Article Text

13C urea breath test (UBT) in the diagnosis of Helicobacter pylori: why does it work better with acid test meals?
1. D Pantoflickova1,
2. D R Scott2,
3. G Sachs2,
4. G Dorta1,
5. A L Blum1
1. 1Gastroenterology Department, CHUV, Lausanne, Switzerland
2. 2Department of Physiology, UCLA School of Medicine, Los Angeles, CA and CURE: Digestive Diseases Research Center, UCLA and VA Greater Los Angeles Health Care System at West LA, Los Angeles, CA, USA
1. Correspondence to:
Dr D Pantoflickova, Department of Gastroenterology, CHUV-BH10, Switzerland;
dpantofl{at}hotmail.com

## Abstract

Background: Acid test meals may improve the accuracy of the 13C urea breath test (UBT). This has been attributed to changes in gastric emptying rather than to the effects of gastric pH on Helicobacter pylori urease.

Aims: To determine whether enhancement of 13CO2 excretion in the UBT in H pylori infected volunteers by acidification of a test meal is due to a delay in gastric emptying.

Methods: Urease activity in vitro was measured in intact bacteria and in bacterial homogenates. Urease activity in vivo was assessed by means of the UBT. Eleven H pylori infected subjects underwent UBTs with neutral Ensure (pH 7.0), acidified Ensure (pH 3.0), and apple juice (pH 3.0). Gastric emptying was assessed by 13C sodium acetate breath test.

Results: From pH 7 to pH 3, the in vitro urease activity of intact bacteria increased sixfold. In contrast, urease activity of bacterial homogenates was inactivated by low pH. In vivo, urease activity, as measured by the UBT 20 minutes after meal ingestion, was higher with apple juice (δ 13CO2=21.1; p=0.03) and acidified Ensure (δ 13CO2=25.5; p=0.01) than with neutral Ensure (δ 13CO2=12.5). Gastric emptying was faster with apple juice (Tmax=36.7 (8) minutes) but not with acidified Ensure (Tmax=63.3 (5) minutes; p=0.06) than with neutral Ensure (Tmax=65.0 (3) minutes; p=0.04).

Conclusions: The higher UBT found with acidified compared with neutral test meals was independent of the emptying rates of the test meals but may have been due to medium acidity dependent activation of intra-bacterial urease in intact H pylori.

• Helicobacter pylori
• urea breath test
• gastric emptying
• gastric acidity
• urease activity
• UBT, 13C urea breath test
• DOB, delta over baseline
• Tmax, time of maximal gastric emptying rate
• T 1/2, half emptying time

## Statistics from Altmetric.com

The 13C urea breath test (UBT) is widely used for determination of H pylori infection.1 On average, this test is quite accurate2 but false negative results are obtained in conditions which are frequently encountered in H pylori infected subjects, such as atrophic gastritis3 and acid suppression by acid pump inhibitors.4–8 It was argued that the accuracy of the UBT may be improved by prolonging the contact of the test meal with H pylori urease.9,10 As citric acid slows gastric emptying and improves the accuracy of the UBT, it was concluded that citric acid had a favourable effect because of its slowing effect on gastric emptying.9–11 Acidification of test meals has become a standard recommendation.1

In this study we examined an alternative mechanism for the effect of acidification by citric acid. The activity of intra-bacterial urease, which represents the major compartment of H pylori urease,12 depends on the medium pH, as medium acidity activates urea entry via activation of UreI, the urea channel in the inner membrane of H pylori. This increased urea entry in acidic medium results in increased urease activity,13,14 and hence increased acidity of the bacterial environment is predicted to increase the UBT.

## MATERIALS AND METHODS

### Assessment of H pylori urease activity at different pH values in vitro

#### Growth of H pylori

H pylori strain ATCC43504 was grown on blood agar plates (BBL TSA 5% sheep blood; Becton-Dickenson, Franklin Lakes, New Jersey, USA) in a microaerobic atmosphere (5% O2, 10% CO2, 85% N2) at 37.0° C for 24 hours. Cells from one plate were harvested in 300 μl of 1 mM phosphate buffer, pH 7.0, from which aliquots were taken for the urease assay described below.

#### Urease assay in vitro

The method chosen was release of radioactive 14CO2 from 14C urea to resemble the UBT, as previously described.15 The incubation medium contained 5 mM labelled 14C urea with a specific activity of 10 μCi/μmol. In the experiments we used 100 mM sodium phosphate buffer at pH 3.0, 5.0, and 7.0 containing NaCl 138 mM, KCl 5 mM, CaCl2 1 mM, MgCl2 2 mM, glucose 10 mM, and glutamine 1 mM. pH was adjusted with HCl. pH did not change over the 30 minute measurement period carried out at 37°C and urease activity was linear with time. Incubation was carried out in a tube sealed with a rubber stopper that contained a well with a filter paper soaked in 0.5 N KOH. 14CO2 generated was trapped in the well following injection of 5 N H2SO4 into the medium. Counting was done in an LKB scintillation counter. Protein determination was performed by the Lowry method.16 Urease activity was measured as a function of varying medium pH, both in a bacterial homogenate generated by French press treatment and in intact bacteria. The former gives total urease activity, the latter activation of intra-bacterial urease.

### Assessment of H pylori urease activity at different pH values in vivo

#### Study population

Eleven H pylori positive subjects (eight men, three women; age range 26–36 years) participated in this crossover randomised study. Eleven H pylori negative subjects served as controls (six men, five women; age range 23–36 years). All were healthy with no history of gastrointestinal disease or other illness, and no current gastrointestinal symptoms. The volunteers were not taking any medications other than oral contraceptives. Subjects who had taken antibiotics, proton pump inhibitors, H2 receptor antagonists, and bismuth containing preparation in the four weeks before enrolment were excluded from the study. The study was conducted according to the Declaration of Helsinki and the rules of Good Clinical Practice.

#### Diagnosis of H pylori infection

The presence or absence of H pylori infection was diagnosed at enrolment by an enzyme linked immunosorbent assay (ELISA; Roche, Switzerland) and a standard UBT. Fasting blood samples were collected for determination of specific IgG anti-H pylori antibodies. The standard 13C UBT was performed using 100 mg of 13C urea and orange juice as a test meal, as previously described.2 Baseline and 30 minute breath samples were analysed by isotope ratio mass spectrometer (PDZ Europa, Crewe, UK). The diagnostic cut off value for the standard 13C-UBT was δ 13CO2=5.0% delta over baseline (DOB).17

#### Study protocol

Each volunteer underwent three different UBTs in a crossover randomised design. After an overnight fast, the following test meals were administered: standard Ensure (237 ml, 250 kcal, 9 g protein, 40 g carbohydrate, 6 g fat; pH 6.8), Ensure acidified with 4 g of citric acid (237 ml, 250 kcal, 40 g carbohydrate, 9 g protein, 6 g fat; pH 3.0), and apple juice (200 ml, 45 kcal, 11 g carbohydrate, 0 g protein, 0 g fat; pH 3.0). A commercial apple juice was used. Its composition, as stated by the manufacturer, was: citric acid (4 g), ascorbic acid (2 g), malic acid (2 g), and small amounts of isocitric acid and cis-acotinic acid with a resulting pH of 3.

Breath samples were collected into plastic breath collectors (Alimenterics, Inc., New York, USA) at baseline and at 10 minute intervals over 60 minutes after administration of the test meal with 13C urea. 13C labelled urea (99%) was administered in a solution of distilled water (5 ml) immediately after ingestion of a test meal. 13C/12C ratio was measured by the Laser Assisted Ratio Analyzer (LARA System; Alimenterics Inc.).17 Values are expressed as δ excess over baseline per ml. The diagnostic cut off value for our laboratory is δ 13C02=5.0% DOB.17

### Gastric emptying studies

Gastric emptying of test meals was assessed by 13C sodium acetate breath test in 11 H pylori positive volunteers. Breath samples for gastric emptying studies were obtained at baseline and at 10 minute intervals over 120 minutes after administration of a test meal with 150 mg of 13C labelled sodium acetate.1813CO2 excretion was measured, as above, by the Laser Assisted Ratio Analyzer System (Alimenterics Inc.). The results were expressed as percentage of 13C dose recovered (13C-PDR). 13C-PDR was calculated according to the formula of Schoeller and colleagues.19

$Math$ $Math$

where 13δ = ((Rt/RPDB) −1) ×103 − ((R0/RPDB)−1) ×103 (Rt=13C:12C in the sample at time t; R0=13C:12C in the sample at baseline; RPDB=13C:12C in international standard PeeDeeDelemnite=0.0112372); CO2 is CO2 production rate; D is the dose of substrate administered; Mr is molecular weight of the substrate; P is 13C atom % excess; and n is the number of labelled carbon positions.19 Total CO2 production rate was assumed to be 5 mmol/m2 of body surface/min.

Body surface area was calculated according to the formula of Haycock and colleagues.20

$Math$

where H is height (m) and W is weight (kg).

### Statistical methods

DOB values obtained in UBT studies were analysed using Sigma Stat 2.0 software (SPSS Inc., Chicago, Illinois, USA). The Komolgorov-Smirnov test was applied to test the normality of the distribution of the data. As the data were not normally distributed, the overall effect of different test meals on 13CO2 excretion was assessed by the Friedman repeated measures analysis of variance on ranks. To isolate the group or groups that differed from the others, we used Dunnett’s method as the all pairwise comparison procedure. Pairwise comparisons were made using the Wilcoxon signed rank test. Data are presented as group median values with interpercentile range (10–90 centiles). Values of p<0.05 were considered significant.

The data obtained by the 13C acetate breath test were fitted by a mathematic formula using the least squares procedure.18 The formula is given by:

$Math$

where y is the percentage of 13CO2 excretion per hour; t is time in hours; and a, b, c are constants.

Cumulative 13C recovery data were fitted to the formula:

$Math$

which is the inverse of the scintigraphic radioactivity decay curve.21 From these constants, the time of maximal gastric emptying rate Tmax = b/c and the half emptying time T½ = −1/k ln(1–2−1/β), were calculated. Non-linear regression analysis of these formulas was performed using a commercial software program (NLREG 5.0, Brentwood, Tennessee, USA).

## RESULTS

### In vitro urease activity at different medium pH values

The results of the urease assays at pH values of 7.0, 5.0, and 3.0 are shown in fig 1. Urease activity of intact bacteria is low at pH 7.0, rises steeply at pH 5.0, and remains essentially the same at pH 3.0. The opposite is true for urease activity assayed in bacterial homogenates where activity is maximal at pH 7.0, declines steeply at pH 5.0, and is absent at pH 3.0. These data are in accordance with previously published results showing that the enzyme has a pH optimum close to neutrality and is irreversibly inactivated below pH 4.5 in bacterial homogenates.14 In contrast, urease activity of the intact bacteria increases with acidification of the medium due to increased urea permeability via UreI, allowing a large increase in urea access to intra-bacterial urease.14

Figure 1

In vitro urease activity in intact (n=3) and homogenised Helicobacter pylori (n=3). In the intact organism, urease activity increased sixfold as the medium pH declined whereas urease activity of the bacterial homogenate decreased with a fall in medium pH. Values are mean (SEM).

### 13C UBT studies

All 11 H pylori negative and 11 H pylori positive subjects completed the study. Group median δ values are shown in fig 2. Median 13CO2 excretion rates at 10, 20, and 30 minutes were lower with standard Ensure at a pH of 6.8 than with acidified Ensure and apple juice, both initially at pH 3 (p<0.05) (fig 2). There were no significant differences in 13CO2 excretion between the three 13C UBT protocols at 40 minutes (fig 2). The amount of 13CO2 excretion was higher with Ensure than with both acidic test meals only during the 50 minute and 60 minute periods of the study (p<0.05) (fig 2). Hence the major effect of acidification occurred between 0 and 30 minutes, presumably because of buffering of the administered acidity, blunting the stimulation of urea entry, with buffering by the citric acid in the acidified Ensure being more effective than buffering in apple juice.

Figure 2

In vivo urease activity with test meals (see methods) in Helicobacter pylori positive and negative subjects. Median δ 13CO2 values of H pylori positive subjects are shown in table 1. *p<0.05 versus acidified Ensure pH 3.0; †p<0.05 versus apple juice and acidified Ensure pH 3.0. Filled symbols depict values in H pylori positive subjects; open symbols in H pylori negative subjects.

All H pylori negative subjects had 13CO2 excretion values below the cut off value of 5.0% DOB at all collection times (fig 2).

### Gastric emptying studies

The results are shown in fig 3. Gastric emptying was faster with apple juice (Tmax=36.7 (8) minutes) but not with acidified Ensure (Tmax=63.3 (5) minutes; p=0.06) than with neutral Ensure (Tmax=65.0 (3) minutes; p=0.04).

Figure 3

Gastric emptying (see methods) of the three types of test meals. Data are medians of 11 Helicobacter pylori positive subjects. Arrows indicate the time of Tmax. *p=0.03 and 0.04 versus Ensure pH 6.8 and acidified Ensure pH 3.0, respectively. 13C-PDR, percentage of 13C dose recovered.

Similarly, the T½ of gastric emptying was lower with apple juice (71.7 (9) minutes; p<0.05) than with both Ensure meals (acidified Ensure 103.4 (7) minutes; neutral Ensure 105.6 (6) minutes) (fig 3). Acidification of the Ensure meal did not slow gastric emptying, contrary to the hypothesis explaining the effect of citric acid on CO2 excretion9.

## DISCUSSION

We have confirmed previous observations in H pylori infected subjects10,11,22 that had shown that acidification of a test meal by citric acid enhanced 13CO2 excretion in a UBT. Apple juice which contains a variety of organic acids23 also increases 13CO2 excretion.

It has been known since the classical studies by Hunt and Knox,24 and recently been described by Shiotani and colleagues,25 that organic acids, in particular citric acid, slow gastric emptying in a dose dependent manner. This has led to the hypothesis that citric acid increases 13CO2 excretion because it slows gastric emptying and thus prolongs contact of the test meal with H pylori and its urease.9,10,22

To test the gastric emptying hypothesis, we performed gastric emptying tests in H pylori positive volunteers who had undergone UBTs. Ensure meals were emptied more slowly than apple juice which had a pH identical to the acidified Ensure meal. The emptying rates of the neutral and Ensure test meal acidified with citric acid were identical. The difference in gastric emptying between Ensure test meals and apple juice is explained by the different caloric density of the test meals.26 Ensure as a meal with higher caloric density empties less rapidly than apple juice. Thus when using a test meal of comparable caloric content but different pH, the rate of gastric emptying is determined by its caloric density and not by its pH. The observations reported here show that slowed gastric emptying does not explain enhancement of the UBT induced by acidification of the test meal.

Our in vitro results suggest that the pH at the gastric surface activates H pylori urease by activating urea entry via UreI, and that this effect is the main determinant of the acidity dependent 13CO2 excretion rate in vivo.

In vitro, external acidity leads to stimulation of urease activity in intact bacteria (fig 1) due to facilitation of urea access across the bacterial inner membrane to intra-bacterial urease via the proton gated urea channel UreI.13,14 Thus when acid in the form of apple juice or acidified test meals is given to H pylori infected volunteers, urease is stimulated due to acidification of the bacterial environment, and increased 13CO2 excretion in vivo occurs rapidly (fig 2) compared with neutral pH meals. As similar 13CO2 excretion was observed with an Ensure test meal containing only citric acid as an acidifier and apple juice, a test meal which had an identical pH but contained a large variety of organic acids, especially malic acid,23 we suggest, in contrast with previous authors,9 that it is the pH and not the type of organic acid that determines the magnitude of the 13CO2 response.

It has previously been observed that administration of neutral Ensure led to an increase in intragastric pH to ∼4.0 and to increased plasma gastrin levels.27 One hour after this neutral Ensure meal, intragastric pH started to fall. Hence maximal stimulation of intra-bacterial urease activity would be delayed by one hour compared with acidified Ensure.27 This explains the rapid increase in 13CO2 excretion with acidified Ensure compared with neutral Ensure, as observed here.

Prior studies have shown a dose-response effect with citric acid but no dose-response effect with ascorbic acid solutions in spite of their similar pH values.9 In addition, endogenous stimulation of acid secretion by pentagastrin appeared not to have an effect on urease activity.9 It has therefore been argued that acid does not play a direct role in 13CO2 excretion in UBTs.28 The lack of effect of ascorbic acid compared with citric acid can be explained by the lower buffering capacity of the monocarboxylic ascorbate compared with the tricarboxylic citric acid. In addition, ascorbic acid is a potent inhibitor of a variety of bacterial ureases, including H pylori urease.29 Hence the dose dependent effect of citric acid can be explained by its increased buffering capacity and its ability to maintain a low pH in the bacterial environment30 and perhaps also by its ability to cross the diffusional resistance of the gastric mucus at higher concentrations.31,32 Pentagastrin injection increases gastric acid secretion but does not necessarily reduce pH at the site of bacterial colonisation. Juxtamucosal pH has been claimed to be maintained at a relatively neutral pH during pentagastrin stimulated acid secretion pH 1, unless acid is applied intraluminally.33 Administration of pentagastrin also leads to increased mucus gel thickness.34

The UBT may give false negative data in H pylori infected subjects with atrophic gastritis3 and, in particular, during treatment with proton pump inhibitors.4–8 When these subjects undergo UBTs with meals containing citric acid, the diagnostic accuracy of the test is improved.10,11,22 It has been argued that a false negative UBT in subjects on proton pump inhibitor therapy is due to a decreased bacterial load.28 However, more likely, as proton pump inhibitor therapy increases intragastric pH, intra-bacterial urease activity will decrease.35 Intragastric titration from pH 3 to 5 and to 7 led to a decrease in urease activity in vivo, as assessed by the 13C-UBT.36

Acidification increased H pylori load in gastric aspirates from chronic proton pump inhibitor treated subjects.37 Chronic proton pump inhibitor treatment may therefore result in a decreased bacterial load. However, immediate improvement in the rapid urease test at pH 7.0 found with acidification cannot be due to an immediate change in the bacterial load.37–39 Thus the explanation for the decrease in UBT during proton pump inhibitor treatment is more likely due to the pH increase induced by proton pump inhibitors and a decrease in intra-bacterial urease activity because of inactivation of urea entry into the organism. After chronic therapy, a decrease in the number of bacteria may also contribute to the decrease in UBT. Acidification of the test meal is therefore desirable for accurate UBT results, particularly in subjects on proton pump inhibitor therapy.

In conclusion, we suggest that activation of the cytoplasmic urease by urea entry into the cytoplasm of H pylori rather than gastric emptying may explain the increased 13CO2 excretion when an acid test meal is given.

Table 1

Median δ 13CO2 values (10–90 centiles) of Helicobacter pylori positive subjects

## Acknowledgments

These studies were supported in part by NIH grant Nos DK53642, DK41301, DK46917, and DK58333, and by the Swiss National Fund No 31-43240.98.

View Abstract

## Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.