点击显示 收起
Dr. von Haunersches Kinderspital, Ludwig-Maximilians-University Munich, Munich, Germany
Children's Hospital, University of Vienna, Vienna, Austria
ABSTRACT
A new rapid one-step immunochromatographic test using monoclonal antibodies for detection of Helicobacter pylori antigen in stool in children was evaluated on coded stool samples from 159 children (mean age, 9.7 ± 5.0 years; 118 from Munich, 41 from Vienna): 86 children were H. pylori infected defined by positive culture and/or 2 other positive tests ([13C]urea breath test, histology, rapid urease test), and 73 children showed concordant negative results. Seventy-nine patients (12.1 ± 3.8 years; 42 from Munich; 37 from Vienna) were tested 6 to 8 weeks after anti-Helicobacter pylori therapy with urea breath test and stool test. In Munich, all 160 tests (118 pre- and 42 posttreatment) were independently read by two observers. Equivocal results were excluded for calculation of sensitivity and specificity but were considered as false to assess accuracy. The two observers in Munich agreed in 63 out of 65 positive and 89 out of 95 negative results, while eight times (5.0%) they judged the test as equivocal. Pretreatment and posttreatment results for sensitivity were 88.1% (79.2 to 94.1) and 88.9% (51.8 to 99.7), specificity 88.1% (77.8 to 94.1) and 93.9% (85.2 to 98.3), and accuracy 83.5% and 81.5%, respectively. We conclude that the new monoclonal immunochromatographic quick test shows a good interobserver agreement, but equivocal results occur in 5%. Performance is comparable before and after therapy. The test may become a good alternative in children in settings where a [13C]urea breath test or a reliable enzyme immunoassay stool test are not available.
INTRODUCTION
Helicobacter pylori infection is the major cause of peptic ulcer disease and chronic gastritis and is almost always acquired in early childhood. For the diagnosis of H. pylori infection, gastrointestinal endoscopy with concordant results of biopsy based methods (culture, histology, and rapid urease test) is considered to be the "gold standard." Several noninvasive methods for the detection of H. pylori infection are available. In children, tests must be reliable in all age groups (8). Most serological tests show a low sensitivity in young children (5, 7, 8, 19). The [13C]urea breath test (UBT) gives an excellent performance, in both, adults and children, but specificity decreases in very young children, and collection of exhaled air is difficult in this age group (2, 6). Recently an enzyme immunoassay (EIA) based on polyclonal antibodies was developed for detection of H. pylori antigen in stool. Results of different studies showed conflicting results indicating large test to test variability, both pre- and posttreatment in children and adults (4, 10, 13, 16-18). In contrast, stool EIA based on monoclonal antibodies showed excellent results, with very high sensitivity and specificity (9, 13). The Immunocard STAT! HpSA (Meridian Bioscience Europe) is a novel one-step immunochromatographic quick test based on detection of monoclonal antibodies to H. pylori antigen in feces. The aim of this study was to evaluate this test for detection of H. pylori infection in a large number of children before and after H. pylori eradication therapy in comparison to a well-defined H. pylori status established by the results of invasive diagnostic techniques and the UBT.
MATERIALS AND METHODS
Patients. For the evaluation prior to first therapy, 159 children (80 girls, 79 boys, mean age 9.7 ± 5.0 years) were enrolled in two pediatric hospitals (Munich, n = 118; Vienna, n = 41). All children underwent upper gastrointestinal endoscopy because of abdominal symptoms suggestive of organic disease. None of the children had been treated for H. pylori infection in the past. Children were excluded if they took antibiotic or acid-suppressive drugs (proton pump inhibitors, H2-receptor antagonists, antacids, bismuth preparations) within 4 weeks prior to testing, if they had diarrhea, or if the H. pylori status was not clearly defined as described below. In our centers, about 1 out of 9 children undergoing upper endoscopy is H. pylori infected. To have a meaningful number of H. pylori-infected patients in relation to the total number of tests performed, we included all H. pylori-positive children who fulfilled the inclusion criteria and selected randomly an equal number of noninfected children from each center.
For the posttreatment evaluation, 79 patients (mean age 12.1 ± 3.8 years; Munich, n = 42; Vienna, n = 37) were tested 6 to 8 weeks after anti-H. pylori therapy.
The study was approved by the local ethics committees, and informed consent was obtained by the parents and children, if appropriate.
Definition of H. pylori status. During upper endoscopy, biopsies from the gastric antrum and corpus were taken from every child for histological examination, formalin-fixed, stained with hematoxylin-eosin and modified Giemsa, and viewed for the presence of H. pylori by local pathologists who were blinded for the results of the other tests performed. For the rapid urease test (n = 157) and for bacterial culture (n = 153), one antral specimen each was obtained. Biopsies for culture were transported to the local microbiological laboratory in transport media and were processed within 4 h. The UBT (n = 150) was performed as previously described (6). Briefly, after a fasting period of at least 4 h, a baseline breath sample was obtained using a breath bag or, in very young children, a face mask. The children drank 150 ml of apple juice (pH 3.4); thereafter, they received 20 ml of juice containing 75 mg 13C-labeled urea and then drank 30 ml of pure apple juice to flush the tracer from the mouth. Children <3 years old ingested only a total of 80 to 100 ml apple juice. Another breath samples was obtained 30 min after tracer application. The expired air was transferred into 10-ml vacutainers. The breath samples were analyzed by isotope ratio mass spectrometry. The test was considered as positive when the over baseline value after 30 min was 5.
Pretreatment H. pylori status was defined as positive if culture and/or at least two of the other applied methods (histology, rapid urease test, UBT) gave positive results. A negative H. pylori status was considered if all performed tests gave concordant negative results. For ethical reasons and in accordance with the consensus statement of the pediatric task force group on H. pylori infection, endoscopy was not performed in order to monitor the success of H. pylori therapy (1). Instead, posttreatment UBT was performed 6 to 8 weeks after therapy and the results were compared with the results of the stool test.
Stool antigen test. Parents of children scheduled for endoscopy were asked to bring a stool sample of their child at the time of the procedure or to send it in by mail within 3 days after endoscopy before any therapy was initiated. All samples were stored at –20°C until analyzed.
The stool antigen test is a rapid immunoassay, based on a lateral flow chromatography technique using monoclonal antibodies for the qualitative detection of H. pylori antigens in human stool. The sample diluents prevents nonspecific reactions. The strip in the cassette contains, above the sample port, red latex particles bound to the H. pylori monoclonal antibody and blue latex particles bound to a control protein. In the presence of H. pylori antigens, the red particles will be bound to the antigen during the flow and captured by a line of the same monoclonal antibodies, while the blue particles will be bound by a line of control antibody.
The tests were performed at the two local sites according to manufacturer's recommendations. In brief, a small portion (5 to 6 mm diameter) of stool specimen was transferred into the sample diluent using an applicator stick and mixed properly for 15 to 20 s. Thereafter, the tip of the vial was broken off and 4 drops were dispensed into the round window at the lower end of the device. Results were read after exactly 5 min. A sample was considered as negative, when only one blue control band appeared, and as positive, when a distinguishable pink-red line was seen in addition to the blue control band. Any line or color change appearing after 5 min had no diagnostic value. If a faint pink-red line occurred, the result was called equivocal.
All tests were performed on coded stool samples, which means that the persons performing and reading the test were blinded for the H. pylori status of the children. In Munich, all 160 tests were independently read by two investigators (D. Antos and N. Konstantopoulos).
Statistics. The statistical analysis was performed using SPSS (Statistical Package for Social Sciences version 11.5; SPSS Inc., Chicago, Illinois). Sensitivity and specificity with confidence intervals, and the likelihood ratio for a positive and negative stool test result were calculated against the defined H. pylori status as gold standard (3). Equivocal results were excluded for calculation of sensitivity, specificity, and positive and negative likelihood ratio, but they were considered as false results for calculation of accuracy. The Chi-square test was used to evaluate differences between the results pre- and posttreatment and between the three different age groups.
RESULTS
According to the predefined criteria, 86 (54%) of the 159 patients were H. pylori-infected and 73 were noninfected. Culture was positive in 80 children, while the remaining six infected children had positive results for histology and rapid urease test. Of the 86 children with a positive H. pylori status, histology and rapid urease test were positive in 85 children each. The UBT was performed in 86 out of 86 H. pylori-positive children and gave a positive result in all (sensitivity 100%). The proportions of infected children in the two centers Munich and Vienna were 53% and 54%, respectively. When the children were divided into three age groups, it was obvious that there were fewer infected children in the youngest age group (Table 1).
Of the 79 patients tested 6 to 8 weeks posttreatment, 9 (11.4%) showed a positive UBT result and were considered as treatment failures, while the remaining 70 children had clearly negative test results.
The two independent observers in Munich agreed in 63 of 65 positive (97.0%) and 89 of 95 negative results (93.6%). Eight times (5.0%), the reading was judged as equivocal by the two investigators due to the weak expression of the pink-red line. In Vienna, 4 of 78 results (5.1%) were judged as equivocal by the investigator. The total number of equivocal results was 12 of 238 (5%); 8 of 12 occurred in the pretreatment and the remaining in the posttreatment samples.
Table 2 gives the results for sensitivity, specificity, accuracy, and likelihood ratios for the pretreatment and the posttreatment. After excluding the equivocal cases, there were 8 false-positive and 10 false-negative test results in the pretreatment group and 4 false-positive and 1 false-negative test result in the posttreatment group. For calculation of accuracy, equivocal results were considered as false results, which means an additional six false-positive and two false-negative test results in the pretreatment group and four false-positive results in the posttreatment group. There were no significant differences in the performance of the test results prior to and after therapy. With respect to age, no significant differences were seen in the performance of the test between the three different age groups, both pre- and posttreatment (data not shown).
DISCUSSION
To our best knowledge, this is the first study applying this novel rapid one-step immunochromatographic assay to establish the diagnosis of H. pylori infection and to monitor the success of anti-H. pylori therapy in children.
As a gold standard to compare with for the pretreatment evaluation, we used strict predefined criteria based on results of invasive diagnostic methods (histology, rapid urea test, culture) plus the UBT. Inclusion in this study required at least three of the four tests performed with conclusive results. For the posttreatment evaluation, we used only the results of the UBT as comparison, which is considered to be an adequate method to monitor success of therapy in both adults (14) and children (1). In most children, abdominal symptoms disappeared or improved after treatment, therefore it would have been unethical to perform endoscopy in children for research purposes only. In addition, the pretreatment results confirm our previous experience that the UBT is 100%, sensitive regardless of age, if the test is performed according to a strict protocol (6). A weakness of the UBT is false-positive results in infants and children below 6 years of age (6). In the posttreatment group, only 4 of 79 children fell into this age category, and all of them had a negative UBT and stool test. Therefore, we feel quite confident that the UBT is an adequate gold standard for the posttreatment group to compare with.
Compared to the UBT, stool tests are obviously more convenient, particularly in pediatric patients. Stool samples can be obtained from children without their active collaboration. Another advantage of the test is its performance by individual caregivers without the need of a laboratory or any equipment. In addition, the test is less costly and time consuming than the UBT. The result is available within 10 min. The sample can be easily sent in by mail. Neither keeping the samples at room temperature for up to 5 days nor storage at –20°C seems to influence the accuracy of the stool tests (9, 13). Some stool samples for this study had been frozen up to 36 months, but the performance of the test did not depend on the length of storage time, confirming previous studies (23). Two previous studies using the same office-based test performed in adults from Germany and Taiwan also used frozen, not fresh, stool samples (11, 22). Therefore, we cannot judge whether freezing decreases the accuracy of the results.
The interobserver agreement was excellent: in 95% of the 160 tests performed in Munich, both independent observers judged either positive or negative. However, in 5% the test line was so weakly visible that they judged the results as equivocal. In Vienna, only one investigator read the results; however, she reported also 5% of the test as "cannot decide." These not classifiable results occurred in both infected and noninfected children and in the pre- and posttreatment evaluation. Ten of the 12 equivocal results occurred in H. pylori-negative patients. We do not have any specific information from the producer about the sample diluents, which contain ingredients for the prevention of nonspecific reactions. We excluded all children with diarrhea at the time of stool sampling and those taking acid-suppressive drugs or antibiotics during the 4 weeks prior to testing, since these medications may give false-negative results (15, 20). In the German study in adult patients, two investigators independently judged the intensity of the bands, but the number of discrepant or equivocal results were not reported (11). The authors reported that the intensity of the band seen in the near patient test correlated nicely with the optical density obtained by the EIA test (Premier Platinum HpSA EIA; Meredian Diagnostic) in the same stool sample (11). Therefore, it seems likely that patients with a low bacterial antigen load may have either borderline/equivocal or misclassified test results by both methods.
The rapid monoclonal test correctly classified 203 of the 238 pre- and posttreatment stool samples, giving an overall accuracy of 85% with a sensitivity of 88.3% and specificity of 91%. The performance for sensitivity, specificity, and accuracy of the test was independent of the location (Munich or Vienna; data not shown) and the pre- and posttreatment status (Table 2). The likelihood ratio pretreatment for a positive test result was 7.4 and for a negative test 0.1; both values fulfill the criteria for a good diagnostic test (3). We did not give the results for the positive and negative predictive values, because, in contrast to the likelihood ratio, these figures depend on the prevalence of the H. pylori infection in the population studied. In our pretreatment population, about half of the patients were infected due to the selection process. If we would have taken all consecutive children undergoing upper endoscopy in our unit we would have to investigate about 800 children to end up with the same number of H. pylori-positive patients. Our results confirm the two studies in adults using the same near patient test which also showed no significant differences in the test characteristics before and after treatment (11, 22). Leodolter et al. performed the test in 100 adult patients, half of them after treatment for H. pylori infection (11). While the overall specificity was excellent (98.3%) with only one false-positive test among 58 noninfected patients, the sensitivity was 83% in the untreated patients and 73% in the posttreatment group. In a second study from Taiwan, 253 adults were investigated, 167 pre- and 86 posttreatment, 118 out of 253 were H. pylori positive (22). The sensitivity and specificity before therapy was 94.6% and 89.3% and after treatment 100% and 93.3%, respectively. The better sensitivity in the Taiwanese study may be explained by the fact that most infected patients had peptic ulcer disease. These patients probably have a higher bacterial load than the children in our study with gastritis only.
The test results of quick test, which is based on monoclonal antibodies, are comparable to results in children obtained with the EIA stool test based on polyclonal antibodies. Testing the same stool samples with different production lots of the polyclonal test indicated intercharge variation (13). This may explain the wider range reported for the sensitivity of the test (10, 16, 18, 21). In contrast, H. pylori stool EIA based on monoclonal antibodies showed consistently excellent results (9, 12, 13). Due to the production process, monoclonal antibodies should have less intercharge variability, making the test more reliable. The site of the antigen detection of this test is proprietary. Therefore, lower accuracy of this test compared to the monoclonal EIA may be due to the antigen/antibodies used or due to other yet-unknown reasons.
The performance of diagnostic tests for detection of H. pylori in young children is always a crucial point (8). In most studies the sensitivity of the test cannot be judged due to the small number of infected children (9, 16). In this study, only 9 stool samples were obtained from infected children below 6 years of age, all pretreatment (Table 1). The test gave a positive result in 7 samples (77.8%). Due to the small number, we cannot draw any conclusion for the sensitivity of the test in toddlers and preschool children.
In conclusion, the office-based one-step immunochromatographic test using monoclonal antibodies for diagnosis of H. pylori infection is convenient, with results available within 10 min. The interobserver agreement is good, but equivocal results occur in 5% of the cases. As in adults, the accuracy of the rapid stool test is comparable before and 6 to 8 weeks after therapy but seems inferior to the laboratory based monoclonal EIA. In settings where the monoclonal EIA or the UBT are not available, feasible, or too expensive, this office-based test may become a good noninvasive alternative to assess the H. pylori status in children.
ACKNOWLEDGMENTS
The study was supported by Meridian Bioscience Europe, who provided free test kits for the determination of stool antigen, and by the Child Health Foundation (Stiftung "Kindergesundheit").
REFERENCES
Drumm, B., S. Koletzko, and G. Oderda. 2000. Helicobacter pylori infection in children: a consensus statement. J. Pediatr. Gastroenterol. Nutr. 30:207-213.
Imrie, C., M. Rowland, B. Bourke, and B. Drumm. 2001. Limitations to carbon 13-labeled urea breath testing for Helicobacter pylori in infants. J. Pediatr. 139:734-737.
Jaeschke, R., G. Guyatt, and D. L. Sackett. 1994. Users' guides to the medical literature. III. How to use an article about a diagnostic test. A. Are the results of the study valid Evidence-Based Medicine Working Group. JAMA 271:389-391.
Kabir, S. 2003. Review article: clinic-based testing for Helicobacter pylori infection by enzyme immunoassay of faeces, urine and saliva. Aliment. Pharmacol. Ther. 17:1345-1354.
Khanna, B., A. Cutler, N. R. Israel, M. Perry, A. Lastovica, P. I. Fields, and B. D. Gold. 1998. Use caution with serologic testing for Helicobacter pylori infection in children. J. Infect. Dis. 178:460-465.
Kindermann, A., H. Demmelmair, B. Koletzko, S. Krauss-Etschmann, B. Wiebecke, and S. Koletzko. 2000. Influence of age on 13C-urea breath test results in children. J. Pediatr. Gastroenterol. Nutr. 30:85-91.
Kindermann, A., N. Konstantopoulos, N. Lehn, H. Demmelmair, and S. Koletzko. 2001. Evaluation of two commercial enzyme immunoassays, testing immunoglobulin G (IgG) and IgA responses, for diagnosis of Helicobacter pylori infection in children. J. Clin. Microbiol. 39:3591-3596.
Koletzko, S., and A. Feydt-Schmidt. 2001. Infants differ from teenagers: use of non-invasive tests for detection of Helicobacter pylori infection in children. Eur. J. Gastroenterol. Hepatol. 13:1047-1052.
Koletzko, S., N. Konstantopoulos, D. Bosman, A. Feydt-Schmidt, E. A. van Der, N. Kalach, J. Raymond, and H. Russmann. 2003. Evaluation of a novel monoclonal enzyme immunoassay for detection of Helicobacter pylori antigen in stool from children. Gut 52:804-806.
Konstantopoulos, N., H. Russmann, C. Tasch, T. Sauerwald, H. Demmelmair, I. Autenrieth, and S. Koletzko. 2001. Evaluation of the Helicobacter pylori stool antigen test (HpSA) for detection of Helicobacter pylori infection in children. Am. J. Gastroenterol. 96:677-683.
Leodolter, A., K. Wolle, U. Peitz, A. Schaffranke, T. Wex, and P. Malfertheiner. 2004. Evaluation of a near-patient fecal antigen test for the assessment of Helicobacter pylori status. Diagn. Microbiol. Infect. Dis. 48:145-147.
Leodolter, A., U. Peitz, M. P. Ebert, K. Agha-Amiri, and P. Malfertheiner. 2002. Comparison of two enzyme immunoassays for the assessment of Helicobacter pylori status in stool specimens after eradication therapy. Am. J. Gastroenterol. 97:1682-1686.
Makristathis, A., W. Barousch, E. Pasching, C. Binder, C. Kuderna, P. Apfalter, M. L. Rotter, and A. M. Hirschl. 2000. Two enzyme immunoassays and PCR for detection of Helicobacter pylori in stool specimens from pediatric patients before and after eradication therapy. J. Clin. Microbiol. 38:3710-3714.
Malfertheiner, P., F. Megraud, C. O'Morain, A.-P. S. Hungin, R. Jones, A. Axon, D. Y. Graham, and G. Tytgat. 2002. Current concepts in the management of Helicobacter pylori infection-the Maastricht 2-2000 Consensus Report. Aliment. Pharmacol. Ther. 16:167-180.
Manes, G., A. Balzano, G. Iaquinto, C. Ricci, M. M. Piccirillo, N. Giardullo, A. Todisco, M. Lioniello, and D. Vaira. 2001. Accuracy of stool antigen test in posteradication assessment of Helicobacter pylori infection. Dig. Dis. Sci. 46:2440-2444.
Megraud, F. 2005. Non-invasive tests to detect Helicobacter pylori infection in children and adolescents: results of a multicentric European study. J. Pediatr. 146:198-203.
Oderda, G., A. Rapa, D. Marinello, B. Ronchi, and A. Zavallone. 2001. Usefulness of Helicobacter pylori stool antigen test to monitor response to eradication treatment in children. Aliment. Pharmacol. Ther. 15:203-206.
Oderda, G., A. Rapa, B. Ronchi, P. Lerro, M. Pastore, A. Staiano, G. L. de'Angelis, and P. Strisciuglio. 2000. Detection of Helicobacter pylori in stool specimens by non-invasive antigen enzyme immunoassay in children: multicentre Italian study. BMJ 320:347-348.
Oliveira, A. M. R., G. A. Rocha, D. M. Queiroz, E. N. Mendes, C. A. De, T. C. Ferrari, and A. M. Nogueira. 1999. Evaluation of enzyme-linked immunosorbent assay for the diagnosis of Helicobacter pylori infection in children from different age groups with and without duodenal ulcer. J. Pediatr. Gastroenterol. Nutr. 28:157-161.
Roggero, P., A. Bonfiglio, S. Luzzani, A. Valade, E. Cataliotti, G. Corno, M. C. Garlaschi, E. Carissimi, F. Mosca, and V. Carnelli. 2002. Helicobacter pylori stool antigen test: a method to confirm eradication in children. J. Pediatr. 140:775-777.
van Doorn, O. J., D. K. Bosman, B. W. van't Hoff, J. A. Taminiau, F. J. ten Kate, and A. van der Ende. 2001. Helicobacter pylori stool antigen test: a reliable non-invasive test for the diagnosis of Helicobacter pylori infection in children. Eur. J. Gastroenterol. Hepatol. 13:1061-1065.
Wu, I. C., H. L. Ke, Y. C. Lo, Y. C. Yang, C. H. Chuang, F. J. Yu, Y. C. Lee, C. M. Jan, W. M. Wang, and D. C. Wu. 2003. Evaluation of a newly developed office-based stool test for detecting Helicobacter pylori: an extensive pilot study. Hepatogastroenterology 50:1761-1765.
Yee, Y. K., K. T. Yip, T. L. Que, K. K. Chang, K. F. Li, C. K. Lee, S. W. Wong, S. F. Lau, and M. L. Szeto. 2002. Efficacy of enzyme immunoassay for the detection of Helicobacter pylori antigens in frozen stool specimens: local validation. Aliment. Pharmacol. Ther. 16:1739-1742.