Detected and overlooked cervical spine injury among comatose trauma patients: from the Pennsylvania trauma outcomes study

Full access

Object

A rule for identifying patients with traumatic brain injury (TBI) who are at exceptionally low risk of cervical spine injury might be clinically useful. The goal in this study was to research case records to determine whether such a rule would be practicable.

Methods

The Pennsylvania Trauma Outcomes Study database was used to find patients with TBI in whom Glasgow Coma Scale (GCS) scores at admission were 8 or less. Cases of cervical spine injury were identified from diagnostic codes. Associations between these injuries and a variety of clinical variables were tested using chi-square analysis. The probability of a cervical spine injury in these patients was modeled by logistic regression. Decision tree models were constructed and statistical determinants of overlooked cervical spine injury were examined.

The prevalence of cervical spine injury among 41,142 cases of TBI was 8%. The mechanism of injury, presence of thoracolumbosacral (TLS) spinal, limb and/or facial fracture, patient age, GCS score at admission, and the presence of hypotension were all factors associated with cervical spine injury. These were incorporated into the following logistic regression model: probability of cervical spine injury = 1/(1 + exp[4.030 − 0.417*mechanism − 0.264*age − 0.678*TLS − 0.299*limb + 0.218*GCS score − 0.231*hypotension − 0.157*facial]).

This model yielded a rule for clearance of 28% of cases, with a negative predictive value (NPV) of 97%. Decision tree analysis yielded an easily stated rule for clearance of 24% of cases, with an NPV of 98.2%. The prevalence of overlooked cervical spine injury among all patients with severe TBI was 0.3%; the prevalence of overlooked cervical spine injury among patients in whom it was later diagnosed was 3.9%. Overlooked cervical spine injury was less common among patients with associated TLS fractures (odds ratio 0.453, 95% confidence interval 0.245–0.837).

Conclusions

No acceptable rule for relaxation of vigilance in the search for cervical spine injury among patients with severe TBI has been identified. Levels of provider vigilance and consequent rates of overlooked cervical spine injury can be affected by environmental cues and presumably by other behavioral and organizational factors.

Abbreviations used in this paper:CI = confidence interval; GCS = Glasgow Coma Scale; ICD.9.CM = International Classification of Diseases, Ninth Revision; NPV = negative predictive value; OR = odds ratio; PTOS = Pennsylvania Trauma Outcomes Study; ROC = receiver operating characteristic; TBI = traumatic brain injury; TLS = thoracolumbosacral.

Object

A rule for identifying patients with traumatic brain injury (TBI) who are at exceptionally low risk of cervical spine injury might be clinically useful. The goal in this study was to research case records to determine whether such a rule would be practicable.

Methods

The Pennsylvania Trauma Outcomes Study database was used to find patients with TBI in whom Glasgow Coma Scale (GCS) scores at admission were 8 or less. Cases of cervical spine injury were identified from diagnostic codes. Associations between these injuries and a variety of clinical variables were tested using chi-square analysis. The probability of a cervical spine injury in these patients was modeled by logistic regression. Decision tree models were constructed and statistical determinants of overlooked cervical spine injury were examined.

The prevalence of cervical spine injury among 41,142 cases of TBI was 8%. The mechanism of injury, presence of thoracolumbosacral (TLS) spinal, limb and/or facial fracture, patient age, GCS score at admission, and the presence of hypotension were all factors associated with cervical spine injury. These were incorporated into the following logistic regression model: probability of cervical spine injury = 1/(1 + exp[4.030 − 0.417*mechanism − 0.264*age − 0.678*TLS − 0.299*limb + 0.218*GCS score − 0.231*hypotension − 0.157*facial]).

This model yielded a rule for clearance of 28% of cases, with a negative predictive value (NPV) of 97%. Decision tree analysis yielded an easily stated rule for clearance of 24% of cases, with an NPV of 98.2%. The prevalence of overlooked cervical spine injury among all patients with severe TBI was 0.3%; the prevalence of overlooked cervical spine injury among patients in whom it was later diagnosed was 3.9%. Overlooked cervical spine injury was less common among patients with associated TLS fractures (odds ratio 0.453, 95% confidence interval 0.245–0.837).

Conclusions

No acceptable rule for relaxation of vigilance in the search for cervical spine injury among patients with severe TBI has been identified. Levels of provider vigilance and consequent rates of overlooked cervical spine injury can be affected by environmental cues and presumably by other behavioral and organizational factors.

Abbreviations used in this paper:CI = confidence interval; GCS = Glasgow Coma Scale; ICD.9.CM = International Classification of Diseases, Ninth Revision; NPV = negative predictive value; OR = odds ratio; PTOS = Pennsylvania Trauma Outcomes Study; ROC = receiver operating characteristic; TBI = traumatic brain injury; TLS = thoracolumbosacral.

If the inline PDF is not rendering correctly, you can download the PDF file here.

Article Information

Contributor Notes

Address reprint requests to: Joseph H. Piatt Jr., M.D., Section of Neurosurgery, St. Christopher's Hospital for Children, Erie Avenue at Front Street, Philadelphia, Pennsylvania 19134-1095. email: joseph.piatt@tenethealth.com.
Headings
Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 406 329 10
PDF Downloads 157 46 2
EPUB Downloads 0 0 0
PubMed
Google Scholar