The seizures observed in this age-group are difficult to classify because of their atypical features and, in particular, because the state of awareness and responsiveness often cannot be properly assessed (Watanabe et al., 2002; Kramer, 1999). They are often expressed by uncharacteristic features, which makes the distinction between focal and generalized seizures difficult.

Generalized tonic-clonic seizures are rare (Brunon et al., 1978; Sillanpää, 1973), and a detailed analysis of those attacks that are apparently generalized often reveals a lack of synchrony and symmetry of the motor phenomena that may predominate successively in different parts of the body, thus giving the impression that its components are more or less independent (Dalla Bernardina et al., 1978a). The generalized seizures also tend to be purely or predominantly tonic or clonic, rather than consisting of a regular tonic-clonic sequence. A unilateral predominance in the amplitude of the ictal electroencephalogram (EEG) discharge and irregularities in their rhythm are common, and their hemispheric synchronization is often imperfect (Gastaut et al., 1974c; Gastaut and Broughton, 1972).

Partial seizures are much more common (Cavazzuti et al., 1984; Aicardi and Chevrie, 1983; Dalla Bernardina, 1978a), but, often, their expression is less clearly indicative of a localized discharge than that in older children. A relatively common type is marked by head and/or eye deviation toward one side and is accompanied by a predominantly tonic contraction involving only half the body or, more commonly, both sides symmetrically or asymmetrically (Watanabe and Okumara, 2000; Duchowny, 1987). These authors termed the attacks “complex partial seizures of infancy”; they were manifested by head turning; tonic, often lateralized motor phenomena; and apparent loss of contact. However, such features are often encountered in early seizures with or without alteration of consciousness, and many seizures are limited to hypertonia or hypotonia; mild convulsive movements; simple staring; and autonomic manifestations, such as flushing of the face, pallor, perioral cyanosis, or disturbances of respiratory or cardiac rhythms. Such seizures are classified as “hypomotor seizures” by some investigators (Hamer et al., 1999; Acharya et al., 1997) (see Chapter 10). Precise video-EEG study in 32 infants (Rathgeb et al., 1998) confirmed that the main features were motor phenomena that involved mostly the head, face, and trunk; arrest of activity; and autonomic symptoms. The degree of awareness is often impossible to determine. Such atypical seizures may alternate with or may evolve into more characteristic attacks as the infants grow older; in addition, lesions of the temporal lobe that initially are atypical in expression have been shown to develop later into typical temporal lobe seizures (Aicardi, 1970). Because of the difficulty in assessing the level of consciousness, the distinction between simple and complex partial seizures as defined by the International League Against Epilepsy (ILAE) classification of 1985 (Commission on Classification and Terminology of the ILAE, 1985) is not valid in this age group.

The EEG patterns associated with atypical partial seizures are quite variable (Aicardi, 1994a) in morphology and localization. In some patients, they may not be detectable on the scalp. Focal discharges have been recorded in infants with bilateral motor phenomena or diffuse autonomic manifestations (Aicardi and Chevrie, 1982b; Dalla Bernardina et al., 1978a), thereby indicating that partial seizures can easily be
mistaken for generalized ones; differentiation may be impossible without simultaneous EEG recordings. Motor phenomena tend to be associated with anterior EEG discharges, and arrest and autonomic features with posterior discharges.

The relative lack of organization of the infantile seizures probably results from the incomplete development of the central nervous system (CNS), with the variable excitability being caused by changes in synaptic organization and the maturation of neurotransmitters, as well as by imperfect myelination and function of the interhemispheric connections (Holmes and Stafström, 1997; Moshé, 1987). However, the infant brain can occasionally produce massive epileptic myoclonus with bilateral, irregular spike-waves and, very rarely, even atypical absences (Aicardi, 1995).

The etiology of infantile seizures is dominated by the presence of brain lesions (Aicardi and Chevrie, 1982b). The nature, localization, and extent of organic brain damage partly determine the ictal symptomatology. For example, tuberous sclerosis is not commonly responsible for seizure types other than infantile spasms in the first year of life (Pampiglione and Harden, 1973). Most of lesions responsible for early infantile epilepsy are prenatal in origin, and they are often quite extensive, even when they give rise to focal seizures (Chevrie and Aicardi, 1977). However, a close relationship exists between the lateralization of the lesions and that of seizures when both are unilateral or predominantly unilateral. The responsible lesions include brain damage of prenatal and perinatal origin, brain malformations, and diffuse or localized cortical dysplasias. Gross malformations were found in 57 patients in a large series in which 175 infants were thought to have acquired brain damage of prenatal or perinatal origin (Chevrie and Aicardi, 1977). The overall incidence of brain malformations would probably be higher with modern neuroimaging, which was not available in this series. Magnetic resonance imaging (MRI) has shown that cortical dysplasias and other developmental abnormalities are among the most common causes of infantile epilepsy (Guerrini et al., 1996c; Kuzniecky et al., 1992a; Palmini et al., 1991d; Andermann and Straszak, 1982) (see Chapter 19).

From a neuropathologic standpoint, the damage caused by prenatal and perinatal factors is mainly represented by ulegyrias and clastic lesions, such as the porencephalies. However, gyral abnormalities can be induced by prenatal factors such as anoxia and infections (e.g., cytomegalovirus).

The risk of early epilepsy is significantly increased following an abnormal birth. It is 7.2% even after low forceps delivery, and it reaches 11.3% following midcavity forceps. It is also increased by uterine dysfunction during labor and instances in which the lowest recorded fetal heart rate drops below 60 beats per minute. Conversely, the induction of labor, abnormalities of the cord, polyhydramnios, breech delivery, and abruptio placentae had no predictive value (Nelson and Ellenberg, 1986).

Metabolic causes of infantile seizures are uncommon. Hypocalcemia is unusual beyond the neonatal period (Oki et al., 1991), but this should be systematically investigated. Pyridoxine dependency, however, should always be kept in mind, even in those patients past the neonatal period. Recent reports (Baxter and Aicardi, 1999; Mikati et al., 1991; Goutières and Aicardi, 1985; Bankier et al., 1983; Stephenson and Byrne, 1983) have noted that the disorder may manifest at up to 18 to 24 and even 36 months of age with any type of seizures, including infantile spasms (Aicardi, 1999b). A test of the therapy for this condition (injection of 50 to 200 mg of pyridoxine) is the only way of establishing the diagnosis; therefore, this should be systematically performed.

Other metabolic causes include biotinidase deficiency, which can be easily diagnosed and treated by the administration of 10 to 20 mg of biotin, and neuroglycopenia resulting from a deficiency of the glucose transporter GLUT-1 (De Vivo et al., 1995), which is also treatable by the ketogenic diet. The latter justifies lumbar puncture to demonstrate low cerebrospinal fluid (CSF) sugar even when no cause is found for the seizures as the blood sugar level is normal.

In a large series by Chevrie and Aicardi (1977) and one by Cavazzuti et al. (1984), symptomatic cases were more common than cryptogenic (or idiopathic) ones. This was especially applicable to infantile spasms, in which 60% of cases were thought to be symptomatic and 40% were considered cryptogenic. However, the significance of these figures is open to discussion because the more severe symptomatic cases may be selectively included in such series, which originated from third-referral centers. On the other hand, they were collected at a time when modern neuroimaging was not available, and a number of CNS lesions could have been missed.

More recent work (Abe et al., 2000; Watanabe and Okumura, 2000; Vigevano et al., 1994) suggests that a significant proportion of infantile seizures may not be due to brain lesions and that, thus, they may have
a more favorable prognosis. At least some such cases appear to be familial or at least to occur in patients with a strong family history of seizures of various types, thereby supporting the role of genetic factors in the origin of epileptic phenomena in the first 2 years of life. More than half the patients studied by Watanabe et al. (1990) had a family history of epilepsy or febrile convulsions, and “benign infantile convulsions” (Vigevano et al., 1992a, 1992b; Malafosse, 1990) appear to be dominantly inherited. Likewise, genetic factors are probably important in cases of epilepsy not manifested by infantile spasms. In one large series (Chevrie and Aicardi, 1977), a positive family history of seizures was present in 28% of infants with uncharacteristic seizures versus in 16% of those with infantile spasms, and the proportion of positive family history was higher in those with apparently generalized seizures than in those with partial seizures, especially when only nonsymptomatic cases were considered.

Thus, genetic factors likely are more important in the etiology of infantile seizures than was previously thought. Such factors may be of major importance in some syndromes of early epilepsy, some of which are apparently of monogenic origin (Hirose, 2002; Bate and Gardiner, 1999), or they may represent a significant factor in cases with multifactorial inheritance, in which several genes and environmental factors are involved.

In several large series of infantile seizures published in the late 1970s and early 1980s (Cavazzuti et al., 1984; Chevrie and Aicardi, 1971, 1975, 1977), classification was limited to broad categories defined according to the dominant types of seizures. Only a limited number of epilepsy syndromes were recognized, especially West syndrome (or the presence of infantile spasms), with other cases being defined by the occurrence of uncharacteristic, (apparently) generalized or partial seizures.

In a study of 437 cases of chronic seizure disorders beginning before the age of 1 year, Chevrie and Aicardi (1977) classified 230 of them as “infantile spasms,” whereas 207 could not be ascribed to specific syndromes. A few of the 207 cases were probably myoclonic epilepsies or an early form of LGS, but the vast majority could be divided only into partial (57 cases) and generalized (99 cases) seizures; 51 patients with seizures lasting 30 minutes or more were classified as “status epilepticus.” Cavazzuti et al. (1984) similarly divided their patients into the following three groups: infantile spasms (183 cases), status epilepticus (66 cases), and “others” (138 cases). The last group included 51 infants with partial seizures and 87 with generalized seizures.

In a study of 504 children with epilepsy who had their first seizure between the ages of 28 days and 36 months, Dalla Bernardina et al. (1982) used a more complex classification. They ascribed 163 patients to the group “epileptic encephalopathies,” which included West syndrome and LGS, with several subdivisions. They classified 189 as “partial epilepsies,” also with various subgroups, and 80 as “generalized epilepsies,” 43% of which were “myoclonic epilepsies.” Only 72 cases remained unclassified. Only 37 (8%) of the patients in this series had a generalized epilepsy of a type other than myoclonic. Because the upper age limit was set at 3 years, more precise classification might have been possible than in the two preceding studies, which were limited to infants younger than 1 year. Clearly, a large proportion of the cases of LGS or myoclonic epilepsy cannot be diagnosed before 1 year of age, and this also applies to most of the subtypes of partial epilepsies reported by Dalla Bernardina et al. (1982). Though complex partial seizures are difficult to identify because of the problem of ascertaining the level of consciousness of infants ictally, these have been increasingly recognized (Watanabe et al., 1990; Duchowny et al., 1988; Dalla Bernardina et al., 1982). Most are probably associated with organic brain damage (Duchowny, 1992; Duchowny et al., 1988), and they have a poor prognosis that likely is related to the diffuse distribution of the lesions.

More recent studies have emphasized the existence of recognizable epilepsy syndromes in infants younger than 2 years. However, the definition of syndromes in this age group may be difficult because of the rapidly changing state of brain maturation, which is responsible for changes in seizure types and the associated features in the same patient. Thus, an epilepsy that, at 6 months, is expressed by infantile spasms can be expressed a few months later by atonic falls and/or atypical absences. More commonly, early epilepsies are initially manifested by atypical, abortive, or unclassifiable seizures that are later replaced by fits that are more specific for the accepted epilepsy syndromes, such as LGS. As a result, categorization of a syndrome is often impossible for a long period, consequently leading to prognostic uncertainties.

Some of the epilepsy syndromes that can occur in the first 2 years of life may be early forms (or the initial manifestations) of classic syndromes that also exist
in older children, and these are described in other chapters of this book (see Chapters 3, 10, and 17). However, as the chapter indicated earlier, LGS often becomes characteristic only after 2 years of age and very seldom in the first year of life. The diagnosis of severe myoclonic epilepsy (Dravet syndrome) can be suspected in infants on the early occurrence of convulsions that are often prolonged and unilateral and that are triggered by mild fevers and then recur within a short period of time (usually less than 2 to 3 months). Confirmation of the diagnosis, however, has to await the appearance of myoclonias and/or other “minor” seizures in the second or third year of life. The diagnosis of “benign” myoclonic epilepsy rests, in principle, on the exclusive occurrence of brief isolated myoclonic jerks without other types of seizures, except febrile convulsions, but a definitive confirmation is possible only later with the appearance of no other types of paroxysms. A recessive form of infantile myoclonic epilepsy with seizures that may persist into adulthood has been reported in an Italian family and has been mapped to chromosome 16p13 (Zara et al., 2000).