Because epilepsy is a group of conditions and not a single homogeneous disorder, and because seizures may be symptoms of both diverse brain disorders and an otherwise normal nervous system, it is neither possible nor desirable to develop inflexible guidelines for what constitutes a “standard” or “minimal” set of diagnostic tests. Furthermore, it is not always clear from the history that the patient has even had a seizure or has epilepsy. Thus, the role of the physician is threefold: (a) to determine whether epilepsy or seizures exist and not some alternative diagnosis; (b) to define, if possible, an underlying cause; and (c) to optimize treatment.

Epilepsy is primarily a historical diagnosis, and the initial assessment and approach to management are based, in large part, on the clinical history, especially on an accurate description of the event in question. Information should be obtained from the patient when possible, as well as from other individuals (especially family members) who have observed typical attacks. Memory is always highly selective, however, and the abrupt and unexpected nature of seizures—especially the first—may render witnessed accounts unreliable. The patient’s recollection may also be limited by alteration of consciousness and postictal amnesia. Important focal features may be overlooked, and estimates of seizure duration are almost invariably exaggerated. Auras are often particularly difficult to characterize from the patient’s description because the subjective experiences are only vaguely recalled (“a feeling in my head”) or conveyed using terms that different individuals use to mean quite different things (e.g., dizzy, shaking, chill, numbness, spacey). Finally, there are important age-related differences in seizure semiology that physicians must take into account. Even with these limitations, physicians must train themselves to take a systematic history that does not overlook important features (Table 1). Family members of patients who have recurrent episodes may succeed in capturing one or more of them on videotape, and this can provide invaluable information not otherwise available. As Aicardi and Taylor conclude (Chapter 70), “Clinical diagnosis is an intellectual process whereby all available sources of information, from the purely clinical to the highly technical, are integrated with a view to arrive at a meaningful conclusion.”

The clinical data from the history and physical examination should allow a reasonable determination of probable diagnosis, seizure and epilepsy classification, and the likelihood of an underlying pathologic condition of the brain. Based on these considerations, diagnostic testing should be undertaken selectively. Thus, a normal child with brief lapses of attention whose symptoms can be reproduced in the office by hyperventilation is a very different patient than is a middle-aged man who has developed partial seizures and seems to be getting progressively worse. It must be remembered, too, that diagnostic testing is never a substitute for clinical judgment and observation through follow-up and reexamination.

A great number of both routine and highly specialized diagnostic tests are available to aid physicians in the evaluation of patients with known or suspected seizures. Routine blood tests are rarely diagnostically useful in healthy children and adults but are necessary in newborns and in older patients with acute or chronic systemic disease to detect abnormal electrolyte, glucose, calcium, and magnesium values and impaired liver or kidney function.

Monitoring antiepileptic drug (AED) levels depends on the clinical situation. Leppik (Chapter 71) suggests that levels be determined at steady state after good seizure control has been attained to establish a useful benchmark. This can be repeated at annual follow-up as needed, for example, at the time of renewal of a driver’s license, to ensure that adequate levels are being maintained. Other times when AED levels may be indicated are when a patient has breakthrough seizures, experiences toxicity, or is being treated with other drugs known to influence AED metabolism (especially when such comedication is altered).

The electroencephalogram (EEG) remains the most useful diagnostic test when a seizure disorder is considered (Chapters 73,74,75). An EEG should generally be obtained in every case as an aid to diagnosis or to assist in classifying the type of seizure or epilepsy syndrome. In some instances, EEG findings will aid in prognosis and in determining the need for treatment. Buzáki and Traub (Chapter 72) provide an excellent discussion of the physiologic basis of EEG activity and introduce new concepts concerning EEG oscillatory patterns. They suggest that the term desynchronization is misleading and that synchronous EEG activities should be characterized in terms of slow and fast oscillatory patterns. The same oscillatory patterns may be responsible for generating both interictal (Chapter 73) and ictal (Chapter 74) discharges. Bleck et al. (Chapter 75) discuss the greatly expanded use of EEG in intensive care unit (ICU) settings, which has led to major improvements in ICU care. Tassinari and Rubboli (Chapter 77) advocate the use of polygraphic recordings in selected patients to help differentiate sleep-related phenomena from epileptic seizures. Visual qualitative analysis of seizure-related EEG discharges has relied mainly on the experience and expertise of the electroencephalographer. Wong and Lopes da Silva (Chapter 76) describe and advocate the use of quantitative methods to characterize electrical and electromagnetic signals responsible for interictal and ictal discharges. The magnetoencephalogram
(MEG) (discussed in Chapter 78) can provide additional information concerning the spatial features of the epileptogenic brain region, although it is unlikely that MEG will ever be routinely available. These and other quantitative analyses are complementing traditional visual assessment of EEG and other physiologic data and providing additional information that is clinically useful in selected situations. Long-term monitoring (Chapters 92 and 93) with simultaneous collection of EEG data and videotaped behavior has allowed improved characterization of seizure semiology and definitive separation of epileptic from nonepileptic behaviors. It has also opened the possibility of surgical intervention to far greater numbers of patients. LeVan and Gottman (Chapter 95) summarize the essential role played by computers in data reduction and analysis of physiologic data obtained during both routine EEG and long-term monitoring.