Prevalence and incidence of epilepsy in the Nordic countries

Method

We have undertaken a review of original articles based on searches in PubMed up to and including 1 January 2015. We searched for the terms epilepsy and epidemiology with the aid of the Boolean operator AND in combination with each of the Nordic countries separately. The search returned 141 hits for Norway, 262 for Sweden, 199 for Finland (including the Åland Islands, for which there were no hits), 213 for Denmark (including six for Greenland and one for the Faroe Islands) and 29 for Iceland.

The search results were reviewed in light of the title and abstract. In total, we found 38 original articles in English, Norwegian or Danish in which prevalence and/or incidence rates were reported. We found no articles in Finnish, Icelandic or Swedish.

An additional search with the terms «epilepsy AND (incidence OR prevalence)» for each of the Nordic countries returned 163 hits for Norway, 296 for Sweden, 247 for Finland (including the Åland Islands), 243 for Denmark (including six for Greenland and one for the Faroe islands) and 30 for Iceland. The additional search did not identify any further relevant original articles. The search results were cross-checked against reference lists in key English-language review articles and commentaries (6 – 9).

Prevalence

Prevalence is defined as the proportion of individuals in a given population who have the diagnosis in question at a given point in time (the prevalence day). Prevalence is commonly reported as a percentage or thousandth part of the total population in the relevant area.

Nordic studies that have investigated the prevalence of epilepsy are shown in Table 1 (10 – 38). Eight of these studies include all age groups (10 – 13, 26 – 29), but only four of them were restricted to active epilepsy (10 – 13). The procedure for identification of patients was identical in the four studies of active epilepsy, with hospital-based searches and retrospective review of patient records. Unfortunately, they used varying definitions of epilepsy.

One of these studies had been conducted in Iceland more than 50 years ago. Unsurprisingly, it found a significantly lower prevalence rate (3.4/1 000) than the three others (10). This finding is in line with an older study from Northern Norway (prevalence 3.5/1 000), which included all age groups, but was not restricted to active epilepsy (26).

The three remaining studies reported prevalence rates of 4.8/1 000 (12), 6.5/1 000 (13) and 7.6/1 000 (11). The study with the lowest prevalence rate stemmed from Iceland. Here, untreated patients who had been seizure-free for more than one year were excluded (12). The highest prevalence rate was found in the Faroe Islands, despite exclusion of all patients who had been seizure-free for more than five years, irrespective of treatment. As an explanation, the author noted that the patient records had frequently been written by personnel who had no particular interest in epilepsy, and that this may have resulted in overdiagnosis. The final study stemmed from Norway. In it, active epilepsy was defined as ongoing treatment and/or at least one seizure over the last five years (13). To date, this remains the only study that has used the most recent guidelines for epidemiological epilepsy research from the International League Against Epilepsy (ILAE) and their comprehensive proposals for amendment of the terminology from 2010 (39, 40).

Eight studies investigated the prevalence of active epilepsy in children (14 – 21). The prevalence rates range from 3.2 to 5.1 per 1 000 inhabitants. When considering that the prevalence of epilepsy increases with age (12), it is unsurprising that the study reporting the highest prevalence rates did not include the youngest children (the age group 0 – 5 years) (21).

Prevalence figures for active epilepsy in adults have been reported in four Nordic studies (22 – 25), two of which stem from Norway (24, 25). We believe there is reason to assume that both of these have a high sensitivity and specificity with regard to the identification of patients, since they used population-based surveys combined with individual interviews and hospital searches/retrospective record reviews respectively. The cohorts investigated, however, were small. Their prevalence rates of active epilepsy amounted to 6.7 (24) and 8.2 per 1 000 (25) respectively. The two other Nordic studies were based on larger populations, with prevalence rates found to be 5.5 (23) and 6.3 per 1 000 (22) respectively. In the latter study, all included patients were clinically examined by the authors, and we may thus assume that it has high specificity.

Incidence

Incidence is defined as the proportion of individuals in a certain population who are diagnosed with the condition is question within a given period of time. Incidence is commonly reported as the number of new cases per 100 000 persons per year. Nordic studies that have investigated the incidence of epilepsy are shown in Table 2 (10, 11, 14, 15, 18, 22, 26 – 28, 41 – 49). Because such studies are often difficult to undertake in practice, there are fewer studies of incidence than of prevalence. To ensure the highest possible sensitivity and specificity, incidence studies ought to be prospective in nature (9). Moreover, it should be noted whether the study also includes single, unprovoked seizures or whether it is restricted to the ILAE definition of epilepsy (40, 50). This is especially relevant for incidence studies, since many of them register all those who are recorded with a first-time epileptic seizure and report this figure, without waiting to see whether the patient suffers a second seizure. The definitions of epilepsy used in the studies we are referring to in this article are summarised in Table 3 (10 – 38, 41 – 49).

Only five prospective incidence studies of epilepsy (41 – 45) have been undertaken in the Nordic countries, of which only two include all age groups. The study with the largest population basis, that included single, unprovoked seizures (42), found an annual incidence of 34/100 000. A prospective study from Iceland found an incidence of 33/100 000. Here, single unprovoked seizures had been excluded (41).

A smaller prospective study among children found an annual incidence of epilepsy of 53/100 000 (44). A prospective study of adult patients found an incidence of 56/100 000. Here, single unprovoked seizures had been included (45).

Discussion

In 2011, ILAE published new guidelines for epidemiological research on epilepsy (39), because substantially differing methodologies had made comparisons of studies difficult (6, 7). Epidemiological studies commonly define epilepsy as a minimum of two unprovoked epileptic seizures during a period of more than 24 hours (39, 51). This remained ILAE’s definition until 2014, when it was expanded to also include a single unprovoked seizure if it is part of an epilepsy syndrome or if the risk of recurring seizures is estimated to exceed 60 % (50).

Studies that include single unprovoked and/or acute symptomatic seizures will naturally result in a higher prevalence of epilepsy than studies that have deemed such seizures as not qualifying for the diagnosis.

According to the ILAE definition, active epilepsy is a condition involving ongoing treatment with antiepileptic medication and/or at least one epileptic seizure over a specified period of time, commonly the last 2 – 5 years. It is recommended that epidemiological studies restrict their reporting to active epilepsy (39, 51). However, there is still room for some discrepancy – the shorter the period that has elapsed since the last seizure, the more patients will be excluded from the group with active epilepsy. For those who have been seizure-free for more than ten years and have gone without antiepileptic medication for more than five years, the ILAE has introduced the concept of «resolved epilepsy» (50). This means that their epilepsy is in remission, not that the disease has been cured. This concept has not yet been applied in epidemiological studies.

Another cause of substantially different findings is the method used for identification of patients. No method is totally watertight, and recommending a single specific procedure is therefore difficult. Studies that use a population-based questionnaire may have limitations caused by low response rates and selection bias. Moreover, such studies may for practical reasons be restricted to smaller populations or delimited age groups, as are door-to-door surveys.

Studies based on registered diagnostic codes may provide for larger study populations, although there is a risk of overestimating the prevalence of the disease. In addition, it is difficult to restrict such studies to active epilepsy. A study from Denmark showed that approximately 20 % of all patients who were registered with an epilepsy code did not fulfil ILAE’s definition of epilepsy (52). In our recent study from Buskerud county we made similar findings (13). A retrospective review of patient records to validate the diagnosis would naturally increase the specificity, but this is conditional on the correctness of the information in the records and that the registrations have been made by competent personnel.

Age variation in the study population is another problem. Only a handful of studies from the Nordic countries include all age groups. Studies that are limited to a specific age group can only with difficulty be extrapolated to the entire population, since it has previously been proven that there is an elevated prevalence of epilepsy in the oldest section of the population (12). This trend is also evident in the material we have collected for this article. In studies that have investigated the prevalence of active epilepsy in children and adolescents (0 – 19 years), the prevalence rate varies from 3.2 to 4.1 per 1 000 (14 – 20), while the prevalence of active epilepsy among adults is higher (5.5 – 8.2/1 000) (22 – 25). In the Nordic countries, no isolated studies of prevalence among the elderly have been made, nor has this been investigated in the adolescent group.

As regards incidence, there are too few studies in our material and their results are too divergent to permit any conclusions regarding the age distribution. It is known, however, that the incidence curve for epilepsy is double-humped, with a peak early in life and a new increase in the oldest age groups (53).

Unfortunately, some stigma remains attached to the diagnosis of epilepsy (54). A wish to keep this diagnosis secret on the part of some people in certain communities or age groups may cause underreporting. This source of error may perhaps be greater in older studies, and perhaps also in studies undertaken in rural areas.

Conclusion

Epidemiological research on epilepsy is not an exact science. All the methods in use include sources of error. An adequate description of the methodology and its limitations, as well as a shared definition of epilepsy, is essential in all such research efforts. Studies that are based on updated guidelines and classifications of epilepsy remain scarce – not only in the Nordic countries, but also globally.

On the basis of this literature review, it is reasonable to assume that the real prevalence of active epilepsy in the Nordic countries amounts to approximately 6 per 1 000, i.e. that approximately 30 000 individuals in Norway suffer from active epilepsy. In light of the few prospective studies that have been undertaken, the annual incidence can be estimated to 30 – 60 new cases per 100 000 inhabitants.