1.     How do you decide which patients with abdominal trauma need a CT?

Setting aside those patients with obvious abdominal injuries, for whom most every Emergency Physician would order an abdominal CT (or send directly to the OR with trauma surgeons, depending on hemodynamic stability), the answer to this question is not so straightforward.

A physician’s physical examination of a patient has been found to be notoriously unreliable in detecting intra-abdominal injury (Schurink, 1997), especially in patients with altered mental status–secondary to either injury and/or intoxication–or with distracting injuries (Ferrera, 1998).  Given this, and the increased availability and accuracy of CT,  scans for abdominal trauma have increased greatly in number.  However, such scanning, despite its ability to readily identify injuries, comes with a cost, both financial and health-related. Outside of the problem of discovering incidental findings, radiation exposure is a real concern given that a 45 year-old undergoing a pan-scan incurs an estimated lifetime radiation-induced mortality risk of 1 in 1250 (Brenner, 2004).

So, how then to limit the number of abdominal CTs? A 2011 ACEP clinical policy on adult patients with acute blunt abdominal trauma looked for any clinical predictors that would allow a clinician to identify patients at low risk who may not need an abdominal CT.  Based on a literature review that included studies published through August 2008, they were able only to make a Level C recommendation that reads as follows:

“Patients with isolated abdominal trauma for whom occult abdominal injury is being considered are at low risk for adverse outcome and may not need abdominal CT scanning if the following are absent: abdominal tenderness, hypotension, altered mental status (GCS < 14), costal margin tenderness, abnormal chest radiograph,  hematocrit < 30% and hematuria (> 25 RBC/HPF) (Diercks, 2011).”

This recommendation was based largely on a Class II study by Holmes, et al., which evaluated adult patients with blunt torso trauma and who underwent CT scanning. Based on their findings they derived and then validated a clinical prediction rule for identifying any intra-abdominal injury.  The rule they developed was as written above, except that they also included the presence of a femur fracture as an indication to scan. They found these criteria together to have a sensitivity of 98.1% & 95.8%, and an NPV of 99.3% & 98.6% in the derivation and validation phases, respectively. Using these rules they found they could have avoided use of CT in one third of their patients (Holmes, 2009).

Other studies reviewed were considered Class III studies. One of these found that the following clinical criteria were sufficient to rule out any intra-abdominal injury, with a sensitivity and NPV of 99% if a patient met all the criteria: GCS > 13, normal abdominal exam (no tenderness or guarding), normal FAST and AP Chest x-ray, and normal lab data (Hct > 36%, WBC < 10,000/mm^3, AST < 50 IU/L).  The caveat was that this combination was found in only 11% of patients with a negative CT, and so is limited in its ability to reduce the number of CTs performed (Poletti, 2004).

An additional, prospective study attempting to establish a diagnostic algorithm for patients with blunt abdominal trauma has been published since the clinical policy working group did its literature search. Deunk, et al., derived (but have not yet validated) a rule that had a 97% sensitivity in a population of high-energy, blunt trauma patients. In their algorithm, if a patient had: stable ABCs/Vital signs (defined as RR between 10 and 29, SaO2 > 95, HR < 120, BP > 90); a GCS > 9 with no anisocoria or open skull fracture; a normal abdominal, pelvic, back and extremity exam; a base excess of > -3; and a normal FAST, CXR, Pelvic XR, and LS Spine xrays, then no abdominal CT was required. They found this would reduce CT usage by 22%.  The algorithm has not yet been validated (Deunk, 2010).

While intriguing, most of these rules are not helpful when a decision needs to be made right away about whether or not to scan a patient, as they require lab data to be returned and/or for non-standard trauma x-rays to be performed.  However, with point of care hemoglobin/hematocrit testing (or with a blood gas hematocrit that is returned quickly) and urine dip-stick availability, the criteria suggested by the ACEP guidelines could reasonably be performed in a trauma bay.

Bottom line: there are no data to date that conclusively answer this question or that allow for Level A or even Level B recommendations to be made.  Some physicians argue that all blunt trauma patients should be scanned; however, for those looking to find a way to reduce the number of unnecessary scans and their attendant radiation risks, the ACEP guidelines may be helpful in assessing the likelihood of intra-abdominal injuries.

2.  Do you order chest CTs on stable patients with focal chest wall tenderness and a normal AP Chest X-ray?

This is another commonly encountered scenario.  Of course, there are no studies that answer this question exactly.  However, there are several that examine imaging in chest trauma and the likelihood of injury based on various clinical parameters.

First, it’s worth noting that a supine AP chest radiograph, the canonical trauma x-ray, is not terribly sensitive for pneumothorax, with some studies reporting as low a sensitivity as 36-48% (Wilkerson, 2010).   A normal trauma chest x-ray does not rule out chest pathology.  Indeed, some studies suggest that because of this, all or almost all blunt trauma patients should undergo chest CT (Trupka, 1997).

A few studies have looked at what might indeed rule in or rule out significant chest injuries. One such study found that three factors–subcutaneous emphysema, lung contusion, rib fracture–as seen either on clinical exam and/or CXR were each independent predictors of occult pneumothorax (Ball, 2005).  Similarly, a literature review of predictors of intra-thoracic injury after blunt trauma found that the presence of an abnormal chest x-ray, an abnormal physical exam (chest/lung exam and/or coma), or an abnormal chest ultrasound was a significant predictor for chest injury seen in CT, and hence should prompt further imaging (Brink, 2009).

Other studies have tried to establish a decision instrument that would permit clinicians to forgo chest CT in blunt trauma patients.  In one, the authors evaluated 12 clinical parameters that may identify patients with no risk factors of severe intra-thoracic injury and found that the following seven parameters had a 99.3% sensitivity for identifying such injuries:  age > 60, chest pain, chest wall tenderness, distracting injury, intoxication, abnormal alertness/mental status, and rapid deceleration injury (fall > 20 feet or MVC > 40 mph).  The absence of any of these negated the need for further imaging, and would have resulted in a 14% reduction in chest imaging in the patients studied.  While this decision instrument holds promise for a modest reduction in CTs, it still needs to be validated (Rodriguez, 2011).

Still other studies suggest that a CXR and abdominal film may suffice in identifying clinically significant chest injuries, or that chest ultrasound is sufficiently sensitive for pneumothoraces (Wilkerson, 2010). Unfortunately, no rule or decision instrument  for effectively ruling out significant chest injury–as exists for the cervical spine and head trauma–has yet been successfully validated and embraced. Even as researchers work to create such a rule, there is debate among physicians as to the clinical significance of “occult” findings (i.e., those that are seen on CT and not on CXR),  such as occult pneumothoraces, subtle lung contusions, and non-displaced rib fractures, as most of these are treated with observation only (Yadav, 2010).

All these considerations will surely keep the debate of whether and when to scan the chest of a blunt trauma patient interesting. In the interim, physicians will need to continue to use their best judgment, perhaps guided in part by some of the partially studied parameters as listed above, in making this decision.

3. Do you pan scan in stable trauma patients. If so, when and why?

Not shockingly, the literature is split on this as well, with trauma surgeons coming down more on the side of pan scanning and emergency physicians attempting to be more selective in their scanning practices (Gupta, 2011; Tillou, 2009).  The pros for scanning are generally considered to be reduction in the number of missed injuries, faster identification (and hence treatment) of clinically significant (but perhaps not clinically apparent) injuries, and faster discharge for those who have a negative study.  Cons are primarily cost and radiation which, as stated above, is an estimated lifetime radiation-induced mortality risk of 1 in 1250 for a 45 year-old who is pan-scanned.

In one study advocating the pan-scan, Huber-Wagner found that integration of whole-body CT into early trauma care significantly increased the probability of survival in patients with poly-trauma.  The results of this study, however, are difficult to interpret, as they were based on derived, expected mortality values versus actual mortality.  There was actually no significant difference in mortality between those who were pan-scanned and those who weren’t. The pan-scanned group had a higher anticipated mortality rate (based on their injury severity scores) than the non-scanned group, and so the actual mortality rate was interpreted as reflecting improved relative survival for this group (Huber-Wagner, 2009).

Another article that strongly advocated pan-scanning multi-trauma patients was by Salim, et al., who claimed that scanning is appropriate/necessary in those trauma patients with significant mechanism even if they are “evaluable” (i.e. conscious) and have no visibly obvious external injury. This group found that scanning these patients changed treatment in 19% of the cases (Salim, 2006).  Though this article is widely referenced, it has some major methodological flaws, as pointed out by Snyder (2008) and others. Most notably, though the patients were “evaluable,” the patients weren’t actually evaluated. That is, no physical exam was performed. They did not assess for chest wall, abdominal, or cervical spine tenderness, nor did they ask whether the patient had pain anywhere. Hence, no clinical judgment was applied at all as to whether the patient required a scan other than whether there was clearly visible external injury. Also concerning was that several patients had a GCS between 5 and 8, some even requiring intubation. It is unclear how these patients were deemed “evaluable.”

In support of selective scanning is a study by Gupta, et al., who found that though some injuries may be missed by avoiding a pan-scan, few of these are critical.  He had both trauma surgeons and emergency physicians indicate which scans they desired for a given trauma patient, and evaluated the outcomes. ED physicians tended to want fewer studies, and consequently missed more injuries, though few of these required a change in pre-defined critical actions. Nevertheless, the surgeon and emergency physician authors of the paper disagreed on the importance of missing these injuries (Gupta, 2011).

4. Do you have clinical situations in which a negative FAST exam precludes the need for a CT?

According to ATLS, a FAST is indicated only in unstable trauma patients with an unidentifiable cause of hypotension, as only hemodynamically stable patients are able to have a CT scan due to its time-consuming nature. Hence, in just such an unstable patient, a (positive) FAST does indeed preclude the need for a CT.

However, ATLS is considered out of date with respect to these mandates. Helical CTs available today are much faster than the machines of a decade ago, thus allowing some marginally stable patients to also have CT imaging performed. Similarly, the use of the FAST exam has expanded to most trauma patients, irrespective of their hemodynamic stability.  Thus, today, stable or moderately stable patients may be found to have positive FASTs, yet these patients will generally get a CT before a decision is made about whether to take the patient to the OR. Still, the grossly unstable patient with a positive FAST will likely bypass the scanner en route to the OR.

Are there other scenarios when we might we safely use FAST as the sole abdominal imaging technique in evaluating a trauma patient? One group of authors state that if the patient is stable, has a negative initial FAST exam, and has no known possible risk factors for intra-abdominal injury (i.e., rib, pelvic, or spinal fracture; brief hypotension; hematuria; intoxication; persistent base deficit; head injury; distracting injury; or abdominal tenderness), then such a patient may forgo CT scanning and instead undergo serial examinations (Dunham,2008).

Of course, others disagree. A Cochrane review evaluating trauma algorithms that include ultrasound exams in patients with blunt abdominal trauma found that there is insufficient evidence from randomized controlled trials to justify promotion of ultrasound-based clinical pathways in diagnosing patients with suspected abdominal trauma (Stengel, 2005). Others have studied its use specifically in hemodynamically stable trauma patients and found it to have an unacceptably low sensitivity of 41% in this population (Natarajan, 2010).

Outside of the issue of its use in stable patients, it is important to remember that the FAST exam has limitations in identifying abdominal injuries: while good at identifying moderate hemoperitoneum, it is poor at detecting lesser amounts of blood in the peritoneum, bowel injuries, retroperitoneal bleeding, diaphragmatic tears, or bony injuries (Rhea, 2004). Hence, in a stable patient, one must have a very low suspicion for abdominal injury based on risk factors as listed above (and in question 1), to use FAST as the sole imaging technique.