One of the five 2014 American College of Surgeons’ Choosing Wisely recommendations is to avoid routing whole-body CT imaging of trauma patents, also known as the ‘pan-CT’. Until now, no validated decision instrument existed to help guide clinicians decide whether to obtain a chest CT in the setting of blunt trauma. This month, Dr. Robert Rodriguez and the multi-institutional NEXUS Chest CT research team published a paper describing the derivation and validation of 2 decision instruments in PLOS Medicine [1].
Methods
Prospective multicenter observational study at 8 urban level-1 trauma centers enrolling all patients who sustained blunt trauma and underwent a chest xray (CXR) and/or chest CT. Injury outcome measures were divided into clinically “major” and “minor” injuries.
[su_spoiler title=”Major Injuries” style=”fancy” icon=”caret”]- Aortic or great vessel injury (all considered major)
- Ruptured diaphragm (all considered major)
- Pneumothorax: received evacuation procedure (chest tube or other procedure)
- Hemothorax: received drainage procedure (chest tube or other procedure)
- Sternal fracture: received surgical intervention
- Multiple rib fracture: received surgical intervention or epidural nerve block
- Pulmonary contusion: received mechanical ventilation (including non-invasive ventilation) primarily for respiratory failure within 24 h for management
- Thoracic spine fracture: received surgical intervention
- Scapular fracture: received surgical intervention
- Mediastinal or pericardial hematoma: received drainage procedure
- Esophageal injury: received surgical intervention
- Tracheal or bronchial injury: received surgical intervention
- Pneumothorax: no evacuation procedure but observed as inpatient >24 h significance
- Hemothorax: no drainage procedure but observed as inpatient for >24 h
- Sternal fracture: no surgical intervention
- Multiple rib fracture: no surgical intervention or epidural nerve block
- Pulmonary contusion or laceration: no mechanical ventilation but observed >24 h
- Thoracic spine fracture: no surgical intervention
- Scapular fracture: no surgical intervention
- Mediastinal or pericardial hematoma: no surgical intervention
- Esophageal injury: no surgical intervention
- Tracheal or bronchial injury: no surgical intervention
Using 14 clinical criteria, the “Chest CT-Major” and “Chest CT-All” decision instruments were developed to screen for (major only) vs (major + minor) injuries, respectively, in the hopes of decreasing chest CT utilization.
[su_spoiler title=”Inclusion criteria” style=”fancy” icon=”caret”]- Age >14 years
- Presenting to the ED for blunt trauma occurring within 6 hours of arrival
- Having chest imaging (CXR or CT) in the ED
Results
Two validated decision instruments were described, Chest CT-Major and Chest CT-All, with 6 and 7 criteria, respectively.
NEXUS Chest Decision Instrument (Chest CT-Major): To help rule out clinically MAJOR injuries |
---|
1. Abnormal chest x-ray
|
2. Distracting injury (same as in NEXUS c-spine definition)
|
3. Chest wall tenderness*
|
4. Sternal tenderness* |
5. Thoracic spine tenderness* |
6. Scapular tenderness* |
* One might combine these 4 criteria into 1 criterion of any thoracic tenderness (imagine a burrito wrap rule).
NEXUS Chest Decision Instrument (Chest CT-All): To help rule out clinically MAJOR and MINOR injuries |
---|
All 6 criteria from Chest CT-Major above plus… |
7. Rapid deceleration
|
Validation Phase Results
A total of 11,477 patients were enrolled with a median age of 46 y (6,002 derivation phase, 5,475 validation phase).
Injuries | Sensitivity | Specificity | NPV | Neg LR |
---|---|---|---|---|
MAJOR INJURIES | ||||
• Chest CT-Major | 99.2% (95.4-100%) | 31.7% (29.9-33.5%) | 99.9% (99.3-100%) | 0.03 (0.04-0.19) |
• Chest CT-All | 99.2% (95.4-100%) | 20.8% (19.2-22.4%) | 99.8% (98.9-100%) | 0.04 (0.06-0.29) |
MAJOR and MINOR INJURIES | ||||
• Chest CT-Major | 90.7% (88.3-92.8%) | 37.9% (35.8-40.1%) | 91.8% (89.7-93.6%) | 0.24 (0.19-0.31) |
• Chest CT-All | 95.4% (93.6-96.9%) | 25.5% (23.5-27.5%) | 93.9% (91.5-95.8%) | 0.18 (0.13-0.25) |
Values listed with 95% confidence intervals in parentheses
LR = Likelihood ratio
What Did the Decision Instruments Miss?
Chest CT-Major
- Major Injuries: Missed 1 of 120 patients (80 year old man who fell 7 stairs and a subarachnoid hemorrhage, who also sustained an isolated pneumothorax requiring a chest tube)
- Minor Injuries: Missed 64 of 691 patients (included rib fractures, sternal fracture, pulmonary contusion, thoracic spine fracture, scapular fracture, pneumothorax, >1 minor injury)
Chest CT-All
- Major Injuries: Also missed same patient above (1 of 120 patients): 80 year old man who fell 7 stairs and a subarachnoid hemorrhage, who also sustained an isolated pneumothorax requiring a chest tube
- Minor Injuries: Missed 31 of 691 patients (included rib fractures, sternal fracture, pulmonary contusion, thoracic spine fracture, scapular fracture, pneumothorax, >1 minor injury)
Conclusion
The NEXUS Chest CT research team should be applauded for rigorously addressing a common issue of Chest CT utilization for blunt trauma patients in the era of pan-CT’ing. Developing a 2-tier decision instrument, the Chest CT-Major and Chest CT-All, helps institutions address whether they wish to use a more risk-tolerant or risk-averse approach, respectively. More specifically, do they want to detect clinically “major” injuries only or also clinically minor injuries that do not require acute surgical interventions?
This study provides a validated tool to safely supplement clinician judgment in the decision to obtain chest CT imaging for blunt trauma patients.
The decision instruments:
- Chest CT-Major: Abnormal chest -ray + distracting injury + thoracic tenderness
- Chest CT-All: Above + deceleration mechanism
Bottom lines:
- Both decision instruments have >99% sensitivity to detect clinically major injuries
- The Chest CT-All instrument has a >95% sensitivity to detect major AND minor injuries.
- Appropriate use of these decision tools may safely allow clinicians to forego 25-37% of chest CTs.
[su_box title=”Q&A with the first author, Dr. Rodriguez” style=”glass” box_color=”#1a80b6″ title_color=”#ffffff”]
Robert Rodriguez, MD
Professor of Clinical Emergency Medicine
Residency Research Director
University of California, San Francisco
Why do we need a decision instrument on ordering chest CT’s in blunt trauma?
Even though the prevalence of significant thoracic injury has remained constant, the use of chest CT (especially as part of a pan-scan) in adult blunt trauma evaluation has increased substantially over the past decade. The problems with this increased CT use include higher costs (charges for chest CT are approximately $3,600 per study), radiation exposure (approximately 5-7 millisieverts – which is equivalent to about 350 CXRs), and increased time in the ED. Notably, this radiation exposure translates to real risk of cancer induction, especially in younger trauma patients. The vast majority of these CT studies fail to show injuries that change patient management. We sought to derive instruments to guide safe, selective use of chest CT in blunt trauma.
Why 2 decision instruments (DI) instead of 1?
We created 2 DIs because we recognized that clinicians (especially trauma surgeons vs EM) have different viewpoints on the need to diagnose minor injuries. CT-All picks up over 95% of these minor injuries with a specificity of approximately 25%. CT-Major picks up over 90% of minor injuries with greater specificity (>37%), allowing clinicians to forego imaging in a greater percentage of patients.
Any pitfalls that you invision in the practical implementation of these DI’s?
Remember that these DIs are meant to be a one-way directive tool. Similar to other decision instruments like the NEXUS cervical spine tool, they should only be used to rule-out the need for chest CT and never to indicate that you should get CT.
[/su_box]
Reference
- Rodriguez RM, Langdorf MI, Nishijima D, et al. 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(10): e1001883. PMID: 26440607 [Open Access PDF]