Efficacy of fecal helicobacter pylori(HP)determination in patients with nonvariceal acute upper gastrointestinal bleeding(NVAUGIB)

GASTROENTEROLOGY Vol. 118, No.4

A506 AGA ABSTRACTS

2706 USEFULNESS OF URINE·BASED ELISA FOR DETECTION OF IGG ANTIBODIES TO HEUCOBACTER PYWRI IN CHILDREN. Seiichi Kato, Tetsuya Tachikawa, Kyoko Ozawa, Mutsuko Konno, Yasuo Oyake, Yutaka Nakazato, Hitoshi Tajiri, Masumi Okada, Takuji Fujisawa, Kazuie linuma, Tohoku Univ Sch of Medicine, Sendai, Japan; Otsuka Pharm Co , Ltd, Tokushima, Japan; Sapporo Kosei Gen Hosp, Sapporo, Japan; Hachinohe Red Cross Hosp, Hachinohe, Japan; Social Insurance Ohmiya Gen Hosp, Ohmiya, Japan; Osaka Univ Med Sch, Suita, Japan; Wakayama Rosai Hosp, Wakayama, Japan; Kurume Univ Sch of Medicine, Kurume, Japan. Background: Serum enzyme-linked immunosorbent assay (ELISA) kits have been widely used for non-invasive testing to detect Helicobacter pylori infection. However, more non-invasive tests are desirable for diagnosis in children or for use in epidemiological studies. Our aim was to investigate usefulness of recently developed urine-based ELISA kit (URINELISA H. pylori Antibody) in children. Methods: Random voided urine specimens and sera collected from 816 children (0-15 years) were measured using two commercially available serum-based ELISA kits and the urine-based kit for IgG antibodies to H. pylori. Based on results with the two serum-based ELISA kits, the sensitivity, specificity and accuracy of the urine-based ELISA were evaluated. With regard to false-positive and false-negative results, urinary IgG concentration was studied in some specimens. Results: Both two serum ELISAs were positive in 41 children and were negative in 666, who were enrolled in this study. Overall sensitivity, specificity, and accuracy of urine-based ELISA were 85.4%, 95.5%, and 94.9%, respectively: age group of 1-6 years showed high sensitivity (100%) and specificity (94.7%-99.3%). Urinary IgG concentrations and IgG/creatinine levels were higher in false-positives (p
e

2708 SWALLOWED STRING SUPERCEDES STOMACH SCOPING FOR H. PYWRI ISOLATION? Amy L. Samuels, Helen M. Windsor, Grace Y. Ho, Luke D. Goodwin, Barry J. Marshall, Dept of Microbiology, Univ of Western Australia, Nedlands, Australia; Dept of Medicine, Univ of Western Australia, Nedlands, Australia; Gastroenterology Dept, Sir Charles Gairdner Hosp, Nedlands, Australia. INTRODUCTION: Since antibiotic resistant H. pylori has emerged as a clinical problem there is a need for reliable non-invasive techniques to collect H. pylori for culture. The aim of this study was to develop improved methods for isolation of H. pylori from a swallowed gastric string. PATIENTS AND METHODS: 40 patients whose H. pylori status had been determined by endoscopic biopsy or C I 4 urea breath test swallowed a capsule containing 90cm of nylon string (Enterotest). After one hour the string was retrieved. The distal 60cm of string was cut in two. Half was placed in brain heart infusion broth (BHill) and half was briefly washed in saline before placement in BHIB. Two aliquots from each BHIE mixture

were plated onto 3 different selective media plates. The BHIB was removed from the string and the broth suspensions were concentrated 30x by centrifugation. Two aliquots of each resuspended pellet were plated onto the 3 selective media - Skirrows agar; and Wilkins-Chalgren agar & Colistin-Nalidixic acid agar both with Dent selective supplement. Plates were incubated for 7 days at 37"C in an atmosphere with 10% CO 2 , RESULTS: String test cultures were successful in 32 of 33 patients known to be infected with H. pylori (97%, CI = 84-99%). 7 patients were determined to be negative for H. pylori by UBT and were also negative for the string test. In an anonymous questionnaire, 73% of patients preferred the string test to endoscopy. CONCLUSION: High isolation rates can be achieved using this non invasive method if selective media are used and if the string is washed prior to plating out to minimize the nasopharageal contaminants. Colonies of H. pylori isolated using this technique can be used to determine antibiotic sensitivity levels as well as for molecular epidemiological studies where patients might refuse an endoscopy. This non invasive, safe procedure can easily be performed by clinical research staff or in general practice. 2709 WHAT HAPPENS TO CLOTEST RAPID UREASE TESTS OVER THE WEEKEND? Barry 1. Marshall, Simon O. Chairman, Helen M. Windsor, Grace Y. Ho, Luke D. Goodwin, Gastroenterology Dept, Sir Charles Gairdner Hosp, Nedlands, Australia; Dept of Medicine, Univ of Western Australia, Nedlands, Australia. INTRODUCTION: Gel-based rapid urease tests such as the CLOtest continue to provide the endoscopist with convenient and accurate diagnosis of H. pylori. The sensitivity of the CLOtest rapid urease test is reportedly highest when read at 24 hours. This time point is inconvenient when tests are performed on Friday and must be read again on Saturday. It was our observation that yellow CLOtests rarely changed colour between Saturday and Monday. Therefore there was probably little disadvantage to the Monday reading. AIM: To see if CLOtests changed colour between 24 and 72 hours when left at room temperature (17-22°C) over the weekend. METHODS: A digital photography system ("CLOcam") was implemented using a Kodak 323 webcam with a 640 x 480 pixel resolution, connected to the Internet. The pictures were retrieved remotely from the Internet location ..www.hpylori.com.au.. (Figure). During Friday morning endoscopy sessions, consenting patients had five mucosal biopsies taken for: CLOtest (1), histology (2), and culture (2). The CLOtests were warmed to 30-35°C for 6 hours then kept at room temperature for the next 64 hours. "CLOcam" snapshots were taken at least every hour. Digital images at the 24 hour and 72 hour time points were examined by an independent observer. Histology was examined blind after staining with H&E and Toluidine blue. Culture was performed on specimens which were either histology or CLOtest positive. RESULTS: Over eight weekends, 55 consecutive CLOtests were studied from Friday lists. At 24 hours, 18 CLOtests were red and 37 were yellow. Between 24 and 72 hours one other CLOtest changed from yellow to red. Histology from the "late" CLOtest showed occasional spiral organisms but culture was negative. CONCLUSION: Positive CLOtests at 24 hours always remain red. Occasional tests which change from yellow to red between 24 and 72 hours (I case in this series) may be true positives from patients with patchy colonization or sparse organisms. In this study 97% of negative CLOtests remained yellow at room temperature for 72 hours, allowing Friday tests to be read on Monday morning. 2710 EFFICACY OF FECAL HELICOBACTER PYLORI(HP)DETER. MINA TION IN PATIENTS WITH NONVARICEAL ACUTE UPPER GASTROINTESTINAL BLEEDING(NVAUGm). Diego Lopez-Penas, Antonio Naranjo-Rodriguez, Angel Gonzalez-Galilea, Antonio Hervas-Molina, Fernando Lopez-Rubio, Fernando RodriguezLopez, Gonzalo Mino-Fugarolas, U C Aparato Digest H U Reina Sofia, Cordoba. Spain; S Anatomy Patologica H U Reina Sofia, Cordoba, Spain; S Microbiology H U Reina Sofia, Cordoba, Spain. BACKGROUND.Several methods are available to detect HP infection in clinical practice.It has been suggested that sensitivity of these tests is lower for patients with NVAUGIB.A new test to detect HP antigens in stool samples has been recently developed,using an enzyme immunoassay(HpSA).Its accuracy in bleeding patients with melaena has not been well defined.AIMS.To determine the accuracy of HpSA test in patients with NVAUGIB and to compare HpSA test with standard tests for detection of HP infection.PATIENTS AND METHODS.32 patients with NVAUGIB were included.Endoscopy was performed within 12h after admission.The cause of bleeding was duodenal ulcer 15,gastric ulcer I D.gastric mucosae acute lesions 5 and peptic esofagitis 2.Patients were tested for HP infection by measurement of serum immunoglobulin G and antral and corpus biopsy specimens for rapid urease test and histology with H&E and Giemsa stain; l3C-Urea breath test was performed within 2 days after endoscopy and stool samples were collected on the same day to perform HpSA test.A patient was considered infected by HP if three out of the five methods were positive.Sensitivity,specificity,positive predictive value and negative predictive value of the different tests were compared.RESULTS.23 patients

April 2000

AGAA507

out of 32 were infected by HP.The results of the different tests for the detection of HP are shown in the table.The results taking account only patients with gastric or duodenal ulcer are included in brackets.22 patients having a positive HpSA test had melaena.2 patients with HP infection had a negative rapid urease test and a positive HpSA test.CONCLUSIONS.HpSA test is technically feasible in patients with melaena.This method shows a good sensitivity and positive predictive value compared to the others tests, and it could be considered as a complement to urease test in order to reduce false-negative results.

Sensitivity Specificity PPV NPV

Urease test

Histology

Urea Breath Test

Serological test

HpSA

91.3(90) 100(100) 100(100) 81.8(71.4)

73.9(70) 100(100) 100(100) 60(45.4)

86.9(85) 100(100) 100(100) 75(62.5)

100(100) 11.1(20) 74.1(83.3) 100(100)

88(100) 14.2(NA) 78.5(80) 25(NA)

Results inpatients with gastric orduodenal ulcer inbrackets.NAnot available.

2711 INTRAGASTRIC PH AFFECTS THE RESULTS OF THE BC-UREA BREATH TEST (UBT) IN THE DIAGNOSIS OF H. PYWRI (HP). DrahaPantoflickova, Edward A. Lew,GianDorta,JosephR. Pisegna, Gordon V. Ohning, LarryV. Faller, DavidScott, John H. Walsh,GeorgeSachs,Andre L. Blum, CHUV, Lausanne, Switzerland; Yale Univ Med Ctr, New Haven, CT; CURF1Digestive Diseases Research Ctr, Los Angeles, CA. Background:UBT is based on the ability of intact Hp urease to hydrolyze urea. Highestactivationof Hp urease occurs at pH values between 2.5-6.2. Aims:To evaluate effect of intragastric pH and rate of gastric emptying of a test meal on the diagnostic accuracy of the UBT. Methods: Eleven Hp positive volunteers underwent UBT during intragastric titration to pH 7.0, 5.0, and 3.0. Eleven other Hp positive and eleven Hp negative subjects underwentthree UBTs with neutral Ensure (pH 7.0), acidifiedEnsure (pH 3.0) and applejuice (pH 3.0) as test meals. Gastric emptying of test meals was assessed in six of the Hp positive subjects by sodium acetate breath test (SABT). Results: All Hp negatives had [) values below the cut-off value ([)excess :s; 6.0 per mil) at all collection times. [)values listed in the table aregroup medians of Hp positives. Conclusions:!. The accuracy of the UBT increased with decreasing intragastric pH while gastric emptying of the test meal did not play an essential role. 2. Our results are best explained by acid dependent activation of intra-bacterial urease of Hp. 3. We suggest that administration of an acid test meal may improve the diagnostic performanceof UBT in subjects with PPI therapy. UBTs

Titration studies pH S' 3.0 5.0 7.0

17.2 12.5 6,8'

Test meal studies Test meal pH Ensure Acid.Ensure Apple juice

7.0 3.0 3.0

S'

SABTs Gastric emptying T""

16.8 214 16.7

51.3 56.1 30.0'

'Calculated at30min sample. 'p<0.05 pH 7vspH 3 and vspH 5.'p<0.05 apple juice vsEnsure and vsacidified Ensure. FN=false negative tests.

2712 VALIDATION OF A COMMERCIAL ELISA TO DETECT ANTICAGA ANTIBODIES IN CHll..DREN WITH H. PYWRIINFECTION. DulcieneMm Queiroz, Gifone A. Rocha, Andreia Me Rocha, Edilberto N. Mendes, AnfrisinaSt Carvalho,Ana Mmf Nogueira, Adriana Santos, Univ Fed de Minas Gerais, Belo Horizonte, Brazil. Although several studies employing ELISA or Western blot have been performed to establish CagA status in children, we are unaware of studies validating commercial ELISAs for CagA in this population. This would be important since there are evidences that ELISA has poor accuracy for the diagnosis of H.pylori infection in youngchildren. We studied 115children (51 boys, mean age 9.2 yrs., rangefrom 2 to 16 yrs.). Among them, 62 were H. pylori-positive (21 with duodenal ulcer) and 53 were H. pylori-negative. H. pylori positivity was confirmed by positive culture alone or both preformed ureasetest andhistology and H. pylori negativity whenall testswerenegative. Genomic DNAs from bacterial strains were PeR-amplified using 2 sets of oligonucleotide primers previously described by Kelly et al., (1994)and Peek et al. (1995). Sera wereassayedfor anti-CagA antibodies by usinga commercial ELISA (Helicobacter p120, CagA ELISA, Viva Diagnostika, Hurth, Germany) and the assaywas performed according to the recommendations of the manufacturer of the kit. The meanage as well as the numberof boys were significantly higher (p=0.OOO2 and p=0.02, respectively) in duodenal ulcer children than in H. pylori-positive children without duodenal ulcer and H. pylori-negative children. The cagA gene as detected by Peek's plus Kelly's primers or just Peek's or Kelly's was present more frequently in H. pylori strainsisolated fromduodenal ulcerchildren (20,95.2%)than in thoseisolated from H. pylori-positive children without duodenal ulcer (22, 57.9%) (p=O.OO6). The sensitivity, specificity and positive and negative predictive values of ELISAto detectanti-CagA antibodies were95.2%(82.6% - 99.2%), 88.6% (79.0% - 94.3%), 8!.6% (67.5% - 90.8%) and 97.2%(89.4% - 99.5%), respectively. There was no difference in the sensitivity of the test between H.

pylori-positive children with (100%) and without duodenal ulcer (90.9%) (p=0.48). Whenchildren werestratified by age, 2 to 6 yrs.,7 to 11yrs. and 12 to 16 yrs. antibodies anti-CagA were detected more frequently in the older ones (p=0.001) even when the duodenal ulcer children were excluded from the analysis (p=O.Ol). In conclusion, ELISA showed high accuracy for the determination of CagA statusin children and it can be useful in epidemiological studies.

2713 EVALUATION OF A NOVEL CONTINUOUS REAL TIME BC UREA BREATH ANALYZER. COMPARISON TO UREASE TEST AND HISTOPATHOLOGICAL DETECTION OF H. PYLORI. H. Shirin,O. Shevah, G. Kenet, Y. Wardi, M. Shahmurov, R. Bruck, Dept of Gastroenterology andPathology, The E Wolfson Medical Ctr,Holon, Israel; S F Moss, St Luke's-Roosesvelt Hosp Cent, Columbia Univ, New York, NY; Avni, Dept of Gastroeneterology, The E Wolfson Medical Ctr, Holon, Israel. It has been reported recently that the novel 13C breathanalyzer (Breath ID™, Oridion, Israel) is a reliable tool to detect H. pylori witha shortertestingtime period. Basedon Molecular Correlation Spectrometry, the BreathID continuously measures 13CO/2C02 concentrations from the patient's breath and establishes the 13CO/2C02 ratio. The aim of the current study was to determine the diagnostic valueof this new urea breathtest (UBT) in comparison to the endoscopic diagnostic criteria of histology and rapid urease test (RUT) defined as the gold standard. Methods: In the ongoing study, 74 consecutive patients (mean age 58.9 y) underwent upper gastrointestinal endoscopy for upperabdominal signsand symptoms. Atendoscopy, thepresence of H. pylori was examined by biopsies from the gastric antrum and body for rapid urease test and histologic assessment. After recovering from the endoscopic examination, the patients were analyzed for H. pylori by BreathID. Aftermeasuring the baseline, 75 mg C-ureaand4 g citricaciddissolved in 200 ml waterweregivento subjects. BreathID continuously sampled the subject's breathfor 20 min,viaa nasalcannulalikedevice. TheDOBwasmeasured and displayed on the BreathID screen in real time with a threshold of 5 DaB. Results: Overall agreement for BreathID vs RUT (CUTest, Temmler Pharma, Germany) was 92.6% and with biopsywas 91.3%.The gold standard identified H. pylori in 35 patients (47.3%). Complete agreement between the gold standard test and the BreathID was observed in 95.9% (71n4) of patients (34/35 positive and 37/39 negative). Disagreement was found in 3 patients. Two were negative by the goldstandard and positive by BreathID and one H. pylori positive patientby the gold standard, was found negative by the urea breathtest.The sensitivity of BreathID was97% andthe specificity 95%when compared with the gold standard. Conclusions: The BreathID urea breathtest for the detection of H. pylori has comparable sensitivity and specificity to the claims of the currently available UBT tests. Furthermore, BreathID has the advantages of ease of use with minimal medical stuff requirement, real time rapid results (20 min maximum) and low cost, which make the BreathID preferable to other UBT assays. 2714

DIFFERENT RANITIDINE DOSES HAVE EQUAL NEGATIVE EFFECT ON THE RESULTS OF UREA BREATH TEST (UBT). Daniela Tracci, Monica Pivari, Patrizia Zentilin, Pietro Dulbecco, Laura Tessieri, Giuliana Bisso, Maria R. Mele, Vincenzo Savarino, DI M I Gastroenterology Unit, Genoa, Italy. The negative impact of ranitidine on the results of UBT is morecontroversial thanthatof PPIs,but the conversion frompositive to negative UBT in patients takingthis H2-blocker has alsobeen reported. However, few dataare available addressing the relevance of the drug dosage in this negative effect. We assessed the rate of UBT conversion after 14 days of therapy with four different dosesofranitidine.Methods. We recruited 120dyspeptic patients (44 M, 76 F, meanage 54 yrs) with H. pyloriinfection ascertained on the basisof the concomitant positive resultsof Clo-test, histology and UBT.Nonepatient had undergone gastric surgery or had ulcer bleeding. Patients takingPPIs,H2 antagonists, bismuth compounds or antibiotics in the four weekspriorto entry were excluded. Our patients received randomly either ranitidine 300 mg hs, ranitidine 300mg mane,ranitidine 150mg bid or ranitidine 300mg bid for 14 days. Breathtestings were repeated on days 14 and 21 after the beginning of treatment. 13C-UBT was performed afterovernight fasting andbreathsamples were collected at baseline and 30' afteringestion of 75 mg l3C-urea dissolved in 150mL 0.033mollLcitricacid (Cortex Italia, Milan, Italy). 13C enrichment was measured by means of IRMS and a [) value higher than 5%0 was considered positive. Chi-square test was used for statistical analysis. Results. The 4 groups were well matched regarding the main demographic characteristics. Two patients dropped out of the study.The sensitivity of l3C-UBT was modified as in the following table. Our findings show that all the dosages of ranitidine we used are equally able to convert UBT results from positive to negative after 2 weeksof therapy, without statistical difference among them. Conclusions. We confirm that ranitidine affects negatively UBT results, in analogy with the effects of PPls on this test. This negative impact is not dependent on the drug dose. In somecasesthe UBT conversion lastsfor more than 7 days and patients shouldremainoff for 2 weeks prior to testing. Ranitidine dosages

UBT negonday14

UBT neg onday21

Ran 300 mghs Ran 300 mgmane Ran 150 mgbid Ran 300 mgbid

3/30 (10%) 4/30(13%) 5/28 (17%) 3/30 (10%)

0/30 3/30 (10%) 2/28 (7%) 1/30(3%)