. Am J Dig Dis 1972;18: 426-440. Lewin KJ, Dowling F, Wright JP, Taylor KB. Gastric morphology and serum gastrin levels in pernicious anemia. Gut 1976;17: 551-560. Stockbrugger RW, Menon GG, Beilby JOW, Mason RR, Cotton PB. Gastroscopic screening in 80 patients with pernicious anemia. Gut 1983;24:1141-1147. Kekki M, Varis K, Pohjanpola H, Isokoski M, Ihamaki T, Siurala M. Course of antrum and body gastritis in pernicious anemia families. Dig Dis Sci 1983;28:698-704. Owen DA. Gastritis and duodenitis. In: Appelman I-ID, ed.
GASTROENTEROLOGY Vol. 100, No. 2
Pathology of the esophagus, stomach and duodenum. New York: Churchill Livingstone, 1984:37-77. 23. Blaser MJ, Perez-Perez GI, Lindenbaum J, Schneidman D, Van Deventer G, Marin-Sorensen M, Weinstein WM. Association of Helicobacter pylori infection with specific gastroduodenal pathology. Rev Infect Dis 1990 (in press).
Received April 11,199O. Accepted August 17,199O. Address requests for reprints to: Cornelius P. Dooley, M.D., Section of Gastroenterology, Department of Medicine, USC School of Medicine, 2025 Zonal Avenue, Los Angeles, California 90033.
1991;100:328-332
Helicobacfer pylori Infection in Pernicious Anemia: A Prospective Controlled Study TSE-LING FONG, CORNELIUS P. DOOLEY, MARGARITA DEHESA, HARTLEY COHEN, RALPH CARMEL, PATRICK L. FITZGIBBONS, GUILLERMO I. PEREZ-PEREZ, and MARTIN J. BLASER Departments of Medicine and Pathology, University of Southern California School of Medicine, Los Angeles, California, and the Division of Infectious Diseases, Vanderbilt University School of Medicine, Nashville, Tennessee
Although some authors believe that Helicobacter pylori is the etiologic agent in chronic nonspecific gastritis, it has also been suggested that the bacterium colonizes inflamed mucosa as a secondary event. This study documents the prevalence of Zf. pylori in 28 patients with pernicious anemia and compares the findings with those of a group of 28 age-, race-, and sex-matched asymptomatic control subjects. All subjects underwent endoscopy with biopsy of the gastric antrum and corpus. A sample of serum was obtained before endoscopy for determination of antibodies (immunoglobulin A and immunoglobulin G) to H. pylori. The prevalence of H. pylori (by biopsy) in patients with pernicious anemia was significantly less than that in controls (11% vs.71%, P < 0.0001).All patients with pernicious anemia had abnormalities of corpus histology (inflammation and/or atrophy). In addition, 50% of patients with pernicious anemia had a lymphocytic infiltration of the antrum. All controls with H. pyZori had gastritis, 50% having active chronic gastritis. Atrophic changes of the corpus were more commonly found in patients with pernicious anemia (75% vs. 7%, P < 0.0001). Serology and biopsy results correlated poorly in the patients with pernicious anemia: all 5 patients with positive serology results had negative biopsy results, whereas all 3 patients with positive cultures on biopsy had negative serological studies. In conclusion, patients with pernicious anemia are protected from infection with H. pylori, and H. pylori does not passively colonize mucosa inflamed by an unrelated process.
M
any investigators believe (l-4) that Helicobacter pylori is the etiologic agent in chronic, nonspe-
cific gastritis
(often termed
type B gastritis).
Con-
versely, a number of investigators (5,6)suggest that the bacterium may incidentally colonize gastric epithelium in the setting of nonspecific gastritis induced by other causes. Attention has been focused on certain specific forms of gastritis (including pernicious anemia and Menetrier’s disease) in an effort to resolve this controversy. It is argued that H. pylori should be found with equal frequency in specific and nonspecific forms of gastritis if the inflammation is merely permissive for H. pylori colonization. At the least, one might anticipate that the prevalence of H. pylori infection in specific forms of gastritis matches that of the background population, and indeed such infection is common in the asymptomatic population (7). Results of preliminary studies (8-11) suggest that H. pylori is uncommon in specific forms of gastritis. However, these studies are flawed because they lack appropriate population controls. The purposes of this study were to estimate the prevalence of H. pylori infection in patients with pernicious anemia and to compare the findings with those of a group of matched asymptomatic subjects. Methods Subjects Thirty-eight patients with suspected pernicious anemia [low-serum vitamin B,, level, megaloblastic anemia) attending the hematology clinic at the Los Angeles CountyUniversity of Southern California Medical Center were asked to participate in this study. Of these, 32 agreed to
o 1991 by the American Gastroenterological oom5005/91/$3.00
Association
February 1991
H. PYLORI
participate but one subsequently failed to keep all appointments. The final diagnosis of pernicious anemia was based on the presence of a low serum B,, level and at least one of the following: (a) an abnormal Schilling test that corrected with administration of intrinsic factor; (b) absence of intrinsic factor in gastric secretions; and (c) presence of antibodies directed against intrinsic factor in the patient’s serum. Based on these criteria, 3 of the study patients were found not to have pernicious anemia (2 were subsequently shown to have intestinal malabsorption). Of the remaining 28 patients, 12 were male, the mean age (*SD) was 59 2 14 years; 20 were Hispanic, 5 were black, and 3 were white. The mean pH (?SD) in gastric aspirates available from 25 of these patients was 7.3 2 0.4 (range, 6.3-8.0). For each case, one age-, sex-, and race-matched asymptomatic volunteer was selected from a large group of subjects studied previously at this institution (7,12). A control subject was randomly selected from a group with the age (2 3 years), sex, and race characteristics of the case. All participants in the study gave written informed consent under protocols approved by the Research Committee of the Los Angeles County-University of Southern California Medical Center.
INFECTION IN PERNICIOUS ANEMIA 329
The identification of H. pylori in all specimens by reviewing the H&E stains (7~3).
was made
Serological Techniques Serum samples were examined for H. pylori by an enzyme-linked immunosorbent assay described by PerezPerez et al. (14). In brief, the antigen used was a pool of the sonicated isolates of H. pylori from five patients with gastritis. Serum samples were diluted 1:50 for the immunoglobulin (1g)A assay and 1:800 for the IgG assay. The second antibody was peroxidase-conjugated goat anti-human IgA or IgG. Analysis of Data A subject was considered to be infected with H. pylori only if the bacterium was cultured from, or observed in, biopsy specimens. Differences between groups were assessed by the x2 test; P < 0.05 was considered significant. Results
Gastrointestinal Endoscopy All subjects underwent endoscopy and biopsy of the upper gastrointestinal tract. Before the endoscopy, 5 to 8 mL of venous blood was withdrawn and the resultant serum was stored at -4°C. During the endoscopy, paired biopsy specimens were obtained routinely from the prepyloric gastric antrum and the gastric corpus. Separate sterile biopsy forceps were used for each site. One biopsy specimen from each pair was placed in transport medium for subsequent microbiological assessment. The remaining biopsy specimen was fixed in neutral buffered formalin and retained for subsequent histological study. Incidental lesions found at endoscopy also underwent biopsy for histological evaluation. Microbiological
Techniques
Immediately after endoscopy, the biopsy specimens were removed from the transport medium, plated onto both chocolate and Campylobacter-selective agars (Scott Laboratories, Los Angeles, CA), and cultured at 37°C under microaerobic conditions (Campypak, BBL Microbiology Systems, Becton, Dickinson and Co., Cockeysville, MD). The plates were checked every 2 to 3 days, and identification of H. pylori was made on the basis of the morphological features of the colony and Gram’s stain, and positive reactions to oxidase, catalase, and urease. Histological Studies Gastric biopsy specimens were stained with H&E and graded for gastritis based on the inflammatory cell infiltrate according to previously published criteria (7). Chronic gastritis denoted a lymphocytic infiltrate and active chronic gastritis described a neutrophilic infiltrate in addition to chronic gastritis. In addition to inflammation, atrophy and intestinal metaplasia were also noted when present.
Endoscopy Endoscopy showed unsuspected gastric lesions in six of the study participants. A gastric polypoid lesion was discovered at endoscopy in one of the patients with pernicious anemia. This was subsequently removed by endoscopic snare and was found to contain a focus of carcinoma in situ. In the control group, four subjects had gastric erosions and one had subepithelial hemorrhages. Prevalence
of Helicobacter
pylori
H. pylori was detected in only three (11%) of the patients with pernicious anemia compared with 20 (71%) of the control group (P < 0.0001). H. pylori was identified by culture only in the three infected pernicious anemia patients. Conversely, H. pylori was identified by culture and stain in 13, stain only in 6, and culture only in 1 of the control group. Two of the 3 pernicious anemia patients with H. pylori also had autoimmune cytopenias (15), a finding not noted in the remaining 25 patients. Gastric Histology Gastric histology was abnormal in all patients with pernicious anemia. Twenty-five of the 28 patients with pernicious anemia were found to have a chronic inflammatory infiltrate in gastric biopsy specimens. Of these, the inflammation was confined to the antrum in 6 cases, to the corpus in 11 cases, and involved both antrum and corpus in 8 cases (Table 1). Inflammation was noted in 2 (antral in 1, corpus in 1) of the 3 patients infected with H. pylori, the third
330
GASTROENTEROLOGY Vol. 100, No. 2
FONG ET AL.
Table I. Summary of Biopsy Findings N
Antrum Inflammation
Pernicious anemia H. pylori (+) HpyIori (-) Controls H. pylori (+)” H. pylon’(- )
corpus
Atrophy
Inflammation
3 25
1 13
0 0
1 18
20 8
18 (7) 1
0 2
20 (7) 2
Atrophy 3 18 1 1
“Numbers in par&theses refer to number with active chronic gastritis at each site. A total of 10 subjects had this lesion (4 in both antrum and corpus).
having atrophic changes only (Table 1). H. pylori was cultured from both antrum and corpus in these 3 patients. Atrophic changes (glandular atrophy and/or intestinal metaplasia) were noted in corpus biopsies from 21 (75%) of the patients with pernicious anemia compared with 2 (7%) of the control group (P < 0.0001, Table 1). Eighteen of these 21 had intestinal metaplasia of the corpus. Intestinal metaplasia of the antrum was found in 2 patients with pernicious anemia (both with atrophic changes in the corpus). One of these 2 also had antral gastritis. H. pylon’ was found in biopsy specimens of both the antrum and corpus from 17 of the 20 infected controls. The remaining 3 had H. pylori in corpus biopsy specimens only. A chronic inflammatory infiltrate was noted in both antrum and corpus in 18 of the 20 controls with H. pylori. The remaining 2 had inflammation of the corpus only (Table 1) and also had H. pylori in corpus biopsy specimens only. Inflammation was observed in 2 additional control subjects who were not found to have H. pylori on culture and stain. One of these had inflammation of the corpus only, the other having inflammation of both antrum and corpus (Table 1). Active chronic gastritis was found only in control subjects infected with H. pylori: 10 of the 20 with H. pylori had this lesion (Table 1). Intestinal metaplasia was not noted in biopsy samples from the control group. Serology
Disparate serological results were noted in the pernicious anemia patients: none of the three patients with H. pylori on gastric biopsy specimens had H. pylori antibodies; four of the remaining patients had IgA antibodies; one of these patients also had IgG antibody directed against H. pylori; one additional patient had IgG antibody directed against the bacterium. Both gastric histology and the clinical features of these five patients were indistinguishable from those of the other 23 patients with pernicious anemia.
Serum was available from 18 of the 20 controls with, and all 8 controls without, H. pylori. In those infected, IgG and IgA antibodies directed against H. pylori were noted in 17 and 15 of the 18 samples, respectively. Of the 8 subjects who did not have H. pylori on gastric biopsy samples, 3 had IgG and IgA antibodies directed against the bacterium. Interestingly, 2 of these 3 were noted to have gastritis on biopsy (Table 1). Discussion This study demonstrates that H. pylori is an infrequent finding in patients with pernicious anemia and confirms the data of previous studies (8-10). The previous studies were retrospective and therefore used the histological technique only. The prevalence of H. pylori in these studies ranged from 3.5% to 21%, but tissue samples were available from the corpus only in many of the patients (8-10). This study is an advance on previous studies in that it was prospective, used both microbiological and histological techniques to diagnose infection with H. pylori, examined tissue from both antrum and corpus in all patients, and used matched controls. The finding of H. pylori in 11% of patients with pernicious anemia contrasted starkly with a prevalence of 71% in the age-, race-, and sex-matched controls. This high prevalence in the controls is accounted for by the advanced age of the study group and the high proportion of Hispanics (7,12). Therefore, H. pylori is found much less commonly in patients with pernicious anemia than in the background population. This suggests that the pathological process in pernicious anemia protects the gastric mucosa from colonization by H. pylori. It is not known why this might occur. Intestinal metaplasia, which is a common sequela in pernicious anemia, would mitigate against colonization by H. pylon’. However, intestinal metaplasia was an uncommon finding in the antra of our patients with pernicious anemia. Achlorhydria, a universal finding in pernicious anemia, might also prevent colonization by H. pylori, perhaps by allowing overgrowth of competing organisms. Conversely, the immunological processes involved in the pathogenesis of pernicious anemia might cause alterations in the glycoprotein calyx of the gastric epithelial cell and interfere with specific bacterial attachment mechanisms (16). There are no data to confirm or refute this hypothesis. The infrequent finding of H. pylori in patients with pernicious anemia adds further support to the contention that the bacterium is the causative agent of chronic nonspecific gastritis. We found chronic non-
February
1991
specific gastritis in 22 of the 28 control subjects. Twenty of these 22 (91%) were found to harbor H. pylori, confirming a consistent relationship between the bacterium and the histological lesion found by numerous centers throughout the world (l-3). Conversely, pernicious anemia is the consequence of a specific form of gastritis and H. pylori is negatively associated with this lesion. In addition, Ormand et al. (11) and Drumm et al. (17) were unable to find H. pylori in any patients with other types of specific gastritis, including M&&trier’s disease, eosinophilic gastritis, and Crohn’s gastritis. Therefore, it would appear that the relationship between H. pylori and the so-called chronic, nonspecific gastritis is unique. Our findings make it extremely unlikely that Zf. pylori is a sequela to gastritis, as has been suggested (6). As expected, all patients with pernicious anemia had evidence of the pathological process in the proximal stomach (18). Loss of gastric secretory function was demonstrated by the findings of achlorhydria and absence of intrinsic factor in gastric aspirates. Inflammation of the antrum (lymphocytic infiltration only) was noted in 50% of the patients with pernicious anemia. Antral gastritis has been noted in previous studies (l&19-21) and varies in frequency from 21% to 100%. The significance of antral inflammation in pernicious anemia is unknown. It may be an extension of the pathological processes at work in the corpus. Alternatively, it may be chronic, nonspecific gastritis superimposed on the pathological process of pernicious anemia (22). The lack of association of this antral gastritis with H, pylori and the lack of a neutrophilic infiltrate in these cases provide strong evidence against this hypothesis. The serological results also deserve comment. In the control group, there was a close correlation between antibody status and results of gastric biopsies, as has been previously validated (7,14,17). Conversely, the serological results in the patients with pernicious anemia were often in conflict with biopsy data. The serological response to H. pylori observed in our patients was predominantly IgA. In a separate serological study, demonstrating a negative association between H. pylori infection and pernicious anemia, 7 (17.5%) of 40 patients had IgA antibodies to H. pylori only (23). Seropositivity in IgA, but not IgG, is uncommon in 23.pylori-infected individuals (7,14,17). We are unable to provide an explanation for this finding in pernicious anemia. In summary, our study confirms previous retrospective reports that H. pylori is an infrequent finding in patients with pernicious anemia. Our controlled data show that the prevalence of the infection in pernicious anemia is significantly less than that of the background population. We conclude that the associ-
H. PYLORI INFECTION
IN PERNICIOUS
ANEMIA
331
ation between H. pylori and chronic, nonspecific gastritis is unique and that the pathological process in pernicious anemia protects against H. pylori infection.
References 1. Dooley CP, Cohen H. The clinical significance of Campylobocterpylori. Ann Intern Med 1988;108:70-79. 2. Marshall BJ, McGechie DB, Rogers PA, Glancy RJ. Pyloric Compylobacter infection and gastroduodenal disease. Med J Aust 1985;142:439-444. 3. Graham DY, Klein PD. Campylobocter pyloridis gastritis: The past, the present, and speculations about the future. Am J Gastroenterol 1987;82:283-286. 4. Blaser MJ. Helicobacterpylori and the pathogenesis of gastroduodenal inflammation. J Infect Dis 1990;161:1237-1241. 5. Varis K. Gastritis-a misused term in clinical gastroenterology. Stand J Gastroenterol 1988;23(Suppl 155):53-58. 6. Peterson WL, Lee E, Feldman M. Relationship between Campylobacterpylori and gastritis in healthy humans after administration of placebo or indomethacin. Gastroenterology 1988;95: 1185-1197. PL, Bauer M, Appleman MD, 7. Dooley CP, Cohen H, Fitzgibbons Perez-Perez GI, Blaser MJ. Prevalence of Helicobacter pylori and histologic gastritis in asymptomatic persons. N Engl J Med 1989;321:1562-1566. 8. O’Connor HJ, Axon ATR, Dixon MF. Campylobacter-like organisms unusual in type A (pernicious anemia) gastritis (letter). Lancet 1984;2:1092. 9. Gonzalez JD, Sancho FJ, Sainz S, Such J, Fernandez M, Mones anemia (letter). Xiol J. Campylobacfer pylori and pernicious Lancet 1988;1:57. 10. Flejou J-F, Bahame P, Smith AC, Stockbrugger RW, Rode J, Price AB. Pernicious anemia and Campylobacfer-like organisms: is the gastric antrum resistant to colonization? Gut 1989;30:60-64. JE, Talley NJ, Shorter CR, Conley CR, Wilson WR, 11. Ormand Phillips SF. Campylobacter pylori prevalence in specific forms of gastritis: further evidence supporting a pathogenic role for C. pylori in chronic antral gastritis (abstr). Gastroenterology 1989; 96:A378. P, Perez-Perez GI, 12. Dehesa M, Dooley CP, Cohen H, Fitzgibbons Blaser MJ. High prevalence of Campylobacter pylori in an asymptomatic Hispanic population (abstr). Gastroenterology 1989;96:A115. PL, Dooley CP, Cohen H, Appleman MD. Preva13. Fitzgibbons lence of gastric metaplasia, inflammation and Campylobacter pylori in the duodenum of a normal population. Am J Clin Path01 1988;90:711-714. GI, Dworkin BM, Chodos JE, Blaser MJ. Campylo14. Perez-Perez bacterpylori antibodies in humans. Ann Intern Med 1988;109: 11-17. AP, Sacks Y, Carmel R. Autoimmune cytopenias in 15. Rabinowitz pernicious anemia: a report of four cases and review of the literature. Eur J Haematol 1990;44:18-23. 16. Evans DG, Evans DJ Jr, Moulds JJ, Graham DY. N-acetylneuraminyllactose-binding fibrillar hemagglutinin of Campylobacter pylori: a putative colonization factor antigen. Infect Immun 1988;56:2896-906. GI, Blaser MJ, Sherman PM. Intrafamil17. Drumm B, Perez-Perez ial clusteringofHeZicobacterpylori. N Engl J Med 1990:322:359363. RG, Mackay IR. A reappraisal of the nature and 18. Strickland
332
19.
20.
21.
22.
FONG ET AL.
significance of chronic atrophic gastritis. Am J Dig Dis 1972;18: 426-440. Lewin KJ, Dowling F, Wright JP, Taylor KB. Gastric morphology and serum gastrin levels in pernicious anemia. Gut 1976;17: 551-560. Stockbrugger RW, Menon GG, Beilby JOW, Mason RR, Cotton PB. Gastroscopic screening in 80 patients with pernicious anemia. Gut 1983;24:1141-1147. Kekki M, Varis K, Pohjanpola H, Isokoski M, Ihamaki T, Siurala M. Course of antrum and body gastritis in pernicious anemia families. Dig Dis Sci 1983;28:698-704. Owen DA. Gastritis and duodenitis. In: Appelman I-ID, ed.
GASTROENTEROLOGY Vol. 100, No. 2
Pathology of the esophagus, stomach and duodenum. New York: Churchill Livingstone, 1984:37-77. 23. Blaser MJ, Perez-Perez GI, Lindenbaum J, Schneidman D, Van Deventer G, Marin-Sorensen M, Weinstein WM. Association of Helicobacter pylori infection with specific gastroduodenal pathology. Rev Infect Dis 1990 (in press).
Received April 11,199O. Accepted August 17,199O. Address requests for reprints to: Cornelius P. Dooley, M.D., Section of Gastroenterology, Department of Medicine, USC School of Medicine, 2025 Zonal Avenue, Los Angeles, California 90033.
