special article 1995;45:1411-1413 Practice parameter: The electroencephalogram in the evaluation of (Summary statement)

Report of the Quality Standards Subcommittee of the American Academy of Neurology

Overview. The Quality Standards Subcommittee ticles remained. We reviewed these articles in their (QSS) of the American Academy of Neurology entirety.l (AAN) develops practice parameters for neurolo- Literature classification criteria. In general gists to use in evaluating clinical disorders, partic- terms, a test may be considered clinically more use- ularly with regard to selecting diagnostic proce- ful than competing modalities if it effectively (ie, dures and treatment. In choosing topics for practice with adequate sensitivity andor specificity) or efi- parameters, the QSS considers several factors- ciently (ie, with acceptable risk and cost) separates prevalence of disease, frequency of procedures, patients into distinct diagnostic groups or provides costs, membership requests, controversy, urgency, prognostic or therapeutic information. We com- external constraints, and others. AAN members re- pared the literature on EEG in headache patients quested a practice parameter on the use of the elec- against the following generally accepted criteria for troencephalogram (EEG) in the evaluation of pa- determining the clinical usefulness of a diagnostic tients with headache. test: 1. A comparison of the test with accepted inde- Justification. Practice variations have created the pendent diagnostic criteria (ie, a “gold standard”). need for development and use of practice parame- Potential diagnostic criteria against which to com- ters. The neurologist has long been a resource for pare the EEG in headache patients include clinical patients with headache by excluding serious under- criteria for the diagnosis of headache disorders, the lying causes, confirming a correct diagnosis, and state of the patient after follow-up, and the results planning an effective management strategy. If the of head-imaging studies. EEG is used for patients with headache in some ge- 2. A comparison that includes normal matched ographic areas or clinical settings and not in oth- controls, patients with a spectrum of disease (eg, ers, patient outcome from the use of this test and mild to severe), and individuals with commonly how it relates to the neurologist’s role in headache confused disorders. Ideally, the EEG studies should management must be explored. include normal subjects without headache, patients with different headache subtypes (eg, Description of the process. We performed a with aura, migraine without aura, tension MEDLINE search for articles published between ), and individuals with structural cause 1966 and 1994 using the key words “electroenceph- of headache. alography” and “headache.” Search terms were ex- 3. A description of the setting in which the com- ploded as appropriate. We also performed text word parison is done, so that a judgment of potential bi- searches using these terms. Additionally, we did a ases can be made. secondary search of the bibliographies of these arti- 4. A blinded interpretation of the EEG as well as cles to identify articles published before 1966. a measure of the reproducibility and observer vari- Using this strategy, we found 90 articles in which ation of the test. Disagreement between EEG inter- authors discussed the EEG and headache diagnosis preters is common. or management. After excluding case reports, case 5. A sensible definition of normal as applied to series of subjects selected because of abnormal the test. Many EEG findings formerly thought to EEGs, abstracts with insufficient information for be abnormal have subsequently been shown to be analysis, and reviews without original data, 40 ar- normal variants. I See also page 1263 1

Approved by the Quality Standards Subcommittee July 17, 1994. Approved by the Practice Committee July 28, 1994. Approved by the Executive Board September 23, 1994. Address correspondence and reprint requests to Joanne F. Okagaki, American Academy of Neurology, Suite 335,2221 University Avenue SE,Minneapo- lis, MN 55414. The background paper to this practice parameter, published in this issue, is also available upon request from the American Academy of Neurology office (phone 612/623-2439).

July 1995 NEUROLOGY 45 1411 6. A description of the techniques that permits 80% to 91%. Although the relatively high sensitivi- exact replication. Standard techniques for perform- ties and specificities reported suggest that the H- ing an EEG, including the number of channels and response may be effective at distinguishing mi- recording electrodes, accepted montages, and com- graine patients from controls, and possibly mi- puter interpretation, have changed. graineurs from tension headache sufferers, it is im- portant to recall that in all these studies the H-re- Results. Summary. Headache disorders are clini- sponse was compared with established clinical cri- cal syndromes defined by historical criteria. The teria. One cannot conclude from any of these stud- EEG is not included in the diagnostic criteria of the ies that the H-response, or any other EEG abnor- International Headache Society for migraine or mality, is more effective or efficient than the neuro- other major headache categories. The majority of logic history and neurologic examination at deter- headache sufferers do not have an identifiable mining whether a patient has headaches. Thus, al- structural lesion that explains their . An EEG though the presence of the H-response may have could be considered useful in the evaluation of interesting implications for the underlying patho- headache if abnormalities were observed that (1) genesis of migraine, it adds nothing to the diagnos- separate persons with and without headache, po- tic evaluation of the headache patient. tentially shedding light on the underlying patho- Does the EEG define headache subgroups? A physiology of headaches; (2) define subgroups asso- few of the studies attempted to define headache ciated with different natural histories or responses subgroups based on the EEG. Some only identified to therapy; or (3) effectively identify patients with a findings that by modern criteria would be consid- definable structural etiology. ered normal, while others failed to provide con- The articles reviewed ranged in publication date vincing evidence that EEG can discern headache from 1941 to 1994. One article met all six of the lit- type. erature criteria described in the preceding section, Migraine and may coexist in some pa- three articles met five criteria, 11 met four, nine tients. Postictal headache in patients with epilepsy met three, seven met two, seven met one, and two may have many of the same clinical features as id- met none. iopathic migraine headaches. Most patients with Many of these studies have major flaws, includ- epilepsy and headache do not present a diagnostic ing (1)biases such as referral bias (not population- dilemma. In patients presenting with headache in based) or verification bias (selecting controls after a whom atypical associated symptomatology makes a normal EEG); (2) poorly controlled (most studies disorder reasonably probable, epileptiform were uncontrolled or, when controlled, were not activity on an EEG would significantly raise the age-hex-matched); (3) not blinded; (4) high ob- probability of epilepsy. Even though little informa- server variability (similar abnormality rates, differ- tion is available, it is reasonable to assume that ent abnormalities); and (5) archaic criteria for nor- some of the symptoms that increase the probability malcy (patterns originally considered abnormal of a seizure disorder are atypical migrainous aura subsequently described in normal subjects, eg, pos- and episodic loss of consciousness. terior slow waves of youth). In addition, the EEG Does the EEG identify patients whose headaches had different rates of abnormality in headache sub- have a structural cause? The paucity of literature types but was not helpful in distinguishing them. that addresses this question suggests that the EEG Authors could not relate EEG findings to headache does not effectively identify headache patients with severity or response to proposed treatment. These underlying structural lesions and should not be studies fail to show an association of EEG with a considered an effective screen for a malignant malignant cause for headache, different response to headache etiology. The sensitivity of EEG is consid- treatment, or difference in prognosis. erably less than that of head CT or MRI in identify- Is there an increased prevalence of EEG abnor- ing patients with brain tumors. Assuming the malities in headache patients? Studies designed to availability of CT or MRI, there is little justifica- determine whether headache patients have an in- tion for the use of EEG when an intracranial struc- creased prevalence of EEG abnormalities report tural lesion is suspected. conflicting results. The only well-matched con- Conclusion. No study has consistently demon- trolled studies to address this question either found strated that the EEG improves diagnostic accuracy that there was no increased prevalence of EEG ab- for the headache sufferer. The EEG has not been normalities in patients with headache or found that convincingly shown to identify headache subtypes, there was a prominent photic driving response. nor has it been shown to be an effective screening Prominent photic driving at high flash frequen- tool for structural causes of headache. cies (H-response) in migraine patients is the most consistently reported difference between headache Recommendation. We therefore recommend the patients and controls. All the studies describing the following: photic driving response used specialized EEG tech- The EEG is not useful in the routine evaluation niques, including wave analyzers and spectral of patients with headache (guideline). This does not analysis. The reported sensitivity of the H-response exclude the use of EEG to evaluate headache pa- varied from 26% to loo%, and the specificity from tients with associated symptoms suggesting a

1412 NEUROLOGY 45 July 1995 seizure disorder, such as atypical migrainous aura or episodic loss of consciousness. Assuming head- DE F I M T I0 N S imaging capabilities are readily available, EEG is not recommended to exclude a structural cause for Classification of evidence headache (option). Class I. Evidence provided by one or more well-designed randomized controlled clinical trials, including overviews Future research. Numerous interesting and po- (meta-analyses) of such trials. tentially fruitful areas of research on the EEG in the setting of chronic headache remain. Studies Class 11. Evidence provided by one or more well-de- comparing the effectiveness of EEG with that of signed clinical studies such as case-control studies, co- head imaging in screening for intracranial pathol- hort studies, and so forth. ogy in patients with headache, taking cost and pa- Class 111. Evidence provided by expert opinion, nonran- tient preferences into account, would be worth- domized historical controls, or one or more case reports. while. Additionally, studies designed to determine whether specific EEG characteristics, such as Strength of recommendations prominent photic driving, identify subsets of head- ache patients with distinctive responses to therapy Standards. Generally accepted principles for patient would be useful. management that reflect a high degree of clinical cer- tainty (ie, based on class I evidence or, when circum- stances preclude randomized clinical trials, overwhelm- ing evidence from class 11 studies that directly addresses Acknowledgments the question at hand or from decision analysis that di- rectly addresses all the issues). The Quality Standards Subcommittee wishes to thank members of the American Academy of Neurology's Member Reviewer Net- Guidelines. Recommendations for patient management work, as well as members of the Headache and Facial Pain Sec- that may identify a particular strategy or range of man- tion and the Clinical Neurophysiology Section, for reviewing agement strategies and that reflect moderate clinical cer- this practice parameter. Special thanks go to coauthors Michael tainty (ie, based on class I1 evidence that directly ad- K. Greenberg, MD (who also served as project facilitator), and dresses the issue, decision analysis that directly ad- Gary S. Gronseth, MD. dresses the issue, or strong consensus of class 111 evi- dence). The following professional associations reviewed and com- mented on this practice parameter: American Association for Practice options or advisories. Other strategies for the Study of Headache, American EEG Society, American Medi- patient management for which the clinical utility is un- cal EEG Association, American Academy of Family Physicians, certain (ie, based on inconclusive or conflicting evidence American College of Physicians, and American Society of Inter- or opinion). nal Medicine. Practice parameters. Results, in the form of one or more specific recommendations, from a scientifically Quality Standards Subcommittee: Jay Rosenberg, based analysis of a specific clinical problem. MD (Chair); Milton Alter, MD, PhD; Thomas N. Byrne, MD; Jasper R. Daube, MD; Gary Franklin, MD, MPH; Michael L. Goldstein, MD; Michael K. Greenberg, MD (Facilitator, Coauthor); Douglas J. This statement is provided as an educational service Lanska, MD; Shrikant Mishra, MD, MBA; Ger- of the American Academy of Neurology. It is based maine L. Odenheimer, MD; George Paulson, MD; on an assessment of current scientific and clinical in- Richard A. Pearl, MD; and James Stevens, MD. formation. It is not intended to include all possible proper methods of care for a particular neurologic problem or all legitimate criteria for choosing to use a specific procedure. Neither is it intended to exclude Reference any reasonable alternative methodologies. The AAN recognizes that specific patient care decisions are the 1. Gronseth GS, Greenberg MK. The utility of the electro- prerogative of the patient and the physician caring encephalogram in the evaluation of patients presenting with for the patient, based on all of the circumstances in- headache: a review of the literature. Neurology 1995;45: volved. 1263-1267.

July 1995 NEUROLOGY 45 1413 Practice parameter: The electroencephalogram in the evaluation of headache (Summary statement) Report of the Quality Standards Subcommittee of the American Academy of Neurology Neurology 1995;45;1411-1413 DOI 10.1212/WNL.45.7.1411

This information is current as of July 1, 1995

Updated Information & including high resolution figures, can be found at: Services http://n.neurology.org/content/45/7/1411.citation.full

Permissions & Licensing Information about reproducing this article in parts (figures,tables) or in its entirety can be found online at: http://www.neurology.org/about/about_the_journal#permissions Reprints Information about ordering reprints can be found online: http://n.neurology.org/subscribers/advertise

Neurology ® is the official journal of the American Academy of Neurology. Published continuously since 1951, it is now a weekly with 48 issues per year. Copyright Copyright 1995 by the American Academy of Neurology. All rights reserved. Print ISSN: 0028-3878. Online ISSN: 1526-632X.