Helicobacter DNA in bile: correlation with hepato-biliary diseases


Correspondence to: Dr C. A. Fallone, Division of Gastroenterology, Room R 2.28, McGill University Health Centre, Royal Victoria Hospital, 687 Pine Avenue, West Montreal, Quebec, H3A 1A1, Canada. E-mail: cfallone@lan1.molonc.mcgill.ca


Background : Helicobacter has been identified in isolated cases of hepato-biliary diseases, but its role in the pathogenesis of these conditions remains unclear.

Aim : To determine whether Helicobacter could be detected in bile obtained at endoscopic retrograde cholangiopancreatography, and to evaluate the prevalence of this infection in patients with hepato-biliary diseases.

Methods : Bile was collected from 125 patients with various hepato-biliary diseases undergoing endoscopic retrograde cholangiopancreatography. Among them, 75 were diagnosed with biliary stones, 15 with pancreatico-biliary malignancies and four with primary sclerosing cholangitis. The detection of Helicobacter in DNA extracted from these bile samples was performed using Helicobacter genus-specific primers (capable of detecting 100–1000 organisms/mL).

Results : Helicobacter was detected in all positive controls. Only three samples had polymerase chain reaction inhibitors. All remaining bile samples (122 patients with hepato-biliary diseases) were negative for Helicobacter DNA.

Conclusions : Helicobacter can be detected in bile samples using polymerase chain reaction. This infection, however, was not present in any of our patients diagnosed with gallstones or hepato-biliary malignancies, raising doubt as to the possible association between Helicobacter and these entities. Given the low sample size of patients with primary sclerosing cholangitis, more studies are required to determine whether an association exists with this condition.


Helicobacter pylori is a Gram-negative, spiral-shaped, motile micro-organism that has been shown to play a causative role in the pathogenesis of chronic gastritis and peptic ulcer disease. 1 Infection with this bacterium has also been linked to the development of gastric adenocarcinoma and mucosa-associated lymphoid tissue lymphoma. 2–4 Numerous other Helicobacter species have subsequently been isolated from sites other than the stomach, including the gastrointestinal tract, liver and biliary tree of animals. H. hepaticus and H. bilis have been detected in liver tissue from mice with chronic active hepatitis and hepatocellular cancer. 5,   6 Similarly, H. canis has been isolated from the liver of a dog with multifocal necrotizing hepatitis. 7

In humans, H. pylori DNA has been detected and partially sequenced in bile samples obtained via trans-hepatic aspiration.8 In this study, H. pylori DNA was found in three patients with either gastric or pancreatic cancer. Based on the H. pylori urease B gene and immunohistochemical staining, Kawaguchi et al. also demonstrated an organism closely resembling H. pylori in a resected gall-bladder specimen.9 The possibility of Helicobacter infection as a risk factor for gallstone formation has been suggested by Figura et al., who showed that H. pylori-infected patients have specific antibodies and some also have H. pylori antigens and genomic material in bile.10, 11 Fox et al. identified H. bilis and H. pullorum in bile and gall-bladder tissue from Chileans with chronic cholecystitis, adding further support to this hypothesis.12 Recently, H. pylori (confirmed with Southern blot hybridization or partial 16S sequencing) was found by polymerase chain reaction (PCR) in liver samples from isolated cases of patients with primary sclerosing cholangitis, primary biliary cirrhosis and various other chronic liver diseases.13–15 However, the prevalence of this infection in patients with such conditions, and the pathogenic significance of these findings, remain unknown.

The objectives of this study were to determine whether Helicobacter species could be detected using PCR on bile obtained from endoscopic retrograde cholangiopancreatography, and to evaluate the prevalence of this infection in patients with hepato-biliary diseases, such as cholelithiasis and pancreatico-biliary malignancies. This would allow the evaluation of the potential pathogenic role of Helicobacter species in hepato-biliary diseases in humans.



Consecutive patients undergoing endoscopic retrograde cholangiopancreatography for various indications at three McGill University teaching hospitals (Montreal General, Royal Victoria and Jewish General Hospitals) were asked to participate in this study. Patient demographics and diagnoses were recorded and bile samples were collected during the procedure. Bile was aspirated after selective cannulation of the common bile duct, and stored at − 20 °C. The protocol and informed consent form were approved by the institutions' Research and Ethics Committee.

DNA extraction

Patient bile samples were thawed at room temperature; 1 mL of each specimen was diluted with an equal amount of sterile distilled water, centrifuged at 18 000 g for 15 min, and the supernatant was discarded. This has the effect of concentrating any bacterial cells in the pellet, and of removing some of the inhibitors present in bile. The pellet was subsequently mixed with 200 µL of sterile distilled water and run through the FastDNA kit (Bio101 Inc, Q-BIOgene, Carlsbad, CA, USA) as per the manufacturer's instructions. The DNA thus extracted was stored at 4–8 °C until PCR testing.

In order to determine the lower limit of detection of the extraction protocol, bile was spiked with serial dilutions of live H. pylori cells (ATCC strain 43504). Although this is not an entero-hepatic Helicobacter, as we tested for genus-specific rather than species-specific Helicobacter, it serves as a good control. This strain had previously been grown on charcoal-containing medium, under microaerophilic conditions, and cells were suspended in phosphate-buffered saline. A standard suspension of 108 colony-forming units (cfu)/mL (estimated by McFarland standards) was prepared. Ten-fold serial dilutions of the H. pylori cell suspension were made and used to spike bile samples to final concentrations of 1–106 cfu/mL of bile. These spiked samples were then mixed with equal volumes of water and centrifuged at 13 500 r.p.m. In order to ensure that no Helicobacter cells or DNA was lost with the discarded supernatant, the supernatant from these spiked samples was kept and processed in parallel. DNA extraction was performed for both pellet and supernatant from these spiked samples, in the manner described above. The entire process was also tested without prior centrifugation. Patient bile samples were batched; with each batch, positive extraction controls consisting of bile spiked with 102, 103 and 104 cells of H. pylori were included, in addition to negative controls consisting of water only.

DNA amplification

Template DNA was subjected to PCR using Helicobacter genus-specific primers C97 (5′-GCTATGACGGGTATCC-3′) and C98 (5′-GATTTTACCCCTACACCA-3′), which amplify a 398-base pair segment within the conserved 16S rRNA gene. The PCR mix (50 µL) contained 0.1 mm of each of the four deoxynucleotide triphosphates (dNTPs), 2 µm of each of the two primers, 1.5 mm MgCl2, 2 U taq polymerase and 10× reaction buffer (both from Sigma Chemicals, St Louis, MO, USA) and 5 µL of template DNA. The PCR was conducted in a Gene Amp PCR system 2700 (Applied Biosystems, Foster City, CA, USA) under the following conditions: 94 °C for 5 min; 35 cycles at 94 °C for 1 min, 58 °C for 1 min and 72 °C for 3 min. The PCR products were separated in a 1% agarose gel stained with ethidium bromide. The expected PCR products were visualized with a gel documentation system (Syngene, Frederick, MD, USA).

Testing for the presence of PCR inhibitors in bile

After each of the patient samples had been tested by PCR, a second set of experiments was performed in order to ensure that inhibition of PCR did not occur. DNA extracted from each of the patient bile samples was spiked with a standard concentration of genomic DNA, and subsequently tested with PCR for the presence of Helicobacter DNA.

Statistical analysis

The proportion of patients with hepato-biliary diseases with Helicobacter detected in bile serves as an estimate of the prevalence of this infection in these conditions. Ninety-five per cent confidence intervals based on this estimate were calculated using the normal approximation of the binomial distribution. We could not base our sample size estimates on results from the literature, as pilot data analysis demonstrated the absence of any Helicobacter DNA. We thus calculated the required sample size based on the 95% confidence interval whose upper margin ruled out a clinically important percentage association between pancreatico-biliary conditions and the presence of Helicobacter species. In the event that no patients were Helicobacter positive, a sample of 100 patients with pancreatico-biliary conditions would result in a prevalence association between 0.02% and 5%.


One hundred and twenty-five patients were included in this study. Fifty-nine per cent of these patients were female, and the age ranged from 18 to 95 years (mean, 60.5 ± 18.8 years). Twenty-seven per cent (27/99) of patients were on antibiotics prior to endoscopic retrograde cholangiopancreatography. The diagnoses of the individual patients are listed in Table 1. The majority of patients had gallstones, with most of the remainder having malignancies and a few primary sclerosing cholangitis.

Table 1.  Patient diagnoses
Diagnosisn  = 125
  1. CBD, common bile duct.

 CBD stones only27
 Gall-bladder stones only8
 Both CBD and gall-bladder stones40
 Ampullary carcinoma5
 Pancreatic carcinoma6
Biliary stricture13
Primary sclerosing cholangitis4

Characteristics of the assay

The template DNA extracted from bile samples spiked with H. pylori cells yielded strong PCR bands of the expected lengths, with a lower detection limit of 1000 cells; the assay was occasionally able to detect down to 100 cells/mL of bile, but these bands were fainter and not consistently seen. The sensitivity of our assay was thus determined to be 1000 cells/mL of bile. Parallel testing of the DNA extracted from the supernatant and the pellet fractions of these spiked samples revealed that detection was 2 logs more sensitive on the pellet fraction of the specimens.

PCR inhibition occurred in only three patients. All other patient DNA templates spiked with genomic DNA gave strong bands, confirming that significant PCR inhibition did not occur. All negative controls produced no detectable bands on gel electrophoresis.

Helicobacter detection in patients

None of the 122 patients (the three patients with PCR inhibition were excluded) had Helicobacter detected in their bile. The 95% confidence intervals of the estimates of prevalence according to the underlying diagnoses are shown in Table 2. It should be noted that the upper limit of the 95% confidence interval for Helicobacter prevalence in the bile of patients with gallstones does not exceed 7%. The confidence intervals were less narrow in patients with underlying biliary malignancies, and particularly wide in patients with primary sclerosing cholangitis, because of the low sample size for patients with this condition.

Table 2.  Point prevalence and 95% confidence intervals (CI) of Helicobacter species in hepato-biliary disease
Diagnosisn  = 122 Helicobacter detected 95% CI
  1. CBD, common bile duct; GB, gall-bladder.

 CBD only27 0–19
 GB only8 0–53
 Both39 0–13
 Cholangio/ampulla  carcinoma9 0–48
 Pancreatic carcinoma5 0–72
Biliary stricture1300–36
Primary sclerosing cholangitis400–81


This study has demonstrated the ability of PCR to detect Helicobacter organisms in human bile with a sensitivity of 100–1000 organisms/mL. Only three of the samples tested were inhibitory for PCR, and extraction controls run with each batch confirmed that the extraction efficiency was consistent throughout the study. A negative result using this assay implies that there are not a significant number of organisms present in bile in this group of patients. Other studies that have identified Helicobacter in patients with similar hepato-biliary conditions did not express the sensitivity of their assays, but our assay is 2 logs more sensitive than detection by standard microscopy with Gram stain or silver stain (one bacterium seen per high-power field equates with 105 bacteria/mL). The use of endoscopic retrograde cholangiopancreatography in order to obtain bile for the detection of such organisms is thus an adequate method for this purpose, which is less invasive than obtaining liver tissue or surgical specimens. Endoscopic retrograde cholangiopancreatography also avoids contamination with H. pylori from the stomach, as the sampling devices are inserted inside the endoscope and hence never traverse the stomach.

This study is the first large North American investigation addressing the issue of Helicobacter in bile and, although it is a negative study, the results are important. None of the 122 patients who underwent endoscopic retrograde cholangiopancreatography for hepato-biliary conditions had Helicobacter present in bile. This suggests that Helicobacter is not a significant cause of biliary problems in general. This seems to be the case, especially for conditions in which there was a sufficiently large number of patients to be confident about the point prevalence estimates; for example, in the 74 patients with gallstone disease, in whom the observed prevalence was 0%, the corresponding 95% confidence interval was 0–7%. This reflects an accurate prevalence estimate of Helicobacter species in the bile of this population. Hence, Helicobacter, at least in terms of its presence in bile obtained from the common bile duct, is not significantly associated with gall-bladder stones in a North American population. A similar conclusion can be drawn, although with less confidence, with regard to any association with pancreatico-biliary malignancies. The small number of samples in the other diagnostic groups, however, limit the conclusions that can be drawn due to the fact that their 95% confidence intervals are excessively wide. For example, even though no Helicobacter was identified in the bile of the four patients with primary sclerosing cholangitis included in this study, this does not confidently exclude an association. Such a potential association should be further studied because H. pylori has been detected in some patients with primary sclerosing cholangitis,13 and a novel Helicobacter has also been found in cotton top tamarins with chronic colitis.16 This animal is an excellent model for ulcerative colitis, a condition known to be associated with primary sclerosing cholangitis. In addition, several biliary infections, such as Cryptosporidium and human immunodeficiency virus, can cause a clinical and endoscopic picture similar to primary sclerosing cholangitis. Hence, it is certainly plausible that an infectious agent present in the gastrointestinal tract may gain access to the biliary tree via the ampulla of Vater and be responsible, at least in part, for the development of primary sclerosing cholangitis.

Our results are similar to those of Mendez-Sanchez et al., who did not demonstrate an association between gallstone disease and Helicobacter in Mexicans,17 but different from those of Fox et al., who detected Helicobacter species in gall-bladder tissue from Chileans with chronic cholecystitis.12 A possible explanation for this discrepancy is that Helicobacter may be adherent to the gall-bladder or bile duct mucosa and, by testing the bile, one is not collecting the adherent Helicobacter organisms. This would lead to a false negative result with the use of our assay. However, when searching for H. pylori in gastric juice, the organisms are detected with PCR despite the fact that Helicobacter are adherent to the gastric mucosa.18, 19 Nevertheless, in future studies, both bile and biliary mucosa should be examined for the presence of this organism. It is also possible that the organisms preferentially live in the bile canaliculi, as the bile in this location is biochemically different from that in the common bile duct. Sampling with endoscopic retrograde cholangiopancreatography may then miss a canalicular infection. The other possible explanation for this discrepancy is that there are different infection prevalences in different geographical regions. For example, the rate of H. pylori infection in Chile and other developing countries is higher than that in Canada.20 Perhaps these other Helicobacter species follow similar epidemiological tendencies. Regional variations in the distribution of bile-resistant Helicobacter species have also been suggested as a possible cause of inconsistent results.21 In addition, some studies have used enzyme-linked immunoabsorbent assay to identify Helicobacter species in bile samples.22 This method is not as specific in detecting Helicobacter as PCR, and may lead to false positive results due to cross-reactivity between different bacterial organisms or proteins.

In summary, Helicobacter DNA can be detected in bile with the use of PCR. Helicobacter was not present in any of the bile samples of our patient populations with common bile duct stones, gall-bladder stones and pancreatico-biliary malignancies. Certain populations (for example, those with primary sclerosing cholangitis) must be studied more extensively in a larger sample before a firm conclusion can be drawn with regard to the association and possible role of Helicobacter in this condition.


The authors are indebted to Michael Purdy, Deborah Ross and Catherine Wheeler for their able technical assistance. Funding for this study was provided by the Canadian Helicobacter Study Group. CAF, MB and ANB are research scholars of the Fonds de la Recherche en Santé du Québec.