Report of the Clinical Protocol Committee: Development of ...

11 downloads 0 Views 155KB Size Report
Thomas Anthony, MD, Todd Baron, MD, Sebastiano Mercadante, MD,. Sylvan Green, MD, Dennis ...... Sebastian S, Johnston S, Geoghegan T,. Torreggiani W ...
Vol. 34 No. 1S July 2007

Journal of Pain and Symptom Management S49

Special Article

Report of the Clinical Protocol Committee: Development of Randomized Trials for Malignant Bowel Obstruction Thomas Anthony, MD, Todd Baron, MD, Sebastiano Mercadante, MD, Sylvan Green, MD, Dennis Chi, MD, John Cunningham, MD, FACP, FACG, Anne Herbst, MD, Elizabeth Smart, and Robert S. Krouse, MD, FACS Department of Surgery (T.A.), University of Texas, Southwestern Medical Center, and Veterans Affairs North Texas Health Care System (T.A.), Dallas, Texas; Gastroenterology and Hepatology Department (T.B.), Mayo Clinic-Rochester, Rochester, Minnesota; Oncology Department (S.M.), La Maddalena Cancer Center, Palermo, Italy; Department of Biometry (S.G.), Arizona Cancer Center, University of Arizona, Tucson, Arizona; Department of Surgery (D.C.), Memorial Sloan-Kettering Cancer Center, New York, New York; Division of Gastroenterology (J.C.) and Department of Surgery (A.H., R.S.K.), University of Arizona College of Medicine, Tucson, Arizona; National Coalition for Cancer Survivorship (E.S.), Silver Spring, Maryland; and Southern Arizona Veterans Affairs Health Care System (R.S.K.), Tucson, Arizona, USA

Abstract Malignant bowel obstruction (MBO) is a commonly encountered palliative care problem. There have been very few comparative trials in this area, and consequently there is very little clinical evidence upon which therapy can be rationally based. The purpose of this paper is to highlight the discussion and decision-making process that was undertaken by the Clinical Protocol Subcommittee during the development of a proposed clinical trial of best medical care versus surgical or endoscopic treatment for MBO. The development of the proposed clinical trials followed an orderly process. The first step taken was a discussion of a specific definition for MBO. Once agreed upon, this definition helped identify inclusion and exclusion criteria for the proposed trial. This was followed by an extensive literature review, which helped define both surgical and endoscopic approaches to MBO as well as what constituted best medical care. An extensive discussion was then undertaken concerning the best outcome measure of success for medical, surgical, and endoscopic interventions. All of the above steps culminated in two proposed protocols, one for MBO of the small intestine distal to the ligament of Treitz and a second for colonic obstructions. The small intestinal trial is designed to compare surgical intervention versus best medical care, whereas the colonic trial seeks to compare surgery with endoscopically-placed intraluminal stents coupled with best medical

Address reprint requests to: Thomas Anthony, MD, VA North Texas Health Care System, 5323 Harry Hines Blvd., Dallas, TX 75390-9155, USA. E-mail: [email protected] Ó 2007 U.S. Cancer Pain Relief Committee Published by Elsevier Inc. All rights reserved.

Accepted for publication: April 18, 2007.

0885-3924/07/$esee front matter doi:10.1016/j.jpainsymman.2007.04.011

S50

Anthony et al.

Vol. 34 No. 1S July 2007

care. J Pain Symptom Manage 2007;34:S49eS59. Ó 2007 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Key Words Malignancy, bowel obstruction, clinical trial

Introduction Malignant bowel obstruction (MBO) is a common palliative care problem for which there is no defined algorithm of care. A major goal of the International Conference on MBO was to develop the framework for clinical trials that would help address some of the questions surrounding this clinical dilemma, including the efficacy of operative and nonoperative forms of palliation. The two trials proposed by the Clinical Protocol Committee are based on location of the intestinal obstruction. The first trial intends to include patients with small bowel obstructions distal to the ligament of Treitz (LT) as well as multisite colonic blockages. These patients will be randomized to undergo either operative decompression or best medical management. The second trial will include patients with solitary obstructions of the colon and rectum, and patients will be randomized to undergo either surgical decompression or endoscopic stent placement. The following article describes the background for, and clinical issues relevant to, these proposed trials for treatment of MBO. The committee has addressed many of the major issues and come to reasonable consensus, including an algorithm for a future trial. This paper hopes to clarify the thought processes and decision making for the conclusions drawn related to these palliative care trials.

Definition of Malignant Bowel Obstruction and Development of Inclusion and Exclusion Criteria One of the more difficult tasks facing the International Conference on MBO and the Clinical Protocol Committee was deciding how to define MBO for a clinical trial, and secondarily specifying inclusion and exclusion criteria. Patients with underlying malignancy who are additionally suffering from signs and symptoms

of bowel obstruction represent a difficult diagnostic and therapeutic challenge. In distinction to nonmalignant bowel obstruction, there are multiple complicating factors that are unique to the individuals with underlying malignancy. These additional factors may be entirely or partially responsible for the clinical picture of MBO. Altered motility from tumor infiltration of the celiac and myenteric plexus, anticholinergic medication, opioids, and electrolyte imbalances due to malignancy (e.g., hypercalcemia, hypokalemia) can all contribute to a bowel obstruction-like pattern for patients with advanced cancer. These factors need to be considered when assessing a patient for the possibility of MBO. The committee felt that the first step toward developing a clinical trial was to specify the criteria that would be used to define patients for inclusion in the studies. In fact, variability in inclusion criteria and definitions is one of the major problems in the MBO literature, and one reason why there is such wide variation in findings. The committee’s definition criteria for MBO are outlined in Table 1. First, the patient must have clear clinical criteria of a bowel obstruction. This is based on history, physical examination, and radiological criteria. In addition, the committee felt it was important that each patient have a computed tomography (CT) scan of the abdomen to clearly document findings related to obstruction (dilated proximal bowel, a transition point, decompressed distal bowel, etc.) as well as either indirect or direct evidence of cancer-related findings, such as carcinomatosis, evidence

Table 1 Definition of Malignant Bowel Obstruction 1. Clinical evidence of a bowel obstruction (via history/ physical/radiographic examination). 2. Bowel obstruction beyond the LT. 3. Intra-abdominal primary cancer with incurable disease. 4. Non-intra-abdominal primary cancer with clear intraperitoneal disease.

Vol. 34 No. 1S July 2007

Randomized Trial Development for MBO

of metastatic disease, or ascites. Second, because the location of an MBO can determine the treatment options, the committee felt it necessary to concentrate on specific parts of the intestine. Obstructions proximal to the LT are commonly treated endoscopically, either with dilatation, stenting, ablation, or a combination of techniques. The committee felt that there was little controversy concerning the endoscopic approach as the primary treatment modality for obstructions in these locations. There was general agreement that treatment decisions became less clear for the obstructions beyond the LT. Therefore, obstructions beyond the LT became the focus of the proposed clinical trials. Third, prior documentation of an intraperitoneal cancer (e.g., colorectal, ovarian, or gastric cancer, carcinoid, etc.) and incurable disease would be required for inclusion in the study. Fourth, for the rare, but known patient with a primary nonintraperitoneal cancer (e.g., lung or breast cancer) and a bowel obstruction, there would have to be clear evidence, either via a previous procedure or radiological studies, of intraperitoneal metastases. Bowel obstruction is occasionally the first presentation of a patient with intra-abdominal malignancy. Retrospective and autopsy series have suggested that intestinal obstruction occurs in 5%e51% of ovarian cancer patients and in 10%e28% of gastrointestinal cancers.1 Obstruction is also thought to be more common among late-stage tumors. Despite the association with late-stage disease, many individuals presenting with obstruction can be cured, as can individuals with obstructions due to limited local recurrence of disease. Because individuals who present either with primary obstructing cancers or with limited local recurrence can potentially be treated for cure, it was deemed inappropriate for these patients to be included in a trial that included possible randomization to a noncurative treatment. For patients with MBO in the context of prior history of cancer treatment, the question frequently arises whether or not the patient has a recurrence of their cancer as the etiology for the obstruction. Bowel obstruction from malignancy often occurs as a slowly progressive process, with gradual worsening of symptoms over an extended period of time. Only very

S51

infrequently is emergency surgery required.2 It is widely recognized that not all patients with a prior history, or even current evidence, of intra-abdominal malignancy, who present with bowel obstruction, will have a malignant etiology for this obstruction. Adhesions from prior surgery, radiation enteritis, chronic ischemia, inflammatory bowel disease, and various infections are some examples of potential nonmalignant sources of mechanical obstruction. In these situations, the likelihood that an obstruction is due to a recurrence of malignancy is related to a number of factors, including the original stage of the malignancy, time elapsed since surgery, and the presence of other signs or symptoms of recurrent/metastatic disease. None of these historical factors are, however, sufficient to differentiate MBO from nonmalignant etiologies. Studies have estimated that benign causes of bowel obstruction will be present in 3%e48% of patients who present with obstruction after abdominal surgery for cancer.3e5 The clinical presentation of patients with nonmalignant obstruction and MBO is often indistinguishable. In the majority of patients, physical examination does not provide help in differentiating benign from malignant obstruction. It is frequently difficult to identify the source of an obstruction even with radiologic evaluations. Plain films can offer confirmatory evidence of bowel obstruction, but rarely provide enough information to define the source of bowel obstruction. CT and magnetic resonance imaging (MRI) have been used to help evaluate both the location and the likely etiology of bowel obstruction. Both CT and MRI are very good with respect to differentiation of partial and complete bowel obstruction.6,7 These studies have, however, been much less accurate in defining the etiology of the obstruction. With improving technology, there have been more encouraging preliminary reports concerning the accuracy of radiological differentiation of benign and malignant obstruction. Low and colleagues, for example, recently reported sensitivity, specificity, and accuracy of approximately 90% in identifying the etiology of the obstruction using gadolinium-enhanced MRI and criteria that included the presence of an obstructing mass, focal mural thickening, or evidence of disseminated abdominal tumor.8 At this point, however, neither CT nor MRI

S52

Anthony et al.

possesses a sufficient specificity to reliably differentiate benign from malignant causes of obstruction. Therefore, despite improvements in imaging technologies, intraoperative pathologic findings remain the gold standard for etiologic diagnosis. Given that preoperative definition of the etiology of an obstruction in the setting of prior intra-abdominal cancer is often impossible, requiring defining the underlying etiology of the obstruction prior to patient enrollment is impractical. However, since this trial focuses on patients with advanced incurable cancer, the etiology of the obstruction becomes a less important consideration. Indeed, it is reasonable to treat patients with nonmalignant obstructions nonoperatively because there is evidence that this is unlikely to cause complications.9,10 Randomization of patients will also provide some assurance of equal representation of both nonmalignant and malignant obstructions in each of the proposed treatment arms. On the basis of this reasoning, a consensus was obtained that clear documentation of an obstruction due to tumor was not necessary for patient inclusion. Proposed inclusion and exclusion criteria for a study are outlined in Table 2. Overall, these criteria reflect the desire to be as specific as possible to clearly identify patients for inclusion without making criteria so stringent as to hamper accrual to the trial.

Rationale for Treatment Algorithms for MBO Many treatments have been used in an attempt to palliate MBO that is distal to the LT. These treatments can be broadly classified as surgical, endoluminal, and medical. There is, at this time, no standard approach to choosing treatment for many presentations of MBO. The primary goal in performing the proposed clinical trials is to specifically assess two potential therapeutic options for treatment of MBO. Upon presentation to the hospital, the normal treatment course for a patient with a history of advanced cancer and signs and symptoms consistent with a bowel obstruction is to undergo intravenous (IV) hydration, be allowed nothing by mouth, and to undergo

Vol. 34 No. 1S July 2007

Table 2 Additional Study Inclusion and Exclusion Criteria Inclusion criteria 1. Compatible with definition of MBO. 2. Admission to hospital. 3. Patients must be candidates for an intra-abdominal procedure and the attendant anesthetic technique required. 4. Proxy identified and involved in consent process. 5. Age $18 years old. 6. Performance status (Karnofsky >60% one week prior to admission). Exclusion criteria 1. Patients in whom treatment for cure is still possible. 2. Patients with obstruction requiring emergency surgery, including sepsis, signs of perforation, or ‘‘acute’’ abdomen on examination. 3. Patients who are incapable of tolerating a surgical procedure based on clinical status due to cancer status or other underlying medical problems. 4. Actively dying based on defined criteria. 5. Bowel obstruction that resolves with supportive measures within 48 hours of admission. 6. Inability to obtain informed consent.

nasogastric tube decompression. A portion of patients will improve to the point of allowing resumption of oral intake with these standard interventions. Although there are no data regarding the percentage of patients likely to improve with simple supportive measures, the committee believes that a much greater percentage will actually require further treatment. Accordingly, the committee has proposed that after 48 hours of supportive treatment and with a lack of symptom resolution, consenting patients would undergo randomization. This would give physicians time to ensure that proper pre-study testing had been accomplished, to participate in discussions with patients and families concerning treatment and other end-of-life issues, and to obtain agreement concerning appropriateness of inclusion of the patient in the study by both the surgical and gastroenterological teams. Patients would be randomized to undergo surgical decompression versus either endoscopic stenting or best medical care depending on the location of their obstruction. Specifically, obstruction of the small bowel and multisite colon would be randomized to either surgical intervention or best medical care. Patients with solitary obstructions distal to the cecum would be candidates for randomization between surgical decompression and endoluminal stenting (see Figs. 1 and 2).

Vol. 34 No. 1S July 2007

Randomized Trial Development for MBO

S53

SMALL BOWEL ALGORITHM Patients with small bowel obstruction 48-72°

Eligible/Agrees to randomization

Eligible/Does not agree to randomization

Consent

Consent

Best Medical Careb

Surgerya

No Improvement 5-7d

Continued Best Medical Careb

Surgerya

Observation

Fig. 1. Flow diagram of proposed clinical trial for small bowel obstruction and multisite large bowel obstruction (asurgical therapy will be determined by operating surgeon at the time of exploration; bBest Medical Care as outlined in Table 3).

masses on physical examination, ascites, carcinomatosis (as opposed to isolated local recurrence), multiple points of obstruction, advanced disease, and poor clinical status all negatively influence the likelihood of successful surgical outcome.2,11 The main outcome measure for most of these studies, however, was survival, not improvement in quality of life (QOL). All of these poor prognostic factors are surrogates for disease volume or late time-course, which may portend poor survival, but do not necessarily equate with inability to palliate symptoms or QOL. Based on this

Treatment Options for MBO Surgical Treatment for MBO Surgical intervention is one of the most commonly used solutions for patients with an MBO. Surgical decision making must take into account all the standard considerations in performing major operations on potentially frail individuals, and additionally weigh the likelihood of successful palliation. There are no clear guidelines to help identify patients who are most likely to benefit from attempted surgical palliation of MBO. Studies have variously suggested that palpable intra-abdominal

LARGE BOWEL ALGORITHM Patients with large bowel obstruction 48-72°

Eligible/Agrees to randomization

Eligible/Does not agree to randomization

Consent

Consent

Surgerya

No Improvement Stent 5-7d Continued Best Medical Careb

Surgerya

Observation

Fig. 2. Flow diagram of proposed clinical trial for solitary large bowel obstruction (asurgical therapy will be determined by operating surgeon at the time of exploration; bBest Medical Care as outlined in Table 3).

S54

Anthony et al.

limited clinical information, our group felt that exclusion of any of these subgroups was unwarranted. When dealing with MBO, the type of operation performed is in large measure dictated by the intraoperative findings. Aggressive resection of the tumor to relieve the obstruction is usually unwarranted. The simplest surgical technique to restore intestinal continuity is often the best strategy. This may include creation of stomas, intestinal bypass, resection of short segment of bowel, or a combination of these techniques. With increasing laparoscopic expertise, it is also possible that many of these patients could be treated with a minimally invasive approach. Minimization of operating time and maximization of palliation often are competing issues, and the highest levels of surgical judgment need to be used to ensure the desired outcome. A surgical arm to a clinical trial of treatment for MBO will not be able to standardize surgery to a single option, at least at this time. Given the wide range of individual patient characteristics and the myriad surgical options, the committee felt it best not to recommend any standardized approach to surgical intervention. Small case series have suggested that the use of percutaneous decompressive procedures (either gastrostomy or jejunostomy) performed by interventional radiologists or endoscopists may be of benefit to patients with MBO.12,13 Although there was some dissension among committee members, it was felt that these types of procedures differed substantially when compared with traditional surgical procedures in terms of anesthetic requirement, morbidity, patient acceptance, and outcome. Therefore, these procedures, for the purpose of the trial, will not be considered options solely for the surgical arm of the trial, but can be used as an adjunctive treatment on either arm of this trial. There are a number of possible negative outcomes for patients undergoing attempted surgical correction of MBO, including potential morbidity, mortality, and failure to achieve palliation. The rates of surgical morbidity and mortality have been extremely variable. Morbidity rates from 0% to 90% and 30-day mortality rates from 0% to 41% have been reported in recent series.2,14,15 These rates suggest that imperfect surgical selection criteria are being

Vol. 34 No. 1S July 2007

applied to patients with MBO. Therefore, a secondary goal of the proposed study will be to develop criteria that will help better select patients for surgical intervention. Additionally, not all patients who are taken to the operating room for palliation of MBO will achieve this goal. Studies have suggested that between 32% and 80% of patients taken to the operating room will be successfully palliated.16,17 These studies are, however, hampered by their retrospective nature and the lack of uniform criteria for judging success. Endpoints that have been used include survival, tolerance of oral intake at discharge, overall QOL, ability to return home, and absence of recurrent obstruction.11 Even when these endpoints are met, a significant percentage of these individuals develop reobstruction (10%e50%), although the time to reobstruction remains poorly characterized.18,19 The permanence of palliation should, therefore, be considered as an additional endpoint for a trial of palliation. A recent Cochrane systematic review focusing on the role of surgery to treat MBO concluded that the role of surgery remains controversial.1 After evaluation of the available literature, the committee concurred with this assessment. Although the literature suggests that symptoms can be controlled in many individuals, the quality of this literature does not allow for assessment of selection bias or accurate assessment of potential negative outcomes, including morbidity, mortality, and rates of reobstruction.

Best Medical Care for MBO Up until 1985, surprisingly little information was available in support of nonoperative approaches to MBO. In that year, Baines and colleagues reported on the effectiveness of several nonoperative measures in palliating patients with MBO.20 Since that time, a number of studies have reported on a variety of additional nonoperative (largely pharmacologic) measures to treat MBO. Currently, palliative care practitioners use a combination of medications, usually including analgesics, antiemetics, and antisecretory agents. Best medical care for this study would entail choices from each of these medication groups (Table 3). In addition, treatments can include insertion of a percutaneous endoscopic gastrostomy tube (PEG)

Vol. 34 No. 1S July 2007

Randomized Trial Development for MBO

Table 3 Best Medical Care Antisecretory Somatostatin analoguesa Anticholinergics Corticosteroids Antiemetics Haloperidol Prochlorperazine Pain medications Morphine Methadone PEG/TPN per team decision a

Mandatory for Best Medical Care arm.

or total parenteral nutrition (TPN) at the discretion of the research teams. Analgesic Agents. Pain is a prominent complaint in patients with MBO. Pain may be characterized as either continuous or colicky in nature. There are very limited data that pertain to pain control in patients with MBO. Anecdotal evidence suggests that for continuous pain, opioids such as continuous IV morphine infusions are required.2 Due to the depressive effect on bowel motility, these agents may also relieve colicky pain. However, if colicky pain persists after opioid infusion, many experts advocate the addition of an anticholinergic agent, such as hyoscine butylbromide (HB).21 This agent is more useful than hyoscine hydrobromide because it does not cross the bloodbrain barrier and is, therefore, less sedating. One prospective, randomized trial has reported that both scopolamine butylbromide and octreotide (the longer-acting somatostatin analogue) had favorable effects on colicky and continuous pain in the setting of MBO.22 Antiemetic Agents. Two of the most debilitating symptoms for patients with MBO are nausea and vomiting. A recent systematic review of the effectiveness of antiemetic therapy in the setting of obstructed cancer patients found very variable rates of successful antiemetic therapy.23 Specifically, much higher rates of response were reported in uncontrolled trials (75%e93%) compared with randomized controlled trials (response rates 23%e36% for nausea and 18%e52% for vomiting). Empiric therapy was as effective as therapy based on the patient’s clinical picture. The evidence in

S55

favor of the use of many commonly used agents, such as haloperidol, prochlorperazine, cyclizine, and olanzepine, was characterized as weak to nonexistent. There was some evidence to support the use of metoclopramide (compared with placebo), levosulpride (a dopamine antagonist) versus metoclopramide, and serotonin antagonists compared with traditional antiemetics. The overall evidence in support of antiemetics was surprisingly weak.23 Steroids are often used in patients with MBO and are widely thought to have antiemetic properties. They appeared to be somewhat effective in resolution of obstruction, but there are no data concerning their effectiveness as antiemetics. Resolution of bowel obstruction with corticosteroids has been the subject of a Cochrane review. This review concluded that when only randomized trials were considered, there was a nonsignificant trend for the resolution of obstruction when corticosteroids were used (dexamethasone 6e16 mg/ day given intravenously).24 Antisecretory Agents. The antisecretory agent that has shown the most promise in the treatment of MBO is octreotide. Many of the negative consequences of bowel obstruction can be linked to gut hormones such as vasoactive intestinal polypeptide (VIP). Increased secretion of VIP during episodes of obstruction has been linked to increased bowel hyperemia, edema, luminal secretion, and nonproductive peristaltic activity.25,26 Octreotide is effective at suppressing VIP, along with several other stimulatory gut hormones. Octreotide can be administered subcutaneously or intravenously using either continuous or intermittent dosing schedules. In the bowel, octreotide decreases the secretion of water, sodium, and chloride. Octreotide also improves electrolyte and water absorption, suppresses stimulatory gut and pancreatic secretions, decreases bowel peristaltic activity, and reduces splanchnic blood flow. The general inhibitory effect of octreotide may, therefore, help break the cycle of secretion, distension, and bowel hyperactivity that complicates bowel obstruction.27 A limited number of studies of the effectiveness of octreotide in the treatment of MBO have been performed. Successful utilization of octreotide was reported in three cases series.28e30 Although differing doses were reported in each

S56

Anthony et al.

study, octreotide had a positive effect on symptoms. Vomiting was alleviated (usually within one hour), and nasogastric tubes were either removed or volumes aspirated were dramatically reduced. Importantly, almost no side effects were reported. Based on the successes reported in these case series, larger prospective trials were undertaken. Three prospective studies have been reported to date.22,31,32 The results of these studies verified the encouraging results noted in the original case series. Specifically, vomiting was controlled in 39/51 (76%) patients; if a nasogastric tube was used prior to instituting octreotide therapy, these tubes were most often removable or aspirated volumes significantly reduced. In a prospective, randomized, follow-up study comparing 0.3 mg/day dosing of octreotide with 60 mg/day of scopolamine butylbromide in 17 patients (seven inpatients and 10 home care patients) with inoperable MBO, both treatments were successful in diminishing pain and allowing nasogastric tube removal. The volume of secretion on Treatment Days 2 and 3 was significantly lower in the patients treated with octreotide.22 A second prospective randomized trial in 18 patients with inoperable obstruction compared a 0.3 mg/day dose of octreotide to 60 mg/day HB. This study found that octreotide was more effective at controlling gastrointestinal symptoms: patients treated with octreotide had fewer episodes of emesis and decreased intensity of nausea compared with HB-treated patients.31 A final larger randomized, controlled, double-blind trial of 68 patients, assigned patients to receive either chlorpromazine (15e25 mg/day) plus HB (60e80 mg/day) or octreotide (0.6e0.8 mg/ day). Nausea, vomiting, anorexia, and fatigue were all noted to be better for patients receiving octreotide.32 Based on the preliminary information presented above, the Clinical Protocol Committee felt that octreotide is a potential alternative to surgery in patients with MBO. Although the dose and duration of treatment varied among studies, the expert opinion was for a dose of 0.2 mg three times a day. It was felt that it was reasonable to expect a significant response by Day 5. It should be noted that a patient who develops signs and symptoms requiring operation or in whom there is no symptomatic

Vol. 34 No. 1S July 2007

improvement within 5e7 days on best medical care could cross over to the surgical intervention arm (see Fig. 1).

Endoluminal Stenting of Colonic Lesions for Palliation of MBO For obstructions located distal to the cecum, surgical treatment will be compared with endoluminal stenting. Experience with the use of expandable stents to palliate obstruction is accumulating. Expandable stents can be placed as proximal as the cecum and as distal as the rectum, as long as there is a margin of about 2 cm from the anus. In addition to the palliation of primary colon cancers, colorectal stents can be used to treat extrinsically compressing processes such as pelvic cancers, regional metastases, and lymphadenopathy. Contraindications to colorectal stent placement include known or suspected colonic ischemia or perforation, multiple sites of obstruction, and standard endoscopy contraindications. The presence of any of these factors will preclude patient entry into the trial. Comprehensive Reviews of Colonic Stenting. The efficacy and safety of colorectal stents have been systematically reviewed. Clinical success in the palliative stent patients was achieved in 302/336 (90%). Complications included perforation, migration, reobstruction, bleeding, and pain. Interventional mortality rate was 1% (3/598). Stent migration was observed in 10% of patients (54/551) and was managed with stent removal or replacement or operation. Reobstruction resulted from tumor ingrowth (62%; 32/52), fecal impaction (25%; 13/52), and migration (13%; 7/52) and was more common among palliative patients. Bleeding occurred in 5% (27/525). In most cases, bleeding resolved spontaneously within 24 hours.33 A pooled analysis by Sebastian et al. evaluated self-expanding metal stents for the management of patients with malignant colorectal obstructions.34 In palliative patients (n ¼ 791), clinical success was defined as colonic decompression within 48 hours without need for additional intervention. The technical success rate in palliative patients was 93%. Cumulative technical and clinical success rates for palliated patients were 93% and 91%, respectively.34 Most cases of reobstruction can

Vol. 34 No. 1S July 2007

Randomized Trial Development for MBO

be treated endoscopically. Based on the encouraging rate of palliation and the low rate of complications, it was felt that this would be an appropriate treatment option to compare with surgical intervention for patients with colonic obstructions.

Role of an Observational Arm There are many publications related to MBO, but there is little known about the natural history for these patients. Although there is clear equipoise related to the major treatments for many patients with MBO, making an RCT important, it also would be beneficial to include an observational arm to such a study. There are many patients who may not be eligible for this study but who could still provide important information. This may include patients who are ineligible due to poor overall status, those patients wishing no further treatment, or those patients with a strong preference for a particular treatment option (i.e., those not wishing to be randomized). The committee felt that following these patients could yield useful information. Information concerning treatment variability, treatment-related morbidity, survival, and QOL could still be obtained albeit in an uncontrolled manner. This cohort database would be reflective of current standard practice and provide an additional comparison group for the treatments under study.

Statistical Considerations The statistical approaches for this study are based on multiple considerations. More importantly, an adequate primary outcome measure must be devised to clearly delineate benefit or equality from the two treatment arms. As determined during the conference, a composite measure of QOL outcomes that can be clearly documented should be used as the primary outcome measure. This would be the number of ‘‘good days,’’ defined as days out of the hospital without a nasogastric tube or hydration. This would mean that the patients are likely able to take in enough by mouth without needing additional IV supplementation, as well as were not living with problems related to nausea and vomiting. The number of patients necessary for such

S57

a study would also depend on the meaningful difference in ‘‘good days.’’ The median survival of patients with MBO is approximately 10 weeks, and therefore, it was determined that a reasonable difference would be two weeks. If one can consider such a limited life span, the conference participants reasoned that two weeks of improved or alleviated symptoms represent an important difference in outcome. The null hypothesis would be that the two treatment groups are equivalent (difference in number of good days is zero). Available data from Memorial Sloan-Kettering Cancer Center show a mean number of good days (in 10 weeks) of 33 and standard deviation of 25. To have 90% power to detect a mean difference between treatment groups of 10 good days would require 133 patients per group, based on a t-test at the twosided 0.05 significance level. This corresponds to an effect size of 0.4. All patients who are randomized would be included in the analysis using an intention-to-treat approach. Note that for detecting a difference of 14 days or greater (with 90% power), the required sample size would be approximately 70 patients per group.

Summary MBO is a common problem that is faced by clinicians, patients, and families. Preliminary evidence suggests that there are various approaches to the malady, especially depending on the location of obstruction. No comparative trials have been performed to date that have assessed the potential role of these therapeutic approaches in an unselected population of patients with MBO. This paper outlines proposed clinical trials designed to compare different approaches for care of patients facing this problem. As with the vast majority of issues in palliative care, there is scant clinical evidence to guide therapy for these challenging patients. Current therapy is based on empiric decision making, resulting in highly variable applications and outcomes. Progress will only be possible through the conduct of carefully controlled, prospective, randomized trials that are carefully analyzed and reported. A trial comparing therapies for MBO will help to define effective therapy, identify selection criteria,

S58

Anthony et al.

and serve as a model for future studies in palliative care.

References 1. Feuer DJ, Broadley KE. Surgery for the resolution of symptoms in malignant bowel obstruction in advanced gynaecological and gastrointestinal cancer. [Cochrane Review]. In: The Cochrane Library, Vol. 4. Chichester, UK: John Wiley & Sons, Ltd., 2004. 2. Ripamonti C, Bruera E. Palliative management of malignant bowel obstruction. Int J Gynecol Cancer 2002;12:135e143. 3. Ketcham AS, Hoye RC, Pilch YH, Morton DL. Delayed intestinal obstruction following treatment for cancer. Cancer 1970;25(2):406e410. 4. Osteen RT, Guyton S, Steele G Jr, Wilson RE. Malignant intestinal obstruction. Surgery 1980; 87(6):611e615. 5. Legendre H, Vahhuyse F, Caroli-Bose FX, Pector JC. Survival and quality of life after palliative surgery for neoplastic gastrointestinal obstruction. Eur J Surg Oncol 2001;27:364e367. 6. Maglinte DD, Heitkamp DE, Howard TJ, Kelvin FM, Lappas JC. Current concepts in imaging of small bowel obstruction. Radiol Clin North Am 2003;41(2):263e283. 7. Delabrousse E, Destrumelle N, Brunelle S, et al. CT of small bowel obstruction in adults. Abdom Imaging 2003;28(2):257e266. 8. Low RN, Chen SC, Barone R. Distinguishing benign from malignant bowel obstruction in patients with malignancy: findings at MR imaging. Radiology 2003;228(1):157e165. 9. Bastounis E, Hadjinikolaou L, Ioannou N, et al. Somatostatin as adjuvant therapy in the management of obstructive ileus. Hepatogastroenterology 1989;36:538e539. 10. Mercadante S, Avola G, Maddaloni S, et al. Octreotide prevents the pathological laterations of bowel obstruction in cancer patients. Support Care Cancer 1996;4:393e394. 11. Krouse RS. Surgical management of malignant bowel obstruction. Surg Oncol Clin N Am 2004; 13:479e490. 12. Cannizzaro R, Bortoluzzi F, Valentini M, et al. Percutaneous endoscopic gastrostomy as a decompressive technique in bowel obstruction due to abdominal carcinomatosis. Endoscopy 1995;27(4): 317e320. 13. Cunningham MJ, Bromberg C, Kredentser DC, Collins MB, Malfetano JH. Percutaneous gastrostomy for decompression in patients with advanced gynecologic malignancies. Gynecol Oncol 1995; 59(2):273e276.

Vol. 34 No. 1S July 2007

14. Higashi H, Shida H, Ban K, et al. Factors affecting successful palliative surgery for malignant bowel obstruction due to peritoneal dissemination from colorectal cancer. Jpn J Clin Onc 2003;33(7): 357e359. 15. Yazdi GP, Miedema BW, Humphrey LJ. High mortality after abdominal operation in patients with large-volume malignant ascites. J Surg Oncol 1996;62(2):93e96. 16. Turnbull AD, Guerra J, Starnes HF. Results of surgery for obstructing carcinomatosis of gastrointestinal, pancreatic, or biliary origin. J Clin Oncol 1989;7(3):381e386. 17. Lund B, Hansen M, Lundvall F, et al. Intestinal obstruction in patients with advanced carcinoma of the ovaries treated with combination chemotherapy. Surg Gynecol Obstet 1989;169(3):213e218. 18. Solomon HJ, Atkinson KH, Coppleson JV, et al. Complications in the management of ovarian cancer. Aust N Z J Obstet Gynaecol 1983;23(2):65e68. 19. Phillip J, Do T, Hacker N, Grant P, Lickiss N. Outcomes in patients with GIT obstructiond62 episodes in 33 patients in a 12 month period. [abstract]. Int J Gynecol Cancer 1997;7(Supp 2):86. 20. Baines M, Oliver DJ, Carter RL. Medical management of intestinal obstruction in patients with advanced malignant disease. Lancet 1985;2(8462): 990e993. 21. Ripamonti C, Twycross R, Baines M, et al. Clinical-practice recommendations for the management of bowel obstruction in patients with end-stage cancer. Working Group of the European Assiciation for Palliative Care. Support Care Cancer 2001;9(4): 223e233. 22. Ripamonti C, Mercadente S, Groff L, et al. Role of octreotide, scopolamine butylbromide and hydration in symptom control of patients with inoperable bowel obstruction and nasogastric tubes: a prospective randomized trial. J Pain Symptom Manage 2000;19:23e34. 23. Glare P, Pereira G, Kristjanson LJ, Stockler M, Tattersall M. Systematic review of the efficacy of antiemetics in the treatment of nausea in patients with far-advanced cancer. Support Care Cancer 2004; 12(6):432e440. 24. Feuer DJ, Broadley KE. Corticosteroids for the resolution of malignant bowel obstruction in advanced gynaecological and gastrointestinal cancer. [Cochrane Review]. In: The Cochrane Library, Vol. 4. Chichester, UK: John Wiley & Sons, Ltd., 2004. 25. Basson MD, Fielding LP, Bilchik AJ, et al. Does vasoactive intestinal polypeptide mediate the pathophysiology of bowel obstruction? Am J Surg 1989; 157(1):109e115. 26. Neville R, Fielding LP, Cambria RP, Modlin I. Vascular responsiveness in obstructed gut. Dis Colon Rectum 1991;34(3):229e235.

Vol. 34 No. 1S July 2007

Randomized Trial Development for MBO

27. Ripamonte C. How to use octreotide for malignant bowel obstruction. J Support Oncol 2004;2(4): 1e8. 28. Mercadante S, Maddaloni S. Octreotide in the management of inoperable gastrointestinal obstruction in terminal cancer patients. J Pain Symptom Manage 1992;7(8):496e498. 29. Khoo D, Riley J, Waxman J. Control of emesis in bowel obstruction in terminally ill patients. Lancet 1992;339(8789):375e376. 30. Steadman K, Franks A. A woman with malignant bowel obstruction who did not want to die with tubes. Lancet 1996;347(9006):944. 31. Mercadante S, Ripamonti C, Casuccio A, Zecca E, Groff L. Comparison of octreotide and hyoscine butylbromide in controlling gastrointestinal

S59

symptoms due to malignant inoperable bowel obstruction. Support Care Cancer 2000;8:188e191. 32. Mystakidou K, Tsilika E, Kalaidopoulou O, et al. Comparison of octreotide administration vs. conservative treatment in the management of inoperable bowel obstruction in patients with far advanced cancer: a randomized, double-blind, controlled clinical trial. Anticancer Res 2002;22(2B):1187e1192. 33. Khot UP, Lang AW, Murali K, Parker MC. Systematic review of the efficacy and safety of colorectal stents. Br J Surg 2002;89:1096e1102. 34. Sebastian S, Johnston S, Geoghegan T, Torreggiani W, Buckley M. Pooled analysis of the efficacy and safety of self-expanding metal stenting in malignant colorectal obstruction. Am J Gastroenterol 2004;99:2051e2057.