Isolated microorganisms in plastic biliary stents placed for benign and malignant diseases

Introduction

Since its first application in 1968. endoscopic retrograde cholangiopancreatography (ERCP) [1] remains an effective procedure for diagnosis and mainly for treating diseases of biliary and pancreatic tracts. The evolution of other non-invasive imaging techniques, such as magnetic resonance cholangiopancreatography (MRCP) [2] and endoscopic ultrasound (EUS) [3], has limited the diagnostic prospective of ERCP [4]. However, endoscopic sphincterotomy, stone extraction and stent placement have established ERCP as a gold standard treatment for a variety of malignant and benign diseases of biliary and pancreatic ducts [5]. When plastic stents are chosen to maintain bile duct patency, stent occlusion with consequent bile stasis and cholangitis constitutes one of the major late complications [6,7]. The elimination of the anti-microbial barrier of Oddi [8] and the low pressure in common bile duct, due to endoscopic sphincterotomy and endoprosthesis insertion leads to duodenal reflux, allowing bacterial colonization and biofilm formation, resulting to stent occlusion [7,9,10]. This study was performed to identify the spectrum of microbial flora involved in endoprosthesis colonization and biofilm formation, as well as to evaluate their resistance to advanced antibiotic therapy.

Patients and methods

During a 3-year period, 51 plastic biliary stents were extracted from 42 patients. There were 25 female and 17 male patients with a median age of 71 years (range 39-92). Twenty three stents were inserted for treating malignant disease and 28 for benign disease (Table 1). A total of 31 straight polyurethane stents with side flaps (20 had a diameter of 10-Fr and 11 of 7-Fr) and 20 polyurethane pig-tail stents with side holes (7-Fr) (Wilson-Cook Medical Inc.) were extracted. Stent extraction was decided on a scheduled basis up to 3 months after placement for benign disease or if clinical signs of stent clogging such as cholangitis, recurrent jaundice or biliary colic with elevated liver function tests developed.

Table 1

Main indication for the endoscopic intervention

All patients with symptoms of stent occlusion had been administered antibiotics for various lengths of time before ERCP. Patients scheduled for stent removal were given a single dose of antibiotic before ERCP, usually cefuroxime or ciprofloxacin.

Results

Sixteen stents were schedule explanted and 35 because of clinical signs of occlusion. Fifty stents were endoscopically retrieved and one surgically during a Whipple’s procedure. The median stent time to removal was 90 days (range, 9-730).

All our samples were positive for bacterial growth. Monomicrobial growth was present in two stents, whilst polymicrobial in all the remaining.

There were 162 growths from 16 different species. The median number of bacteria per stent was 3 (range, 1-6). There was statistical corellation between stent remaining time and number of organisms isolated (P<0.05) (Fig. 1). Table 2 shows the distribution of different microbial species in the stents. The most frequently encountered organisms were Enterococcus spp (74%), Escherichia coli (E. coli) (62%) and Klebsiella spp (58%).

Figure 1

Scatterplot of stent duration in place vs. number of organisms Pearson correlation of remaining time (days) and N of organisms=0,449 P-Value=0.001

Table 2

Distribution of microbial species isolated in biliary stents

There were no statistical differences between the organisms isolated between straight vs. pig-tail stents (P>0.05) and 10-Fr vs. 7-Fr stents (P>0.05). E. coli was more frequently encountered in benign disease vs. malignant disease (78% vs.43%; P<0.05). There was no statistical difference in the other organisms between malignant and benign disease (P>0.05). Klebsiella spp, Pseudomonas spp, and Candida spp were more frequently isolated in occluded vs. non-occluded stents, 68% vs. 37%, 22% vs. 0 and 40% vs. 6% respectively (P<0.05) (Table 3).

Table 3

Microbial species isolated in biliary stents according to disease and stent characteristics

There was no statistical difference between organisms, as regards stent duration in place (P>0.05) (Table 4).

Table 4

Isolated bacteria and stent duration in place

The screening for potential resistance of bacteria was performed by defining specific type β-lactamases (e.g. ESBLs, AmpC), carvapenemases (e.g. MBLs, Klebsiella pneumoniae carbapenemase-KPC) and resistance to advanced antibiotics (quinolones and vancomycin) (Table 5). Eleven of 32 (34%) isolated E. coli strains, and 4 of 8 (50%) Pseudomonas spp strains were resistant to quinolones. Five of 14 (35%) isolated Enterobacter spp strains expressed Amp-C derepression and were highly resistant. Enterococcus spp and Klebsiella spp had a low resistance rate.

Table 5

Antibiotic resistance of the most frequently isolated organisms

There was no statistical difference in the resistance rate for E. coli and Enterobacter species, between occluded and non-occluded stents (P>0.05), and between malignant and benign diseases (P>0.05).

Discussion

The occlusion of endoscopically placed plastic biliary stents is the most frequent cause of biliary infection and recurrent jaundice, increasing morbidity and health care cost due to the need of stent exchange [9]. The occlusion mechanism is primarily based on bacterial colonization, as intestinal bacterial flora has ascending access to the biliary system due to elimination of the barrier function of the sphincter of Oddi. Subsequent biofilm formation on the inner surface of the stent by amorphous sludge from microbial byproducts, proteins, dietary fibers and biliary salts induce stent lumen encasement [11-14]. The outer surface of the distal end of the stent comes in direct contact with the duodenal lumen and our preparation technique was suggested to obtain samples from the inner surface where the bacteria are attached and the biofilm is formatted. However, a limitation is that contamination via the endoscope, when it is passed through the duodenum, or via cross-transmission between different patients cannot be excluded, despite vigilant disinfection.

A polymicrobial growth in bile or stents is quite common with an incidence ranging from 8 to 67% [15-17]. Nearly all our stents (96%) had polymicrobial growth. This high percentage may be due to the way our samples were obtained. The number of bacteria isolated per stent was in correlation with stent duration in place (P<0.05), and the longer duration of stent (P<0.05). On the contrary, the type of bacteria isolated was not related to stent duration (P>0.05).

Microbial organisms isolated from biliary stents include both aerobic and anaerobic species, as well as fungi [9]. Enterococcus spp, E. coli, and Klebsiella spp are the most common bacteria isolated from the sludge removed from biliary stents. However, the ratio between the isolated organisms varies in different studies, probably depending on either the portion of the stent analysed (proximal or distal part) or the protocol of sampling and microbiological analysis [9].

Another limitation in our study is that patients with symptoms of stent occlusion, and especially cholangitis, received prolonged treatments with antibiotics before stent exchange and that may influence the type of organisms cultured and their sensitivity.

Organisms isolated from the sludge of biliary stents are similar in 47% of patients to those isolated from blood in patients with biliary sepsis [17]. Isolation of similar organisms from blood and from bile shows a wide spectrum from 21-67% of the patients with bacteremia [16,18].

In our study Enterococcus spp (74%), E. coli (62%), and Klebsiella spp (58%) were also the most frequently isolated organisms. Stent diameter (10-Fr vs. 7-Fr) and shape (straight vs. pig-tail) did not play a role in the frequency of organisms. The same bacteria were isolated for benign and malignant disease, with the exception of E. coli. E. coli was more frequently isolated in benign diseases (P<0.05). Analysis of bile and stent samples of patients with gallstone disease has showed E. coli and Gram-negative bacteria to be most commonly found [17,19].

Klebsiella spp, Pseudomonas spp, and Candida spp were more frequently isolated in occluded than non-occluded stents (P<0.05). This may be due to more hospital admissions and more antibiotics used in patients with occluded stents. Pseudomonas spp was not isolated in any non-occluded stent scheduled for extraction.

The incidence of anaerobes in our study was very low (2%). Isolated anaerobic bacteria rates in the literature are controversial because of the difficulties of isolation and proliferation style of some facultative-anaerobic organisms [9,20,21]. However, anaerobic bacteria are suggested to play a significant role in biliary stent clogging [22] and anaerobic therapy is suggested if a biliary-enteric anastomosis is present [23].

The clinical presentation of cholangitis associated with stent occlusion ranges from mild abdominal discomfort and pyrexia to life-threatening septic shock. Antimicrobial therapy is usually empirical. Initial therapy should cover the Enterobacteriaceae, especially E. coli and relief of biliary obstruction is mandatory. Based on pharmacokinetic studies and in vitro susceptibility findings, ciprofloxacin was suggested to be superior to other antibiotics in prophylaxis and treatment of biliary sepsis [24,18]. In our study, E. coli, Klebsiella spp, and Pseudomonas spp were found to be resistant to quinolones in 34%, 20%, and 50% respectively. There was no difference in E. coli resistance between occluded and non-occluded stents (P>0.05). Therefore, quinolones may not be adequate either for prophylaxis in patients scheduled for stent exchange, or for treatment of cholangitis associated with stent occlusion. In Europe, the quinolone resistance rate, for hospitalized patients, ranges from 6% (France) to 20% (Spain) [25]. A recent paper demonstrated that E. coli resistance to quinolones increased significantly (P<0.0005) between 2007 (20.0%) and 2011 (29.2%) in Canadian hospitals [26]. The increased and often unjustified use of quinolones in the community and in hospitalized patients may have contributed to the high resistance rate in our sample.

Interpreting our results (Table 5) we could suggest that antibiotics combined with ESBLs inhibitors like tazobactam, clavulanate or sulbactam might be the first option in patients with cholangitis associated with stent occlusion, which is in agreement with recent literature [23].

Enterococcus species were the most frequently isolated organisms (74%), but their pathogenicity in biliary tract infections remains unclear [17,27]. The need for antibiotic covering against Enterococcus should be limited to immunocompromised patients with symptoms of stent occlusion and those in severe sepsis [28].

Candida species were isolated in 40% of occluded stents. Our results are consistent with Negm et al [15], where a 31% rate of candida infection in aspirated bile, was found in patients with biliary stents. Fungal infection of the biliary tract is difficult to diagnose and the isolation of fungi may represent only colonisation due to contamination or selection of patients [29]. In patients with reccurent cholangitis or sepsis associated with stent occlusion and malignant disease under chemotherapy, fungal infection has to be taken into account, when designing anti-infectious treatment.

In conclusion, Enterococcus species, E. coli, and Klebsiella species are the most frequently isolated organisms in plastic biliary stents. In occluded stents Pseudomonas species and Candida species should be taken into account. Quinolones may not to be adequate for the treatment of cholangitis associated with stent occlusion. In patients under chemotherapy for malignancy and stent occlusion related biliary sepsis, antifungal and enterococcal covering should be considered.