Malignant bowel obstruction (MBO) occurs frequently. Original Articles

Transcription

1 JOURNAL OF PALLIATIVE MEDICINE Volume 14, Number 7, 2011 ª Mary Ann Liebert, Inc. DOI: /jpm Original Articles Management of Patients with Malignant Bowel Obstruction and Stage IV Colorectal Cancer Kimberly Moore Dalal, M.D., 1,5,6 Marc J. Gollub, M.D., 2 Thomas J. Miner, M.D., 3 W. Douglas Wong, M.D., 1 Hans Gerdes, M.D., 4 Mark A. Schattner, M.D., 4 David P. Jaques, M.D., 1 and Larissa K.F. Temple, M.D. 1 Abstract Background: Malignant bowel obstruction (MBO), a serious problem in stage IV colorectal cancer (CRC) patients, remains poorly understood. Optimal management requires realistic assessment of treatment goals. This study s purpose is to characterize outcomes following palliative intervention for MBO in the setting of metastatic CRC. Study Design: Retrospective review of a prospective palliative database identified 141 patients undergoing surgical (OR; n ¼ 96) or endoscopic (GI; n ¼ 45) procedures for symptoms of MBO. Results: Median patient age was 58 years, median follow-up 7 months. Most (63%) had multiple sites of metastases. Computed tomography (CT) scan findings of carcinomatosis ( p ¼ 0.002), ascites ( p ¼ 0.05), and multifocal obstruction with carcinomatosis and ascites ( p ¼ 0.03) significantly predicted the need for percutaneous or open gastrostomy tube, or stoma. Procedure-associated morbidity for 81 patients with small bowel obstruction (SBO) was 37%; 7% developed an enterocutaneous fistula/anastomotic leak. Thirty-day mortality was 6%. Most (84%) patients were palliated successfully; some received additional chemotherapy (38%) or surgery (12%). Procedure-associated morbidity for 60 patients with large bowel obstruction (LBO) was 25%; 11 patients (18%) required other procedures for stent failure, with one death at 30 days. Symptom resolution was >97%. Patients with LBO had improved symptom resolution, shorter length of stay (LOS), and longer median survival than patients with SBO. Conclusions: Patients with MBO and stage IV CRC were successfully palliated with GI or OR procedures. Patients with CT-identified ascites, carcinomatosis, or multifocal obstruction were least likely to benefit from OR procedures. CT plays an important role in preoperative planning. Sound clinical judgment and improved understanding are required for optimal management of MBO. Introduction Malignant bowel obstruction (MBO) occurs frequently in stage IV colorectal cancer (CRC) patients. 1 Although median survival has improved in patients with metastatic CRC, long-term survival for those with MBO is only 4 to 9 months. 2 This problem remains poorly understood, and careful evaluation of individual treatment goals is required to achieve optimal management. Surgical palliation is best defined as an operative procedure done primarily for noncurative relief of symptoms caused by advanced malignancy. A recent prospective study of more than 1000 palliative surgical procedures concluded that relief is possible in select patients, but new or recurrent symptoms limit its durability. 3 Identification of patients with metastatic CRC who may benefit from surgery for MBO is difficult. Currently, insufficient data exist to guide the clinician or facilitate shared decision making. The goal of this study was to evaluate management of MBO in patients with metastatic CRC. Although small bowel obstruction (SBO) and large bowel obstruction (LBO) are clinically different forms of obstruction, both are typical morbid complications of end-stage CRC and, a posteriori (from the patient s point of view) are experienced in much the same 1 Department of Surgery, 2 Department of Radiology, 4 Division of Gastroenterology, Memorial Sloan-Kettering Cancer Center, New York, New York. 3 Department of Surgery, Warren Alpert Medical School, Brown University, Rhode Island Hospital, Providence, Rhode Island. 5 Department of Surgery, David Grant United States Air Force Medical Center, Fairfield, California. 6 Department of Surgery, University of California-San Francisco Medical Center, San Francisco, California. Accepted February 18,

2 MALIGNANT BOWEL OBSTRUCTION STAGE IV CRC 823 way. Therefore, patients with either SBO or LBO were included in this study. By assessing overall morbidity of treatment, success of palliation (defined as symptom relief, restoration of ability to eat, discharge home, durability of relief), and associated patient-specific factors, we believe these data provide important information to physicians and their patients who are confronting either SBO or LBO in the setting of metastatic CRC. Methods One hundred forty-one patients with metastatic CRC and MBO, treated palliatively at Memorial Sloan-Kettering Cancer Center (MSKCC) between January 1, 2000 and June 30, 2005, were identified through the hospital billing database ( ) and the prospectively maintained surgical palliative procedures database ( ). 4 Patients were included in the study if their charts described planned interventions as having palliative intent. Individuals were excluded if they were younger than 18, had resectable metastases, or had appendiceal or anal pathology. Approval for this study was obtained from the MSKCC Institutional Review Board. The charts of patients meeting inclusion criteria were retrospectively reviewed for patient, tumor, and treatment factors. MBO was defined as either SBO (i.e., located between the ligament of Treitz and the ileocecal valve), or LBO (i.e., located between the cecum and the rectum). Two patients with both SBO and LBO were categorized as SBO. Each computed tomography (CT) scan performed within 14 days of a palliative procedure was reviewed by one gastrointestinal radiologist (MG), who was blinded to management of the MBO. CT scans were evaluated for location and number of sites of obstruction, presence of carcinomatosis, and ascites. Interventions were classified as surgical (OR) including exploratory laparotomy, surgical placement of gastrostomy (open g)-tube, stoma, intestinal bypass, bowel resection, lysis of adhesions (LOA); or as endoscopic (GI), including percutaneous endoscopic gastrostomy (PEG) or rectal stent. The outcomes of palliative interventions were evaluated. Surgical complications within 30 days of operation were graded using a previously described system, 5 and 30-day mortality was assessed. The highest level of severity was recorded if a patient suffered more than one procedure-related complication. Hastening of death was defined as death within 30 days resulting from postoperative complications. Length of hospital stay and discharge location (home, skilled nursing facility) were computed. Success of palliation was determined by evaluating symptom resolution and ability to eat at discharge. Additional surgical, medical, or radiation therapies given to manage recurrent or new symptoms at 30 and 100 days were recorded. Patients were followed until death. Statistical methods Descriptive statistics were performed. Patients with MBO were evaluated for symptom resolution based on site of obstruction (SBO versus LBO) and type of intervention (GI versus OR). Clinicopathologic correlates were analyzed using the w 2 or Fisher s exact test, where appropriate. For continuous variables, differences between means were analyzed with the Student s t test. A priori, analyses were planned to address two clinical issues: 1) Can we predict who will require a g-tube/stoma? 2) Can we predict who will be able to sustain nutrition after treatment? Clinical, tumor, CT, and treatment factors were assessed. Univariate comparisons were made using w 2 or Fisher s exact test, where appropriate. Independently associated factors were identified by proportional hazards regression analysis (Cox model). Analyses were performed using the SPSS Statistical Package (SPSS Inc, Chicago, IL). Differences reaching p ¼ 0.05 were considered significant. Results Patient characteristics From January 1, 2000 to June 30, 2005, 141 stage IV CRC patients with MBO were treated operatively or endoscopically. A total of 79 patients had SBO, 60 had LBO, and 2 had both (patients with both SBO and LBO were classified as SBO patients; thus, the SBO cohort ¼ 81). Median follow-up was 7.1 months; overall median survival was 8 months. Patient cohort Table 1 describes the cohort at time of presentation with MBO. The median patient age was 58; 55% were male. The Table 1. Cohort at Time of Palliative Procedure for Malignant Bowel Obstruction (MBO) N ¼ 141* Patient characteristics Age, y, median, range 58 (29 89) Gender (M/F), n (%) 77 (55%) M/64 (45%) F Weight loss >10 lbs in 74 (53%) prior 90 days, n (%) ECOG status, median 1.0 ( 0.8) Albumin, median (g/dl) 3.3 ( 1.2) Hemoglobin, median (g/dl) 10.5 ( 2.0) Site of primary tumor Right colon 42 (30%) Left colon 71 (50%) Rectum 27 (19%) Not specified 1 (0.7%) Prior therapy (any) Chemotherapy 110 (78%) Radiotherapy 44 (31%) Surgery 97 (69%) Characteristics of obstruction Number of metastatic sites** One 52 (37%) Two 50 (36%) Three or more 38 (27%) CT findings at obstruction*** Multiple sites of bowel 28 (25.4%) obstruction Carcinomatosis 46 (41.8%) Ascites 57 (51.8%) Multifocal þ 16 (14.5%) carcinomatosis þ ascites *Two with small bowel obstruction (SBO) and large bowel obstruction (LBO). **One with locally advanced disease. ***Ten patients did not have preoperative CT scans (7 SBO, 3 LBO) and 21 patients (1 SBO, 20 LBO) had CT scans >14 days pre-surgery. ECOG, Eastern Cooperative Oncology Group.

3 824 MOORE DALAL ET AL. majority had a primary colon cancer (80%; 42 right-sided, 71 left-sided) and more than one site of metastatic disease (88, 63%). Most patients had previously received treatment for their colorectal cancer; only 44 (31%) had not undergone surgery. CT scans were done in 110 patients within 14 days of presentation of MBO; these most commonly showed ascites (51.8%) and carcinomatosis (41.8%). Multifocal sites of obstruction, carcinomatosis, and ascites were more frequently seen separately and together in SBO versus LBO (21% versus 2.5%, 0.007). Of the 8 patients with adhesion-related obstructions, 6 had CT scans within 14 days of surgery and only 2 were identified on preoperative CT. Symptoms at presentation are outlined in Table 2. These commonly included nausea and/or vomiting (85%), abdominal pain (76%), or altered gastrointestinal function such as constipation, distension, or bloating (76%). SBO patients were more likely to experience nausea/vomiting (98% versus 67%, p < 0.001); LBO patients were more likely to have abdominal pain (73% versus 80%, p ¼ ns). Procedures Palliative procedures are shown in Table 3. A total of 46 patients (32%) with MBO were treated endoscopically. Endoscopic management was more common in patients with LBO than in those with SBO (47% versus 22%, p ¼ 0.004). A total of 8 patients had MBO secondary to adhesions, and were treated with definitive LOA. The majority of patients who went to the operating room had a definitive procedure (93%, 88/95). A total of 19/96 (20%) patients who went to the operating room had just a stoma. The majority of gastrostomy tubes were inserted by GI (25/31). These were generally for the purpose of palliation (74%, 23/31) only, and were utilized more often in patients with SBO than LBO (36% versus 3%, p < 0.001). Among patients with ascites who were palliated with PEGs, no major ascites leaks were seen. Technical success, defined as no anastomotic leak or need for additional procedure(s) to facilitate eating within 90 days after the procedure, was achieved in 82% (SBO 84% [100%, 18/18 GI PEG; 79%, 49/62 OR]; LBO 80% [59%, 48/60 GI stent; 97%, 32/33 OR]). Palliative procedure utcome Symptom relief of nausea, vomiting, pain. Resolution of symptoms was excellent for all MBO patients (Table 2). Nausea and vomiting resolved in 94% (112/119) within 30 days. Preoperative pain was relieved in 85% (91/107). In SBO patients, 91% experienced relief of nausea/vomiting and 80% reported pain resolution. In LBO patients, 100% experienced relief of nausea/vomiting and 92% reported pain resolution. Restoration of ability to eat. A total of 114 patients (84%) were able to eat at 30 days after discharge. Patients with LBO versus SBO (97% versus 74%, p ¼ 0.001) were more likely able to eat. However, patients treated solely with gastrostomy tubes, although able to eat, were incapable of sustaining nutrition independently (14/22, 64%). Most patients treated operatively (88%, 53/60) particularly those who underwent bypass, stoma creation, resection, or LOA were able to sustain nutrition (94%). Table 2. Efficacy and Durability of Symptom Palliation by GI or OR Procedures for Malignant Bowel Obstruction (MBO) Symptoms Symptom resolution in patients experiencing symptom* New symptoms N (%) Recurrent symptoms N (%) Pain resolved % (n) Resume oral nutrition % (n) Vomiting resolved % (n) Pain %(n) Unable to take PO % (n) Nausea/ Vomiting % (n) Overall (n ¼ 141) 85% (119) 85% (119) 76% (107) 94% (112/119) 96% (114/119) 85% (91/107) 30 d: 18 (13%) 30 d: 25 (18%) 100 d: 11 (8%) 100 d: 18 (13%) SBO (n ¼ 81) 98% (79) 98% (79) 73% (59) 91% (72/79) 74% (57/77) 80% (47/59) 30 d: 10 (12%) 30 d: 18 (22%) 100 d: 4 (5%) 100 d: 11 (14%) 94% (17) 94% (17) 78% (14) 76% (13/17) 18% (3/17) 57% (8/14) 30 d: 0 30 d: 1 (6%) 100 d: 1 (6%) 100 d: 1 (6%) GI (percutaneous g tube; PEG; n ¼ 18) OR (g-tube only; n ¼ 5) 100% (5) 100% (5) 60% (3) 100% (5/5) 60% (3/5) 33% (1/3) 30 d: 0 30 d: 1 (20%) 100 d: d: 0 OR (n ¼ 58) 98% (57) 98% (57) 72% (42) 95% (54/57) 93% (51/55) 90% (38/42) 30 d: 10 (17%) 30 d: 16 (28%) 100 d: 3 (5%) 100 d: 10 (17%) LBO (n ¼ 60) 67% (40) 67% (40) 80% (48) 100% (40/40) 95% (38/40) 92% (44/48) 30 d: 8 (13%) 30 d: 7 (12%) 100 d: 7 (12%) 100 d: 7 (12%) GI (stent; n ¼ 27) 52% (14) 52% (14) 85% (23) 100% (14/14) 93% (13/14) 83% (19/23) 30 d: 6 (22%) 30 d: 3 (11%) 100 d: 6 (22%) 100 d: 1 (4%) OR (n ¼ 33) 79% (26) 79% (26) 76% (25) 100% (26/26) 96% (25/26) 100% (25/25) 30 d: 2 (6%) 30 d: 4 (12%) 100 d: 1 (3%) 100 d: 6 (18%) *Denominator may be less than patients with symptom due to missing information in patient charts. LOS, length of stay; SBO, small bowel obstruction; GI, endoscopic procedure; PEG, percutaneous endoscopic gastrostomy; OR, operative procedure; LBO, large bowel obstruction, PO, per os (orally).

4 MALIGNANT BOWEL OBSTRUCTION STAGE IV CRC 825 Table 3. Palliative Procedures for Patients with Small Bowel Obstruction (SBO) or Large Bowel Obstruction (LBO) MBO (n ¼ 141) SBO (n ¼ 81) LBO (n ¼ 60) Procedure N (%) N (%) N (%) GI G-tube 19 (13%) 18 (22%) 1 (2%) Colorectal stent 27 (19%) 27 (45%) OR Resection only 34 (24%) 18 (22%) 16 (27%) Definitive lysis 8 (6%) 8 (10%) of adhesions Bypass only 21 (15%) 19 (23%) 2 (3%) G-tube þ bypass 6 (4%) 6 (7%) þ/ stoma Stoma 19 (13%) 6 (7%) 13 (22%) G-tube only 6 (4%) 5 (6%) 1 (2%) Ex lap only 1 (1%) 1 (1%) MBO, malignant bowel obstruction; SBO, small bowel obstruction; LBO, large bowel obstruction; GI, endoscopic procedure; PEG, percutaneous endoscopic gastrostomy; OR, operative procedure; Ex lap, exploratory laparotomy. Durability of symptom relief. Thirty percent (42/141) developed new symptoms requiring additional procedures and/or readmission. At 30 (22%) and 100 (15%) days, SBO patients had returned to the hospital for infectious processes requiring percutaneous drains, or for acute renal failure or progression of disease. At 30 (8%) and 100 (12%) days, some LBO patients had developed new symptoms, requiring subsequent procedures or readmission for fever, pain, acute renal failure, or progression of disease. Morbidity and mortality. A total of 45 patients (30%) had 59 complications (Table 4). Ten percent experienced grade 3 or 4 complications. Morbidity was higher in SBO than in LBO patients (37% versus 25%; p ¼ ns). Morbidity after OR was most common in SBO patients secondary to fistula, whereas morbidity after GI was most common in LBO patients secondary to stent failure. Thirty-day postoperative mortality was 4% (6/141) (Table 4). Two patients died from progression of disease. Four others had postoperative complications (sepsis, anastomotic leak, multisystem organ failure, myocardial infarction with sepsis) causing hastening of death. Length of stay (LOS), discharge home. Overall median post-procedure LOS was 8 days (SBO 9 days, LBO 5 days) (Table 4). Operative procedures resulted in 1- to 2-week longer median LOS than endoscopic procedures. For patients without complications, median LOS was 8 days; for those with grade 3 5 complications, median LOS was 16 days. Ninety percent (123/136) of patients with MBO were discharged home. Eighty-eight percent of SBO patients (68/77) were discharged home, 5 with home hospice; 12% were discharged to a nursing facility; and 2 expired before discharge. SBO patients receiving gastrostomy tubes (GI or OR) were less likely to be discharged home than patients undergoing OR procedures. All LBO patients treated with stents and 88% undergoing operative procedures were discharged home. Table 4. Postoperative Outcomes: Length of Stay, Morbidity, and Mortality Overall median LOS 8 days % discharged home 123/136 (90%) Morbidity 0 (none) 98 (70) 1 (local or bedside) 9 (6) 2 (invasive monitoring or 12 (9) IV medication) 3 (operation, IR, intubation, 13 (9) or therapeutic endoscopy) 4 (persistent disability 2 (1) or major organ resection) 5 (death) 6 (4) Technical failure 18% (25/141) SBO 13 Enterocutaneous fistula 6 Failed bypass -> PEG 4 Delayed gastric emptying -> 1 gastrojejunostomy and feeding jejunostomy Exploratory laparotomy -> PEG 2 LBO 12 Failed stent -> resection/ 11 stoma (1), stoma (1), PEG (4), stent (4), dilation (1) Internal hernia 1 IV, intravenous; SBO, small bowel obstruction; PEG, percutaneous endoscopic gastrostomy; Ex lap, exploratory laparotomy; LBO, large bowel obstruction; IR, interventional radiology. Adjuvant therapy, survival. Following the initial palliative procedure, 50% (71/141) received palliative chemotherapy and 4% (6/141) received radiation therapy. SBO patients undergoing OR procedures were more likely to receive chemotherapy (33%) than those having PEG (4%) or open g-tube (1%) placement. In LBO patients, 67% had chemotherapy, 3% had radiation, and 17% had subsequent noncurative operative procedures, regardless of stoma or stent. The median overall survival was 8.1 months, with a median follow-up of 7 months (SBO 3.8 months, LBO 8.2 months). SBO patients had a median survival of 4.7 months (range, months); LBO patients had a median survival of 8.8 months (range months). SBO patients treated with PEG or open g-tube had very short median survival of 1.8 and 5.9 months, respectively, whereas patients receiving more definitive surgical therapy had a median survival of 9.4 months ( p < 0.001). Patients with LBO who had definitive surgical procedures demonstrated significantly better survival (median, 16.8 months) than those undergoing stent placement (median, 7.2 months; p ¼ 0.02). Predictors of outcomes GI treatment was associated with Eastern Cooperative Oncology Group (ECOG) status >1 and presence of carcinomatosis demonstrated on CT ( p ¼ 0.015). Predictors of g-tube, stoma and ability to eat are shown in Table 5. On multivariate analysis, predictors of g-tube or stoma included SBO ( p ¼ 0.050) and CT findings of carcinomatosis ( p < 0.001) (Table 5). On multivariate analysis, independent predictors of inability to eat were CT findings of carcinomatosis ( p ¼ 0.001),

5 826 MOORE DALAL ET AL. Table 5. Multivariate Analyses for Outcomes in Patients Undergoing Palliative Procedures for Malignant Bowel Obstruction (MBO) Univariate Multivariate w 2 P value 95% CI P value Predictors of PEG/open g-tube/stoma Clinical Age (>65 years) Gender Wt loss >10 lbs ECOG > Albumin < Hemoglobin < Tumor Site of tumor Prior chemotherapy Large vs. small bowel obstruction , Single vs. multiple sites of metastases CT findings Multiple sites of bowel obstruction Carcinomatosis , <0.001 Ascites , Multifocal þ carcinomatosis þ ascites , Ability to sustain nutrition at discharge GI vs. OR < , Clinical Age (>65 years) Gender Wt loss >10 lbs ECOG > , Albumin < Hemoglobin < , Tumor Site of tumor Prior chemotherapy , Large vs. small bowel obstruction < , Single vs. multiple sites of metastases CT findings Multiple sites of bowel obstruction , Carcinomatosis < , Ascites < , Multifocal þ carcinomatosis þ ascites , CI, Confidence Interval; PEG, percutaneous endoscopic gastrostomy; OR, operative procedure; GI, endoscopic procedure; CT, computed tomography; LBO, large bowel obstruction; SBO, small bowel obstruction; ECOG, Eastern Cooperative Oncology Group. multifocal obstruction with carcinomatosis and ascites ( p ¼ 0.050), and hemoglobin <10.5 ( p ¼ 0.016). Discussion Our data demonstrate that appropriate palliative intervention for MBO can be achieved in carefully selected patients with metastatic CRC. Patients with SBO and LBO were treated with a variety of palliative procedures. In our series, only 7 patients went to the operating room for g-tube placement or exploration alone. The results show that, with successful relief of symptoms, patients can be discharged home (90%) able to eat (84%). Our multivariate model also suggests that patient features, as well as CT findings, help determine who is likely to benefit from GI rather than OR intervention. Additionally, the results demonstrate that CT is a critical component of preoperative workup in patients with MBO. Because carcinomatosis is one of the most common causes of SBO 2,5 8 and is best identified on CT, and because patients with carcinomatosis and ascites are at higher risk for unsuccessful surgery, undesirable intraoperative scenarios can be avoided with accurate CT interpretation. 9 Our study demonstrates that presence of carcinomatosis, ascites, and multifocal disease significantly impacted whether patients had g-tube or stoma placement. These findings confirm other reports showing that CT is superior to plain radiograph in diagnosing SBO Still others have demonstrated that CT can accurately corroborate intraoperative findings, 14,15 and CT enteroclysis may be even more accurate. 16 Most importantly, however, our study demonstrates specifically how CT can be used in clinical decision making by directing patients to either GI or OR management of MBO. Given that most of these procedures are not emergent, 17 radiographic data may augment decision making and help prevent unsuccessful

6 MALIGNANT BOWEL OBSTRUCTION STAGE IV CRC 827 surgery. Thus, we believe that the use of CT (and maybe CT enteroclysis) is paramount in diagnosing and facilitating treatment decision making for patients with malignant MBO. One-third of the patients in this cohort were treated endoscopically. Our data demonstrate that GI for SBO provided good palliation, but was associated with significantly shorter life expectancy: a median survival of 1.8 months. Although PEGs were technically successful, they palliated nausea/ vomiting in only 82%. Fortunately, these patients often went home 18 able to sip fluids for comfort along with intravenously administered fluids (as most could not consume enough solid food to ensure adequate nutrition). However, LBO patients receiving rectal stents went home quickly, ate well, and had significantly longer survival, corroborating the findings of Adler and Baron that single-site MBO can be successfully palliated by stent placement in 85% to 95% of patients, with lower than 10% morbidity. 19 Although surgery achieved excellent palliation, restoration of enough gastrointestinal function to maintain nutrition, and longer survival, patients with more advanced disease who underwent GI procedures had an average LOS 2 to 3 times shorter, with similar symptom relief. Therefore, utilization of PEGs and stents should be carefully considered. When a patient with MBO consents to a palliative procedure, the question he or she asks most fearfully is, Will I end up with a tube or a bag? In our study, 35% of all MBO patients did receive a g-tube or stoma. G-tubes were placed in 22% (mainly SBO patients), either endoscopically as PEGs, or operatively alone or in combination with another procedure. Stomas were created in 13% (mostly LBO patients). However, 10% of the entire study cohort achieved successful palliation of their obstruction with LOA alone, a reminder that treating MBO may not always require a tube or bag. Nevertheless, the management of MBO entails significant morbidity and mortality. In their study of 1022 palliative procedures, Miner et al. demonstrated a 30-day morbidity and mortality of 29% and 11%, respectively. 3 In our study, morbidity was also 29%, but mortality was much lower (4%). This approaches the range of 5% to 32% reported by Legendre et al. and Makela et al., with death most frequently due to progression of disease. 20,21 Five patients in our cohort had hastening of death because of postoperative complications or disease progression. This lower-than-expected mortality may be a result of the large number of patients treated with colorectal stents. Thus, we emphasize thoughtful patient selection, taking all clinicopathologic factors and CT findings into account in order to delay mortality. It is important to keep in mind that the primary goal in comprehensive care of patients with metastatic disease is symptom palliation rather than cure. 22,23 Extended survival may be a secondary benefit. 24 Unfortunately, however, survival is often used as the sole measure of success. 25 Quality-oflife (QOL) concerns are reported infrequently and, when documented, often vary. 26 Without clearly defined outcome measures, it is difficult to define success. Symptom palliation is judged by achievement and durability of symptom control, improved QOL, and limited morbidity and mortality. 3,26 Nevertheless, recurrence of the primary symptom is noted in 10% to 50% of patients and treatment for additional symptoms is required in 29%, 3,27 demonstrating the limited durability of palliative procedures. Twenty percent of the patients in our study required additional procedures or readmission for recurrent primary symptoms, and 30% for new symptoms. These findings must be clearly communicated during the consent process for an initial palliative procedure. Given advances in chemotherapy, we considered it important to evaluate MBO in the context of associated changes in the course of metastatic CRC. To understand outcomes after palliative surgery in this population, we chose to evaluate both LBO and SBO in our cohort. As the management options differ, we attempted to define differences in presentation, treatment, and outcomes. Our data suggest that there were differences, and that SBO is clearly associated with poorer outcomes. This study, undertaken retrospectively over a 5-year period, has some inherent limitations. Palliative intent was ascertained from electronic medical record progress notes or operative reports. Symptoms and/or symptom resolution were derived from charts, and therefore may not reflect patients experiences. We were unable to assess QOL before and after intervention(s). Furthermore, we captured only those individuals treated endoscopically or operatively, and were not able to obtain the denominator. Identification of all patients presenting to our institution with MBO (including those managed with nasogastric tube alone, medications alone, or hospice, etc.) would be ideal. Unfortunately, we were not able to obtain this data at the time of this study. However, these data may be gathered as part of a national study that is currently being planned. It is certainly the case that some patients with MBO are managed successfully with nasogastric tubes 28 and medicines, including methadone, 29,30 metoclopramide 30 or haloperidol, 5 octreotide with anticholinergics, 31 or steroids 32,33 ; therefore, only a larger study including these additional modes of treatment would provide a complete descriptive assessment of successful management. The results of future prospective randomized clinical trials should help optimize management. Leaders in palliative care are now attempting to implement such trials. 34 We believe that our study presents accurate and meaningful data that will be of significant help in larger prospective trials. Conclusions A diverse armamentarium of GI and OR procedures can be successfully employed to palliate the symptoms of MBO in patients with stage IV CRC, with acceptable morbidity and mortality, including palliation of nausea/vomiting and restoration of the ability to eat. Effective management involves multidisciplinary teamwork between surgical and medical oncologists, gastroenterologists, and radiologists. CT should be performed whenever possible before operating, as identification of ascites, carcinomatosis, or multifocal obstruction indicates that the patient will derive little benefit from surgery. Acknowledgments The authors thank Cynthia Cabral for her assistance in research and collection of crucial data. Her work was a valuable contribution to the successful completion of this study. This study was presented, in part, at the Gastrointestinal Cancers Symposium (co-sponsored by the American

7 828 MOORE DALAL ET AL. Gastroenterological Association, the American Society of Clinical Oncology, the American Society for Therapeutic Radiology and Oncology, and the Society of Surgical Oncology), January 26 28, 2006, San Francisco, California. Author Disclosure Statement No competing financial interests exist. References 1. Davis MP, Nouneh C: Modern management of cancer-related intestinal obstruction. Curr Oncol Rep 2000;2: Aranha GV, Folk FA, Greenlee HB: Surgical palliation of small bowel obstruction due to metastastic carcinoma. Ann Surg 1981;47: Miner TJ, Brennan MF, Jaques DP: A prospective, symptom related, outcomes analysis of 1022 palliative procedures for advanced cancer. Ann Surg 2004;240: Martin RC, Jaques DP, Brennan MF, Karpeh M: Achieving R0 resection for locally advanced gastric cancer: Is it worth the risk of multiorgan resection? J Am Coll Surg 2002;194: Ripamonte C: Management of bowel obstruction in advanced cancer. Curr Opin Oncol 1994;6: Blair SL, Chu DZ, Schwarz RE: Outcome of palliative operations for malignant bowel obstruction in patients with peritoneal carcinomatosis from nongynecological cancer. Ann Surg Oncol 2001;8: Higashi H, Shida H, Ban K, Yamagata S, Masuda K, Imanari T, Yamamoto T: Factors affecting successful palliative surgery for malignant bowel obstruction due to peritoneal dissemination from colorectal cancer. Jpn J Clin Oncol 2003;33: Aabo K, Pedersen H, Bach F, Knudsen J: Surgical management of intestinal obstruction in the late course of malignant disease. Acta Chir Scand 1984;150: Baron TH: Interventional palliative strategies for malignant bowel obstruction. Curr Oncol Rep 2009;11: Maglinte DD, Balthazar EJ, Kelvin FM, Megibow AJ: The role of radiology in the diagnosis of small-bowel obstruction. AJR Am J Roentgenol 1997;168: Maglinte DD, Gage SN, Harmon GH, Kelvin FM, Hage JP, Chua GT, Ng AC, Graffis RF, Chernish SM: Obstruction of the small intestine: Accuracy and role of CT in diagnosis. Radiology 1993;188: Maglinte DD, Kelvin FM, Sandrasegaran K, Nakeeb A, Romano S, Lappas JC, Howard TJ: Radiology of small bowel obstruction: Contemporary approach and controversies. Abdom Imaging 2005;30: Megibow AJ, Balthazar EJ, Cho KC, Medwid SW, Birnbaum BA, Noz ME: Bowel obstruction: Evaluation with CT. Radiology 1991;180: Frager D, Medwid SW, Baer JW, Mollinelli B, Friedman M: CT of small-bowel obstruction: Value in establishing the diagnosis and determining the degree and cause. AJR Am J Roentegenol 1994;162: Matsuoka H, Takahara T, Masaki T, Sugiyama M, Hachiya J, Atomi Y: Preoperative evaluation by magnetic resonance imaging in patients with bowel obstruction. Am J Surg 2002;183: Gollub MJ: Multidetector computed tomography enteroclysis of patients with small bowel obstruction: A volume-rendered surgical perspective. J Comput Assist Tomogr 2005;29: Rubin S: Intestinal obstruction in advanced ovarian cancer: What does the patient want? Gynecol Oncol 1999;75: Scheidbach H, Horbach T, Groitl H, Hohenberger W: Percutaneous endoscopic gastrostomy/jejunostomy (PEG/PEJ) for decompression in the upper gastrointestinal tract. Surg Endosc 1999;13: Adler DG, Baron TH: Endoscopic palliation of colorectal cancer. Hematol Oncol Clin North Am 2002;16: Legendre H, Valhuyse F, Caroli-Bosc FX, Pector JC: Survival and quality of life after palliative surgery for neoplastic gastrointestinal obstruction. Eur J Surg Oncol 2001;27: Makela J, Kiviniemi H, Laitinen S, Kairaluoma MI: Surgical management of intestinal obstruction after treatment for cancer. Eur J Surg 1991;157: Krouse RS, McCahill LE, Easson AM, Dunn GP: When the sun can set on an unoperated bowel obstruction: management of malignant bowel obstruction. J Am Coll Surg 2002;195: [No authors listed] The American College of Surgeons Committee on Ethics. Statement on principles guiding care at the end of life. Bull Am Coll Surg 1998;83: Krouse R: Surgical management of malignant bowel obstruction. Surg Oncol Clin N Am 2004;13: Krebs HB, Goplerud DR: Surgical management of bowel obstruction in advanced ovarian carcinoma. Obstet Gynecol 1983;61: Miner TJ, Jaques DP, Shriver CD: A prospective evaluation of patients undergoing surgery for the palliation of an advanced malignancy. Ann Surg Oncol 2002;9: Miner TJ, Jacques DP, Paty PB, Guillem JG, Wong WD: Symptom control in patients with locally recurrent rectal cancer. Ann Surg Oncol 2003;10: Tang E, Davis J, Silberman H: Bowel obstruction in cancer patients. Arch Surg 1995;130: Davis MP, Nouneh C: Modern management of cancer-related intestinal obstruction. Curr Pain Headache Rep 2001;5: Isbister WH, Elder P, Symons L: Non-operative management of malignant intestinal obstruction. J R Coll Surg Edinb 1990;35: Pandha HW, Waxman J: Octreotide in malignant intestinal obstruction. Anticancer Drugs 1996;7: Baumrucker S: Management of intestinal obstruction in hospice care. Am J Hosp Palliat Care 1998;15: Feuer DJ, Bradley KE, Shepherd JH, Barton DP: Systematic review of surgery in malignant bowel obstruction in advanced gynecological and gastrointestinal cancer. Gynecol Oncol 1999;75: Krouse RS: The international conference on malignant bowel obstruction: A meeting of the minds to advance palliative care research. J Pain Sympt Manage 2007;34:S1 6. Address correspondence to: Larissa K.F. Temple, M.D. Colorectal Service, Department of Surgery Memorial-Sloan Kettering Cancer Center 1275 York Avenue, Room C-1079 New York, NY templel@mskcc.org