18 
When Should Follow-Up Computed Tomography Scans Be Obtained?
BRIEF ANSWER
The evidence supports a level III recommendation for use of serial computed tomography (CT) for clinical management and determination of prognosis following severe traumatic brain injury (TBI). However, use of serial CT scans to improve patient outcome can be supported only as a level III recommendation.
Background
Clearly, CT scans have improved our understanding of the evolution of TBI. In their 1977 seminal study on the value of head CT, French and Dublin1 reported a 52% incidence of new lesions or progression of known lesions in their patients who had two or more CT scans following TBI (class III data). However, no prospective randomized controlled trials have evaluated serial head CT scanning in TBI patients. Identification of the optimal number and timing of follow-up head CT scans in these patients requires further evaluation with respect to outcome-based measures.
Pearl
Although entry criteria and patient populations vary, several studies have documented that more than half of head-injured patients demonstrate new findings on repeat head CT scans.
Literature Review
Determining Prognosis with Serial Computed Tomography Scans
Several studies have provided evidence that serial CT scanning may be helpful in predicting outcome. Kobayashi et al2 (class III data) performed serial CT scanning on 138 patients with severe head injuries [Glasgow Coma Scale (GCS) score <8). Scans were obtained upon admission (within hours of the injury) and after 1, 3, and 7 days, and after 1 month. New findings (not visualized on the initial CT scan) were seen in 91 of the 138 patients, and significant new lesions were found in 60 of the 91 patients. A significant correlation existed between the development of new lesions and poor outcome at the time of discharge. The authors concluded that serial CT scanning can help to predict outcome in severe TBI patients and that serial scans are especially useful when clinical symptoms may be masked by therapy.
Cooper et al3 (class III data) also found that serial CT scanning was useful in prognostic assessments and that it may also be of value in improving the accuracy of clinical assessment. Their study found that follow-up CT scans demonstrated new lesions in 30 of 58 patients (52%) and that the development of new lesions was strongly associated with a bad outcome. The authors could not find any evidence that operation or reoperation indicated by the results of repeat CT scanning actually improved outcome.
Lobato et al4 found that 51% of 587 severely head-injured patients demonstrated CT changes that were significant enough to change their initially assigned category (either worse or better) on the Marshall CT classification scale (class III data). Of the 587 patients, 97% had their first repeat scan within 3 days after injury, with the average interval between the first and second scan being 37 hours. Final outcomes were more closely correlated with the repeat scans than with the initial CT diagnoses. Thus, a pathologic categorization made by using an early repeat scan rather than the initial scan might be most useful for prognostic purposes. The authors recommended follow-up CT scanning within the first 2 days after injury.
Pearl
The appearance of new lesions has been associated with a poor prognosis after TBI, but it has been difficult to show that prompt detection and evacuation of these new lesions improves outcome.
Serial Computed Tomography Scans in Clinical Management
It is well known that lesions accompanying TBI evolve over time. However, even though the literature suggests that sequential CT scanning can help in the diagnosis of new lesions, and even though sequential CT scanning following TBI is a common practice, it is not clear that more expeditious detection of new lesions necessarily improves outcome. To complicate this issue further, the proliferation of CT scanners in general hospitals and improvements in trauma systems and in patient transport practices have reduced the interval between head injury and initial CT scan in many regions. As a result, it has become more likely that an admission scan may demonstrate a lesion very early in its course, before it has reached its most dramatic appearance. Since appropriate timing of follow-up CT scans remains poorly defined except for a few specific lesions, and since different lesions appear to have different propensities to progress, the safest strategy may be to tailor serial scanning to the type of lesions found on the initial CT scan and to the postinjury time at which the initial scan was obtained.
Based on an analysis of the European Brain Injury Consortium database, Servadei et al5 recommended a follow-up CT scan whenever the initial scan reveals evidence of diffuse injury (level III recommendation). They found that one in six such patients demonstrates significant CT evolution of the lesions. When the initial scan revealed Marshall-type II, III, and IV diffuse injuries,6 subsequent scans had high rates of progression to mass lesions (13–14% for types II and III and 20% for type IV). The authors suggested that a CT scan be repeated within 12 hours whenever the initial scan was obtained within 3 hours of injury and within 24 hours in all other instances (level III recommendation). A third scan was recommended on the third day after trauma, although the authors acknowledged that this recommendation was empiric.
Compared with other types of brain injuries, traumatic intracerebral hematomas and epidural hematomas probably have the greatest propensity to expand. Yamaki et al7 (class III data) performed a retrospective review of 48 patients who developed a traumatic intracerebral hemorrhage ≥3 cm in diameter at some point during their acute hospitalization. The initial scan was performed between 0.5 and 6 hours after the head injury. Traumatic intracerebral hematomas >3 cm in diameter developed in 56% of patients within 6 hours, in 81% of patients within 12 hours, and in 100% of patients within 24 hours of the initial trauma. Servadei et al8 (class III data) also found that the lesions that are most prone to enlarge are epidural hematomas and intracerebral hemorrhages. Performance of the initial CT scan within 3 hours of the initial trauma did not exclude rapid evolution of a new mass lesion or expansion of a lesion identified initially. They recommended that a follow-up CT scan be obtained within 12 hours in patients in whom the first scan was obtained within 3 hours of the initial injury.
Pearl
Related posts:
Should I Monitor Jugular Venous Oxygen Saturation?
Are Steroids Indicated in the Treatment of Head Injury?
Do I Need to Decompress Patients with Spinal Cord Injury Right Away?
How Do I Diagnose and Manage SCIWORA?
What Is the Outcome of Patients with Mild, Moderate, or Severe Traumatic Brain Injury?
Should Head-Injured Patients Be Taken to the Nearest Hospital or to a Hospital with Neurosurgical Capability That Is Farther Away?
