- Email: [email protected]
Can a computed tomography scoring system predict the need for surgery in small-bowel obstruction?
The American Journal of Surgery 194 (2007) 780 –784
Presentation
Can a computed tomography scoring system predict the need for surgery in small-bowel obstruction? Kory Jones, M.D., Alicia J. Mangram, M.D., F.A.C.S.*, Ricardo A. Lebron, M.D., Lennard Nadalo, M.D., Ernest Dunn, M.D., F.A.C.S. Methodist–Medical Education, 1441 North Beckley Avenue, Dallas, TX 75203, USA Manuscript received June 19, 2007; revised manuscript September 10, 2007 Presented at the 59th Annual Meeting of the Southwestern Surgical Congress, Rancho Mirage, CA, March 25–29, 2007
Abstract Background: Small-bowel obstruction (SBO) is a common dilemma faced by general surgeons. A timely and accurate diagnosis is crucial, based on history, physical examination, and radiographic studies. A computed tomography (CT) scan has become an increasingly common diagnostic modality. The aim of this study was to define a set of CT criteria that may help determine whether a patient would require operative intervention. Methods: A retrospective chart review was performed over an 18-month period. Patients diagnosed with SBO or partial SBO (PSBO) who had a CT scan performed were included. CT scans were then reviewed independently by a staff radiologist blinded to the clinical outcome. A scoring system based on 7 radiographic criteria was then developed. Statistical analysis was performed on the data. Results: From March 1, 2004, to August 30, 2005, 96 patients were found to have the diagnosis of SBO or PSBO and a documented CT scan. Seventy-four patients had undergone prior intra-abdominal procedures. Fifty-five percent were taken to the operating room (OR) over an average of 1.9 days (range 1–12 days). Seventy-nine percent of these patients went to the OR within the first 24 hours. Lysis of adhesions was performed in 38%, small bowel was resected in 47%, and incarcerated ventral hernias were found in 11%. Ischemic bowel was found in 11% of the cases. The CT scoring system was then correlated with the actual treatment. A score of 8 or higher predicted the need for surgery 75% of the time. When looking at the criteria individually, patients with a CT reading of complete obstruction, dilated small bowel, or free fluid were operated on 77%, 66%, and 65% of the time, respectively. Conclusion: A CT scoring system can successfully predict the need for surgery 75% of the time. Likewise, specific criteria, when present in combination, can predict the need for operative intervention in 79% of cases with SBO. © 2007 Published by Excerpta Medica Inc. Keywords: Small-bowel obstruction; Computed tomography; Scoring system
Small-bowel obstruction (SBO) is a common dilemma faced by general surgeons. In fact, it is an increasing cause of acute abdominal pain seen in the emergency department. A study from England reported an increase of 212% in the number of elderly patients with SBO from 1976 to 1989 [1]. Even the treatment modalities have changed over the years. The old paradigm was to not let the sun set on an obstruction of the small intestine because of the difficulty of distinguishing between a complete obstruction that could compromise the viability of the intestine and a partial ob* Corresponding author. Tel.: ⫹1-214-947-2303; fax: ⫹1-214-947-2361. E-mail address: [email protected] 0002-9610/07/$ – see front matter © 2007 Published by Excerpta Medica Inc. doi:10.1016/j.amjsurg.2007.09.020
struction [2]. Today, there are multiple options concerning the management of SBO. These include emergent operation when bowel ischemia or necrosis is suspected, conservative management with nasogastric suctioning and bowel rest, and, if observation fails, subsequent surgical exploration [3]. The following question remains: what patients can we safely treat conservatively? This is important because many studies have shown that delays in the definitive treatment of complete SBO can result in a significant increase in the need for bowel resection as well as morbidity and mortality [4]. One study showed that the mortality rate for patients with strangulated bowel increased from 8% to 25% if the surgery
A.J. Mangram et al. / The American Journal of Surgery 194 (2007) 780 –784
781
Table 1 Point values for CT scoring system
was delayed more than 36 hours [5]. Fevang et al [6] found that mortality rates for SBO actually decreased from 1962 to 1995 in part because of a decrease in time to treatment. The diagnosis of complete versus partial SBO has traditionally been made based on clinical examination and plain film radiography. More recently, computed tomography (CT) scan has become an increasingly common diagnostic modality. Several studies have evaluated CT scan for its sensitivity and specificity in distinguishing complete versus partial SBO. In a study by Frager et al [7], the sensitivity of diagnosing complete SBO by CT scan was 100% and prevented a 12- to 72-hour delay in surgery. Another study looked at the use of plain films versus CT scan and found the sensitivity and specificity to be 75% and 53% for plain films compared with 92% and 71% for CT. They suggested that CT scan should be considered for use as the primary diagnostic tool for SBO [8]. When seeing a patient with a possible SBO, the general surgeon must balance the clinical findings with radiology studies. There are little data to define what percentage of complete SBO versus partial SBO (PSBO) diagnosed by CT scan will need surgical repair. The aim of this study was to define a set of CT criteria that may help determine whether a patient would require operative intervention.
dict the need for surgical intervention. Multivariate and bivariate analyses were performed on independent variables using the SPSS software for Windows. Pearson chi-square tests were calculated for significance by using an epidemiology computer system analysis. Single variables as well as various combinations were analyzed to identify any additive effects.
Methods With institutional review board approval, a retrospective chart review was performed over an 18-month period at a community-based teaching facility. Patients diagnosed with SBO or PSBO who had a CT scan performed were included in the study. Patients under 18 years of age, with malignant ascites, obstruction caused by malignancy, peritoneal carcinomatosis, or those without CT evidence of SBO or PSBO were excluded. Also, those with an immediate postoperative obstruction were excluded. The presenting history and physical (including types of previous surgeries and comorbidities), radiologic findings, therapeutic interventions, and outcomes were examined. Those patients who presented with probable SBO were scanned after the oral administration of 32 oz of a mixture of gastrogaffin (50 mL), tea (15 mL), and water (1,000 mL). Patients were examined 2 hours after the oral contrast. Those patients with normal renal function were also given 100 mL of intravenous contrast. The contrasted scans were obtained using a 5-mm collimation. Delayed images were needed in selected cases in which high-grade obstruction or ileus was present. All studies were reviewed on an imaging workstation. A staff radiologist blinded to the original reading and the clinical outcome reviewed CT scans independently. A scoring system based on 7 radiographic findings was developed by the radiologist involved in the study. The criteria included dilated small bowel, transition point, ascites, complete obstruction, partial obstruction, closed loop, and/or free air. Point values were added or subtracted based on the findings from the repeat scans (Table 1). Complete obstruction was defined as when gas and/or oral contrast in the proximal small-bowel segment did not pass beyond the transition point into the distal bowel. Repeat CT scans were ordered at the discretion of the attending surgeon. The radiologist then used this scoring system without the benefit of clinical outcome data to pre-
Results From March 1, 2004, to August 30, 2005, 96 patients were found to have the diagnosis of SBO or PSBO and documented CT scan. Fifty-six percent were female. The average age was 63 (⫾18.9). All patients were admitted to the hospital, and CT scans were performed within 12 hours of admission. Repeat CT scans were performed in 36 patients usually within 12 hours of the initial scan. Fifty-five percent were taken to the operating room (OR). Seventynine percent of these patients went to the OR within the first 24 hours. Seventy-five patients had undergone prior intraabdominal procedures, including, 31 hysterectomies, 22 colectomies, 22 cholecystectomies, 15 appendectomies, 6 lysis of adhesions, and 6 small-bowel resections. Forty patients had 2 or more prior procedures. Of the patients with prior surgery that were taken to the OR, lysis of adhesions was performed in 39%, small bowel was resected in 51%, and incarcerated ventral hernias were found in 10%. Twenty-one patients did not have previous procedures, and they all had a CT scan that showed SBO or PSBO. Twelve of these patients were taken to surgery. Three patients had multiple adhesions, 2 had ventral hernias, and 2 had intra-abdominal abscesses. The other 5 were found to have cholecystitis, appendicitis, volvulus, and inflammatory bowel disease, and 1 patient had a single band causing obstruction. Four patients underwent small-bowel resection. All surgical patients were taken to the OR over an average of 1.9 days (range 1–12 days). Of the 25 patients who underwent bowel resection, the length of time to the OR was 1.48 days. Ischemic bowel was found in 11% of the cases. The need for bowel resection was not dependent on the length of the time to surgical therapy. Additionally, obtaining a CT scan did not appear to delay definitive operative treatment. The CT scoring system was then correlated with the actual treatment. A score of 8 or higher predicted the need
Initial CT findings Free air Transition point Complete obstruction Closed loop Free fluid Partial obstruction Repeat CT findings Resolution Improved obstruction Persistent SBO Worsening obstruction Free air
5 3 3 3 3 2 ⫺5 ⫺2 ⫹2 ⫹3 ⫹5
782
A.J. Mangram et al. / The American Journal of Surgery 194 (2007) 780 –784
Table 2 Results of a Pearson chi-square test for statistical analysis CT findings
P value
Transition point Complete obstruction Closed loop Dilated small bowel Free fluid
.166 .001 .198 .128 .302
for surgery 75% of the time. A value of 8 on the scoring system affects the sensitivity versus specificity. A value of 7 on the scoring system would have predicted a higher percentage needing surgery but would have falsely indicated the need for surgery in some cases. When looking at the criteria individually, patients with a CT reading of dilated small bowel or free fluid were operated on 66% and 65% of the time, respectively. Additionally, those patients who had CT findings of complete obstruction in combination with a dilated small bowel went to the OR 79% of the time, and those with a transition point and dilated small bowel were operated on in 76% of the cases. However, these values were not statistically significant. These criteria appeared to be additive, but there was a small sample size. In fact, the only CT finding that statistically predicted the need for surgical intervention was complete obstruction (P ⫽ .001). When partial SBO was identified on a CT scan, 52% were treated successfully without surgery and 48% were taken to the operating room. However, when complete obstruction was read, 6% were treated conservatively and 94% required operative intervention. A CT diagnosis of closed loop obstruction was found in 2 patients, and both went to the OR. This was not statistically significant secondary to the small number (Table 2). Free air was not found on any CT scans. The average length of hospital stay for all patients was 8.3 days (7.2). For those patients who underwent smallbowel resection, the mean length of stay was 11.8 days (⫾8.7). The postoperative morbidity included 7 wound infections, 1 dehiscence, 1 patient with pneumonia, and 1 with sepsis. Two patients died during the postoperative period. Both of these patients were taken to the OR for their SBO in less than 24 hours. In those patients treated conservatively, 1 patient had sepsis, 1 had pneumonia, and 1 developed a deep venous thrombosis. Four patients in the conservative group died because of non–intra-abdominal sources. The average age of those patients who died in both the surgery group and the conservative group was 86 years. Comments Traditionally, an abdominal plain film is the initial diagnostic test for SBO. Often, it has been performed in the emergency department before a surgeon is even consulted. The findings suggestive of SBO are air-fluid levels, smallbowel distention, a paucity of air in the distal colon, and in some instances free air [9,10]. Despite the increased utilization of CT scans in the diagnosis of SBO, there are still those that believe it is an unnecessary adjunct. A retrospective review of patients treated by a single surgeon suggested that there was no added advantage of imaging other than
plain films in deciding which patients would require emergent operation. This surgeon found in his experience partial obstructions could be treated conservatively up to 15 days and that CT and fluoroscopy are only needed in cases of obstruction caused by malignancy [11]. However, now that CT scans have become a more common diagnostic tool, there have been several studies performed to evaluate the utility of CT for acute abdominal pain. A study by Nagurney et al [12] suggested that although abdominal radiographs have some value as a screening tool, most patients are now going to have a CT scan. He suggests that plain abdominal radiographs may be unhelpful and misleading. We certainly concur with this opinion. Less than half of our patients had plain films before CT scanning. The written CT reports almost never referred to the prior plain films, and follow-up plain films were not suggested by the radiologist. In the future, eliminating these plain abdominal films might prove to be cost saving to the patient. Similar data have been presented by Frager et al [7]. He showed in a study of 46 patients with combined clinical and radiographic findings demonstrating complete SBO there was a sensitivity of 48%. In the same patients, CT was 100% sensitive for complete SBO [7]. In our retrospective review, we found that if a radiologist reads a CT scan as a complete obstruction, 94% needed operative intervention. Another concern for the general surgeon is the issue of how to deal with a CT finding of partial SBO. Mallo et al [13], in a MEDLINE search of 15 studies, suggested that a CT scan showing PSBO represents a clinical condition that will resolve without surgical intervention. In our review, we found that of the 80 patients with the CT finding of PSBO, 48% needed operative intervention. In our experience, serial clinical examination has been enhanced with follow-up CT scans at 8 to 10 hours. These scans can suggest progression of contrast into the colon, thus eliminating the need for immediate laparotomy. Failure of the progression of the contrast can affirm the diagnosis of complete SBO with the need for operative repair. A few articles have examined certain CT findings that are particularly suggestive of ischemic bowel and strangulation. Another article by Frager et al [14] found that a diagnosis of ischemia could be based on 2 or more of the following CT findings: bowel-wall thickening, high attenuation of the bowel wall on unenhanced scans, mesenteric edema or fluid, asymmetric bowel-wall enhancement with intravenous contrast, pneumatosis, or portal venous gas. They found 100% sensitivity in their series. A similar study showed that the CT findings of poor or no enhancement of bowel wall, a serrated beak, a large amount of ascites, an unusual course of mesenteric vessels, and diffuse engorgement of mesenteric vasculature increased the accuracy of diagnosing strangulated small bowel. They concluded that using a set of selected, known criteria helped to differentiate simple versus strangulated SBO [15]. Although in our study we were not specifically looking for criteria that would predict intestinal ischemia, we did not identify CT findings that accurately predicted the need for bowel resection or the presence of ischemic bowel. The goal of our study was to identify a set of objective criteria that can be used to aid the surgeon in determining the appropriate treatment of SBO. Proximal dilatation with
A.J. Mangram et al. / The American Journal of Surgery 194 (2007) 780 –784
an identifiable transition point to decompressed bowel is the classic finding. We examined these criteria, along with ascites, a closed loop, and free air, alone and in combination. By using these criteria, the scoring system did appear to supply an additive effect aiding surgeons in their decision-making. This system did not reach statistical significance likely because of the low power of our article. We did find that complete obstruction is the only statistically significant predictor of the need for surgery. This can be especially helpful when faced with a patient with equivocal physical examination findings, but a CT reading of complete obstruction. It should be noted that the need for bowel resection was not dependent on the preoperative finding of complete versus partial obstruction. Interestingly, there were 21 patients who did not have previous surgical procedures. Yet, they had CT readings of SBO or PSBO. It is very common for CT scans of the abdomen to have been obtained for a wide variety of reasons when ordered in the emergency room. The presumptive diagnosis of SBO is seldom the reason for the request. The positive predictive value of the CT of the abdomen is important as it is related to SBO; however, because of the low threshold for ordering CT scans from the emergency room today, many other causes of abdominal pain will be presented to the radiologist (appendicitis, renal stones, pelvic inflammatory disease, and so on) that may present with regional small-bowel dilatation that must be differentiated from SBO. Megibow et al [16] pointed out that CT scans can be useful in patients who have not had prior surgery but present with signs of infection, bowel infarction, or palpable abdominal masses. In our series of 21 patients without prior surgery, 12 went to the operating room. Four of these patients had adhesions; 2 had ventral hernias; 2 had intraabdominal abscesses; and 1 each had acute cholecystitis, acute appendicitis, volvulus, and inflammatory bowel disease. The remaining 9 patients in this group had resolution of their acute conditions without operative intervention. More than likely, these patients had an ileus from the inflammatory processes. Despite the many advances in CT scans and the proven utility as a diagnostic aid in acute abdominal pain, this study ultimately shows that clinical findings are still the mainstay in diagnosing SBO. This should include white blood cell count, electrolytes, history, and, most importantly, physical examination. If a patient has peritoneal signs, he/she should be taken to the OR promptly without the need for radiographic studies. CT scans can be a useful adjunct when the patient has equivocal findings and as follow-up studies. Conclusion A CT scoring system can successfully help the surgeon decide whether or not the patient requires surgery 75% of the time. Likewise, specific criteria, when present in combination, can also aid in the decision-making process for operative intervention in 79% of cases with SBO. However, complete obstruction is the only reliable radiographic indicator for the need for operative intervention. Ultimately, a CT scan can be a guide for treatment, but the final decision should be based on clinical findings.
783
References [1] Faruqi R, Galland RB, Williams JM. An audit of surgical emergencies in the very old. Ann R Coll Surg Engl 1991;73:285– 8. [2] Silen W, Hein M, Goldman L. Strangulation obstruction of the small intestine. Arch Surg 1962;85:137– 45. [3] Foster NM, McGory ML, Zingmond DS, Ko CY. Small bowel obstruction: a population-based appraisal. J Am Coll Surg 2006;203: 170 – 6. [4] Bickell NA, Federman AD, Aufses AH. Influence of time on risk of bowel resection in complete small bowel obstruction. J Am Coll Surg 2005;201:847–54. [5] Shrock TR. Small intestine. In: Way LW, eds. Current surgical diagnosis and treatment. Norwalk, CT: Appleton and Lange; 1988: 561– 85. [6] Fevang BT, Fevang J, Strangeland L, et al. Complication and death after surgical treatment of small bowel obstruction: a 35-year institutional experience. Ann Surg 2000;231:529 –37. [7] Frager D, Medwid SW, Baer JW, et al. CT of small-bowel obstruction: value in establishing the diagnosis and determining the degree and cause. AJR Am J Roentgenol 1994;162:37– 41. [8] Daneshmand S, Hedley CG, Stain SC. The utility and reliability of computed tomography scan in the diagnosis of small bowel obstruction. Am Surg 1999;65:922– 6. [9] Flasar MH, Goldberg E. Acute abdominal pain. Med Clin North Am 2006;90:481–503. [10] Frager D. Intestinal obstruction: role of CT. Gastroenterol Clin North Am 2002;31:777–99. [11] Nauta RJ. Advanced abdominal imaging is not required to exclude strangulation if complete small bowel obstructions undergo prompt laparotomy. J Am Coll Surg 2005;200:904 –11. [12] Nagurney JT, Brown DF, Novelline RA, et al. Plain abdominal radiographs and abdominal CT scans for non-traumatic abdominal pain-added vale? Am J Emerg Med 1999;17:668 –71. [13] Mallo RD, Salem L, Lalani T, Flum DR. Computed tomography diagnosis of ischemia and complete obstruction in small bowel obstruction: a systematic review. J Gastrointest Surg 2005;9:690 – 4. [14] Frager D, Baer JW, Medwid SW, et al. Detection of intestinal ischemia in patients with acute small-bowel obstruction due to adhesions or hernia: efficacy of CT. AJR Am J Roentgenol 1996;166:67–71. [15] Ha HK, Kim JS, Lee MS, et al. Differentiation of simple and strangulated small-bowel obstruction: usefulness of known CT criteria. Radiology 1997;204:507–12. [16] Megibow AJ, Balthazar EJ, Cho KC, et al. Bowel obstruction: evaluation with CT. Radiology 1991;180:313– 8.
Discussion Jacques Heppell, M.D. (Scottsdale, AZ): It is nice to see that a quantitative approach by means of a computed tomography scoring system was tested to predict the need for surgery in small-bowel obstruction. We know that early recognition of intestinal strangulation is very difficult on conventional clinical grounds only. The treatment of choice for complete small-bowel obstruction remains early surgical intervention. However, nonoperative management of selected patients with patients with intestinal obstruction is sometimes clearly indicated, namely in the presence of partial obstruction, obstructive radiation enteritis, in the early postoperative period, and in presence of carcinomatosis or pseudo-obstruction. Most of these conditions except for partial small-bowel obstruction were excluded from your study. This may explain that 55% of your patients required surgery within 2 days of admission. When complete obstruction was found, 94% required a surgical intervention compared with 48% when obstruction was partial or incomplete. Ischemic bowel was found in 11% of cases. In patients who had prior surgery, small-bowel resection was needed in 51% of your cases. You found that the need for
784
A.J. Mangram et al. / The American Journal of Surgery 194 (2007) 780 –784
resection was not dependent on the length of time to surgical intervention, which in your study varied from 1 to 12 days. Irreversible ischemic changes require resection. There is discordance in your data regarding the need for bowel resection. Because the morbidity rate is higher in patients requiring resection, it would be nice to know the indications for resection in your study group and determine if computed tomography scan could predict not only the need for surgery but also the need for bowel resection. There are limitations to your scoring system. Two of the seven criteria, the presence of free air and closed loop, could not be used because of small numbers. Five criteria remained, one being dilated small bowel, which is a prerequisite for diagnosis. Only 40% of the patients had a repeat computed tomography scan to document improved or worsening obstruction. Finally, only 4 criteria remained to be tested: presence of transition point, ascites, and complete or partial obstruction. It is, therefore, not surprising to find that complete obstruction was the only reliable radiographic indicator for the need for surgical intervention. This finding concurs with randomized controlled trials on the use of water-soluble contrast in the management of adhesive small bowel obstruction. Although Gastrographin does not cause resolution of small-bowel obstruction, it can be used as a predictive test for nonoperative resolution and identify patients with complete obstruction needing early surgical intervention, thus reducing hospital stay. In my practice, I found computed tomography scan useful to pick up findings that I may have missed on physical examination such as the presence of incarcerated hernia in an obese patient. Indeed, 10% of your patients were found to have incarcerated hernia at laparotomy. A computed tomography scan can also detect unsuspected portal venous air and pneumobilia. Small-bowel volvulus is a life threatening surgical emergency associated with obstruction. It can be detected on computed tomography scan by the presence of the “spoke wheel” sign. In addition, the presence of intestinal intussusceptions with the classic “bull’s eye” sign can be recognized on computed tomography and lead to a timely surgical intervention. Finally, I would like to congratulate the authors for their
efforts and a study leading to another, I would like to ask them what modifications to their scoring system would be worth studying in a prospective manner to minimize the morbidity of bowel resection in patients with small-bowel obstruction. Dan Margulies, M.D. (Los Angeles, CA): I was concerned about the use of oral contrast in all of these. Can you comment whether you felt this contributed to any of those morbidities including the pneumonia and the deaths. It seems like the finding of complete obstruction could be determined without the use of contrast and because that was the only finding, are you are going to continue using oral contrast in all of your patients? Kory Jones, M.D. (Dallas, TX): Dr. Heppell, our indications for small-bowel resection intraoperatively were based on the feeling of the surgeon that was performing the surgery at the time. Unfortunately, that is not always documented on the operative report. Any evidence of ischemia or bowel necrosis was an indication, but, other than that, it is based on what the surgeon feels like based on the surgery. We did not find in our series any computed tomography findings that would indicate the need for small-bowel resection. Other articles have looked at computed tomography findings that would suggest bowel strangulation or ischemia and had a high sensitivity and specificity using computed tomography scans for that purpose. As far as modifications for perspective review, I believe that we would have to look more carefully at the different criteria with the scoring system and look at those in conjunction with each other and make sure that everybody is using a standardized scoring system if we are going to be looking at this prospectively. Other than that, I am not sure what modifications we would go by, except for increasing our number and making sure we were standardized. As far as the oral contrast, it was not used in all studies, and I am sorry if it seemed that way. The definition for complete obstruction was actually gas and/or oral contrast. So, in those patients, we did not have oral contrast. It was based on whether or not gas pass being a transition point to the distal bowel. And, we did not find any increased morbidities in those patients who did receive oral contrast.
Presentation
Can a computed tomography scoring system predict the need for surgery in small-bowel obstruction? Kory Jones, M.D., Alicia J. Mangram, M.D., F.A.C.S.*, Ricardo A. Lebron, M.D., Lennard Nadalo, M.D., Ernest Dunn, M.D., F.A.C.S. Methodist–Medical Education, 1441 North Beckley Avenue, Dallas, TX 75203, USA Manuscript received June 19, 2007; revised manuscript September 10, 2007 Presented at the 59th Annual Meeting of the Southwestern Surgical Congress, Rancho Mirage, CA, March 25–29, 2007
Abstract Background: Small-bowel obstruction (SBO) is a common dilemma faced by general surgeons. A timely and accurate diagnosis is crucial, based on history, physical examination, and radiographic studies. A computed tomography (CT) scan has become an increasingly common diagnostic modality. The aim of this study was to define a set of CT criteria that may help determine whether a patient would require operative intervention. Methods: A retrospective chart review was performed over an 18-month period. Patients diagnosed with SBO or partial SBO (PSBO) who had a CT scan performed were included. CT scans were then reviewed independently by a staff radiologist blinded to the clinical outcome. A scoring system based on 7 radiographic criteria was then developed. Statistical analysis was performed on the data. Results: From March 1, 2004, to August 30, 2005, 96 patients were found to have the diagnosis of SBO or PSBO and a documented CT scan. Seventy-four patients had undergone prior intra-abdominal procedures. Fifty-five percent were taken to the operating room (OR) over an average of 1.9 days (range 1–12 days). Seventy-nine percent of these patients went to the OR within the first 24 hours. Lysis of adhesions was performed in 38%, small bowel was resected in 47%, and incarcerated ventral hernias were found in 11%. Ischemic bowel was found in 11% of the cases. The CT scoring system was then correlated with the actual treatment. A score of 8 or higher predicted the need for surgery 75% of the time. When looking at the criteria individually, patients with a CT reading of complete obstruction, dilated small bowel, or free fluid were operated on 77%, 66%, and 65% of the time, respectively. Conclusion: A CT scoring system can successfully predict the need for surgery 75% of the time. Likewise, specific criteria, when present in combination, can predict the need for operative intervention in 79% of cases with SBO. © 2007 Published by Excerpta Medica Inc. Keywords: Small-bowel obstruction; Computed tomography; Scoring system
Small-bowel obstruction (SBO) is a common dilemma faced by general surgeons. In fact, it is an increasing cause of acute abdominal pain seen in the emergency department. A study from England reported an increase of 212% in the number of elderly patients with SBO from 1976 to 1989 [1]. Even the treatment modalities have changed over the years. The old paradigm was to not let the sun set on an obstruction of the small intestine because of the difficulty of distinguishing between a complete obstruction that could compromise the viability of the intestine and a partial ob* Corresponding author. Tel.: ⫹1-214-947-2303; fax: ⫹1-214-947-2361. E-mail address: [email protected] 0002-9610/07/$ – see front matter © 2007 Published by Excerpta Medica Inc. doi:10.1016/j.amjsurg.2007.09.020
struction [2]. Today, there are multiple options concerning the management of SBO. These include emergent operation when bowel ischemia or necrosis is suspected, conservative management with nasogastric suctioning and bowel rest, and, if observation fails, subsequent surgical exploration [3]. The following question remains: what patients can we safely treat conservatively? This is important because many studies have shown that delays in the definitive treatment of complete SBO can result in a significant increase in the need for bowel resection as well as morbidity and mortality [4]. One study showed that the mortality rate for patients with strangulated bowel increased from 8% to 25% if the surgery
A.J. Mangram et al. / The American Journal of Surgery 194 (2007) 780 –784
781
Table 1 Point values for CT scoring system
was delayed more than 36 hours [5]. Fevang et al [6] found that mortality rates for SBO actually decreased from 1962 to 1995 in part because of a decrease in time to treatment. The diagnosis of complete versus partial SBO has traditionally been made based on clinical examination and plain film radiography. More recently, computed tomography (CT) scan has become an increasingly common diagnostic modality. Several studies have evaluated CT scan for its sensitivity and specificity in distinguishing complete versus partial SBO. In a study by Frager et al [7], the sensitivity of diagnosing complete SBO by CT scan was 100% and prevented a 12- to 72-hour delay in surgery. Another study looked at the use of plain films versus CT scan and found the sensitivity and specificity to be 75% and 53% for plain films compared with 92% and 71% for CT. They suggested that CT scan should be considered for use as the primary diagnostic tool for SBO [8]. When seeing a patient with a possible SBO, the general surgeon must balance the clinical findings with radiology studies. There are little data to define what percentage of complete SBO versus partial SBO (PSBO) diagnosed by CT scan will need surgical repair. The aim of this study was to define a set of CT criteria that may help determine whether a patient would require operative intervention.
dict the need for surgical intervention. Multivariate and bivariate analyses were performed on independent variables using the SPSS software for Windows. Pearson chi-square tests were calculated for significance by using an epidemiology computer system analysis. Single variables as well as various combinations were analyzed to identify any additive effects.
Methods With institutional review board approval, a retrospective chart review was performed over an 18-month period at a community-based teaching facility. Patients diagnosed with SBO or PSBO who had a CT scan performed were included in the study. Patients under 18 years of age, with malignant ascites, obstruction caused by malignancy, peritoneal carcinomatosis, or those without CT evidence of SBO or PSBO were excluded. Also, those with an immediate postoperative obstruction were excluded. The presenting history and physical (including types of previous surgeries and comorbidities), radiologic findings, therapeutic interventions, and outcomes were examined. Those patients who presented with probable SBO were scanned after the oral administration of 32 oz of a mixture of gastrogaffin (50 mL), tea (15 mL), and water (1,000 mL). Patients were examined 2 hours after the oral contrast. Those patients with normal renal function were also given 100 mL of intravenous contrast. The contrasted scans were obtained using a 5-mm collimation. Delayed images were needed in selected cases in which high-grade obstruction or ileus was present. All studies were reviewed on an imaging workstation. A staff radiologist blinded to the original reading and the clinical outcome reviewed CT scans independently. A scoring system based on 7 radiographic findings was developed by the radiologist involved in the study. The criteria included dilated small bowel, transition point, ascites, complete obstruction, partial obstruction, closed loop, and/or free air. Point values were added or subtracted based on the findings from the repeat scans (Table 1). Complete obstruction was defined as when gas and/or oral contrast in the proximal small-bowel segment did not pass beyond the transition point into the distal bowel. Repeat CT scans were ordered at the discretion of the attending surgeon. The radiologist then used this scoring system without the benefit of clinical outcome data to pre-
Results From March 1, 2004, to August 30, 2005, 96 patients were found to have the diagnosis of SBO or PSBO and documented CT scan. Fifty-six percent were female. The average age was 63 (⫾18.9). All patients were admitted to the hospital, and CT scans were performed within 12 hours of admission. Repeat CT scans were performed in 36 patients usually within 12 hours of the initial scan. Fifty-five percent were taken to the operating room (OR). Seventynine percent of these patients went to the OR within the first 24 hours. Seventy-five patients had undergone prior intraabdominal procedures, including, 31 hysterectomies, 22 colectomies, 22 cholecystectomies, 15 appendectomies, 6 lysis of adhesions, and 6 small-bowel resections. Forty patients had 2 or more prior procedures. Of the patients with prior surgery that were taken to the OR, lysis of adhesions was performed in 39%, small bowel was resected in 51%, and incarcerated ventral hernias were found in 10%. Twenty-one patients did not have previous procedures, and they all had a CT scan that showed SBO or PSBO. Twelve of these patients were taken to surgery. Three patients had multiple adhesions, 2 had ventral hernias, and 2 had intra-abdominal abscesses. The other 5 were found to have cholecystitis, appendicitis, volvulus, and inflammatory bowel disease, and 1 patient had a single band causing obstruction. Four patients underwent small-bowel resection. All surgical patients were taken to the OR over an average of 1.9 days (range 1–12 days). Of the 25 patients who underwent bowel resection, the length of time to the OR was 1.48 days. Ischemic bowel was found in 11% of the cases. The need for bowel resection was not dependent on the length of the time to surgical therapy. Additionally, obtaining a CT scan did not appear to delay definitive operative treatment. The CT scoring system was then correlated with the actual treatment. A score of 8 or higher predicted the need
Initial CT findings Free air Transition point Complete obstruction Closed loop Free fluid Partial obstruction Repeat CT findings Resolution Improved obstruction Persistent SBO Worsening obstruction Free air
5 3 3 3 3 2 ⫺5 ⫺2 ⫹2 ⫹3 ⫹5
782
A.J. Mangram et al. / The American Journal of Surgery 194 (2007) 780 –784
Table 2 Results of a Pearson chi-square test for statistical analysis CT findings
P value
Transition point Complete obstruction Closed loop Dilated small bowel Free fluid
.166 .001 .198 .128 .302
for surgery 75% of the time. A value of 8 on the scoring system affects the sensitivity versus specificity. A value of 7 on the scoring system would have predicted a higher percentage needing surgery but would have falsely indicated the need for surgery in some cases. When looking at the criteria individually, patients with a CT reading of dilated small bowel or free fluid were operated on 66% and 65% of the time, respectively. Additionally, those patients who had CT findings of complete obstruction in combination with a dilated small bowel went to the OR 79% of the time, and those with a transition point and dilated small bowel were operated on in 76% of the cases. However, these values were not statistically significant. These criteria appeared to be additive, but there was a small sample size. In fact, the only CT finding that statistically predicted the need for surgical intervention was complete obstruction (P ⫽ .001). When partial SBO was identified on a CT scan, 52% were treated successfully without surgery and 48% were taken to the operating room. However, when complete obstruction was read, 6% were treated conservatively and 94% required operative intervention. A CT diagnosis of closed loop obstruction was found in 2 patients, and both went to the OR. This was not statistically significant secondary to the small number (Table 2). Free air was not found on any CT scans. The average length of hospital stay for all patients was 8.3 days (7.2). For those patients who underwent smallbowel resection, the mean length of stay was 11.8 days (⫾8.7). The postoperative morbidity included 7 wound infections, 1 dehiscence, 1 patient with pneumonia, and 1 with sepsis. Two patients died during the postoperative period. Both of these patients were taken to the OR for their SBO in less than 24 hours. In those patients treated conservatively, 1 patient had sepsis, 1 had pneumonia, and 1 developed a deep venous thrombosis. Four patients in the conservative group died because of non–intra-abdominal sources. The average age of those patients who died in both the surgery group and the conservative group was 86 years. Comments Traditionally, an abdominal plain film is the initial diagnostic test for SBO. Often, it has been performed in the emergency department before a surgeon is even consulted. The findings suggestive of SBO are air-fluid levels, smallbowel distention, a paucity of air in the distal colon, and in some instances free air [9,10]. Despite the increased utilization of CT scans in the diagnosis of SBO, there are still those that believe it is an unnecessary adjunct. A retrospective review of patients treated by a single surgeon suggested that there was no added advantage of imaging other than
plain films in deciding which patients would require emergent operation. This surgeon found in his experience partial obstructions could be treated conservatively up to 15 days and that CT and fluoroscopy are only needed in cases of obstruction caused by malignancy [11]. However, now that CT scans have become a more common diagnostic tool, there have been several studies performed to evaluate the utility of CT for acute abdominal pain. A study by Nagurney et al [12] suggested that although abdominal radiographs have some value as a screening tool, most patients are now going to have a CT scan. He suggests that plain abdominal radiographs may be unhelpful and misleading. We certainly concur with this opinion. Less than half of our patients had plain films before CT scanning. The written CT reports almost never referred to the prior plain films, and follow-up plain films were not suggested by the radiologist. In the future, eliminating these plain abdominal films might prove to be cost saving to the patient. Similar data have been presented by Frager et al [7]. He showed in a study of 46 patients with combined clinical and radiographic findings demonstrating complete SBO there was a sensitivity of 48%. In the same patients, CT was 100% sensitive for complete SBO [7]. In our retrospective review, we found that if a radiologist reads a CT scan as a complete obstruction, 94% needed operative intervention. Another concern for the general surgeon is the issue of how to deal with a CT finding of partial SBO. Mallo et al [13], in a MEDLINE search of 15 studies, suggested that a CT scan showing PSBO represents a clinical condition that will resolve without surgical intervention. In our review, we found that of the 80 patients with the CT finding of PSBO, 48% needed operative intervention. In our experience, serial clinical examination has been enhanced with follow-up CT scans at 8 to 10 hours. These scans can suggest progression of contrast into the colon, thus eliminating the need for immediate laparotomy. Failure of the progression of the contrast can affirm the diagnosis of complete SBO with the need for operative repair. A few articles have examined certain CT findings that are particularly suggestive of ischemic bowel and strangulation. Another article by Frager et al [14] found that a diagnosis of ischemia could be based on 2 or more of the following CT findings: bowel-wall thickening, high attenuation of the bowel wall on unenhanced scans, mesenteric edema or fluid, asymmetric bowel-wall enhancement with intravenous contrast, pneumatosis, or portal venous gas. They found 100% sensitivity in their series. A similar study showed that the CT findings of poor or no enhancement of bowel wall, a serrated beak, a large amount of ascites, an unusual course of mesenteric vessels, and diffuse engorgement of mesenteric vasculature increased the accuracy of diagnosing strangulated small bowel. They concluded that using a set of selected, known criteria helped to differentiate simple versus strangulated SBO [15]. Although in our study we were not specifically looking for criteria that would predict intestinal ischemia, we did not identify CT findings that accurately predicted the need for bowel resection or the presence of ischemic bowel. The goal of our study was to identify a set of objective criteria that can be used to aid the surgeon in determining the appropriate treatment of SBO. Proximal dilatation with
A.J. Mangram et al. / The American Journal of Surgery 194 (2007) 780 –784
an identifiable transition point to decompressed bowel is the classic finding. We examined these criteria, along with ascites, a closed loop, and free air, alone and in combination. By using these criteria, the scoring system did appear to supply an additive effect aiding surgeons in their decision-making. This system did not reach statistical significance likely because of the low power of our article. We did find that complete obstruction is the only statistically significant predictor of the need for surgery. This can be especially helpful when faced with a patient with equivocal physical examination findings, but a CT reading of complete obstruction. It should be noted that the need for bowel resection was not dependent on the preoperative finding of complete versus partial obstruction. Interestingly, there were 21 patients who did not have previous surgical procedures. Yet, they had CT readings of SBO or PSBO. It is very common for CT scans of the abdomen to have been obtained for a wide variety of reasons when ordered in the emergency room. The presumptive diagnosis of SBO is seldom the reason for the request. The positive predictive value of the CT of the abdomen is important as it is related to SBO; however, because of the low threshold for ordering CT scans from the emergency room today, many other causes of abdominal pain will be presented to the radiologist (appendicitis, renal stones, pelvic inflammatory disease, and so on) that may present with regional small-bowel dilatation that must be differentiated from SBO. Megibow et al [16] pointed out that CT scans can be useful in patients who have not had prior surgery but present with signs of infection, bowel infarction, or palpable abdominal masses. In our series of 21 patients without prior surgery, 12 went to the operating room. Four of these patients had adhesions; 2 had ventral hernias; 2 had intraabdominal abscesses; and 1 each had acute cholecystitis, acute appendicitis, volvulus, and inflammatory bowel disease. The remaining 9 patients in this group had resolution of their acute conditions without operative intervention. More than likely, these patients had an ileus from the inflammatory processes. Despite the many advances in CT scans and the proven utility as a diagnostic aid in acute abdominal pain, this study ultimately shows that clinical findings are still the mainstay in diagnosing SBO. This should include white blood cell count, electrolytes, history, and, most importantly, physical examination. If a patient has peritoneal signs, he/she should be taken to the OR promptly without the need for radiographic studies. CT scans can be a useful adjunct when the patient has equivocal findings and as follow-up studies. Conclusion A CT scoring system can successfully help the surgeon decide whether or not the patient requires surgery 75% of the time. Likewise, specific criteria, when present in combination, can also aid in the decision-making process for operative intervention in 79% of cases with SBO. However, complete obstruction is the only reliable radiographic indicator for the need for operative intervention. Ultimately, a CT scan can be a guide for treatment, but the final decision should be based on clinical findings.
783
References [1] Faruqi R, Galland RB, Williams JM. An audit of surgical emergencies in the very old. Ann R Coll Surg Engl 1991;73:285– 8. [2] Silen W, Hein M, Goldman L. Strangulation obstruction of the small intestine. Arch Surg 1962;85:137– 45. [3] Foster NM, McGory ML, Zingmond DS, Ko CY. Small bowel obstruction: a population-based appraisal. J Am Coll Surg 2006;203: 170 – 6. [4] Bickell NA, Federman AD, Aufses AH. Influence of time on risk of bowel resection in complete small bowel obstruction. J Am Coll Surg 2005;201:847–54. [5] Shrock TR. Small intestine. In: Way LW, eds. Current surgical diagnosis and treatment. Norwalk, CT: Appleton and Lange; 1988: 561– 85. [6] Fevang BT, Fevang J, Strangeland L, et al. Complication and death after surgical treatment of small bowel obstruction: a 35-year institutional experience. Ann Surg 2000;231:529 –37. [7] Frager D, Medwid SW, Baer JW, et al. CT of small-bowel obstruction: value in establishing the diagnosis and determining the degree and cause. AJR Am J Roentgenol 1994;162:37– 41. [8] Daneshmand S, Hedley CG, Stain SC. The utility and reliability of computed tomography scan in the diagnosis of small bowel obstruction. Am Surg 1999;65:922– 6. [9] Flasar MH, Goldberg E. Acute abdominal pain. Med Clin North Am 2006;90:481–503. [10] Frager D. Intestinal obstruction: role of CT. Gastroenterol Clin North Am 2002;31:777–99. [11] Nauta RJ. Advanced abdominal imaging is not required to exclude strangulation if complete small bowel obstructions undergo prompt laparotomy. J Am Coll Surg 2005;200:904 –11. [12] Nagurney JT, Brown DF, Novelline RA, et al. Plain abdominal radiographs and abdominal CT scans for non-traumatic abdominal pain-added vale? Am J Emerg Med 1999;17:668 –71. [13] Mallo RD, Salem L, Lalani T, Flum DR. Computed tomography diagnosis of ischemia and complete obstruction in small bowel obstruction: a systematic review. J Gastrointest Surg 2005;9:690 – 4. [14] Frager D, Baer JW, Medwid SW, et al. Detection of intestinal ischemia in patients with acute small-bowel obstruction due to adhesions or hernia: efficacy of CT. AJR Am J Roentgenol 1996;166:67–71. [15] Ha HK, Kim JS, Lee MS, et al. Differentiation of simple and strangulated small-bowel obstruction: usefulness of known CT criteria. Radiology 1997;204:507–12. [16] Megibow AJ, Balthazar EJ, Cho KC, et al. Bowel obstruction: evaluation with CT. Radiology 1991;180:313– 8.
Discussion Jacques Heppell, M.D. (Scottsdale, AZ): It is nice to see that a quantitative approach by means of a computed tomography scoring system was tested to predict the need for surgery in small-bowel obstruction. We know that early recognition of intestinal strangulation is very difficult on conventional clinical grounds only. The treatment of choice for complete small-bowel obstruction remains early surgical intervention. However, nonoperative management of selected patients with patients with intestinal obstruction is sometimes clearly indicated, namely in the presence of partial obstruction, obstructive radiation enteritis, in the early postoperative period, and in presence of carcinomatosis or pseudo-obstruction. Most of these conditions except for partial small-bowel obstruction were excluded from your study. This may explain that 55% of your patients required surgery within 2 days of admission. When complete obstruction was found, 94% required a surgical intervention compared with 48% when obstruction was partial or incomplete. Ischemic bowel was found in 11% of cases. In patients who had prior surgery, small-bowel resection was needed in 51% of your cases. You found that the need for
784
A.J. Mangram et al. / The American Journal of Surgery 194 (2007) 780 –784
resection was not dependent on the length of time to surgical intervention, which in your study varied from 1 to 12 days. Irreversible ischemic changes require resection. There is discordance in your data regarding the need for bowel resection. Because the morbidity rate is higher in patients requiring resection, it would be nice to know the indications for resection in your study group and determine if computed tomography scan could predict not only the need for surgery but also the need for bowel resection. There are limitations to your scoring system. Two of the seven criteria, the presence of free air and closed loop, could not be used because of small numbers. Five criteria remained, one being dilated small bowel, which is a prerequisite for diagnosis. Only 40% of the patients had a repeat computed tomography scan to document improved or worsening obstruction. Finally, only 4 criteria remained to be tested: presence of transition point, ascites, and complete or partial obstruction. It is, therefore, not surprising to find that complete obstruction was the only reliable radiographic indicator for the need for surgical intervention. This finding concurs with randomized controlled trials on the use of water-soluble contrast in the management of adhesive small bowel obstruction. Although Gastrographin does not cause resolution of small-bowel obstruction, it can be used as a predictive test for nonoperative resolution and identify patients with complete obstruction needing early surgical intervention, thus reducing hospital stay. In my practice, I found computed tomography scan useful to pick up findings that I may have missed on physical examination such as the presence of incarcerated hernia in an obese patient. Indeed, 10% of your patients were found to have incarcerated hernia at laparotomy. A computed tomography scan can also detect unsuspected portal venous air and pneumobilia. Small-bowel volvulus is a life threatening surgical emergency associated with obstruction. It can be detected on computed tomography scan by the presence of the “spoke wheel” sign. In addition, the presence of intestinal intussusceptions with the classic “bull’s eye” sign can be recognized on computed tomography and lead to a timely surgical intervention. Finally, I would like to congratulate the authors for their
efforts and a study leading to another, I would like to ask them what modifications to their scoring system would be worth studying in a prospective manner to minimize the morbidity of bowel resection in patients with small-bowel obstruction. Dan Margulies, M.D. (Los Angeles, CA): I was concerned about the use of oral contrast in all of these. Can you comment whether you felt this contributed to any of those morbidities including the pneumonia and the deaths. It seems like the finding of complete obstruction could be determined without the use of contrast and because that was the only finding, are you are going to continue using oral contrast in all of your patients? Kory Jones, M.D. (Dallas, TX): Dr. Heppell, our indications for small-bowel resection intraoperatively were based on the feeling of the surgeon that was performing the surgery at the time. Unfortunately, that is not always documented on the operative report. Any evidence of ischemia or bowel necrosis was an indication, but, other than that, it is based on what the surgeon feels like based on the surgery. We did not find in our series any computed tomography findings that would indicate the need for small-bowel resection. Other articles have looked at computed tomography findings that would suggest bowel strangulation or ischemia and had a high sensitivity and specificity using computed tomography scans for that purpose. As far as modifications for perspective review, I believe that we would have to look more carefully at the different criteria with the scoring system and look at those in conjunction with each other and make sure that everybody is using a standardized scoring system if we are going to be looking at this prospectively. Other than that, I am not sure what modifications we would go by, except for increasing our number and making sure we were standardized. As far as the oral contrast, it was not used in all studies, and I am sorry if it seemed that way. The definition for complete obstruction was actually gas and/or oral contrast. So, in those patients, we did not have oral contrast. It was based on whether or not gas pass being a transition point to the distal bowel. And, we did not find any increased morbidities in those patients who did receive oral contrast.