Posted: June 5th, 2015
Is magnetic resonance imaging effective in the early detection of Alzheimer’s disease in asymptomatic adults?
School of Health & Life Sciences
Session 2013-2014
Student Name:
Student Matriculation Number:
Programme: Diagnostic Imaging
Level: MSc
Module Code: MMB722342
Module Title: Masters Framework Dissertation
Dissertation Title: An Investigation of the Use of Plain Film Radiography as Compared to CT Scanning in the Evaluation of Potential Cervical Spinal Injuries in Blunt Trauma Patients : A Structured Literature Review
Word Count:
Declaration
“This assignment is my own work. It has not been and will not be presented for assessment for any other module or piece of work which accrues credit for the award for which I am currently studying”. I have also submitted a copy of this work to Turnitin via GCU learn’
Abstract
Background/Aim
Traditionally, suspected cervical spine injury (CSI) in trauma patients has been commonly assessed with plain radiography. However, research conducted over the last two decades revealed that Computed Tomography (CT) is more efficient than plain x-ray films in the detection of CSI. The issue remains controversial and no general consensus has been reached if CT scanning or plain x-ray films should be used to image suspected CSI patients. Therefore, this structured literature review seeks to determine which among the two available modalities of plain radiography and CT scanining is more effective in detecting suspected CSI in patients that have suffered blunt trauma. On the basis of the findings, a recommendation will be put forward.
Methodology
Related literature was obtained by conducting a search of a range of healthcare databases, including Medline, AMED, CINAHL Cochrane library, Science Direct and EMBASE. In addition to this search, which was undertaken in the period 2004-2014, the references in related articles were also searched to amass as much primary quantitative studies as possible. The inclusion/exclusion criteria obtained from the research topic were applied to determine relevant studies. These were then subjected to the process of data extraction based on an individual form for every article. Furthermore, the quality of the included studies was evaluated with QUADAS-2. Since the results of the studies were not homogeneous, meta-analysis was not carried out.
Results
Only four articles were deemed appropriate enough to be included. The strengths and weaknesses of all studies were highlighted via critical assessment. In the case of plain radiography, the sensitivity estimates lacked homogeneity; on the other hand, CT exhibited consistently high estimates. It could therefore be realised that CT is more effective in detecting CSI than plain radiography.
Conclusion
CT can detect CSI more accurately than plain radiography. Nonetheless, the likely risk of injury of the patients influences which imaging method is more suitable for use. CT is recommended for patients whose risk of injury is moderate to high. However, in the case of patients with low risk of injury, CT may be unjustified as it is expensive and exposes patients to increased radiation doses. Yet the sensitivity of the x-ray films is not high enough to eliminate CSI in these cases. Given these drawbacks, identifying the efficiency of plain radiography and CT in the detection of CSI in blunt trauma patients requires systematic multicentre prospective cohort trails to investigate patients with low risk of injury by addressing a number of factors, including radiation dose, time, expenditure and availability of the two methods.
Acknowledgements
Though only my name appears on the cover of this dissertation, a great many people have contributed to its production. I owe my gratitude to all those people who have made this dissertation possible and because of whom my gratitude experience has been that I will cherish forever.
I would like to express my deepest gratitude to my advisors, Mrs. Jan Little and Mrs. Diane Dickson, for their excellent guidance, caring, patience, and providing me with excellent atmosphere for doing the research.
I would like to thank my parents, sisters and brothers as they were always supporting me and encouraging me with their best wishes.
I would also like to thank my wife, Najal Dhafer. She was always there cheering me up and stood by me through the good and bad times.
Finally, I’m grateful to all my friends who were willing to help and give best suggestion.
Table of content
List of Tables
| Number | Name | Page |
| 1 | Classification of CSIs | |
| 2 | Excluded studies | |
| 3 | Participants characteristics. |
List of figures
| Number | Name | Page |
| 1 | Diagram of atlas and axis | |
| 2 | Diagram of typical structure of the subaxial cervical spine vertebrae | |
| 3 | Summary of studies sensitivity. | |
| 4 | Summary of studies specificity. |
List of Abbreviation/Glossary.
| Terms | Abbreviations |
| Allied and Complementary Medicine Database | AMED |
| First cervical vertebra | C1 |
| Second cervical vertebra | C2 |
| Third cervical vertebra | C3 |
| Sixth cervical vertebra | C6 |
| Seven cervical vertebra | C7 |
| Cumulative Index to Nursing and Allied Health Literature | CINAHL |
| Centre for Reviews and Dissemination | CRD |
| Cervical Spine Injury | CSI |
| Cervical Spine Fracture | CSF |
| Computed Tomography | CT |
| Full cervical series | FCS |
| Multi-detector CT | MDCT |
| Medical Subject Heading | MeSH |
| Magnetic Resonance Imaging | MRI |
| National Emergency X-Radiography Utilization Study | NEXUS |
| Reporting Items for Systematic Reviews
|
PRISMA |
| Quality Assessment of Diagnostic Accuracy Studies | QUADAS |
| Randomised Control Trails | RCT |
| United Kingdom | UK |
| United State | US |
Chapter One
1 Introduction
‘Trauma’ comes from the Greek word meaning ‘wound’. As such, a straightforward description of the term ‘trauma’ would be: physical injury caused by an external physical force being applied. The application of this external force may be deliberate (e.g. in the case of physical assault) or accidental (e.g. in the case of a car crash) (Anderson et al, 2010). Anderson et al. (2010) and Benzel and Connolly (2012) stated that two types of trauma exist: penetrating trauma and blunt force trauma. The first of these types occurs when external force actually penetrates the body, for example, being shot or stabbed will result in penetrating trauma. As its name indicates, the application of blunt force, a force that does not penetrate the body, is responsible for blunt force trauma. Falls or motor vehicle crashes are examples of blunt force trauma (Anderson et al, 2010; Quann & Sidwell, 2011; Benzel, & Connolly, 2012).
Injury to the cervical spine must be borne in mind when someone suffers a blunt force trauma (White et al, 2007; Lindsey & Gugala, 2010). Diagnosing cervical spine injury (CSI) quickly and accurately is vital for patients to receive the best possible treatment (Mower et al, 2001; Ball & Watson, 2010; Lindsey & Gugala, 2010). Any delay in diagnosis can lead to serious consequences for both healthcare provider and patients (Ball & Watson, 2010; Lindsey & Gugala, 2010). These consequences include permanently disabling patients which leads to healthcare providers being forced to expend significant amounts on their care (Berlin, 2003; Ball & Watson, 2010). On average, one undiagnosed CSI costs US$2.9 million or US$5 million if the injury also caused neurological complications (Theologis et al, 2014).
The primary evaluation of patients for CSI consists of a comprehensive physical exam and an in-depth assessment of whether the injury meets the criteria to warrant a radiographic examination. If it is deemed necessary to image the cervical spine with radiographic examination then plain radiography is traditionally chosen as a diagnostic test for potential CSI patients (White et al, 2007; Heinzelmann et al, 2008; Kortbeek et al, 2008; Lindsey & Gugala, 2010). It is surprising then that studies by Diaz et al. (2003), Widder et al. (2004), Brohi et al. (2005), Gale et al. (2005), McCulloch et al, 2005 and Mathen et al. (2007) all revealed that plain radiography missed potentially serious injuries due to its relatively low levels of sensitivity (from 52% to 85%). Therefore, with the recent development of newer generation, high speed CT, CT scanners, plain radiography is often being overlooked as the once preferred choice for initial imaging tests of potential cervical spine injuries. (Widder et al, 2004; Theocharopoulos et al, 2009; Blackmore & Avey, 2011; Quann & Sidwell, 2011).
However, CT scanners produce much higher doses of ionising radiation resulting in special consideration being given to the thyroid gland. Also, CT scans are a considerably more expensive procedure when compared to plain film radiography. Such factors do not justify the routine use of CT scanning for the imaging of all suspected CSI patients (White et al, 2007; Kortbeek et al, 2008; Pneumaticos et al, 2014). Furthermore, there is no clear confirmation that using CT scans to diagnose CSIs leads to an increase in the detection of more serious cases than would be detectable by x-ray (Yates et al, 2007).
Currently, the process used to image blunt force trauma patients of CSIs is different from hospital to hospital. Widder et al. (2004) and McCracken et al. (2013) commented on the lack of a widespread set management policy for dealing with this issue. As a result, a structured literature review examining all of the materials and data available concerning how CSIs should be assessed in the first instance is required. This review can thus provide some valuable insights and can potentially be of use to healthcare providers and patients alike. This structured literature review aims to determine whether plain radiography or CT scanning is a more effective method for the detection of blunt force trauma patients for possible CSIs.
Chapter Two
2 Literature review
2.1 The anatomy of the cervical spine
Having a strong understanding of how the anatomy of the cervical vertebrae works is a vital part of correctly diagnosing CSIs as it can help practitioners to better comprehend the potential impact of the circumstances in which an injury has occurred. This can prevent injuries from remaining undiagnosed (Benzel & Connolly, 2012).
Gardner et al. (2005), Pimentel and Diegelmann (2010), and Pneumaticos et al. (2014) described the cervical spine as a complex arrangement of seven vertebrae, several joints and numerous related ligaments and muscles. Gardner et al. (2005) explained that this structure allows the head and neck to move in six directions. The cervical spine conveys the vertebral artery between second cervical vertebra (C2) and sixth cervical vertebra (C6) and the cervical cord is found within the spinal canal (Gardner et al, 2005; Pneumaticos et al, 2014).
The bones of the cervical spine consist of: the atlas first cervical vertebra (C1), the axis C2 (see Figure 1) and the subaxial spine (third cervical vertebra to seventh cervical vertebra (C3-C7)). Flexion/extension of the spine transpires first at the atlanto-occipital joint (found on the base of the skull between the occipital condyles) and the large superior facets of the lateral mass of the atlas (Tubbs et al, 2004; Gardner et al, 2005; Heinzelmann et al, 2008). In contrast to the other vertebrae which all have a vertebral body, the atlas is just a ring that enables the head to rotate through the peg or odontoid process (Gardner et al, 2005; Heinzelmann et al, 2008; Pneumaticos et al, 2014). According to Oliver and Middleditch (2005) and Heinzelmann et al. (2008), approximately half of all cervical spine rotation happens at the atlanto-axial joint. The transverse atlantal ligament (which moves between the lateral masses of the atlas) holds the peg against the front part of the atlas (Tubbs et al, 2004; Gardner et al, 2005), and the atlas also forms a joint with the axis through the base of the lateral mass of the atlas and the superior articular facet of the axis (Gardner et al, 2005). The distinctive feature of the axis is the peg which forms a joint with the atlas and the large spinous process of the atlas. The inferior articular facets are positioned posteriorly, the superior articular facets anteriorly (Gardner et al, 2005; Heinzelmann et al, 2008). Similar to the remaining sections of the subaxial spine, they slope anteroinferiorly. They are connected by the pars interarticularis, described by Gardner et al. (2005) as a narrow tube of bone.
Figure 2 shows the typical structure of the subaxial cervical spine vertebrae. Gardner et al. (2005), Heinzelmann et al. (2008) and Pneumaticos et al. (2014) explained that each disc, paired facet joint and anterior vertebral body carries the load of the head posterolaterally. As stated by Gardner et al. (2005), flexion/extension is enabled by the superoposterior direction of the inferior facets and the anteroinferior direction of the upper facets. They added that lateral flexion is a composite movement that requires rotation. The vertebrae are supported by sturdy ligaments. The posterior and anterior longitudinal ligaments run posterior and anterior to the vertebral body. A number of ligaments are found between the spinous processes posteriorly, notably the ligamentum nuchae which, as Gardner et al. (2005), Pimentel and Diegelmann (2010) have observed, connects the spinous processes running from the external occipital crest to the bifid spinous process of the second vertebrae of the cervical spine to C7. The ligamentum nuchae connects distally with the investing fascia covering the trapezius muscle. Muscles lie over the vertebrae and encircle the spine to provide stability. Gardner et al. (2005) and Pimentel and Diegelmann (2010) explained that the longus coli muscles, the sternocleidomastoid muscle and the strap muscles run the length of the spine anteriorly and the erector spinae groups runs posteriorly, connecting to the occiput.
Figure 1: Diagram of atlas and axis
Adapted from: (Gardner et al, 2005).
Figure 2: Diagram of typical structure of the subaxial cervical spine vertebrae
Adapted from: (Gardner et al, 2005).
2.2 CSIs: epidemiology and aetiology
In the available epidemiologic research on patients with CSIs, the focus is generally on referral centres and admitted patients. Agrawal (2008) argued that this study population may not be representative of patient patterns or trauma treated in the majority of casualty departments. In response, a different range of papers has been examined to ensure a broader spectrum of patients is taken into consideration.
Research by Ozsarlak (2007), Simpson and Harris (2013) and Pneumaticos et al. (2014) revealed that over 50% of all spinal injuries are to the cervical spine, with the total percentage of blunt force trauma patients suffering from CSI ranging from 2% to 4.5%. Ozsarlak (2007) observed that in patients with multiple injuries this percentage increases to 5.9%.
Looby and Flanders (2011) and Pneumaticos et al. (2014) showed that CSIs, inclusive of damage to the ligaments and fractures, are chiefly secondary injuries of neck and head trauma. Fractures at the lower and upper points of the cervical spine make up the majority of CSIs. Looby and Flanders (2011) and Pneumaticos et al. (2014) showed that C6 and C7 vertebral fractures made up around half of all cervical spine fractures, C2 fractures around one third, the C2 peg a little under one sixth, and C1 vertebral fractures around one-tenth. According to Ozsarlak (2007) and Looby and Flanders (2011), if the unstable cervical fracture of a patient is not stabilised they are in danger of suffering a cervical spinal cord injury. Looby and Flanders (2011) stated that 0.7% of spinal cord injuries include no fracture while 3% of cervical spine fractures include no spinal cord injury.
Pneumaticos et al. (2014) showed that women are 400% less likely than men to suffer a CSI because that women are normally less involved in car accidents and sport activitieswhich could cause CSI. The authors also added that individuals in their thirties and sixties are at greatest risk of CSI.
Ozsarlak (2007) showed that 40% of CSIs in blunt force trauma patients are the result of vehicular collisions, 20% are due to falls and 14% to sporting activities. Furthermore, Ozsarlak (2007) and Pneumaticos et al. (2014) showed that older individuals are more likely to suffer a CSI due to a fall, younger individuals due to a sporting injury or vehicular accident.
Heinzelmann et al. (2008) affirmed that CSIs are linked to high rates of permanent neurological complications, morbidity and mortality despite accounting for only a limited number of all injuries. In the USA, for instance, around 5,000 new instances of quadriplegia and 6,000 deaths per annum are the result of CSIs (Ozsarlak, 2007; Pimentel & Diegelmann, 2010). One fifth of the deaths caused by road traffic collisions are due to serious CSI, mostly to the upper cervical spine (Ozsarlak, 2007).
2.3 CSIs: classification, biomechanical instability and pathophysiology
Pimentel and Diegelmann (2010) noted that a number of CSI classification systems are in use but no single one has been uniformly accepted by practitioners or researchers. Ozsarlak (2007), Pimentel and Diegelmann (2010), Looby and Flanders (2011) and Benzel and Connolly (2012) each listed different classification systems for CSIs including fracture instability, morphology, level of injury and trauma mechanism. However, Ozsarlak (2007) and Pimentel and Diegelmann (2010) noted that determining the precise mechanism of trauma in a CSI is challenging and the complex nature of certain single traumas shows the existence of a number of injury mechanisms.
Gardner et al. (2005), Pimentel and Diegelmann (2010), and Pneumaticos et al. (2014) illustrated that the two highest cervical vertebrae are anatomically and biomechanically different from the third to seventh vertebrae. It can be seen that in the majority of research a mixture of classification methods is employed, for example, injury level and mechanism of trauma then morphological description and stability evaluation (Ozsarlak, 2007; Pimentel & Diegelmann, 2010). An overview of the CSIs classified according to injury level, fracture stability and mechanism of trauma can be seen in Table 1.
According to Ozsarlak (2007), the trauma mechanism of 4% of CSIs is hyperextension, of 6% lateral rotation, vertical compression and flexion-rotation account for 12%, extension for 20% to 38% and flexion for 46% to 79%. The dominant motion of the cervical spine is flexion-extension, despite its anatomical properties allowing motion in all planes (Ozsarlak, 2007; Benzel & Connolly, 2012). The trauma mechanism is significantly affected by the direction of the forces responsible for the injury and the situation of the neck and head during impact (Ozsarlak, 2007; Looby & Flanders, 2011; Benzel & Connolly, 2012).
Looby and Flanders (2011) and Benzel and Connolly (2012) affirmed that an evaluation of spinal stability must be used in combination with every classification type. This is because particular treatments for CSI are determined based on the biomechanical and clinical stability/instability of the cervical spine.
In imaging the spine, the status of a spinal trauma as stable or unstable is crucial. Heinzelmann et al. (2008) and Looby and Flanders (2011) stated that a mechanically unstable spinal injury exposed to a normal spectrum of movement and physiologic loading can be subject to damaging deformation. A number of different identification systems for spinal stability have been put forward, the most widely employed being, as noted by Ozsarlak (2007) and Looby and Flanders (2011), the three-column concept devised by Dennis. This theory aids in predicting stability in regards to various patterns of spinal trauma. In the three-column concept, the spinal column is separated into posterior, middle and anterior columns (Ozsarlak, 2007; Looby and Flanders, 2011). The posterior column includes: the ligaments (e.g. the ligamentum flavum, the facet joint capsule and the interspinous and supraspinous ligaments) and the posterior bony elements (e.g. the lamina, the spinous processes and the pedicles). The middle column includes: the posterior annulus, the posterior vertebral body and the posterior longitudinal ligament. The anterior column includes: the anterior annulus fibrosis, the anterior vertebral body and the anterior longitudinal ligament(Ozsarlak, 2007; Looby & Flanders, 2011).
An injury is deemed mechanically unstable if two or more of the three columns are disturbed (Looby & Flanders, 2011; Benzel & Connolly, 2012). According to Looby and Flanders (2011), radiologists must be descriptive when discussing a spinal trauma/fracture and its morphology. By doing so radiologists will be able to accurately convey to the clinician the stability/instability of the trauma and its particular type. Ozsarlak (2007) and Looby and Flanders (2011) enumerated the radiological features of a mechanically unstable spinal fracture as follows: broadening of the intervertebral canal; vertebral body height reduction of more than half; disturbance to the posterior vertebral body line; broadening of the interspinous space, the interpediculate distance and/or the facet joints; kyphosis of more than 20 degree; and translation/displacement of more than 2mm suggesting disturbance of the ligaments. Physicians are able to choose a suitable treatment (surgical or conservative) if they understand the mechanical and neurological stability of an injury, as determined by suitable diagnostic methods (Ozsarlak, 2007; Looby & Flanders, 2011; Benzel & Connolly, 2012 ).
Table.1. Classification of CSIs
|
Mechanism Type Level Stability |
|
| Flexion injury Simple wedge fracture Any level Stable | |
| Flexion teardrop fracture Any level Unstable | |
| Anterior subluxation Any level Stable | |
| Bilateral facet dislocation Any level Unstable | |
| Clay shoveler’s fracture Lower cervical Stable | |
| Flexion-rotation injury | Unilateral facet dislocation Any level Stable |
| Rotatory atlanto-axial dislocation C1–C2 Unstable | |
| Extension injury | Hyperextension dislocation Any level Unstable |
| Hangman’s fracture C2 Stable/unstable | |
| Extension teardrop fracture Any level Stable/unstable | |
| Isolated posterior arch fracture Any level, usually C1 Stable | |
| Anterior arch avulsion fracture C1 Stable | |
| Laminar fracture Any level Stable | |
| Hyperextension-lateral rotation injury | Pillar fracture Any level Stable |
| Pedicolaminar fracture – separation types I, II, III Any level Stable | |
| Pedicolaminar fracture – separation types IV Any level Unstable | |
| Vertical compression injury | Jefferson’s fracture C1 Unstable |
| Burst fracture of vertebral body Lower cervical Stable/unstable | |
| Craniocervical injury | Atlas fractures types I, II, IV C1 Stable |
| Atlas fractures type III C1 Stable/unstable | |
| Atlas fractures types V (Jefferson’s fracture) C1 Unstable | |
| Atlanto-axial subluxation-distraction C1 Unstable | |
| Atlanto-occipital dislocation-distraction C1 Unstable | |
| Odontoid process fracture types I, III C2 Stable | |
| Odontoid process fracture types II, IIA C2 Unstable | |
| Occipital condyle fracture types I, II Occipital bone Stable | |
| Occipital condyle fracture type III Occipital bone Unstable | |
| Adapted from: (Özsarlak, 2007). |
2.3 Blunt trauma patients and cervical spine imaging
2.3.1 Indication for cervical spine imaging
It is widely accepted that it is difficult to diagnose a patient of cervical spine injury or establish whether they should be provided with cervical spine imaging. As a result, a significant amount of research has tried to create suitable criteria for determining whether cervical spine imaging is needed following blunt force trauma. However, it was not until the National Emergency X-Radiography Utilization Study (NEXUS) by Hoffman et al. in 2000 and the Canadian Cervical Spine Rule study by Stiell et al. in 2001 that any generally accepted protocols were established.
Hoffman et al. (2000), Ozsarlak (2007), Tins et al. (2007) Looby and Flanders (2011) and Pneumaticos et al. (2014) illustrated that in the NEXUS for low-risk criteria, trauma patients will require cervical spine radiography unless they meet the NEXUS low-risk criteria which is illustrated in appendix 1. This list of criteria can be used to image close to all unstable CSIs in most patients. The list test has a 99.8% negative predictive value, meaning that if a patient meets all of the criteria they have a very small chance of having a CSI. Ozsarlak (2007) and Looby and Flanders (2011) showed that the sensitivity of the decision rule is 99%, however, the positive predictive value of the test is only 2% given the criteria’s low level (12.9%) of specificity.
The Canadian Cervical Spine Group carried out a similar study. They concluded that the risk of CSI following blunt force trauma is low if the patient meet their criteria which is illustrated in appendix 2 (Stiell et al, 2001; Stiell et al, 2003; Ozsarlak, 2007; Looby & Flanders, 2011). According to Looby and Flanders (2011), these criteria have specificity 42.5% and sensitivity of 99.6% and are more accurate when applied to stable and alert patients. Stiell et al. (2003) and Ozsarlak (2007) showed that the use of these criteria for stable and alert patients lowered the number of cervical spine radiographic exams by up to 40%.
Patients with blunt force spinal trauma will require imaging if they do not satisfy the criteria.
In an emergency department, a patient can be diagnosed of a CSI using plain radiography, a CT scan or magnetic resonance imaging (MRI). Each method has advantages and disadvantages, thus the clinical situation must be taken into account in determining the best choice for the patient (Ozsarlak, 2007; Heinzelmann et al, 2008).
2.3.2.1 Plain film radiography
The widespread availability and reduced costs of plain radiography makes it the initial imaging method for patients who cannot be diagnose of a CSI by clinical evaluation alone (Heinzelmann et al, 2008; Kortbeek et al, 2008; Lindsey & Gugala, 2010). However, technical adequacy, the interpretive abilities of the physician and the amount and type of perspectives obtained affect its success (Grossheim et al, 2009; Lindsey & Gugala, 2010).
A radiograph of a single lateral view is the first image to be taken and, according to Daffner and Hackney (2007), and Chimutengwende-Gordon et al. (2010), this one image is typically insufficient. Lindsey and Gugala (2010) showed that the sensitivity of a single lateral image is from 74% to 86% in patients confirmed with CSIs. A large number of studies, notably the Daffner and Hackney (2007), White et al. (2007), and Heinzelmann et al. (2008), Kortbeek et al. (2008), Lindsey and Gugala (2010) proposed that the current accepted minimum requirement for blunt force trauma patients with CSIs is a full cervical series (FSC) consisting of a lateral view, an open-mouth odontoid view and an anteroposterior view. That said, Chimutengwende-Gordon et al. (2010) and Lindsey and Gugala (2010) claimed that the effectiveness of this imaging relies upon the quality of the radiographs produced, with radiographs of the cervical spine often being inadequate. Indeed, the inability to obtain good quality images accounts for around 94% of the mistakes resulting in a delayed or a missed diagnosis of CSIs (Grossheim et al, 2009). Researchers, such as Daffner and Hackney (2007), White et al. (2007), Grossheim et al. (2009) and Lindsey and Gugala (2010) pointed out the distinction that technically adequate FCSs are able to contribute considerably to the accuracy of plain radiography. Yet Grossheim et al. (2009) highlighted that there can be a high occurrence of interpretation mistakes made by radiologists, physicians and surgeons. However, according to Platzer et al. (2006), serious CSIs sometimes remain undiagnosed even when a FCS is of diagnostic quality and is correctly interpreted. As a result, Richards (2005), Hunter et al. (2014) and Raby et al. (2014) suggested that two oblique views should also be taken to improve indication of the spinal alignment and integrity of the pedicles and facets. Additionally, Chimutengwende-Gordon et al. (2010) suggested a swimmer’s view be included to offer a further detail of the cervicothoracic junction. The drawback to the inclusion of this extra imaging is the rise in time outlay a significantly greater radiation dose and costs. A comparisons between the three-view and the five-view series revealed that the five-view series does not raise injury detection levels but did facilitate a more precise diagnosis (Daffner & Hackney, 2007; White et al, 2007; Grossheim et al, 2009). Kortbeek et al. (2008), Grossheim et al. (2009) Lindsey and Gugala and (2010) Khan et al. (2013) thus concluded that the three-view FCS is sufficient for initial detection and CT scanning or MRI are the preferred methods for further imaging if needed.
Freedman (2005) and Chimutengwende-Gordon et al. (2010), stated that flexion-extension views are able to be obtained if possibly injury to the ligaments has occurred in a patient with focal neck pain but with low misalignment of the lateral view and no indication of fracture or instability. The patient must also be conscious and cooperative. Anglen et al. (2002) and Lindsey and Gugala (2010) argued, however, that serious injury cannot be identified in this way and the process is time-consuming.
The major drawback of plain film radiography is that in many patients it cannot dependably indicate injuries at the cervicothoracic and occipitocervical junctions as they are complex structure and need more sophisticated technique (Powe, 2006; Lindsey & Gugala, 2010).
2.3.2.2 CT scanning
CT is useful for patients who continue to suffer symptoms despite negative x-rays, whose plain film radiographs show pre-vertebral swelling potentially indicating a CSI, and who have unresolved radiographic abnormalities (Kortbeek et al, 2008; Lindsey & Gugala, 2010; Ackland & Cameron, 2012). Current CT scanners include adaptations offering fast and effective imaging in all planes and volume imaging (Lindsey & Gugala, 2010; Ackland & Cameron, 2012; Theologis et al, 2014). Furthermore, according to Theologis et al. (2014), there is lower potential for failing to diagnose displaced transverse fractures (e.g. a type II dens fracture) with new CT scanners as they are equally sensitive in all planes. Brown et al. (2005) reported that 99.3% of all cervical spine fractures were found using only CT with any missed fractures needing no or little treatment. New CT devices are able to adequately assess abnormal angulation, rotation, dislocations, subluxation and intervertebral distances, areas that single slice spiral or axial CT are deficient in (Van Goethem et al, 2005). Technological advances in CT scanners have led to quicker examination times as CT machines are now faster and have improved spatial, temporal and contrast resolution a result of their use of smaller isotropic voxels (Bensch et al, 2004; Van Goethem et al, 2005). McCulloch et al. (2005) highlighted that modern CT scanning is not only more sensitive it is less expensive. Cervical spine CT becomes even less expensive when used as an extension of the primary CT of organs such as the thorax, abdomen and head.
In contrary, the drawbacks of CT as compared with plain radiography are higher exposure to radiation and restricted availability. In addition, Grossheim et al. (2009), Lindsey and Gugala (2010), Sheikh et al. (2012) and Ackland and Cameron (2012) affirmed that CT is not helpful in identifying injuries to ligaments.
2.3.2.3 MRI
Powe (2006) stated that sufficient assessment of ligament injuries in blunt trauma patients is not achieved even by the combined use of plain radiography and CT. MRI provides a useful and non-invasive tool for detecting ligament, neural or disc injuries. Como et al. (2009), Lindsey and Gugala (2010), Ackland and Cameron (2012), and Theologis et al. (2014) explained that MRI is advised for patients with neurological deficit. In such a situation an MRI is a safe and effective way to assess the spinal cord as it can reveal spinal cord odema and compression and epidural haematoma. If a ligament injury is assumed, further MRI is needed (Lindsey & Gugala, 2010). MRI is also needed when there are clinical indications of gaps between spinous processes or focal tenderness or when plain radiography or CT reveals interspinous widening or kyphosis (Pow, 2006; Lindsey & Gugala, 2010). Nonetheless, Lindsey and Gugala (2010) and Theologis et al. (2014) stated that MRI is not optimal for primary imaging of the cervical spine as it is time-consuming, it hinders a patient’s monitoring devices, it is costly and, compared with CT scans and x-rays, provides low-quality images of bony structures.
2.3.2.4 Plain film radiography versus CT
Appropriate imaging for CSIs is still hotly debated despite a widespread acceptance of the necessity of imaging for cervical spine trauma patients (White et al, 2007; Theocharopoulos et al, 2009; Ackland & Cameron, 2012). Trauma imaging methods have changed over the past ten years as the conventional choice of plain film radiography as the primary method for cervical spine assessment is affected by the rise in the availability of high-grade CT machines (Theocharopoulos et al, 2009; Blackmore & Avey, 2011; Ackland & Cameron, 2012).
In the context of plain film radiography, studies by MacDonald et al. (1999) and Nguyen et al. (2005) showed that plain radiography is adequate for cervical spine trauma patients. MacDonald et al. (1990) revealed that three radiograph views identified 99% of medically significant injuries. Nguyen et al. (2005) reported that, provided the radiographs display the full cervical spine (from the occiput to the first thoracic vertebra), plain radiography has specificity of 95% and sensitivity of 93%.
However, many other studies have shown conflicting results. A prospective study of plain radiography and CT on 100 individuals with sustained CSI revealed that plain film radiography missed 17.2% of unstable injuries and 52.3% of total injuries (Diaz et al, 2003).
A study of 437 individuals with blunt force trauma who were subject to CT scanning revealed that a sufficient lateral cervical spine film identified injuries with 53.3% sensitivity. Moreover, 45% of unstable injuries were missed using plain lateral films of the cervical spine and no unstable injuries were missed using CT of the cervical spine (Brohi et al, 2005).
Another study revealed an even higher number of cervical spine fracture missed using plain film radiography only (Gale et al, 2005). CT was adopted as a benchmark in this research and Gale et al. (2005) revealed that three radiograph views had a sensitivity of 32% and a specificity of 99% in identifying cervical spine fractures, that is, thirteen of nineteen fractures were missed.
A report on the NEXUS revealed that 36% of the 224 patients who had a minimum of one secondary injury that plain radiography failed to identify and that was only seen on later CT scans (Barrett et al, 2006).
Furthermore, In a prospective study of 667 patients (60 with CSIs), the patients underwent both plain radiography and CT scanning, revealing specificity of 97.4% and sensitivity of 45% for plain radiography and specificity of 99.5% and sensitivity of 100% for CT scanning (Mathen et al, 2007). In the Mathen et al. (2007) study, CT identified all medically significant injuries.
These striking results have led some researchers to look further into the effectiveness of plain radiography in cervical spine detection and to query its usefulness. In a breakthrough study by McCulloch et al. (2005) on the value of plain radiography as compared with current CT scans, less than half of emergency department X-ray radiographs were determined to be technically adequate. The radiographs as a whole had specificity of 97% and sensitivity of 45%. The technically adequate radiographs had moderately higher levels of specificity (98%) and sensitivity (52%). However, CT had specificity and sensitivity of 98%. Crucially, the technically adequate plain radiographs failed to identify any injury in almost half of the patient group that CT scanning detected injuries in. CT used on its own failed to detect one odontoid fracture. However, this fracture was easy to identify on review of the scans and thus it is probable that human error was responsible here rather than a lack of imaging sensitivity.
These similar results seen with CT scanning have led to CT scanning becoming the first imaging method used in cases of potential CSI (McCulloch et al, 2005; Blackmore & Avey, 2011). According to Blackmore and Avey (2011), CT is easily accessible in trauma centres and many other healthcare institutions and CT scans can be performed quickly. Numerous researchers such as Diaz et al. (2003), Brohi et al. (2005), Gale et al. (2005), McCulloch et al. (2005) Mathen et al. (2007) now believe that plain radiography can be completely replaced by modern CT for patients with cervical trauma and that CT should be considered as gold standard practice for the detection of CSIs and used as the initial diagnostic test for such injuries.
Even so, Tins et al. (2007) and Theocharopoulos et al. (2009) warned that high doses of radiation are linked to CT scans, compared with plain radiography, and thus CT can potentially lead to solid cancers, leukaemia and other negative outcomes. As such, Theocharopoulos et al. (2009) concluded that CT should be used when the advantages of an accurate diagnosis outweigh the risk of radiation effects as compared with other imagining methods.
In response to this consideration, Kortbeek et al. (2008) emphasised that the 2008 Advanced Trauma Life Support Teaching Protocol stressed the continued utilisation of plain radiography as the first imaging test for CSI. CT is to be used only when plain film radiographs suggest an injury or prove inadequate or if the patient presents with pain when the radiographs appear normal. Kortebeek et al. (2008) reported that if a head CT is needed as part of a patient’s evaluation, the scan at or near the posterior of the C2 body should be continued to include the atlanto-axial articulation and to thus, for instance, increase discovery of upper cervical injuries. Additionally, according to White et al. (2007) and Kortebeek et al. (2008), if it is impossible for plain radiographs to be adequate then CT should be utilised.
Plain radiography must be taken into account by physicians in some situations to evaluate CSIs when patients may not be capable of undergoing further imaging methods. For example, when internal mechanical hardware is present in the body of a patient they will not be able to undergo MRI or if a patient is over the weight limit of CT scanners they will not have the option of having a CT scan (Powe, 2006).
Researchers such as White et al. (2007), Kortebeek et al. (2008) and Ackland and Cameron (2012) emphasised that the value of plain radiography should not be ignored. They also added that neutral, open-mouth odontoid, lateral and anterior-posterior views should continue even when a patient will also undergo MRI and CT.
2.4 The contribution of the study
White et al. (2007) and Sheikh et al. (2012) commented that the method for screening trauma patients for potential CSI is the most debated and controversial topic in radiology. Ackland and Cameron (2012) explained that ethical issues prevent the use of randomised controlled trials (RCT) to study this area. This is because it is unethical to expose patients to double doses of radiation by making them undergo both plain film radiography and CT. Thus, researchers in such cases rely on non-RCT to determine the effectiveness of plain radiography as compared with CT in CSI assessment. The large number of non-randomised studies in the current literature share one key limitation: the absence of randomisation in choosing participants which means that substantial differences could exist between study groups and this may have an impact upon a study’s results. Studies by Diaz et al. (2003) and McCulloch et al. (2005), for instance, were biased as they only concerned individuals who were assumed to have a high risk of CSI and this can lead to exaggerating the results.
Kitchenham et al, (2009) stated that a structured literature review would be valuable if the harm or advantages of intervention were uncertain. As several non-randomised control studies in this area of research have been carried out covering various sides of this area and have shown conflicting outcomes, a structured literature review will be more beneficial than other studies. In order to establish a scientific response to the issue of imaging methodology for CSI, This structure literature review seeks to answer the review question which is: which of the two modalities (CT and plain radiography) are more accurate in the detection of CSI.
Chapter Three
3. Methodology.
3.1 Overview:
Research methodology may be categorised into several distinct stages, which governs the advancement of the research itself (Dawson, 2002). In order to deduce a scientific analysis of these procedures in structured reviews, methodical and definitive systems must be observed. These systems may vary in accordance to the research questions they are devised to investigate (White & Schmidt, 2005). In turn, the defined research question determines the process of analysis, and advocates various forms of primary research that may then be undertaken (Gough et al, 2012.( The research question can also determine which secondary studies can be used to examine the results of the primary studies included (Dawson, 2002). However, before identifying appropriate forms of primary research which will be included, it may be advantageous to identify specifically the kind of literature review which will be undertaken. Reviews, which purposes are to obtain, gather and examine evidential data, are specifically defined as ‘Quantitative reviews’ (Gough et al, 2012).
Quantitative research methodology may be sufficient, as stated by Lipowski (2008), to answer research questions, such as those concerned with healthcare interventions; for example diagnosis. This review question refers to the accuracy of medical diagnostic testing. Existing test accuracy research is chiefly concerned with gauging the effectiveness of an evaluation test (‘index test(s)’) and determining the existence of an identified disease or medical condition (Virgili et al, 2009; Leeflang et al, 2013). In order to evaluate the accuracy of such tests. the reference standard is used as a barometer with which to compare Index test(s) results, (Mariska et al, 2008). Patients who show positive results in accordance to the index test, are to be considered as diseased and subsequently become categorised as a percentage group known as ‘sensitivity’. Those patients who show negative results in accordance to the index test are to be categorised as a percentage group referred to as ‘specificity’ (Mariska et al, 2008; Tatsioni et al, 2005).
3.2 Inclusion criteria
The validity and reliability of reviews are dependent on the design of primary studies would be included due to the fact that the design of some primary studies are more robust and rigorous that others (Centre for Reviews and Dissemination (CRD), 2009(. The most significant study design in healthcare intervention is RCT. This notably incudes diagnostic studies (CRD, 2009; Higgins & Green, 2011; Holly et al, 2012). In RCTs, in order to prevent the possible damaging effects on results caused by extreme inconsistencies between participating groups, individuals are selected at random (CRD, 2009). Although this method is still regarded as one of the most effective research processes in healthcare, in many cases RCTs prove to be both impractical and more significantly, unethical, as indicated by diagnosis accuracy research (White & Schmidt, 2005; Higgins & Green, 2011). In the specific case of this research question, due to the unethical nature of purposefully exposing subjects to both x-ray tests and CT scanning, no RCT studies exist. Medical professionals in order to support the research for such investigations use non-RCTs (Reeves, 2008).
Diagnostic accuracy research has some distinctive aspects, which define it when compared to other intervention studies (Whiting et al, 2011). Due to these distinctive aspects, two characteristic non-RCT designs may suitably describe the categories that diagnostic accuracy falls into. The first being a ‘diagnostic cohort study’, which may also be described as a ‘single-gate’ study. This type of diagnostic study utilises a single set of criteria. The second category is the ‘case–control accuracy study’, which may also be called a ‘two-gate’ study; as the name suggests, two groups of criterion are employed (Mariska et al, 2008; CRD, 2009). ‘Single-gate’ and ‘two-gate’ studies may be collectively described a form of cross-sectional study (Higgins et al, 2008; CRD, 2009). In the single-gate study, patients assumed to be suffering from an identified illness are successively chosen to undertake the predetermined or so called index test(s). To formulate the diagnosis, test results from the reference standard and index test(s) are contrasted. (Higgins et al, 2008; CRD, 2009). The accuracy of the diagnosis is ensured by applying the same category of criteria to each of the patient groups. This crucially, is maintained during ‘consecutive enrolment’ (Rutjes et al, 2005; Mariska et al, 2008). This consecutive enrolment uphold a reasonable level of sensitivity and specificity which makes this approach universally reliable despite its non-randomised designs which may affect its reliability and a tendency towards bias (Mariska et al, 2008; CRD, 2009).
In the ‘two-gate’ or case-control accuracy study, patients who suffer from a target illness are selected from alternate populations to those contributing patients who show no signs of the target illness. This is also known as healthy control (Mariska et al, 2008; CRD, 2009). Healthy participants are subjected to criteria differing from the criteria that are applied to the sufferers of target diseases. This method utilises a pair of differing gates, which exposes the research to fundamental bias (Rutjes et al, 2005; CRD, 2009) due to the possibility of inconsistencies between individual patient cases and the consequential risk of overcompensating diagnostic testing (Rutjes et al, 2005; CRD, 2009). Having considered how this method of research can result in inadequate or otherwise flawed data, only studies single-gate were included in this review to specifically reduce the risk of biased results. The resulting outcomes can be therefore of sufficient reliability to be used by practicing medical personnel.
This study focuses on adult patients (16 and over) who were all suspected of suffering from CSI and need a radiological examination. All of the patients had suffered blunt force trauma, prior to this diagnostic imaging, and all of the available example studies were included in order to develop these research results. CSI injuries are relatively rare in children and more likely to occur among adults, making it significant that adult patients were the focus of this study (krekes & letton, 2010; Mortazavi et al, 2011(. Furthermore, the anatomical configurations of children’s spines are incomparable to those of adult patients, suggesting it to be fundamentally flawed to comprise a study that examines both children and adult patients (Looby & Flanders; Mortazavi et al, 2011). Study surveys concerned with this subject matter; tend to distinguish between adults and children when examining pre-existing cases of CSI, which confirms this inconsistency. Moreover, it is supportive of the theory that adult patients should be evaluated independently and categorised separately (Mortazavi et al, 2011).
3.2.3 Types of intervention (Index tests).
The intervention in other health care studies is frequently viewed as the index test(s) in diagnostic test accuracy (CRD, 2009). Accuracy review research can involve narrow questions including only one index test or can be broader as such, having a comparison of a performance between two index tests or more. In comparative diagnostic tests accuracy, Higgins et al. (2008) has recommended that recent test within clinics should be recognised as an index test. The purpose of this is to distinguish the names between the different varieties of index tests. As such, the term comparator test can be utilised upon the current test that are being used in practice. In order to diminish any confusion during the review, emphasises are placed where reference standard will never be the comparator.
The current question of the review was restricted to the comparison of the CT performance (index test due to it being recent) with the plain radiography’s performance (comparator test because it is a traditional technology) within the attempt to detect CSI. As such, all the studies have benchmarked the utilisation of plain film radiography alongside with the CT scanner’s performance within the evaluation of blunt trauma patients.
In determining whether the target disease or condition occurs in the participating patient, an approved method is to use the reference standard (CRD, 2009; Leeflang et al, 2013). This research therefore, contains studies that utilised the reference standard to identify the presence or absence of CSI. The reference standard in such cases is radiologists interpretation of the CT and plain radiographs and/or medical records in the discharge departments.
When evaluating the accuracy of radiological testing, it is crucial to establish whether or not the index tests are sufficiently ethical and relevant by identifying their sensitivity and specificity. By establishing these characteristics, it is possible to gauge how effective and dependable the test results may be (CRD, 2009). Thus, the foremost area of interest in the outcome of this testing was the sensitivity and specificity of each study case; and was considered in traditional radiography testing as well as CT.
The sensitivity and specificity data is required to be populated using 2 x 2 table (see appendix. 3). The table illustrates the correlation between data from the reference standard and the data obtained from the index tests; and are presented at a specified diagnostic threshold. This is the point at which results are classified as positive or negative.
3.3 Exclusion criteria
As this review focuses primarily on adult patients who are suspected with CSI after blunt trauma, studies which obtain results from pediatric patients, were excluded.
Other studies which were excluded from the review were study cases that did not adhere to the universal standard of radiology testing, which states that; three plain of plain film radiographs, is the accepted minimum. The three projections are: (i) Anteroposterior view, (ii) lateral view and (iii) open mouth view.
Studies which failed to include a comparison between CT scanning and plain radiography in their research were also not considered.
Due to the nature of the research, which is explicitly concerned with the cervical spine in its entirety, research that only examined partial areas of the cervical spine e.g. the upper vertebrae or lower vertebrae, was discounted.
In the case of some studies, insufficient evidence for the adopting of a reference standard to confirm the presence or absence of the target injury was presented, thus invalidating those research.
3.4 Identification of eligible studies
The subject of the research governs the appropriate use of specific electronic databases. (CRD, 2009). Medline and Embase are the favoured electronic databases used in health care interventions. They are considered adequate chronological sources for investigating diagnostic accuracy tests. (Higgins et al, 2008; CRD, 2009; Beynon et al, 2013). Medline and Embase are designed so that researchers may easily locate relevant information by searching with topic headings or central phrasing. (Higgins et al, 2008). Synonyms, equivalent medical language and reference codes, may also be accessed using reference databases. These are helpful when decoding acronyms and searching for the definitions of obscure medical terms (Higgins et al, 2008; Beynon et al, 2013). Therefore, in the search for relevant articles for this review, these two databases have been used initially. However, a range of various databases including Science Direct, Cochrane collaboration, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Allied and Complementary Medicine Database (AMED) and Pubmed, were also used to displace any search bias
It is imperative that the main concepts of the research question are explicitly stated, so that they may be referenced as subject headings or key words in the search process (Tasioni et al, 2005(. The relevant subheadings that have been identified to be applied to diagnostic testing research are as follows: (i) index test(s) (ii) target condition (Tasioni et al, 2005; Virgili et al, 2009). The index tests that are explored in this research question are: plain radiography and Computed Tomography, whereas, Cervical Spine Injuries (CSI) is the target condition of this review.
In the initial phase of research, the identified subheadings and keywords were entered into the previously mentioned database search engines. This was an integral part of the research process and identified significant research data.
The following are the search terms that define the index test:
– Computed Tomography, CT, helical CT, Multi-detector CT or (MDCT).
The following are the search terms that define the comparator test:
– plain radiography, plain film radiography.
The following are the search terms that define the target condition:
– Cervical spine injury, CSI, cervical spine fracture or (CSF).
Boolean operators were used in the search. The OR Boolean function was particularly useful to combine the terms that has similar meaning within the study such as “Computed Tomography” with “CT”. An operator such as AND was used to restrict the search through merging the words that specify the topic such as “plain radiography” merging with Computed Tomography “OR” CT.
As advocated by Mariska et al. (2008), the Boolean operator NOT did not feature in this research. This was specifically due to the inadequate accuracy record of Boolean operator NOT which can result in missing some significant studies.
Tasioni et al. (2005) and Beynon et al. (2013) stated that due to the ambiguous subject titling and inconsistent research filing, the search for diagnostic accuracy studies may prove to be unreliable. To compensate for this weakness, delimiter and search filters should be avoided during the database search process in order that all significant research data may be harvested. Therefore, delimiters have specifically been eschewed were possible, during this research process. The studies examined in this review were provided from the years spanning from 2004 to 2014. It may be noted that in the mentioned period, significant improvements were made to both plain film radiology and CT scanning equipment and technique. The effectiveness and precision of the scanning machines during the period (between 2004 and 2014) can truly reflect the current practice, when compared to the old machines. These issues of improvements have been proven in various research projects such as Fong (2007), Kalender et al. (2007) and Lindsey & Gugala (2010) which make the usage of this delimiter reasonable.
Another delimiter must also be noted, that these research documents were only originally published in English and Arabic as they are the languages spoken by the author.
3.5 Data collection and analysis
Relevant data was located using the database systems. The quantities of each result were subsequently recorded. The virtual management system Refworks was used to categorise and document the research articles. Is must be noted that there were some crossovers and duplications of search results, owing to the fact that identical search words were entered into equivalent database search engines. These were duly recorded in the Refworks filing database. The second phase was to screen the titles and abstracts of all retrieved articles in order to check for eligible articles. This initial screening refined the search results in accordance to the inclusion and exclusion criteria to reach to desirable content results.
The reference lists from all relevant articles were scrutinised to identify any other possible relevant studies. Finally, the full text version of the potentionally relevant articles were obtained for reading in their entirety to identify those finally for inclusion.
Cochrane Collaboration provided a useful data extraction form (See appendix. 4) which can be used for all health intervention studies including diagnostic test accuracy (Higgins et al, 2008). This form has been used to extract the most significant information from all included studies. This data extraction form where devised in accordance of each of the articles, to further improve the selection of relevant studies. This phase also aided in identifying exclusion criteria. Digital computer tools were used to develop the data extraction forms. An advantage of this is the ability to easily reproduce the form models for future tests.
Due to the heterogeneity of the studies’ results and differences in the characteristics of included articles, no meta-analysis was used in this review. Moreover, no blinding was applied during the data extraction phase as there was only one reviewer.
Heterogeneity is one of the defining features of diagnostic accuracy, which was prevalent in the results of the primary studies (Whiting et al, 2003; Whiting et al., 2011). The Quality Assessment of Diagnostic Accuracy Studies (QUADAS) was created to evaluate the attributes of primary studies in 2003, in response to the heterogeneity factor (Whiting et al, 2003; Whiting et al, 2011; Reltsma et al, 2012). Since its creation, QUADAS may be considered one of the most successful and widely used quality testing tools in diagnostic test studies. Over five hundred studies contain references to QUADAS and it also features in over two hundred diagnostic reviews contained in the Database of Abstract and Reviews of Effects (DARE). (Whiting et al, 2011). The Cochrane Collaboration, Agency for Healthcare Research and Quality, UK National Institute for Health and Clinical Excellence have all commended the use of QUADAS (Whiting et al, 2011). However, there are records of disparaging reviews of this system (Whiting et al, 2011; Reltsma et al, 2012). In response to this criticism, nine respected authorities took it upon themselves to purposefully advance QUADAS-2 to be used in their work developing diagnostic testing techniques, and in doing so dispelled any negative criticism of the QUADAS tool. The more recent QUADAS-2 is widely considered to be the most significant tool of its kind. This review has made use of this excellent reputation, track record and dependability for quality results.
QUADAS.2 may be divided into 4 categories. These are illustrated in appendix 5. They are as follows: the first being ‘Patient selection’, the second being ‘Index test(s)’, the third being ‘reference standard’ and the fourth flow and timing. The four domains are being investigated based on their potential bias. However, the first three domains are investigated based on applicability. ‘Signaling questions” are used in order to determine the biasness. The three answers are “yes”, “no” or “unclear”. If “yes” is selected then there is a “low” possibility of biasness, while, if “no’ is chosen then a “high” potential of bias exists. “Unclear” can be only used when insufficient information is recorded (Whiting et al, 2011; Reltsma et al, 2012). The applicability part does not involve singling questions. The applicability judgment can be directly defined as “low”, “high”, or “unclear” (Whiting et al, 2011).
CHAPTER Four
4 Result:
4.1 Selection of Articles
Two hundred and thirty five applicable records were identified during the preliminary databases research. Once the duplicated results had been removed a total of one hundred and fifty seven studies where identified. Once the screening process of the titles and abstracts of one hundred and fifty studies had been undertaken, and documents containing data identified in the exclusion criteria had been subtracted from these search results, a remainder of eleven relevant documents were identified. Further screening of the reference list of potentially relevant articles did not alter the final number of significant studies. After reading and analysing the full text of the eleven articles in accordance with the inclusion and exclusion criteria, four articles remained. As shown in appendix 6, a Reporting Items for Systematic Reviews (PRISMA) flow diagram, illustrated the research procedure.
A final four studies were included in this review as they all met the inclusion and exclusion criteria previously stated in methodology section. The research articles used in this review are as follows:
Widder, S., Doig, C., Burrowes, P., Larsen, G., Hurlbert, R.J. & Kortbeek, J.B. 2004, “Prospective evaluation of computed tomographic scanning for the spinal clearance of obtunded trauma patients: preliminary results”, Journal of Trauma and Acute Care Surgery, Vol. 56, no. 6, pp. 1179-1184.
McCulloch, P.T., France, J., Jones, D.L., Krantz, W., Nguyen, T., Chambers, C., Dorchak, J. & Mucha, P. 2005, “Helical Computed Tomography Alone Compared with Plain Radiographs with Adjunct Computed Tomography to Evaluate the Cervical Spine After High-Energy Trauma”, Journal of Bone and Joint Surgery, Vol. 87, no. 11, pp. 2388-2394.
Mathen, R., Inaba, K., Munera, F., Teixeira, P.G., Rivas, L., McKenney, M., et al. 2007, “Prospective evaluation of multislice computed tomography versus plain radiographic cervical spine clearance in trauma patients”, The Journal of Trauma, Vol. 62, no. 6, pp. 1427-1431.
Bailitz, J., Starr, F., Beecroft, M., Bankoff, J., Roberts, R., Bokhari, F., et al. 2009, “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”, The Journal of Trauma, Vol. 66, no. 6, pp. 1605-1609.
As illustrated in table 2, seven articles were excluded from the original eleven articles, owing to the fact that they failed to meet the entire requisite inclusion and exclusion criterion.
| Excluded Articles | Reason for exclusion |
|
Open mouth odontoid view and Anteriorposterior view were not obtained |
|
Open mouth odontoid view was not obtained |
|
Included children population |
|
Included children population |
|
The of Lack of reference standard |
|
The of Lack of reference standard |
|
Included children population |
Table. 2. Excluded studies.
4.2 Quality assessment of included studies
As stated previously, QUADAS-2 has been used to assess the quality of included studies in both risk of bias and applicability concerns. Due to the illogical format of results given by the QUADAS-2 tool, an alternative method of data presentation is recommended by some research bodies (Whiting et al, 2011; Reltsma et al, 2012 ). As shown in the appendix 7; a system of symbols and table results clearly present the data, without ambiguousness.
Reltsma et al (2012) stated that table results presented in this way do not adequately encapsulate the data results. A more efficient method would be to comprise a dual graph diagram representing the results of quality assessment. Refer to appendixes 8 and 9 which illustrate the QUADQS-2 data.
The assessment of bias may be divided into four categories, as previously stated (i) patient section (ii), index test(s), (iii) reference standard, (iv) flow timing.
In spite of the consecutive selection process of participants and the assurance of including both negative and positive patients, patients selection bias was identified in the QUADAS-2 results in three studies (Widder, 2004; McCulloch et al, 2005; Mathen, 2007). Widder (2004) and McCulloch (2005) studies were notably bias, due to their focusing on specifically high-energy trauma patients who also displayed high severity scores. They examined patients who were prone to experiencing a high level of CSI due to the nature and severity of their injuries. Shown by Mathen (2007), the enrolment of patients from Level 1 trauma wards, caused heavily partisan results, in spite of every trauma category being included in the study. Noll (2006) argued that the care offered in American Medical Institutions directly affected the ranking of the hospital on a scale approved by medical profession, such as the American College of Surgeons. It is possible that bias occurred in this study, owing to the fact that patients suffering from an increased susceptibility of CSI were treated specifically in Level 1 trauma centres. The study performed by Bailitz et al. (2009) may be considered to be more fairly representative. This reduction in bias causing factors may be attributed to the patient admittance system; including a variety of different hospitals i.e. public teaching hospital, urban hospital and Level 1 Trauma centres.
The index tests bias part is similar to ‘blinding’ in other medical intervention surveys. The index test results may be manipulated if the standard reference is pre-known (Whitening et al, 2011). This suggests that physicians with knowledge of an index test must be blinded to other index tests as well as the results of the reference standard and vice versa. Widder (2004), McCulloch (2005) and Bailitz et al. (2009) undertook testing with diminished jeopardy of index test bias, by insuring that physicians interpreting the result of plain radiographs were blinded to the results obtained in CT testing. Furthermore, all physicians were also blinded to the results of the reference standard. Mathen (2007) used a partial blinding as physicians interpreted CT results were blinded to plain radiographs results, however, the same physicians interpreted plain radiographs after the CT results were pre-known. This approach of interpretation can cause an overestimation of plain radiographs sensitivity and increases the bias of plain radiographs outcomes.
An assumption of one hundred per cent sensitivity of reference standard is made, so that test accuracy estimates can be fairly made. The supposition is also made that inadequate classification of test results is the cause of inconsistencies between the reference standard and the index test results (Worster & Carpenter, 2008; Whitening et al, 2011). Worster and Carpenter (2008) declared that impartiality is forfeited when a diagnostic test is used as a means of making a medical decision or forming a clinical opinion. Index tests become assimilated with the reference standard, in examples such as these, intrinsically causing incorporation bias. The overestimation of diagnostic accuracy is the intrinsic problem with incorporation bias (Worster & Carpenter, 2008; Wu et al, 2014). This means that the correct results of a test, which is undertaken with integral incorporation bias, could be expected to outdo those results gained from a test which has no infiltration of with incorporation bias (Worster & Carpenter, 2008; Wu et al, 2014). As the reference standard in all included studies were provided by medical professionals and referencing to clinical records, in the cases of CT scanning and plain radiography. This means the included studies contain incorporation bias and inadequate reference standard.
Bias was considered to be a threat when long periods of time relapsed between the reference standard and the index testing process; when patients received differing reference standards or when patients were omitted from the research data (Wade et al, 2013). Every one of the four examined studies utilised the same reference standard, and all of the patients’ resulting data as was submitted for analysis. Adequate time has relapsed between the index testing and reference standard, due to the fact that the data was collected prospectively. These factors of ‘flow’ and ‘Timing’ significantly reduce the risk of bias.
Applicability concerns are used within the QUADAS-2 for the assessment of patients’ selection, index test(s) and reference standard.
During the course of selecting the patients, there may be issues regarding the applicability. The issue may be that the patients that have been chosen within the study are not similar to those as set up by the review question regarding the seriousness of the target condition, demographic features, existence of differential diagnosis and setting of the research (Whitening et al, 2011). Within such instances, concerns would be focused on the patients selection’s applicability within the studies by (Widder, 2004; McCulloch, 2005; Mathen, 2007). This is because the three studies have enrolled patients with serious injury that were assumed to have a high risk of CSI. However, the objective of the review would include and categorised the various risk levels to all suspected patients with CSI. The study conducted by Bailitz et al. (2009) has included suspected CSI patients that are categorised into low, moderate and high risk. Therefore, this study would have a more prominent applicability and generalisability are is practical compared with other studies.
The variations within the technology that conducts the test, interpretation and execution would play their respective roles of affecting the estimates of the accuracy of the diagnostic test within the index test applicability. Index test may be different from those determined within the review question (Whitening et al, 2011; Wade et al, 2013). The index tests used by all the four studies were identical (CT and plain film radiography). These were stated within the review question and as such, there were minimal concern regarding the index tests’ applicability.
There are concerns regarding the reference standard applicability where the target condition that was determined by the reference standard was not equivalent to the question (Whitening et al, 2011; Leeflang et al, 2013). The target condition (CSI) within the four studies that were determined by the reference standard (radiologists interpretation and/or medical records) is similar to the review question and this minimises the applicability concerns regarding reference standard.
4.3 Study characteristics.
Appendix.11 shows a summary of the most significant study characteristics of each included studies in this structured literature review.
A total of 2681 adult patients were included in this review. The lowest sample size within the study was 102 participants, where 77% and 23% were males and females respectively. The largest sample size consisted of 1505 participants, where 72% and 28% were males and females respectively (Bailitz et al, 2009). There were 407 participants in the studies conducted by McCulloch et al. (2005), where 67% and 33% were males and females respectively. Mathen’s (2007) study consists of 667 participants with 70% and 30% were males and females respectively.
The average age of the participants were 35 years, 40.3 years, 35.4 years and 37 years for Wideder (2004), McCulloch (2005), Mathen (2007) and Poitz et al, (2009) respectively. The duration of studies were between five to 36 months. Widder (2004) and Bailitz et al. (2009) have participants that are over 16 years old whereas McCulloch (2005) only have participants that are over 18 years old. However, Mathen (2007) stated that adult patients were recruited in his study without specifying the age range of the participants.
Widder (2004) did not present sufficient information regarding the mechanism of injuries. Motor vehicle accident (MVA) was found to be the most prominent mechanism of injuries within the other three studies. The study conducted by McCulloch et al. (2005) demonstrated that 71% of patients are injured due to MVAs, 20% due to fall and 9% due to other conditions. Mathen’s (2007) found 48% of patients injured due to MVAs, 14.4% due to fall and 13.5% and 1% due to other conditions. The Bailitz et al. (2009) study has found 40% of patients injuries are due to MVAs, 25% due to battery, 20% due to fall and pedestrian versus auto. Table.3 depicted the characteristics of all participants in all four studies.
| Study | Age range | Average age | Male/female percentage | Duration of study | Mechanism of injuries |
| Widder (2004) | 16 and older | 35 years | 77% Males
23% females |
36 months | Un-provided |
| McCulloch et al. (2005) | 18 and older | 40.3 years | 67% Males
33% Femals |
18 months | MVAs= 71%
Fall= 20% Other= 9 |
| Mathen (2007) | Un- provided | 35.4 years | 72% Males
28% Females |
5 months | MVAs= 48%
Pedestrian versus auto= 14.4% Falls= 13.5% Others= 1%
|
| Bailitz et al. (2009) | 16 and older | 37 years | 70% Males
30% Females |
23 months | MVAs= 40%
Falls= 20% Pedestrian versus auto= 9% |
Table.3. Participants characteristics.
4.3.2 Interventions (Index tests(.
CT was used to assess all the participants within the four studies. However, there are various features and protocols within using the CT index test. Therefore, each study may have a slightly different result. The CT scanning types and protocols which were used each of the all four included studies are illustrated in appendix 10.
Plain film radiography is the alternative index test (intervention) which is also known as a comparator test. This test is differs from a CT test where the type of plain radiography machines that were used as a comparator test would be deemed insignificant in these cases. Therefore, attention should be based upon the attainment of the amount of views and the type. All the four studies have documented 3 views mainly; anteroposterior, odontoid, lateral a minimum standard. Studies conducted by Widder (2004) McCulloch et al. (2005) and Mathen (2007) use the swimmer’s view as and when directed by radiologists.
The radiologists have interpreted the results of CT and plain radiographs within the four studies.
Reference standard was applied within all the studies in order to determine the existence of CSI. Studies conducted by Widder (2004) and Bailitz et al. (2009) have used radiologists interpretation and final diagnosis of medical records during discharge as the reference standard. The reference standard used by McCulloch et al. (2005) study is the radiologists’ interpretation of the entire radiographic studies as well as all clinical data. The reference standard used in the study conducted by Mathen (2007) is the radiologist interpretation of CT images and all medical data.
4.3.4.1 Sensitivity
According to all the studies, CT sensitivity was significantly higher than plain radiography (see figure.2); sensitivity estimation of CT was documented as 100% within three studies (Widder, 2004; Mathen, 2007; Bailitz et al, 2009) while one of the studies (McCulloch et al, 2005) recorded it as 98%. The aggregate sensitivity of plain radiography was recorded within a range from 36% to 45%. A study conducted by Bailitz et al. (2009) has categorised the patients into the high, moderate and low risk of injury. The high, moderate and low categories are recorded as 46%, 37% and 25% respectively for plain radiography sensitivity. False negative CT was recorded on one study by McCulloch et al. (2005). By contrast, all the studies have recorded false negative results from plain radiography.
Figure.2. Summary of studies sensitivity.
4.3.4.2Specificity
Patients that were diagnosed were particularly chosen by Bailitz et al. (2009). This means that the calculation for the specificity of plain radiography and CT could not be achieved. Three other studies that have demonstrated the specificity of plain radiography and CT were closely correlated. The range of the specificity of CT was from 98% to 100%. Meanwhile, the range of specificity of plain radiography was from 97% to 98%. The two studies have acknowledged false negative results that are related to CT (McCulloch et al, 2005; Mathen, 2007). At the same time, there are three studies (Widder, 2004; McCulloch et al, 2005; Mathen, 2007) that have discovered false negative findings that are related to plain radiography. Summary of the specificity of plain radiography and CT is shown in figure 3.
Figure.2. Summary of studies specificity.
Chapter Five
5. Discussion
5.1 The main results
The superiority of CT over plain radiography in detecting CSI in blunt trauma patients has been confirmed by every one of the four included studies. An increasing body of evidence points towards the unsuitability of plain radiography in the discovery of CSI in comparison to CT. Does this mean that CT should be used instead of plain radiography to detect potential CSI in blunt trauma patients? To be able to draw definitive conclusions about the clinical significance of these findings, the data derived from the included studies must be analysed more closely.
5.2 Findings based on current evidence and clinical implications
According to Hunter et al. (2014), the likely risk of injury in blunt trauma patients should be taken into account when assessing CSI in these patients, in order to be able to decide which imaging technique is more appropriate. As proposed by Blackmore and Avey (2011) (see appendix 12), a likelihood of cervical spine fracture greater than 10% should denote high risk patients, while a likelihood of between 4-10% and lower than 4% should be considered to denote moderate and low risk patients, respectively. Taking note of this classification, the included studies can be more closely analysed by discussing their results separately, depending on whether they focused solely on patients at high CSI risk or whether they included all potential CSI patients, at any risk level.
5.2.1 Studies on blunt trauma patients at high CSI risk
Patients at high CSI risk constituted the focus of three of the included studies, namely those conducted by Widder (2004), McCulloch et al. (2005) and Mathen (2007). All of them observed that plain radiography often does not detect CSI. This is all the more serious as many such injuries were significant from a clinical viewpoint and had to be treated by surgical stabilisation. Of the three studies, the largest sample of patients (N=667) was used by Mathen (2007), who conducted a comparison between (MDCT) and the three typical views of plain x-ray films. The results revealed that MDCT detected all the injuries (60 of 60 cases), whilst the plain radiographs detected only 27 (45%) and misdiagnosed in 33 (55%) out of the 60 injury cases. Among the injuries that plain radiography failed to detect, 15 were clinically important, 5 necessitating surgery, while the rest of them necessitated either long-term collar or halo stabilisation. In defence of plain radiography, Grossheim et al. (2009) argued that the high rate of undetected or overlooked injuries is likely due to inappropriate x-ray films or incorrect interpretation of the images. Inappropriate x-ray films may be caused by overlapping items, like a nasogastric line, or by imprecise positioning of the patients (Richard, 2005). Nonetheless, even in the absence of technical problems, plain radiography can fail to detect CSI in blunt trauma patients. McCulloch et al. (2005) reported that in 48% of cases the radiographs were appropriate. Moreover, to determine how sensitive these appropriate radiographs were, the researchers conducted a sub-analysis and found that, in spite of radiograph adequacy, in 12 of the cases the injuries were still not detected (52% sensitivity).
The results obtained by all three studies (Widder, 2004; McCulloch et al, 2005; Mathen, 2007) suggest that the use of plain radiography in high risk trauma patients does not always have an optimal outcome. Additionally, even if their technical adequacy is ensured, plain radiographs can still generate incorrect assessments of CSI in this category of patients. The results of these studies corroborate those attained by Holmes and Akkinepalli (2005) who carried out a meta-analysis of a number of 3834 patients over seven studies. They reported overall sensitivity estimates of 52% and 98% for plain radiography and CT, respectively. Other studies have also highlighted that CT is more accurate in detecting CSI compared to plain radiography (Nguyen and Clark, 2005; SharmaMc, 2007).
Two key conclusions can be derived from these results. The first one is that, in order to obtain more accurate imaging, many high risk patients are subjected to CT scanning after plain radiography for better visualisation of some areas, which increases not only the radiographic imaging costs, but also the exposure to ioniing radiation (McCulloch et al., 2005; Blackmore and Avey, 2011). The treatment of patients with serious injuries may be delayed or disrupted if appropriate radiographs cannot be produced or if they have to be repeated (Blackmore and Avey, 2011). The second conclusion is that the presence of injuries cannot be disproven based on the absence of abnormalities on appropriate radiographs. As warned by Widder (2004) and McCulloch et al. (2005), diagnosis may be deferred by exclusive reliance on plain radiography for the detection of CSI, resulting in disability or even death. In this context, examining high CSI risk patients with CT scanning first may be a more efficient strategy. In comparison to plain radiography, the examination time is considerably reduced as the CT scanning of the cervical spine can be undertaken at the same time as the scanning of other parts of the body, such as the head, pelvis or abdomen (McCulloch et al, 2005; Daffner, 2006). Crucially, injured patients could be treated faster thanks to the high sensitivity of CT. Munera et al. (2012) emphasised that, in addition to speeding up surgical stabilisation, early detection of CSI can also afford the option of steroid therapy, which is efficient only if it is applied immediately after the injury was sustained.
5.2.2 Studies on all suspected CSI blunt trauma patients
The study conducted by Bailitz et al. (2009) was the only one of the four included studies that investigated blunt trauma patients at all levels of risk. Of the 1505 patients that constituted the study sample, 50 (3.3%) had injuries of clinical importance. All were detected by CT (100% sensitivity), while only 18 were detected by plain radiography (36% sensitivity). Among these 50 patients, 15 were at high CSI risk, 19 at moderate risk and 16 at low risk. Of these, only 7 high risk (46% sensitivity), 7 moderate risk (37% sensitivity) and 4 low risk patients (25% sensitivity) had their injuries correctly detected by plain radiography. It can be implied from these results that all blunt trauma patients considered to necessitate diagnostic radiography should be assessed via CT scanning instead of plain radiography. On the other hand, some researchers have advocated that patients at moderate and high CSI risk should be differentiated from those at low risk (Theocharopoulos, 2009; Hunter et al, 2014). However, all the studies discussed in this review and others in the related literature concur that a CT scan of the cervical spine is the optimal choice in the case of patients at a moderate and high risk of CSI.
As far as patients at low risk of injury are concerned, it is still uncertain whether the use of good quality radiography is appropriate. Resolving this matter calls for a comprehensive investigation of all the available studies in the related literature. Apart from the above mentioned study conducted by Bailitz et al. (2009) which included a few number of patients suspected with low risk of CSI which makes it un-representative to clinical practice due to this small number. Searching in the literature yield two more studies (Mower et al, 2001; Griffen et al, 2003) which examined the effectiveness of plain radiography in the discovery of CSI in adult blunt trauma patients at low risk and have a bigger sample size. The research sample employed by Mower et al. (2001) consisted of 34,069 blunt trauma patients from 21 different trauma care centres. Plain radiography detected CSI in 818 patients (2.4%), but missed it in 320 patients. Most of the misdiagnosed injuries were associated with cases where the interpretation of the radiographs was inadequate. Consequently, adjunctive imaging studies were used to make an injury diagnosis. CSI was overlooked by appropriate radiography in only 23 cases (0.07%), while just three patients with possibly unstable fractures were misdiagnosed. Based on their findings, the authors concluded that the likelihood that a CSI will be misdiagnosed on acceptable radiographs is minimal and furthermore that the use of CT is unnecessary in patients at low risk, for whom plain radiography is suitable. Accordingly, Griffen et al. (2003) did not include in their analysis the patients whose radiographs were deemed inappropriate. Nevertheless, plain radiography was still unsuccessful in detecting CSI in 41 out of 116 patients (35%). In these cases, the attending radiologist had not identified any abnormality on the radiographs and therefore additional assessment with CT was not recommended.
The examination of these three studies suggests that the existing literature on this topic is still not sufficiently comprehensive. All studies have certain limitations in terms of methodology that could affect the validity of their evaluations of the performance of the diagnostic modalities. The study by Bailitz et al. (2009) was the only one that was prospective. This design has an advantage over the retrospective deign because the investigators have the ability to control patients recruitment in accordance to clinical characteristics which could increase the quality of the outcomes (CRD, 2009). Furthermore, all studies applied ineffectively or did not specify clearly the use of a certified reference standard. For instance, Mower et al. (2001) did not apply any gold standard systematically, but rather focused on the risk management logs which may have overlooked significant injuries, hence indicating that plain radiography is more sensitive than it actually is. Similarly, Griffen et al. (2001) and Bailitz et al. (2009) failed to use a consistent gold standard beyond CT in their comparisons of radiographs to routine CT. As a result, the accuracy of CT might have been exaggerated, making it seem more efficient than plain radiography. In addition, Bailitz et al. (2009) were the only ones to blind the radiologists. What is more, there is uncertainty in three of the trials as to the manner in which the differences in the final radiology reports were managed, if at all, or if radiologists established an agreement on the final reads. Last but not least, precise definition of “abnormal” radiographs were offered only by one study, that conducted by Mower et al. (2001).
The limitations of these studies, however, do not diminish the fact that plain radiography has a low efficiency in the detection of CSI. In the three trials focusing on the comparison of plain radiography with CT, sensitivity estimates exhibited a great deal of variation (36-65%), adequate radiographs being the only ones associated with higher values (Mower et al, 2001; Griffen et al, 2001; Bailitz et al, 2009). Even when the focus was only on clinically significant injuries and despite the small number of cases and the variation allowed in the definition of clinically significant, the efficiency of plain radiography remained low. Although many of the abnormalities missed by plain radiography were detected by CT, it is uncertain what the clinical implications of these missed abnormalities are.
There has been a staggering increase in reliance on CT scanning, including of the cervical spine, in the past ten years. Beforehand, negative plain radiographs were taken as an indicator that cervical collars could be removed (Boone and Brunberg, 2008; Lee et al, 2009). In light of the poor sensitivities as documented by the previously mentioned studies, the oversight of a considerably high number of “clinically significant injuries” would have been a likely possibility, causing the blunt trauma victims to become paralysed. Needless to say, this did not happen. The fact that some CT results might point to the presence of actually non-existent abnormalities could explain this situation. Furthermore, as noted by Hunter et al. (2014), the prognosis of injuries deemed according to CT criteria as “unstable” or “clinically significant” might be better than expected when no treatment is administered. Regrettably, the accuracy of plain radiography in patients at low risk has not been directly evaluated in any trials. Though scant, the existing evidence indicates that plain radiography does not have a high enough sensitivity to eliminate the possibility of CSI in blunt trauma patients. To determine conclusively how efficient plain radiography and CT are in detecting CSI in cases of blunt trauma, consistent prospective trials targeting solely patients at low risk of CSI must be systematically conducted.
5.3 Additional considerations:
Since injury oversight may have severe implications, sensitivity plays a key role in deciding which imaging method is most appropriate for detecting CSI in blunt trauma patients. Nonetheless, the imaging techniques selection should also take into account a number of additional factors, including radiation exposure, costs involved, duration of examination and equipment availability.
As mentioned by Theocharpoulos (2009), subjecting the cervical spine to CT scanning entails a high degree of risk as the area containing the organ with the greatest sensitivity to radiation, the thyroid gland, is exposed to a high dose of ionising radiation. In a study on potential CSI patients, Rybicki et al. (2002) conducted a comparison between the radiation dose of CT and that of plain radiography. The mean skin dose for CT and plain radiography was 27.2 mGy and 2.75 mGy, respectively. Furthermore, the authors used an ionisation chamber to determine the amount of radiation that the thyroid gland was exposed to, revealing that plain radiography delivered 14 times less radiation than CT. However, the study had a drawback in that a single-detector helical CT scanner was employed to determine both skin and thyroid radiation exposure. Unlike this type of scanner, the new MDCT technology necessitates less radiation. Theocharopoulos (2009) used MDCT and reported that an appropriate radiation dose in the case of plain radiography and CT was equal to 0.050 mSv and 3.8 mSv, respectively. However, even though CT scans involves exposure to higher levels of radiation, the author argued that the benefit to risk ratio of CT alone is more favourable than that of plain radiography or radiography combined with CT. The author further added that this holds true even in the case of low risk but highly radiosensitive patients such as women in their early twenties. The conclusion drawn by Theocharopoulos (2009) was that the use of a higher dose of radiation in CT than in plain radiography or radiography combined with CT improves diagnostic accuracy and therefore outweighs the negatives, whilst also being suitable for administration to any patient, irrespective of sex, age or fracture risk. Nonetheless, both studies have a weakness in that they did not specify the precise risk of future neoplasm, which is essential and thus requires further investigation.
Blackmore and Avey (2011) carried out a cost-effectiveness analysis which concluded that CT was the optimal method in the case of patients at high risk of CSI, not only reducing costs, but also improving health by preventing paralysis. The cost-effectiveness of CT was explained by the two authors in terms of the fewer inappropriate examinations as well as the extremely high medical and financial expenditure that paralysis incurs. Moreover, given the high sensitivity of CT and the uncommon nature of fractures, the likelihood of patients becoming paralysed as a result of overlooked injuries is minimal. By contrast, the lifetime medical and financial costs of paralysis are significant – between $525,000 and $950,000, and are further increased by social costs like wage loss (Blackmore and Avey, 2011). Paralysis has serious implications not only financially, but also health-wise. Blackmore and Avey (2011) also noted that sensitivity analysis testing of estimates imprecision clearly showed that CT had a higher performance than radiography in patients at high risk. Moreover, the authors reported that, in the case of patients at moderate risk, the costs of CT exceeded those of radiography, but made up for the high costs in the quality-adjusted life ($25,000 per year). However, in patients at low risk, radiography was better than CT as the costs of the latter were unjustified (Blackmore and Avey, 2011).
In a study of 156 suspected CSI trauma patients, Daffner (2001) revealed that a CT scan could be completed, on average, in 12 minutes, whereas plain radiography required 22 minutes. The difference of 10 minutes was attributed by the author to the necessity of having to repeat inappropriate radiographs. Similarly, McCulloch et al. (2005) reported that a CT scan of the cervical spine was completed in three-quarters of the time required to produce plain radiographs. These studies suggest that CT is more time-effective than radiography in assessing potential CSI blunt trauma patients.
Most new emergency and trauma centres are adequately equipped with both CT and plain radiography. However, in rural or isolated regions, particularly in developing countries, suspected CSI patients continue to be assessed via plain radiography because CT equipment is not available (Ackland and Cameron, 2012).
5.4 Limitations
The present review has some limitations. One significant limitation is that the study may not be entirely objective as only one researcher reviewed the existing studies (Higgins and Green, 2011). Under ideal circumstances, the literature should be explored by a minimum of two reviewers. This is essential particularly during the process of selecting the studies as the decisions regarding which studies to include and which not can be quite subjective, irrespective of the fact that the inclusion criteria are pre-established (CRD, 2009). Moreover, only four articles fulfilled the inclusion/exclusion criteria, which is an additional limitation as possibly relevant studies may have been left out.
Yet another limitation of this study is the inability to obtain some relevant studies, as a result of lack of funding or resources, such as studies written in languages other than English or Arabic and with no available translation, as well as conference abstracts and unpublished articles. Hence, it is possible that pertinent studies were excluded. Furthermore, RCTs represent the gold standard framework for primary healthcare intervention studies and therefore the absence of RCTs because of patient safety considerations could undermine the validity of the studies included (CRD, 2009; Higgins and Green, 2011). An additional limitation is that the discrepancies in the findings of the studies included hindered the performance of statistical pooling and meta-analysis.
Chapter Six
6. Conclusions and recommendations
The findings gathered by this study support the fact that CT is more effective in detecting CSI than plain radiography. All the included studies reported that plain radiography often missed CSI, many of which were clinically significant and necessitated treatment like surgical stabilisation.
The heightened sensitivity, time-effectiveness and the high costs associated with an incorrect diagnosis seem to indicate that CT is the best imaging method for suspected CSI patients. Nonetheless, insufficient evidence exists to indicate that CT is optimal for trauma patients at all risk levels. This is due to the fact that the sample of patients chosen by Widder (2004), McCulloch et al. (2005) and Mathen (2007) were all at high risk and therefore the findings of these authors do not apply to minor trauma patients at low risk of CSI. On the other hand, blunt trauma patients at all risk levels were included in the study by Bailitz et al. (2009). However, the findings of this study suggest that that CT is much more effective than plain radiography in all risk level patients. The results obtained in cases of moderate to high risk can be accepted as they are supported by the results obtained not only by the studies included but also by many other studies in the literature. On the other hand, the results are not as pertinent in the case of patients at low risk due to the small size of the patient sample employed. The CT costs and health risk involved in exposing a large proportion of patients to high radiation doses are greater than the benefits of identifying additional injuries in low risk patients.
Thus, it can be concluded that it is recommended that the risk of CSI of the patients should be taken into account when deciding which imaging method to use. CT is optimal in the case of patients at moderate and high risk. As far as low risk patients are concerned, although insufficient, the evidence indicates that the sensitivity of radiography is not high enough to eliminate the possibility of CSI in adult blunt trauma patients. Hence, identifying the efficiency of plain radiography and CT in detecting CSI in blunt trauma patients requires systematic multicentre prospective cohort trails to investigate patients at low risk of injury by addressing a number of factors, including radiation dose, time, costs and availability of the two methods.
7.References
Ackland, H.M. & Cameron, P.A. 2012, “Cervical spine clearance in trauma patients”, The Indian Journal of Neurotrauma, Vol. 9, no. 2, pp. 7 9-8 4.
Agrawal, A. 2008, “Cervical Spine Clearance: A Review and Understanding of the Concepts”, Journal of Nepal Medical Association, Vol. 47, no. 4, pp. 244-250.
Anderson, P.A., Gugala, Z., Lindsey, R.W., Schoenfeld, A.J. & Harris, M.B. 2010, “Clearing the cervical spine in the blunt trauma patient”, The Journal of the American Academy of Orthopaedic Surgeons, Vol. 18, no. 3, pp. 149-159.
Anglen, J., Metzler, M., Bunn, P. & Griffiths, H. 2002, “Flexion and extension views are not cost-effective in a cervical spine clearance protocol for obtunded trauma patients”, The Journal of Trauma, Vol. 52, no. 1, pp. 54-59.
Bailitz, J., Starr, F., Beecroft, M., Bankoff, J., Roberts, R., Bokhari, F., et al. 2009, “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”, The Journal of Trauma, Vol. 66, no. 6, pp. 1605-1609.
Ball, C. & Watson, D. 2010, “A 12 month clinical audit of cervical spine imaging in multiply injured and intubated patients”, The British Journal of Radiology, Vol. 83, no. 987, pp. 257-260.
Barrett, T.W., Mower, W.R., Zucker, M.I. & Hoffman, J.R. 2006, “Injuries missed by limited computed tomographic imaging of patients with cervical spine injuries”, Annals of Emergency Medicine, Vol. 47, no. 2, pp. 129-133.
Bensch, F.V., Kiuru, M.J., Koivikko, M.P. & Koskinen, S.K. 2004, “Spine fractures in falling accidents: Analysis of multidetector CT findings”, European Radiology, Vol. 14, no. 4, pp. 618-624.
Benzel, E.C. & Connolly, P.J. 2012, The cervical spine, 5th edn, Lippincott Williams & Wilkins, Philadelphia.
Berlin, L. 2003, “CT versus radiography for initial evaluation of cervical spine trauma: What is the standard of care?”, American Journal of Roentgenology, Vol. 180, no. 4, pp. 911-915.
Beynon, R., Leeflang, M., McDonald, S., Eisinga, A., Mitchell, R.L., Whiting, P. & Glanville, J.M. 2013, “Search strategies to identify diagnostic accuracy studies in MEDLINE and EMBASE”, Cochrane Database Syst Rev, vol. 9.
Blackmore, C.C. & Avey, G.D. 2011, “21 imaging of the spine in victims of trauma.” in: Evidence-based imaging, Springer, pp. 357-368.
Bowling, A. 2009, Research methods in health: Investigating health and health services, 3rd edn, Open University Press, Philadelphia .
Broder, J., Fordham, L.A. & Warshauer, D.M. 2007, “Increasing utilization of computed tomography in the pediatric emergency department, 2000–2006”, Emergency Radiology, Vol. 14, no. 4, pp. 227-232.
Brohi, K., Healy, M., Fotheringham, T., Chan, O., Aylwin, C., Whitley, S., et al. 2005, “Helical computed tomographic scanning for the evaluation of the cervical spine in the unconscious, intubated trauma patient”, Journal of Trauma-Injury, Infection, and Critical Care, Vol. 58, no. 5, pp. 897-901.
Brown, C.V., Antevil, J.L., Sise, M.J. & Sack, D.I. 2005, “Spiral computed tomography for the diagnosis of cervical, thoracic, and lumbar spine fractures: Its time has come”, Journal of Trauma-Injury, Infection, and Critical Care, Vol. 58, no. 5, pp. 890-896.
Centre for Reviews and Dissemination 2009, Systematic reviews: CRD’s guidance for undertaking reviews in health care. York: CRD, University of York.
Chimutengwende-Gordon, M., Sidhu, J., Khan, W.S. & Maruthainar, N. 2010, “Advanced trauma life support radiographic trauma series: Part 1–the cervical spine radiograph”, Journal of Perioperative Practice, Vol. 20, no. 9, pp. 322-428.
Como, J.J., Diaz, J.J., Dunham, C.M., Chiu, W.C., Duane, T.M., Capella, J.M., et al. 2009, “Practice management guidelines for identification of cervical spine injuries following trauma: Update from the eastern association for the surgery of trauma practice management guidelines committee”, The Journal of Trauma, Vol. 67, no. 3, pp. 651-659.
Daffner, R.H. 2000, “Cervical radiography for trauma patients: A time-effective technique?”, American Journal of Roentgenology, Vol. 175, no. 5, pp. 1309-1311.
Daffner, R.H. 2004, “Controversies in cervical spine imaging in trauma patients”, Emergency Radiology, Vol. 11, no. 1, pp. 2-8.
Daffner, R., Sciulli, R., Rodriguez, A. & Protetch, J. 2006, “Imaging for evaluation of suspected cervical spine trauma: A 2-year analysis”, Injury, Vol. 37, no. 7, pp. 652-658.
Daffner, R.H. & Hackney, D.B. 2007, “ACR appropriateness criteria< sup>®</sup> on suspected spine trauma”, Journal of the American College of Radiology, Vol. 4, no. 11, pp. 762-775.
Dawson, C. 2002, Practical research methods: a user-friendly guide to mastering research techniques and projects, How To Books Ltd, Oxford.
Hilton, D. 2004, Systematic reviews in health care – a practical guide, Public Health Association of Australia, Richmond.
Diaz Jr, J.J., Gillman, C., Morris Jr, J.A., May, A.K., Carrillo, Y.M. & Guy, J. 2003, “Are five-view plain films of the cervical spine unreliable? A prospective evaluation in blunt trauma patients with altered mental status”, Journal of Trauma and Acute Care Surgery, Vol. 55, no. 4, pp. 658-664.
Fong, C. 2007, “The big picture: Advances in radiology”, Nursing Management (Springhouse), Vol. 38, no. 12, pp. 52-53.
Freedman, I., van Gelderen, D., Cooper, D.J., Fitzgerald, M., Malham, G., Rosenfeld, J.V., et al. 2005, “Cervical spine assessment in the unconscious trauma patient: A major trauma service’s experience with passive flexion-extension radiography”, Journal of Trauma-Injury, Infection, and Critical Care, Vol. 58, no. 6, pp. 1183-1188.
Gale, S., Gracias, V., Reilly, P. & Schwab, C.W. 2005, “The inefficiency of plain radiography to evaluate the cervical spine after blunt trauma”, Journal of Trauma-Injury, Infection, and Critical Care, Vol. 58, no. 1, pp. 215-219.
Gardner, A., Grannum, S. & Porter, K. 2005, “Cervical spine trauma”, Trauma, Vol. 7, no. 3, pp. 109-121.
Gough, D., Thomas, J. & Oliver, S. 2012, “Clarifying differences between review designs and methods”, Syst Rev, Vol. 1, no. 1, pp. 28.
Griffen, M.M., Frykberg, E.R., Kerwin, A.J., Schinco, M.A., Tepas, J.J., Rowe, K., et al. 2003, “Radiographic clearance of blunt cervical spine injury: Plain radiograph or computed tomography scan?”, The Journal of Trauma: Injury, Infection, and Critical Care, Vol. 55, no. 2, pp. 222-227.
Harrison, P. & Cairns, C. 2008, “Clearing the cervical spine in the unconscious patient”, Oxford Journals, Vol. 8, no. 4, pp. 117-120. Available at: http://ceaccp.oxfordjournals.org/content/8/4/117.full.pdf+html [Accessed 20th June 2014].
Heinzelmann, M., Eid, K., & Boos, N. 2008, “Cervical spine injuries”. In: Spinal Disorders, eds. N. Boos. & M, Aebi, Springer Berlin Heidelberg, New York, pp. 825-881.
Higgins, J.P., Green, S. & Collaboration, C. 2008, Cochrane handbook for systematic reviews of interventions, Wiley Online Library.
Higgins, J.P.T. & Green, S., 2011. Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0. [online]. The Cochrane Collaboration. Available at: http://www.cochrane-handbook.org/ [Accessed 10th June 2014].
Hoffman, J.R., Mower, W.R., Wolfson, A.B., Todd, K.H. & Zucker, M.I. 2000, “Validity of a set of clinical criteria to rule out injury to the cervical spine in patients with blunt trauma”, New England Journal of Medicine, Vol. 343, no. 2, pp. 94-99.
Holly, C. & Salmond, S.W. 2012, Comprehensive systematic review for advanced nursing practice, Springer Publishing Company, New York.
Holmes, J.F. & Akkinepalli, R. 2005, “Computed tomography versus plain radiography to screen for cervical spine injury: A meta-analysis”, Journal of Trauma-Injury, Infection, and Critical Care, Vol. 58, no. 5, pp. 902-905.
Hunter, B.R., Keim, S.M., Seupaul, R.A. & Hern, G. 2014, “Are plain radiographs sufficient to exclude cervical spine injuries in low-risk adults?”, The Journal of Emergency Medicine, Vol. 46, no. 2, pp. 257-263.
Kalender, W.A. & Kyriakou, Y. 2007, “Flat-detector computed tomography (FD-CT)”, European Radiology, Vol. 17, no. 11, pp. 2767-2779.
Khan, M., Chong, S.T., Mazza, M.B., Hoff, C. & Albeiruti, A. 2013, “Clearing the posttraumatic adult cervical spine”, Contemporary Diagnostic Radiology, Vol. 36, no. 8, pp. 1-7.
Kitchenham, B., Pearl Brereton, O., Budgen, D., Turner, M., Bailey, J. & Linkman, S. 2009, “Systematic literature reviews in software engineering–a systematic literature review”, Information and software technology, Vol. 51, no. 1, pp. 7-15.
Kortbeek, J.B., Al Turki, S.A., Ali, J., Antoine, J.A., Bouillon, B., Brasel, K., et al. 2008, “Advanced trauma life support, 8th edition, the evidence for change”, The Journal of Trauma, Vol. 64, no. 6, pp. 1638-1650.
Kreykes, N.S. & Letton, J.,Robert W. 2010, “Current issues in the diagnosis of pediatric cervical spine injury”, Seminars in Pediatric Surgery, Vol. 19, no. 4, pp. 257-264.
Lee, J., Kirschner, J., Pawa, S., Wiener, D.E., Newman, D.H. & Shah, K. 2010, “Computed tomography use in the adult emergency department of an academic urban hospital from 2001 to 2007”, Annals of Emergency Medicine, Vol. 56, no. 6, pp. 591-596.e1.
Leeflang, M.M., Deeks, J.J., Takwoingi, Y. & Macaskill, P. 2013, “Cochrane diagnostic test accuracy reviews”, Systematic reviews, Vol. 2, no. 1, pp. 82-87.
Lindsey, R.W. & Gugala, Z. 2010, “Clearing the cervical spine.” in: Spine trauma, eds. V. Patel et al, Springer, Galveston, pp. 11-31.
Lipowski, E.E. 2008, “Developing great research questions”, American Journal of Health-System Pharmacy : AJHP : Official Journal of the American Society of Health-System Pharmacists, Vol. 65, no. 17, pp. 1667-1670.
Looby, S. & Flanders, A. 2011, “Spine trauma”, Radiologic Clinics of North America, Vol. 49, no. 1, pp. 129-163.
MacDonald, R., Schwartz, M., Mirich, D., Sharkey, P. & Nelson, W. 1990, “Diagnosis of cervical spine injury in motor vehicle crash victims: How many X-rays are enough?”, Journal of Trauma and Acute Care Surgery, Vol. 30, no. 4, pp. 392-397.
Mathen, R., Inaba, K., Munera, F., Teixeira, P.G., Rivas, L., McKenney, M., et al. 2007, “Prospective evaluation of multislice computed tomography versus plain radiographic cervical spine clearance in trauma patients”, The Journal of Trauma, Vol. 62, no. 6, pp. 1427-1431.
Mariska M.G. Leeflang, Jonathan J. Deeks, Constantine Gatsonis, Patrick M.M. Bossuyt, on behalf of the Cochrane Diagnostic Test Accuracy Working Group & Cochrane Diagnostic Test Accuracy Working Group 2008, “Systematic Reviews of Diagnostic Test Accuracy”, Annals of Internal Medicine, Vol. 149, no. 12, pp. 889-897.
McCulloch, P.T., France, J., Jones, D.L., Krantz, W., Nguyen, T., Chambers, C., et al. 2005, “Helical computed tomography alone compared with plain radiographs with adjunct computed tomography to evaluate the cervical spine after high-energy trauma”, Journal of Bone and Joint Surgery, Vol. 87, no. 11, pp. 2388-2394.
Mortazavi, M., Gore, P.A., Chang, S., Tubbs, R.S. & Theodore, N. 2011, “Pediatric cervical spine injuries: A comprehensive review”, Child’s Nervous System, Vol. 27, no. 5, pp. 705-717.
McCracken, B., Klineberg, E., Pickard, B. & Wisner, D.H. 2013, “Flexion and extension radiographic evaluation for the clearance of potential cervical spine injures in trauma patients”, European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Society, Vol. 22, no. 7, pp. 1467-1473.
Mower, W.R., Hoffman, J.R., Pollack, C.V., Zucker, M.I., Browne, B.J., Wolfson, A.B., et al. 2001, “Use of plain radiography to screen for cervical spine injuries”, Annals of Emergency Medicine, Vol. 38, no. 1, pp. 1-7.
Munera, F., Rivas, L.A., Nunez Jr, D.B. & Quencer, R.M. 2012, “Imaging evaluation of adult spinal injuries: Emphasis on multidetector CT in cervical spine trauma”, Radiology, Vol. 263, no. 3, pp. 645-660.
Nguyen, G.K. & Clark, R. 2005, “Adequacy of plain radiography in the diagnosis of cervical spine injuries”, Emergency Radiology, Vol. 11, no. 3, pp. 158-161.
Noll, B. 2006, “National Inventory of Hospital Trauma Centers”, Journal of Trauma Nursing, Vol. 13, no. 3, pp. 148-149.
Oliver, J. & Middleditch, A. 2005, Functional anatomy of the spine: Alison Middleditch, Jean Oliver, Butterworth-Heinemann, Oxford.
Ong, A.W., Rodriguez, A., Kelly, R., Cortes, V., Protetch, J. & Daffner, R.H. 2006, “Detection of cervical spine injuries in alert, asymptomatic geriatric blunt trauma patients: Who benefits from radiologic imaging?”, The American Surgeon, Vol. 72, no. 9, pp. 773-777.
Özsarlak, Ö. 2007, “Cervical trauma.” in: Spinal imaging, eds. J. Goethem et al, Springer, New York, pp. 279-324.
Phal, P.M. & Anderson, J.C. 2006, “Imaging in spinal trauma”, Seminars in Roentgenology, Vol. 41, no. 3, pp. 190-195.
Pimentel, L. & Diegelmann, L. 2010, “Evaluation and management of acute cervical spine trauma”, Emergency Medicine Clinics of North America, Vol. 28, no. 4, pp. 719-738.
Platzer, P., Hauswirth, N., Jaindl, M., Chatwani, S., Vecsei, V. & Gaebler, C. 2006, “Delayed or missed diagnosis of cervical spine injuries”, The Journal of Trauma, Vol. 61, no. 1, pp. 150-155.
Pneumaticos, S.G., Triantafyllopoulos, G.K. & Giannoudis, P.V. 2014, “Injuries of the cervical spine”, European Surgical Orthopaedics and Traumatology: The EFORT Textbook, , pp. 717-741.
Powe, C.B. 2006, “Cervical spine clearance in the blunt trauma patient: A review of current management strategies”, Journal of Trauma Nursing : The Official Journal of the Society of Trauma Nurses, Vol. 13, no. 2, pp. 80.
Quann, J.T. & Sidwell, R.A. 2011, “Imaging of the cervical spine in injured patients”, Surgical Clinics of North America, Vol. 91, no. 1, pp. 209-216.
Raby, N., Berman, L., Morley, S. & de Lacey, G. 2014, Accident and emergency radiology: A survival guide, 3rd edn, Elsevier Health Sciences, Philadelphia.
Reeves, B.C. 2008, “Principles of research: Limitations of non-randomized studies”, Surgery (Oxford), Vol. 26, no. 3, pp. 120-124.
Reitsma, J.B., Moons, K.G., Bossuyt, P.M. & Linnet, K. 2012, “Systematic reviews of studies quantifying the accuracy of diagnostic tests and markers”, Clinical chemistry, Vol. 58, no. 11, pp. 1534-1545.
Rutjes, A.W., Reitsma, J.B., Vandenbroucke, J.P., Glas, A.S. & Bossuyt, P.M. 2005, “Case–control and two-gate designs in diagnostic accuracy studies”, Clinical chemistry, Vol. 51, no. 8, pp. 1335-1341.
Richards, P.J. 2005, “Cervical spine clearance: A review”, Injury, Vol. 36, no. 2, pp. 248-269.
Rybicki, F., Nawfel, R.D., Judy, P.F., Ledbetter, S., Dyson, R.L., Halt, P.S., et al. 2002, “Skin and thyroid dosimetry in cervical spine screening: Two methods for evaluation and a comparison between a helical CT and radiographic trauma series”, American Journal of Roentgenology, Vol. 179, no. 4, pp. 933-937.
Sharma, O.P., Oswanski, M.F., Yazdi, J.S., Jindal, S. & Taylor, M. 2007, “Assessment for additional spinal trauma in patients with cervical spine injury”, The American Surgeon, Vol. 73, no. 1, pp. 70-73.
Sheikh, K., Belfi, L.M., Sharma, R., Baad, M. & Sanelli, P.C. 2012, “Evaluation of acute cervical spine imaging based on ACR Appropriateness Criteria®”, Emergency radiology, Vol. 19, no. 1, pp. 11-17.
Simpson, A.K. & Harris, M.B. 2013, “Cervical spine clearance.” in: Orthopedic traumatology, eds. M. Sethi et al, Springer, New York, pp. 23-39.
Stiell, I.G., Clement, C.M., McKnight, R.D., Brison, R., Schull, M.J., Rowe, B.H., et al. 2003, “The canadian C-spine rule versus the NEXUS low-risk criteria in patients with trauma”, New England Journal of Medicine, Vol. 349, no. 26, pp. 2510-2518.
Stiell, I.G., Wells, G.A., Vandemheen, K.L., Clement, C.M., Lesiuk, H., De Maio, V.J., et al. 2001, “The canadian C-spine rule for radiography in alert and stable trauma patients”, Jama, Vol. 286, no. 15, pp. 1841-1848.
Tatsioni, A., Zarin, D.A., Aronson, N., Samson, D.J., Flamm, C.R., Schmid, C. & Lau, J. 2005, “Challenges in systematic reviews of diagnostic technologies”, Annals of Internal Medicine, vol. 142, no. 12_Part_2, pp. 1048-1055.
Theocharopoulos, N., Chatzakis, G. & Damilakis, J. 2009, “Is radiography justified for the evaluation of patients presenting with cervical spine trauma?”, Medical Physics, Vol. 36, no. 10, pp. 4461-4470.
Theologis, A.A., Dionisio, R., Mackersie, R., McClellan, R.T. & Pekmezci, M. 2014, “Cervical spine clearance protocols in level 1 trauma centers in the united states”, Spine, Vol. 39, no. 5, pp. 356-361.
Tins, B. & Cassar-Pullicino, V. 2007, “Controversies in “clearing” trauma to the cervical spine”, Seminars in ultrasound, CT and MRIElsevier, , pp. 94.
Tubbs, R.S., Salter, E.G. & Oakes, W.J. 2004, “The accessory atlantoaxial ligament”, Neurosurgery, Vol. 55, no. 2, pp. 400-404.
Van Goethem, J.W.M., Maes, M., Özsarlak, Ö., van den Hauwe, L. & Parizel, P.M. 2005, “Imaging in spinal trauma”, European Radiology, Vol. 15, no. 3, pp. 582-590.
Virgili, G., Conti, A.A., Murro, V., Gensini, G.F. & Gusinu, R. 2009, “Systematic reviews of diagnostic test accuracy and the Cochrane collaboration”, Internal and emergency medicine, vol. 4, no. 3, pp. 255-258.
Wade, R., Corbett, M. & Eastwood, A. 2013, “Quality assessment of comparative diagnostic accuracy studies: Our experience using a modified version of the QUADAS‐2 tool”, Research Synthesis Methods, Vol. 4, no. 3, pp. 280-286.
White, A. & Schmidt, K. 2005, “Systematic literature reviews”, Complementary therapies in medicine, Vol. 13, no. 1, pp. 54-60.
White, A.P., Kerr, S., Mendel, R.C., Hannallah, D. & Vaccaro, A.R. 2007, “Imaging update on cervical spinal trauma, instability screening, and clearance”, Seminars in Spine Surgery, Vol. 19, no. 2, pp. 98-105.
Whiting, P., Rutjes, A.W., Reitsma, J.B., Bossuyt, P.M. & Kleijnen, J. 2003, “The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews”, BMC medical research methodology, Vol. 3, no. 1, pp. 25-78.
Whiting, P.F., Rutjes, A.W.S., Westwood, M.E., Mallett, S., Deeks, J.J., Reitsma, J.B., Leeflang, M.M.G., Sterne, J.A.C., Bossuyt, P.M.M. & QUADAS-2 Group 2011, “QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies”, Annals of Internal Medicine, Vol. 155, no. 8, p. 529-W-161.
Widder, S., Doig, C., Burrowes, P., Larsen, G., Hurlbert, R.J. & Kortbeek, J.B. 2004, “Prospective evaluation of computed tomographic scanning for the spinal clearance of obtunded trauma patients: preliminary results”, Journal of Trauma and Acute Care Surgery, Vol. 56, no. 6, pp. 1179-1184.
Worster, A. & Carpenter, C. 2008, “Incorporation bias in studies of diagnostic tests: How to avoid being biased about bias”, Cjem, Vol. 10, no. 2, pp. 174-175.
Wu, L., Zhang, Y. & Zeng, X. 2013, “The QUADAS-2 tool for the quality assessment of diagnostic accuracy study: An introduction”, Journal of Hubei University of Medicine, Vol. 32, no. 3, pp. 201-208.
Yates, D., Aktar, R., Hill, J. & Guideline Development Group 2007, “Assessment, investigation, and early management of head injury: Summary of NICE guidance”, BMJ (Clinical Research Ed.), Vol. 335, no. 7622, pp. 719-720.
Appendixes
Appendix 1: NEXUS low-risk criteria
| Criteria | Number |
| There is no indication of intoxication. | 1 |
| No posterior mid-line cervical spine tenderness | 2 |
| No painful distracting injuries | 3 |
| No focal neurological shortfall and a normal alertness level | 4 |
Appendix 2: Canadian Cervical Spine Group low-risk Criteria
| Criteria | Number |
| Patient is fully alert with a Glasgow Coma Scale score less than 15 | 1 |
| Patient is not subject to any high-risk factors, such as paresthesia in the hands and feet | 2 |
| · Patients being under the age of 65 or the absence of a dangerous trauma mechanism (e.g. a fall of less than 3m or a high-speed motor vehicle, motorbike or bicycle collision);
|
3 |
| A lack of mid-line cervical tenderness and a delay in the onset of neck pain); | 4 |
| Patient has the ability to actively move their neck 45 degrees to the right and left | 5 |
Appendix 3. A 2 x 2 table for data collection
Adapted from (Centre for Reviews and Dissemination, 2009).
| Reference Standard. | |||
| True | False | ||
|
Result of index tests |
Positive | TP | FP |
| Negative | FN | TN | |
| Total | TP+ FN | FP+TN | |
Where TP = True positive
TN = True negative
FP = False positive
FN = False negative
Sensitivity = TP / TP + FN
The proportion of people with the target condition who have a positive test result.
Specificity = TN / TN + FP
The proportion of people without the target condition who have a negative test result.
Appendix 4: Data extraction tool
(adapted from Higgins et al, 2008).
| FACTORS | ASSESSMENT | COMMENTS | |||||||
| TYPE OF STUDY | |||||||||
| 1. Is the study described as randomized? | Yes | Unclear | No | ||||||
| NB. Please answer “No” if the study is a crossover | |||||||||
| or quasi-randomized trial. | Exclude | ||||||||
| PARTICIPANTS | |||||||||
| 2. Were participants diagnosed as patients | Yes | Unclear | No | ||||||
| with disease of interest? | |||||||||
| Exclude | |||||||||
| Subgroups available? | |||||||||
| 3. Were participants of the prespecified age? | Yes | Unclear | No | ||||||
| NB: Please answer „Yes“, If mix age participants | |||||||||
| i.e. both >18 years and < 18 years are included | Exclude | ||||||||
| and state it as comments. | |||||||||
| No: If only < 18 years. | |||||||||
| INTERVENTIONS | |||||||||
| 4. Were comparison groups treated with | Yes | Unclear | No | ||||||
| prespecified intervention in one group and | |||||||||
| control intervention in other group? | |||||||||
| NB: study can have 3 arms e.g. CT arm, CT+RT | Exclude | ||||||||
| (CMT) arm or RT arm, if so please cross „Yes“ and | |||||||||
| state it as comments. | |||||||||
| OUTCOMES | |||||||||
| 5. Did the study report prespecified | Yes | Unclear | No | ||||||
| outcomes? | |||||||||
| Exclude | |||||||||
| FINAL DECISION | |||||||||
| 1 X “No” | = EXCLUDE | ||||||||
| 1 X “Unclear” = UNCLEAR | |||||||||
Appendix 5. QUADAS-2: A Revised Tool for the Quality Assessment of Diagnostic Accuracy Studies (adapted from Whiting et al., 2011).
| Domain 1: Patient selection | |
| A. Risk of bias | |
| Describe methods of patient selection: | |
| Was a consecutive or random sample of patients enrolled? | Yes/No/Unclear |
| Was a case-control design avoided? | Yes/No/Unclear |
| Did the study avoid inappropriate exclusions? | Yes/No/Unclear |
| Could the selection of patients have introduced bias? | RISK: LOW/HIGH/UNCLEAR |
| B. Concerns regarding applicability | |
| Describe included patients (prior testing, presentation, intended use of index test and setting): | |
| Is there concern that the included patients do not match the review question? | CONCERN: LOW/HIGH/UNCLEAR |
| Domain 2: Index test(s) (if more than 1 index test was used, please complete for each test) | |
| A. Risk of bias | |
| Describe the index test and how it was conducted and interpreted: | |
| · Were the index test results interpreted without knowledge of the results of the reference standard? | Yes/No/Unclear |
| · If a threshold was used, was it pre-specified? | Yes/No/Unclear |
| Could the conduct or interpretation of the index test have introduced bias? | RISK: LOW/HIGH/UNCLEAR |
| B. Concerns regarding applicability | |
| Is there concern that the index test, its conduct, or interpretation differ from the review question? | CONCERN: LOW/HIGH/UNCLEAR |
| Domain 3: Reference standard | |
| A. Risk of bias | |
| Describe the reference standard and how it was conducted and interpreted: | |
| · Is the reference standard likely to correctly classify the target condition? | Yes/No/Unclear |
| · Were the reference standard results interpreted without knowledge of the results of the index test? | Yes/No/Unclear |
| Could the reference standard, its conduct, or its interpretation have introduced bias? | RISK: LOW/HIGH/UNCLEAR |
| B. Concerns regarding applicability | |
| Is there concern that the target condition as defined by the reference standard does not match the review question? | CONCERN: LOW/HIGH/UNCLEAR |
| Domain 4: Flow and timing | |
| A. Risk of bias | |
| Describe any patients who did not receive the index test(s) and/or reference standard or who were excluded from the 2×2 table (refer to flow diagram):
Describe the time interval and any interventions between index test(s) and reference standard: |
|
| · Was there an appropriate interval between index test(s) and reference standard? | Yes/No/Unclear |
| · Did all patients receive a reference standard? | Yes/No/Unclear |
| · Did patients receive the same reference standard? | Yes/No/Unclear |
| · Were all patients included in the analysis? | Yes/No/Unclear |
| Could the patient flow have introduced bias? | RISK: LOW/HIGH/UNCLEAR |
Appendix 6. PRISMA Flow Diagram
| Records identified through database searching (n = 235 ) |
| Additional records identified through other sources (n = 0 ) |
| Records after duplicates removed (n = 157 ) |
| Records screened (n = 157 ) |
| Records excluded (n = 146 ) |
| Full-text articles assessed for eligibility (n = 11 ) |
| Full-text articles excluded (n = 7 ) |
| Studies included in qualitative synthesis (n = 4 ) |
| Studies included in quantitative synthesis (meta-analysis) (n = 0 ) |
Appendix 7. Tabular presentation for QUADAS-2 results
| Study | RISK OF BIAS | APPLICABILITY CONCERNS | |||||
| PATIENT SELECTION | INDEX TESTS | REFERENCE STANDARD | FLOW AND TIMING | PATIENT SELECTION
|
INDEX TEST | REFERENCE STANDARD | |
| Nguyen & Clark (2005).
|
L | J | L | J | L | J | J |
| McCulloch et al. (2005). | L | J | L | J | L | J | J |
| Mathen (2007)
|
L | L | L | J | L | J | J |
| Bailitz et al. (2009). | J | J | L | J | J | J | J |
Low Risk J L High Risk ? Unclear
Appendix 8. Proportion of studies of low, high or unclear risk of bias.
Appendix 9. Proportion of studies of low, high or unclear concern regarding applicability
.
Appendix 10. CT scanning types and used protocols.
| CT scanning types and used protocols | Study |
| General Electric Speed Helical CT scanner was used. The resulting pictures being captured were through 2.5mm thickness cuts with 3.75mm rotations | Widder (2004). |
| General Electric Speed Helical CT scanner was used. The images from the occiput to T4 were within the trauma protocol such that is includes 2.5*2-mm cuts with sagittal reformats and standard coronal | McCulloch et al. (2005). |
| MCTs was used (from occiput to T1) using a four channel CT scanner (Phillips MX 8000, Bothell, WA) with a collimation of 2 mm. There has been a remade of Coronal and sagittal reformation images using 1.5-mm to 2-mm intervals originated from the axial sources images through utilizing a standard workstation. | Mathen (2007). |
| Not provided | Bailitz et al (2009). |
Appendix 11. Summary of studies’ characteristics.
| Study
|
Sample size | Index test | Comparator test | Reference standard | Sensitivity of index test (CT) | Sensitivity of comparator test (plain radiography) | Specificity of index test (CT) | Specificity of comparator test (plain radiography) |
| Widder (2004) | 102 | Helical CT | Plain radiography with minimum 3-views | Radiologists interpretation of CT and final diagnosis in medical record at discharge | 100% | 39% | %100 | %98 |
| McCullon et al. (2005) | 407 | Helical CT | Plain radiography with minimum 3-views | Radiologists interpretation of all radiographic studies and all clinical data was the reference standard | 98% | %45 | %98 | %97 |
| Mathen (2007) | 667 | MDCT | Plain radiography with minimum 3-views | Radiologist interpretation of CT scan and all clinical data | %100 | %45 | %99.5 | %97.4 |
| Bailitz et al. (2009) | 1505 | MDCT | Plain radiography with minimum 3-views | Radiologists interpretation of CT and final diagnosis in medical record at discharge | 100% | %36 | ? | ? |
Appendix 12: Determination of CSI risk in accordance to the probability of CSF.
| Risk level | Mechanism and features | Probability of fracture |
| Patients with high risk of CSI | – Patients with focal neurological deficit
– Severe head injury, – High-energy mechanism and age older than 50 years. |
Greater than 10% |
| Patients with high moderate of CSI | – High-energy mechanism and age 50 years of younger
– Moderate –energy mechanism and older than 50 years. |
Between (4-10%). |
| Patients with low risk of CSI | – Moderate-energy mechanism and 50 years older
or younger – low energy mechanism. |
Less than 4 %. |
Expert paper writers are just a few clicks away
Place an order in 3 easy steps. Takes less than 5 mins.