Mechanisms of Biliary Stent Clogging: Confocal Laser Scanning and Scanning Electron Microscopy

• Introduction
• Patients and Methods
• Confocal Laser Scanning Microscopy
• Scanning Electron Microscopy
• Results
• Confocal Laser Scanning Microscopy
• Scanning Electron Microscopy
• Discussion
• Conclusions
• References

Background and Study Aims: Endoscopic insertion of plastic biliary endoprostheses is a well-established treatment for obstructive jaundice. The major limitation of this technique is late stent occlusion. In order to compare events involved in biliary stent clogging and identify the distribution of bacteria in unblocked stents, confocal laser scanning (CLS) and scanning electron microscopy (SEM) were carried out on two different stent materials - polyethylene (PE) and hydrophilic polymer-coated polyurethane (HCPC).
Patients and Methods: Ten consecutive patients with postoperative benign biliary strictures were included in the study. Two 10-Fr stents 9 cm in length, one made of PE and the other of HCPC, were inserted. The stents were electively exchanged after 3 months and examined using CLS and SEM.
Results: No differences were seen between the two types of stent. The inner stent surface was covered with a uniform amorphous layer. On top of this layer, a biofilm of living and dead bacteria was found, which in most cases was unstructured. The lumen was filled with free-floating colonies of bacteria and crystals, surrounded by mobile laminar structures of mucus. An open network of large dietary fibers was seen in all of the stents.
Conclusions: The same clogging events occurred in both PE and HCPC stents. The most remarkable observation was the identification of networks of large dietary fibers, resulting from duodenal reflux, acting as a filter. The build-up of this intraluminal framework of dietary fibers appears to be a major factor contributing to the multifactorial process of stent clogging.

Introduction

Endoscopic insertion of plastic biliary endoprostheses is a well-established treatment for obstructive jaundice. The major limitation of this technique is late stent occlusion after a median period of 3-6 months, which necessitates stent exchange [1] [2]. Stent obstruction is caused by biliary sludge, which consists of crystals of calcium bilirubinate and calcium palmitate, as well as proteins, mucopolysaccharides, cholesterol crystals, and bacteria [3] [4].

It is generally assumed that the initial event in stent blockage is adherence of proteins and bacteria to the inner wall of the stent, forming a biofilm. Subsequently, biliary components such as calcium bilirubinate and calcium fatty acid soaps precipitate, since bacteria secrete β-glucuronidase and phospholipases [5] [6] [7]. The bacteria are probably already introduced during transpapillary placement of the stent. In addition, after the stent is in place, bacteria can colonize the biliary tract due to reflux from the duodenum.

The attachment of bacteria to the inner wall of the stent may depend on the surface properties of the polymers used [8]. Various materials have been used in the manufacture of stents: polyethylene, polyurethane, and Teflon. In-vitro studies have found a direct relation between the friction coefficient and the amount of encrusted material [5] [9]. Teflon has the lowest friction coefficient and therefore the maximum potential for preventing stent clogging. A hydrophilic polymer coating was shown to be effective in reducing bacterial adherence in in-vitro studies [10]. The Hydromer stent not only has a smooth texture, but also a coating that absorbs water and provides a hydrophilic sheath. As bacteria initially attach through hydrophobic interactions, this coating may reduce the amount of bacterial adhesion and thus increase the period of stent patency. However, the positive results reported in several in-vitro studies were not confirmed in prospective clinical trials [11] [12] [13] [14].

In the present study, confocal laser scanning (CLS) and scanning electron microscopy (SEM) were carried out in two different stent materials - polyethylene and hydrophilic polymer-coated polyurethane - in order to compare the events involved in stent clogging and identify the distribution of (living and dead) bacteria in unblocked biliary stents.

Patients and Methods

From February 2002 to December 2002, 10 consecutive patients with benign postoperative biliary strictures were included in the study. There were five men and five women, with a mean age of 50 years (range 30 - 64 years). Two 10-Fr stents 9 cm long were inserted in patients with a postoperative bile duct stenosis: one standard polyethylene stent (PE) (Wilson-Cook, Winston-Salem, North Carolina, USA) and one hydrophilic polymer-coated polyurethane stent (HCPC) (Biosearch, Somerville, New Jersey, USA). All of the patients had an indication for the insertion of two biliary plastic stents, based on enrollment in a patient management protocol for the treatment of benign stenosis. This protocol consists of a total of 1 year of stenting with elective stent exchanges every 3 months, while increasing the number of inserted stents in a cumulative fashion. The HCPC stent was soaked in water for 5 min before use. Elective stent exchange was carried out as a standard treatment after 3 months, to prevent cholangitis due to clogging. None of the patients received prophylactic antibiotics. The stents were then examined using CLS within 30 min of stent removal and also using SEM.

Confocal Laser Scanning Microscopy

Rings of approximately 5 mm were cut from the stents - at the distal end, in the center, and at the proximal end of the stent. Specimens were stained for 15 min with SYTO 9 and propidium iodide as a live/dead stain, rinsed in buffer, and imaged. The live/dead viability test used relies on the fact that membranes of dead cells are permeable to many dyes that cannot cross them in the living state. SYTO 9 is a very effective stain, with minimum nonspecific binding during staining of complex communities [15]. A Leica SP2 confocal microscope was used for imaging. Excitation was carried out with the 488-nm line of the argon-ion laser. SYTO 9 was detected between 500 nm and 535 nm, and propidium iodide was detected between 650 nm and 700 nm. For additional structural information, the reflected image of the 488-nm line was acquired.

Images were acquired in a 512 × 512 format (8-bit). An HCL PL APO 20.0 × 0.70 Imm/Corr ultraviolet lens was used, with a zoom factor of between 1 and 5. The pinhole was set at a diameter of one airy disk, corresponding to a z-resolution of approximately 3 μm. Data stacks were generated over a depth of up to 120 μm, with a step size of 2 μm. Using the Leica software, images of SYTO 9 and propidium iodide were merged, and stereo pairs were generated from the three-dimensional image stacks.

Scanning Electron Microscopy

The specimen were fixed in McDowell’s fixative for at least 48 h, dehydrated, and finally dried with hexamethyl disilazane. The dried specimens were mounted on stubs and coated with approximately 10 nm gold. The specimens were imaged with a Philips SEM 525 device, operated at either 5 or 10 kV and equipped with an Orion frame-grabber.

Results

The endoprostheses from each group were analyzed using both confocal laser scanning and scanning electron microscopy. None of the patients presented with symptoms of stent clogging, and none of the patients received prophylactic antibiotics. All analyzed stents were patent, as judged by eye. There were no differences in the amounts or distributions of sludge and bacteria between the polyethylene and hydrophilic-coated polyurethane stents on either CLS or SEM.

Confocal Laser Scanning Microscopy

Both the polyethylene and hydrophilic polymer-coated polyurethane stents showed a similar phenomenon: the layer attached to the internal wall of the stent was a highly reflective amorphous layer (with a thickness of approximately 15 - 30 μm), which was extremely uniform in thickness and distribution in each stent. This layer formed the substrate for the attached bacteria.

In some cases, a structured biofilm was observed, with dead bacteria (red) in the top layer, which is in direct contact with the bile, while covering the underlying living (green) cells (Figure [1]). However, in most cases, this layer was without any structure: dead and living bacterial cells were interspersed, or cloudy areas with living bacteria covered with dead bacteria were seen (Figure [2]). Crystals were occasionally found embedded in these biofilms. Particularly in and around these unstructured biofilms, mucoid-like sheets were found, which in many cases extended into the lumen of the stent. The thickness and density of the unstructured biofilms were more irregular than in the structured biofilms. As far as could be observed, the biofilm attached to the stent wall never extended more than 100 μm into the lumen of the stent. Taking into account the inner diameter of the stent, the reduction in the luminal radius and effect on bile flow were negligible.

The lumen of all the stents examined was filled with free-floating mixed colonies of living and dead bacteria. The size of these colonies varied considerably, from approximately 50 μm in cross-section to a major part of the stent diameter (2.5 mm). These colonies were aggregated in a substance with a slightly higher reflectance than that of the surrounding bile. Small crystals (5 μm or less) were often found dispersed in these loosely attached, floating colonies. If large, free-floating crystals were found (up to 150 μm), in many cases they - or at least their fringes - were covered with a mixture of living and dead bacteria.

In a number of cases, these free-floating colonies were surrounded by laminar structures (probably mucus) - part of a highly mobile complex network of sometimes anastomosing structures with highly reflective surfaces. These structures were sometimes attached to the wall of the stent. In cross-section the size of these empty spaces varied from 50 μm to over 500 μm. These spaces were empty, since neither fluorescent nor reflected signals were present. This is best seen in the three-dimensional image shown in Figure [3].

Reflux of dietary fibers was found in all of the stents. These fibers formed a more rigid, tangled network, which created a kind of filter (Figure [4]).

Scanning Electron Microscopy

Directly on the inner surface of both types of stent, the same amorphous layer of uniform thickness as that observed with CLS was seen (Figure [5 a]). This layer is loosely attached to the wall of the stent. It is consequently easily detached from the stent after dehydration and drying. This layer was covered by bacteria and yeasts, starting to form a biofilm (Figure [6], asterisk). The yeasts can be recognized by their spherical shape and size (Figure [6], arrows). The shape and structure of this layer was variable. Sometimes, a clear structure could be identified, with the bacteria oriented perpendicular to the stent wall and spaces in the biofilm that were empty of bacteria (Figure [5 c]). More often, no distinct orientation was found, with mucus and clusters of micro-organisms dispersed.

A network of amorphous material - probably mucus - in which colonies of bacteria were embedded was also observed (Figure [7]).

Crystals were present in some of the stents, whereas they were absent in others. These crystals were found both organized into biofilms and dispersed throughout the mucus and biomaterial (Figure [5 b]).

Reflux material from the gut was found over the whole length of the stent, often covered with micro-organisms (Figure [8], arrow).

The HCPC stent also showed a very porous wall in the endoprosthesis, in which contrast additives and small holes were seen (Figure [5 a]), whereas the polyethylene stent showed a solid plastic surface.

In contrast to the results of confocal microscopy, which showed a volume filled with highly flexible and transparent structures (in fluorescence and reflected mode), scanning electron microscopy gave the impression that this biomaterial, in whatever shape, always forms a solid obstruction. This is due to a fundamental difference in the imaging methods - confocal laser scanning allows optical sectioning, whereas SEM images only the surfaces.

Discussion

The first finding in this study, identified on both CLS and SEM, was the presence of an amorphous layer covering the stent wall that has not previously been described. In earlier studies, it was mostly blocked stents that were studied, and CLS was not used. Investigation of patent stents has the advantage that the initial stages of clogging can be studied. Particularly in the final stage of occlusion, when biliary flow is almost absent, significant transformations in the clogged material may take place. Various studies have reported amorphous material in the stent lumen, but there have been no reports of an organized layer [6] [15].

In the present study, both CLS and SEM were used to collect information from the same specimens, using two microscopic techniques and integrating the findings. The most important difference between CLS and SEM is that CLS provides fluorescently labeled in-vivo images and noninvasive structural imaging, with a facility for three-dimensional reconstruction. This makes it possible to examine complex, organized structures such as biofilms without disruption or fixation. The live/dead viability test used relies on the fact that membranes of dead cells are permeable to many dyes (in this case propidium iodide, resulting in a red color) that cannot cross them in the living state, whereas other dyes are accumulated by living cells (in this case SYTO 9, resulting in a green color). This staining system is intended for use with pelagic bacteria, and its role in mixed-species biofilm has not been established. However, an impression of the distribution of live and dead cells can be obtained. SYTO 9 has been shown to be a very effective stain, with minimal nonspecific binding during staining of complex biofilm communities [15].

An interesting finding in the present study was that in a number of stents, living bacteria were attached to the stent wall, covered by a layer of dead bacteria. However, this distribution was not uniform in all of the stents, nor in any single stent or between stents. This finding implies that living bacteria may be protected by dead bacteria that form a physical barrier to bacterial penetration, for example. The finding may also explain why antibiotics do not prolong stent patency in clinical trials [16].

Another finding was a mobile, cloudy network of mucus, bacteria, and crystals, including large empty spaces, which was best visualized in three-dimensional CLS images. The fact these spaces were genuinely empty was concluded from the absence of both fluorescent and reflected signals. The occurrence of crystals appears to be a patient-dependent factor. After an amorphous layer on the stent surface, growth of the biofilm connects bacteria to other bacteria, mucus, and crystals, forming microcolonies that are also connected to each other in highly flexible three-dimensional networks.

The most remarkable observation was the presence of large numbers of plant fibers that had refluxed from the duodenum. In the three-dimensional pictures, as well as in the SEM pictures, tangled networks of vegetable material acting as a filter were found. Large fibers may cause sudden obstruction of the stent lumen, independently of accumulation due to the amount of biliary sludge. This view is in accordance with the findings of a recent study, which also reported on the causal role of duodenobiliary reflux and suggested changes in stent design in order to prevent reflux [17].

Since two types of stent were placed (and removed) simultaneously in all of the patients, it was possible to study the effect of different stent materials on the formation of the biofilm. In this study, standard biliary stents made of polyethylene were used, as well as a hydrophilic polymer-coated polyurethane stent. In theory, the coated stent may prevent bacterial adhesion, since bacteria initially attach through hydrophobic interactions. However, no differences in bacterial growth between the two types of unblocked stent were observed in this study. This finding is in accordance with clinical studies in which no differences in stent patency were observed [13] [14].

Two stents placed in parallel were used in this study. Although doubling the surface area of the stent lumen might have influenced the overall patency rate, as suggested by the fact that none of the patients presented with symptoms of stent obstruction, the underlying mechanism of stent clogging must be the same regardless of the number of stents used.

In the past, many investigators have used SEM techniques for biofilm studies, in which the fixation procedures involved can induce morphological changes. CLS is able to demonstrate complex networks in aqueous biofilms by nondisruptive techniques. Both of these microscopic techniques provided an explicit and additive impression of the clogging mechanism. SEM implies that the samples are fixed and subsequently coated with gold, with a static image being obtained only from the outside of the structures of the biofilm. Consequently, the development of biofilm as imaged with SEM suggests that bacterial growth and biofilm formation significantly contribute to stent obstruction. Using CLS, only staining is applied, combined with reflection. No material leaked from the stent rings during staining, implying that the structures inside the stent are imaged in situ. In addition, truly three-dimensional information is obtained from the total material inside the stent and not only from its outer boundaries. Using this true three-dimensional information from CLS, it was proved that the structures concerned are open, transparent, and extremely flexible, forming at most a minor obstruction to bile flow.

It is tempting to speculate that the development of the amorphous layer and the biofilm provide a medium to which dietary fibers adhere, building up a progressively dense network that obstructs bile flow and traps more and more biomaterial until the stent lumen becomes totally occluded. However, given the fragility of the biofilm as demonstrated by CLS, this appears improbable. It seems more likely that large fibers are trapped mechanically in the relatively small stent lumen; however, no definite conclusions can be drawn on the basis of the present data.

Conclusions

Similar clogging events occurred in both the polyethylene and the hydrophilic polymer-coated polyurethane stents. The thickness of the biofilm is so small in comparison with the inner diameter of the stent that it has negligible effects on bile flow. The most notable observation was the identification of networks of large dietary fibers resulting from duodenal reflux, acting as some sort of filter. This suggests that dietary fibers are a major contributing factor to stent clogging. It also explains why different stent surface materials and the administration of antibiotics or bile salts do not have any effect on stent patency.

References

• 1 Huibregtse K, Tytgat G N. Palliative treatment of obstructive jaundice by transpapillary introduction of large bore bile duct endoprosthesis.  Gut. 1982;  23 371-375
  CrossrefPubMedSearch in Google Scholar
• 2 Davids P H, Groen A K, Rauws E A. et al . Randomised trial of self-expanding metal stents versus polyethylene stents for distal malignant biliary obstruction.  Lancet. 1992;  340 1488-1492
  CrossrefPubMedSearch in Google Scholar
• 3 Groen A K, Out T, Huibregtse K. et al . Characterization of the content of occluded biliary endoprostheses.  Endoscopy. 1987;  19 57-59
  Thieme ConnectPubMedSearch in Google Scholar
• 4 Sung J Y, Leung J W, Shaffer E A. et al . Bacterial biofilm, brown pigment stone and blockage of biliary stents.  J Gastroenterol Hepatol. 1993;  8 28-34
  PubMedSearch in Google Scholar
• 5 Coene P P, Groen A K, Cheng J. et al . Clogging of biliary endoprostheses: a new perspective.  Gut. 1990;  31 913-917
  CrossrefPubMedSearch in Google Scholar
• 6 Leung J W, Ling T K, Kung J L, Vallance-Owen J. The role of bacteria in the blockage of biliary stents.  Gastrointest Endosc. 1988;  34 19-22
  CrossrefPubMedSearch in Google Scholar
• 7 Speer A G, Cotton P B, Rode J. et al . Biliary stent blockage with bacterial biofilm: a light and electron microscopy study.  Ann Intern Med. 1988;  108 546-553
  CrossrefPubMedSearch in Google Scholar
• 8 McAllister E W, Carey L C, Brady P G. et al . The role of polymeric surface smoothness of biliary stents in bacterial adherence, biofilm deposition, and stent occlusion.  Gastrointest Endosc. 1993;  39 422-425
  CrossrefPubMedSearch in Google Scholar
• 9 Dowidar N, Kolmos H J, Matzen P. Experimental clogging of biliary endoprostheses: role of bacteria, endoprosthesis material, and design.  Scand J Gastroenterol. 1992;  27 77-80
  PubMedSearch in Google Scholar
• 10 Jansen B, Goodman L P, Ruiten D. Bacterial adherence to hydrophilic polymer-coated polyurethane stents.  Gastrointest Endosc. 1993;  39 670-673
  PubMedSearch in Google Scholar
• 11 Binmoeller K F, Seitz U, Seifert H. et al . The Tannenbaum stent: a new plastic biliary stent without side holes.  Am J Gastroenterol. 1995;  90 1764-1768
  PubMedSearch in Google Scholar
• 12 Van Berkel A M, Boland C, Redekop W K. et al . A prospective randomized trial of Teflon versus polyethylene stents for distal malignant biliary obstruction.  Endoscopy. 1998;  30 681-686
  PubMedSearch in Google Scholar
• 13 Costamagna G, Mutignani M, Rotondano G. et al . Hydrophilic Hydromer-coated polyurethane stents versus uncoated stents in malignant biliary obstruction: a randomized trial.  Gastrointest Endosc. 2000;  51 8-11
  CrossrefPubMedSearch in Google Scholar
• 14 Van Berkel A M, Bruno M J, Bergman J J. et al . A prospective randomized study of hydrophilic polymer-coated polyurethane versus polyethylene stents in distal malignant biliary obstruction.  Endoscopy. 2003;  35 478-482
  Thieme ConnectPubMedSearch in Google Scholar
• 15 Lawrence J R, Neu T R, Swerhone G DW. Application of multiple parameter imaging for the quantification of algal, bacterial and exopolymer components of microbial biofilms.  J Microbiol Methods. 1998;  32 253-261
  CrossrefPubMedSearch in Google Scholar
• 16 Sung J J, Sollano J D, Lai C W. et al . Long-term ciprofloxacin treatment for the prevention of biliary stent blockage: a prospective randomized study.  Am J Gastroenterol. 1999;  94 3197-3201
  CrossrefPubMedSearch in Google Scholar
• 17 Weickert U, Venzke T, Konig J. et al . Why do bilioduodenal plastic stents become occluded. A clinical and pathological investigation on 100 consecutive patients?.  Endoscopy. 2001;  33 786-790
  Thieme ConnectPubMedSearch in Google Scholar

M. Bruno, M. D., Ph. D.

Dept. of Gastroenterology and Hepatology C2 - 220

Academic Medical Center · Meibergdreef 9 · 1105 AZ Amsterdam · The Netherlands

Fax: +31-20-6917033

Email: m.j.bruno@amc.uva.nl

References

• 1 Huibregtse K, Tytgat G N. Palliative treatment of obstructive jaundice by transpapillary introduction of large bore bile duct endoprosthesis.  Gut. 1982;  23 371-375
  CrossrefPubMedSearch in Google Scholar
• 2 Davids P H, Groen A K, Rauws E A. et al . Randomised trial of self-expanding metal stents versus polyethylene stents for distal malignant biliary obstruction.  Lancet. 1992;  340 1488-1492
  CrossrefPubMedSearch in Google Scholar
• 3 Groen A K, Out T, Huibregtse K. et al . Characterization of the content of occluded biliary endoprostheses.  Endoscopy. 1987;  19 57-59
  Thieme ConnectPubMedSearch in Google Scholar
• 4 Sung J Y, Leung J W, Shaffer E A. et al . Bacterial biofilm, brown pigment stone and blockage of biliary stents.  J Gastroenterol Hepatol. 1993;  8 28-34
  PubMedSearch in Google Scholar
• 5 Coene P P, Groen A K, Cheng J. et al . Clogging of biliary endoprostheses: a new perspective.  Gut. 1990;  31 913-917
  CrossrefPubMedSearch in Google Scholar
• 6 Leung J W, Ling T K, Kung J L, Vallance-Owen J. The role of bacteria in the blockage of biliary stents.  Gastrointest Endosc. 1988;  34 19-22
  CrossrefPubMedSearch in Google Scholar
• 7 Speer A G, Cotton P B, Rode J. et al . Biliary stent blockage with bacterial biofilm: a light and electron microscopy study.  Ann Intern Med. 1988;  108 546-553
  CrossrefPubMedSearch in Google Scholar
• 8 McAllister E W, Carey L C, Brady P G. et al . The role of polymeric surface smoothness of biliary stents in bacterial adherence, biofilm deposition, and stent occlusion.  Gastrointest Endosc. 1993;  39 422-425
  CrossrefPubMedSearch in Google Scholar
• 9 Dowidar N, Kolmos H J, Matzen P. Experimental clogging of biliary endoprostheses: role of bacteria, endoprosthesis material, and design.  Scand J Gastroenterol. 1992;  27 77-80
  PubMedSearch in Google Scholar
• 10 Jansen B, Goodman L P, Ruiten D. Bacterial adherence to hydrophilic polymer-coated polyurethane stents.  Gastrointest Endosc. 1993;  39 670-673
  PubMedSearch in Google Scholar
• 11 Binmoeller K F, Seitz U, Seifert H. et al . The Tannenbaum stent: a new plastic biliary stent without side holes.  Am J Gastroenterol. 1995;  90 1764-1768
  PubMedSearch in Google Scholar
• 12 Van Berkel A M, Boland C, Redekop W K. et al . A prospective randomized trial of Teflon versus polyethylene stents for distal malignant biliary obstruction.  Endoscopy. 1998;  30 681-686
  PubMedSearch in Google Scholar
• 13 Costamagna G, Mutignani M, Rotondano G. et al . Hydrophilic Hydromer-coated polyurethane stents versus uncoated stents in malignant biliary obstruction: a randomized trial.  Gastrointest Endosc. 2000;  51 8-11
  CrossrefPubMedSearch in Google Scholar
• 14 Van Berkel A M, Bruno M J, Bergman J J. et al . A prospective randomized study of hydrophilic polymer-coated polyurethane versus polyethylene stents in distal malignant biliary obstruction.  Endoscopy. 2003;  35 478-482
  Thieme ConnectPubMedSearch in Google Scholar
• 15 Lawrence J R, Neu T R, Swerhone G DW. Application of multiple parameter imaging for the quantification of algal, bacterial and exopolymer components of microbial biofilms.  J Microbiol Methods. 1998;  32 253-261
  CrossrefPubMedSearch in Google Scholar
• 16 Sung J J, Sollano J D, Lai C W. et al . Long-term ciprofloxacin treatment for the prevention of biliary stent blockage: a prospective randomized study.  Am J Gastroenterol. 1999;  94 3197-3201
  CrossrefPubMedSearch in Google Scholar
• 17 Weickert U, Venzke T, Konig J. et al . Why do bilioduodenal plastic stents become occluded. A clinical and pathological investigation on 100 consecutive patients?.  Endoscopy. 2001;  33 786-790
  Thieme ConnectPubMedSearch in Google Scholar

M. Bruno, M. D., Ph. D.

Dept. of Gastroenterology and Hepatology C2 - 220

Academic Medical Center · Meibergdreef 9 · 1105 AZ Amsterdam · The Netherlands

Fax: +31-20-6917033

Email: m.j.bruno@amc.uva.nl