Effect of Inpatient Electroencephalography on Clinical Decision Making: EEG Is More Valuable Than Findings Suggest

To the Editor:

We read with interest the findings of the retrospective investigation of the role of electroencephalography (EEG) in the treatment of hospitalized patients by Harmon and colleagues in the December 2013 issue of The Journal of the American Osteopathic Association.¹ In their cohort of 200 inpatients, EEG was found to rarely contribute to clinical decision making and in no case resulted in a change in diagnosis or management. We have some concerns that if the findings are not viewed in the context of the overall value of EEG, these results may lead clinicians to believe that EEG is not as useful a diagnostic modality as was previously believed and could result in decreased orders for EEGs when clinically appropriate.

Although EEG ordered apart from specific indications may not always contribute to clinical decision making, there remain many instances in which EEG is a valuable clinical tool. For example, EEG remains the most useful laboratory test for the classification of seizures and specific epileptic syndromes.² A generalized seizure and a partial seizure with rapid secondary generalization may be very difficult to distinguish clinically, but the seizure type may be readily determined with an EEG if a recording is made during the onset of a seizure. In some cases, EEG may be the only modality that can conclusively distinguish a seizure from a pseudo-seizure. Although computed tomography and magnetic resonance imaging are now generally used to assess focal lesions, focal slowing visible on EEG recordings can sometimes reveal pathology in a specific brain region that would not be detected on structural imaging.²

In addition, EEG is one of the most helpful modalities available for diagnosing and monitoring delirium. The fluctuating state of awareness in delirium is accompanied by characteristic EEG changes, and the varying levels of a patient's attention parallel the slowing of background EEG rhythms.³ To some degree, the same type of findings may apply to patients in a coma; in some such cases, continuous or frequent EEG monitoring in the intensive care unit setting may be of value.⁴ Slowing of EEG rhythms that accompany dementia can also progress as dementia advances, providing a modality for monitoring disease severity. Electroencephalography may also be quite useful for conclusively distinguishing dementia from pseudodementia (a syndrome in which dementia is mimicked by depression or other psychiatric disorders).²

Quantitative EEG, the transformation of selected EEG activity such as frequency or voltage by Fourier analysis into numeric values, which are often mapped, was not considered in the study. Quantitative EEG may provide valuable information in several clinical conditions. For example, quantitative EEG has been shown to be useful in the assessment of mild traumatic brain injury, even in the absence of brain abnormalities on magnetic resonance images.⁵ Quantitative EEG analysis of frequency and mean theta power may be sensitive to the early presence of subjective cognitive dysfunction and might be useful in the initial evaluation of patients suspected of having dementia, as well as in estimating the degree of cognitive deterioration over time.⁶ In addition, there is some evidence that distinctive electrophysiologic profiles may be associated with different psychiatric disorders.⁷

It should be noted that because the study by Harmon and colleagues¹ involved only inpatients undergoing EEG, the implications of the results cannot be extended beyond hospitalized patients. In our experience, a large percentage of EEGs—possibly the majority—are performed on outpatients. Therefore, it is important for physicians to consider that these findings may not be applicable to patients who undergo EEG as a whole.

Part of the problem with EEG is that the EEG signal is quite sensitive to many variables (eg, metabolic, vascular, endocrinologic) that affect central nervous system function. As a result, EEG findings are often nonspecific. However, this aspect of EEG should not cause clinicians to avoid using EEG in situations in which it could be helpful. As suggested by Harmon and colleagues,¹ clearer guidelines for the use of EEG should be promulgated. We agree that further education regarding indications for EEG is needed to reduce health care costs, but physicians must continue to use this valuable diagnostic modality when appropriate. Certainly, as the authors suggest, further research on the effectiveness of EEG (particularly, in our opinion, quantitative EEG, which has much unexplored clinical potential) is warranted.