Frequency and characteristics of cognitive impairment in individuals with calcified neurocysticercosis

Abstract

Objective

Neurocysticercosis (NCC) is the most common helminthic infection of the central nervous system (CNS). When NCC calcifies, it causes symptoms including memory impairment. Here we will assess the frequency and characteristics of cognitive impairment using screening tools in patients with calcified NCC.

Methods

We performed a cross-sectional study in participants of an ongoing cohort of patients with calcified NCC, who were administered the Mini-Mental State Examination (MMSE) and Memory Alteration Test (MAT) simultaneously. We analyzed these test scores in relation to sociodemographic, clinical, and neuroimaging characteristics.

Results

We included 336 individuals, with an average age of 43.6 +/− 14.8 years, 56% of whom were female. The median scores for MMSE and MAT were 27 (IQR 25-28) and 41 (IQR 36-45), respectively. We observed cognitive impairment in 43 (12.8%) participants with MMSE and 90 (26.8%) with MAT. Among the 57 participants aged 60 years and older, 17 (29.8%) had cognitive impairment on MMSE and 29 (50.9%) on MAT. Lower scores were associated with older age (p < 0.001 for both tests), and the presence of left temporal lobe calcifications (p < 0.001 for both), associations that persisted after multivariate adjustment (p=0.017 for MMSE and p=0.043 for MAT).

Conclusion

In this cohort, we found that low cognitive scores are frequently observed in patients with calcified NCC (12.8% by MMSE, 26.8% by MAT) and are associated with older age and calcifications in the left temporal lobe. Further studies with appropriate control groups are needed to assess these associations beyond screening tools and examine other contributing factors.

1. INTRODUCTION

Neurocysticercosis (NCC), caused by the larval stage of the tapeworm Taenia solium, is the most common helminthic central nervous system (CNS) infection in the world. It is endemic in many regions of Latin America, sub-Saharan Africa, and parts of Asia, with its highest incidence found in developing countries [1–3]. In Peru, the prevalence of NCC is estimated at 10-20% in the general population and can exceed 30% in individuals with epilepsy in endemic areas [4–6]. It typically begins as a cyst, evading the host’s immune response and establishing in the brain. Once identified, the cyst goes through a series of changes, and ultimately may regress or leave behind a calcified lesion. These calcifications are associated with the development of epilepsy and other chronic neurological conditions [2,6–8]. The most common symptoms in individuals with calcified NCC include seizures, headaches, and memory-related issues [1,2,9,10].

Cognitive impairment has been described in various studies on NCC, although the methods for its detection and the frequency reported have varied. An example of this is the study by Carabin et al., which found that the reported frequency of altered mental state and psychiatric symptoms in several NCC studies ranged from 0.5% to 74.9%, highlighting the variability of data [9]. More recent studies on cognitive impairment in NCC have used small sample sizes. While Del Brutto et al. found no significant difference in cognitive performance between patients with calcified NCC and healthy controls, Rodrigues et al. and Ciampi de Andrade et al. observed lower cognitive performance in participants with NCC regarding executive functions, verbal and nonverbal memory, constructive praxis, and verbal fluency, compared to healthy controls[11–14].

We assessed the frequency and characteristics of cognitive impairment in a large cohort of patients with calcified NCC in a reference center for NCC in Peru using the screening tools Mini-Mental State Examination (MMSE) and the Memory Alteration Test (MAT).

2. METHODS

2.1. Study design

We conducted a cross-sectional study on cognitive function in patients with calcified NCC recruited from the outpatient clinic at the Instituto Nacional de Ciencias Neurológicas (National Institute of Neurological Sciences) from Lima, Peru, a reference center for NCC in the country. Patients were participants of a cohort study initiated in 2011, aimed at characterizing the clinical evolution and identifying factors associated with new symptomatic episodes in patients with calcified NCC.

2.2. Study population

The parent cohort included patients with a diagnosis of calcified NCC, defined by a brain computed tomography (CT) scan showing only calcified lesions. For this study we included participants aged 12 years or older, with calcified NCC diagnosed via CT scan, and with available MMSE and MAT scores available in their files. Our exclusion criteria were: evidence of viable NCC (either a viable or degenerating lesion on neuroimaging, or a >10 positive ratio in the enzyme-linked immunosorbent assay used for detecting circular antigens of Taenia solium (Ag-ELISA)), primary epilepsy, persistent or progressive intracranial hypertension, focal persistent or established neurological deficit, pregnancy, refusal to undergo a CT scan if prior imaging was unavailable, and missing information regarding the location or number of calcified lesions on imaging. All the participants included had a negative Ag-ELISA ratio or had a positive ratio below 10 and showed a tendency to diminish. It is worth noting that we only collected the location of the lesions, not their volume.

2.3. Study procedures

For the cohort study, during the recruitment session the research team gathered general information regarding age, gender, years of educational study, age of onset of first neurological symptom, type of first neurological symptom, seizure type experienced, history of status epilepticus, history of neurosurgery, ventriculoperitoneal (VP) shunt history, history of previous active lesion, use of antiepileptic drugs (AEDs), type of AED, number and location of calcified lesions, presence of perilesional edema, and the presence of various cognitive impairment risk factors (traumatic head injury, hypertension, diabetes mellitus, human immunodeficiency virus (HIV) diagnosis, cardiovascular disease, alcohol consumption, and smoking history). We include all these variables in the analyses for this secondary study.

The MMSE and the MAT were administered simultaneously to the patients in the cohort study by the study doctors, previously trained for this task. These assessments were conducted either during the recruitment session or at one of the follow-up visits, which occurred every 3 months during the first year of follow-up and every 6 months thereafter.

The MMSE is a widely used cognitive test worldwide. It provides a quantitative assessment of cognitive function, with a high score of 30 and a cutoff score of 24, based on the study by Robles-Arana in the Peruvian population, and has good sensitivity and specificity for participants with any educational level. This test is composed of five cognitive areas: orientation (10 points), registration (3 points), attention and calculation (5 points), delayed recall (3 points), and language and praxis (9 points) [15,16].

The MAT is used to screen for mild cognitive impairment in the general population, is especially useful in patients with low levels of education, and its highest score is 50. We are using a cutoff score of 37, used in other studies by Custodio et al. in Peruvian populations with similar characteristics. Also, the study analyzed the five memory-related domains: temporal orientation (10 points), semantic memory (5 points), recent memory (15 points), remote memory (10 points), and free/cued recall (10 points) [17].

All participants underwent both cognitive assessment testing. Both tests have been validated for their use in Spanish speaking countries, and some specific adaptations have been made for the Peruvian population [18–20].

2.4. Data analysis

Data was analyzed using R version 4.3.0 (2023-04-21). Numerical variables were reported with measures of central tendency and dispersion, median and interquartile range (IQR). Categorical variables were reported as frequencies and percentages. We assessed the frequency of cognitive impairment using the MAT and the MMSE.

For the bivariate analysis, we treated test scores as continuous variables. We included all the variables mentioned before, such as sex and age, and explored those variables with statistical significance while ensuring there was no risk of collinearity. We reported p-values with a 0.05 significance level. To explore the bivariate relationships and measure their potential as predictors to include in multiple regression, we employed generalized additive models (GAM) to assess non-linear associations. In the multivariate analysis, we used GAM to evaluate independent associations between cognitive test scores and the selected predictors. GAM allowed us to account for potential non-linear relationships between predictors and outcomes, ensuring a flexible and robust model. Additionally, we evaluated each adjusted model to verify collinearity using the variance inflation factor (VIF<10) from the performance package in R. To assess and control for potential confounding by sociodemographic variables such as age and gender, these factors were included in the model adjustment. The study also evaluated the inclusion of variables related to risk factors for cognitive impairment, aiming to maintain the most parsimonious model possible. Domain-specific analyses were conducted for each cognitive area using separate regression models, adjusting for age, gender, and lesion characteristics.

2.5. Ethics

The information for this study was extracted from the study “Cohort study for characterizing the clinical evolution and associated factors to new symptomatic episodes in patients with only calcified neurocysticercosis. Lima, Peru, 2011-2027”, approved by the main Institutional Review Board of the Instituto Nacional de Ciencias Neurológicas and the Universidad Peruana Cayetano Heredia (last approved protocol by SIDISI code 59136-CIEI-INCN). All participants provided written informed consent to enter this study. Data extracted were coded and anonymized to ensure confidentiality and protect participant privacy.

3. RESULTS

3.1. Study population

Of the 621 participants in the original cohort, 285 were removed due to exclusion criteria detailed in Figure 1. Finally, 336 participants were included in our final analysis.

Figure 1: study flowchart regarding exclusion criteria.

Ag-ELISA: antigen enzyme-linked immunosorbent assay; NCC: neurocysticercosis; CT scan: computed tomography scan

*Exclusion criteria for this study, either Mini-Mental State Examination or Memory Alteration Test absent

3.2. Baseline characteristics

The baseline characteristics of the participants are summarized in table 1. The average age was 43.6 ± 14.8 years, 56% were female, the median duration of symptoms before recruitment was 10 years (IQR 5 – 20). The median number of calcified lesions was 3 (IQR 1 – 7). We identified 27 patients (8.0%) with pericalcification edema. Additionally, 39 patients (11.6%) had undergone neurosurgery (including ventriculoperitoneal shunts, endoscopic surgery for intraventricular cysts, and other surgeries). The median MMSE score was 27 (IQR 25 – 28) and the median MAT score was 41 (IQR 36 – 45). Cognitive impairment was found in 43 patients (12.8%) by MMSE and 90 (26.8%) by MAT. Among the 57 patients aged 60 years and older, 17 (29.8%) had cognitive impairment by MMSE and 29 (50.9%) by MAT.

Table 1:

characteristics of patients in the calcified neurocysticercosis cohort

Characteristics	N (%)
Gender
Feminine	190 (56.5)
Masculine	146 (43.5)
Age (years)*	43.6 ± 14.8
Over 60 years old	57 (16.9)
Years of study*	10.8 ± 3.98
Time since symptom onset (years)†	10 (5-20)
Number of calcified lesions †	3 (1-7)
Edema
Present	27 (8.0)
Absent	309 (92.0)
Previous neurosurgery
None	297 (88.4)
Yes	39 (11.6)
Type of seizure (ever experienced)
None	79 (23.5)
Focal	52 (15.5)
Focal progression to BTCS	57 (17.0)
Generalized	148 (44.0)
Antiepileptic drug used
Phenytoin	78 (38.2)
Carbamazepine	76 (37.3)
Valproic acid	40 (19.6)
Clobazam	7 (3.4)
Phenobarbital	3 (1.5)
Test scores †
MMSE	27 (25-28)
MAT	41 (36-45)
Risk factors for cognitive impairment
Traumatic brain injury	69 (20.6)
Hypertension	19 (5.7)
Diabetes Mellitus	7 (2.1)
Stroke	3 (0.9)
Alcohol consumption	37 (11.0)
Smoking history	10 (3.0)

^*Mean ± standard deviation

^†Median (IQR)

BTCS: bilateral tonic-clonic seizure, MMSE: Mini-Mental State Examination, MAT: Memory Alteration Test

Of the 336 participants, 146 had tests on the recruitment session and of these, 21% had cognitive impairment according to MMSE, and 36% by MAT. One hundred and ninety took the tests during follow-up, with 20% and 6.8% showing cognitive impairment according to MMSE and MAT, respectively.

Regarding the relation between MMSE and MAT outcomes, 31 patients had cognitive impairment according to MMSE and MAT, while 234 did not have it according to both. We observed discrepancies in 71 participants (12 were MMSE positive, MAT negative, and 59 were MMSE negative, MAT positive). We found an agreement of 78.9% and a kappa coefficient of 0.35 (0.24 – 0.47) for inter-rater reliability.

Concerning the cognitive domains of both tests, the age over 60 years was related to lower scores. Other factors that were related to lower scores in several domains on both tests were male sex and the presence of left temporal lobe calcifications. With respect to seizures, 257 participants (76.5%) reported a history of seizures of any type, with only 15 (5.7%) experiencing seizures within 72 hours before cognitive testing. Among these, 148 (44.0%) reported generalized seizures, 52 (15.5%) focal seizures, and 57 (17.0%) focal progressing to bilateral tonic-clonic seizures.

A total of 204 participants were taking AEDs at the time of testing, with the most commonly used medications being phenytoin (38.2%), carbamazepine (37.3%), valproic acid (19.6%), clobazam (3.4%), and phenobarbital (1.5%). Regarding cognitive impairment risk factors, 19 patients (5.7%) had hypertension, 7 (2.1%) diabetes mellitus, 3 (0.9%) had a stroke, 69 (20.6%) had traumatic brain injury, 37 (11%) consumed alcohol, 10 (3%) smoked, and none were HIV-positive.

3.3. Bivariate analysis

We compared cognitive test scores between individuals who had seizures within 72 hours prior to testing (MMSE median 27, IQR 26–28; MAT median 41, IQR 36–44) and those without recent seizures (MMSE median 27, IQR 25–28; MAT median 42, IQR 36–46), and found no significant differences (p = 0.400 for MMSE; p = 0.900 for MAT). Patients using AEDs at the time of testing had slightly lower MMSE scores (median 27, IQR 25–28) compared to those not on AEDs (median 27, IQR 26–29), with this difference approaching marginal significance (p = 0.056). No significant differences were observed in MAT scores (p = 0.300).

In the MAT, male patients (median 40, IQR 34–45) scored lower than females (median 42, IQR 37–46), reaching marginal significance (p = 0.050). Patients with a history of neurosurgery also had significantly lower MAT scores (median 37, IQR 31–41) compared to those without surgery (median 42, IQR 36–45; p = 0.016). These differences were not observed in MMSE scores Years of education were strongly associated with higher scores on both MMSE and MAT (p < 0.001), while age was negatively associated with scores on both tests (p < 0.001). Time since the first seizure was significant for MMSE (p = 0.032) but not for MAT (p = 0.065). A negative association was also observed between both MMSE and MAT scores and time since the first neurological symptom (p < 0.001). The number of calcified lesions was not associated with cognitive performance (p = 0.146 for MMSE; p = 0.814 for MAT). Non-linear relationships are illustrated in Supplemental Material 1. No significant MMSE differences were observed in patients with traumatic brain injury, stroke, diabetes, hypertension, smoking, or alcohol use.

For MAT scores, hypertensive patients showed a marginally lower performance (median 39, IQR 29–44) than normotensive individuals (median 41, IQR 36–45; p = 0.094). No significant differences were observed for traumatic brain injury, stroke, diabetes, smoking, or alcohol history. Regarding lesion location (Table 2), MMSE scores were lower in patients with bilateral hemisphere lesions (median 27, IQR 24–28) compared to right hemisphere lesions (median 28, IQR 27–29; p < 0.001, GAM). For parietal lobe lesions, bilateral involvement was associated with lower MMSE scores (median 27, IQR 24–28) compared to those without lesions (median 27, IQR 26–28; p = 0.029). Temporal lobe lesions were linked to lower MMSE scores (median 27, IQR 24–28 vs. 27, IQR 26–28; p = 0.005), with left temporal lobe lesions showing the greatest effect (median 25.5, IQR 24–27; p < 0.001).

Table 2:

Relationship between test scores according to cognitive testing and location of calcified lesions

			MMSE			MAT
			Median	IRQ	p-value*	Median	IRQ	p-value*
Hemispheres	Yes		27	25 – 28		41	36 - 45
		Left	27	26 – 28	0.212	43	35 - 47	0.892
		Right	28	27 - 29	Ref.	42	38 - 45	Ref.
		Both	27	24 - 28	<0.001	41	36 - 44	0.166
Frontal	No		27	26 - 28	Ref.	40	35 - 45	Ref.
	Yes		27	25 - 28	0.493	41.5	36 - 46	0.243
		Left	27	25 - 28		42	36 - 47
		Right	27	26 - 29		42	35 - 45
		Both	27	25 - 28		41	37 - 45
Parietal	No		27	26 - 28	Ref.	42	36 - 46	Ref.
	Yes		27	24 - 28	0.022	41	36 - 45	0.298
		Left	26	24 - 28	0.105	41	35 - 44
		Right	27.5	24 - 28	0.362	42	37 - 45
		Both	27	24 - 28	0.029	41	37 - 44
Occipital	No		27	25 - 28	Ref.	42	36 - 46	Ref.
	Yes		27	25 - 28	0.610	41	36 - 45	0.800
		Left	27	24.5 - 28		40	35 - 45
		Right	28	27 - 29		41	37 - 44
		Both	27	24 - 28		41	34 - 45
Temporal	No		27	26 - 28	Ref.	42	37 - 46	Ref.
	Yes		27	24 - 28	0.005	40	35 - 44	0.003
		Left	25.5	24 - 27	<0.001	37	29 - 43	<0.001
		Right	27	25 - 28	0.196	41	36 - 45	0.142
		Both	28	25 - 29	0.776	41	37 - 44	0.469

^*Unadjusted analysis

MMSE: Mini-Mental State Examination; MAT: Memory Alteration Test

In the MAT, temporal lobe lesions were associated with lower scores (median 40, IQR 35–44) compared to individuals without temporal lesions (median 42, IQR 37–46; p = 0.003), with left temporal lobe lesions having the strongest impact (median 37, IQR 29–43; p < 0.001). No significant associations were observed for parietal lobe lesions.

3.4. Multiple regression analysis

Our multiple regression analyses were conducted using two models: one adjusting for the minimum available variables and another adjusting for cognitive risk factors, as detailed in Table 3. We identified two predictors with a statistically significant impact on cognitive performance across both screening tests. The first predictor was age, which exhibited a significant non-linear effect on cognitive performance in both MMSE and MAT (p < 0.001 for both models), with scores declining as age increased. The second predictor was the presence of left temporal lobe lesions. In the MMSE, patients with left temporal lesions scored on average 1.2 points lower than those without lesions when adjusting for gender and age (β = −1.20; 95% CI: −2.30 to −0.23; p = 0.017), and 1.6 points lower when additionally adjusting for cognitive risk factors (β = −1.60; 95% CI: −2.80 to −0.46; p = 0.007). For the MAT, these patients scored 2.59 points lower than those without lesions when adjusting for gender and age (β = −2.59; 95% CI: −5.10 to −0.10; p = 0.043), and 3.2 points lower when adjusting for cognitive risk factors (β = −3.20; 95% CI: −6.10 to −0.30; p = 0.032).

Table 3:

Multivariate Analysis of Predictors Associated with MMSE and MAT Scores

	MMSE (unadjusted) †			MMSE (adjusted with the minimum variables) †1			MMSE (adjusted with risk factors) †2			MAT (unadjusted) ‡			MAT(adjusted with the minimum variables) ‡1			MAT (adjusted with risk factors) ‡2
Predictors	Coef.	95% CI *	P value	Coef.	95% CI *	P value	Coef.	95% CI *	P value	Coef.	95% CI *	P value	Coef.	95% CI *	P value	Coef.	95% CI *	P value
Intercept	26.5		<0.001	26.5		<0.001	26.5		<0.001	41		<0.001	41		<0.001	26.5		<0.001
Gender
Female	—	—		—	—		—	—		—	—		—	—		—	—
Masculine	0	−0.66, 0.65	0.900	−0.05	−0.66, 0.57	0.879	0.03	−0.75, 0.81	0.900	−1.7	−3.3, 0.0	0.050	−1.78	−3.3, −0.25	0.023	−1.2	−3.1, 0.77	0.200
Presence of lesion in temporallobes
No	—	—	—	—	—	—	—	—		—	—	—	—	—		—	—
Both	−0.17	−1.3, 0.97	0.800	−0.02	−1.1, 1.1	0.879	−0.53	−1.8, 0.72	0.400	−1.1	−4.0, 1.8	0.500	−0.78	−3.4, 1.9	0.565	−1.8	−4.9, 1.3	0.300
Right	−0.63	−1.6, 0.32	0.200	−0.54	−1.5, 0.34	0.253	0.05	−1.1, 1.2	0.900	−1.8	−4.3, 0.60	0.140	−1.9	−4.2, 0.36	0.101	0.04	−2.8, 2.9	0.900
Left	−2.1	−3.1, −1.1	<0.001	−1.2	−2.3, −0.23	0.017	−1.6	−2.8, −0.46	0.007	−4.9	7.5, −2.2	<0.001	−2.59	−5.1, −0.10	0.043	−3.2	−6.1, −0.3	0.032
Smoothed age effect			<0.001			<0.001			<0.001			<0.001			<0.001			<0.001
History of traumatic brain injury	−0.41	1.2, 0.39	0.300	—	—	—	−0.47	−1.4, 0.46	0.300	−1.0	−3.1, 1.0	0.300	—	—	—	−0.81	−3.1, 1.5	0.500
History of hypertension	0.19	−1.2, 1.6	0.800	—	—	—	1.6	0.00, 3.3	0.052	−3.0	−6.6,0.52	0.095	—	—	—	0.59	−3.5,4.7	0.8
History of diabete smell it us	1	−1.3, 3.3	0.400	—	—	—	1.8	−0.81, 4.4	0.200	3.5	−2.3.9.2	0.200	—	—	—	6.1	−0.42, 13	0.068
History of stroke	−1.6	−5.1, 1.8	0.400	—	—	—	−1.9	−7.5, 3.7	0.5	0.38	8.4,9.1	0.900	—	—	—	10	−3.7, 24	0.200
History of alcohol consumption	0.87	−0.18, 1.9	0.100	—	—	—	0.63	−0.48, 1.7	0.300	1.0	−1.6, 3.6	0.400	—	—	—	0.82	−1.9, 3.5	0.600
Smoking history	1.3	−0.56, 3.2	0.400	—	—	—	0.81	−1.1, 2.7	0.400	1.4	−3.2, 6.0	0.600	—	—	—	1.1	−3.6, 5.8	0.700

^*CI = Confidence interval

^†1MMSE information model: df = 7; Log-likelihood = −815; AIC = 1,645; BIC = 1,675; R2(adj.) =0.138; Deviance explained= 15.3%; Deviance = 2,603; Residual df = 326; No. Obs. = 333

^†2MMSE information model: df = 12; Log-likelihood = −496; AIC = 1,018; BIC = 1,061; R2(adj.) =0.138; Deviance explained= 15.3%; Deviance = 1,487; Residual df = 194; No. Obs. = 206

^‡1MAT information model: df = 7; Log-likelihood = −1,115; AIC = 2,246; BIC = 2,276; R2(adj.) = 0.182; Deviance explained = 19.6%; Deviance = 15,828; Residual df = 326; No. Obs. = 333

^‡2MAT information model: df = 12; Log-likelihood = −682; AIC = 1,390; BIC = 1,434; R2(adj.) = 0.182; Deviance explained = 19.6%; Deviance = 9,072; Residual df = 194; No. Obs. = 206

MMSE: Mini-Mental State Examination; MAT: Memory Alteration Test; Coef: coefficient.

Additionally, male participants scored 1.78 points lower than females on the MAT when adjusting only for age and gender (β = −1.78; 95% CI: −3.30 to −0.25; p = 0.023), but this difference was not significant after adjusting for cognitive risk factors (β = −1.20; 95% CI: −3.10 to 0.77; p = 0.200). Finally, none of the cognitive risk factors were independently associated with test scores.

4. DISCUSSION

In this study, 43 patients (12.8%) met criteria for cognitive impairment according to their MMSE scores, while 90 patients (26.8%) met criteria according to MAT scores. Among participants aged ≥60 years, frequency was 29.8% with MMSE and 50.9 with MAT. As our study did not include a control group, direct comparison to a reference population was not possible; however, our findings can be contrasted with previous reports from similar settings. Segovia et al. reported an 8.6% prevalence of cognitive impairment in Peru among individuals aged ≥60 years using the MMSE [21], and Belón Hercilla et al. found a prevalence of 21% in adults aged ≥55 years in Arequipa [22]. Both estimates are lower than the 29.8% frequency observed in our study for the same age group. For the MAT, Sanchez et al. reported a prevalence of mild cognitive impairment at 17.9% in individuals aged ≥60 [23], substantially lower than the 50.9% observed in our cohort.

In our population, the MMSE and MAT yielded markedly different rates of cognitive impairment, with MAT identifying a greater proportion of cases. Several factors likely explain this discrepancy. A recent review of cognitive assessment tools for mild cognitive impairment reported that MAT has one of the highest sensitivities for detecting such cases, outperforming the MMSE, which demonstrates lower sensitivity in this context [24]. Additionally, MAT was designed for use in populations with lower educational attainment, making it more suitable for detecting impairment in our cohort [17]. The two tools also differ in the cognitive domains they assess: MAT focuses on episodic and semantic memory and was specifically developed to detect amnesic mild cognitive impairment (aMCI), whereas MMSE primarily evaluates orientation, language, praxis, and general aspects of memory. As a result, MAT is better at discriminating between cognitively healthy individuals and those with probable aMCI [17]. However, as both are screening instruments that target specific domains, they share the inherent limitation of potentially missing cognitive impairment affecting untested domains.

The observed relationships between age, years of education, and cognitive scores are consistent with established evidence. Higher educational attainment is associated with improved performance on cognitive tests, likely reflecting greater cognitive reserve, which facilitates the use of alternative neural networks to compensate for brain pathology [25,26]. Conversely, advancing age is a well-recognized risk factor for cognitive decline due to cumulative effects of neurodegenerative changes, cerebrovascular disease, sensory impairment, and reduced processing speed [27,28]. These factors contribute to lower scores in older individuals, even in the absence of overt dementia.

As expected, those who received the screening tests on their recruitment day had higher odds of cognitive impairment in both tests, compared to those who received them on follow-up. This can be explained by the fact that the NCC treatment in related to an improvement of symptoms, or that they had more years of education [29].

A history of neurosurgery was associated with lower MAT scores. This may reflect postoperative cognitive decline, as reported in patients with normal pressure hydrocephalus following ventriculoperitoneal shunt placement [30].

Bilateral parietal lobe lesions are associated with lower cognitive scores, reflecting the role of these regions in visuospatial processing, numerical cognition, and integration of sensory information. Disruption of parietal-temporal networks can further impair attention and memory, making bilateral involvement particularly detrimental to overall cognitive function [31,32].

We also identified that the domains of remote memory, free/cued memory of MAT and orientation of MMSE were influenced by the presence of lesions in the left temporal lobe. Lesions in this area were associated with lower scores in both tests in the bivariate analysis, and in multiple regression after adjustment by confounders. An explanation for this finding is that the temporal lobes are key to the formation and retrieval of memory, so its damage will cause memory issues [33]. Also, we know that the left temporal lobe has a high number of neuronal connections with the rest of the hemisphere, and a lesion in that area will have a great impact on cognitive function. This has been called the disruption of fronto-parietal-temporal networks and called a possible cause of cognitive dysfunction in NCC [34]. Other diseases that affect this area also manifest cognitive impairment, like temporal lobe epilepsy [35]. This can be part of the explanation for our finding that patients with lesions in the left hemisphere are associated with lower scores when compared to those with lesions in the right hemisphere. This finding could probably not be proposed before due to the lower number of patients evaluated [11,29,36].

Importantly, epilepsy can cause cognitive impairment, particularly when seizures are frequent, as this may lead to neuronal loss and lasting dysfunction across multiple cognitive domains. Antiepileptic treatment can improve cognition by decreasing the seizure frequency; however, some medications—especially older-generation AEDs, such as those used in this study—can themselves contribute to cognitive impairment as a side effect [37].

Our study has several limitations. The fact that the tests were administered at different time points (at recruitment or during follow-up) may include variance in disease evolution, as shown by higher scores in follow-up cases, and could also affect the temporal correlation with imaging, which was mostly done by the cohort study protocol before recruitment. There are some potential structural associations that we did not evaluate in this population, such as hippocampal atrophy [14]. Additionally, we could not rule out all the factors that influence cognitive impairment, including depression, electrolyte disturbances, thyroid diseases, vitamin B12 deficiency and infections. We have used basic screening tools for cognitive function, which means we could have not detected some cases of cognitive impairment expressed in areas not covered by these tests. Finally, our study design as cross-sectional did not allow us to compare our findings with age and gender matched control.

In conclusion, cognitive impairment was frequent in this series of individuals with calcified NCC in both the MMSE and MAT tests (12.8% and 26.8%, respectively). In participants older than 60 years, the proportions of cognitively impaired patients increased to 29.8% for the MMSE and 50.9% for the MAT. These proportions are larger than those reported in other studies of similar populations, and considering the low frequency of risk factors for cognitive impairment in the studied population, the most likely explanation for this difference is the effect of NCC and epilepsy on cognitive functions. Among the factors studied, years of education and age were the most influential in both test scores. We also found that patients with left temporal lobe calcifications had lower cognitive scores compared to those without temporal lesions, likely due to the various cognitive functions of the temporal lobe, including memory, language, and attention. Patients with calcified NCC must be screened for cognitive impairment, especially those with lesions on temporal lobes. Further research in this topic using more detailed neuropsychological batteries and control populations and considering other cognitive disorder etiologies is needed, to accurately assess the magnitude of cognitive impairment in this population.

Highlights.

• Memory-related issues are a common, understudied complaint in neurocysticercosis

• Cognitive impairment was more prevalent in calcified neurocysticercosis

• Calcifications in the left temporal lobe are associated with lower cognitive scores