Cognitive profile in Egyptian multiple sclerosis patients has no correlation with serum neurofilament level

Background Grey matter loss is thought to be the primary reason of cognitive disability in MS, with trans-synaptic axonal degeneration acting a supportive role. This research sought to evaluate cognitive profile of Egyptian multiple sclerosis patients and find out if it has a correlation with serum neurofilament or not. Methods This was a cross-sectional research performed on a total of 60 patients with MS and 30 healthy controls. BICAMS battery of neuropsychological tests was used which includes SDMT, CVLT and BVMT. Serum NFLs using ELISA technique. Results Mean ± SD of NFL in RRMS was 82.25 ± 170.9, in PPMS was 22.08 ± 7.26, in SPMS was 95.82 ± 187.5, and in con-trol group was 56.65 ± 125.4, there was high statistical substantial variations among the different groups while there was non-statistical variation between RRMS and PPMS groups, also there was no variation between PPMS and SPMS with regard to serum level of NFL. There is no significant correlation between the NFL and different cognitive tests. Conclusion Since sNfL did not strongly connect with cognitive function in MS patients, it is possible that it cannot be used as a substitute indicator for neuropsychological state in these groups.


Introduction
Multiple sclerosis (MS) is one of autoimmune diseases common in females more than males.It is a multifocal disease with various symptoms that may be motor, cerebellar, sensory, visual, and sphincter and many other rare symptoms [1].
One of the most important symptoms that gain attention nowadays is cognitive impairment.It may occur separately as cognitive relapse which needs attention and treatment by corticosteroids.Cognitive impairment affects all types of multiple sclerosis either in remitting relapsing type (RRMS), primary progressive and secondary progressive.It may occur early and start with the early pathogenic process of MS or it may occur lately with progression of the disease [2].Some MS patients do not report any cognitive symptoms along their disease course which means that its pathogenesis remains elusive and not related to the course of the disease [3].
Multiple biomarkers are now available for activity and progression of the disease and associated cognitive affection.Neurofilament light chain (NFL) is considered one of the novel biomarkers for assessment of activity of the disease.It also increases in progressive stages of the disease [4].
Cognitive problems in patients with multiple sclerosis have not received much attention up to this point due to potential difficulties in identification in the usual clinical setting.Between 40 and 70 percent of MS patients experience cognitive impairment (CI) even in the early stages of the disease [5].
Cognitive abnormality in MS has been associated with both white matter (demyelinated lesions and white matter that appears normal) and grey matter (cerebral cortex, deep nuclei), with white matter loss being associated with deficits in working memory and mental speed of processing and grey matter atrophy being associated with deficits in verbal memory [6].
In this research, we sought to evaluate cognitive profile in Egyptian multiple sclerosis patients and find out if it has correlation with serum neurofilament or not.Methods 60 multiple sclerosis patients are diagnosed clinically and radiologically according to McDonald's criteria [7] and 30 healthy controls.MS patients were being split in to two groups; 30 patients with PMS and 30 with RRMS, then the PMS group was further subdivided in to PPMS (15 patients) and SPMS (15 patients).
Complete history taking, including educational level, disease duration, age of onset, number of relapse, the current DMT used, compliance with the drug, switching or not.Disability assessment using expanded disability status scale (EDSS) [8].Timed 25 walk test (25WT) and 9 peg hole test (9PHT).Self-report of cognitive complains using perceived deficit questionnaire (PDQ) [9].
Neuropsychological tests were done for assessment of cognition using Arabic version brief international cognitive assessment for MS (BICAMS).Symbol digit modality test is part of the BICAMS battery, reversed visuospatial memory (BVMTR) exam and the California verbal learning test (CVLT) [10].
Depression anxiety and stress scale 21 (DASS) score for evaluation of depression, stress and anxiety.Fatigue scale for motor and cognitive functions (FSMC) with its mental and physical subscale [11].
The amount of neurofilament light chain (NFL) in human serum was measured using the sandwich-ELISA method.The micro-ELISA plate that comes with this kit has been pre-coated with a human NEFL-specific antibody.The micro-ELISA plate wells are filled with benchmarks, specimens, and a particular antibody.After that, an avidin-horseradish peroxidase (HRP) mixture and a biotinylated detecting antibody specific for Human NEFL are added to each microplate well.Free pieces are removed from the wash.Into each well, the substrate solution is poured.The only wells that will be blue in color include human NEFL, biotinylated detect antibody, and avidin-HRP conjugated.When stop solution is added, the enzyme-substrate process is halted, and the color becomes yellow.The optical density (OD) is determined spectrophotometrically at a wavelength of 450 nm ± 2 nm.The degree of human NEFL directly correlates with the variation in OD value.The amount of human NFL in the samples is determined by comparing the OD of the samples to the reference curve.

Statistical analysis of the data
Data have been fed to the laptop and analyzed the use of IBM SPSS software program package deal model 20.0.(Armonk, NY: IBM Corp) Qualitative information have been defined the use of quantity and percent.The Shapiro-Wilk was used to confirm the normality of distribution of variables.Significance of the acquired consequences changed into judged on the 5% level.
The used checks were Chi-square test for categorical variables, to examine among specific groups, Fisher's Exact or Monte Carlo correction for Chi-square while extra than 20% of the cells have predicted much less than 5, one-way ANOVA test for generally disbursed quantitative variables, to examine among extra than groups, and post hoc test (Tukey) for pairwise comparisons, Kruskal-Wallis test for no longer generally disbursed quantitative variables, to examine among extra than studied groups and post hoc test (Dunn's) for pairwise comparisons, Mann-Whitney test for no longer generally disbursed quantitative variables, to examine among studied groups, Student's t-test for generally disbursed quantitative variables, to examine among studied groups, Spearman coefficient to correlate among disbursed no longer generally quantitative variables and linear regression analysis to hit upon the maximum independent/ affecting element for affecting different cognitive tests.

Demographic data and clinical features
The mean age of RRMS patients was 33.33 ± 6.84 years old, compared to 41.43 ± 8.89 years for PMS and 35.50 ± 8.70 years for the control group.There were non-statistically substantial variations between the three groups with regard to each of the education level and there was a statistical variation between the three groups with regard to gender, with the majority of RRMS patients being females (Table 1).
There were great statistically substantial variations between the RRMS group and subdivided groups of PMS regarding the age of onset of disease, where it was at a lower age among the RRMS group, followed by the SPMS group, while it was at an older age in the PPMS group, furthermore, as regards disease duration; there was high statistically substantial variations between the three groups where it was of longest duration among SPMS patients, also as regards EDSS; there were high statistically substantial variations between groups where it was lower in the RRMS group and higher mean in SPMS (Table 2).
Table 3 shows that there were high statistically substantial variations between the RRMS group and PMS as regards each of 25 FWT, right 9PHT, and Left 9PHT which were higher among patients with PMS than in RRMS patients.
Regarding DASS score, it was noticed that the total score was significantly high in the RRMS group with a median 27.5 (p value = 0.013), including its subscales, anxiety, stress and depression (Table 4).
The same for FSMC which showed a statistically significant increase in both RRMS and PMS in comparison to the control group (p values = 0.001, 0.001, respectively) (Table 5).Furthermore, there were high statistical substantial variations between three groups with regard to each of immediate recall, short term tall, short term cued, long term cued and long term total which were of higher level in RRMS group, and with regard to total BVMT; mean ± SD.Was 17.27 ± 9.05, 13.07 ± 7.54 and 24.30 ± 3.64 in RRMS, PMS and control groups, respectively, with high statistical substantial variations between three groups.

Correlation between SDMT and EDSS in each groups
Group I (RRMS) showed high substantial negative connection between SDMT and age, BMI, EDSS and relapses number, and in group II (PMS); there were non-significant correlation between SDMT and other variables of patients except as regards EDSS, there was negative significant correlation between SDMT and EDSS (Fig. 1).

Regression analysis for cognitive tests
Univariate analysis (Table 7) showed a significant effect of several parameters on SDMT such as type of MS, age and age of onset, educational level and EDSS.On multivariate it was found that EDSS has significant and independent effect on SDMT.The same different parameters affecting total recall immediate and total BVMT test; type of MS especially RRMS, age of patients, female and male gender, and EDSS are factors affect total immediate recall (Fig. 2) Age of patients and EDSS only parameters that affect total BVMT significantly and EDSS is the only factor affecting it independently (Fig. 3).

Neurofilament levels and its correlations with different parameters
Figure 4 shows that mean ± SD.Of NFL in RRMS was 82.25 ± 170.9, in PPMS was 22.08 ± 7.26, in SPMS was 95.82 ± 187.5, and in control group was 56.65 ± 125.4,there was high statistical substantial variations between the different groups while there were non-statistical variations between RRMS and PMS groups.There were variations between PPMS and SPMS as regards serum level of NFL.Also there were no any significant correlation between serum neurofilament level and any cognitive tests used (Table 8).
There were non-statistical substantial variations between both switched or ongoing DMT and the serum level of NFL in both MS groups (Table 9).

Discussion
MS is common in middle-aged females as in the current study with statistical variations between three groups as regards gender where most patients with RRMS were females, the median age of RRMS patients was higher than the mean age in PMS in agreement with literatures [3,12].
Furthermore, in our study, there was high statistical substantial variations between RRMS group and subdivided groups of PMS as regards age of onset of disease where it was lower among RRMS patients and the oldest age in PPMS group.Furthermore, with regard to disease duration; it was of longest duration among SPMS patients, also with regard to EDSS, it was lower in RRMS group and highest in PPMS.
In consistence with our findings, the study of Hussein et al. 12 there were 400 confirmed cases of multiple sclerosis, including two-thirds women and one-third men.The average age at illness start was 28.42 ± 8.48.The patients' average age was 32.59 ± 9.41 years, while the average age of illness start was 28.42 ± 8.48 years.
Egyptian research by Hashim et al. [13] which was conducted at Cairo University revealed that the average age of illness beginning in females was 27.7 ± 7.99 years old, whereas it was 29.02 ± 2.63 years old in males.
As opposed to that, the research of Filippatou et al. [14] reported that age differences between RRMS patients and controls were not statistically substantial, although PMS patients were older than the other groups.
In the present investigation, the mean NFL in RRMS was higher than PMS but not statistically significant.There were significant statistical variations between MS groups and control group.There were variations between PPMS and SPMS as regards serum level of NFL but not statistically significant.In comparison with the study of Aktas et al. [15] which reported that the median NFLs value was 16.02 pg/mL, whereas those with SPMS had greater significant values (U = 67.0,p = 0.038, r = − 0.308; RRMS: median = 12.00 pg/mL, SPMS: median = 20.00pg/ mL).
Another research by Bridel et al. [16] revealed that the mean ± SD. of NFL in healthy control was 7.1 (2.9), the mean ± SD. of NFL in RRMS was 14.4 (9.8), in PPMS was 14.5 (5.8), in SPMS was 13.1 (7.6).At baseline, NFLs was comparable among MS subtypes but greater in all MS subtypes compared to HC.
In the current study, there were non-statistical substantial variations between both switched and ongoing disease-modifying therapy and the serum level of NFL in both MS groups.Early research suggests that sNfL levels may be able to distinguish between various therapies at the level of patient groups [17] In one analysis of Novakova et al. [18] after an average follow-up of 12 months, patients who switched between disease-modifying medicines with equivalent effectiveness had stable NFLs levels, compared to patients who advanced to therapies with greater efficacies.
Patients beginning highly active immunotherapies had greater NFLs levels at treatment onset than those starting on mild/moderate treatments, confirming, and extending these results.This causes a bigger relative reduction once therapy starts [19,20] so, baseline sNfL levels were able to predict the number of future therapy changes as well as therapy intensifications.These results are also consistent with the research of Bridel et al. [16] which revealed that mean NFLs was higher in all MS subtypes when compared to HC (p < 0.0001) and was more favorably correlated with age in HC (r = 0.70, p < 0.001).Median NFLs was reduced in untreated RRMS and treated RRMS (p = 0.036) and higher in HC (p 0.001) at follow-up compared to baseline.HC (50.0%), untreated RRMS (51.4%), treated RRMS (33.3%),SPMS (45.0%), and PPMS (46.2%) all showed differences in NFLs levels at follow-up that were more than 20% from baseline values.
In the present research, there was no statistically substantial relationship between the serum concentration of NFL and any of the following variables in either group of MS patients: age, BMI, duration of illness, EDSS, or age of disease onset; however, there was a substantial inverse relationship between the serum level of NFL and the number of relapses in the PMS group of patients.
In compliance with our results, the investigation of Aktas et al. [15] found no association between NFLs either the RRMS or SPMS subsamples, nor with age, sex, educational level, EDSS score, age at illness start, subtype of MS, immunotherapy categorization, time since last relapse, time since last changes in immunomodulatory medication.They found also no correlation between serum neurofilaments and different cognitive tests as our results.Other results found a correlation between CI and neurofilament level in CSF sample mainly which may be  more accurate than serum samples [21,22].Other studies proved that neurofilament level were elevated in MS patients and correlated with CI in MS patients [23,24].Different cognitive tests used, and different sample sizes may be the factor behind this controversies.
Most of studies used different method for detection neurofilament which is single-molecule array (Simoa) assay and quantified in picograms per milliliter in contrary to ours using ELISA methods which may affect results so, still questionable if NFL is a marker for axonal damage or degeneration and progression [25].
In line with our results of cognitive tests, Montaser et al. [17] found a very substantial variation between the subgroups of MS patients and the control group with respect to SDMT, with SPMS being more impacted than RRMS.In the present study, in group I (RRMS); there was high substantial negative relationship between SDMT and age, BMI, EDSS and number of relapses, and in group II (PMS); there were non-significant correlation between SDMT and other variables of patients except as regards EDSS, there was negative substantial connection between SDMT and EDSS.
Vázquez-Marrufo et al. [26] in which a statistical investigation revealed that several factors had meaningful connections.The moderate connection between SDMT and EDSS (r = − 0.679, p = 0.0009) was noteworthy.

Table 5 Comparison between the three studied groups according to FSMC
Means with common letters are not significant (i.e., means with different letters are significant) IQR: inter-quartile range, SD: standard deviation F: F for one-way ANOVA test, pairwise comparison bet. each 2 groups was done using post hoc test (Tukey) p: p value for comparing between the three studied groups p 1 : p value for comparing between Group I and Group II p 2 : p value for comparing between Group I and Group III   In addition to above findings, we found that as regards parameters affecting total recall immediate; type of MS especially RRMS, age of patients, female and male gender, and EDSS are factors affect total recall immediate, while as regards parameters affecting total BVMT among MS patients; we found that age of patients and EDSS only parameters that affect total BVMT.
In a cross-sectional study of Hassanshahi et al. [27] which sought to assess spatial perception, visual processing speed, memory, and visual learning in MS patients according to age, gender, and educational attainment, no substantial variation was found in the mean scores of the dependent variables (JLO, SDMT, and BVMR-T scores) according to the classes of independent factors (sex, education status) (P > 0.05).Age was a confounding variable, but it had no effect (P > 0.05).Additionally, there was no substantial connection between gender and education level (P > 0.05).Age, gender, and Fig. 2 Multivariate linear regression analysis for the parameters affecting total recall immediate (n = 60) for total patients Fig. 3 Multivariate linear regression analysis for the parameters affecting total BVMT (n = 60) for total patients education level had no discernible impact on memory, visual learning, visual processing speed, or spatial perception, according to the study's findings.
In agreement with the outcomes produced by Caparel-liDáquer et al. [28] who revealed that Men and higher education groups had the greatest marks on the judgement of line orientation (JLO) test's right response.The findings may not be consistent because of the various sample populations and sizes.
A tiny sample size was one of the research's drawbacks.The length of the condition, cultural circumstances, and lifestyle may also have had an impact on the cognitive test findings.Also this research is defective at determining the exact duration of drugs received for treatment.

Conclusions
Considering all of the aforementioned factors, our findings imply that serum NFLs concentrations do not seem to be a surrogate biomarker for cognitive function and neuropsychiatric symptoms in persons with relatively mild clinical manifestations and no acute disease activity.In light of its potential use in clinical settings, the sensitivity of NFLs as a single metric for such complex functional results is called into doubt, particularly in small samples outside of large scientific studies.

4 Fig. 4
Fig.4 Comparison of the four investigated groups based on serum NFL.Group I: RRMS, Group IIA: PPMS, Group IIB: SPMS, Group III: Control

Table 1
Comparison of the three analyzed groups based on demographic informationMeans with common letters are not significant (i.e., means with different letters are significant) IQR inter-quartile range, SD standard deviation, χ 2 Chi-square test F F for one-way ANOVA test, pairwise comparison bet. each 2 groups was done using Post Hoc Test (Tukey) p: p value for comparing between the three studied groups p 1 : p value for comparing between RRMS and PMS p 2 : p value for comparing between RRMS and control p 3 : p value for comparing between PMS and control * Statistically significant at p ≤ 0.05

Table 2
Comparison of the three study groups based on several clinical dataMeans/medians with common letters are not significant (i.e., means/ medians with different letters are significant) IQR inter-quartile range, SD standard deviation, χ 2 : Chi-square test F: F for one-way ANOVA test, pairwise comparison bet. each 2 groups was done using post hoc test (Tukey) H: H for Kruskal-Wallis test, pairwise comparison bet. each 2 groups was done using post hoc test (Dunn's for multiple comparisons test) Sig. bet.grps p 1 < 0.001 * , p 2 < 0.001 * , p 3 = 0.412

Table 3
Comparison of the two groups under investigation regarding 25 WT and 9PHTIQR inter-quartile range, SD standard deviation, U Mann-Whitney test p: p value for comparing between the two studied groups * Statistically significant at p ≤ 0.05

Table 4
Comparison between the three studied groups according to DASS scoreMedians with common letters are not significant (i.e., medians with different letters are significant) IQR inter-quartile range, SD standard deviation H: H for Kruskal-Wallis test, pairwise comparison bet. each 2 groups was done using post hoc Test (Dunn's for multiple comparisons test)

Table 6
Comparison between the studied groups with subdivisions regarding BICAMs batteryMeans with common letters(a-a)(b-bc) are not significant (i.e., means with different letters(a-b) are significant BVMT brief visuospatial memory test, SDMT symbol digit modality test, CVLT California verbal learning test, IQR inter-quartile range, SD standard deviation F: F for one-way ANOVA test, pairwise comparison bet.Each 2 groups was done using post hoc test (Tukey)H: H for Kruskal-Wallis test, pairwise comparison bet.Each 2 groups was done using post hoc test (Dunn's for multiple comparisons test) p: p value for comparing between the four studied groups, p 1 : p value for comparing between RRMS and PPMS p 2 : p value for comparing between RRMS and SPMS, p 3 : p value for comparing between RRMS and control p 4 : p value for comparing between PPMS and SPMS, p 5 : p value for comparing between PPMS and control , p 2 < 0.001 * , p 3 = 0.061, p 4 = 0.553, p 5 < 0.001 * , p 6 < 0.001 * *

Table 7
Univariate and multivariate linear regression analysis for the parameters affecting SDMT (n = 60) for total patients DASS depression, anxiety and stress score, FSMC fatigue score for motor and cognitive functions, B unstandardized coefficients, C.I confidence interval, LL lower limit, UL upper limit # All variables with p < 0.05 was included in the multivariate * Statistically significant at p ≤ 0.05

Table 8
Correlation between serum NFL with different clinical variables in each group r s : Spearman coefficient

Table 9
Relation between serum NFL with different clinical variables in each group SD standard deviation, H H for Kruskal-Wallis test, U: Mann-Whitney test p: p value for relation between serum NFL and different clinical variables No.