Clinical Manifestations Associated with Neurocysticercosis: A Systematic Review

Background The clinical manifestations of neurocysticercosis (NCC) are poorly understood. This systematic review aims to estimate the frequencies of different manifestations, complications and disabilities associated with NCC. Methods A systematic search of the literature published from January 1, 1990, to June 1, 2008, in 24 different electronic databases and 8 languages was conducted. Meta-analyses were conducted when appropriate. Results A total of 1569 documents were identified, and 21 included in the analysis. Among patients seen in neurology clinics, seizures/epilepsy were the most common manifestations (78.8%, 95%CI: 65.1%–89.7%) followed by headaches (37.9%, 95%CI: 23.3%–53.7%), focal deficits (16.0%, 95%CI: 9.7%–23.6%) and signs of increased intracranial pressure (11.7%, 95%CI: 6.0%–18.9%). All other manifestations occurred in less than 10% of symptomatic NCC patients. Only four studies reported on the mortality rate of NCC. Conclusions NCC is a pleomorphic disease linked to a range of manifestations. Although definitions of manifestations were very rarely provided, and varied from study to study, the proportion of NCC cases with seizures/epilepsy and the proportion of headaches were consistent across studies. These estimates are only applicable to patients who are ill enough to seek care in neurology clinics and likely over estimate the frequency of manifestations among all NCC cases.


Introduction
Neurocysticercosis (NCC) is primarily found in countries with poor sanitation and hygiene and improper slaughterhouse services. However, due to globalization and immigration, NCC is increasingly being reported in developed countries [1]. Humans become infected by ingesting Taenia solium eggs that later develop into oncospheres. These larvae can migrate to any organ in the body, but most reports have focused on cysts located in the Central Nervous System (CNS), eyes, muscles or subcutaneous tissues. The larvae have been found in several locations in the CNS. This diversity of locations is believed to partly explain the range of NCC's clinical manifestations. In addition, the signs and symptoms associated with NCC depend on the larvae's number, developmental stage (active, transitional or calcified), on the duration of the infection and the host's immune response [2].
Seizures and epilepsy are considered to be the most common manifestations of NCC. However, several other neurological disorders can also occur [3]. Unfortunately, these less common manifestations are rarely recognized as being linked to NCC, especially in low resource countries where imaging technology is scarce [4]. Thus, data on the full range of clinical expression of NCC are lacking, although such data are essential to accurately estimate the burden of NCC on different communities. This systematic review aims to estimate the frequency of the main clinical manifestations associated with NCC.

Methods
A systematic search of the literature, including documents published from January 1,1990 to June 1, 2008, was conducted to capture data on clinical manifestations associated with NCC.

Search strategy and data source
PubMed, Commonwealth Agricultural Bureau (CAB) Abstracts, and 23 international databases were searched for data on NCC manifestations. Articles published in Chinese, English, French, Portuguese, Spanish, Italian, Romanian and German were searched. Two different searches were launched to cover both clinical manifestations and mortality associated with NCC infection. For the clinical manifestations, our search strategy in PubMed included terms: "Cysticercosis/complications" [MeSH] OR "Cysticercosis/history" [MeSH] OR ''Cysticercosis/pathology" [MeSH] OR "Cysticercosis/psychology" [MeSH] OR "Cysticercosis/radiography" [MeSH] OR "Cysticercosis/radionuclide imaging" [MeSH] OR "Cysticercosis/ultrasonography" [MeSH]. CAB Abstracts and the international search engines were queried using the following keywords: ''Taenia solium'', ''taeniasis'' or ''taeniosis'', ''cysticercosis'', and ''neurocysticercosis''. One Thesis in Medicine from Burkina Faso was identified through contacts in Sub-Saharan Africa and was included.
For mortality associated with NCC, PubMed was searched using the terms: ''cysticercosis/mortality'' [MeSH] OR "neurocysticercosis/mortality" [MeSH]. In CAB Abstracts and the international search engines the keywords ''neurocysticercosis and mortality'' were used.

Inclusion and exclusion criteria
Documents reporting valid (defined as an absence of major biases, see later), original data on clinical manifestations associated with NCC were eligible for inclusion. Books and conference abstracts were excluded because they were unlikely to have sufficient details on the methodology used.
All documents retrieved were screened based on the title and the abstract. The exclusion criteria for phase I were: 1) wrong agent; 2) animal data only; 3) no original data on the frequency of NCC's clinical manifestations; 4) case series with less than 20 participants; 5) review article without original data; and 6) editorials or letters to the editors without original data. Documents without abstracts were included in the next phase. After phase I, all eligible full text documents were reviewed qualitatively (phase II) and quantitatively (phase III). The exclusion criteria for phase II were identical to those used in phase I in addition to: 1) high potential for information bias (defined as no neuroimaging (CT-scans or MRI) or autopsies used for the diagnosis of NCC); 2) high potential for selection bias (defined as the study of volunteers or less than 80% of patients with imaging and NCC); or 3) all available data were from before 1990 or after June 1, 2008. The quantitative data from documents included after phase II were extracted in phase III.
Articles reporting the proportion of epilepsy cases with lesions of NCC were excluded from the current study and reported in another article [5].

Data extraction
Data on studies' characteristics, methodological quality and frequency of clinical manifestations and mortality were collected. Data extraction was conducted independently by at least two investigators. A third investigator checked a random sample of 10% of all of the entries. Discrepancies were resolved through discussion until a consensus was reached. The screening process (phase I) was performed in an ExcelH spreadsheet (Microsoft Corp., Redmond, WA). Methodological factors (phase II) and frequency data (phase III) were recorded in standardized electronic forms of a data extraction tool which was developed in AccessH (Microsoft Corp., Redmond, WA) specifically for this review (available from the authors on request). Authors of primary studies were contacted when the article being reviewed contained missing or unclear information on the study design or results.

Data synthesis and analysis
Whenever two or more different studies described the same clinical manifestation, we conducted a meta-analysis and estimated the pooled proportion of the given clinical manifestation among people with NCC. For these analyses, studies reporting ''seizures'' and those reporting ''epilepsy'' were combined, as most reports did not discriminate between the two. The definition of epilepsy is the occurrence of at least two unprovoked seizures separated by at least 24 hours [6].
As there was great variability in the characteristics of the included documents, results were expressed as random-effects models using proportion with 95% confidence intervals (95% CI) [7]. The Freeman-Tukey double arcsine transformation was used for pooled estimates of proportion and corresponding 95% CI from the random-effects model [8][9]. The Cochran's Q test was used to assess homogeneity across studies and the I 2 index was used to summarize the total variability in proportion due to between-study variation [10]. Random-effect models were used due to important heterogeneity between studies. A sensitivity analysis was conducted by estimating the pooled proportion after omitting one study at a time. The analysis was performed with the R META package (Version 0.8-2; Guido Schwarzer in R-META metagen function) from R statistical software (R Development Core Team, www.R-project.org). No study had considerable effect on the pooled estimate and results from the sensitivity analyses are not presented. A mixed-effects regression model was used to determine if the age group (children vs adults) significantly influenced the estimated percentage of seizures and epilepsy among people with NCC.

Literature search
A total of 1569 documents were identified in phase I. Figure 1 shows the number of papers identified in each database and included in each phase and the reasons for exclusions. After phase I, nearly three-quarters of the articles were excluded. An additional 383 articles were excluded during phase II, most of which (n = 200) did not have manifestation data or did not use neuroimaging to diagnose NCC. Fourteen Chinese articles could not be traced and were excluded. Finally, 11 articles were

Author Summary
Neurocysticercosis is an infection of the brain with the flatworm Taenia solium which is normally transmitted between humans and pigs. Sometimes, humans can infect other humans and the larva of the parasite can go the brain, causing the disease neurocysticercosis. There has never been a systematic review of what clinical signs are found among people with neurocysticercosis. We conducted a thorough review of the literature to answer this question. We reviewed 1569 and 21 were of a sufficient quality to be included in the final analysis. Among neurocysticercosis patients who are seeking care in neurology clinics, about 79% have seizures/epilepsy, 38% severe headaches, 16% focal deficits and 12% signs of increased intracranial pressure. Several other symptoms were also reported in less than 10% of patients. People with neurocysticercosis who seek care in neurology clinics show a whole range of manifestations. Clinicians should be encouraged to consider neurocysticercosis in their differential diagnosis when a patient presented with one of the symptoms described in this review. This would ultimately improve the estimates of the frequency of symptoms associated with neurocysticercosis.

Literature search on death rate associated with NCC
A total of 45 publications were identified. Of those, 27 were excluded in phase I (seven case reports, 14 reviews or letters, two animal studies, four without death data). Eighteen studies were read in full. Of those, two were case reports, nine did not provide an estimate of death rate and three were reviews. Table 2 reports the four studies included.

Distribution of manifestations among people with NCC attending neurology clinics
People diagnosed with NCC presented with a wide range of clinical manifestations. Definitions and measurement methods for manifestations were very rarely provided (Table 1).  [11] was an outlier, which explains the very large 95% confidence interval for this estimate among adults. The distribution of manifestations in patients with active and inactive lesions was presented in three studies [12][13][14] (Table 4). There was a considerably higher proportion of patients with inactive NCC who presented with seizures/epilepsy (.88%) as compared with patients with active lesions (60-63%). Conversely, the proportion with signs or symptoms of intracranial hypertension was approximately 25% in patients with active lesions but was not found in patients with inactive NCC lesions. Hydrocephalus and meningitis were reported by only one author and were more frequent with active lesions.

Distribution of manifestations in people with NCC attending an imaging clinic
In one study from Brazil [15], 236 patients seen at an imaging clinic had lesions suggestive of NCC of which 219 were inactive lesions. Lesions suggestive of other pathologies were found in 48 (20.3%) of the cases with suspected NCC. The distribution of manifestations was 30% with epilepsy, 51% with headaches, 8% with focal motor/sensory deficits. There were 35% with ''other'' symptoms found among patients with a CT-scan due to suspected neoplasia or stroke.
Among 231 patients with NCC lesions seen in a neuroradiology department, 87 (38%) were asymptomatic, incidental findings in trauma patients and cases of suspected cerebrovascular disease [13]. All of these 87 had inactive lesions. This study is unique because it reports on the possible clinical spectrum of NCC.

Distribution of manifestations among autopsied patients
In a review of 901 autopsies conducted in a department of pathology and anatomy at the University Hospital of Paraná, Brazil, the authors reported on 28 cases of NCC, with medical charts available for 27 of the cases [16]. Of those, 13 were asymptomatic, nine had seizures as a complicating factor of the clinical picture prior to death, four had increased intracranial pressure, one had meningitis, one had a cerebrovascular form and

Proportion of NCC cases seeking care who died
The proportion of cases of NCC who sought care and subsequently died were 2.3% (2/88) of adult patients with active lesions in South Africa [17], 18.5% (5/27), 2.0% (1/50) and 1.6% (2/122) of pediatric patients in India [18][19] and Mexico [20], respectively, 3.2% (1/31) of adult patients with stroke in Ecuador [21], 5.3% (2/38) of patients with active lesions in Portugal [14], and 0.9% (1/112) of patients in Houston, Texas [22]. Most deaths were associated with complications of shunt surgery for the treatment of hydrocephalus. The duration of patient follow-up and referrals to other facilities were not reported, which limits the interpretation of the data. In a study of 27 autopsied patients in Brazil, the NCC lesions were considered the cause of death in 30% of the autopsied cases [16].

Duration of disease at the time of seeking medical care
Several of the studies reported the time from the onset of symptoms to seeking medical care, with an average of 56.8 weeks in adult patients in South Africa [17], a median of 3.5 months (range 0-492 months) in patients in the United States [22], a median of 2 days (range ,1 day to 8.75 years) in children presenting to the emergency room in the United States [23], and a range of 5 days to 20 years in inpatients seeking care in China [24]. Percentages of 77.3% [25] and 80% [26] of patients had sought care within one year, 92.8% within three years [27]. The time from onset of symptoms to seeking and receiving medical care will also vary depending on the type of manifestations and the local medical services' capacity.

Proportion of NCC cases among people with specific manifestations
In people presenting with glioma [32], malignant hematological disease [33], or non-aneurysmal subarachnoid hemorrhage [34], the proportions with NCC were 18.6%, 6.2% and 4.0%, respectively ( Table 5). The proportion of NCC reported in people with stroke was 7.4% [21]. Some correlation between the location of the cyst and the focal neurological deficit was found in all NCC cases. These studies can only be used to encourage physicians to add NCC to their list of differential diagnoses when such a manifestation occurs, especially in endemic countries.

Association between manifestation and NCC
The odds ratio of NCC and cerebral glioma was estimated to be 7.63 (95%CI: 2.03-31.09) when cases of glioma were compared to  Manifestations of Neurocysticercosis www.plosntds.org age-sex-socioeconomic status matched, previously healthy, head trauma controls [32]. The odds ratio of the relation between NCC and malignant hematological diseases was estimated to be 3.54 (95%CI: 1.17-9.79) when autopsied cases were compared to autopsied cases without any type of neoplasm [33].

Discussion
This study is the first systematic review of clinical manifestations associated with NCC, which can have a wide spectrum of neurologic and psychiatric manifestations including seizures, epilepsy, headache, cerebrovascular disorders, motor deficits and depression [35].
More than three-quarters of symptomatic NCC patients seen in neurological clinics present with seizures or epilepsy. Although definitions of these conditions were very rarely provided, the estimate was surprisingly consistent across studies as a result of including only studies of a certain quality, making them more comparable to one another and the results more valid. Several recent review papers have reported percentages of NCC cases presenting with seizures and epilepsy varying from 70% to 90% [36][37][38][39][40][41][42].
The proportion of NCC cases seen in neurological clinics with seizures/epilepsy was higher in children than adults. In a review paper of NCC in childhood from India, the authors reported that from 70% to 90% of children with NCC present with seizures [43], which agrees very well with our finding. However, these results may also reflect the fact that more children with seizures/ epilepsy are referred to facilities with CT as compared to adults. In addition, if adults tend to be referred to neurology clinics for a larger spectrum of neurological disorders, this would reduce the proportion of seizure/epilepsy observed.
The next most common manifestation was headaches, at a frequency of approximately one-third of symptomatic NCC patients. The between-study estimates were more variable than what was seen for seizures/epilepsy, but were still reasonably consistent. This is surprising, since no study provided a definition for headaches. The proportion of pediatric patients with headaches was similar to that in adults but lower than the estimate for all ages combined. Measuring headaches in toddlers and young children is especially challenging since most of them cannot communicate their symptoms [44].
The effect of NCC on altered mental state and psychiatric symptoms remains poorly described. However, in the studies that were included here, they were the presenting manifestations in about 5% of cases of NCC, except for one study [11], where 52% were found to have depression at presentation. Had the studies also included psychiatry clinics, these estimates may have been higher. The proportion of NCC cases with symptoms of or increased intracranial pressure was similar between children and adults. This could be due to the fact that papilledema, which is more common among children, was included in this category of symptoms.
All of the publications found in this review reported on patients with symptomatic NCC seen in neurology clinics where imaging  was available. Therefore, the distribution of manifestations overestimates the true frequency of NCC-associated disease, since patients who are asymptomatic or with only mild symptoms are unlikely to be seen in neurology clinics. Indeed, in two studies which were conducted in neuroimaging departments, about 35% of all NCC cases were asymptomatic [14][15]. In an autopsy study, nearly 50% of cases of NCC did not have symptoms noted in their medical charts [16]. There is a lack of knowledge on the proportion of NCC cases who will develop symptoms, when in the course of disease specific symptoms occur, and the frequency with which the manifestations change over time. In a study conducted in Mexico, 9.1% of randomly-selected residents without neurological symptoms were found to have NCC based on CT-scan examinations [45]. In another study conducted in Honduras with sampling based on EITB results, 31 of 148 participants (21%) had lesions of NCC. Of these, 26 (18%) showed no manifestations, two had headaches, two had epilepsy, and one had dizziness [46]. The authors demonstrated that EITB had very poor accuracy in detecting NCC, which would suggest that sampling based on EITB may not introduce any important selection bias. If this is the case, then we could conclude that 16.1% (5/31) of people in that community had prevalent NCC symptoms. Unfortunately, in none of those studies was the history of manifestations reported. Assessing the distribution of manifestations among people with active and inactive lesions can inform us somewhat about the natural history of NCC. Seizures and epilepsy were more frequent among patients with calcified lesions. Those with active lesions were more likely to present with increased intracranial pressure, hydrocephalus or meningitis. If properly defined, the term ''epilepsy'' would be used to include only persons with unprovoked, recurrent seizures [6]. Thus, any cases of epilepsy that were a result of NCC would, by definition, have to occur in persons with inactive lesions, otherwise they would be acute symptomatic seizures. The higher proportion of seizures/epilepsy in those with inactive lesions may also reflect that NCC and epilepsy may be cooccurring conditions rather than be causally linked.
The duration of NCC-associated disease remains unknown. This review of the literature only allowed the estimation of the time between the first recorded or reported symptom and medical care. Some patients will never seek care and the duration of disease will remain unknown since NCC can only be accurately diagnosed with imaging. Once patients are in care, in case of active disease, cysticercosis will be treated and the symptoms will most likely disappear, although in the case of seizures, they may persist beyond the period of active disease.
Death was reported in only a few studies. It has been reported that neurologic deterioration in patients with NCC may be a lifethreatening event with numerous causes and diverse clinical presentations [40]. In one study, the principal concurrent conditions listed as contributing to death included hydrocephalus, cerebral edema, cerebral compression, and epilepsy/convulsions [28]. The methods to estimate death rates were so heterogeneous that they could not be combined. In order to estimate the global burden of NCC, using a country-specific case fatality rate would be more helpful.
Various uncommon clinical manifestations have been reported in numerous case reports, illustrating that clinical manifestations associated with NCC are non-specific and pleomorphic [35,40]. However, in this review, only case series that had more than 20 participants were included meaning that rarer manifestations are not included. Another important limitation is the lack of definitions of the outcomes of interest. In addition, it is possible that some researchers chose to report on only a certain set of symptoms and not on others.
Alertness of medical staff is needed to better recognize and diagnose NCC, including providing symptoms' definitions to improve our knowledge of its clinical spectrum. Other limitations are inherent to NCC itself and result from the difficulty of diagnosing NCC even with neuro-imaging [47]. Brain calcifications or granulomas which represent the most frequently observed feature in NCC are also common in tuberculosis, sarcoidosis and toxoplasmosis [48]. These lesions may lead to false positive NCC diagnoses and biased estimates of symptoms' distribution [46]. This systematic review of the literature shows that NCC imposes a heavy burden in endemic communities causing a wide range of neurological, neuropsychological and psychiatric manifestations and even premature death. Some clinical manifestations have an insidious onset and a slow progression, making their diagnosis difficult and often delayed. Hence, NCC should be kept in mind when confronted with any neurological manifestation in patients with histories of residing in endemic areas. When the clinical presentation suggests NCC infection, it is critical to perform a neuroimaging examination. The development of modern, affordable, valid diagnostic procedures and tests is needed to improve understanding of all the clinical manifestations of NCC and its epidemiology. A highly sensitive, specific and inexpensive diagnostic tool will represent a big step in gaining insight into the morbidity and mortality caused by NCC and will help to accurately estimate its global burden.