Prevalence of Helicobacter pylori infection in long-term hemodialysis patients

Transcription

1 original article & 29 International Society of Nephrology Prevalence of Helicobacter pylori infection in long-term hemodialysis patients Mitsushige Sugimoto 1, Kyoko Sakai 2, Masakazu Kita 3, Jiro Imanishi 3 and Yoshio Yamaoka 1 1 Section of Gastroenterology, Department of Medicine, Michael E. DeBakey Veterans Affairs Medical Center, Baylor College of Medicine, Houston, Texas, USA; 2 Department of Gastroenterology and Hepatology, Nara City Hospital, Nara, Japan and 3 Department of Microbiology, Kyoto Prefectural University of Medicine, Kyoto, Japan Patients on hemodialysis often have gastrointestinal complications; however, it is unclear if Helicobacter pylori infection is present in these patients. Here we determined the prevalence of H. pylori infection in 539 Japanese hemodialysis patients by measuring serum anti-h. pylori IgG antibodies. Endoscopy was performed on 299 of these patients and the results were compared to 4 patients with normal renal function who had also undergone endoscopy and sero-testing. A second cohort of 478 dialysis patients, within the original group, was checked serologically for H. pylori infection three times over a four-year observation period. The prevalence of infection in these patients was significantly lower than in those patients with normal renal function, irrespective of the clinical outcomes. The prevalence of H. pylori infection significantly decreased as the duration of dialysis increased, particularly within the first four years following initiation of dialysis. About one-third of patients on dialysis for less than four years became serologically negative for H. pylori infection within this observation period. Our study suggests that although longterm dialysis patients have low prevalence of H. pylori, they still have significant gastroduodenal diseases, such as peptic ulcers, that require endoscopic follow-up. Kidney International (29) 75, 96 13; doi:1.138/ki.28.58; published online 8 October 28 KEYWORDS: Helicobacter pylori; hemodialysis; chronic renal failure; peptic ulcer Correspondence: Yoshio Yamaoka, Section of Gastroenterology, Department of Medicine, Michael E. DeBakey Veterans Affairs Medical Center (111D), Baylor College of Medicine, Rm 3A-32, 22 Holcombe Blvd, Houston, Texas 773, USA. yyamaoka@bcm.tmc.edu Received 15 April 28; revised 11 August 28; accepted 26 August 28; published online 8 October 28 Helicobacter pylori (H. pylori) is a spiral-shaped Gramnegative flagellate bacterium. Chronic H. pylori infection has close associations with gastrointestinal diseases, such as peptic ulcer (PU), gastric hyperplastic polyps, gastric adenoma, gastric cancer, and gastric mucosa-associated lymphoid tissue lymphoma. 1 3 Recently, increased evidence suggests that the development of some extragastrointestinal disorders including idiopathic thrombocytopenic purpura, chronic idiopathic urticaria, and iron deficiency anemia are also related with H. pylori infection of gastric mucosa. 4 6 In these infected patients, the eradication of H. pylori is recommended as the first-line therapy for the prevention and cure of gastroduodenal diseases. 1,7 Although there is no significant evidence that H. pylori infection is directly associated with progression of renal dysfunction, patients receiving chronic hemodialysis and continuous ambulatory peritoneal dialysis (CAPD) often incur gastrointestinal troubles over their long treatment period Patients with chronic renal failure (CRF) present with various clinical symptoms, including nausea, dyspepsia, appetite loss, epigastric discomfort, and heartburn, as well as histological, physiological, and functional disorders of the gastrointestinal system. 18,19 These symptoms not only decrease the quality of life of patients, but also may affect their nutrition status, thus contributing to the development of malnutrition, which is a potent predictor of morbidity and mortality in patients with CRF. Moreover, 25 75% of patients with CRF undergoing dialysis suffer from a number of gastrointestinal lesions and their complications (for example, gastric erosions, PUs, angiodysplasia, and gastrointestinal bleeding). 8 17,2 H. pylori infection is crucial in many gastrointestinal conditions not only in individuals with normal renal function, but also in CRF patients receiving chronic dialysis, CAPD, and kidney transplant. Recently, although associations between patients with CRF and the prevalence of H. pylori infection have been reported, 9 14,21 59 their results are still controversial, probably due to small number of studies and short duration periods. Only four previous studies examined more than 2 patients (maximum of 322 patients 2 ), and no previous studies examined patients with mean dialysis duration periods of 41 months. In addition, 96 Kidney International (29) 75, 96 13

2 M Sugimoto et al.: H. pylori infection and dialysis o r i g i n a l a r t i c l e only few reports evaluated the relationship between H. pylori infection and dialysis treatment duration. 41,49 Therefore, we aimed to clarify the prevalence of H. pylori infection in more than 5 patients receiving dialysis with a mean duration period of more than 8 years. Moreover, we first performed a follow-up survey to assess H. pylori infection in the same patient cohort receiving dialysis for 4 years. RESULTS Patients A total of 539 patients receiving dialysis at Tojinkai Hospital, Kyoto, Japan during April 1997 were enrolled in this study. A total of 4 patients with normal renal function who underwent endoscopy at the University Hospital of Kyoto Prefectural University of Medicine, Kyoto, Japan between January 1996 and April 1997 were also enrolled in this study. Among them 121 patients visited the hospital for annual health check irrespective of gastrointestinal symptoms and remaining patients for investigation of gastrointestinal symptoms. All patients were determined H. pylori status by serological test using anti-h. pylori immunoglobulin G (IgG) antibody at the beginning of this study in Accuracy of serological test for patient receiving dialysis treatment The detection of H. pylori status by anti-h. pylori IgG antibody in patients receiving dialysis treatment might be influenced by the impaired immune system, and might be judged as false-negative cases. To check the accuracy of this serological test, we performed 13 C-urea breath test as well as two invasive tests using biopsy specimens obtained during endoscopy (H. pylori culture and rapid urease test) for randomly selected 7 patients who had either positive or negative serological test. When at least one of alternative tests yielded positive results, we judged the patients as H. pylori positive and compared with the serological test. The specificity, sensitivity, positive predictive value, negative predictive value, and accuracy of serological test were 94.1, 97.2, 97., 94.6, and 95.7%, respectively, confirming that this serological test was reliable and accurate even for patients receiving dialysis treatment. Prevalence of H. pylori infection in patients receiving dialysis at the beginning of the study Demographic and clinical characteristics of patients enrolled at the beginning of the study (1997) are summarized in Table 1. The basement diseases in chronic dialysis patients were chronic glomerulonephritis (CGN) including IgA nephropathy, membranous nephropathy, minimal change, and antineutrophil cytoplasmic antibody-related nephritis (n ¼ 388), diabetes (n ¼ 13), and others including lupus nephritis, polycystic kidney, chronic pyelonephritis, and nephrosclerosis (n ¼ 48). There were no significant differences in age, sex, and symptom scores among different basement diseases of dialysis (Table 1). Gastrointestinal symptoms also had no association with H. pylori status and dialysis duration (data not shown). The prevalence of H. pylori infection and H. pylori IgG levels in patients with diabetic nephropathy was significantly higher than in the CGN group (P ¼.1 for each) (Table 1). In addition, the treatment period in patients with diabetic nephropathy was 4.9±.4 years, which was significantly shorter than those with the CGN group (9.2±.4 years, Po.1). The prevalence of H. pylori infection in patients receiving dialysis (mean duration of 8.4±.3 years) was 48.6% (95% confidence interval (CI): %, 262 of 539), which was significantly lower than in all patients with normal renal function (78.5%, 95% CI: %, 314 of 4, Po.1) as well as patients with normal renal function who visited the hospital for annual health check (69.4%, 95% CI: %, 84 of 121, P ¼.1) (Table 1). Importantly, the prevalence of H. pylori infection in patients receiving dialysis treatment for less than 1 year (74.1%, %, 4 of 54) was similar to annual health check patients with normal renal function (P ¼.1). The mean duration of dialysis treatment in H. pylori-positive patients (6.62±.4 years) was significantly shorter than in negative patients (9.45±.4 years, Po.1). Table 1 Patient demographics Patients with normal renal function Dialysis patients Annual health check Total CGN Diabetic nephritis Others Total P-value* P-value** n Sex (M/F) 59/62 223/ /169 55/48 23/25 296/ Age 56.6± ± ±.6 6.4± ± ± H. pylori+ (%) 84 (69.4) 314 (78.5) 179 (46.1) 62 (6.2) 21 (43.8) 262 (48.6).3 o.1 H. pylori IgG (U/ml) 22.3 ( ) 26.4 ( ) 8.7 ( ) 2.2 ( ) 8.9 ( ) 15.1 ( ).19 o.1 Dialysis period (year) NA NA 9.2±.4 4.9±.4 9.2± ±.3 o.1 Symptom scores (%) NT NT 1 (1 1) 1 (1 1) 1 (1 1) 1 (1 1) ±.7 (13.5) 1.4±.1 (15.2) 1.1±.1 (5.4) 1.4±.1 (13.) CGN, chronic glomerulonephritis, NA, not available, NT, not tested. Age and dialysis period are reported as mean±sd, and H. pylori IgG levels and symptom scores are presented as median (25 75% quartiles), mean±sd and percentage of patients with symptom. Patients with normal renal function did not test the symptom scoring, but only presence/absence of symptom. *Analyzed differences among three different basement diseases of hemodialysis. **Analyzed differences among dialysis patients and patients with normal renal function. Kidney International (29) 75,

3 o r i g i n a l a r t i c l e M Sugimoto et al.: H. pylori infection and dialysis Table 2 Endoscopic findings in dialysis patients and patients with normal renal function Patients with normal renal function Dialysis patients Diseases Annual health check Total CGN Diabetic nephritis Others Total n Gastritis 112 (92.6%)* 187 (46.8%)* 154 (75.5%) 49 (84.5%) 3 (81.1%) 233 (77.8%) Gastric ulcer 5 (4.1%) 116 (29.%) 24 (11.7%) 6 (1.3%) 4 (1.8%) 34 (11.4%) Duodenal ulcer 4 (3.3%) 72 (18.%) 15 (7.4%) 1 (1.7%) 3 (8.1%) 19 (6.4%) Gastric cancer (.%) 18 (4.5%) 3 (1.5%) 2 (3.4%) (.%) 5 (1.7%) Others (.%) 7 (2.5%) 8 (3.9%) (.%) (.%) 8 (2.7%) CGN, chronic glomerulonephritis. Others included patients with mucosa-associated lymphoid tissue lymphoma, submucosal tumor, and post-gastrectomy. *Po.5 (compared the prevalence with total of dialysis patients). Table 3 Relationship between endoscopic findings and H. pylori infection Patients with normal renal function Dialysis patients Diseases Annual health check CGN Diabetic nephritis Others Total P-value Gastritis 75/112 (67.%) 67/154 (43.5%) 28/49 (57.1%) 15/3 (5.%) 11/233 (47.2%) o.1 Gastric ulcer 5/5 (1%) 18/24 (75.%) 5/6 (83.3%) 2/4 (5.%) 25/34 (73.5%).189 Duodenal ulcer 4/4 (1%) 9/15 (6.%) /1 (%) 1/3 (33.3%) 1/19 (52.6%).77 Gastric cancer / (%) 1/3 (33.3%) 1/2 (5.%) / (%) 2/5 (4.%) Others / (%) 3/8 (37.5%) / (%) / (%) 3/8 (37.5%) Total 84/121 (69.4%) 98/24 (48.%) 34/58 (58.6%) 18/37 (48.6%) 15/299 (5.2%) o.1 CGN, chronic glomerulonephritis. Others included patients with mucosa-associated lymphoid tissue lymphoma, submucosal tumor, and post-gastrectomy. The data were demonstrated as number of H. pylori-positive patients per number of total patients with gastroduodenal disease (percentage of H. pylori-positive patients). P-value analyzed the comparison with annual health check patients with normal renal function and dialysis patients. Endoscopical finding in patients at the beginning of the study Endoscopic examination was performed in 299 patients receiving dialysis treatment as well as all patients with normal renal function. There were no significant differences in age, sex, dialysis period, and H. pylori status between dialysis patients with and without receiving endoscopy (data not shown). Endoscopic findings were divided into five diseases: simple gastritis or normal appearing mucosa without ulcer and malignancy (gastritis), gastric ulcer, duodenal ulcer, gastric cancer, and others (Table 2). Endoscopic examination of dialysis patients revealed gastritis in 233 patients (77.8%), gastric ulcer in 34 (11.4%), duodenal ulcer in 19 (6.4%), and gastric cancer in 5 (1.7%) patients (Table 2). In patients with normal renal function, only 46.8% (187 of 4) had gastritis (Po.1) probably due to the fact that patients with normal renal function were selected from patients who received endoscopy in the University Hospital where the frequency of patients visiting the hospital only for annual health check without symptoms was low. In fact, patients with normal renal function for annual health check revealed gastritis in 112 patients (92.6%), gastric ulcer in 5 (4.1%), and duodenal ulcer in 4 (3.3%) patients. As dialysis patients were not selected by the presence of gastrointestinal symptoms, we compared data of dialysis patients with those of only 121 patients with normal renal function for annual health check in subsequent analyses. Importantly, the prevalence of H. pylori infection was significantly lower in dialysis patients than in patients with normal renal function, irrespective of the clinical outcomes (Table 3). For example, among patients with gastritis alone, the prevalence of H. pylori infection was still significantly lower in dialysis patients than in patients with normal renal function (67. and 47.2%, Po.1). In addition, importantly, approximately half of the dialysis patients with duodenal ulcer (52.6%) were H. pylori negative although it is the consensus that most of the duodenal ulcer patients are infected with H. pylori (Table 3). In H. pylori-positive patients, 26.7% (4 of 15) of dialysis patients and only 1.7% (9 of 87) of patients with normal renal function developed severe gastroduodenal diseases (P ¼.1). In addition, 15.2% (22 of 145) of H. pylori-negative patients still developed severe gastroduodenal diseases, whereas none of H. pylori-negative patients with normal renal function ( of 9) developed severe gastroduodenal diseases. Overall, we could conclude that lower prevalence of H. pylori infection in dialysis patients than in patients with normal renal function was not due to different patterns of clinical outcomes. The frequency of PU (duodenal ulcer and gastric ulcer) in patients receiving dialysis treatment less than 1 years was similar between H. pylori infected and uninfected patients (23.7%, 28 of 118; 2.2%, 19 of 94). However, the frequency in patients receiving dialysis treatment equal or more than 1 years was significantly higher in H. pylori-infected patients (18.9%, 7 of 37) than in uninfected patients (4.%, 2 of 5) (P ¼.2). In addition, most ulcers in uninfected patients were healing to the red scar stage, and active ulcers were rare (data not shown). When dialysis patients were divided into two groups based on the presence or absence of gastroduodenal symptoms, there were no significant differences of the prevalence of each disease between two groups irrespective of H. pylori infection (Table 4). 98 Kidney International (29) 75, 96 13

4 M Sugimoto et al.: H. pylori infection and dialysis o r i g i n a l a r t i c l e Table 4 Gastroduodenal symptoms in relation to endoscopic findings and H. pylori infection in dialysis patients H. pylori positive, %(n=36) Patients with symptoms H. pylori negative, %(n=34) Total, %(n=7) H. pylori positive, %(n=226) Patients without symptoms H. pylori negative, %(n=243) Total, %(n=469) Gastritis Gastric ulcer Duodenal ulcer Gastric cancer 6 3 Others Others included patients with mucosa-associated lymphoid tissue lymphoma, submucosal tumor, and post-gastrectomy. Follow-up studies with anti-h. pylori IgG antibodies Of 539 dialysis patients, 46 patients in 1999 and 15 patients in 21 did not have the serological test due to death, withdrawn agreement, or change of hospital. In the remaining 478 dialysis patients, the basement diseases were CGN (n ¼ 346), diabetes (n ¼ 89), and others (n ¼ 43). We only analyzed 478 dialysis patients in subsequent analyses for comparison of three time points (1997, 1999, and 21). At the beginning of the study (1997), the prevalence of H. pylori infection gradually decreased until 4 years after starting dialysis, irrelevant of basement diseases, followed by a plateau (Figure 1a). At follow-up survey of the same patients at 2 (1999) and 4 years (21) after initiation, the prevalence of H. pylori infection in relation to dialysis duration showed similar patterns observed in 1997 (Figure 1b). This pattern showed that the prevalence of H. pylori infection decreased in the first 4 years of dialysis and plateauing after 5 years of treatment. The decreased patterns of the prevalence of H. pylori infection according to treatment period were similar irrespective of the basement diseases (data not shown). The prevalence of H. pylori infection in 1997, 1999, and 21 was 51.6% (95% CI: %, 247 of 478), 42.9% ( %, 25 of 478), and 38.3% ( %, 183 of 478), respectively, indicating that the infection rate gradually decreased during dialysis treatment (Figure 2a). In other words, 26.7% ( %, 66 of 247) of dialysis patients naturally cured H. pylori infection during the observation periods of 4 years, in particular diabetes patients (36.8%, %, 21 of 57) (Figure 2a; Table 5). When dialysis patients were divided into three groups according to treatment periods at the beginning of the study in 1997 ( 4 5 9, and equal or more than 1 years), the prevalence of H. pylori infection in patients receiving equal or more than 5 years of dialysis treatment showed that the patients did not naturally cure H. pylori infection (Figure 2b). A total of 36.2% ( %, 42 of 116) of patients receiving 4 years of dialysis treatment in 1997 naturally cured H. pylori infection during 4 years of observation (Figure 2b). In agreement with the prevalence of H. pylori infection, anti-h. pylori IgG antibody levels in 1997 (2. U/ml, 25 75% quartiles: ) were significantly decreased during 2 and 4 years of observation in 1999 (12.3, , Po.1) and in 21 (11.2, , Po.1; Figure 3a). Anti-H. pylori IgG antibody levels in patients receiving 4 years of dialysis treatment in 1997 (21., ) was also significantly H. pylori infection rate (%) H. pylori infection rate (%) Year decreased in 1999 (16.3, , Po.1) and in 21 (15.4, , Po.1; Figure 3b). In contrast, anti-h. pylori IgG antibody levels in patients receiving equal or more than 5 years of dialysis treatment in 1997 were unchanged during 2 and 4 years of observation (data not shown). DISCUSSION The prevalence of H. pylori infection in patients receiving CAPD, chronic dialysis, or kidney transplant was reported to be equal or lower compared to subjects with normal renal function in various populations. 9 13,21 59 Patients receiving dialysis had significantly lower prevalence of H. pylori Treatment period 8 1 Treatment period Year Figure 1 Prevalence of H. pylori infection. (a) Within the first 4 years of treatment, the infection rate was decreased irrelevant of basement membrane diseases causing chronic renal failure. The decreased patterns of the prevalence of H. pylori infection according to treatment period were similar irrespective of the basement diseases. Prevalence of H. pylori infection in different treatment periods: 1997, 1999, and 21 in 478 patients. (b) Infection rates of H. pylori in 1999 and 21 were similar to those observed in 1997 among patients and different treatment periods. *Po.5 (vs patients with treatment period of less than 1 year). Kidney International (29) 75,

5 o r i g i n a l a r t i c l e M Sugimoto et al.: H. pylori infection and dialysis H. pylori infection rate (%) H. pylori infection rate (%) P <.1 P < infection in five studies, 22,27,28,4,49 whereas only one study reported that dialysis patients had significantly higher infection rates. 14 In the present study, we investigated the largest number of dialysis patients with a variety of treatment periods, and found that the prevalence of H. pylori infection in dialysis patients with CRF was significantly lower than in patients with normal renal function. The infection rate gradually decreased with increased dialysis treatment periods. Treatment period CGN Diabetes Others 4 years 5 1 years 11 yrars Figure 2 Changes in H. pylori infection rate in 1997, 1999, and 21 among total patients, CGN patients, and patients with diabetes and others, and among different treatment period groups. (a) The H. pylori infection rate was gradually decreased irrespective of age, sex, and basement diseases according to chronic dialysis treatment. H. pylori infection in patients with diabetic nephropathy was significantly higher than that in the CGN group. *Po.5 (vs CGN and others groups). Changes in H. pylori infection rate in 1997, 1999, and 21 among different treatment period groups. (b) The H. pylori infection rate was significantly decreased according to chronic dialysis treatment in patients with a treatment period of less than 4 years. *Po.5 (vs infection rate in 1997). Table 5 The change of the prevalence of H. pylori infection in dialysis patients in a follow-up survey Year Total CGN Diabetic nephritis Others n % n % n % n % CGN, chronic glomerulonephritis; +, H. pylori infection positive;, H. pylori infection negative. By the 4-year follow-up period, we demonstrated that H. pylori infection decreased in dialysis patients among different treatment periods, and the decrease was observed only in patients receiving dialysis for equal or less than 4 years. Surprisingly, more than one-third (36.2%) of these patients receiving equal or less than 4 years of dialysis treatment had naturally cured H. pylori infection without eradication therapies with proton pump inhibitors and antibiotics. In Japan, renal transplantation is not well established compared to Europe and the United States. Therefore, patients receiving dialysis for a prolonged period (more than 1 years) are not rare. Only a few previous reports have evaluated the relationship between H. pylori infection and dialysis treatment period, 41,49 and those reports demonstrate that the mean dialysis duration in H. pylori-positive patients was significantly shorter than in uninfected patients, as confirmed in this study. Nakajima et al. 49 reported that the prevalence of H. pylori infection markedly decreased when the treatment duration was 2 years or more. However, in all these previous studies, only one time point was observed, making it impossible to confirm whether the same patients naturally cured the infection after the dialysis. Although 37 (83.7%) of 43 previous reports had no significant differences in the prevalence of H. pylori infection among CRF patients and individuals with normal renal function, most of those previous reports enrolled less than 1 dialysis patients with an observation period of less than 1 year. 9 14,21 59 A short dialysis treatment window might not reveal important findings, such as the observed decreased incidence of H. pylori within the first 4 years. We performed a meta-analysis of previous report data 9 14,21 59 and data in this study to offset type 2 errors in determining prevalence of H. pylori infection. The prevalence of H. pylori infection in patients with renal dysfunction receiving chronic dialysis and CAPD was 43.4% (95% CI: %, 1284 of 2962) and 34.8% ( %, 113 of 325), respectively. The prevalence in overall patients with renal dysfunction was 41.8% ( %), 1742 of 4171), which was significantly lower than in the individuals with normal renal function both with and without gastrointestinal symptoms (48.9%, %, 1284 of 2622, Po.1). Although the sample size of patients receiving CAPD and kidney transplant was small, we concluded that patients with renal dysfunction have a low prevalence of H. pylori infection. There are at least three explanations as to why dialysis patients have low prevalence of H. pylori infection: (1) blood urea levels as well as urea nitrogen levels in gastric secretions are higher in dialysis patients than in patients with normal renal function, and high urea levels inhibit H. pylori growth in the stomach; 6 (2) H. pylori might be eradicated upon antibiotic treatment, both because antibiotics are commonly used during the initial treatment periods, and because antibiotic concentrations are higher in patients with renal failure; (3) Patients receiving dialysis have higher levels of proinflammatory cytokines, including interleukin-1b, -6, -8, 1 Kidney International (29) 75, 96 13

6 M Sugimoto et al.: H. pylori infection and dialysis o r i g i n a l a r t i c l e ELISA (anti-h. pylori IgG) U/ml P <.1 P <.1 U/ml P <.1 P <.1 P <.1 P <.1 1 ELISA (anti-h. pylori IgG) Figure 3 Four years follow-up of anti-h. pylori IgG antibody. All patients receiving dialysis (a) and patients with dialysis period of less than 4 years (b). In all patients receiving dialysis and patients with a treatment period of less than 4 years, the median anti-h. pylori IgG antibody level in 1997 was significantly higher than in 1999 and 21. and tumor necrosis factor, from activated inflammatory cells infiltrating the gastric mucosa. 61 As a result, gastric atrophy progresses, accompanied by increased ph, and finally H. pylori are not able to live in gastric mucosa. 62 In this study, there was no significant difference in the prevalence of H. pylori infection between patients with normal renal function and patients receiving less than 1 year of dialysis treatment. Therefore, urea concentration and antibiotic usage are unlikely contributors to decrease the prevalence of infection. One possible explanation for decreased H. pylori prevalence is severe inflammation resulting from dialysis, and progression to gastric atrophy. However, further studies to investigate the histopathology of gastric inflammation and atrophy are needed. Dialysis patients may have a higher risk of gastric mucosal damages compared with individuals with normal renal function by systemic and/or local chronic circulatory failure, 63,64 hypergastrinemia, 42 high ammonia levels, 26 and enhanced inflammation. Moreover, most dialysis patients also receive anticoagulant drugs for hypertension and brain infarction and/or heparin treatment. Therefore, patients often incurred massive hemorrhage from either PUs or gastric erosions. Importantly, we found that approximately half of the dialysis patients with duodenal ulcer were H. pylori negative although it is the consensus that most of the duodenal ulcer patients are infected with H. pylori. In addition, 15.2% (22 of 145) of H. pylori-negative patients still developed severe gastroduodenal diseases, whereas none of H. pylori-negative patients with normal renal function ( of 9) developed severe gastroduodenal diseases. Sotoudehmanesh et al. 1 also reported that H. pylori-negative dialysis patients frequently revealed any erosions in stomach and/or duodenum. These data indicate that not only H. pyloripositive dialysis patients, but also H. pylori-negative dialysis patients should receive endoscopic examination to prevent potentially destructive gastrointestinal events. In general, gastroduodenal symptoms are one of major triggers to find out PU and neoplastic diseases in healthy individuals. However, long-term dialysis patients often report decreased sensory disturbance, and in the present study the gastrointestinal symptoms of dialysis patients had no association with endoscopically monitored gastrointestinal diseases, which is in agreement with a previous report. 57 Gastrointestinal symptoms in dialysis patients might be caused not only by H. pylori infection, but also by high urea levels, decline of gastrointestinal motility, and amyloid protein deposition. In the present study, there were no significant differences of the prevalence of each disease between dialysis patients with and without gastroduodenal symptoms irrespective of H. pylori infection (Table 4). Dialysis patients may frequently develop PUs and erosion without gastrointestinal symptoms, and therefore, patients receiving dialysis irrespective to gastroduodenal symptoms should undergo endoscopic evaluation. One important question is whether H. pylori eradication therapies are necessary for H. pylori-infected dialysis patients. Thedatathatmorethanone-thirdofthepatientsreceiving equalorlessthan4yearsofdialysistreatmenthadnaturally cured H. pylori infection, and that approximately half of the dialysis patients with duodenal ulcer were H. pylori negative indicate that the eradication therapy might be meaningless. However, we showed that the frequency of PU in patients receiving dialysis treatment equal or more than 1 years was significantly higher in H. pylori-infected patients than in uninfected patients and most ulcers in uninfected patients were healing to the red scar stage, and active ulcers were rare. We therefore recommend receiving eradication therapies for H. pylori-infected dialysis patients, especially for patients receiving dialysis treatment for long time (for example, 5 years or more) where the chance of natural eradication becomes rare. We also demonstrated that the prevalence of H pylori infection in diabetic nephropathy patients was significantly higher than in the CGN group. Although the data on the prevalence of H. pylori infection in patients with type 2 diabetes remain contradictory, 65,66 a higher infection rate may be present due to an impaired immune system. 67 However, in the present study, patients with diabetes showed a decreased infection rate after the start of dialysis; therefore immunosuppression observed in diabetic patients should be considered independent of H. pylori infection. In conclusion, we considered that although the prevalence of H. pylori in Japanese patients receiving dialysis decreased Kidney International (29) 75,

7 o r i g i n a l a r t i c l e M Sugimoto et al.: H. pylori infection and dialysis in about 6% of patients within the first 4 years of dialysis, long-term dialysis patients irrespective of gastric symptoms and H. pylori infection should receive endoscopic check-ups and essential medication of acid-inhibitory drugs (for example, proton pomp inhibitors) to reduce the chance of developing severe gastroduodenal diseases. Moreover, we recommend receiving eradication therapies for H. pyloriinfected dialysis patients, especially for patients receiving dialysis for 5 years or more where the chance of natural eradication becomes rare. The clinical protocol for a detailed endoscopic check-up and eradication therapy for dialysis patients should be evaluated in future studies. MATERIALS AND METHODS Patients A total of 539 patients receiving chronic dialysis and 4 patients with normal renal function were enrolled in this study. The criteria for control patients were plasma creatinine levels less than 1.2 mg/1 ml and BUN levels less than 25 mg/1 ml. We administered a questionnaire regarding the severity of common gastrointestinal symptoms including dyspepsia, vomiting, nausea, epigastralgia, and heartburn. The dialysis patients were asked to rank the severity of the symptoms from 1 to 5 based on the frequencies of the symptoms (1, asymptomatic; 2, mild symptoms: 1 3 days a month; 3, moderate symptoms: 1 3 days a week; 4, severe symptoms: 4 6 days a week; and 5, very severe symptoms: everyday). No patients enrolled in the study received previous treatment for H. pylori infection. The research design was approved by the ethics committee and informed consent was obtained from all patients. H. pylori infection criteria and follow-up of anti-h. pylori IgG antibody by ELISA In all 539 dialysis patients and 4 patients with normal renal function enrolled in this study, H. pylori status was determined using a serum enzyme-linked immunosorbent assay (ELISA) test (Eiken Chemical, Tokyo, Japan) at the beginning of the study (1997). In addition, in 478 patients receiving chronic dialysis, follow-up ELISA test was also performed 2 and 4 years after (1999 and 21). A positive reaction was considered 42 U/ml. When were 15 2 U/ml (borderline), the ELISA test was repeated within 1 month. If the titer was still within 15 2 U/ml, H. pylori infection was evaluated by 13 C-urea breath test, culture, and/or rapid urease test. When at least one of alternative tests was positive, we judged the patients as H. pylori infection positive. Endoscopic check-up Information about gastroduodenal diseases was obtained from 299 patients receiving hemodialysis and all 4 patients with normal renal function. Endoscopy was performed for patients receiving chronic dialysis between May 1996 and April 1998 and for patients with normal renal function between January 1996 and April H. pylori-related diseases were diagnosed by endoscopy and histopathology. Endoscopic findings were divided into five diseases: simple gastritis or normal appearing mucosa without ulcer and malignancy (gastritis), gastric ulcer, duodenal ulcer, gastric cancer, and others. PUs with red scars were categorized as ulcers, but those with white scars (completely cured) were considered simple gastritis. Data analysis Statistical differences in demographic characteristics (mean age, men/women), prevalence of H. pylori, frequency of gastrointestinal diseases, and treatment period among different disease groups were determined by one-way analysis of variance or the w 2 -test. To determine whether anti-h. pylori IgG levels differed between the three time points, Wilcoxon s signed-rank test was used when significant differences were observed by Friedmann s test. Statistically significant differences in symptom scores were determined using the Mann Whitney U-test when significant differences were observed by the Kruskal Wallis test. Parametric data are presented as mean±s.d., and nonparametric data as median (25 75% quartiles). All P-values were two sided, and P-values o.5 were considered statistically significant. DISCLOSURE All the authors declared no competing interests. ACKNOWLEDGMENTS The project described was supported by grant number R1 DK62813 from National Institutes of Health (NIH). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH. We thank Dr Yoshiko Masugi, Dr Koji Okino, and all staffs in Tojinkai hospital, Kyoto, Japan for obtaining informed consent and for providing samples. We also thank Dr Takeshi Okanoue and Dr Shoji Mitsufuji (Kyoto Prefectural University of Medicine, Kyoto, Japan) for providing samples. We especially thank Dr Tadashi Kodama (Second Okamoto Hospital, Kyoto, Japan) for providing prudent advice. REFERENCES 1. Uemura N, Okamoto S, Yamamoto S et al. Helicobacter pylori infection and the development of gastric cancer. N Engl J Med 21; 345: Hopkins RJ, Girardi LS, Turney EA. Relationship between Helicobacter pylori eradication and reduced duodenal and gastric ulcer recurrence: a review. Gastroenterology 1996; 11: Wotherspoon AC, Doglioni C, de Boni M et al. Antibiotic treatment for low-grade gastric MALT lymphoma. Lancet 1994; 343: Gasbarrini A, Franceschi F, Tartaglione R et al. Regression of autoimmune thrombocytopenia after eradication of Helicobacter pylori. Lancet 1998; 352: Tebbe B, Geilen CC, Schulzke JD et al. Helicobacter pylori infection and chronic urticaria. J Am Acad Dermatol 1996; 34: Annibale B, Marignani M, Monarca B et al. Reversal of iron deficiency anemia after Helicobacter pylori eradication in patients with asymptomatic gastritis. Ann Intern Med 1999; 131: Take S, Mizuno M, Ishiki K et al. The effect of eradicating helicobacter pylori on the development of gastric cancer in patients with peptic ulcer disease. Am J Gastroenterol 25; 1: Ala-Kaila K. Upper gastrointestinal findings in chronic renal failure. Scand J Gastroenterol 1987; 22: Sezer S, Ibis A, Ozdemir BH et al. Association of Helicobacter pylori infection with nutritional status in hemodialysis patients. Transplant Proc 24; 36: Sotoudehmanesh R, Ali Asgari A, Ansari R et al. Endoscopic findings in end-stage renal disease. Endoscopy 23; 35: Prakash J, Agrawal BK. Upper gastrointestinal mucosal lesions in chronic renal failure. Indian J Gastroenterol 1991; 1: Gheissari A, Rajyaguru V, Kumashiro R et al. Gastrointestinal hemorrhage in end stage renal disease patients. Int Surg 199; 75: Tsai CJ, Hwang JC. Investigation of upper gastrointestinal hemorrhage in chronic renal failure. J Clin Gastroenterol 1996; 22: Khedmat H, Ahmadzad-Asl M, Amini M et al. Gastro-duodenal lesions and Helicobacter pylori infection in uremic patients and renal transplant recipients. Transplant Proc 27; 39: Milito G, Taccone-Gallucci M, Brancaleone C et al. Assessment of the upper gastrointestinal tract in hemodialysis patients awaiting renal transplantation. Am J Gastroenterol 1983; 78: Milito G, Taccone-Gallucci M, Brancaleone C et al. The gastrointestinal tract in uremic patients on long-term hemodialysis. Kidney Int 1985; 17: S157 S Musola R, Franzin G, Mora R et al. Prevalence of gastroduodenal lesions in uremic patients undergoing dialysis and after renal transplantation. Gastrointest Endosc 1984; 3: Kidney International (29) 75, 96 13

8 M Sugimoto et al.: H. pylori infection and dialysis o r i g i n a l a r t i c l e 18. Var C, Gultekin F, Candan F et al. The effects of hemodialysis on duodenal and gastric mucosal changes in uremic patients. Clin Nephrol 1996; 45: Kang JY, Ho KY, Yeoh KG et al. Peptic ulcer and gastritis in uraemia, with particular reference to the effect of Helicobacter pylori infection. J Gastroenterol Hepatol 1999; 14: Wee A, Kang JY, Ho MS et al. Gastroduodenal mucosa in uraemia: endoscopic and histological correlation and prevalence of helicobacterlike organisms. Gut 199; 31: Offerhaus GJ, Kreuning J, Valentijn RM et al. Campylobacter pylori: prevalence and significance in patients with chronic renal failure. Clin Nephrol 1989; 32: Shousha S, Arnaout AH, Abbas SH et al. Antral Helicobacter pylori in patients with chronic renal failure. J Clin Pathol 199; 43: Ala-Kaila K, Vaajalahti P, Karvonen AL et al. Gastric Helicobacter and upper gastrointestinal symptoms in chronic renal failure. Ann Med 1991; 23: Davenport A, Shallcross TM, Crabtree JE et al. Prevalence of Helicobacter pylori in patients with end-stage renal failure and renal transplant recipients. Nephron 1991; 59: Loffeld RJ, Peltenburg HG, vd Oever H et al. Prevalence of Helicobacter pylori antibodies in patients on chronic intermittent haemodialysis. Nephron 1991; 59: Neithercut WD, Rowe PA, el Nujumi AM et al. Effect of Helicobacter pylori infection on intragastric urea and ammonium concentrations in patients with chronic renal failure. J Clin Pathol 1993; 46: Jaspersen D, Fassbinder W, Heinkele P et al. Significantly lower prevalence of Helicobacter pylori in uremic patients than in patients with normal renal function. J Gastroenterol 1995; 3: De Vecchi AF, Quatrini M, Boni F et al. Epidemiology of Helicobacter pylori in dialysis patients. Perit Dial Int 1995; 15: Luzza F, Imeneo M, Maletta M et al. Helicobacter pylori-specific IgG in chronic haemodialysis patients: relationship of hypergastrinaemia to positive serology. Nephrol Dial Transplant 1996; 11: Seyrek N, Kocabas E, Hazar S et al. Helicobacter pylori antibodies in patients on chronic hemodialysis. Nephron 1996; 72: Krawczyk W, Gorna E, Suwala J et al. Frequency of Helicobacter pylori infection in uremic hemodialyzed patients with antral gastritis. Nephron 1996; 74: Abu Farsakh NA, Roweily E, Rababaa M et al. Brief report: evaluation of the upper gastrointestinal tract in uraemic patients undergoing haemodialysis. Nephrol Dial Transplant 1996; 11: Hruby Z, Myszka-Bijak K, Gosciniak G et al. Helicobacter pylori in kidney allograft recipients: high prevalence of colonization and low incidence of active inflammatory lesions. Nephron 1997; 75: Antoniou S, Dimitriadis A, Kliridou M et al. Prevalence of Helicobacter pylori antibodies in CAPD patients. Nephron 1997; 75: Moustafa FE, Khalil A, Abdel Wahab M et al. Helicobacter pylori and uremic gastritis: a histopathologic study and a correlation with endoscopic and bacteriologic findings. Am J Nephrol 1997; 17: Ozgur O, Boyacioglu S, Ozdogan M et al. Helicobacter pylori infection in haemodialysis patients and renal transplant recipients. Nephrol Dial Transplant 1997; 12: Munoz de Bustillo E, Sanchez Tomero JA, Sanz JC et al. Eradication and follow-up of Helicobacter pylori infection in hemodialysis patients. Nephron 1998; 79: Fabrizi F, Martin P, Dixit V et al. Epidemiology of Helicobacter pylori in chronic haemodialysis patients using the new RIBA H. pylori SIA. Nephrol Dial Transplant 1999; 14: Araki H, Miyazaki R, Matsuda T et al. Significance of serum pepsinogens and their relationship to Helicobacter pylori infection and histological gastritis in dialysis patients. Nephrol Dial Transplant 1999; 14: Misra V, Misra SP, Dwivedi M et al. Decreased sensitivity of the ultrarapid urease test for diagnosing Helicobacter pylori in patients with chronic renal failure. Pathology 1999; 31: Yildiz A, Besisik F, Akkaya V et al. Helicobacter pylori antibodies in hemodialysis patients and renal transplant recipients. Clin Transplant 1999; 13: Gur G, Boyacioglu S, Gul C et al. Impact of Helicobacter pylori infection on serum gastrin in haemodialysis patients. Nephrol Dial Transplant 1999; 14: Huang JJ, Huang CJ, Ruaan MK et al. Diagnostic efficacy of (13)C-urea breath test for Helicobacter pylori infection in hemodialysis patients. Am J Kidney Dis 2; 36: Emir S, Bereket G, Boyacioglu S et al. Gastroduodenal lesions and Helicobacter pylori in children with end-stage renal disease. Pediatr Nephrol 2; 14: Aguilera A, Codoceo R, Bajo MA et al. Helicobacter pylori infection: a new cause of anorexia in peritoneal dialysis patients. Perit Dial Int 21; 21(Suppl 3): S152 S Schoonjans R, Van VB, Vandamme W et al. Dyspepsia and gastroparesis in chronic renal failure: the role of Helicobacter pylori. Clin Nephrol 22; 57: Fabbian F, Catalano C, Bordin V et al. Esophagogastroduodenoscopy in chronic hemodialysis patients: 2-year clinical experience in a renal unit. Clin Nephrol 22; 58: Tsukada K, Miyazaki T, Katoh H et al. Seven-day triple therapy with omeprazole, amoxycillin and clarithromycin for Helicobacter pylori infection in haemodialysis patients. Scand J Gastroenterol 22; 37: Nakajima F, Sakaguchi M, Amemoto K et al. Helicobacter pylori in patients receiving long-term dialysis. Am J Nephrol 22; 22: Marsenic O, Peco-Antic A, Perisic V et al. Upper gastrointestinal lesions in children on chronic haemodialysis. Nephrol Dial Transplant 23; 18: Strid H, Simren M, Stotzer PO et al. Delay in gastric emptying in patients with chronic renal failure. Scand J Gastroenterol 24; 39: Lopez T, Quesada M, Almirall J et al. Usefulness of non-invasive tests for diagnosing Helicobacter pylori infection in patients undergoing dialysis for chronic renal failure. Helicobacter 24; 9: Nakajima F, Sakaguchi M, Oka H et al. Prevalence of Helicobacter pylori antibodies in long-term dialysis patients. Nephrology (Carlton) 24; 9: Al-Mueilo SH. Gastroduodenal lesions and Helicobacter pylori infection in hemodialysis patients. Saudi Med J 24; 25: Trimarchi H, Forrester M, Schropp J et al. Low initial vitamin B12 levels in Helicobacter pylori positive patients on chronic hemodialysis. Nephron Clin Pract 24; 96: c28 c Blusiewicz K, Rydzewska G, Rydzewski A. Gastric juice ammonia and urea concentrations and their relation to gastric mucosa injury in patients maintained on chronic hemodialysis. Rocz Akad Med Bialymst 25; 5: Lui SL, Wong WM, Ng SY et al. Seroprevalence of Helicobacter pylori in Chinese patients on continuous ambulatory peritoneal dialysis. Nephrology (Carlton) 25; 1: Lentine KL, Parsonnet J, Taylor I et al. Associations of serologic markers of infection and inflammation with vascular disease events and mortality in American dialysis patients. Clin Exp Nephrol 26; 1: Altay M, Turgut F, Akay H et al. Dyspepsia in Turkish patients on continuous ambulatory peritoneal dialysis. Int Urol Nephrol 28; 4: Gladziwa U, Haase G, Handt S et al. Prevalence of Helicobacter pylori in patients with chronic renal failure. Nephrol Dial Transplant 1993; 8: Hwang IR, Kodama T, Kikuchi S et al. Effect of interleukin 1 polymorphisms on gastric mucosal interleukin 1beta production in Helicobacter pylori infection. Gastroenterology 22; 123: Wesdorp RI, Falcao HA, Banks PB et al. Gastrin and gastric acid secretion in renal failure. Am J Surg 1981; 141: Block GA, Raggi P, Bellasi A et al. Mortality effect of coronary calcification and phosphate binder choice in incident hemodialysis patients. Kidney Int 27; 71: Nakamura S, Sasaki O, Nakahama H et al. Clinical characteristics and survival in end-stage renal disease patients with arteriosclerosis obliterans. Am J Nephrol 22; 22: Anastasios R, Goritsas C, Papamihail C et al. Helicobacter pylori infection in diabetic patients: prevalence and endoscopic findings. Eur J Intern Med 22; 13: Bener A, Micallef R, Afifi M et al. Association between type 2 diabetes mellitus and Helicobacter pylori infection. Turk J Gastroenterol 27; 18: Bytzer P, Talley NJ, Leemon M et al. Prevalence of gastrointestinal symptoms associated with diabetes mellitus: a population-based survey of 15, adults. Arch Intern Med 21; 161: Kidney International (29) 75,