In summary: EEG is used only as an ancillary test and is not one of the 3 main clinical criteria of brain death (in most US states).
From (Waters et al. 2004):
The diagnosis of brain death in most European countries and in the US, is defined as the total and irreversible loss of all brain functions (the ‘whole’ brain concept).
Only in the UK is brain death defined as a non‐functional state of the brainstem in which all the signs of brainstem activity are absent (the ‘lower’ brain concept). In this context, examination of cortical and subcortical bioelectrical activity is not necessary in the UK and therefore an EEG and other technical tests need not be performed.
(Wijdicks et al. 2010) gives a description of the state of brain-death determination in the United States.
Most US states have adopted the Uniform Determination of Death Act. That act, however, defers to "accepted medical standards" in defining determination of death.
In 1995, the American Academy of Neurology published a practice parameter to establish these medical standards. The practice parameters emphasized three necessary clinical findings to confirm brain death:
- coma with a known cause,
- absense of brainstem reflexes, and
(The 1995 practice parameter is available here, and (Wijdicks 1995) has more background information.)
The EEG is listed in the practice parameter as an ancillary test. "In adults, ancillary tests are not needed for the clinical diagnosis of brain death and can not replace a neurologic examination. Physicians ordering ancillary test should appreciate the disparities between tests and the potential for false-positives (i.e. the test suggests brain death, but the patient does not meet the clinical criteria)." (Wijdicks et al. 2010)
The EEG (along with other confirmatory tests) is recommended in certain situations where the clinical criteria are known to be less reliable (severe facial trauma, preexisting pupillary abnormalities, toxic levels of drugs, sleep apnea).
In the 2010 review of the effectiveness of the clinical criteria established in 1995, they conclude:
In adults, recovery of neurologic function has not been reported after the clinical diagnosis of brain death has been established using the criteria given in the 1995 ANN practice parameter.
Despite this apparent practical effectiveness, they do emphasize that the guidance given in the 1995 practice parameter and repeated in 2010 is not evidence-based:
Many of the details of the clinical neurological examination to determine brain death cannot be established by evidence-based methods. The detailed brain death evaluation protocol that follows is intended as a useful tool for clinicians. It must be emphasized that this guidance is opinion-based. Alternative protocols may be equally informative.
Waters, C. E., G. French, and M. Burt. "Difficulty in brainstem death testing in the presence of high spinal cord injury." British journal of anaesthesia 92, no. 5 (2004): 760-764.
Wijdicks, Eelco FM. "Determining brain death in adults." Neurology 45, no. 5 (1995): 1003-1011.
Wijdicks, Eelco FM, Panayiotis N. Varelas, Gary S. Gronseth, and David M. Greer. "Evidence-based guideline update: Determining brain death in adults Report of the Quality Standards Subcommittee of the American Academy of Neurology." Neurology 74, no. 23 (2010): 1911-1918.