There is no room for error when diagnosing brain death. Incorrect diagnosis causes inappropriate withdrawal of care. However, delayed or missed diagnoses lead to futile care and lost opportunities for organ donation.
American neurologists do not have a consistent rationale for accepting brain death as death, nor a clear understanding of diagnostic tests for brain death –Joffe 2012
Because brain death determinations are infrequent, general neurologists taking on hospital services occasionally who are called upon to declare a patient brain dead will rarely be able to acquire enough experience. –Wijdicks 2013
As with many uncommon situations that are unique to critical illness, we cannot always rely on specialist consultation. Critical care practitioners must develop a firm grasp of this diagnosis. This post will explore some diagnostic conundrums in brain death diagnosis. The radionuclide flow scan is emphasized because it is the most commonly used tool to sort out difficult cases.
Overview of approach to brain death
Pursuing a diagnosis of brain death too early may lead to premature diagnostic closure. The initial approach to the patient should focus on supportive care and thoughtful investigation of coma. Brain death evaluation should be pursued in patients with no brain activity on neurologic examination, a mechanism of injury compatible with brain death (e.g. cardiac arrest, massive intracranial hemorrhage), and neuroimaging consistent with brain death (e.g. herniation, diffuse edema).
In straightforward cases, brain death may be established on the basis of a bedside neurologic examination and apnea test (1). However, in some situations the patient may not be stable enough to tolerate apnea (e.g. due to hypoxemic respiratory failure). There are other situations in which the apnea test may provide incorrect information (e.g. intoxication, C-spine damage, COPD with chronic hypercapnia). Finally, some situations prevent performance of a complete neurologic examination (e.g. severe facial or ocular trauma). In all of these cases an ancillary test is required, most often a radionuclide flow scan.
Movements observable after brain death
Brain death liberates the spinal cord of any supervision or suppression from the brain, causing spinal motor reflexes to run amok. This produces a variety of bizarre movements, often in response to stimulation. For example:
- Triple flexion: probably the most common movement observed. Stimulation of the feet causes flexion at the ankles, knees, and hips. This may appear to be a volitional withdrawal response to painful stimulation of the feet.
- Myokymia: localized quivering of a muscle (may involve facial or ocular muscles).
- Lazarus sign: flexion at the waist.
These movements are alarming to family members and medical personnel who aren't experienced with brain death. It's common for a brain death diagnosis to be inappropriately delayed or missed due to such movements.
Brain death mimics
Brain death mimics can be divided into roughly two groups.
Wrong-diagnosis mimics are the most concerning. These are patients with an unrecognized alternative diagnosis which causes the patient to appear brain dead:
- Unrecognized intoxication (e.g., tricyclics, lidocaine, baclofen, barbiturates, paralytics, anticholinergics, organophosphates)
- Locked-in syndrome
- High C-spine injury
- Fulminant Guillian-Barre Syndrome
These patients have conditions which are often treatable and consistent with long-term survival. Accurate diagnosis is essential. One important clue to wrong-diagnosis mimics is that they are generally associated with normal neuroimaging.
Wrong-severity mimics do indeed have catastrophic, irreversible brain damage. They have no likelihood of a meaningful recovery. However, they aren't brain-dead due to the persistence of some brainstem functions.
An example would be a patient who has profound anoxic brain injury with marked cerebral edema. The patient has an isoelectric EEG for several days with no confounding factors (e.g. no sedation or hypothermia). The patient has no cranial nerve reflexes or responsiveness to pain. However, the patient does continue to generate a respiratory rate on apnea examination. This patient has no likelihood of awakening from coma. However, the patient is not brain dead due to the absence of apnea. The injury isn't quite severe enough to qualify for brain death.
Basics of the radionuclide flow scan
Underlying physiology: why is there no flow?
The lack of blood flow to the brain is explained as shown below. Regardless of the initial injury, eventually brain death occurs due to a spiral of progressive intracranial hypertension, tissue damage, and edema. The final common pathway is profoundly elevated intracranial pressure with intracranial circulatory arrest.
Details of flow scan
The radionuclide flow scan involves injection of a radiotracer dye into a peripheral vein. If there is any perfusion to brain tissue, radionuclide will enter brain tissue. Deposition of radiotracer is monitored over a period of two hours. Brain death causes lack of blood flow to the brain, causing an “empty skull sign.”
Various types of flow scans are available. The most widely used is a two-dimensional flow scan which involves radionuclide imaging in the frontal and lateral projections. This has the advantage that the scanner is portable and may be brought to the ICU. Logistically this is convenient because it is fairly rapid, inexpensive, and may be available around the clock in larger hospitals.
Other types of flow scanning include CT-SPECT, which is a three-dimensional nuclear flow scan. This post will focus on two-dimensional (“planar”) radionuclide scanning because this is the most commonly utilized confirmatory test in the United States currently.
The flow scan remains accurate despite medications
One problem which is increasingly common in the opioid epidemic is patients who present with apparent brain death and cerebral edema following cardiac arrest due to intoxication. Brain death declaration traditionally requires the allowance of >5 half-lives to facilitate dissipation of any sedative drug. This is a flawed concept, because in overdose the half-life of drugs often becomes extended and unpredictable. On a more pragmatic level, it is often unknown what drug(s) the patient may be intoxicated with.
A major advantage of the flow scan is that it remains accurate in the face of medications (Donohoe 2012). For patients with possible intoxication, a flow scan may allow evaluation of brain death without delay. This is an advantage compared to EEG, which can be flattened by some medications (particularly barbiturates).
Some additional benefits of a flow scan also include:
- It provides objective evidence which is permanently available on the record.
- Formal interpretation by a neuro-radiologist provides an un-biased “second opinion.”
Performance of radionuclide flow scans
Radionuclide flow studies are ~78% sensitive for brain death (Joffe 2010). Unfortunately, some patients who are brain dead as defined by clinical criteria will have some brain perfusion (2). The most common explanation for this phenomenon is that these patients are indeed in the process of brain death, but they haven't progressed quite far enough to completely abolish cerebral perfusion. Repeat perfusion scan after waiting (e.g. 12 hours) will often reveal complete absence of flow.
Some authors recommend delaying the initial perfusion scan for ~6 hours after discovering clinical finding of brain death to avoid this sort of false-negative scan. Lack of 100% sensitivity emphasizes that a perfusion scan shouldn't be obtained if it isn't indicated. If the clinical exam is sufficient to establish brain death and there aren't any confounding factors, then a perfusion scan may only serve to add confusion.
Another cause of persistent blood flow in patients who are clinically brain dead is anatomic abnormalities which prevent pressure accumulation in the skull (e.g. ventricular drain or skull defect; Wijdicks 2010). Such patients may reach a state of clinical brain death with severe tissue damage, but fail to accumulate enough pressure to cause intracranial circulatory arrest.
Lack of blood flow to the brain is 100% specific for catastrophic brain injury with inability to regain consciousness. A good-quality flow scan showing absence of flow is universally accepted as confirming brain death within the appropriate clinical context.
There are a few case reports of patients with undetectable blood flow who did subsequently demonstrate minimal brainstem activity (e.g., respiratory drive)(3). All of these patients had sustained catastrophic brain damage incompatible with consciousness. Thus, a two-dimension flow scan may rarely miss patients with a trickle of perfusion into the brainstem (wrong-severity mimics as discussed above). This phenomenon is extraordinarily rare and may relate to the quality of images obtained (for example, the flow scan shown above appears somewhat under-exposed). It may also relate to suboptimal imaging of the medulla using the flow scan:
Approach to using the perfusion scan
Like any diagnostic test, a flow scan is only useful when applied in the correct clinical context. The presence of impaired flow may be consistent with the early phases of brain death, so in this case a repeat study should be considered in 12-24 hr. A normal scan should prompt an aggressive search for wrong-diagnosis brain death mimics.
Crazy idea: does the definition of brain death require revision?
Our definition of brain death arose in 1968 from an ad hoc committee at Harvard Medical School. They set out to define clinical criteria which, when met, would indicate that the patient would never awaken from coma:
Given the tools at their disposal, they did a brilliant job. Neither CT nor MRI scanning existed in 1968. To be sure that a patient wouldn't wake up, it was necessary to define brain death to include lack of any brainstem function. By doing so, they ensured that the injury was so catastrophic that there was zero likelihood of recovery.
However, current radiologic techniques allow us to define brain death more precisely. In order to establish the presence of irreversible coma, what is truly required is necrosis of the cerebral cortex (supratentorial brain tissue). Without the cortex, awareness is impossible – even if the brainstem and cerebellum are working perfectly. In 1968 it was impossible to test regional brain viability. However, with advances in neuroradiology (e.g. CT-SPECT), it is increasingly possible to identify patients with no blood flow to the brain cortex. Regardless of brainstem perfusion and function, these patients are irreversibly comatose and arguably brain dead.
Redefining brain death may seem like a radical concept, but it's not a new idea (Rothstein 1993). The belief that irreversible coma is synonymous with death is an old and widely accepted concept (Joffe 2012). The scanning technology required to reliably determine this is improving rapidly. If a more precise definition of brain death could be validated and accepted, this might increase the number of lives saved by organ transplantation while reducing nonbeneficial care.
- Brain death diagnosis should be considered only after patient stabilization and evaluation for other possible causes of coma.
- Spinal reflexes may produce strange movements among brain dead patients, often causing confusion.
- The most concerning mimics of brain death are missed alternative diagnoses (e.g. intoxication, C-spine injury), because these may be treatable.
- Radionuclide flow scan may be used in confusing situations. The scan isn't entirely sensitive, because limited flow may continue in the early phases of brain death. However, lack of blood flow allows for a confident diagnosis of brain death (high specificity).
- Organ donation in the ED (EMCrit Podcast 52)
- Poisoning and the diagnosis of brain death (The Poison Review)
- Brain death (Radiopaedia)
- Brain Death (LITFL)
- Controversies in Brain Death (Martin Smith, SMACC-Dub)
- This is true in the United States, but in some countries a confirmatory test is required.
- This creates a bit of a problem regarding what the gold-standard is. One possibility is that these patients aren't truly brain dead (i.e. the clinical exam is nonspecific), whereas the other possibility is that they are brain dead (i.e. the perfusion scan is insensitive).
- Venkatram 2015. This article has been retracted because the interpretation of the scans in Case 2 were interpreted incorrectly. However, the central points of the article still seem to remain valid particularly with regards to Case 1. Another case report of this phenomenon was reported by Joffe 2010.