September
1992
prevalence rate of H. pylori antibodies in their blood.‘,* Adults in that region have the highest mortality rates from gastric cancer documented in the world today and a very high prevalence rate of intestinal metaplasia of the gastric mucosa.3 These same children have elevated blood levels of pepsinogen C, a clear sign of antral gastritis.* These findings clearly suggest that H. pylori may be involved in the etiology of this disease. The children in this highrisk area also have greater potential of endogenous nitrosation than children in low-risk areas, strongly suggesting that N-nitroso carcinogens are synthesized in the gastric microenvironment in subjects living in the highlands.* On the other hand, in the lowlands of Costa Rica the rates of gastric cancer and intestinal metaplasia are several times lower than those in the highlands. But the children of such lowlands have the same high prevalence of H. pylori infection and high serum concentrations of pepsinogen C. These additional observations in our opinion provide scientific evidence that H. pylori-associated antral gastritis (DAG) in children of the highlands correlates with high rates of intestinal metaplasia, gastric ulcers, and gastric carcinoma in adults. In the lowlands, however. the association with such diseases is not found. Similar situations are also present in other populations at very low risk for gastric carcinoma such as Yemen.4 Holcombe reported high rates of H. pylori infection since childhood and a high prevalence of gastritis in Africa but a very low prevalence of intestinal metaplasia and gastric carcinoma.5 The recent scientific evidence available in our opinion clearly adds weight to the concept that DAG and MAG are distinct entities even though H. pylori may be a factor in both diseases. PELAYO CORREA, M.D. JOHN H. YARDLEY, M.D.
Department of Pathology Louisiana State University Medical Center 1901 Perdido Street New Orleans, Louisiana 70112-1393 1.
Sierra R, Oshima H, Muiioz N, et al. Exposure to N-nitrosamines and other risk factors for gastric cancer in Costa Rican children. In: O’Neill IK, Chen S, Bartsch H, eds. Relevance to humans of N-nitroso compounds, tobacco smoke and mycotoxins. Lyon, France: International Agency for Research on Cancer, 1991:162-167. 2. Sierra R, Mufioz N, Peiia S. Antibodies to Helicobacter pylori and pepsinogen levels in children from Costa Rica. Cancer Epid Biom Prev 1992 (in press). 3. Salas J. Lesiones precancerosas de1 est6mago en Costa Rica. Patologia (MBxico) 1977;15:63-79. 4. El-Guneid A, El-Sherif AM, Murray-Lyon IM, Shousa S. Effect of chewing Quat on mucosal histology and prevalence of Helicobacter pylori in the esophagus, stomach and duodenum of Yemeni patients. Histopathology 1991;19:437-443. 5. Holcombe C. Helicobacter pylori: the African enigma. Gut 1992:33:429-431,
Histological Classification of Chronic Gastritis: An Iconoclastic View Dear Sir: 1 have read with great interest the article by Correa and Yardley’ and the editorial by Rubin’ dealing with chronic gastritis. Being a fully trained gastroenterologist and a fully trained pathologist, 1 agree with most of the comments made by Dr. Rubin, especially with those on the need of a close collaboration between pathologists and gastroenterologists and on the need for pathologists to learn more about clinical gastroenterology. 1 have, however, some problems with the definition of chronic gastritis given by Dr. Rubin in which he states that “chronic gastritis” implies
CORRESPONDENCE
1117
histological evidence of infiltration with inflammatory cells. 1 have also some problems with the statements of Drs. Correa and Yardley and Dr. Rubin on endoscopy. The definition of chronic gastritis can not depend solely on the presence of inflammatory cells. Inflammation is a very complex reaction involving not only infiltration of tissues with inflammatory cells but also alterations of epithelial cells (as a result of damage or reaction], vascular alterations, and the production and release of various mediators. Damage of structures (at the microscopic or ultrastructural level) induced by an etiologic agent is followed by a reaction that can predominantly consist of infiltration of the tissue by inflammatory cells, either polymorphs or lymphocytes or both. For lymphocytes either the B cell components or the T cells may predominate or the reaction may rather be a mixture of both. Infiltration by leukocytes may, however, be minimal or even absent. Infiltration by inflammatory cells depends on a series of events characterized by the expression of molecules on cell surfaces such as major histocompatibility complex class 11molecule? and the production and release of inflammatory mediators. These processes are now extensively studied in the gastrointestinal tract and in other organ systems, and pathologists should be aware of them. Pathologists should include the new developments of immunohistochemistry and cellular biology in the study of such complex processes as gastritis and take these new developments into account when trying to formulate definitions. The stomach may be particular insofar that normally the number of inflammatory cells (lymphocytes and/or plasma cells] within the epithelium and lamina propria is very small. Some authors4 have claimed that lymphoid tissue is absent in normal gastric mucosa, but intraepithelial lymphocytes can normally be found,5 and we have found even lymphoid follicles in otherwise normal fundic mucosa of newborns in occasional autopsy cases studied by an en face technique. In the other segments of the gastrointestinal tract lymphocytes and plasma cells are always present in the epithelium and lamina propria of the mucosa (as they are in small numbers in the gastric mucosa) as components of the gut associated lymphoepithelial tissue. Their presence is not necessarily an indication of a pathological chronic inflammation but rather the manifestationofa normal reaction towards a continuous challenge by intraluminal antigens. This reaction could therefore even be called “physiological inflammation.” In the stomach, the mucosal defense system is different from the other segments, most probably because of the production of acid and the differences in challenge. This does however not mean that challenges such as drugs (nonsteroidal anti-inflammatory drugs) or reflux of duodenal contents can not induce chronic damage. If this damage does not result in an inflammatory infiltrate, this is most likely because of the lack of production of chemotactic or other mediators. The absence or paucity of leukocytes does however not mean that there is no inflammation. Inflammation can indeed involve the immune system or nonspecific factors such as secretion and vascular alterations. The reactive gastritis of the Sydney system6 or the reflux gastritis according to Dixon et al.’ is a good example of such a reaction. As stated by Dr. Correa, the topographic distribution of the lesions might be important for the reaction and the natural history of gastritis. As such, “superficial gastritis” may still be important be it in another definition. The proliferation zone of the stomach epithelium has been shown to be localized in the depth of the crypts at the junction with the glands. In this area the mucous neck cells can be found, and these cells can undergo mitosis. Furthermore, it has been shown that the dividing cells can differentiate into epithelial cells of other types.’ Damage of this cell compartment will have consequences other than damage of surface epithelial cells or glandular cells. Major or irreversible damage of the proliferation zone might lead to atrophy.
1118
GASTROENTEROLOGY Vol. 103. No. 3
CORRESPONDENCE
Endoscopy is important for the diagnosis and classification of gastritis and at least for adequate sampling and description. It has proven to be of a major benefit in acute gastritis and also in some chronic conditions, as illustrated by the identification of congestive gastropathy. After all, endoscopy is, like pathology, a morphological approach and therefore, like pathology, no more objective nor uniformly pathognomonic. It is, however, my belief that a critical use of both methods is needed in close collaboration to further unravel the dilemmas presented by chronic gastritis, and some of them are beautifully described by Dr. Rubin to come finally to a classification including the whole spectrum of gastric lesions and to help a rational therapeutic approach. KAREL GEBOES Department of Pathology University Hospital Minderbroedersstraat 12 3000 Louvain, Belgium 1 Correa P, Yardley JH. Grading and classification
2. 3.
4.
5.
6.
7.
a.
of chronic gastritis: one American response to the Sydney system. Gastroenterology 1992;102:355-359, Rubin CE. Histological classification of chronic gastritis: an iconoclastic view. Gastroenterology 1992;102:360-361. Steiniger B, Falk P, Lohmuller M, Van der Meide PH. Class II MHC antigens in the rat digestive system. Normal distribution and induced expression after interferon-gamma treatment in vivo. Immunology 1989;68:507-513. Wotherspoon AC, Ortiz-Hidalgo C, Falzon MR, Isaacson PG. Helicobacter pylori-associated gastritis and primary B-cell gastric lymphoma. Lancet 1991;338:1175-1176. Haot J, Delos M, Wallez L, Hardy N, Lenzen B, Jouret-Mourin A. Intraepithelial lymphocytes in inflammatory gastric pathology. Acta Endosc 1986;16:187-188. Misiewicz JJ, Tytgat GNJ, Goodwin CS, Price A, Sipponen P, Strickland R, Cheli R. The Sydney system: a new classification of gastritis. Sydney, Australia: World Congress of Gastroenterology, 1990. Dixon MF, O’Connor HJ, Axon ATR, King RFJG, Johnston D. Reflux gastritis: distinct histopathological entity? J Clin Path01 1986;39:524-530. Helander HF. The cells of the gastric mucosa. Int Rev Cytol 1981;70:217-298.
HLA DRw8 and Primary Biliary Cirrhosis Dear Sir: Primary biliary cirrhosis (PBC) is regarded as an autoimmune disorder with an immunogenetic background. We wish to add a comment to the interesting article reported recently in GASTRO-
ENTEROLOGY by Manns et al.’ They investigated the association of PBC with HLA in the German population and confirmed a significant increase of HLA DRw8, which was previously reported in white PBC patients from North America.* We have also found a strong association of the disease with HLA DRw8 in Japanese patients (relative risk = 11.98; corrected P < 0.001).3 Table 1 shows a summary of statistical analysis using the x2 test or Fisher’s Exact Test and the P values. The presence of HLA DRw8 may be a risk factor common to all races for the development of PBC, as HLA B27 is for ankylosing spondylitis. Serologically defined DRw8 can be separated into four molecules with different amino acid sequences and six at the DNA level. Further investigations are needed of HLA DRw8 typing on a molecular basis, which may help to identify the factors triggering this autoimmune disease and the autoantigens involved. TAKASHI MAEDA, M.D. SABURO ONISHI, M.D. TOSHIJI SAIBARA, M.D. SHINJI IWASAKI, M.D. YASUTAKE YAMAMOTO,
M.D. First Department of Internal Medicine Kochi Medical School Nankoku, Kochi 783, Japan Manns MP, Bremm A, Schneider PM, Notghi A, Gerken G, Prager-Eberle M, Stradmann-Bellinghausen B, Meyer zum Btischenfelde KH, Rittner C. HLA DRw8 and complement C4 deficiency as risk factors in primary biliary cirrhosis. Gastroenterology 1991;101:1367-1373. Gores GJ, Moore SB, Fisher LD, Powell FC, Dickson ER. Primary biliary cirrhosis: association with class II major histocompatibility complex antigens. Hepatology 1987;7:889-893. Maeda T, Onishi S, Miyamoto T, Shingai R, Tanaka H, Iwasaki S, Ueda H, Nakata S, Iwamura S, Miyazaki M, Saibara T, Yamamoto Y. Association of primary biliary cirrhosis with HLA DRw8. Acta Hepatol Jpn 1992;33 (in press).
Reply. The highly significant increase of HLA DRw8 in Japanese patients with PBC as reported by Dr. Maeda et al. is of particular interest. It seems as if the genetic background in PBC is identical at least for white and Japanese patients. Possibly the genetic background is similar in all races as suggested by Dr. Maeda et al. However, information from other races is necessary. Furthermore, data on HLA class III antigens, i.e., complement factors C2, C4, and Bf, would be very interesting because we have described an increased frequency of C4A-QO alleles in PBC.’ We also agree with Dr. Maeda et al. that the introduction of molecular biological
Table 1. Summarv of Statistical Results Using the J Test or Fisher’s Exact Test and Relative Risk PBC
HLA locus A3 All Aw24 (9) B27 B39 (16) B51 (5) cw4 DR2 DRw8
patients
[N = 26-31) l/31 5/31 23/31 l/31 5/31 lo/31 O/26 5/28 22/28
(3.2%) (16.1%) (74.2%) (3.2%) (16.1%) (32.2%) (0.0%) (17.8%) (78.6%)
Normal controls (N = 116-136) O/136 39/136 84/136 O/136 6/136 23/136 12/118 44/128 30/128
F. Fisher’s Exact Test; x2, x2 test. “P values were corrected using the following formula: Corrected
(0.0%) (28.7%) (61.8%) (0.0%) (4.4%) (16.9%) (10.1%) (34.4%) [23.4%)
Relative risk
0.478 1.779 4.167 2.339 0.451 11.976
P = P value X No. of HLA Antigens Tasted.
P value (two-sided) 0.186 0.152 0.193 0.196 0.033 0.053 0.125 0.088
1992
prevalence rate of H. pylori antibodies in their blood.‘,* Adults in that region have the highest mortality rates from gastric cancer documented in the world today and a very high prevalence rate of intestinal metaplasia of the gastric mucosa.3 These same children have elevated blood levels of pepsinogen C, a clear sign of antral gastritis.* These findings clearly suggest that H. pylori may be involved in the etiology of this disease. The children in this highrisk area also have greater potential of endogenous nitrosation than children in low-risk areas, strongly suggesting that N-nitroso carcinogens are synthesized in the gastric microenvironment in subjects living in the highlands.* On the other hand, in the lowlands of Costa Rica the rates of gastric cancer and intestinal metaplasia are several times lower than those in the highlands. But the children of such lowlands have the same high prevalence of H. pylori infection and high serum concentrations of pepsinogen C. These additional observations in our opinion provide scientific evidence that H. pylori-associated antral gastritis (DAG) in children of the highlands correlates with high rates of intestinal metaplasia, gastric ulcers, and gastric carcinoma in adults. In the lowlands, however. the association with such diseases is not found. Similar situations are also present in other populations at very low risk for gastric carcinoma such as Yemen.4 Holcombe reported high rates of H. pylori infection since childhood and a high prevalence of gastritis in Africa but a very low prevalence of intestinal metaplasia and gastric carcinoma.5 The recent scientific evidence available in our opinion clearly adds weight to the concept that DAG and MAG are distinct entities even though H. pylori may be a factor in both diseases. PELAYO CORREA, M.D. JOHN H. YARDLEY, M.D.
Department of Pathology Louisiana State University Medical Center 1901 Perdido Street New Orleans, Louisiana 70112-1393 1.
Sierra R, Oshima H, Muiioz N, et al. Exposure to N-nitrosamines and other risk factors for gastric cancer in Costa Rican children. In: O’Neill IK, Chen S, Bartsch H, eds. Relevance to humans of N-nitroso compounds, tobacco smoke and mycotoxins. Lyon, France: International Agency for Research on Cancer, 1991:162-167. 2. Sierra R, Mufioz N, Peiia S. Antibodies to Helicobacter pylori and pepsinogen levels in children from Costa Rica. Cancer Epid Biom Prev 1992 (in press). 3. Salas J. Lesiones precancerosas de1 est6mago en Costa Rica. Patologia (MBxico) 1977;15:63-79. 4. El-Guneid A, El-Sherif AM, Murray-Lyon IM, Shousa S. Effect of chewing Quat on mucosal histology and prevalence of Helicobacter pylori in the esophagus, stomach and duodenum of Yemeni patients. Histopathology 1991;19:437-443. 5. Holcombe C. Helicobacter pylori: the African enigma. Gut 1992:33:429-431,
Histological Classification of Chronic Gastritis: An Iconoclastic View Dear Sir: 1 have read with great interest the article by Correa and Yardley’ and the editorial by Rubin’ dealing with chronic gastritis. Being a fully trained gastroenterologist and a fully trained pathologist, 1 agree with most of the comments made by Dr. Rubin, especially with those on the need of a close collaboration between pathologists and gastroenterologists and on the need for pathologists to learn more about clinical gastroenterology. 1 have, however, some problems with the definition of chronic gastritis given by Dr. Rubin in which he states that “chronic gastritis” implies
CORRESPONDENCE
1117
histological evidence of infiltration with inflammatory cells. 1 have also some problems with the statements of Drs. Correa and Yardley and Dr. Rubin on endoscopy. The definition of chronic gastritis can not depend solely on the presence of inflammatory cells. Inflammation is a very complex reaction involving not only infiltration of tissues with inflammatory cells but also alterations of epithelial cells (as a result of damage or reaction], vascular alterations, and the production and release of various mediators. Damage of structures (at the microscopic or ultrastructural level) induced by an etiologic agent is followed by a reaction that can predominantly consist of infiltration of the tissue by inflammatory cells, either polymorphs or lymphocytes or both. For lymphocytes either the B cell components or the T cells may predominate or the reaction may rather be a mixture of both. Infiltration by leukocytes may, however, be minimal or even absent. Infiltration by inflammatory cells depends on a series of events characterized by the expression of molecules on cell surfaces such as major histocompatibility complex class 11molecule? and the production and release of inflammatory mediators. These processes are now extensively studied in the gastrointestinal tract and in other organ systems, and pathologists should be aware of them. Pathologists should include the new developments of immunohistochemistry and cellular biology in the study of such complex processes as gastritis and take these new developments into account when trying to formulate definitions. The stomach may be particular insofar that normally the number of inflammatory cells (lymphocytes and/or plasma cells] within the epithelium and lamina propria is very small. Some authors4 have claimed that lymphoid tissue is absent in normal gastric mucosa, but intraepithelial lymphocytes can normally be found,5 and we have found even lymphoid follicles in otherwise normal fundic mucosa of newborns in occasional autopsy cases studied by an en face technique. In the other segments of the gastrointestinal tract lymphocytes and plasma cells are always present in the epithelium and lamina propria of the mucosa (as they are in small numbers in the gastric mucosa) as components of the gut associated lymphoepithelial tissue. Their presence is not necessarily an indication of a pathological chronic inflammation but rather the manifestationofa normal reaction towards a continuous challenge by intraluminal antigens. This reaction could therefore even be called “physiological inflammation.” In the stomach, the mucosal defense system is different from the other segments, most probably because of the production of acid and the differences in challenge. This does however not mean that challenges such as drugs (nonsteroidal anti-inflammatory drugs) or reflux of duodenal contents can not induce chronic damage. If this damage does not result in an inflammatory infiltrate, this is most likely because of the lack of production of chemotactic or other mediators. The absence or paucity of leukocytes does however not mean that there is no inflammation. Inflammation can indeed involve the immune system or nonspecific factors such as secretion and vascular alterations. The reactive gastritis of the Sydney system6 or the reflux gastritis according to Dixon et al.’ is a good example of such a reaction. As stated by Dr. Correa, the topographic distribution of the lesions might be important for the reaction and the natural history of gastritis. As such, “superficial gastritis” may still be important be it in another definition. The proliferation zone of the stomach epithelium has been shown to be localized in the depth of the crypts at the junction with the glands. In this area the mucous neck cells can be found, and these cells can undergo mitosis. Furthermore, it has been shown that the dividing cells can differentiate into epithelial cells of other types.’ Damage of this cell compartment will have consequences other than damage of surface epithelial cells or glandular cells. Major or irreversible damage of the proliferation zone might lead to atrophy.
1118
GASTROENTEROLOGY Vol. 103. No. 3
CORRESPONDENCE
Endoscopy is important for the diagnosis and classification of gastritis and at least for adequate sampling and description. It has proven to be of a major benefit in acute gastritis and also in some chronic conditions, as illustrated by the identification of congestive gastropathy. After all, endoscopy is, like pathology, a morphological approach and therefore, like pathology, no more objective nor uniformly pathognomonic. It is, however, my belief that a critical use of both methods is needed in close collaboration to further unravel the dilemmas presented by chronic gastritis, and some of them are beautifully described by Dr. Rubin to come finally to a classification including the whole spectrum of gastric lesions and to help a rational therapeutic approach. KAREL GEBOES Department of Pathology University Hospital Minderbroedersstraat 12 3000 Louvain, Belgium 1 Correa P, Yardley JH. Grading and classification
2. 3.
4.
5.
6.
7.
a.
of chronic gastritis: one American response to the Sydney system. Gastroenterology 1992;102:355-359, Rubin CE. Histological classification of chronic gastritis: an iconoclastic view. Gastroenterology 1992;102:360-361. Steiniger B, Falk P, Lohmuller M, Van der Meide PH. Class II MHC antigens in the rat digestive system. Normal distribution and induced expression after interferon-gamma treatment in vivo. Immunology 1989;68:507-513. Wotherspoon AC, Ortiz-Hidalgo C, Falzon MR, Isaacson PG. Helicobacter pylori-associated gastritis and primary B-cell gastric lymphoma. Lancet 1991;338:1175-1176. Haot J, Delos M, Wallez L, Hardy N, Lenzen B, Jouret-Mourin A. Intraepithelial lymphocytes in inflammatory gastric pathology. Acta Endosc 1986;16:187-188. Misiewicz JJ, Tytgat GNJ, Goodwin CS, Price A, Sipponen P, Strickland R, Cheli R. The Sydney system: a new classification of gastritis. Sydney, Australia: World Congress of Gastroenterology, 1990. Dixon MF, O’Connor HJ, Axon ATR, King RFJG, Johnston D. Reflux gastritis: distinct histopathological entity? J Clin Path01 1986;39:524-530. Helander HF. The cells of the gastric mucosa. Int Rev Cytol 1981;70:217-298.
HLA DRw8 and Primary Biliary Cirrhosis Dear Sir: Primary biliary cirrhosis (PBC) is regarded as an autoimmune disorder with an immunogenetic background. We wish to add a comment to the interesting article reported recently in GASTRO-
ENTEROLOGY by Manns et al.’ They investigated the association of PBC with HLA in the German population and confirmed a significant increase of HLA DRw8, which was previously reported in white PBC patients from North America.* We have also found a strong association of the disease with HLA DRw8 in Japanese patients (relative risk = 11.98; corrected P < 0.001).3 Table 1 shows a summary of statistical analysis using the x2 test or Fisher’s Exact Test and the P values. The presence of HLA DRw8 may be a risk factor common to all races for the development of PBC, as HLA B27 is for ankylosing spondylitis. Serologically defined DRw8 can be separated into four molecules with different amino acid sequences and six at the DNA level. Further investigations are needed of HLA DRw8 typing on a molecular basis, which may help to identify the factors triggering this autoimmune disease and the autoantigens involved. TAKASHI MAEDA, M.D. SABURO ONISHI, M.D. TOSHIJI SAIBARA, M.D. SHINJI IWASAKI, M.D. YASUTAKE YAMAMOTO,
M.D. First Department of Internal Medicine Kochi Medical School Nankoku, Kochi 783, Japan Manns MP, Bremm A, Schneider PM, Notghi A, Gerken G, Prager-Eberle M, Stradmann-Bellinghausen B, Meyer zum Btischenfelde KH, Rittner C. HLA DRw8 and complement C4 deficiency as risk factors in primary biliary cirrhosis. Gastroenterology 1991;101:1367-1373. Gores GJ, Moore SB, Fisher LD, Powell FC, Dickson ER. Primary biliary cirrhosis: association with class II major histocompatibility complex antigens. Hepatology 1987;7:889-893. Maeda T, Onishi S, Miyamoto T, Shingai R, Tanaka H, Iwasaki S, Ueda H, Nakata S, Iwamura S, Miyazaki M, Saibara T, Yamamoto Y. Association of primary biliary cirrhosis with HLA DRw8. Acta Hepatol Jpn 1992;33 (in press).
Reply. The highly significant increase of HLA DRw8 in Japanese patients with PBC as reported by Dr. Maeda et al. is of particular interest. It seems as if the genetic background in PBC is identical at least for white and Japanese patients. Possibly the genetic background is similar in all races as suggested by Dr. Maeda et al. However, information from other races is necessary. Furthermore, data on HLA class III antigens, i.e., complement factors C2, C4, and Bf, would be very interesting because we have described an increased frequency of C4A-QO alleles in PBC.’ We also agree with Dr. Maeda et al. that the introduction of molecular biological
Table 1. Summarv of Statistical Results Using the J Test or Fisher’s Exact Test and Relative Risk PBC
HLA locus A3 All Aw24 (9) B27 B39 (16) B51 (5) cw4 DR2 DRw8
patients
[N = 26-31) l/31 5/31 23/31 l/31 5/31 lo/31 O/26 5/28 22/28
(3.2%) (16.1%) (74.2%) (3.2%) (16.1%) (32.2%) (0.0%) (17.8%) (78.6%)
Normal controls (N = 116-136) O/136 39/136 84/136 O/136 6/136 23/136 12/118 44/128 30/128
F. Fisher’s Exact Test; x2, x2 test. “P values were corrected using the following formula: Corrected
(0.0%) (28.7%) (61.8%) (0.0%) (4.4%) (16.9%) (10.1%) (34.4%) [23.4%)
Relative risk
0.478 1.779 4.167 2.339 0.451 11.976
P = P value X No. of HLA Antigens Tasted.
P value (two-sided) 0.186 0.152 0.193 0.196 0.033 0.053 0.125 0.088
