- A site dedicated to provide information about brain death

This information is based on the American Acadamy of Neurology Guidelines.

Disclaimer: All information provided is felt to be accurate, but accuracy and completeness can not be guaranteed.. You must check with your state, hospital, or medical director for local policies and procedures. Use of this material is at your own risk.

General Considerations and Prerequisites:

Brain death examination is usually performed by neurologists or neurosurgeons and is a direct consequence of their involvement in the care of patients with an acute brain injury. Unavailability of these specialists in the neurosciences, however, still allows clinical determination of brain death by a medical or surgical intensivist, anesthesiologist, or a pediatrician. There is a general consensus that transplant surgeons should not be involved in the clinical diagnosis or indications for confirmatory tests. It is equally obvious that physicians who diagnose brain death should have experience in order to reduce errors or the neglect of key elements of the neurologic examination.

Any physician who is asked to determine brain death needs to personally review the CT scan of the brain in detail. Typically a large size intracranial hemorrhage, massive ischemic stroke with brain shift, multiple hemorrhagic contusions, large extradural or subdural hematoma, diffuse brain edema with absent basal cisterns, fissures and sulci, pontine hemorrhage with obstructive hydrocephalus are noted, all readily explaining coma and loss of brain stem reflexes. One should not proceed with examination if a CT scan is not available or if the CT scan is completely normal. Only in patients seen very early after cardiac or respiratory resuscitation and sometimes in patients with fulminant meningoencephalitis the CT scan may be normal but repeated CT studies should show diffuse, global brain edema. Be as it may, when patients with large, destructive lesions in the brain are seen in the Emergency Department factors such as, sedatives or alcohol may still play a role and confound clinical examination. When CT scan shows a destructive lesion but not much of brain shift or edema it is advised to repeat the CT scan and often progression of the abnormalities are seen. This is more likely observed in patients treated with anticoagulants and epidural hematomas, particularly those located in the posterior fossa. In the event of a repeatedly normal CT scan examination, cerebrospinal fluid could reveal diagnostic findings such as pleocytosis, increased erythrocyte count, or positive Gram stain. CSF should be collected for a PCR specifically directed towards herpes simplex or rabies.

The most impressive confounder of brain death is hypothermia. The diagnosis of brain death cannot be reliably made until the core temperature has reached 32°C. (It has led to the dictum "a patient is not dead until the patient is warm and dead.") Patients with marked hypothermia have lost the ability to shiver, feel as cold as a toad, often display tachycardia, and may have lost all brain stem reflexes when the temperatures dip into the 28°C range. Accidental hypothermia is frequent and often involves young survivors of mountaineering accidents or rescues from submersion into ice water. Occasionally hypothermia is additionally associated with ingestion of drugs such as opioids, barbiturates, benzodiazepines, phenothiazines, tricyclic antidepressants, and Lithium but exposure to cold is necessary. None of these patients should have a neurologic examination until rewarmed using heating blankets. When no perfusing cardiac rhythm is present, rewarming through cardiopulmonary bypass is the only option.

Obviously, drug ingestion may seriously confound neurologic examination, although there have been only a few reports (tricyclic antidepressants) of patient examinations that have clearly mimicked the diagnosis of brain death. An important differentiating clinical feature is that pupil response to light persists in the overwhelming majority of poisoning. Drug screens may be helpful but may miss certain drug types. Alcohol, barbiturates, antiepileptics, benzodiazepines, antihistamines, antidepressants, antipsychotics, stimulants such as amphetamines, narcotics, analgesics, and many of the cardiovascular drugs can all be traced in most toxicological screens. In addition, several drugs induce acid-base abnormalities that should already point to a possible intoxication. Respiratory acidosis is associated with opiates, ethanol, barbiturates, and anesthetics. Metabolic acidosis is common in acetaminophen, ethanol and methanol, as well as ethylene glycol, salicylates, isoniazid, cyanide, cocaine, strychnine and papaverine. However, alcohol seems to be the most common confounder. Alcohol has a plasma half-life of 10 mL/hour and the legal alcohol limit for determination of brain death is 800-1500 mg/L.

A reasonable approach in possibly drugged patients who may actually be brain dead is as follows:

  • One should administer the antidote naloxone and flumazenil first when opioids and benzodiazepines are considered.
  • Brain death can be declared if a screening test reveals traces of drug below the therapeutic range.
  • When a drug or poison cannot be quantified but seems highly likely from history, one should not make the diagnosis of brain death but observe the patient for at least four times the excretion half-life assuming no half-life prolongation due to additional organ dysfunction.

In any other case in which a suspicion of a drug intoxification remains high but cannot be identified one should continue to observe the patient for at least two days for a change in brain stem reflexes and motor response. If none are observed, a confirmatory test, preferably a cerebral angiogram can be performed if transplantation is still considered.

A common clinical problem is the presence of serum pentobarbital levels in patients with traumatic brain injury. Barbiturates have been used in these patients to reduce increased intracranial pressure. Again, it is assumed that the clinical diagnosis of brain death can be made if the serum levels are less than the therapeutic range. (Most laboratories would be less than 5 mcg/mL.)

Acute metabolic derangement and endocrine crisis can mimic brain death but more often diffuse cerebral edema, extensive demyelination, or anoxic ischemic injury, is a consequence of these derangements. Examples are brain edema in fulminant hepatic failure and ketotic hyperglycemia. Many severe abnormalities such as hyponatremia, hypernatremia, hyperglycemia or hypoglycemia, hypothyroidism, pan- hypopituitarism or Addison’s disease may decrease the level of consciousness and confound neurologic examination, but a complete loss of brain stem reflexes is seldomly observed.

Only when all these above-mentioned confounders are excluded, should one proceed with a clinical examination to determine brain death.


Neurologic examination of patients suspected of brain death needs to be precise and document absolutes. Nonetheless, pitfalls remain and misinterpretations are sources for discussions between physicians. Sometimes facial injury may make assessment very difficult.

Several components need to be examined.

1) Coma. Patients should lack all evidence of responsiveness due to complete loss of consciousness. The depth of coma is usually examined by using standard painful stimuli, compressing on the supraorbital nerve (localized on the medial part of the eyebrow ridge), forceful nailbed pressure, or temporomandibular joint compression. Any other pain stimulus may be insufficient and leads to different interpretations. One should resist the use of painful stimuli such as sternal rubbing, twisting of forearm or nipples, or simply applying pinpricks on several parts of the extremities. Eye opening or motor response to pain or voice should be repeatedly absent. But in occasional circumstances a pain stimulus may generate a quick, slow response in the extremities. Typically they are seen in the arms and may involve brief flexion in the fingers or a minimal eyelid elevation. The pathways are unclear but they should be considered generated from the spinal cord. Its clinical differentiation from a retained motor response requires expertise but the absence of a coordinated pathological flexion or extension is typical and many of these movements disappear after repeated stimulation.

2) Absence of brain stem reflexes. First the pupils are examined. No Response to a bright light source such as a torch light is documented in both eyes but if any uncertainty exists a magnifying glass should be used. Patients who are brain dead have typically mid position pupils that are between 4-6 mm in size. Topical ocular installation of drugs and trauma to the bulbus oculi may cause pupil abnormalities, and some patients do have preexisting abnormalities which will make the examination unreliable. The ocular movements are best tested using ice water injection. The oculocephalic reflex can be tested first. Fast turning of the head to both sides should not produce any eye movement. However when some conjunctival swelling has occurred in trauma this may be difficult to elicit. Vestibular testing with ice water should be a standard procedure. The head is elevated to 30 degrees to allow the horizontal canal to become vertical. Injection of 50-cc ice water into the external auditory canal will not result in a tonic movement to the side of the stimulus (in coma the slow component remains and the fast nystagmus component disappears). Both sides need to be tested with an interval of five minutes. Caloric testing can be muted when blood or cerumen in the auditory canal and therefore direct inspection is necessary before the test is performed. This test is followed by examination of facial sensation and facial motor response. Absent corneal reflex can be confirmed with a throat swab and no blinking should be seen. Absent grimacing to pain is confirmed with deep pressure on condyles at the level of the temporomandibular joint. The jaw reflex should also be absent (tapping on the chin with a reflex hammer). The bulbar function is best tested by examining cough response to bronchial suctioning. Simply moving the endotracheal tube to and fro is not an adequate stimulus. A catheter should be inserted into the trachea and advanced to the level of the carina and followed by suctioning trying to elicit a cough response. The gag reflex remains an unreliable test. If these brain stem reflexes are absent, one can proceed with the apnea test.

3) The apnea test. This procedure is based on disconnection of the ventilator and the principle of apneic oxygenation. Preoxygenation is necessary to eliminate the nitrogen stores and facilitate oxygen transport. Typically PaCO2 rises in the first minutes due to equilibration of the arterial carbon dioxide with mixed central venous carbon dioxide followed by a rapid rise of PaCO2 due to metabolic production of the carbon dioxide. It is estimated that approximately 3-6 mmHg/minute rise of PaCO2 occurs after disconnection from the ventilator. The target PCO2 is 60 mmHg but several additional precautions are necessary before proceeding with the test. Hypothermia should be corrected from 32°, which is a prerequisite for a determination of brain death, to normothermia 36°-37°. Patients who are hypotensive-systolic blood pressure less than 90 mmHg- should receive a fluid bolus or an increase in their intravenous dopamine. Preoxygenation aiming at a PO2 of at least 200 mmHg as well as correcting hypocapnia aiming at a normal PaCO2 (40mmHg) is necessary before disconnecting the ventilator. Usually ten minutes of 100% oxygen with a FiO2 of 1 is sufficient to have the PO2 rise to these levels. After disconnection of the ventilator adequate oxygen source using a catheter placed at the carina (6 L/minute) is needed. For approximately eight minutes (8 times 3 mm Hg is 24 mmHg rise from baseline results in the desired target value) the patient’s abdomen and chest are observed for movements. If no movements are seen a new blood gas is drawn. When the PCO2 has reached 60 mmHg the apnea test is positive and in combination with exclusion of confounding factors, absent brain stem reflexes and coma, the clinical diagnosis of brain death can be made. During the apnea test continuous monitoring of oxygen saturation, pulse, and blood pressure is needed. If a hypotension occurs the apnea test should be aborted. Typically hypotension occurs in a poorly oxygenated patient or when respiratory acidosis becomes profound. An immediate blood gas needs to be performed before reconnecting to the ventilator, commonly documenting an already significant elevated PaCO2. Alternative methods of the apnea test using potential dangerous carbon dioxide augmentation as well as using the ventilator and continuous positive airway pressure should be discouraged. False readings can be seen when the mechanical ventilator sensors are used to document apnea. If the PCO2 has not risen to the desired target range, the apnea test should be repeated, this time for ten minutes.


All information provided is felt to be accurate, but accuracy and completeness can not be guarenteed.. You must check with your state, hospital, or medical director for local policies and proceedures. Use of this material is at your own risk.

Clinical Exam
Confirmatory tests
Frequently Asked Questions