Nonoperative Management of Blunt Abdominal Trauma: Have We Gone Too Far?

Focused Abdominal Sonography for Trauma

Abdominal ultrasonography is essential in the unstable patient for rapid detection of hemoperitoneum. The hemodynamically unstable patient with a positive focused abdominal sonography for trauma (FAST) examination should undergo laparotomy immediately. More generally, FAST had a sensitivity of 83.3% and specificity of 99.7% in 1,540 patients (1,227 with blunt injuries, 313 with penetrating injuries) [6]. The study was even more accurate in the evaluation of patients with precordial or transthoracic wounds (sensitivity 100%, specificity 99.3%) and hypotensive patients with blunt abdominal trauma (sensitivity 100%, specificity 100%) [6]. Sonography is an extension of the physical examination and should not interfere with the primary and secondary survey. The advantages of this technique are that it is cost-effective, noninvasive, and rapid; avoids ionizing radiation; and can be repeated as many times as needed. In fact, repeating the scan increases the likelihood of diagnosing intra-abdominal injuries. Blackbourne et al. performed a prospective observational study of 547 patients, finding an increase in sensitivity from 31.1% on the initial FAST to 72.1% (p < 0.001) on secondary ultrasonographic examination performed within 24 h of admission. The specificity was 99.8% for the initial scan and 99.8% for the secondary examination [7].

There are some limitations of FAST. It is unreliable in children, as this population generally has solid organ injury without hemoperitoneum on presentation [8]. Also, FAST is operator-dependent. The number of examinations a surgeon must perform to achieve acceptable accuracy ranges from 100 to 400 [9]. In addition, the false-negative rate has been reported to be higher in the setting of pelvic fracture, thoracolumbar spine fractures, hematuria, and rib fractures, settings in which the risk of associated intra-abdominal injury is higher [10].

Computed Tomography

Computed tomography scanning allows accurate diagnosis of solid organ injury and has facilitated the evolution of nonoperative treatment. A CT scan is both sensitive and specific; it has 98% accuracy in grading solid organ injuries and quantifying hemoperitoneum and allows evaluation of the retroperitoneum [11]. However, the use of CT scan in trauma has some caveats. It is not appropriate for unstable patients, has a risk of inducing renal failure secondary to contrast nephropathy, and misses 15% of hollow viscus injuries [11–13].

Diagnostic Peritoneal Lavage

Diagnostic peritoneal lavage has 98.5% accuracy. The study is sensitive, but not specific. It will detect hemoperitoneum reliably, but 25% of laparotomies were nontherapeutic when DPL was applied exclusively [14].

The general indications for laparotomy on the basis of DPL are 10 mL of gross blood, recovery of foreign material or intestinal contents, or a red blood cell count >100,000/mm³. The ratio of red to white blood cells has been suggested to be useful in the diagnosis of hollow viscus injury [15].

Advantages and Disadvantages of Nonoperative Management

The obvious advantage of nonoperative management is avoidance of the short-term risks of laparotomy: Anesthesia, iatrogenic injury, longer convalescence, and longer stays. The long-term risks of laparotomy also are avoided: Small bowel obstruction, incisional hernia, or overwhelming post-splenectomy infection in patients undergoing splenectomy. However, there are serious risks in the event of failure: Delayed bleeding or missed injuries. The key decision for selection of a patient for nonoperative management of solid organ injury is hemodynamic stability. As we discuss later, if this cardinal principle is not followed, patients are placed at risk unnecessarily [16].

Why is nonoperative management of solid abdominal organ injury so often successful? The simplistic explanation is that we now understand that in most patients, injuries of the liver and spleen will heal. Furthermore, the incidence of full-thickness intestinal injury, a difficult injury to diagnose, is low. In a multicenter study, Fakhry et al. [12] reported full-thickness hollow viscus injury in only 0.3% of 227,972 blunt-trauma admissions. However, although these injuries are infrequent, failure to recognize them in a timely fashion is catastrophic. In this particular series, neither physical examination nor any diagnostic test was reliable. Computed tomography missed 13% of intestinal perforations. In these patients, the mortality rate increased with the time to treatment. With prompt diagnosis and treatment (within 8 h), the mortality rate was only 2%, whereas delays of 8–16 h produced a 9% mortality rate (four-fold increase), and the mortality rate reached 31% (a 15-fold increase) when these injuries were discovered >24 h after admission [12]. Ten percent to twenty percent of patients with lap-belt marks on the abdomen or Chance fractures of the lumbar spine have associated intestinal or mesenteric injury [17]. Any suspicion of hollow viscus injury, change in hemodynamic status, change in abdominal pain pattern, increase in abdominal pain, or unexplained hyperamylasemia should alert the surgeon to the need for operation. In the case of moderate to significant free intra-abdominal fluid without solid organ injury, intestinal injury must be suspected [13]. In addition, as increasing numbers of solid organ injuries are detected in a patient with blunt trauma, the incidence of hollow viscus injury increases [18].

Blunt Hepatic Injury

The liver is injured frequently as a result of both blunt and penetrating thoracoabdominal trauma. Patients with hepatic injury generally present either with hemodynamic decompensation or hemodynamic stability. The key decision point is: “Is the patient stable enough for a CT scan?” If the answer is yes, the patient generally can be managed nonoperatively. Irrespective of the grade of the liver injury, if the patient is hemodynamically stable and the abdominal examination does not reveal an indication for operation, the patient can be observed [19,20]. High-grade hepatic injury, particularly when isolated to the right lobe or the “split liver,” will do well with observation [21]. In both adults and children, observation is successful in greater than 80% of blunt hepatic injuries [22]. As described by Richardson et al., reporting on 1,842 liver injuries treated over 25 years, the mortality rate for such injuries has declined significantly over the past several decades. The overall rate has fallen from 19% to 9%, and the rate related to liver injury has dropped from 12% to 5% [22]. Death from hepatic injury remains attributable to bleeding in 85% of cases. This important paper by Richardson et al. suggested four reasons for the decline in deaths from hepatic injury. First, better operative techniques are available for the management of major hepatic and venous injuries when patients require laparotomy. Second, damage control principles have been applied. Third, the utilization of angiography and embolization, both as a tool to avoid laparotomy and as a perioperative aid in hemorrhage control, has reduced blood loss. Fourth, surgeons operated much less often, despite the fact that the incidence of grade IV and V hepatic injury was consistent at 16% over the years of the study; and the authors suggested that observation prevented both the surgeon and the patient from getting into trouble in the operating room in attempts to control major liver injury and retrohepatic venous injury.

Wider use of interventional radiology and gastroenterology methods—angiography and embolization, percutaneous drainage of bilomas and abscess, and placement of biliary stents—also has facilitated successful nonoperative management and treatment of complications of liver injury. Approximately one fourth (25–27%) of hepatic injuries managed without operation will require an intervention to control bleeding or complications [23]. Interestingly, in this paper by Carillo et al., even when patients developed such complications, 85% were managed without laparotomy. Extravasation or an arterial blush on CT is a strong predictor of failure of observation with both hepatic and splenic injury [24–27]. Hemodynamic stability largely dictates whether angiography and embolization or laparotomy is the appropriate intervention.

When the patient is hemodynamically unstable secondary to a liver injury, prompt surgical exploration is mandatory. When operation is needed, it almost invariably means severe injuries to the liver (American Association for Surgery of Trauma Organ Injury Scale grades IV and V) and presents a difficult challenge (Table 1) [28]. The corollary is that grade IV and V liver injuries necessitate operation in as many as 55% of cases [29,30]. Given the success of packing and damage control, the trend over the last 20 years has been to minimize the amount and duration of surgery [22]. In addition, several reports from two decades ago documented a 55% mortality rate associated with hepatic resection for trauma [31,32]. Thus, a damage control approach and liver packing when needed, hepatorrhaphy, tissue sealants, and devices such as the argon-beam coagulator became the standard tools in the majority of cases.

When operating on a patient with liver injury, apply the simplest maneuver that controls the bleeding. If packing is effective in an unstable patient, perform an abbreviated laparotomy applying damage control principles. If bleeding persists, attempt control with tissue sealants and packing. These techniques certainly are effective when no major vascular structures are involved. On the other hand, when major vessels are injured or a large amount of parenchyma is compromised with persistent bleeding, nonanatomic or anatomic hepatic resections and direct vessel/duct repairs should be considered. If packing and hepatorrhaphy fail to control bleeding, the surgeon must change the operative plan quickly. It is critical to make the decision to perform anatomic or nonanatomic resection early enough that the need for blood products and the complications of extensive transfusion (coagulopathy, hypothermia, acidosis) will be reversible. The definitive operation must then be completed expeditiously with appropriate expertise, with the goal of minimizing the physiologic insult from extensive transfusion (Table 2). As recently as 2004, the operative mortality rate for grade IV and V injuries was 66%, with the majority of the deaths being from uncontrolled bleeding [30]. Furthermore, Menegaux et al. [33] reported that a conservative operative approach, avoiding anatomic or nonanatomic resection, resulted in a higher operative mortality rate, 34% vs. 24%. More recently, Polanco et al. [29] reported an overall mortality rate of 17.8% in patients with complex liver injury, with a rate of only 7% being related directly to the liver injury (bleeding), in a series of 56 hepatic resections in patients with complex blunt liver injuries (grades III, IV, and V). One third of these patients had associated major venous injury (inferior vena cava, hepatic or portal vein). These 56 patients were part of a larger group of 215 patients with complex hepatic injury. Two-thirds of the 215 patients required operation; only one-third were stable enough for observation. Strong et al. and Tsugawa et al. reported mortality rates of 8% and 24%, respectively, with anatomic resection for liver injury [34,35]. These studies confirm that 15–20% of blunt liver injuries are high grade. Such injuries often necessitate operation because of hemodynamic instability; these operations are challenging technically. Hepatic resection, both anatomic and nonanatomic, is required uncommonly, but has a role in the management of complex injuries and can be performed safely when appropriate expertise is available.

Blunt Splenic Injury

The spleen is the abdominal organ most commonly injured in such a way as to necessitate laparotomy for blunt injury. Splenectomy remained the standard of care for all splenic injuries for most of the 20th Century on the basis of several assumptions that we now know to be incorrect: That splenectomy had no consequence; that the spleen could not heal; and that delayed splenic rupture was common after injury. The spleen's role in protection from infection was demonstrated first in 1919 in a classic paper by Morris and Bullock [36]. A higher mortality rate was demonstrated in asplenic rats challenged with rat plague bacillus.

As with injuries of the liver, management of blunt splenic injury has changed dramatically over the past 20 years. A report on infants in 1952 [37], followed by Singer's review in 1973 of 2,795 asplenic patients [38], documented the adverse long-term immunologic consequences of splenectomy. This led to the concept of operative splenic preservation (splenorrhaphy) for the following two decades. By the 1980s, nonoperative management of blunt splenic injury was common in children. The development of CT provided an invaluable tool for the diagnosis and management of blunt injury, allowing evolution of nonoperative management. More than 90% of children with blunt splenic injury can be managed nonoperatively: The failure rate of observation in this situation is only 1–2% [39]. Nonoperative management developed more slowly in adults, with the criteria for safe nonoperative management being less well defined. Currently, the majority of blunt splenic injuries in adults are managed without operation [24,40]. In contrast to children, 15–30% of adults require immediate operation because of hemodynamic instability. In addition, the failure rate of observation in adults is 2–13%, higher than in children [24].

Hemodynamic stability is the key criterion for deciding on whether to use observation for blunt splenic injury in adults. Controversy exists regarding the impact of the grade of injury, the quantity of hemoperitoneum, the associated injuries, and patient age on the safety of nonoperative management (Table 3). The infrequency of grade IV and V splenic injuries at a single institution has made evaluation of factors critical to their management difficult. The Eastern Association for the Surgery of Trauma (EAST) Practice Management Guidelines state that “the severity of splenic injury (based on grade of splenic injury and quantity of hemoperitoneum) are not contraindications to nonoperative management” [19]. Although this statement may be true, large databases evaluating the natural history of grade IV and V splenic injury in adults confirm that these patients generally undergo splenectomy, and the failure rate of observation is high [16,40,41]. Watson et al., using the National Trauma Data Bank, reported 3,085 adults sustaining severe (Abbreviated Injury Scale score IV or V) blunt splenic injury from 1997 to 2003 [41]. Sixty-percent of these patients with high-grade injury went directly to the operating room. Nonoperative management was attempted in 40.5% of patients with these high-grade injuries, but ultimately failed in 54.6% of these. Failure of nonoperative management was associated with low admission systolic blood pressure, higher Injury Severity Score (ISS), and longer hospital and intensive care unit (ICU) length of stay.

Through the Multi-Institutional Trials Committee of EAST, 1,488 adult patients with blunt splenic injury at 27 trauma centers in 1997 were examined retrospectively [40]. Nonoperative failure was defined as an ultimate requirement for laparotomy after admission to the floor or ICU for observation. Sixty-one percent of the patients were admitted for observation, and failure of observation occurred in 10.8% of patients. The 39% of patients going immediately to the operating room had lower blood pressure, higher heart rate, lower Glasgow Coma Scale score, and higher ISS than the patients who were observed successfully. Interestingly, the patients who failed observation fell between these two groups, as judged by physiologic parameters. The risk of failure increased by grade: I = 5%; II = 10%; III = 20%; IV = 33%; and V = 75%. It is essential to realize that the failure rate in grade V injuries represented only 5% of these injuries, as 95% of the patients went directly to the operating room. Adding the grade V injuries requiring immediate operation and those that failed observation, 98.5% of these injuries ultimately necessitated laparotomy.

Does it make sense to attempt to observe an adult with a grade V splenic injury in view of the natural history of the injury? Increasing hemoperitoneum was an independent predictor of early or late need for laparotomy. In addition, within each grade, as the quantity of hemoperitoneum increased, the risk of failure of observation increased. The conclusions from this paper are: Higher-grade splenic injuries have more associated hemoperitoneum, and the risk of operation for blunt splenic injury in adults is determined by hemodynamic status, ISS, injury grade, and quantity of hemoperitoneum. Of the patients who failed observation, 61% did so in the first 24 h.

Were these patients mistriaged? What were the factors that predicted failure? These questions were addressed in a subsequent study [16]. Ninety-seven of the 1,488 adults with blunt splenic injury failed nonoperative management. Charts, including all hospital records, progress notes, nurse notes, emergency department records, CT reports, and prehospital and air medical reports with all identifying information removed, were requested on those patients. Eighty of the charts were returned and reviewed in detail. The group who failed observation could be separated on the basis of hemodynamic parameters. Forty-four percent of patients were always hemodynamically stable. Thirty-one percent were responders who may have had one or two episodes of hypotension or tachycardia, but then had stable vital signs. Twenty-five percent of patients admitted for observation who failed were hemodynamically unstable. Patients who failed observation declared this failure by declining hematocrit, greater abdominal pain, persistent tachycardia, or physician concern about another intra-abdominal injury. Importantly, 15% of the patients demonstrated failure of observation with acute hemodynamic decompensation. Ten patients failing observation of blunt splenic injury died (12% mortality rate). The mortality rate differed significantly within groups according to hemodynamics: stable, 3%; responders, 9%; and unstable, 39%. The ISS was significantly different in the stable (ISS = 22) and unstable (ISS = 31) patients; the ISS was 28 in the responders. Sixty percent of the deaths were preventable, attributable to delayed diagnosis and treatment of abdominal injuries. In this series, three patients exsanguinated in the hospital, two of whom never underwent operation. The factors producing preventable deaths were observation of unstable patients who clearly required operation; misreading of CT studies; and a 42% false-negative rate in the FAST studies (which were not always performed) [16]. This study clearly demonstrated that failure of nonoperative management does not involve solely delayed laparotomy; some patients will die as a consequence.

The role of angiography/embolization (A/E) in the management of blunt splenic injury remains controversial. Several studies have reported high success rates (>80%) and low failure rates (2–5%) [25,42,43], whereas other reports have shown little improvement utilizing A/E [26,44,45]. Patients more likely to fail observation despite A/E have high-grade splenic injury and significant hemoperitoneum, hypotension, or extravasation on arteriography [26]. Understand that a minority of patients undergo A/E—approximately 5–25% of those patients observed. As noted in the EAST studies, the multicenter paper from Western Trauma Association confirmed that patients with grade V injuries usually have already undergone splenectomy and are only uncommonly candidates for A/E. Only 4 of 140 patients undergoing A/E at four busy Level I trauma centers in the WTA study had grade V splenic injuries [42]. In addition, complications of A/E occur in 20% of patients and include failure to control bleeding (11–15%), missed injuries, and splenic abscesses [26,42,45].

Despite findings in earlier studies, older age is not a contraindication to nonoperative management [46]. In the multicenter series from EAST, 15% of patients were 55 years of age or older. Similar proportions of younger and older patients went directly to the operating room (41% vs. 38%). However, the mortality rate for patients older than 55 years was significantly greater than for younger patients (43% vs. 23%) [40]. This observation has been corroborated in other studies: Age as an independent predictor of death [46]. The failure rate for observation of splenic injury was 19% for those 55–64 years old, 27% for those 65–74 years old, and 28% for those >75 years old. This trend toward a higher failure rate with age was not statistically significant, however. The mortality rate was high regardless of management in these older patients, and failure of nonoperative management was associated with significantly longer hospital and ICU stays [46].

Conclusions

The paradigm shift in the management of blunt abdominal trauma has been to lesser invasiveness. Avoidance of a laparotomy with its short-term and long-term risks is of obvious benefit to the patient. Management of blunt splenic injury with observation and splenic preservation will avoid the lifelong risk of overwhelming postsplenectomy infection. The majority of blunt hepatic and splenic injuries can be managed nonoperatively. Most patients with blunt hepatic injury are hemodynamically stable and can be observed, irrespective of the grade of injury. On the other hand, high-grade hepatic injury (grades IV and V) often necessitates operation or management of complications. When hepatic injury mandates laparotomy because of hemodynamic instability, the operation is challenging technically. The risks of failure of nonoperative management of blunt splenic injury include preventable deaths. Safe nonoperative management requires adherence to cardinal surgical principles, examination and re-examination of the patient, and fastidious clinical decision-making.