Non-invasive (non-endoscopic) tests

Serology (laboratory enzyme-linked immuno-sorbent assay [ELISA]) is widely available and relatively inexpensive but is not the test of choice in most circumstances. Serologic testing can determine whether a patient has had H. pylori infection in the past, but cannot definitively identify the presence of current infection. A positive result may be obtained even after the infection has cleared. The sensitivity and specificity are about 80%; consequently, serology risks missing approximately one of every five cases of H. pylori infection or, conversely, administering unnecessary treatment to patients who are no longer infected.

Urea breath testing is an excellent non-invasive method of assessing the presence of H. pylori via the urease activity of the organism (Fig 5.1). In a patient with infection, the patient ingests urea labelled with either the isotope ¹³C (which is non-radioactive) or the isotope ¹⁴C (using a low dose so there is low radioactivity). If infection is present, there is breakdown of the urea via urease produced by the bacteria; the excessive labelled carbon dioxide produced can be quantified in expired breath. The ¹⁴C test is contraindicated in pregnant women and children. This technique offers excellent positive and negative predictive values, regardless of the background prevalence of H. pylori. Moreover, the test is useful both before and after treatment. Accurate results are also more often achieved in the setting of upper-gastrointestinal bleeding with this test. Sensitivity is reduced by medications such as proton pump inhibitors, bismuth and antibiotics, which decrease the H. pylori density on which the test is based. Patients should be off acid suppression medications for at least 1 week before testing by the urea breath test for infection if possible; if the test is negative and the patient is or recently has been on acid suppression drugs, testing with an alternative method or repeating the test later should be considered.

Figure 5.1 Schematic representation of the ¹⁴C and ¹³C breath test. After ingestion of labelled urea, urea is metabolised by urease and labelled CO₂ is produced. This is exhaled via the lungs and the concentration of labelled carbon atoms can be determined in the exhaled air. The value at 20 or 30 min may be used to define a positive urea breath test (upper curve). An early peak can occur in both Helicobacter pylori-positive and -negative cases from hydrolysis of urea by mouth flora.

The faecal antigen test is excellent (and equivalent in terms of accuracy to the urea breath test). As with other types of testing, the faecal antigen test can produce false-negative results in the presence of proton pump inhibitor therapy.

Endoscopic tests

H. pylori infection can be diagnosed at endoscopy by a screening test, the rapid urease test, which has a high sensitivity and specificity (over 95%) (Fig 5.2). It involves placing a biopsy in a pH-sensitive medium containing urea. Urease produced by the bacteria splits the urea to ammonia that changes the pH and produces a colour change. The various commercially available biopsy urease tests have comparable diagnostic accuracy, with excellent specificity (93% to 100%) and very good sensitivity (89–98%). It should be noted, however, that the sensitivity is reduced in patients who have taken proton pump inhibitors within the preceding 1–2 weeks. In addition, biopsy testing may yield false-negative results in patients with ulcers that are actively bleeding, and therefore all bleeding ulcer patients should be re-evaluated, preferably using urea breath testing if initially biopsy tests for the infection are negative.

Figure 5.2 A rapid urease test—a gastric biopsy is placed in the pH-sensitive medium. Top and bottom tests: no colour change. Middle test: change to magenta within 3 hours, indicating the presence of urease (and hence H. pylori).

Gastric histology offers excellent sensitivity and specificity, but is much more costly than biopsy urease testing. Moreover, the accurate detection of H. pylori depends on the site, number, and size of gastric biopsies. To optimise the diagnostic yield, multiple biopsy samples should be obtained from both the antrum and gastric body. All samples can be placed into a single container to reduce costs. As in the case of biopsy urease testing, the sensitivity of histology is decreased in the presence of proton pump inhibitor therapy such that the diagnosis of H. pylori infection will probably be missed in more than one-third of patients. The main advantages of this technique are its ability to check for pathological changes associated with infection (such as atrophy or intestinal metaplasia) and to diagnose mucosa-associated lymphoid tissue (MALT) lymphoma. Usually, the gastroenterologist will take samples for histology if there is cause for concern based on the visual findings or if a gastric ulcer is present.

Culture is not routinely available, but is valuable for examining patterns of antimicrobial resistance in patients who have failed multiple courses of therapy for H. pylori infection.

Testing initially

Testing for H. pylori should be conducted only if the clinician is planning to offer treatment to all patients who have positive results. Any patient with a history of peptic ulcer or gastric cancer (including MALT lymphoma) must have H. pylori excluded if not already checked.

Asymptomatic patients are not usually tested except in special situations. If a patient has a strong family history of gastric cancer, testing can be performed to reassure the patient and treat if infected. Patients about to be started on a long-term non-steroidal anti-inflammatory drug (NSAID) or low-dose aspirin may be offered testing, as eradication of H. pylori in this setting reduces (but does not abolish) the risk of subsequent peptic ulcer. Patients with unexplained documented iron deficiency anaemia after a comprehensive evaluation (to exclude other causes including colonic pathology) may be offered testing and treatment for H. pylori as this can resolve the problem. H. pylori has also been linked to immune thrombocytopenic purpura, so testing in this situation may be considered. H. pylori testing is not recommended for patients with symptoms of gastroesophageal reflux disease (e.g. heartburn) or for oesophagitis.

Testing post-therapy

Once treatment of H. pylori has been completed, universal testing to confirm eradication (i.e. cure) of the infection is neither cost-effective nor practical. Current guidelines recommend post-treatment testing to confirm eradication in specific groups of patients: namely, those with H. pylori–associated peptic ulcer, persistent dyspepsia after treatment (as you need to decide whether to try retreating the infection or offer another empirical treatment approach), those with MALT lymphoma, and those who have undergone endoscopic mucosal resection for early gastric cancer. The urea breath or stool antigen test is recommended in these settings at least 4 weeks after H. pylori therapy has been completed (re-testing earlier gives false negative results). Serology cannot be used to confirm eradication of H. pylori infection.

Treatment failure

If first-line treatment fails, the infection is usually more difficult to eradicate. The best chance of success is with the first prescription given. Normally if first-line therapy fails a different regimen is prescribed next.

The most important predictors of treatment failure are poor adherence to prescribed regimens and antibiotic resistance. Adherence may be limited by the complexity of treatment regimens, the patient’s incomplete understanding of the rationale for treatment, the side effects of medications, or the patient’s assumption that the medication can be stopped once the symptoms have improved. Promotion of good adherence is critical, as rates of successful eradication of H. pylori are falling. Effective communication with patients is the key to success. The clinician should emphasise that medications must be taken exactly as prescribed to minimise the risk of treatment failure. The patient should also be given an explanation of how non-adherence can lead to the development of antibiotic resistance.

Antibiotic resistance is a crucial determinant of treatment outcome. Trends document increasing rates of clarithromycin resistance and relatively stable but high rates of metronidazole resistance. Regardless of the regimen, however, the likelihood of H. pylori resistance is elevated in patients previously treated with either metronidazole or a macrolide.

If two attempts to treat the infection fail, specialist consultation is advised. Culture of the organism and testing for antibiotic sensitivities may be considered. Salvage treatment regimens include (1) a 10-day course of a standard dose of proton pump inhibitor, amoxicillin (1 g), and levofloxacin (250 mg), all given twice daily, or (2) a proton pump inhibitor twice daily, rifabutin (300 mg once daily), and amoxicillin (1 g twice daily) for 10 days.