Neurocysticercosis and cognitive impairment among people with epilepsy in Taenia solium endemic regions of rural southern Tanzania: A hospital-based cross-sectional study in mental health clinics of selected sites in Tanzania

for Kongwa participants than Chunya, with a statistically significant association (95% CI: 1.75, 156; p = 0.037). Additionally, having and education was associated with a 91% reduction in the odds of cognitive impairment (OR = 0.09) compared to no education, which was also statistically significant (95% CI: 0.01, 0.33; p = 0.002). There was no association between cognitive impairment and NCC. Conclusion: Our study found a 22.2% prevalence of NCC among PWE. Cognitive impairment was present in 11.8% of PWE but was not significantly associated with NCC. Socioeconomic and educational factors may play a larger role in cognitive impairment among PWE.


Introduction
Epilepsy is a chronic neurological disorder that occurs throughout the life course, increasing with age with a peak in early and late life, and manifests with different types of epileptic seizures and other comorbidities [1].
Epilepsy can be idiopathic (genetic and/or with no obvious brain lesions) or secondary to obvious brain lesions as well as metabolic, infectious, immunological, and drug/alcohol-related causes.The most comprehensive data on the burden of epilepsy overall comes from the 2016 GBD study of epilepsy that estimates almost 46 million patients with all-active epilepsy (both idiopathic and secondary epilepsy globally) with an age-standardized prevalence of 621.5 per 1,00,000 population [2].Causes of epilepsy in low-income and middle-income countries (LMICs) are different with infectious disease still contributing a large part.In our study, we focus on the most prevalent helminthic cause of secondary epilepsy in LMICs, which is Taenia solium (neuro) cysticercosis/taeniasis (TSCT).
Neurocysticercosis (NCC) is the cause of approximately 30 % of epilepsy cases in TSCT endemic areas as was the case in our study region in Southern Tanzania [3,4].T. solium egg ingestion through contaminated hands, food, or water due to a lack of sanitation and hygiene may lead to NCC, the most important helminthic infection of the central nervous system.Approximately 80 % of NCC lesions are clinically asymptomatic but can become symptomatic at any time.The location of the cysts in the brain (intraparenchymal or extraparenchymal), the number and size of the cysts, and whether the cyst is viable or inactive, together with the host's immune response, determine the neurological manifestations of NCC [5][6][7][8].The most common neurological manifestations of NCC are epileptic seizures, either in their acute or chronic form, i.e., epilepsy [5,6,9].The mechanisms of recurrent epileptic seizures in the context of NCC-associated epilepsy include not only the space-occupying lesion itself, but also the overall inflammatory process with its diverse immune responses, perifocal edema and/or gliosis, calcifications, which have been shown to possess epileptogenic properties themselves, and hippocampal sclerosis, which can lead to temporal lobe epilepsy and/or memory impairment [9,10].
Epilepsy is not only defined by epileptic seizures but also by comorbidities of a cognitive and psychological nature and their social, cultural, economic, and political implications.Dhakal and Bobrin (2023) understand cognition or cognitive function as follows: "Cognition is the mental action or process of acquiring knowledge and understanding through thought, experience, and the senses.It encompasses various aspects of high-level intellectual functions and processes such as attention, memory, knowledge, decision-making, planning, reasoning, judgment, perception, comprehension, language, and visuospatial functions.Cognitive processes use existing knowledge and generate new knowledge" [11].
In simple terms, a person with cognitive impairment has difficulty remembering, picking up new information, focusing, or making judgments that have an impact on their daily lives [12].Cognitive impairment ranges from mild to severe and its transition into dementia can be fluid.Epilepsy can lead to cognitive impairment, which has also been shown in different African countries, including Tanzania.Kariuki et al. (2014) within a multi-country African study defined cognitive impairments as a person's awareness of person, place, and time and their ability to follow standardized instructions during the neurological examination [13].Numerous combinations of variables, such as early onset of epilepsy, frequency, intensity, and length of seizures, as well as medications especially polytherapy of ASM, may affect cognitive function in PWE [14,15].The NCC itself can also contribute to cognitive impairment, especially in the context of acute symptomatic seizures and epilepsy.Studies from Latin America reported cognitive impairment to be more pronounced in PWE and NCC compared to PWE without NCC and healthy controls [16][17][18].However, in a study in Zambia, using a neuropsychological test battery, there was no difference in cognitive impairment between PWE with and without NCC [19].
Clinical examination and the evaluation of the mental state most times is the first approach to diagnosing cognitive impairment and are often corroborated by the patient's personal history and the collateral history of a family member.Ideally, this should be combined with standardized neuropsychological testing, which needs appropriate tools for the social and cultural context.There are several known and validated tools like the Mini-Mental State Examination (MMSE), and the Montreal Cognitive Assessment (MoCA).However, data from LMICs is patchy and both tests have been validated in sub-Saharan Africa (SSA) [20][21][22].This includes Tanzania, where both MMSE and MoCA have been used in individuals with traumatic brain injury admitted to a tertiary hospital [21].The MMSE was also used in Tanzania for the identification of dementia among the elderly in a low-literacy setting in rural Tanzania and Cameroon among the elderly rural population [20,22].Others include the Selective Reminding Test (SRT) and Spatial Recall Test (SPART) [19], which have also been validated in the African context in Zambian and even translated into the local Chewa language [19].
However, the vast majority of cognitive assessment tools have been designed for high-income countries (HICs) settings and the best cognitive assessment tool in a particular setting might have low validity and reliability in another setting with different socio-cultural aspects as well as literacy rates [23].In the current study, cognitive assessment was performed as part of the neurological examination using contextualized, standardized questions with trained assessors.Overall, this study was conducted to assess the prevalence of NCC among PWE in a rural setting of southern Tanzania, to describe their cognitive impairment and its associated risk factors.

Study design and setting
This was a multicenter hospital-based cross-sectional study.Study participants were PWE living in the catchment areas of Kongwa and Chunya district hospitals in Dodoma and Mbeya districts in rural Southern Tanzania and attending mental health clinics for the management of their epilepsy.This study was a part of the CYSTINET-Africa project and study sites were the same as described by Makasi et al. (2023) [24].
Chunya District is part of the Mbeya region in the southern highlands of Tanzania, covering an area of 17,505 km 2 .The population of Chunya was 3,44,471 inhabitants, with males amounting to 1,76,457.The annual population growth of Chunya is 8.2 % according to Census 2022 [25].Ethnic groups residing in the Chunya district include Kimbu, Bungu, Guruka (semi-Safwa), Nyakusa, Nyamwezi, Sukuma, Barbaig, and Hehe.The main source of income for people here is subsistence farming.Chunya district's popularity is based on mining activities.Mostly gold mining is carried out by small-scale miners.There is a total of 33 healthcare facilities: 29 dispensaries, 3 health centers, and 1 district hospital.There are around 30 public health facilities, and 3 are private health facilities.
Kongwa District is part of the Dodoma region of Tanzania, covering an area 3,958 km 2 .The population is 4,43,867 inhabitants of which males are 214,475 and the annual population growth for the district is 3.7 % [25].The main source of income for Kongwa is small scale farming, but the district is also popular for livestock farming.The two main dominant ethnic groups in Kongwa are the Gogo and Kaguru.There is a total of 72 healthcare facilities: 66 dispensaries, 5 health centers, and 1 district hospital.There are around 63 public health facilities, including one is owned by the Prison system, one held by a parastatal entity, 3 are private health facilities, and 4 are owned by faithbased organizations.
Kongwa and Chunya share comparable traits when it comes to raising free-range pigs.Chunya's mining, forestry, and fishing industries are indicative of a lower poverty level than Kongwa's, yet official data was not available to support this.

Procedures
For Kongwa District Hospital, one researcher (CM) and the doctor in charge of the mental health clinic (MN) attended to all individuals who reported at the mental health clinic from July 2020 to April 2021.The approach was the same in Chunya District Hospital with one researcher (SN) and the doctor in charge of the mental health clinic (DM) being involved in participant recruitment.Mental health clinics in Tanzania serve different kind of mental health problems.Patients attending the mental health clinics include individuals with depression, schizophrenia, bipolar disorders and other brain disorders like dementia and epilepsy.We visited mental health clinics in Kongwa and Chunya district hospitals and for our study only included PWE.We therefore carried out a convenience sampling in which all people who were already known to the clinic with a diagnosis of epilepsy or newly diagnosed patients were invited to take part in the study.Epilepsy was defined as a health condition marked by repetitive (two or more) epileptic seizures that are not due to any acute intracranial or extracranial pathologies [26,27].In this study, active epilepsy was defined as being on ASM or having had the most recent seizure during the last five years [26,27].

Data collection methods and tools
Potential participants were informed about the research project and given some time to decide on their participation; informed consent was then obtained from each participant before being included in the study.This was followed by the completion of the eligibility form which also mentioned the inclusion and the exclusion criteria.Included were PWE on or off ASM attending the mental health clinic; aged 18 years and above; residing in the study catchment area, i.e. in Kongwa and Chunya districts; agreeing to participate in the study and complying with the study procedures, including cerebral computed tomography (CT) examination (initial CT examination and follow-up) and treatment of NCC, if required.Excluded were PWE with a current history of mental health disorders such as depression or schizophrenia; those mentally incapacitated not being able to follow instructions; those with a history of major mental health disorders; those seriously ill (physically) and pregnant women (based on history, clinical examination and, in case of uncertainty, a pregnancy test).
Data was collected with electronic data collection software using a portable mobile device, i.e.Tablet computer.The KoboToolbox software program [28] was used to construct an electronic database and to collect the relevant information within the larger CYSTINET-Africa study.Participants who qualified to be included in the study were presented with a demographic questionnaire, an in-depth neurological questionnaire and a neurological examination questionnaire.At each site, a welltrained researcher and medical practitioner (CM for Kongwa and SN for Chunya) attended to patients during the clinic days and recruited suitable participants according to the inclusion and exclusion criteria mentioned above.All three tools (demographic questionnaire, in-depth neurological questionnaire and neurological examination questionnaire which also included the mental state examination part) were applied by the same practitioner at each study site using the KoboToolbox software.Mental state (orientation, appearance, behavior, cooperation, speech, mood, and cognition (appropriate variables for our context were chosen from the standard full mental state examination [29]) and neurological (cranial nerves, muscle strength, muscle tone, reflexes, sensation, and coordination) examinations were performed.Before using the electronic questionnaires, CM and SN who are medical doctors working with CYSTINET-Africa, received extra training from doctors with neurological experience (ASW, DS).The clinical cognitive assessment was performed by CM and SN as part of the mental state examination and the categories as mentioned in Table 1 were established.
The assessor (CM or SN) reported whether it was normal or abnormal and in case of any declared abnormality, it had to be fully explained via the KoboToolbox.Results were discussed with the study team, foremost with ASW, a trained neurologist.Unfortunately, the assessment tool did not give room for grading of the severity.

Serological diagnosis of cysticercosis and cerebral computed tomography examination
Blood samples from all recruited PWE (n = 223) collected from Kongwa and Chunya were shipped to Mbeya Referral Hospital.The blood samples were then examined for cysticercosis (CC) Ab and Ag using the LDBio Western blot (LDBio Diagnostics, Lyon, France) and the monoclonal Ab-based B158/B60 ELISA (apDia, Turnhout, Belgium), respectively.
PWE who were positive in either of the CC serological tests were offered a CT examination.Those who were found to have other neurological signs/symptoms in addition to epilepsy during the neurological examination were also invited for CT examination, even if they were seronegative.Those who were found to have active cysts in their brain

Table 1
Components of the clinical cognitive assessment during mental state examination.

Category
Description (s) Question(s) asked

Simple arithmetic cognitive assessment
Simple arithmetic considering an individual's age, education status, and the possibility of illiteracy.
Evaluation involved asking math problems (basic operations in mathematics).

Recent memory cognitive assessment
Recent event recall whereby the participant was asked about recent experiences or occurrences.
The questions asked were about participants' breakfast, activities, and visitors during the day and the previous day.

Remote memory cognitive assessment
Famous personalities or particular childhood events The most common item/question asked was about the history of previous presidents of Tanzania.
were assessed for treatment and, if eligible, treated within the CYSTINET-Africa study.

Diagnosis of neurocysticercosis
Participants from Kongwa were taken to Dodoma Mbijiwe Diagnostic Center and those from Chunya to the Zonal Referral Hospital in Mbeya for their CT examination.A GE Revolution CT scanner with 32 slices and a 1.5 mm slice thickness was used to perform the CT examination at Mbeya Zonal Referral Hospital and a Philips CT scanner with 6 slices and a 2.5 mm slice thickness was used in Dodoma Mbijiwe Diagnostic Center.The CT scans were examined by a neuroradiologist with NCC expertise (CR).The locations of the lesions were classified as extraparenchymal (intraventricular or subarachnoid) or parenchymal (frontal lobe, temporal lobe, parietal lobe, occipital lobe, cerebellum, brainstem).We diagnosed NCC based on the 2017 updated Del Brutto criteria.Using exposure, clinical, and neuroimaging criteria, NCC was categorized as either definitive or probable [30].The clinical study team (CM, DS, CR, BJN, and ASW), including a neurologist (ASW), discussed all patients seen with NCC and decided together on a management plan.

Data analysis
Data compilation was done through the software program Kobo-Toolbox using a Tablet computer.Data was uploaded to the server located centrally at NIMR-Muhimbili in Dar es Salaam whenever the Tablet computer was connected to power.This was followed in regular intervals by data quality checks, data cleaning and identification of missing variables.Participants with missing variables due to missed neurological/mental state examination could not be included in the cognitive impairment analysis, i.e. from the original 223 participants, only 221 were included in the analysis.
Data were analyzed descriptively where count and proportion were used to show the distribution of the demographic, clinical, and social characteristics, neurological outcomes, and neuroradiological presentation of the parasite's stages.Chi-square test/Fisher's exact test was used to test for a significant difference between demographic, clinical, and social characteristics on the one hand and the different recruitment sites on the other hand, between neurological examination results and normal/abnormal cognition and between CC status and normal/ abnormal cognition.The prevalence of PWE presenting with NCC was calculated from those who went for CT scan only.The CC status was used for stratification of the cognition, which was expressed as normal/ abnormal.If one or more of the different components, arithmetic cognitive, remote cognitive, and recent cognitive (Table 1), was abnormal according to educational level, which was ascertained in each patient during history taking, cognitive impairment ( = abnormal cognition) was diagnosed.Other mental state deficits such as disorientation were not used in the categorization of the cognitive impairment but related to whether or not cognitive impairment was present.Data analysis was done by using Stata version 14 (StataCorp.2016.Stata Statistical Software: Release 14. College Station, TX: StataCorp LLC).
Logistic regression was used to assess the association between potential predictors and cognitive function impairment.A univariate logistic regression model was used to estimate crude odds ratios for each of the potential predictors.Multivariate logistic regression was used to assess the association between factors associated with cognitive impairment while controlling for confounding variables.The factors included in the multivariate logistic regression were selected using the stepwise selection (forward and backward) method to choose the best model.The Akaike Information Criterion (AIC) was used to compare models in each step until we reached the best model that produced the minimum AIC.

Ethics statement
Ethical approval was granted by The Kilimanjaro Christian Medical College Research Ethics and Review Committee (CRERC); Certificate No 2450.The CYSTINET-Africa study proposal received ethical clearance from the National Institute for Medical Research (NIMR) with reference number NIMR/HQ/R.8a/Vol.IX/ 2529 and the Technical University of Munich, Klinikum rechts der Isar, Ethics Committee with reference number 537/18.Approval to conduct this study in the selected sites was obtained after meetings with local administrative authorities to explain the study's aims and procedures.Individual written informed consent was sought from all participants involved in the study.

Results
A total of 510 patients attending the mental health clinics were seen consecutively, with 223 PWE enrolled in the study (Fig. 1).All study participants were recruited from the mental health clinics of Kongwa and Chunya district hospitals between July 2020 and April 2021.Two hundred and eighty-seven people were diagnosed with other mental health disorders including depression, schizophrenia, bipolar disorders, and other mental health disorders.Only PWE (n = 223) were recruited and offered serological testing for CC positivity/negativity of which 25 (11.2 %) were CC positive (either Ag or Ab or both).However, only 221 PWE underwent neurological/mental state examination and therefore PWE who were positive on serology, none of which underwent CT examination, and 22 PWE who were negative on CC serology; only one of which showed neurological signs/symptoms in addition to their epilepsy and therefore underwent neuroimaging and was diagnosed with NCC (Fig. 1).

Sociodemographic characteristics and cysticercosis/ neurocysticercosis prevalence of people with epilepsy
A total of 223 PWE were recruited from both, Kongwa and Chunya mental health clinics; whereby, 172 (77.1 %) came from Kongwa.Overall, 118 (52.9 %) were female and about two-thirds (67.7 %) of the recruited participants were below 40 years of age, 182 (82.4 %) were farmers (peasants), 108 (48.4 %) were single and 77 (34.5 %) did not have formal education.The distribution of age, sex, and religion was very similar between PWE from Kongwa and Chunya.However, compared to Chunya, Kongwa had more than twice as many PWE without formal education (p = 0.002) and twice as many PWE who were farmers (p < 0.001).Also, more people in Chunya had CC+Ag-tests compared to those in Kongwa (17.6 % vs 3.5 %; p < 0.001), although people in Kongwa and Chunya had similar percentages of CC+Ab-tests (7.6 % vs 7.8 %).Overall, in our study population of PWE from southern Tanzania, 11.2 % had serologically diagnosed CC (8.7 % in Kongwa vs 19.6 % in Chunya; p = 0.030) and 22.2 % had neuroradiologically diagnosed NCC (11.5 % in Kongwa vs 50.0 % in Chunya; p = 0.024) (Table 2).The 8 PWE with NCC had a median age of 37 with a range of 23-65.Five (62.5 %) were female and they were all farmers.

Outcomes of mental state and neurological examination in people with epilepsy
Mental state and neurological examinations were performed in 221 PWE; two participants did not attend to it.Cognitive impairment as defined above (Table 1) was diagnosed in 11.8 % (26/221) of PWE (Table 3).Of those, impairment in the categories "arithmetic cognitive", "recent cognitive" and "remote cognitive" were found in 23, 2, and 26 PWE, respectively.Abnormal results were seen in at least one, two or three categories in 16 PWE, 8 PWE, and 2 PWE, respectively.Considering other mental state components, significant differences were found with behavior (34.6 % vs 0.5 %, p < 0.001), cooperation (50.0 % vs 0.0 %, p < 0.001), and speech (53.8 % vs 2.6 %, p < 0.001) when comparing PWE with cognitive impairment to those without cognitive impairment (Table 4).Neurological examination including cognitive assessment showed abnormal results altogether in 28 PWE (26 PWE with and 2 PWE without cognitive impairment; Tables 3 and 4).

Cognitive impairment in people with epilepsy and its relationship with cysticercosis and neurocysticercosis
Twenty-five PWE were CC+ (one or both serological test results positive) and 196 CC-(both CC serological tests negative).For the 26

Table 3
Cognitive impairment and/or neurological signs/symptoms in people with epilepsy (N = 28).*Speech impairment is based on neurological grounds and is different from speech impairment mentioned in Table 4 which is part of the mental state examination.; p = 0.5) (Table 5).A total of 36 participants were examined with CT, of whom 18 (50.0%) were positive on any of the CC serological tests or both.Of those 18 PWE who were CC+, 7 (38.9%) showed NCC-typical lesions on neuroimaging.One hundred and ninety-six PWE were negative on both of the CC serological tests, and 18 PWE who were CC-went for CT examination because they had neurological signs/symptoms of which one was diagnosed with NCC (Fig. 1 and Table 3).There were 8 PWE with NCC; 7 of them were CC+ and had no cognitive impairment, and only one of them who was CC-was found to have cognitive impairment (Fig. 1 and Table 6).The total number of NCC-associated lesions for all 8 PWE with NCC was 240, including active, mixed, and inactive NCC.Altogether, there were 79 lesions in the vesicular, 6 in the degenerative, and 155 in the calcified stage.Overall, one patient had active-stage lesions only, four had mixed lesions and three had calcified lesions only.The one patient with cognitive impairment had only calcified lesions (n = 76).The other seven patients without cognitive impairment had in the median 22 lesions (IQR 14-36) (Table 6).The relationship between the different stages of the parasite, positive/negative CC serological results, and presence/absence of cognitive impairment is given in Table 6.

Prevalence of cysticercosis and neurocysticercosis
Our current study showed that the prevalence of CC and NCC among PWE recruited from mental health clinics in southern Tanzania was 11.2 % and 22.2 %, respectively.The prevalence was drawn from only those who went for CT examination (including PWE who were CC+ and those with neurological signs/symptoms as suggested by the assessing physician).Results from a study in the districts of southern highlands of Tanzania showed more than 30 % of PWE admitted to Vwawa, Tukuyu, and Ifisi district hospitals, had NCC lesions in their brain (38 % in Vwawa, 32 % in Tukuyu, and 31 % in Ifisi) [31].Results from an epilepsy-clinic-based study in northern Tanzania showed 10.4 % of the recruited PWE with probable and 3.3 % with definitive NCC [3].A pooled analysis from three countries and four different study sites in PWE with NCC in Tanzania (Dar es Salaam and Haydom (rural Tanzania)), Uganda, and Malawi showed a variation of NCC prevalence estimates in PWE among sites ranging from 2.0 % (95 % CI 0.4 % to 3.6 %) in Dar es Salaam to 17.5 % (95 % CI 12.4 % to 22.6 %) in Haydom [6].The studies in Uganda and Malawi were community-based, and the two studies in Tanzania were hospital-based.A systematic review and meta-analysis of 12 studies from Bolivia, Brazil, Columbia, Ecuador, Honduras, Peru, India, Burkina Faso, the Republic of South Africa, and Tanzania showed that the pooled prevalence of NCC among PWE of all ages was 29.0 % (95 % CI: 22.9 %-35.5 %) [4].Considering local and more global study results, the prevalence of NCC in PWE in our study seems to be lying more in the middle.
Interestingly, CC and NCC were more prevalent in Chunya compared to Kongwa, although Kongwa seems to have a lower socioeconomic status.Cysticercosis and NCC depend on the presence of T. solium taeniasis in the community as its presence contributes to the contamination of the environment with the infective eggs.It could be that in Chunya because of the gold-digging activities and the related social activities a lot of (infected) pork is consumed and with the still rather low level of hygiene and freely roaming pigs, this could lead to a perpetuation of infection in terms of T. solium taeniasis but also CC and NCC (see introduction).Although there is an association between socioeconomic status and the prevalence of CC/NCC, the relationship is complex, and multiple factors contribute to the spread of the disease.Cross-cutting targets like full access to at least basic water supply, sanitation, and hygiene need to be complemented by education and access to healthcare services [32].A study by Makasi et al. (2023) showed low knowledge of the relation between T. solium and epilepsy in both districts of Kongwa and Chunya [24].Also, the state of the health services in both districts is essentially similar.Health facilities in both districts cannot perform immunodiagnostic tests for CC or diagnostic neuroimaging for NCC.This is because of a lack of equipment (serological tests and neuroimaging) and a skilled health workforce for T. solium-related health issues.This calls for increased public health activities around TSCT using the One Health approach, considering aspects of human, animal and environmental health for joint agenda setting.

Cognitive impairment in people with epilepsy and those with neurocysticercosis
Our current study shows that 11.8 % of all recruited PWE were found to have cognitive impairment.This is rather low compared to numbers from resource-rich settings which have been quoted with 70-80 % [33,34].Cognitive diagnostic tools under those circumstances may be more sophisticated and likely pick up more nuanced cognitive impairment, including deficits in attention, memory, language and visuospatial orientation, amongst other deficits, whereas our cognitive assessment was very basic, part of the overall clinical examination and non-standardized.Also, cognitive impairment in PWE depends on seizure type, onset of seizures, intake of ASM, co-morbidities, brain lesions (such as in NCC), and other factors [33,34].Interestingly, NCC was not found to have an association with cognitive impairment in our study, which could be due to only 8 PWE being diagnosed with NCC.The tests that were used to assess cognitive impairment were part of the mental state examination, which, in turn, was part of the neurological/clinical examination.We did not use specific cognitive function tests, because the study was not a priori designed for comprehensive cognitive assessment but was part of the larger CYSTINET-Africa project, where the emphasis was on the burden of NCC in people with and without epilepsy.Our chosen tests broadly target memory function but are not comprehensive and may have been prone to bias based on the examiner's experience, deficits in orientation and attention, lack of cooperation, and mood impairment in participants during the examination (Table 4), and the individual's education level.Indeed, an association between the outcome of cognitive impairment and education level was found in our study (Tables 7 and 8).
In general, there is a lack of data concerning cognitive impairment in SSA, and data get even more scarce when it comes to cognitive impairment in PWE, let alone in PWE and NCC.There seems to be only one study from SSA assessing cognitive performance in PWE with and without NCC in Zambia [19].Within this cohort, epilepsy was found to be associated with cognitive impairment and reduced quality of life, and PWE due to NCC showed similar performance compared to PWE due to other causes.Despite a rather comprehensive assessment of cognitive impairment with a test battery of 4 tests, including Mini-Mental State Examination (MMSE), Digit Span, Selective Reminding Test (SRT), Spatial Recall Test (SPART), and Test Battery of Attentional Performance (TAP), the study design showed several limitations, including a small cohort of PWE compared to our study (n = 47 versus n = 223) [19].More comprehensive assessments come from South America with larger study populations with up to 224 NCC cases [35][36][37].According to these studies, NCC seems to be associated with a wide spectrum of cognitive abnormalities, ranging from lower performance levels in single tests to manifest dementia.Additionally, cognitive impairment seems more pronounced in the active vesicular stage and with a greater number of lesions, which was confirmed by an Indian study assessing 43 NCC cases [38].Nevertheless, cognitive test batteries and control groups showed large variations in those studies, including matching with healthy controls only, without comparison to PWE due to other causes.Consequently, there is a need for further studies addressing this data gap on cognitive impairment in NCC, especially in SSA.
This call poses a challenge, as there is a general lack of data on cognitive assessment in SSA, and also if one applies a broader, non-NCC, and non-epilepsy-specific context.Overall, there are only a few studies with a structured assessment of cognitive function in SSA, and many of them were conducted in the context of (cerebral) malaria and with a focus on children and the use of specific pediatric assessment scales [ 39,40].Moreover, these studies are not evenly distributed but concentrated on a few countries like Kenya [41][42][43][44].In general, applied cognitive tests and definitions of cognitive impairment in SSA vary widely [22,45,46].Some studies did not use specific cognitive assessments at all but based cognitive impairment on a person's orientation to person, place, and time, and their ability to follow standardized instructions only [13].Like in our study, this is probably due to the resource-limited setting, but also reflecting the heterogeneity of study designs and study populations, which originate in the heterogeneity of the African continent itself, including a huge variety in ethnicities, cultures, socio-economic status, and, last but not least, languages and dialects.In this context, it is of note, that the majority of the applied cognitive assessment tools such as MMSE, SRT, SPART, and others in SSA, do not originate from local settings but are transferred from highincome countries and are standardized and validated in populations of high-income settings, many of them without validation in specific African populations.Moreover, the frequently used MMSE for example, requires reading and calculating abilities, which poses a challenge in some patient groups, especially in rural African populations [19,47,48].Only in a subset of studies, adaptions of the applied tools with respect to the local context have been made [49,50].Moreover, there are only a few studies with more sophisticated tools such as full batteries of cognitive tests designed and validated to meet the needs of the population from SSA.One consists of a neuropsychological assessment battery for the measurement of cognitive function in children from Africa (Tanzania and Ghana) and Asia (Bangladesh), including general intelligence, executive function, and development of literacy skills [49], and the other one is the African Neuropsychology Battery, which includes different memory tests and was used among individuals of African descent in Congo and USA [45].Overall, the studies showed that the test batteries or specific sub-tests thereof produced valid and reliable results within African populations and between different neurodegenerative diseases, but the performance of the chosen tests between different cultures still meets with challenges [45,49].

Risk factors associated with cognitive impairment in people with epilepsy
Risk factors for cognitive impairment in PWE have been identified but have mainly been assessed in HICs.They include age, age of onset of epilepsy, type of epilepsy/epilepsy syndrome, duration of epilepsy, seizure frequency, intake of ASM and its duration, type of ASM and any combination (polypharmacy), comorbidities, brain lesions and others [14,15,33,34].In the univariate regression analysis in our study, there was a suggestion that location, education, NCC (curiously with an inverse relationship), and herbal treatment may be associated with cognitive impairment.However, in the multivariate logistic regression analysis, this could only be confirmed for location and education.Age, sex, as well as epilepsy-, seizure-, NCC-and ASM-related factors did not seem to play a role in our study population, although, as stated above, our study population of PWE with cognitive impairment (n = 26) and PWE and NCC (n = 8) was rather small.There was also a suggestion that herbal treatment may be associated with cognitive impairment in PWE.In the setup of Chunya and Kongwa, knowledge about the causes of epilepsy is rather low; it is often believed that epilepsy is caused by evil spirits, a curse or witchcraft [24].Therefore, PWE frequently opts for traditional treatment, delaying treatment with ASM, but the association with herbal treatment was not robust.However, the association between location and education was robust and can be explained by the lower socioeconomic status of Kongwa which may entail a lower overall education level, which, in turn, may reflect on the outcome of cognitive testing.Some school-based education seems to be relevant to answer the questions related to the category "simple arithmetic cognitive assessment" and "remote memory cognitive assessment" but not so much related to "recent memory cognitive assessment" (Table 1) and indeed, PWE with cognitive impairment performed worse in the former two test categories (Table 5).This indicates that education may play a role in the outcome of cognitive testing in PWE, but more sophisticated cognitive testing in a larger SSA population would be necessary to ascertain relevant risk factors for cognitive impairment in PWE in SSA including in the context of NCC.

Strengths and limitations
There is hardly any data on cognitive impairment in PWE in SSA most likely because of the dearth of suitable, contextualized cognitive assessment tools, trained psychologists, psychiatrists or neurologists and a prevailing lack of interest in brain health in many LMICs, which still have their major focus on infectious diseases, although noncommunicable diseases including brain disorders are on a steep rise [51].In HICs, it has been demonstrated that cognitive impairment in PWE is substantial and there is no reason to assume that this is not the case in LMICs [14,15,33,34].Moreover, data on PWE and NCC in terms of cognitive impairment in SSA is scarce, although studies from South America indicate a substantial burden of cognitive impairment in PWE and NCC (see discussion).Our study draws attention to mental comorbidities in PWE with and without NCC in SSA with a focus on cognitive impairment and calls for more action in the context of brain health.However, our study has clear limitations which include the rather small sample size of PWE with cognitive impairment (n = 26) and PWE with NCC (n = 8), although drawn from a rather large study population of 223 PWE.This also did not allow for further analysis such as the impact of brain lesions, e.g.vesicular lesions only, mixed lesions, and calcifications only, on cognitive impairment in PWE and NCC.In addition, our cognitive assessment tools were part of the mental state examination, which, in turn, was part of the clinical/neurological examination and ideally needs specialists or at least trained examiners to perform the examination, the latter was the case in our study.Still, variability between examiners is likely.Also, our approach to the cognitive assessment in our patients was rather pragmatic, designing three test categories suitable for the mental health clinic setting and the study population in a rather non-standardized way, which may have underestimated the true burden of cognitive impairment in our population and may have been influenced by confounders such as school education.Our clinical approach to the diagnosis of cognitive impairment in PWE was not backed by a psychological test battery as relevant cognitive tests so far have not been validated in our study population and cognitive assessment in our PWE was not among the primary aims of the larger CYSTINET-Africa study (see methods).On a positive note, PWE with a serological or clinical indication for NCC were able to access neuroimaging, which is rare in rural areas of SSA, and hence our prevalence estimate of NCC in PWE can be assumed to be robust.

Conclusion
Our study showed that 22.2 % of PWE in southern Tanzania recruited from a mental health clinic in a T. solium endemic setting had NCC and 11.8 % of all PWE had cognitive impairment.We could not establish a relationship between NCC and cognitive impairment, which could be due to the small sample size of PWE with NCC (n = 8) and/or our nonstandardized approach to cognitive assessment, containing questions in the categories of "simple arithmetic cognitive assessment", "recent memory cognitive assessment" and "remote memory cognitive assessment", only, and is performed as part of the mental state examination during the overall clinical/neurological examination.The NCC prevalence in PWE seems aligned with what is expected in T. solium endemic areas, whereas the prevalence of cognitive impairment in our study population of PWE seems rather underestimated compared to data from other geographic regions.There is a dearth of locally developed or sufficiently adapted cognitive assessment tools in countries of SSA, although brain disorders, including cognitive impairment, are on a steep rise in LMICs.Our publication should also serve as a call for action to decision-makers to invest in brain health and take firm measures against the rising burden of neurologic/psychiatric disorders ( = brain disorders), including epilepsy, NCC, and cognitive impairment/dementia.

Ethics Statement
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.

Fig. 1 .
Fig. 1.Patients' recruitment flowchart in mental health clinics and their major outcomes in terms of neurological signs/symptoms, neuroimaging results and presence/absence of cognitive impairment.

7 )
Cognitive impairment and impaired speech and ataxic gait 1 (3.6)Cognitive impairment and hemiplegia and impaired speech* 1 (3.6)Cognitive impairment and weakness of one hand and impaired speech 1

Table 2
Sociodemographic and clinical characteristics of people with epilepsy (N = 223 ɸ ).

Table 3 )
. Of the 25 PWE who were CC positive, only 18 PWE went for a CT scan, and 7 individuals did not show up for their CT scan examination appointment.On the other hand, only 18 out of 23 PWE who were CC negative but showed other neurological signs/symptoms on examination underwent CT examination.The rest did not show up for the examination without giving any reason.Eight PWE showed NCC-typical lesions.Altogether, 26 people showed cognitive impairment: 4

Table 4
Distribution of normal/abnormal mental state and neurological examination results in people with epilepsy with and without cognitive impairment (N = 221 ɸ ).
ɸ missing data for 2 PWE; *abnormal speech in this context is within mental state examination and different from speech impairment in Table3, which is based on neurological impairment; ** Fisher's exact test; *** Cognitive impairment has been included among neurological signs/symptoms (based on Table3) and therefore statistical comparison between PWE with and without cognitive impairment seems invalid (NA = not applicable).

Table 5
Distribution of cognitive impairment in people with epilepsy stratified by cysticercosis serology results (N = 221 ɸ ).

Table 6
Neuroradiological presentation of parasitic stages and number of cysts in people with epilepsy and neurocysticercosis with and without cognitive impairment and stratified by CC status (N = 8).

Table 7
Univariate logistic regression of potential risk factors for cognitive impairment in people with epilepsy (N = 221).

Table 8
Multivariate logistic regression of potential risk factors for cognitive impairment in people with epilepsy (N = 221 ɸ ).
ɸ missing results of two PWE; OR = Odds Ratio; CI = Confidence Interval.