Digestive Diseases and Sciences, Vol. 36, No. 4 (April 1991), pp. 424--430
Pathogenetic Role of Helicobacter pylori Duodenal Ulcer Disease
in
Multivariate Analysis of Factors Affecting Relapse WAI-MO HUI, JOANA HO, and SHIU-KUM LAM
The pathogenesis of duodenal ulcer disease is multifactorial and the contribution of Helicobacter pylori in relation to the other factors to the release of duodenal ulcer is unknown. To investigate this, we studied 147 patients with endoscopically proven healed ulcers. These patients were randomized to receive either placebo, misoprosto1200 tzg or misoprosto1300 t.~gfour times daily, and clinical, personal, physiological and endoscopic characteristics were obtained prospectively. Endoscopy was performed at the active phase of the ulcer and when the ulcer healed. Biopsies were taken from the antrum to assess histologically for: (1) the activity of gastritis as assessed by the degree of polymorph infiltration, (2) the degree of chronic inflammation by the degree of chronic inflammatory cells infiltration and degree of mucosal degeneration, and (3) bacteriologically for the presence of H. pylori. The severity of the gastritis and the bacterial density were graded independently by two pathologists. The patients were assessed at two-month intervals for 12 months or until the ulcer relapsed. The results demonstrated that the relapse rates of duodenal ulcer were similar in the three treatment groups. The relapse rate was higher in the group with higher density of the bacteria (P < 0.05). The degree of gastritis did not affect the relapse rate o f duodenal ulcer in either the placebo or misoprostol group or in all patients combined. Stepwise logistic regression analysis identified that increased duodenal inflammation, male sex, early-onset disease, and H. pylori adversely affected relapse of the ulcer. We conclude that multiple factors affect the relapse of duodenal ulcer and H. pylori is one of them. KEY WORDS: Helicobacter pylori; duodenal ulcer; relapse.
The pathogenetic role of H. pylori in duodenal ulcer disease is controversial (1, 2). While histological, microbiological (3-5), and immunological studies (6-8) demonstrated an association between the tWO conditions, there is a lack of direct evidence that the bacteria play a pathogenetic role. We previously have demonstrated the occurrence and frequency of the Manuscript received July 9, 1990; revised manuscript received November 2, 1990; accepted November 8, 1990. From the Departments of Medicine and Pathology, University of Hong Kong, Queen Mary Hospital, Hong Kong. Address for reprint requests: Hui Wai Mo, Department of Medicine, University of Hong Kong, Queen Mary Hospital, Pokfulam Road, Hong Kong.
424
bacteria remained unchanged with healing of duodenal ulcer (9). Recent interests have been centered on the association of the bacteria and the recurrence of duodenal ulcer. Treatment of duodenal ulcers with bismuthate salt results in clearance of the bacteria and is accompanied by improvement in gastritis as well as a lower relapse rate of duodenal ulcer (10-12). Whether this lower relapse rate is related to baCtericidal effect of bismuthate salt (13, 14) or other mechanisms (15-17) remains controversial. Moreover, the contribution, if any, of the bacteria in relation to the other factors that have been reported to affect duodenal ulcer relapse is unclear. The following study was Digestive Diseases and Sciences, Vol. 36, No. 4 (April 1991)
0163-2116/91/0400-0424506.50/09 1991PlenumPublishingCorporation
H. pylori IN ULCER DISEASE conducted to examine the density of the bacteria and other factors in relation to relapse of duodenal ulcer.
MATERIALS AND METHODS Patients. The patients-were recruited from the general medical wards and the outpatient clinic of the University Department of Medicine, Hong Kong. They were excluded if they had a concurrent medical problem, particularly renal disease, cardiovascular disease, diabetes mellitus, chronic obstructive airway disease, pyloric stenosis, previous gastric surgery, pregnancy, or any other treatment for duodenal ulcer other than antacids over the past 14 days, The patients with duodenal ulcer documented on endoscopy were randomized to receive: (1) placebo, (2) misoprostol 200 ;xg four times a day, or (3) misoprostol 300 Ixg four times a day, (G.D. Searle & Company, Chicago, Illinois) in a double-blind randomized manner to assess its efficacy in the treatment of duodenal ulcer. Forty-six prospectively selected clinical, personal, physiological, and endoscopic characteristics were assessed for each patient (Table 1). These had been reported previously (18). In particular, the endoscopic appearance of duodenal inflammation surrounding the ulcer was scored as gross hyperemia, granularity, and/or edema = 3, moderate =2, mild = 1, absent = 0. Healing of the duodenal ulcer was assessed endoscopically and at each endoscopy at least two antral and two fundal biopsy specimens were taken using Olympus biopsy forceps, cup size 5 • 5 mm, (Olympus Corporation of America, New Hyde Park, New York) until the ulcer healed or up to a maximum of 12 weeks. The present study was a follow-up of the study on healing of duodenal ulcer. The results on the healing of duodenal ulcer and gastritis have been reported in detail previously (19, 20). Written informed consent was obtained from all patients and the protocol was approved by the Ethics Committee of University of Hong Kong. Follow-up. The patients with healed duodenal ulcer and who were willing to participate in the study was assessed regularly at two-month intervals up to 12 months. During 9 this period they received no maintenance treatment and were advised to abstain from smoking, alcohol, and analgesic intake. They were followed up at two-month intervals for symptoms and endoscopy was performed at four-month intervals or whenever ulcer symptoms recurred. Recurrence of ulcer was defined as reappearance of a duodenal ulcer at endoscopy, however small, and this marked the end of the study for the individual patient. Endoscopy was performed whenever they had symptoms or at four-month intervals even when they were asymptomatic. Assessment of Gastric Mucosal Inflammation. The gastric biopsies were sectioned and stained with hematoxylin and eosin. The gastric mucosal inflammation of assessed in terms of the activity of gastritis and degree of chronic inflammation by the degree of polymorph infiltration and the degree of mononuclear cell infiltration and mucosal degeneration, respectively, and graded as nil, mild, moderate, and severe. This has been reported in detail previously (21). Both parameters were assessed independently, and improvement of gastritis was considered to Digestive Diseases and Sciences, Vol. 36, No. 4 (April 1991)
have occurred if specimens obtained at serial biopsies indicated transition to a better grade when compared with the initial specimens. Assessment of H. pylori in Antrum. The sections were stained also with Warthin-Starry silver stain and were examined for the presence of H. pylori identified as small curvilinear bacilli on the surface epithelium. The number of bacteria were graded as: 0, no characteristic bacteria; 1, occasional spiral organisms found after meticulous search; 2, scattered bacteria in most high-power fields or scattered groups of numerous bacteria; 3, numerous bacteria in most high-power fields. All clinical and endoscopic information was withheld from the pathologists, and the sections were read independently by two pathologists. Statistical Analysis. Analysis of variance, • test with Yates' correction, and Mann-Whitney test were used as appropriate (22-24). The relapse of duodenal ulcer was analyzed using life-table analysis. To evaluate the factors that may affect the relapse of duodenal ulcer, the stepwise logistic regression analysis was performed on the potential factors, including clinical, personal, endoscopic, and physiological characteristics; H. pylori; degree of activity; and chronic inflammation of the gastritis using the BMDP Statistical program (BMDP Statistical Software, Inc., 1964, Los Angeles, California) (25). To avoid overoptimistic biases in the assessment of the variable, the criteria for the selection ofF-to remove was 0.10 and F-to enter was 0.05. RESULTS Characteristics of patients. Thirty-one, 59, and 57 patients with healed ulcer in the placebo, misoprostol 200-jxg, and 300-~xg treatment group, respectively participated in the study on the recurrence of duodenal ulcer. The characteristics of the patients were comparable except those patients with ulcers healed by misoprostol had a significantly lower 2-hr postprandial gastrin level when compared to control (Table 1). On follow-up, 6 (19%), 16 (31%), and 12 (21%) patients defaulted from the placebo, misoprostol 200Ixg, and 300-1xg groups, respectively, and their data were included for analysis up to the time of default. The biopsy specimens were inadequate for the assessment of the severity of gastritis and density of bacteria in eight (5%) and five patients (3%), respectively. Therefore the number of patients assessed was less than the number entered into the study. For comparison, the bacterial density of nil and mild was graded together as mild, and moderate and severe were graded as severe. Relapses of duodenal ulcer. The relapse rates at six months were 33%, 46%, and 47%, respectively in patients initially healed with placebo, misoprostol 200 Ixg and 300 ~xg (Figure 1). There was no significant difference among the three groups. As the relapse of the duodenal ulcer was similar before
425
H I U ET A L TABLE 1. CHARACTERISTICSOF PATIENTSTREATEDWITH PLACEBO,MISOPROSTOL(MISO) 200 gtg AND 300 lxg FOUR TIMES A DAY, WITH HEALED ULCER Placebo
Miso 200
Miso 300
P
N
31
59
57
Age (years) Sex (% male) Onset age (year) Early (age
Pathogenetic Role of Helicobacter pylori Duodenal Ulcer Disease
in
Multivariate Analysis of Factors Affecting Relapse WAI-MO HUI, JOANA HO, and SHIU-KUM LAM
The pathogenesis of duodenal ulcer disease is multifactorial and the contribution of Helicobacter pylori in relation to the other factors to the release of duodenal ulcer is unknown. To investigate this, we studied 147 patients with endoscopically proven healed ulcers. These patients were randomized to receive either placebo, misoprosto1200 tzg or misoprosto1300 t.~gfour times daily, and clinical, personal, physiological and endoscopic characteristics were obtained prospectively. Endoscopy was performed at the active phase of the ulcer and when the ulcer healed. Biopsies were taken from the antrum to assess histologically for: (1) the activity of gastritis as assessed by the degree of polymorph infiltration, (2) the degree of chronic inflammation by the degree of chronic inflammatory cells infiltration and degree of mucosal degeneration, and (3) bacteriologically for the presence of H. pylori. The severity of the gastritis and the bacterial density were graded independently by two pathologists. The patients were assessed at two-month intervals for 12 months or until the ulcer relapsed. The results demonstrated that the relapse rates of duodenal ulcer were similar in the three treatment groups. The relapse rate was higher in the group with higher density of the bacteria (P < 0.05). The degree of gastritis did not affect the relapse rate o f duodenal ulcer in either the placebo or misoprostol group or in all patients combined. Stepwise logistic regression analysis identified that increased duodenal inflammation, male sex, early-onset disease, and H. pylori adversely affected relapse of the ulcer. We conclude that multiple factors affect the relapse of duodenal ulcer and H. pylori is one of them. KEY WORDS: Helicobacter pylori; duodenal ulcer; relapse.
The pathogenetic role of H. pylori in duodenal ulcer disease is controversial (1, 2). While histological, microbiological (3-5), and immunological studies (6-8) demonstrated an association between the tWO conditions, there is a lack of direct evidence that the bacteria play a pathogenetic role. We previously have demonstrated the occurrence and frequency of the Manuscript received July 9, 1990; revised manuscript received November 2, 1990; accepted November 8, 1990. From the Departments of Medicine and Pathology, University of Hong Kong, Queen Mary Hospital, Hong Kong. Address for reprint requests: Hui Wai Mo, Department of Medicine, University of Hong Kong, Queen Mary Hospital, Pokfulam Road, Hong Kong.
424
bacteria remained unchanged with healing of duodenal ulcer (9). Recent interests have been centered on the association of the bacteria and the recurrence of duodenal ulcer. Treatment of duodenal ulcers with bismuthate salt results in clearance of the bacteria and is accompanied by improvement in gastritis as well as a lower relapse rate of duodenal ulcer (10-12). Whether this lower relapse rate is related to baCtericidal effect of bismuthate salt (13, 14) or other mechanisms (15-17) remains controversial. Moreover, the contribution, if any, of the bacteria in relation to the other factors that have been reported to affect duodenal ulcer relapse is unclear. The following study was Digestive Diseases and Sciences, Vol. 36, No. 4 (April 1991)
0163-2116/91/0400-0424506.50/09 1991PlenumPublishingCorporation
H. pylori IN ULCER DISEASE conducted to examine the density of the bacteria and other factors in relation to relapse of duodenal ulcer.
MATERIALS AND METHODS Patients. The patients-were recruited from the general medical wards and the outpatient clinic of the University Department of Medicine, Hong Kong. They were excluded if they had a concurrent medical problem, particularly renal disease, cardiovascular disease, diabetes mellitus, chronic obstructive airway disease, pyloric stenosis, previous gastric surgery, pregnancy, or any other treatment for duodenal ulcer other than antacids over the past 14 days, The patients with duodenal ulcer documented on endoscopy were randomized to receive: (1) placebo, (2) misoprostol 200 ;xg four times a day, or (3) misoprostol 300 Ixg four times a day, (G.D. Searle & Company, Chicago, Illinois) in a double-blind randomized manner to assess its efficacy in the treatment of duodenal ulcer. Forty-six prospectively selected clinical, personal, physiological, and endoscopic characteristics were assessed for each patient (Table 1). These had been reported previously (18). In particular, the endoscopic appearance of duodenal inflammation surrounding the ulcer was scored as gross hyperemia, granularity, and/or edema = 3, moderate =2, mild = 1, absent = 0. Healing of the duodenal ulcer was assessed endoscopically and at each endoscopy at least two antral and two fundal biopsy specimens were taken using Olympus biopsy forceps, cup size 5 • 5 mm, (Olympus Corporation of America, New Hyde Park, New York) until the ulcer healed or up to a maximum of 12 weeks. The present study was a follow-up of the study on healing of duodenal ulcer. The results on the healing of duodenal ulcer and gastritis have been reported in detail previously (19, 20). Written informed consent was obtained from all patients and the protocol was approved by the Ethics Committee of University of Hong Kong. Follow-up. The patients with healed duodenal ulcer and who were willing to participate in the study was assessed regularly at two-month intervals up to 12 months. During 9 this period they received no maintenance treatment and were advised to abstain from smoking, alcohol, and analgesic intake. They were followed up at two-month intervals for symptoms and endoscopy was performed at four-month intervals or whenever ulcer symptoms recurred. Recurrence of ulcer was defined as reappearance of a duodenal ulcer at endoscopy, however small, and this marked the end of the study for the individual patient. Endoscopy was performed whenever they had symptoms or at four-month intervals even when they were asymptomatic. Assessment of Gastric Mucosal Inflammation. The gastric biopsies were sectioned and stained with hematoxylin and eosin. The gastric mucosal inflammation of assessed in terms of the activity of gastritis and degree of chronic inflammation by the degree of polymorph infiltration and the degree of mononuclear cell infiltration and mucosal degeneration, respectively, and graded as nil, mild, moderate, and severe. This has been reported in detail previously (21). Both parameters were assessed independently, and improvement of gastritis was considered to Digestive Diseases and Sciences, Vol. 36, No. 4 (April 1991)
have occurred if specimens obtained at serial biopsies indicated transition to a better grade when compared with the initial specimens. Assessment of H. pylori in Antrum. The sections were stained also with Warthin-Starry silver stain and were examined for the presence of H. pylori identified as small curvilinear bacilli on the surface epithelium. The number of bacteria were graded as: 0, no characteristic bacteria; 1, occasional spiral organisms found after meticulous search; 2, scattered bacteria in most high-power fields or scattered groups of numerous bacteria; 3, numerous bacteria in most high-power fields. All clinical and endoscopic information was withheld from the pathologists, and the sections were read independently by two pathologists. Statistical Analysis. Analysis of variance, • test with Yates' correction, and Mann-Whitney test were used as appropriate (22-24). The relapse of duodenal ulcer was analyzed using life-table analysis. To evaluate the factors that may affect the relapse of duodenal ulcer, the stepwise logistic regression analysis was performed on the potential factors, including clinical, personal, endoscopic, and physiological characteristics; H. pylori; degree of activity; and chronic inflammation of the gastritis using the BMDP Statistical program (BMDP Statistical Software, Inc., 1964, Los Angeles, California) (25). To avoid overoptimistic biases in the assessment of the variable, the criteria for the selection ofF-to remove was 0.10 and F-to enter was 0.05. RESULTS Characteristics of patients. Thirty-one, 59, and 57 patients with healed ulcer in the placebo, misoprostol 200-jxg, and 300-~xg treatment group, respectively participated in the study on the recurrence of duodenal ulcer. The characteristics of the patients were comparable except those patients with ulcers healed by misoprostol had a significantly lower 2-hr postprandial gastrin level when compared to control (Table 1). On follow-up, 6 (19%), 16 (31%), and 12 (21%) patients defaulted from the placebo, misoprostol 200Ixg, and 300-1xg groups, respectively, and their data were included for analysis up to the time of default. The biopsy specimens were inadequate for the assessment of the severity of gastritis and density of bacteria in eight (5%) and five patients (3%), respectively. Therefore the number of patients assessed was less than the number entered into the study. For comparison, the bacterial density of nil and mild was graded together as mild, and moderate and severe were graded as severe. Relapses of duodenal ulcer. The relapse rates at six months were 33%, 46%, and 47%, respectively in patients initially healed with placebo, misoprostol 200 Ixg and 300 ~xg (Figure 1). There was no significant difference among the three groups. As the relapse of the duodenal ulcer was similar before
425
H I U ET A L TABLE 1. CHARACTERISTICSOF PATIENTSTREATEDWITH PLACEBO,MISOPROSTOL(MISO) 200 gtg AND 300 lxg FOUR TIMES A DAY, WITH HEALED ULCER Placebo
Miso 200
Miso 300
P
N
31
59
57
Age (years) Sex (% male) Onset age (year) Early (age
