Epilepsia, 44(7):950–955, 2003 Blackwell Publishing, Inc. C 2003 International League Against Epilepsy
Cysticercosis as a Major Risk Factor for Epilepsy in Burundi, East Africa ∗ †Georges Nsengiyumva, ∗ Michel Druet-Cabanac, ∗ Bienvenue Ramanankandrasana, ∗ Bernard Bouteille, ‡L´eopold Nsizabira, and ∗ Pierre-Marie Preux ∗ Institute of Neuroepidemiology and Tropical Neurology (EA 3174), Limoges, France; and †National Institute of Public Health, and ‡Department of Neurology, University Hospital of Kamenge, Bujumbura, Burundi
Summary: Purpose: Human cysticercosis is a direct consequence of infection by Taenia solium larvae (Cysticercus cellulosae). Results of studies on the impact of neurocysticercosis on epilepsy in Africa are inconsistent. The objective was to evaluate the role of cysticercosis in epilepsy in Burundi. Methods: A prevalent matched case–control design was used in the Kiremba area, Burundi, between March and April 2001. One case with epilepsy was matched to two control subjects, according to their age. Cases were subjects who had shown at least two unprovoked epileptic seizures within a 24-h time range and who lived in the Kiremba area. The control subjects also lived in Kiremba and had neither neurologic illness nor kinship with the people with epilepsy. Seropositiv-
ity for cysticercosis was the exposure variable. Three hundred twenty-four prevalent cases, with onset of epilepsy between 1950 and 2000, and 648 age-matched controls were included. Results: This study found a link between cysticercosis infestation and the occurrence of epilepsy (odds ratio, 3.8; 95% confidence interval, 2.5–5.1). Conclusions: The study highlighted the importance of cysticercosis in the area of Kiremba, as 31.5% of the control subjects screened positive for this parasite. The attributable risk for cysticercosis was 50% (95% confidence interval, 42–57) in this population. Key Words: Africa—Burundi—Cysticercosis— Epilepsy—Matched case–control study.
Cysticercosis in humans results from the infection by the larval stage (cysticercus) of the human tapeworm, Taenia solium. Humans are the only definitive host, but human fecal excretions disseminate eggs in the environment, and then eggs infect pigs and sometimes humans, causing cysticercosis. It is most frequent in many developing countries with pig breeding under poor sanitary conditions. Cysticercosis can provoke neurologic symptoms during encystation in the nervous system (1). Studies conducted in Latin America or Southeast Asia have shown that cysticercosis is a major cause of seizures and one of the leading causes of late-onset epilepsy (after 20 years). The role of this parasite in epilepsy has not been clearly evaluated in African patients. The data available on this subject provided conflicting results (2–4), with some authors finding a positive link (5,6), while others found no link at all (7,8). Cysticercosis is highly endemic in Burundi, where foci of infection have been identified with
a high seroprevalence rate among people with epilepsy (9,10). The objective of this study was to evaluate the importance of cysticercosis as a risk factor for the occurrence of epilepsy within an area of high cysticercosis endemicity, known as Kiremba, located in the northern part of Burundi. SUBJECTS AND METHODS The study was done in March and April 2001, in the territorial division of Kiremba in the north of Burundi. The Kiremba division (surface area, 245 km2 ) comprises 46 hills with 70,000 inhabitants. This is a rural area with extensive pig-breeding practices. The division had five primary care centers and one hospital of 200 beds. A matched case–control design was used. The cases were subjects with epilepsy, and the controls had no epilepsy. The exposure factor studied was cysticercosis. Clinical and paraclinical examinations were carried out in the people with epilepsy and the controls in the Kiremba hospital. The study was carried out under the auspices of Burundi’s Ministry of Public Health, which in Burundi has the role of a national ethical commitee.
Accepted March 1, 2003. Address correspondence and reprint requests to Pr. P-M. Preux at Institut de Neuro´epid´emiologie et de Neurologie Tropicale (EA3174), School of Medicine, 2, rue du Docteur Marcland, 87025 Limoges Cedex, France. E-mail:
[email protected]
950
CYSTICERCOSIS AND EPILEPSY IN BURUNDI Inclusion The medical personnel working in these towns had, in the weeks preceding the study, sensitized the village leaders, people with epilepsy, and their families to their participation in the study. Informed verbal consent was obtained after information and discussion in the local language from each included subject. People with epilepsy were defined as subjects having had at least two unexplained, unprovoked seizures occurring within a time range >24 h (all types of generalized or partial seizures were included) (11). The controls had no neurologic illness. Both cases and controls had been living in the Kiremba area for ≥2 years. Matching criteria Two controls were matched against one subject with epilepsy for the age (±5 years). Cases and controls had no blood relationship. A physician who verified the clinical inclusion and exclusion criteria examined each subject. Collecting data Local physicians in Burundi gathered the data. A standardized data-collection sheet was translated into the local dialect and was applied to both cases and controls. This data sheet was greatly inspired from the questionnaire for investigation of epilepsy in tropical countries (12). It gathered information on the subjects’ identity; the demographics; the history; the family history, and occurrence of epilepsy in first- or second-degree relatives; the epilepsy risk factors [cranial traumas, severe childhood illnesses (severe measles, cerebral malaria, meningitis, encephalitis)]; the risk of contracting cysticercosis (lifestyle, sanitary habits, water supply, pig breeding, use of veterinary control services during slaughtering of pigs, and dietary habits related to the consumption and cooking of pork meat). Clinical examination A neurologist confirmed the diagnosis after a thorough history taking and a complete neurologic clinical examination and classified the seizures according to the classification defined by the Commission on Classification and Terminology of the International League Against Epilepsy (13). Motor and sensory problems, cerebellar, pyramidal, and extrapyramidal syndromes, as well as cranial nerve palsies were thoroughly evaluated in both the cases and the controls. The physician looked for the dates of onset of seizure disorders, the date of the latest seizure, checked whether the epilepsy was active (at least one seizure in the last 5 years), and listed the antiepileptic drugs (AEDs) used. Cutaneous lesions specific to cysticercosis (e.g., subcutaneous nodules) were thoroughly searched for by inspection and palpation. These examinations were carried out in both cases and controls.
951
Electroencephalographic evaluation An EEG was performed on people with epilepsy with a portable digital device (Medatec Brainnet II, Medatec Medical Data Technology, Belgium). Those recordings were made according to the same standardized procedure and by a technician trained in the use of such device. The recording was 15–20 min long. The electrodes were mounted according to the standard 10-20 system by using standard procedures, with reference electrode as C1 or bilateral ear electrodes. Hyperventilation was performed in all cases. However, photic stimulation was unavailable. The results of the recordings were digitally archived for later review by reviewers who were blinded to clinical information, in the Limoges University Hospital’s Neurophysiologic Department, Limoges, France. Serologic evaluation A 10-ml blood sample was obtained from each subject by using nonanticoagulated Vacutainer tubes to perform serologic examination for cysticercosis. Each sample was then centrifuged, and the sera transferred into cryofreezing Nunc tubes, and immediately put into liquid nitrogen containers. Then they were preserved in dry ice in an isothermal container ready to be flown to Limoges, France. Cysticercosis serologies were done in all the sera in the same laboratory by the same technician. The screening for cysticercosis was done by using the ELISA (enzyme-linked immunosorbent assay) method. The antigen used for coating the microtiter plates was a crude soluble extract of T. solium cysticerci. Absorption was measured at 405 nm with a LP 400 spectrophotometer (Diagnostic Pasteur). The reaction threshold was 0.400. Values ≥0.400 were considered to be positive for cysticercosis. This method has a sensitivity of 86% and a specificity of 92%, as described by Houinato et al. (14). Statistical analysis Data were analyzed by using Epi-Info 5.01b software (Centers for Disease Control, French version: National School of Public Health, 1992) and Statview 5.0 software (SAS Institute, Inc., Cary, NC, U.S.A.). Quantitative variables were described by using mean and standard deviation (SD). The frequency comparisons were done with the Pearson χ 2 test. The mean comparisons were done by using the Mann–Whitney or Kruskal–Wallis tests. Exposure proportions of the cases and the controls were calculated, and matched odds ratio (OR) and its confidence interval were estimated. The 95% confidence interval was estimated. A multivariate analysis was made by using logistic regression. The matching variable (age) was included in the model. The initial model held relevant factors that were statistically associated with epilepsy with a preserving threshold (p < 0.25); the model was manually simplified in a descending way, according to Hosmer and Lemeshow’s (15) method, and Wald’s test was used to evaluate nonsignificant variables. First-degree Epilepsia, Vol. 44, No. 7, 2003
952
G. NSENGIYUMVA ET AL.
interactions were evaluated concerning significant factors kept in the final model. The minimum number of subjects needed for the study (two controls for each case, with a 5% α error risk and a power of 80%, in a region where the prevalence of cysticercosis in the controls was estimated at 25% at the time of an early survey conducted in Kiremba within the few months before the present study) was 380 triplets, hence 1,140 subjects. RESULTS Three hundred twenty-four triplets were included, representing 324 epilepsy cases and 648 controls, giving a total of 972 subjects. Four hundred seventy-six (49.0%) were males, and 496 (51.0%) were females. No significant difference was found between the cases and the controls on the matching criteria of age [the mean age was 26.0 years (SD, 15.4 years) for cases and 26.8 years (SD, 14.9 years) for controls]. Cysticercosis Among the 972 subjects examined, 397 were positive for ELISA cysticercosis serology. This figure represents 40.8% of the study population (44.3% were males, and 37.5% were females; p = 0.03). The average age was 29.8 years (SD, 16.0 years) in those who had cysticercosis and 24.3 years (SD, 13.9 years) in those who did not have cysticercosis. Main risk factors significantly associated with a positive ELISA serology were age (p < 0.001), occupation (45% of the farmers were positive, this was significantly higher than the other occupational activities; p < 0.001) and pig breeding close to dwelling places (p = 0.03). There was no significant difference according to serologic status concerning the other risk factors for cysticercosis [lack of toilet at home (p = 0.18), lack of public water tap near dwelling places (p = 0.10), eating habits related to pork meat consumption and cooking (p = 0.08)]. Subcutaneous nodules were found only in 4.7% of the seropositive subjects. Epilepsy The mean age of people with epilepsy was 26.0 years (SD, 15.4 years); females, 25.5 years (SD, 14.2 years), and males, 26.5 years (SD, 16.5 years). The sex ratio (M/F) was 1:1. The distribution of the epileptic seizures was 55% generalized seizures and 45% simple partial or complex partial seizures. In this study, 99.7% of the subjects had active epilepsy. The mean evolution period for epilepsy was 9 years (SD, 7 years). Seventy-seven percent of people with epilepsy reported seizure onset before age 20 years. Clinical examination revealed burns for 6.8% of people with epilepsy and subcutaneous nodules for 4.8%. Thirteen cases had neurologic signs: amyotrophy of a leg (two), aphasia (one), cerebellar syndrome (three), facial dysmorphy (one), hypoacousia (one), oculomotor palsy (one), and psychomotor retardation (four). A family hisEpilepsia, Vol. 44, No. 7, 2003
tory of epilepsy was found in 41.6% of the people with epilepsy (first and second degrees). Personal history of head injury was found in only two patients, and history of severe childhood diseases in 28 subjects. No form of antiepileptic treatment had been set up for 83.4% of the people with epilepsy. EEGs were carried out on 250 people with epilepsy. One result could not be read because of excessive artifact. Analysis of those records did not reveal any abnormality in 118 (47.4%) cases. Thirty-one (12.4%) cases had abnormal records in showing EEG seizure activity. Of these, 17 (54.8%) had generalized seizures, and 14 (45.2%) had partial onset with or without or secondary generalization. Nonepileptic EEG abnormalities were found in 100 (40.2%) of the records (isolated abnormality of background activity, focal slowing, or generalized slowing). Comparisons between cases and controls No significant difference was found between people with epilepsy and control subjects regarding gender (p = 0.12) and average dwelling period in Kiremba (p = 0.60). Comparisons of other factors between people with epilepsy and controls are in Table 1. Relation between cysticercosis and epilepsy The matched OR was 3.8 (95% confidence interval, 2.7–5.1). Of the people with epilepsy, 59.6%, and 31.5% of the controls had cysticercosis (p < 0.001). After adjustments on other variables according to a multivariate analysis, the association between cysticercosis and epilepsy remained significant. Adjusted OR was 4.1 (95% confidence interval, 3.0–5.6). Results of the final model are displayed in Table 2. Cysticercosis was still a risk factor, regardless of the type of epileptic seizure (generalized seizures: matched OR was 3.8, 95% confidence interval: 2.5–5.8; partial or secondary generalized seizures: matched OR was 3.6; 95% confidence interval, 2.1–5.7). Analysis of EEG recordings did not reveal any specific pattern regarding the serologic status for cysticercosis. Computation of etiologic fraction for epilepsy in cysticercosis-seropositive subjects was 73.4% (95% confidence interval, 64–80) using univariate OR, and the population-attributable risk for cysticercosis was 50% (95% confidence interval, 42–57). DISCUSSION In Burundi, several foci of cysticercosis have been identified, but few studies have investigated the possible relation between cysticercosis and epilepsy (9,10). Cysticercosis and epilepsy represent two major health problems in Burundi because of their serious medical, social, cultural, and economic implications. The present study searched for a relation between cysticercosis infestation and epilepsy. This was the first matched case–control study carried out on this theme in this country with such a large number of cases.
CYSTICERCOSIS AND EPILEPSY IN BURUNDI
953
TABLE 1. Comparisons between cases and controls (univariate), Kiremba, Burundi, 2001
Sex (male) Indoor toilets Water pool nearby Pig-breeding family Pig breeding nearby Pork meat consumption Roasted cooking of pork meat Family history of epilepsy Family history of neurologic disease Personal history of cranial trauma Severe disease during childhood Seropositivity for cysticercosis
Cases (n = 324)
Controls (n = 648)
OR
95% CI
p Value
170 (52.5%) 1 (0.3%) 86 (26.5%) 80 (24.7%) 72 (22.2%) 266 (82.1%) 44 (13.6%) 116 (35.8%) 17 (5.2%) 2 (0.6%) 28 (8.6%) 193 (59.6%)
306 (47.2%) 27 (4.2%) 249 (38.4%) 98 (15.1%) 142 (21.9%) 508 (78.4%) 105 (16.2%) 102 (15.7%) 29 (4.7%) 8 (1.2%) 27 (4.2%) 204 (31.5%)
1.2 0.1 0.6 1.8 1.3 1.3 0.9 3.8 1.1 0.5 2.1 3.8
0.9–1.7 0.0–0.4 0.4–0.7 1.3–2.6 0.9–1.8 0.9–1.8 0.6–1.4 2.7–5.7 0.5–2.1 0.1–2.2 1.2–3.8 2.7–5.1
0.10
