Diffuse axonal injury.

Object. The authors investigated the ramifications of producing diffuse axonal injury (DAI) by lateral head rotation in a rat model. Methods. Using a special injury-producing device, the rat’s head was rapidly rotated 90 ̊ in the coronal plane at an angular velocity of at least 753.13 rad/second and an angular acceleration of at least 1.806 105 rad/second2; the rotation was complete within 2.09 msec. There were no statistically significant changes in PO2, PCO2, pH, or blood pressure values at 5, 15, or 60 minutes after head rotation compared with their respective preinjury baseline values. The rats exhibited posttraumatic behavior suppression for an average of 12.6 minutes. The mortality rate was 17%. The rats that survived had diffuse subarachnoid hemorrhage around the brainstem and upper cervical cord, but no obvious brain contusion. In sections stained with silver or hematoxylin and eosin, axonal swelling and bulblike protrusions at the axonal axis were observed in the medulla oblongata, midbrain, upper cervical cord, and corpus callosum between 6 hours and 144 hours postinjury. The axonal injuries were most severe in the brainstem and were accompanied by parenchymal bleeding. The density of bulblike axonal protrusions peaked 6 hours postinjury in the medulla oblongata and 24 hours postinjury in the midbrain. Conclusions. Rapid lateral head rotation can produce DAI characterized by severe damage to the rat brainstem. We believe that the core of the methods and conclusions are fundamentally unsound, particularly with regard to the biomechanics of injury. It is well established that rotational acceleration is the predominant mechanism that causes tissue deformations within the brain in humans leading to DAI. Mass effects of the large brain of humans play an essential role in this process, in which the deforming brain can be thought of as literally pulling itself apart during the acceleration–deceleration period. To produce equivalent tissue strains in animals with smaller brain masses, the accelerations must be much greater. The authors failed to mention that injury-producing accelerations across different sized brains has been extensively examined2,4 and is summarized by the standard known as the Holbourn scaling relationship: ̈s = (Mref/ Ms)̈ref, where Mref and Ms denote the reference and scaled brain masses, respectively, and ̈ref and ̈s denote the reference and scaled rotational accelerations, respectively. This scaling relationship has previously been demonstrated in animal models of rotational acceleration. To replicate the tissue strains leading to DAI in humans, accelerations had to be increased 500% for a 140-g brain of a baboon1 and 630% for a 90-g brain of a pig.3,5 By extension, the inertial forces necessary to produce equivalent tissue strains in the less-than-2-g brain of a rat would be unachievable, and the accelerations would have to approach 8000% of the level required for DAI to occur in the brain of a human. The applied forces and accelerations reported by He and colleagues were, however, at least 10 times less than anticipated by the scaling relationship. In addition, insufficient details were provided to determine the accuracy of the accelerations reported. Finally, it is important to note that the signature anatomical characteristic of DAI in humans is axonal damage distributed throughout the large white matter domains. But the lissencephalic brain of a rodent has a remarkable paucity of white matter, and as such, it could be argued that clinically relevant DAI cannot be produced in rodents by any means. Note that if one’s goal is simply to study traumatic axonal injury, this pathology has already been convincingly demonstrated in a number of established rodent brain impact models. DAVID F. MEANEY, PH.D. SUSAN S. MARGULIES, PH.D. DOUGLAS H. SMITH, M.D. University of Pennsylvania Philadelphia, Pennsylvania