The Diagnosis of Helicobacter pylori Infection in Uremic Patients Peter A. Rowe, MD, MRCP, Adil M. EI Nujumi, MB, MRCP, Craig Williams, MRCP, MRCPath, Stephen Dahill, MB, ChB, J. Douglas Briggs, FRCP, and Kenneth E. L. McColl, MD, FRCP • Two urease-based tests-the urease slide test and the radiolabeled urea breath test, are commonly used for the diagnosis of Helicobacter pylori infection of the stomach. The reliability of these tests in chronic uremia was compared with serological tests for H pylori antibodies, and with direct detection of the organism by microscopy or culture of gastric antral biopsies. Twenty-seven patients with chronic renal failure and dyspepsia underwent upper gastrointestinal endoscopy. Twelve of these patients (46%) were judged to be infected with H pylori on the basis of identification of the organism on microscopy or culture of antral biopsy. Both urease-based tests were able to determine H pylori status, despite the markedly increased concentrations of urea in the gastric juice found in chronic renal failure. The urease slide test performed on antral biopsies obtained at endoscopy proved reliable in determining H pylori status with no false-positive nor false-negative results after 20 minutes and 24 hours of incubation. The 14C-urea breath test also differentiated the infected from the uninfected patients. The 20-minute 14C02 excretion (kg %dose/mmol CO2 x 100) ranged from 50 to 834 in the H pylori-infected patients, compared with 0.3 to 27 in the H pylori-noninfected patients (P < 0.0001); the 90-minute values ranged from 88 to 398 in the former, compared with 1 to 79 in the laUer (P < 0.0001). The excretion of 14C02 (derived from bacterial hydrolysis of ingested 14C-urea) was higher in all the uremic patients compared with nonuremic controls, and in half of the H pylori-noninfected uremic patients there was a late increase in 14C02 excretion. On serological testing, nine of 11 patients in the H pylori-infected group were found to have an antibody titer of 20 U/mL or greater, but none of the H pylori-noninfected group, giving a sensitivity of 82% and a specificity of 100%. Tests based on detection of the urease activity of H pylori remain a reliable means of determining H pylori status in chronic renal failure, although a higher cutoff value to indicate infection in the 14C-urea breath test is needed. We recommend a value of 50 kg %dose/mmol CO2 x 100 at 20 minutes as the upper limit of normal, rather than the value of 20 kg %dose/mmol CO2 X 100 that is normally used in subjects with normal renal function. © 1992 by the National Kidney Foundation, Inc. INDEX WORDS: Renal failure; Helicobacter pylori; diagnosis; 14C-urea breath test; urease slide test; CLO test.
I
NFECTION OF THE gastric mucosa with Helieobaeter pylori (formerly Campylobaeter pylori) is an important cause of gastritis l and predisposes to duodenal ulcer disease. 2 The bacterium is remarkable because of its high urease activity that splits urea in the gastric juice into ammonia and carbon dioxide. 3 The ammonia neutralizes gastric acid in the immediate vicinity of the organism and may assist its survival in the inhospitable environment of the stomach. Infection with H pylori causes a decrease in urea concentration and an increase in ammonium concentration in the gastric juice. 4 The rate of ammonia production by H pylori is limited by
From the Renal Unit and Departments of Medicine and Therapeutics, Microbiology, and Histopathology, Western Infirmary, Glasgow, Scotland. Received April 8, 1992; accepted in revised form July 17, 1992. Supported by a grant from the Biomedical Research Committee ofthe Scottish Home and Health Department, andfrom the Research Support Group of the Greater Glasgow Health Board. Address reprint requests to Peter A. Rowe, MD, MRCP, Renal Unit, Western Irifirmary, Glasgow, GIl 6NT, Scotland. © 1992 by the National Kidney Foundation, Inc. 0272-6386/92/2006-0007$3.00/0 574
the availability ofurea,5 and urea concentrations in the gastric juice are increased in patients with chronic renal failure. 4 Diagnosis of H pylori infection can be made by demonstrating the gram-negative bacteria in biopsy specimens obtained at endoscopy, or by culture of gastric antral biopsies. Microscopy requires tissue-processing, and the bacterium is microaerophilic and time-consuming to culture, which delays diagnosis. Two other commonly used more rapid tests rely on the bacterium's high urease activity. These are the urease slide test,6 which relies on demonstrating urease activity in an antral biopsy obtained at endoscopy, and the radio labeled urea breath test, 7 which depends on increased respiration of radiolabeled CO 2 derived from splitting of ingested radiolabeled urea. These tests have been shown to be reliable in patients with normal renal function,6-8 but could be affected by the high endogenous urea concentration in the gastric juice of patients with chronic renal failure. High concentrations of urea in gastric juice could saturate bacterial urease, and delay hydrolysis of the small quantity of administered radiolabeled urea, or alternatively, lead to high ammonia concentrations that could inhibit bacterial urease activity.
American Journal of Kidney Diseases, Vol XX, No 6 (December), 1992: pp 574-579
HELICOBACTER PYLORI IN CHRONIC RENAL FAILURE
We have conducted a prospective study of unselected dyspeptic patients with chronic renal failure, to evaluate the diagnostic utility of the radiolabeled urea breath test and urease slide test.
METHODS Patients with chronic renal failure were recruited from those with end-stage renal failure on renal replacement therapy or from those attending the predialysis clinic. The criterion for inclusion was dyspeptic symptoms of sufficient severity to warrant diagnostic endoscopy. Patients already receiving treatment with any acid-inhibiting agent or with antibiotics discontinued this therapy at least I week before the tests were performed. Each subject underwent a radiolabeled urea breath test within I week of a standard diagnostic upper gastrointestinal endoscopy. On the day of endoscopy, a serum sample was obtained for measurement of H pylori antibodies by an enzyme-linked immunoadsorbent assay for IgG (Helico-G kit, Porton, Cambridge, UK).9 All endoscopies were performed by one investigator (K.E.L.M.), in the morning following an overnight fast. On entering the stomach, 10 mL of resting gastric juice was aspirated by means of a trap in the suction line and sent for estimation of pH, ammonium, and urea concentrations as previously described. to After completion of the examination, three biopsies were taken from the gastric antrum on the greater curvature 2 cm from the pylorus. One was used for the urease slide test, and the other two were submitted for bacteriological culture and microscopic examination. The urease slide test was performed using a commercially available kit (CLO test, Delta West, Bentley, Western Australia). A gastric antral biopsy is placed in the preprepared well on the slide (containing urea), and if H pylori is present in the biopsy, the ammonia generated produces a change in the colour of a pH indicator. The test was interpreted by the investigator performing the endoscopy who was at that time unaware of the results of other investigations. The slide was inspected for any color change after 20 minutes and again after 24 hours of incubation at 30°C. The radiolabeled urea breath test relies on the bacterial hydrolysis of orally administered 14C-urea, with subsequent respiratory excretion of I'C02 . 7 After an overnight fast, patients were asked to clean their teeth using a clean toothbrush and tap water, without swallowing. They then drank a test meal consisting of 250 mL of Ensure Plus (Abbott, Maidenhead, UK), followed by 0.4 MBq of 14C-urea dissolved in 25 mL of water. After this, they cleaned their teeth again and rinsed out their mouths. This cleansing procedure is routinely performed to minimize any hydrolysis of the administered radiolabeled urea by the normal mouth flora (which include urease-producing bacteria)." Breath samples were collected at lO-minute intervals for 2 hours, and the excretion of I'C02 per mmol respired CO2 was calculated. Counts were corrected for body weight and expressed as a percentage of the activity of the original dose. Microscopy of the gastric biopsies was performed by one investigator (S.D.), who was unaware of the clinical details or the results of the other tests. Tissue samples were fixed in
575 formalin and a paraffin section stained with Cresyl Fast Violet for detection of H pylori. Gastric biopsies were transferred to the microbiology laboratory in sterile saline and immediately plated out onto horse blood agar plates containing Skirrow's selective supplement (Mast, UK). The plates were examined after 3 and 5 days of incubation at 37°C in a microaerophilic atmosphere. Isolates were identified by their Gram stain appearance and by their urease, oxidase, and catalase reactions. Patients were judged to be infected with H pylori if the organism was demonstrated in the antral biopsy by either microscopy or culture. Control values were obtained from subjects with normal renal function who had previously been investigated for dyspepsia. Statistical methods were analysis of variance and unpaired t test where data were normally distributed, and by the corresponding nonparametric tests (Kruskal-Wallis and MannWhitney U tests) where distribution was not normal. The breath test curves were analyzed by a repeated measures analysis of variance (BMDP P2V).
RESULTS Twenty-seven patients with chronic renal failure (15 males, 12 females) were studied. Twelve (five males, seven females) were found to be infected with H pylori, a prevalence of 46%. Microscopic examination demonstrated the presence of H pylori in 11 of the 12, while in the final case, microscopy was negative but culture was positive. Culture was positive in six of the 12 patients thus judged to be infected with H pylori and negative in the other 15 patients in whom the organism was not seen on microscopy. Data are incomplete for four noninfected patients; in two patients, no gastric aspirate was obtained and no breath test was performed; in one patient, no gastric aspirate was obtained; and in one patient, no breath test was performed. There was no significant difference in the distribution of the sexes, nor in the type and duration of renal failure treatment between the H pylori-infected and -noninfected groups (Table 1). Nor was there any difference in the current or past use of acid-inhibitory agents between the two groups. The H pylori-infected patients tended to be older, as previously reported,12.13 but this did not reach statistical significance. The only significant difference between the two groups was in the presence or absence of their own teeth. In the infected group, eight of 12 had no teeth, compared with only four of the 15 in the noninfected group (x 2 = 4.32; P < 0.05). The serum urea concentrations were increased to a similar extent in the infected and noninfected
ROWE ET AL
576 Table 1. Patient Characteristics H pylori-
H pylori-Infected Sex Male Female Renal failure treatment No. on CAPD No. on hemodialysis No. predialysis Median age in years (range) Median duration of dialysis in months (range) Median serum urea mmol/L (range) mg/dL (range) Median serum creatinine I'moi/L (range) mg/dL (range)
Noninfected
5 7
10 5
7 3 2
11 3
57.1 (34-73)
48.5 (27-67)
(0-100)
20.5 (0-142)
15
21.3 (12.0-31.2) 59.7 (33.6-87.4)
23.8 (10.7-32.2) 66.7 (30.0-90.2)
891 (287-1,664) 10.1 (3.2-18.8)
981 (612-1,700) 11.1 (6.9-19.2)
creased quickly in the first 10 minutes and then plateaued, before decreasing slowly in the remaining 60 minutes of the 2-hour test. By contrast, in the noninfected uremic patients, 14C02 excretion increased slowly after 30 minutes in six of the 11, but remained at a steady low level in the other patients. These time trends are significantly different both within each group (F = 2.27, P = 0.012) and between groups (F = 3.9, P < 0.0001). There was no overlap between the infected and noninfected uremic groups before 120 minutes. The infected uremic patients' curves slope down toward the end of the 2-hour test, and the noninfected uremic patients' slope up, so the best separation occurs in the early phase between 20 and 60 minutes. At 20 minutes, the median value
Abbreviation: CAPD, continuous ambulatory peritoneal dialysis.
patients (Table 1). The gastric aspirate urea concentrations in the H pylori-infected patients (median, 1.9; range, 0.2 to 8.7 mmol/L) were much lower (P < 0.0002) than in the noninfected patients (median, 13.8; range, 5.4 to 20.8 mmol/ L). The urea concentration in the gastric aspirate correlated with the serum urea in the noninfected patients (r = 0.53, P = 0.05), but not in the infected patients. As expected, the concentrations of ammonia in the gastric aspirate were much higher (median, 17.5; range, 2.7 to 43.1 mmol/ L) in the infected patients (P = 0.0015) than in the noninfected patients (median, 4.0; range, 0.05 to 12.3 mmol/L). There was no significant difference in gastric pH between the two groups, although the range varied widely (infected group: median, 2.9; range, 1.1 to 7.0; noninfected group: median, 2.2; range, 1.2 to 7.7). In every patient judged infected with H pylori, the urease slide test was clearly positive after both 20 minutes and 24 hours of incubation. In the noninfected group, the test was negative in every patient at both 20 minutes and after 24 hours of incubation. The results of the 14C_urea breath test are shown in Fig 1. In the uremic patients, the values obtained from the H pylori-infected group were much higher than the values obtained from the noninfected group (F = 28.1, P < 0.0001). In the infected uremic patients, excretion of 14C02 in-
500 400 300 0 0
......
200
-- -
>< 100 0'"
() 0
E E
0 50
....... (I)
en
0
40
-c
~
01
::.c::
30
----
20 10
- --
""]
0 0
30
60
90
120
Time (minutes) Fig 1. 14C02-urea breath test curves from patients with chronic renal failure (CRF; -----) and controls with normal renal function (--). e, H pylori-infected patients; 0, noninfected patients. Data are median values with interquartile ranges. Note the 10-fold difference in the scale used on the y-axiS for the infected patients (upper panel) compared with the noninfected patients (lower panel). The curves in the H pylori-infected patients are significantly different from those in the noninfected (CRF: F = 28.1, P < 0.0001; controls: F = 149.1, P < 0.0001).
HELICOBACTER PYLORI IN CHRONIC RENAL FAILURE
(range) in the infected uremic group was 207 (50 to 834) compared with 12 (0.3 to 27) kg %dose/ mmol CO2 X 100 (P < 0.0001) in the noninfected uremic group (Fig 2), and at 90 minutes the corresponding values were 177 (88 to 398) and 28.5 (1 to 79), respectively (P < 0.0001). The uremic patients infected with H pylori had somewhat higher levels of 14C02 excretion than those with normal renal function, but this was only just statistically significant (F = 4.29, P = 0.05). In contrast, the curves from the uremic noninfected patients were strikingly different (F = 8.76, P = 0.008) from those obtained from the noninfected patients with normal renal function, where no patient showed a late increase in 14C02 excretion. In three of six patients with a late increase in 14C02 excretion, organisms other than H pylori were noticed in the antral biopsy, but none were observed in the five without such a late increase. The results of serological tests were available in 22 patients. A titer of greater than or equal to 20 U/mL was interpreted as positive. Titers of H pylori IgG were significantly (P = 0.002) higher in the infected group (median, 46 U/mL; range, 6.2 to 197) than in the noninfected group (median, 9.5 U/mL; range 5 to 19.30). Titers greater than or equal to 20 U /mL were present in nine of the 11 patients with H pylori infection, but in none of the noninfected patients, giving a sensitivity of 82% and a specificity of 100%. Endoscopy was normal in five of the 12 H pylori-infected patients. Of the other seven, hiatus hernia was diagnosed in three, and one each had petechiae in the stomach and duodenum, duodenitis, gastroesophageal inflammation and petechial hemorrhage, and moderate chronic gastritis. In seven of the 12 H pylori-noninfected patients, no abnormality was found. Of the remaining five patients, one each had esophagitis and gastric stasis (a diabetic), hiatus hernia and esophagitis, scattered gastric petechiae, mild superficial gastritis, and mild duodenitis. DISCUSSION
Before the value of a diagnostic test can be judged, it is essential to make an accurate diagnosis. The method used in this study to establish the diagnosis of H pylori infection in the majority of patients was a careful microscopic examination of an antral biopsy. 8 Bacteriological culture was
577 1000
•
800
"' !::J 600
• •• ••
C
0 0
E
0400 N
> ~
"' !::J 600 c
'E
0400 0>
200 0
•
:
~ Chronic renal failure
Normal renal function
Fig 2. Individual values of 14C02 excretion corrected for respired CO2 and body weight at 20 minutes (upper panel) and 90 minutes (lower panel), in the H pyloriinfected (e) and -noninfected (0) CRF patients and in controls with normal renal function. Median value (interquartile range) in the infected CRF patients at 20 minutes was 207 (91.7 to 436.8), compared with 12 (5 to 17.25) in the noninfected CRF patients. There is no overlap between the infected and noninfected groups at 20 minutes.
positive in 50% of the patients judged to be infected, a sensitivity similar to that reported previously.8 Our data show that despite the higher gastric juice urea levels in patients with chronic uremia, diagnostic tests based on H pylori's high urease activity remain reliable. The 14C-urea breath test accurately discriminated between H pylori-infected and -noninfected uremic patients, with a 20-minute value greater than 50 indicating infection. This cutoff level in uremic patients is higher than the value of 20 usually used for subjects with normal renal function, since the excretion of radiolabeled 14C02 was greater in all subjects with chronic renal failure. The higher levels of 14C02 excretion observed could be due to the induction of urease activity in the bacterium by the prevailing higher
578
levels of urea in the gastric aspirate 4 or, alternatively, the presence of other urease-producing organisms. The late increase in 14C02 excretion seen in the renal failure patients without H pylori infection suggests that the ingested radiolabeled urea is being hydrolyzed by other bacteria. This could occur with colonization of the upper gastrointestinal tract by other urease-producing bacteria such as Proteus species. Factors that predispose to bacterial colonization include hypochlorhydria, which is common in chronic renal failure, 14 and impaired gastrointestinal motility, which might be a sequel to chronic constipation, the use of the peritoneal cavity for dialysis, or a minor degree of uremic autonomic neuropathy. Colonization of the upper gastrointestinal tract could also help explain the relative difficulty in culturing H pylori. as this is more difficult when other bacteria are present. The finding that H pylori infection was associated with an edentulous state is of interest, since the teeth are colonized by other urease-producing organisms. II Ingestion of normal mouth flora could inhibit the ability of H pylori to colonize the stomach, and since older patients are less likely to possess their own teeth, this could also help explain the increasing prevalence of H pylori infection with increasing age. 12 ,13 We were not able to identify any particular characteristics of the H pylori-noninfected patients who showed a late increase in 14C02 excretion. There were no differences in the type or duration of renal failure treatment, in the number of episodes of chronic ambulatory peritoneal dialysis peritonitis, or in the use of acid-inhibitory agents when compared with those H pylori-noninfected patients without a late increase, and there was no difference in fasting gastric pH. The numbers were too small to rule out the possibility that urease organisms from the teeth of the H pylorinoninfected patients had been able to colonize the upper gastrointestinal tract, although it is of interest that microorganisms other than H pylori were found in the antral biopsy specimens of three of the six with a late increase, and in none of the remainder. The urease slide test also proved reliable in determining H pylori status. Despite the late increase in the urease activity shown in the 14C_ urea breath test in the H pylori-noninfected
ROWE ET AL
patients, there was no suggestion of any late falsepositive results in the urease slide test. H pylori lives under the mucous layer and closely adheres to the epithelial surface in the gastric antrum. Consequently, only the urease activity of organisms present on the antral biopsy will be detected by the urease slide test, and the result is thus not affected by other urease-producing bacteria in the lumen of the stomach or small bowel. Previous studies of H pylori infection in renal failure have relied on microscopic examination or serological testing. 15-19 The presence of H pylori infection in 46% of our patients was similar to the prevalence observed (52%) in healthy subjects without dyspepsia, 13 and slightly less than the reported prevalence (62%) in dyspeptic subjects without renal failure. 2o This figure closely agrees with the prevalence of 44% reported by Offerhaus et al,16 but is higher than the 34% described by Davenport et al. I8 Both studies relied on serological tests for H pylori antibodies-a method of diagnosis shown by our data to be less sensitive than urease-based tests. This study demonstrates that urease-based tests can accurately diagnose H pylori infection in uremic patients, with suitable adjustment of the normal range for the 14C02-urea breath test. Our results suggest that H pylori infection is no more frequent in chronic renal failure than in healthy volunteers. It is therefore unlikely that it is any more important a factor in causing gastroduodenitis in chronic renal failure than in the healthy population. Reliable diagnosis of H pylori infection by means of urease-based tests will permit further studies of the mechanisms underlying gastritis and dyspepsia in chronic renal failure. ACKNOWLEDGMENT The authors would like to thank Jennifer Harwood for measuring the gastric juice ammonia, and the Department of Nuclear Medicine at the Gardiner Institute for performing the 14C-urea breath tests.
REFERENCES I. Rauws EAJ, Langenberg W, Houthoff HJ, et al: Cam-
pylobacter pyloridis-associated chronic active gastritis. Gastroenterology 94:33-40, 1988 2. Axon ATR: Helicobacter pylori in the pathogenesis of peptic ulcer-Evidence in favour, in Malfertheiner P, Ditschuneit H (eds): Helicobacter pylori, Gastritis and Peptic Ulcer. Berlin, Germany, Springer-Verlag, 1990, pp 345-357 3. Bode G, Malfertheiner P, Lennhardt G, et al: Virulence factors of Helicobacter pylori-Ultrastructural features, in
HELICOBACTER PYLORI IN CHRONIC RENAL FAILURE Malfertheiner P, Ditschuneit H (eds): Helieobaeter pylori, Gastritis and Peptic Ulcer. Beriin, Germany, Springer-Veriag, 1990, pp 63-73 4. Kim H, Park C, Jang WI, et al: The gastric juice urea and ammonia levels in patients with Campylobaeter pylori. Am J Clin PathoI94:187-191, 1990 5. Chittajallu RS, Neithercut WD, Macdonald AMI, et aI: Effect of increasing H elieobaeter pylori ammonia production by urea infusion on plasma gastrin concentrations. Gut 32: 21-24, 1991 6. Marshall BJ, Warren JR, Francis GJ, et aI: Rapid urease test in the management of Campylobaeter pyloridis-associated gastritis. Am J Gastroenterol 82:200-210, 1987 7. Dill S, Payne-James JJ, Misiewicz JJ, et aI: Evaluation of 13C-urea breath test in the detection of H elieobaeter pylori and in monitoring the effect of tripotassium dicitratobismuthate in non-ulcer dyspepsia. Gut 31:1237-1241, 1990 8. Logan RPH, Polson RJ, Misiewicz JJ, et aI: Simplified single sample 13carbon urea breath test for Helieobaeter pylori: Comparison with histology, culture, and ELISA serology. Gut 32:1461-1464, 1991 9. Mendall MA, Goggin P, Bellhouse A, et aI: Performance of a commercially available serological test for Helieobaeter pylori and its relation to age. Gut 32:AI257, 1991 (abstr) 10. Neithercut WD, Milne A, Chittajallu RS, et al: The detection of Helieobaeter pylori infection of the gastric mucosa by measurement of gastric aspirate ammonium and urea concentrations. Gut 32:973-976, 1991 II. Bowden GHW, Ellwood DS, Hamilton IR: Microbial ecology of the oral cavity, in Alexander M (ed): Advances in Microbial Ecology, vol 3. New York, NY, Plenum, 1979, pp 135-192
579 12. Sitas F, Forman D, Yarnell JWG, et al: Helieobaeter pylori infection rates in relation to age and social class in a population of Welsh men. Gut 32:25-28, 1991 13. Graham DY, Malaty HD, Evans DG, et al: Epidemiology of Helieobaeter pylori in an asymptomatic population in the United States. Effect of age, race, and socioeconomic status. Gastroenterology 100: 1495-150 I, 1991 14. Muto S, Muramaya N, Asano Y, et al: Hypergastrinaemia and achlorhydria in chronic renal failure. Nephron 40:143-148, 1985 15. Shousha S, Keen C, Parkins RA: Gastric metaplasia and Campylobaeter pylori infection of duodenum in patients with chronic renal failure. J Clin Pathol 42:348-351, 1989 16. Offerhaus GJA, Kreuning J, Valentijn RM, et al: Campylobaeter pylori: Prevalence and significance in patients with chronic renal failure. Clin Nephrol 32:239-241, 1989 17. Shousha S, Bull TB, Parkins RA: Duodenal ultrastructure in patients with chronic renal failure with a comment on the incidence of Campylobaeter pylori infection .. Ultrastruct Pathol 14:1-10, 1990 18. Davenport A, Shallcross TM, Crabtree JE, et al: Prevalence of Helieobaeter pylori in patients with end stage renal failure and renal transplant recipients. Nephron 59:597-601, 1991 19. Wee A, Kang JY, Ho MS, et aI: Gastroduodenal mucosa in uraemia: Endoscopic and histological correlation and prevalence of Helieobaeter-like organisms. Gut 31: 1093-1096, 1990 20. Tytgat GNJ, Noach LA, Rauws EAJ: Isgastroduodenitis a cause of chronic dyspepsia? Scand J Gastroenterol 26: 33-39, 1991 (suppl 182)
I
NFECTION OF THE gastric mucosa with Helieobaeter pylori (formerly Campylobaeter pylori) is an important cause of gastritis l and predisposes to duodenal ulcer disease. 2 The bacterium is remarkable because of its high urease activity that splits urea in the gastric juice into ammonia and carbon dioxide. 3 The ammonia neutralizes gastric acid in the immediate vicinity of the organism and may assist its survival in the inhospitable environment of the stomach. Infection with H pylori causes a decrease in urea concentration and an increase in ammonium concentration in the gastric juice. 4 The rate of ammonia production by H pylori is limited by
From the Renal Unit and Departments of Medicine and Therapeutics, Microbiology, and Histopathology, Western Infirmary, Glasgow, Scotland. Received April 8, 1992; accepted in revised form July 17, 1992. Supported by a grant from the Biomedical Research Committee ofthe Scottish Home and Health Department, andfrom the Research Support Group of the Greater Glasgow Health Board. Address reprint requests to Peter A. Rowe, MD, MRCP, Renal Unit, Western Irifirmary, Glasgow, GIl 6NT, Scotland. © 1992 by the National Kidney Foundation, Inc. 0272-6386/92/2006-0007$3.00/0 574
the availability ofurea,5 and urea concentrations in the gastric juice are increased in patients with chronic renal failure. 4 Diagnosis of H pylori infection can be made by demonstrating the gram-negative bacteria in biopsy specimens obtained at endoscopy, or by culture of gastric antral biopsies. Microscopy requires tissue-processing, and the bacterium is microaerophilic and time-consuming to culture, which delays diagnosis. Two other commonly used more rapid tests rely on the bacterium's high urease activity. These are the urease slide test,6 which relies on demonstrating urease activity in an antral biopsy obtained at endoscopy, and the radio labeled urea breath test, 7 which depends on increased respiration of radiolabeled CO 2 derived from splitting of ingested radiolabeled urea. These tests have been shown to be reliable in patients with normal renal function,6-8 but could be affected by the high endogenous urea concentration in the gastric juice of patients with chronic renal failure. High concentrations of urea in gastric juice could saturate bacterial urease, and delay hydrolysis of the small quantity of administered radiolabeled urea, or alternatively, lead to high ammonia concentrations that could inhibit bacterial urease activity.
American Journal of Kidney Diseases, Vol XX, No 6 (December), 1992: pp 574-579
HELICOBACTER PYLORI IN CHRONIC RENAL FAILURE
We have conducted a prospective study of unselected dyspeptic patients with chronic renal failure, to evaluate the diagnostic utility of the radiolabeled urea breath test and urease slide test.
METHODS Patients with chronic renal failure were recruited from those with end-stage renal failure on renal replacement therapy or from those attending the predialysis clinic. The criterion for inclusion was dyspeptic symptoms of sufficient severity to warrant diagnostic endoscopy. Patients already receiving treatment with any acid-inhibiting agent or with antibiotics discontinued this therapy at least I week before the tests were performed. Each subject underwent a radiolabeled urea breath test within I week of a standard diagnostic upper gastrointestinal endoscopy. On the day of endoscopy, a serum sample was obtained for measurement of H pylori antibodies by an enzyme-linked immunoadsorbent assay for IgG (Helico-G kit, Porton, Cambridge, UK).9 All endoscopies were performed by one investigator (K.E.L.M.), in the morning following an overnight fast. On entering the stomach, 10 mL of resting gastric juice was aspirated by means of a trap in the suction line and sent for estimation of pH, ammonium, and urea concentrations as previously described. to After completion of the examination, three biopsies were taken from the gastric antrum on the greater curvature 2 cm from the pylorus. One was used for the urease slide test, and the other two were submitted for bacteriological culture and microscopic examination. The urease slide test was performed using a commercially available kit (CLO test, Delta West, Bentley, Western Australia). A gastric antral biopsy is placed in the preprepared well on the slide (containing urea), and if H pylori is present in the biopsy, the ammonia generated produces a change in the colour of a pH indicator. The test was interpreted by the investigator performing the endoscopy who was at that time unaware of the results of other investigations. The slide was inspected for any color change after 20 minutes and again after 24 hours of incubation at 30°C. The radiolabeled urea breath test relies on the bacterial hydrolysis of orally administered 14C-urea, with subsequent respiratory excretion of I'C02 . 7 After an overnight fast, patients were asked to clean their teeth using a clean toothbrush and tap water, without swallowing. They then drank a test meal consisting of 250 mL of Ensure Plus (Abbott, Maidenhead, UK), followed by 0.4 MBq of 14C-urea dissolved in 25 mL of water. After this, they cleaned their teeth again and rinsed out their mouths. This cleansing procedure is routinely performed to minimize any hydrolysis of the administered radiolabeled urea by the normal mouth flora (which include urease-producing bacteria)." Breath samples were collected at lO-minute intervals for 2 hours, and the excretion of I'C02 per mmol respired CO2 was calculated. Counts were corrected for body weight and expressed as a percentage of the activity of the original dose. Microscopy of the gastric biopsies was performed by one investigator (S.D.), who was unaware of the clinical details or the results of the other tests. Tissue samples were fixed in
575 formalin and a paraffin section stained with Cresyl Fast Violet for detection of H pylori. Gastric biopsies were transferred to the microbiology laboratory in sterile saline and immediately plated out onto horse blood agar plates containing Skirrow's selective supplement (Mast, UK). The plates were examined after 3 and 5 days of incubation at 37°C in a microaerophilic atmosphere. Isolates were identified by their Gram stain appearance and by their urease, oxidase, and catalase reactions. Patients were judged to be infected with H pylori if the organism was demonstrated in the antral biopsy by either microscopy or culture. Control values were obtained from subjects with normal renal function who had previously been investigated for dyspepsia. Statistical methods were analysis of variance and unpaired t test where data were normally distributed, and by the corresponding nonparametric tests (Kruskal-Wallis and MannWhitney U tests) where distribution was not normal. The breath test curves were analyzed by a repeated measures analysis of variance (BMDP P2V).
RESULTS Twenty-seven patients with chronic renal failure (15 males, 12 females) were studied. Twelve (five males, seven females) were found to be infected with H pylori, a prevalence of 46%. Microscopic examination demonstrated the presence of H pylori in 11 of the 12, while in the final case, microscopy was negative but culture was positive. Culture was positive in six of the 12 patients thus judged to be infected with H pylori and negative in the other 15 patients in whom the organism was not seen on microscopy. Data are incomplete for four noninfected patients; in two patients, no gastric aspirate was obtained and no breath test was performed; in one patient, no gastric aspirate was obtained; and in one patient, no breath test was performed. There was no significant difference in the distribution of the sexes, nor in the type and duration of renal failure treatment between the H pylori-infected and -noninfected groups (Table 1). Nor was there any difference in the current or past use of acid-inhibitory agents between the two groups. The H pylori-infected patients tended to be older, as previously reported,12.13 but this did not reach statistical significance. The only significant difference between the two groups was in the presence or absence of their own teeth. In the infected group, eight of 12 had no teeth, compared with only four of the 15 in the noninfected group (x 2 = 4.32; P < 0.05). The serum urea concentrations were increased to a similar extent in the infected and noninfected
ROWE ET AL
576 Table 1. Patient Characteristics H pylori-
H pylori-Infected Sex Male Female Renal failure treatment No. on CAPD No. on hemodialysis No. predialysis Median age in years (range) Median duration of dialysis in months (range) Median serum urea mmol/L (range) mg/dL (range) Median serum creatinine I'moi/L (range) mg/dL (range)
Noninfected
5 7
10 5
7 3 2
11 3
57.1 (34-73)
48.5 (27-67)
(0-100)
20.5 (0-142)
15
21.3 (12.0-31.2) 59.7 (33.6-87.4)
23.8 (10.7-32.2) 66.7 (30.0-90.2)
891 (287-1,664) 10.1 (3.2-18.8)
981 (612-1,700) 11.1 (6.9-19.2)
creased quickly in the first 10 minutes and then plateaued, before decreasing slowly in the remaining 60 minutes of the 2-hour test. By contrast, in the noninfected uremic patients, 14C02 excretion increased slowly after 30 minutes in six of the 11, but remained at a steady low level in the other patients. These time trends are significantly different both within each group (F = 2.27, P = 0.012) and between groups (F = 3.9, P < 0.0001). There was no overlap between the infected and noninfected uremic groups before 120 minutes. The infected uremic patients' curves slope down toward the end of the 2-hour test, and the noninfected uremic patients' slope up, so the best separation occurs in the early phase between 20 and 60 minutes. At 20 minutes, the median value
Abbreviation: CAPD, continuous ambulatory peritoneal dialysis.
patients (Table 1). The gastric aspirate urea concentrations in the H pylori-infected patients (median, 1.9; range, 0.2 to 8.7 mmol/L) were much lower (P < 0.0002) than in the noninfected patients (median, 13.8; range, 5.4 to 20.8 mmol/ L). The urea concentration in the gastric aspirate correlated with the serum urea in the noninfected patients (r = 0.53, P = 0.05), but not in the infected patients. As expected, the concentrations of ammonia in the gastric aspirate were much higher (median, 17.5; range, 2.7 to 43.1 mmol/ L) in the infected patients (P = 0.0015) than in the noninfected patients (median, 4.0; range, 0.05 to 12.3 mmol/L). There was no significant difference in gastric pH between the two groups, although the range varied widely (infected group: median, 2.9; range, 1.1 to 7.0; noninfected group: median, 2.2; range, 1.2 to 7.7). In every patient judged infected with H pylori, the urease slide test was clearly positive after both 20 minutes and 24 hours of incubation. In the noninfected group, the test was negative in every patient at both 20 minutes and after 24 hours of incubation. The results of the 14C_urea breath test are shown in Fig 1. In the uremic patients, the values obtained from the H pylori-infected group were much higher than the values obtained from the noninfected group (F = 28.1, P < 0.0001). In the infected uremic patients, excretion of 14C02 in-
500 400 300 0 0
......
200
-- -
>< 100 0'"
() 0
E E
0 50
....... (I)
en
0
40
-c
~
01
::.c::
30
----
20 10
- --
""]
0 0
30
60
90
120
Time (minutes) Fig 1. 14C02-urea breath test curves from patients with chronic renal failure (CRF; -----) and controls with normal renal function (--). e, H pylori-infected patients; 0, noninfected patients. Data are median values with interquartile ranges. Note the 10-fold difference in the scale used on the y-axiS for the infected patients (upper panel) compared with the noninfected patients (lower panel). The curves in the H pylori-infected patients are significantly different from those in the noninfected (CRF: F = 28.1, P < 0.0001; controls: F = 149.1, P < 0.0001).
HELICOBACTER PYLORI IN CHRONIC RENAL FAILURE
(range) in the infected uremic group was 207 (50 to 834) compared with 12 (0.3 to 27) kg %dose/ mmol CO2 X 100 (P < 0.0001) in the noninfected uremic group (Fig 2), and at 90 minutes the corresponding values were 177 (88 to 398) and 28.5 (1 to 79), respectively (P < 0.0001). The uremic patients infected with H pylori had somewhat higher levels of 14C02 excretion than those with normal renal function, but this was only just statistically significant (F = 4.29, P = 0.05). In contrast, the curves from the uremic noninfected patients were strikingly different (F = 8.76, P = 0.008) from those obtained from the noninfected patients with normal renal function, where no patient showed a late increase in 14C02 excretion. In three of six patients with a late increase in 14C02 excretion, organisms other than H pylori were noticed in the antral biopsy, but none were observed in the five without such a late increase. The results of serological tests were available in 22 patients. A titer of greater than or equal to 20 U/mL was interpreted as positive. Titers of H pylori IgG were significantly (P = 0.002) higher in the infected group (median, 46 U/mL; range, 6.2 to 197) than in the noninfected group (median, 9.5 U/mL; range 5 to 19.30). Titers greater than or equal to 20 U /mL were present in nine of the 11 patients with H pylori infection, but in none of the noninfected patients, giving a sensitivity of 82% and a specificity of 100%. Endoscopy was normal in five of the 12 H pylori-infected patients. Of the other seven, hiatus hernia was diagnosed in three, and one each had petechiae in the stomach and duodenum, duodenitis, gastroesophageal inflammation and petechial hemorrhage, and moderate chronic gastritis. In seven of the 12 H pylori-noninfected patients, no abnormality was found. Of the remaining five patients, one each had esophagitis and gastric stasis (a diabetic), hiatus hernia and esophagitis, scattered gastric petechiae, mild superficial gastritis, and mild duodenitis. DISCUSSION
Before the value of a diagnostic test can be judged, it is essential to make an accurate diagnosis. The method used in this study to establish the diagnosis of H pylori infection in the majority of patients was a careful microscopic examination of an antral biopsy. 8 Bacteriological culture was
577 1000
•
800
"' !::J 600
• •• ••
C
0 0
E
0400 N
> ~
"' !::J 600 c
'E
0400 0>
200 0
•
:
~ Chronic renal failure
Normal renal function
Fig 2. Individual values of 14C02 excretion corrected for respired CO2 and body weight at 20 minutes (upper panel) and 90 minutes (lower panel), in the H pyloriinfected (e) and -noninfected (0) CRF patients and in controls with normal renal function. Median value (interquartile range) in the infected CRF patients at 20 minutes was 207 (91.7 to 436.8), compared with 12 (5 to 17.25) in the noninfected CRF patients. There is no overlap between the infected and noninfected groups at 20 minutes.
positive in 50% of the patients judged to be infected, a sensitivity similar to that reported previously.8 Our data show that despite the higher gastric juice urea levels in patients with chronic uremia, diagnostic tests based on H pylori's high urease activity remain reliable. The 14C-urea breath test accurately discriminated between H pylori-infected and -noninfected uremic patients, with a 20-minute value greater than 50 indicating infection. This cutoff level in uremic patients is higher than the value of 20 usually used for subjects with normal renal function, since the excretion of radiolabeled 14C02 was greater in all subjects with chronic renal failure. The higher levels of 14C02 excretion observed could be due to the induction of urease activity in the bacterium by the prevailing higher
578
levels of urea in the gastric aspirate 4 or, alternatively, the presence of other urease-producing organisms. The late increase in 14C02 excretion seen in the renal failure patients without H pylori infection suggests that the ingested radiolabeled urea is being hydrolyzed by other bacteria. This could occur with colonization of the upper gastrointestinal tract by other urease-producing bacteria such as Proteus species. Factors that predispose to bacterial colonization include hypochlorhydria, which is common in chronic renal failure, 14 and impaired gastrointestinal motility, which might be a sequel to chronic constipation, the use of the peritoneal cavity for dialysis, or a minor degree of uremic autonomic neuropathy. Colonization of the upper gastrointestinal tract could also help explain the relative difficulty in culturing H pylori. as this is more difficult when other bacteria are present. The finding that H pylori infection was associated with an edentulous state is of interest, since the teeth are colonized by other urease-producing organisms. II Ingestion of normal mouth flora could inhibit the ability of H pylori to colonize the stomach, and since older patients are less likely to possess their own teeth, this could also help explain the increasing prevalence of H pylori infection with increasing age. 12 ,13 We were not able to identify any particular characteristics of the H pylori-noninfected patients who showed a late increase in 14C02 excretion. There were no differences in the type or duration of renal failure treatment, in the number of episodes of chronic ambulatory peritoneal dialysis peritonitis, or in the use of acid-inhibitory agents when compared with those H pylori-noninfected patients without a late increase, and there was no difference in fasting gastric pH. The numbers were too small to rule out the possibility that urease organisms from the teeth of the H pylorinoninfected patients had been able to colonize the upper gastrointestinal tract, although it is of interest that microorganisms other than H pylori were found in the antral biopsy specimens of three of the six with a late increase, and in none of the remainder. The urease slide test also proved reliable in determining H pylori status. Despite the late increase in the urease activity shown in the 14C_ urea breath test in the H pylori-noninfected
ROWE ET AL
patients, there was no suggestion of any late falsepositive results in the urease slide test. H pylori lives under the mucous layer and closely adheres to the epithelial surface in the gastric antrum. Consequently, only the urease activity of organisms present on the antral biopsy will be detected by the urease slide test, and the result is thus not affected by other urease-producing bacteria in the lumen of the stomach or small bowel. Previous studies of H pylori infection in renal failure have relied on microscopic examination or serological testing. 15-19 The presence of H pylori infection in 46% of our patients was similar to the prevalence observed (52%) in healthy subjects without dyspepsia, 13 and slightly less than the reported prevalence (62%) in dyspeptic subjects without renal failure. 2o This figure closely agrees with the prevalence of 44% reported by Offerhaus et al,16 but is higher than the 34% described by Davenport et al. I8 Both studies relied on serological tests for H pylori antibodies-a method of diagnosis shown by our data to be less sensitive than urease-based tests. This study demonstrates that urease-based tests can accurately diagnose H pylori infection in uremic patients, with suitable adjustment of the normal range for the 14C02-urea breath test. Our results suggest that H pylori infection is no more frequent in chronic renal failure than in healthy volunteers. It is therefore unlikely that it is any more important a factor in causing gastroduodenitis in chronic renal failure than in the healthy population. Reliable diagnosis of H pylori infection by means of urease-based tests will permit further studies of the mechanisms underlying gastritis and dyspepsia in chronic renal failure. ACKNOWLEDGMENT The authors would like to thank Jennifer Harwood for measuring the gastric juice ammonia, and the Department of Nuclear Medicine at the Gardiner Institute for performing the 14C-urea breath tests.
REFERENCES I. Rauws EAJ, Langenberg W, Houthoff HJ, et al: Cam-
pylobacter pyloridis-associated chronic active gastritis. Gastroenterology 94:33-40, 1988 2. Axon ATR: Helicobacter pylori in the pathogenesis of peptic ulcer-Evidence in favour, in Malfertheiner P, Ditschuneit H (eds): Helicobacter pylori, Gastritis and Peptic Ulcer. Berlin, Germany, Springer-Verlag, 1990, pp 345-357 3. Bode G, Malfertheiner P, Lennhardt G, et al: Virulence factors of Helicobacter pylori-Ultrastructural features, in
HELICOBACTER PYLORI IN CHRONIC RENAL FAILURE Malfertheiner P, Ditschuneit H (eds): Helieobaeter pylori, Gastritis and Peptic Ulcer. Beriin, Germany, Springer-Veriag, 1990, pp 63-73 4. Kim H, Park C, Jang WI, et al: The gastric juice urea and ammonia levels in patients with Campylobaeter pylori. Am J Clin PathoI94:187-191, 1990 5. Chittajallu RS, Neithercut WD, Macdonald AMI, et aI: Effect of increasing H elieobaeter pylori ammonia production by urea infusion on plasma gastrin concentrations. Gut 32: 21-24, 1991 6. Marshall BJ, Warren JR, Francis GJ, et aI: Rapid urease test in the management of Campylobaeter pyloridis-associated gastritis. Am J Gastroenterol 82:200-210, 1987 7. Dill S, Payne-James JJ, Misiewicz JJ, et aI: Evaluation of 13C-urea breath test in the detection of H elieobaeter pylori and in monitoring the effect of tripotassium dicitratobismuthate in non-ulcer dyspepsia. Gut 31:1237-1241, 1990 8. Logan RPH, Polson RJ, Misiewicz JJ, et aI: Simplified single sample 13carbon urea breath test for Helieobaeter pylori: Comparison with histology, culture, and ELISA serology. Gut 32:1461-1464, 1991 9. Mendall MA, Goggin P, Bellhouse A, et aI: Performance of a commercially available serological test for Helieobaeter pylori and its relation to age. Gut 32:AI257, 1991 (abstr) 10. Neithercut WD, Milne A, Chittajallu RS, et al: The detection of Helieobaeter pylori infection of the gastric mucosa by measurement of gastric aspirate ammonium and urea concentrations. Gut 32:973-976, 1991 II. Bowden GHW, Ellwood DS, Hamilton IR: Microbial ecology of the oral cavity, in Alexander M (ed): Advances in Microbial Ecology, vol 3. New York, NY, Plenum, 1979, pp 135-192
579 12. Sitas F, Forman D, Yarnell JWG, et al: Helieobaeter pylori infection rates in relation to age and social class in a population of Welsh men. Gut 32:25-28, 1991 13. Graham DY, Malaty HD, Evans DG, et al: Epidemiology of Helieobaeter pylori in an asymptomatic population in the United States. Effect of age, race, and socioeconomic status. Gastroenterology 100: 1495-150 I, 1991 14. Muto S, Muramaya N, Asano Y, et al: Hypergastrinaemia and achlorhydria in chronic renal failure. Nephron 40:143-148, 1985 15. Shousha S, Keen C, Parkins RA: Gastric metaplasia and Campylobaeter pylori infection of duodenum in patients with chronic renal failure. J Clin Pathol 42:348-351, 1989 16. Offerhaus GJA, Kreuning J, Valentijn RM, et al: Campylobaeter pylori: Prevalence and significance in patients with chronic renal failure. Clin Nephrol 32:239-241, 1989 17. Shousha S, Bull TB, Parkins RA: Duodenal ultrastructure in patients with chronic renal failure with a comment on the incidence of Campylobaeter pylori infection .. Ultrastruct Pathol 14:1-10, 1990 18. Davenport A, Shallcross TM, Crabtree JE, et al: Prevalence of Helieobaeter pylori in patients with end stage renal failure and renal transplant recipients. Nephron 59:597-601, 1991 19. Wee A, Kang JY, Ho MS, et aI: Gastroduodenal mucosa in uraemia: Endoscopic and histological correlation and prevalence of Helieobaeter-like organisms. Gut 31: 1093-1096, 1990 20. Tytgat GNJ, Noach LA, Rauws EAJ: Isgastroduodenitis a cause of chronic dyspepsia? Scand J Gastroenterol 26: 33-39, 1991 (suppl 182)
