August 2011, VOLUME 197
NUMBER 2

Recommend & Share

August 2011, Volume 197, Number 2

Neuroradiology/Head and Neck Imaging

Original Research

Screening Cervical Spine CT in a Level I Trauma Center: Overutilization?

+ Affiliations:
1 Department of Radiology, Henry Ford Health System, 2799 W Grand Blvd, Detroit, MI 48202.

2 Department of Emergency Medicine, Henry Ford Health System, Detroit, MI.

3 Department of Neurosurgery, Henry Ford Health System, Detroit, MI.

Citation: American Journal of Roentgenology. 2011;197: 463-467. 10.2214/AJR.10.5731

ABSTRACT
Next section

OBJECTIVE. The objective of our study was to analyze the use of screening cervical spine CT performed after trauma and establish the opportunity of potentially avoidable studies when evidence-based clinical criteria are applied before imaging.

MATERIALS AND METHODS. All cervical spine CT examinations performed in the emergency department of a level 1 trauma center between January and December 2008 on adult patients with trauma were analyzed; 1589 studies were evaluated. Radiology reports and clinical data were reviewed for the presence of fracture or ligamentous injury and for the mode of injury. We also looked for documentation of clinical criteria used to perform the CT study. In particular, we looked for mention of posterior midline cervical tenderness, focal neurologic deficit, level of alertness, evidence of intoxication, and clinically apparent distracting injury. These five criteria were established by the National Emergency X-Radiography Utilization Study (NEXUS) to identify patients with a low probability of cervical spine injury who consequently needed no cervical spine imaging.

RESULTS. Of the 1589 studies reviewed, 41 (2.6%) were positive for an acute cervical spine injury and 1524 (95.9%) were negative. The remaining 24 studies (1.5%) were indeterminate on the initial CT examination but subsequent imaging and clinical follow-up failed to show acute injury. Of the 1524 examinations with no acute injury, 364 (23.9%) had no documentation of any of the five NEXUS low-risk criteria.

CONCLUSION. The strict application of the NEXUS low-risk criteria could potentially reduce the number of screening cervical spine CT examinations in the setting of trauma in more than 20% of cases, thereby avoiding a significant amount of unnecessary radiation and significant cost.

Keywords: cervical spine, emergency radiology, NEXUS criteria, trauma

Blunt trauma with potential cervical spine injury is a frequent reason for presentation to emergency departments across the United States, with more than 1 million patients treated annually [1]. Cervical spine injury is estimated to occur in 2–10% of such cases, with more than 10,000 new cases each year [24]. The cost associated with these injuries is great, with an estimated $3.4 billion spent annually [2, 5]. A delay or failure to diagnose cervical spine injuries can produce disastrous consequences for patients, and as such, emergency physicians often have a low threshold for ordering cervical spine imaging [6, 7].

With the increasing emphasis on cost-effective health care delivery, the cost-benefit trade-off of advanced imaging is under increased scrutiny. One other downside to more widespread use of diagnostic imaging is increased patient radiation exposure. Although minimizing radiation exposure without compromising patient care is fundamental to the practice of radiology, the issue has recently become even more important as public awareness of medical radiation has increased. In the past year, The New York Times and other news organizations have published numerous articles discussing the potential dangers of medical radiation. One way to reduce radiation exposure is to reduce the number of patients undergoing medical imaging.

In 2000, the National Emergency X-Radiography Utilization Study (NEXUS) [6] established low-risk criteria to identify patients with a low probability of cervical spine injury who consequently needed no cervical spine imaging. To meet the NEXUS criteria [6], a patient must have the following: no tenderness at the posterior midline of the cervical spine; no focal neurologic deficit; normal level of alertness; no evidence of intoxication; and no clinically apparent, painful injury that might distract the patient from the pain of a cervical spine injury. Since its publication in 2000, the NEXUS clinical criteria have become one of the standard practices for determining the need for cervical spine imaging in trauma patients.

While the NEXUS investigators looked at the need for conventional radiography of the cervical spine, in recent years CT has supplanted conventional radiography as the standard of care for assessing cervical spine injury [4, 8]. According to the American College of Radiology (ACR) Appropriateness Criteria [9], the most appropriate imaging examination in a trauma patient with suspected cervical spine injury meeting the necessary clinical criteria (NEXUS or Canadian C-Spine Rule [CCR]) is unenhanced CT. For low-risk patients who do not meet the necessary clinical criteria, the criteria state that no imaging is indicated [9].

Although previous studies have evaluated the NEXUS criteria, few studies have addressed the performance and impact of the NEXUS criteria after publication in 2000 [1, 10]. The purpose of this study was to assess the ordering practices of emergency department physicians at a major level 1 trauma center when ordering screening cervical spine CT studies for blunt trauma cases and to assess whether strict application of the NEXUS criteria could have reduced the number of unnecessary studies.

Materials and Methods
Previous sectionNext section
Study Group

A search of radiology information systems was performed to find all cervical spine CT examinations performed in the Henry Ford Health System in the year 2008 of patients older than 18 years and with the following terms in the presenting history: “trauma,” “MVA (motor vehicle accident),” “rule out fracture,” and “assault.” A total of 2224 studies were identified. Of these, 419 were excluded because they were performed at satellite facilities that are not classified as level 1 trauma centers.

The remaining 1805 studies were then evaluated for the following exclusion criteria: patient had no documented trauma despite indication given on CT request, patient presented as an outpatient or as an inpatient (i.e., not to the emergency department), patient presented with remote trauma (> 48 hours before presentation), patient presented with penetrating injuries, and patient presented for follow-up examinations of a known fracture. An additional 216 studies were excluded on the basis of these criteria, so a total of 1589 examinations were included for the present analysis.

A detailed investigation of the radiographic and electronic-based medical records of these patients was performed independently by a second-year radiology resident and a third-year radiology resident to identify cervical spine fracture, dislocation, or subluxation. For our purposes, a positive study was one in which the radiologist's dictation indicated a fracture of any type, a dislocation, or subluxation based on the CT findings. A negative study had none of these findings. An indeterminate study was one in which the radiologist suggested a finding may be related to trauma or to another cause. In these cases, further imaging and medical records were reviewed to confirm the finding.

Historical and physical examination data from the emergency department documentation were evaluated for the presence of the five NEXUS criteria: focal neurologic deficit, midline cervical spine tenderness, intoxication status, distracting injury, and mental status as defined by the Glasgow coma scale (GCS). The patient was considered to have normal mental status if he or she was documented to be alert and oriented to person, place, and time or if there was no documentation of GCS. In addition, information regarding paravertebral cervical tenderness and painful or decreased cervical range of motion was also collected. Although this information is not part of the NEXUS criteria, the CCR does include the ability to actively rotate one's neck when determining the need for imaging. This additional clinical information, combined with the NEXUS criteria, will be referred to as the “liberalized NEXUS criteria” for the purpose of this study. Information pertaining to demographics, initial triage level, mechanism of injury, other pertinent radiologic findings, and patient disposition was also obtained.

The institutional review board approved this study.

Statistical Analysis

A chi-square test was used to determine the statistical significance of the difference between the incidence of detected injury in patients with no documented NEXUS criteria and the incidence of detected injury in those with documented NEXUS criteria.

Results
Previous sectionNext section

Of the 1552 patients evaluated, 921 (59.3%) were male and 631 (40.7%) were female. Thirty patients underwent multiple CT examinations for a repeat trauma during a separate examination: 24 patients twice, five patients three times, and one patient four times. The mean age of all patients was 43.4 years (age range, 18–100 years).

The initial triage level for all studies was in line with what would be expected for a level 1 trauma center: 696 patients (43.8%) triaged as level 1; 831 (52.3%), as level 2; and 56 (3.5%), as level 3. For the remaining six patients (0.4%), a triage level was not recorded. Triage acuity is a system of categorization of patients according to their need for emergent medical intervention. All patients presenting to the emergency department are assessed by a triage nurse and assigned a triage acuity on the basis of the chief complaint, physiologic criteria, airway assessment, ECG findings, neurologic signs and symptoms, GCS, location and mechanism of injury, and department-specific protocols.

The mechanism of injury primarily involved three types: fall (381/1589, 24%), assault (477, 30%), and motor vehicle crash (599, 37.7%). Pedestrian-versus–motor vehicle injuries (70, 4.4%) and a variety of miscellaneous injuries (62, 3.9%) made up the remaining of the studies.

Of the 1589 cervical spine CT examinations performed, 41 (2.6%) were positive for an acute cervical spine injury and 1524 (95.9%) were negative. Twenty-four studies (1.5%) were indeterminate on the initial CT examination but subsequent imaging and clinical follow-up failed to show an acute injury (Table 1).

Of the 1524 examinations with no acute injury, 364 (23.9%) had no documentation of any of the five NEXUS criteria and 308 (20.2%) documented no NEXUS criteria in addition to no record of limited or painful range of motion or posterolateral or paraspinal tenderness (i.e., the liberalized NEXUS criteria) (Table 1).

Of all the studies performed of patients with no positive NEXUS criteria (372 studies), 364 (97.8%) were negative for acute cervical spine injury, four (1.1%) were positive, and four (1.1%) were indeterminate but again failed to show acute injury either clinically or on follow-up imaging (Fig. 1). Of those patients with no documented liberalized NEXUS criteria (316 studies), the results were as follows: 308 (97.5%), negative; four (1.3%) positive; and four (1.3%), indeterminate.

Discussion
Previous sectionNext section

The goal of the NEXUS low-risk criteria is to identify trauma patients with a low probability of cervical spine injury, thereby sparing those patients from likely unnecessary cervical spine imaging [1, 6, 10, 11]. To meet the NEXUS criteria, a patient must have the following: no tenderness at the posterior midline of the cervical spine; no focal neurologic deficit; a normal level of alertness; no evidence of intoxication; and no clinically apparent, painful injury that might distract him or her from the pain of a cervical spine injury [6]. These criteria were validated in 2000 with a study in the United States involving 34,069 trauma patients. The results of that study showed the criteria to have a sensitivity of 99% and a specificity of 12.9% for identifying patients with cervical spine injury, although more recent studies have questioned those numbers [1, 6, 10].

TABLE 1: Fracture Results

In 2001, a second decision rule, the CCR was published. This rule uses three high-risk criteria (age ≥ 65 years, dangerous mechanism, paresthesias in extremities), five low-risk criteria (simple rear-end motor vehicle collision, sitting position in emergency department, ambulatory at any time, delayed onset of neck pain, absence of midline cervical spine tenderness), and the ability of the patient to actively rotate his or her neck to determine the need for cervical spine radiography [11]. The ACR accepts both criteria in their Appropriateness Guidelines [9] as a means of screening patients before imaging of the cervical spine.

In 2002, the American Association of Neurologic Surgeons made the following recommendation [12]:

Radiographic assessment of the cervical spine is not recommended in trauma patients who are awake, alert, and not intoxicated, who are without neck pain or tenderness, and who do not have significant associated injuries that detract from their general evaluation.

Yet, even with these recommendations, emergency physicians often have a low threshold for ordering cervical spine imaging [6, 7]. This fact may relate to the disastrous consequences that a missed cervical spine injury can lead to; however, it is important to note that with a highly sensitive test such as CT, both disastrous injuries and clinically inconsequential injuries will be detected. Still, although the cost—to the patient, physician, and society—is great for a missed cervical spine injury, liberal screening with CT has its costs as well, both financially and in terms of radiation exposure to patients. In fact, studies in Europe and the United States estimate that CT examinations make up approximately 15% of all imaging studies, but they account for 75% of the total population radiation dose [13]. Thus, even a small reduction in the number of cervical spine CT examinations could have a dramatic impact when extrapolated to the entire United States. For this reason, periodic studies, such as this one, are important to assess the effectiveness and utilization of both the clinical criteria and the radiologic studies being performed.

In our study, strict application of the NEXUS criteria before cervical spine imaging would have decreased the number of negative cervical spine CT studies by 23.9% (364 fewer studies). Applying a more liberal NEXUS criteria to include the presence or absence of pain, limited range of motion, or posterolateral cervical spine tenderness would have still resulted in a 20.2% reduction in the number of negative studies (308 fewer studies).

figure
View larger version (22K)

Fig. 1 Flow diagram illustrates breakdown of study subjects according to National Emergency X-Radiography Utilization Study (NEXUS) [6] low-risk criteria and CT results.

The study also looked at what effect imaging only patients with positive NEXUS criteria would have had on the incidence of detected injury. Of all patients who underwent CT for a potential cervical spine injury, the incidence of injury was only 2.6%. When only patients without NEXUS criteria were included, the incidence decreased to 1.1%. When only patients with positive NEXUS criteria were included, the incidence increased to 3%. Thus, the incidence of detected injury was almost three times greater when patients without NEXUS criteria were removed (p < 0.05). Although only a small increase was gained by excluding patients without NEXUS criteria (2.6–3%), this finding is likely a simple reflection of the very low overall incidence of injury in this patient population, an expected finding with a very sensitive but less specific test (i.e., NEXUS criteria).

Although this study was not specifically designed to test the actual performance of the NEXUS criteria, of the patients undergoing CT, the sensitivity of the NEXUS criteria was found to be only 90.2%, with a positive predictive value of 3%. This is significantly less than the sensitivity of 99% cited in the NEXUS study [6]. However, this figure is similar to those cited by Dickinson et al. [1] in 2004 and Stiell et al. [10] in 2003 who reported sensitivities of 92.7% and 90.7%, respectively.

The obvious downside to more strict application of clinical criteria before ordering screening CT examinations is the potential for more missed injuries. In this study, four patients with cervical spine injury had no documented NEXUS or liberalized NEXUS criteria in their charts. The four potentially missed injuries included a nondisplaced fracture of the transverse process of C1 with extension to a lateral mass, an isolated nondisplaced fracture of the transverse process and lamina of C7, a fracture of the lamina of C6, and an isolated fracture of the anterior process of C1. The mechanisms of injury in these cases were pedestrian-versus–motor vehicle collision, fall, and motor vehicle crash (two cases). Two of these patients were discharged from the emergency department with cervical collars, one was admitted and then discharged with a cervical collar, and one was admitted and then discharged with only physical therapy. No fractures without documented NEXUS criteria were unstable or required surgical intervention. Interestingly, if the CCR had been applied to those fractures that were missed by the NEXUS criteria, only one additional patient would have been detected (age ≥ 65 years). The remaining cases would have been missed by the CCR as well.

Because this study was based on a retrospective review of charts, the data are limited to information that was recorded by medical personnel at the time of each patient's presentation. An example of this limitation is that the presence or absence of posterior midline tenderness was recorded with much greater consistency than was pain with or limitation of with movement and posterolateral or paravertebral tenderness. An additional limitation, also related to the retrospective nature of the study, was the potential for reviewer bias because of the nonblinded review of medical records. However, given the objective nature of the information being collected, we think that this did not significantly affect the results.

A second major limitation, determining the presence or absence of a distracting injury, is partly because of the retrospective nature of the study and partly because of the inherent subjectivity of the NEXUS criteria. In the original NEXUS study, the authors chose not to define a “distracting injury” because they thought that doing so would be misleading. Instead, the authors “allowed the clinicians to judge whether the patients had an injury that could produce distracting pain” [6]. However, in a 2001 study testing the performance of the individual NEXUS criteria, the NEXUS authors, although still not defining a “distracting injury,” did provide examples of such, which included the following [14]:

(1) any long bone fracture; (2) visceral injury requiring surgical consultation; (3) a large laceration, degloving injury, or crush injury; (4) large burns; (5) any other injury producing acute functional impairment.”

The authors [14] went on to state that:

Physicians may classify any injury as distracting if it is thought to have potential to impair the patient's ability to appreciate other injuries.

Because of the retrospective nature of this study, the presence of distracting injuries had to be determined primarily by injuries that were radiographically apparent. Although some distracting injuries such as lacerations were certainly missed, many nondistracting injuries were almost certainly included because all bone fractures, including nasal bone and finger fractures, were considered distracting.

The way in which patients are often triaged in emergency departments is a third limitation. Quite frequently, when a patient arrives, a triage physician initially examines the patient and determines the need for medical imaging. However, the results of this encounter are sometimes not recorded and by the time a full evaluation is completed, the examination has changed. Thus, although the NEXUS criteria may have been present at the time the study was ordered, those findings may no longer be present at the time the physical examination findings are documented. This workflow also introduces the possibility of bias if the CT results are available before the recorded examination.

Despite the limitations of this study, the evidence is fairly convincing that even with its wide acceptance as a clinical screening tool for cervical spine injury, many patients continue to be imaged despite having no documentation of NEXUS criteria. This is exemplified by the approximately 24% of our patients without documented NEXUS criteria who underwent screening CT examinations in 2008. Thus, if clinicians had strictly adhered to the NEXUS criteria and the 2002 recommendations of the American Association of Neurologic Surgeons, nearly one fourth of patients undergoing CT would have been spared unnecessary imaging. The potential savings of this reduction, both in terms of radiation dose and health care money, are quite significant.

The question of “appropriate” utilization, the issue our study attempts to address, must take into account not only the potentially unnecessary studies, but also those injuries that would have been missed if those “unnecessary” studies had not been performed. The answer to that question is not so clear. Although a large number of CT examinations (24%) could have potentially been avoided by applying strict clinical criteria, a little over 1% of cervical spine injuries would have been missed. This then begs the question, how many missed fractures are acceptable to save a large number of patients from the financial and radiation burden of unnecessary imaging? Furthermore, because not all fractures lead to spinal instability, what if those missed fractures have little bearing on clinical management and outcome?

Regardless of the answers to the questions posed, with a nation currently focused on saving health care dollars and decreasing radiation exposure, the burden is on both radiologists and clinicians to address appropriate utilization of all imaging. Studies such as this one, although time-intensive, are essential to ensuring quality and safety because they provide a means of periodically evaluating the effectiveness of both clinical screening tools and the radiologic studies being performed. What this study makes clear is that further investigations into ways to increase the “appropriate” utilization of screening radiologic studies is warranted, whether in emergency departments or elsewhere in the health care system.

References
Previous sectionNext section
1. Dickinson G, Stiell IG, Schull M, et al. Retrospective application of the NEXUS low-risk criteria for cervical spine radiography in Canadian emergency departments. Ann Emerg Med 2004; 43:507–514 [Google Scholar]
2. Bailitz J, Starr F, Beecroft M, et al. CT should replace three-view radiographs as the initial screening test in patients at high, moderate, and low risk for blunt cervical spine injury: a prospective comparison. J Trauma 2009; 66:1605–1609 [Google Scholar]
3. Greenbaum J, Walters N, Levy PD. An evidence-based approach to radiographic assessment of cervical spine injuries in the emergency department. J Emerg Med 2009; 36:64–71 [Google Scholar]
4. Griffen MM, Frykberg ER, Kerwin AJ, et al. Radiographic clearance of blunt cervical spine injury: plain radiograph or computed tomography scan? J Trauma 2003; 55:222–227 [Google Scholar]
5. Blackmore CC, Emerson SS, Mann FA, Koepsell TD. Cervical spine imaging in patients with trauma: determination of fracture risk to optimize use. Radiology 1999; 211:759–765 [Google Scholar]
6. Hoffman JR, Mower WR, Wolfson AB, Todd KH, Zucker MI. Validity of a set of clinical criteria to rule out injury to the cervical spine in patients with blunt trauma: National Emergency X-Radiography Utilization Study Group. N Engl J Med 2000; 343:94–99 [Google Scholar]
7. Poonnoose PM, Ravichandran G, McClelland MR. Missed and mismanaged injuries of the spinal cord. J Trauma 2002; 53:314–320 [Google Scholar]
8. Barboza R, Fox JH, Shaffer LET, Opalek JM, Farooki S. Incidental findings in the cervical spine at CT for trauma evaluation. AJR 2009; 192:725–729 [Abstract] [Google Scholar]
9. American College of Radiology Website. American College of Radiology: ACR Appropriateness Criteria. Clinical condition: suspected spine trauma. www.acr.org/SecondaryMainMenuCategories/quality_safety/app_criteria/pdf/ExpertPanelonNeurologicImagingSpineTraumaUpdateinProgressDoc13.aspx. Published 1999. Updated 2009. Accessed August 13, 2010 [Google Scholar]
10. Stiell IG, Clement CM, McKnight RD, et al. The Canadian C-Spine Rule versus the NEXUS low-risk criteria in patients with trauma. N Engl J Med 2003; 349:2510–2518 [Google Scholar]
11. Stiell IG, Wells GA, Vandemheen KL, et al. The Canadian C-spine rule for radiography in alert and stable trauma patients. JAMA 2001; 286:1841–1848 [Google Scholar]
12. [No authors listed]. Radiographic assessment of the cervical spine in asymptomatic trauma patients. Neurosurgery 2002; 50(suppl 3):S30–S35 [Google Scholar]
13. Smith AB, Dillon WP, Lau BC, et al. Radiation dose reduction strategy for CT protocols: successful implementation in neuroradiology section. Radiology 2008; 247:499–506 [Google Scholar]
14. Panacek EA, Mower WR, Holmes JF, Hoffman JR. Test performance of the individual NEXUS low-risk clinical screening criteria for cervical spine injury. Ann Emerg Med 2001; 38:22–25 [Google Scholar]

Address correspondence to B. Griffith ().

Recommended Articles

Screening Cervical Spine CT in a Level I Trauma Center: Overutilization?

Full Access
American Journal of Roentgenology. 2003;180:911-915. 10.2214/ajr.180.4.1800911
Citation | Full Text | PDF (43 KB) | PDF Plus (117 KB) 
Full Access, , ,
American Journal of Roentgenology. 2000;174:713-717. 10.2214/ajr.174.3.1740713
Abstract | Full Text | PDF (1644 KB) | PDF Plus (319 KB) 
Full Access
American Journal of Roentgenology. 2001;177:677-679. 10.2214/ajr.177.3.1770677
Abstract | Full Text | PDF (1394 KB) | PDF Plus (1482 KB) 
Full Access
American Journal of Roentgenology. 2012;198:W417-W425. 10.2214/AJR.11.8150
Abstract | Full Text | PDF (950 KB) | PDF Plus (1003 KB) 
Full Access, , , ,
American Journal of Roentgenology. 2009;192:725-729. 10.2214/AJR.08.1420
Abstract | Full Text | PDF (467 KB) | PDF Plus (531 KB) 
Full Access
American Journal of Roentgenology. 1990;155:465-472. 10.2214/ajr.155.3.2117342
Abstract | PDF (1465 KB) | PDF Plus (589 KB)