- Email: [email protected]
Polyethylene stent blockage: a porcine model
Polyethylene stent blockage: a porcine model Nicolas Maillot, MD, Philippe Aucher, PhD, Stephane Robert, MD, Jean Pierre Richer, MD, Didier Bon, MD, Christian Moesch, PhD, Ghislaine Grollier, MD, PhD, Jacques Irani, MD, Michel Carretier, MD, Michel Beauchant, MD Poitiers and Limoges, France
Background: Endoscopic insertion of biliary stents is a useful treatment for obstructive jaundice resulting from unresectable tumors of the pancreas and biliary tree. The main drawback is the recurrence of jaundice due to clogging. The aim of this study was to establish an experimental model of polyethylene stent clogging in large white pigs. Methods: A straight polyethylene stent of 5F (group I), 7F (group II) or 10F size (group III) was inserted in the common bile duct. Animals were killed at 2 months, or earlier if physical signs suggesting stent clogging occurred. Chemicophysical analysis of stent deposition combined stereomicroscopy and identification of the contents by means of Fourrier transform infrared spectroscopy. Bacteriologic analyses included identification of aerobic and anaerobic bacteria and measurement of β-glucuronidase, lecithinase and lipase activities. Results: Physical signs suggesting stent obstruction or death occurred in 8 of 8 animals in group I, 11 of 12 in group II, and 2 of 8 in group III (p < 0.001). The proportion of mucoprotein in the stent contents tended to fall with increasing stent diameter (mean 82%, 58% and 47% for 5F, 7F and 10F, respectively), whereas wheat starch and calcium bilirubinate content increased with increasing stent diameter (9% and 4%, 18% and 10%, and 29% and 23% for 5F, 7 F, and 10F, respectively), although none of these differences were statistically significant. A variety of bacteria were cultured from the stent deposits, including anaerobic strains. Clostridium species were associated with the highest enzyme activities. Conclusions: In this model the major component of early stent deposits was mucoprotein, and numerous aerobic and anaerobic bacteria were isolated. Formation of calcium bilirubinate was a late phenomenon and poorly related to bacterial enzymatic activities. (Gastrointest Endosc 2000;51:12-8.)
Endoscopic insertion of biliary stents is a useful and well-tolerated treatment for obstructive jaundice resulting from unresectable tumors of the pancreas and biliary tree.1 Stent insertion relieves jaundice and pruritus and improves quality of life.2 The main drawback is the recurrence of jaundice or cholangitis due to clogging. Self-expanding metal stents offer durable patency but are expensive. Clogging of polyethylene stents is due to sludge adhesion. The main sludge components are proteins, fibers and bacteria.1 Bacterial adherence to the stent wall may be the initial event in stent blockage,3 but Received September 25, 1998. For revision February 4, 1999. Accepted June 23, 1999. From the Laboratory of Multivisceral Transplantation, INRA Le Magneraud, Department of Surgery, Laboratory of Microbiology and Liver Unit, University Hospital, Poitiers, and Department of Hygiene, Limoges University, Limoges, France. Presented in part at the Digestive Disease Week, San Francisco, California, May 1996. Supported by grants from Laboratoires Houdé France, Wilson Cook Medical France and Ligue Poitou-Charentes contre le Cancer. Reprint requests: Michel Beauchant, MD, Service d’Hépato-gastroentérologie, BP577, 86021 Poitiers, France. Copyright © 2000 by the American Society for Gastrointestinal Endoscopy 0016-5107/2000/$12.00 + 0 37/1/100920 12
GASTROINTESTINAL ENDOSCOPY
antibiotic prophylaxis has given variable results.4-6 To our knowledge, an experimental model of stent clogging has never been extensively described, although such a model might prove useful for the study of mechanisms leading to stent clogging and subsequently the development of new drug treatments7 or stent designs aimed at preventing or retarding blockage. The aim of this study was to establish an experimental pig model of polyethylene stent clogging. METHODS Animals We used 48 healthy large white pigs of both genders, aged 10 to 12 weeks and weighing 20 to 35 kg. They were kept in a conventional closed housing system at 22° ± 2°C and given a piglet diet (INRA 2213; Ets Arrive, St Fulgent, France) and tap water ad libitum. The study was approved by our institutional review committee. The animals were cared for in accordance with university and national guidelines (French Ministry of Agriculture license number for experimental studies: no. 04355) and National Institutes of Health guidelines. General anesthesia was administered as previously described.8 Briefly, food was withheld overnight, and preoperative tranquilization was obtained by rapid atraumatic nasal administration of 0.2 mg/kg midazolam. Anesthesia was induced with halothane and 100% oxygen. A 20-gauge VOLUME 51, NO. 1, 2000
Polyethylene stent blockage: a porcine model
plastic catheter (Insyte W; Becton Dickinson Vascular Access Inc, Sandy, Utah) was inserted into an ear vein. Atropine sulfate 10 µg/kg was given intravenously to reduce pharyngeal and tracheal secretion and prevent postintubation bradycardia. Neuromuscular blockade was induced by injection of suxamethonium iodide, 1 mg/kg. An orotracheal tube (Portex 6-6.5; Portex SA, Berck sur mer, France) was immediately inserted. Epidural anesthesia (lidocaine 1% and 1:200,000 epinephrine) was followed by discontinuation of halothane and maintenance of general anesthesia (1.5% isoflurane in 50% nitrous oxide–50% oxygen). Epidural anesthesia was used to avoid the need for morphine analgesia during surgery. Lansoprazole (30 mg orally, once daily for 5 days; Houdé Laboratories, Paris La Défense, France) was used to prevent bleeding from peptic stress ulcers. After laparotomy, transversal choledochotomy was performed and a straight polyethylene stent with side flaps (Cotton-Leung biliary stent; Wilson-Cook Medical, Winston-Salem, N.C.) was inserted. Animals were assigned to receive a 5F (group I), 7F (group II) or 10F stent (group III). The stent was attached to the posterior wall of the common bile duct (one external suture with Prolene 4.0) to avoid a high rate of spontaneous migration and emerged through the major papilla. The bile duct was ligated to the stent. In the control group, gallbladder bile was collected transhepatically after laparotomy and immediately stored at –80°C for bacteriologic analysis. The animals were monitored daily. They received orally 3157 kcal per day and wheat starch was the main component of their feeding (40%). Physical signs that suggested stent clogging were as follows: altered general status with weight loss (instead of an expected weight gain of 5 kg/week) and anorexia, jaundice, and/or clinical sepsis. Animals were killed within 12 hours of onset of these criteria; the remaining animals were killed at 2 months. In case of sudden death, autopsy was performed within 12 hours. Autopsy focused on macroscopic evidence of bile stasis proximal to the stent, i.e., hepatomegaly with a dark-green appearance or a dilated common bile duct proximal to the stent. The stent was extracted surgically under aseptic conditions and immediately cut into 2 equal parts. The upper (proximal) part was stored at –80°C for bacteriologic analysis, and the lower (distal) part was stored at room temperature for chemicophysical analysis of its desiccated contents. Chemicophysical analysis Chemicophysical analysis of the distal part of the prosthesis was carried out as previously reported.9 The surface and longitudinal sections were examined with a stereomicroscope (Wild M3Z, magnification × 6.5 to 80; Wild Heerbrugg Ltd., Heerbrugg, Switzerland). Because stent deposits were not homogeneously distributed on the luminal wall of the stent, the site with the thickest layer of deposits was localized and quantification was done with an eyepiece micrometer and expressed as percentage luminal obstruction. Samples of the deposits with different structural aspects were carbon coated and examined by means of scanning electron microscopy and energydispersive x-ray analysis (EDAX) (Jeol JFM 35 apparatus VOLUME 51, NO. 1, 2000
N Maillot, P Aucher, S Robert, et al.
coupled to an EDAX 9100/60 x-ray analyzer; Jeol Ltd., Tokyo, Japan). Fourier transform infrared spectroscopy of the desiccated deposits was carried out with a Bruker IF 28 CS spectrophotometer (Bruker Analytische Messtechnik GmbH, Rheinstetten, Germany). The samples were prepared by using the KBr pellet technique (3 mm in diameter). Compounds were identified and their proportions were determined by comparison with the reference spectra of pure compounds and their mixtures. Reference spectra of calcium phosphates (amorphous carbonated calcium phosphate, carbonate apatite and brushite), calcium carbonates (calcite, aragonite and vaterite), cholesterol monohydrate and mucoproteins were taken from the Bruker FTIR library for the analysis of urinary stones (Birsy search software).10 The other reference spectra (calcium salts of fatty acids, bile acids and bilirubin) were obtained with compounds from a commercial supplier (Fluka Chemie AG, Buchs, Switzerland) or were synthesized in our laboratory.9,11 Polarizing light microscopy (Nikon Labophot 2-Pol equipped with a photomicrographic HFX-DX system; Nikon Corporation, Tokyo, Japan) was used to detect birefringent elements (mainly wheat starch granules). Other microchemical analyses9 were done to detect phosphate anions in mixtures primarily containing calcium bilirubinate or to discriminate between protein and traces of blood. Bacteriologic analysis One milliliter of sterile saline was injected through the lumen of the proximal part of the stent and gently vortexed. The crude extract was diluted and spread on the following agar plates: meat-yeast agar (VL Sanofi Pasteur, Gentilly, France) enriched with sheep blood (5%) and meat-yeast agar with blood and gentamicin (40 µg/mL). After 72 hours of incubation in aerobic and anaerobic conditions, viable bacteria were counted and identified using ATB strips (bioMérieux, Lyon, France). Bacterial lecithinase and lipase activities were studied by inoculating egg yolk agar plates (EYA, Oxoid, France). β-Glucuronidase and several esterase activities were measured at pH 7 using the API Zym system (bioMérieux). A score of 0 to 5 was assigned by comparison with a color chart. Bile collected from the control group was analyzed in the same way. Statistical analysis The prevalence of stent clogging was analyzed according to stent size and the presence of major bacterial species by means of the log rank test. Relative amounts of mucoproteins, wheat starch and calcium bilirubinate were compared within the 3 groups by using Spearman’s correlation test. Relationship between physical signs of obstruction or death and bacterial isolates was assessed using the Fisher exact test. The level of significance was set at p < 0.05 in all tests (SAS stat 6.10; SAS Institute Inc., Cary, N.C.).
RESULTS Twelve of the 48 pigs were not included in the final analysis because of spontaneous migration of GASTROINTESTINAL ENDOSCOPY
13
N Maillot, P Aucher, S Robert, et al.
A
Polyethylene stent blockage: a porcine model
B
C
Figure 1. Characteristics of deposits observed in clogged stents by scanning electron micrography. A, Wheat starch and some cellulose fibers were the main compounds in a 7F stent deposit (orig. mag. ×1000); B, deposits located on the inner surface of a 10F stent: cholesterol crystals entrapped by mucoproteins (orig. mag. ×1000) and (C) at higher magnification, fluffy appearance of the surface of the deposit, corresponding to bacteria entrapped in mucoproteins (orig. mag. ×10,000).
Table 1. Clinical outcome of 28 pigs after surgical stent insertion No. of animals
Group I (n = 8)
Presence of physical signs suggesting stent obstruction Weight loss Jaundice Sepsis Sudden death Median time to death (days, range) No. of animals killed at 2 mo (up to 70 days) without physical signs of obstruction
8 8 5 1 3 7 (4 to 16) 0
Group II (n = 12)
Group III (n = 8)
11 9* 4 2 4 11 (5-28) 1
2 2 2 0 2 63† (14 to 70) 6
*Weight loss was the only sign suggesting clogging in 2 of these animals. of stent patency was significantly longer in group III than in group I plus II (p < 0.001, log rank test).
†Duration
the stent (n = 8) or because an autopsy could not be carried out immediately (n = 4). Twenty-eight of the remaining animals received a biliary stent: 8 in group I (5F stent), 12 in group II (7F stent) and 8 in group III (10F stent). The control group comprised 8 animals without stents and was used for bacteriologic analysis of gallbladder bile. Outcome Clinical outcome is shown in Table 1. Among the physical signs suggesting clogging, weight loss alone was noted in only 2 animals in group II. In 5 of them, the stent was not clogged and there was no evidence of bile stasis at laparotomy. The duration of stent patency was significantly longer in group III than in groups I and II combined (p < 0.001). The remaining 7 animals (1 in group II and 6 in group III) were killed at 2 months, and macroscopic evidence of bile stasis was noted in 4. 14
GASTROINTESTINAL ENDOSCOPY
Chemicophysical analysis The median percentage obstruction of the stent (distal part, desiccated contents) was 10% in the 3 groups. No deposits were seen in 8 stents: 1 group I stent, 6 group II stents and 1 group III stent. At autopsy or death, 5 had no evidence of stent obstruction despite the presence of clinical criteria of clinical stent clogging: 3 of them had clinical jaundice, increased bilirubin level and common bile duct dilatation and 2 of them weight loss only. Chemical analysis and scanning electron microscopy revealed that the major components were mucoproteins, wheat starch and calcium bilirubinate (Table 2; Fig. 1). Food fibers were rare and were only detected by scanning electron microscopy (Fig. 1A). Mucoprotein deposition tended to fall with increasing stent diameter (mean 82% in group I, 58% in group II and 47% in group III), whereas wheat starch deposition increased with increasing stent diameter VOLUME 51, NO. 1, 2000
Polyethylene stent blockage: a porcine model
N Maillot, P Aucher, S Robert, et al.
Table 2. Chemicophysical analysis of desiccated deposits by means of Fourier transform infrared spectroscopy Group I (n = 8) Stents without deposits Composition of stent deposits (mean % dry weight)* Mucoprotein Wheat starch Calcium bilirubinate Calcite Apatite Amorphous carbonated calcium phosphate Calcium palmitate Cholesterol Sodium ursodeoxycholate
Group II (n = 12)
Group III (n = 8)
1
6
1
82 9 4 2 0 4 0 0 0
58 18 10 2 4 2 2 0 5
47 29 23 0 0 0 0 1 0
*No significant differences were observed among the 3 groups; food fibers (cellulose) were rare and were only detected by scanning electron microscopy.
Table 3. Organisms cultured from surgically extracted stents (all were infected) Group I (n = 8) Aero-anaerobic bacteria Gram negative rod (n = 22) E coli Other Enterobacteria Gram positive cocci (n = 53) Staphylococcus spp Streptococcus spp Enterococcus spp Anaerobic bacteria Gram positive rod (n = 21) C perfringens Other gram positive rod Gram negative rod (n = 21) Bacteroides spp Other gram negative rod Gram positive cocci (n = 1) Peptostreptococcus spp Yeast (n = 1) Candida albicans
Group III (n = 8)
5 1
8 4
2 2
8 8 2
2 9 10
0 6 8
6 0
9 2
2 2
3 1
9 2
4 1
1
0
0
0
1
0
(9% in group I, 18% in group II and 29% in group III), as did calcium bilirubinate deposition (4% in group I, 10% in group II and 23% in group III), although none of these differences were statistically significant. The highest levels of calcium bilirubinate were observed in 2 group III stents (60% and 70% of stent deposits). Bacteriologic findings All of the gallbladder bile samples from the control group were sterile. Bacteriologic analysis could not be done on 2 group II stents because of technical problems. The remaining 26 stents yielded 117 bacterial isolates (mean 4, range 1 to 11 isolates per stent; mean density 1 × 105 organisms/mL; Fig. 1C) and 1 yeast (Table 3). Among the 75 aerobes and facultative anaerobes, the most frequent species were VOLUME 51, NO. 1, 2000
Group II (n = 10)
Escherichia coli (15 stents, 13%), Staphylococcus spp (10 stents, 9%), Streptococcus spp (23 stents, 20%) and Enterococcus spp (20 stents, 17%). Among the 42 strict anaerobes, Clostridium perfringens was identified in 17 stents (15%) and Bacteroides spp in 16 (14%). Physical signs suggesting obstruction or death were significantly associated with the presence of E coli (p = 0.014) and strict anaerobes (p = 0.028). Both types of organism were identified in all the animals that died from stent clogging. Results did not differ in terms of number of bacterial isolates and presence of strict anaerobes whether analysis was performed at autopsy (median 6 isolates, range 1 to 11, and presence of anaerobes in 8 of 9 stents) or at death (median 3 isolates, range 1 to 7, and presence of anaerobes in 11 of 17 stents). Enzyme activities are shown in Table 4. E coli had low enzyme GASTROINTESTINAL ENDOSCOPY
15
N Maillot, P Aucher, S Robert, et al.
Polyethylene stent blockage: a porcine model
Table 4. Enzymatic activities of the bacterial isolates Esterase C4
Esterase lipase C8
2.7 2.3 2.8 0.3 2.8 0.6
2.4 1.9 3.4 0 1.8 1.1
Staphylococcus spp (n = 10) Streptococcus spp (n = 23) Enterococcus spp (n = 20) E coli (n = 17) Clostridium spp (n = 19) Bacteroides spp (n = 16)
Lipase C14 0.2 0.1 0.03 0 0 0
β-Glucuronidase 0 2.1 0 0.1 3.0 1.1
Enzymatic activities are given as the mean colorimetric score (ranging from 0 to 5). A positive lecithinase activity was noted in 12 of 19 Clostridium spp and in none of the other isolates.
activity; β-glucuronidase activity was mainly associated with Streptococcus spp (mean score 2.1), Clostridium spp (mean score 3.0) and Bacteroides spp (mean score 1.1). Bacterial β-glucuronidase activity was observed irrespective of the presence (7 of 10; 3 stents removed at 2 months) or absence of calcium bilirubinate (14 of 16; 3 stents removed at 2 months) in the stent deposits. Lecithinase activity was noted in 12 of 19 Clostridium spp isolates and in none of the other isolates. DISCUSSION The main component of early stent deposits in this model was mucoproteins. A variety of bacteria were also cultured from the stent deposits, including anaerobic strains rarely described in humans. To our knowledge, few experimental models of stent clogging have been described.7,12,13 Preliminary results obtained with a cat model suggested that ciprofloxacin might prevent stent obstruction, but the composition of the stent deposits was not given.7 We used a pig model, because these large animals are more suitable for the study of the 10F polyethylene stents commonly used in the clinical setting. Pigs are also an established model for the study of bile salts, phospholipid metabolism and biliary protein secretion.14-16 Physical signs suggesting stent clogging or sudden death were rare with the 10F stents after 2 months of follow-up, in keeping with clinical results. Ten French stents are clearly superior to 7F or 8F stents in clinical practice,1 although median patency rarely exceeds 4 to 6 months.17 Internal diameter is a major determinant of patency and is more important than stent design itself.18,19 In the clinical setting, calcium bilirubinate entrapped in mucoproteins is the main component of stent deposits, but in humans, stents were usually studied at the time of obstruction, much later than in this study. In pigs early stent clogging was related to mucoproteins. In a placebo-controlled study, Smit et al.6 investigated the preventive efficacy of aspirin and doxycycline. Stents were rou16
GASTROINTESTINAL ENDOSCOPY
tinely removed at 2 months, the shortest time for the study of stent deposits in humans. The composition of the stent deposits was similar to that in our study, with a protein content of 48% to 57% and numerous microorganisms (mainly Enterococcus spp and E coli). No food fibers were detected. Moreover, only the distal part of the stent was studied, the proximal part being kept for an extensive bacteriologic analysis; this may have explained why in 5 cases no deposits were found despite clinical criteria suggesting stent clogging, although intermittent obstruction by food could not be ruled out. A variety of bacteria were cultured from the stent deposits, including anaerobic strains rarely described in humans.6,9,20,21 To our knowledge the presence of anaerobic strains in human stent deposits has rarely been studied.20,22 In this study stents were extracted surgically under sterile conditions, allowing complete bacteriologic study. This was not the case in human studies in which stents were removed endoscopically. The maximal half-day delay in autopsy did not influence the bacteriologic results because the number of isolates and anaerobes were comparable between stents studied at death and at autopsy. The abundance of bacteria in our study was probably related to the fact that the stents were inserted in the common bile duct so as to traverse the sphincter of Oddi, thereby allowing free communication with the duodenal contents,23 a condition that corresponds to the clinical setting when stents are inserted endoscopically. In a hypothetical model of gallstone formation, accumulation of mucin gel entraps cholesterol monohydrate crystals, forming biliary sludge and macroscopic cholesterol gallstones.24,25 In the absence of cholesterol hypersecretion, it is often calcium bilirubinate that precipitates out and leads to pigmented stones. In our pig model, calcium bilirubinate and cholesterol crystals were also trapped by the abundant mucoproteins. In the clinical setting, where stent clogging occurs within 3 to 6 months, occluded stents contain a heterogeneous, amorphous, orange-brown or blackish-brown deposit,1,3 VOLUME 51, NO. 1, 2000
Polyethylene stent blockage: a porcine model
the composition of which resembles that of pigmented stones.22 This was the case for 2 animals with 10F stents killed at 2 months. In our study anaerobic strains showed the highest β-glucuronidase activity, whereas E coli had low enzyme activity. We used a similar β-glucuronidase method as in previous human studies in which β-glucuronidase activity has been linked to E coli1,3 and C perfringens.20 These activities may theoretically lead to an excess of pigment deposit, which was the main cause of late clogging of the largest (10F) stents in our model. However, because these activities were also noted in bacteria cultured from stents without calcium bilirubinate deposit, it is reasonable to suggest that enzymatic activity of pig bile played a determinant role, as for pigment stone formation in humans.26 Lecithinase and lipase activities presented in Table 4 are those mainly associated with other pigment deposits such as palmitate; however, palmitate deposits were rare in pig stents. Infrared spectroscopy cannot characterize more precisely the nature of the mucoprotein deposition. These mucoproteins could be of bacterial origin or, because of the large amount responsible for early stent clogging, secreted by the gallbladder or the biliary tree epithelium. It has recently been suggested that mucin hypersecretion is caused by both inflammation27 and bacteria.28 The stent itself may also promote mucoprotein hypersecretion by inducing severe inflammatory changes with fibrosis and ulcerative lesions.29 Our results appear to explain why systemic antibiotics such as doxycycline and amoxicillin failed to have any preventive effect in randomized trials.5,6 The conditions for successful antibiotic prophylaxis of stent blocking would appear to be an extremely broad spectrum of activity and high biliary concentrations, but no such compounds are currently available. Oral quinolones4 or nonabsorbable antibiotics might theoretically delay the onset of symptoms of stent clogging, although they are ineffective on anaerobic strains. In conclusion, although this model is not entirely satisfactory, we suggest that this pig model might be helpful for the understanding of mechanisms related to early polyethylene stent clogging. Antibiotics are theoretically not helpful due to the abundance of bacteria identified within the first weeks following stent insertion. We suggest that the presence of anaerobic strains has been largely underestimated in humans. Future studies should concentrate on the origin of mucoprotein hypersecretion. Concerning brown pigment stone formation, the main drawback of this animal model is that follow-up is limited to 3 months because of the rapid growth and prohibitive weight of these animals. VOLUME 51, NO. 1, 2000
N Maillot, P Aucher, S Robert, et al.
ACKNOWLEDGMENT We thank Pierre Lortholary, Catherine Henry and William Hebrard for technical assistance, Professor Denis Sautereau for reviewing the manuscript and David Young for editorial assistance. REFERENCES 1. Sung JJY, Chung SCS. Endoscopic stenting for palliation of malignant biliary obstruction: a review of progress in the last 15 years. Dig Dis Sci 1995;40:1167-73. 2. Ballinger AB, McHugh M, Catnach EM, Alstead EM, Clark ML. Symptom relief and quality of life after stenting for malignant bile duct obstruction. Gut 1994;35:467-70. 3. Speer AG, Cotton PB, Rode J, Seddon AM, Neal CR, Holton J, et al. Biliary stent blockage with bacterial biofilm: a light and electron microscopy study. Ann Intern Med 1988;108:546-53. 4. Barrioz T, Ingrand P, Besson I, De Lédinghen V, Silvain C, Beauchant M. Randomised trial of prevention of biliary stent occlusion by ursideoxycholic acid plus norfloxacin. Lancet 1994;344:581-2. 5. Ghosh S, Palmer KR. Prevention of biliary stent occlusion using cyclical antibiotics and ursodeoxycholic acid. Gut 1994;35:1757-9. 6. Smit JM, Out MMJ, Groen AK, Huibregtse K, Jansen PLM, Van Marle J, et al. A placebo-controlled study on the efficacy of aspirin and doxycycline in preventing clogging of biliary endoprostheses. Gastrointest Endosc 1989;35:485-9. 7. Libby ED, Morck D, McKay S, Lam K, Olson M, Leung JWC. Ciprofloxacin prevents stent blockage in an animal model [abstract]. Gastroenterology 1994;106:A346. 8. Lacoste L, Karayan J, Bouquet S, Rouffineau J, Péchier JM, Carritez JC, et al. Pharmacokinetics of lidocaine with epinephrine in piglets following epidural anaesthesia. Lab Anim 1996;30:228-33. 9. Moesch C, Sautereau D, Cessot F, Berry P, Mounier M, Gainant A, et al. Physicochemical and bacteriological analysis of the contents of occluded biliary endoprostheses. Hepatology 1991;14:1142-6. 10. Moesch C, Sautereau D, Berry P, Cessot F, Gainant A, Raby C, et al. Library of reference infrared spectra for the analysis of gallstones [abstract]. Gastroenterology 1991;100:A776. 11. Berthelot M, Cornu G, Daudon M, Helbert M, Laurence C. Computer-aided infrared analysis of urinary calculi. Clin Chem 1987;33:2070-3. 12. Geoghegan JG, Branch MS, Costerton JW, Pappas TN, Cotton PB. Biliary stent occludes earlier if the distal tip is in the duodenum in dogs. Gastroenterol Endosc 1991;37:A257. 13. Sung JY, Leung JWC, Shaffer EA, Lam K, Costerton JW. Ascending infection of the biliary tract after surgical sphincterotomy and biliary stenting. J Gastroenterol Hepatol 1992;7:240-5. 14. Alvaro D, Cantafora A, Attili AF, Ginanni Corradini S, De Luca C, Minervini G, et al. Relationships between bile salts hydrophobicity and phospholipid composition in bile of various animal species. Comp Biochem Physiol 1986;83B:551-4. 15. Legrand-Defretin V, Juste C, Corring T, Rerat A. Enterohepatic circulation of bile acids in pigs: diurnal pattern and effect of a reentrant biliary fistula. Am J Physiol 1986;250:G295-G301. 16. Toth JL, Harvey PRC, Upadyha GA, Strasberg SM. Albumin absorption and protein secretion by the gallbladder in man and in pig. Hepatology 1990;12:729-37. 17. Davids PHP, Groen AK, Rauws EA, Tytgat GNJ, Huibregtse K. Randomised trial of self-expanding metal stents versus polyethylene stents for distal malignant biliary obstruction. Lancet 1992;340:1488-92. GASTROINTESTINAL ENDOSCOPY
17
N Maillot, P Aucher, S Robert, et al.
18. Coene PPLO, Groen AK, Out MMJ, Tytgat GNJ, Huibregtse K. Clogging of biliary endoprostheses: a new perspective. Gut 1990;31:913-7. 19. Sung JJY, Chung S, Tsui CP, Co AL, Li AKC. Omitting sideholes in biliary stents does not improve drainage of the obstructed biliary system: a prospective randomized trial. Gastrointest Endosc 1994;40:321-5. 20. Dowidar N, Kolmos HJ, Lyon H, Matzen P. Clogging of biliary endoprostheses: a morphologic and bacteriologic study. Scand J Gastroenterol 1991;26:1137-44. 21. Molinari G, Pugliese V, Schito GC, Guzman CA. Bacteria involved in the blockage of biliary stents and their susceptibility to antibacterial agents. Eur J Microb Infect Dis 1996;15:88-92. 22. Leung JWC, Ling TKW, Kung JLS, Vallance-Owen J. The role of bacteria in the blockage of biliary stents. Gastrointest Endosc 1988;34:19-22. 23. Cetta FM. Bile infection documented as initial event in the pathogenesis of brown pigment biliary stones. Hepatology 1986;6:482-9.
Polyethylene stent blockage: a porcine model
24. Afdhal N, Niu N, Gantz D, Small DM, Smith BF. Bovine gallbladder mucin accelerates cholesterol monohydrate crystal growth in model bile. Gastroenterology 1993;106:1515-23. 25. Carey MC, Cahalane MJ. Whiter biliary sludge? Gastroenterology 1988;95:508-23. 26. Ho KJ, Hsu SC, Chen JS, Ho LHC. Human biliary β-glucuronidase: correlation of its activity with deconjugation of bilirubin in the bile. Eur J Clin Invest 1986;16:361-7. 27. Sasaki M, Nakanuma Y, Kim YS. Expression of apomucins in the intrahepatic biliary tree in hepatolithiasis differs from that in nomal liver and extrahepatic biliary obstruction. Hepatology 1998;27:46-53. 28. Choi JW, Klinkspoor JH, Lee SP. Lipopolysaccharide from E Coli stimulates mucin secretion by cultured dog gallbladder epithelial cells [abstract]. Gastroenterology 1997;112:A1242. 29. Karsten TM, Coene PPL, van Gulik TM, Bosma A, van Marle J, James J, et al. Morphologic changes of extrahepatic bile ducts during obstruction and subsequent decompression by endoprosthesis. Surgery 1992;111:562-8.
Submission of FINAL manuscript on diskette Gastointestinal Endoscopy strongly encourages the submission of final manuscripts on disk. Although files created with WordPerfect are preferred, please send your final manuscript in any electronic format. On your disk, please indicate computer system (e.g., IBM, MacIntosh) and word processing software used (e.g., WordPerfect 6.1).
18
GASTROINTESTINAL ENDOSCOPY
VOLUME 51, NO. 1, 2000
Background: Endoscopic insertion of biliary stents is a useful treatment for obstructive jaundice resulting from unresectable tumors of the pancreas and biliary tree. The main drawback is the recurrence of jaundice due to clogging. The aim of this study was to establish an experimental model of polyethylene stent clogging in large white pigs. Methods: A straight polyethylene stent of 5F (group I), 7F (group II) or 10F size (group III) was inserted in the common bile duct. Animals were killed at 2 months, or earlier if physical signs suggesting stent clogging occurred. Chemicophysical analysis of stent deposition combined stereomicroscopy and identification of the contents by means of Fourrier transform infrared spectroscopy. Bacteriologic analyses included identification of aerobic and anaerobic bacteria and measurement of β-glucuronidase, lecithinase and lipase activities. Results: Physical signs suggesting stent obstruction or death occurred in 8 of 8 animals in group I, 11 of 12 in group II, and 2 of 8 in group III (p < 0.001). The proportion of mucoprotein in the stent contents tended to fall with increasing stent diameter (mean 82%, 58% and 47% for 5F, 7F and 10F, respectively), whereas wheat starch and calcium bilirubinate content increased with increasing stent diameter (9% and 4%, 18% and 10%, and 29% and 23% for 5F, 7 F, and 10F, respectively), although none of these differences were statistically significant. A variety of bacteria were cultured from the stent deposits, including anaerobic strains. Clostridium species were associated with the highest enzyme activities. Conclusions: In this model the major component of early stent deposits was mucoprotein, and numerous aerobic and anaerobic bacteria were isolated. Formation of calcium bilirubinate was a late phenomenon and poorly related to bacterial enzymatic activities. (Gastrointest Endosc 2000;51:12-8.)
Endoscopic insertion of biliary stents is a useful and well-tolerated treatment for obstructive jaundice resulting from unresectable tumors of the pancreas and biliary tree.1 Stent insertion relieves jaundice and pruritus and improves quality of life.2 The main drawback is the recurrence of jaundice or cholangitis due to clogging. Self-expanding metal stents offer durable patency but are expensive. Clogging of polyethylene stents is due to sludge adhesion. The main sludge components are proteins, fibers and bacteria.1 Bacterial adherence to the stent wall may be the initial event in stent blockage,3 but Received September 25, 1998. For revision February 4, 1999. Accepted June 23, 1999. From the Laboratory of Multivisceral Transplantation, INRA Le Magneraud, Department of Surgery, Laboratory of Microbiology and Liver Unit, University Hospital, Poitiers, and Department of Hygiene, Limoges University, Limoges, France. Presented in part at the Digestive Disease Week, San Francisco, California, May 1996. Supported by grants from Laboratoires Houdé France, Wilson Cook Medical France and Ligue Poitou-Charentes contre le Cancer. Reprint requests: Michel Beauchant, MD, Service d’Hépato-gastroentérologie, BP577, 86021 Poitiers, France. Copyright © 2000 by the American Society for Gastrointestinal Endoscopy 0016-5107/2000/$12.00 + 0 37/1/100920 12
GASTROINTESTINAL ENDOSCOPY
antibiotic prophylaxis has given variable results.4-6 To our knowledge, an experimental model of stent clogging has never been extensively described, although such a model might prove useful for the study of mechanisms leading to stent clogging and subsequently the development of new drug treatments7 or stent designs aimed at preventing or retarding blockage. The aim of this study was to establish an experimental pig model of polyethylene stent clogging. METHODS Animals We used 48 healthy large white pigs of both genders, aged 10 to 12 weeks and weighing 20 to 35 kg. They were kept in a conventional closed housing system at 22° ± 2°C and given a piglet diet (INRA 2213; Ets Arrive, St Fulgent, France) and tap water ad libitum. The study was approved by our institutional review committee. The animals were cared for in accordance with university and national guidelines (French Ministry of Agriculture license number for experimental studies: no. 04355) and National Institutes of Health guidelines. General anesthesia was administered as previously described.8 Briefly, food was withheld overnight, and preoperative tranquilization was obtained by rapid atraumatic nasal administration of 0.2 mg/kg midazolam. Anesthesia was induced with halothane and 100% oxygen. A 20-gauge VOLUME 51, NO. 1, 2000
Polyethylene stent blockage: a porcine model
plastic catheter (Insyte W; Becton Dickinson Vascular Access Inc, Sandy, Utah) was inserted into an ear vein. Atropine sulfate 10 µg/kg was given intravenously to reduce pharyngeal and tracheal secretion and prevent postintubation bradycardia. Neuromuscular blockade was induced by injection of suxamethonium iodide, 1 mg/kg. An orotracheal tube (Portex 6-6.5; Portex SA, Berck sur mer, France) was immediately inserted. Epidural anesthesia (lidocaine 1% and 1:200,000 epinephrine) was followed by discontinuation of halothane and maintenance of general anesthesia (1.5% isoflurane in 50% nitrous oxide–50% oxygen). Epidural anesthesia was used to avoid the need for morphine analgesia during surgery. Lansoprazole (30 mg orally, once daily for 5 days; Houdé Laboratories, Paris La Défense, France) was used to prevent bleeding from peptic stress ulcers. After laparotomy, transversal choledochotomy was performed and a straight polyethylene stent with side flaps (Cotton-Leung biliary stent; Wilson-Cook Medical, Winston-Salem, N.C.) was inserted. Animals were assigned to receive a 5F (group I), 7F (group II) or 10F stent (group III). The stent was attached to the posterior wall of the common bile duct (one external suture with Prolene 4.0) to avoid a high rate of spontaneous migration and emerged through the major papilla. The bile duct was ligated to the stent. In the control group, gallbladder bile was collected transhepatically after laparotomy and immediately stored at –80°C for bacteriologic analysis. The animals were monitored daily. They received orally 3157 kcal per day and wheat starch was the main component of their feeding (40%). Physical signs that suggested stent clogging were as follows: altered general status with weight loss (instead of an expected weight gain of 5 kg/week) and anorexia, jaundice, and/or clinical sepsis. Animals were killed within 12 hours of onset of these criteria; the remaining animals were killed at 2 months. In case of sudden death, autopsy was performed within 12 hours. Autopsy focused on macroscopic evidence of bile stasis proximal to the stent, i.e., hepatomegaly with a dark-green appearance or a dilated common bile duct proximal to the stent. The stent was extracted surgically under aseptic conditions and immediately cut into 2 equal parts. The upper (proximal) part was stored at –80°C for bacteriologic analysis, and the lower (distal) part was stored at room temperature for chemicophysical analysis of its desiccated contents. Chemicophysical analysis Chemicophysical analysis of the distal part of the prosthesis was carried out as previously reported.9 The surface and longitudinal sections were examined with a stereomicroscope (Wild M3Z, magnification × 6.5 to 80; Wild Heerbrugg Ltd., Heerbrugg, Switzerland). Because stent deposits were not homogeneously distributed on the luminal wall of the stent, the site with the thickest layer of deposits was localized and quantification was done with an eyepiece micrometer and expressed as percentage luminal obstruction. Samples of the deposits with different structural aspects were carbon coated and examined by means of scanning electron microscopy and energydispersive x-ray analysis (EDAX) (Jeol JFM 35 apparatus VOLUME 51, NO. 1, 2000
N Maillot, P Aucher, S Robert, et al.
coupled to an EDAX 9100/60 x-ray analyzer; Jeol Ltd., Tokyo, Japan). Fourier transform infrared spectroscopy of the desiccated deposits was carried out with a Bruker IF 28 CS spectrophotometer (Bruker Analytische Messtechnik GmbH, Rheinstetten, Germany). The samples were prepared by using the KBr pellet technique (3 mm in diameter). Compounds were identified and their proportions were determined by comparison with the reference spectra of pure compounds and their mixtures. Reference spectra of calcium phosphates (amorphous carbonated calcium phosphate, carbonate apatite and brushite), calcium carbonates (calcite, aragonite and vaterite), cholesterol monohydrate and mucoproteins were taken from the Bruker FTIR library for the analysis of urinary stones (Birsy search software).10 The other reference spectra (calcium salts of fatty acids, bile acids and bilirubin) were obtained with compounds from a commercial supplier (Fluka Chemie AG, Buchs, Switzerland) or were synthesized in our laboratory.9,11 Polarizing light microscopy (Nikon Labophot 2-Pol equipped with a photomicrographic HFX-DX system; Nikon Corporation, Tokyo, Japan) was used to detect birefringent elements (mainly wheat starch granules). Other microchemical analyses9 were done to detect phosphate anions in mixtures primarily containing calcium bilirubinate or to discriminate between protein and traces of blood. Bacteriologic analysis One milliliter of sterile saline was injected through the lumen of the proximal part of the stent and gently vortexed. The crude extract was diluted and spread on the following agar plates: meat-yeast agar (VL Sanofi Pasteur, Gentilly, France) enriched with sheep blood (5%) and meat-yeast agar with blood and gentamicin (40 µg/mL). After 72 hours of incubation in aerobic and anaerobic conditions, viable bacteria were counted and identified using ATB strips (bioMérieux, Lyon, France). Bacterial lecithinase and lipase activities were studied by inoculating egg yolk agar plates (EYA, Oxoid, France). β-Glucuronidase and several esterase activities were measured at pH 7 using the API Zym system (bioMérieux). A score of 0 to 5 was assigned by comparison with a color chart. Bile collected from the control group was analyzed in the same way. Statistical analysis The prevalence of stent clogging was analyzed according to stent size and the presence of major bacterial species by means of the log rank test. Relative amounts of mucoproteins, wheat starch and calcium bilirubinate were compared within the 3 groups by using Spearman’s correlation test. Relationship between physical signs of obstruction or death and bacterial isolates was assessed using the Fisher exact test. The level of significance was set at p < 0.05 in all tests (SAS stat 6.10; SAS Institute Inc., Cary, N.C.).
RESULTS Twelve of the 48 pigs were not included in the final analysis because of spontaneous migration of GASTROINTESTINAL ENDOSCOPY
13
N Maillot, P Aucher, S Robert, et al.
A
Polyethylene stent blockage: a porcine model
B
C
Figure 1. Characteristics of deposits observed in clogged stents by scanning electron micrography. A, Wheat starch and some cellulose fibers were the main compounds in a 7F stent deposit (orig. mag. ×1000); B, deposits located on the inner surface of a 10F stent: cholesterol crystals entrapped by mucoproteins (orig. mag. ×1000) and (C) at higher magnification, fluffy appearance of the surface of the deposit, corresponding to bacteria entrapped in mucoproteins (orig. mag. ×10,000).
Table 1. Clinical outcome of 28 pigs after surgical stent insertion No. of animals
Group I (n = 8)
Presence of physical signs suggesting stent obstruction Weight loss Jaundice Sepsis Sudden death Median time to death (days, range) No. of animals killed at 2 mo (up to 70 days) without physical signs of obstruction
8 8 5 1 3 7 (4 to 16) 0
Group II (n = 12)
Group III (n = 8)
11 9* 4 2 4 11 (5-28) 1
2 2 2 0 2 63† (14 to 70) 6
*Weight loss was the only sign suggesting clogging in 2 of these animals. of stent patency was significantly longer in group III than in group I plus II (p < 0.001, log rank test).
†Duration
the stent (n = 8) or because an autopsy could not be carried out immediately (n = 4). Twenty-eight of the remaining animals received a biliary stent: 8 in group I (5F stent), 12 in group II (7F stent) and 8 in group III (10F stent). The control group comprised 8 animals without stents and was used for bacteriologic analysis of gallbladder bile. Outcome Clinical outcome is shown in Table 1. Among the physical signs suggesting clogging, weight loss alone was noted in only 2 animals in group II. In 5 of them, the stent was not clogged and there was no evidence of bile stasis at laparotomy. The duration of stent patency was significantly longer in group III than in groups I and II combined (p < 0.001). The remaining 7 animals (1 in group II and 6 in group III) were killed at 2 months, and macroscopic evidence of bile stasis was noted in 4. 14
GASTROINTESTINAL ENDOSCOPY
Chemicophysical analysis The median percentage obstruction of the stent (distal part, desiccated contents) was 10% in the 3 groups. No deposits were seen in 8 stents: 1 group I stent, 6 group II stents and 1 group III stent. At autopsy or death, 5 had no evidence of stent obstruction despite the presence of clinical criteria of clinical stent clogging: 3 of them had clinical jaundice, increased bilirubin level and common bile duct dilatation and 2 of them weight loss only. Chemical analysis and scanning electron microscopy revealed that the major components were mucoproteins, wheat starch and calcium bilirubinate (Table 2; Fig. 1). Food fibers were rare and were only detected by scanning electron microscopy (Fig. 1A). Mucoprotein deposition tended to fall with increasing stent diameter (mean 82% in group I, 58% in group II and 47% in group III), whereas wheat starch deposition increased with increasing stent diameter VOLUME 51, NO. 1, 2000
Polyethylene stent blockage: a porcine model
N Maillot, P Aucher, S Robert, et al.
Table 2. Chemicophysical analysis of desiccated deposits by means of Fourier transform infrared spectroscopy Group I (n = 8) Stents without deposits Composition of stent deposits (mean % dry weight)* Mucoprotein Wheat starch Calcium bilirubinate Calcite Apatite Amorphous carbonated calcium phosphate Calcium palmitate Cholesterol Sodium ursodeoxycholate
Group II (n = 12)
Group III (n = 8)
1
6
1
82 9 4 2 0 4 0 0 0
58 18 10 2 4 2 2 0 5
47 29 23 0 0 0 0 1 0
*No significant differences were observed among the 3 groups; food fibers (cellulose) were rare and were only detected by scanning electron microscopy.
Table 3. Organisms cultured from surgically extracted stents (all were infected) Group I (n = 8) Aero-anaerobic bacteria Gram negative rod (n = 22) E coli Other Enterobacteria Gram positive cocci (n = 53) Staphylococcus spp Streptococcus spp Enterococcus spp Anaerobic bacteria Gram positive rod (n = 21) C perfringens Other gram positive rod Gram negative rod (n = 21) Bacteroides spp Other gram negative rod Gram positive cocci (n = 1) Peptostreptococcus spp Yeast (n = 1) Candida albicans
Group III (n = 8)
5 1
8 4
2 2
8 8 2
2 9 10
0 6 8
6 0
9 2
2 2
3 1
9 2
4 1
1
0
0
0
1
0
(9% in group I, 18% in group II and 29% in group III), as did calcium bilirubinate deposition (4% in group I, 10% in group II and 23% in group III), although none of these differences were statistically significant. The highest levels of calcium bilirubinate were observed in 2 group III stents (60% and 70% of stent deposits). Bacteriologic findings All of the gallbladder bile samples from the control group were sterile. Bacteriologic analysis could not be done on 2 group II stents because of technical problems. The remaining 26 stents yielded 117 bacterial isolates (mean 4, range 1 to 11 isolates per stent; mean density 1 × 105 organisms/mL; Fig. 1C) and 1 yeast (Table 3). Among the 75 aerobes and facultative anaerobes, the most frequent species were VOLUME 51, NO. 1, 2000
Group II (n = 10)
Escherichia coli (15 stents, 13%), Staphylococcus spp (10 stents, 9%), Streptococcus spp (23 stents, 20%) and Enterococcus spp (20 stents, 17%). Among the 42 strict anaerobes, Clostridium perfringens was identified in 17 stents (15%) and Bacteroides spp in 16 (14%). Physical signs suggesting obstruction or death were significantly associated with the presence of E coli (p = 0.014) and strict anaerobes (p = 0.028). Both types of organism were identified in all the animals that died from stent clogging. Results did not differ in terms of number of bacterial isolates and presence of strict anaerobes whether analysis was performed at autopsy (median 6 isolates, range 1 to 11, and presence of anaerobes in 8 of 9 stents) or at death (median 3 isolates, range 1 to 7, and presence of anaerobes in 11 of 17 stents). Enzyme activities are shown in Table 4. E coli had low enzyme GASTROINTESTINAL ENDOSCOPY
15
N Maillot, P Aucher, S Robert, et al.
Polyethylene stent blockage: a porcine model
Table 4. Enzymatic activities of the bacterial isolates Esterase C4
Esterase lipase C8
2.7 2.3 2.8 0.3 2.8 0.6
2.4 1.9 3.4 0 1.8 1.1
Staphylococcus spp (n = 10) Streptococcus spp (n = 23) Enterococcus spp (n = 20) E coli (n = 17) Clostridium spp (n = 19) Bacteroides spp (n = 16)
Lipase C14 0.2 0.1 0.03 0 0 0
β-Glucuronidase 0 2.1 0 0.1 3.0 1.1
Enzymatic activities are given as the mean colorimetric score (ranging from 0 to 5). A positive lecithinase activity was noted in 12 of 19 Clostridium spp and in none of the other isolates.
activity; β-glucuronidase activity was mainly associated with Streptococcus spp (mean score 2.1), Clostridium spp (mean score 3.0) and Bacteroides spp (mean score 1.1). Bacterial β-glucuronidase activity was observed irrespective of the presence (7 of 10; 3 stents removed at 2 months) or absence of calcium bilirubinate (14 of 16; 3 stents removed at 2 months) in the stent deposits. Lecithinase activity was noted in 12 of 19 Clostridium spp isolates and in none of the other isolates. DISCUSSION The main component of early stent deposits in this model was mucoproteins. A variety of bacteria were also cultured from the stent deposits, including anaerobic strains rarely described in humans. To our knowledge, few experimental models of stent clogging have been described.7,12,13 Preliminary results obtained with a cat model suggested that ciprofloxacin might prevent stent obstruction, but the composition of the stent deposits was not given.7 We used a pig model, because these large animals are more suitable for the study of the 10F polyethylene stents commonly used in the clinical setting. Pigs are also an established model for the study of bile salts, phospholipid metabolism and biliary protein secretion.14-16 Physical signs suggesting stent clogging or sudden death were rare with the 10F stents after 2 months of follow-up, in keeping with clinical results. Ten French stents are clearly superior to 7F or 8F stents in clinical practice,1 although median patency rarely exceeds 4 to 6 months.17 Internal diameter is a major determinant of patency and is more important than stent design itself.18,19 In the clinical setting, calcium bilirubinate entrapped in mucoproteins is the main component of stent deposits, but in humans, stents were usually studied at the time of obstruction, much later than in this study. In pigs early stent clogging was related to mucoproteins. In a placebo-controlled study, Smit et al.6 investigated the preventive efficacy of aspirin and doxycycline. Stents were rou16
GASTROINTESTINAL ENDOSCOPY
tinely removed at 2 months, the shortest time for the study of stent deposits in humans. The composition of the stent deposits was similar to that in our study, with a protein content of 48% to 57% and numerous microorganisms (mainly Enterococcus spp and E coli). No food fibers were detected. Moreover, only the distal part of the stent was studied, the proximal part being kept for an extensive bacteriologic analysis; this may have explained why in 5 cases no deposits were found despite clinical criteria suggesting stent clogging, although intermittent obstruction by food could not be ruled out. A variety of bacteria were cultured from the stent deposits, including anaerobic strains rarely described in humans.6,9,20,21 To our knowledge the presence of anaerobic strains in human stent deposits has rarely been studied.20,22 In this study stents were extracted surgically under sterile conditions, allowing complete bacteriologic study. This was not the case in human studies in which stents were removed endoscopically. The maximal half-day delay in autopsy did not influence the bacteriologic results because the number of isolates and anaerobes were comparable between stents studied at death and at autopsy. The abundance of bacteria in our study was probably related to the fact that the stents were inserted in the common bile duct so as to traverse the sphincter of Oddi, thereby allowing free communication with the duodenal contents,23 a condition that corresponds to the clinical setting when stents are inserted endoscopically. In a hypothetical model of gallstone formation, accumulation of mucin gel entraps cholesterol monohydrate crystals, forming biliary sludge and macroscopic cholesterol gallstones.24,25 In the absence of cholesterol hypersecretion, it is often calcium bilirubinate that precipitates out and leads to pigmented stones. In our pig model, calcium bilirubinate and cholesterol crystals were also trapped by the abundant mucoproteins. In the clinical setting, where stent clogging occurs within 3 to 6 months, occluded stents contain a heterogeneous, amorphous, orange-brown or blackish-brown deposit,1,3 VOLUME 51, NO. 1, 2000
Polyethylene stent blockage: a porcine model
the composition of which resembles that of pigmented stones.22 This was the case for 2 animals with 10F stents killed at 2 months. In our study anaerobic strains showed the highest β-glucuronidase activity, whereas E coli had low enzyme activity. We used a similar β-glucuronidase method as in previous human studies in which β-glucuronidase activity has been linked to E coli1,3 and C perfringens.20 These activities may theoretically lead to an excess of pigment deposit, which was the main cause of late clogging of the largest (10F) stents in our model. However, because these activities were also noted in bacteria cultured from stents without calcium bilirubinate deposit, it is reasonable to suggest that enzymatic activity of pig bile played a determinant role, as for pigment stone formation in humans.26 Lecithinase and lipase activities presented in Table 4 are those mainly associated with other pigment deposits such as palmitate; however, palmitate deposits were rare in pig stents. Infrared spectroscopy cannot characterize more precisely the nature of the mucoprotein deposition. These mucoproteins could be of bacterial origin or, because of the large amount responsible for early stent clogging, secreted by the gallbladder or the biliary tree epithelium. It has recently been suggested that mucin hypersecretion is caused by both inflammation27 and bacteria.28 The stent itself may also promote mucoprotein hypersecretion by inducing severe inflammatory changes with fibrosis and ulcerative lesions.29 Our results appear to explain why systemic antibiotics such as doxycycline and amoxicillin failed to have any preventive effect in randomized trials.5,6 The conditions for successful antibiotic prophylaxis of stent blocking would appear to be an extremely broad spectrum of activity and high biliary concentrations, but no such compounds are currently available. Oral quinolones4 or nonabsorbable antibiotics might theoretically delay the onset of symptoms of stent clogging, although they are ineffective on anaerobic strains. In conclusion, although this model is not entirely satisfactory, we suggest that this pig model might be helpful for the understanding of mechanisms related to early polyethylene stent clogging. Antibiotics are theoretically not helpful due to the abundance of bacteria identified within the first weeks following stent insertion. We suggest that the presence of anaerobic strains has been largely underestimated in humans. Future studies should concentrate on the origin of mucoprotein hypersecretion. Concerning brown pigment stone formation, the main drawback of this animal model is that follow-up is limited to 3 months because of the rapid growth and prohibitive weight of these animals. VOLUME 51, NO. 1, 2000
N Maillot, P Aucher, S Robert, et al.
ACKNOWLEDGMENT We thank Pierre Lortholary, Catherine Henry and William Hebrard for technical assistance, Professor Denis Sautereau for reviewing the manuscript and David Young for editorial assistance. REFERENCES 1. Sung JJY, Chung SCS. Endoscopic stenting for palliation of malignant biliary obstruction: a review of progress in the last 15 years. Dig Dis Sci 1995;40:1167-73. 2. Ballinger AB, McHugh M, Catnach EM, Alstead EM, Clark ML. Symptom relief and quality of life after stenting for malignant bile duct obstruction. Gut 1994;35:467-70. 3. Speer AG, Cotton PB, Rode J, Seddon AM, Neal CR, Holton J, et al. Biliary stent blockage with bacterial biofilm: a light and electron microscopy study. Ann Intern Med 1988;108:546-53. 4. Barrioz T, Ingrand P, Besson I, De Lédinghen V, Silvain C, Beauchant M. Randomised trial of prevention of biliary stent occlusion by ursideoxycholic acid plus norfloxacin. Lancet 1994;344:581-2. 5. Ghosh S, Palmer KR. Prevention of biliary stent occlusion using cyclical antibiotics and ursodeoxycholic acid. Gut 1994;35:1757-9. 6. Smit JM, Out MMJ, Groen AK, Huibregtse K, Jansen PLM, Van Marle J, et al. A placebo-controlled study on the efficacy of aspirin and doxycycline in preventing clogging of biliary endoprostheses. Gastrointest Endosc 1989;35:485-9. 7. Libby ED, Morck D, McKay S, Lam K, Olson M, Leung JWC. Ciprofloxacin prevents stent blockage in an animal model [abstract]. Gastroenterology 1994;106:A346. 8. Lacoste L, Karayan J, Bouquet S, Rouffineau J, Péchier JM, Carritez JC, et al. Pharmacokinetics of lidocaine with epinephrine in piglets following epidural anaesthesia. Lab Anim 1996;30:228-33. 9. Moesch C, Sautereau D, Cessot F, Berry P, Mounier M, Gainant A, et al. Physicochemical and bacteriological analysis of the contents of occluded biliary endoprostheses. Hepatology 1991;14:1142-6. 10. Moesch C, Sautereau D, Berry P, Cessot F, Gainant A, Raby C, et al. Library of reference infrared spectra for the analysis of gallstones [abstract]. Gastroenterology 1991;100:A776. 11. Berthelot M, Cornu G, Daudon M, Helbert M, Laurence C. Computer-aided infrared analysis of urinary calculi. Clin Chem 1987;33:2070-3. 12. Geoghegan JG, Branch MS, Costerton JW, Pappas TN, Cotton PB. Biliary stent occludes earlier if the distal tip is in the duodenum in dogs. Gastroenterol Endosc 1991;37:A257. 13. Sung JY, Leung JWC, Shaffer EA, Lam K, Costerton JW. Ascending infection of the biliary tract after surgical sphincterotomy and biliary stenting. J Gastroenterol Hepatol 1992;7:240-5. 14. Alvaro D, Cantafora A, Attili AF, Ginanni Corradini S, De Luca C, Minervini G, et al. Relationships between bile salts hydrophobicity and phospholipid composition in bile of various animal species. Comp Biochem Physiol 1986;83B:551-4. 15. Legrand-Defretin V, Juste C, Corring T, Rerat A. Enterohepatic circulation of bile acids in pigs: diurnal pattern and effect of a reentrant biliary fistula. Am J Physiol 1986;250:G295-G301. 16. Toth JL, Harvey PRC, Upadyha GA, Strasberg SM. Albumin absorption and protein secretion by the gallbladder in man and in pig. Hepatology 1990;12:729-37. 17. Davids PHP, Groen AK, Rauws EA, Tytgat GNJ, Huibregtse K. Randomised trial of self-expanding metal stents versus polyethylene stents for distal malignant biliary obstruction. Lancet 1992;340:1488-92. GASTROINTESTINAL ENDOSCOPY
17
N Maillot, P Aucher, S Robert, et al.
18. Coene PPLO, Groen AK, Out MMJ, Tytgat GNJ, Huibregtse K. Clogging of biliary endoprostheses: a new perspective. Gut 1990;31:913-7. 19. Sung JJY, Chung S, Tsui CP, Co AL, Li AKC. Omitting sideholes in biliary stents does not improve drainage of the obstructed biliary system: a prospective randomized trial. Gastrointest Endosc 1994;40:321-5. 20. Dowidar N, Kolmos HJ, Lyon H, Matzen P. Clogging of biliary endoprostheses: a morphologic and bacteriologic study. Scand J Gastroenterol 1991;26:1137-44. 21. Molinari G, Pugliese V, Schito GC, Guzman CA. Bacteria involved in the blockage of biliary stents and their susceptibility to antibacterial agents. Eur J Microb Infect Dis 1996;15:88-92. 22. Leung JWC, Ling TKW, Kung JLS, Vallance-Owen J. The role of bacteria in the blockage of biliary stents. Gastrointest Endosc 1988;34:19-22. 23. Cetta FM. Bile infection documented as initial event in the pathogenesis of brown pigment biliary stones. Hepatology 1986;6:482-9.
Polyethylene stent blockage: a porcine model
24. Afdhal N, Niu N, Gantz D, Small DM, Smith BF. Bovine gallbladder mucin accelerates cholesterol monohydrate crystal growth in model bile. Gastroenterology 1993;106:1515-23. 25. Carey MC, Cahalane MJ. Whiter biliary sludge? Gastroenterology 1988;95:508-23. 26. Ho KJ, Hsu SC, Chen JS, Ho LHC. Human biliary β-glucuronidase: correlation of its activity with deconjugation of bilirubin in the bile. Eur J Clin Invest 1986;16:361-7. 27. Sasaki M, Nakanuma Y, Kim YS. Expression of apomucins in the intrahepatic biliary tree in hepatolithiasis differs from that in nomal liver and extrahepatic biliary obstruction. Hepatology 1998;27:46-53. 28. Choi JW, Klinkspoor JH, Lee SP. Lipopolysaccharide from E Coli stimulates mucin secretion by cultured dog gallbladder epithelial cells [abstract]. Gastroenterology 1997;112:A1242. 29. Karsten TM, Coene PPL, van Gulik TM, Bosma A, van Marle J, James J, et al. Morphologic changes of extrahepatic bile ducts during obstruction and subsequent decompression by endoprosthesis. Surgery 1992;111:562-8.
Submission of FINAL manuscript on diskette Gastointestinal Endoscopy strongly encourages the submission of final manuscripts on disk. Although files created with WordPerfect are preferred, please send your final manuscript in any electronic format. On your disk, please indicate computer system (e.g., IBM, MacIntosh) and word processing software used (e.g., WordPerfect 6.1).
18
GASTROINTESTINAL ENDOSCOPY
VOLUME 51, NO. 1, 2000