Время 25-foot Проходное испытание независимо связано с когнитивной функцией при множественном склерозе: перекрестное исследование.

Introduction

Multiple sclerosis (MS) is an inflammatory demyelinating disease of the CNS that afflicts 2.8 million people worldwide [ 1 ]. Cognitive impairment (CI) has been recognized as a symptom of MS for as long as the disease has been diagnosed [ 2 ]; however, in recent years, it has garnered more attention and has been investigated as a core hallmark of MS [ 3 ]. It has been estimated that CI affects 65% of people with multiple sclerosis (PwMS) during the course of the disease [ 4 ], and has been associated with increased rates of unemployment and lower quality of life among these patients [ 5 ]. In addition, CI is an invaluable prognostic marker, as studies have revealed that CI at the time of diagnosis is associated with poorer outcomes and predicts disability progression [ 6 ], cortical thinning [ 7 ], and death [ 8 ].

The correlates of CI in PwMS have been the focus of many investigations. CI is influenced by demographic factors, such as age and education. Studies have demonstrated a positive correlation between age and CI [ 9 ], while higher education is linked to a lower risk of CI, with more educated patients often retaining a better verbal capacity [ 10 ]. CI is also influenced by the characteristics of MS, such as disease duration and level of disability as measured by the Expanded Disability Status Scale (EDSS) [ 11 ]; however, reports in the literature are heterogeneous with respect to EDSS and, at best, point to a moderate correlation [ 12 ].

With that said, evidence regarding the correlation of CI and motor function in MS is sparse. The link between CI and motor dysfunction _including both upper limb and gait dysfunction_ has been delineated in both healthy aging adults [ 13 ] and several neurological diseases, such as Alzheimer’s and Parkinson’s disease [ 14 , 15 ]. Furthermore, cognitive rehabilitation in MS has improved motor function [ 16 ]. However, variables relating to physical function, such as the 9-hole peg test (9-HPT) and the timed 25-foot walk test (T25FW), have not been investigated extensively in relation to cognition in PwMS. Limited evidence suggests that both 9-HPT and T25FW correlate with cognitive performance [ 17 , 18 ] and that the T25FW, but not the 9-HPT, is predictive of cognitive worsening in PwMS [ 18 , 19 ]. Yet, these studies did not account for confounders in their analyses, used a limited number of domain-specific cognitive assessment measures, or did not attempt to compare motor function assessment scales in their relationship with cognition.

CI in PwMS significantly impacts patient care, underscoring the need for precise assessment and targeted treatment strategies. Recognizing the factors that influence CI is essential for optimizing treatment approaches in this population. Given the limited evidence on the association and predictive value of physical function concerning CI, this study aimed to identify the factors and predictors of cognitive domains most likely to be impaired in PwMS using the Brief International Cognitive Assessment in Multiple Sclerosis (BICAMS) with a specific focus on physical function tests, such as the T25FW and the 9-HPT. This focus allows for a more detailed understanding of how physical function may correlate with and predict cognitive performance in PwMS, informing better management of CI in clinical settings.

Methods

This cross-sectional study was conducted at Kashani Hospital in Isfahan, Iran, from September 2023 to September 2024. The study adhered to the ethical principles outlined in the Declaration of Helsinki, and was approved by the ethics committee of Isfahan University of Medical Sciences under registration code IR.ARI.MUI.REC.1401.061. All participants provided informed consent prior to enrollment. This report follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (Additional file 1 ) [ 20 ].

Participants

Adult participants aged 18 or older who met the diagnostic criteria for MS based on the revised McDonald criteria [ 21 ] were eligible for enrollment. Patients diagnosed with MS at least 1 year earlier with no relapses during that time were included. Patients with other neurological conditions or chronic diseases, including gastrointestinal, cardiac, liver, renal, or respiratory issues, were excluded, as were those who were pregnant or breastfeeding. No limitations were imposed on the type or severity of MS or the type of treatment received.

Study variables and endpoints

Demographic variables, including age, gender, and years of education, were recorded. Characteristics of the disease, including MS onset type (relapsing–remitting MS (RRMS) or progressive MS (PMS)), disease duration, type of disease-modifying therapies (DMTs), and EDSS [ 22 ] were assessed and documented by a neurologist. Next, the BICAMS was administered to evaluate the cognitive functions of PwMS. The BICAMS comprises three tests: the Symbol Digit Modalities Test (SDMT), the California Verbal Learning Test-second edition (CVLT-II), and Brief Visuospatial Memory Test-Revised (BVMT-R) [ 23 ]. The motor functions of PwMS were also evaluated by administering the 9-HPT and the T25FW. The tests utilized in this study are described below.

Administered tests

Symbol digit modalities test

In this test, participants are shown a stimulus page displaying a key with nine number–symbol pairs. Over 90s, they must verbally identify the correct number corresponding to each target symbol. Performance is scored based on the total number of accurate responses, yielding scores from 0 to 110 [ 24 , 25 ]. A validated Persian translation of SDMT was used [ 26 ]. The oral version was used instead of the verbal version according to the BICAMS.

California verbal learning test-second edition

The CVLT-II is an auditory/verbal learning and memory test. A validated Persian translation [ 26 ] of the test was administered, assessing memory through a 16-item word list divided into four semantic categories. Participants were exposed to the list and then asked to recall the words across five trials. The total number of words recalled in each trial was then summed to calculate the total recall score, which ranges from 0 to 80 [ 27 ].

Brief visuospatial memory test-revised

The BVMT-R is a measure of visuospatial memory. During the test, six abstract designs are shown for 10s. Afterward, the display is removed, and the patients are asked to reproduce the designs on paper. Each design is scored from 0 to 2 based on accuracy and location, with total scores ranging from 0 to 12. There are three learning trials, and the primary outcome is the total score across these trials, ranging from 0 to 36.

9-Hole peg test

The 9-HPT evaluates manual dexterity by requiring participants to place and remove nine pegs, one at a time, into nine holes as quickly as possible. This task is repeated with both the dominant and non-dominant hands, with each hand tested twice. The time taken to complete the test with each hand is recorded, averaged, and reported as the final measure of performance [ 28 ].

Timed 25-foot walk test

In the T25FW test, which evaluates the function of the lower extremities, the patient is positioned at the starting point of a marked 25-foot course and instructed to walk the 25 feet as quickly and safely as possible. The task is repeated immediately by having the patient walk back the same distance. The average duration of the trial is then calculated and reported [ 28 ].

Statistical analysis

All analyses were conducted using Stata version 18. The normal distribution was checked using the Shapiro–Wilk test. Normally distributed variables were expressed as mean (standard deviation), while non-normal variables were reported as median (interquartile range). Correlations between SDMT, CVLT-II, and BVMT-R and age, years of education, disease duration, EDSS, T25FW, and 9-HPT were investigated using Pearson’s ( r ) (for data with normal distribution) or Spearman’s ( ρ ) (for data with non-normal distribution) correlation coefficients. All correlations are depicted in a correlation matrix. A hierarchical regression model was employed to investigate the independent predictors of each cognitive test. In step 1, covariates (age, gender, education, disease duration, EDSS) were included based on prior literature [ 29 – 31 ] and significant correlations with the forward stepwise selection method. In step 2, motor function tests (T25FW, 9-HPT) were entered using a forward stepwise selection method to prevent overfitting. To assess multicollinearity, we computed the Variance Inflation Factor (VIF) for each predictor. A VIF value below five was considered indicative of low multicollinearity. Regression diagnostics were conducted to identify potential outliers and influential data points. We examined standardized residuals, Cook’s distance, and leverage values. Data points with Cook’s distance > 1 or high leverage were further investigated to determine their impact on the model. A p value less than 0.05 was deemed statistically significant in all analyses.

Results

Population characteristics

In total, 125 PwMS were enrolled in this study. PwMS had a mean age of 37.58 (8.10) years, were mostly female (86.4%), and had 16 years of education. The patients had a median disease duration of 8 (4–11) years and a median EDSS of 1.5 (1–2.5). Most of the PwMS presented with the course of RRMS (89.6%) rather than progressive MS (10.4%). The most frequently administered DMTs were interferon-a (30.4%) and rituximab (19.2%). The included participants had a median T25FW score of 5.36 (4.89–6.17) and a median 9-HPT score of 22.5 (20.94–25.82). In addition, the patients had a median SDMT score of 48 (42–53), a mean CVLT-II score of 48.78 (8.85), and a median BVMT-R score of 26 (20–29) (Table  1 ). Table 1 Characteristics of the included population Variables PwMS ( n  = 125) Demographics Age 37.58 (8.10)* Gender Male 17 (13.6%) Female 108 (86.4%) Education years 16 (12–16)^ MS characteristics  Type of MS   Relapsing–remitting 112 (89.6%)   Progressive 13 (10.4%)   Disease Duration 8 (4–11)^   EDSS 1.5 (1–2.5)^  Disease modifying therapies   None 9 (7%)   Azathioprine 1 (0.8%)   Dimethyl fumarate 15 (12%)   Fingolimod 5 (4%)   Glatiramer acetate 2 (1.6%)   Interfron-a 38 (30.4%)   Natalizumab 2 (1.6%)   Ocrelizumab 14 (11.2%)   Rituximab 24 (19.2%)   Terifluromide 15 (12%)  Motor function   T25FW 5.36 (4.89–6.17)^   9-HPT 22.5 (20.94–25.82)^   Dominant hand 9-HPT 22 (20–24.46)^   None-dominant hand 9-HPT 24 (21–26)^  Cognitive function   SDMT 48 (42–53)^   CVLT-II 48.78 (8.85)*   BVMT-R 26 (20–29)^ *Mean (standard deviation). ^median (Q1–Q3). EDSS, expanded disability status scale; SDMT, symbol digit modalities test; CVLT-II, California verbal learning test-II; BVMT-R, brief visuospatial memory test-revised; T25FW, timed 25-foot walk test; 9-HPT, 9-hole peg test

Table 1 Characteristics of the included population Variables PwMS ( n  = 125) Demographics Age 37.58 (8.10)* Gender Male 17 (13.6%) Female 108 (86.4%) Education years 16 (12–16)^ MS characteristics  Type of MS   Relapsing–remitting 112 (89.6%)   Progressive 13 (10.4%)   Disease Duration 8 (4–11)^   EDSS 1.5 (1–2.5)^  Disease modifying therapies   None 9 (7%)   Azathioprine 1 (0.8%)   Dimethyl fumarate 15 (12%)   Fingolimod 5 (4%)   Glatiramer acetate 2 (1.6%)   Interfron-a 38 (30.4%)   Natalizumab 2 (1.6%)   Ocrelizumab 14 (11.2%)   Rituximab 24 (19.2%)   Terifluromide 15 (12%)  Motor function   T25FW 5.36 (4.89–6.17)^   9-HPT 22.5 (20.94–25.82)^   Dominant hand 9-HPT 22 (20–24.46)^   None-dominant hand 9-HPT 24 (21–26)^  Cognitive function   SDMT 48 (42–53)^   CVLT-II 48.78 (8.85)*   BVMT-R 26 (20–29)^ *Mean (standard deviation). ^median (Q1–Q3). EDSS, expanded disability status scale; SDMT, symbol digit modalities test; CVLT-II, California verbal learning test-II; BVMT-R, brief visuospatial memory test-revised; T25FW, timed 25-foot walk test; 9-HPT, 9-hole peg test

Characteristics of the included population

*Mean (standard deviation). ^median (Q1–Q3). EDSS, expanded disability status scale; SDMT, symbol digit modalities test; CVLT-II, California verbal learning test-II; BVMT-R, brief visuospatial memory test-revised; T25FW, timed 25-foot walk test; 9-HPT, 9-hole peg test

Correlates of cognitive functions

Age was not significantly associated with SDMT, CVLT-II, or BVMT-R scores, and disease duration was only correlated with CVLT-II ( ρ  = − 0.18, p value = 0.042). On the other hand, the strongest associations observed were those between T25FW and cognitive domains, with SDMT ( ρ  = -0.45, p value < 0.0001), BVMT-R ( ρ  = − 0.29, p value  =  0.001), and CVLT-II ( ρ  = − 0.28, p value = 0.002) showing the strongest correlations. The next was 9-HPT, which was also significantly associated with SDMT ( ρ  = − 0.43, p value < 0.0001), CVLT-II ( ρ  = − 0.32, p value  =  0.0003), and BVMT-R ( ρ  = − 0.26, p value = 0.004). EDSS was also correlated with all three, with SDMT ( ρ  = − 0.41, p value < 0.0001), CVLT-II ( ρ  = − 0.21, p value = 0.0209), and BVMT-R ( ρ  = − 0.19, p value = 0.031) showing the strongest associations, respectively. Finally, years of education were most strongly associated with BVMT-R ( ρ  = 0.24, p value = 0.006), CVLT-II ( ρ  = 0.24, p value = 0.0072), and SDMT ( ρ  = 0.22, p value = 0.0123), respectively (Table  2 ; Fig.  1 ). Table 2 Correlation matrix between cognitive domains and demographic variables, MS characteristics, and motor function Variables SDMT BVMT-R CVLT-II Age Coefficient − 0.16 − 0.08 − 0.06* p value 0.0769 0.378 0.5060 Education Coefficient 0.22 0.24 0.24 p value 0.0123 0.006 0.0072 Disease Duration Coefficient − 0.16 − 0.10 − 0.18 p value 0.078 0.2708 0.042 EDSS Coefficient − 0.41 − 0.19 − 0.21 p value < 0.0001 0.031 0.0209 T25FW Coefficient − 0.45 − 0.29 − 0.28 p value < 0.0001 0.001 0.002 9HPT Coefficient − 0.43 − 0.26 − 0.32 p value < 0.0001 0.004 0.0003 Bold values represent significant results. Data are Spearman’s rank correlation coefficients unless stated by *Pearson’s correlation coefficient. EDSS, expanded disability status scale; SDMT, symbol digit modalities test; CVLT-II, California verbal learning test-II; BVMT-R, brief visuospatial memory test-revised; T25FW, timed 25-foot walk test; 9-HPT, 9-hole peg test Fig. 1 Correlations between cognitive domains assessed by BICAMS and timed 25-foot walk test scores

Table 2 Correlation matrix between cognitive domains and demographic variables, MS characteristics, and motor function Variables SDMT BVMT-R CVLT-II Age Coefficient − 0.16 − 0.08 − 0.06* p value 0.0769 0.378 0.5060 Education Coefficient 0.22 0.24 0.24 p value 0.0123 0.006 0.0072 Disease Duration Coefficient − 0.16 − 0.10 − 0.18 p value 0.078 0.2708 0.042 EDSS Coefficient − 0.41 − 0.19 − 0.21 p value < 0.0001 0.031 0.0209 T25FW Coefficient − 0.45 − 0.29 − 0.28 p value < 0.0001 0.001 0.002 9HPT Coefficient − 0.43 − 0.26 − 0.32 p value < 0.0001 0.004 0.0003 Bold values represent significant results. Data are Spearman’s rank correlation coefficients unless stated by *Pearson’s correlation coefficient. EDSS, expanded disability status scale; SDMT, symbol digit modalities test; CVLT-II, California verbal learning test-II; BVMT-R, brief visuospatial memory test-revised; T25FW, timed 25-foot walk test; 9-HPT, 9-hole peg test

Correlation matrix between cognitive domains and demographic variables, MS characteristics, and motor function

Bold values represent significant results. Data are Spearman’s rank correlation coefficients unless stated by *Pearson’s correlation coefficient. EDSS, expanded disability status scale; SDMT, symbol digit modalities test; CVLT-II, California verbal learning test-II; BVMT-R, brief visuospatial memory test-revised; T25FW, timed 25-foot walk test; 9-HPT, 9-hole peg test

Correlations between cognitive domains assessed by BICAMS and timed 25-foot walk test scores

Predictors of cognitive functions

Table 3 delineates the results of the multivariable stepwise regression performed to estimate the predictors of each cognitive domain. The model constructed to predict SDMT was the most successful among the three during the first ( R 2  = 22.35%, p value < 0.001) and the second step ( R 2  = 38.15%, p value < 0.001). EDSS (beta coefficient = − 3.340, p value < 0.001) and disease duration (beta coefficient = − 0.294, p value = 0.030) were found to contribute to SDMT in the first step significantly. In the second step, both EDSS (beta coefficient = − 1.418, p value = 0.029) and disease duration (beta coefficient = − 0.258, p value = 0.033) were retained as significant predictors. In addition, T25FW emerged as a significant factor and was the strongest predictor among all variables (beta coefficient = − 3.100, p value < 0.001), yet 9-HPT was not independently associated with the SDMT in step 2. Table 3 Hierarchical stepwise regression for assessing the relationship between cognitive function and demographic and clinical variables Cognitive domain Coefficient, p value R 2 , Model p value Beta SDMT Step 1 EDSS − 3.34, < 0.001 22.35%, < 0.001 − 0.44 Disease Duration − 0.29, 0.030 − 0.18 Step 2 EDSS − 1.42, 0.029 38.15%, < 0.001 − 0.19 T25FW − 3.10, < 0.001 − 0.47 Disease Duration − 0.26, 0.033 − 0.15 CVLT-II Step 1 Education 0.68, 0.020 11.56%, 0.002 0.21 EDSS − 1.54, 0.029 − 0.19 Disease Duration − 0.31, 0.043 − 0.18 Step 2 Education 0.77, 0.005 24.02%, < 0.001 0.24 EDSS 0.31, 0.691 0.038 Disease Duration − 0.28, 0.048 − 0.16 T25FW − 2.91, < 0.001 − 0.42 BVMT-R Step 1 Education 0.63, 0.005 8.62%, 0.004 0.25 Disease Duration − 0.25, 0.033 − 0.19 Step 2 Education 0.60, 0.004 22.70%, < 0.001 0.24 Disease Duration − 0.23, 0.037 − 0.17 T25FW − 1.98, < 0.001 − 0.38 EDSS, expanded disability status scale; SDMT, symbol digit modalities test; CVLT-II, California verbal learning test-II; BVMT-R, brief visuospatial memory test-revised; T25FW, timed 25-foot walk test; 9-HPT, 9-hole peg test

Table 3 Hierarchical stepwise regression for assessing the relationship between cognitive function and demographic and clinical variables Cognitive domain Coefficient, p value R 2 , Model p value Beta SDMT Step 1 EDSS − 3.34, < 0.001 22.35%, < 0.001 − 0.44 Disease Duration − 0.29, 0.030 − 0.18 Step 2 EDSS − 1.42, 0.029 38.15%, < 0.001 − 0.19 T25FW − 3.10, < 0.001 − 0.47 Disease Duration − 0.26, 0.033 − 0.15 CVLT-II Step 1 Education 0.68, 0.020 11.56%, 0.002 0.21 EDSS − 1.54, 0.029 − 0.19 Disease Duration − 0.31, 0.043 − 0.18 Step 2 Education 0.77, 0.005 24.02%, < 0.001 0.24 EDSS 0.31, 0.691 0.038 Disease Duration − 0.28, 0.048 − 0.16 T25FW − 2.91, < 0.001 − 0.42 BVMT-R Step 1 Education 0.63, 0.005 8.62%, 0.004 0.25 Disease Duration − 0.25, 0.033 − 0.19 Step 2 Education 0.60, 0.004 22.70%, < 0.001 0.24 Disease Duration − 0.23, 0.037 − 0.17 T25FW − 1.98, < 0.001 − 0.38 EDSS, expanded disability status scale; SDMT, symbol digit modalities test; CVLT-II, California verbal learning test-II; BVMT-R, brief visuospatial memory test-revised; T25FW, timed 25-foot walk test; 9-HPT, 9-hole peg test

Hierarchical stepwise regression for assessing the relationship between cognitive function and demographic and clinical variables

EDSS, expanded disability status scale; SDMT, symbol digit modalities test; CVLT-II, California verbal learning test-II; BVMT-R, brief visuospatial memory test-revised; T25FW, timed 25-foot walk test; 9-HPT, 9-hole peg test

CVLT-II was the next most successfully predicted cognitive domain in the first ( R 2  = 11.56%, p value = 0.002) and second steps ( R 2  = 24.02%, p value < 0.001). EDSS (beta coefficient = − 1.540, p value = 0.029), education (beta coefficient = 0.677, p value = 0.020), and disease duration (beta coefficient = − 0.310, p value = 0.043) were found to be significant predictors in the first step. In the second step, education (beta coefficient = 0.773, p value = 0.005) and disease duration (beta coefficient = − 0.283, p value = 0.048) were retained as significant predictors, yet EDSS lost its significance. In addition, T25FW was also found to be a significant predictor and was the most strongly associated with CVLT-II (beta coefficient = − 2.910, p value < 0.001), yet 9-HPT was not independently associated with the CVLT-II.

BVMT-R was the least successfully predicted in the first ( R 2  = 8.62%, p value = 0.004) and second steps ( R 2  = 22.70%, p value < 0.001). Education (beta coefficient = 0.628, p value = 0.005) and disease duration (beta coefficient = -0.253, p value = 0.033) were significant predictors of BVMT-R in the first step. In the second step, education (beta coefficient = 0.595, p value = 0.004) and disease duration (beta coefficient = − 0.229, p value = 0.037) both maintained their significance, and T25FW, again, emerged as the most predictive variable (beta coefficient = − 1.984, p value < 0.001), and again the 9-HPT was not independently associated with BVMT-R.

Discussion

This cross-sectional study examined the correlation and predictive value of physical function in relation to CI, along with predictors of different cognitive domains in a sample of PwMS. Among the performance-based measures, the T25FW showed the most positive associations with cognitive performance, followed by the 9-HPT, which was also significantly correlated with cognitive functions. Among the predictive models, SDMT was the most accurately predicted, and BVMT-R was the least accurately predicted cognitive domain. Furthermore, the T25FW was found to be the most significant predictor in all three cognitive domains, while the 9-HPT did not reach significance in any.

In this study, we observed that T25FW is a more reliable correlate of cognition compared to EDSS. EDSS showed a moderate but consistent negative association with all the cognitive domains, reflecting that greater physical disability is associated with poorer cognitive performance. This is in line with the findings of Caneda et al. [ 32 ] and Virgilio et al. [ 33 ]. As a global measure of neurological disability, EDSS encompasses a wide range of functional impairments that may account for its moderate impact on cognition in isolation [ 22 ]. In all regression models, EDSS was a significant predictor, especially for SDMT and CVLT-II, but its influence diminished when T25FW was considered. This supports the view that while EDSS reflects general overall disability, it may not capture specific dynamic features of motor function that are more directly related to cognitive outcomes. This loss of EDSS predictive strength upon adding the T25FW underlines the possibility that cognitive impairment may relate more directly to current physical limitation rather than to the global disability score represented by EDSS.

By contrast, T25FW was the strongest predictor of cognitive performance in all fit models, outperforming EDSS in its association with BICAMS measures. The T25FW test directly captures the motor speed and coordination functions inextricably linked to many cognitive processes [ 34 ], such as attention [ 35 ], speed of processing [ 36 ], and memory [ 37 ]. Importantly, EDSS has been characterized as time-consuming and expensive [ 38 ], while T25FW requires minimal temporal and spatial effort [ 39 ]. As a result, T25FW represents a more practical tool for clinical settings. Although this study specifically focused on cognitive associations, prior work has indicated that T25FW is a more sensitive and useful measure of disability outcomes compared to EDSS due to the former's better indication of subtle changes in physical function [ 40 , 41 ]. Such findings suggest that while EDSS provides useful information regarding overall disability, T25FW may function as a clinically effective and sensitive measure of change in mobility as well as having significant associations with cognition. It could be valuable as an alternative or supplement when physical and cognitive functions are being routinely assessed with EDSS in settings to enhance the responsiveness and efficiency of clinical assessments. However, our results contrast with a recent cross-sectional study, which found no correlation between T25FW and BICAMS measures [ 42 ], indicating the need for further research to clarify these relationships.

The correlation of manual dexterity with CI, however, is not as robust. Significant correlations of the 9-HPT were found with all three BICAMS cognitive domains, indicating that fine motor skills may be associated with cognitive performance, particularly in the speed of processing and memory. This is consistent with recent studies that observed correlations between BICAMS and the 9-HPT [ 42 , 43 ]. It did not, however, appear in the regression analysis as a significant predictor of cognitive outcomes. Hence, though 9-HPT is related to cognitive function, its impact perhaps is nominal compared to that brought about by other factors, such as total disability and mobility, which are captured by EDSS and T25FW. This lack of predictive significance perhaps reflects shared variance with these broader measures. It would imply that 9-HPT is useful in assessing motor function but adds no independent value to the prediction of cognitive decline.

In a similar study, Benedict et al. [ 36 ] investigated the association between motor function and cognition using motor function as the independent variable, attempting to identify which measures of cognition are mostly associated with motor function. They found both 9-HPT and T25FW to be independently associated with SDMT, with T25FW also independently associated with CVLT-II as well. Similar to our findings, the strongest relationship among cognitive domains was with the SDMT, yet we found no association between 9-HPT and cognition. This could be due to our inclusion of both 9-HPT and T25FW as predictors, with T25FW having a stronger association that rendered 9-HPT redundant. This analysis highlights the practicality of T25FW compared to 9-HPT.

The robustness of this association between T25FW and cognition compared to 9-HPT may extend to other motor assessment scales as well. While the relationship between other scales such as the 6-min walk test (6MWT), six-spot step-test (SSST), and the 12-item MS walking scale and cognition has not been explored, the SSST is superior to the 6MWT and the 6MWT is superior to T25FW test in detecting walking impairments [ 44 ]. 6MWT has also been linked to global cognitive function in heart failure patients and healthy older adults [ 45 , 46 ], and SSST is associated with processing speed in MS [ 47 ], yet further studies are required to both assess the association between 6MWT and cognition in MS and to identify the motor scale most closely related to the patients cognitive status.

Education was linearly positively related to cognitive performance on all the BICAMS, therefore, providing the idea that higher levels of education may be protective against a decline in cognitive function. In the current study, education has emerged as a significant predictor through regression analyses, mostly for BVMT-R and CVLT-II, thus pointing to possible protective processes in memory and to a lesser extent in processing abilities. These findings are in line with the theory of cognitive reserve, where higher educational attainment may provide a better ability to tolerate or compensate for neurological changes accompanying disease progression [ 48 ]. They are also consistent with the findings of Rimkus et al. [ 49 ] and Estrada-López et al. [ 10 ] regarding the protective effect of higher levels of education on cognitive dysfunction in PwMS and its predictive value for future CI [ 33 ]. Interestingly, although education was less strongly predictive than variables, such as T25FW, the consistency of its association underlined the role of lifelong learning and intellectual engagement in maintaining cognitive health. A relationship between education and cognitive function such as this might suggest that individuals with a higher educational background experience a form of resilience to cognitive decline, perhaps due to greater neural plasticity or more effective cognitive strategies. These findings provide a rough outline of the gains that are possible through educational achievement and suggest that cognitive stimulation and learning interventions could be of value even when formal education has been completed, especially for populations at risk of cognitive impairment.

Whereas one might intuitively think that longer disease duration should be related to poorer cognitive performance, the initial correlational analysis revealed no significant association between disease duration and any of the BICAMS cognitive measures. Where such an association is absent in any model, it may indicate that having the disease for a longer period does not directly predict lower scores on these cognitive domains, emphasizing that the length of disease exposure may not solely drive cognitive impairment in this population. It could be that physical functioning, disability progression, or individual variability in disease pathology is a more powerful factor in cognitive outcome than disease duration per se. However, it did emerge as a modest yet significant predictor for each cognitive domain in multivariable regression models. That is to say, although isolated, it may not relate directly, its influence becomes more lucid when interlinked with other factors, such as disability and mobility. This subtle role of duration suggests that the development of disease-related influences in cognition may be related more to associated physical and functional changes over time than to the mere passage of time with the disease. However, there are discrepancies between studies regarding the correlation between disease duration and cognitive impairment. For example, Achiron et al. [ 50 ] reported a higher proportion of cognitive impairment with increased disease duration, while Lynch et al. [ 51 ] found no correlation between the two. In contrast, Virgilio et al. [ 33 ] found disease duration with significant predictive value for future CI. These findings, therefore, reinforce the need to focus on disability management and the preservation of functional abilities as a means of potentially mitigating cognitive decline, even in persons with long-standing diseases.

Limitations

This study has some limitations worth considering. The cross-sectional nature and the sample size of the study limit the generalizability of our findings; thus, further verification is required for the generalizability of these results in the diverse population of PwMS. Moreover, the potential influences of unmeasured confounders include comorbid conditions, lifestyle differences, medication use, and white matter lesion load that were not considered, which may alter the associations found between physical and cognitive function. Our cognitive assessment was based on the BICAMS, which the National MS Society recommends as an alternative to complete neuropsychological assessments, yet future studies should use more extensive cognitive batteries such as the Minimal Assessment of Cognitive Function in MS (MACFIMS) as it includes an executive function domain as well. Although BICAMS is a robust and reliable tool for evaluating cognitive functions in PwMS [ 52 , 53 ], incorporating a general cognitive screening tool like the Montreal Cognitive Assessment (MoCA) in future studies could offer a more comprehensive evaluation of cognitive function. Memory was assessed using BVMT-R and CVLT-II, which primarily measure visuospatial and verbal learning, respectively. This selection was based on its brevity, clinical feasibility, and extensive validation of BICAMS in PwMS. However, additional measures such as short-term recall, long-term recall, and recognition from the BVMT-R and Rey Auditory Verbal Learning Test (RAVLT) could provide a more detailed evaluation of memory performance and should be addressed in future studies. According to the BICAMS, an oral version of the SDMT was used, yet a written version may have shown a more significant association with the 9-HPT or T25FW due to graphokinetic factors. This may also explain why some studies have found significant associations between manual dexterity and SDMT [ 17 ].

Conclusion

This study highlights the connection between physical function and cognitive performances in PwMS. The T25FW emerged as the strongest predictor of cognitive performance, showing a positive relationship with all BICAMS tests, while EDSS was also related to cognitive scores but less effectively than T25FW. These findings suggest that T25FW could be considered as a sensitive and clinically practical alternative to EDSS while having a stronger association with cognition. We recommend that future, longitudinal studies investigate this relationship and account for a wide array of cofounders to discover more reliable associations.

Supplementary Information

Additional file 1.

Additional file 1.