Journal of Gastroenterology and Hepatology (1992) I, 569-57 1
ALIMENTARY T R A C T A N D PANCREAS
Field evaluation of a rapid, simple and inexpensive urease test for the detection of Helicobacter pylori P. H . KATELARIS*, D. G. L O W E , + P. N O R B U * A N D M. J. G. F A R T H I N G * Departments of *Gastroenterology and +Pathology St Bartholomew's Hospital, London, U K and *Doeguling Tibetan Hospital Mundgod, Karnataka India
Abstract A modified rapid urease test (MRU test) for the detection of Helicobacter pylori was evaluated under field conditions during an endoscopic survey in rural India and compared with a commercially available urease test (CLO test) and with histology. Of 195 consecutive subjects who underwent upper gastrointestinal endoscopy, 153 (78.5%) were positive for Helicobacter pylori when tested by the CLO test and/or histology. The sensitivity and specificity of the MRU test relative to this was 97.4 and 95.2%, respectively when the test was read over a 3 h period. The MRU test was positive in 77.4, 89.0, 93.8 and 96.6% of cases at 1, 5, 20 and 60 min, respectively, compared with 2.7, 14.4,48.6 and 71.2% of cases for the CLO test at the same time. The accuracy of the MRU test was thus similar to that of other methods for the detection of Helicobacter pylori. Furthermore, it gave a positive diagnosis more rapidly than other tests, in most cases before the subject had left the endoscopy suite. T h e MRU test is extremely simple to prepare and read and costs less than E0.05 per test compared with 22.26 for a CLO test. It is suitable for use in clinical or epidemiological work and especially where cost factors are critical.
Key words: Helicobacter pylori, rapid urease test, upper gastrointestinal endoscopy.
INTRODUCTION There are now several validated methods for the detection of Helicobacter pylori infection of the gastric mucosa. These are: urease tests, Gram's stain, histology, culture, breath tests and serology.'q2Each has advantages, disadvantages and is applicable to particular situations. The value of each test must be assessed with regard to the following parameters: sensitivity and specificity, time elapsed for a positive result, availability, cost, convenience and simplicity of use, invasiveness and risks of the test to the subject. Rapid urease tests, of which there are a number, compare favourably with other available methods for the detection of H . pylori when judged by these criteria. These tests detect urease activity present in biopsy specimens. Helicobacter pylori produces urease in abundance. When a biopsy is incubated in a medium containing urea and a pH sensitive colour marker, urease hydrolyses urea to carbon dioxide and ammonia causing a rise in pH and a change in the colour of the medium. The published results of the evaluation of different methods for the detection of H. pylori are usually obtained in the ideal circumstances of a research setting in a teaching hospital and may not reflect the per-
formance of a test in routine or less ideal settings. The aim of this study was to evaluate a modified rapid urease test (MRU test) under field conditions and to compare it with histology and a commercially available rapid urease test (CLO test, Delta West Ltd, Bentley, Western Australia).
METHODS Preparation of MRU test Each test consisted of 0.5 m L of 10% (wtlvol) unbuffered urea in distilled water (pH 6.8) to which was added one drop of a 1% phenol red (free acid) suspension in a clear 0.7 mL Eppendorf tube. A gastric biopsy was placed in the tube. A positive test was indicated by rapid colour change of the media surrounding the biopsy from yellow to magenta and a plume of magenta colour trailing upwards from the biopsy, followed by rapid generalized colour change throughout the media. T o achieve rapid positive results it was important not to shake the tube after placing the biopsy in it. This test differs from previous reports where either a lesser concentration of
Correspondence: Dr Peter Katelaris, Department of Gastroenterology, St Bartholomew's Hospital, West Smithfield, London EClA 7BE, United Kingdom. Accepted for publication 9 June 1992.
P. H . Katelaris et al.
570 urea,3 a greater volume of urea solution4 or a greater concentration of phenol red5 were used and the pH titrated to 6.5-6.8. These tests have been regarded as positive if a colour change occurred within 1 or ~ v e r n i g h t .In ~ the current study the pH was not subsequently adjusted after addition of phenol red and the test was read over 24 h to determine the time course for positive results. The test should appear bright yellow before its use; orange or magenta discolouration indicates that the pH is too high (above 6.8) and that the pH of the distilled water in use should be measured. In this case the pH would require titration with HCI to pH 6.5-6.8, or until the yellow colour appeared. The urea solution may be stored at 4°C and the test prepared on the day of use to ensure colour stability. Alternatively, the test may be made then stored at -20°C but the completed test should not be stored at 4°C because of discolouration that occurs with time. In this study, MRU tests were prepared daily from stock reagents and warmed to room temperature before use.
Field study The test was evaluated during an epidemiological survey of peptic ulcer disease in a rural community in India. One hundred and ninety-five consecutive endoscopies were performed by a single endoscopist. Two biopsies were taken from the antrum 2 cm from the pylorus anteriorly, one for the MRU test and the other for the CLO test. Two biopsies were taken for histology using haematoxylin and eosin and cresyl fast violet stains and these samples were examined by a single pathologist. Urease tests were read independently over a 24 h period and the time elapsed until a positive result was recorded. Histology was rated as positive or negative for H. pylori. Positive specimens were graded as having few (l), moderate numbers (2), or many (3) organisms present using the Sydney System classification of gastritk6 Ethical approval was obtained from the administration of the Doeguling Hospital and informed consent was obtained from each subject before endoscopy.
both. Using these data the MRU test identified 150/153 positives and 39/42 negative specimens when read during a 24 h period. This represents a sensitivity of 98.0% and a specificity of 92.8% when read during this time. However only two of the MRU tests became positive after more than 3 h. Of these, one was a false-positive. As the test is very rapid, when a time limit for positivity of 3 h is used, the sensitivity becomes 97.4% and the specificity 95.2%. The CLO test alone identified 146/153 positive specimens and (by definition) all negative samples and thus had a sensitivity of 95.4% and a specificity of ‘100%’ after 24 h. T h e positive predictive value of the MRU test at 3 h was 98.7% and the negative predictive value was 90.9% and for the CLO test ‘100%’ and 85.7%, respectively at 24 h. One late positive result by the CLO test was negative by both histology and MRU test and although defined as a true-positive probably represented a falsepositive for this test. Histology alone identified 109/153 positive samples yielding a sensitivity of 71.2% and a specificity of ‘100%’. The MRU test was positive in 77.4% of cases within 1 min and 93.8% within 20min compared with 2.7 and 48.6% of cases for the CLO test at the same times. Of the CLO tests 4.8% did not become positive until between 12 and 24 h. Only 3.4% of positive MRU tests were not positive by 60 min. The time course for positive results for both tests is shown in Fig. 1. There was a Lorrelation between the histological grade of samples and the time to positive of CLO tests. Of samples graded 3 (many organisms), 39.4% were CLO test positive within 5 min compared with 8.6% of those graded 1-2 (P < 0.001). As nearly all MRU tests were positive by this time, no such correlation was observed with this test.
DISCUSSION The sensitivity and specificity of the results for the MRU test were comparable with those reported for other methods of detecting H. ~ y l o r i . *These ,~ results were achieved during a field study in non-ideal circumstances. Furthermore, the gold standard for comparison as defined in this study was biased against a greater sensitivity and
Definition of positive and negative results There is no absolute gold standard for comparison of methods for the detection of H. pylori as no test has both 100% sensitivity and specificity and published results vary. Published sensitivity and specificity for CLO test over 24 h are approximately 95 and ~OOYO’-~ and for histology about 90 and loo%, respectively, depending on the number of sections, types of stains used and the experience of the observer.”-” For this study a sample was defined as a true-negative if both CLO test and histology were negative and a true-positive if either CLO test or histology were positive. ~ ~ _ _ _ _ _ __
0
RESULTS Of 195 subjects, 153 (78.5%) were positive by the CLO test and/or histology and 42 (21.5%) were negative by
1 rnin
5 rnin
2011117 60 min
3h
3-12 h
12-24h
Time
Figure 1 Time elapsed before a positive result for the MKU test (0)and CLO test (A).
Field evaluation of a rapid urease test specificity of the MRU test compared with the CLO test or histology, as any positive MRU test that was negative by the other two was regarded as a false-positive. As the MRU test was more sensitive than both the other two tests considered individually, the MRU test can be stated to be at least as good and probably better than the results reported here. The MRU test is extremely sinple to prepare requiring neither specific laboratory skills or equipment. T h e speed of the test probably relates to the fact that the urea solution is unbuffered. This rapidity with which positive results are apparent is convenient to the endoscopist as most positive tests will be apparent before the end of the endoscopy and nearly all before the subject has left the endoscopy suite. As no tests became positive after 12 h and only two after 3 h (one of which was a false-positive) it is appropriate to limit the duration that the test is read to 3 h. This test does not contain a bacteriostatic agent but as positives are rapidly detected it is likely that only preformed urease is being detected. Contaminants such as Proteus and Pseudomonas spp. may produce low levels of urease over time and this may theoretically lead to false-positives if read later. Although this did not appear to be a problem it further supports limiting the time to read the test to 3 h. The unit cost to make the MRU test is approximately f0.05 compared with a retail price of f2.26 for the CLO test. The MRU test is a rapid, reliable, simple and inexpensive method for detecting H. pylori infection. It is useful in a clinical or research setting where it is considered desirable to detect H. pylori infection during endoscopy, and especially where cost factors are critical.
ACKNOWLEDGEMENTS Dr Katelaris was supported by The Smith & Nephew Foundation. Professor Farthing gratefully acknowledges financial support by the Wellcome Trust. The authors
571
thank Dr G. Tippett, R. Hamlyn and the staff of the Doeguling Tibetan Hospital for valuable assistance.
REFERENCES 1. PETERSON W. L. Helicobacter pylori and peptic ulcer disease. N . Engl. J. Med. 1991; 324: 1043-8. 2. MEGRAUD F. Comparison of different tests for Campylobacter pylori. Scand. J. Gastroenterol. 1988; 23 (Suppl. 142): 64-8. T. J., GAL.A. & LEEA. Campylobacter 3. HAZEIS.S., BORODY pyloridis Gastritis I: Detection of urease as a marker of bacterial colonisation and gastritis. Am. J. Gastroenterol. 1987; 82: 292-6. 4. ARVIND A. S., COOKR. S., TABAQCHALI S. & FARTHING M. J. G. One-minute endoscopy room test for Campylobacter pylori. Lancet 1988; i: 704. 5. THILLAINAYAGAM A. V., ARVINDA. S., COOKK. S., HARRISON 1. G., TABAQCHAI.~ S. & FARTHING M. J. G. The diagnostic efficiency of an ultrarapid endoscopy room test for Helicobacter pylori. Gut 1991; 32: 467-9. 6. PRICEA. B. The Sydney System: Histological division. J. Gastroenterol. Hepatol. 1991; 6: 209-22. 7. MARSHALL. B. J., WARRENJ. R., FRANCIS G. J., LAKGTOK S. R., GOOI)WIN C. S. & BLINCOW E. D. Rapid urease test in the management of Campylobacter pyloridis-associated gastritis. Am. J . Gastroenterol. 1987; 92: 200-10. 8. MORRIS A., MCINTYKE D., ROSET & NICHOI.SOK G. Rapid diagnosis of Campylobacter pyloridis infection. Lancet 1986; i: 149. 9. BORROMEO M., LAMRERT J. R. & PINKART K. J. Evaluation of ‘CLO-test’ to detect Carnpylobacter pyloridis in gastric mucosa. 3.Clin. Pathol. 1987; 40: 462-3. 10. MAIIAKE., KEMP J., WESTBLOMU et al. Evaluation of staining methods for identifying Campylobacter pylori. Am. J. Clin. Pathol. 1988; 90: 450-3. 11. McNu1.n C. A. M., DEKTJ. C., UFF J. C. ef al. The prevalence of Campylobacter pylori in 1447 patients at endoscopy. A m . 3.Gastroenterol. 1988; 83: 1035.
ALIMENTARY T R A C T A N D PANCREAS
Field evaluation of a rapid, simple and inexpensive urease test for the detection of Helicobacter pylori P. H . KATELARIS*, D. G. L O W E , + P. N O R B U * A N D M. J. G. F A R T H I N G * Departments of *Gastroenterology and +Pathology St Bartholomew's Hospital, London, U K and *Doeguling Tibetan Hospital Mundgod, Karnataka India
Abstract A modified rapid urease test (MRU test) for the detection of Helicobacter pylori was evaluated under field conditions during an endoscopic survey in rural India and compared with a commercially available urease test (CLO test) and with histology. Of 195 consecutive subjects who underwent upper gastrointestinal endoscopy, 153 (78.5%) were positive for Helicobacter pylori when tested by the CLO test and/or histology. The sensitivity and specificity of the MRU test relative to this was 97.4 and 95.2%, respectively when the test was read over a 3 h period. The MRU test was positive in 77.4, 89.0, 93.8 and 96.6% of cases at 1, 5, 20 and 60 min, respectively, compared with 2.7, 14.4,48.6 and 71.2% of cases for the CLO test at the same time. The accuracy of the MRU test was thus similar to that of other methods for the detection of Helicobacter pylori. Furthermore, it gave a positive diagnosis more rapidly than other tests, in most cases before the subject had left the endoscopy suite. T h e MRU test is extremely simple to prepare and read and costs less than E0.05 per test compared with 22.26 for a CLO test. It is suitable for use in clinical or epidemiological work and especially where cost factors are critical.
Key words: Helicobacter pylori, rapid urease test, upper gastrointestinal endoscopy.
INTRODUCTION There are now several validated methods for the detection of Helicobacter pylori infection of the gastric mucosa. These are: urease tests, Gram's stain, histology, culture, breath tests and serology.'q2Each has advantages, disadvantages and is applicable to particular situations. The value of each test must be assessed with regard to the following parameters: sensitivity and specificity, time elapsed for a positive result, availability, cost, convenience and simplicity of use, invasiveness and risks of the test to the subject. Rapid urease tests, of which there are a number, compare favourably with other available methods for the detection of H . pylori when judged by these criteria. These tests detect urease activity present in biopsy specimens. Helicobacter pylori produces urease in abundance. When a biopsy is incubated in a medium containing urea and a pH sensitive colour marker, urease hydrolyses urea to carbon dioxide and ammonia causing a rise in pH and a change in the colour of the medium. The published results of the evaluation of different methods for the detection of H. pylori are usually obtained in the ideal circumstances of a research setting in a teaching hospital and may not reflect the per-
formance of a test in routine or less ideal settings. The aim of this study was to evaluate a modified rapid urease test (MRU test) under field conditions and to compare it with histology and a commercially available rapid urease test (CLO test, Delta West Ltd, Bentley, Western Australia).
METHODS Preparation of MRU test Each test consisted of 0.5 m L of 10% (wtlvol) unbuffered urea in distilled water (pH 6.8) to which was added one drop of a 1% phenol red (free acid) suspension in a clear 0.7 mL Eppendorf tube. A gastric biopsy was placed in the tube. A positive test was indicated by rapid colour change of the media surrounding the biopsy from yellow to magenta and a plume of magenta colour trailing upwards from the biopsy, followed by rapid generalized colour change throughout the media. T o achieve rapid positive results it was important not to shake the tube after placing the biopsy in it. This test differs from previous reports where either a lesser concentration of
Correspondence: Dr Peter Katelaris, Department of Gastroenterology, St Bartholomew's Hospital, West Smithfield, London EClA 7BE, United Kingdom. Accepted for publication 9 June 1992.
P. H . Katelaris et al.
570 urea,3 a greater volume of urea solution4 or a greater concentration of phenol red5 were used and the pH titrated to 6.5-6.8. These tests have been regarded as positive if a colour change occurred within 1 or ~ v e r n i g h t .In ~ the current study the pH was not subsequently adjusted after addition of phenol red and the test was read over 24 h to determine the time course for positive results. The test should appear bright yellow before its use; orange or magenta discolouration indicates that the pH is too high (above 6.8) and that the pH of the distilled water in use should be measured. In this case the pH would require titration with HCI to pH 6.5-6.8, or until the yellow colour appeared. The urea solution may be stored at 4°C and the test prepared on the day of use to ensure colour stability. Alternatively, the test may be made then stored at -20°C but the completed test should not be stored at 4°C because of discolouration that occurs with time. In this study, MRU tests were prepared daily from stock reagents and warmed to room temperature before use.
Field study The test was evaluated during an epidemiological survey of peptic ulcer disease in a rural community in India. One hundred and ninety-five consecutive endoscopies were performed by a single endoscopist. Two biopsies were taken from the antrum 2 cm from the pylorus anteriorly, one for the MRU test and the other for the CLO test. Two biopsies were taken for histology using haematoxylin and eosin and cresyl fast violet stains and these samples were examined by a single pathologist. Urease tests were read independently over a 24 h period and the time elapsed until a positive result was recorded. Histology was rated as positive or negative for H. pylori. Positive specimens were graded as having few (l), moderate numbers (2), or many (3) organisms present using the Sydney System classification of gastritk6 Ethical approval was obtained from the administration of the Doeguling Hospital and informed consent was obtained from each subject before endoscopy.
both. Using these data the MRU test identified 150/153 positives and 39/42 negative specimens when read during a 24 h period. This represents a sensitivity of 98.0% and a specificity of 92.8% when read during this time. However only two of the MRU tests became positive after more than 3 h. Of these, one was a false-positive. As the test is very rapid, when a time limit for positivity of 3 h is used, the sensitivity becomes 97.4% and the specificity 95.2%. The CLO test alone identified 146/153 positive specimens and (by definition) all negative samples and thus had a sensitivity of 95.4% and a specificity of ‘100%’ after 24 h. T h e positive predictive value of the MRU test at 3 h was 98.7% and the negative predictive value was 90.9% and for the CLO test ‘100%’ and 85.7%, respectively at 24 h. One late positive result by the CLO test was negative by both histology and MRU test and although defined as a true-positive probably represented a falsepositive for this test. Histology alone identified 109/153 positive samples yielding a sensitivity of 71.2% and a specificity of ‘100%’. The MRU test was positive in 77.4% of cases within 1 min and 93.8% within 20min compared with 2.7 and 48.6% of cases for the CLO test at the same times. Of the CLO tests 4.8% did not become positive until between 12 and 24 h. Only 3.4% of positive MRU tests were not positive by 60 min. The time course for positive results for both tests is shown in Fig. 1. There was a Lorrelation between the histological grade of samples and the time to positive of CLO tests. Of samples graded 3 (many organisms), 39.4% were CLO test positive within 5 min compared with 8.6% of those graded 1-2 (P < 0.001). As nearly all MRU tests were positive by this time, no such correlation was observed with this test.
DISCUSSION The sensitivity and specificity of the results for the MRU test were comparable with those reported for other methods of detecting H. ~ y l o r i . *These ,~ results were achieved during a field study in non-ideal circumstances. Furthermore, the gold standard for comparison as defined in this study was biased against a greater sensitivity and
Definition of positive and negative results There is no absolute gold standard for comparison of methods for the detection of H. pylori as no test has both 100% sensitivity and specificity and published results vary. Published sensitivity and specificity for CLO test over 24 h are approximately 95 and ~OOYO’-~ and for histology about 90 and loo%, respectively, depending on the number of sections, types of stains used and the experience of the observer.”-” For this study a sample was defined as a true-negative if both CLO test and histology were negative and a true-positive if either CLO test or histology were positive. ~ ~ _ _ _ _ _ __
0
RESULTS Of 195 subjects, 153 (78.5%) were positive by the CLO test and/or histology and 42 (21.5%) were negative by
1 rnin
5 rnin
2011117 60 min
3h
3-12 h
12-24h
Time
Figure 1 Time elapsed before a positive result for the MKU test (0)and CLO test (A).
Field evaluation of a rapid urease test specificity of the MRU test compared with the CLO test or histology, as any positive MRU test that was negative by the other two was regarded as a false-positive. As the MRU test was more sensitive than both the other two tests considered individually, the MRU test can be stated to be at least as good and probably better than the results reported here. The MRU test is extremely sinple to prepare requiring neither specific laboratory skills or equipment. T h e speed of the test probably relates to the fact that the urea solution is unbuffered. This rapidity with which positive results are apparent is convenient to the endoscopist as most positive tests will be apparent before the end of the endoscopy and nearly all before the subject has left the endoscopy suite. As no tests became positive after 12 h and only two after 3 h (one of which was a false-positive) it is appropriate to limit the duration that the test is read to 3 h. This test does not contain a bacteriostatic agent but as positives are rapidly detected it is likely that only preformed urease is being detected. Contaminants such as Proteus and Pseudomonas spp. may produce low levels of urease over time and this may theoretically lead to false-positives if read later. Although this did not appear to be a problem it further supports limiting the time to read the test to 3 h. The unit cost to make the MRU test is approximately f0.05 compared with a retail price of f2.26 for the CLO test. The MRU test is a rapid, reliable, simple and inexpensive method for detecting H. pylori infection. It is useful in a clinical or research setting where it is considered desirable to detect H. pylori infection during endoscopy, and especially where cost factors are critical.
ACKNOWLEDGEMENTS Dr Katelaris was supported by The Smith & Nephew Foundation. Professor Farthing gratefully acknowledges financial support by the Wellcome Trust. The authors
571
thank Dr G. Tippett, R. Hamlyn and the staff of the Doeguling Tibetan Hospital for valuable assistance.
REFERENCES 1. PETERSON W. L. Helicobacter pylori and peptic ulcer disease. N . Engl. J. Med. 1991; 324: 1043-8. 2. MEGRAUD F. Comparison of different tests for Campylobacter pylori. Scand. J. Gastroenterol. 1988; 23 (Suppl. 142): 64-8. T. J., GAL.A. & LEEA. Campylobacter 3. HAZEIS.S., BORODY pyloridis Gastritis I: Detection of urease as a marker of bacterial colonisation and gastritis. Am. J. Gastroenterol. 1987; 82: 292-6. 4. ARVIND A. S., COOKR. S., TABAQCHALI S. & FARTHING M. J. G. One-minute endoscopy room test for Campylobacter pylori. Lancet 1988; i: 704. 5. THILLAINAYAGAM A. V., ARVINDA. S., COOKK. S., HARRISON 1. G., TABAQCHAI.~ S. & FARTHING M. J. G. The diagnostic efficiency of an ultrarapid endoscopy room test for Helicobacter pylori. Gut 1991; 32: 467-9. 6. PRICEA. B. The Sydney System: Histological division. J. Gastroenterol. Hepatol. 1991; 6: 209-22. 7. MARSHALL. B. J., WARRENJ. R., FRANCIS G. J., LAKGTOK S. R., GOOI)WIN C. S. & BLINCOW E. D. Rapid urease test in the management of Campylobacter pyloridis-associated gastritis. Am. J . Gastroenterol. 1987; 92: 200-10. 8. MORRIS A., MCINTYKE D., ROSET & NICHOI.SOK G. Rapid diagnosis of Campylobacter pyloridis infection. Lancet 1986; i: 149. 9. BORROMEO M., LAMRERT J. R. & PINKART K. J. Evaluation of ‘CLO-test’ to detect Carnpylobacter pyloridis in gastric mucosa. 3.Clin. Pathol. 1987; 40: 462-3. 10. MAIIAKE., KEMP J., WESTBLOMU et al. Evaluation of staining methods for identifying Campylobacter pylori. Am. J. Clin. Pathol. 1988; 90: 450-3. 11. McNu1.n C. A. M., DEKTJ. C., UFF J. C. ef al. The prevalence of Campylobacter pylori in 1447 patients at endoscopy. A m . 3.Gastroenterol. 1988; 83: 1035.
