Prevalence, Incidence, and Etiology of Epilepsies in Rural Honduras: The Salamá Study

Authors


Address correspondence and reprint requests to Dr. M.T. Medina at Instituto de Neurociencias, Col. Tepeyac, Calle Yoro, Edificio Mater Dei, Tegucigalpa, Honduras. E-mail: marcotmedina@yahoo.com or marco@gbm.hn

Abstract

Summary: Purpose: Determination of epilepsy etiology in population-based studies is difficult because of the high cost of diagnostic tests. However, cost-effectiveness may be proven if preventive public-health strategies can be established from the test results. We report an epilepsy population-based study using clinical and laboratory techniques.

Methods: A medical team administered an epilepsy survey to 88% of the residents by census in the rural county of Salamá, Honduras. Ninety of 100 participants identified with active epilepsy underwent a neurologic examination, video-electroencephalography (video-EEG), brain computed tomography (CT) scan, and serum enzyme-linked immunoelectrotransfer blot (EITB) for cysticercosis. Final diagnoses were based on the International League Against Epilepsy classifications for seizures and epilepsy syndromes. Combined epidemiologic, clinical, video-EEG, neuroimaging, and serum EITB assays were used for the diagnosis of epilepsy etiologies.

Results: Among 6,473 residents surveyed, 151 persons with epilepsy (prevalence rate, 23.3/1,000) were identified, 100 of whom had active epilepsy (15.4/1,000) on the prevalence day. Incidence was determined to be 92.7/100,000. Partial seizures with or without secondary generalization were common (92.2%). Symptomatic epilepsy (62%) was primarily due to neurocysticercosis (37%), perinatal brain damage (8%), post-traumatic (3%), and poststroke (2%). Eight percent were idiopathic, and 30% were cryptogenic (unknown cause).

Conclusions: Symptomatic epilepsies primarily explained the high prevalence and incidence of epilepsy in Salamá. Integration of video-EEG and brain CT scan with clinical-epidemiologic evaluation was critical for determination of epilepsy etiology. Establishment of specific programs for continuation of epidemiologic surveillance, education, intervention, and long-term follow-up will benefit the Salamá region.

Epilepsy is a worldwide health problem (1–3). Recent analyses concluded that Latin America, especially Central America, has high epilepsy prevalence rates that appear to be secondary to a high frequency of symptomatic “preventable” epilepsies (4,5). Comparison data are not always available because few worldwide epidemiologic studies used standardized case-ascertainment criteria (1,3,6–8). Primarily because of the high costs and/or multiple logistic factors, the vast majority of previous epidemiologic studies were two-phased and did not use neuroimaging or electroencephalography (EEG) to determine the etiology and syndromic classification of the epilepsies (9).

The International League Against Epilepsy (ILAE) has published standardized diagnostic criteria for epileptic seizures and syndromes (10,11) and recently proposed the use of a five-axes approach to include ictal phenomena, seizure type, syndrome, etiology, and impairment (12). The cost-effectiveness of obtaining complete seizure-evaluation information will be demonstrated by establishment of specific health programs to reduce the frequency of the symptomatic epilepsies in the long term. With these recommendations and criteria in mind, we performed a house-to-house survey to establish the prevalence, incidence, and etiology of epilepsies by using clinical evaluations, video-electroencephalography (video-EEG), brain computed tomography (CT) scans, and serum enzyme-linked immunoelectrotransfer blot (EITB) assays for cysticercosis in the rural county of Salamá, Honduras.

METHODS

Population

In 1995, analyses of morbidity in 11 counties in Eastern Honduras identified epilepsy as the tenth leading cause of medical consultation (13). Because of the favorable logistic conditions of Salamá County (Fig. 1), this rural area was chosen for a population-based study of epilepsy. The county had a total land surface of 342 km2. Twenty-nine of its 31 communities have fewer than 640 inhabitants each (range, 10–2,422); the county has no major city. A census was previously done by the government in 1994 (CEFASA or Family Health Census), which reported an eligible county population of 7,384.

Figure 1.

Localization of Salamá County in Olancho, Honduras.

For our study, we performed a census simultaneous with house-to-house epilepsy-screening, registering the total population by age and sex. We used the standard age and sex strata classification used by the Secretary of Health and the World Health Organization (WHO). We were assisted by the fact that one of the physicians involved in this study (Dr. Martínez) had lived in this community for 18 years. Another physician (Dr. Durón) lived in this community during a portion of the study period.

Primary occupations in Salamá are in agriculture, forestry, livestock, coffee, and small business. Its 31 communities are accessible by nonpaved roads; the literacy percentage is 88%; and three governmental health centers are in the area. Even though 10 different ethnic groups with their own dialects reside in Honduras, all residents in Salamá are mestizos and speak Spanish only. According to the Secretary of Health, the prevalence of poverty as measured by the index of unsatisfied basic needs for this community was elevated at 27% (13).

Bioethics

Persons identified with probable or definite active or inactive epilepsy by the survey were included in the in-depth epilepsy study. Approval for the epilepsy study protocol and consent form was given by the Secretary of Health of Honduras.

Case definitions

The definitions and classifications for epileptic seizures proposed by the ILAE (10,11) were used, as well as some additional concepts to make the definitions operative (8).

Study phases

Fig. 2 shows the phases of the study.

Figure 2.

Methods of the study by phases.

Phase I. House-to-house survey and clinical evaluation

Initially a house-to-house survey was done by some of the authors, including one board-certified neurologist/epileptologist/neurophysiologist, two board-certified public health specialists, and two general practitioners. The nonneurologist health care team members were trained before the survey in the classification of seizures by teaching sessions and observing video-EEG recordings. Multiple family members often were available in each household and could provide information about the patient. A few initial general demographic and medical history questions were asked by the investigators to determine which individual had the most knowledge of the patient, family, and the patient's seizures. That person became the key informant and was often a relative of the patient and/or a witness to a seizure event. The key informant was then asked to describe the clinical history of the patient.

We used a similar approach for other epidemiologic studies we have conducted in Honduras (14–16). We have found that in 80% of cases, the mother, grandmother, or other female relatives were the best key informants (Medina, unpublished data).

To identify persons by survey with a history of seizures in their lifetimes, any seizure-like events were classified according to the ILAE (8). The screening survey instrument used had been previously submitted to a pilot test in 20 Honduran families and was primarily based on four questions translated into Spanish: (a) Has someone at home lost consciousness or fallen unconscious? (b) Has someone ever been disconnected from the surroundings or without movement and staring? (c) Has someone had involuntary uncontrollable movements or strange sensations in their limbs or any part of the body? and (d) Has someone ever had convulsions, seizures, or epilepsy?

This is a modified version of the questionnaire used in 1996 by Aziz et al. (17) to detect the characteristics of convulsive and nonconvulsive epileptic seizures. Aziz and other authors, including Placencia et al. (18), have modified the WHO questionnaire published in 1981 to adapt to local situations and preferences. Some authors used modified WHO questionnaires, but, with some exceptions (17–19), did not publish the modifications (20,21). Acceptable sensitivities and specificities have been reported (>90% and 50%, respectively) (18–21).

We used our modified questionnaire for a National Prevalence Study in Honduras, which included a population of 135,126, and found it effective (22). The survey instrument also included registration of demographic data and the temporal or definite classification of events under study for epileptic seizures. Inhabitants of Salamá County were considered residents if they had lived in the county for ≥6 months preceding the prevalence day (April 30, 1997).

Patients with possible or probable epilepsy who consented to participate were thoroughly evaluated by a complete history and examination at local health centers in Salamá County. A family history including a pedigree was completed. After this interview, the relatives or observers of the event in question were asked to describe the seizures. Finally, they were asked to identify the patient's seizure by watching a 10-min video-EEG that included all the seizure types according to the ILAE Classification (10). These evaluations were initially done by nonneurologists, and at a later date, were validated by the epileptologist. In some cases, these evaluations were performed at home when physical and mental disabilities or transportation issues were a problem. Epilepsy cases were considered active if any seizures had occurred during the last 5 years, and inactive, if greater than 5 years had elapsed.

Phase II. Neuroimaging, electroencephalographic, and immunologic studies

Patients who were identified clinically as having active epilepsy were invited to have a video-EEG, brain CT scan, and serum EITB assay for cysticercosis. Local health and governmental institutions were involved in the organization and transportation of patients 160 km (99.2 miles) to the capital city of Tegucigalpa for their studies. Transportation costs and studies were paid by the project and the local government. Every patient had (a) a brain CT scan (GE Sytec 2000i) with and without contrast using 5-mm slices; (b) a 30-min video-EEG with photo stimulation and hyperventilation, following the criteria of the 10-20 International System, with tracings recorded on an eight-channel Grass EEG machine and in video by a split-screen system (sleep-induced EEGs were performed in three cases); and (c) serum EITB assays by a parasitologist for cysticercosis by using the specifications of Immunetics Inc. and the Centers for Disease Control and Prevention. These neurocysticercosis (NCC) results were published separately (23).

Phase III. Final seizure, syndromic, and etiologic classification

All semiologic and neurologic findings were reviewed and classified per ILAE criteria (10,11). A combination of epidemiologic, clinical, neuroimaging, and laboratory criteria proposed by Sánchez et al. (24) and Del Brutto et al. (25) were used for the diagnosis of NCC. The criteria used for symptomatic epilepsy were those proposed by the ILAE (8) and Roger et al. (26).

Phase IV. Treatment and follow-up plan

After the final diagnosis was made, each patient received information on the diagnosis and a plan for treatment and follow-up.

Data quality and analysis

A final review for data validation was performed by the specialists composing the study team. Brain CT scans were separately interpreted by a study-blinded board-certified neuroradiologist. Data entry was done in Microsoft Excel and verified by two of the researchers. Analyses were done with the Statistical Package for Social Sciences (SPSS for Windows, version 6.1).

RESULTS

The initial screening survey questionnaire was administered in 1997 to 88% (6,473) of the total eligible population of 7,384 persons living in Salamá County. The population was very cooperative, and the house-to-house-survey was done in 11 days. The population was 51% female; 56% were younger than 20 years. This is representative of the general demographic characteristics of the population in Honduras. Our modified epilepsy questionnaire had a sensitivity of 99.2% and a specificity of 75%.

The total number of houses in the county was 1,328. All were visited by the survey team, and 187 (14%) were uninhabited. A house was not considered uninhabited until the survey team had visited it 3 times and confirmed (with the assistance of neighbors and health community volunteers) that the residents were not in the community. In this community, people migrate to major cities or other locations during months that coffee, corn, or beans are being harvested.

Classification of population suspected of having seizures

In the survey, initially a total number of 272 persons had a history of events compatible with epileptic seizures (Table 1). Of these, 24 (8.8%) were secondary to nonepileptic events (migraine, syncope, other systemic causes of loss of consciousness, nonepileptic abnormal movements, schizophrenia, circulatory problems in the limbs, and pseudoseizures). A seizure diagnosis was rejected in six cases after a video-EEG, a brain CT scan, and an evaluation by the epileptologist were done. These were cases with symptoms suggestive of epilepsy but ultimately diagnosed as hemiplegic migraine (one case), migraine with aura (three cases), syncope (one case), and Gilles de La Tourette syndrome (one case).

Table 1. Classification of the 272 persons initially suspected of having seizures
 
Cases
Initial
group (%)
Prevalence
rate/1,000
95%
CI
  1. aThis category includes the only two patients diagnosed with neonatal seizures.

All epileptic seizures24891.238.333.5–43.1
All epilepsies15155.523.319.6–27.0
Activea10036.815.412.4–18.5
Inactive 5118.8 7.9 5.7–10.0
Single seizure 5821.3 9.0 6.7–11.3
Febrile seizures 3914.3 6.04.1–7.9
Simple 3312.1 5.13.4–6.8
Complex  6 2.2 0.90.2–1.7
Pseudoseizures  7 2.6 1.10.3–1.9
Nonepileptic events 17 6.3 2.61.4–3.9

Criteria for epileptic seizures were met by 248 (91.2%) persons. Final classification and prevalence rates of all cases with epileptic seizures are shown in Table 1. The most frequent seizure group (55.5%) met criteria for active or inactive epilepsy. The second most frequent group (21.3%) had isolated seizure events (prevalence rate, 9.0/1,000). Attributable causes of isolated seizures were diagnosed as alcohol withdrawal (8.9%), NCC confirmed by brain CT scan (6.5%), brain trauma (4.3%), eclampsia (2.1%), and encephalitis (2.1%). However, the etiology of the isolated seizures was undetermined in 76% of cases. The third most frequent category of seizures (14.3%) was febrile seizures (prevalence rate, 6.0/1,000) and most patients (84.6%) met criteria for simple febrile seizures (1). Of the entire population surveyed, only two patients had a history of neonatal seizures (prevalence rate, 0.3/1,000), and both now experience active epilepsy.

Epilepsy prevalence

The overall prevalence of epilepsy was 23.3/1,000 [95% confidence interval (CI), 19.6–27.0). The prevalence of active epilepsy was 15.4/1,000 (95% CI, 12.4–18.5) and inactive epilepsy was 7.9/1,000 (95% CI, 5.7–10.0; Table 1). Overall prevalence rates were 25.2/1,000 in females and 21.3/1,000 in males. The prevalence of active epilepsy (Table 2) was higher in females (17.1/1,000) than in males (13.7/1,000), but was not statistically significant (χ21= 1.24; p = 0.2653). However, a significantly greater prevalence was found in women age 20 years or older (χ21= 6.30; p = 0.0121) as compared with men.

Table 2. Age distribution and prevalence (cases per 1,000) of active epilepsy
Age
groups (yr)
Total
population
Total
cases

Prevalence
Total male
population
Total male
cases
Prevalence
in males
Total female
population
Female
cases
Prevalence
in females
 0–91,89518 9.5 9441111.7951 7 7.4
10–191,7423218.4 8481720.08941516.8
20–29  9011718.9 453 715.54481022.3
30–39  6911115.9 325 412.3366 719.1
40–49  4891224.5 233 2 8.62561039.1
50–59  334 618.0 150 1 6.7184 527.2
60–69  230 313.0 111 1 9.0119 216.8
70–79  123 1 8.1  52 0 0.0 71 114.1
80+   68 0 0.0  25 0 0.0 43 0 0.0
Total6,473100 15.43,141 4313.73,332  5717.1

Fifty percent of all epileptic seizures occurred in patients younger than 10 years. In patients with active epilepsy, 50% were younger than 20 years at examination, and 18% were younger than 10 years. Fifty-two percent of patients with active epilepsy had onset before age 10 years, and 77% had onset before age 20 years. In general, prevalence rates higher than 10/1,000 were present in most groups, with peaks in males in the age intervals from 10 to 19 years, and in females, from 40 to 49 years (Table 2).

Epilepsy incidence

The incidence of epilepsy in this prospective study was calculated on a door-to-door basis for the 1996 to 1997 period. For the determination of incidence in the following years, a capture-recapture method was established through a permanent surveillance program by the Secretary of Health. The denominator used was the survey population of 1997. New cases are now being studied with the same methods including video-EEG and brain CT scan. Epilepsy incidence by year was 92.7/100,000 (six cases; 95% CI, 18.5–166.9) in 1996–1997, 61.8 (four cases; 95% CI, 1.2–122.4) in 1998, 15.4 (one case; 95% CI, 0.0–45.7) in 1999, 30.9 (two cases; 95% CI, 0.0–73.7) in 2000, 77.2 (five cases; 95% CI, 9.5–145.0) in 2001, and 61.8 (four cases; 95% CI, 1.2–122.4) in 2002 to 2003. Differences between these rates were not statistically significant.

Seizure and syndromes classification for active epilepsy

Of the 90 persons with active epilepsy who consented to diagnostic studies, 92.2% had simple or complex partial seizures with or without secondary generalization (Fig. 3). Patients most commonly had symptomatic epilepsies (62%) primarily due to NCC (37%). According to Del Brutto's criteria (25), all patients had definitive criteria for NCC (presence of a major + one minor + one epidemiologic criterion), with the majority of patients by history having had or been exposed to tapeworms. Detailed data on the epilepsy due to NCC will be reported in a separate article. Other causes of symptomatic epilepsies are detailed in Table 3.

Figure 3.

Electroclinical classification of seizure types in 90 participants with active epilepsy.

Table 3. Etiology of active epilepsy in the patients studied by age groups
Etiology/Age groups0–910–1920–2930–3940–4950–5960+Total%
  1. aThe “Other” category includes a tumor (probable meningioma), noncysticercotic granuloma, chronic alcoholism sequelae, chronic hydrocephalus, progressive myoclonic epilepsy (probable Unverricht–Lundborg disease).

Cryptogenic 7 8 3 3 23127 30.0
Idiopathic 3 4  7  7.8
All symptomatic 721 8 6 93256 62.2
Neurocysticercosis 413 5 3 61133 36.6
Perinatal brain damage  5 1 1  7  7.8
Poststroke  1 1 2  2.2
Cortical dysplasia 1  1  2  2.2
Post-traumatic 2 1  3  3.3
Postmeningitis/encephalitis  1  1  2  2.2
Multifactorial  1  1  2  2.2
Othera  1  1 12  5  5.6
All etiologies24551815219590100.0

Eight percent had idiopathic epilepsies, and 30% had cryptogenic epilepsies (unknown cause). Idiopathic epilepsies were primarily present in persons younger than 20 years. Symptomatic and epilepsies of unknown cause were found in all age groups. The idiopathic epilepsies found were benign childhood epilepsy with centrotemporal spikes (4.4%), childhood epilepsy with occipital paroxysms (1.1%), and childhood absence epilepsy (1.1%). One patient had myoclonic–astatic epilepsy (1.1%) that met criteria for idiopathic epilepsy, even though the actual classification considers this syndrome as symptomatic or cryptogenic (8,11). One case of progressive myoclonic epilepsy was probably due to Unverricht–Lundborg disease (1.1%).

The overall contribution of the diagnostic studies performed can be measured by positive or abnormal findings. With the electroclinical diagnostic approach of the ILAE (10,11), which used interictal as well as ictal video-EEG findings, 84% of our active epilepsy cases demonstrated either specific focal or generalized epileptiform activities and/or focal slow activities that correlated with the seizure semiology and neuroimaging findings. In five cases, interictal findings were very specific [i.e., the presence of centrotemporal spikes with normal background in four cases with partial rolandic epilepsy (idiopathic epilepsy) and the presence of occipital paroxysms in a case typical of childhood epilepsy with occipital paroxysms (idiopathic partial epilepsy)]. Although most of our significant EEG abnormalities were interictal, in two cases, we recorded ictal discharges related to myoclonias and absences.

Significant findings were present in 60% of brain CT scans and in 30% of blood EITB assays for cysticercosis. Brain calcifications that were probably due to NCC were found in three (3.3%) additional cases and were considered incidental findings.

Past history

Historical data that could be related to the increased risk for active epilepsy were family history of epilepsy (67%); febrile seizures (8%) and neonatal seizures (2.2%); personal (14%) and household (37%) history of taeniosis; midwife-assisted delivery (79%); delivery abnormalities including forceps use, prolonged or precipitate labor, or abnormal bleeding before or during delivery (32%); head trauma (7%); and neuroinfection (2%).

Treatment

Fourteen patients with active epilepsy had never visited a doctor for treatment; 7.7% had visited their doctor, but an epilepsy diagnosis had not been made. In retrospect, these were cases with simple, complex partial, and absence seizures. Forty-two (46.7%) patients with active epilepsy were receiving AED treatment. Although ∼20% of patients with active and inactive epilepsy in the world are treated (27), 46.7% of patients with active epilepsy in Salamá were receiving treatment. This higher percentage of treated epileptics was primarily due to 1994 census reports identifying the high frequency of epilepsy and a local community physician (Dr. Martinez) becoming an advocate for this group of patients. The percentage of treated persons would be lower if we considered epilepsy patients with both active and inactive epilepsy. Fifty-eight percent of those treated were receiving monotherapy with phenytoin (PHT) or phenobarbital (PB). The remainder received carbamazepine (CBZ) or valproate (VPA) in combination with PHT or PB. Seventy-six percent of patients had had seizures in the last 12 months, and 47% had had status epilepticus sometime during the course of their epilepsy.

Interventional and follow-up program

After the results of the epidemiologic and clinical study were known, they were discussed with the Secretary of Health to design strategies for community interventions, epidemiologic surveillance, and follow-up of the cohort. A full report of the details of the interventional program will be published separately as results become available.

DISCUSSION

Epilepsy is highly prevalent in the rural county of Salamá, Honduras. This is one of the first studies from a developing country that shows epilepsy etiologies in a population-based study. We found that the vast majority of epilepsies were symptomatic, which contrasts with industrialized countries, where idiopathic epilepsies are reported to be more frequent (28,29). NCC was the primary cause of symptomatic epilepsy.

Our field study showed that the percentage of persons with epileptic seizures identified in the first phase of the Salamá study (4.2%) is lower than that in other population-based studies from Karachi, Pakistan (6%; 17) and Medellín, Colombia (9.7%; 30). Differences may be due to methodologic aspects, especially the training and experience of the persons involved in the survey, and to financial support. All of our study patients had a full evaluation regardless of ability to pay. The higher percentage of cases considered suggestive of epilepsy in other studies may reflect lower specificity of the screening. This results in more suspected cases and more excluded cases from the final data. When sensitivities and specificities are reported, most report sensitivities >90% but variable specificities as low as 50.8% (18). Our questionnaire had a sensitivity of 99.2% and a specificity of 75%.

The overall epilepsy prevalence rate (23.3/1,000 inhabitants) and the prevalence of active epilepsy (15.4/1,000) in Salamá are high compared with data of developed industrialized countries but similar to other Latin-American data (4,5). In the former, overall prevalence rates range from 5/1,000 to 8/1,000. Most studies published in Central and South America show prevalence rates ranging from 7/1,000 to 57/1,000 (4,5) and include: 57/1,000 in a Panama rural county, 22/1,000 in Panama City (31,32), 8.5 to 52/1,000 in Guatemala (4), and 21.4/1,000 in Medellín, Colombia (30). Other studies have reported rates of 3/1,000 to 15/1,000 in Africa (33) and 9.9/1,000 in Pakistan (17). These studies appear not to be strictly comparable because of methodologic differences and local regional characteristics (1,3,7). The Salamá incidence rate also was high (92.7/100,000), with rates usually reported from 30 to 50/100,000 person-years in industrialized countries (1,3,6).

Most persons in population studies with active epilepsy have an onset of seizures in the first two decades of life (1,3,6). This was true in Salamá County. The prevalence peaks from the second to the fifth decade of life. Contrary to this finding, it has been reported that in industrialized countries like the United States, the prevalence rate is higher in the young, levels off, and then increases again in the older population (older than 65 years) (1,3). Rates in the older population were lower in Salamá. Possibly this is a result of a lower life expectancy and high mortality associated with lack of access to optimal care, resulting in inadequate control of seizures and status epilepticus.

The epidemiology of single isolated seizures varies in previous studies. We found a prevalence of 8.9/1,000. A similar study from Pakistan reported a prevalence of 2/1,000 (17). Although a cumulative incidence is used, the risk of having a single seizure is 3.6% in Rochester, Minnesota (34). The rate of febrile seizures in the total population from Salamá is low (6.0/1,000) when compared with that in Rochester, Minnesota data (1,6), which reports a rate of 5% in the childhood population and 2% in the total population (2). Our low prevalence of neonatal seizures probably confirms that most seizures are subtle and difficult to distinguish from normal movements (35).

Epidemiologic studies on epilepsy must consider the methodologic gaps and variables influencing the prevalence rates (7). For example, a retrospective review of medical records may be an inaccurate method to determine epilepsy prevalence, because patients may never consult a doctor because of the subtle nature of some seizure symptoms. The Salamá governmental clinic's registry of persons with epilepsy, when compared with the house-to-house survey, underreported active epilepsy by 64%. Problems with access to diagnostic studies and social stigma can delay the diagnosis and also contribute to underreporting of epilepsy (15,36). For these reasons, the most current sensitive and specific method for detecting epilepsy is community screening by an appropriate questionnaire, complete clinical evaluations, and the use of diagnostic procedures/tests.

By combining clinical and diagnostic findings, 98% of Salamá's active epilepsy patients had their seizures classified. The video-EEG contributed to the diagnosis 84% of the time. Although the MRI has greater sensitivity (37,38), brain CT neuroimaging studies helped clarify the etiology of active epilepsy in 60% of the Salamá cases. Although most studies in developed industrialized countries report a higher proportion of primary generalized epilepsies (1,6), recent reports from developing countries report a higher frequency of partial seizures with or without secondary generalization, which may indicate a high incidence of symptomatic epilepsy (2,5). Our data are consistent with this and also indicate that many epilepsies in developing countries are probably preventable. NCC is highly endemic in Latin America, and previous reports from hospital-based studies in the region document NCC as a common cause of epilepsy and epileptic seizures (39,40). Although not as frequent, perinatal brain damage appears to be responsible for 8% of the active symptomatic epilepsy in our study. Although the methods differed, perinatal brain damage has been reported to be the cause of 0.5–14% of the epilepsies in the world, with higher prevalences in underdeveloped countries (2,4,5).

Another problem detected by the current study is the high percentage of patients with active epilepsy who are not taking their prescribed treatment. Because 76% had seizures in the last 12 months with a high frequency of status epilepticus, concerns can be raised with regard to the access to treatment in this surveyed region. A similar situation has been found in other regions of Latin America, Asia, and Africa (27). Cost-effective access to AED treatments, education about medication side effects and stigma, as well as self-withdrawal from AED treatment with or without the use of alternative medicines need to be included in a comprehensive management plan (41–43).

The use of standardized definitions/criteria and algorithms/protocols with proven clinical effectiveness is needed for accurate case ascertainment for epilepsy studies (44). Neuroimaging studies, video-EEG, and laboratory studies are critical in determining the etiology of the epilepsies. By using these investigative criteria and techniques in our Salamá study, the high prevalence and incidence of epilepsy was found to be explained primarily by symptomatic epilepsies. This led to the establishment of programs in epidemiologic surveillance, education, intervention, and long-term follow-up specifically planned to benefit the people of the Salamá region.

Acknowledgments

Acknowledgment:  We thank the patients, their families, and all the Salamá communities for their participation in this study. We also thank Rafael Aguilar-Estrada, Arnold Thompson, Sofía Dubón, Francis Barahona, Marisabel Rivera, Francisco Ramírez, and Ana L. Sánchez from the National Autonomous University of Honduras for their assistance during this study. Funding for transportation, brain CT scans, and video-EEGs was provided by the World Bank through the project Nutrition and Health under the Secretary of Health. The Mayor of Salamá, Rafael Zúniga, provided assistance for transportation of patients and did collaborative work with the Secretary of Health on the establishment of the interventional program. The Karolinska International Research Training (KIRT) and Inger Ljungström provided support for the immunologic studies. Centro de Neurodiagnóstico provided video-EEG material for patient interviews and education, and technical assistance for EEG studies. The Japanese International Cooperation Agency (JICA), Dr. M. Sone from the Shizuoka Swine and Poultry Research Station (Shizuoka, Japan), Dr. H. Sakai from the Laboratory of Parasitology, Department of Disease Control of the Graduate School of Veterinary Medicine, Hokkaido University (Sapporo, Japan), and Dr. Peter Schantz from the division of Parasitic Disease, National Center for Infectious Disease, Centers for Disease Control and Prevention (Atlanta, Georgia, U.S.A.) provided the funding and technical assistance for the EITB assays. Lenín Banegas from the headquarters of the Secretary of Health in Salamá contributed to the field study and follow-up program.

Ancillary