Article, Surgery

Diagnosis of bowel and mesenteric injuries in blunt abdominal trauma: a prospective study

Original Contributions

Diagnosis of bowel and mesenteric injuries in blunt abdominal trauma: a prospective study

Fabrice Menegaux MDa,*, Christophe Tre’sallet MDa, Marylin Gosgnach MDb,

Quang Nguyen-Thanh MDa, Olivier Langeron MDb, Bruno Riou MD, PHDb

aDepartment of General Surgery, Groupe Hospitalier Pitie’-Salpe^trie`re, Assistance Publique-Ho^pitaux de Paris (AP-HP),

Universite’ Pierre et Marie Curie (Paris VI), 75651 Paris, France

bDepartment of Anaesthesiology, Critical Care, Emergency Medicine and Surgery, Groupe Hospitalier Pitie’-Salpe^trie`re, Assistance Publique-Ho^pitaux de Paris (AP-HP), Universite’ Pierre et Marie Curie (Paris VI), 75651 Paris, France

Accepted 15 May 2005

Abstract

Purpose: Currently, nonoperative management is the procedure of choice for solid organ injury in patients with a blunt abdominal trauma. Missed blunt bowel and mesenteric injuries (BBMIs) are possible because diagnosis is difficult. The aim of our study was to test a new algorithm for BBMI diagnosis using Abdominal ultrasonography (AUS), computed tomography (CT), and diagnostic peritoneal lavage (DPL).

Methods: We reviewed cases of blunt abdominal injuries over a 10-year period, then we designed an algorithm that was prospectively tested in Hemodynamically stable patients over a 2-year period. An abnormal AUS led to helical CT. When the CT showed more than 2 findings suggestive of BBMI, laparotomy was performed. In case of 1 or 2 Abnormal CT findings, we performed a DPL and calculated the ratio of white blood cells (WBCs) to red blood cells (WBC/RBC ratio) in the lavage fluid and divided this by the WBC/RBC ratio in peripheral blood. A ratio of 1 or higher was considered positive for BBMI, and a laparotomy was immediately performed. Patients with a ratio of less than 1 were managed nonoperatively.

Results: In the retrospective study, 26 (1%) of 2126 patients admitted to our trauma center for blunt trauma had a BBMI, including 15 (58%) diagnosed after a median delay of 24 hours. In the prospective study, 531 patients were admitted for blunt trauma with multiple injuries, including

131 with abdominal trauma. Computed tomography was performed in 40 patients. There were 2 criteria or more of BBMI in 1 patient, 0 criteria in 27 patients (with an uneventful follow-up), and

1 or 2 criteria in 12 patients who had DPL with a median ratio of 0.82 (ranges, 0.03-9). Five patients had a ratio of 1 or higher. They underwent immediate laparotomy. In all 5 cases, BBMI was found. The 7 patients who had a ratio of less than 1 were observed in ICU and treated for extra-abdominal injuries. No BBMI injury was missed in these patients. The accuracy of the algorithm was 100% (95% confidence interval, 0.99-1.00).

T Corresponding author. Service de Chirurgie Ge’ne’rale, Ho^pital de la Pitie’, 47-83 boulevard de l’ho^pital, 75651 Paris Cedex 13, France. Tel.: +33 1 42 17 66 32; fax: +33 1 42 17 66 07.

E-mail address: [email protected] (F. Menegaux).

0735-6757/$ – see front matter D 2006 doi:10.1016/j.ajem.2005.05.014

Conclusion: The proposed algorithm (based on AUS, CT, and DPL) had a high accuracy to diagnose BBMI while requiring the performance of DPL in only a few (2%) patients.

D 2006

Introduction

Injury to the bowel or mesentery during abdominal blunt trauma is rare, occurring in less than 1% of blunt Trauma victims [1-3]. This low incidence explains the limited experience of most trauma teams with this injury. Further- more, the identification of blunt hollow viscus injury is notoriously difficult in patients with multiple injuries in whom physical examination may be unrealistic. In recent years, because of marked advancements in technology, computed tomography (CT) has become the modality of choice to assess blunt abdominal injuries [4], leading to a dramatic increase in nonoperative management of solid organ injury [5,6]. Concomitantly, the possibility of missed blunt bowel and mesenteric injuries (BBMIs) is theoreti- cally increased by use of CT even if the overall incidence of missed injury is quite low and should not influence decisions concerning eligibility for nonoperative manage- ment [7]. Patients with missed BBMI can develop sepsis, multiple organ failure, and could potentially progress to death.

The aim of our study was to increase the accuracy of diagnosis of BBMI and to shorten the delay between the time of trauma and the diagnosis. For that purpose, we reviewed the cases of blunt abdominal injuries in our level I trauma center, seen over a 10-year period, and then we proposed an algorithm that was tested over a 2-year period in a prospective analyses.

Patients and methods

Retrospective study

An evaluation of BBMI in blunt trauma patients was retrospectively performed by examination of our trauma registry from January 1991 to December 2000. Initial assessment of blunt trauma patients was standardized. Immediately after their arrival in the hospital, physical examination was performed. During resuscitation and within 30 minutes, all patients underwent imaging assessment at the bedside as previously described [8], including chest and pelvic x-rays, and Abdominal ultrasonography . Hemodynamic instability with a large amount of free intraperitoneal fluid led to urgent laparotomy. Hemodynam- ically stable patients underwent additional Plain radiography of the spine and selected bones according to physical examination and most of them had an abdominal CT (helical CT since 1998). They were admitted for observation in the intensive care unit (ICU) and were considered candidates for nonoperative management. Assessment of vital signs,

physical examinations, fluid losses and requirements, labo- ratory tests, and additional AUS or abdominal CT was performed as necessary. Indications for delayed laparotomy included unexplained hemodynamic instability or sepsis, and abdominal x-ray or CT findings suggestive of BBMI.

Algorithm

An algorithm was developed for hemodynamically stable patients with blunt abdominal trauma. It was developed after reviewing the results of our retrospective study and the literature, and was agreed upon by consensus of an expert panel of surgeons and anesthesiologists.

If the initial AUS was normal, patients with multiple trauma had the assessment of other injuries and AUS was repeated 4 hours later. If the second AUS was normal, no further investigation was performed. An abnormal AUS in hemodynamically stable patients led to helical CT. We also performed a CT in patients having a direct abdominal trauma with a seat belt sign [9] even if the initial AUS was normal and in patients with a high-velocity trauma [10]. Specific findings indicative of BBMI at CT scan were pneumoperitoneum, unexplained intraperitoneal fluid, bow- el wall thickening, mesenteric fat streaking, and mesenteric hematoma. When the CT showed more than 2 findings suggestive of BBMI, a laparotomy was undertaken. In case of 1 or 2 abnormal CT findings, we performed a diagnostic peritoneal lavage (DPL) through a short Midline incision under local or general anesthesia. One liter of Ringer lactate solution was instilled into the peritoneal cavity through a catheter. Ten minutes later, peritoneal fluid was returned, and (after at least 700 mL back) a sample of effluent was analyzed for cell count, white blood cells (WBCs), and red blood cells (RBCs), and search for food fibers and bacteria. Diagnostic peritoneal lavage was completed with a cell count ratio dividing the WBC and the RBC ratio (WBC/RBC ratio) in the lavage fluid by the WBC/RBC ratio in peripheral blood [11]. A ratio of 1 or higher was considered positive for BBMI, and a laparotomy was immediately performed. Patients with a ratio of less than 1 were managed nonoperatively in the ICU (Fig. 1). The presence of food fibers and/or bacteria in the lavage fluid was also considered as positive for BBMI and led to urgent laparotomy.

Prospective study

This algorithm was prospectively tested during a 2-year period, from January 2001 to December 2002, in all hemodynamically stable patients with blunt abdominal trauma. A patient was considered stable when arterial

Fig. 1 Algorithm for the management of patients with blunt abdominal trauma. Cell count ratio is calculated by dividing the WBC/RBC ratio in the lavage fluid by the WBC/RBC ratio in peripheral blood.

systolic blood pressure was greater than 90 mm Hg on admission or after transfusion of a maximum amount of 4 RBC units in the first 24 hours. Patients were managed in the ICU by the trauma team, and Clinical decisions were obtained by consensus.

Statistical analysis

Data are expressed as mean F SD for normally distributed variables, or median (extremes) for non-gaussian variables. Using CT findings and cell count ratio, we assessed the diagnostic efficiency of the algorithm. Sensi- tivity, specificity, positive and negative predictive values, and accuracy (defined as the sum of concordant cells divided by the sum of all cells in the 2-by-2 table) and their 95% confidence interval (CI) were calculated, regarding the diagnosis of overall BBMI and the diagnosis of BBMI requiring therapeutic laparotomy. The odds ratios and their 95% CIs were calculated using the Confidence Interval Analysis 2.0 software (BMJ Books 2000, UK).

Results

Retrospective study

From January 1991 to December 2000, 2126 patients were admitted to our trauma center for blunt trauma. Among these patients, 769 (36%) had abdominal trauma, including

26 (1.2%) with a BBMI that was diagnosed at urgent

laparotomy in 11 patients or after a median delay of 24 hours (range, 8-720 hours) in 15 patients (58%).

There was no useful information for BBMI diagnosis to be gained in the group of patients who had their bowel injury diagnosed during laparotomy for massive hemoper- itoneum and hemodynamic instability. These patients had no preoperative Diagnostic procedure (2 patients) or a single AUS that showed only the presence of a variable amount of fluid in the abdomen (9 patients).

Among patients who had a delayed diagnosis of BBMI, 20% (3/15) were transferred from another hospital. Immedi- ate AUS was normal in 10 patients (67%) and showed a small or moderate amount of free intraperitoneal fluid in 4 patients and a significant hemoperitoneum (N500 mL) in 1 patient. A second AUS was done in 6 patients and showed free intraperitoneal fluid in all 6. A CT scan was performed in 11 patients for suspicious AUS (n = 2), abnormality on physical examination (n = 6), or both (n = 3). In 2 patients, the CT scan showed only extra-digestive signs (1 obstruction of the iliac artery from an intimal tear, 1 renal hematoma). In 1 patient, CT scan was initially considered as normal, but on review showed small bowel wall thickening. The most common finding was unexplained intraperitoneal fluid (n = 7) that was the only abnormality in 4 patients. Pneumoperitoneum was found in 2 patients, bowel wall thickening in 3 patients, and small Bowel distension in patients. Several abnormal CT findings were found in 3 patients and all 3 had unexplained intraperitoneal fluid. Four patients did not have CT scan because of abdominal

Fig. 2 Prospective study (cell count ratio is calculated by dividing the WBC/RBC ratio in the lavage fluid by the WBC/RBC ratio in peripheral blood).

guarding with fever, and there was significant hemoperito- neum on AUS in 2 patients transferred from another hospital, a pneumoperitoneum diagnosed on abdominal x-ray, and repeated episodes of Small bowel obstruction with sepsis

1 month after the trauma. All these patients underwent laparotomy. Some of them had several injuries: a small bowel perforation in 11, a duodenum perforation in 1, and mesenteric tears in 8 with small bowel or colon devascula- rization in 7 and 1 patients, respectively. Resection of the ruptured or ischemic bowel was performed in all cases but 4: exteriorization of the perforated bowel as a stoma in 3 unstable patients and 1 Primary repair of a perforation. In patients who had bowel resection, immediate anastomosis was performed in 6 and a double stoma in 5 because of hemodynamic instability during operation. Two patients (13%) died from extra-abdominal lesions. Among survi- vors, 2 patients developed Intra-abdominal abscess and needed iterative surgery. Six patients with a stoma survived. They had a closure of the stoma in a mean delay of 3.1 F 1.8 months after the initial surgical procedure.

Prospective study

From January 2001 to December 2002, 531 patients were admitted in our center for blunt trauma with multiple injuries, including 131 with abdominal trauma (25%). In this population, BBMI was diagnosed in 6 patients (1.1%). Thirty-four patients (26%) had immediate surgery for hemodynamic instability with significant hemoperito- neum. No severe BBMI was observed in these patients.

Seventy-seven patients had a second AUS and 40 had a CT scan that showed more than 2 criteria for BBMI in 1 case (3 criteria). This patient underwent surgery, which revealed a mesenteric tear with active bleeding. Suture repair was performed with an uneventful postoperative course. Twelve patients (9% of abdominal trauma and 2% of the entire population of blunt trauma) had 1 or

2 CT findings suggestive of BBMI. They all had DPL without incident. Neither food fibers nor bacteria were found in the lavage fluid. The median lavage to peritoneal blood count ratio was 0.82 (range, 0.03-9). Five patients had a ratio of 1 or higher. They underwent laparotomy with a median delay of 18 hours (range, 6-24 hours). All 5 patients had a BBMI. Two operations were therapeutic with suture for active bleeding from a mesenteric tear (ratio = 2) and a small bowel resection with 1-stage anastomosis for a bowel perforation (ratio = 9). In 3 cases, there were mesenteric hematoma and tears without active bleeding (ratio = 1.2, 1.6, and 3), and laparotomy was nontherapeutic. Therefore, positive pre- dictive value of our algorithm dropped from 100% (95% CI, 61%-100%) to 50% (95% CI, 20%-81%) for BBMI

diagnosis and usefulness of laparotomy for BBMI, respectively. Sensitivity, specificity, and negative predic- tive value were 100% in both studies. Postoperative course was uneventful in all cases. The 7 patients who had a ratio of less than 1 were observed in ICU and treated for extra-abdominal injuries. None of them developed Abdominal infection. No BBMI injury was

missed in patients who were treated nonoperatively (95% CI, 0%-0.7%) (Fig. 2).

Discussion

Our retrospective study illustrates the difficulties in diagnosis of BBMI because 58% were missed at initial evaluation. This injury is rare, and physical examination is unreliable in multiple blunt trauma patients who do not require urgent laparotomy. Although technical improve- ments increase the reliability of AUS, CT, and DPL, their individual sensitivity remains poor in the diagnosis of BBMI in hemodynamically stable patients [12]. Our prospective study using a combination of investigations showed that BBMI diagnosis can be improved using a very low rate of DPL (2%).

In stable blunt trauma patients, the trauma team has to balance 2 risks related to a delayed diagnosis of BBMI that are associated with significant increases in morbidity and mortality [12], or an unnecessary and nontherapeutic laparotomy with its own specific complications [13,14]. Because of our poor results in the retrospective study, and the lack of a definitive diagnostic procedure for BBMI, we decided to combine helical CT with DPL in a prospective diagnostic study. We deliberately chose not to use laparos- copy in our protocol because this procedure requires general anesthesia and has not been evaluated in Multiple trauma patients, especially in patients with severe brain injury. Moreover, its accuracy in the diagnosis of BBMI has been reported to be poor [15]. Therefore, this procedure cannot be currently recommended as a safe and accurate tool to assess patients with blunt abdominal trauma [16].

Computed tomographic scan is considered as the most reliable method for diagnosis of BBMI, although a series reports a 10% to 15% false-negative rate [12,17,18]. The use of oral and rectal Contrast material is controversial because the risk of aspiration may be significant, and this seems to be of small diagnostic gain [19]. The CT signs suggestive of BBMI have been described [20,21]. The most specific sign of bowel rupture is extravasated oral contrast material, but this is rarely observed. Other findings include pneumo- peritoneum, bowel wall thickening, mesenteric hematoma, mesenteric fat streaking, free intraperitoneal fluid, especially in an interloop location, and bowel wall enhancement with intravenous contrast. Malhotra et al [22] compared CT findings between 2 groups of patients who had from blunt trauma, with or without BBMI: unexplained intraperitoneal fluid was the most common sign but had a very poor specificity. We observed a similar result because this sign was found in 7 (64%) of 11 CT and was the only abnormality in 4 patients. The diagnostic value of pneumo- peritoneum remains controversial. Kane et al [23] reported that among 18 patients with free intraperitoneal gas on CT scan, only 4 (22%) had gastrointestinal perforation. Hamilton et al [24] reviewed 118 CT scans in patients with

blunt abdominal trauma: none of the 7 patients with pneumoperitoneum showed a bowel perforation, and none of the 3 patients with hollow viscus injury had intraperito- neal free gas. Malhotra et al [22] showed that the combination of CT findings increased the number of true positive, and thus elaborated an algorithm based on the number of CT findings: patients with a Normal CT scan did not need any further exploration, and those with 2 or more CT findings underwent laparotomy. Patients with 1 CT finding needed DPL, but criteria for positive DPL were not defined.

Another issue is that DPL has some rare but serious complications. An interesting result of our algorithm using DPL and cell count ratio was that only a few patients (2% in our series) needed this invasive procedure with a 100% specificity and sensitivity for BBMI diagnosis. However, the positive predictive value dropped to 50% for usefulness of laparotomy. To improve this positive predictive value, the threshold of the ratio could be raised, maybe from 1 to 1.5, but with a risk to decrease sensitivity.

Our study had some limitations because of the small numbers. Because BBMI is a very rare injury, the only way to precisely define the ideal threshold should be an evaluation in a large prospective multi-institutional study, using a receiver operating characteristic curve methodology [25].

The diagnosis of BBMI remains a challenge for physicians taking care of patients with blunt abdominal trauma. Currently, the most efficient approach is to combine different diagnostic modalities to improve their accuracy. The judicious use of commonly available and noninvasive investigations (AUS, CT) may reduce the risk of unnecessary laparotomies. Consequently, invasive diag- nostic modalities such as DPL can be reserved for selected patients.

References

  1. Watts DD, Fakhry SM, EAST Multi-Institutional Hollow Viscus Injury Research Group. Incidence of hollow viscus injury in blunt trauma: an analysis from 275,557 trauma admissions from the EAST multi-institutional trial. J Trauma 2003;54:289 – 94.
  2. Allen GS, Moore FA, Cox CS, et al. Hollow visceral injury and blunt trauma. J Trauma 1998;45:69 – 78.
  3. Wisner DH, Yong C, Blaisdell FW. Blunt intestinal injury. Arch Surg 1990;125:1319 – 23.
  4. Feliciano DV. Diagnostic modalities in abdominal trauma. Surg Clin North Am 1991;71:241 – 56.
  5. Croce MA, Fabian TC, Menke PG, et al. Non operative management of Blunt hepatic trauma is the treatment of choice for hemodynam- ically stable patients. Ann Surg 1995;221:744 – 55.
  6. Knudson MM, Lim RC, Oakes DD, Jeffrey RB. Non operative management of blunt liver injuries in adults: the need for continued surveillance. J Trauma 1990;30:1494 – 500.
  7. Miller PR, Croce MA, Bee TK, Malhotra AK, et al. Associated injuries in blunt solid organ trauma: implications for missed injury in nonoperative management. J Trauma 2002;53:238 – 44.
  8. Peytel E, Menegaux F, Cluzel P, et al. Initial imaging assessment of severe blunt trauma. Intensive Care Med 2001;27:1756 – 61.
  9. Chandler CF, Lane JS, Waxman KS. Seatbelt sign following blunt trauma is associated with increased incidence of Abdominal injury. Am Surg 1997;10:885 – 8.
  10. American College of Surgeon’s Committee on TRAUMA (ASCOT). Resources for Optimal care of the injured patient. Chicago7 American College of Surgeon; 1990.
  11. Fang JF, Chen RJ, Lin BC. Cell count ratio: new criterion of diagnostic peritoneal lavage for detection of hollow organ perforation. J Trauma 1998;45:540 – 4.
  12. Fakhry SM, Watts DD, Luchette FA, EAST Multi-Institutional Hollow Viscus Injury Research Group. Current Diagnostic approaches lack sensitivity in the diagnosis of perforated blunt small bowel injury: analysis from 275,557 trauma admissions from the EAST multi- institutional HVI trial. J Trauma 2003;54:295 – 306.
  13. Renz BM, Feliciano DV. Unnecessary laparotomies for trauma: a prospective study of morbidity. J Trauma 1995;38:350 – 6.
  14. Morrison JE, Wisner DH, Bodai BI. Complications after negative laparotomy for trauma: long-term follow-up in a health maintenance organization. J Trauma 1996;41:509 – 13.
  15. Elliott DC, Rodriguez A, Moncure M, et al. The accuracy of Diagnostic laparoscopy in trauma patients: a prospective, controlled study. Int Surg 1998;83:294 – 8.
  16. Leppaniemi AK, Elliott DC. The role of laparoscopy in blunt abdominal trauma. Ann Med 1996;28:483 – 9.
  17. Fang JF, Chen RJ, Lin BC, et al. Small bowel perforation: is Urgent surgery necessary? J Trauma 1999;47:515 – 20.
  18. Livingston DH, Lavery RF, Passannante MR, et al. Free fluid on Abdominal computed tomography without solid organ injury after blunt abdominal injury does not mandate celiotomy. Am J Surg 2001;182:6 – 9.
  19. Tsang BD, Panacek EA, Brant WE, Wisner DH. Effect of oral administration for abdominal computed tomography in the evaluation of acute blunt trauma. Ann Emerg Med 1997;30:7 – 13.
  20. Novelline RA, Rhea JT, Bell T. Helical CT of abdominal trauma. Radiol Clin North Am 1999;37:591 – 612.
  21. Brody JM, Leighton DB, Murphy BL, et al. CT of blunt trauma bowel and mesenteric injury: typical findings and pitfalls in diagnosis. Radiographics 2000;20:1525 – 36.
  22. Malhotra AK, Fabian TC, Katsis SB, et al. Blunt bowel and mesenteric injuries: the role of screening computed tomography. J Trauma 2000;48:991 – 1000.
  23. Kane NM, Francis IR, Burney RE, et al. Traumatic pneumo- peritoneum. Implications of computed tomography diagnosis. Invest Radiol 1991;26:574 – 8.
  24. Hamilton P, Rizoli S, McLellan B, Murphy J. Significance of intra- abdominal extraluminal air detected by CT scan in blunt abdominal trauma. J Trauma 1995;39:331 – 3.
  25. Hanley JA, Mc Neil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29 – 36.