PAPER. Defining the Role of Computed Tomography in Blunt Abdominal Trauma

Transcription

1 Defining the Role of Computed Tomography in Blunt Abdominal Trauma Use in the Hemodynamically Stable Patient With a Depressed Level of Consciousness Jay D. Pal, MD, PhD; Gregory P. Victorino, MD PAPER Hypothesis: Controversy exists regarding the use of diagnostic peritoneal lavage (DPL) vs computed tomography (CT) in the evaluation of blunt abdominal trauma. It has been suggested that one role for DPL is to diagnose bowel injuries in hemodynamically stable patients with an unreliable abdominal examination result. Our hypothesis is that CT is specific and sensitive for diagnosing hollow viscus injuries and is therefore an appropriate diagnostic modality in the hemodynamically stable blunt trauma patient with an unreliable abdominal examination result due to a depressed level of consciousness. Design: Retrospective consecutive case review. Setting: An urban level II trauma center. Patients: The medical records of 1388 consecutive patients admitted between January 1, 1991, and December 31, 2000, were reviewed. Inclusion criteria included blunt trauma patients who were hemodynamically stable (defined as a systolic blood pressure 90 mm Hg) with unreliable abdominal examination results secondary to a depressed level of consciousness (Glasgow Coma Scale score 11). Main Outcome Measures: Hollow viscus injury diagnosed by CT and missed diagnosis of hollow viscus injury by CT. Results: Of 1388 patients who met entry criteria, 87 had hollow viscus injuries; CT identified 85 of these Computed tomography diagnosed intestinal injury with a sensitivity of 97.7%, specificity of 98.5%, and an overall accuracy of 99.4%. Conclusion: At our institution, CT is a reliable and accurate diagnostic modality when used to evaluate hollow viscus injuries in the hemodynamically stable blunt trauma patient with an unreliable abdominal examination result due to a depressed level of consciousness. Arch Surg. 2002;137: Author s Note: Due to the peer review process, the title of the paper was changed. The initial title of the abstract that was presented at the Pacific Coast Surgical Association meeting was, Should Diagnostic Peritoneal Lavage (DPL) for Blunt Trauma Be of Historical Interest Only? From the Department of Surgery, University of California San Francisco, East Bay, Oakland. CONTROVERSY EXISTS regarding the use of computed tomography (CT) vs diagnostic peritoneal lavage (DPL) in the evaluation of blunt abdominal trauma. Diagnostic peritoneal lavage has been criticized for being an overly sensitive and invasive diagnostic tool. 1-4 Computed tomography has been criticized as an expensive study that has the potential to miss hollow viscus 1,3,5 The concern about hollow viscus injuries is increased due to an unreliable physical examination result in patients with a depressed level of consciousness. These patients may not be able to communicate regarding abdominal pain or identify abdominal tenderness, which may be important first clues of an occult hollow viscus injury. Some authors recommend DPL under such circumstances, whereas others suggest that CT is reliable. 2,3,6-8 We reviewed our experience in hemodynamically stable blunt trauma patients with a depressed level of consciousness. In part owing to the technological improvements and our increased experience with CT, we propose that CT is an appropriate singular diagnostic modality for evaluation of this diagnostically challenging patient population. MATERIALS AND METHODS The study patient population was drawn from admissions to the trauma service at Alameda County Medical Center in Oakland, Calif, from January 1, 1991, to December 31, Alameda County Medical Center is a university-based level II trauma center. A retrospec- 1029

2 Table 1. Patient Demographics* Demographic BIT No BIT P Value Age, y 36 ± ± M/F, % 59/41 55/45.33 ISS 15 ± 9 14 ± 8.48 GCS score 8 ± 3 9 ± 4.41 LOS, d 9 ± 6 11 ± 8.19 *Data are presented as mean ± SEM unless otherwise indicated. BIT indicates blunt intestinal trauma; ISS, Injury Severity Score; GCS, Glasgow Coma Scale; and LOS, length of stay. Table 2. Computed Tomographic Findings in Patients With Bowel Injuries Finding No. of Patients Sensitivity, % Specificity, % Free fluid Pneumoperitoneum Bowel-wall thickening Mesenteric hematoma tive review of the trauma registry, which includes patient demographics, emergency department assessment and treatment records, in-hospital radiographic evaluation, and surgical intervention, was undertaken to identify patients who met the entry criteria. Patients who sustained blunt abdominal trauma and were hemodynamically stable, defined as a systolic blood pressure greater than 90 mm Hg, and demonstrated a depressed mental status as defined by a Glasgow Coma Scale score of less than 11 were considered for enrollment. Initial abdominal and pelvic CT scans were performed within 2 hours after assessment by the trauma service (GE High Speed Advantage CT Scanner; General Electric Medical Co, Milwaukee, Wis). Axial scans with 1-cm cuts were obtained from the diaphragm to the femoral heads after intravenous infusion of Hypaque (Nocamed, Princeton, NJ) contrast after a delay of 70 seconds. Oral contrast was administered at the discretion of the trauma surgeon. The CT scans were initially reviewed by the trauma service with or without the assistance of attending radiologists. Subsequent to initial review, staff radiologists retrospectively reviewed all CT scans. Radiographic signs considered to be suggestive of blunt intestinal injuries were pneumoperitoneum, hemoperitoneum, extravasation of oral contrast, bowel-wall thickening, and focal hematomas. A CT scan result was considered positive if the radiologist s final transcribed report stated that the scan was suggestive of intestinal A patient was considered to have a hollow viscus injury if there was a bowel perforation documented in the operative report. Mesenteric injuries or serosal tears were not included in this group. The sensitivity, specificity, positive and negative predictive values, and the overall accuracy were calculated. RESULTS During the 10-year study period, 1388 consecutive patients met the entry criteria and had complete medical records available for review. There were 87 hollow viscus injuries documented, and CT identified 85 of these The mechanism of injury for these patients and distribution of bowel injuries are as follows: Injury Mechanism Patients, No. Motor vehicle crash 44 Fall 19 Assault 12 Bicycle 7 Automobile vs pedestrian 3 Other 2 Patient demographics of the study population are given in Table 1. There were no significant differences between the entire population and those patients who sustained intestinal The radiology reports were reviewed, and injuries by location were as follows: Location Jejunum 23 Ileum 21 Jejunal mesentery 17 Ileal mesentery 17 Right colon 12 Mesocolon 5 Duodenum 4 Transverse colon 2 Left colon 1 Numerous patients had multiple radiographic findings suggestive of bowel The most common location of injury diagnosed by CT in these patients was the jejunum, with the ileum a close second. Radiographic findings were also extracted from the radiology reports (Table 2). The most common radiographic finding in patients with intestinal injuries was free fluid in the peritoneum. Isolated free fluid was also the most common finding in patients without intestinal injuries and, thus, the least specific of radiographic signs. The most specific finding was pneumoperitoneum. The extravasation of oral contrast was always associated with intestinal However, the administration of oral contrast was not uniform throughout the study period because there existed a recent trend away from the use of oral contrast in blunt abdominal trauma patients. The locations of injuries found at laparotomy were as follows: Location Jejunum 47 Ileum 45 Right colon 15 Duodenum 8 Left colon 3 Transverse colon 1 There were no nontherapeutic laparotomies in this series. The more severely injured patients often sustained multiple intestinal The most common site of injury was the distal small bowel. The associated solid organ injuries were as follows: Organ Spleen 27 Liver 23 Kidney 13 Pancreas

3 Lacerations and hematomas of the liver and spleen were most commonly associated with blunt intestinal In the 2 patients in whom the CT scan did not identify a bowel injury, the injuries were identified clinically or by an additional CT scan within 24 hours of admission. The initial scans of these 2 patients were interpreted as having no obvious signs of a hollow viscus injury. On retrospective review and further discussion with staff radiologists, the CT scans were thought to demonstrate subtle signs of hollow viscus injuries, including bowel-wall thickening and free fluid. Seventeen CT scans reviewed by staff radiologists were interpreted as having signs of intestinal Isolated free fluid was the most commonly reported finding in 13 of 17 cases. All 17 patients were treated nonoperatively based on clinical judgment and had uncomplicated hospital courses, with a mean observation time of 6±5 days. These patients are thought not to have sustained blunt intestinal The DPLs were completed in only 3 patients enrolled in this study. All results of the 3 DPLs were positive. All 3 patients also had CT scans that were suggestive of hollow viscus injury, and in each case this was documented at laparotomy. The overall incidence of blunt intestinal injury in this series is 6%. Computed tomography diagnosed intestinal injury with a sensitivity of 97.7% and specificity of 98.5%. The positive and negative predictive values are 83.3% and 99.8%, respectively. The overall accuracy is 99.4%. COMMENT The delayed diagnosis of a hollow viscus injury can be a devastating complication and may lead to sepsis, multiorgan failure, and death. Trauma surgeons make every effort to diagnose these injuries in a timely fashion and often rely on DPL and/or CT to aid in their evaluation. The controversy lies in knowing what study to use and when. A criticism of CT is that it can miss hollow viscus 1,9 In the neurologically intact patient, a natural back-up system exists where the clinician can discover a bowel injury missed by CT through physical examination. However, in patients with a depressed level of consciousness whose physical examination result is unreliable, this back-up system does not exist. Therefore, to promptly diagnose all bowel injuries, it has been suggested that DPL may be superior to CT in hemodynamically stable patients with a depressed level of consciousness. 7,8 The DPL is regarded by some to be the standard screening test for intra-abdominal 10 However, DPL it is not without its disadvantages. A significant drawback of DPL is the high rate of nontherapeutic laparotomies. 2 In addition, DPL is invasive and may require several hours for the tests to be adequately sensitive for identifying small bowel 11 Because of improvements in imaging technology and our increased experience with CT, we hypothesized that CT is a reliable and accurate diagnostic modality in the hemodynamically stable patient with a depressed level of consciousness. We reviewed our experience in this clinical situation and found CT to have a sensitivity of 97.7%, a specificity of 98.5%, and an overall accuracy of 99.4%. These results are comparable to a recent study that looked at helical CT scans of all hemodynamically stable blunt trauma patients regardless of level of consciousness. 12 In their study, CT detected 64 of 68 bowel injuries for a sensitivity of 94%. However, in their study, CT had an accuracy of 86%, which was slightly lower than what we found. Improvement in diagnostic imaging technology may be one reason for increased sensitivity of CT as seen in our study. Later model scanners have less motion artifact and better resolution that may allow more accurate diagnoses. The sensitivity of CT as a screening tool for blunt intestinal injury can also be increased when certain diagnostic criteria for a positive scan result are used. 12,13 When radiographic signs, such as hemoperitoneum, bowel-wall thickening, extravasation of oral contrast, free fluid, and focal hematomas, are used as diagnostic clues, the sensitivity of CT is thought to be comparable to or greater than DPL The most frequent finding on CT in blunt abdominal trauma was free fluid. However, this finding was not sensitive or specific (80% and 42%, respectively). Pneumoperitoneum was much more specific (98%), but was not sensitive (78%). Bowel-wall thickening performed poorly when used to identify bowel injuries and had a sensitivity of 55% and a specificity of 47%. These figures are comparable to the study by Killeen et al 12 in which free fluid had a sensitivity of 76% and specificity of 39%, pneumoperitoneum had a sensitivity of 76% and specificity of 98%, and bowel-wall thickening had a sensitivity of 44% and specificity of 47%. Given the high diagnostic accuracy of CT in this study, of concern is the subgroup of 17 patients who were interpreted by attending radiologists as having intestinal injuries but did not undergo surgical exploration. Despite CT diagnosis of bowel injury, these patients were treated nonoperatively based on clinical judgment. Thirteen of these patients had isolated free fluid on CT. The presence of isolated free fluid has been suggested to be a strong indication for laparotomy because of therapeutic laparotomy rates of 94% and 54% found in 2 small studies 15,19 of 31 and 34 patients each. However, a more recent study 20 of 90 blunt trauma patients with isolated free fluid on CT showed that only 8% of these patients had intestinal injuries and 92% did not. The mean observation period for these patients was 8 days. The authors concluded that isolated free fluid on CT does not mandate celiotomy. The patients from our study who belonged to this group were observed for a mean of 6 days. All were discharged without consequence and are believed not to have sustained bowel We agree that isolated free fluid on CT does not mandate laparotomy. The major limitation of the current study is the inability to directly compare the results of DPL and CT scan in the evaluation of blunt trauma patients. Small patient populations have limited previous studies that have attempted to make this comparison partly because the incidence of blunt intestinal injury is approximately 5%. Also inherent in our retrospective design is the inability to assess improvement in patient care using our CT evaluation criteria. This study is also confounded by the variability owing to having several trauma attending physicians, 1031

4 each with a different management guideline, and the bias secondary to the predilection for CT at our institution. Another limitation is the lack of follow-up data on patients with negative CT scan results. However, this group of patients was observed in-house for a mean of 11 days, which is a reasonable period to confidently rule out hollow viscus We conclude that at our institution CT is a reliable, sensitive, and specific diagnostic modality when used to evaluate hollow viscus injuries in the hemodynamically stable blunt trauma patient with an unreliable abdominal examination result due to a depressed level of consciousness. We find additional diagnostic modalities, such as DPL, to be unnecessary and are not routine in our evaluation of this patient population. This paper was presented at the 73rd Annual Meeting of the Pacific Coast Surgical Association, Las Vegas, Nev, February 17, 2002, and is published after peer review and revision. The discussions are based on the originally submitted manuscript and not on the revised manuscript. Corresponding author and reprints: Gregory P. Victorino, MD, Department of Surgery, University of California San Francisco, East Bay, 1411 E 31st St, Oakland, CA REFERENCES 1. Davis JW, Hoyt DB, Mackersie RC, McArdle MS. Complications in evaluating abdominal trauma: diagnostic peritoneal lavage versus computerized axial tomography. J Trauma. 1990;30: Drost TF, Rosemurgy AS, Kearney RE, Roberts P. Diagnostic peritoneal lavage: limited indications due to evolving concepts in trauma care. Am Surg. 1991;57: Marx JA, Moore EE, Jorden RC, Eule J Jr. Limitations of computed tomography in the evaluation of acute abdominal trauma: a prospective comparison with diagnostic peritoneal lavage. J Trauma. 1985;25: Otomo Y, Henmi H, Mashiko K, et al. New diagnostic peritoneal lavage criteria for diagnosis of intestinal injury. J Trauma. 1998;44: Meyer DM, Thal ER, Weigelt JA, Redman HC. Evaluation of computed tomography and diagnostic peritoneal lavage in blunt abdominal trauma. J Trauma. 1989; 29: Henneman PL, Marx JA, Moore EE, Cantrill SV, Ammons LA. Diagnostic peritoneal lavage: accuracy in predicting necessary laparotomy following blunt and penetrating trauma. J Trauma. 1990;30: Wisner DH, Chun Y, Blaisdell FW. Blunt intestinal injury: keys to diagnosis and management. Arch Surg. 1990;125: Frame SB, Browder IW, Lang EK, McSwain NE Jr. Computed tomography versus diagnostic peritoneal lavage: usefulness in immediate diagnosis of blunt abdominal trauma. Ann Emerg Med. 1989;18: Marx JA, Moore EE, Jorden RC, Eule J Jr. Limitations of computed tomography in the evaluation of acute abdominal trauma: a prospective comparison with diagnostic peritoneal lavage. J Trauma. 1985;25: Ochsner MG, Knudson MM, Pachter HL, et al. Significance of minimal or no intraperitoneal fluid visible on CT scan associated with blunt liver and splenic injuries: a multicenter analysis. J Trauma. 2000;49: Root HD, Keizer PJ, Perry JF Jr. The clinical and experimental aspects of peritoneal response to injury. Arch Surg. 1967;95: Killeen KL, Shanmuganathan K, Poletti PA, Cooper C, Mirvis SE. Helical computed tomography of bowel and mesenteric J Trauma. 2001;51: Taylor CR, Degutis L, Lange R, Burns G, Cohn S, Rosenfield A. Computed tomography in the initial evaluation of hemodynamically stable patients with blunt abdominal trauma: impact of severity of injury scale and technical factors on efficacy. J Trauma. 1998;44: Harris HW, Morabito DJ, Mackersie RC, Halvorsen RA, Schecter WP. Leukocytosis and free fluid are important indicators of isolated intestinal injury after blunt trauma. J Trauma. 1999;46: Cunningham MA, Tyroch AH, Kaups KL, Davis JW. Does free fluid on abdominal computed tomographic scan after blunt trauma require laparotomy? J Trauma. 1998;44: Donohue JH, Federle MP, Griffiths BG, Trunkey DD. Computed tomography in the diagnosis of blunt intestinal and mesenteric J Trauma. 1987;27: Sherck J, Shatney C, Sensaki K, Selivanov V. The accuracy of computed tomography in the diagnosis of blunt small-bowel perforation. Am J Surg. 1994;168: Blow O, Bassam D, Butler K, Cephas GA, Brady W, Young JS. Speed and efficiency in the resuscitation of blunt trauma patients with multiple injuries: the advantage of diagnostic peritoneal lavage over abdominal computerized tomography. J Trauma. 1998;44: Brasel KJ, Olson CJ, Stafford RE, Johnson TJ. Incidence and significance of free fluid on abdominal computed tomographic scan in blunt trauma. J Trauma. 1998; 44: Livingston DH, Lavery RF, Passannante MR, et al. Free fluid on abdominal computed tomography after blunt injury does not mandate celiotomy. Am J Surg. 2001;182:6-9. DISCUSSION Felix D. Battistella, MD, Sacramento, Calif: Initial reports in the 1980s found CT scans lacked the sensitivity needed to reliably diagnose bowel injuries after blunt trauma. This study adds to the growing body of experience that bowel injuries can be accurately diagnosed with the newer generation of highresolution helical CT scanners. In the hands of experienced physicians, CT scans have a reported sensitivity of 88% to 94%; this study found a 98% sensitivity. However, after hours, scans are frequently interpreted by nonradiologists or less-experienced physicians. Initial misinterpretations that are corrected the next morning can lead to delays in diagnosing bowel Delays in treating bowel injuries, even as short as 8 hours according to a recently published study, are associated with a significant increase in morbidity and mortality. In your study, you used the radiologist s final interpretation to determine the sensitivity of CT scan. Did you find any differences between the initial and final interpretations of the CT scans? If so, were there delays in diagnosing and treating bowel injuries? DPL has the advantage of giving surgeons prompt, easy-to-interpret information with low morbidity. My second question deals with the 17 patients who had scans that were interpreted by the staff radiologist as having signs of intestinal injury but who did not have an operation. How did you make the decision to manage these patients with an unreliable physical exam nonoperatively? Were the CT findings missed initially and discovered after the patients had been followed for a period during which their neurologic status improved such that you were comfortable with the reliability of their physical examination? What is your recommended treatment for the patient who has an unreliable exam and free intraabdominal fluid without evidence of solid organ injury? I believe helical CT scans have a high sensitivity for blunt bowel injury, especially when interpreted by experienced physicians who can identify the sometimes subtle findings associated with intestinal However, a mentor once told me that the answer to a question in the title of a paper is almost always no. I suggest this applies to your paper. Despite the success of CT scan, I m not ready to eliminate DPL as a diagnostic tool. George C. Velmahos, MD, Los Angeles, Calif: Although I fully agree that the role of DPL is nearly none for hemodynamically stable patients, my major concern is the hemodynamically unstable patient. I saw in the authors algorithm that they substituted DPL with FAST [focused abdominal sonography for trauma]. In our hands, and despite the great results that are reported in the literature, FAST is not that reliable. So, I would like to ask the authors, in a patient with multiple injuries and hemodynamic instability, are they ready to dismiss the abdomen as the cause of the hemodynamic instability based on a negative FAST without using DPL? 1032

5 William P. Schecter, MD, San Francisco, Calif: I would like to commend the authors for bringing a clinically critical issue to our attention and that is, how can we exclude hollow viscus injury during the nonoperative management of patients with solid organ injury and free intraperitoneal fluid? This is an excellent study that supports the view that the current generation of helical CT scanners is correct most of the time. The problem is, every time one of these studies is presented, there are always a couple of patients with hollow viscus injury that are missed, and I continue to remain concerned during the nonoperative management of these patients for fear of missing this injury. We need to get an even better imaging study that will do for abdominal surgery what the CT scanner has done for neurosurgery. I don t think we are quite there yet, even with the current generation of helical CT scanners. James J. Peck, MD, Portland, Ore: In the 17 patients who had false-positives, would you do a DPL? Would that help you to make the diagnosis? Would you repeat the CT scan? Repeat CT scan 24 hours later is helpful in detecting evolving inflammation. How many of these cases did you explore for free fluid, but found actually the reason the CT scan was positive was a splenic hilar injury or some other bleed and the intestinal injury was incidental? Lawrence A. Danto, MD, Stockton, Calif: I am sure Dr Organ intended for this paper to be controversial. There is another way of looking at these studies, not just simply as finding indications for operation but also finding indications for observation. In my experience, one of the greatest values of DPL has been to absolutely indicate or identify the patient safe to observe. I wonder if you looked at this in your review. Gail Tominaga, MD, Honolulu, Hawaii: I have a few questions regarding the technique of CT scanning. What type of scanner was used on these patients and were the CT cuts every 10 mm, 5 mm, or 3 mm? Did you use oral contrast and, if so, how long did you wait after giving the oral contrast to scan the patients? Also, according to your algorithm at the beginning, you performed CT scans in all hemodynamically stable patients. Do you think there is a role for a FAST or something else beside CT scan in all hemodynamically stable blunt trauma patients? Claude H. Organ, Jr, MD, Oakland, Calif: The program committee chair, Dr Wilson, will permit Dr Victorino to close and I stand in my closing remarks just to say that I am not an author, but a sponsor of this. I thought after reading the paper it would create some discussion. I would like to thank the program committee for allowing a PGY-1 [postgraduate-year 1] like Dr Pal to present this to this group. It will stimulate other residents to do as he and Dr Sheldon did earlier today in these presentations. There is a negative side to doing reviews. All PGY-1s should do it, because they need to have the experience of looking at incomplete, poorly dictated operative reports and, second, to be able to give the cacography award each year to the trauma surgeon with the worst handwriting. Dr Victorino: Dr Battistella brought up the question about how we extracted the data from the radiologists, and it is in fact the final dictated report. So there may be some circumstances where the trauma team would read the CT scan in the middle of the night and the radiologist would alter that in the morning. We did not extract that information. We definitely agree that time is of the essence in diagnosing these We do know that in the 2 patients whose injuries we did miss, we picked these up within 24 hours and there were no sequelae from those delays. Dr Battistella also questioned the 17 false-positive patients and how we managed these patients. These patients were managed based on clinical judgment. Most of these patients had minor mechanisms, and the attending surgeon at that time felt that the mechanism was not of severe enough force to cause an intestinal injury. I will say that 13 of these patients had isolated free fluid as their only finding on CT scan. Initially, the problem of isolated free fluid on CT scans was thought to mandate surgical exploration based on 2 studies. These studies had positive therapeutic exploration rates of 90% and 50%. Subsequently, a larger multicenter study has been completed and found 90 patients with isolated free fluid. They reported that 92% of these patients did not have hollow viscus We like to observe these patients with isolated free fluid as the only finding on CT scan. We admit them for observation, follow the white count and temperature curve, and may repeat the CT scan if indicated. The next question was regarding FAST examination in a hemodynamically unstable patient. At our institution we have had good success with this. We have obtained a portable Sonosite, and our chief residents do a very good job of diagnosing the hemoperitoneum in the hemodynamically unstable patient. The FAST exam is repeatable in equivocal cases. To my knowledge, we have not missed any cases of hemoperitoneum in unstable patients. In those patients where the FAST exam is negative, they do go on to get a CT scan. Dr Schecter mentioned his nervousness in observing this particular patient population. I can t agree more. I get very nervous observing these patients, and we attempt to convey that nervousness to our residents. They are very diligent in following these patients and ensure that we don t miss any of these The next question was also about the 17 patients who were diagnosed falsely positive by CT scan and if we would recommend DPL or repeat CT scan in this situation. As I mentioned before, we would admit these patients for observation and if there were any question, we would repeat the CT scan. We would not recommend a DPL. There are some authors in the literature, however, recommending that DPL and CT scan are complementary in this particular scenario. A question was asked if any patients were explored for a solid organ injury and then subsequently a hollow viscus injury was found. That did not occur in any of our patients. There was a question about what type of scanner we use. We use the GE high-speed helical scanner. We do not use oral contrast. We have gotten away from that. We think that administration of oral contrast may add unnecessary time to getting these patients properly diagnosed and that it does not appreciably improve the radiographic diagnoses of bowel It may also increase the aspiration risk. 1033