|
|
| |||||||||||||||||||||||||||||||||||||||||
Background Esophageal obstruction due to cancer can produce debilitating dysphagia. Rapid palliation is usually possible with endoscopic placement of a plastic esophageal prosthesis, but this device has a high rate of complications. A new alternative is a metal-mesh stent that collapses to 3 mm in diameter at placement but can then expand up to 16 mm.
Methods Patients with esophageal carcinoma (39 patients) or malignant extrinsic obstruction (3 patients) were randomly assigned to treatment with either a plastic prosthesis (16 mm in diameter) or an expansile metal-mesh stent. The patients were evaluated every six weeks until death. The degree of palliation was expressed as a dysphagia score and a Karnofsky performance score.
Results Complications were significantly less frequent with the metal stents than with the plastic prostheses (no complications vs. nine; P<0.001). The dysphagia and Karnofsky scores improved significantly and to a similar degree in both treatment groups. The most common causes of recurrent dysphagia were migration of the plastic prostheses (five patients) and ingrowth or overgrowth of the metal stents by tumor (five patients). The rates of reintervention were similar in both treatment groups, as were the 30-day mortality rates. The period of hospitalization after placement of a prosthesis was significantly longer in the group given plastic prostheses than in the group given metal stents (mean ±SE, 12.5 ±2.1 vs. 5.4 ±1.0 days; P = 0.005). Despite their higher initial cost, the metal stents were cost effective because of the absence of fatal complications and the decrease in the hospital stay.
Conclusions Expansile metal stents are a safe and cost-effective alternative to conventional plastic endoprostheses in the treatment of esophageal obstruction due to inoperable cancer.
Laser therapy, radiotherapy, and conventional plastic prostheses have all been used for palliation. Radiotherapy achieves an acceptable degree of palliation of dysphagia in approximately 40 percent of patients,3,4 but it may take two months4. Laser therapy is not suitable for extrinsic compression, is difficult to perform if the narrowed region is long and tortuous, and requires repeated treatments with expensive equipment that is not always available. Esophageal prostheses provide rapid relief of dysphagia and are useful in patients in poor nutritional condition.
Recently, a new class of endoprostheses has become available that promises reduced complications. An expansile tube of metal mesh is compressed and then inserted while restrained on a delivery device of small diameter. When the stent is positioned in the stricture, the restraining device is removed and the mesh expands against the stricture. Metal stents can be easily inserted because their diameter is small. Since they expand to a large diameter, they offer the prospect of rapid palliation of symptoms. There have been several uncontrolled studies of successful insertion of metal stents for malignant esophageal obstruction5,6,7,8,9,10,11. We therefore compared a conventional plastic prosthesis with a metal stent in a randomized prospective trial.
Methods
Eligibility and Randomization
Adults with dysphagia due to inoperable malignant obstruction of the esophagus or the esophagogastric junction were eligible for the study. An obstruction was considered inoperable if he or she had local tumor extension, distant metastases, a recurrence after previous gastric or esophageal resection, or poor general health due to serious concomitant disease. Patients were also excluded if they had a tumor within 2 cm of the upper esophageal sphincter, a tracheoesophageal fistula, or an initial Karnofsky performance score below 30.
Forty-two consecutive patients (35 men and 7 women) with malignant obstruction of the esophagus or the esophagogastric junction who were seen at three hospitals between November 1991 and September 1992 (21 patients in Kassel, 13 in Bad Hersfeld, and 8 in Berlin) entered the study. Nine other patients were excluded: two had a tumor within 2 cm of the upper esophageal sphincter, four had a tracheoesophageal fistula, and three had an initial Karnofsky score below 30. The patients enrolled were randomly assigned to treatment with a plastic prosthesis (n = 21) or a metal stent (n = 21) according to a computer-generated random-number chart.
The study was conducted in accordance with the ethical standards set forth in the Declaration of Helsinki (1983). Informed consent was obtained from all patients.
Placement of the Devices
Plastic Prostheses
We used the largest conventional plastic prosthesis (Wilson-Cook, Winston-Salem, N.C.), which has an outer diameter of 16 mm and an inner diameter of 12 mm. Each stricture had to be dilated with a 20-mm balloon catheter (Rigiflex, Microvasive, Watertown, Mass.) before the prosthesis was placed. If the stenosis was very tight, we serially dilated the stricture with balloon catheters with diameters of 10, 15, and 20 mm in separate sessions before the day of placement. The prosthesis was placed with a Savary dilator-introducer passed over a guide wire (Amplatz-Extra Stiff, Wilson-Cook) under fluoroscopic monitoring. All patients required general anesthesia for placement of the plastic prosthesis.
Metal Stents
Because of the small diameter of the assembly with the metal stent, it is not necessary to dilate the stricture to the same extent as for a plastic prosthesis, and a variety of approaches can be used depending on the length and the degree of the stenosis. In our study, 14 patients underwent balloon dilation of their strictures (usually to a diameter of 10 mm) before stent placement, 3 patients had laser therapy, and 4 received no pretreatment before stent placement. The stents that we used (Wallstents, Schneider AG, Bulach, Switzerland) are 3 mm in diameter and 15 cm long when constrained; they expand to 16 mm in diameter and shorten to 10 cm when released. Each radiopaque stent was placed in the stricture under fluoroscopic monitoring, and the restraining membrane was retracted, releasing the stent in the stricture (Figure 1). Details of this type of stent and its release have been described elsewhere12. All patients were awake but sedated during stent placement by endoscope (GIF-XQ20 and GIF-PQ20, Olympus, Tokyo, Japan).
|
The patients were evaluated every six weeks until death. Follow-up telephone calls were alternated with outpatient visits. The general health of the patients was assessed with the Karnofsky score. The ability to swallow was expressed as a dysphagia score modified from the score of Mellow and Pinkas13: a score of 0 denoted the ability to eat a normal diet; 1, the ability to eat some solid food; 2, the ability to eat semisolids only; 3, the ability to swallow liquids only; and 4, complete dysphagia. Each prosthesis or stent was assumed to be functioning successfully if the dysphagia score decreased by at least one grade. The patients were not asked to avoid specific foods, but they were asked to eat slowly, to chew thoroughly, and to drink carbonated beverages after eating.
Cost-Effectiveness Analysis
Cost-effectiveness analysis was performed as described by Detsky and Naglie14. The retail cost of the metal stent used in this study was approximately $1,500, and that of the plastic prosthesis $130: the incremental cost of a metal stent was therefore $1,370. Cost effectiveness was determined according to two factors: the number of premature deaths that were direct results of complications of stent placement, and the number of days of hospitalization for recurrent dysphagia and treatment of complications.
Statistical Analysis
All analyses were performed on an intention-to-treat basis, and the results were expressed as means ±SE. Differences between the treatment groups were determined by the Mann-Whitney test and Fisher's exact test, as appropriate. Comparisons of the groups' dysphagia scores and Karnofsky scores relied on differences from base-line values. Differences in survival were tested with life-table analysis (Kaplan-Meier method). The outcomes of the two treatments were compared by the log-rank test.
Results
Characteristics of the Patients
The two treatment groups were similar in their clinical characteristics (Table 1). The majority of patients (88 percent) were followed until death. Three patients in the plastic-prosthesis group and two in the metal-stent group were still alive at the latest evaluation (minimal follow-up period, 6 and 6 1/2 months, respectively).
|
Prostheses or stents could be placed in 41 of the 42 patients. We were unable to insert a plastic prosthesis in one patient with a carcinoma of the gastric cardia and a very tight stricture, despite two attempts and serial dilation with balloon catheters. This patient underwent endoscopic gastrostomy. All metal stents were successfully inserted.
The overall 30-day mortality was 29 percent in the plastic-prosthesis group and 14 percent in the metal-stent group (P = 0.29) (Table 2). Three of the deaths in the plastic-prosthesis group were related to placement; all deaths in the metal-stent group were related to disease progression. The mean length of follow-up of patients surviving for more than 30 days was similar in the two groups. According to life-table analysis, there was no significant difference between the groups in survival (P = 0.35). There was a long-term improvement in dysphagia in both groups in all patients with lesions of the upper thoracic esophagus. Improvement occurred in the majority of patients with involvement of the lower esophagus and the cardia (91 percent in the plastic-prosthesis group and 92 percent in the metal-stent group). The dysphagia scores and the Karnofsky scores (determined every six weeks until death) improved significantly from base line in both groups; there was no significant difference between the groups in the magnitude of improvement from base line at any point.
|
Complications of device placement and functioning were significantly more frequent in the plastic-prosthesis group (P<0.001) (Table 3), resulting in a procedure-related mortality of 14 percent (two deaths due to perforation and one due to aspiration pneumonia). One patient with a perforation required a chest tube because of tension pneumothorax and died two days later. Another patient had pleural effusions and fever after prosthesis placement and died five days later despite antibiotic therapy; an autopsy showed rupture of the esophagus into the mediastinum. One patient had aspiration pneumonia after placement of the prosthesis and died five days later despite antibiotic therapy. One patient with a perforation survived for more than 30 days with conservative treatment.
|
Ingrowth of tumor through the mesh of the metal stent caused recurrent dysphagia in three patients. The ingrowth was treated by ablation of the tumor inside the stent with a laser. Tracheoesophageal fistulas developed in two patients with metal stents during the late stages of their disease; plastic prostheses (internal diameter, 10 mm in one patient and 12 mm in the other) were inserted through the stents. One of these patients had previously received laser therapy for tumor ingrowth; however, the fistula developed above the stent and did not pass through the mesh. In the other patient, the tracheoesophageal fistula passed through the mesh of the stent. Tumor overgrowth developed in two patients, at the distal end of the metal stent in one and at the proximal end in the other. The distal overgrowth was treated by implantation of a second, overlapping stent, and the proximal overgrowth by endoscopic laser ablation. Food-bolus impaction occurred twice in one patient with a plastic prosthesis and in three patients with metal stents, and was treated by endoscopic disimpaction. The frequency of recurrent dysphagia was similar in the two treatment groups; it occurred in seven patients (33 percent) in each. The mean number of reinterventions (Table 4) was slightly higher in the plastic-prosthesis group but not significantly so. The total number of days spent in the hospital for treatment of complications and dysphagia was higher in the plastic-prosthesis group.
|
Three patients died prematurely as a direct consequence of complications of placement of plastic prostheses. The incremental cost-effectiveness ratio for the prevention of these premature deaths was therefore $457 -- that is, metal stents cost $457 more per premature death avoided (incremental cost of $1,370/3 deaths). When the cost-effectiveness ratio was determined according to the number of days spent in the hospital for recovery from placement and for treatment of complications, the use of metal stents decreased the overall cost, provided the cost of hospitalization exceeded $193 a day.
Discussion
Palliation is often difficult to achieve in patients with esophageal obstruction due to cancer. We conducted a controlled trial in which patients received either an expansile metal-mesh stent or a conventional, rigid plastic prosthesis. Our results are remarkable for the absence of procedure-related complications in those who received the expansile metal stents. Such a low complication rate is unusual for insertion of esophageal prostheses. The success rate and perforation rate among the patients given plastic prostheses in our study are comparable to the rates in other recent studies15,16,17,18. Recurrent dysphagia due to device migration and blockage occurred in 33 percent of those who received plastic prostheses, a proportion slightly higher than in one previous report16 and lower than in another15. Long-term palliation of symptoms in this group was similar to that in previous studies15,16.
Metal stents are technically easier to place in the esophagus than in the biliary tree because the endoscope and delivery catheter used in placement are less angulated19. Less dilation is required than for plastic prostheses, reducing complications. Our protocol required in-hospital observation after placement of the metal stents because the nature and incidence of complications were unknown. Our experience suggests that these stents may be placed on an outpatient basis in many patients, which should further reduce costs. Since metal stents like those we used cannot be removed or repositioned after placement, great care must be taken to place them correctly.
Tumor ingrowth is a problem that is peculiar to metal stents because of the spaces between their struts. The extent of ingrowth is a function of the rate of tumor growth and the length of survival. The incidence of tumor ingrowth in this study was comparable to the incidence of ingrowth through metal stents placed for biliary obstruction20,21. Laser ablation of ingrowth significantly reduced dysphagia, as has been reported in another study11. Treatment of ingrowth was straightforward and could usually be performed on an outpatient basis. The rigid structure of the plastic prosthesis leads to a high rate of initial complications but does not permit tumor ingrowth. As the tumor mass increases, however, the prosthesis may become displaced. The resulting dysphagia is more difficult to treat than ingrowth and requires extraction and replacement of a prosthesis that has migrated. In contrast, a metal stent may become enmeshed in the tumor and continue to exert pressure against it, making migration unlikely.
Metal stents compare favorably with laser therapy for the palliation of dysphagia in malignant esophageal obstruction. In a recent study, the success rate of laser therapy was 77 percent initially and dropped to 63 percent when patients were followed until death16. The rate of serious complications of this treatment ranges from 4 to 9 percent16,18. Laser therapy requires three to five sessions for initial treatment and periodic (usually monthly) retreatment to maintain luminal patency,15 but repeated trips to the hospital are a serious burden to patients with a limited time to live16. As the disease progresses, outpatient laser therapy becomes increasingly difficult22. In one study, only 40 percent of patients could be treated in this way16. Radiotherapy and combined therapy (chemotherapy plus radiation) achieve effective palliation in patients with esophageal cancer, but the incidence of serious side effects is high. In one recent study,23 severe side effects occurred in 44 percent of patients receiving combination therapy, life-threatening side effects occurred in 20 percent, and the ability to swallow improved in only 58 percent. This is a serious shortcoming since the quality of life remaining to these patients depends to a large degree on their ability to swallow24. It is possible that combination therapy may be appropriate for some patients, with placement of a metal stent as initial therapy to relieve the dysphagia, rapidly followed by chemotherapy or radiation to decrease tumor mass and prolong patency of the stent. Promising results have been obtained with a combination of laser therapy and external-beam radiotherapy25.
Expandable metal stents are in a state of evolution. Early devices had higher complication rates19 than those used in this study (Wallstents). Possible developments include coating the metal mesh with a membrane to prevent tumor ingrowth,26 but preliminary studies indicate that migration of the stent or disruption of the membrane by tumor ingrowth may result10,27. In conclusion, metal stents are a cost-effective and safer alternative to conventional prostheses in the palliation of malignant esophageal obstruction.
Supported by a research study award from the Deutscher Akademischer Austauschdienst (to Dr. Vakil while in Germany).
Source Information
From Medizinische Klinik I (K.K.) and the Institut fur Rontgendiagnostik (H.-J.W.), Stadtische Kliniken, Kassel, Germany; Innere Abteilung IV, Krankenhaus Neukolln, Berlin, Germany (N.B.); the Medizinische Klinik, Kreiskrankenhaus, Bad Hersfeld, Germany (M.K.); and the University of Rochester Medical Center, Rochester, N.Y. (N.V.). Presented in part at the plenary session of the American Society of Gastrointestinal Endoscopy, Boston, May 19, 1993.
Address reprint requests to Dr. Knyrim at Med. Klin. I., Stadt. Kliniken, Monchebergstr. 41, 3500 Kassel, Germany.
References
| |||||||||||||||||||||||||||||||||||||||||
Related Letters:
Expansile Stents in Esophageal Cancer
Coia L. R., Ahmad N., Rosenthal S. A., Bennett J. R., Dakkak M., Vakil N., Knyrim K.
Extract |
Full Text
N Engl J Med 1994;
330:790-791, Mar 17, 1994.
Correspondence
Endoscopic Palliation of Malignant Dysphagia
McCray R. S., Wong R. C.K., Van Dam J.
Extract |
Full Text
N Engl J Med 1997;
336:294-295, Jan 23, 1997.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
