Occult cytomegalovirus cholangitis as a potential cause of cholestatic complications after orthotopic liver transplantation? A study of cytomegalovirus DNA in bile


  • The data presented in this article were partly presented as an oral presentation at the 2012 meeting of the American Association for the Study of Liver Diseases.

  • Daniel Nils Gotthardt was supported by a grant from the German Research Foundation.

  • The authors have no conflict of interest to declare.

Address reprint requests to Daniel Nils Gotthardt, M.D., Department of Internal Medicine IV, University Hospital of Heidelberg, Im Neuenheimer Feld 410, Heidelberg 69120, Germany. Telephone: +49-6221-56-6538; FAX: +49-6221-56-6727; E-mail: daniel_gotthardt@med.uni-heidelberg.de


Cholestatic complications, important causes of morbidity and mortality after orthotopic liver transplantation (OLT), often have an unclear etiology. Human cytomegalovirus (CMV) infections occur in immunosuppressed patients and can be detected in blood samples. However, CMV analyses of body fluids and biopsies are more sensitive. Here we evaluated whether a CMV analysis of bile could reveal occult CMV cholangitis. We evaluated OLT patients undergoing endoscopic retrograde cholangiography (ERC) for suspected biliary complications after OLT at a tertiary care center. Biliary CMV DNA levels were measured with real-time polymerase chain reaction. A nonanastomotic biliary lesion (NABL) group consisted of patients with nonanastomotic strictures (NASs) at the time of ERC (n = 59) and patients with normal ERC findings but microscopic biliary lesions in biopsy samples (n = 12). The anastomotic stricture (AS) group comprised patients with ASs only (n = 53). In all, 124 OLT patients underwent 240 ERC procedures. Biliary CMV DNA was detected in 14 of the 124 patients and was more frequently found in the NABL group (12/71 for the NABL group versus 2/53 for the AS group, P = 0.02). Concurrent sampling of CMV DNA in blood yielded negative results. Biliary CMV was more frequently detected in patients with a positive recipient status (13/73 or 17.8% versus 1/44 or 2.3%, P < 0.05). There was no significant difference in the incidence of biliary CMV between patients with a high-risk CMV status and patients with a low-risk CMV status. The median interval between OLT and biliary CMV detection was 8.4 months (range = 0.4-212.8 months). In conclusion, biliary CMV was detected in a substantial number of patients after OLT and was significantly associated with NASs or microscopic biliary lesions. A potential occult CMV infection could, therefore, be considered as a contributory etiological factor in the development of biliary complications. Liver Transpl 19:1142–1150, 2013. © 2013 AASLD.


anastomotic stricture




endoscopic retrograde cholangiography


nonanastomotic biliary lesion


nonanastomotic stricture


orthotopic liver transplantation


polymerase chain reaction

See Editorial on Page 1062

Biliary complications are a major cause of morbidity and mortality in patients after orthotopic liver transplantation (OLT). The incidence of biliary complications after cadaveric OLT ranges from 10% to 30%.[1-3] Mortality rates secondary to biliary complications range from 8% to 15%.[4] However, the etiology of these complications is often unknown, and further research is required to establish appropriate treatments or prevention strategies.

The most common biliary complications are strictures of the biliary tract, bile leaks, and gallstones. Strictures can occur at the site of the anastomosis [anastomotic strictures (ASs)] or in any other part of the donor biliary tract [nonanastomotic strictures (NASs)]. The former are associated with the surgical technique and local ischemia, and if they are treated with dilatation, stenting, and, in some cases, surgical revision, the long-term outcomes are good.[4] In contrast, the latter can be difficult to treat with interventional therapy. The reported graft failure rates among patients with NASs are as high as 50%.[5]

Cytomegalovirus (CMV) is the most common viral infectious pathogen in OLT recipients.[6] Infections can occur in various organs, such as the liver, gut, and lungs, or can manifest as systemic infections. CMV can be detected via serological analysis or through the identification of CMV DNA by polymerase chain reaction (PCR) in infected tissue.[7]

Before CMV prophylaxis, CMV infection was a risk factor for biliary complications after OLT.[8, 9] It is well known that despite prophylaxis, CMV infection is associated with increased graft loss and mortality after OLT.[10] However, studies examining the association between CMV and biliary complications after OLT despite the maintenance of CMV prophylaxis are scarce. Some studies suggest that CMV infection has no effect on the formation of NASs or biliary strictures in general after OLT.[11, 12]

The objective of this study was to examine the role of CMV in the development of biliary complications after OLT through an investigation of the presence of CMV DNA in bile.


Patients and Endoscopic Interventions

This study was conducted in accordance with the Declaration of Helsinki and was approved by the local ethics committee. Informed consent in writing was obtained from each patient.

Between September 2006 and February 2011, OLT patients who were referred to the University Hospital of Heidelberg for endoscopic retrograde cholangiography (ERC) because of suspected biliary complications were recruited for this study. This prospective study investigated the bile composition in patients after OLT according to the biliary complications.

Patients received cyclosporine A and tacrolimus combined with methylprednisolone and/or mycophenolate mofetil as immunosuppressive therapy according to the Heidelberg OLT protocol. Patients received CMV prophylaxis with valganciclovir for 3 months after liver transplantation except when both the donor and the recipient were CMV-negative. All patients received transplants from deceased donors; transplants involving non–heart-beating donors are not permitted in Germany. In addition, no ABO-incompatible transplantation was performed in the present study. Patients with primary or secondary sclerosing cholangitis were not included in the present study because they undergo hepaticojejunostomy and are not treated with ERC at our center. The study cohort comprised patients with viral hepatitis, alcoholic cirrhosis, hepatocellular carcinoma, and other indications.

Furthermore, records were checked for any prior stenosis of the hepatic artery or steal phenomena confirmed on angiography; data on immunosuppression, CMV episodes after OLT, donor age, and cold ischemia times were recorded from patients' files.

Each patient underwent a physical examination and laboratory investigations, including an assessment of cholestatic parameters before ERC. ERC was performed by 1 of 2 experienced endoscopists under conscious sedation with midazolam, propofol, and/or short-acting opiates. A therapeutic duodenoscope (TJF160R or TJF160VR, Olympus Corp., Tokyo, Japan) was used for ERC, and selective cannulation of the recipient bile duct was performed with either a guide wire (0.035-inch Jagwire, Boston Scientific, Natick, MA) or a short-nose sphincterotome (Fluorotome, Boston Scientific). No patient showed endoscopic evidence of duodenal lesions (eg, ulcers).

Bile samples were obtained after selective intubation before any therapeutic procedure as previously described.[13] Patients positive for biliary CMV were tested with concurrent pp65 antigen and serum CMV DNA analyses. For patients who underwent repeat ERC for suspected cholestasis during the study, a new bile sample was taken and analyzed for CMV DNA. The endpoints were defined as retransplantation, death, and the last follow-up visit.

Patients were retrospectively assigned to 1 of 2 groups: a nonanastomotic biliary lesion (NABL) group or an AS group. The NABL group comprised patients with NABLs seen during ERC or liver biopsy at any time. An NABL was defined as the presence of an NAS during ERC or regular ERC findings combined with histological signs of biliary damage and chronically elevated alkaline phosphatase and gamma-glutamyltransferase levels. The biliary lesions found in liver biopsy samples included cholangitis (with inflammatory cells around and in the wall and within the lumina of small ducts, degenerative changes in the epithelium, and ductular reactions) and cholestasis (with bilirubinostasis and/or cholate stasis as well as ductular metaplasia). If patients had macroscopic findings of NASs, the liver biopsy results did not change the classification. In addition, we analyzed the data separately for patients with macroscopic lesions and patients with microscopic lesions. The AS group consisted of patients with ASs. Follow-up was continued until May 2011. Patients were excluded if they had only gallstones; in addition, we excluded patients who underwent regular ERC and biopsy that did not indicate any histological evidence of biliary damage and patients who underwent regular ERC without liver biopsy during the study.

Bile Analysis

The bile samples were analyzed with CMV PCR, and after extraction of the DNA, the quantity of DNA was assessed with DNA NanoDrop.

Nucleic acid was isolated from 200 μL of untreated bile with a QIAamp blood kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. A TaqMan real-time PCR assay was performed that targeted the UL 86 region in the CMV genome. For a quantitative analysis of CMV DNA, 5 μL of extracted nucleic acids was amplified with forward primer CMV1 (5′-CAG CCT ACC CGT ACC TTT CCA-3′) and reverse primer CMV2 (5′-GCG TTT AAT GTC GTC GCT CAA-3′), and they were detected with a 5′-FAM-TTC TAC TCA AAC CCC ACC ATC TGC GC-TAMRA-3′ probe. PCR was performed in a reaction volume of 20 μL with a ready-to-use master mix (Roche Diagnostics, Mannheim, Germany) containing Taq DNA polymerase and deoxyribonucleotide triphosphates. Amplification and detection were performed on a LightCycler 480 instrument (Roche Diagnostics) with a thermocycling profile of 95°C for 5 minutes followed by 50 cycles at 95°C for 5 seconds and 60°C for 20 seconds. The detection limit was determined to be 50 copies/mL. In all PCR assays, a positive control with a plasmid containing the CMV UL 86 gene as well as a negative control were used. Before the extraction of DNA from patient samples, a plasmid containing a gene of the alga Volvox carteri was added to each sample and amplified during PCR in parallel to CMV in order to detect any inhibition during the PCR process. Additionally, a CMV DNA quantification standard (3-fold) was used for all assays to allow quantification of the amplified CMV DNA from patient samples. Quantified CMV DNA was expressed as copies per milliliter. PCR was considered valid when all the controls met all the quality criteria.

To evaluate the loss of CMV DNA during the extraction from bile, 2 CMV-negative bile samples were spiked with standardized amounts of the first World Health Organization international standard for human CMV (National Institute for Biological Standards and Controls code 09/162). Serial dilutions containing CMV in amounts ranging from 104 to 102 IU/mL were prepared and extracted as described. Quantitative real-time PCR was performed as described, and the detected amounts of CMV DNA were compared with the initial amount that was spiked into the bile sample. No relevant loss of CMV DNA during the extraction process was observed.

Bile samples from 30 patients with gallstone disease, which were acquired in a manner similar to that used for bile samples from patients after OLT, were all found to be negative for CMV DNA.

Blood CMV Detection

CMV in ethylene diamine tetraacetic acid plasma was routinely assessed via the detection of pp65 antigen in circulating polymorphonuclear leukocytes. For the quantitative CMV pp65 antigenemia assay, approximately 8 mL of an ethylene diamine tetraacetic acid blood sample was used to isolate leukocytes. Then, 500,000 leukocytes were carefully spun on a slide with a CytoSpin centrifuge. Cells were fixed and stained with an anti-CMV pp65 mouse monoclonal antibody and were then washed and further incubated with an anti-mouse, immunoglobulin G, fluorescein isothiocyanate–labeled antibody. Slides were analyzed with ultraviolet microscopy, and CMV pp65 antigen–positive cells were counted.

Data obtained from hospital records included results of serum analyses for CMV DNA or pp65 within the 21 days before or after the initial ERC procedure as well as the CMV status of the transplant recipient and the donor. In addition, for patients positive for biliary CMV DNA, available serum samples, obtained within the 3 weeks before or after positive ERC findings, were analyzed for CMV DNA.

Statistical Analysis

Continuous data were compared with the nonparametric Mann-Whitney U test. The actuarial survival rate was estimated with the Kaplan-Meier product limit estimator. Differences between the actuarial estimates were tested with the log-rank test. Frequency differences were compared with the chi-square test. If frequencies were less than 5, Fisher's exact test was used. P < 0.05 was considered statistically significant. All analyses were performed with PASW Statistics 18.0 (SPSS, Inc., Chicago, IL).


Patients and Follow-Up

Between September 2006 and February 2011, 139 liver transplant patients referred for ERC because of suspected cholestasis successfully underwent bile sampling during ERC. Among the 71 patients in the NABL group, 59 had macroscopic biliary lesions, and 12 had microscopic biliary lesions. The AS group comprised 53 patients (Fig. 1). Fifteen patients were excluded: 3 patients had only gallstones on ERC, and the remaining 12 patients had normal ERC findings with no microscopic biliary lesions or did not undergo liver biopsy within the study period. No biliary CMV was detected among the excluded patients. Table 1 shows the baseline patient demographics.

Figure 1.

Patient flow chart.

Table 1. Patient Demographics
 NABLs (n = 71)ASs (n = 53)P ValueAll Patients (n = 124)Macroscopic Biliary Lesions (n = 59)P ValueOnly Microscopic Biliary Lesions (n = 12)P Value
  1. NOTE: The P values were calculated in comparison with the AS group.

  2. a

    The data are presented as means and standard deviations.

  3. b

    Arterial perfusion–related problems included arterial thrombosis, steal phenomena, and relevant stenosis treated by angiography.

Number of ERC procedures per patienta2.0 ± 1.31.8 ± 1.20.3431.9 ± 1.22.1 ± 1.30.3231.9 ± 1.20.752
Age (years)a52.8 ± 9.954.05 ± 10.90.91953.5 ± 10.353.3 ± 10.30.95352.3 ± 7.10.612
Sex: male/female (n/n)49/2243/100.12792/3240/190.1089/30.632
CMV status (n)        
Donor-positive/recipient-negative: high risk1180.9701980.83630.755
Donor-negative/recipient-positive or donor-positive/recipient-positive: medium risk40330.39172350.80050.120
Donor-negative/recipient-negative: low risk1690.34326120.62440.437
Unknown43 74 0 
Occurrence of CMV episode after OLT (n)23130.34036180.48050.231
Biopsy-proven acute rejection (n)20100.23230170.21930.938
Arterial perfusion–related problems (n)b22130.42935210.20410.399
Donor age (years)a56.554.80.52355.856.0 ± 18.40.65358.8 ± 16.00.409
Cold ischemia time (hours)a11.110.90.70511.011.2 ± 2.80.57910.4 ± 3.00.741
Time after OLT (months)a30.4 ± 57.528.2 ± 42.20.01840.0 ± 126.233.8 ± 61.00.12013.6 ± 33.00.008
Laboratory parameters before ERCa        
Bilirubin (mg/dL)3.7 ± 4.12.5 ± 3.00.0393.2 ± 3.73.2 ± 3.70.2146.3 ± 4.90.001
Glutamic oxaloacetic transaminase (IU/L)104.0 ± 133.392.0 ± 138.30.49098.9 ± 135.199.3 ± 141.00.974127.1 ± 87.30.027
Glutamic pyruvic transaminase (IU/L)149.3 ± 214.1101.0 ± 180.00.110128.6 ± 201.0131.8 ± 175.00.204235.1 ± 347.40.097
Alkaline phosphatase (IU/L)350.7 ± 293.1347.4 ± 349.30.230349.3 ± 320.5343.4 ± 295.20.305387.6 ± 292.30.298
Gamma-glutamyltransferase (IU/L)819.0 ± 1007.2555.7 ± 680.70.055706.5 ± 889.0783.6 ± 1056.50.141993.4 ± 728.60.030
C-reactive protein (mg/L)34.6 ± 47.423.4 ± 41.40.06029.9 ± 45.236.8 ± 51.00.13723.9 ± 21.20.045
White blood cells (/nL)8.8 ± 6.76.4 ± 3.50.1177.7 ± 5.78.7 ± 7.00.1469.4 ± 5.40.065

The mean follow-up period was 29.0 ± 16.6 months. There was no significant difference in follow-up between the NABL group (27.2 ± 15.8 months) and the AS group (31.3 ± 17.4 months, P = 0.14). The average total number of ERC procedures was 1.9 ± 1.2, total of 240 ERC in all patients. Successful bile sampling was performed for all patients. There was no significant difference in the total number of ERC procedures performed per patient between the NABL group (2.0 ± 1.3) and the AS group (1.8 ± 1.1, P = 0.34).

During follow-up, 26 patients died: 19 in the NABL group and 7 in the AS group (P = 0.06). All deaths were due to graft failure. Eleven of the 15 patients who underwent retransplantation were from the NABL group, and 4 were from the AS group (P = 0.18). The NABL group had a significantly higher number of graft losses resulting in death or retransplantation (26 versus 10, P = 0.03).

Biliary CMV Detection

Table 2 compares the biliary CMV status. Biliary CMV was detected more frequently in the NABL group versus the AS group (12 versus 2, P = 0.02; Fig. 2A). After the NABL group was divided according to the presence of macroscopic or microscopic biliary lesions, biliary CMV was detected in 8 of the 59 patients with macroscopic biliary lesions (P = 0.14) and in 4 of the 12 patients with microscopic biliary lesions (P = 0.008) but in only 2 of the 53 patients with ASs (Fig. 2B). The detected levels of biliary CMV were between 50 and 83,000 copies/mL. At the baseline, CMV DNA was detected in 5 patients from the NABL group and in 1 patient from the AS group. During the follow-up, biliary CMV DNA was found in 7 additional patients from the NABL group and in another patient from the AS group. Four patients with biliary CMV had only histological biliary lesions. The median interval between OLT and the detection of biliary CMV was 8.4 months (range = 0.4-212.8 months).

Figure 2.

Biliary CMV in the NABL and AS groups. (A) The numbers of patients who were positive for biliary CMV in the NABL and AS groups were significantly different (P = 0.02). (B) The percentages of patients who were positive for biliary CMV DNA are indicated according to the presence of macroscopic or microscopic lesions detected by ERC or histology only. Eight of 59 patients with macroscopic biliary lesions were positive for CMV, whereas only 2 of 53 patients with ASs were (P = 0.14); 4 of 12 patients with microscopic biliary lesions were positive for CMV, whereas only 2 of 53 patients with ASs were (P = 0.008).

Table 2. Comparisons by the Biliary CMV Status
 Biliary CMV StatusP Value
Positive (n = 14)Negative (n = 110)
  1. a

    The data are presented as means and standard deviations.

Number of ERC procedures per patienta2.4 ± 1.11.9 ± 1.20.06
Age (years)a54.4 ± 14.052.9 ± 9.40.11
Sex: male/female (n/n)12/280/300.30
CMV status (n/N)   
Donor-positive/recipient-negative: high risk1/1418/1060.34
Donor-negative/recipient-positive or donor-positive/recipient-positive: medium risk13/1460/103<0.001
Donor-negative/recipient-negative: low risk0/1425/1030.15
Occurrence of CMV episode after OLT (n)6300.23
Biopsy-proven acute rejection (n)3271.00
Arterial perfusion–related problems (n)1340.11
Donor age (years)a59.9 ± 16.755.3 ± 17.00.39
Cold ischemia time (hours)a10.7 ± 3.111.0 ± 2.80.91
Laboratory parameters before ERCa   
Bilirubin (mg/dL)3.5 ± 4.93.1 ± 3.60.97
Glutamic oxaloacetic transaminase (IU/L)85.9 ± 76.1100.5 ± 141.00.91
Glutamic pyruvic transaminase (IU/L)134.4 ± 150.5127.9 ± 206.10.84
Alkaline phosphatase (IU/L)342.9 ± 292.6350.1 ± 325.10.98
Gamma-glutamyltransferase (IU/L)650.6 ± 653.1713.6 ± 916.80.93
C-reactive protein (mg/L)23.0 ± 37.330.8 ± 46.20.63
White blood cells (/nL)6.6 ± 4.010.7 ± 5.90.75

Four of the 14 patients with biliary CMV (28.6%) underwent repeat ERC with successful biliary sampling during follow-up. Biliary CMV was detected again in the patient from the AS group 5.6 months after the first detection. Four patients with biliary CMV (28.6%) experienced graft failure during the study, whereas 32 of the 110 patients with a negative biliary CMV status (29.1%) did (P = 0.97).

The detection of biliary CMV DNA was not significantly associated with CMV episodes after OLT, biopsy-proven acute rejection, problems related to arterial perfusion, donor age, or cold ischemia time (Table 2). There was no significant difference in the immunosuppression received at the time at which ERC was performed (data not shown).

According to the DNA NanoDrop test, there was no significant difference in the total DNA quantity in bile between the NABL group (13.7 ± 17.3 ng/μL) and the AS group (9.2 ± 8.2 ng/μL, P = 0.24). Furthermore, there was no significant difference in the total DNA quantity in bile between patients who were biliary CMV–positive (18.32 ± 30.2 ng/μL) and patients who were biliary CMV–negative (11.1 ± 10.9 ng/μL, P = 0.66).

For 7 patients who were biliary CMV–positive, liver biopsy samples were available within 2 weeks of ERC; CMV DNA could not be detected in any of the samples by either immunohistochemistry or PCR.

Serum CMV Detection and CMV Serostatus of Donors and Recipients

Blood pp65 analyses 21 days before or after the baseline were performed for 47 of the 71 patients in the NABL group and for 32 of the 53 patients in the AS group. Blood pp65 was present in 1 patient from each group. Both patients tested negative for biliary CMV DNA. For 34 of the 110 patients who were biliary CMV–negative, hospital records showed negative CMV DNA findings in serum. All but 1 of the patients who tested positive for biliary CMV had negative results for pp65 and for CMV DNA in serum; for the 1 patient who tested positive for CMV DNA in bile, pp65 results were not available within 3 weeks, but he was negative for CMV DNA in serum.

The CMV status for the transplant recipients and donors was obtained for 117 of the 124 patients. Biliary CMV was more frequently detected in patients at intermediate risk (donor-positive/recipient-positive or donor-negative/recipient-positive; Table 2). In addition, biliary CMV was more frequently detected in patients with a recipient-positive status (13/73 or 17.8% versus 1/44 or 2.2%, P < 0.05). There was no significant association between biliary CMV detection and a high-risk (donor-positive/recipient-negative) or low-risk CMV status (donor-negative/recipient-negative). Nine of the 53 patients with a positive donor status (17.0%) tested positive for biliary CMV, whereas 5 of the 53 patients with a negative donor status (9.4%) did (P = 0.32).


To our knowledge, this is the first study to investigate CMV DNA in biliary samples from living humans. Biliary CMV DNA was detected by PCR in a substantial number of OLT patients presenting with suspected cholestasis, and it was not detectable in bile from patients suffering from choledocholithiasis only. Positive findings for biliary CMV were significantly associated with macroscopic or histological NABLs. Thus, CMV might play a significant role in the development of nonanastomotic biliary complications.

We also showed that biliary CMV DNA was detected despite negative blood results for pp65 antigen and CMV DNA. This discrepancy between blood CMV detection and ongoing CMV activity in transplant patients has previously been described in CMV colitis.[14] On the one hand, it may be argued that there is a lack of sensitivity for testing pp65 antigen and CMV DNA in serum.[15-17] On the other hand, if we assume the absence of viremia, it can be hypothesized that an occult, ongoing CMV infection of the bile duct system could account for this discrepancy. Furthermore, in patients who underwent lung transplantation, the detection of CMV in bronchoalveolar lavage fluid along with the detection of CMV in blood was associated with the development of bronchiolitis obliterans; moreover, the detection of CMV in tissue biopsy samples is considered most sensitive in patients with inflammatory bowel disease.[18, 19] These findings may be due to local replication of CMV and damage without an indication by a positive result for CMV in blood. This variety of special situations of isolated detection might argue against a special tropism of CMV to bile ducts and instead imply special risk factors for local replication. The identification of these risk factors might be performed in further studies.

CMV disease in OLT patients most frequently presents as CMV syndrome, with the constellation of fever, neutropenia and/or thrombocytopenia, or as CMV hepatitis, as evidenced by elevated levels in liver function tests.[10] Both are associated with viremia.[20-22] However, like other herpes viruses, CMV is known to establish latency in, for example, blood leukocytes. Murine models suggest that CMV latency occurs in epithelial and endothelial cells. In particular, hepatic sinusoidal endothelial cells can harbor CMV and lead to reactivation in the liver.[23] However, in addition to reactivation, CMV can be shed continuously from restricted host sites at low levels.[24] Specifically, endothelial and epithelial cells are associated with chronic virus shedding. Studies have reported that CMV can be detected for years in urine and oral secretions of asymptomatic pediatric patients.[25, 26] Furthermore, CMV is shed in breast milk during the postpartum period.[27]

Although chronic virus shedding can be asymptomatic and not associated with overt disease in immunocompetent hosts, transplant patients may experience localized inflammatory processes resulting from latent CMV shedding. For instance, latent CMV infections have been associated with transplant vascular sclerosis and restenosis after peripheral vascular graft procedures and vasculopathy of cardiac allograft transplants.[28, 29] Furthermore, latent CMV infections histologically detected in tubular epithelial cells of renal allografts have been associated with poor long-term graft function.[30] In general, patients with biliary complications are grouped according to the presence of ASs and NASs, and patients with histologically detected lesions are analyzed separately. This grouping is based on clinical relevance because NASs can be treated by endoscopic or radiological interventions. Patients with only histologically detected lesions are the most difficult to characterize. Therefore, in the present study, we combined patients with classic NASs and histologically detected lesions into a single group because we aimed to include patients with damage to bile ducts of any size (large ducts and/or small ducts); similarly to patients with primary sclerosing cholangitis, the damage in these patients can be restricted to the small ducts only. However, further studies are required to determine additional differences and similarities between these groups. A subgroup analysis showed that biliary CMV detection was more frequent in each subgroup versus patients with ASs only.

We suggest that occult biliary CMV infections may play a critical role in the development of nonanastomotic cholestatic complications in OLT patients. Chronic CMV latency in biliary epithelial cells and viral shedding may induce chronic inflammation and, consequently, lead to fibrotic reactions in the bile duct system. Because screening for CMV in OLT patients is based on symptomatic viremia, an occult biliary CMV infection may also explain the lack of evidence associating CMV infections with the formation of nonanastomotic cholestatic complications within OLT patients within the era of CMV prophylaxis.

In this study, the median interval between OLT and biliary CMV detection was 8.4 months. In 5 patients, biliary CMV was detected more than 12 months after OLT. Biliary CMV detection was associated with a recipient-positive CMV status but not with a high-risk or donor-positive status. Thus, if we assume the presence of occult CMV cholangitis, our findings indicate that biliary CMV emerges most likely during CMV reactivation rather than infection. On the basis of the results of serum, urine, and histological sampling, several studies have described a plateau in the cumulative incidence of CMV manifestation 6 months after OLT.[31-33] The authors suggest that this stable incidence may be caused by the routine tapering of immunosuppression 3 to 4 months after OLT, which is in accordance with the Heidelberg protocol for liver transplantation. Consequently, prolonged CMV prophylaxis is currently being debated for OLT patients at high risk for CMV infection or reactivation in order to decrease the risks of graft loss and death.

Our study did not examine the sensitivity or specificity of biliary CMV sampling. Biliary CMV was detected in 6 patients during the initial ERC procedure. The remaining positive biliary CMV findings were revealed during a second procedure (4 patients) or during a third procedure (4 patients) with successful biliary sampling. Further research is required to determine whether these findings reflect the course of the disease or represent the sensitivity of this test. Among the 124 patients included in the study, 2 patients who were positive for serum pp65 antigen tested negative for biliary CMV. Because cholestasis and cholangitis cause cell damage, PCR testing may be more likely to provide false-positive results in probes with a large quantity of DNA. However, there was no significant difference in the total quantity of DNA in bile between positive and negative bile samples. Hence, the results of this study indicate that biliary CMV sampling can be a useful diagnostic tool, particularly in OLT patients with suspected cholestasis, and may reveal an occult, ongoing CMV infection or reactivation.

In conclusion, we successfully demonstrated that biliary CMV could be detected in a substantial number of OLT patients. Furthermore, biliary CMV is significantly associated with NASs or microscopic biliary lesions. Therefore, CMV infection should be considered as a possible contributing etiological factor to nonanastomotic biliary complications along with well-established factors such as arterial complications, donor age, and cold ischemia time. Further research is required to investigate the sensitivity, specificity, and clinical relevance of biliary CMV sampling. Because of the high morbidity and graft failure rates with NASs, improvements in treatment options and preventive strategies are desperately needed.