1. Introduction
Multiple sclerosis (MS) is a chronic inflammatory demyelinating disease of the central nervous system that frequently involves the visual pathway. Optic neuritis (ON) is a common manifestation of MS, occurring in up to 50–70% of patients during the disease course [ 1 , 2 ]. Even in the absence of overt ON, subclinical demyelination of the optic nerve and visual pathways is often present in MS, contributing to subtle visual dysfunction that may be unrecognised [ 3 , 4 , 5 ]. Because the visual system can be readily evaluated using both structural and functional measures, it has emerged as a useful model for studying MS-related neuroinflammation and neurodegeneration in vivo [ 4 , 6 ]. In particular, damage to the anterior visual pathway provides insight into broader MS pathology, given that retinal axons and optic nerve myelin are frequently targets of the disease from its earliest stages [ 7 ].
Visual evoked potentials (VEPs) are well-established non-invasive neurophysiological tools for assessing the integrity of visual pathways [ 8 ]. In MS, conventional full-field pattern-reversal VEPs have long been used to detect optic nerve lesions that may be clinically silent, thereby providing evidence of dissemination in space, even when patients have no visual symptoms [ 9 , 10 ]. The P100 component of pattern-reversal VEP, a positive waveform peaking at approximately 100 ms, is of particular clinical relevance. Delayed P100 latency in patients with MS is highly sensitive to optic nerve demyelination or incomplete remyelination and can be observed even in eyes without a history of acute ON [ 8 , 11 ]. Thus, VEPs serve as an objective means to uncover subclinical visual pathway involvement in MS, complementing other diagnostic modalities, such as MRI. The diagnostic relevance of VEP has recently been reaffirmed by the 2024 revision of the McDonald criteria, which for the first time incorporates the optic nerve as a fifth central nervous system topography and endorses VEP as objective evidence of optic nerve involvement for demonstrating dissemination in space [ 12 ]. In addition to their diagnostic utility, VEPs are increasingly recognized for their potential use in monitoring diseases and as outcome measures in clinical trials aimed at neuroprotection or repair [ 7 , 9 ]. Cross-sectional studies have demonstrated that prolongation of VEP latency correlates with greater MRI lesion load and even with brain atrophy, suggesting that slower visual conduction may reflect a broader impact of MS on central nervous system integrity [ 4 , 13 ]. VEP abnormalities have also been associated with clinical disability; for example, patients with prolonged P100 latencies or reduced amplitudes tend to have higher Expanded Disability Status Scale (EDSS) scores and worse functional outcomes [ 3 , 14 ]. Longitudinal assessment of VEPs could offer a means to track subclinical disease progression in MS; however, to date, only a few studies have addressed this in detail [ 3 , 6 , 7 ]. Earlier investigations suggested that serial evoked potential testing might predict future outcomes. For example, Fuhr et al. [ 15 ] reported that VEP and motor EP changes over time paralleled the clinical evolution of MS, and Kallmann et al. [ 16 ] reported that patients with abnormal baseline EPs developed greater disability during follow-up. Despite these examples, no consensus has been reached on how to incorporate serial VEP monitoring into routine practice, largely because robust longitudinal data linking VEP changes to definitive clinical or radiological outcomes are limited. In particular, the question remains whether progressive latency prolongation or amplitude loss on VEP truly anticipates an upcoming disability or MRI changes, or whether these measures simply mirror damage that has already occurred.
This exploratory study aimed to assess the clinical utility of longitudinal VEP measurements in relapsing–remitting MS (RRMS), focusing on whether latency and amplitude changes show distinct associations with concurrent inflammatory activity and disability.
2. Materials and Methods
2.1. Study Design and Participants
This multicenter retrospective cohort study was conducted using data from three MS centers in Turkey: Sakarya Training and Research Hospital (Sakarya), Bursa Yüksek İhtisas Training and Research Hospital (Bursa), and Bağcılar Training and Research Hospital (Istanbul). All patients were diagnosed with RRMS according to the 2017 revised McDonald criteria [ 17 ]. In accordance with national regulations for multicenter retrospective studies, institutional permission was first obtained from the administration of each participating hospital, after which a single ethics approval covering all three centers was granted by the Sakarya University Faculty of Medicine Ethics Committee (approval number: 150, 19 November 2024). The study was conducted in accordance with the Declaration of Helsinki. Owing to the retrospective design of the study, the requirement for written informed consent was waived by the Ethics Committee of Sakarya University Faculty of Medicine. From an initial pool of 2980 screened records, 83 patients met the inclusion criteria: age 18–65 years, at least two full-field pattern-reversal VEP assessments performed ≥1 year apart, and complete clinical and radiological data. Exclusion criteria were as follows: comorbid conditions affecting visual or neurological function (e.g., diabetes mellitus, Graves’ disease, uveitis), baseline EDSS score > 5.5, incomplete clinical or imaging data, use of nonstandard immunosuppressive therapies or multiple disease-modifying treatments within 6 months before or between VEP assessments, MS relapse or suspected optic neuritis within 3 months prior to VEP testing, and any episode of optic neuritis occurring between the two VEP assessments. The patient selection process is illustrated in Figure 1 .
2.2. Clinical and Radiological Data Collection
Demographic variables included age and sex. Clinical variables comprised disease duration, EDSS score, and annualized relapse rate (ARR). ARR was calculated as the number of clinically confirmed relapses occurring between the first and second VEP assessments divided by the length of that interval in years; thus, ARR and longitudinal VEP changes refer to the same observation window, and the associations reported are concurrent rather than predictive. EDSS was recorded at the first VEP, at the second VEP, and at the last available clinical visit. Treatment type and duration during the interval between VEP assessments were also recorded. Disease-modifying therapies (DMTs) used during this interval were grouped, in line with published efficacy classifications and the ECTRIMS/EAN guideline framework [ 18 , 19 ], as platform or moderate-efficacy (interferon-β, glatiramer acetate, teriflunomide, dimethyl fumarate) or high-efficacy (fingolimod, natalizumab, cladribine, ocrelizumab) agents. Baseline MRI variables included the presence of lesions in the brainstem, spinal cord, and upper cervical cord (C1–C4), total T2 lesion burden, and the presence of gadolinium-enhancing lesions. Brain and spinal cord MRI at all three centers were acquired on 1.5 Tesla MRI systems (GE Healthcare, Chicago, IL, USA) using a standardized MS protocol comprising axial and sagittal T2-weighted/FLAIR sequences and pre- and post-gadolinium T1-weighted sequences; the same field strength and acquisition protocol were used across sites, and images were rated by the treating neurologists at each center. Baseline MRI was the diagnostic scan performed within a maximum of 6 months of the first VEP, and baseline MRI data were available for all 83 included patients, as complete imaging data were an inclusion criterion. Cerebrospinal fluid (CSF) analysis, performed at diagnosis within the same time window, included oligoclonal band (OCB) status and IgG index; CSF data were available in 72 patients.
2.3. Visual Evoked Potentials (VEPs)
VEP assessments were performed routinely, independent of clinical suspicion of relapse or optic neuritis, using a Keypoint Classic electromyography/evoked potential system (Alpine Biomed ApS, Skovlunde, Denmark). All three centers used identical devices and standardized stimulus displays in accordance with a shared protocol based on the International Society for Clinical Electrophysiology of Vision (ISCEV) guidelines, ensuring consistency in stimulus presentation and recording across sites [ 8 ].
Pattern-reversal stimulation was delivered using a high-contrast checkerboard pattern (8 × 8 checks), with each check subtending approximately 1° of visual angle at a viewing distance of 100 cm. Recordings were obtained under controlled conditions, including standardized ambient lighting, room temperature (22–24 °C), and stable fixation. Electrodes were placed according to the international 10–20 EEG system, with the active electrode at Oz, reference at Fz, and ground at Cz. Monocular full-field stimulation was applied, and P100 latency and N75-P100 peak-to-peak amplitudes were measured bilaterally.
Given the retrospective design and absence of a locally derived normative dataset, abnormality thresholds were defined based on ISCEV-compliant reference values. A P100 latency > 118 ms and amplitude < 5 µV were considered abnormal [ 8 , 20 ]; these cut-offs defined cross-sectional (single-timepoint) abnormality. Separately, longitudinal change between the two assessments was evaluated: latency prolongation was defined as an increase of ≥2 ms from baseline in at least one eye (a longitudinal-change threshold, distinct from the cross-sectional cut-off above), whereas amplitude reduction was defined as a decrease of >0.3 µV from baseline or a value below 5 µV in at least one eye. Latency and amplitude abnormalities were assessed at the individual-eye level, and a patient was classified as showing latency prolongation or amplitude reduction if either eye met the respective threshold; the two classifications were defined independently and were not mutually exclusive. Associations between longitudinal changes in VEP parameters (latency and amplitude) and disability outcomes were evaluated using correlation analyses.
2.4. Statistical Analysis
Statistical analyses were performed using IBM SPSS Statistics for MacOS (version 30.0; IBM Corp., Armonk, NY, USA). Categorical variables are reported as frequencies and percentages, and continuous variables as medians (range). The Kolmogorov–Smirnov test was used to assess normality. Mann–Whitney U tests were used to compare continuous variables, and Pearson’s chi-square or Fisher’s exact tests were used to compare categorical variables. Spearman’s rank correlation was used to evaluate the relationship between the VEP changes and EDSS worsening. Logistic regression analysis was performed to identify factors independently associated with latency prolongation. The discriminative ability of longitudinal latency change for concurrent on-study relapse activity (ARR > 0 during the inter-assessment interval) was assessed by receiver-operating-characteristic (ROC) analysis, reporting the area under the curve (AUC) with bootstrap 95% confidence intervals. To account for the non-independence of the two eyes, a generalized estimating equation (GEE) logistic model with an exchangeable working correlation was fitted at the eye level. Between-group differences in ordinal outcomes were additionally quantified using Hodges–Lehmann median differences with 95% confidence intervals. Confirmed disability worsening was defined a priori as an EDSS increase of ≥1.5 points from a baseline of 0, ≥1.0 point from a baseline of 1.0–5.5, or ≥0.5 points from a baseline > 5.5 at the second VEP, sustained at the last available visit; progression independent of relapse activity (PIRA) was defined as confirmed disability worsening in the absence of any relapse during the inter-assessment interval, and relapse-associated worsening (RAW) as confirmed worsening with at least one relapse, following established definitions [ 21 , 22 ]. To address potential confounding by treatment, sensitivity analyses were performed in which (i) patients receiving moderate-efficacy versus high-efficacy DMT were compared, (ii) the principal associations were re-examined within each DMT stratum, and (iii) DMT efficacy tier was entered as a covariate in the logistic, GEE, and linear regression models. Statistical significance was defined as a two-tailed p -value of <0.05. Because of the exploratory nature of the study, results of secondary and sensitivity analyses are reported without adjustment unless stated otherwise and should be interpreted as hypothesis-generating.
3. Results
3.1. Demographic and Baseline Characteristics
A total of 83 patients with RRMS were included, of whom 79.5% were female, with a median age of 36 years (range, 18–60). The median disease duration was 6.0 years, and the median baseline EDSS score was 1.5 (range, 0–5.5). The median interval between the initial and follow-up VEP assessments was 2.0 years, during which the median ARR was 0.3. At baseline, median P100 latency was 116 ms in the right eye and 114 ms in the left eye, with corresponding amplitudes of 9.0 µV and 9.5 µV. Baseline MRI revealed brainstem lesions in 32.5% of patients, spinal cord lesions in 57.8%, and gadolinium-enhancing lesions in 38.6%. During the inter-assessment interval, 54 patients (65.1%) received platform or moderate-efficacy therapies—most commonly dimethyl fumarate (22.9%) and teriflunomide (19.3%), 24 patients (28.9%) received high-efficacy therapies (fingolimod, natalizumab, cladribine, or ocrelizumab), and 5 (6.0%) were untreated ( Table 1 , Supplementary Table S3 ).
3.2. Longitudinal Changes in VEP Latency and Amplitude and Their Distinct Clinical Associations
During follow-up, 36 patients (43.4%) developed P100 latency prolongation in at least one eye. This group exhibited a markedly higher ARR during the same interval compared with those without latency prolongation ( p < 0.001). Baseline MRI findings revealed a lower lesion burden in the latency prolongation group, including fewer brainstem lesions, spinal cord lesions, and upper cervical cord involvement ( p = 0.047, p = 0.005, and p = 0.038, respectively). Gadolinium-enhancing lesions were also less frequent in this group ( p = 0.004). In parallel, the median CSF IgG index was lower among patients with latency prolongation ( p = 0.035). No significant differences were observed between groups in terms of disease duration, baseline EDSS score, or EDSS worsening ( Table 2 ).
A total of 47 patients (56.6%) exhibited a significant reduction in P100 amplitude during follow-up. Compared with patients without amplitude reduction, this group had shorter baseline right-eye latency (median 112 ms vs. 119.5 ms, p = 0.028), while disease duration and ARR were similar between groups. Despite comparable baseline EDSS scores, patients with amplitude reduction had higher EDSS scores both at the second VEP assessment ( p = 0.017) and at final follow-up ( p = 0.028), although the magnitude of the difference was small (0.5 EDSS points). No significant differences were observed between groups in baseline MRI or CSF parameters ( Table 3 , Figure 2 ).
3.3. Correlation and Multivariate Analyses
Changes in P100 latency were not significantly correlated with EDSS worsening (all p > 0.15). Among eight Spearman correlation tests performed across VEP parameters and time intervals, a nominally significant inverse correlation was observed between left-eye P100 amplitude change and EDSS change between the first and second VEP assessments ( p = 0.036); however, this association did not survive Bonferroni correction for multiple comparisons, and should therefore be interpreted as an exploratory finding. No significant correlations were identified for other VEP parameters or time intervals ( Table 4 ). Logistic regression analysis including age, disease duration, baseline EDSS score, and ARR identified ARR as the only variable independently associated with latency prolongation ( p = 0.005). No variables were independently associated with amplitude reduction. In a GEE logistic model accounting for inter-eye correlation, latency prolongation remained associated with the ARR ( p < 0.001). Longitudinal latency change discriminated patients with concurrent on-study relapse activity with an AUC of 0.72 (95% CI 0.60–0.83), whereas baseline latency did not (AUC 0.52) ( Figure 3 ). Latency prolongation and amplitude reduction overlapped only partially; 23 patients showed both, 13 latency prolongation only, 24 amplitude reduction only, and 23 neither, indicating that the two classifications were largely independent of each other. The higher follow-up EDSS accompanying amplitude reduction corresponded to a Hodges–Lehmann median difference of 0.5 points (95% CI 0.0–0.5), and the higher relapse activity accompanying latency prolongation to a median ARR difference of 0.5 (95% CI 0.2–1.0). In a linear regression model of EDSS at the second VEP adjusted for baseline EDSS, ARR, and DMT efficacy tier, amplitude reduction was associated with a non-significant 0.29-point higher EDSS (95% CI −0.02 to 0.60, p = 0.069), indicating that the unadjusted group difference was attenuated after accounting for baseline disability.
3.4. Formally Defined Disability Worsening and PIRA
Using the pre-specified definition, confirmed disability worsening between the first and second VEP that was sustained at the last visit occurred in 8 patients (9.6%): 3 (3.6%) fulfilled the criteria for PIRA and 5 (6.0%) for RAW. All three PIRA events and 7 of the 8 confirmed worsening events occurred in patients with amplitude reduction (PIRA: 3/47 vs. 0/36, Fisher’s exact p = 0.25; confirmed worsening: 7/47 [14.9%] vs. 1/36 [2.8%], p = 0.13), whereas none of the PIRA events occurred in patients with latency prolongation (0/36 vs. 3/47, p = 0.25). Among the 40 relapse-free patients, EDSS at the second VEP did not differ significantly between those with and without amplitude reduction ( p = 0.068), and the amplitude–EDSS correlation was not significant in this subgroup ( p = 0.31). Given the very small number of events, these analyses are descriptive only and do not permit inference regarding amplitude reduction as a marker of PIRA.
3.5. Sensitivity Analyses According to Disease-Modifying Therapy
Patients receiving high-efficacy DMT ( n = 24) had a longer disease duration ( p = 0.014), a higher baseline EDSS ( p = 0.015), and more frequent gadolinium-enhancing ( p = 0.005) and spinal cord lesions ( p = 0.006) than those receiving moderate-efficacy DMT ( n = 54), consistent with treatment escalation in patients with more active disease ( Supplementary Table S3 ). During the inter-assessment interval, high-efficacy DMT was associated with fewer patients experiencing relapses ( p = 0.049), a lower ARR ( p = 0.050), and a lower frequency of latency prolongation ( p = 0.012), with a median left-eye latency change of −5 ms versus +1 ms ( p = 0.004); the frequency of amplitude reduction ( p = 0.23) and the change in EDSS ( p = 0.34) did not differ between tiers. The association between latency prolongation and ARR was present within both strata (moderate-efficacy: median ARR 0.66 vs. 0.0, p = 0.022; high-efficacy: 1.0 vs. 0.0, p = 0.034), and ARR remained associated with latency prolongation after adding DMT tier to the multivariable logistic model (OR 2.73, 95% CI 1.29–5.79, p = 0.009; high-efficacy DMT OR 0.44, 95% CI 0.13–1.57, p = 0.21) and to the eye-level GEE model (OR 2.39, 95% CI 1.40–4.05, p = 0.001). The AUC of longitudinal latency change for concurrent relapse activity was 0.71 in the moderate-efficacy and 0.62 in the high-efficacy stratum. In contrast, the higher EDSS at the second VEP among patients with amplitude reduction was observed in the moderate-efficacy stratum ( p = 0.007) but not in the smaller high-efficacy stratum ( p = 0.40). In an exploratory logistic model, high-efficacy DMT was associated with lower odds of amplitude reduction (OR 0.27, 95% CI 0.08–0.87, p = 0.028), independent of age, disease duration, baseline EDSS, and ARR. The five untreated patients were too few for separate analysis.
3.6. Baseline VEP Abnormalities
At baseline, prolonged P100 latency (>118 ms) was present in 31.3% of patients and was associated with a higher frequency of spinal cord lesions ( p = 0.032), but not with ARR or EDSS measures at baseline or during follow-up. Reduced baseline amplitude was associated with higher ARR during follow-up ( p = 0.044), a higher frequency of visual onset symptoms ( p = 0.029), older age ( p = 0.045), and fewer gadolinium-enhancing lesions ( p = 0.015). No significant differences were observed between groups in EDSS values at baseline or at the last assessment ( Supplementary Table S1 ).
4. Discussion
In this multicenter retrospective cohort study, we longitudinally evaluated conventional VEP metrics in 83 patients with RRMS. Our findings suggest that longitudinal latency and amplitude changes carry different clinical associations in RRMS. P100 latency prolongation was associated with higher relapse activity during the same observation window, with ARR remaining the sole variable independently associated with latency prolongation in multivariable analysis, and this association was robust to stratification and adjustment for DMT efficacy tier. In contrast, amplitude reduction showed a weaker, group-level association with disability: EDSS at follow-up was 0.5 points higher despite similar baseline scores, but the corresponding correlation analysis yielded only a nominally significant association that did not survive Bonferroni correction for eight simultaneous comparisons, and the group difference was attenuated after adjustment for baseline EDSS. EDSS worsening was not associated with latency changes. Post hoc power analysis further indicated that the study was underpowered for the amplitude–EDSS correlation, so the association requires confirmation in a larger cohort. When a formal definition of PIRA was applied, only three events were identified; although all occurred in patients with amplitude reduction, the number of events is far too small to support any inference, and our data therefore neither establish nor exclude a link between amplitude decline and PIRA. Baseline analyses provided additional context. Prolonged latency was associated with greater spinal cord involvement, whereas reduced amplitude was associated with older age and higher relapse rates; however, neither baseline abnormality was associated with EDSS measures at baseline or during follow-up. In addition, patients with latency prolongation exhibited lower lesion burden and lower CSF IgG index at baseline.
The two principal VEP metrics, P100 latency and amplitude, are generally considered to reflect different pathophysiological aspects of MS. A prolonged P100 latency indicates slowed axonal conduction caused by demyelination along visual pathways [ 8 , 11 ], whereas a reduced P100 amplitude signifies axonal loss or persistent conduction block, reflecting irreversible damage [ 9 , 23 ]. For example, in acute optic neuritis, both latency delay and amplitude attenuation are commonly observed; latency may partially recover with remyelination, but amplitude loss often persists in cases of axonal degeneration [ 11 , 23 ]. These complementary parameters provide a unique window for subclinical demyelination and neurodegeneration. Notably, patients in the progressive phase of MS often exhibit more pronounced abnormalities in both measures, which is consistent with cumulative axonal loss [ 7 , 10 ]. Our findings are compatible with this dual-pathway interpretation, although we emphasize that the present study did not measure inflammation or neurodegeneration directly with independent biomarkers (e.g., gadolinium enhancement during follow-up, serum neurofilament light chain, or OCT), and the mechanistic attribution therefore remains an interpretation of the observed associations rather than a demonstrated conclusion. Latency prolongation was associated with concurrent relapse activity, but not with disability worsening, which is consistent with its known sensitivity to demyelination [ 8 , 11 ]. This association was robust to the non-independence of the two eyes in a GEE model and to adjustment for DMT tier, and yielded moderate discriminative accuracy for concurrent on-study relapse activity, suggesting that serial latency monitoring could serve as a practical correlate of inflammatory activity, pending prospective validation. The weaker, group-level association between amplitude reduction and disability would be in keeping with the concept that axonal loss underlies irreversible functional decline in MS [ 13 , 23 ], but the small effect size, the lack of significance after correction for multiple comparisons, and the attenuation after adjustment for baseline EDSS mean that our data provide only preliminary support for this interpretation. The observation that latency prolongation emerged in patients with fewer spinal cord lesions and a lower IgG index at baseline may suggest that focal optic pathway demyelination may evolve somewhat independently of global disease burden [ 4 , 24 ], although this finding also requires replication. Whether amplitude decline captures ‘silent progression’ or PIRA [ 21 , 25 ] could not be adequately tested here: with a formal PIRA definition, only three events were identified over a median of 2 years, and larger cohorts with longer follow-up will be needed to address this question.
These results are in line with those of previous reports that emphasized the prognostic value of VEP abnormalities in predicting future clinical outcomes. Miletic-Drakulic et al. [ 14 ] demonstrated that preserved VEP responses were predictive of achieving NEDA-3 over 10 years, whereas significant unilateral delays were linked to less favorable disease trajectories. Similarly, Vecchio et al. [ 3 ] reported that prolonged latency in non-neuritic eyes at diagnosis predicted subsequent disability and lesion accrual. Although baseline VEP abnormalities were not associated with disability worsening in our study, longitudinal changes, particularly amplitude decline, showed a nominal association with concurrent worsening of the EDSS score. These observations are consistent with, but do not prove, an added value of serial monitoring over single-point assessment; the exploratory nature of the correlation findings underscores the need for prospective validation.
Importantly, several classical studies, such as those by Fuhr et al. [ 15 ] and Kallmann et al. [ 16 ], have shown that early evoked potential abnormalities across multiple modalities can predict long-term outcomes. However, their designs typically rely solely on baseline testing. Our findings raise the possibility that monitoring dynamic changes in VEP parameters over time may offer a more nuanced assessment of the disease course, although the present design cannot establish predictive value. This approach is also supported by more recent work highlighting that VEP metrics correlate with MRI parameters such as lesion volume and brain atrophy, and may outperform standard clinical measures in detecting diffuse or subclinical pathology [ 6 , 13 ]. The reaffirmed diagnostic role of VEP within the 2024 revised McDonald criteria further underscores the value of standardized, serial optic pathway assessment in contemporary MS care [ 12 ].
Treatment is an important potential confounder in any observational study of inflammatory activity. In our cohort, patients receiving high-efficacy DMT had more active disease at baseline, in keeping with treatment escalation, yet experienced fewer relapses and less frequent latency prolongation during follow-up, and their left eye latency tended to shorten rather than lengthen. These observations are consistent with reports that effective immunotherapy is accompanied by stabilization or partial recovery of VEP latency [ 20 ]. Importantly, the latency-relapse association was present within both DMT strata and persisted after adjustment for DMT tier, arguing against confounding by treatment intensity as the sole explanation for this finding. The amplitude–EDSS association, by contrast, was confined to the moderate-efficacy stratum, and high-efficacy DMT was associated with lower odds of amplitude reduction in an exploratory model. Given the small size of the high-efficacy subgroup, these stratified results should be regarded as hypothesis-generating; nevertheless, they suggest that treatment intensity may influence amplitude evolution and should be accounted for in future prospective studies.
This study has several limitations that should be considered when interpreting its findings. First, the retrospective design and modest sample size constrain generalizability. Post hoc power analysis indicated that the study was adequately powered for the primary logistic regression analysis (power > 99% for OR = 2.91, α = 0.05); however, the correlation analyses were underpowered (power ≈ 56% for ρ = −0.231 at α = 0.05, ≈26% after Bonferroni). Accordingly, the amplitude-related correlation findings should be regarded as hypothesis-generating; prospective studies enrolling at least 145–236 participants will be required to confirm these associations with adequate statistical confidence. Second, the follow-up period (median 2 years) was relatively short for capturing long-term disability trajectories, and the observed between-group EDSS differences (0.5 points) fall below the commonly accepted minimal clinically important difference of approximately 1.0 point, warranting cautious clinical interpretation. Relatedly, because relapses, EDSS change, and VEP change were all ascertained over the same interval, our results describe concurrent associations and cannot establish that VEP changes predict subsequent relapses or disability. Formally defined PIRA events were too few to test the relationship between amplitude decline and relapse-independent progression, and the retrospective design did not allow confirmation of disability worsening at a fixed 3- or 6-month interval. Furthermore, inflammation and neurodegeneration were not measured directly with independent biomarkers, so the pathophysiological interpretation of latency and amplitude changes remains inferential. Third, the amplitude reduction threshold used in this study (a decrease of >0.3 µV from baseline or a value below 5 µV) was operationally defined rather than derived from a validated normative dataset, and its clinical significance relative to established ISCEV criteria requires further evaluation. Fourth, optical coherence tomography (OCT) data were unavailable, precluding structure–function correlation analyses that could have strengthened the interpretation of amplitude changes as a marker of axonal loss [ 26 , 27 ]. In addition, although sensitivity analyses stratified by and adjusted for DMT efficacy tier yielded consistent results for the latency-relapse association, treatment allocation was not randomized, the high-efficacy subgroup was small, and residual confounding by indication (patients with more active disease receiving more potent therapy), by individual agents within each tier, and by treatment duration cannot be excluded; these analyses were not pre-specified and should be considered exploratory. Fifth, although patients with recent optic neuritis were excluded, subclinical episodes occurring between assessments could not be completely ruled out. Sixth, abnormality thresholds were derived from published reference values rather than locally validated norms, which may introduce measurement variability across sites; likewise, although MRI was acquired with a uniform field strength and protocol, images were rated locally rather than by a central reader. Additionally, the use of a relatively large checkerboard check size may have reduced protocol sensitivity compared with the smaller check sizes recommended in current ISCEV guidelines, potentially underestimating the prevalence of subclinical demyelination. Finally, the independent analysis of each eye may introduce intra-subject correlation effects that we addressed in a sensitivity analysis using a generalized estimating equation model, which confirmed the latency-relapse association; nonetheless, larger cohorts are needed to extend such eye-level models to disability outcomes. Prospective multimodal studies with larger cohorts, extended follow-up, locally established reference values, and pre-specified correction for multiple comparisons are warranted to confirm and extend these findings.
5. Conclusions
In this exploratory multicenter retrospective cohort, longitudinal P100 latency prolongation was independently associated with concurrent relapse activity, and this association was robust to adjustment for disease-modifying therapy, whereas amplitude decline showed only a weak group-level association with disability that did not survive correction for multiple comparisons and was attenuated after adjustment for baseline EDSS. These findings are compatible with, but do not demonstrate, distinct inflammatory and neurodegenerative substrates for the two VEP parameters, and the small number of formally defined PIRA events precluded any conclusion regarding relapse-independent progression. Serial VEP assessment may merit further evaluation as a complementary tool within multimodal MS monitoring, but prospective studies with larger cohorts, longer follow-up, and independent biomarkers of inflammation and neurodegeneration are required before longitudinal VEP changes can inform individualized management decisions.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
Supplementary Materials
The following supporting information can be downloaded from https://www.mdpi.com/article/10.3390/medicina62091750/s1 : Supplementary Table S1 (Baseline VEP abnormalities and their clinical associations), Supplementary Table S2 (Distribution of disease-modifying therapies used between VEP assessments ( n = 83)), and Supplementary Table S3 (Clinical characteristics, relapse activity, and longitudinal VEP outcomes according to disease-modifying therapy efficacy tier ( n = 78; five untreated patients excluded)).
Author Contributions
Conceptualization, A.T. and N.K.İ.; methodology, A.T., M.S. and S.Ö. (Samet Öncel); formal analysis, A.T. and S.Ö. (Samet Öncel); investigation, S.Ö. (Samet Öncel), S.Ö. (Sami Ömerhoca), S.A., E.Ç. and H.K.; data curation, S.A., E.Ç. and H.K.; writing—original draft preparation, S.Ö. (Samet Öncel) and A.T.; writing—review and editing, A.T., M.S. and N.K.İ.; supervision, A.T. and N.K.İ. All authors have read and agreed to the published version of the manuscript.
Institutional Review Board Statement
The study was conducted in accordance with the Declaration of Helsinki. Following institutional permission from each participating hospital, a single ethics approval covering all three centers (Sakarya Training and Research Hospital, Bursa Yüksek İhtisas Training and Research Hospital, and Bağcılar Training and Research Hospital) was granted by the Ethics Committee of Sakarya University Faculty of Medicine (approval number 150, dated 19 November 2024).
Informed Consent Statement
Patient consent was waived due to the retrospective design of the study; the waiver was approved by the Ethics Committee of Sakarya University Faculty of Medicine (Approval No. 150, 19 November 2024). All patient data were anonymized prior to analysis.
Data Availability Statement
The data presented in this study are available on request from the corresponding author. The anonymized minimal dataset supporting the reported findings can also be provided to the editorial office as Supporting Information for internal evaluation upon request.
Conflicts of Interest
Author Abdulkadir Tunç was employed by the company Bloocell Health Technologies Industry and Trade Joint Stock Company. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
References
1. Hojjati S.M. Zarghami A. Hojjati S.A. Baes M. Optic neuritis, the most common initial presenting manifestation of multiple sclerosis in northern Iran Casp. J. Intern. Med. 2015 6 151 155 PMC4650790 26644882
2. Miller D. Barkhof F. Montalban X. Thompson A. Filippi M. Clinically isolated syndromes suggestive of multiple sclerosis, part I: Natural history, pathogenesis, diagnosis, and prognosis Lancet Neurol. 2005 4 281 288 10.1016/S1474-4422(05)70071-5 15847841
3. Vecchio D. Barbero P. Galli G. Virgilio E. Naldi P. Comi C. Cantello R. Prognostic Role of Visual Evoked Potentials in Non-Neuritic Eyes at Multiple Sclerosis Diagnosis J. Clin. Med. 2023 12 2382 10.3390/jcm12062382 36983382 PMC10058458
4. Martinez-Lapiscina E.H. Arnow S. A Wilson J. Saidha S. Preiningerova J.L. Oberwahrenbrock T. Brandt A.U. E Pablo L. Guerrieri S. Gonzalez I. Retinal thickness measured with optical coherence tomography and risk of disability worsening in multiple sclerosis: A cohort study Lancet Neurol. 2016 15 574 584 10.1016/S1474-4422(16)00068-5 27011339
5. Klistorner A. Graham E.C. Yiannikas C. Barnett M. Parratt J. Garrick R. Wang C. You Y. Graham S.L. Progression of retinal ganglion cell loss in multiple sclerosis is associated with new lesions in the optic radiations Eur. J. Neurol. 2017 24 1392 1398 10.1111/ene.13404 28799222
6. Oertel F.C. Krämer J. Motamedi S. Keihani A. Zimmermann H.G. Dimitriou N.G. Condor-Montes S. Bereuter C. Cordano C. Abdelhak A. Visually Evoked Potential as Prognostic Biomarker for Neuroaxonal Damage in Multiple Sclerosis From a Multicenter Longitudinal Cohort Neurol. Neuroimmunol. Neuroinflamm. 2023 10 e200092 10.1212/NXI.0000000000200092 36878713 PMC10026703
7. Guerrieri S. Comi G. Leocani L. Optical coherence tomography and visual evoked potentials as prognostic and monitoring tools in progressive multiple sclerosis Front. Neurosci. 2021 15 692599 10.3389/fnins.2021.692599 34421520 PMC8374170
8. Odom J.V. Bach M. Brigell M. Holder G.E. McCulloch D.L. Mizota A. Tormene A.P. International Society for Clinical Electrophysiology of Vision ISCEV standard for clinical visual evoked potentials: (2016 update) Doc. Ophthalmol. 2016 133 1 9 10.1007/s10633-016-9553-y 27443562
9. Leocani L. Guerrieri S. Comi G. Visual evoked potentials as a biomarker in multiple sclerosis and associated optic neuritis J. Neuroophthalmol. 2018 38 350 357 10.1097/WNO.0000000000000704 30106802
10. Leocani L. Rovaris M. Boneschi F.M. Medaglini S. Rossi P. Martinelli V. Amadio S. Comi G. Multimodal evoked potentials to assess the evolution of multiple sclerosis: A longitudinal study J. Neurol. Neurosurg. Psychiatry 2006 77 1030 1035 10.1136/jnnp.2005.086280 16735397 PMC2077734
11. Klistorner A. Arvind H. Nguyen T. Garrick R. Paine M. Graham S. O’DAy J. Grigg J. Billson F. Yiannikas C. Axonal loss and myelin in early ON loss in postacute optic neuritis Ann. Neurol. 2008 64 325 331 10.1002/ana.21474 18825673
12. Saidha S. Green A.J. Leocani L. Vidal-Jordana A. Kenney R.C. Bsteh G. Outteryck O. Thompson A. Montalban X. Coetzee T. The use of optical coherence tomography and visual evoked potentials in the 2024 McDonald diagnostic criteria for multiple sclerosis Lancet Neurol. 2025 24 880 892 Erratum in Lancet Neurol. 2025 , 24 , e13. https://doi.org/10.1016/S1474-4422(25)00349-7 10.1016/S1474-4422(25)00275-3 40975103
13. Covey T.J. Golan D. Zarif M. Bumstead B. Buhse M. Kaczmarek O. Sergott R. Wilken J. Sima D.M. Van Hecke W. Individual differences in visual evoked potential latency are associated with variance in brain tissue volume in people with multiple sclerosis: An analysis of brain function-structure correlates Mult. Scler. Relat. Disord. 2022 68 104116 10.1016/j.msard.2022.104116 36041331
14. Miletic-Drakulic S. Miloradovic I. Jankovic V. Azanjac-Arsic A. Lazarevic S. VEP Score of a Left Eye Had Predictive Values for Achieving NEDA-3 over Ten Years in Patients with Multiple Sclerosis Sensors 2022 22 8849 10.3390/s22228849 36433445 PMC9696926
15. Fuhr P. Borggrefe-Chappuis A. Schindler C. Kappos L. Visual and motor evoked potentials in the course of multiple sclerosis Brain 2001 124 2162 2168 10.1093/brain/124.11.2162 11673318
16. Kallmann B.A. Fackelmann S. Toyka K.V. Rieckmann P. Reiners K. Early abnormalities of evoked potentials and future disability in patients with multiple sclerosis Mult. Scler. 2006 12 58 65 10.1191/135248506ms1244oa 16459720
17. Thompson A.J. Banwell B.L. Barkhof F. Carroll W.M. Coetzee T. Comi G. Correale J. Fazekas F. Filippi M. Freedman M.S. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria Lancet Neurol. 2018 17 162 173 10.1016/S1474-4422(17)30470-2 29275977
18. A Samjoo I. Worthington E. Drudge C. Zhao M. Cameron C. A Häring D. Stoneman D. Klotz L. Adlard N. Efficacy classification of modern therapies in multiple sclerosis J. Comp. Eff. Res. 2021 10 495 507 10.2217/cer-2020-0267 33620251
19. Montalban X. Gold R. Thompson A.J. Otero-Romero S. Amato M.P. Chandraratna D. Clanet M. Comi G. Derfuss T. Fazekas F. ECTRIMS/EAN guideline on the pharmacological treatment of people with multiple sclerosis Mult. Scler. 2018 24 96 120 10.1177/1352458517751049 29353550
20. Saridas F. Hojjati F. Alizada S. Lazrak S. Koc E.R. Turan O.F. Effects of disease-modifying therapies on remyelination in multiple sclerosis; evaluation via visual evoked potential test Mult. Scler. Relat. Disord. 2024 91 105850 10.1016/j.msard.2024.105850 39232395
21. Kappos L. Wolinsky J.S. Giovannoni G. Arnold D.L. Wang Q. Bernasconi C. Model F. Koendgen H. Manfrini M. Belachew S. Contribution of relapse-independent progression vs relapse-associated worsening to overall confirmed disability accumulation in typical relapsing multiple sclerosis in a pooled analysis of 2 randomized clinical trials JAMA Neurol. 2020 77 1132 1140 10.1001/jamaneurol.2020.1568 32511687 PMC7281382
22. Lublin F.D. A Häring D. Ganjgahi H. Ocampo A. Hatami F. Čuklina J. Aarden P. Dahlke F. Arnold D.L. Wiendl H. How patients with multiple sclerosis acquire disability Brain 2022 145 3147 3161 10.1093/brain/awac016 35104840 PMC9536294
23. You Y. Klistorner A. Thie J. Gupta V.K. Graham S.L. Axonal loss in a rat model of optic neuritis is closely correlated with visual evoked potential amplitudes using electroencephalogram-based scaling Investig. Ophthalmol. Vis. Sci. 2012 53 3662 10.1167/iovs.12-9843 22723598
24. Alshowaeir D. Yiannikas C. Klistorner A. Multifocal visual evoked potential (mfVEP) and pattern-reversal visual evoked potential changes in patients with visual pathway disorders: A case series Neuroophthalmology 2015 39 220 233 10.3109/01658107.2015.1074253 27928359 PMC5122941
25. Giovannoni G. Butzkueven H. Dhib-Jalbut S. Hobart J. Kobelt G. Pepper G. Sormani M.P. Thalheim C. Traboulsee A. Vollmer T. Brain health: Time matters in multiple sclerosis Mult. Scler. Relat. Disord. 2016 9 S5 S48 10.1016/j.msard.2016.07.003 27640924
26. Tavazzi E. Jakimovski D. Kuhle J. Hagemeier J. Ozel O. Ramanathan M. Barro C. Bergsland N. Tomic D. Kropshofer H. Serum neurofilament light chain and optical coherence tomography measures in MS: A longitudinal study Neurol. Neuroimmunol. Neuroinflamm. 2020 7 e737 10.1212/NXI.0000000000000737 32424064 PMC7251512
27. Petzold A. Balcer L.J. Calabresi P.A. Costello F. Frohman T.C. Frohman E.M. Martinez-Lapiscina E.H. Green A.J. Kardon R. Outteryck O. Retinal layer segmentation in multiple sclerosis: A systematic review and meta-analysis Lancet Neurol. 2017 16 797 812 10.1016/S1474-4422(17)30278-8 28920886