Stool antigen for the diagnosis of Helicobacter pylori infection in cirrhosis: comparative usefulness of three different methods

Transcription

1 Aliment Pharmacol Ther 2003; 17: doi: /j x Stool antigen for the diagnosis of Helicobacter pylori infection in cirrhosis: comparative usefulness of three different methods X. CALVET*, M. QUESADA, M. ROSELLÓ, F. SALCEDA, I. SANFELIU, B. DALMAU* & M. GIL* *Unitat de Malalties Digestives and Laboratori de Microbiologia, Corporació Parc Taulí, Sabadell, Barcelona, Spain Accepted for publication 30 November 2002 SUMMARY Background: Helicobacter pylori infection may lead to peptic ulcer disease, and causes significant morbidity in patients with cirrhosis. The measurement of H. pylori antigens in human stools has been proposed as a valuable, non-invasive, diagnostic tool. A number of tests have recently been commercialized. However, very few data are available on their reliability in patients with cirrhosis. Aim: To evaluate the usefulness of three new tests HpSA (Meridian Diagnostics Inc., Cincinnati, OH, USA), Simple H. pyl (OPERON S.A., Zaragoza, Spain) and FemtoLab H. pylori (Connex, Martinsried, Germany) in the diagnosis of H. pylori infection in cirrhotic patients. Methods: H. pylori infection was determined in 79 cirrhotic patients (48 men, 31 women; age range, years; mean, 62 ± 11 years) by concordance of histology and urea breath test. The sensitivity, specificity and positive and negative predictive values of each stool test were calculated. Results: According to the reference method, the sensitivities of HpSA, Simple H. pyl and FemtoLab H. pylori immunoassays were 76%, 87% and 78%, respectively, and their specificities were 93%, 62% and 79%, respectively. Conclusions: Faecal tests are non-invasive and easy-toperform tools for the diagnosis of H. pylori infection. However, their sensitivity and specificity seem to be non-optimal in patients with cirrhosis. INTRODUCTION Peptic ulcer is a frequent finding in cirrhotic patients, and its complications result in significant morbidity and mortality. 1 Helicobacter pylori infection is the major cause of peptic ulcer in both cirrhotics and the general population. 2, 3 Although there are no data on cirrhotic patients, the eradication of H. pylori infection would probably lead to the almost total prevention of recurrent peptic ulcer and its complications, as it does in noncirrhotic individuals. 4 7 In addition, H. pylori may also increase the risk of hepatic encephalopathy, although this topic remains extremely controversial Therefore, the correct diagnosis and treatment of H. pylori Correspondence to: Dr X. Calvet, Unitat de Malalties Digestives, Corporació Parc Taulí, Parc Taulí, s/n, Sabadell, Barcelona, Spain. xcalvet@cspt.es infection is important in patients with liver cirrhosis and peptic ulcer. The search for an ideal test to diagnose H. pylori infection has been only partially successful. The perfect test would be highly reliable, inexpensive and noninvasive. Until recently, serology and the 13 C-urea breath test were the only non-invasive tests available. However, the inaccuracy of serology has led recent consensus reports to advise against its general use. 13 In contrast, the 13 C-urea breath test is very sensitive and specific, but expensive, as it usually requires trained personnel to obtain the breath samples, and both 13 C-urea and the instrumentation necessary to interpret the samples are costly. 14 Recently, it has been reported that H. pylori antigens can be measured in human stools. The stool test may become a valuable, non-invasive, diagnostic tool. Three new tests have been commercialized: Premier Platinum Ó 2003 Blackwell Publishing Ltd 727

2 728 X. CALVET et al. HpSA (HpSA, Meridian Diagnostics Inc., Cincinnati, OH, USA), Simple H. pyl (OPERON S.A., Zaragoza, Spain) and FemtoLab H. pylori (Connex, Martinsried, Germany). Excellent results have been reported with all these tests in dyspeptic patients However, these new stool antigen diagnostic tests have not been evaluated in patients with cirrhosis. Therefore, the objective of this study was to evaluate the reliability of these new stool antigen detection tests for the diagnosis of H. pylori infection in cirrhosis. PATIENTS AND METHODS Patients Between September 1998 and June 2000, 101 consecutive cirrhotic patients were included in a previously reported study aimed to evaluate the reliability of multiple diagnostic tests for H. pylori infection in patients with cirrhosis. The results of this study have been described elsewhere. 19 The inclusion criteria were liver cirrhosis diagnosed by biopsy or by clinical, biochemical and echographic criteria, as well as upper gastrointestinal endoscopy for screening or follow-up of oesophageal varices. The exclusion criteria were an age of < 18 years, refusal to participate in the study or treatment with proton pump inhibitors or antibiotics in the 4 weeks before the diagnostic tests. None of the patients had been previously tested or treated for H. pylori infection. The study conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the ethical committee of the Corporació Parc Taulí. Patients gave informed consent to participate in the study. In this study, two antral and two corporeal gastric biopsies were obtained for histology. Samples were stained with haematoxylin and eosin and Giemsa. The pathologist who evaluated the samples was unaware of the other diagnostic tests. A 13 C-urea breath test was performed in accordance with the standard European protocol. 20 Finally, a faecal sample was collected and frozen at ) 70 C until analysis. All samples were obtained within 2 weeks of endoscopy. For this study, we selected patients in whom both histology and urea breath test had been performed and who also had a frozen faecal sample. Seventy-nine patients fulfilled these criteria. The mean age was 62.4 ± 10.5 years (range, years); 48 patients were male and 31 were female. Cirrhosis was secondary to virus C infection in 44 cases, virus B infection in two, alcoholic in 29 and cryptogenic in four. Sixty-four patients were designated as Child A, 14 as Child B and two as Child C according to the Child Pugh classification. 21 Determination of H. pylori status H. pylori status was determined from the results of histology and urea breath test. Patients with a positive histology (either of the antrum or gastric body) and a positive urea breath test were considered to be infected. Patients with negative scores on all tests were considered to be uninfected. Finally, those with discordant results were considered to be undetermined and were excluded from the study. Determination of faecal antigen Premier Platinum HpSA (Meridian Diagnostics Inc., Cincinnati, OH, USA) was performed in accordance with the manufacturer s specifications. Readings were performed at 450/630 nm. Tests with an absorbance superior or equal to 0.12 were considered to be positive, those below 0.1 negative, and those between 0.1 and 0.12 indeterminate. FemtoLab H. pylori (Connex, Martinsried, Germany) uses monoclonal antibodies to H. pylori as capture antibodies adsorbed to microwells. In brief, a small portion of the stool specimen was mixed with a sample diluent and was then centrifuged for 5 min at 7000 r.p.m.; 50 ll of the supernatant and 50 ll of peroxidase-conjugated monoclonal antibodies were added to the microwells and incubated for 1 h at room temperature in a mechanical shaker. The wells were then washed to remove unbound material. Substrate was added and incubated for 10 min. In the presence of H. pylori antigens, colour developed. Stop solution was added and the results were read by spectrophotometry. In accordance with the manufacturer s instructions, specimens with absorbance values at 450/620 nm of 0.15 were considered to be positive, and those with values below 0.15 negative. Simple H. pyl (OPERON S.A., Zaragoza, Spain) is a rapid immunochromatographic test using a single monoclonal antibody named L1. A diluted faecal sample is placed in an immunochromatographic support and the result is read after 5 min. A positive test will display a red line in the reading area next to a blue control line. By contrast,

3 STOOL TEST FOR H. PYLORI DIAGNOSIS IN CIRRHOSIS 729 the sample is considered to be negative when only the control blue line develops. If no line appears in the reading area, the test is considered to be null. Statistical analysis The sensitivity, specificity and positive and negative predictive values and their 95% confidence intervals were calculated according to standard methods. The receiver operating characteristic curve was plotted and the values of the maximum sensitivity and specificity were determined for the quantitative tests. Calculations were performed using conventional software (SPSS 10 for Windows). RESULTS Faecal samples were available in 79 patients. Fortyeight patients tested positive for H. pylori infection using haematoxylin and eosin or Giemsa stain and 47 by 13 C-urea breath test. According to the established gold standard, 45 of these patients were positive, 29 negative and five undetermined. The sensitivity, specificity and positive and negative predictive values of the three techniques are shown in Table 1. As the diagnostic reliability of the different tests was sub-optimal, we performed a detailed post hoc analysis of the patients with false negative or false positive results. Few coincidences were observed, with only two of 13 patients with at least one false positive result being positive in all faecal tests analysed and two of 14 patients with at least one false negative result being negative in all faecal tests analysed. Receiver operating characteristic curves were plotted for the two quantitative enzyme-linked immunoabsorbent assay methods (FemtoLab H. pylori and Premier Platinum HpSA). The area under the curve was for FemtoLab H. pylori and for Premier Platinum HpSA (Figure 1). Changing the cut-off value did not clearly improve the results of either test. Figure 1. Receiver operating characteristic curve of quantitative stool tests (FemtoLab H. pylori and Premier Platinum HpSA). The area under the curve was for FemtoLab H. pylori and for Premier Platinum HpSA. Changing the cut-off value did not clearly improve the results of either of the tests. DISCUSSION The results of this study suggest that the reliability of the faecal H. pylori test in patients with cirrhosis is limited. Differences between individual tests were minor. Simple H. pyl showed a slightly better sensitivity but a very low specificity. Premier Platinum HpSA showed an excellent specificity but a low sensitivity, and FemtoLab H. pylori had relatively low values for both sensitivity and specificity. These results contrast with the excellent values found with these tests in dyspeptic 17, 18, patients. The reasons for these different results between cirrhotics and dyspeptic patients remain unclear. Geographical variations in the H. pylori antigen pattern could account for the reduced reliability of the test in our area. However, all stool antigen tests have been previously tested locally and all showed excellent results, although the polyclonal 18, 23, 25, 26 tests seemed to be more irregular. Although patients were asked specifically about antisecretory drugs or antibiotic use before inclusion, Table 1. Sensitivity, specificity and positive and negative predictive values of the faecal tests Sensitivity (%) Specificity (%) PPV (%) NPV (%) Simple H. pyl 87 (73 95) 62 (42 79) 78 (64 88) 75 (53 90) FemtoLab H. pylori 78 (63 88) 79 (60 91) 85 (70 94) 70 (51 84) Premier Platinum HpSA 76 (60 87) 93 (76 99) 94 (80 99) 71 (54 84) CI, confidence interval; NPV, negative predictive value; PPV, positive predictive value.

4 730 X. CALVET et al. surreptitious use of these drugs cannot be totally ruled out as a cause of the false negative results. We hypothesized that the decreased sensitivity could be attributed to an intrinsic or pharmacological factor that reduces the gastric H. pylori density or inhibits one of the steps in the immunodiagnostic test. If this were the case, we would expect false negative results of faecal tests to be found in only a few patients, and all faecal tests would be negative in these patients. However, the analysis showed great heterogeneity. Only two of the 14 patients with at least one false negative faecal test result were negative in all the tests analysed; therefore, the poor results were unlikely to be due to the presence of an inhibitory factor. This finding makes it unlikely that certain drugs commonly used in cirrhotics (for example, lactulose, which alters the colonic microflora and faecal ph) or changes secondary to portal hypertension (which may change the density or distribution of H. pylori infection) may have affected the reliability of the tests. A final possible explanation is that the poor performance of the faecal tests might not be specific to cirrhotic patients, but related to a general lack of reliability of tests of this kind. The publication of reports showing the sub-optimal performance of these tests is of great scientific and clinical interest, as there is a strong risk of publication bias favouring studies with good results. On the basis of the current published data, the Maastricht consensus report suggests that the efficacy of faecal tests is similar to that of the highly reliable urea breath test, and recommends them as alternatives for non-invasive diagnosis in primary care. 13 Probably, we should be more prudent before proposing faecal tests as a trustworthy alternative. Indeed, there is a risk of repeating the experience of serology for H. pylori. For years, most of the published papers showed excellent results. However, finally, and in spite of a considerable body of evidence, the Maastricht consensus report acknowledged what many of us had, in fact, suspected: that serology is often unreliable and should not be used in clinical practice. 13 Whatever the reason, all faecal tests performed worse than histology or the urea breath test. 19 However, these tests are cheap, non-invasive and easy to perform. As the number of patients in our study was limited and the confidence intervals wide, further evaluation is required before removing faecal tests from the list of H. pylori diagnostic techniques in patients with chronic liver disease. ACKNOWLEDGEMENTS This study was supported by grants from the Societat Catalana de Digestologia and the Corporació Parc Taulí. The manufacturers of the diagnostic devices did not fund the study; however, the providers were aware that the diagnostic reliability of the samples would be evaluated. The kits were kindly supplied without charge (Simple H. pyl, OPERON S.A., Zaragoza, Spain) or sold at reduced fees (FemtoLab H. pylori, Connex, Martinsried, Germany, and Premier Platinum HpSA, Meridian Diagnostics, Cincinnati, OH, USA). REFERENCES 1 Bonnevie O. Causes of death in duodenal and gastric ulcer. Gastroenterology 1977; 73: Calvet X, Navarro M, Gil M, et al. Epidemiology of peptic ulcer disease in cirrhotic patients: role of Helicobacter pylori infection. Am J Gastroenterol 1998; 93: Siringo S, Burroughs AK, Bolondi L, et al. Peptic ulcer and its course in cirrhosis: an endoscopic and clinical prospective study. J Hepatol 1995; 22: van der Hulst RW, Rauws EA, Koycu B, et al. Prevention of ulcer recurrence after eradication of Helicobacter pylori: a prospective long-term follow-up study. Gastroenterology 1997; 113: Vergara M, Casellas F, Saperas E, et al. Helicobacter pylori eradication prevents recurrence from peptic ulcer haemorrhage. Eur J Gastroenterol Hepatol 2000; 12: Rokkas T, Karameris A, Mavrogeorgis A, et al. Eradication of Helicobacter pylori reduces the possibility of rebleeding in peptic ulcer disease. Gastrointest Endosc 1995; 41: Jaspersen D, Koerner T, Schorr W, et al. Helicobacter pylori eradication reduces the rate of rebleeding in ulcer hemorrhage. Gastrointest Endosc 1995; 41: Dasani BM, Sigal SH, Lieber CS. Analysis of risk factors for chronic hepatic encephalopathy: the role of Helicobacter pylori infection. Am J Gastroenterol 1998; 93: Gubbins GP, Moritz TE, Marsano LS, et al. Helicobacter pylori is a risk factor for hepatic encephalopathy in acute alcoholic hepatitis: the ammonia hypothesis revisited. The Veterans Administration Cooperative Study Group No Am J Gastroenterol 1993; 88: Rinaldi V, Zullo A, Diana F, et al. Helicobacter pylori, hyperammonemia, and hepatic encephalopathy: is there a correlation? [Letter]. Am J Gastroenterol 1997; 92: Vascónez C, Elizalde JI, Llach J, et al. Helicobacter pylori, hyperammonemia and subclinical portosystemic encephalopathy: effects of eradication. J Hepatol 1999; 30: Calvet X, Nogueras C, Roqué M, et al. Helicobacter pylori is not a risk factor for hepatic encephalopathy. Dig Liver Dis 2001; 33:

5 STOOL TEST FOR H. PYLORI DIAGNOSIS IN CIRRHOSIS Malfertheiner P. Current concepts in the management of Helicobacter pylori infection the Maastrich consensus report 2, Aliment Pharmacol Ther 2002; 16: Marchildon PA, Ciota LM, Zamaniyan FZ, et al. Evaluation of three commercial enzyme immunoassays compared with the 13 C urea breath test for detection of Helicobacter pylori infection. J Clin Microbiol 1996; 34: Agha-Amiri K, Peitz U, Mainz D, et al. A novel immunoassay based on monoclonal antibodies for the detection of Helicobacter pylori antigens in human stool. Z Gastroenterol 2001; 39: Archimandritis A, Giontzis A, Kourtesas D, et al. Diagnosis of Helicobacter pylori infection with a stool assay: is it useful after eradication treatment? Am J Gastroenterol 2001; 96: Lakner M, Leier S, Dehnert S, et al. A novel near-patient test for the direct detection of H. pylori antigens in stool. Gastroenterology 2001; 120(Suppl. 1): A491 A492(Abstract). 18 Salceda F, Calvet X, Sanfeliu I, et al. Evaluación de un nuevo test rápido para la detección de Helicobacter pylori en heces. Med Clin (Barc) 2002; 118: Calvet X, Sanfeliu I, Musulén E, et al. Evaluation of Helicobacter pylori diagnostic methods in patients with liver cirrhosis. Aliment Pharmacol Ther 2002; 16: Domínguez Munoz JE, Leodolter A, Sauerbruch T, et al. A citric acid solution is an optimal test drink in the 13 C-urea breath test for the diagnosis of Helicobacter pylori infection. Gut 1997; 40: Pugh RNH, Murray-Lyon IM, Dawson JL, et al. Transection of the esophagus for bleeding esophageal varices. Br J Surg 1973; 60: Haindl D, Benesch H, Finck A, et al. A novel enzyme inmunoassay for the direct detection of H. pylori antigens in stool. Gastroenterology 2001; 120(Suppl. 1): A492 A493(Abstract). 23 Vaira D, Malfertheiner P, Megraud F, et al. Diagnosis of Helicobacter pylori infection with a new non-invasive antigenbased assay. HpSA European Study Group. Lancet 1999; 354: Roselló M, Quesada M, Saufeliu I, et al. Evaluación de un nuevo método inmunodiagnóstico (FemtoLab) para Helicobacter pylori en heces. Enf Infecc Microbiol Clin 2002; 20: Calvet X, Feu F, Forné M, et al. The evaluation of a new immunoenzyme analysis for the detection of Helicobacter pylori infection in stool samples. Gastroenterol Hepatol 1999; 22: Forné M, Domínguez J, Fernández-Bañares F, et al. Accuracy of an enzyme immunoassay for the detection of Helicobacter pylori in stool specimens in the diagnosis of infection and post-treatment check-up. Am J Gastroenterol 2000; 95: