FULL-LENGTH ORIGINAL RESEARCH
Association between epilepsy and cysticercosis and toxocariasis: A population-based case–control study in a slum in India
Address correspondence to Gagandeep Singh, Department of Neurology, Dayanand Medical College & Hospital, Tagore Nagar, Civil Lines, Ludhiana-141001, Punjab, India. E-mail: firstname.lastname@example.org
Purpose: To assess the association between epilepsy and exposure to the parasites, Toxocara canis and Taenia solium in a slum-community in India.
Methods: A door-to-door survey to determine the prevalence of epilepsy was carried out by trained field workers. For every case, one age- and gender-matched control was selected from the same community. Serologic evaluation was carried out to detect antibodies against T. canis and T. solium.
Key Findings: The crude prevalence of active epilepsy was 7.2 per 1,000. We enrolled 114 people with active epilepsy and 114 controls. The prevalence of antibodies to T. canis was similar in people with active epilepsy (4.7%; 5 of 106 people) and in controls (5.7%; 6 of 106 people). The prevalence of antibodies to T. solium was 25.5% (27 of 106) in people with active epilepsy, significantly higher than in controls (12.3%; 13 of 106 cases; p = 0.02). Adjusted conditional (fixed-effects) logistic regression estimated an odds ratio of 2.8 (95% confidence interval 1.2–6.8) for detection of T. solium antibodies. Nineteen people with active epilepsy demonstrated evidence of neurocysticercosis (NCC) on magnetic resonance imaging (MRI), including 7 (36.5%) with solitary cysticercus granuloma.
Significance: Our findings do not support an association between epilepsy and exposure to T. canis in the community studied. A significant association between T. solium exposure and epilepsy was observed. Of those with active epilepsy and evidence of NCC on MRI, a large proportion demonstrated solitary cysticercus granuloma.
Epilepsy is one of the most common neurologic disorders worldwide. The prevalence of active epilepsy is estimated to be between 6.2 and 7.6 per 1,000 population in developed countries (Granieri et al., 1983; Hauser et al., 1991) and 5–10 per 1,000 population in resource-poor countries (Aziz et al., 1994; Preux & Druet-Cabanc, 2005; Ngugi et al., 2010). The high prevalence in resource-poor countries has been partly attributed to a greater frequency of a variety of infectious diseases, including helminthic infestations (Singh & Prabhakar, 2008).
Neurocysticercosis (NCC), caused by the larval stage of the cestode helminth, Taenia solium is the most common parasitic infestation of the brain (García et al., 2003). It is the main cause of late-onset epilepsy (or epileptic seizures) in some resource-poor countries (Medina et al., 1990; Pal et al., 2000). Several population-based, cohort and case–control studies, supported by clinical, pathologic, and experimental data have confirmed a causal association between T. solium cysticercosis and epilepsy (Del Brutto et al., 2005; Montano et al., 2005; Nicoletti et al., 2005; Rajshekhar et al., 2006; Prasad et al., 2008). These association studies have been largely based on serologic evidence of exposure to the parasite T. solium using the enzyme-linked immunoelectrotransfer blot (EITB) assay (Tsang et al., 1989).
Human infestation with Toxocara species (including Toxocara canis, or dog roundworm, and T. catis, or cat roundworm) can occur in any of the following four types: visceral larva migrans, ocular larva migrans, covert toxocariasis, and common toxocariasis, although most human infestations are silent (Rubinsky-Elefant et al., 2010). Risk factors for human exposure are dog ownership, pica, and contact with soil contaminated with dog feces. Many early, poorly controlled studies suggested an association between T. canis exposure and epilepsy (Critchley et al., 1982). More recently, case–control studies in Bolivia, Burundi, and Italy also suggested a statistically significant association between T. canis exposure and epilepsy (Nicoletti et al., 2002, 2007, 2008). Evidence of a brain lesion due to T. canis is limited to a few case reports (Mikhael et al., 1974; Hill et al., 1985); therefore, it is not clear whether T. canis is a risk factor for epilepsy. We undertook a population-based, case–control study to determine the association between both T. solium and T. canis exposure and epilepsy in a slum community in India.
We conducted a population-based, case–control study between June 2010 and February 2011 by carrying out a door-to-door survey for epilepsy with confirmation of diagnosis of epilepsy using a two-step protocol. The two explanatory variables were serologic evidence of exposure to T. solium and T. canis.
The study was carried out in the Jamalpur urban field practice area of the Department of Social and Preventive Medicine, Dayanand Medical College, in Ludhiana, an industrial city in Punjab state in Northwest India. The city has a population of approximately 3 million. The registered population (inhabitants for at least 5 years) of the survey area was 15,750 and comprised 60% ethnic native Punjabis and 40% migrant laborers from elsewhere in India. Many people in the survey area shared dwellings (e.g., 5–6 families in small, 1–2 room tenements), and piped water was available to few; others used shared or separate hand-pumps. Primary health care was provided by an urban health center run by staff from a teaching medical college in the city (one physician and two auxiliary nurse midwifes), but many people preferred to visit private physicians for their health care needs.
Selection of cases and controls
Cases were selected in two phases:
Phase I: A door-to-door survey was conducted by two field workers using an epilepsy screening questionnaire adapted from previous epidemiologic studies of epilepsy (Placencia et al., 1992; Wang et al., 2003; Gourie-Devi et al., 2004). It was validated in 142 individuals, from the same geographical area, who attended a neurology clinic and underwent an epileptologic assessment by a neurologist (GS), yielding a sensitivity of 0.83 and specificity of 0.84. Demographic components were adapted from a previous neuroepidemiologic study (Meneghini et al., 1991). The socioeconomic status of the families was recorded on a modified Udai Pareek Scale (Pareek & Trivedi, 1979). Two additional structured questionnaires, one for risk factors for epilepsy and the other for assessment of risk of T. canis or T. solium infestations, were designed and were completed by subjects. Field workers were especially recruited for the study and were intensively trained in survey methods over a 2-month period.
Phase II: An epileptologist conducted a complete neurologic assessment in all individuals who screened positive between July 2010 and April 2011. Those with inactive epilepsy, nonepileptic seizures, and single seizures were excluded. Those with active epilepsy underwent sleep and awake digitized electroencephalography (EEG) examinations (Natlink Traveller, Biologic, Mundelein, IL, U.S.A.; 101 of 114 cases; 88.6%) and specialized epilepsy-protocol magnetic resonance imaging (MRI; 1.5 Tesla Magneto, Avento, 18 Channel; Siemens, Erlangen, Germany; 93 of 114 cases; 81.6%). When MRI showed evidence suggestive of active or inactive neurocysticercosis, previously proposed diagnostic criteria were applied to establish a diagnosis of cysticercotic infestation (Del Brutto et al., 2001). Clinical, EEG, and imaging data for all cases were double-entered into structured proformas and maintained in a database.
For each case, an age- and gender-matched, healthy (with no known neurologic disorder or history of seizures or epilepsy) control (n = 114) was randomly selected from the same study area. Age matching was ±2 years for age >10 years and ±1 year for age ≤10 years. All controls completed the screening questionnaire for epilepsy and were excluded if they screened positive. The controls also completed the questionnaires for risk factors.
Classification of epilepsies
Epilepsy was defined and categorized according to the epidemiologic criteria of the International League Against Epilepsy (ILAE) (Commission on Epidemiology and Prognosis, ILAE, 1993). Active epilepsy was defined as having had at least two epileptic seizures, including one in the previous 5 years, regardless of any antiepileptic drug (AED) treatment (Commission on Epidemiology and Prognosis, ILAE, 1993).
Blood sampling and serologic evaluations
Venous blood samples, collected from 106 cases (93%) and 114 controls were immediately separated and frozen at −80°C. Later, the sera were assayed to detect antibodies against T. canis using commercially available enzyme-linked immunosorbent assay (ELISA) (In Vitro Diagnostic Research, Carlsbad, CA, U.S.A.) and T. solium using an enzyme-linked EITB assay (Immunetics, Boston, MA, U.S.A.; Brunello et al., 1986; Tsang et al., 1989) at the Department of Microbiology, Dayanand Medical College. The T. canis ELISA used an excretory/secretory (ES) antigen from Toxocara larvae to screen for serum immunoglobulin G (IgG) antibodies. The Toxocara ES antigen-based ELISA for the detection of IgG antibodies was 78–91% sensitive and 86–93% specific for the diagnosis of toxocariasis (Speiser & Gottstein, 1984; Jacquier et al., 1991). The results were read using an ELISA reader at 450/650–620 nm. Absorbance reading ≥0.3 optical density (OD) units was considered positive. Toxocara canis–ELISA positive sera were further subjected to an immunoblot (Toxocara Immunoblot IgG; LDBIO Diagnostics, Lyon, France) assay (Magnaval et al., 1991). The immunoblot used Toxocara ES antigens separated into low molecular weight (24–35 kDa) and high molecular weight (70–90 and 200 kDa) bands. The presence of two or more low molecular weight bands was considered positive. Antibodies to antigens of low molecular weight (24–35 kDa) detected with western blot assay were considered highly specific for toxocariasis, thus avoiding problems of cross-reactivity with other helminthiasis (Magnaval et al., 1991; Rubinsky-Elefant et al., 2010). The EITB assay was conducted for the detection of T. solium IgG antibodies in serum with antigen-bearing nitrocellulose membrane using alkaline phosphatase as substrate. Antibodies against any of the six glycoprotein antigens of molecular weights 50, 42–39, 24, 21, 18, and 14 kDa were considered positive (Tsang et al., 1989).
Data were analyzed using STATA version 9 (StataCorp, College Station, TX, U.S.A.). Univariate comparison for the explanatory variables (T. canis and T. solium seropositivity), various baseline socioeconomic parameters, and risk factors for infection in cases and controls were first undertaken using the McNemar’s test (for matched case–control studies). A p-value of <0.05 was considered significant. Those variables for which p was <0.1 were entered into a multivariate analysis using fixed-effects conditional logistic regression to estimate an adjusted odds ratio.
The study was approved by the local institutional ethics committee. Informed consent was obtained from cases and controls (or parents or legal guardians in the case of children under the age of 12 years).
Sample size estimation
We assumed a population seroprevalence of T. canis exposure to be 20% based on epidemiologic surveys previously undertaken in India (Malla et al., 2002; Dar et al., 2008). An estimated 240 subjects (120 each of cases and controls) were required to detect an odds ratio (OR) of 2.0 with 80% power at a two-sided level of significance of 5%.
Of 151 individuals who were screen positive, 37 were excluded; 20 had inactive epilepsy, 2 had single seizures, 8 had febrile seizures, and 7 had nonepileptic seizures. Therefore, 114 cases (69 [61%] male) with active epilepsy remained. The crude prevalence of active epilepsy in the surveyed area was 7.2 per 1,000 population. Eight persons (7%) with epilepsy declined to provide blood samples, 13 (11.4%) did not undergo EEG examination, and in 21 (18%) cases, MRI could not be performed for various reasons (refusal, 14; anxiety, 3; pregnancy, 1; age <5 years, 3).
Demographic, socioeconomic, and risk factor profiles of the cases and controls are provided in Table 1 (see Tables S1 and S2 in Supporting Information). Sera of 7 of the 106 people with epilepsy (7%) and 8 controls (8%) demonstrated anti-T. canis antibodies using ELISA. Toxocara canis-immunoblot assay conducted on ELISA-positive sera confirmed exposure to the parasite in five cases and six controls. Taenia solium-EITB assay was positive in 27 (25%) of the 106 cases in whom it was tested and 13 (12%) of 106 matched controls (p = 0.02).
Table 1. Comparison of selected socioeconomic variables in cases and controls
|Caste|| || || |
| Upper||31 (29.2)||35 (33)||0.49 (ns)|
| Artisan or lower||75 (70.8)||71 (67)|
|Self/Husband’s occupation|| || || |
| Business or service||43 (40.6)||41 (38.7)||0.75 (ns)|
| Laborer||63 (59.4)||65 (61.3)|
|Education|| || || |
| Above matric||10 (9.4)||16 (15.1)||0.21 (ns)|
| Below matrica||96 (90.6)||90 (84.9)|
|Occupation|| || || |
| Working at home||68 (64.2)||62 (58.5)||0.35 (ns)|
| Serving outside for money||38 (35.9)||44 (41.5)|
|Mother’s education|| || || |
| Above matric||4 (3.8)||1 (0.9)||0.22 (ns)|
| Below matrica||102 (96.2)||105 (99.1)|
|Family type|| || || |
| Joint||16 (15.1)||20 (18.9)||0.48 (ns)|
| Nuclear||90 (84.9)||86 (81.1)|
|Family size|| || || |
| Large or medium||52 (49.1)||44 (41.5)||0.29 (ns)|
| Small||54 (50.9)||62 (58.5)|
|House ownership|| || || |
| Own houseb||95 (89.6)||78 (73.6)||0.006|
| Rented||11 (10.4)||28 (26.4)|
|Household assets|| || || |
| 10 or more||21 (19.8)||1 (0.9)||1.00 (ns)|
| Below 10||85 (80.2)||105 (99.1)|
|Type of houses|| || || |
| Puccac||64 (60.4)||58 (54.7)||0.43 (ns)|
| Kachad or mixed||42 (39.6)||48 (45.3)|
|Number of rooms|| || || |
| One||25 (23.6)||41 (38.7)||0.02|
| Two or three||81 (76.4)||65 (61.3)|
|Drinking water facility|| || || |
| Piped||95 (89.6)||82 (77.4)||0.02|
| Own or common hand pump||11 (10.4)||24 (22.6)|
|Nonvegetarian food consumption||66 (62.3)||49 (46.2)||0.02|
|Pork consumption||7 (6.6)||2 (1.9)||0.12 (ns)|
|Washing hands after defecation||104 (98.1)||106 (100)||1.00 (ns)|
|Washing hands before eating meals||104 (98.1)||106 (100)||1.00 (ns)|
|Feeding of stray dogs||36 (33.9)||56 (52.8)||0.01|
Conditional fixed-effects logistic regression analysis, adjusted for those baseline parameters for which p < 0.1 in the univariate analysis (Table 1), estimated an adjusted OR of 2.8 (95% confidence interval [CI] 1.2–6.8, p = 0.02) for seropositive status for T. solium. Two sensitivity analyses to account for the missing patients for whom the results of the EITB assay were not available were carried out: the ORs were recalculated, first assuming that all the missing cases were EITB positive and then assuming that they were all EITB negative. The recalculated ORs were 3.8 (95% CI 1.6–8.8; p = 0.002) and 2.8 (95% CI, 1.1–6.7; p = 0.02), respectively. The adjusted odds ratio for exposure to T. canis was 0.82 (95% CI 0.2–3.8; p = 0.8). Antibodies to both T. solium (using EITB) and T. canis (with immunoblot) were simultaneously detectable in three cases and one control.
MRI was undertaken in 93 (82%) of the 114 cases. Of 19 people (17%) with active epilepsy who had evidence of NCC on MRI, 7 (37%) had a solitary cysticercus granuloma, 7 (37%) solitary calcification, and 5 (26%) had multiple active and inactive calcified parenchymal cysticerci. Eight (42%) of the 19 with MRI evidence of NCC were seropositive for T. solium antibodies and 2 of the 19 were Toxocara ELISA positive (of which one was confirmed positive by Toxocara canis-immunoblot assay) (Table 2).
Table 2. Correlation between imaging findings and serologic studies in people with epilepsy and evidence of neurocysticercosis (n = 19) on MRI
|Solitary cysticercus granuloma||7 (36.8)||1|| ||2|
|Solitary calcification||7 (36.8)|| || ||4|
|Multiple lesions||5 (26.3)||1||1||2|
The population assessed was carefully selected, as there were stray dogs and pigs in the community and it typically represented a low socioeconomic area. Most inhabitants were daily wage workers, often migrant laborers. Hence, the community represented a population that was at high risk for acquiring both T. solium and T. canis infections. We were able confirm an association between T. solium seropositive status and epilepsy, but no such association was found between T. canis seropositive status and epilepsy. This is in contrast with recent data from rural-community studies in Bolivia and Burundi and a hospital-based study from Italy, which suggested an association between T. canis exposure and epilepsy (Nicoletti et al., 2002, 2007, 2008). In the Bolivian study, the adjusted OR was 2.70 (95% CI 1.4–5.2), in the Burundi study it was 2.1 (95% CI 1.2–3.8), and in the Italian study it was 3.9 (95% CI, 1.9–8.0). There are many possible explanations for the differences in the findings of our study and those of the previous studies, including differences in population attributes (such as genetic make-up, risk-factor profiles and behaviors, and low endemicity) and methodologic characteristics (e.g., exclusion of single seizures or inactive epilepsy). In two of the previous studies, there appeared to be a high risk of exposure in the community (12% in Bolivian study and 51% in Burundian study) compared with our population in which the seropositivity rate among controls was 5.7% (Nicoletti et al., 2002, 2007). The baseline seropositivity rates for the community studied have not been determined, but rates in other communities in India are in the order of 6–33% (Malla et al., 2002; Mirdha & Khokar, 2002; Traub et al., 2002, 2005; Dar et al., 2008). Hence, while estimating sample size a baseline seropositivity rate of 20% was assumed.
Our study confirms the association between T. solium parasite exposure and epilepsy in the field setting. The association has been established in rural South India and in several rural studies in Latin America (Del Brutto et al., 2005; Montano et al., 2005; Nicoletti et al., 2005; Rajshekhar et al., 2006). Seropositivity was higher in people with active epilepsy in our study (25%) than in the study from rural South India (13%) (Rajshekhar et al., 2006), although there were differences in serologic methods employed in the two studies.
We used MRI to study the association between epilepsy and cysticercosis. Computerized tomography (CT) was used in the population-based studies from Latin America (including Bolivia, Peru, and Ecuador) and in these studies, high frequencies of single or multiple calcifications were observed in both people with epilepsy and asymptomatic individuals (Del Brutto et al., 2005; Montano et al., 2005; Nicoletti et al., 2005). In the field studies from Latin America, single or multiple calcific lesions were found in about two thirds of people with NCC. The remainder were either live, active (vesicular) cysticerci or, very rarely, single cysticercus granulomas (Del Brutto et al., 2005; Montano et al., 2005; Nicoletti et al., 2005). In our study, more than one third of all NCC cases had solitary cysticercus granuloma and another one third had calcific NCC. In the study from South India, 7% of people had solitary cysticercus granuloma (Rajshekhar et al., 2006), whereas in a study from a pig-farming rural community, about one fourth of people with NCC had solitary granuloma (Prasad et al., 2008). Therefore, a single cysticercus lesion in the granulomatous stage appears to be a common finding in field studies from India but not from Latin America. This could be due to differences in the imaging methods used. Two of the Latin American field studies used routine (noncontrast) CT scanning, whereas one used contrast-enhanced CT scans. It is possible that granulomas were missed on routine CT scans, as the administration of contrast is required to visualize granulomas on CT scans. On the other hand, the larger proportion of solitary cysticercus granulomas in our study as well as other studies from India than in studies from Latin America could be due to geographic and ethnic differences between the two populations (Del Brutto et al., 2005; Montano et al., 2005; Nicoletti et al., 2005; Rajshekhar et al., 2006; Prasad et al., 2008). Indeed, several hospital-based studies from India have shown a high frequency of solitary cysticercus granulomas in people with seizures and epilepsy (Chandy et al., 1989; Murthy & Subba Reddy, 1998). Solitary cysticercus granulomas constitute a relatively lower proportion (approximately 20%) of hospital-based cohorts of NCC from Latin America (Del Brutto, 1995). Theoretically, a solitary granuloma represents degeneration occurring at a relatively early stage in evolution of the cysticercus compared with a live, active cyst, which may suggest immune evasive mechanisms developed by the cysticercus over long periods. It may be that a lower disease burden and lesser degree of exposure to the parasite in the Indian subcontinent leads to a relatively early degeneration of the cyst and granuloma formation (Garcia et al., 2010).
Of interest, of the 27 T. solium–seropositive cases, MRI (undertaken in 96%) demonstrated evidence of NCC in 8 (30.8%). This could be due to infestation in the past that has since resolved, extraneural cysticercosis (e.g., in the muscle or skin), or residual-calcified NCC, which could be missed on the MRI.
Limitations of the study include a potential selection bias and the potential statistical under powering of the study. The particular population in our study was selected mainly due to convenience, as catchment area formed the field practice area of the medical college and this may represent a selection bias. Another potential source of bias might be the large proportion of migrant population, the proportion of which is higher than the average proportion of interstate migrants in India (14%) (Government of India, Ministry of Home Affairs, 2001). We do, however, feel that these are not sufficient to have changed significantly the results. While calculating the sample size for the study, we assumed a baseline seropositivity rate of 20%. Because we found a low seroprevalence for T. canis, it is possible that the chosen sample size was insufficient to detect a difference between cases and controls. However, in view of the lower seropositivity in cases than in controls, it is unlikely that a larger sample size would have suggested an association between T. canis exposure and epilepsy.
This population-based survey of active epilepsy in a slum area in India failed to confirm an association between exposure to T. canis and epilepsy. A significant association between T. solium exposure and epilepsy, however, confirms that this parasite is an important risk factor for epilepsy. The imaging spectrum of NCC in the community in India comprises a high proportion of individuals with solitary cysticercus granuloma, a finding that is in contrast to the spectrum observed in Latin America, where calcific lesions are the most common finding. The reasons for the high frequency of solitary granulomas in India need to be determined.
This research project was supported by the Indian Council of Medical Research vide Project No. 5/4-5/19/Neuro/2008-NCD-I. JWS is supported by the Marvin Weil Epilepsy Research Fund and is based at UCLH/UCL, which receives a proportion of funding from the United Kingdom Department of Health’s National Institute for Health Research Biomedical Research Centres’ funding scheme. We gratefully acknowledge Dr. Gail Bell, for expertly reviewing the manuscript; Pratibha for data entry; and field workers Manpreet and Rajwinder for the survey. Dr. Andy Hall, London School of Tropical Medicine and Hygiene and Dr. A. Nicoletti, University of Catania, Italy offered suggestions in planning the project.
GS is a member of Infectious Diseases Society of America (IDSA), Guidelines Committee for management of cysticercosis and has received grants from the Indian Council of Medical Research for various cysticercosis-related projects. The other authors have no conflicts in relation to this work. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.