Free
Critical Care Medicine  |   June 2017
Clinical Judgment Is Not Reliable for Reducing Whole-body Computed Tomography Scanning after Isolated High-energy Blunt Trauma
Author Notes
  • From the Grenoble Alpes Trauma Center, Pôle Anesthésie-Réanimation, CHU Grenoble Alpes, Grenoble, France (T.M., J-F.P., P.B.); Université Grenoble Alpes, Grenoble, France (T.M., V.B., J-L.B., G.F., J-F.P., P.B.); INSERM U1216, Grenoble, France (T.M., J-F.P., P.B.); Département de Médecine d’Urgence, CHU Grenoble Alpes, Grenoble, France (V.B., L.S.); Pôle Recherche, CHU Grenoble Alpes, Grenoble, France (J-L.B.); Pole Imagerie, CHU Grenoble Alpes, Grenoble, France (G.F., F.T.); Divisions of Trauma Anesthesiology and Critical Care Medicine, Department of Anesthesiology, R Adams Cowley Shock Trauma Center, Baltimore, Maryland (S.M.G.).
  • T.M. and V.B. contributed equally to this article.
    T.M. and V.B. contributed equally to this article.×
  • Submitted for publication September 19, 2016. Accepted for publication February 17, 2017.
    Submitted for publication September 19, 2016. Accepted for publication February 17, 2017.×
  • Address correspondence to Dr. Bouzat: Pôle d’Anesthésie-Réanimation, CHU Grenoble Alpes, F-38000 Grenoble, France. PBouzat@chu-grenoble.fr. Information on purchasing reprints may be found at www.anesthesiology.org or on the masthead page at the beginning of this issue. Anesthesiology’s articles are made freely accessible to all readers, for personal use only, 6 months from the cover date of the issue.
Article Information
Critical Care Medicine / Clinical Science / Critical Care / Radiological and Other Imaging / Technology / Equipment / Monitoring / Trauma / Burn Care
Critical Care Medicine   |   June 2017
Clinical Judgment Is Not Reliable for Reducing Whole-body Computed Tomography Scanning after Isolated High-energy Blunt Trauma
Anesthesiology 6 2017, Vol.126, 1116-1124. doi:10.1097/ALN.0000000000001617
Anesthesiology 6 2017, Vol.126, 1116-1124. doi:10.1097/ALN.0000000000001617
Abstract

Background: The purpose of this study was to test the diagnostic performance of clinical judgment for the prediction of a significant injury with whole-body computed tomography scanning after high-energy trauma.

Methods: The authors conducted an observational prospective study in a single level-I trauma center. Adult patients were included if they had an isolated high-energy injury. Senior trauma leaders were asked to make a clinical judgment regarding the likelihood of a significant injury before performance of a whole-body computed tomography scan. Clinical judgments were recorded using a probability diagnosis scale. The primary endpoint was the diagnosis of a serious-to-critical lesion on the whole-body computed tomography scan. Diagnostic performance was assessed using receiver operating characteristic analysis.

Results: Of the 354 included patients, 127 patients (36%) had at least one injury classified as abbreviated injury score greater than or equal to 3. The area under the receiver operating characteristic curve of the clinical judgment to predict a serious-to-critical lesion was 0.70 (95% CI, 0.64 to 0.75%). The sensitivity of the clinical judgment was 82% (95% CI, 74 to 88%), and the specificity was 49% (95% CI, 42 to 55%). No patient with a strict negative clinical examination had a severe lesion (n = 19 patients). The sensitivity of the clinical examination was 100% (95% CI, 97 to 100%) and its specificity was 8% (95% CI, 5 to 13%).

Conclusions: Clinical judgment alone is not sufficient to reduce whole-body computed tomography scan use. In patients with a strictly normal physical examination, whole-body computed tomography scanning might be avoided, but this result deserves additional study in larger and more diverse populations of trauma patients.

What We Already Know about This Topic
  • Whole-body computed tomography (CT) scanning is performed to evaluate the extent of injury in patients who have sustained high-energy trauma. In such patients, the use of clinical judgment by expert physicians in reducing the need for CT scans was determined.

What This Article Tells Us That Is New
  • The diagnostic performance of clinical judgment in the prediction of the presence of significant lesions on computed tomography (CT) scan was modest and was considered to be insufficient. However, in patients with a completely normal physician examination, clinical judgment predicted the lack of a lesion on CT scan.

  • The data suggest that clinical judgment is not sufficient to reduce the need for whole-body CT scans in patients with high-energy trauma.

WHOLE-BODY computed tomography (CT) scanning is the imaging modality of choice for the diagnosis of injuries after severe trauma.1  It provides an exhaustive assessment of posttraumatic lesions and prevents missing occult lesions2  with a beneficial effect on mortality.3,4  In stable patients with suspected severe injury because of high-energy trauma, a whole-body CT scan is increasingly performed to diagnose occult lesions. However, occult injuries are not always clinically relevant,5–7  and the benefits of whole-body CT scanning on mortality have been shown in physiologically unstable trauma patients.8  Recently, a systematic review failed to identify any randomized controlled trials that studied whole-body CT scanning in stable patients with high-energy trauma.9  Moreover, x-ray exposes patients to the risk of radiation-induced cancer,10  and the infusion of an iodine agent is associated with contrast-induced nephropathy.11  Hence, the possible beneficial effects of a whole-body CT scan in stable, high-energy trauma patients may be balanced by long-term safety issues. To avoid futile radiation, the clinical judgment of the physician may screen patients at risk for relevant posttraumatic lesions.12,13  This judgment combines the circumstances of the trauma with clinical examination and bedside ultrasonography. However, the predictive value of this clinical judgment remains unknown for the detection of serious, severe, or critical injuries. The aim of our study was to estimate the diagnostic accuracy of clinical judgment for the prediction of a significant injury on the whole-body CT scan. We hypothesized that clinical judgment alone can obviate the need for whole-body CT scanning.
Materials and Methods
Study Design and Patients
We conducted a prospective observational study in a level-I trauma center (Grenoble University Hospital, Grenoble, France) from June 2014 to June 2015. The Regional Institutional Ethics Committee (Comité d’Ethique des Centres d’Investigation Clinique de l’Inter-Région Rhône-Alpes-Auvergne, institutional review board No. 5708) approved the study design and, given its observational nature, waived the requirements for written informed consent from each patient.
Inclusion criteria were patients older than 15 yr, stable from the prehospital field to the hospital admission, and admitted in the trauma center for suspected severe blunt trauma because of a high-energy mechanism. Stability was defined by a Glasgow Coma Scale (GCS) score strictly greater than 13, an oxygen saturation measured by pulse oximetry (Spo2) strictly greater than 92% without oxygen supply or mechanical ventilation, and a systolic blood pressure (SBP) strictly more than 90 mmHg without fluid resuscitation or vasopressor. A high-energy trauma was defined according to the Vittel criteria14 : fall more than 6 m, ejection, projection, blast, crushed injuries, the presence of a deceased or severe trauma patient in the same accident, or an accident with an estimated high velocity.15,16  All of the patients had a whole-body CT scan within 6 h of their initial injury according to our institutional protocol.17  Noninclusion criteria were unstable patients defined as having the following: (1) an SBP less than 90 mmHg and/or the need for more than 20 ml/kg intravenous fluid therapy and/or vasopressor; (2) an Spo2 less than 92% without oxygen supply; (3) the use of mechanical ventilation; or (4) a GCS less than or equal to 13.
Study Protocol and Data Collection
Consecutive patients were included at their admission to the emergency department based on their trauma history and physiologic stability. The following clinical data were prospectively collected: age; sex; mechanism of injury; vital variables (heart rate, SBP, Spo2, and GCS) on admission; the Mechanism, Glasgow, Age, and Arterial Pressure score18 ; the new injury severity score with detailed abbreviated injury score (AIS)19,20 ; the need for surgery, disposition of the patient after the emergency department (home, standard surgical ward, or intensive care unit [ICU]); length of hospital stay; in-hospital mortality; and mortality predicted by the Trauma Injury Severity Score.21  The physical examination was performed by the attending physician and was detailed for each body area (head, neck, face, thorax, abdomen, spine, and limbs). The result of the bedside ultrasonography was also reported. Ultrasonography consisted of a focused assessment with sonography for trauma on admission.22  According to the trauma history, clinical examination, and bedside ultrasonography, the physician was asked to report in a dedicated patient file his/her judgment for a significant posttraumatic lesion according to a probability diagnosis scale: 0 = sure of the absence of a lesion; 1 = suspect the absence of a lesion; 2 = suspect the presence of a lesion; and 3 = sure of the presence of a lesion.23  Before whole-body CT scan completion, the attending physician was also asked to predict the disposition of the patient after the emergency department, including home, standard surgical ward, or ICU. The attending physician was the senior trauma leader (anesthesiologist or emergency physician) in charge of the trauma bay.
Whole-body CT Scan
Each patient was transported to the radiology department to undergo a whole-body CT scan. CT scans were conducted using Philips Brilliance 40, Philips Brilliance 64 (Philips Medical Systems, The Netherlands), or Siemens Sensation 16 (Siemens Healthcare, Germany). Whole-body CT protocol consisted of the following acquisitions: (1) a nonenhanced encephalic CT scan; (2) a nonenhanced CT scan of the neck, from the base of the skull to the level of the second thoracic vertebra; (3) a contrast-enhanced CT scan of the thorax, abdomen, and pelvic regions from the level of the sixth cervical vertebra to the lesser trochanter; and (4) a contrast-enhanced scan of the abdomen from the diaphragmatic dome to the lesser trochanter. Arms were placed above the head after the CT scan of the head and neck. Optionally, mainly when a severe cervical trauma was suspected, a CT angiography of arterial supraaortic vessels was added. Regarding contrast-medium injection, a 120-ml bolus of iso-osmolar, nonionic iodinated contrast material (350 mg of iodine per milliliter, Iohexol [Omnipaque 350; GE Healthcare, United Kingdom]), followed by a saline flush of 40 ml, was injected into an antecubital vein at a flow rate of 4 ml/s. The data acquisition was initiated 6 s after 100-Hounsfield unit (HU) attenuation in the descending thoracic aorta. After an additional delay of 45 s, the abdominal portal venous-enhanced phase was acquired. For supraaortic vessel acquisition, an additional injection of 120 ml of the same contrast medium was performed, and acquisition was triggered at 75-HU attenuation in the ascending aorta. Image reconstruction included systematically 1-mm-thick axial CT images for all of the body parts in appropriate windows and level settings, 1-mm contiguous reconstructed images in axial and sagittal plane and bone settings (window, 2500 HU; level, 500 HU) of the cervical, thoracic, and lumbar spines and the pelvis. Whole-body CT scans were immediately evaluated and documented by a radiologic resident, supervised by a trauma-dedicated senior radiologist.
Endpoints
The primary outcome was the diagnosis of a serious, severe, and/or critical lesion using whole-body CT scan. Injury severity was defined by at least one injury classified as AIS greater than or equal to 3, because a score greater than or equal to 3 is regarded as a serious injury. All of the lesions were classified by two trained physicians (T.M. and V.B.) using the AIS 2005 catalog (https://www.aaam.org/) and the final report of the senior radiologist. In case of disagreement regarding the classification of a lesion, an expert in trauma care (P.B.) was responsible for final classification. Secondary outcomes were characteristics of the population with a serious-to-critical lesion after high-energy trauma and disposition after the emergency department to home, standard surgical ward, or ICU.
Study Size
To have a good diagnostic accuracy, we expected an area under the receiver operating characteristic curve equal to 0.8 for the clinical judgment to predict a significant posttraumatic injury. The number of patients to be included was set at 350 patients to obtain an acceptable 95% CI between 0.75 and 0.84.
Statistical Analysis
Descriptive statistics included frequencies and percentages for categoric variables and mean ± SD or median values (25th to 75th percentiles) for continuous variables according to their distribution. The diagnostic performance of physician judgment to predict a clinically relevant lesion on whole-body CT scan was evaluated using receiver operating characteristic (ROC) analysis. The ROC curve was analyzed using both nonparametric and parametric methods. The nonparametric analysis was conducted using the method of DeLong et al.,24  and 95% CIs were calculated using both asymptotic normal approximation and bootstrap resampling from 1,000 replicates. The parametric analysis was conducted by fitting a binomial model by maximum likelihood.
The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and positive and negative likelihood ratios (LR+ and LR–) were also reported for the clinical judgment considering two categories of patients, those with a high suspicion of severe lesion (probability diagnosis scale 2 and 3) and those with a low suspicion (probability diagnosis scale 0 and 1). The same parameters were calculated for the physical examination alone.
To offer insight into clinical consequences, decision curve analysis was also performed for the clinical judgment and the physical examination.25  These curves determine the relationship between a chosen predicted probability threshold and the relative value of false-positive and false-negative results to obtain a value of net benefit using the test at that threshold. Comparisons between the group of patients with AIS greater than or equal to 3 lesions and the group of patients with no significant lesion were performed using a chi-square test (or Fisher exact test, where appropriate) for categoric variables and a Student’s t test for continuous variables (or a nonparametric Mann–Whitney U test, where appropriate). Statistical analysis was performed with the software STATA 13.0 (Stata Corp, USA). All of the tests were two tailed, and a P value of 0.05 or less was considered statistically significant.
Results
Within the study period, 367 consecutive patients with a suspicion of severe blunt trauma after a high-energy mechanism according to the Vittel criteria were eligible. Flowchart of the study population is presented in figure 1, and 354 patients with high-energy trauma were finally included. Patient characteristics are presented in table 1. Whole-body CT scan revealed 127 patients (36%) with at least one injury classified as AIS greater than or equal to 3. Overall, 171 severe lesions were identified, and 36 patients (28%) had more than one severe injury. Lesions were more frequently located in the thorax (40%), head (16%), limbs (15%), and spinal (13%) areas and were less frequently located in the abdomen (7%), pelvic bone (5%), neck (2%), and facial (2%) areas. We found three patients with a critical lesion (AIS = 5), including two patients with a severe traumatic brain injury and one patient with a thoracic injury. Findings for the physical examination are detailed in table 2. On admission, 19 patients (5%) had a normal physical examination. After the emergency department, 126 patients (47%) were discharged to home within 24 h. Fifty-seven patients (16%) were admitted to the ICU, and 167 patients (36%) were admitted to the standard surgical ward with a median length of stay in the hospital of 6 days (range, 2 to 12 days). Seventy-one patients required surgery within 24 posttraumatic hours, including orthopedic surgery for 51 patients, spinal surgery for 14 patients, maxillofacial surgery for four patients, laparoscopy for one patient, and neurosurgery for one patient. One patient required a splenic embolization.
Table 1.
Patient Characteristics (n = 354 Patients)
Patient Characteristics (n = 354 Patients)×
Patient Characteristics (n = 354 Patients)
Table 1.
Patient Characteristics (n = 354 Patients)
Patient Characteristics (n = 354 Patients)×
×
Table 2.
Results of Physical Examination
Results of Physical Examination×
Results of Physical Examination
Table 2.
Results of Physical Examination
Results of Physical Examination×
×
Fig. 1.
Flowchart of the study population. AIS = abbreviated injury score; CT = computed tomography.
Flowchart of the study population. AIS = abbreviated injury score; CT = computed tomography.
Fig. 1.
Flowchart of the study population. AIS = abbreviated injury score; CT = computed tomography.
×
Primary Outcome
Of the 354 patients, clinical judgment was reported for 338 patients. Sixteen patients with missing data were excluded from the analysis (complete case analysis). The area under the curve (AUC) for clinical judgment to predict a significant lesion on the whole-body CT scan was 0.70 (95% CI, 0.64 to 0.75; fig. 2). Using a bootstrap technique, AUC was 0.70 (95% CI, 0.64 to 0.75). The ROC curve fitted to the binomial model with its 95% CI is presented in figure 3 (AUC = 0.73 [95% CI, 0.67 to 0.79]). A total of 210 patients had a high suspicion for a severe lesion (probability diagnosis scale 2 and 3), and 128 patients had a low suspicion (probability diagnosis scale 0 and 1). In the low-suspicion group, 21 patients had at least one lesion classified as AIS greater than or equal to 3. Four patients had a severe lesion in the limbs area only. Eighteen severe occult lesions were found in 17 patients (13%), because one patient had two AISs at greater than or equal to 3 lesions. In the high-suspicion group, only 97 of 210 patients had severe lesions. Hence, the sensitivity of the clinical judgment was 82% (95% CI, 74 to 88%) and the specificity was 49% (95% CI, 42 to 55%). The NPV of the clinical judgment was 84% (95% CI, 76 to 90%), and the PPV was 46% (95% CI, 39 to 53%). LR+ was 1.6 (95% CI, 1.4 to 1.9), and LR– was 0.4 (95% CI, 0.2 to 0.6). Regarding the performance of the physical examination, no patient with a strict negative clinical examination had a severe lesion (n = 19 patients), whereas 127 of 334 patients with a positive clinical examination had a severe injury. Hence, the sensitivity of the clinical examination was 100% (95% CI, 97 to 100%). However, its specificity was only 8% (95% CI, 5 to 13%). The NPV was 100% (95% CI, 83 to 100%), and the PPV was 15% (95% CI, 10 to 20%). LR+ was 1.1 (95% CI, 1.05 to 1.14), and LR– was 0.
Fig. 2.
Empirical receiver operating characteristic (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. The area under the ROC curve (AUC ROC) was 0.70 (95% CI, 0.64 to 0.75).
Empirical receiver operating characteristic (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. The area under the ROC curve (AUC ROC) was 0.70 (95% CI, 0.64 to 0.75).
Fig. 2.
Empirical receiver operating characteristic (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. The area under the ROC curve (AUC ROC) was 0.70 (95% CI, 0.64 to 0.75).
×
Fig. 3.
Smooth receiver operating characteristics (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. This curve was generated by fitting the binomial distribution. The gray area represents the 95% CI of the ROC curve. The area under the ROC curve (AUC ROC) was 0.73 (95% CI, 0.67 to 0.79).
Smooth receiver operating characteristics (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. This curve was generated by fitting the binomial distribution. The gray area represents the 95% CI of the ROC curve. The area under the ROC curve (AUC ROC) was 0.73 (95% CI, 0.67 to 0.79).
Fig. 3.
Smooth receiver operating characteristics (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. This curve was generated by fitting the binomial distribution. The gray area represents the 95% CI of the ROC curve. The area under the ROC curve (AUC ROC) was 0.73 (95% CI, 0.67 to 0.79).
×
Decision curves are presented in figure 4. At the acceptable probability threshold of 10% (one patient with a positive CT scan for every 10 patients), physical examination is more beneficial than clinical judgment: patients with abnormal physical examination benefit from systematic whole-body CT scanning at this threshold. With this strategy, 5 of 100 CT scans are avoided without increasing the number of missing injuries. On the other hand, a strategy based on clinical judgment leads to 30 futile CT scans in a population of 100 patients. For a probability threshold at 20% (one patient with a significant lesion for every five patients), a strategy based on a high suspicion after clinical judgment is more beneficial than a strategy based on physical examination. With this strategy, 1 of 100 CT scans are avoided without increasing the number of missing injuries.
Fig. 4.
Net benefit curves for the clinical judgment and the physical examination. These curves determine the relationship between a chosen predicted probability threshold (x-axis) and the relative value of false-positive and false-negative results to obtain a value of net benefit (y-axis) using the test at that threshold. At the acceptable threshold of 10%, there is no net benefit for the clinical judgment as compared with the physical examination.
Net benefit curves for the clinical judgment and the physical examination. These curves determine the relationship between a chosen predicted probability threshold (x-axis) and the relative value of false-positive and false-negative results to obtain a value of net benefit (y-axis) using the test at that threshold. At the acceptable threshold of 10%, there is no net benefit for the clinical judgment as compared with the physical examination.
Fig. 4.
Net benefit curves for the clinical judgment and the physical examination. These curves determine the relationship between a chosen predicted probability threshold (x-axis) and the relative value of false-positive and false-negative results to obtain a value of net benefit (y-axis) using the test at that threshold. At the acceptable threshold of 10%, there is no net benefit for the clinical judgment as compared with the physical examination.
×
Secondary Outcomes
Using univariate analysis, age, GCS, and the presence of an abnormal physical sign on admission were variables associated with the presence of a significant lesion on a whole-body CT scan (table 3). The predicted disposition before whole-body imaging was available for 300 patients. Fifty-eight of 100 patients with a prediction of home were finally discharged to the home, 29 of 52 patients with a prediction of ICU were admitted to the ICU, and 121 of 148 patients with a prediction of standard ward were finally admitted to a standard surgical ward. Hence, sensitivities for the clinical judgment were 58% (95% CI, 48 to 68%), 56% (95% CI, 41 to 70%), and 82% (95% CI, 75 to 88%) to predict final disposition at home, in the ICU, and in the surgical ward, respectively. Specificities were 95% (95% CI, 90 to 97%), 92% (95% CI, 88 to 95%), and 59% (95% CI, 51 to 67%), respectively.
Table 3.
Comparison between the Group of Patients with a Lesion Classified as AIS ≥ 3 and the Group of Patients with a Lesion Classified as AIS < 3
Comparison between the Group of Patients with a Lesion Classified as AIS ≥ 3 and the Group of Patients with a Lesion Classified as AIS < 3×
Comparison between the Group of Patients with a Lesion Classified as AIS ≥ 3 and the Group of Patients with a Lesion Classified as AIS < 3
Table 3.
Comparison between the Group of Patients with a Lesion Classified as AIS ≥ 3 and the Group of Patients with a Lesion Classified as AIS < 3
Comparison between the Group of Patients with a Lesion Classified as AIS ≥ 3 and the Group of Patients with a Lesion Classified as AIS < 3×
×
Discussion
In a cohort of stable patients with a suspicion of severe injury after high-energy blunt trauma, we observed a poor diagnostic performance of clinical judgment to predict a significant lesion on whole-body CT scan. Interestingly, the physical examination had excellent sensitivity, and 0 of the 19 patients with a normal examination on admission had a clinically relevant lesion after whole-body imaging. These results suggest that whole-body CT scan remains mandatory to assess occult injuries after high-energy trauma, except in cases of a strictly normal physical examination.
In our population, the incidence of severe lesions was high, because one third of patients presented with at least one lesion classified as AIS greater than or equal to 3. Conversely, the median new-injury severity score was low, demonstrating a large variability among injury severity in our cohort. Hence, accurate and early detection of these lesions is the main goal of the attending physician and argues for systematic indication of whole-body CT scan on admission. This imaging technique is increasingly used by trauma teams across the world despite well-known long-term issues and no strong scientific evidence.26  Although the true incidence is unknown, the risk of radiation-induced cancer is well established. All of the authors agree with it even if they fail to objectively quantify its incidence.10,27  Another concern is the occurrence of contrast-induced nephropathy. Although this problem is rare and usually reversible, it is an avoidable complication of whole-body CT scanning nonetheless.11  As a result, physicians in the trauma bay face a dilemma between the need for diagnosing all possible lesions and reducing the radiation dose.
The clinical judgment of the physician on admission was not accurate to predict severe lesions, as shown by the AUC with a lower limit of its 95% CI equal to 0.64. Intuition integrates a subjective and individual assessment of trauma history with clinical findings to build clinical judgment. This concept is described as a gut feeling,12  and its role was demonstrated for the in-hospital triage of trauma patients.13  In our study, clinical judgment missed 13% of severe lesions and had poor sensitivity (82%). Indeed, clinical judgment regarding severe injuries and the need for lifesaving interventions is fraught with difficulty. In a recent study comparing clinical judgment with an automated analysis of photoplethysmograph waveforms from pulse oximetry, pulse oximeter algorithms consistently outperformed provider assessments.28  Clinical judgment can lead to errors in decision-making; such errors have been associated with up to two thirds of preventable hemorrhagic deaths after injury.29  Interestingly, clinical variables including heart rate, systolic arterial pressure, and GCS also had poor performance to predict the need of initial imaging assessment in 400 consecutive patients with severe trauma.30  This indicates that adding physiologic variables does not improve the performance of clinical judgment. Occult lesions were mostly located in the thorax and head areas. In the existing literature, the incidence of occult lesions is variable, from 3 to 14%, depending on the type of trauma patients, the definition of lesion severity, and the type of suspected lesion.2,5,6,31  Our findings were similar to those of Shannon et al.,31  who found that the clinical suspicion of injury was poorly associated with imaging findings. In other studies, the high incidence of occult severe lesions might be explained by pulmonary contusions, which might be asymptomatic.7  Accordingly, these lesions were considered to be minor if not associated with mechanical ventilation.32 
In the case of an absolutely normal physical examination, we did not find severe lesions in our population (NPV = 100%). This indicates that a CT scan would not be indicated in this category of trauma patients. This result was additionally confirmed by a decision curve showing net benefit of a strategy based on physical examination at a threshold probability of 10%. In our study, a normal physical examination was defined as the absence of any sign of trauma, including pain or hematomas. For instance, these trauma patients belong to the category of patients admitted for a suspicion of severe trauma after a road traffic accident with one deceased patient in the same accident. Despite the absence of posttraumatic symptoms, these patients are considered at risk for having a potentially lethal lesion according to trauma history.33  We suggest that these patients could be safely monitored without total body imaging thanks to physical examination and bedside ultrasonography. As a result, this approach allows for a limitation of the radiation dose and the avoidance of an iodine agent. Conversely, the specificity of the physical examination was poor, demonstrating that a positive physical examination did not predict severe posttraumatic lesions. For instance, acute pain after limb fractures may mask another vital lesion, which was described as the concept of distracting injuries.34  Nevertheless, distracting injuries were not associated with increased occult abdominal lesions and did not challenge the relevancy of a clinical examination for the diagnosis of significant abdominal injury.35  In this study, we found few occult severe abdominal injuries (n = 2 of 354 patients).
Another role of the whole-body CT scan is to help the decision for final disposition. In our study, 11 patients (3%) estimated to be discharged to the home before CT scan were finally directed to the standard ward, and 23 patients (6%) initially directed to the standard ward required an admission to the ICU. These results also confirmed the prominent role of a whole-body CT scan for the medical management of trauma patients. According to the clinical judgment for the occurrence of a severe lesion, the attending physician also failed to accurately predict final disposition after admission to the emergency department. Innovative CT technology, based on adaptive statistical iterative reconstruction, is developed for dose reduction without loss of image quality. In trauma patients, application of iterative reconstruction allowed for a 23% reduction of the radiation dose.36 
We acknowledge several limits of our study. First, a limited proportion of patients had a strictly normal physical examination (5%). The generalization of our results on a larger population deserves additional studies focusing on patients with a normal physical examination. Second, we only conducted an observational study. However, we performed a diagnosis study following the standard international recommendation to assess the diagnosis performance of clinical judgment.37  Third, 13 patients were excluded because of the absence of a whole-body CT scan. These patients had no CT scan and could not be analyzed for the primary and secondary outcomes. However, we considered our recruitment to be exhaustive, because approximately 500 patients are admitted each year in the trauma system of the Northern French Alps, which includes 17 trauma centers.38  Fourth, we focused on the prediction of serious-to-critical posttraumatic lesions. However, minor injuries could also be responsible for late morbidity, and our study is limited to the exploration of immediate consequences of major lesions. Finally, the study population was adult trauma patients, and our results cannot be extrapolated to a pediatric population.
Conclusions
The clinical judgment of an attending physician is not sufficiently accurate to reduce the use of routine whole-body CT scanning after high-energy traumatic injuries. In patients with a physical examination completely devoid of any signs of traumatic injury, whole-body CT scanning might be avoided, but the results from this investigation deserve additional study in larger and more diverse populations of trauma patients.
Research Support
Support was provided solely from institutional and/or departmental sources.
Competing Interests
The authors declare no competing interests.
References
ATLS Subcommittee, American College of Surgeons’ Committee on Trauma, International ATLS Working Group: Advanced trauma life support (ATLS(R)), 9th edition. J Trauma Acute Care Surg 2013; 74:1363–6 [PubMed]
Salim, A, Sangthong, B, Martin, M, Brown, C, Plurad, D, Demetriades, D . Whole body imaging in blunt multisystem trauma patients without obvious signs of injury: results of a prospective study. Arch Surg 2006; 141:468–75 [Article] [PubMed]
Huber-Wagner, S, Lefering, R, Qvick, LM, Körner, M, Kay, MV, Pfeifer, KJ, Reiser, M, Mutschler, W, Kanz, KG ; Working Group on Polytrauma of the German Trauma Society: Effect of whole-body CT during trauma resuscitation on survival: A retrospective, multicentre study. Lancet 2009; 373:1455–61 [Article] [PubMed]
Caputo, ND, Stahmer, C, Lim, G, Shah, K . Whole-body computed tomographic scanning leads to better survival as opposed to selective scanning in trauma patients: A systematic review and meta-analysis. J Trauma Acute Care Surg 2014; 77:534–9 [Article] [PubMed]
Sierink, JC, Saltzherr, TP, Russchen, MJ, de Castro, SM, Beenen, LF, Schep, NW, Goslings, JC . Incidental findings on total-body CT scans in trauma patients. Injury 2014; 45:840–4 [Article] [PubMed]
Seah, MK, Murphy, CG, McDonald, S, Carrothers, A . Incidental findings on whole-body trauma computed tomography: Experience at a major trauma centre. Injury 2016; 47:691–4 [Article] [PubMed]
Deunk, J, Poels, TC, Brink, M, Dekker, HM, Kool, DR, Blickman, JG, van Vugt, AB, Edwards, MJ . The clinical outcome of occult pulmonary contusion on multidetector-row computed tomography in blunt trauma patients. J Trauma 2010; 68:387–94 [Article] [PubMed]
Huber-Wagner, S, Biberthaler, P, Häberle, S, Wierer, M, Dobritz, M, Rummeny, E, van Griensven, M, Kanz, KG, Lefering, R ; TraumaRegister DGU: Whole-body CT in haemodynamically unstable severely injured patients: A retrospective, multicentre study. PLoS One 2013; 8:e68880 [Article] [PubMed]
Van Vugt, R, Keus, F, Kool, D, Deunk, J, Edwards, M . Selective computed tomography (CT) versus routine thoracoabdominal CT for high-energy blunt-trauma patients. Cochrane Database Syst Rev 2013: CD009743
Beatty, L, Furey, E, Daniels, C, Berman, A, Tallon, JM . Radiation exposure from CT scanning in the resuscitative phase of trauma care: A level one trauma centre experience. CJEM 2015; 17:617–23 [Article] [PubMed]
Colling, KP, Irwin, ED, Byrnes, MC, Reicks, P, Dellich, WA, Reicks, K, Gipson, J, Beilman, GJ . Computed tomography scans with intravenous contrast: Low incidence of contrast-induced nephropathy in blunt trauma patients. J Trauma Acute Care Surg 2014; 77:226–30 [Article] [PubMed]
Cabrera, D, Thomas, JF, Wiswell, JL, Walston, JM, Anderson, JR, Hess, EP, Bellolio, MF . Accuracy of ‘my gut feeling:’ comparing system 1 to system 2 decision-making for acuity prediction, disposition and diagnosis in an academic emergency department. West J Emerg Med 2015; 16:653–7 [Article] [PubMed]
Newgard, CD, Nelson, MJ, Kampp, M, Saha, S, Zive, D, Schmidt, T, Daya, M, Jui, J, Wittwer, L, Warden, C, Sahni, R, Stevens, M, Gorman, K, Koenig, K, Gubler, D, Rosteck, P, Lee, J, Hedges, JR . Out-of-hospital decision making and factors influencing the regional distribution of injured patients in a trauma system. J Trauma 2011; 70:1345–53 [Article] [PubMed]
Riou, B, Thicoïpé, M, Atain-Kouadio, P, Carli, P . Comment évaluer la gravité?, Actualités en réanimation préhospitalière: le traumatisé grave. Paris, SFEM Editions 2002; 115–28.
Cotte, J, Courjon, F, Beaume, S, Prunet, B, Bordes, J, N’Guyen, C, Contargyris, C, Lacroix, G, Montcriol, A, Kaiser, E, Meaudre, E . Vittel criteria for severe trauma triage: Characteristics of over-triage. Anaesth Crit Care Pain Med 2016; 35:87–92 [Article] [PubMed]
Babaud, J, Ridereau-Zins, C, Bouhours, G, Lebigot, J, Le Gall, R, Bertrais, S, Roy, PM, Aubé, C . Benefit of the Vittel criteria to determine the need for whole body scanning in a severe trauma patient. Diagn Interv Imaging 2012; 93:371–9 [Article] [PubMed]
Bouzat, P, Broux, C, Ageron, FX, Thony, F, Arvieux, C, Tonetti, J, Gay, E, Rancurel, E, Payen, JF . Trauma network for severely injured patients. Ann Fr Anesth Reanim 2013; 32:531–4 [Article] [PubMed]
Sartorius, D, Le Manach, Y, David, JS, Rancurel, E, Smail, N, Thicoïpé, M, Wiel, E, Ricard-Hibon, A, Berthier, F, Gueugniaud, PY, Riou, B . Mechanism, Glasgow Coma Scale, age, and arterial pressure (MGAP): A new simple prehospital triage score to predict mortality in trauma patients. Crit Care Med 2010; 38:831–7 [Article] [PubMed]
Lavoie, A, Moore, L, LeSage, N, Liberman, M, Sampalis, JS . The New Injury Severity Score: a more accurate predictor of in-hospital mortality than the Injury Severity Score. J Trauma 2004; 56:1312–20 [Article] [PubMed]
Stevenson, M, Segui-Gomez, M, Lescohier, I, Di Scala, C, McDonald-Smith, G . An overview of the injury severity score and the new injury severity score. Inj Prev 2001; 7:10–3 [Article] [PubMed]
Schluter, PJ, Nathens, A, Neal, ML, Goble, S, Cameron, CM, Davey, TM, McClure, RJ . Trauma and Injury Severity Score (TRISS) coefficients 2009 revision. J Trauma 2010; 68:761–70 [Article] [PubMed]
Shackford, SR, Rogers, FB, Osler, TM, Trabulsy, ME, Clauss, DW, Vane, DW . Focused abdominal sonogram for trauma: The learning curve of nonradiologist clinicians in detecting hemoperitoneum. J Trauma 1999; 46:553–64 [Article] [PubMed]
Hyacinthe, AC, Broux, C, Francony, G, Genty, C, Bouzat, P, Jacquot, C, Albaladejo, P, Ferretti, GR, Bosson, JL, Payen, JF . Diagnostic accuracy of ultrasonography in the acute assessment of common thoracic lesions after trauma. Chest 2012; 141:1177–83 [Article] [PubMed]
DeLong, ER, DeLong, DM, Clarke-Pearson, DL . Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988; 44:837–45 [Article] [PubMed]
Moons, KG, Altman, DG, Reitsma, JB, Ioannidis, JP, Macaskill, P, Steyerberg, EW, Vickers, AJ, Ransohoff, DF, Collins, GS . Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): Explanation and elaboration. Ann Intern Med 2015; 162:W1–73 [Article] [PubMed]
Ahmadinia, K, Smucker, JB, Nash, CL, Vallier, HA . Radiation exposure has increased in trauma patients over time. J Trauma Acute Care Surg 2012; 72:410–5 [Article] [PubMed]
Hui, CM, MacGregor, JH, Tien, HC, Kortbeek, JB . Radiation dose from initial trauma assessment and resuscitation: Review of the literature. Can J Surg 2009; 52:147–52 [PubMed]
Mackenzie, CF, Gao, C, Hu, PF, Anazodo, A, Chen, H, Dinardo, T, Imle, PC, Hartsky, L, Stephens, C, Menaker, J, Fouche, Y, Murdock, K, Galvagno, S, Alcorta, R, Shackelford, S ; ONPOINT Study Group: Comparison of decision-assist and clinical judgment of experts for prediction of lifesaving interventions. Shock 2015; 43:238–43 [Article] [PubMed]
O’Reilly, D, Mahendran, K, West, A, Shirley, P, Walsh, M, Tai, N . Opportunities for improvement in the management of patients who die from haemorrhage after trauma. Br J Surg 2013; 100:749–55 [Article] [PubMed]
Peytel, E, Menegaux, F, Cluzel, P, Langeron, O, Coriat, P, Riou, B . Initial imaging assessment of severe blunt trauma. Intensive Care Med 2001; 27:1756–61 [Article] [PubMed]
Shannon, L, Peachey, T, Skipper, N, Adiotomre, E, Chopra, A, Marappan, B, Kotnis, N . Comparison of clinically suspected injuries with injuries detected at whole-body CT in suspected multi-trauma victims. Clin Radiol 2015; 70:1205–11 [Article] [PubMed]
Rodriguez, RM, Langdorf, MI, Nishijima, D, Baumann, BM, Hendey, GW, Medak, AJ, Raja, AS, Allen, IE, Mower, WR . Derivation and validation of two decision instruments for selective chest CT in blunt trauma: a multicenter prospective observational study (NEXUS Chest CT). PLoS Med 2015; 12:e1001883 [Article] [PubMed]
Sasser, SM, Hunt, RC, Faul, M, Sugerman, D, Pearson, WS, Dulski, T, Wald, MM, Jurkovich, GJ, Newgard, CD, Lerner, EB . Guidelines for field triage of injured patients: Recommendations of the National Expert Panel on Field Triage, 2011. MMWR Recomm Rep 2012; 61:1–20 [PubMed]
Cason, B, Rostas, J, Simmons, J, Frotan, MA, Brevard, SB, Gonzalez, RP . Thoracolumbar spine clearance: Clinical examination for patients with distracting injuries. J Trauma Acute Care Surg 2016; 80:125–30 [Article] [PubMed]
Rostas, J, Cason, B, Simmons, J, Frotan, MA, Brevard, SB, Gonzalez, RP . The validity of abdominal examination in blunt trauma patients with distracting injuries. J Trauma Acute Care Surg 2015; 78:1095–101 [Article] [PubMed]
Kahn, J, Grupp, U, Kaul, D, Böning, G, Lindner, T, Streitparth, F . Computed tomography in trauma patients using iterative reconstruction: Reducing radiation exposure without loss of image quality. Acta Radiol 2016; 57:362–9 [Article] [PubMed]
Ray, P, Le Manach, Y, Riou, B, Houle, TT . Statistical evaluation of a biomarker. Anesthesiology 2010; 112:1023–40 [Article] [PubMed]
Bouzat, P, Ageron, FX, Brun, J, Levrat, A, Berthet, M, Rancurel, E, Thouret, JM, Thony, F, Arvieux, C, Payen, JF ; TRENAU Group: A regional trauma system to optimize the pre-hospital triage of trauma patients. Crit Care 2015; 19:111 [Article] [PubMed]
Fig. 1.
Flowchart of the study population. AIS = abbreviated injury score; CT = computed tomography.
Flowchart of the study population. AIS = abbreviated injury score; CT = computed tomography.
Fig. 1.
Flowchart of the study population. AIS = abbreviated injury score; CT = computed tomography.
×
Fig. 2.
Empirical receiver operating characteristic (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. The area under the ROC curve (AUC ROC) was 0.70 (95% CI, 0.64 to 0.75).
Empirical receiver operating characteristic (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. The area under the ROC curve (AUC ROC) was 0.70 (95% CI, 0.64 to 0.75).
Fig. 2.
Empirical receiver operating characteristic (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. The area under the ROC curve (AUC ROC) was 0.70 (95% CI, 0.64 to 0.75).
×
Fig. 3.
Smooth receiver operating characteristics (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. This curve was generated by fitting the binomial distribution. The gray area represents the 95% CI of the ROC curve. The area under the ROC curve (AUC ROC) was 0.73 (95% CI, 0.67 to 0.79).
Smooth receiver operating characteristics (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. This curve was generated by fitting the binomial distribution. The gray area represents the 95% CI of the ROC curve. The area under the ROC curve (AUC ROC) was 0.73 (95% CI, 0.67 to 0.79).
Fig. 3.
Smooth receiver operating characteristics (ROC) curve of the clinical judgment to predict a serious-to-severe lesion on whole-body computed tomography scanning after high-energy trauma. This curve was generated by fitting the binomial distribution. The gray area represents the 95% CI of the ROC curve. The area under the ROC curve (AUC ROC) was 0.73 (95% CI, 0.67 to 0.79).
×
Fig. 4.
Net benefit curves for the clinical judgment and the physical examination. These curves determine the relationship between a chosen predicted probability threshold (x-axis) and the relative value of false-positive and false-negative results to obtain a value of net benefit (y-axis) using the test at that threshold. At the acceptable threshold of 10%, there is no net benefit for the clinical judgment as compared with the physical examination.
Net benefit curves for the clinical judgment and the physical examination. These curves determine the relationship between a chosen predicted probability threshold (x-axis) and the relative value of false-positive and false-negative results to obtain a value of net benefit (y-axis) using the test at that threshold. At the acceptable threshold of 10%, there is no net benefit for the clinical judgment as compared with the physical examination.
Fig. 4.
Net benefit curves for the clinical judgment and the physical examination. These curves determine the relationship between a chosen predicted probability threshold (x-axis) and the relative value of false-positive and false-negative results to obtain a value of net benefit (y-axis) using the test at that threshold. At the acceptable threshold of 10%, there is no net benefit for the clinical judgment as compared with the physical examination.
×
Table 1.
Patient Characteristics (n = 354 Patients)
Patient Characteristics (n = 354 Patients)×
Patient Characteristics (n = 354 Patients)
Table 1.
Patient Characteristics (n = 354 Patients)
Patient Characteristics (n = 354 Patients)×
×
Table 2.
Results of Physical Examination
Results of Physical Examination×
Results of Physical Examination
Table 2.
Results of Physical Examination
Results of Physical Examination×
×
Table 3.
Comparison between the Group of Patients with a Lesion Classified as AIS ≥ 3 and the Group of Patients with a Lesion Classified as AIS < 3
Comparison between the Group of Patients with a Lesion Classified as AIS ≥ 3 and the Group of Patients with a Lesion Classified as AIS < 3×
Comparison between the Group of Patients with a Lesion Classified as AIS ≥ 3 and the Group of Patients with a Lesion Classified as AIS < 3
Table 3.
Comparison between the Group of Patients with a Lesion Classified as AIS ≥ 3 and the Group of Patients with a Lesion Classified as AIS < 3
Comparison between the Group of Patients with a Lesion Classified as AIS ≥ 3 and the Group of Patients with a Lesion Classified as AIS < 3×
×