22.2 Diffuse Axonal Injury and Cerebral Contusions (03.I.1–2)

Blueprint items 03.I.1–2 sit next to extra-axial hematoma on the outline but behave differently at the bedside. Diffuse axonal injury (DAI) is a tissue diagnosis of shear that the CT scanner often undercalls. Cerebral contusions are bruises you can watch grow on the 24- and 72-hour scans. Both live under the same secondary-injury rules: do not let the patient become hypotensive or hypoxic, do not give steroids, and treat intracranial pressure when the Brain Trauma Foundation (BTF) 4th-edition numbers say to treat.

Independent OpenExamPrep material for this ABIM Neurocritical Care topic cites BTF 4th edition (Carney and colleagues, Neurosurgery 2017), the surgical TBI guidelines (Bullock and colleagues, Neurosurgery 2006), and CRASH (Lancet 2004/2005). It is not a BTF or ABIM publication.

Diffuse axonal injury: grades and imaging

High-energy angular acceleration shears axons at interfaces of different density. The clinical signature is immediate coma without a mass lesion large enough to explain the exam. Patients who talk and then deteriorate usually have an expanding extra-axial or contusional clot, not isolated DAI.

The Adams–Gennarelli pathologic grades remain the exam scaffold:

Grade is assigned by the highest structure involved. Gray–white junction plus callosum without brainstem is grade 2, not grade 3. Isolated callosal blood after a linear impact can also be a contusion or a vascular injury; DAI is a pattern, not a single voxel.

CT shows DAI only when microhemorrhages are large enough: small hyperdensities at the gray–white junction, in the corpus callosum, or in the dorsolateral midbrain, sometimes with tiny amounts of intraventricular blood. A normal CT does not exclude DAI. Gradient-echo (GRE) and especially susceptibility-weighted imaging (SWI) detect far more hemorrhagic shear lesions. Diffusion-weighted imaging and diffusion-tensor imaging pick up nonhemorrhagic axonal injury. Do not tell families that “the CT is clean, so the coma is metabolic” until an MRI has been considered and confounders (sedation, seizure, extra-axial clot) are addressed.

DAI itself is not a lesion you evacuate. Care is supportive: airway, sodium, fever, seizure control when indicated, ICP monitoring when the BTF indications are met, and early rehabilitation. Brainstem (grade 3) lesions and a persistently low GCS after confounders are cleared predict a worse functional outcome, but early withdrawal based on a single SWI collage is how self-fulfilling prophecy enters TBI care.

Cerebral contusions: blossom and when to operate

Contusions are hemorrhagic bruises of cortex and subcortical white matter. They favor the orbitofrontal gyri and temporal poles, where brain strikes the sphenoid and cribriform ridges (coup–contrecoup). A small admission hematoma is not a final volume. Blossoming — expansion of hemorrhage and edema — is expected over 24–72 hours. That is why a GCS 14 patient with a 12 mL temporal contusion at 02:00 can be herniating at 18:00.

TBI-associated coagulopathy, residual warfarin or a factor Xa inhibitor, antiplatelet agents, and thrombocytopenia all increase expansion. Reverse what you can as you would for spontaneous ICH (Chapter 4), obtain an early stability CT, and keep platelet count and fibrinogen in a hemostatic range if an operation is coming. Hypertonic saline or mannitol treats mass effect and ICP, not the bruise itself.

Surgical thresholds from the 2006 TBI surgical guidelines (still the numbers boards use) are conceptual, not a substitute for the exam in front of you:

• Evacuate a parenchymal lesion that causes progressive deterioration referable to that lesion, medically refractory intracranial hypertension, or clear CT mass effect.
• A patient with GCS 6–8, a frontal or temporal contusion larger than 20 mL, midline shift of 5 mm or more, and/or cisternal compression is a surgical candidate.
• Any parenchymal hematoma larger than 50 mL is generally evacuated in a salvageable patient.
• Posterior-fossa contusions have a lower volume threshold because the brainstem and fourth ventricle sit in a small box.

Bifrontal decompressive craniectomy for diffuse swelling is a different operation. DECRA (NEJM 2011) showed worse outcome with early DC for diffuse injury. RESCUEicp (NEJM 2016) used DC as last-tier ICP rescue and reduced mortality at the cost of more vegetative and severely disabled survivors. Do not confuse those trials with evacuation of a blossomed temporal contusion.

GCS, BTF intracranial pressure, and secondary injury

The Glasgow Coma Scale still defines TBI severity for protocols:

Record the motor score and whether the patient is intubated or eye-swollen. A GCS of 3T is not the same conversation as a GCS of 3 with fixed pupils from herniation.

BTF 4th-edition numbers that belong on this blueprint item:

ICP monitoring is recommended in salvageable severe TBI (GCS 3–8) with an abnormal CT, and it may be used when the CT is normal if two or more of age over 40 years, unilateral or bilateral motor posturing, and systolic pressure under 90 mm Hg are present. Hyperosmolar therapy is mannitol or hypertonic saline for ICP crisis; neither is a standing infusion for every contusion. Prophylactic hypothermia is not a BTF recommendation for improving outcome.

Secondary injury is the part you can still prevent after the crash. Hypotension drops cerebral perfusion pressure through a damaged autoregulatory curve. Hypoxia adds ischemic axonal injury on top of mechanical shear. In the field and in the CT scanner, those two insults matter as much as the Marshall CT grade. The same logic forbids prolonged uncontrolled fever and hyponatremia, but hypotension and hypoxia are the pair the outline expects you to name.

Early post-traumatic seizures in severe TBI are reduced by 7 days of an antiseizure medication (the evidence was phenytoin; levetiracetam is the usual substitute). Prophylaxis does not prevent late epilepsy and is not automatically continued for weeks in blunt injury. Penetrating injury is the exception discussed in the next section.

CRASH: corticosteroids harm TBI

The CRASH trial (Edwards, Roberts, and colleagues, Lancet 2004 for 14-day death; 2005 for 6-month outcome) randomized 10,008 adults with GCS 14 or less within 8 hours of injury to a 48-hour high-dose methylprednisolone infusion (2 g load, then 0.4 g per hour) or placebo. Death at 2 weeks was 21.1% with steroid versus 17.9% with placebo (relative risk 1.18). Death at 6 months was 25.7% versus 22.3%. Harm was not confined to a single subgroup. That result retired corticosteroids for post-traumatic cerebral edema.

Do not import NASCIS megadose methylprednisolone from outdated spinal-cord protocols into head injury. Do not give dexamethasone “for CT swelling” after blunt TBI. Steroids still have roles in tumor edema and some inflammatory diseases; they are the wrong drug for traumatic contusions and DAI.

A worked 48-hour timeline

A helmetless motorcyclist is GCS 6 at the scene, systolic pressure 78 mm Hg, oxygen saturation 84%. Those two numbers already dominate prognosis. After resuscitation he is GCS 7T. Admission CT shows punctate gray–white hemorrhages and a 15 mL left temporal contusion without surgical mass effect. You place an ICP monitor because this is severe TBI with an abnormal CT, keep ICP at or below the treat-above-22 rule, hold CPP near 60–70 mm Hg, reverse an INR of 1.9, and obtain a 6-hour stability CT. MRI the next day shows additional SWI lesions in the splenium and dorsolateral midbrain (DAI grade 3) plus blossoming of the temporal contusion to 28 mL with 6 mm of shift. The ICP is now 28 mm Hg on osmotherapy. That is an operative contusion, not a steroid indication. CRASH stays on the shelf.