Portal de Periódicos da CAPES

Helicobacter pylori: drowning in a pool of blood?

1999; Elsevier BV; Volume: 49; Issue: 3 Linguagem: Inglês

10.1016/s0016-5107(99)70024-6

1097-6779

Loren Laine, Hartley Cohen,

Gastrointestinal disorders and treatments

Diagnosis of Helicobacter pylori infection is crucial in the short-term and long-term management of patients with bleeding ulcers. If a patient with ulcer bleeding requires urgent surgical intervention, knowledge of the H pylori status of the patient may help guide the choice of procedures (i.e., a simple procedure such as oversewing in H pylori–positive patients vs full-blown ulcer surgery in H pylori-negative patients). Even if the acute bleeding episode resolves without further bleeding in the short term, one third of patients will have recurrent bleeding within the next 1 to 2 years without medical or surgical treatment.1Laine L. Helicobacter pylori and complicated ulcer diseases.Am J Med. 1996; 100: 52S-59SAbstract Full Text PDF PubMed Google Scholar Before the recognition of H pylori's role in the pathogenesis of ulcer disease, our options after a major bleeding episode were either to place the patient on indefinite maintenance therapy (usually an H2 receptor antagonist at half dose) or to perform an acid-lowering surgical procedure. However, a number of studies have documented that eradication of H pylori, without any further maintenance therapy, results in an extremely low rate of recurrent bleeding, thus preventing the need for surgery or a lifetime of medical therapy.1Laine L. Helicobacter pylori and complicated ulcer diseases.Am J Med. 1996; 100: 52S-59SAbstract Full Text PDF PubMed Google Scholar Patients with bleeding ulcers not associated with H pylori will generally require long-term antisecretory therapy to decrease the risk of recurrent bleeding. Thus, documentation of H pylori status in a patient presenting with a bleeding ulcer is mandatory to determine future management. H pylori diagnostic tests may be divided into those requiring endoscopy to obtain biopsies of the gastric mucosa and those not requiring endoscopy. Endoscopic tests include culture, histology, and rapid urease tests (RUTs). Culture is not usually relied on for diagnosis of H pylori infection because the organism is fastidious and the sensitivity of culture often is lower than that of other tests. Although considered the gold standard for the diagnosis of H pylori, histologic evaluation is not without pitfalls related to observer error, sampling error, the density of H pylori infection, and staining methods. For an experienced pathologist, hematoxylin and eosin stain is adequate for routine clinical purposes, unless very few organisms are noted (possible false-positive result) or no organisms are seen in the presence of more than a mild increase in inflammatory cell infiltrate (possible false-negative).2Cohen H Laine L. Endoscopic methods for the diagnosis of Helicobacter pylori.Aliment Pharmacol Ther. 1997; 11: 3-9Crossref PubMed Scopus (59) Google Scholar In either circumstance, additional special stains (e.g., Giemsa, Genta, silver) should be used. Some pathologists routinely use special stains initially because the organisms are more easily visualized. Histologic assessment of inflammatory cell infiltrate is also very helpful in the diagnosis of H pylori because the presence of chronic active gastritis (mononuclear cells and neutrophils) is virtually diagnostic of H pylori.3Cutler AF Havstad S Ma CK Blaser MJ Perez-Perez GI Schubert TT. Accuracy of invasive and noninvasive tests to diagnose Helicobacter pylori infection.Gastroenterology. 1995; 109: 136-141Abstract Full Text PDF PubMed Scopus (513) Google Scholar Because of their simplicity, accuracy, and relatively rapid determination (1 to 24 hours) of H pylori status, RUTs are generally considered the initial endoscopic test of choice. RUTs are based on the urease-producing activity of H pylori: when urease is present, urea in the tests is hydrolyzed, releasing ammonia and raising the pH. A color change caused by a pH indicator in the tests denotes a positive result. Commercially available tests include the CLOtest and Hpfast, agar gel tests in which H pylori urease hydrolyzes urea in the gel, and the PyloriTek in which biopsy specimens are placed on a urea-impregnated pad and, if H pylori (and thus urease) is present, ammonia is produced and passes through a membrane to cause a color change in an overlying pH indicator. Comparative studies of the three tests show no significant differences in sensitivity or specificity at the final reading, although at 1 hour the Pyloritek has a significantly higher sensitivity than the agar gel tests.4Laine L Lewin D Naritoku W Estrada R Cohen H. Prospective comparison of commercially available rapid urease tests for the diagnosis of Helicobacter pylori.Gastrointest Endosc. 1996; 44: 523-526Abstract Full Text Full Text PDF PubMed Scopus (100) Google Scholar Sensitivity of a single CLOtest (the most widely studied of the three tests) is not affected by incubation at 37° C or by the size or number of biopsy specimens.2Cohen H Laine L. Endoscopic methods for the diagnosis of Helicobacter pylori.Aliment Pharmacol Ther. 1997; 11: 3-9Crossref PubMed Scopus (59) Google Scholar Sensitivities of the RUTs in patients not treated for H pylori can be expected to be approximately 90%; specificities are 95% to 100%.2Cohen H Laine L. Endoscopic methods for the diagnosis of Helicobacter pylori.Aliment Pharmacol Ther. 1997; 11: 3-9Crossref PubMed Scopus (59) Google Scholar, 4Laine L Lewin D Naritoku W Estrada R Cohen H. Prospective comparison of commercially available rapid urease tests for the diagnosis of Helicobacter pylori.Gastrointest Endosc. 1996; 44: 523-526Abstract Full Text Full Text PDF PubMed Scopus (100) Google Scholar However, sensitivities vary in different studies, even when done at the same institution, and range from approximately 80% to 95%. Urea breath tests are non-endoscopic tests which, like biopsy-based tests, reflect the actual bacterial load and have reported sensitivities and specificities of approximately 90% to 95%. Tests which measure the bacterial load should not be performed within 4 weeks of therapy with bismuth compounds or antibiotics that might temporarily diminish bacterial density. Proton pump inhibitors also suppress H pylori, but an interval of 2 weeks between the last dose of proton pump inhibitor and H pylori diagnostic testing seems adequate.5Laine L Estrada R Trujillo M Knigge K Fennerty MB. The effect of proton pump inhibitor therapy on diagnostic testing for Helicobacter pylori.Ann Intern Med. 1998; 129: 547-550Crossref PubMed Scopus (183) Google Scholar In contrast, antibody tests do not necessarily reflect current bacterial load, and their presence may indicate prior rather than current infection. Results vary from test to test and location to location and, overall, antibody tests are less accurate than the urea breath tests or endoscopic tests: a recent meta-analysis of antibody tests reported a sensitivity of 85% and a specificity of 79%.6Loy CT Irwig LM Katelaris PH Talley NJ. Do commercial serologic kits for Helicobacter pylori infection differ in accuracy? A meta-analysis.Am J Gastroenterol. 1996; 91: 1138-1144PubMed Google Scholar Another non-endoscopic test is the detection of H pylori antigen in stool using anti-H pylori antibodies. Preliminary studies suggest that this test has sensitivity and specificity comparable to endoscopic-based tests before treatment, but specificity after therapy may be decreased.7Makristathis A Pasching E Schutze K Wimmer M Rotter ML Hirschl AM. Detection of Helicobacter pylori in stool specimens by PCR and antigen EIA.Gut. 1998; 42 ([abstract]): A47PubMed Google Scholar, 8Trevisani L Sartori L Galvani F Dentale A Caselli M Ruina M et al.Evaluation of a new enzyme immunoassay for detecting Helicobacter pylori in faeces.Gut. 1998; 42 ([abstract]): A47-A48PubMed Google Scholar When an ulcer is discovered at upper endoscopy, we recommend taking endoscopic biopsy specimens for RUT and banking additional specimens for histologic examination. There is no contraindication to taking several specimens from uninvolved gastric mucosa at the end of an emergency endoscopic procedure for ulcer bleeding, barring a severe bleeding diathesis or a patient in whom the procedure must be terminated because of hemodynamic instability. Although blood antibody testing is often used as a screening test, we do not believe that antibody testing should replace endoscopic biopsies in the patient with a complicated ulcer because of antibody testing's lower specificity. For example, if a patient has a false-positive antibody test and is treated for H pylori, when post-treatment testing reveals no H pylori, the physician will mistakenly believe that the cause of the bleeding has been removed and no further preventive therapy is necessary; in this case, the patient will have a high risk of recurrent bleeding. Therefore, we believe that establishing the diagnosis with certainty at the time of the bleeding episode is crucial. H pylori infection is reported in most studies in approximately 80% to 100% of patients with duodenal ulcers and 70% to 90% of patients with gastric ulcers. Recent U.S. data, however, indicate that the prevalence may be lower.9Peterson WL Ciociola AA Sykes DL McSorley DJ Webb DD. Ranitidine bismuth citrate plus clarithromycin is effective for healing duodenal ulcers, eradicating H. pylori and reducing ulcer recurrence.Aliment Pharmacol Ther. 1996; 10: 251-256Crossref PubMed Scopus (109) Google Scholar, 10Laine L Hopkins RJ Girardi LS. Has the impact of Helicobacter pylori therapy on ulcer recurrence in the United States been overstated: a meta-analysis of rigorously designed trials.Am J Gastroenterol. 1998; 93: 1409-1415PubMed Google Scholar Furthermore, the prevalence of H pylori infection has been reported to be lower in patients with complicated ulcers than in those with uncomplicated ulcers. Hosking et al.11Hosking SW Yung MY Chung SC Li AKC. Differing prevalence of Helicobacter in bleeding and nonbleeding ulcers [abstract.Gastroenterology. 1992; 102: A85Google Scholar reported that 71 of 102 patients (71%) presenting with bleeding duodenal ulcers were H pylori-positive, whereas 112 of 121 patients (93%) presenting with nonbleeding duodenal ulcers during the same period had H pylori infection (p < 0.01). Thirty-nine of sixty-four patients (61%) with bleeding gastric ulcers were H pylori-positive, whereas 30 of 40 patients (75%) with nonbleeding gastric ulcers had H pylori infection (difference not significant). Importantly, Hosking et al.11Hosking SW Yung MY Chung SC Li AKC. Differing prevalence of Helicobacter in bleeding and nonbleeding ulcers [abstract.Gastroenterology. 1992; 102: A85Google Scholar assessed H pylori status with only an RUT. Jensen et al.12Jensen DM You S Pelayo E Jensen ME. The prevalence of Helicobacter pylori and NSAID use in patients with severe UGI hemorrhage and their potential role in recurrence of ulcer bleeding.Gastroenterology. 1992; 102 ([abstract]): A90Google Scholar reported an H pylori prevalence rate of only 72% in patients with bleeding duodenal ulcers and 79% in patients with bleeding gastric ulcers. Again, a potentially imperfect gold standard was used: serologic testing, biopsy (test not stated), or urea breath test was used, but the numbers tested by each method were not reported. Lee et al.,13Lee JM Breslin NP Fallon C O'Morain CA The rapid urease test lacks sensitivity in bleeding peptic ulcer disease.Gastroenterology. 1997; 112 ([abstract]): A195Abstract Full Text PDF Scopus (805) Google Scholar using a gold standard including RUT, histology, and culture, subsequently found that H pylori was significantly lower in 55 patients with bleeding duodenal ulcers than in 69 patients with nonbleeding ulcers (73% vs 93%). What are potential explanations for the low prevalence of H pylori infection in patients with complicated ulcers? The use of nonsteroidal anti-inflammatory drugs (NSAIDs) would seem the most likely explanation. However, in the studies cited above a significantly increased prevalence of NSAID use was not documented in the H pylori–negative patients. Certainly, surreptitious NSAID use is common in patients presenting with bleeding14Lanas A Sekar MC Hirschowitz BI. Objective evidence of aspirin use in both ulcer and nonulcer upper and lower gastrointestinal bleeding.Gastroenterology. 1992; 103: 862-869Abstract Full Text PDF PubMed Scopus (220) Google Scholar and may still explain many of the H pylori–negative patients with ulcer bleeding. Others have suggested that patients with H pylori–negative ulcers represent a more virulent ulcer diathesis, with a higher chance of complications.15McColl KEL El-Nujumi AM Chittajallu RS Dahill SW Dorrian CA El-Omar E et al.A study of the pathogenesis of Helicobacter pylori–negative chronic duodenal ulceration.Gut. 1993; 34: 762-768Crossref PubMed Scopus (136) Google Scholar If this is true, then a lower H pylori prevalence would be expected among patients with complicated ulcers than those with uncomplicated ulcers. Another possibility to consider is whether bleeding somehow decreases the sensitivity of H pylori diagnostic testing. Lai et al.,16Lai KC Hui WM Lam SK. Bleeding ulcers have high false negative rates for antral Helicobacter pylori when tested with urease test.Gastroenterology. 1996; 110 ([abstract]): A167Abstract Full Text PDF PubMed Scopus (410) Google Scholar in a 1996 abstract, were the first to suggest that the sensitivity of RUT (type unspecified) might be decreased in patients with bleeding ulcers. RUT was positive in about 60% of patients with bleeding ulcers; most of the remainder had positive serologic tests for H pylori but no other diagnostic testing (e.g., histologic examination or urea breath test) was performed. In the study referred to above, Lee et al.,13Lee JM Breslin NP Fallon C O'Morain CA The rapid urease test lacks sensitivity in bleeding peptic ulcer disease.Gastroenterology. 1997; 112 ([abstract]): A195Abstract Full Text PDF Scopus (805) Google Scholar in addition to finding a lower prevalence of H pylori infection in patients with bleeding ulcers, also reported that RUT (type unspecified) had a lower sensitivity in patients with bleeding ulcers (75% with bleeding vs 98% with nonbleeding ulcers). The results presented in their abstract, however, do not rule out that bleeding causes all biopsy tests to decrease in sensitivity. Colin et al.17Colin R Bigard MA Notteghem B Brazier F Berkelmans I Bretagne JF et al.Poor sensitivity of direct tests for detection of Helicobacter pylori on antral biopsies in bleeding ulcers (BU).Gastroenterology. 1997; 112 ([abstract]): A93Google Scholar examined RUT, histology, and culture in 182 patients with bleeding ulcers, apparently using serology as their gold standard. They reported sensitivities of the three biopsy tests as 31%, 26%, and 25%, respectively; surprisingly they stated that sensitivity decreased with a longer time between the bleeding episode and obtaining the biopsy, a finding which seems to run counter to the theory that the presence of blood decreases biopsy test sensitivity. Nevertheless, Colin et al.17Colin R Bigard MA Notteghem B Brazier F Berkelmans I Bretagne JF et al.Poor sensitivity of direct tests for detection of Helicobacter pylori on antral biopsies in bleeding ulcers (BU).Gastroenterology. 1997; 112 ([abstract]): A93Google Scholar suggested that all direct tests for H pylori had decreased sensitivity in the setting of ulcer bleeding. Archimandritis et al.18Archimandritis A Tzivras M Souyioultzis S Papaparaskevas I Apostolopoulos P Delladetsima I et al.High rates of false negative rapid urease test (CLO) in patients with upper gastrointestinal bleeding (UGB).Gut. 1997; 41 ([abstract]): A76Google Scholar evaluated 55 patients with upper GI bleeding and reported positive H pylori serology in 89%, positive histology in 56%, and positive CLOtest in 40%. The use of serology as a gold standard and the lack of nonbleeding ulcer controls limit the strength of the conclusions in these studies. In this issue of Gastrointestinal Endoscopy, Tu et al.19Tu TC Lee CL Wu CH Chen TK Chan CC Huang SH et al.Comparison of invasive and noninvasive tests for detecting Helicobacter pylori infection in bleeding peptic ulcers.Gastrointest Endosc. 1999; 49: 302-306Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar studied patients presenting with upper GI bleeding within 3 days of endoscopy. They only included patients with H pylori infection as determined by an acceptable gold standard. Serologic testing had a sensitivity of 97%, whereas the sensitivity of the CLOtest and histology biopsy tests was only 64% and 73%. Interestingly, the sensitivity of the urea breath test which, like the biopsy tests, measures bacterial load, was 94%. After dividing the patients into those with blood and those without blood in the antrum at the time of endoscopy, they found that sensitivities of histology, culture, urea breath test, and serology were comparable between the two groups and that the sensitivity of the CLOtest was lower in the patients with blood identified at endoscopy. However, the difference between the two groups did not achieve traditional levels of statistical significance using a two-tailed test; the p value was approximately 0.07. Because all patients in the study had bleeding within 3 days of endoscopy, the fact that the sensitivity of CLOtest was higher in those without blood seen at the time of endoscopy suggests a transient effect of blood on RUT. However, the overall group of patients (with and without blood seen at endoscopy) had a lower sensitivity with biopsy-based testing than with non-endoscopic testing, again raising the possibility that blood may decrease the diagnostic yield of all endoscopic biopsy tests. Other studies have attempted to determine whether bleeding alters the diagnostic yield of RUT in experimental models. We assessed both the CLOtest and Pyloritek in parallel endoscopic biopsies with one biopsy specimen soaked in blood and the comparator biopsy specimen not soaked in blood.20Laine L Siddom O Emami S Estrada R Cohen H. The effect of blood on rapid urease testing of gastric mucosal biopsy specimens.Gastrointest Endosc. 1998; 47: 141-143Abstract Full Text Full Text PDF PubMed Scopus (25) Google Scholar The number of positive tests for the blood-soaked and standard biopsy specimens were comparable at all times for both RUTs. Discordant results between the blood-soaked and standard specimens were seen in 17 of 400 test comparisons (4%): 8 in which only the blood-soaked specimen was positive and 9 in which only the standard specimen was positive. Thus, we concluded that “contamination” of biopsy specimens with blood does not alter RUT results. Perry et al.21Perry M Vakil N Cutler AF. Admixture with blood or ETOH does not explain false negative urease tests in H. pylori–positive active upper GI bleeders.Gastroenterology. 1997; 112 ([abstract]): A257Google Scholar assessed the in vitro effect of whole blood on two different RUT kits (CLOtest, Hpfast). When adding urease solution to the RUTs, blood enhanced the performance of RUT as compared with saline, and no false-negative results occurred, providing further support for the notion that admixture with blood cannot explain false-negative RUTs in patients with upper GI bleeding. In contrast, Lee et al.22Lee JG Lam K Solnich J Leung JW Hsu R. The effects of storage temperature, incubation conditions, and contaminants on the sensitivity of CLO, Hpfast, and Pyloritek for detection of H. pylori.Gastrointest Endosc. 1997; 45 ([abstract]): AB95Abstract Full Text PDF Scopus (2) Google Scholar reported that a mixture of blood, gastric juice, and bile decreased the sensitivity of the three different RUTs, although mixtures of blood and gastric juice or blood and bile did not affect the sensitivity. They also stated that the presence of blood obscured the color change with CLOtest and Hpfast and interfered with the test interpretation. Finally, Leung et al.23Leung WK Sung JJY Siu KLK Chan FKL Ling TKW Cheng AFB. False-negative biopsy urease test in bleeding ulcers caused by the buffering effects of blood.Am J Gastroenterol. 1998; 93: 1914-1918Crossref PubMed Scopus (64) Google Scholar examined the effect of blood on a non-commercial RUT. They reported that the color change of the pH indicator was reduced by blood and progressively suppressed by higher concentrations of serum albumin and suggested that serum albumin interferes with the color change of RUT by buffering the pH changes. This seems to contradict the findings of Lee et al.22Lee JG Lam K Solnich J Leung JW Hsu R. The effects of storage temperature, incubation conditions, and contaminants on the sensitivity of CLO, Hpfast, and Pyloritek for detection of H. pylori.Gastrointest Endosc. 1997; 45 ([abstract]): AB95Abstract Full Text PDF Scopus (2) Google Scholar who suggested that blood only decreases sensitivity in the presence of both gastric juice and bile. Clearly all of the above studies have potential flaws or shortfalls in their design, and they provide often contradictory results. The clinical studies, albeit imperfect, do raise the possibility that blood may somehow decrease the accuracy of RUT; they also raise the possibility that other biopsy tests could be decreased in sensitivity. Well-designed studies published in full manuscript form will be required to settle the issue. How can we fit this confusing jumble of studies, most presented only in abstract form, into our clinical practice? Frankly, this information should have little effect on our clinical practice, even if bleeding is documented to decrease the sensitivity of the RUT. The RUT has an excellent specificity, approaching 100%, and there is no evidence that specificity is decreased with bleeding. Thus, a positive test still can be relied on to rule in the diagnosis of H pylori and a majority of H pylori–positive patients with bleeding will still have a positive RUT. Biopsy specimens for histologic examination should always be obtained when attempting to diagnose H pylori at endoscopy because false-negative RUTs occur in 10% or more of patients. Furthermore, histologic examination provides important additional information if no evidence of H pylori is seen: when marked inflammatory cell infiltration (including neutrophils) is present, the diagnosis of H pylori infection can be made in an untreated patient with a high degree of certainty. Thus, in patients undergoing endoscopic testing for H pylori diagnosis, we recommend performing an RUT and banking biopsy specimens to be sent for histologic examination if the RUT is negative. If both RUT and histologic examination are negative and no inflammatory cell infiltration is present, we would perform no further testing; if chronic active gastritis is present we would consider this diagnostic of H pylori. If there is a question on the histologic examination (e.g., mild-moderate mononuclear cell infiltrate without neutrophils) we would also order a non-endoscopic test in patients with a bleeding ulcer because of the importance of a correct diagnosis and the possibility of a decreased sensitivity with biopsy-based testing. An analogous situation is present in patients undergoing testing after H pylori therapy. Individual endoscopic biopsy tests have lower sensitivities after therapy, leading to the recommendation that multiple biopsies and more than one method of testing be used.24European Helicobacter pylori Study Group Technical annex: tests used to assess Helicobacter pylori infection.Gut. 1997; 41: S10-S18PubMed Google Scholar Some may argue that an antibody test alone should be used to diagnose H pylori in patients with ulcer bleeding. The addition of a biopsy during endoscopy is far more costly than an antibody test (additional professional fee, additional charges for biopsy forceps and endoscopy suite) even if only the inexpensive RUT is performed. However, because of the poorer specificity of the antibody tests and the possibility that a physician may mistakenly attribute a bleeding ulcer to H pylori and stop all maintenance therapy inappropriately, we believe that the cost of endoscopic biopsy testing is justified in the high-risk and high-cost subset of ulcer patients who present with bleeding.