J Neurol Neurosurg Psychiatry 2001;70:127–129 127 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.70.1.127 on 1 January 2001. Downloaded from

SHORT REPORT

Isolated medulla oblongata function after severe traumatic injury

E F M Wijdicks, J L D Atkinson, H Okazaki

Abstract reflexes were absent. Brain CT documented a The objective was to report the first large left epidural supratentorial haematoma in pathologically confirmed case of partly addition to subarachnoid in the basal functionally preserved medulla oblongata cisterns and and intraparen- in a patient with catastrophic traumatic chymal haemorrhage in both cerebellar pedun- brain injury. cles and . Laboratory tests including A patient is described with epidural serum alcohol concentration and urinary toxi- haematoma with normal and cology screen were unremarkable. He was and a retained coughing taken to surgery as an emergency for left crani- reflex brought on only by catheter suc- otomy and removal of epidural haematoma. tioning of the carina. Multiple contusions in the thalami and pons were found but POSTOPERATIVE COURSE the medulla oblongata was spared at After evacuation of the epidural haematoma he necropsy. remained comatose. Repeat brain CT showed a In conclusion, medulla oblongata function new large epidural haematoma in the posterior may persist despite rostrocaudal deterio- fossa with a stellate haematoma in the pons and ration. This comatose state (“medulla newly imaged contusions in both thalami. man”) closely mimics brain death. Neurological examination by one of the ( 2001;70:127–129) J Neurol Neurosurg Psychiatry attending physicians showed lack of conscious- Keywords: brain death; head injury; apnoea test; ness, virtually absent brain stem reflexes (with outcome specific attention to vertical eye movement and blinking), and no motor response to pain. A cough response was momentarily noted by one Department of Any neurological catastrophe may result in of the attending nurses but no recognisible Neurology, Neurologic- brain death and its transition is typically deter- response was seen with movement of the

Neurosurgical mined by clinical neurological examination endotracheal tube. A brief episode of hypoten- http://jnnp.bmj.com/ Intensive Care Unit, alone.1 In many countries of the world, sion with a systolic blood pressure of 80 mm Saint Marys Hospital, confirmatory tests are optional in adults but Hg was noted. 1216 Second Street SW remain obligatory in children and the new- Phenylephrine was administered to ensure 55905 Rochester, 2 Minnisota, USA born. Therefore, clinical examination should adequate . No diabetes insipidus was E F M Wijdicks be unequivocal and precise. It should include seen. An apnoea test was performed. Within testing of respiratory drive after maximal minutes after disconnection a spontaneous Department of stimulation of the respiratory centres with regular respiratory eVort with tidal volumes Neurosurgery hypercarbia using apnoeic oxygenation tech- between 300–500 ml was recorded at a Pco of on October 2, 2021 by guest. Protected copyright. J L D Atkinson 2 niques. Only a few cases have been described in 34 mm Hg. Re-examination of the patient (by Department of which spontaneous breathing occurs during EFMW) showed absent brain stem reflexes but Anatomic Pathology apnoea testing.3 a notable discrepancy between upper and lower and Neuropathology, We report on a patient with an isolated, tracheal stimulation and response. Up and Mayo Clinic and barely detectable, cough response and sponta- down movement of the endotracheal tube pro- Foundation, 200 First neous breathing of several days’ duration until duced no response. Deep catheter suctioning Street SW, Rochester, MN 55905, Minnesota, withdrawal of support. This comatose state of the carina or mainstem bronchus elicited a USA closely mimics brain death. reproducible, but faint, cough response. This H Okazaki clinical condition persisted for 48 hours. Two Case description additional apnoea tests with roughly 12 hour Correspondence to: This patient was extracted from a car after a intervals produced similar results. The family Professor EFMWijdicks, Department of Neurology, motor vehicle accident. During transport to the decided to withdraw support and the patient Mayo Clinic, 200 First Street hospital he remained unconscious, possibly was placed on a T piece with 4 l oxygen. For an SW, Rochester, Minnesota had an apnoeic episode, developed bradycar- additional 8 hours the patient had normal res- 55905, 507–284–2511, USA [email protected] dia, and was intubated. On arrival, Glasgow piratory drive but was intermittently tachyp- coma sum score was 3, the left pupil was noeic with rates varying from 20–30/minute. Received 3 February 2000 dilated (diameter of 8 mm), fixed to light, the Oxygenation measured by pulse oxymetry and in revised form 20 May 2000 right pupil was normal size (diameter of 4 mm) remained normal. Blood pressures stabilised at Accepted 27 June 2000 with a normal light response, but oculocephalic 90–100 mm Hg systolic pressure without the

www.jnnp.com 128 Wijdicks, Atkinson, Okazaki J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.70.1.127 on 1 January 2001. Downloaded from

Serial macroscopical brain slices show multiple trauma induced haemorrhagic lesions in the , thalamus, and pons but with sparing of the medulla oblongata.

use of pharmacological support. Respiratory showed no intrinsic lesions save for minimal arrest within minutes occurred after extubation perivascular haemorrhages along the fourth after withdrawal of support. ventricular surface. Additionally, there was herniation of the left NEUROPATHOLOGY lateral hemispheric brain tissue through the At necropsy, there were multiple closed skull large craniotomy defect, up to 1.5 cm in height, fractures involving the bilateral parieto- accompanied by evidence of recent venous- occipital sutures, the , the type infarction, most notable at the marginal petrous portion of the temporal bones, and the zones of the herniated tissue which had been roof of the right . Superficial contusions compressed against the bony edge of the crani- were noted in the ventral surfaces of the frontal otomy defect. There were no residual epidural

lobes; the anterior, ventral, and lateral aspects or subdural haemorrhages in this area. http://jnnp.bmj.com/ of the temporal lobes; splenium; and the However, there was a large (16×9×2 cm) ventral surfaces of the cerebellar hemispheres. fresh epidural haemorrhage making a mild The lesions central to the clinical presenta- indentation of both the occipital and posterior, tion of this case were found in the rostral and particularly lateral temporal, aspects of the cer- middle brain stem constituting deep contusive ebrum. lesions (figure). There was mild shift of the cerebral midline Most rostrally, haemorrhages were found in structures from left to right. A mild bowing of

the hypothalami and paramedian thalami, the paramedian above the cor- on October 2, 2021 by guest. Protected copyright. bilaterally and in the left lateral thalamus and pus callosum was seen and limited bilateral . In the , smaller uncal herniation without any significant mesial haemorrhages were dispersed across the cross temporal lobe herniation, or rostral or caudal sections, whereas at the upper and midpontine compression of the midbrain or upper pons. levels, they coalesced to form a large central Histological examination showed no well confluent area measuring 3×3cmatits developed areas of axonal disruption except in maximal dimension. There was a limited the vicinity of contusion or postoperative areas rupture into the fourth ventricle. In the lower (APP and neurofilament immunostains). Eosi- pons, small haemorrhages were largely limited nophilic degeneration among scattered neocor- to the midline area. tical and hippocampal was seen. Also present were haemorrhages in the inte- rior of the dorsal vermis, the and Discussion dentate nuclei of both cerebellar hemispheres, Our case is unique because we document the separate from the mild cortical lesions de- presence of retained coughing with a vigorous scribed previously in the ventral surface of the stimulus, normal breathing drive and oxygena- cerebellum. tion, and normal blood pressure despite the The medulla showed only a mild extrinsic loss of other reflexes. Typically, compression of its posterolateral aspects at its medullary destruction loss is a final event in middle level by a mild tonsillar herniation but rostrocaudal herniation syndromes; however,

www.jnnp.com Isolated medulla oblongata function after severe traumatic brain injury 129 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.70.1.127 on 1 January 2001. Downloaded from

its function loss may take several additional patient rather than lead to the usual expected hours to complete. In our case, progression demise of haemodynamic systems and cardiac from the initial impact did not involve medul- arrest. lary structures. Three categories of pitfalls in the diagnosis A few cases of spontaneous breathing during of brain death have been identified by Pallis the apnoea test in patients who have fulfilled all and Harley.6 These are failure to meet precon- other clinical criteria of brain death have been ditions and exclusions, diYculty with interpret- published. Notably, Ropper et al reported four ation of signs, and failure to elicit them appro- patients with sparing of only the medulla priately. In our case, these pitfalls were clearly oblongata “indicated by normal respirations” apparent during evaluation in the intensive and these patients “survived for several days to care unit. In addition to these pitfalls, the pres- 3 weeks with stable unsupported blood pres- ence of a locked in syndrome needed to be sures.” Details on cough responses and excluded through absence of responsiveness, necropsy confirmation of sparing of the vertical eye movements, and blinks on com- medulla oblongata were not described.3 Only mand.6 The declaration of brain death requires one instance of breathing eVorts while fulfilling academic precision and failure to do so may other criteria for brain death was found in our lead to a false diagnosis of brain death.1 Clues series of 145 apnoea procedures but apnoea, in to retained brain stem function are normal both cases, was documented hours later.4 blood pressure and faint coughing during suc- Other cases of breathing during the apnoea test tioning while pontomesencephalic reflexes are have been explained by spinal reflex activity absent. resulting in intercostal retraction and shoulder We speculate that this patient may represent movements but with minimal ineVective air some of the recently reported cases of “chronic movement.3 brain death.” In his provocative review, Shew- Partial retained function of the medulla mon collected patients from the literature who oblongata precludes the diagnosis of brain apparently fulfilled the clinical criteria of brain death. The traumatic brain lesions in our case death but had a prolonged somatic survival.7 illustration are characteristic of a severe impact Intensive care was continued because family in which primary brainstem lesions are more members were strongly motivated by religious commonly found. In fact, a recent MR imaging reasons, failed to accept the hopelessness of study documented 64% brainstem lesions of this state, or support was prolonged to have a 61 consecutive patients with traumatic brain pregnancy come to term. The tendency of injury and a fatal outcome in patients with these cases to “survive” with simple support bilateral pontine lesions.5 Although not par- systems casts doubt on the somatic disintegra- ticularly of primary relevance to our case, we tion after brain death. However, we suspect think that the brainstem lesions are primary that in some of these accumulated patients, because the confluent distribution in the cough reflexes and breathing drive were not midbrain is diVerent from the typical second- specifically mentioned as absent or remained ary linear type of haemorrhage which occurs untested. Failure to recognise persistent me- along the midline perforators in the midbrain. dulla function may not only lead to misdiagno- No evidence of anterior-posterior elongation of sis of brain death, but may also incorrectly the midbrain caused by lateral compression of imply that the brain and brain stem are not

the mesial temporal lobe herniation was found. critical organs for function of physiological sys- http://jnnp.bmj.com/ In addition, the admission CT demonstrated tems. rostral brainstem haemorrhages. Our patient example not only reinforces the 1 Wijdicks EFM. Determining brain death in adults. Neurol- ogy 1995;45:1003–101. need for apnoea testing but also careful evalua- 2 Ashwal S. Brain death in the newborn. Current perspec- tion of cough reflexes. Failure to examine tives. Clin Perinatol 1997;24:859–882. cough responses with bronchial suctioning or 3 Ropper AH, Kennedy SK, Russel L. Apnea testing in the diagnosis of brain death. J Neurosurg 1981;55:942–6. deferral of a formal apnoea test (for example, in 4 Goudreau JL, Wijdicks EFM, Emery S. Complications dur- ing apnea testing in the determination of brain death: pre- patients with neurogenic pulmonary oedema) on October 2, 2021 by guest. Protected copyright. disposing factors. Neurology 2000;55:1045–80. may lead to incorrect diagnosis of brain death. 5 Firsching R, Woischneck D, Diedrich M, et al. Early The presence of an intact or partly damaged magnetic resonance imaging of brainstem lesions after severe head injury. J Neurosurg 1998;89:707–12. medulla oblongata may result in normal 6 Pallis C, Harley DH. ABC of , 2nd ed. London: BMJ. regulation of cardiovascular, respiratory, and 7 Shewmon DA. Chronic brain death: meta-analysis and con- autonomic function and may stabilise the ceptual consequences. Neurology 1998;51:1538–45.

www.jnnp.com