Home
JournalsCollections
For Authors For Reviewers For Editorial Board Members
Article Processing Charges Open Access
Ethics Advertising Policy
Editorial Policy Resource Center
Company Information Contact Us
OPEN ACCESS

A Multipathogen Bile Sample-based PCR Assay Can Guide Empirical Antimicrobial Strategies in Cholestatic Liver Diseases

  • Michael Jahn1,* ,
  • Mustafa K Özçürümez2 ,
  • Sebastian Dolff3 ,
  • Hana Rohn3,
  • Dominik Heider4 ,
  • Alexander Dechêne5 ,
  • Ali Canbay2 ,
  • Peter M. Rath6 and
  • Antonios Katsounas2 
Journal of Clinical and Translational Hepatology   2022;10(5):788-795

doi: 10.14218/JCTH.2021.00337

Received:

Revised:

Accepted:

Published online:

 Author information

Citation: Jahn M, Özçürümez MK, Dolff S, Rohn H, Heider D, Dechêne A, et al. A Multipathogen Bile Sample-based PCR Assay Can Guide Empirical Antimicrobial Strategies in Cholestatic Liver Diseases. J Clin Transl Hepatol. 2022;10(5):788-795. doi: 10.14218/JCTH.2021.00337.

Abstract

Background and objectives

Polymerase chain reaction (PCR) techniques provide rapid detection of pathogens. This pilot study evaluated the diagnostic utility and clinical impact of multiplex real-time PCR (mRT-PCR, SeptiFast) vs. conventional microbial culture (CMC) in bile samples of patients with chronic cholestatic liver diseases (cCLDs), endoscopic retrograde cholangio-pancreatography (ERCP), and peri-interventional-antimicrobial-prophylaxis (pAP).

Methods

We prospectively collected bile samples from 26 patients for microbiological analysis by CMC and mRT-PCR. Concordance of the results of both methods was determined by Krippendorff's alpha (α) for inter-rater reliability and the Jaccard index of similarity.

Results

mRT-PCRbile and CMCbile results were concordant for only Candida albicans (α=0.8406; Jaccard index=0.8181). mRT-PCRbile detected pathogens in 8/8 cases (100%), CMCbile in 7/8 (87.5%), and CMCblood in 5/8 (62.5%) with clinical signs of infection. mRT-PCRbile, CMCbile, and CMCblood had identical detection results in 3/8 (37.5%) with clinical signs of infection (two Klebsiella spp. and one Enterococcus faecium). The total pathogen count was significantly higher with mRT-PCRbile than with CMCbile (62 vs. 31; χ2=30.031, p<0.001). However, pathogens detected by mRT-PCRbile were more often susceptible to pAP according to the patient infection/colonization history (PI/CH) and surveillance data for antibiotic resistance in our clinic (DARC). Pathogens identified by mRT-PCRbile and resistant to pAP by PI/CH and DARC were likely to be clinically relevant.

Conclusions

mRT-PCR in conjunction with CMCs for bile analysis increased diagnostic sensitivity and may benefit infection management in patients with cholestatic diseases. Implementation of mRT-PCR in a bile sample-based diagnostic routine can support more rapid and targeted use of antimicrobial agents in cCLD-patients undergoing ERCP and reduce the rate/length of unnecessary administration of broad-spectrum antibiotics.

Keywords

ERCP, PCR, Antibiotic prophylaxis, Cholestasis, Cholangitis

Introduction

Endoscopic retrograde cholangio-pancreatography (ERCP) is a mainstay of therapeutic procedures in the vast majority of chronic cholestatic liver diseases (cCLDs), which are associated with either primary or secondary obstructions of the biliary system, leading to reduced or disrupted bile flow. Regardless of the cause, stasis of bile supports exponential growth of micro-organisms, which enter the biliary tree via the portal venous system or by ascent from the intestine and cause increased intrabiliary pressure as well as cholangiovenous reflux along with bacteremia or fungemia.1

ERCP-related interventions like balloon dilatation, stenting, lavage or cast and stone extractions help to restore the biliary drainage and are recommended by numerous treatment guidelines for cCLD, e.g., ischemic-type biliary lesions (ITBLs)2 or primary sclerosing cholangitis (PSC).3–5 Antibiotic prophylaxis prior to ERCP is not routinely recommended, as it does not significantly reduce the risk of the subsequent emergence of cholangitis in unselected patients.6,7 However, peri-interventional antimicrobial prophylaxis (pAP) should be administered in patients undergoing ERCP if the likelihood of achieving the best outcome, i.e. complete biliary drainage, is small. This recommendation refers to complex clinical conditions caused by cCLD, especially in association with hilar tumors, immunosuppression, and pancreatic pseudocysts communicating with the pancreatic duct.8,9 Currently, local recommendations for pAP still follow relatively heterogeneous guidelines on empirical use of antimicrobial agents. Moreover, reliable results of conventional microbial cultures (CMC) require up to 48 h. Microbiological tests detecting cholangitis-causing pathogens more rapidly would contribute toward more timely and specific administration of antimicrobials. To this end, multiplex real-time PCR (mRT-PCR) assays are promising to complement CMC performance and improve overall diagnostic ability.10–12

Until 2019, SeptiFast (Roche Diagnostics, GmbH, Mannheim, Germany) was a commercially available mRT-PCR assay able to detect 20 different bacterial or fungal micro-organisms commonly involved in systemic bloodstream infections (Table 1).11 The original scope of SeptiFast was to rapidly identify bacterial and fungal DNA in the bloodstream of patients suspected of having sepsis. The test was designed to run directly, requiring no prior incubation or preculture preparation of blood samples. It has also been shown to be a reliable method for diagnostic evaluation of bacteremia in association with endocarditis or periprosthetic joint infections, particularly in patients receiving antibiotic prophylaxis.13 Rapid identification of bacterial and fungal pathogens may be of huge benefit in patients with cCLD who are at risk of, or suspected of having, severe cholangitis and require broad-spectrum antibiotic prophylaxis/treatment and further prompt medical intervention, e.g., ERCP. Our particular interest in investigating the diagnostic utility of SeptiFast using bile samples from patients with cCLD was that this step may reduce the rate/length of unnecessary use of broad-spectrum antibiotics. The aim was to contribute toward antimicrobial stewardship and promote targeted treatment earlier in the clinical management pathway. To the best of our knowledge, this is the first study to evaluate the diagnostic utility of an mRT-PCR assay of bile (mRT-PCRbile) as a complementary test to CMC of bile samples (CMCbile) obtained during ERCP and required to provide robust antimicrobial susceptibility information.

Table 1

Pathogens (n=20) included in the SeptiFast mRT-PCRbile panel

Gram-positive bacteria
1. CoNS
2. Staphylococcus aureus
3. Streptococcus pneumoniae
4. Streptococcus spp.a
5. Enterococcus faecium
6. Enterococcus faecalis

Methods

This study was conducted at the University Hospital Essen from May 2016 to August 2018. During that time, bile fluid samples were obtained from within the biliary tract during ERCP in patients with cCLDs. Bile (1.5 mL) was directly collected into a 2.6 mL EDTA monovette (S-Monoette 2.6 mL K3E; Sarstedt AG and Co KG, Nümbrecht, Germany). The residual 2–5 mL bile was used for inoculation of blood culture sets (BD Bactec Lyitc/10 Anaerobe/F and BD Bactec Plus Aerobe/F9240 systems; Becton, Dickinson Co., Sparks, USA). All bile samples were immediately sent to our microbiological laboratory for DNA amplification and bile culture analysis.

Pathogen detection by mRT-PCR of bile samples

We used SeptiFast kits and a LightCycler® 2.0 real-time PCR instrument (Roche Diagnostics GmbH, Mannheim, Germany) for DNA assays. The LightCycler SeptiFast Test, which was taken off the market a few years ago, is a semi-automated real-time PCR system that was designed for simultaneous detection of DNA of most sepsis-relevant bacterial species (Table 1) in EDTA-preserved blood.14 Primers, PCR conditions, and formulas are patent protected. The patent information are disclosed in PCT publications WO97/46707, WO97/46712, and WO98/46714) and include descriptions of the methods we used.15,16 DNA was extracted from the bile samples, processed via mechanical lysis, and purified following the manufacturer’s instructions. For real-time PCR amplification in parallel reactions for Gram-positive and Gram-negative bacteria and fungi we used 50 µL volumes of each DNA extract.17 Detection and identification of the automatically amplified products was performed with specific hybridization probes.11

As specified by the manufacturer, the analytical sensitivity of SeptiFast is 100 colony-forming units per mL (CFU/mL) for coagulase-negative Staphylococci (CoNS), Streptococcus spp. and Candida glabrata, and 30 CFU/mL for all the other measured micro-organisms.17,18Table 1 lists the bacteria and fungi that are detectable by SeptiFast. Measures for prevention of contamination were as follows: extraction of DNA in a safety cabinet, use of MGRADE disposables, and extraction and amplification of DNA in separate rooms. An internal control was added in the extraction step to each sample. In each extraction series, a negative control was included. As positive control, the reagent controls were used. PCR results were available within 4.5 h after the specimens were processed in the laboratory.

Pathogen detection with CMC in bile and/or blood samples

For bile culture analysis, the automated microbial detection platform BacT/Alert 3D system (BioMerieux, Marcy l’Etoile, France) was used, which colorimetrically detects the growth of micro-organisms based on their CO2 production. Bile Samples were declared as negative if no growth of micro-organisms was detectable after 5 days of incubation. In positive bile cultures, subsamples were taken for Gram staining, plate culturing, and subsequent analysis. Identification and determination of the micro-organisms’ antibiotic susceptibilities were tested with the VITEK2 system (BioMerieux). Results of bile cultures were obtained within 72 h. Because of an insufficient amount of bile collected during ERCP, six patients had to be excluded from further consideration, as their samples could be subjected to microbial diagnostics either by CMCbile (n=3) or mRT-PCRbile (n=3) only. Thus, bile samples of 26 of 32 initially enrolled patients were included in the final analysis. In patients developing systemic infections (e.g., cholangitis), conventional microbial culture of blood samples (CMCblood) was also performed. CMCblood samples were processed with the BACTEC 9240 system (BD Bactec Lyitc/10 Anaerobe/F and BD Bactec Plus Aerobe/F 9240 systems; Becton, Dickinson and Co., Sparks, MD, USA). After the inoculation of 5–10 mL of blood, further analysis was similar to the bile culture procedure.

Statistics

Statistical analysis was performed with SPSS (version 21.0; IBM Corp., Armonk, NY, USA). For descriptive statistics, absolute and relative frequencies were calculated for categorical variables. Patients suspected of post-ERCP cholangitis were compared with clinically inapparent patients for systemic infections. Inferential statistics included Fisher’s exact tests for categorical variables and Mann-Whitney U test for continuous variables. Results were considered significant when p ≤ 0.05.

To describe the diagnostic concordance between mRT-PCRbile and CMCbile results, we used Krippendorff’s alpha test (α) to estimate the interrater reliability,19 and the Jaccard index to measure similarity.20,21 Reliability was indicated when α >0.800; non-reliability was indicated when α <0.667.19 The Jaccard index does not impose any weights. It assigns a value of 1 in case of match and 0 otherwise. Hafnia spp. and Enterococcus casseliflavus (n=2) were not considered for concordant analysis, as they were not part of the SeptiFast panel. McNemar chi-square tests at a significance level of 5% were performed to compare the overall detection rates of micro-organisms with CMCbile and mRT-PCRbile. The McNemar chi-square test does not test for independence, but for consistency in responses across two variables. It is generally used with paired data, and can directly compare counts.22

Ethical approval

This study was performed in accordance following the ethical principles and standards of the 1964 Helsinki declaration and the guidelines of the International Conference for Harmonization for Good Clinical Practice. The study was approved by the local institutional review board (IRB: “Ethik-Kommission am Universitätsklinikum Essen”, 18-8482-BO).

Results

Samples from 26 patients with cCLDs like PSC (n=14/26; 54%), biliary disorders after liver transplantation (n=10/26; 38%) and secondary sclerosing cholangitis (SSC; n=2/26; 8%) were included in the final analysis. The median age, body mass index, and Charlson comorbidity index were 56 years, 25 kg/m2 and 3.5. Most frequent reasons for medication intake (Table 2) were immunosuppressive therapies after LTx (n=10), maintenance therapy for PSC (n=4), hypertension (n=11), and diabetes (n=5). It should be noted that none of the patients was pregnant. Overall, eight patients (n=8/26; 31%) developed clinical signs of post-ERCP cholangitis; of those, seven patients (n=7/8; 88%) were on immunosuppressive therapy after liver transplantation (Table 2).

Table 2

Characteristics of patients with and without clinical signs and symptoms of colangitis infection at admission

CharacteristicAll patients n (%) or median (IQR) [min-max]No clinical symptoms n (%) or median (IQR) [min-max]Clinical symptoms n (%) or median (IQR) [min-max]p-value
Patients26188NA
Age, years56 (46–62) [29–78]47 (43–61) [29–78]59 (55–65) [50–72]0.090*
Body mass index, kg/m225 (23–26) [20–36]25 (23–27) [20–36]25 (24–26) [20–27]0.933*
Sex
  Male19 (73.1)14 (77.8)5 (62.5)0.635**
  Female7 (26.9)4 (22.2)3 (37.5)0.635**
Chronic cholestatic liver disease
  PSC14 (53.8)14 (77.8)0<0.000**
Post-LTx10 (38.5)3 (16.7)7 (87.5)0.001**
  Post-LTx/no ITBL3 (11.5)1 (5.6)2 (25.0)0.215**
  Post-LTx/ITBL Type I3 (11.5)03 (37.5)0.022**
  Post-LTx/ITBL Type II1 (3.8)1 (5.6)01.000**
  Post-LTx/ITBL Type III3 (11.5)1 (5.6)2 (25.0)0.215**
SSC2 (7.7)1 (5.6)1 (12.5)0.529**
Chronic disease
  Charlson comorbidity index3.5 (2.2–5.7) [0–9.0]3.0 (2.0–4.7) [0–9.0]4.0 (3.0–7.2) [2.0–8.0]0.216*
  Immunosuppression14 (53.8)7 (38.9)7 (87.5)0.028**
  Hypertension11 (42.3)6 (33.3)5 (62.5)0.246**
  Diabetes mellitus5 (19.2)2 (11.1)3 (37.5)0.330**
Pre-interventional antibiotic prophylaxis
  Ciprofloxacin18 (69.2)14 (77.8)4 (50.0)0.197**
  Metronidazole19 (73.1)14 (77.8)5 (62.5)0.635**
  Ceftriaxon4 (15.4)3 (16.7)1 (12.5)1.000**
  Piperacillin/tazobactam3 (11.5)03 (37.5)0.022**
  Linezolid2 (7.7)02 (25.0)0.086**
  Cefixim1 (3.8)1 (5.6)01.000**
Microbiological findings
  CMCbile20 (76.9)13 (72.2)7 (87.5)0.628**
    Detection of ≥2 pathogens10 (38.5)6 (33.3)4 (50.0)0.664**
  SeptiFast mRT-PCRbile25 (96.2)17 (94.4)8 (100.0)1.000**
    Detection of ≥2 pathogens21 (80.8)14 (77.8)7 (87.5)1.000**

Pathogen detection mRT-PCR vs. CMC in bile samples

One or more micro-organisms were identified in 20 patients by CMCbile (n=20/26; 77%) and in 25 patients by mRT-PCRbile (n=25/26; 96%). CMCbile and mRT-PCRbile concurrently detected numerous pathogens (n=29) (Table 3, Fig. 1). They were Candida albicans 9×, Enterococcus faecalis 6×, Enterococcus faecium 6×, Klebsiella pneumoniae/oxytoca 4×, Escherichia coli 1×, Streptococcus spp. 1×, Pseudomonas aeruginosa 1×, and CoNS 1×.

Table 3

Micro-organisms detected in bile samples by mRT-PCRbile and/or CMCbile

PathogenBile sample
J-indexIR
nOverall positive samplesmRT-PCR posCMC posmRT-PCR | CMC
pos|pospos|negneg|posneg|neg
E. faecalis2611107641150.54540.5714
C. albicans2611119920150.81810.8406
E. faecium2611116650150.54540.5714
K. pneumoniae/oxytoca26884440190.50000.5750
Streptococcus spp26771160190.14280.1307
E. coli26542131210.20000.2609
E. cloacae/aerogenes26550050210.0000−0.0851
P. aeruginosa26221110240.50000.6531
CoNS26221110240.50000.6531
S. aureus26110010250.00000.0000
S. maltophilia26110010250.00000.0000
Total286646231293322230.45310.5514
Pathogen detection in bile samples by mRT-PCR and CMC.
Fig. 1  Pathogen detection in bile samples by mRT-PCR and CMC.

mRT-PCRbile, multiplex real-time PCR of bile samples; CMCbile, conventional microbial culture of bile samples; CMCblood, conventional microbial culture of blood samples; CoNS, Coagulase-negative staphylococci; P. aeruginosa, Pseudomonas aeruginosa; C. albicans, Candida albicans; E. coli, Escherichia coli; E. faecalis, Enterococcus faecalis; Kl. Pneumoniae/oxytoca: Klebsiella pneumoniae, Klebsiella oxytoca; E. faecium: Enterococcus faecium; E. cloacae/aerogenes: Enterobacter cloacae, Enterobacter aerogenes; Streptococcus spp.: Streptococcus species; Staph. aureus: Staphylococcus aureus; S. maltophilia: Stenotrophomonas maltophilia. P, patient. The figure shows which method successfully identified each pathogen in each patient. While two positive hits were delivered via CMCbile only (yellow heading), 34 positive hits were delivered via mRT-PCRbile only (blue heading). Highlighted frames indicate patients with clinical signs of infection. Yellow frames (P2, P10, P11, P13, and P17) indicate symptomatic patients with negative blood cultures, whereby most of these cases (seven of 11) have only positive hits by mRT-PCRbile and were missed by CMCbile. Red frames indicate pathogens that were concurrently detected by CMCblood and CMCbile or mRT-PCRbile (P16, P23, P24, and P25). Black bold frames indicate cases with positive blood cultures and clinical signs of infections, whereby CMCbile or mRT-PCRbile, detected additional pathogens that were not detected by CMCblood (P 16, P 23, P 24, P 25). Especially C. albicans was frequently detected in bile samples, but not in blood cultures of patients with signs of clinical infection, e.g., P 23, P24, and P25.

A total of 33 pathogens were exclusively detected by mRT-PCRbile and not by CMCbile (Table 3, Fig. 1). They were Streptococcus spp. 6×, Enterococcus faecium 5×, Enterobacter cloacae/aerogenes 5×, Enterococcus faecalis 4×, Klebsiella pneumoniae/oxytoca 4×, Escherichia coli 3×, Candida albicans 2×, Pseudomonas aeruginosa 1×, CoNS 1×, S. aureus 1×, and Stenotrophomonas maltophilia 1×. Only two pathogens, Enterococcus faecalis 1× and Escherichia coli 1×, were detected by CMCbile and not by mRT-PCRbile. CMCbile and mRT-PCRbile concurrently detected more than one pathogen in approximately one-third of all cases (n=9/26; 35%). Moreover, mRT-PCRbile identified multiple micro-organisms in 12 additional cases.

All patients (n=26) received pAP. The vast majority (n=18/26, 69.2%) received ciprofloxacin and metronidazole (Table 2), and bile culture predominantly identified pathogens that were not covered by those antibiotics (n=21/31, 67.7%). The pathogens were Candida albicans 9×, Enterococcus faecalis 6×, Entercoccus faecium 5×, Streptococcus anginosus 1× (Supplementary Table 1). In contrast, mRT-PCRbile frequently detected pathogens that were likely to be susceptible to the administered antibiotics (n=24/62; 38.7%). They were Klebsiella pneumoniae/oxytoca 8×, Enterobacter cloacae/aerogenes 5×, Escherichia coli 4×, Pseudomonas aeruginosa 2×, CoNS 2×, Enterococcus faecium 1×, Stenotrophomonas maltophilia 1×, and Staphylococcus aureus 1× (Supplementary Table 1). Here, susceptibility was suggested by surveillance data for antibiotic resistance at our hospital and previous patient history and microbiological records.

Concordance of mRT-PCR and CMC results in bile samples

The total pathogen count detected by mRT-PCRbile (n=62; Table 3) was significantly higher than that detected by CMCbile (n=31 as shown in Table 3; χ2 = 30.031, p<0.001 as shown in Table 4). CMCbile detected 33.3% (n=7) and mRT-PCRbile 95.2% (n=20) of all Gram-negative bacteria (n=21) found in bile samples. These results reveal a remarkable difference of 61.2% in detection outcome between CMCbile and mRT-PCRbile. Furthermore, CMCbile detected 46.8% (n=15) and mRT-PCRbile 96.8% (n=31) of all Gram-positive bacteria (n=32) found in bile samples. These results illustrate an equally remarkable difference of 50% in detection outcome between CMCbile and mRT-PCRbile. In contrast, the difference in detection outcomes for fungi by CMCbile vs. mRT-PCRbile was small (Table 4, Fig. 1), 81.8% (n=9) vs. 100% (n=11). The concordance of mRT-PCRbile and CMCbile achieved statistical significance only for detection of Candida albicans, with a Krippendorf’s alpha inter-rater reliability of 0.8406 and a Jaccard index of similarity of 0.8181. Detection of other pathogens had only modest or low interrater reliability and similarity between mRT-PCRbile and CMCbile (Table 4).

Table 4

Detection of Gram-positive bacteria, Gram-negative bacteria, and fungi in bile samples by mRT-PCR and CMC

Test systemIsolates, n (%)
Gram-positive bacteriaGram-negative bacteriaFungiAll
mRT-PCR31/32 (96.8)20/21 (95.2)11/11 (100)62/64 (96.8)
CMC15/32 (46.8)7/21 (33.3)9/11 (81.8)31/64 (48.4)
χ2=30.031, p<0.001

Patients with clinical symptoms of infection (cholangitis)

Among patients with clinical signs of infection (n=8/26, 31%), mRT-PCRbile detected pathogens in bile in all cases, CMCbile in seven (87.5%) and CMCblood in five (62.5%). CMCbile, mRT-PCRbile, and CMCblood delivered identical results in three of these patients (n=3/8, 37.5%). They were Klebsiella spp. 2× and Enterococcus faecium 1×. In patients with positive CMCblood, four had identical hits between CMCblood and mRT-PCRbile (n=4/5, 80%), Klebsiella spp. 2×, Enterococcus faecium 1×, and Enterobacter spp. 1×, whereas three had identical hits between CMCblood and CMCbile (n=3/5, 60%), Klebsiella pneumoniae 2× and Enterococcus faecium 1× (Fig. 1).

Discussion

Apart from comparable performance in identification of Candida albicans, we found major discrepancies between both methods, i.e., mRT-PCRbile and CMCbile, in their ability to detect micro-organisms in bile samples. The distribution of organisms identified by CMCbile, mRT-PCRbile, or both methods is shown in Figure 1. Statistical comparisons between the CMCbile and mRT-PCRbile outcomes stratified by pathogen were precluded by the small number of results and patients. However, there are certain noteworthy findings. The majority of the Candida albicans isolates (n=9/11) were detected by both CMCbile and mRT-PCRbile, with only two isolates added by mRT-PCRbile only. In contrast, all the Streptococcus spp. isolates (n=7/7) were detected by mRT-PCRbile, with only one isolate identified by both CMCbile and mRT-PCRbile. CMCbile identified Enterococcus in one additional case and Escherichia coli in another, whereas mRT-PCRbile alone failed to detect those pathogens. Overall, regardless Gram stain classification, detection rates for both unique microbes and microbial clusters were significantly higher with mRT-PCRbile.

The findings are in line with previous studies comparing these methods using various specimens. In blood samples, higher detection rates for unique bacteria or fungi as well as combined microbial clusters, were reported for mRT-PCRbile compared with CMCblood.23,24 Of note, Sancho-Tello et al.13 described higher detection rates for micro-organisms in different purulent fluids, including bile samples, for mRT-PCR compared with CMC. A modest concordance between the two detection techniques was reported. Tajeddin et al.25 used a PCR denaturing gradient gel electrophoresis to assay bile samples from patients with biliary tract disorders. Compared with CMC, that PCR method also delivered positive single and multiple microbial hits at higher frequencies in bile fluids. However, in urine samples of patients with suspected urinary tract infection, Lehmann et al.12 reported equally high pathogen detection rates with mRT-PCR and with CMC.

To the best of our knowledge, this is the first study to use mRT-PCR to test bile samples from patients with chronic cholestatic diseases. Only 1.5 mL of bile fluid was needed for analysis of micro-organisms with mRT-PCR compared with the 2–5 mL that is required for CMCbile. Furthermore, the SeptiFast mRT-PCR assay results were available within 4–5 h compared with the 24–72 h required to generate CMCbile reports along with antibiograms after specimen sampling. The SeptiFast mRT-PCR assay detects DNA of micro-organisms present at low counts. Thus, it is not surprising that mRT-PCR assays have had a high specificity/accuracy performance in various clinical contexts in previous studies.13,14,18 Therefore, mRT-PCR assays seem to be valuable to rule out fungobilia and bactobilia in bile samples. At the same time, the non-quantitative detection of DNA fragments, e.g., from degraded pathogens, colonizing micro-organisms, nonpathogens, or facultative pathogens, render the interpretation of positive mRT-PCR results challenging, and led to a sensitivity in previous studies that lagged far behind its specificity.13,14,18 In our cohort, the small number of patients developing clinical signs of cholangitis (n=8/26, 30.7%) limited the interpretation of the true sensitivity of both methods.

In line with previous reports, the most frequently detected micro-organisms by either mRT-PCR or CMCbile were Candida albicans, Enterococcus faecium and Enterococcus faecalis.26–28 In our cohort, pAP did not cover those pathogens (0/26 of Candida albicans cases and 2/26 of Enterococcus spp. Cases) (Table 3, Supplementary Table 1). Despite that in only 1/8 patients suspected of chalngitis, Enterococcus faecium was isolated in CMCblood specimen, and Candida albicans and Enterococcus faecalis were not detected in any CMCblood specimens from patients with clinical signs of cholangitis. Therefore, we assume that the detection of Enterococcus spp. and Candida albicans in bile samples are of low clinical significance, and rather represent colonization of the bile tract that requires no specific antibiotic treatment in the absence of clinical signs of acute cholangitis. However, that assumption cannot be extrapolated to other microbial strains that are considered susceptible to pAP. mRT-PCRbile detected such pathogens at higher rates compared with CMCbile in our cohort (Table 3, Supplementary Table 1). That is in line with previous studies reporting higher microbial detection rates for DNA detection kits vs. conventional culture methods in patients receiving antibiotics in different clinical settings.13,24,29 The pAP in our cohort must be considered effective against bacteria like Escherichia coli, Pseudomonas aeruginosa or Klebsiella pneumoniae in most cases (Supplementary Table 1). Those microbial strains are not only common causative pathogens for acute cholangitis,30 but also seem to promote polymicrobial bile infections because of their biofilm forming abilities.25 In our patients with clinical signs of cholangitis, mRT-PCRbile and CMCbile concurrently identified Klebsiella spp., but CMCbile failed to detect Enterobacter cloacae, which was identified by mRT-PCRbile and CMCblood (Fig. 1).

With regard to the high specificity previously reported for DNA detection kits, the lack of detection of pathogens in bile samples by mRT-PCRbile represents, in light of our data, a true negative result, and is unlikely an effect of brief prophylactic administration of broad-spectrum antibiotics. Although this observation/postulation already guides our empiric practice in post-interventional antibiotic management of patients with cCLDs, it lacks broad evidence, and therefore should be further addressed in large studies in the future. To date, only a handful of studies reported clear benefits for rapid molecular diagnostics (RMD) in antibiotic treatment management.31 Our single-center experience suggests that it is safe to discontinue standard antibiotic prophylaxis with ciprofloxacin or ceftriaxone if mRT-PCRbile detects no pathogens other than Enterococcus spp. or Candida albicans in the patient´s bile fluid. According to previous research, significant shortcomings, e.g., low sensitivity and limited antibiotic resistance detection in whole blood of patients with suspected blood stream infection and/or sepsis, represent reasons for discontinuation of RMD production by companies.31 However, conjunctional use of mRT-PCR assays and CMC in bile samples seems more performant if longitudinally applied and if antimicrobial prophylaxis/therapy for patients with cCLD takes into account the patient infection/colonization history (PI/CH) and the surveillance data for antibiotic resistance-in our clinic (DARC). In this study, longitudinal bile sample-based CMC results, which existed only for a small minority of patients, were excluded from analysis because of a lack of parallel longitudinal mRT-PCRbile tests. Finally, but important, mRT-PCRbile cannot detect anaerobic micro-organisms; thus, it is not suitable to guide decisions regarding the use of metronidazole as part of post-interventional antibiotic treatment strategies. Including anaerobic bacterial species in the mRT-PCRbile panel would presumably be of major interest in the evaluation of bile samples in similar clinical settings and would render this spectrum of detectable pathogens more complete for covering the clinical picture. Intervention studies of the potential added value of mRT-PCR assays in bile diagnostics of comparable populations, such as patients with cCLD undergoing ERCP, and focusing on clinically relevant outcomes such as reducing the risk for cholangiosepsis and/or the selective pressure for antibiotic resistance are needed. To this end, researchers should be aware of other commercially available assays such as VYOO (SIRS-Lab, Jena, Germany) and the SepsiTest (Molzym, Bremen, Germany), which have already been successfully tested as molecular diagnostics of sepsis in critically ill patients and therefore seem promising for application as RMD alternatives to SeptiFast in bile.32

Supporting information

Supplementary Table 1

Microbiological findings of bile and blood cultures and the applied peri-interventional antibiotics in each patient.

(DOCX)

Abbreviations

cCLDs: 

chronic cholestatic liver diseases

CFU/mL: 

colony-forming units per mL

CMC: 

conventional microbial culture

CoNS: 

coagulase-negative Staphylococci

DARC: 

data for antibiotic resistance in our clinic

ERCP: 

endoscopic retrograde cholangio-pancreatography

ITBLs: 

ischemic-type biliary lesions

IQR: 

interquartile range

LTx: 

liver transplantation

mRT-PCR: 

multiplex real-time PCR

pAP: 

peri-interventional-antimicrobial-prophylaxis

PCR: 

polymerase chain reaction

PI/CH: 

patient infection/colonization history

PSC: 

primary sclerosing cholangitis

RMD: 

rapid molecular diagnostics

SSC: 

secondary sclerosing cholangitis

Declarations

Acknowledgement

We acknowledge support by the Open Access Publication Fund of the University of Duisburg-Essen.

Data sharing statement

The data used to support the findings of this study are included within the supplementary information file, and are also available from the corresponding author upon request.

Funding

None to declare.

Conflict of interest

The authors have no conflict of interests related to this publication.

Authors’ contributions

Conceptualization (AK), methodology (AK, AC, MJ), software (MKÖ), validation (SD, PMR, AC), formal analysis (AK, DH, MJ), investigation (AK, SD, AD, MJ), resources (AK, AD, PMR), data curation (HR, MKÖ, DH, MJ), writing the original draft (MJ, AK), review and editing (MJ, SD, HR, AK), visualization (AK), supervision (HR, AC), and project administration (AK).

References

  1. Hanau LH, Steigbigel NH. Acute (ascending) cholangitis. Infect Dis Clin North Am 2000;14(3):521-546 View Article PubMed/NCBI
  2. Buis CI, Hoekstra H, Verdonk RC, Porte RJ. Causes and consequences of ischemic-type biliary lesions after liver transplantation. J Hepatobiliary Pancreat Surg 2006;13(6):517-524 View Article PubMed/NCBI
  3. European Society of Gastrointestinal Endoscopy; European Association for the Study of the Liver. Role of endoscopy in primary sclerosing cholangitis: European Society of Gastrointestinal Endoscopy (ESGE) and European Association for the Study of the Liver (EASL) Clinical Guideline. J Hepatol 2017;66(6):1265-1281 View Article PubMed/NCBI
  4. Chapman R, Fevery J, Kalloo A, Nagorney DM, Boberg KM, Shneider B, et al. Diagnosis and management of primary sclerosing cholangitis. Hepatology 2010;51(2):660-678 View Article PubMed/NCBI
  5. EASL Clinical Practice Guidelines. EASL Clinical Practice Guidelines: Management of cholestatic liver diseases. J Hepatol 2009;51(2):237-267 View Article PubMed/NCBI
  6. Bai Y, Gao F, Gao J, Zou DW, Li ZS. Prophylactic antibiotics cannot prevent endoscopic retrograde cholangiopancreatography-induced cholangitis: a meta-analysis. Pancreas 2009;38(2):126-130 View Article PubMed/NCBI
  7. Brand M, Bizos D, O‘Farrell P. Antibiotic prophylaxis for patients undergoing elective endoscopic retrograde cholangiopancreatography. Cochrane Database Syst Rev 2010:CD007345 View Article PubMed/NCBI
  8. Domagk D, Oppong KW, Aabakken L, Czakó L, Gyökeres T, Manes G, et al. Performance measures for endoscopic retrograde cholangiopancreatography and endoscopic ultrasound: A European Society of Gastrointestinal Endoscopy (ESGE) Quality Improvement Initiative. United European Gastroenterol J 2018;6(10):1448-1460 View Article PubMed/NCBI
  9. Tang X, Gong W, Jiang B. Antibiotic prophylaxis for GI endoscopy. Gastrointest Endosc 2015;81(6):1503-1504 View Article PubMed/NCBI
  10. Corless CE, Guiver M, Borrow R, Edwards-Jones V, Fox AJ, Kaczmarski EB. Simultaneous detection of Neisseria meningitidis, Haemophilus influenzae, and Streptococcus pneumoniae in suspected cases of meningitis and septicemia using real-time PCR. J Clin Microbiol 2001;39(4):1553-1558 View Article PubMed/NCBI
  11. Lehmann LE, Hunfeld KP, Emrich T, Haberhausen G, Wissing H, Hoeft A, et al. A multiplex real-time PCR assay for rapid detection and differentiation of 25 bacterial and fungal pathogens from whole blood samples. Med Microbiol Immunol 2008;197(3):313-324 View Article PubMed/NCBI
  12. Lehmann LE, Hauser S, Malinka T, Klaschik S, Weber SU, Schewe JC, et al. Rapid qualitative urinary tract infection pathogen identification by SeptiFast real-time PCR. PLoS One 2011;6(2):e17146 View Article PubMed/NCBI
  13. Sancho-Tello S, Bravo D, Borrás R, Costa E, Muñoz-Cobo B, Navarro D. Performance of the LightCycler SeptiFast test Mgrade in detecting microbial pathogens in purulent fluids. J Clin Microbiol 2011;49(8):2988-2991 View Article PubMed/NCBI
  14. Dark P, Blackwood B, Gates S, McAuley D, Perkins GD, McMullan R, et al. Accuracy of LightCycler(®) SeptiFast for the detection and identification of pathogens in the blood of patients with suspected sepsis: a systematic review and meta-analysis. Intensive Care Med 2015;41(1):21-33 View Article PubMed/NCBI
  15. Hansen WL, Bruggeman CA, Wolffs PF. Evaluation of new preanalysis sample treatment tools and DNA isolation protocols to improve bacterial pathogen detection in whole blood. J Clin Microbiol 2009;47(8):2629-2631 View Article PubMed/NCBI
  16. Wellinghausen N, Kochem AJ, Disqué C, Mühl H, Gebert S, Winter J, et al. Diagnosis of bacteremia in whole-blood samples by use of a commercial universal 16S rRNA gene-based PCR and sequence analysis. J Clin Microbiol 2009;47(9):2759-2765 View Article PubMed/NCBI
  17. LightCycler SeptiFast Test Package Insert. Roche Diagnostics, 2006. View Article PubMed/NCBI
  18. Chang SS, Hsieh WH, Liu TS, Lee SH, Wang CH, Chou HC, et al. Multiplex PCR system for rapid detection of pathogens in patients with presumed sepsis - a systemic review and meta-analysis. PLoS One 2013;8(5):e62323 View Article PubMed/NCBI
  19. Hayes AF, Krippendorff K. Answering the call for a standard reliability measure for coding data. Commun Methods Meas 2007;1:77-89 View Article PubMed/NCBI
  20. Real R, Vargas JM. The probabilistic basis of Jaccard’s index of similarity. Syst Biol 1996;45(3):380-385 View Article PubMed/NCBI
  21. Jaccard P. The distribution of the flora in the alpine zone. New Phytol 1912;11(2):37-50 View Article PubMed/NCBI
  22. McNEMAR Q. Note on the sampling error of the difference between correlated proportions or percentages. Psychometrika 1947;12(2):153-157 View Article PubMed/NCBI
  23. Burdino E, Ruggiero T, Allice T, Milia MG, Gregori G, Milano R, et al. Combination of conventional blood cultures and the SeptiFast molecular test in patients with suspected sepsis for the identification of bloodstream pathogens. Diagn Microbiol Infect Dis 2014;79(3):287-292 View Article PubMed/NCBI
  24. Yanagihara K, Kitagawa Y, Tomonaga M, Tsukasaki K, Kohno S, Seki M, et al. Evaluation of pathogen detection from clinical samples by real-time polymerase chain reaction using a sepsis pathogen DNA detection kit. Crit Care 2010;14(4):R159 View Article PubMed/NCBI
  25. Tajeddin E, Sherafat SJ, Majidi MR, Alebouyeh M, Alizadeh AH, Zali MR. Association of diverse bacterial communities in human bile samples with biliary tract disorders: a survey using culture and polymerase chain reaction-denaturing gradient gel electrophoresis methods. Eur J Clin Microbiol Infect Dis 2016;35(8):1331-1339 View Article PubMed/NCBI
  26. Voigtländer T, Negm AA, Schneider AS, Strassburg CP, Manns MP, Wedemeyer J, et al. Secondary sclerosing cholangitis in critically ill patients: model of end-stage liver disease score and renal function predict outcome. Endoscopy 2012;44(11):1055-1058 View Article PubMed/NCBI
  27. Kirchner GI, Scherer MN, Obed A, Ruemmele P, Wiest R, Froh M, et al. Outcome of patients with ischemic-like cholangiopathy with secondary sclerosing cholangitis after liver transplantation. Scand J Gastroenterol 2011;46(4):471-478 View Article PubMed/NCBI
  28. Gelbmann CM, Rümmele P, Wimmer M, Hofstädter F, Göhlmann B, Endlicher E, et al. Ischemic-like cholangiopathy with secondary sclerosing cholangitis in critically ill patients. Am J Gastroenterol 2007;102(6):1221-1229 View Article PubMed/NCBI
  29. Pilecky M, Schildberger A, Knabl L, Orth-Höller D, Weber V. Influence of antibiotic treatment on the detection of S. aureus in whole blood following pathogen enrichment. BMC Microbiol 2019;19(1):180 View Article PubMed/NCBI
  30. Kwon JS, Han J, Kim TW, Oh JH, Kwon HH, Jung JT, et al. Changes in causative pathogens of acute cholangitis and their antimicrobial susceptibility over a period of 6 years. Korean J Gastroenterol 2014;63(5):299-307 View Article PubMed/NCBI
  31. D‘Onofrio V, Salimans L, Bedenić B, Cartuyvels R, Barišić I, Gyssens IC. The clinical impact of rapid molecular microbiological diagnostics for pathogen and resistance gene identification in patients with sepsis: A systematic review. Open Forum Infect Dis 2020;7(10):ofaa352 View Article PubMed/NCBI
  32. Schreiber J, Nierhaus A, Braune SA, de Heer G, Kluge S. Comparison of three different commercial PCR assays for the detection of pathogens in critically ill sepsis patients. Med Klin Intensivmed Notfmed 2013;108(4):311-318 View Article PubMed/NCBI