- Email: [email protected]
Palliation of Malignant Bile Duct Obstruction with Metallic Biliary Endoprostheses: Technique, Results, and Complications
Palliation of Malignant Bile Duct Obstruction with Metallic Biliary Endoprostheses: Technique, Results, and Complications1 Michael J. Lee, M D Steven L. Dawson, M D Peter R. Mueller, M D Thonten L. Krebs, M D Sanjay Saini, M D Peter F. Hahn, MD, PhD
Index terms: Bile ducts, neoplasms, 768.32, 768.33 * ducts, prostheses, 768.1229 Bile ducts, stenosis or obstruction, 768.2886
JVIR 1992; 3:665-671
From the Department of Radiology, Massachusetts General Hospital and Harvard Medical School, 32 Fruit St, Boston, MA 02114. Received January 22, 1992; revision requested March 27; revision received June 22; accepted June 24. Address reprint requests to M.J.L. SCVIR, 1992
Expandable metallic stents were placed in 34 patients with pathologically proved malignant bile duct obstruction to determine ease of insertion, benefits of a one-stageinsertion, and cost-effectiveness relative to conventional plastic stents. Thirty-eight strictures, ranging in length from 1to 7 cm (mean, 3.2 cm), were present in the 34 patients. Strictures were located in the lower common bile duct (n = 22), middle of the common bile duct (n = 6), and hilar confluence (n = 10).In 13patients (38%)metallic stents were placed at the time of initial biliary drainage (one-stageprocedure), while the remaining patients underwent stent placement within 1-7 days of biliary drainage (two-stageprocedure). Biliary obstruction was relieved in 31 of 34 patients (91%).Three patients died within 14 days of stent insertion of unrelated causes, without any change in biliary status. Mean duration of follow-upfor all patients was 5.3 months (range, 0.5-14 months). Four episodes of stent occlusion occurred in three patients (12%occlusion rate); each episode was treated successfully. The average length of hospital stay for patients who underwent a one-stage procedure was 13 days (range, 3-33 days) and was 20 days (range, 9-42 days) for patients who underwent a twostage procedure. The facility of one-step insertion, low occlusion rate, and the many strategies available for treatment of occluded stents make metallic stents an attractive alternative to conventional plastic stents in palliating patients with malignant biliary obstruction. SELF-EXPANDING metallic stents have been used in the vascular system ( I ) , tracheobronchial tree (2), and urinary tract (3).More recently, they have been used in the biliary tree (4-9). These expandable metallic stents offer a new alternative in palliating patients with malignant bile duct obstruction. However, important questions remain to be answered regarding the use of these more expensive metallic stents instead of less expensive conventional stents. Some of these questions include, How technically difficult are they to insert? Does insertion of metal stents in a one-stage procedure confer any extra benefit? However, the most critical question is, DOmetal stents confer any advantages over plastic stents that justify the routine use of metal
stents in patients with biliary obstruction? Early reports described a disappointingly high occlusion rate for metallic stents ( > 20%)(7,s).However, more recent reports have described low incidences of occlusion (43). We now report our experience with selfexpanding metallic endoprostheses in 34 patients with malignant bile duct obstruction and attempt to answer the questions posed above.
PATIENTS AND METHODS Forty-nine metallic stents were placed in 34 patients (gender ratio, 20 men to 14 women; age range, 45-87 years; mean age, 62 years) with malignant bile duct obstruction. All 34
665
666
Journal of Vascular and Interventional Radiology
November 1992
patients were jaundiced, 18 patients had pruritus, and eight were septic at the time of biliary drainage. Serum total bilirubin (normal, 0-1.0 mg/dL [O-17.1 ymol/L]) and alkaline phosphatase (normal, 45 to 115 U/L) levels were elevated in all patients prior to biliary drainage and stent placement. Serum total bilirubin levels ranged from 2.3 to 33 mg/dL (39.3564.3 kmol/L) (mean, 16.4 mg/dL [280.4 kmol/LI) and serum alkaline phosphatase levels ranged from 181 to 1,956 U/L (mean, 506 U/L). Twenty-five of 34 patients had undergone unsuccessful endoscopic drainage, while nine patients were referred from outside institutions with percutaneous biliary drainage catheters in situ. Biliary obstruction was due to pancreatic carcinoma (n = 141, metastatic disease (n = lo), cholangiocarcinoma (n = 6), and local spread from gallbladder carcinoma (n = 2), gastric carcinoma (n = I),and hepatocellular carcinoma (n = 1). Thirtyeight malignant strictures, ranging in length from 1to 7 cm (mean, 3.2 cm), were present in the 34 patients. The strictures were located in the lower common bile duct (22 of 38 strictures), in the middle of the common bile duct (six of 38), and in the region of the hilar confluence (10 of 38). Pathologic confirmation of malignancy was obtained in all patients by means of either needle biopsy (16 of 34 patients), surgery (15 patients), or percutaneous biliary brushing (three patients). The Wallstent endoprosthesis (Schneider, Plymouth, Minn) was used in all patients. This stent is composed of braided stainless steel monofilaments and is 6.8 cm in length when expanded to its full diameter of 1cm. The stent is compressed onto a 7-F delivery catheter and covered by a plastic membrane that rolls back to release the stent when the membrane is hydraulically activated. Our technique for stent insertion did not differ substantially from the techniques described by other authors (5-7). Antibiotics (1g of ampicillin and 80 mg of gentamicin) were routinely administered intravenously prior to stent placement.
All patients underwent percutaneous biliary drainage performed in standard fashion before stent placement. Internal biliary drainage was achieved with all patients at the time of the initial drainage. Four patients with hilar lesions required both left- and right-sided biliary drainage. In 13 patients (38%),metallic endoprostheses were placed in a one-stage procedure a t the time of initial biliary drainage, while in the remaining patients, stents were placed within 1-7 days (mean, 3 days) of the biliary drainage when pathologic confirmation of malignancy became available. Metal stents were placed a t the time of initial biliary drainage if a confirmed diagnosis of malignancy was available. The stent delivery catheter was introduced over a heavy-duty guide wire (0.038-inch Arnplatz extra stiff; Cook, Bloomington, Ind). An 8-F sheath was first placed over the stent delivery catheter but was not actually introduced into the liver. The sheath was introduced over the stent delivery catheter if problems arose in deploying the stent, so that the stent, if partially open, could be withdrawn through the sheath without injuring the liver. The stent was ~ositioned across the obstructing lesion so that the stent overlapped the proximal and distal ends of the stricture. Further stents were deployed as necessary to achieve this goal (Fig 1).After deployment, the stent was dilated at the site of the stricture with an 8-mm balloon catheter in patients in whom a prominent waist in the stent was seen a t the site of stricture (30 of 38 strictures). An 8-F external catheter was left in situ for a mean of 4 days (range, 1-10 days) and was removed after contrast material injection confirmed stent patency and good position of the stent vis-a-visthe stricture. Follow-UDinformation was obtained on 21 patients by examining hospital records if patients were followed up a t our institution or by contacting the patient's family or primary doctor if the patient was followed up a t another institution. Reports of further admissions to hos-
pital, redevelopment of jaundice, and levels of serum alkaline phosphatase and bilirubin, where available, were noted.
RESULTS A total of 49 stents were required for the 38 strictures. In 22 patients, a single stent was sufficient to bridge the stricture, while in 12 patients, more than one stent was necessary. Relief of obstruction with disappearance of jaundice and pruritus occurred in 3 1 of 34 patients (91%) within 2-6 weeks of stent insertion. Bilirubin and alkaline phosphatase levels were available for 30 of 34 patients due to the fact that some patients returned to other institutions for further care. These 30 patients had either normal (10 patients) or decreasing levels (20 patients) of total bilirubin (range, 0.3-28 mg/dL; mean, 5.6 mg/dL [range, 5.1-478.8 kmol/L; mean, 95.8 kmol/LI) or alkaline phosphatase (range, 75-900 U/L; mean, 274 UIL) a t the time of biochemical testing (range, 1day to 6 months after stent placement; mean, 36 days). Three patients died within 2 weeks of stent insertion.
Patient Survival The total length of follow-up for all patients averaged 5.3 months (range, 0.5-14 months). Twenty-four of 34 patients (71%)died between 3 days and 14 months (mean, 3.0 months) after stent insertion. Three of these 24 patients died within 30 days of stent insertion. One patient had a pulmonary embolus 2 days after stent insertion, while the remaining two patients had extensive metastatic disease and developed both liver and renal failure. Twenty-three of the 24 patients required no further intert on vention and had ~ a t e n stents clinical follow-uiat the time of death. Metallic stent occlusion due to tumor overgrowth occurred in one patient, requiring further percutaneous intervention. Ten of 34 patients (29%)are alive and well with functioning stents a t
Lee et a1
667
- -
Volume 3 Number 4
an average of 6.5 months after insertion (range, 3-10 months), although two of these patients required further intervention for stent occlusion. Average length of hospital stay for patients who underwent a one-stage procedure was 13 days (range, 3-33 days) and was 20 days (range, 9-42 days) for patients who underwent a two-stage procedure. Two of the 13 patients who underwent a one-stage procedure were in poor clinical condition prior to stent placement, while eight of the 21 patients who underwent a two-stage procedure were in poor clinical condition prior to stent placement.
Short-term Complications Three ~ a t i e n t had s com~lications related to percutaneous bikary drainage. Biliary sepsis occurred in two patients and was successfully treated with antibiotics. A third patient with a coagulopathy experienced a decrease in hematocrit after percutaneous biliary drainage, which required red cell transfusion, vitamin Kinjection, and fresh frozen plasma to stabilize the patient's condition. Insertion of metal stents was consequently delayed until each patient's condition was stable.
C. d. Figure 1. Images of a patient with metastatic portal adenopathy causing biliary obstruction. Two stents were required to fully bridge the stricture. (a)Transhepatic cholangiography reveals a stricture within the middle of the common bile duct without distal passage of contrast material into the duodenum. Surgical clips (arrow) mark the approximate level of the proximal end of the stricture. (b) A Wallstent was deployed across the stricture but migrated distally during insertion so that only a small portion of the stent was proximal to the stricture at the level of the surgical clips (arrow). Contrast material injection did not reveal any proximal filling of the biliary tree supporting this conclusion. (c) A second Wallstent was deployed to provide adequate overlap proximal to the stricture. An 8-mm balloon (arrows) was used to inflate the metal stents a t the area of the stricture where a waist was present. (d) After balloon dilation the stents (arrows) have opened up to a large luminal diameter. An 8-F external pigtail catheter was left in situ for 3 days and removed after contrast material injection revealed good drainage into the duodenum. The patient died 5 months after stent insertion of disease progression without evidence of stent malfunction.
Technical Difficulties Technical complications occurred early in our experience and included stent delivery in the wrong location and failure of stent deployment due to incomplete retraction of the rolling membrane. Deployment of stents in the wrong location occurred in one patient. The stent was deployed too distally across the ampulla and migrated into the duodenum and then into the stomach. A new stent was deployed in the correct location, and the stent in the stomach was removed endoscopically. Incomplete retraction of the rolling membrane occurred with five stents. These stents were removed and fresh stents inserted. Long-term Complications One episode of stent occlusion occurred in two patients with hilar
668
Journal of Vascular and Interventional Radiology
November 1992
strictures, and two episodes of occlusion occurred in one patient with a low common bile duct stricture. Thus, our total occlusion rate was 12% (four episodes of occlusion in 34 patients). Stent occlusion was due to tumor overgrowth in one patient with a hilar stricture (at 5 months after stent insertion). The patient was in terminal condition, and a n internallexternal catheter was placed through the stent for the remainder of the patient's life (1month). Stent occlusion was due to debris and inspissated bile in the remaining two patients (Fig 2). Both patients were older than 80 years of age and were admitted with cholangitis and dehydration at 1and 2 months after stent insertion. Treatment consisted of intravenous antibiotics, rehydration with intravenous fluids, and short-term percutaneous drainage of the biliary tree. Stents were cleared of debris with the use of a balloon catheter, which was inflated and passed up and down through the stent to reestablish patency (Fig 2). When the acute septic episode had abated, the drainage catheters were removed. A second episode of occlusion occurred in one of these patients 2 months later and was treated in the same manner. Both patients are presently alive and well (6 and 8 months after the first episode of occlusion) with functioning stents and without further episodes of occlusion.
DISCUSSION Conventional plastic endoprostheses for palliation of malignant biliary obstruction are associated with two main problems, stent migration and occlusion. While the problem of stent migration has largely been solved by the provision of anchoring sutures, stent occlusion remains a difficult problem. Mueller et a1 (lo), in a multicenter review, found that occlusion occurred in 23% of patients on longterm follow-up, while Mendez et a1 (11)believed that the problems with plastic biliary endoprostheses were of sufficient magnitude to warrant that
Comparison of Wallstent Endoprosthesis Occlusion Rates No. of Author Year Patients Gillams et a1 (7) 1990 45 Lammer et a1 (9) 1990 61 Lameris et a1 (4) 1991 69 Adam et a1 (5) 1991 41
their use be completely abandoned. However, Dick et al(12) in a series of 100 patients from one institution reported a lower occlusion rate of 12.3%, but the total long-term complication rate was 17.5% when the migration rate (5.2%)was considered. Against this background skepticism regarding the use of plastic stents, there has been increasing interest in the use of metallic endoprostheses. These metallic endoprostheses, when deployed across a biliary stricture, open up to provide a luminal diameter (24-30 F) significantly larger than that of plastic endoprostheses. This large luminal diameter offers the theoretical advantage of longerterm patency vis-a-vis the smaller luminal diameter of the ~ l a s t i cendoprosthesis. In addition, migration does not occur with metallic stents because they become incorporated and fixed in the wall of the common bile duct. Earlier reports, however, detailing the use of metallic biliary endoprostheses, did not justify this initial optimism. Gillams et a1 (7) reported an occlusion rate of 42% for the Wallstent endoprosthesis, while Irving et a1 (8) reported an occlusion rate approaching 50% for the Gianturco stent. However, more recent reports on the Wallstent endoprosthesis by Lammer et a1 (91, Lameris et a1 (4), and Adam et al(5) are more encouraging (Table). These authors reported stent occlusion rates of 13%, 1096, and 7%, respectively. In the series reported by Lameris et al (41, stent occlusion mainly occurred in ~ a t i e n t with s hilar strictures. Our long-term occlusion rate, a t 12%, was similar to the low occlusion rates found by previous authors (4,5,9).
Occlusion Rate (%I 42 13 10 7
Unlike Lameris et a1 (4),we did not see a large occlusion rate for patients with hilar strictures; however, the number of patients with hilar strictures in this series was small (10 patients), and statistical inferences cannot be made from this. Although metal stents cannot be removed once inserted, there are many stratagems available for treatment of metallic stent occlusion, as there are with plastic stent occlusion. The sine qua non for management of endoprosthesis occlusion is the placement of a guide wire through the center of the endoprosthesis. This can be technically difficult with occluded plastic endoprostheses because of the small luminal diameter (13). However, the large luminal diameter (24-30 F) of the metallic endoprostheses facilitates easy passage of a guide wire. Once this has been achieved, many therapeutic options are available. If stent occlusion is due to inspissated bile, the patient is treated with intravenous fluids and antibiotics and the metal stent is cleared of debris by passing a balloon catheter through the lumen of the stent (Fig 2). A temporary drainage catheter is left in situ until any associated cholangitis has improved and contrast material injection shows a patent stent. The drainage catheter is then removed. Alternatively, if the patient is judged to be in terminal condition, a long-term internal-external catheter can be left in situ. If the metal stent becomes occluded by tumor. a new metal stent can be inserted percutaneously through the occluded stent. a long-term internal-external catheter can be placed, or a plastic stent can be inserted endoscopically through the occluded metallic stent. u
Lee et a1
669
Volume 3 Number 4
Figure 2. Images from an 82-year-old man with ampullary carcinoma in whom a metallic stent was inserted. The patient returned 2 months later with jaundice and sepsis. (a)Transhepatic cholangiography shows complete occlusion of the stent by debris (arrows). (b) A guide wire was manipulated easily down through the wide central lumen of the stent and into the duodenum. (c) An 8-mm balloon catheter (arrows) was passed up and down through the stent to clear it of debris. (d) Contrast material injection after the stent is cleared shows free flow of contrast material into the duodenum. An external catheter was left in situ for 5 days while the patient was rehydrated and treated with intravenously administered antibiotics. The external catheter was then removed after repeated contrast material injection confirmed stent patency. Two months later a similar event occurred, which was treated in the same manner. The patient is currently doing well with a functioning stent (4 months after second stent occlusion).
Tumor overgrowth occurred in one of our patients early on in this series and was due to a combination of
stent shortening and poor initial stent placement because of limited biliary purchase above the proximal
end of the tumor. We elected to place a long-term internal-external catheter through the occluded stent be-
670
Journal of Vascular and Interventional Radiology
November 1992
cause of the patient's terminal condition. Stent shortening normally occurs with Wallstent metallic endoprostheses as a means of achieving maximal luminal diameter, that is, the stent shortens as it increases its luminal diameter. Therefore, stent shortening coupled with the proximity of the tumor to the proximal end of the stent can lead to tumor overgrowth. This can be compensated for by ensuring a wide margin of error when placing metallic stents so that proximal and distal ends overlap the stricture as much as possible. Stent occlusion due to debris and inspissated bile occurred in two patients, both of whom were severely dehydrated and older than 80 years. Both patients had cholangitis and jaundice at presentation. Stent occlusion was considered due to severe dehydration with bile inspissation in both of these patients. A further episode of stent occlusion occurred in one of these patients due to similar causes. Both patients underwent another drainage procedure during the acute episodes, and the stent was cleared of debris by using a balloon catheter. There was no evidence of tumor overgrowth in either of these patients, and internal-external catheters were removed after the acute episodes had subsided. Insertion of the metallic Wallstent endoprosthesis, although associated with a steeper learning curve than that associated with the plastic endoprosthesis, is in our experience, easier once familiarity with the technique is gained. The transhepatic tract does not have to be dilated to 14 or 16 F as is needed with plastic endoprostheses. Indeed, the 7-F delivery catheter makes the procedure more comfortable for the patient and facilitates metal stent insertion at the time of initial percutaneous biliary drainage, provided a diagnosis of malignant bile duct obstruction has been confirmed. Metal stents were inserted in 13 such patients (38%)at the time of percutaneous biliary drainage in this series. This subgroup of patients had a shorter hospital stay (mean, 13 days; range, 3-33
days) compared with the remaining 21 patients who had metal stents inserted 1-7 days after percutaneous biliary drainage (mean, 20 days; range, 9-42 days). While realizing that a contributing factor to the longer hospital stay of patients who underwent a two-stage procedure was the poor clinical condition of many of these patients, the possibility of metal stent insertion at the time of initial biliary drainage is a desirable advantage over plastic stents. In addition, comparison of metal stent occlusion rates from this (12%)and other recent series (10% [41, 7% [51, and 13% [91) with those reported for plastic stents (23% [lo] and 12.3% [12]) suggests that metallic stents may have lower occlusion rates than plastic stents. In addition, stent migration, which can be a problem with plastic stents, does not occur with metallic stents. The extra cost of the Wallstent metallic stent ($1,000) compared with conventional plastic stents ($60-$160) may be offset if low occlusion rates are confirmed in further studies with longterm follow-up and if metal stents are inserted in a one-step procedure to shorten hospital stay. We recognize some limitations of our study. The majority of patients (71%) in our study died within 14 months (mean, 3 months) of stent placement. This was due to a combination of the poor clinical condition of many patients prior to stent placement and to the aggressive nature of their diseases. We did not try to select patients for metal stent placement who might have had a longer life expectancy; rather, metal stents were placed in all patients with malignant disease undergoing biliary drainage during the time of this study. Another limitation of our study includes the lack of biochemical follow-up in all patients. Clinical follow-up was available in all patients through contact with the patients' doctor or family. However, because patients came from a national and international referral base, access to biochemical data was limited. The ease of insertion and facility
for one-step insertion give the metallic Wallstent endo~rosthesisdistinct advantages over piastic stents. Although only short-term follow-up is available at the present time, the modestlv low occlusion rates in this and other recent series compare favorably with those of plastic stents. Despite the permanency of metallic stents once inserted, the ease of insertion and the many stratagems available for treatment of stent occlusion make metallic stents an attractive alternative to plastic stents for palliation of malignant bile duct obstruction. References 1. Kimihiko K, Hideo U, Tetsuya Y, et al. Iliac artery stenosis and occlusion: preliminary results of treatment with Gianturco expandable metallic stents. Radiology 1990; 177: 799-802. 2. Wallace MJ, Charnsangavej C, Ogawa K, et al. Tracheobronchial tree: expandable metallic stents used in experimental and clinical applications. Radiology 1986; 158:309-312. 3. Milroy EJG, Chapple CR, El-Din A, Wallsten H. A new stent for the treatment of urethral strictures. J Uro11989; 141:1120-1122. 4. Lameris JS, Stoker J , Nijs HGT, et al. Malignant biliary obstruction: percutaneous use of self-expandable stents. Radiology 1991; 179:703707. 5. Adam A, Chetty N, Roddie M, Yeung E, Benjamin IS. Self-expandable stainless steel endoprostheses for treatment of malignant bile duct obstruction. AJR 1991; 156:321325. 6. Coons HG. Self-expanding stainless steel biliary stents. Radiology 1989; 170:979-983. 7. Gillams A, Dick R, Dooley JS, Wallsten H, El-Din A. Self-expandable stainless steel braided endoprosthesis for biliary strictures. Radiology 1990; 174:137-140. 8. Irving JD, Adam A, Dick R, Dondelinger RF, Lunderquist A, Roche A. Gianturco expandable metallic biliary stents: results of a European clinical trial. Radiology 1989; 172: 321-326. 9. Lammer J , Klein GE, Kleinert R, Hausegger K, Einspieler R. Obstructive jaundice: use of expandable
Lee et a1
671
Volume 3 Number 4
metal endoprosthesis for biliary drainage. Radiology 1990; 177:789792. 10. Mueller PR, Ferrucci J T J r , Teplick SK, et al. Biliary stent endoprosthesis: analysis of complications in 113 patients. Radiology 1985; 156: 637-639. 11. Mendez G Jr, Russell E, LePage JR, Guerra J J , Posniak RA, Trefler M.
Abandonment of endoprosthetic drainage technique in malignant biliary obstruction. AJR 1984; 143: 617-622. 12. Dick BW, Gordon RL, LaBerge JM, Doherty MM, Ring EJ. Percutaneous transhepatic placement of biliary endoprostheses: results in 100 consecutive patients. JVIR 1990; 1:97-100.
13. Lee MJ, Mueller PR, Saini S, Morrison MC, Brink JA, Hahn PF. Occlusion of biliary endoprostheses: presentation and management. Radiology 1990; 176:531-534.
Index terms: Bile ducts, neoplasms, 768.32, 768.33 * ducts, prostheses, 768.1229 Bile ducts, stenosis or obstruction, 768.2886
JVIR 1992; 3:665-671
From the Department of Radiology, Massachusetts General Hospital and Harvard Medical School, 32 Fruit St, Boston, MA 02114. Received January 22, 1992; revision requested March 27; revision received June 22; accepted June 24. Address reprint requests to M.J.L. SCVIR, 1992
Expandable metallic stents were placed in 34 patients with pathologically proved malignant bile duct obstruction to determine ease of insertion, benefits of a one-stageinsertion, and cost-effectiveness relative to conventional plastic stents. Thirty-eight strictures, ranging in length from 1to 7 cm (mean, 3.2 cm), were present in the 34 patients. Strictures were located in the lower common bile duct (n = 22), middle of the common bile duct (n = 6), and hilar confluence (n = 10).In 13patients (38%)metallic stents were placed at the time of initial biliary drainage (one-stageprocedure), while the remaining patients underwent stent placement within 1-7 days of biliary drainage (two-stageprocedure). Biliary obstruction was relieved in 31 of 34 patients (91%).Three patients died within 14 days of stent insertion of unrelated causes, without any change in biliary status. Mean duration of follow-upfor all patients was 5.3 months (range, 0.5-14 months). Four episodes of stent occlusion occurred in three patients (12%occlusion rate); each episode was treated successfully. The average length of hospital stay for patients who underwent a one-stage procedure was 13 days (range, 3-33 days) and was 20 days (range, 9-42 days) for patients who underwent a twostage procedure. The facility of one-step insertion, low occlusion rate, and the many strategies available for treatment of occluded stents make metallic stents an attractive alternative to conventional plastic stents in palliating patients with malignant biliary obstruction. SELF-EXPANDING metallic stents have been used in the vascular system ( I ) , tracheobronchial tree (2), and urinary tract (3).More recently, they have been used in the biliary tree (4-9). These expandable metallic stents offer a new alternative in palliating patients with malignant bile duct obstruction. However, important questions remain to be answered regarding the use of these more expensive metallic stents instead of less expensive conventional stents. Some of these questions include, How technically difficult are they to insert? Does insertion of metal stents in a one-stage procedure confer any extra benefit? However, the most critical question is, DOmetal stents confer any advantages over plastic stents that justify the routine use of metal
stents in patients with biliary obstruction? Early reports described a disappointingly high occlusion rate for metallic stents ( > 20%)(7,s).However, more recent reports have described low incidences of occlusion (43). We now report our experience with selfexpanding metallic endoprostheses in 34 patients with malignant bile duct obstruction and attempt to answer the questions posed above.
PATIENTS AND METHODS Forty-nine metallic stents were placed in 34 patients (gender ratio, 20 men to 14 women; age range, 45-87 years; mean age, 62 years) with malignant bile duct obstruction. All 34
665
666
Journal of Vascular and Interventional Radiology
November 1992
patients were jaundiced, 18 patients had pruritus, and eight were septic at the time of biliary drainage. Serum total bilirubin (normal, 0-1.0 mg/dL [O-17.1 ymol/L]) and alkaline phosphatase (normal, 45 to 115 U/L) levels were elevated in all patients prior to biliary drainage and stent placement. Serum total bilirubin levels ranged from 2.3 to 33 mg/dL (39.3564.3 kmol/L) (mean, 16.4 mg/dL [280.4 kmol/LI) and serum alkaline phosphatase levels ranged from 181 to 1,956 U/L (mean, 506 U/L). Twenty-five of 34 patients had undergone unsuccessful endoscopic drainage, while nine patients were referred from outside institutions with percutaneous biliary drainage catheters in situ. Biliary obstruction was due to pancreatic carcinoma (n = 141, metastatic disease (n = lo), cholangiocarcinoma (n = 6), and local spread from gallbladder carcinoma (n = 2), gastric carcinoma (n = I),and hepatocellular carcinoma (n = 1). Thirtyeight malignant strictures, ranging in length from 1to 7 cm (mean, 3.2 cm), were present in the 34 patients. The strictures were located in the lower common bile duct (22 of 38 strictures), in the middle of the common bile duct (six of 38), and in the region of the hilar confluence (10 of 38). Pathologic confirmation of malignancy was obtained in all patients by means of either needle biopsy (16 of 34 patients), surgery (15 patients), or percutaneous biliary brushing (three patients). The Wallstent endoprosthesis (Schneider, Plymouth, Minn) was used in all patients. This stent is composed of braided stainless steel monofilaments and is 6.8 cm in length when expanded to its full diameter of 1cm. The stent is compressed onto a 7-F delivery catheter and covered by a plastic membrane that rolls back to release the stent when the membrane is hydraulically activated. Our technique for stent insertion did not differ substantially from the techniques described by other authors (5-7). Antibiotics (1g of ampicillin and 80 mg of gentamicin) were routinely administered intravenously prior to stent placement.
All patients underwent percutaneous biliary drainage performed in standard fashion before stent placement. Internal biliary drainage was achieved with all patients at the time of the initial drainage. Four patients with hilar lesions required both left- and right-sided biliary drainage. In 13 patients (38%),metallic endoprostheses were placed in a one-stage procedure a t the time of initial biliary drainage, while in the remaining patients, stents were placed within 1-7 days (mean, 3 days) of the biliary drainage when pathologic confirmation of malignancy became available. Metal stents were placed a t the time of initial biliary drainage if a confirmed diagnosis of malignancy was available. The stent delivery catheter was introduced over a heavy-duty guide wire (0.038-inch Arnplatz extra stiff; Cook, Bloomington, Ind). An 8-F sheath was first placed over the stent delivery catheter but was not actually introduced into the liver. The sheath was introduced over the stent delivery catheter if problems arose in deploying the stent, so that the stent, if partially open, could be withdrawn through the sheath without injuring the liver. The stent was ~ositioned across the obstructing lesion so that the stent overlapped the proximal and distal ends of the stricture. Further stents were deployed as necessary to achieve this goal (Fig 1).After deployment, the stent was dilated at the site of the stricture with an 8-mm balloon catheter in patients in whom a prominent waist in the stent was seen a t the site of stricture (30 of 38 strictures). An 8-F external catheter was left in situ for a mean of 4 days (range, 1-10 days) and was removed after contrast material injection confirmed stent patency and good position of the stent vis-a-visthe stricture. Follow-UDinformation was obtained on 21 patients by examining hospital records if patients were followed up a t our institution or by contacting the patient's family or primary doctor if the patient was followed up a t another institution. Reports of further admissions to hos-
pital, redevelopment of jaundice, and levels of serum alkaline phosphatase and bilirubin, where available, were noted.
RESULTS A total of 49 stents were required for the 38 strictures. In 22 patients, a single stent was sufficient to bridge the stricture, while in 12 patients, more than one stent was necessary. Relief of obstruction with disappearance of jaundice and pruritus occurred in 3 1 of 34 patients (91%) within 2-6 weeks of stent insertion. Bilirubin and alkaline phosphatase levels were available for 30 of 34 patients due to the fact that some patients returned to other institutions for further care. These 30 patients had either normal (10 patients) or decreasing levels (20 patients) of total bilirubin (range, 0.3-28 mg/dL; mean, 5.6 mg/dL [range, 5.1-478.8 kmol/L; mean, 95.8 kmol/LI) or alkaline phosphatase (range, 75-900 U/L; mean, 274 UIL) a t the time of biochemical testing (range, 1day to 6 months after stent placement; mean, 36 days). Three patients died within 2 weeks of stent insertion.
Patient Survival The total length of follow-up for all patients averaged 5.3 months (range, 0.5-14 months). Twenty-four of 34 patients (71%)died between 3 days and 14 months (mean, 3.0 months) after stent insertion. Three of these 24 patients died within 30 days of stent insertion. One patient had a pulmonary embolus 2 days after stent insertion, while the remaining two patients had extensive metastatic disease and developed both liver and renal failure. Twenty-three of the 24 patients required no further intert on vention and had ~ a t e n stents clinical follow-uiat the time of death. Metallic stent occlusion due to tumor overgrowth occurred in one patient, requiring further percutaneous intervention. Ten of 34 patients (29%)are alive and well with functioning stents a t
Lee et a1
667
- -
Volume 3 Number 4
an average of 6.5 months after insertion (range, 3-10 months), although two of these patients required further intervention for stent occlusion. Average length of hospital stay for patients who underwent a one-stage procedure was 13 days (range, 3-33 days) and was 20 days (range, 9-42 days) for patients who underwent a two-stage procedure. Two of the 13 patients who underwent a one-stage procedure were in poor clinical condition prior to stent placement, while eight of the 21 patients who underwent a two-stage procedure were in poor clinical condition prior to stent placement.
Short-term Complications Three ~ a t i e n t had s com~lications related to percutaneous bikary drainage. Biliary sepsis occurred in two patients and was successfully treated with antibiotics. A third patient with a coagulopathy experienced a decrease in hematocrit after percutaneous biliary drainage, which required red cell transfusion, vitamin Kinjection, and fresh frozen plasma to stabilize the patient's condition. Insertion of metal stents was consequently delayed until each patient's condition was stable.
C. d. Figure 1. Images of a patient with metastatic portal adenopathy causing biliary obstruction. Two stents were required to fully bridge the stricture. (a)Transhepatic cholangiography reveals a stricture within the middle of the common bile duct without distal passage of contrast material into the duodenum. Surgical clips (arrow) mark the approximate level of the proximal end of the stricture. (b) A Wallstent was deployed across the stricture but migrated distally during insertion so that only a small portion of the stent was proximal to the stricture at the level of the surgical clips (arrow). Contrast material injection did not reveal any proximal filling of the biliary tree supporting this conclusion. (c) A second Wallstent was deployed to provide adequate overlap proximal to the stricture. An 8-mm balloon (arrows) was used to inflate the metal stents a t the area of the stricture where a waist was present. (d) After balloon dilation the stents (arrows) have opened up to a large luminal diameter. An 8-F external pigtail catheter was left in situ for 3 days and removed after contrast material injection revealed good drainage into the duodenum. The patient died 5 months after stent insertion of disease progression without evidence of stent malfunction.
Technical Difficulties Technical complications occurred early in our experience and included stent delivery in the wrong location and failure of stent deployment due to incomplete retraction of the rolling membrane. Deployment of stents in the wrong location occurred in one patient. The stent was deployed too distally across the ampulla and migrated into the duodenum and then into the stomach. A new stent was deployed in the correct location, and the stent in the stomach was removed endoscopically. Incomplete retraction of the rolling membrane occurred with five stents. These stents were removed and fresh stents inserted. Long-term Complications One episode of stent occlusion occurred in two patients with hilar
668
Journal of Vascular and Interventional Radiology
November 1992
strictures, and two episodes of occlusion occurred in one patient with a low common bile duct stricture. Thus, our total occlusion rate was 12% (four episodes of occlusion in 34 patients). Stent occlusion was due to tumor overgrowth in one patient with a hilar stricture (at 5 months after stent insertion). The patient was in terminal condition, and a n internallexternal catheter was placed through the stent for the remainder of the patient's life (1month). Stent occlusion was due to debris and inspissated bile in the remaining two patients (Fig 2). Both patients were older than 80 years of age and were admitted with cholangitis and dehydration at 1and 2 months after stent insertion. Treatment consisted of intravenous antibiotics, rehydration with intravenous fluids, and short-term percutaneous drainage of the biliary tree. Stents were cleared of debris with the use of a balloon catheter, which was inflated and passed up and down through the stent to reestablish patency (Fig 2). When the acute septic episode had abated, the drainage catheters were removed. A second episode of occlusion occurred in one of these patients 2 months later and was treated in the same manner. Both patients are presently alive and well (6 and 8 months after the first episode of occlusion) with functioning stents and without further episodes of occlusion.
DISCUSSION Conventional plastic endoprostheses for palliation of malignant biliary obstruction are associated with two main problems, stent migration and occlusion. While the problem of stent migration has largely been solved by the provision of anchoring sutures, stent occlusion remains a difficult problem. Mueller et a1 (lo), in a multicenter review, found that occlusion occurred in 23% of patients on longterm follow-up, while Mendez et a1 (11)believed that the problems with plastic biliary endoprostheses were of sufficient magnitude to warrant that
Comparison of Wallstent Endoprosthesis Occlusion Rates No. of Author Year Patients Gillams et a1 (7) 1990 45 Lammer et a1 (9) 1990 61 Lameris et a1 (4) 1991 69 Adam et a1 (5) 1991 41
their use be completely abandoned. However, Dick et al(12) in a series of 100 patients from one institution reported a lower occlusion rate of 12.3%, but the total long-term complication rate was 17.5% when the migration rate (5.2%)was considered. Against this background skepticism regarding the use of plastic stents, there has been increasing interest in the use of metallic endoprostheses. These metallic endoprostheses, when deployed across a biliary stricture, open up to provide a luminal diameter (24-30 F) significantly larger than that of plastic endoprostheses. This large luminal diameter offers the theoretical advantage of longerterm patency vis-a-vis the smaller luminal diameter of the ~ l a s t i cendoprosthesis. In addition, migration does not occur with metallic stents because they become incorporated and fixed in the wall of the common bile duct. Earlier reports, however, detailing the use of metallic biliary endoprostheses, did not justify this initial optimism. Gillams et a1 (7) reported an occlusion rate of 42% for the Wallstent endoprosthesis, while Irving et a1 (8) reported an occlusion rate approaching 50% for the Gianturco stent. However, more recent reports on the Wallstent endoprosthesis by Lammer et a1 (91, Lameris et a1 (4), and Adam et al(5) are more encouraging (Table). These authors reported stent occlusion rates of 13%, 1096, and 7%, respectively. In the series reported by Lameris et al (41, stent occlusion mainly occurred in ~ a t i e n t with s hilar strictures. Our long-term occlusion rate, a t 12%, was similar to the low occlusion rates found by previous authors (4,5,9).
Occlusion Rate (%I 42 13 10 7
Unlike Lameris et a1 (4),we did not see a large occlusion rate for patients with hilar strictures; however, the number of patients with hilar strictures in this series was small (10 patients), and statistical inferences cannot be made from this. Although metal stents cannot be removed once inserted, there are many stratagems available for treatment of metallic stent occlusion, as there are with plastic stent occlusion. The sine qua non for management of endoprosthesis occlusion is the placement of a guide wire through the center of the endoprosthesis. This can be technically difficult with occluded plastic endoprostheses because of the small luminal diameter (13). However, the large luminal diameter (24-30 F) of the metallic endoprostheses facilitates easy passage of a guide wire. Once this has been achieved, many therapeutic options are available. If stent occlusion is due to inspissated bile, the patient is treated with intravenous fluids and antibiotics and the metal stent is cleared of debris by passing a balloon catheter through the lumen of the stent (Fig 2). A temporary drainage catheter is left in situ until any associated cholangitis has improved and contrast material injection shows a patent stent. The drainage catheter is then removed. Alternatively, if the patient is judged to be in terminal condition, a long-term internal-external catheter can be left in situ. If the metal stent becomes occluded by tumor. a new metal stent can be inserted percutaneously through the occluded stent. a long-term internal-external catheter can be placed, or a plastic stent can be inserted endoscopically through the occluded metallic stent. u
Lee et a1
669
Volume 3 Number 4
Figure 2. Images from an 82-year-old man with ampullary carcinoma in whom a metallic stent was inserted. The patient returned 2 months later with jaundice and sepsis. (a)Transhepatic cholangiography shows complete occlusion of the stent by debris (arrows). (b) A guide wire was manipulated easily down through the wide central lumen of the stent and into the duodenum. (c) An 8-mm balloon catheter (arrows) was passed up and down through the stent to clear it of debris. (d) Contrast material injection after the stent is cleared shows free flow of contrast material into the duodenum. An external catheter was left in situ for 5 days while the patient was rehydrated and treated with intravenously administered antibiotics. The external catheter was then removed after repeated contrast material injection confirmed stent patency. Two months later a similar event occurred, which was treated in the same manner. The patient is currently doing well with a functioning stent (4 months after second stent occlusion).
Tumor overgrowth occurred in one of our patients early on in this series and was due to a combination of
stent shortening and poor initial stent placement because of limited biliary purchase above the proximal
end of the tumor. We elected to place a long-term internal-external catheter through the occluded stent be-
670
Journal of Vascular and Interventional Radiology
November 1992
cause of the patient's terminal condition. Stent shortening normally occurs with Wallstent metallic endoprostheses as a means of achieving maximal luminal diameter, that is, the stent shortens as it increases its luminal diameter. Therefore, stent shortening coupled with the proximity of the tumor to the proximal end of the stent can lead to tumor overgrowth. This can be compensated for by ensuring a wide margin of error when placing metallic stents so that proximal and distal ends overlap the stricture as much as possible. Stent occlusion due to debris and inspissated bile occurred in two patients, both of whom were severely dehydrated and older than 80 years. Both patients had cholangitis and jaundice at presentation. Stent occlusion was considered due to severe dehydration with bile inspissation in both of these patients. A further episode of stent occlusion occurred in one of these patients due to similar causes. Both patients underwent another drainage procedure during the acute episodes, and the stent was cleared of debris by using a balloon catheter. There was no evidence of tumor overgrowth in either of these patients, and internal-external catheters were removed after the acute episodes had subsided. Insertion of the metallic Wallstent endoprosthesis, although associated with a steeper learning curve than that associated with the plastic endoprosthesis, is in our experience, easier once familiarity with the technique is gained. The transhepatic tract does not have to be dilated to 14 or 16 F as is needed with plastic endoprostheses. Indeed, the 7-F delivery catheter makes the procedure more comfortable for the patient and facilitates metal stent insertion at the time of initial percutaneous biliary drainage, provided a diagnosis of malignant bile duct obstruction has been confirmed. Metal stents were inserted in 13 such patients (38%)at the time of percutaneous biliary drainage in this series. This subgroup of patients had a shorter hospital stay (mean, 13 days; range, 3-33
days) compared with the remaining 21 patients who had metal stents inserted 1-7 days after percutaneous biliary drainage (mean, 20 days; range, 9-42 days). While realizing that a contributing factor to the longer hospital stay of patients who underwent a two-stage procedure was the poor clinical condition of many of these patients, the possibility of metal stent insertion at the time of initial biliary drainage is a desirable advantage over plastic stents. In addition, comparison of metal stent occlusion rates from this (12%)and other recent series (10% [41, 7% [51, and 13% [91) with those reported for plastic stents (23% [lo] and 12.3% [12]) suggests that metallic stents may have lower occlusion rates than plastic stents. In addition, stent migration, which can be a problem with plastic stents, does not occur with metallic stents. The extra cost of the Wallstent metallic stent ($1,000) compared with conventional plastic stents ($60-$160) may be offset if low occlusion rates are confirmed in further studies with longterm follow-up and if metal stents are inserted in a one-step procedure to shorten hospital stay. We recognize some limitations of our study. The majority of patients (71%) in our study died within 14 months (mean, 3 months) of stent placement. This was due to a combination of the poor clinical condition of many patients prior to stent placement and to the aggressive nature of their diseases. We did not try to select patients for metal stent placement who might have had a longer life expectancy; rather, metal stents were placed in all patients with malignant disease undergoing biliary drainage during the time of this study. Another limitation of our study includes the lack of biochemical follow-up in all patients. Clinical follow-up was available in all patients through contact with the patients' doctor or family. However, because patients came from a national and international referral base, access to biochemical data was limited. The ease of insertion and facility
for one-step insertion give the metallic Wallstent endo~rosthesisdistinct advantages over piastic stents. Although only short-term follow-up is available at the present time, the modestlv low occlusion rates in this and other recent series compare favorably with those of plastic stents. Despite the permanency of metallic stents once inserted, the ease of insertion and the many stratagems available for treatment of stent occlusion make metallic stents an attractive alternative to plastic stents for palliation of malignant bile duct obstruction. References 1. Kimihiko K, Hideo U, Tetsuya Y, et al. Iliac artery stenosis and occlusion: preliminary results of treatment with Gianturco expandable metallic stents. Radiology 1990; 177: 799-802. 2. Wallace MJ, Charnsangavej C, Ogawa K, et al. Tracheobronchial tree: expandable metallic stents used in experimental and clinical applications. Radiology 1986; 158:309-312. 3. Milroy EJG, Chapple CR, El-Din A, Wallsten H. A new stent for the treatment of urethral strictures. J Uro11989; 141:1120-1122. 4. Lameris JS, Stoker J , Nijs HGT, et al. Malignant biliary obstruction: percutaneous use of self-expandable stents. Radiology 1991; 179:703707. 5. Adam A, Chetty N, Roddie M, Yeung E, Benjamin IS. Self-expandable stainless steel endoprostheses for treatment of malignant bile duct obstruction. AJR 1991; 156:321325. 6. Coons HG. Self-expanding stainless steel biliary stents. Radiology 1989; 170:979-983. 7. Gillams A, Dick R, Dooley JS, Wallsten H, El-Din A. Self-expandable stainless steel braided endoprosthesis for biliary strictures. Radiology 1990; 174:137-140. 8. Irving JD, Adam A, Dick R, Dondelinger RF, Lunderquist A, Roche A. Gianturco expandable metallic biliary stents: results of a European clinical trial. Radiology 1989; 172: 321-326. 9. Lammer J , Klein GE, Kleinert R, Hausegger K, Einspieler R. Obstructive jaundice: use of expandable
Lee et a1
671
Volume 3 Number 4
metal endoprosthesis for biliary drainage. Radiology 1990; 177:789792. 10. Mueller PR, Ferrucci J T J r , Teplick SK, et al. Biliary stent endoprosthesis: analysis of complications in 113 patients. Radiology 1985; 156: 637-639. 11. Mendez G Jr, Russell E, LePage JR, Guerra J J , Posniak RA, Trefler M.
Abandonment of endoprosthetic drainage technique in malignant biliary obstruction. AJR 1984; 143: 617-622. 12. Dick BW, Gordon RL, LaBerge JM, Doherty MM, Ring EJ. Percutaneous transhepatic placement of biliary endoprostheses: results in 100 consecutive patients. JVIR 1990; 1:97-100.
13. Lee MJ, Mueller PR, Saini S, Morrison MC, Brink JA, Hahn PF. Occlusion of biliary endoprostheses: presentation and management. Radiology 1990; 176:531-534.