| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Vascular and Interventional Radiology |
1 From the Department of Gastroenterology, Institut de Malalties Digestives (V.P., A.C., J.M.B., J.L., F.F., S.N.), and Vascular and Interventional Radiology Unit, Centre de Diagnòstic per la Imatge, (M.I.R., X.M.), Hospital Clínic, Villarroel 170, 08036 Barcelona, Spain. Received September 14, 2001; revision requested November 15; revision received January 10, 2002; accepted February 28. Supported in part by grants from the Ministerio de Ciencia y Tecnoloía (SAF00-0038), Fondo de Investigación Sanitaria (01/0104-02), and the Agència d’Avaluació de Tecnologia i Recerca Mèdiques of the Generalitat de Catalunya (026/16/2000). V.P. is a research fellow from the Institut d’Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS). Address correspondence to S.N. (e-mail: snavarro@clinic.ub.es).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Patients with biliary obstruction due to inoperable primary carcinoma of the pancreas, gallbladder, or bile ducts or regional lymph node metastases were included. Evaluated outcomes included technical and therapeutic success rates, morbidity and 30-day mortality rates, hospital stay length and readmission, biliary reobstruction, and overall survival rates. Data were analyzed according to both the intention-to-treat principle and the treatment actually administered. Univariate (Kaplan-Meier method) and multivariate (Cox model) analyses were performed.
RESULTS: After randomization, 28 patients were assigned to receive a percutaneous self-expanding metal stent and 26 patients to receive a 12-F endoscopic polyethylene prosthesis. The technical success rates of both implantation procedures were similar (percutaneous, 75% [21 of 28 patients]; endoscopic, 58% [15 of 26 patients]; P = .29), whereas therapeutic success was higher in the percutaneous group (71% [20 of 28 patients] vs 42% [11 of 26 patients]; P = .03). However, major complications were more common in the percutaneous group (61% [17 of 28 patients] vs 35% [nine of 26 patients]; P = .09) but did not account for differences in 30-day mortality rates (percutaneous, 36% [10 of 28 patients]; endoscopic, 42% [11 of 26 patients]; P = .83). Overall median survival was significantly higher in the percutaneous group than in the endoscopic group (3.7 vs 2.0 months; P = .02). Cox regression analysis enabled identification of placement of the percutaneous self-expanding metal stent as the only independent predictor of survival (relative risk, 2.19; 95% CI: 1.11, 4.31; P = .02).
CONCLUSION: Placement of a percutaneous self-expanding metal stent is an alternative to placement of an endoscopic polyethylene endoprosthesis in patients with malignant biliary obstruction.
© RSNA, 2002
Index terms: Bile ducts, interventional procedures, 766.1229, 766.1267, 774.1269 • Bile ducts, stenosis or obstruction, 766.1229, 766.1267, 774.1269 • Pancreas, interventional procedures, 774.1229, 774.1267, 774.1269 • Pancreas, therapeutic radiology, 774.1229, 774.1267, 774.1269
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
There are two main options for palliative relief of biliary obstruction: surgical decompression by using biliary enteric anastomosis and nonsurgical methods, in which an internal-external catheter or, more important, an endoprosthesis, is placed by way of either transhepatic or transduodenal routes. At the time this article was written, four randomized trials in which surgical bypass was compared with either endoscopic (4–6) or percutaneous stent placement (7) had been published. Although no significant differences in overall survival were observed, in some of these studies (4,5), stent placement was associated with lower procedure-related mortality rates, lower major complication rate, and a shorter median total hospital stay than surgery. However, long-term results with respect to late gastric outlet obstruction and hospital readmission seem to favor surgical bypass (4,8).
In cases in which a nonsurgical approach is favored, endoscopic stent placement has been the most popular palliative procedure. However, to the best of our knowledge, only one report of a randomized trial in which endoscopic and percutaneous stent insertion were compared (9) has been published. In this study, published more than 10 years ago, percutaneous biliary drainage with a rigid external catheter was shown to be less successful than endoscopic insertion in relieving jaundice and was associated with a higher 30-day mortality rate due to complications of hepatic puncture, that is, hemorrhage and bile leak (9). These clear-cut results may have been responsible, at least in part, for the shortage of additional investigations in which both approaches are compared; consequently, endoscopic stent placement became the standard of care for palliation in patients with malignant biliary obstruction (8).
Improvements in technology in the past few years might be a reason to change this assumption. In fact, development of self-expanding metal stents for biliary application avoids the permanent presence of an external catheter, thus decreasing the incidence of major complications (ie, cholangitis, bile peritonitis, hemorrhage) and tube-related problems (ie, dislodgement, migration, and blockage), as well as improving the patient’s psychological condition (8). In addition to these meaningful advantages, larger-diameter metal stents seem to have longer patency than straight polyethylene prostheses, a fact that might decrease the repeat intervention rate and shorten total hospital stay (10–12). However, to our knowledge, the effect of metal stents on results obtained with percutaneous biliary drainage has not yet been properly evaluated.
The purpose of our study was to compare percutaneous self-expanding metal stents with conventional endoscopic polyethylene endoprostheses for the treatment of malignant biliary obstruction by means of a prospective randomized clinical trial.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The target study population consisted of patients with biliary obstruction due to primary carcinoma of the pancreas, gallbladder, or bile ducts, or to regional lymph node metastases. All patients were considered for surgery but could be rejected because of disease extent (metastases, portal vein or hepatic arterial obstruction, or invasion of both hepatic lobes) or severe disability due to age or associated diseases. Diagnosis was based on symptoms at presentation and imaging findings, with pathologic confirmation whenever possible.
Patients were excluded from the trial if they had massive tumoral hepatic infiltration (multiple nodules in both lobes or diffuse infiltration), had undergone previous biliary drainage, had a performance status score (according to the Eastern Co-operative Oncology Group scale) greater than 2, or had refused study participation. Factors known to constitute a limitation or contraindication for one specific stent placement method (ie, clotting abnormalities for percutaneous treatment or gastric outlet obstruction for endoscopic drainage) were not considered exclusion criteria.
The protocol was approved by the ethics committee of the Hospital Clinic, and written informed consent was obtained from each patient. Interim analysis was planned after inclusion of half of the calculated sample size. Criteria for study interruption were significant differences in either 30-day mortality or probability of overall survival observed in such preliminary analysis.
Study Design
Randomization was performed after complete patient work-up, which included blood analysis (hematology; biochemistry, including hepatic and renal tests, and clotting tests), chest radiography, abdominal ultrasonography (US), and/or computed tomography. Patients were stratified into two groups according to the level of biliary obstruction (distal and proximal with respect to the cystic duct), and subsequently randomized, by means of sealed envelopes containing computer-generated random numbers, to receive a percutaneous self-expanding metal stent or an endoscopic polyethylene endoprosthesis. Placement of a plastic stent was the preferred method of endoscopic palliation in our geographic area at the time this trial was designed, because of its low price and ease of performance; consequently, it was performed in the endoscopic group.
The protocol allowed crossover of patients into the other treatment group during the first hospital admission after failure of the primary randomized option. After initial success, patients with recurrence of biliary obstruction during follow-up were considered for prosthesis replacement by using the same procedure previously used.
The main end point of the study was 30-day mortality. The study was designed to evaluate whether, even if no significant differences in this parameter were observed between stent placement procedures, differences in other variables studied favored one approach. Therefore, assuming a 20% 30-day mortality rate in the percutaneous group, a minimum of 50 patients per group was required to demonstrate that the procedures were equivalent (less than 10% difference), with an 
level of .20 and a ß error of .05. Other outcomes were technical success in stent insertion; therapeutic success, described as a decrease in serum bilirubin level of greater than 20% of the preprocedure value (9); procedure-related morbidity during the hospital stay (failure of successful biliary drainage was not recorded as a complication, although it may have been implicated as the cause of death); length of postrandomization hospital stay, defined as number of days spent in the hospital from randomization until first discharge; length of postprocedural hospital stay, defined as the number of hospital days from stent placement until first discharge; length of total hospital stay, defined as the number of hospital days from randomization until death (including both initial stay and readmissions); probability of avoiding readmission from discharge until first readmission; probability of freedom from biliary obstruction due to stent placement until redevelopment of cholestasis due to biliary obstruction; and probability of overall survival from randomization until death.
Treatment and Follow-up
All patients received 200 mg/12 h of intravenously administered a prophylactic antibiotic (ciprofloxacin; Baycip, Barcelona, Spain) before and usually 24–48 hours after the procedure. Coagulation tests were performed before the procedure, and any clotting abnormality was corrected. Stents were inserted by radiologists (M.I.R. and X.M.) and gastroenterologists (J.M.B. and J.L.) with substantial experience (at least 50 stent placements each prior to trial initiation).
Percutaneous technique.—Percutaneous transhepatic cholangiography was performed to localize the site of biliary obstruction, and then a guide wire was placed through the stenosis (single-needle technique). At that time, first a 5-F catheter (William Cook Europe, Denmark) and thereafter a 7–8-F catheter (PBN Medical, Denmark), with its distal end in the duodenum for internal-external drainage, was placed. The catheter was fixed to the skin and connected to a sealed bag, where bile was collected for the first 24–48 hours. A self-expanding metal stent (Wallstent; Boston Scientific, Boston, Mass) mounted on a delivery system was then placed. After progressive withdrawal of the constricting membrane, the prosthesis expanded to 0.8–1.0 cm in diameter (30 F) and shortened from 10.0 to 6.8 cm (13). Dilation was usually performed after stent insertion. An external drainage catheter was left in place for 1 day after stent placement.
Endoscopic technique.—Stent placement was attempted during the endoscopic procedure in all cases. Stents were polyethylene, with a 12-F external diameter (Biliary Drainage Tube PBD; Olympus Optical, Hamburg, Germany). They were inserted over a guide wire and coaxial catheter by means of a duodenoscope (model TJF 100/130; Olympus Optical) by using standard techniques (14). Papillotomy was performed in all patients by means of electrocautery. Strictures were not systematically dilated ahead of time.
After adequate prosthesis insertion, patients were discharged and followed up in the outpatient clinics. Postprocedural surveillance consisted of recording patient medical history and performing physical examination and laboratory studies 15 days after stent placement and every 2 months thereafter. If jaundice or cholestatic symptoms reappeared, or if a substantial increase in serum bilirubin concentration (greater than 20% of the postprocedural value) was observed, abdominal US was performed to confirm biliary obstruction. When patients were unable to attend medical appointments because of clinical deterioration, their relatives were interviewed by telephone at 1-month intervals until the patient’s death. Physicians providing follow-up were not blinded to the type of stent placed.
Cost Analysis
Cost was analyzed with regard to survival. Costs were calculated for each patient, including length of total hospital stay (U.S. $240 per day) and all prostheses used ($1,400 for a metal stent and $285 for a polyethylene prosthesis). Prices were estimated according to the 2000 charge list of the Hospital Clínic. Indirect costs, such as time lost from work or transportation charges, were not factored into the analysis (15).
Statistical Methods
Continuous variables were expressed as means ± SDs and compared by performing the Student t test. Categoric variables were compared by means of the 
2 test, with application of Yates correction when necessary.
Data were analyzed according to both the intention-to-treat principle and the treatment actually applied. Probability curves of overall survival and freedom from biliary obstruction and readmission were calculated according to the Kaplan-Meier method and compared by performing the log-rank test. Data regarding patients who were alive, had no evidence of biliary obstruction recurrence, or were not readmitted at the end of the study were not counted. Probability data were expressed as medians.
Cox proportional hazards modeling with forward selection was planned to avoid the confounding effects of baseline characteristics on patient survival and was performed in a two-step approach (16). First, univariate analysis was performed and included any potential prognostic factor: age, sex, tumor type, biliary obstruction level, duodenal invasion, comorbid medical conditions, performance status, serum bilirubin level, alkaline phosphatase level, 
-glutamyltranspeptidase and serum creatinine concentrations, and prothrombin activity. For continuous variables, the cutoff level chosen was their median value. Thereafter, only variables with a P value of less than .10 at univariate analysis, along with stent placement, were introduced in the Cox model.
All P values were two sided, with a P value of less than .05 considered to indicate a statistically significant difference. Analysis was performed in April 2001 by using SPSS software (SPSS, Chicago, Ill).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
When therapeutic success was calculated according to the treatment actually applied, there were no differences between groups (percutaneous, 25 [93%] of 27 patients; endoscopic, 12 [75%] of 16 patients; P = .17). Similarly, the mean serum bilirubin concentration 15 days after stent placement was 4.5 mg/dL ± 4.3 in the percutaneous group and 9.6 mg/dL ± 8.6 in the endoscopic group (P = .05).
Periprocedural Morbidity and Mortality
The major complication rate was higher in the percutaneous group than in the endoscopic group, although this difference was not statistically significant (61% vs 35%, respectively; P = .09) (Table 2).
On an intention-to-treat basis, there was no difference between groups with respect to 30-day mortality rate (percutaneous, 36%; endoscopic, 42%; P = .83) (Table 2). Similar results were obtained when analysis was performed according to the treatment actually applied (percutaneous, five [19%] of 27 patients; endoscopic, seven [44%] of 16 patients; P = .15).
Duration of Hospital Stay
According to the intention-to-treat principle, there were no differences between groups with regard to either postrandomization (P = .87) or postprocedural (P = .12) hospital stay (Table 2). During follow-up, nine (32%) patients from the percutaneous group and five (19%) from the endoscopic group were readmitted. The probability of avoiding readmission was similar in both treatment groups (percutaneous, 8.1 months; endoscopic, 6.6 months; log-rank test, 0.16; P = .68). In addition, there were no differences with regard to the mean number of readmissions per patient (percutaneous, 1.9 ± 1.4; endoscopic, 1.2 ± 0.5; P = .31). Finally, the mean total hospital stay was similar in both treatment groups (percutaneous, 21.5 days ± 17.4; endoscopic, 17.5 days ± 9.8; P = .31).
When results were expressed according to the treatment actually administered, both groups were similar with respect to mean postrandomization stay (percutaneous, 16.7 days ± 8.8; endoscopic, 12.4 days ± 7.1; P = .10). However, postprocedural stays were longer in patients who received a percutaneous stent than in those who received an endoscopic prosthesis (11.9 days ± 8.7; 7.0 days ± 6.4, respectively; P = .056). Twelve (44%) patients from the percutaneous group and two (13%) from the endoscopic group were readmitted. The probability of avoiding readmission was similar in both treatment groups (percutaneous, 5.7 months; endoscopic, 3.9 months; log-rank test, 0.03; P = .87). In addition, although there was no difference with regard to the mean number of readmissions per patient (percutaneous, 1.7 ± 1.2; endoscopic, 1.5 ± 0.7; P = .86), the mean total hospital stay was longer in patients treated with a percutaneous stent than in those treated with an endoscopic prosthesis (14.0 days ± 8.3; 23.4 days ± 16.8, respectively; P = .04).
Recurrence of Biliary Obstruction
After a median follow-up of 2.5 months (range, 1–19 months), biliary obstruction recurred in 12 (43%) patients in the percutaneous group and in 14 (54%) patients in the endoscopic group. Accordingly, there was no difference in the probabilities of freedom from biliary obstruction on an intention-to-treat basis (percutaneous, 6.2 months; endoscopic, 3.4 months; log-rank test, 1.15, P = .28) (Fig 2).
|
|
Patient Survival
By the end of follow-up, all patients had died. On an intention-to-treat basis, the probability of overall survival was significantly higher in the percutaneous group than in the endoscopic group (3.7 vs 2.0 months; log-rank test, 5.13; P = .02) (Fig 3). At univariate analysis, performance status score at entry was the only additional variable associated with patient survival (Table 3). The Cox regression model, including both the stent placement procedure and performance status score, was used to identify placement of a percutaneous self-expanding metal stent as an independent predictor of survival (relative risk, 2.19; 95% CI: 1.11, 4.31; P = .02).
|
|
|
Cost Analysis
According to the intention-to-treat principle, there were no differences in either the mean cost of first admission (percutaneous, $4,970 ± 2,339 vs endoscopic, $4,224 ± 2,275; P = .24) or the cost of overall treatment (percutaneous, $6,368 ± 4,507 vs endoscopic, $4,767 ± 2,485; P = .12).
When analysis was performed according to the treatment actually administered, percutaneous stent placement was significantly more expensive than endoscopic prosthesis placement with respect to either first admission (mean cost, $5,422 ± 2,112 vs $3,234 ± 1,689; P = .001) or overall treatment (mean cost, $7,153 ± 4,215 vs $3,639 ± 2,008; P = .003). However, when these numbers were adjusted for patient survival, differences in the mean cost of overall treatment disappeared (percutaneous, $3,198 ± 4,043 per month alive vs endoscopic, $3,241 ± 2,854 per month alive; P = .97).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The results achieved in the present study are substantially different from those obtained in the only previous randomized trial to our knowledge in which endoscopic and percutaneous drainage were compared (9). The unambiguous conclusion of the former investigation, in which a rigid, external transhepatic catheter was used, limited wider use of the percutaneous approach because of the high mortality rate associated with it. However, a randomized trial in which transhepatic stent placement was compared with surgical bypass in patients with unresectable pancreatic cancer did not have such discouraging results (7). In that study, the incidence of postprocedural complications and 30-day mortality was similar in both treatment groups. Unfortunately, other available information regarding the utility of percutaneous stents lies only in phase II noncomparative analysis that reflects the particular experience of different groups, thus making it difficult to extrapolate their results (17–20).
The apparently low numbers obtained in both therapeutic groups with respect to technical success and effectiveness may depend, at least in part, on the design of the trial and deserves commentary. First, exclusion criteria used in similar studies (4–7,9–12), such as factors known to constitute a limitation or contraindication for one specific procedure, were not considered in this investigation. Second, the intention-to-treat principle followed in our analysis justifies some discrepancies with noncomparative studies, in which only successfully placed prostheses are usually considered. Third, the patient series in our study was composed of patients unfit for surgery because of tumor extent or severe disability, thus representing a population with a poor prognosis. Finally, most patients had lesions that caused proximal biliary obstruction, a condition that usually limits the technical success of a treatment. Unfortunately, this aspect could not be ascertained in our study because of the reduced number of cases remaining in each subset of patients.
The most meaningful results of the current investigation are those regarding long-term outcome. Indeed, the probability of survival in patients treated with a percutaneous metal stent was higher than that in patients treated with an endoscopic plastic endoprosthesis. This fact, which led to termination of the study, seems to have been determined by a combination of events, such as the higher therapeutic success rate, the limited perioperative mortality rate, and the higher stent patency rate. This last parameter was reflected by both lower serum bilirubin concentration after treatment and lower probability of biliary obstruction recurrence in patients in whom a percutaneous stent was placed.
There are few doubts that the major determinant of the excellent results obtained in the percutaneous group, especially when they are compared with those achieved in previous experiences, is the technological improvement in biliary prostheses (18–20). In our investigation, a self-expanding metal stent was used instead of the straight polyethylene endoprosthesis used a few years ago (9,17). Adequate comparisons between both materials are limited to three prospective randomized trials in which stents were placed by means of either endoscopy or a combined percutaneous-endoscopic route (10–12). Investigators in these studies concluded that metal stents offered a higher success rate and longer prosthesis patency than did polyethylene prostheses. In addition, it has been suggested that this more successful approach is even less costly when considering that patients required fewer repeat interventions and readmissions (10), similar to what was observed in the current study.
Although the current analysis was not specifically designed to investigate the effect of biliary stent placement on the patient’s quality of life, data from our study may enable estimation of this important parameter. If the quality of life of patients with malignant biliary obstruction is determined mainly by assessing procedure-related complications, persistence or recurrence of cholestasis, length of total hospital stay, and physical impairment due to cancer progression, and if we exclude the last parameter, which could not be accurately assessed, then our results in patients assigned to percutaneous treatment were, at the very least, equivalent to those in patients in the endoscopic group.
In conclusion, our results suggest that percutaneous drainage with self-expanding metal endoprostheses is an alternative to endoscopic polyethylene stent placement in patients with malignant biliary obstruction, with potential favorable advantages for survival. With consideration that stent characteristics seem to play an important role in long-term outcome, these results argue against the use of rigid polyethylene prostheses rather than exclude the endoscopic route. In that sense, this study provides a rationale for comparing percutaneous and endoscopic self-expanding metal stents to define the better approach to this important clinical issue.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, S.N.; study concepts, all authors; study design, A.C., S.N.; literature research, V.P., A.C.; clinical studies, all authors; data acquisition, V.P., A.C.; data analysis/interpretation, V.P., A.C., S.N.; statistical analysis, A.C.; manuscript preparation, V.P., A.C., S.N.; manuscript definition of intellectual content, S.N.; manuscript editing, V.P., A.C.; manuscript revision/review, S.N.; manuscript final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. Donelli, E. Guaglianone, R. Di Rosa, F. Fiocca, and A. Basoli Plastic Biliary Stent Occlusion: Factors Involved and Possible Preventive Approaches Clin. Med. Res., March 1, 2007; 5(1): 53 - 60. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H.-J. Wagner With or Against the Bile Flow? Radiology, October 1, 2002; 225(1): 7 - 9. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
