Detection of Helicobacter pylori Gastritis by PCR Correlation With Inflammation Scores and Immunohistochemical and CLOtest Findings

Anatomic Pathology / HELICOBACTER PYLORI GASTRITIS Detection of Helicobacter pylori Gastritis by PCR Correlation With Inflammation Scores and Immunohistochemical and CLOtest Findings Jason Weiss, DO, 1,2 Tat-Kin Tsang, MD, 2,3 Xiangwen Meng, MD, PhD, 4 H. Zhang, MD, PhD, 4 Elizabeth Kilner, 1 Edward Wang, PhD, 5 and William Watkin, MD 1,2 Key Words: Helicobacter pylori; Polymerase chain reaction; Gastritis; Histology; Sydney system; Endoscopic biopsies; Immunohistochemistry DOI: 1.139/APMPEP54G7PN958G Abstract We developed a polymerase chain reaction (PCR) assay to detect Helicobacter pylori in gastric and/or gastroesophageal biopsy specimens of adults with dyspepsia, compared the method with immunohistochemical analysis and CLOtest (Ballard Medical Products, Draper, UT), and correlated the results of each test with the histologic features of infection. H pylori was identified in 36 (6%) of 6 patients irrespective of biopsy site and testing method. In the gastric biopsy specimens, PCR detected H pylori in 29 (52%) of 56 cases, including 11 (1%) of 11 immunohistochemically and/or CLOtest-positive cases. PCR-positive gastric biopsy specimens correlated with a higher average cumulative inflammatory score compared with PCR-negative specimens (P =.1). In gastroesophageal biopsy specimens, PCR detected H pylori in 15 (34%) of 44 cases, including 1 (2%) of 5 immunohistochemically positive specimens. PCRpositive gastroesophageal junction biopsy specimens did not correlate with a higher average cumulative inflammatory score. Overall, PCR detected an additional 23 cases negative by immunohistochemical analysis and/or CLOtest. This PCR assay identified a significant number of H pylori infections that would not be detected by immunohistochemical analysis and/or CLOtest. Helicobacter pylori has a significant pathogenic role in gastritis and peptic ulcer disease, as well as in gastric adenocarcinoma, mucosa-associated lymphoid tissue lymphoma, and, possibly, cardiovascular disease. Consequently, the World Health Organization classified this important bacterial species as a class I carcinogen. 1,2 The prevalence of H pylori infection may approach 1% in some developing countries. 3 In the United States, the prevalence is estimated at 3% to 4%, but it is much higher in African Americans and Hispanics. 4 Numerous techniques have been developed to identify H pylori infection, but none is definitive. Less invasive techniques such as serology, urea breath tests, and stool antigen testing have all been used clinically, 5-7 but esophagogastroduodenoscopy (EGD) examination with biopsy is often required for diagnosis. Pathologists are routinely expected to diagnose H pylori infection when interpreting EGD biopsy specimens. Often, the characteristic histologic pattern of chronic active gastritis prompts the pathologist to scrutinize the biopsy specimen for the typical helical morphologic profile of H pylori. Although the bacteria are often visible on routine H&E-stained slides, most laboratories use immunohistochemical analysis and/or special stains (eg, Giemsa) to highlight the organism, 6,8 particularly in patients who have received prior therapy when organisms may be low in number or have altered morphologic features. 9,1 With the advent of molecular techniques, reliable and sensitive methods for the detection of bacterial gene sequences in many different clinical specimens have become possible. Polymerase chain reaction (PCR) assays to diagnose H pylori infection in stool samples, 11 EGD biopsy specimens, 2,1 saliva, 12,13 and dental plaque 12 have all been described. Am J Clin Pathol 28;129:89-96 89 89 DOI: 1.139/APMPEP54G7PN958G 89

Weiss et al / HELICOBACTER PYLORI GASTRITIS In the present study we compared the diagnostic efficacy of a unique multiplex PCR assay for the detection of H pylori with immunohistochemical detection and the CLOtest (Ballard Medical Products, Draper, UT) in gastric and/or gastroesophageal junction biopsy specimens from patients with symptoms of dyspepsia. We evaluated the results of each test with scores of activity and chronic inflammation. Materials and Methods Patients EGD was performed on 6 patients (3 males, 3 females) with symptoms of dyspepsia from June 23 to December 23. For 4 patients, paired biopsy specimens of the gastric antrum/body (gastric) and gastroesophageal junction were available; for 16 patients, only gastric biopsy specimens were available; and for 4 patients, only gastroesophageal junction biopsy specimens were submitted for routine histologic examination. An additional biopsy specimen was taken from each site and submitted for PCR analysis. In the 56 patients who had gastric biopsy specimens, a separate biopsy specimen of the antral mucosa was taken and submitted for CLOtest analysis. The specimens were obtained at Evanston Northwestern Healthcare s Glenbrook Hospital (Glenview, IL) and Evanston Hospital (Evanston, IL). Informed consent was obtained from each patient, and the study was reviewed and approved by the Evanston Northwestern Health Care Institutional Review Board. Histologic Evaluation All biopsy specimens submitted for histopathologic analysis were formalin-fixed (1%) and paraffin-embedded. Two H&E-stained slides were prepared from each paraffinembedded biopsy specimen. The degree of polymorphonuclear neutrophil activity and chronic (mononuclear) inflammatory infiltrate were graded by using the updated Sydney system visual analog scale for normal, mild, moderate, and marked inflammation with a scoring system from to 3, respectively. 8 All biopsy specimens were evaluated for intestinal metaplasia. Every gastroesophageal junction biopsy specimen contained some gastric cardia mucosa on histologic evaluation. Immunohistochemical Staining for H pylori Immunohistochemical staining was performed on 4-µm sections of all formalin-fixed, paraffin embedded biopsy specimens. Antigen retrieval was performed with Protease 1 (Ventana, Tucson, AZ) according to the manufacturer s instructions. A rabbit polyclonal antibody against H pylori (clone CH-2/429, dilution 1:2; Novocastra, Newcastle upon Tyne, England) was used, and staining was performed on an automated stainer (Ventana). A slide prepared from known H pylori infected tissue biopsy specimens was used as a positive control sample. Negative control samples were prepared by omitting the primary antibody. The results were screened manually. The density of H pylori stained organisms was graded by using the updated Sydney system visual analog scale for normal, mild, moderate, and marked with a scoring system of to 3, respectively. 8 CLOtest The CLOtest rapid urease test was performed on all gastric antrum biopsy specimens according to the manufacturer s instructions. A definite magenta color was required to read the test as positive. The test results were interpreted within 24 hours. PCR Analysis DNA extraction was performed according to published methods with small modifications. 14 Primers specific for the H pylori urease A gene, 15,16 26-kd protein antigen, 17 hpaa gene, 18.86-kb DNA fragment, 19 and DNA sequences of 16S ribosomal RNA 17 were used as amplification targets as previously described. 2,21 For each locus, we selected 1 forward primer as the common primer (FC) and 2 reverse primers (R1 and R2), with R2 located inside the amplifying region of R1. Therefore, 2 DNA fragments, one internal to the other, are expected to be amplified in each locus Figure 1A. The primers were combined into 2 groups: one group consisted of all 5 FC primers with 5 R1 primers (FC-R1), and the second group consisted of all 5 FC primers with 5 R2 primers (FC-R2). Thus, a total of 1 DNA fragments could be amplified, in 2 tubes, each containing 5 amplicons. The DNA extract was added to 2 µl of a reaction mixture containing 1 mmol/l of tris(hydroxymethyl)aminomethane hydrochloride (ph 8.3), 5 mmol/l of potassium chloride, 1.6 mmol/l of magnesium chloride,.1% (wt/vol) gelatin,.3 mmol/l of each deoxynucleotide, and.5 µmol/l of each oligonucleotide primer mixture. Immolase DNA polymerase (4. U; Bioline, London, England) was used. The PCR was performed with a PCR thermal cycler (Mastercycler, Eppendorf Scientific, Waterbury, NY), and the PCR conditions were as follows: an initial denaturation of target DNA at 94 C for 5 minutes, followed by 32 to 41 cycles of 94 C for 3 seconds, 56 C for 5 seconds, and 72 C for 1 minute. The final extension step was at 72 C for 1 minutes. The PCR products were analyzed by electrophoresis of a 1-µL aliquot on a 2.% agarose gel, stained with ethidium bromide and visualized after excitation under UV light Figure 1B. Because of the high diversity of DNA sequences in H pylori, in this study, we defined a case as positive for H pylori if 5 of 1 fragments or both DNA fragments from the 9 Am J Clin Pathol 28;129:89-96 9 DOI: 1.139/APMPEP54G7PN958G

Anatomic Pathology / ORIGINAL ARTICLE same locus were amplified. Positive and negative (water blank) control samples were assayed in each run. The specificity of the PCR assay was evaluated by testing with 3 H pylori strains obtained from American Type Culture Collection (Manassas, VA): ATCC4354, ATCC7392 (26695), and ATCC7824 (J99), and 11 common bacterial species (1) Escherichia coli, (2) Enterobacter aerogenes, (3) Enterobacter cloacae, (4) Enterococcus species, (5) viridians group of Streptococcus, (6) Pseudomonas aeruginosa, (7) Serratia species, (8) Klebsiella pneumoniae, (9) methicillinresistant Staphylococcus aureus, (1) Lactobacillus species, and (11) Citrobacter species, which were isolated from clinical samples at Evanston Northwestern Healthcare, Evanston Hospital were selected as templates to assay. The PCR was performed in the conditions described along with H pylori J99, which was used as the positive control. Statistical Analysis Statistical analysis was performed by using the nonparametric Mann-Whitney U test. P values of less than.5 were considered statistically significant. Results H pylori was identified in 36 (6%) of 6 patients when all testing methods and biopsy sites were combined Table 1. In gastric biopsy specimens, H pylori was detected in 29 (52%) of 56 cases by PCR, 11 (2%) by CLOtest, and 11 (2%) by immunohistochemical analysis (Table 1) Table 2. All gastric biopsy specimens positive by CLOtest were positive by immunohistochemical analysis and vice versa. All gastric biopsy specimens positive by immunohistochemical analysis and the CLOtest were positive by PCR (Table 1) Figure 2A. Nine gastric biopsy specimens demonstrated intestinal metaplasia on histologic examination, with 4 showing H pylori gastritis. In gastroesophageal junction biopsy specimens, H pylori was detected in 15 (34%) of 44 cases by PCR and 5 (11%) of 44 cases by immunohistochemical analysis (Table 1) Figure 2B. Only 1 (2%) of 5 immunohistochemically positive gastroesophageal junction biopsy specimens was positive by PCR. Four of the gastroesophageal junction biopsy specimens demonstrated intestinal metaplasia on histologic examination, with only 1 showing H pylori gastritis (Table 1). Two patients were clinically suspected to have Barrett esophagus on EGD; only 1 was confirmed by histologic examination. The results of the 3 specified tests for H pylori infection were compared with the activity and chronic inflammation scores (Table 2). In the gastric biopsy specimens, the average activity and chronic inflammatory scores were significantly greater in PCR-positive than in PCR-negative biopsy A B FC FC-R1 PCR FC-R2 86 kb DNA fragment Urease A gene 16S ribosomal RNA 26 kd protein hpaa gene specimens (P =.2 and P =.4, respectively). Similarly, both scores were higher in immunohistochemically/clotestpositive biopsy specimens than in the negative specimens (P <.1 for each). In the gastroesophageal junction biopsy specimens, the results were conflicting (Table 2). Although the average scores for activity and chronic inflammation were higher in the immunohistochemically positive cases than in immunohistochemically negative cases (P <.1 and P =.3, respectively), there was no significant difference between the activity scores in PCR-positive and PCR-negative biopsy specimens (P =.81). The average scores for chronic inflammation, while significant, were lower in the PCR-positive cases than the PCR-negative cases (P =.2). The results of the cumulative inflammation scores did not differ from those looking at activity and chronic inflammation individually Figure 3. In gastric biopsy specimens, the mean cumulative score of inflammation was significantly greater in R2 R1 DNA template FC-R1 FC-R2 Figure 1 Primer design. A, Diagram of primers designed for each locus. FC, the forward primer, is the common primer; R1 and R2 are the pair of reverse primers. The amplicons FC- R1 and FC-R2 are amplified from each locus. B, Five pairs of DNA bands amplified from the 5 targeted loci specific for Helicobacter pylori. Am J Clin Pathol 28;129:89-96 91 91 DOI: 1.139/APMPEP54G7PN958G 91

Weiss et al / HELICOBACTER PYLORI GASTRITIS Table 1 Summary of Helicobacter pylori Testing Results and Presence of Intestinal Metaplasia Gastric Gastroesophageal Junction Case No./ Immunohistochemical Immunohistochemical Sex/Age (y) Analysis CLOtest PCR IM Analysis PCR IM 1/F/44 + + + + + + 2/F/67 + + + + 3/M/59 + + + + 4/M/82 + + + 5/F/25 + + + 6/M/78 + + 7/M/7 + + + 8/F/55 + + 9/M/72 + + 1/F/41 + + 11/F/3 + + 12/M/53 + + 13/F/64 + + 14/F/18 + + + 15/F/22 + 16/M/77 + + 17/M/53 + 18/F/25 + 19/F/25 + 2/F/76 + 21/F/78 + 22/M/87 + 23/F/53 + 24/M/7 + 25/F/73 26/M/74 27/M/66 + 28/M/51 + 29/M/65 3/F/44 31/F/44 32/F/8 + 33/M/64 34/M/67 + 35/F/72 36/M/52 37/F/37 38/F/46 39/M/57 4/M/2 + 41/M/57 + + + ND ND ND 42/M/72 + + + ND ND ND 43/F/68 + + + ND ND ND 44/F/82 + + + ND ND ND 45/M/61 + + + ND ND ND 46/F/94 + + + + ND ND ND 47/F/52 + ND ND ND 48/F/58 + ND ND ND 49/F/84 + ND ND ND 5/F/49 ND ND ND 51/M/78 ND ND ND 52/M/61 ND ND ND 53/F/77 ND ND ND 54/F/19 ND ND ND 55/M/71 + ND ND ND 56/M/71 + ND ND ND 57/M/47 ND ND ND ND + * 58/M/55 ND ND ND ND + 59/M/57 ND ND ND ND + 6/M/86 ND ND ND ND + * CLOtest, rapid urease test (Ballard Medical Products, Draper, UT); IM, intestinal metaplasia; ND, not done; PCR, polymerase chain reaction; +, positive or present;, negative or absent. * Suspected Barrett esophagus by esophagogastroduodenoscopy. 92 Am J Clin Pathol 28;129:89-96 92 DOI: 1.139/APMPEP54G7PN958G

Anatomic Pathology / ORIGINAL ARTICLE Table 2 Summary of the Average Helicobacter pylori and Inflammatory Scores for Each Biopsy Site and Testing Method No (%) H pylori Neutrophil Mononuclear Testing Method and Site/Results of Tests Score P Score P Score P Immunohistochemical analysis of biopsy specimen Gastroesophageal junction (n = 44) <.1 <.1.3 Positive 5 (11) 2.4 1.6 2.6 Negative 39 (89).2 1.4 Gastric (n = 56) <.1 <.1 <.1 Positive 11 (2) 2.5 1.8 2.7 Negative 45 (8).2.7 PCR of biopsy specimens Gastroesophageal junction (n = 44).47.81.2 Positive 15 (34).1.3 1 Negative 29 (66).3.4 1.9 Gastric (n = 56) <.1.2.4 Positive 29 (52) 1.9 1.5 Negative 27 (48).1.6 CLOtest (n = 56) <.1 <.1 <.1 Positive 11 (2) 2.5 1.8 2.7 Negative 45 (8).2.7 CLOtest, rapid urease test (Ballard Medical Products, Draper, UT); PCR, polymerase chain reaction. cases with H pylori detected by PCR than in negative cases (P =.1). Interestingly, the cumulative inflammation scores were higher, ie, 3 or more, in patients whose gastric biopsy specimens were H pylori positive by immunohistochemical analysis/clotest as compared with patients who had H pylori detected only by PCR. In the gastroesophageal junction biopsy specimens, the mean cumulative score of inflammation was significantly lower in cases with H pylori detected by PCR than in the negative cases (P =.12). The cumulative inflammation scores were higher ( 2) in immunohistochemically positive cases than in PCR-positive cases; however, a majority of these cases were PCR-negative. A CLOTest 11 PCR 18 Discussion H pylori is a major human pathogen for which accurate detection is needed for proper patient management. No one method has sufficient sensitivity and specificity to be considered a gold standard. We developed a unique PCR assay to detect H pylori in endoscopic biopsy specimens, compared the method with standard techniques of immunohistochemical analysis and CLOtest, and correlated the results of each test with the histologic features of H pylori infection. The ability of PCR to detect H pylori in EGD biopsy specimens has been previously demonstrated. 16,18,22,23 However, few investigators have correlated the results with corresponding degrees of inflammation. 2 B IHC PCR 4 1 14 IHC Figure 2 Distribution of Helicobacter pylori positive biopsy specimens by any method according to the testing method and biopsy site. A, Gastric. B, Gastroesophageal junction. The CLOtest was performed only on antral (gastric) biopsy specimens. CLOtest, rapid urease test (Ballard Medical Products, Draper, UT); IHC, immunohistochemical analysis; PCR, polymerase chain reaction. Am J Clin Pathol 28;129:89-96 93 93 DOI: 1.139/APMPEP54G7PN958G 93

Weiss et al / HELICOBACTER PYLORI GASTRITIS A B 6 6 5 5 Cumulative Inflammation Score 4 3 2 1 PCR (n = 27) PCR+ (n = 29) In the gastric biopsy specimens, PCR identified H pylori infection in 1% of patients who had positive CLOtest or immunohistochemical results (Table 1 and Figure 2A). Consistent with the expected increased sensitivity, PCR detected H pylori in an additional 18 patients with negative immunohistochemical and CLOtest results. The majority of these patients had low cumulative inflammation scores and may not have been suspected after routine histologic examinations of specimens to have had H pylori gastritis. Overall, biopsy specimens with a cumulative inflammation score of 4 or more predicted the presence of H pylori by at least 1 testing method, irrespective of biopsy site (Figure 3) Image 1. In gastroesophageal junction biopsy specimens, PCR did not correlate as well with the paired immunohistochemically stained biopsy specimens, with only 2% (1/5) of the immunohistochemically positive biopsy specimens demonstrating H pylori by PCR (Table 1 and Figure 2B). All 4 of the Cumulative Inflammation Score 4 3 2 1 PCR (n = 29) PCR+ (n = 15) Figure 3 Cumulative inflammation scores in cases positive or negative for Helicobacter pylori results by polymerase chain reaction (PCR). A, Results for gastric biopsy specimens. Horizontal bars indicate mean value (P =.1). B, Results for gastroesophageal junction biopsy specimens. Horizontal bars indicate mean value (P =.12). The total score for inflammation is equal to the sum of chronic inflammation (score, -3) and activity (score, -3). White circles represent cases positive for H pylori by immunohistochemical analysis. All gastric biopsy specimens positive by immunohistochemical analysis were also positive by the CLOtest. PCR-negative, immunohistochemically positive cases were positive by PCR on the concurrent gastric sample (Table 1); however, the results from 2 of these cases were not reported because there were no paired gastric biopsy specimens taken for histologic evaluation. This discrepancy between the immunohistochemical and PCR results might be explained by the fact that the distribution of H pylori is more prominent at or near the antrum, with a patchy distribution throughout the remaining gastric mucosa. In addition, the separate gastroesophageal junction biopsy specimens submitted for PCR analysis may have consisted of only squamous esophageal mucosa, which does not provide the necessary environment for colonization. However, all gastroesophageal junction biopsy specimens submitted for histologic evaluation had at least a portion of gastric cardia. PCR detected H pylori in an additional 14 cases that were negative by immunohistochemical analysis. A majority of these cases had low cumulative inflammation scores. 94 Am J Clin Pathol 28;129:89-96 94 DOI: 1.139/APMPEP54G7PN958G

Anatomic Pathology / ORIGINAL ARTICLE A B When organisms were identified by immunohistochemical analysis, irrespective of biopsy site, they had a cumulative inflammatory score of 2 or more (Figure 3). This finding is similar to those of other studies that showed H pylori to be associated with a cumulative inflammation score of 2 or more.2,24,25 Cases detected only by PCR had lower average acute and chronic inflammation scores. By using a different PCR method, Zsikla et al2 demonstrated similar results. It is generally accepted that the probability of identifying H pylori by histologic examination becomes greater as the degree of inflammation, whether acute or chronic, increases. In cases with low numbers of organisms and minimal accompanying inflammation, histologic examination may not be sufficient. PCR methods would increase the likelihood of H pylori detection. Because PCR can amplify millions of copies from just a few target sequences, the possibility of contamination resulting in false-positive results has been suspected.2,26 As with any sensitive laboratory technique, meticulous attention must be applied to ensure no cross-contamination, including during the preanalytic and postanalytic phases. Our study demonstrates that this PCR assay identifies a significant number of H pylori infections that would not be detected with conventional immunohistochemical analysis and/or CLOtest. Antral and gastric body biopsy specimens are better for predicting the correct H pylori status than gastroesophageal junction biopsy specimens. Future studies examining the usefulness of this PCR on formalin-fixed, paraffin-embedded tissue samples are necessary, which would provide greater clinical benefit because fresh tissue samples may not be available for cases that are immunohistochemically negative and have a suspicious inflammatory background. In addition, the clinical significance of identifying H pylori by PCR alone, particularly in cases with low levels of inflammation, remains to be determined, and prospective studies are in progress. From the Departments of 1Pathology and Laboratory Medicine and 3Gastroenterology, 4Research Institute, and 5CORE, Evanston Northwestern Healthcare, Evanston, IL; and 2Northwestern University Feinberg School of Medicine, Chicago. Address reprint requests to Dr Watkin: Dept of Pathology and Laboratory Medicine, Evanston Northwestern Healthcare, 265 Ridge Ave, Evanston, IL 621. References 1. Das S, Suarez G, Beswick EJ, et al. Expression of B7-H1 on gastric epithelial cells: its potential role in regulating T cells during Helicobacter pylori infection. J Immunol. 26;176:339. 2. Zsikla V, Hailemariam S, Baumann M, et al. Increased rate of Helicobacter pylori infection detected by PCR in biopsies with chronic gastritis. Am J Surg Pathol. 26;3:242-248. 3. Asrat D, Nilsson I, Mengistu Y, et al. Prevalence of Helicobacter pylori infection among adult dyspeptic patients in Ethiopia. Ann Trop Med Parasitol. 24;98:181-189. 4. Go MF. Natural history and epidemiology of Helicobacter pylori infection. Aliment Pharmacol Ther. 22;16(suppl 1):3-15. 5. Hooton C, Keohane J, Clair J, et al. Comparison of three stool antigen assays with the 13C-urea breath test for the primary diagnosis of Helicobacter pylori infection and monitoring treatment outcome. Eur J Gastroenterol Hepatol. 26;18:595599. 6. Dunn BE, Cohen H, Blaser MJ. Helicobacter pylori. Clin Microbiol Rev. 1997;1:72-741. Am J Clin Pathol 28;129:89-96 95 DOI: 1.139/APMPEP54G7PN958G 95 95 Image 1 Representative examples of biopsy specimens with cumulative inflammation scores of 4 (neutrophil score, 1; mononuclear score, 3) from the gastroesophageal junction (A, H&E, 1) and the gastric antrum (B, H&E, 2).

Weiss et al / HELICOBACTER PYLORI GASTRITIS 7. Raguza D, Granato CF, Kawakami E. Evaluation of the stool antigen test for Helicobacter pylori in children and adolescents. Dig Dis Sci. 25;5:453-457. 8. Dixon MF, Genta RM, Yardley JH, et al. Classification and grading of gastritis: the updated Sydney System. International Workshop on the Histopathology of Gastritis, Houston 1994. Am J Surg Pathol. 1996;2:1161-1181. 9. Chan WY, Hui PK, Leung KM, et al. Coccoid forms of Helicobacter pylori in the human stomach. Am J Clin Pathol. 1994;12:53-57. 1. Ashton-Key M, Diss TC, Isaacson PG. Detection of Helicobacter pylori in gastric biopsy and resection specimens. J Clin Pathol. 1996;49:17-111. 11. Kabir S. Detection of Helicobacter pylori in faeces by culture, PCR and enzyme immunoassay. J Med Microbiol. 21;5:121-129. 12. Goosen C, Theron J, Ntsala M, et al. Evaluation of a novel heminested PCR assay based on the phosphoglucosamine mutase gene for detection of Helicobacter pylori in saliva and dental plaque. J Clin Microbiol. 22;4:25-29. 13. Ferguson DA Jr, Li C, Patel NR, et al. Isolation of Helicobacter pylori from saliva. J Clin Microbiol. 1993;31:282-284. 14. Sambrook J, Russell DW. Molecular Cloning: A Laboratory Manual. 3rd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 21. 15. Lin TT, Yeh CT, Wu CS, et al. Detection and partial sequence analysis of Helicobacter pylori DNA in the bile samples. Dig Dis Sci. 1995;4:2214-2219. 16. Pacheco N, Mago V, Gomez I, et al. Comparison of PCR and common clinical tests for the diagnosis of H pylori in dyspeptic patients. Diagn Microbiol Infect Dis. 21;39:27-21. 17. Nilsson HO, Taneera J, Castedal M, et al. Identification of Helicobacter pylori and other Helicobacter species by PCR, hybridization, and partial DNA sequencing in human liver samples from patients with primary sclerosing cholangitis or primary biliary cirrhosis. J Clin Microbiol. 2;38:172-176. 18. He Q, Wang JP, Osato M, et al. Real-time quantitative PCR for detection of Helicobacter pylori. J Clin Microbiol. 22;4:372-3728. 19. Aydin O, Egilmez R, Karabacak T, et al. Interobserver variation in histopathological assessment of Helicobacter pylori gastritis. World J Gastroenterol. 23;9:2232-2235. 2. Meng X, Zhang H, Tsang T. Comparison of a novel multiplex PCR and histology method for the diagnosis of Helicobacter pylori [abstract]. Gastroenterology. 24;126:A187. 21. Meng X, Zhang H, Tsang T, et al. Detection of Helicobacter pylori by a novel multiplex PCR assay [abstract]. Gastroenterology. 23;124:A175. 22. Lascols C, Lamarque D, Costa JM, et al. Fast and accurate quantitative detection of Helicobacter pylori and identification of clarithromycin resistance mutations in H pylori isolates from gastric biopsy specimens by real-time PCR. J Clin Microbiol. 23;41:4573-4577. 23. Genta RM, Graham DY. Comparison of biopsy sites for the histopathologic diagnosis of Helicobacter pylori: a topographic study of H pylori density and distribution. Gastrointest Endosc. 1994;4:342-345. 24. Gallo N, Zambon CF, Navaglia F, et al. Helicobacter pylori infection in children and adults: a single pathogen but a different pathology. Helicobacter. 23;8:21-28. 25. Yamamura F, Yoshikawa N, Akita Y, et al. Relationship between Helicobacter pylori infection and histologic features of gastritis in biopsy specimens in gastroduodenal diseases, including evaluation of diagnosis by polymerase chain reaction assay. J Gastroenterol. 1999;34:461-466. 26. Furuta T, Kaneko E, Suzuki M, et al. Quantitative study of Helicobacter pylori in gastric mucus by competitive PCR using synthetic DNA fragments. J Clin Microbiol. 1996;34:2421-2425. 96 Am J Clin Pathol 28;129:89-96 96 DOI: 1.139/APMPEP54G7PN958G