JLE

Epileptic Disorders

MENU

Epilepsy and epileptic disorders, an epidemiological marker? Contribution of descriptive epidemiology Volume 4, numéro 1, March 2002

The epidemiological studies have provided a much clearer understanding of the frequency, worldwide distribution and prognosis of seizures and epilepsy in the general population. The frequency of a disease is defined by the incidence, prevalence and mortality. Mortality will not be considered in this paper because it has already been analysed in a previous paper in this journal [1].

The first major epidemiological study of epilepsy was performed, nearly thirty years ago, in Rochester, Minnesota, using the record linkage system of the Mayo Clinic. Kurland and Hauser, in their first paper [2], estimated the incidence and prevalence of epilepsy in a known population and demonstrated that epilepsy and seizures are one of the most frequent neurological diseases. Their work set the stage for numerous other studies throughout the world examining the frequency of seizures and epilepsy, their causes and their prognosis. The results of these studies ­ some of which will be presented in the first part of this review ­ show many discrepancies in the evaluation of incidence and prevalence rates. Apart from obvious methodological biases, the results of these studies lead us to recognize that epilepsy is universal ­ it can occur in any people ­ and is ubiquitous ­ i.e. it is observed in every country ­ but is not an equally distributed disease. This review considers this unequal distribution and looks into its causes.

Recently, specific distributions in the epidemiology of seizure disorders have become apparent. Population-based studies, in developed countries, have clearly demonstrated a drastic increase of the incidence and the prevalence in the elderly, and a drop in childhood-onset epilepsy. On the other hand, high rates have been reported frequently, but not always, in developing countries.

We would like to show that epileptic disorders could be an epidemiological marker of endemic-epidemic disease in certain countries, the occurrence and the increase of some pathologies in selected populations, or perhaps, the result of preventive measures.

Definitions are essential

During the past 20 years, a large number of original papers and reviews concerning the incidence and prevalence of epileptic disorders and epilepsies have been published [3-11]. Some of these studies are not devoid of methodological problems, because approaches to the study of the epidemiology of epilepsy have become increasingly sophisticated. The improved methods include correct diagnosis, the development of clear definition of the condition studied and the use of correct classifications. These definitions and classifications have been proposed by the ILAE Commission of epidemiology and prognosis of epilepsies [12] and are used in this article (see table: definitions).

Incidence and prevalence studies

Only those studies that contain standardized definitions of the seizure and that are age-adjusted were reviewed. Adjustment for age and age-and-sex, using the direct method, remove the possible influence of the age structure of the study population and increase the comparability of the different studies; the reference population is usually the 1980 USA population.

Incidence studies

Incidence or incidence number is defined as the number of new cases of seizures or epilepsy occurring during a given time interval, usually one year, in a well defined population. Criteria for defining an incident case must be clearly stated, including specification of whether it is based on date of diagnosis (index seizure) or date of onset. Incidence rate is the ratio of observed cases to population at risk, usually expressed as cases per 100 000 persons/year.

Incidence studies must be divided into those that evaluate recurrent unprovoked seizures (epilepsy) and those that include first seizures and/or acute symptomatic seizures and/or isolated seizures.

Incidence studies of first epileptic seizures (table I)

Age-adjusted incidence rates of first epileptic seizure range from 18.9 to 190 [13-22]. In these studies, both provoked and non-provoked seizures and sometimes single or isolated seizures, are usually considered. Such studies are difficult to compare because inclusion criteria are often different. Only two surveys [17, 21] using the same methodology ­ including acute symptomatic seizures, recurrent unprovoked seizures and single seizures ­ produced similar figures, respectively 69.5 and 69.4 per 100 000 inhabitants. Combining incidence rates for single or isolated convulsions and recurrent seizures, from the Rochester study, the incidence rate for all seizures in Rochester has been estimated between 72.9 and 86.1 [19]. In Ecuador, Placencia et al. [18], using a rigorous methodology [23], found a much higher incidence rate of first, non-febrile seizures (incidence: 120-190) than in Western countries. In Martinique, an intermediate rate has been reported [22]. The incidence rate for only acute symptomatic seizures is difficult to obtain because they are often managed more by general practitioners than by neurologists. In Rochester, the rate has been estimated as only 25.5 [24].

Incidence of epilepsy (table II)

The incidence rate of unprovoked seizures is quite similar to that of epilepsy. Included are patients with two or more seizures, but also patients with a single seizure and an abnormal EEG, allowing classification of this seizure within a well described epileptic syndrome.

In developed countries, the age-adjusted incidence of epilepsy ranges from 24 to 53 /100 000 person-years and seems remarkably consistent across geographical area [25-32]. For instance, similar rates have been reported in Rochester, Minnesota [19], Faroe Islands [29] and Iceland [31]. This is particularly true of reports from studies performed over the last twenty years.

Few incidence studies have been carried out in developing countries. The reports by Placencia et al. [18], Lavados et al. [33], and Rwiza et al. [34] have shown higher incidence rates: 113, 122-190 and 73.3 respectively (table III). Two studies carried out in rural areas in Ethiopia [35] and India [36] have reported lower rates, 64 and 49.3 respectively.

Cumulative incidence

Cumulative incidence represents the probability of an individual's risk of developing epilepsy. Two studies have assessed the risk of seizure developing in some age groups, using hospital and EEG records: Juul-Jensen and Folspang [15] found that there was a 1.7% risk of developing afebrile seizures at the age of 80 in Denmark. Cumulative incidence, calculated for provoked and unprovoked seizures, was 5.8 in Geneva county [21] and in Martinique [22]. Hauser et al. [19] noticed that the cumulative incidence of epilepsy was 3.1% by the age of 80. The discrepancy between life-time prevalence and cumulative incidence makes lifetime prevalence a poor indicator of true cumulative incidence.

Prevalence studies

Prevalence is more a reflection of severity or chronicity of epilepsy than its frequency. It is a measure of the interaction of many factors such as incidence, death and remission of the disease with or without sources of medical care and migration of populations. Inclusion criteria should be specified i.e.: active epilepsy, epilepsy in remission with treatment, epilepsy in remission without treatment.

Prevalence rate represents the ratio of identified cases to the total population, usually expressed as cases per 1 000 persons.

Studies of the prevalence of epilepsy may assess different aspects:

­ The lifetime prevalence is the number of individuals who have experienced an epileptic seizure under the study, regardless of treatment or recent seizure activity. It includes patients with a history of active epilepsy or epilepsy in remission. It assumes that once a person has epilepsy, he or she will always be considered affected.

­ The point prevalence describes the number of patients with "active" epilepsy, at a specified time ­ usually a specific day (prevalence day). The point prevalence represents the burden of the disease in a defined population and determines the need for health care and prevention.

­ The period prevalence is the proportion of patients with epilepsy in a given population during a defined time interval ­ usually one year.

Lifetime prevalence studies (table IV)

Overall lifetime prevalence ranges from 3.5 in Norway [27] to 10.7 in Copiah County (USA) [40]. These figures represent only recurrent unprovoked seizures, except in the report from Guam [66] where febrile seizures have been included explaining the much higher rate. Secular trends between 1950 and 1980 in Rochester show a growing rate from 5 to 8.2 [42]. This is probably related to better diagnosis.

Two studies carried out in developing countries have shown similar rates to those observed in the developed world [39, 41]. However, using the capture-recapture method, Debrock et al., have estimated the prevalence of epilepsy as 33.5 and 35.1 per 1000 in two villages (3 134 inhabitants) in the Benin region of Zinvié [43].

Point prevalence studies (tables V and VI)

The point prevalence rate in developed countries, ranges from 3.7 [44] to 8 [29]. Similar features have been found in some developing countries (table VI) but in others, higher rates, ranging from 37 [51] up to 57 [49] were noted. Some of these studies have reported differential rates for urban and rural areas, usually with higher rates in the rural areas.

Epilepsy and epileptic disorders have a specific distribution

Age-specific distribution of seizures and epilepsy. The relative age-specific incidence rates of epilepsy and first seizures, in developed countries, are very similar in all studies and show a U-shaped pattern (figure 1).

Age-specific incidence of epilepsy and unprovoked seizures is consistently high in the youngest groups, decreases after the first years of life and falls again during adolescence. The incidence rate remains low during the adult years. All recent incidence studies in developed countries have shown an increasing incidence in the elderly, reporting incidence rates higher than in children. This bimodal distribution of epilepsy is observed only in industrialized countries.

Many incidence studies have been conducted specifically in children, excluding neonates, with different case diagnosis methods with a mean rate about 70-80 [review, ref. 11]. In many studies age-specific estimates are quite difficult to compare because of small numbers of cases within each age group. In the Rochester study [42], active prevalence increases in each subsequent age group, with the highest rate occurring in the elderly [for review see ref. 11]. In children, lifetime prevalence ranges from 2.75 [42] to 14 [46]. The lifetime prevalence for males is 1.7 or 2.5 times higher than among females. Point prevalence rates range from 2.7 [42] to 5.5 [57], depending upon the definition of the disease, the age of the children, the type and the location of the study. This increased prevalence of epilepsy among males concurs with trends in the incidence of epilepsy and first seizures.

Studies from European countries [29, 30, 45] report a constant prevalence in adults.

Studies from developing countries [18, 33, 41, 48-54] report the highest prevalence in the second and third decade of life, with a lower prevalence rate in the elderly.

Sex-specific incidence and prevalence rates

Most reports show consistently higher rates of first seizures or/and epilepsies in males than in females (sex ratio M/F: 1.2 to 1.7) suggesting that males are at greater risk than females of non-provoked seizures and epilepsy. However, for most but not all incidence studies, sex specific differences in incidence are not statistically significant. This gender difference is not explained.

Epilepsy in developing countries

The burden of epilepsy in developing countries has become obvious as nearly 75% of 50 million patients with epilepsy live in these countries and diagnosis and therapeutic facilities are poor. A large proportion of patients with epilepsy do not get treatment (treatment gap) because drugs are not available or may not be taken for cultural, political and socioeconomic reasons [55]. Neuroimaging facilities may be available in some countries, but patients may be too poor to afford them.

There is no known scientific reason that the clinical aspects and natural evolution of the disease between developed and developing countries should be different. However, epidemiological studies from developing countries show some controversial results.

Incidence has been reported as high as more than 100/100 000 [18, 33], except in a recent study in India which reports a similar rate to that found in Rochester [36]. The most comprehensive study of incidence and prevalence in a developing country was carried out in Ecuador [18]. A validated questionnaire [23], with a specificity of 93% and a sensitivity of 79% was used. The results of this study are the most accurate obtained from the developing world. The incidence was: 190 /100 000, the lifetime prevalence: 14.3 and the prevalence of active epilepsy, 8/100 000.

A much higher prevalence rate of epilepsy in developing than developed countries, had been reported in some early studies. The prevalence of epilepsy is high in tropical countries, particularly in Africa where it varies between 10 and 55 per 1 000, with an estimated mean prevalence of 15/1 000. Similar figures have been reported in Latin America. These high prevalence rates are probably related to genetic or environmental factors. Studies of prevalence in some minority populations are complicated by socio-economic factors. In Ecuador, prevalence was inversely correlated with community ranking by socio-economic class [18]. The prevalence of epilepsy in Pakistan has been reported to be greater in rural than urban communities [53]. Wide disparities ­ from 2.5 to 57/1 000 ­ have been reported, without any clear explanations. However, recent studies shows trends towards similar rates as those observed in developed countries [40, 50-52, 54].

Numerous methodological limitations have been identified and have raised some questions about the accuracy of the results. They include case definition, case diagnosis, classification and selection bias. In the majority of epidemiological studies, the WHO questionnaire was used. It is quite impossible with this questionnaire to provide good classification of seizure type or syndromic classification. Age-specific rates are difficult to compare with those from developed countries, given the differences in age structure of the populations.

Race

There are few studies in which racial differences can be compared. These studies are too often complicated by socio-economic factors. After adjusting for race, incidence of epilepsy is frequently higher in lower socio-economic classes.

A study of children in the New Haven area of Connecticut, reported an incidence rate of epilepsy up to age 15, of 1.7 times greater in blacks than in whites [56].

Preliminary reports from studies of inner-city black communities provide similar prevalence rates to that in other communities in the United States [19, 42, 58]. However, age-specific prevalence for epilepsy was higher in blacks than in whites [40, 57] and Hispanics, [58] especially in adult populations. De Lorenzo [59] has reported a much higher incidence rate of status epilepticus in blacks than in whites in Richmond, Virginia.

Seizure type

The ILAE 1981 classification of epileptic seizures [60], is based on observed clinical symptoms and, when available, on the EEG. The use of the EEG in field studies is often impractical and the classification of seizure type is quite impossible. The EEG contributes to, but does not always confirm the diagnosis of epilepsy.

Many questionnaires have been designed specifically to recognize generalized tonic clonic seizures or absences. Partial seizures are often underreported. Many so-called generalized seizures are probably secondary generalized partial seizures. The categorization depends on the investigator's scale and/or the sophistication of the investigations (EEG, CT scan, MRI). The reported classification should thus be interpreted with caution.

The majority of patients seem to have generalized seizures, and rates as high as 93% have been reported [52]. Once the international classification has been used, the result varies over a small range: 40 to 60% of epileptic seizures have a generalized onset; 32 to 52% are partial and 10% cannot be classified [19].

The epidemiology of status epilepticus is not well known as the majority of studies focused on hospital-based populations, providing biased information. Three population-based studies in controlled community settings are those of De Lorenzo et al. [59] in Richmond (Virginia), another by Hersdorffer et al. [61] in Rochester (Minn) and by Jallon et al. in Geneva [62]. The incidence rate was 41 per 100 000 in Richmond, 18.3 in Rochester and only 16.5 in Geneva. The age distribution is quite similar to that of first seizure incidence.

The international classification of epileptic syndromes [63] is clearly useful for the evaluation of the disease and the impact of a therapeutic strategy, but is rather difficult to use in epidemiological surveys, because it requires the use of more or less sophisticated investigations. In a prevalence survey conducted in children younger than 10 years living in the Okayama prefecture, Japan [64], 79% of cases were properly classified. In contrast, in an NGPSE study [65], after excluding patients with special syndromes, only 34% of cases could be classified.

Only two studies [32, 66] reporting the incidence of all epileptic syndromes can be compared. The incidence by major syndromic groups was respectively as follows: idiopathic localisation related: 1.7 and 0.2, symptomatic localization related: 13.6 and 17.2; idiopathic generalized: 5.6 and 3.7; symptomatic generalized: 1.1 and 1.7 and undetermined: 1.9 and 9.7. The lower rate of the undetermined type in the French study [66] is probably related to the methods of diagnosis.

There are numerous reports on the incidence of specific syndromes (table VII).

Neonatal convulsions have an incidence of one to 3.5 per 1 000 of live births [67-69]. The diagnosis is rather difficult and they are usually not included in community surveys of epilepsy.

There are marked variations worldwide in the cumulative incidence of febrile convulsions. The reported prevalence in children under 5 years is between 2 and 4% in the USA and Europe [87]. In some countries, much higher rates have been reported: Chile: 5.2% [88], Japan: 8% [78], Mariana Islands: 14% [89]. Lower cumulative incidence rates have been reported in China [90] and in some populations of Central or South America [33].

Etiology

The etiology of epilepsies varies largely with the type of study, the age and the geographical area. However, from an epidemiological standpoint, a causal association between an event and an outcome includes a number of characteristics such as time order, biological plausibility and consistency of the observation [91]. Risk factors for seizures and epilepsy are quantified by case-control studies and prospective cohort studies. The key factor in these studies is the comparison group.

In most investigations, a specific cause of seizures or epilepsy has been found in about 1/4 to 1/3 of all cases [10]. A number of factors are considered to be risk factor for seizures and epilepsy and are discussed in most textbooks [91].

The results are usually presented as a percentage of frequency of different causes. Hauser et al. [42] classified 76% of prevalent cases as idiopathic and/or cryptogenic, 6% due to cerebral vascular diseases, 5% to peri-natal neurological deficits at birth, 5% to trauma, 4% to infection, 2% to tumors, and the remainder to other causes. A review of available studies of the etiology of epilepsy in developing countries, shows that no cause can be found in more than half of the cases. The most prominent of the etiological factors that have been usually considered, are infections of the central nervous system ­ of which a major component appears to be cysticercosis and maybe onchocercosis and HIV infection ­ perinatal damage, family history and head injury. The contribution of vascular diseases and tumor appear to be of minor importance. In Ecuador [18], no putative cause for epilepsy was identified in 73 % of the patients. The likely etiologies included birth trauma (9%), head injury (7%), neurocysticercosis (3%) and cerebro-vascular diseases (3%). The exact extent of this contribution remains unclear and will be discussed below.

The Commission of Epidemiology and Prognosis of the ILAE has proposed a risk factors classification [12]. Appropriate categorization of individual cases may require the use of "state of the art" technologies and procedures. Unfortunately this classification is not yet currently used. In the Geneva county, Jallon et al. [21], using this classification, found, of 64.5% unprovoked seizures : 25.2% were related to a stable neurological condition, 9.9% to a progressive pathology and 29.3% were of unknown etiology (22.7% cryptogenic; 6.6% idiopathic).

DEFINITIONS

Concerning the diagnosis of epileptic disorders

* An epileptic seizure is a clinical event, presumed to result from an abnormal, paroxysmal and excessive discharge of a set of neurons in the brain. The clinical manifestation consists of sudden abnormal and transitory phenomena in relation to the cortical zones affected by the discharge. Such changes include distortion of awareness, motor, sensory, autonomic or mental events, perceived by the patient or an observer, with or without loss of consciousness.

* Epilepsy is defined as a condition characterized by two or more recurrent epileptic seizures unprovoked by any immediately identified cause.

* A single or isolated seizure is the term used when one or more seizures occur in a period not exceeding 24 hours. Some seizures remain isolated as a result of the nature of the disease; others are inhibited by the initiation of treatment. Without treatment, these patients might otherwise have subsequent seizures, but they cannot be considered as having epilepsy because they have had only one seizure.

* Status epilepticus is defined by the occurrence of either an isolated epileptic seizure lasting more than 30 minutes or a succession of epileptic seizures with no recuperation of cortical function during an interval of at least 30 minutes. An episode of status epilepticus should be considered as a single event.

* "Active" epilepsy is defined as epilepsy in a patient who has had at least one epileptic seizure in the previous five years from the time of diagnosis, regardless of antiepileptic drug treatment. "Active", as defined by the ILAE Commission of epidemiology and prognosis, means that only people with recurrent seizures in the last five years prior to the study or on antiepileptic drugs, have been included. In the past, some authors have restricted the definition of active epilepsy to recurrent seizures in a period shorter than five years, e.g., three or even two years.

* Febrile seizure is an epileptic seizure, as defined above, that occurs in childhood after the age of one month, associated with febrile illness, not caused by a CNS infection, without prior neonatal seizure history, or history of previous unprovoked seizure or an acute symptomatic seizure.

* Neonatal seizures are epileptic seizures as defined above, that occur in the first four weeks of life.

* Non-epileptic seizures are clinical behaviours that are assumed to be unrelated to an abnormal discharge of a set of cortical neurons; they include disturbances of brain function such as vertigo, syncope, abnormal movements, nocturnal episodes, transient global amnesia, migraine, or enuresis.

* Non-epileptic psychogenic seizures are sudden non-epileptic behavioural disturbance presumed to be of psychogenic origin.

Concerning the etiology of seizures or epilepsy

Provoked seizures are also called acute symptomatic seizures or seizures related to a specific situation. They occur in close temporal association with an acute systemic, metabolic or toxic insult (infection, tumor, stroke, cranial trauma, intracerebral hemorrhage or acute intoxication or alcohol withdrawal). They are often isolated epileptic events, but they may also be recurrent or even evolve into a status epilepticus when the acute condition recurs, for instance in alcohol ­ or drug ­ withdrawal seizures.

Unprovoked seizures are those presumably symptomatic but without an acute precipitating insult. Two major groups have been identified:

Remote symptomatic unprovoked seizures occur in relation to a well-demonstrated pre-existing condition, substantially increasing the risk for epileptic seizures. Two subgroups have been distinguished: the first as a result of conditions resulting in a static brain disorder or condition. Such cases occur in individuals with epilepsy related to a pre-existing insult to the CNS, such as infection, head injury, stroke, which results in a static lesion. The second as a result of progressive CNS disorders (low grade tumour, infection, slow virus infections, HIV infection, parasitosis, autoimmune disease, identified metabolic disease, neurodegenerative disease).

Unprovoked seizures of unknown etiology. In these cases, no clear antecedent can be detected. Distinction is made, in line with the ILAE Commission of Terminology [39], between idiopathic and cryptogenic epilepsies or seizures.

* Idiopathic epilepsy includes some partial or generalized epileptic syndromes with particular clinical characteristics and with specific EEG findings. They are seen in normal patients without cerebral abnormalities. The presence of epilepsy is believed to be due to a lowered seizure threshold, with perhaps a genetic basis, whether a family history of seizures is present or not.

* Cryptogenic epilepsy is the term reserved for partial or generalized epilepsy not satisfying criteria for idiopathic, epilepsy, and in which neither the history, clinical examination, nor ancillary investigations reveal a lesion accounting for the seizure. Cryptogenic epilepsies are first based on negative criteria whereas symptomatic and idiopathic epilepsies are defined by positive criteria.

Received January 10, 2002

CONCLUSION

Epidemiology of epilepsy and seizures disorders is well known

The epidemiological study of epilepsy remains difficult for a number of reasons. The most important reason is that epilepsy is sometimes unrecognized and is often concealed by the patient or his/her family. Even when suspected, epilepsy is not so easy to diagnose. The medical assessment and knowledge of the disease vary from one country to another in relation to the health infrastructure of the country in which the survey has been carried out. In this way, the etiology of the seizures is often not very well elucidated because it needs a minimum of investigations that were not available. This specific aspect of the epidemiology of epilepsy is a cause of conflicting results. Many epidemiological studies on epilepsy have been published, but lack of rigorous definitions of the disease, differences in the methods of case diagnosis, classification of seizures and/or epileptic syndromes and the evaluation of risk factors impedes meaningful comparisons.

However, valid information has been provided by many studies in a diversity of settings. A number of problems ­ in large part due to methodological problems ­ qualify the interpretation of the results. There is a clear lack of well conducted incidence studies in developing countries. Population ­ based and prospective studies must be encouraged to assess the geographical differences and the contribution of some risk factors to the geographical distribution of incidence. On the other hand, there is a plethora of prevalence studies in developing countries, and there is poor justification for further studies of prevalence of epilepsy apart from the research of a definite risk factor in a given area. The distribution of epilepsy and epileptic seizures, following the ILAE classification of epileptic syndrome and risk factors, must be encouraged.

Seizures and epilepsy could be considered as epidemiological markers

The current data on epidemiology of epilepsy lead us to recognize that the disease is universal (i.e. it can occur in any people) and ubiquitous (i.e. it is observed in every country), but is not equally distributed.

The incidence rate for epilepsy or unprovoked seizures in industrialized countries ranges from 20 to 70/100 000. Information available from developing countries, suggests that the incidence may be higher than 100/100 000. The difference in prevalence rates could be epidemiologically significant. Particular regional distribution of the disease could reflect local risk factors such as infectious diseases (parasitosis), post-natal conditions, peri-natal and lifelong cranial trauma.

There is little doubt that neurocysticercosis is one of the most frequent cause of seizures and epilepsy in adults in some tropical and subtropical zones. Seizures are the most common symptom in 70-90% of patients with neurocysticercosis [92]. Many studies from Latin America [92-95], India [96] and some areas in Africa [97], have shown that infection by the larvae of the pork tapeworm Taenia solium is an important cause of epileptic seizures in endemic communities. A Mexican study reported that neurocysticercosis accounted for 50% of cases of late onset epilepsy [94]. Poor hygiene and living conditions allowing pigs access to human faeces add to the risk factors. The disease is more frequent in the countryside. In New Guinea, introduction of pigs resulted in cysticercosis, which in turn resulted in a so-called "epidemic of burns" [98, 99].

The occurrence of epileptic seizures during onchocercal infestation has been suspected, but epidemiological studies are necessary to confirm the relation between onchocerciasis and epilepsy [100]. Two matched studies conducted in the Central African Republic [101] and Mali [102] found some relationship, although it was not statistically significant. Various risk factors including genetic factors and socio-economic factors could explain the trend toward a higher incidence of epileptic seizures in zones of high endemicity of onchocerciasis.

Seizures ­ and especially febrile convulsions in children ­ could occur in the acute stage of Plasmodium falciparum infestation. The issue of unprovoked seizures following cerebral malaria remains controversial.

The age distribution in developed countries clearly shows that the disease is more frequent in children and in the elderly. This specific distribution has not been observed in developing countries. Secular trends demonstrate that the incidence of epilepsy in children, in developed countries, is decreasing [103] whereas it is increasing drastically in the elderly [104]. Two studies involving only children, have shown a decrease in the incidence with time. In Västerbotten, Sweden, an initial study carried out in 1975 [105] had shown a rate of 134/100 000; twenty years later, a second study [106] reported a rate of 88.8. In the Rochester study [19], between 1954 and 1984, the age - specific incidence of epilepsy has varied from 136 to 79 for 0-1 year of age, from 95 to 53 for 1-4 years and from 55 to 52 for 5-9 years. In the majority of the studies, about 50% of cases of epilepsy started in childhood and adolescence, although in the English cohort [107], only 25% of the patients had their first seizure before the age of 15.

In the elderly, four studies have clearly demonstrated an increasing rate, higher than that observed in the first decade (more than 150-200 /100 000) [17-22]. The incidence rate of unprovoked seizures in the elderly persons reported in Rochester increases from 63.4 to 111 between 1935 and 1984, although the average incidence rate for all age groups has remained more or less constant at about 45 per 100 000 [19]. The incidence for men was always higher (97.3 to 128.5) than that for women (37.4 to 106). The probability that an 80 year-old person has had or will develop epilepsy in the course of his/her life is about 3.5% while that for a 20 year-old person is only 1%.

Several factors may be involved in this pattern, such as antenatal and prenatal care for children and the increase in life expectancy, the risk factors linked with age such as stroke, dementia, cerebral tumors and the medical facilities offered to the elderly [108].

These epidemiological features led us to think that seizures and epilepsy could represent epidemiological markers of early intervention during management of childbirth and childhood development, and some specific age-related or geographical risk factors.