This is part of a recurring series examining landmark articles in Emergency Medicine, in the style of ALiEM’s 52 Articles.
Haydel et al., NEJM 2000. Indications for Computed Tomography in Patients with Minor Head Injury.
1. This two-phase prospective trial sought to “derive and validate a set of clinical criteria that could be used to identify patients with minor head trauma in whom a head CT (HCT) could be foregone.”
2. “Minor head trauma” was defined as: loss of consciousness in a patient presenting with GCS 15, intact cranial nerves, and normal strength and sensation in all extremities.
3. The presence of at least one of the following: Headache, vomiting, age >60 years old, drug/alcohol intoxication, short-term memory deficit, evidence of trauma above the clavicles, or post-traumatic seizure, was found to be 100% sensitive for subsequently positive HCT (presence of cerebral contusion, subdural or epidural hematoma, depressed skull fracture, or subarachnoid or intraparenchymal hemorrhage).
Since the advent of CT in the 1970s, many studies have sought to determine the most efficient and best use of this powerful tool for minor head trauma. Early studies of HCT on patients with GCS 13-15 showed 17-20% incidence of positive findings and recommended HCT for all, but subsequent studies have shown positive HCT findings in GCS 15 patients to be as low as 6-9%. Historically in the USA, ~66% of annual head trauma is “minor,” ~10% of this group will have positive HCTs, and ~1% will require neurosurgery. This begs the question: is there a subset of minor head trauma patients for whom HCT adds little or no value? Could we save time, money, and risk of complications for all participants if we could identify these patients based on clinical presentation?
The study occurred at a large urban Level 1 trauma center from 1997 to 1999, and was split into two phases. Phase 1 enrolled 520 consecutive patients with minor head trauma who were over 3 years old, presenting <24 hours since the trauma, getting a HCT already, and had “minor head trauma,” defined as a loss of consciousness in a patient with a GCS of 15, normal cranial nerves, and normal strength and sensation in all extremities. No patients who did not lose consciousness or who declined HCT were included, and the CT scan was considered “positive” if it showed a cerebral contusion, subdural or epidural hematoma, subarachnoid, depressed skull fracture, or intraparenchymal hemorrhage.
Seven clinical variables from Phase 1 were found to correlate with positive HCT findings, and their predictive value was prospectively assessed in Phase 2, which enrolled 909 patients with the same inclusion criteria as Phase 1. Importantly, the Phase 2 patients still received normal trauma care, with HCTs ordered at the discretion of the providers whether or not any of the Phase 1 variables were present; the researchers were simply validating the Phase 1 criteria again.
● Phase 1 (520 patients)
○ 36 patients (6.9%) had positive HCTs
○ Predictive Variables: Headache, vomiting, age >60, post-traumatic seizure, short-term memory deficit, drug/alcohol intoxication, and evidence of trauma above the clavicles. Of note, "short-term memory deficit + drug/alcohol intoxication + evidence of trauma above the clavicles” were the strongest predictors: If they had scanned ONLY the patients who had all 3 of these combined, the number of scans would have decreased 31%, and sensitivity would still have been 94%
● Phase 2 (909 patients)
○ 57 patients (6.3%) had positive HCTs
○ All patients with positive findings had at least 1 of the seven Phase 1 variables (sensitivity 100% [95% CI 95-100], specificity 25%, NPV 100%)
○ All 212 patients (23.3%) with ZERO Phase 1 variables had negative HCTs (20-26%, 95% CI; NPV 100%)
● Of 93 patients from both Phases who had positive HCTs (total % → % obs, % surgery):
○ Cerebral Contusion: 47% → 100%, 0%
○ SDH: 38% → 94%, 6%
○ SAH: 14% → 100%, 0%
○ Epidural: 10% → 78%, 22%
○ Depressed Skull Fracture: 11% → 80%, 20%
● Historically, lots of head trauma (66%) is minor, with few (~10%) of these patients having positive HCTs, and even fewer (~1%) requiring neurosurgery. So there is fat to trim.
● In this study alone, if the criteria derived in Phase 1 had been applied to the Phase 2 patients (ie, “do not scan if zero variables are present”), the number of HCTs would have decreased by 22% with no additional missed findings. This certainly has broader implications when considering the trajectory of our healthcare spending as a percentage of GDP…. which is essentially like the SpaceX Falcon 9 rocket, which has 9 liquid oxygen engines and can generate 1.5 million pounds of thrust at sea level.
○ The study quotes one estimate that a 10% reduction in the number of HCTs in minor head trauma patients would save >$20,000,000 per year
● This was the first study to derive predictors that were 100% sensitive for positive HCTs, but it is important to note a few caveats:
○ The 95% Confidence Interval of “95-100%” for the sensitivity of their variables indicates that when generalized to the great big world, there is a chance that these predictors will no longer be perfect
○ “Positive CT findings” obviously does not necessarily equate with morbidity or mortality. Whether we should try to find CT findings or those lesions that require intervention is a broader, and more controversial, topic. This study simply sought to attain 100% sensitivity for HCT findings with a high degree of confidence. They provide no information on the clinical significance of these findings as far as mortality or functional outcome.
● Bottom line: Holding the clinical significance of lesions and a discussion on the sensitivity of HCT aside, discharging patients home after minor head trauma with a negative HCT and a normal neurological exam is generally supported in the literature. If the variables these investigators are promoting help to identify patients who are exceedingly unlikely to benefit from receiving a HCT, and it saves everyone money, time and some risk of complications, then let’s consider it the next time someone comes in after being struck in the face by a feather.
Level of Evidence:
Based on the ACEP grading system for therapeutic questions this study was graded a level I.
Resident Reviewer: Dr. Anatoly Kazakin
Faculty Reviewer: Dr. Matt Siket
Haydel et al. Indications for Computed Tomography in Patients with Minor Head Injury. N Engl J Med. 2000 July;343(2);100-105.