0016-5107/92/3803-0357$03.00 GASTROINTESTINAL ENDOSCOPY Copyright © 1992 by the American Society for Gastrointestinal Endoscopy
Symptoms, gastritis, and Helicobacter pylori in patients referred for endoscopy Timothy T. Schubert, MD Arlene B. Schubert, BS Chan K. Ma, MD Detroit, Michigan
Acute Helicobacter pylori infection is associated with dyspeptic symptoms but chronic infection has not clearly been shown to cause symptoms. To define further the role of H. pylori infection and gastritis in dyspepsia, we interviewed all patients about to undergo upper endoscopy, recorded the primary indication for endoscopy, noted the endoscopic findings, and obtained antral biopsies. Among non-ulcer patients there was a strong correlation of acute gastritis with H. pylori. Gastritis and H. pylori increased with age, and non-steroidal anti-inflammatory drug use correlated with normal histology. Neither H. pylori concentration nor gastritis grade correlated with gender, use of alcohol and tobacco, indication for endoscopy, or symptoms (epigastric pain, nausea, vomiting, bloating, belching, heartburn, halitosis, and flatulence). (Gastrointest Endosc 1992;38:357-360)
Acute Helicobacter pylori gastritis, acquired by deliberate ingestion or spontaneously, has been associated with dyspeptic symptoms including epigastric burning, distention or bloating, belching, episodic nausea, flatulence, and halitosis. 1- a An association of dyspeptic symptoms with chronic infection is less clear, and most persons with the infection are asymptomatic. 4 In an analysis of consecutive patients both with and without ulcers, we previously observed an association between the concentration of H. pylori and both the severity of gastritis and the frequency of ulceration. 5 DeLuca6 has proposed that non-ulcer dyspepsia (NUD) is a result of "active" gastritis (infiltration with polymorphonuclear white blood cells). Because endoscopic findings do not predict histologic gastritis, antral biopsy has been proposed for evaluation of patients with NUD to determine the presence of H. pylori7 and the degree of gastritis. 8 If H. pylori were a significant cause of NUD, we postulated that the concentration of H. pylori should correlate with the activity of gastritis and that patients with active gastritis and greater numbers of H. pylori would be the most symptomatic. To determine the Received July 2, 1991. For revision September 24, 1991. Accepted December 6, 1991. From the Division of Gastroenterology and Department of Pathology, Henry Ford Hospital, Detroit, Michigan. Reprint requests: Timothy T. Schubert, MD, Gastroenterology K7, Henry Ford Hospital, 2799 W. Grand Boulevard, Detroit, Michigan 48202. This work was presented in part at the American Society for Gastrointestinal Endoscopy annual meeting, May 21, 1991, New Orleans. VOLUME 38, NO.3, 1992
relation between dyspeptic symptoms, activity of gastritis and H. pylori density, we questioned consecutive patients immediately before endoscopy regarding dyspeptic symptoms and obtained antral biopsies for H. pylori and histologic examination at the time of endoscopy. Ultimately, we hoped to identify a group of patients with striking symptoms and histology who would be most likely to benefit from treatment. METHODS
Consecutive consenting patients referred to the senior author (T. S.) for upper endoscopy from November 1988 to September 1990 were questioned regarding the presence of upper abdominal pain, nausea, bloating, belching, halitosis, vomiting, and heartburn any time in the 4 weeks preceding endoscopy. The primary indication for endoscopy was noted, and at endoscopy abnormalities were recorded. Two antral biopsies were obtained for histology. Excluded from analysis were those unable to give a history (N = 11), those with duodenal or gastric ulcers (N = 128), prior antrectomy (N = 18), on antibiotics in the prior 4 weeks (N = 41), gastric or esophageal cancer (N = 8), active bleeding, or on anticoagulants (N = 12). Histological evaluation for H. pylori was made after staining of the tissue by the Warthin Starry (WS) method. A single pathologist (C. M.), blinded as to the patient's symptoms and the endoscopic findings, identified the organism on the mucosal surface or within the foveolae using a 40power lens and searching multiple fields. The density of H. pylori was graded as follows: no organisms seen = 0, occasional or few organisms seen = 1, and many organisms seen = 2.
357
Gastritis was evaluated on H&E stained sections. Following the Sydney convention, normal (N) biopsies were those with no acute inflammatory cells and none or only occasional chronic (mononuclear) cells. The interpretation was "chronic gastritis" if there was an excess of mononuclear cells but no polymorphonuclear cells. When polymorphonuclear cells were seen the interpretation was "active chronic gastritis" (ACG). SAS Institute Inc. package was utilized. Wilcoxon rank sum test was used to evaluate the association between the WS and gastritis categories and each of the categorical study variables (gender, smoking, alcohol, non-steroidal anti-inflammatory drug use, and each dyspeptic symptom). Spearman rank correlation was used to correlate WS grade with level of gastritis and to correlate age with the WS grades and levels of gastritis. Stepwise logistic regression was used for multivariate analysis. A p < 0.05 was regarded as significant.
There was a significant degree of correlation (r = 0.653) between the grade of H. pylori and the level of gastritis (p < 0.001) (Fig. 1). Among the 156 patients in whom the histology was normal, the WS was positive only twice. Among the 141 with chronic gastritis, WS = 0 in 109 (77%), WS = 1 in 21 (15%), and WS = 2 in 11 (8%). Among the 177 with ACG, 134 (76%) had a positive WS. Among those with WS = 1, 73% had ACG and with WS = 2, 86% had ACG. NSAID use tended to be associated with less inflammation on univariate analysis and on multivariate analysis (accounting for age) this reached statistical significance (p < 0.05) (Fig. 2). The same trend was observed with H. pylori concentration but this did not achieve statistical signinficance. The primary indications for endoscopy were divided into four groups: epigastric pain, 243 (51%); esophageal,85 (18%) (including dysphagia, non-cardiac chest pain, and refractory reflux); blood loss, 82 (17%) (including acute and chronic), and other, 62 (13%). There was no significant difference in the WS grade nor in the gastritis grade among the four groups (Fig. 3). H. pylori concentration (WS grade) was compared with each of seven gastrointestinal symptoms (Table 2). Symptoms were of similar frequency in those without H. pylori, with low grade, and with higher grade of H. pylori. There appeared to be no symptom which was more likely to occur in H. pylori-positive compared with H. pylori-negative patients. Likewise, the
RESULTS
There were 474 patients included in the analysis. The mean age of the study population was 52.6 years (SD, 16.9 years) with a range of 15 to 90 years. There were 226 females (47.7%) and 248 males (52.3%). The ethnic composition was: whites, 224 (47.2%); blacks, 232 (49.0%); middle eastern, 8 (1.7%); Hispanic, 6 (1.3%), and far eastern, 4 (0.8%). Forty-one percent were smokers and 42% had used alcohol in the previous 4 weeks. Thirty-four percent of patients had taken either aspirin or non-aspirin non-steroidal antiinflammatory drugs (NSAIDs) in the preceding 4 weeks. H. pylori was absent (grade 0) in 305 patients (64.3%) and present in 169 patients (35.6%): 83 patients had grade 1 (17.5%) and 86 had grade 2 (18.1 %) (Table 1). Gender, smoking, and alcohol use were similar for the three grades of H. pylori but there was an increase in age with increasing H. pylori grade (p < 0.05). Histology was normal in 156 patients (32.9%), showed chronic gastritis in 141 (29.7%), and ACG in 177 patients (37.3%). Increasing age was associated with more severe gastritis (p < 0.05). Those having normal histology were younger than those with chronic gastritis who in turn were younger than those with ACG (Table 1).
100 N=86
80
....I::
Gastritis
60
"e
~
40
•
None
D
Chronic
EJ
Active
20
a 0 1 2 Warthin Starry Grade
Figure 1. Relation of Warthin Starry grade and gastritis.
Table 1. Demographic distribution by WS density and grade of gastritis Gastritis gradeb
Density of H. pylori on WS"
Number of patients Age (yr) Gender (female) (%) Smokers (%) EtOH (%)
0
1
2
p
N
CG
ACG
p
305 51.2 46.6 42.1 44.1
83 55.1 45.8 42.0 38.6
86 55.3 53.5 38.6 40.5
0.01 0.39 0.70 0.72
156 49.0 44.2 45.4 47.7
141 52.5 48.2 41.7 38.6
177 55.8 50.3 37.7 41.0
0.001 0.27 0.37 0.27
" 0 = no H. pylori, 1 = few or occasional H. pylori, 2 = many H. pylori. b N, none or occasional mononuclear cell; CG, chronic gastritis; ACG, active chronic gastritis.
358
GASTROINTESTINAL ENDOSCOPY
presence or activity of gastritis did not correlate with any of the symptoms (Table 2).
Others have noted histologic gastritis to correlate with H. pylori rather than NSAIDs. 12 We expected those in whom the primary indication for endoscopy was epigastric pain to have a greater frequency of H. pylori and active gastritis. We found neither H. pylori nor gastritis differed according to the primary indication for endoscopy. Contrary to the DeLuca hypothesis,6 "no active gastritis, no dyspepsia," there appeared to be no greater incidence of symptoms associated with active gastritis compared with those without gastritis or those with chronic gastritis. Prior to the current focus on H. pylori, gastritis was not found to correlate with symptoms in patients without ulcers. 13.14 Whether symptoms are associated with chronic H. pylori infection has been controversial. One report claims a "remarkably consistent pattern of symptoms" in patients subsequently found to have histologic gastritis and Campylobacter pyloridis. 9 The pattern of symptoms noted consisted of sternal or epigastric burning or discomfort, belching, reflux, epigastric distention, periodic nausea, flatulence, and halitosis. Others have reported more frequent abdominal pain and vomiting among H. pylori-positive nonulcer patients but no association of "ructus," other symptoms, smoking, alcohol consumption, or dietary habits. 15 Among 59 patients referred for endoscopy, another group found heartburn and burping more frequent among H. pylori-positive patients and loss of appetite less common but no difference in epigastric or retrosternal pain, nausea, vomiting, or belching,16 One study showed only "burping" to be associated with H. pylori in patients referred for endoscopy17 whereas another group found no correlation of symptoms with H. pylori. 18 In our endoscopic population there appeared to be no symptom or constellation of symptoms which occurred more frequently in H. pylori-positive patients or in patients with high concentrations of H. pylori compared with those without H. pylori. Endoscopic studies have reported variable frequencies of H. pylori among those with dyspepsia compared with controls. Rokkas et aU 9 noted a 45% prevalence
DISCUSSION
The present study confirmed the well-established association of H. pylori prevalence with gastritis9 • 10 as well as the known increase of H. pylori prevalence and activity of gastritis with age. l1 There was no correlation of H. pylori or gastritis with gender or with use of tobacco or alcohol in the preceding 4 weeks. NSAID use was associated with lesser grades of gastritis. 45 40 % of Patients on NSAID
25 15 5
o
o
none chronic active
2
Warthin Starry
Gastritis Grade
Figure 2. Relation of NSAID use to Warthin Starry density and gastritis grade. Gastritis Grade
% Warthln Starry
WS Gastritis
lSI
o
0
None
60 50
I!lI
1 2
CG ACG
40 30 20 10 0
a
b
c
a
d
a - eplgastrlc paln b - esophageal
b
c
d
c - blood loss d - other
Figure 3. Warthin Starry density and gastritis grade according to indication for endoscopy.
Table 2. Symptoms by WS density and grade of gastritis Gastritis gradea
Density of H. pylori on WS Symptom Pain Nausea Bloating Belching Heartburn Vomiting Flatulence Halitosis
0
1
2
p
N
CG
ACG
p
59.9 48.7 42.4 38.6 35.0 24.0 32.8 19.7
63.9 48.2 43.4 39.8 40.9 18.1 29.1 19.8
61.6 40.7 45.9 41.2 36.2 18.6 32.1 17.6
0.60 0.28 0.60 0.67 0.95 0.17 0.77 0.84
61.9 49.4 43.2 37.7 40.3 28.4 36.2 21.3
60.3 51.4 44.3 36.4 30.2 17.0 28.6 20.6
60.5 41.8 42.4 42.9 38.4 20.3 31.6 16.6
0.79 0.12 0.86 0.31 0.14 0.10 0.40 0.32
a N, none or occasional mononuclear cell; CG, chronic gastritis; ACG, active chronic gastritis, polymorphonuclear and mononuclear cells.
VOLUME 38, NO.3, 1992
359
of H. pylori among patients with non-ulcer dyspepsia compared with 13% prevalence among controls (patients undergoing gastroscopy for iron deficiency anemia). The control group was younger and had predominantly females compared with the NUD group. Others have shown no difference. Two reports, one studying Hispanics in Los Angeles 20 and another studying Nigerians,21 found no difference in H. pylori prevalence among dyspeptics and controls but the H. pylori prevalence in these populations was over 80%. A study from a city hospital in Kansas City found H. pylori prevalence to be 40% in both controls and those with NUD. 22 Epidemiologic studies have suggested a relation of H. pylori and dyspepsia. One, done in the United States, demonstrated patients regularly using antacids had a 45% prevalence of H. pylori by serology compared with 14% prevalence among those who did not use antacids. 23 In a large epidemiology study in China, dyspepsia was likewise more common in persons with H. pylori by serology.24 These studies suggested a relationship between symptoms and H. pylori, but it is possible that the difference was due to increased numbers of individuals with ulcers among the H. pylori group or (conceivably) that taking antacids predisposes to H. pylori infection. Results of treatment studies also conflict, although the problems in study design preclude firm conclusions. 25 ,26 Trials of bismuth compounds in NUD have shown improvement of symptoms in those in whom H. pylori has been eradicated and worsening in those in whom H. pylori persists or develops.27 Others have failed to show significant symptom improvement in patients with H. pylori and NUD with eradication of H. pylori. 28, 29 Uncontrolled trials demonstrate apparent long-term benefit with eradication ofthe organism in patients with NUD when compared with those in whom the organism was not eradicated. 30 Clearly, welldesigned prospective double-blind studies are needed to determine if there is benefit in treating patients with NUD and H. pylori gastritis.
REFERENCES 1. Marshall B, Armstrong J, McGechie D, Glancy R. Attempt to fulfil Koch's postulates for pyloric Campylobacter. Med J Aust 1985;142:436-9. 2. Morris, A, Nicholson G. Ingestion of Campylobacter pyloridis causes gastritis and raised fasting gastric pH. Am J GastroenteroI1987;82:192-9. 3. Frommer D, Carrick J, Lee A, Hazell S. Acute presentation of Campylobaeter pylori gastritis. Am J Gastroenterol 1988;83:1168-71. 4. Dooley C, Cohen H, Fitzgibbons P, et al. Prevalence of Helieobacter pylori infection and histologic gastritis in asymptomatic persons. N Engl J Med 1989;321:1562-6. 5. Alam K, Schubert T, Bologna SD, Chan K. Density of Helieobaeter pylori-a measure of virulence [Abstract 55]? Am J Gastroenterol1990;85:1228. 6. DeLuca V. No acid, no polys-no "active" gastritis, no dyspepsia. A proposal. J Clin GastroenteroI1989;11:127-31.
360
7. Antonescu C, Marshall B, Frierson H, Guerrant R, McCallum R. How does the endoscopist decide which patients should be biopsied for H. pylori [Abstract 56]? Am J Gastroenterol 1990;85:1229. 8. Vaira D, Holton J, Osborn J, D'Anna L, et al. Endoscopy in dyspeptic patients: is gastric mucosal biopsy useful? Am J Gastroenterol 1990;85:701-4. 9. Rauws E, Langenberg W, Houthoff H, Zanen J, Tytgat GNJ. Campylobaeter pyloridis-associated chronic antral gastritis.: a prospective study of its prevalence and the effects of antibacterial treatment. Gastroenterology 1988;94:33-40. 10. Goodwin C, Armstrong J, Marshall B. Campylobaeter pyloridis, gastritis, and peptic ulceration. J Clin Pathol 1986;39:353-65. 11. Greenberg R, Bank S. The prevalence of Helieobaeter pylori in nonulcer dyspepsia; importance of stratification according to age. Arch Intern Med 1990;150:2053-5. 12. Laine L, Marin-Sorenson M, Weinstein W. Helieobaeter pylori prevalence and mucosal injury in gastric ulcers: relationship to chronic nonsteroidal anti-inflammatory drug ingestion. Gastroenterology 1991;100:A103. 13. Siurala M, Sipponen P, Kekki M. Chronic gastritis: dynamic and clinical aspects. Scand J Gastroenterol 1985;20(suppl 109):69-76. 14. Cheli R, Perasso A, Giacosa A. Dyspepsia and chronic gastritis. Hepatogastroenterol. 1983;30:21-3. 15. Andersen L, Elsborg L, Justesen T. Campylobaeter pylori in peptic ulcer disease. III. Symptoms and paraclinical and epidemiologic findings. Scand J GastroenteroI1988;23:347-50. 16. Loffeld R, Adang R. Computerized history taking for the assessment of Helieobacter pylori in non-ulcer dyspepsia. Rev Esp Enf Digest 1990;70S(suppl1):70-1. 17. Marshal B, Warren J. Unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration. Lancet 1984;1:1311-5. 18. Borsch G, Schmidt G, Wegner M, et al. Campylobaetor pylori:: prospective analysis of clinical and histological factors associated with colonization of the upper gastrointestinal tract. Eur J Clin Invest 1988;18:133-8. 19. Rokkas T, Pursey C, Uzoechina E, et al. Campylobaeter pylori and non-ulcer dyspepsia. Am J GastroenteroI1987;82:1149-52. 20. Dooley C, Arandia D, Dehesa M, Cohen H, Fitzgibbons P. H pylori is not an etiologic factor in nonulcer dyspepsia [Abstract] Am J GastroenteroI1990;85:1232. 21. Holcombe C, Thorn C, Kaluba J, Lucas S. Non-ulcer dyspepsia and Helieobaeter pylori in Northern Nigerians. Rev Esp Enferm Dig 1990;78(suppl1):71. 22. Schubert T, Schnell G. Prevalence of Campylobaeter pylori in patients undergoing upper endoscopy. Am J Gastroenterol 1989;84:637-42. 23. Skoglund M, Whalen J. Correlation of C. pylori and upper GI symptoms. Gastroenterology 1988;94:A241. 24. Yang H, Zhou D. Seroepidemiology of Campylobaeter pylori infection in China. Rev Esp Enferm Dig 1990;78(suppl1):130. 25. Veldhuyzen van Zanten S, Tytgat K, Jalali S, Goodacre R, Hunt R. Can gastritis symptoms be evaluated in clinical trials? An overview of treatment of gastritis, nonulcer dyspepsia and Campylobacter-associated gastritis. J Clin Gastroenterol 1989;11:496-501. 26. Talley N. Campylobaeter pylori and symptoms: Is there a cause and effect relationship [Letter]? Gastroenterology 1989;96:1228. 27. Rokkas T, Pursey C, Uzoechina E, et al. Non-ulcer dyspepsia and short term De-Nol therapy: a placebo controlled trial with particular reference to the role of Campylobaeter pylori. Gut 1988;29;138691. 28. Loffeld R, Potters H, Stobberingh E, Flendrig J, Van Spreeuwel J, Arends J. Campylobacter associated gastritis in patients with non-ulcer dyspepsia: a double blind placebo controlled trial with colloidal bismuth subcitrate. Gut 1989;30:1206-12. 29. Patchett S, Beattie S, Leen E, Keane C, O'Morain C. Helicobacter in non ulcer dyspepsia a controlled trial. Rev Esp Enferm Dig 1990;78(suppl1):122. 30. Marshall B, Hoffman S, Boyd C, McCallum R. Does eradication of H. pylori have a long-term benefit in patients with dyspepsia and gastritis [Abstract]? Am J Gastroenterol 1990;85:1237.
GASTROINTESTINAL ENDOSCOPY
Symptoms, gastritis, and Helicobacter pylori in patients referred for endoscopy Timothy T. Schubert, MD Arlene B. Schubert, BS Chan K. Ma, MD Detroit, Michigan
Acute Helicobacter pylori infection is associated with dyspeptic symptoms but chronic infection has not clearly been shown to cause symptoms. To define further the role of H. pylori infection and gastritis in dyspepsia, we interviewed all patients about to undergo upper endoscopy, recorded the primary indication for endoscopy, noted the endoscopic findings, and obtained antral biopsies. Among non-ulcer patients there was a strong correlation of acute gastritis with H. pylori. Gastritis and H. pylori increased with age, and non-steroidal anti-inflammatory drug use correlated with normal histology. Neither H. pylori concentration nor gastritis grade correlated with gender, use of alcohol and tobacco, indication for endoscopy, or symptoms (epigastric pain, nausea, vomiting, bloating, belching, heartburn, halitosis, and flatulence). (Gastrointest Endosc 1992;38:357-360)
Acute Helicobacter pylori gastritis, acquired by deliberate ingestion or spontaneously, has been associated with dyspeptic symptoms including epigastric burning, distention or bloating, belching, episodic nausea, flatulence, and halitosis. 1- a An association of dyspeptic symptoms with chronic infection is less clear, and most persons with the infection are asymptomatic. 4 In an analysis of consecutive patients both with and without ulcers, we previously observed an association between the concentration of H. pylori and both the severity of gastritis and the frequency of ulceration. 5 DeLuca6 has proposed that non-ulcer dyspepsia (NUD) is a result of "active" gastritis (infiltration with polymorphonuclear white blood cells). Because endoscopic findings do not predict histologic gastritis, antral biopsy has been proposed for evaluation of patients with NUD to determine the presence of H. pylori7 and the degree of gastritis. 8 If H. pylori were a significant cause of NUD, we postulated that the concentration of H. pylori should correlate with the activity of gastritis and that patients with active gastritis and greater numbers of H. pylori would be the most symptomatic. To determine the Received July 2, 1991. For revision September 24, 1991. Accepted December 6, 1991. From the Division of Gastroenterology and Department of Pathology, Henry Ford Hospital, Detroit, Michigan. Reprint requests: Timothy T. Schubert, MD, Gastroenterology K7, Henry Ford Hospital, 2799 W. Grand Boulevard, Detroit, Michigan 48202. This work was presented in part at the American Society for Gastrointestinal Endoscopy annual meeting, May 21, 1991, New Orleans. VOLUME 38, NO.3, 1992
relation between dyspeptic symptoms, activity of gastritis and H. pylori density, we questioned consecutive patients immediately before endoscopy regarding dyspeptic symptoms and obtained antral biopsies for H. pylori and histologic examination at the time of endoscopy. Ultimately, we hoped to identify a group of patients with striking symptoms and histology who would be most likely to benefit from treatment. METHODS
Consecutive consenting patients referred to the senior author (T. S.) for upper endoscopy from November 1988 to September 1990 were questioned regarding the presence of upper abdominal pain, nausea, bloating, belching, halitosis, vomiting, and heartburn any time in the 4 weeks preceding endoscopy. The primary indication for endoscopy was noted, and at endoscopy abnormalities were recorded. Two antral biopsies were obtained for histology. Excluded from analysis were those unable to give a history (N = 11), those with duodenal or gastric ulcers (N = 128), prior antrectomy (N = 18), on antibiotics in the prior 4 weeks (N = 41), gastric or esophageal cancer (N = 8), active bleeding, or on anticoagulants (N = 12). Histological evaluation for H. pylori was made after staining of the tissue by the Warthin Starry (WS) method. A single pathologist (C. M.), blinded as to the patient's symptoms and the endoscopic findings, identified the organism on the mucosal surface or within the foveolae using a 40power lens and searching multiple fields. The density of H. pylori was graded as follows: no organisms seen = 0, occasional or few organisms seen = 1, and many organisms seen = 2.
357
Gastritis was evaluated on H&E stained sections. Following the Sydney convention, normal (N) biopsies were those with no acute inflammatory cells and none or only occasional chronic (mononuclear) cells. The interpretation was "chronic gastritis" if there was an excess of mononuclear cells but no polymorphonuclear cells. When polymorphonuclear cells were seen the interpretation was "active chronic gastritis" (ACG). SAS Institute Inc. package was utilized. Wilcoxon rank sum test was used to evaluate the association between the WS and gastritis categories and each of the categorical study variables (gender, smoking, alcohol, non-steroidal anti-inflammatory drug use, and each dyspeptic symptom). Spearman rank correlation was used to correlate WS grade with level of gastritis and to correlate age with the WS grades and levels of gastritis. Stepwise logistic regression was used for multivariate analysis. A p < 0.05 was regarded as significant.
There was a significant degree of correlation (r = 0.653) between the grade of H. pylori and the level of gastritis (p < 0.001) (Fig. 1). Among the 156 patients in whom the histology was normal, the WS was positive only twice. Among the 141 with chronic gastritis, WS = 0 in 109 (77%), WS = 1 in 21 (15%), and WS = 2 in 11 (8%). Among the 177 with ACG, 134 (76%) had a positive WS. Among those with WS = 1, 73% had ACG and with WS = 2, 86% had ACG. NSAID use tended to be associated with less inflammation on univariate analysis and on multivariate analysis (accounting for age) this reached statistical significance (p < 0.05) (Fig. 2). The same trend was observed with H. pylori concentration but this did not achieve statistical signinficance. The primary indications for endoscopy were divided into four groups: epigastric pain, 243 (51%); esophageal,85 (18%) (including dysphagia, non-cardiac chest pain, and refractory reflux); blood loss, 82 (17%) (including acute and chronic), and other, 62 (13%). There was no significant difference in the WS grade nor in the gastritis grade among the four groups (Fig. 3). H. pylori concentration (WS grade) was compared with each of seven gastrointestinal symptoms (Table 2). Symptoms were of similar frequency in those without H. pylori, with low grade, and with higher grade of H. pylori. There appeared to be no symptom which was more likely to occur in H. pylori-positive compared with H. pylori-negative patients. Likewise, the
RESULTS
There were 474 patients included in the analysis. The mean age of the study population was 52.6 years (SD, 16.9 years) with a range of 15 to 90 years. There were 226 females (47.7%) and 248 males (52.3%). The ethnic composition was: whites, 224 (47.2%); blacks, 232 (49.0%); middle eastern, 8 (1.7%); Hispanic, 6 (1.3%), and far eastern, 4 (0.8%). Forty-one percent were smokers and 42% had used alcohol in the previous 4 weeks. Thirty-four percent of patients had taken either aspirin or non-aspirin non-steroidal antiinflammatory drugs (NSAIDs) in the preceding 4 weeks. H. pylori was absent (grade 0) in 305 patients (64.3%) and present in 169 patients (35.6%): 83 patients had grade 1 (17.5%) and 86 had grade 2 (18.1 %) (Table 1). Gender, smoking, and alcohol use were similar for the three grades of H. pylori but there was an increase in age with increasing H. pylori grade (p < 0.05). Histology was normal in 156 patients (32.9%), showed chronic gastritis in 141 (29.7%), and ACG in 177 patients (37.3%). Increasing age was associated with more severe gastritis (p < 0.05). Those having normal histology were younger than those with chronic gastritis who in turn were younger than those with ACG (Table 1).
100 N=86
80
....I::
Gastritis
60
"e
~
40
•
None
D
Chronic
EJ
Active
20
a 0 1 2 Warthin Starry Grade
Figure 1. Relation of Warthin Starry grade and gastritis.
Table 1. Demographic distribution by WS density and grade of gastritis Gastritis gradeb
Density of H. pylori on WS"
Number of patients Age (yr) Gender (female) (%) Smokers (%) EtOH (%)
0
1
2
p
N
CG
ACG
p
305 51.2 46.6 42.1 44.1
83 55.1 45.8 42.0 38.6
86 55.3 53.5 38.6 40.5
0.01 0.39 0.70 0.72
156 49.0 44.2 45.4 47.7
141 52.5 48.2 41.7 38.6
177 55.8 50.3 37.7 41.0
0.001 0.27 0.37 0.27
" 0 = no H. pylori, 1 = few or occasional H. pylori, 2 = many H. pylori. b N, none or occasional mononuclear cell; CG, chronic gastritis; ACG, active chronic gastritis.
358
GASTROINTESTINAL ENDOSCOPY
presence or activity of gastritis did not correlate with any of the symptoms (Table 2).
Others have noted histologic gastritis to correlate with H. pylori rather than NSAIDs. 12 We expected those in whom the primary indication for endoscopy was epigastric pain to have a greater frequency of H. pylori and active gastritis. We found neither H. pylori nor gastritis differed according to the primary indication for endoscopy. Contrary to the DeLuca hypothesis,6 "no active gastritis, no dyspepsia," there appeared to be no greater incidence of symptoms associated with active gastritis compared with those without gastritis or those with chronic gastritis. Prior to the current focus on H. pylori, gastritis was not found to correlate with symptoms in patients without ulcers. 13.14 Whether symptoms are associated with chronic H. pylori infection has been controversial. One report claims a "remarkably consistent pattern of symptoms" in patients subsequently found to have histologic gastritis and Campylobacter pyloridis. 9 The pattern of symptoms noted consisted of sternal or epigastric burning or discomfort, belching, reflux, epigastric distention, periodic nausea, flatulence, and halitosis. Others have reported more frequent abdominal pain and vomiting among H. pylori-positive nonulcer patients but no association of "ructus," other symptoms, smoking, alcohol consumption, or dietary habits. 15 Among 59 patients referred for endoscopy, another group found heartburn and burping more frequent among H. pylori-positive patients and loss of appetite less common but no difference in epigastric or retrosternal pain, nausea, vomiting, or belching,16 One study showed only "burping" to be associated with H. pylori in patients referred for endoscopy17 whereas another group found no correlation of symptoms with H. pylori. 18 In our endoscopic population there appeared to be no symptom or constellation of symptoms which occurred more frequently in H. pylori-positive patients or in patients with high concentrations of H. pylori compared with those without H. pylori. Endoscopic studies have reported variable frequencies of H. pylori among those with dyspepsia compared with controls. Rokkas et aU 9 noted a 45% prevalence
DISCUSSION
The present study confirmed the well-established association of H. pylori prevalence with gastritis9 • 10 as well as the known increase of H. pylori prevalence and activity of gastritis with age. l1 There was no correlation of H. pylori or gastritis with gender or with use of tobacco or alcohol in the preceding 4 weeks. NSAID use was associated with lesser grades of gastritis. 45 40 % of Patients on NSAID
25 15 5
o
o
none chronic active
2
Warthin Starry
Gastritis Grade
Figure 2. Relation of NSAID use to Warthin Starry density and gastritis grade. Gastritis Grade
% Warthln Starry
WS Gastritis
lSI
o
0
None
60 50
I!lI
1 2
CG ACG
40 30 20 10 0
a
b
c
a
d
a - eplgastrlc paln b - esophageal
b
c
d
c - blood loss d - other
Figure 3. Warthin Starry density and gastritis grade according to indication for endoscopy.
Table 2. Symptoms by WS density and grade of gastritis Gastritis gradea
Density of H. pylori on WS Symptom Pain Nausea Bloating Belching Heartburn Vomiting Flatulence Halitosis
0
1
2
p
N
CG
ACG
p
59.9 48.7 42.4 38.6 35.0 24.0 32.8 19.7
63.9 48.2 43.4 39.8 40.9 18.1 29.1 19.8
61.6 40.7 45.9 41.2 36.2 18.6 32.1 17.6
0.60 0.28 0.60 0.67 0.95 0.17 0.77 0.84
61.9 49.4 43.2 37.7 40.3 28.4 36.2 21.3
60.3 51.4 44.3 36.4 30.2 17.0 28.6 20.6
60.5 41.8 42.4 42.9 38.4 20.3 31.6 16.6
0.79 0.12 0.86 0.31 0.14 0.10 0.40 0.32
a N, none or occasional mononuclear cell; CG, chronic gastritis; ACG, active chronic gastritis, polymorphonuclear and mononuclear cells.
VOLUME 38, NO.3, 1992
359
of H. pylori among patients with non-ulcer dyspepsia compared with 13% prevalence among controls (patients undergoing gastroscopy for iron deficiency anemia). The control group was younger and had predominantly females compared with the NUD group. Others have shown no difference. Two reports, one studying Hispanics in Los Angeles 20 and another studying Nigerians,21 found no difference in H. pylori prevalence among dyspeptics and controls but the H. pylori prevalence in these populations was over 80%. A study from a city hospital in Kansas City found H. pylori prevalence to be 40% in both controls and those with NUD. 22 Epidemiologic studies have suggested a relation of H. pylori and dyspepsia. One, done in the United States, demonstrated patients regularly using antacids had a 45% prevalence of H. pylori by serology compared with 14% prevalence among those who did not use antacids. 23 In a large epidemiology study in China, dyspepsia was likewise more common in persons with H. pylori by serology.24 These studies suggested a relationship between symptoms and H. pylori, but it is possible that the difference was due to increased numbers of individuals with ulcers among the H. pylori group or (conceivably) that taking antacids predisposes to H. pylori infection. Results of treatment studies also conflict, although the problems in study design preclude firm conclusions. 25 ,26 Trials of bismuth compounds in NUD have shown improvement of symptoms in those in whom H. pylori has been eradicated and worsening in those in whom H. pylori persists or develops.27 Others have failed to show significant symptom improvement in patients with H. pylori and NUD with eradication of H. pylori. 28, 29 Uncontrolled trials demonstrate apparent long-term benefit with eradication ofthe organism in patients with NUD when compared with those in whom the organism was not eradicated. 30 Clearly, welldesigned prospective double-blind studies are needed to determine if there is benefit in treating patients with NUD and H. pylori gastritis.
REFERENCES 1. Marshall B, Armstrong J, McGechie D, Glancy R. Attempt to fulfil Koch's postulates for pyloric Campylobacter. Med J Aust 1985;142:436-9. 2. Morris, A, Nicholson G. Ingestion of Campylobacter pyloridis causes gastritis and raised fasting gastric pH. Am J GastroenteroI1987;82:192-9. 3. Frommer D, Carrick J, Lee A, Hazell S. Acute presentation of Campylobaeter pylori gastritis. Am J Gastroenterol 1988;83:1168-71. 4. Dooley C, Cohen H, Fitzgibbons P, et al. Prevalence of Helieobacter pylori infection and histologic gastritis in asymptomatic persons. N Engl J Med 1989;321:1562-6. 5. Alam K, Schubert T, Bologna SD, Chan K. Density of Helieobaeter pylori-a measure of virulence [Abstract 55]? Am J Gastroenterol1990;85:1228. 6. DeLuca V. No acid, no polys-no "active" gastritis, no dyspepsia. A proposal. J Clin GastroenteroI1989;11:127-31.
360
7. Antonescu C, Marshall B, Frierson H, Guerrant R, McCallum R. How does the endoscopist decide which patients should be biopsied for H. pylori [Abstract 56]? Am J Gastroenterol 1990;85:1229. 8. Vaira D, Holton J, Osborn J, D'Anna L, et al. Endoscopy in dyspeptic patients: is gastric mucosal biopsy useful? Am J Gastroenterol 1990;85:701-4. 9. Rauws E, Langenberg W, Houthoff H, Zanen J, Tytgat GNJ. Campylobaeter pyloridis-associated chronic antral gastritis.: a prospective study of its prevalence and the effects of antibacterial treatment. Gastroenterology 1988;94:33-40. 10. Goodwin C, Armstrong J, Marshall B. Campylobaeter pyloridis, gastritis, and peptic ulceration. J Clin Pathol 1986;39:353-65. 11. Greenberg R, Bank S. The prevalence of Helieobaeter pylori in nonulcer dyspepsia; importance of stratification according to age. Arch Intern Med 1990;150:2053-5. 12. Laine L, Marin-Sorenson M, Weinstein W. Helieobaeter pylori prevalence and mucosal injury in gastric ulcers: relationship to chronic nonsteroidal anti-inflammatory drug ingestion. Gastroenterology 1991;100:A103. 13. Siurala M, Sipponen P, Kekki M. Chronic gastritis: dynamic and clinical aspects. Scand J Gastroenterol 1985;20(suppl 109):69-76. 14. Cheli R, Perasso A, Giacosa A. Dyspepsia and chronic gastritis. Hepatogastroenterol. 1983;30:21-3. 15. Andersen L, Elsborg L, Justesen T. Campylobaeter pylori in peptic ulcer disease. III. Symptoms and paraclinical and epidemiologic findings. Scand J GastroenteroI1988;23:347-50. 16. Loffeld R, Adang R. Computerized history taking for the assessment of Helieobacter pylori in non-ulcer dyspepsia. Rev Esp Enf Digest 1990;70S(suppl1):70-1. 17. Marshal B, Warren J. Unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration. Lancet 1984;1:1311-5. 18. Borsch G, Schmidt G, Wegner M, et al. Campylobaetor pylori:: prospective analysis of clinical and histological factors associated with colonization of the upper gastrointestinal tract. Eur J Clin Invest 1988;18:133-8. 19. Rokkas T, Pursey C, Uzoechina E, et al. Campylobaeter pylori and non-ulcer dyspepsia. Am J GastroenteroI1987;82:1149-52. 20. Dooley C, Arandia D, Dehesa M, Cohen H, Fitzgibbons P. H pylori is not an etiologic factor in nonulcer dyspepsia [Abstract] Am J GastroenteroI1990;85:1232. 21. Holcombe C, Thorn C, Kaluba J, Lucas S. Non-ulcer dyspepsia and Helieobaeter pylori in Northern Nigerians. Rev Esp Enferm Dig 1990;78(suppl1):71. 22. Schubert T, Schnell G. Prevalence of Campylobaeter pylori in patients undergoing upper endoscopy. Am J Gastroenterol 1989;84:637-42. 23. Skoglund M, Whalen J. Correlation of C. pylori and upper GI symptoms. Gastroenterology 1988;94:A241. 24. Yang H, Zhou D. Seroepidemiology of Campylobaeter pylori infection in China. Rev Esp Enferm Dig 1990;78(suppl1):130. 25. Veldhuyzen van Zanten S, Tytgat K, Jalali S, Goodacre R, Hunt R. Can gastritis symptoms be evaluated in clinical trials? An overview of treatment of gastritis, nonulcer dyspepsia and Campylobacter-associated gastritis. J Clin Gastroenterol 1989;11:496-501. 26. Talley N. Campylobaeter pylori and symptoms: Is there a cause and effect relationship [Letter]? Gastroenterology 1989;96:1228. 27. Rokkas T, Pursey C, Uzoechina E, et al. Non-ulcer dyspepsia and short term De-Nol therapy: a placebo controlled trial with particular reference to the role of Campylobaeter pylori. Gut 1988;29;138691. 28. Loffeld R, Potters H, Stobberingh E, Flendrig J, Van Spreeuwel J, Arends J. Campylobacter associated gastritis in patients with non-ulcer dyspepsia: a double blind placebo controlled trial with colloidal bismuth subcitrate. Gut 1989;30:1206-12. 29. Patchett S, Beattie S, Leen E, Keane C, O'Morain C. Helicobacter in non ulcer dyspepsia a controlled trial. Rev Esp Enferm Dig 1990;78(suppl1):122. 30. Marshall B, Hoffman S, Boyd C, McCallum R. Does eradication of H. pylori have a long-term benefit in patients with dyspepsia and gastritis [Abstract]? Am J Gastroenterol 1990;85:1237.
GASTROINTESTINAL ENDOSCOPY
