Clinical practice - breath tests

Clinical practice - breath tests G D Bell Gastroenterology Department, Sunderland District General Hospital, Sunderland, UK The underlying principle of the two non-invasive radio-labelled urea breath tests is similar. Both are positive when the patient's stomach is colonised by Helicobacter pylori because the organism's urease enzyme splits the orally administered urea isotope to labelled CO 2 which is then detected in the expired breath. The tests thus reflect active infection and are ideally suited to monitoring the success or failure of different eradication therapies as well as studying rates of acquisition and re-infection/late recrudescence post treatment. [ 13 C]-urea should always be used in children since it is the stable non-radioactive isotope but the [ 14 C]-urea breath test is suitable for most adults, since the dose of radioactivity is minimal. The [ 13 C]-urea breath test (UBT-13) was originally described by David Graham and colleagues from Texas in 1987 1.In writing this chapter, I have relied heavily on British experience 214. Recent reviews provide a detailed account of these tests 8-10 - 14. Urease content of H. pylori H. pylori is the urease-producing bacteria par excellence. The extraordinarily active H. pylori urease, first described by Tytgat's group from Holland 15, is the basis of the observation that patients infected with the organism tend to have higher levels of gastric ammonia and lower levels of gastric urea than uninfected individuals 16. The fact that the H. pylori urease splits urea into ammonia and carbon dioxide (via bicarbonate), and hence causes a detectable change in ph, is the basis of correspondence to: a\\ ^g so-called rapid urease tests performed upon endoscopically Professor G D Bell,,., i 117 Gastroenterology obtained gastric biopsy material 17. Department Sunderland It occurred to us and others 1-2 that if bacterial urease activity could be Royal Hospital, Medical detected using a relatively crude chemical test on unprocessed gastric sciences Faculty, bi O p SV material, how much simpler, more accurate and elegant, it would Sunderland University,, T T,...., 7 Kayii Rd, Sunderland be to measure H. pylort urease activity using some form of non-mvasive SR4 rn> UK isotopically labelled urea breath test. British Medical Bulletin 1998,54 (No. 1): 187-193 C The British Coundl 1998

Helicobacter infection Measurement of 14 CO 2 Most District General hospitals will have a liquid scintillation counter to enable the clinician to measure 14 CO 2 excretion in the patient's breath. To collect the breath, samples at time 0 and at various time intervals after oral administration of the [ 14 C]-urea substrate, the patient exhales through a tube of anhydrous calcium chloride into a vial containing 2 mmol of the CO 2 trapping agent hyamine hydroxide in 2 ml of ethanol with phenolphthalein as indicator. When 2 mmol of exhaled CO 2 has been trapped by the hyamine hydroxide, the vial is decolourised. Then, 10 ml of liquid scintillant is added and the 14 C activity measured by liquid scintillation counting. A 1% standard is also counted and the activity expressed as a percentage of the administered dose per millimole of expired CO 2. This is multiplied by body weight (kg) to allow for endogenous CO 2 production. It may be necessary for the patient to exhale into the hyamine hydroxide for 1-3 min to achieve decolonisation of the solution (and, by inference, the trapping of 2 mmol of CO 2 ). Marshall's group 18 employ only 1 mmol of hyamine in their trapping fluid, which obviously halves the time the patient is required to blow into the solution, since only 1 mmol of CO 2 is trapped. We now use 0.1 MBq of [ 14 C]-urea for our UBT-14 test having initially used 4 times this amount 2. Using the data of Stubbs and Marshall 20, the effective dose a patient might receive from a single UBT-14 would vary between 4 micro Sieverts for a H. pylori negative man to 8 micro Sieverts for a H. pylori positive woman, which is (at maximum) less then 1/500 of that a patient receives when a barium meal is performed 10. Measurement of 13 CO 2 As described in much greater detail elsewhere 10 ' 14, the natural abundance of 14 CO 2 is extremely low, while that of 13 CO 2 is relatively quite high. 13 C is a natural non-radioactive isotope and can be measured relative to 12 C by gas isotope ratio mass spectroscopy. Because 13 C is expressed as a ratio of 12 C, the volume of expired CO 2 for any given sample is not critical, and the analysis can be performed on less than 0.1 ml of exhaled CO 2. The 13 C/ 12 C ratio is usually expressed as parts per thousand (per ml) relative to an international primary standard known as PDP calcium carbonate 10 ' 14. Certain plant materials, such as corn and sugar cane, are enriched with 13 C. Thus, during the day, the natural abundance of 13 CO 2 normally fluctuates slightly in response to dietary consumption of corn, sugar cane, etc. Following ingestion of [ 13 C]-urea, patients infected by H. 188 British Medical Bulletin 1998,54 (No. 1)

14 C and 13 C urea breath tests pylori nearly always have a 13 C/ 12 C ratio in their breath which greatly exceeds that of PDP calcium carbonate, thus giving a positive result. One of the great advantages of the UBT-14 and UBT-13 tests over an endoscopically obtained, biopsy-based test other than the obvious ones of being considerably cheaper and non-invasive is that it is possible to define a cut-off between a positive and a negative result. One way to determine this arbitrary cut-off for both tests is to take the UBT result (be it UBT-14 or UBT-13) of a large number of patients whose 'gold standard' gastric biopsy results are unequivocally H. pylori negative on culture, histology and the rapid urease test and calculate a normal range plus or minus 3 standard deviations from the mean and call this the upper limit of normal. Previously, in a study described by Logan et al 13, the cut-off in adults was said to be 4.9 per ml. It has now been realised that, in many children and some adults, the true cut-off is about 3.5 per ml. As with the UBT-14 test, the amount of [ 13 C]-urea used to conduct the test has been falling. When Graham and colleagues from Texas first described the test in 1987 1, they opted to use 200 mg of [ 13 C]-urea. Now the 'Standard European Method' for the UBT-13 test employs just 100 mg of [ 13 C]-urea without loss of sensitivity or specificity 13. Our own group conducted a comparison of the UBT-14 test with the UBT-13 test using just 75 mg of the latter and found an excellent correlation between the two methods 21. Test meals versus citric acid drink and timing of breath samples for the UBT-13 and UBT-14 tests Up until very recently, most groups have agreed that giving some form of suitable test meal before administering either [ 14 C]-urea or [ 13 C]-urea was important because it delayed gastric emptying and prolonged the time that the bacterial urease is exposed to the radio-labelled urea 8 ' 10-14. Good results had been obtained in preliminary studies also using a citric acid drink 23 ' 24 instead of a standard test meal and so Malfertheiner and colleagues decided to study this more systematically 22. Eighty patients (mean age 40 years) presenting for routine OGD with dyspeptic symptoms were studied 33. All had gastric biopsies taken from both antral and body regions of the stomach and H. pylori status checked by: (i) rapid urease test; (ii) histology; and (iii) bacterial culture. Patients were classified as being H. pylori positive if either bacterial culture or both rapid urease and histology were positive. The UBT-13 test was conducted in all patients after an overnight fast on 3 consecutive days. On each study day, a different test meal was given in a randomised order: (i) 200 ml 0.1 N citric acid solution, with the addition of 25 mg British Medical Bulletin 1998;54 (No. 1) 189

Helicobacter infection Fig. 1 Curves obtained with the three 13 C- UBT procedures in the same patient population (n = 48). On each 13 C-UBT, the same 13 C- urea dose (75 mg) but different test meal was given. Results expressed as means (SEM). Reproduced from Domingguez- Munoz et ap 2 with the permission of the publishers of Gut 24 20 16 12 8 4 0 15 30 Time (min) Citric acid o Meritene Ensure/Calogen 45 60 saccharin as sweetener; (ii) 250 ml of a standard semiliquid meal (Meritene, Wander Pharma 237 kcal; 5 g fat, 20 g protein and 28 g carbohydrate); and (iii) a semiliquid fatty meal consisting of 50 ml Ensure (Abbott) and 50 ml of Calogen (SHS Pharma) which was calculated to contain 275 kcal; 26.7 g fat, 2 g protein, 6.7 g carbohydrate. At 10 min after ingestion of the test meal or citric acid drink, a baseline exhaled breath sample was collected; thereafter, 75 mg of [ 13 C]- urea dissolved in 50 ml water was given orally (time 0). Further breath samples were taken at 15, 30, 45 and 60 min. The UBT-13 test was considered positive if the delta value was 4. Of the patients, 48/80 (60%) were H. pylori positive and 32/80 (40%) H. pylori negative. All 3 breath test procedures were well tolerated, but the citric acid drink was generally considered to be the most pleasant to take 22. The citric acid drink was 24 times cheaper than any of the semiliquid meals and had the further advantage that it could easily be stored at 4 C for weeks on end. The study also confirmed it was not necessary to turn the patient on to each side or head down for a few minutes after ingestion of the labelled urea solution. Figure 1 shows curves obtained in H. pylori positive patients by applying all 3 tests 22. The maximum delta value obtained with the citric acid drink was significantly higher than that obtained with Meritene or Ensure/Calogen (see Fig. 2) and, furthermore, the delta peak was obtained earlier with the citric acid drink than the two test meals (see Fig. 3). It is presumed that the citric acid solution delays gastric emptying because of its low ph (3.0-3.5). In clinical practice, sampling at time 0 and 30 min seemed optimal and gave a high diagnostic accuracy 22. In the case of the [ 14 C]-urea breath test, we originally took breath samples at 10, 20, 30, 40, 50, 60, 80, 90 and 120 min and computed the 2 h area under the curve (2 h AUC) 2. In recent years, we have opted to use the 40 min post [ 14 C]-urea breath sample, because this gave the best correlation with the 2h AUC 25 ' 26. The cut-off we employ to separate H. pylori 190 British Medical Bulletin 1998;54 (No. 1)

14 C and 13 C urea breath tests Fig. 2 Maximal delta values (peaks) obtained with all three 13 C-UBT procedures with different test meals. Results are expressed as means (SEM). Reproduced from Domingguez-Munoz etap 2 with the permission of the publishers of Gut. a. o 30 27 24 21 18 15 12 9 6 3 0 Citric acid Meritene Ensure/Calogen Test meal positive from negative patients is a 2 h AUC of 40, which is equivalent to a 40 min value of 0.416% of administered dose/mmol CO 2 multiplied by body weight in kg 25. Using the single 40 min sample following [ 14 C]-urea, the Amsterdam group separated their H. pylori positive and negative subjects with a sensitivity of 95% and specificity of 98% 19. There is a commercially available [ 13 C]-urea breath test kit and the breath samples may be sent off to a central laboratory for analysis with a 48 h turnaround time for the results. A single UBT-13 test costs about 28 per patient compared with 14 per patient for the UBT-14 test. Clinical and research applications of urea breath testing Assessment of different anr/-helicobacter eradication regimens The greatest role for the UBTs lies in the documentation as to whether or not a particular eradication regimen has or has not successfully eradicated a patient's H. pylori infection. It can take many months, and Fig. 3 Time to delta peak for all three 13 C- UBT procedures with different test meals. Results are expressed as means (SEM). Reproduced from Domingguez-Munoz et ap 2 with the permission of the publishers of Gut. c 1 Time to <i en DU 50 r 40 30 20 10 u j- - ^^ P Citric acid < 0.0011 I JMeritane Test meal p< 0.001 NS Ensure/Calogen British Medical Bulletin 1998;54 (No. 1) 191

Helicobacter infection in some cases years, for the serology tests to become negative post H. pylori eradication and, for this reason, breath testing and not serial serology testing is the best way to monitor the efficacy of different forms of eradication therapy. Using the UBT, it became rapidly clear that many so-called successful eradications were no more than a temporary clearance 3 ' 5 and many so-called 're-infections' following apparently successful eradication were examples of prolonged suppression of the organism followed by recrudescence and not true re-infection 26. The reader is reminded that H. pylori eradication is defined as a negative test taken at least 1 month after completing the candidate course of eradication therapy. The patient should not have had any antibiotics or bismuth containing compounds in the month prior to retesting. All 3 of the PPIs currently available in the UK (omeprazole, lansoprazole and pantoprazole) can produce false negative results to the UBT test and so should be stopped at least 2 and preferably 4 weeks prior to breath testing 6-10 ' 11 ' 14. Study of H. pylori acquisition and re-infection rates The two UBT tests are ideally suited to the task of studying the rate of re-infection or late recrudescence following apparently successful eradication therapy 26. In Ipswich, we have been following a cohort of over 2000 adult patients with apparently successful eradication of their infection for periods of up to 8 years. After the first year, the re-infection rate was less than 0.25 % per patient per year 26. As discussed elsewhere in this issue, re-infection rates in small children and mentally handicapped adults living in homes may be much higher. Acknowledgements The author is grateful to Professor Peter Malfertheiner and Gut for permission to reproduce Figures 1-3 and also for the generous support of the Suffolk Branch of the BDF and the Astra Foundation. References 1 Graham DY, Klein PD, Evans DJ. Campylobacter pylori detected non-invasively by the l3 C-urea breath test. Lancet 1987; ii: 1174-7 2 Bell GD, Weil J, Harrison G et a\. M C-urea breath test analysis, a non-invasive test for Campylobacter pylori in the stomach. Lancet 1987; i: 1367-8 3 Weil J, Bell GD, Gant P, Jones PH, Trowell JE, Harrison G. 'Eradication' of Campylobacter pylori: are we being misled? Lancet 1988; ii: 1245 192 British Medial Bulletin 1998;54 (No. 1)

14 C and 1J C urea breath tests 4 Weil J, Bell GD, Powell KU et al. Heltcobacter pylori infection treated with a tripotassium dicitrato bismuthate and metronidazole combination. Aliment Pharmacol Ther 1990; 4: 651-7 5 Bell GD. Anti-Helicobacter pylori therapy: clearance, elimination or eradication? Lancet 1991; 337:311 6 Weil J, Bell GD, Powell KU et al. Omeprazole and Helicobacter pylori: temporary suppression rather than true eradication. Aliment Pharmacol Ther 1991; 5: 309-13 7 Bell GD, Powell K, Burridge SM et al. Experience with 'Triple' axvn-helicobacter pylori eradication therapy: side effects and the importance of testing the pre-treatment isolate for metronidazole resistance. Aliment Pharmacol Ther 1992; 6: 427-35 8 Bell GD, Weil J. Detection of Helicobacter pylori by the 14 C-urea breath test. In: Rathbone BJ, Heatley RV. (Eds) Helicobacter pylori and Gastroduodenal Disease. Oxford: Blackwell, 1992; 74-87 9 Bell GD, Powell KU, Burridge SM et al. Helicobacter pylori eradication: efficacy and side effect profile of a combination of omeprazole, amoxycillin and metronidazole compared with 3 different forms of bismuth containing triple therapy. QJM 1993; 86: 743-50 10 Bell GD. Urea breath tests. In: Northfield TC, Mendall M, Coggin PM. (Eds) Helicobacter pylori Infection. Pathophysiology, Epidemiology and Management. Dordecht: Kluwer, 1993; 127-38 11 Bell GD, Powell KU, Burridge SM et al. Rapid eradication of Helicobacter pylori in ulcer disease. Aliment Pharmacol Ther 1995; 9: 41-6 12 Logan RPH, Poison RJ, Misiewicz JJ et al. Simplified single sample u Carbon urea breath test for Helicobacter pylori: comparison with histology, culture and ELJSA serology. Gut 1991; 32: 13 Logan RPH, Dill S, Bauer FE et al. The European 13 C-urea breath test for the detection of Helicobacter pylori. Eur ] Gastroenterol Hepatol 1991; 3: 915-21 14 Logan RPH. Detection of Helicobacter pylori by the 13 C-urea breath test. In: Rathbone BJ, Heatley RV. (Eds) Helicobacter pylori and Gastroduodenal Disease. Oxford: Blackwell, 1992; 88-106 15 Langenberg ML, Tytgat GN, Schipper MEI et al. Campylobacter-likt organisms in the stomachs of patients and healthy individuals. Lancet 1984; i: 1348 16 Marshall BJ, Langton SR. Urea hydrolysis in patients with Campylobacter pyloridis infection. Lancet 1986; i: 965-6 17 McNulty CAM, Wise R. Rapid diagnosis of Campylobacter associated gastritis. Lancet 1985; ii: 1443-4 18 Marshall BJ, Surveyor I. Carbon- 14 urea breath test for the diagnosis of Campylobacter pyloriassociated gastritis. } NucI Med 1988; 29: 11-6 19 Rauws EAJ, Royen EAV, Langenberg W et al. 14 C-urea breath test in C. pylori gastritis. Gut 1989; 30: 798-803 20 Stubbs JB, Marshall BJ. Radiation dose estimates for the carbon 14-labelled urea breath test. / Nucl Med 1993; 34: 821-5 21 Bell GD, Powell K, Weil J etal. u C-urea breath test for Helicobacter pylori infection. Gut 1991; 32: 551-2 22 Domingguez-Munoz JE, Leodolter A, Sauerbach T, Malfertheiner P. A citric acid solution is the optimal test drink in the 13 C-urea breath test for the diagnosis of Helicobacter pylori infection. Gut 1997; 40: 459-62 23 Braden B, Duan LP, Caspary WF, Lembecke B. More convenient 13 C-urea breath test modifications still meet the criteria for valid diagnosis of Helicobacter pylori infection. Z Gastroenterol 1994; 32: 198-202 24 Eggers RH, Kulp A, Tegler R et al. A methodological analysis of the 13 C-urea breath test for detection of Helicobacter pylori infections: high sensitivity and specificity within 30 min using 75 mg of 13 C-urea. Eur j Gastroenterol Hepatol 1990; 2: 437-44 25 Weil J, Bell GD, Harrison G. 14 C-urea breath test for C. pylori. Gut 1989; 30: 1656-7 26 Bell GD, Powell KU. Helicobacter pylori reinfection after apparent eradication - the Ipswich experience. Scand J Gastroenterol 1996; 31 (Suppl 215): 96-104 British Medical Bulletin 1998,54 (No. 1) 193