Key Points
Question
What proportion of clinically defined relapses is not associated with radiological activity in patients with relapsing-remitting multiple sclerosis?
Findings
Among 637 clinical events identified as relapses in this cohort study, approximately one-fourth were not associated with new T2 lesions or gadolinium-enhanced T1 lesions on brain and spinal cord magnetic resonance imaging (MRI) and were termed acute clinical events with stable MRI (ACES). ACES were more likely in patients with longer disease duration, receiving highly effective disease-modifying therapies, and presenting with fatigue.
Meaning
The findings highlight the importance of comprehensive brain and spinal cord MRI for classifying clinical events, which may impact therapeutic decisions and randomized clinical trial design considerations.
This cohort study evaluates clinical events identified as relapses not associated with T2 lesions or gadolinium-enhanced T1 lesions on magnetic resonance imaging.
Abstract
Importance
Understanding the association between clinically defined relapses and radiological activity in multiple sclerosis (MS) is essential for patient treatment and therapeutic development.
Objective
To investigate clinical events identified as relapses but not associated with new T2 lesions or gadolinium-enhanced T1 lesions on brain and spinal cord magnetic resonance imaging (MRI).
Design, Setting, and Participants
This multicenter observational cohort study was conducted between January 2015 and June 2023. Data were extracted on June 8, 2023, from the French MS registry. All clinical events reported as relapses in patients with relapsing-remitting MS were included if brain and spinal cord MRI was performed within 12 and 24 months before the event, respectively, and 50 days thereafter with gadolinium injection.
Exposures
Events were classified as relapses with active MRI (RAM) if a new T2 lesion or gadolinium-enhanced T1 lesion appeared on brain or spinal cord MRI or as acute clinical events with stable MRI (ACES) otherwise.
Main Outcomes and Measures
Factors associated with ACES were investigated; patients with ACES and RAM were compared regarding Expanded Disability Status Scale (EDSS) course, relapse rate, confirmed disability accrual (CDA), relapse-associated worsening (RAW), progression independent of relapse activity (PIRA), and transition to secondary progressive (SP) MS, and ACES and RAM rates under each disease-modifying therapy (DMT) were estimated.
Results
Among 31 885 clinical events, 637 in 608 patients (493 [77.4%] female; mean [SD] age, 35.8 [10.7] years) were included. ACES accounted for 166 (26.1%) events and were more likely in patients receiving highly effective DMTs, those with longer disease duration (odds ratio [OR], 1.04; 95% CI, 1.01-1.07), or those presenting with fatigue (OR, 2.14; 95% CI, 1.15-3.96). ACES were associated with significant EDSS score increases, lower than those found for RAM. Before the index event, patients with ACES experienced significantly higher rates of relapse (relative rate [RR], 1.21; 95% CI, 1.01-1.46), CDA (hazard ratio [HR], 1.54; 95% CI, 1.13-2.11), and RAW (HR, 1.72; 95% CI, 1.20-2.45). Patients with ACES were at significantly greater risk of SP transition (HR, 2.58; 95% CI, 1.02-6.51). Although RAM rate decreased with DMTs according to their expected efficacy, ACES rate was stable across DMTs.
Conclusions and Relevance
The findings in this study introduce the concept of ACES in MS, which accounted for one-fourth of clinical events identified as relapses.
Introduction
Relapses are a major clinical end point in patients with multiple sclerosis (MS), with a purely clinical definition: a monophasic clinical episode with patient-reported symptoms and objective findings typical of MS, developing acutely or subacutely, lasting more than 24 hours, and occurring in the absence of fever or infection.1 They should be distinguished from pseudorelapses, defined as a worsening of neurological symptoms corresponding to those of a past relapse, related to an alteration in the neuronal conduction in a context of infection or hyperthermia.2 Although relapses are theoretically expected to be caused by an acute immune attack of the central nervous system associated with the appearance of new T2 lesions and gadolinium-enhanced T1 lesions on magnetic resonance imaging (MRI), the link between clinical relapses and radiological activity has been scarcely studied to date, and only regarding gadolinium-enhanced T1 lesions.3,4
A clinical-radiological paradox emerges from randomized clinical trials: highly effective disease-modifying therapies (DMTs) have a greater impact on radiological activity than on clinical relapse rate. For instance, the Ocrelizumab versus Interferon Beta-1a in Relapsing Multiple Sclerosis (OPERA)5 and Ofatumumab versus Teriflunomide in Multiple Sclerosis (ASCLEPIOS)6 studies demonstrated that ocrelizumab and ofatumumab halved the clinically adjudicated relapse rate, whereas the rate of new T2 lesions was divided by 5, and the rate of gadolinium-enhanced T1 lesions by 20. Moreover, both randomized clinical trials5,6,7,8 and observational studies9,10 have revealed the existence of a floor effect, a minimum threshold of annualized relapse rate (around 0.10 to 0.20 relapses per year) that cannot be overcome even with highly effective DMTs. In real-life practice, particularly with highly effective DMTs, patients may experience clinical events that fulfill the relapse definition but without radiological activity on brain and spinal cord MRI. This apparent paradox might be explained by a discrepancy between the clinical definition of a relapse and the true phenomenon underlying it, which may not always result in the appearance of new T2 lesions or gadolinium-enhanced T1 lesions. In a large observational cohort of patients with MS, we aimed to determine the proportion of acute clinical events diagnosed as relapses that were not associated with radiological activity on brain and spinal cord MRI, and to compare the characteristics and clinical course of patients according to their MRI activity after the clinical event.
Methods
Legal
Patients registered in the Observatoire Français de la Sclérose en Plaques (OFSEP) provided written informed consent for the use of their data in research projects. OFSEP was approved by the French data protection agency (Commission Nationale de l’Informatique et des Libertés) and Comité de Protection des Personnes. The study was declared compliant with Méthodologie de reference 004 of the French data protection agency. The present report follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Patients
Data were extracted on June 8, 2023, from the OFSEP11 database, which collects data from patients with MS in 41 French centers. For each patient, clinical and imaging data are collected retrospectively at the first visit, then prospectively during routine follow-up visits, using a dedicated software (the European Database on Multiple Sclerosis).12
Inclusion criteria were (1) a clinical event reported as a relapse, occurring between January 1, 2015, and June 8, 2023; (2) in patients with a diagnosis of relapsing-remitting MS according to current criteria, either Poser or McDonald1,13,14,15; (3) with a reference MRI of the brain within 12 months and of the whole spinal cord within 24 months before the event, and (4) a brain and spinal cord MRI with gadolinium injection performed within 50 days after. Events were excluded if reported as isolated optic neuritis or if 2 or more clinical events occurred between preevent and postevent MRI.
Data Extraction
For each index event, data were extracted regarding patient demographic characteristics (sex, age, and date of MS onset), event characteristics (date, symptoms, and corticosteroids use), brain and spinal cord MRI (date, T2 lesion load, new or enlarged T2 lesion, gadolinium-enhanced T1 lesion, and reported worsening compared to the previous one), and DMT. For each patient with 1 or more included events, data were extracted for their entire follow-up regarding relapse occurrence, neurological disability measured by the Expanded Disability Status Scale (EDSS),16 transition to secondary progressive (SP) disease according to Lublin definition,17 and DMTs prescribed.
Event Classification
All clinical events meeting the definition of a relapse according to the treating neurologist were considered: patient-reported symptoms and objective findings typical of MS, lasting more than 24 hours, and occurring without fever or infection. Index events were classified as relapses with active MRI (RAM) if the postevent brain or spinal cord MRI indicated worsening, a new or enlarged T2 lesion, or a gadolinium-enhanced T1 lesion compared to the preevent MRI; otherwise they were classified as acute clinical events with stable MRI (ACES). Clinical events and MRIs were not centrally reviewed.
EDSS Score, Disability Accrual Events, and DMT Classification
EDSS score before the index event was the last measured in the preceding 12 months, EDSS score during the index event was the first measured within the following 3 months, and EDSS score after the index event was the first measured between 3 and 12 months thereafter. Confirmed disability accumulation (CDA) was defined as an EDSS worsening (an increase ≥1.5 for a reference EDSS score of 0, ≥1 for a reference EDSS score from 1 to 5.0, or ≥0.5 for a reference EDSS score ≥5.5), confirmed by a further measurement made more than 6 months later, with the CDA date set at that of the first EDSS worsening. The reference EDSS score was defined using a roving approach: its value was set at the first EDSS measured and could change afterwards, either to a lower EDSS score confirmed by another measurement made more than 3 months later or to a higher EDSS score after a CDA event.18 CDAs were classified as relapse-associated worsening (RAW) if 1 or more clinical relapses had been reported in the interval between the CDA event and the reference EDSS measurement or 30 days after the CDA event, and as progression independent of relapse activity (PIRA) otherwise.19,20 DMTs were classified as moderately effective (interferon, glatiramer, teriflunomide, and dimethyl fumarate [DMF]) or highly effective (anti-CD20 [rituximab, ocrelizumab, ofatumumab], natalizumab, alemtuzumab, and fingolimod).
Statistical Analyses
Continuous variables were described by means with SDs or medians with IQRs for time variables, and categorical variables by counts and percentages. The association of variables with the probability of classifying the index event as ACES vs RAM was analyzed using a univariate logistic regression based on a generalized estimating equation approach, then a multivariable model was constructed with the covariates significantly associated with the probability of ACES. Longitudinal models for EDSS score and relapse rate were built using mixed-effects models—a linear and a Poisson model, respectively—adjusted for ACES or RAM classification of the index event, disease duration, covariates associated with ACES, and with a patient random effect. Occurrence of CDA, PIRA, and RAW was studied using Andersen-Gill survival models, adapted to recurrent events, adjusted for covariates associated with ACES, and restricted to patients with 2 or more EDSS measurements during the considered period. Time to SP transition was analyzed using a Cox model, with baseline set at the date of the index event, and adjusted for disease duration. Estimates of the annualized rate of RAM and ACES for each DMT were calculated by multiplying the observed annualized relapse rate by the observed proportion of RAM or ACES among index events under that DMT, and inference was conducted using bootstrap over 10 000 resamples, clustered by patient and stratified by DMT.
EDSS score, relapses, CDA/PIRA/RAW, and time to SP transition models were restricted to the dataset retaining one index event per patient. Two-tailed P values below .05 were considered statistically significant. Analyses were performed using R version 4.0.3 (R Foundation), with lmerTest and geepack packages. More details on the model construction are provided in section 1 in Supplement 1. Sensitivity analyses were conducted to evaluate the impact of choosing a 50-day cutoff for the time interval between the index event and postevent MRI; analyses were replicated in cohorts using cutoffs of less than 30 days and less than 90 days.
Results
A total of 637 clinical events were included, occurring in 608 patients (493 [77.4%] female; mean [SD] age, 35.8 [10.7] years) (eFigure in Supplement 1) and after a median (IQR) disease duration of 2.7 (0.9-8.4) years. Postevent brain and spinal cord MRI was reported to be stable for 166 (26.1%) clinical events, consequently classified as ACES. Among the 471 events classified as RAM, MRI identified a new or enlarged T2 lesion and a gadolinium-enhanced T1 lesion in 233 (49.5%) cases, a new or enlarged T2 lesion without gadolinium-enhanced T1 lesion in 140 (29.7%), a gadolinium-enhanced T1 lesion without new or enlarged T2 lesion in 78 (16.6%), and worsening with no further details in 20 (4.2%). Of 244 index events for which information was available (when brain and spinal cord MRIs were not performed simultaneously), radiological activity was localized to the brain in 86 (35.2%) cases, to the spinal cord in 71 (29.1%), and to both the brain and spinal cord in 87 (35.7%). Compared to excluded clinical events, those that were included occurred in younger patients, with shorter disease duration, different symptoms, and more frequently with spinal cord involvement (eTable 1 in Supplement 1).
Characteristics Associated With ACES
In univariate analyses, classification of the index event as ACES was significantly more likely in patients with longer disease duration (odds ratio [OR] for each additional year, 1.06; 95% CI, 1.04-1.09), presenting with motor impairment (OR, 1.64; 95% CI, 1.14-2.35) or fatigue (OR, 2.70; 95% CI, 1.65-4.40), with higher EDSS score before the index event (OR for each additional point, 1.27; 95% CI, 1.12-1.43), or receiving highly effective DMT (OR compared to untreated patients, 7.32; 95% CI, 4.52-11.87) (Table 1). ACES were significantly less likely in patients presenting with sensory impairment (OR, 0.63; 95% CI, 0.42-0.95). Regarding DMTs, ACES were significantly more likely in patients receiving teriflunomide, fingolimod, natalizumab, and anti-CD20; the effect size increased with the expected DMT efficacy, reaching an inverted ratio of 72.7% of ACES for anti-CD20.
Table 1. Association of Demographic, Clinical, Magnetic Resonance Imaging (MRI), and Disease-Modifying Therapy (DMT) Characteristics With Classification of the Index Event as Acute Clinical Events With Stable MRI (ACES).
| Variable | No. (%)a | Odds ratio (95% CI)b | P value | |
|---|---|---|---|---|
| RAM | ACES | |||
| Total | 471 (73.9) | 166 (26.1) | NA | NA |
| Sex | ||||
| Male | 105 (72.9) | 39 (27.1) | 1 [Reference] | .78 |
| Female | 366 (74.2) | 127 (25.8) | 0.94 (0.62-1.43) | |
| Age at MS onset, y | ||||
| Median (IQR) odds ratio for each additional year | 28.0 (24.0-36.0) | 29.0 (23.0-36.8) | 1.00 (0.99-1.02) | .65 |
| Disease duration at the time of the index event, y | ||||
| Median (IQR) odds ratio for each additional year | 2.0 (0.8-6.8) | 4.9 (1.7-12.1) | 1.06 (1.04-1.09) | <.001 |
| Calendar year of the index event | ||||
| Median (IQR) odds ratio for each additional year | 2018 (2016-2020) | 2018 (2016-2020) | 0.98 (0.91-1.07) | .69 |
| Season of the index event | ||||
| Winter | 112 (70.4) | 47 (29.6) | 1 [Reference] | NA |
| Spring | 117 (77.0) | 35 (23.0) | 0.72 (0.44-1.18) | .20 |
| Summer | 133 (76.4) | 41 (23.6) | 0.72 (0.44-1.18) | .19 |
| Fall | 109 (71.7) | 43 (28.3) | 0.95 (0.59-1.54) | .83 |
| Follow-up duration after the index event, y | ||||
| Median (IQR) odds ratio for each additional year | 3.2 (1.5-5.1) | 3.4 (1.6-4.8) | 0.97 (0.90-1.05) | .48 |
| Symptoms (possibly >1) | ||||
| Walking difficulties | ||||
| No | 357 (75.6) | 115 (24.4) | 1 [Reference] | NA |
| Yes | 114 (69.1) | 51 (30.9) | 1.42 (0.96-2.09) | .08 |
| Motor impairment | ||||
| No | 325 (77.2) | 96 (22.8) | 1 [Reference] | NA |
| Yes | 146 (67.6) | 70 (32.4) | 1.64 (1.14-2.35) | .01 |
| Sensory impairment | ||||
| No | 93 (66.4) | 47 (33.6) | 1 [Reference] | NA |
| Yes | 378 (76.1) | 119 (23.9) | 0.63 (0.42-0.95) | .03 |
| Bladder and bowel dysfunction | ||||
| No | 413 (73.4) | 150 (26.6) | 1 [Reference] | NA |
| Yes | 58 (78.4) | 16 (21.6) | 0.76 (0.42-1.37) | .36 |
| Oculomotor impairment | ||||
| No | 452 (73.6) | 162 (26.4) | 1 [Reference] | NA |
| Yes | 19 (82.6) | 4 (17.4) | 0.58 (0.20-1.74) | .33 |
| Vestibulocochlear symptoms | ||||
| No | 453 (73.5) | 163 (26.5) | 1 [Reference] | NA |
| Yes | 18 (85.7) | 3 (14.3) | 0.47 (0.14-1.57) | .22 |
| Facial motor impairment | ||||
| No | 463 (74.2) | 161 (25.8) | 1 [Reference] | NA |
| Yes | 8 (61.5) | 5 (38.5) | 1.82 (0.60-5.5) | .29 |
| Facial sensory impairment | ||||
| No | 443 (73.5) | 160 (26.5) | 1 [Reference] | NA |
| Yes | 28 (82.4) | 6 (17.6) | 0.61 (0.26-1.45) | .26 |
| Speech impairment | ||||
| No | 464 (73.9) | 164 (26.1) | 1 [Reference] | NA |
| Yes | 7 (77.8) | 2 (22.2) | 0.80 (0.17-3.90) | .78 |
| Optic neuritis | ||||
| No | 456 (74.4) | 157 (25.6) | 1 [Reference] | NA |
| Yes | 15 (62.5) | 9 (37.5) | 1.75 (0.76-4.03) | .19 |
| Fatigue | ||||
| No | 429 (76.5) | 132 (23.5) | 1 [Reference] | NA |
| Yes | 42 (55.3) | 34 (44.7) | 2.70 (1.65-4.40) | <.001 |
| EDSS score before the index event (<12 mo before) | ||||
| Mean (SD) odds ratio for each additional point | 1.5 (1.0-2.1) | 2.0 (1.0-3.5) | 1.27 (1.12-1.43) | <.001 |
| Missing | 147 | 34 | NA | NA |
| EDSS score during the index event (0 to 3 mo after) | ||||
| Mean (SD) odds ratio for each additional point | 2.0 (1.5-3.0) | 2.5 (2.0-4.0) | 1.16 (1.03-1.31) | .02 |
| Missing | 129 | 52 | NA | NA |
| EDSS score after the index event (3 to 12 mo after) | ||||
| Mean (SD) odds ratio for each additional point | 2.0 (1.0-3.0) | 2.5 (1.5-4.0) | 1.19 (1.07-1.33) | <.001 |
| Missing | 112 | 31 | NA | NA |
| EDSS score worsening after the index event | ||||
| No | 196 (70.3) | 83 (29.7) | 1 [Reference] | NA |
| Yes | 72 (71.3) | 29 (28.7) | 0.98 (0.58-1.63) | .93 |
| Missing | 203 | 54 | NA | NA |
| Time interval between preevent brain MRI and index event, mo | ||||
| Median (IQR) odds ratio for each additional mo | 5.1 (2.7-7.8) | 5.4 (3.2-8.1) | 1.02 (0.97-1.08) | .40 |
| Time interval between preevent spinal cord MRI and index event, mo | ||||
| Median (IQR) odds ratio for each additional mo | 5.6 (3.1-9.3) | 5.7 (3.4-9.9) | 1.00 (0.97-1.04) | .93 |
| Time interval between index event and postevent brain and spinal cord MRI, mo | ||||
| Median (IQR) odds ratio for each additional mo | 0.7 (0.3-1.2) | 0.8 (0.4-1.2) | 1.24 (0.85-1.80) | .26 |
| Brain MRI T2 lesion load | ||||
| None | 8 (80.0) | 2 (20.0) | 1 [Reference] | NA |
| 1-4 | 48 (78.7) | 13 (21.3) | 1.08 (0.20-5.73) | .92 |
| 5-8 | 27 (71.1) | 11 (28.9) | 1.66 (0.30-9.12) | .56 |
| ≥9 | 335 (73.5) | 121 (26.5) | 1.46 (0.31-6.98) | .63 |
| Missing | 53 | 19 | 1.40 (0.27-7.22) | .69 |
| Spinal cord MRI T2 lesion load | ||||
| None | 39 (66.1) | 20 (33.9) | 1 [Reference] | NA |
| 1 | 64 (69.6) | 28 (30.4) | 0.84 (0.42-1.68) | .62 |
| ≥2 | 364 (75.7) | 117 (24.3) | 0.63 (0.36-1.12) | .12 |
| Missing | 4 | 1 | 0.49 (0.05-4.63) | .53 |
| DMT level received at the time of the index event | ||||
| No DMT | 231 (84.0) | 44 (16.0) | 1 [Reference] | NA |
| Moderately effectivec | 180 (78.3) | 50 (21.7) | 1.46 (0.93-2.30) | .10 |
| Highly effectived | 51 (42.1) | 70 (57.9) | 7.32 (4.52-11.87) | <.001 |
| Other DMTe | 9 (81.8) | 2 (18.2) | 1.21 (0.27-5.45) | .80 |
| DMT received at the time of the index event | ||||
| None | 231 (84.0) | 44 (16.0) | 1 [Reference] | NA |
| Interferon | 33 (76.7) | 10 (23.3) | 1.64 (0.75-3.56) | .21 |
| Glatiramer | 58 (82.9) | 12 (17.1) | 1.10 (0.54-2.20) | .80 |
| DMF | 45 (83.3) | 9 (16.7) | 1.05 (0.48-2.28) | .91 |
| Teriflunomide | 44 (69.8) | 19 (30.2) | 2.24 (1.20-4.19) | .01 |
| Fingolimod | 26 (44.1) | 33 (55.9) | 6.61 (3.61-12.11) | <.001 |
| Natalizumab | 14 (43.8) | 18 (56.2) | 6.58 (3.03-14.31) | <.001 |
| Anti-CD20 | 6 (27.3) | 16 (72.7) | 18.94 (6.47-55.48) | <.001 |
| Otherf | 14 (73.7) | 5 (26.3) | 1.91 (0.67-5.43) | .22 |
| Received corticosteroids after the index event | ||||
| No | 136 (78.2) | 38 (21.8) | 1 [Reference] | NA |
| Yes | 335 (72.4) | 128 (27.6) | 1.35 (0.89-2.05) | .16 |
Abbreviations: DMF, dimethyl fumarate; EDSS, Expanded Disability Status Scale; MS, multiple sclerosis; NA, not applicable; RAM, relapses with active MRI.
Percentages are calculated per row to show the raw probabilities of ACES for each level of categorical variables.
From univariate generalized estimating equations logistic regressions.
DMF, glatiramer acetate, interferon, and teriflunomide.
Alemtuzumab, anti-CD20 (rituximab, ocrelizumab, ofatumumab), fingolimod, and natalizumab.
Azathioprine (n = 4), cladribine (n = 2), methotrexate (n = 1), mycophenolate mofetil (n = 2), and other (n = 2).
Alemtuzumab (n = 8), azathioprine (n = 4), cladribine (n = 2), methotrexate (n = 1), mycophenolate mofetil (n = 2), and other (n = 2).
In the multivariable analysis, which included 456 index events with no missing data, ACES remained significantly associated with disease duration (OR for each additional year, 1.04; 95% CI, 1.01-1.07), fatigue (OR, 2.14; 95% CI, 1.15-3.96), and DMTs, while associations with EDSS score before the index event, motor impairment, and sensory impairment became nonsignificant (Table 2).
Table 2. Multivariable Analysis of the Probability of Classification of the Index Event as an Acute Clinical Event With Stable Magnetic Resonance Imaging (ACES).
| Odds ratio (95% CI)a | P value | |
|---|---|---|
| Disease duration at the time of the index event | ||
| For each additional year | 1.04 (1.01-1.07) | .02 |
| EDSS score before the index event | ||
| For each additional point | 1.05 (0.89-1.23) | .57 |
| Motor impairment | ||
| Yes vs no | 1.35 (0.81-2.23) | .25 |
| Sensory impairment | ||
| Yes vs no | 0.67 (0.41-1.10) | .12 |
| Fatigue | ||
| Yes vs no | 2.14 (1.15-3.96) | .02 |
| DMT received at the time of the index event | ||
| None | 1 [Reference] | NA |
| Interferon | 2.66 (1.11-6.37) | .03 |
| Glatiramer | 1.38 (0.61-3.12) | .44 |
| DMF | 1.40 (0.50-3.94) | .52 |
| Teriflunomide | 2.53 (1.14-5.61) | .02 |
| Fingolimod | 5.93 (2.91-12.05) | <.001 |
| Natalizumab | 7.73 (3.13-19.07) | <.001 |
| Anti-CD20 | 13.96 (4.79-40.69) | <.001 |
| Otherb | 1.46 (0.43-4.98) | .54 |
Abbreviations: DMF, dimethyl fumarate; DMT, disease-modifying therapy; EDSS, Expanded Disability Status Scale; NA, not applicable.
From a multivariable generalized estimating equations logistic regression.
Alemtuzumab (n = 8), azathioprine (n = 4), cladribine (n = 2), methotrexate (n = 1), mycophenolate mofetil (n = 2), and other (n = 2).
Disease Course Comparison Between Patients With ACES vs Patients With RAM
Retaining only 1 index event per patient, 608 patients were included in further analyses. Results of each model are provided in eTables 2-6 in Supplement 1.
EDSS Score Change After the Index Event
EDSS score significantly increased after both RAM and ACES (Figure 1A); the mean changes in EDSS score after ACES vs RAM were 0.45 (95% CI, 0.29 to 0.61) and 0.72 (95% CI, 0.62 to 0.81) in the period from 0 to 3 months, 0.15 (95% CI, 0.00 to 0.30) and 0.40 (95% CI, 0.31 to 0.50) in the period from 3 to 12 months, and 0.16 (95% CI, 0.02 to 0.31) and 0.36 (95% CI, 0.26 to 0.46) in the period of more than 12 months, respectively, with a significant difference in the EDSS score change between ACES and RAM in the 3 periods.
Figure 1. Comparison of Patients With Acute Clinical Events With Stable MRI (ACES) and Relapses With Active MRI (RAM).
The mean change in Expanded Disability Status Scale (EDSS) score in each time period relative to the index event is predicted from the EDSS mixed-effects linear model for a mean patient. Relapse and confirmed disability accrual (CDA), relapse-associated worsening (RAW), and progression independent of relapse activity (PIRA) rates are unadjusted observed values. Cumulative incidence curves of transition to secondary progressive (SP) disease were obtained using the Kaplan-Meier method. Whiskers represent 95% CIs. MRI indicates magnetic resonance imaging.
Relapse Rate
The relapse rate before the index event was significantly higher in patients with ACES (relative rate [RR], 1.21; 95% CI, 1.01-1.46). However, after the index event, there was no significant difference in relapse rate between patients with ACES and those with RAM (RR, 0.91; 95% CI, 0.70-1.18) (Figure 1B).
CDA, PIRA, and RAW
Before the index event, among the 356 patients with 2 or more EDSS measurements, 162 CDA events occurred, 124 of which (76.5%) were RAW and 38 of which (23.5%) were PIRA. The CDA hazard was significantly higher in patients with ACES (hazard ratio; HR, 1.54; 95% CI, 1.13-2.11), driven by a significantly higher RAW hazard (HR, 1.72; 95% CI, 1.20-2.45) and no significant difference in PIRA (HR, 1.09; 95% CI, 0.55-2.14) (Figure 1C). After the index event, 200 CDA events occurred in 539 patients with 2 or more EDSS measurements, 130 of which (65.0%) were RAW and 70 of which (35.0%) were PIRA; there was no significant difference in CDA (HR, 1.05; 95% CI, 0.75-1.46), RAW (HR, 0.87; 95% CI, 0.56-1.34), and PIRA (HR, 1.43; 95% CI, 0.84-2.40) hazards.
SP Transition
Transition to a SP disease occurred in 9 of 153 patients (5.9%) in the ACES group and 10 of 455 (2.2%) in the RAM group. There was a significantly higher hazard of SP transition in patients with ACES (HR 2.58; 95% CI, 1.02-6.51) (Figure 1D).
Estimation of ACES and RAM Rates According to DMT
The estimated mean annualized rate of ACES was 0.14 events per person-year and of RAM was 0.41 events per person-year. As expected, the RAM rate was significantly higher than the overall mean in untreated patients (RR, 1.42; 95% CI, 1.32-1.53) or those receiving glatiramer acetate (RR, 1.41; 95% CI, 1.14-1.73) and significantly lower in those receiving highly effective DMTs (ie, fingolimod [RR, 0.47; 95% CI, 0.32-0.63], natalizumab [RR, 0.34; 95% CI, 0.20-0.51], and anti-CD20 [RR, 0.07; 95% CI, 0.00-0.16]) (Figure 2). The ACES rate did not significantly differ from the overall mean for most DMTs, with significant differences only in patients receiving fingolimod (RR, 1.67; 95% CI, 1.26-2.15).
Figure 2. Estimated Annualized Rate of Relapses With Active MRI (RAM) and Acute Clinical Events With Stable MRI (ACES) Under Each Disease-Modifying Therapy (DMT).
Horizontal dashed lines indicate overall mean rate. Whiskers indicate 95% CIs. MRI indicates magnetic resonance imaging.
Sensitivity Analyses
In the cohort with a postevent MRI performed within 30 days, 404 events were included, 104 of which (25.7%) were ACES. Analyses in this cohort were consistent with the main analyses, with slight differences related to the loss of statistical power (eTable 7 in Supplement 1). In the cohort with a postevent MRI performed within 90 days, ACES accounted for 303 (29.5%) of the 1027 included events. Events were significantly more likely to be classified as ACES when the postevent MRI was delayed and in patients with higher EDSS scores or lower spinal cord T2 lesion loads, whereas the association with disease duration was no more significant. Conclusions were identical for the analyses of EDSS score change; relapse rate; CDA, RAW, and PIRA; and ACES or RAM rate under each DMT, but the risk of SP transition was not significantly higher in patients with ACES.
Discussion
In this cohort study, one-fourth of clinical events reported as relapses in patients with MS were not associated with radiological activity on brain and spinal cord MRI. These clinical events, which we called ACES, were significantly more likely in patients with longer disease duration, presenting with fatigue, or receiving highly effective DMTs. EDSS scores were significantly increased after ACES, but this increase was significantly lower than that found for RAM. Compared to those with RAM, patients with ACES had significantly higher rates of relapse, CDA, and RAW before the index event, and no significant differences were found thereafter, or regarding PIRA before and after the index event. Patients with ACES were at significantly higher risk of transition to SP disease compared to patients with RAM. The estimated rate of RAM was lower in patients receiving DMTs, in accordance with their expected efficacy, while the ACES rate remained relatively stable across DMTs.
To our knowledge, only 2 studies3,4 have explored the association between clinical relapses and radiological activity, but these focused solely on gadolinium-enhanced T1 lesions. Approximately 40% of patients were found to have no gadolinium-enhanced T1 lesions after a clinical relapse, and 1 of the studies also found a significantly lower probability of gadolinium-enhanced T1 lesions in patients with longer disease duration. Both studies had small population sizes, and neither investigated T2 lesions.
We propose several hypotheses that could explain ACES, classified as active or inactive. In active hypotheses, the clinical event might be related to an inflammatory attack of the central nervous system and therefore qualified as relapse, but without new T2 lesions nor gadolinium-enhanced T1 lesions. This could be explained by false-negative MRI results, particularly when the comparison between MRI results is more complicated (eg, in patients with numerous and confluent T2 lesions) or when MRI is done at distance from the relapse. Other forms of inflammatory activity could be involved: cortical lesions21,22 (which require high-field MRI23 and adapted sequences24,25 to be detected), acute worsening of chronic active lesions,26 or diffuse inflammatory processes.27 The significant disability increase after ACES and the shorter time to SP transition support the idea of an active inflammation in these events, which might be qualified as subacute progression phenomena. In inactive hypotheses, symptoms in ACES might be unrelated to MS inflammatory activity but rather to pseudorelapses with less obvious triggers (asymptomatic viral infection, warm weather, or other unrecognized causes), to a subjective worsening of MS symptoms linked to another condition (eg, fatigue and depression), to another neurological condition mimicking a relapse,28 or might be an early sign of neurodegenerative processes in patients with exhausted compensatory mechanisms. Future studies focusing on MRI (potentially using automated algorithms for MRI comparison) and biological biomarkers (such as serum neurofilament light chain and serum glial fibrillary acidic protein) could provide an interesting insight into the underlying causes of ACES.
The implications of our findings are important for the design of randomized clinical trials and real-life practice. In randomized clinical trials, it is crucial to better determine the target of the therapies tested: if DMTs primarily aim to reduce the rate of relapses with radiological activity (which is probably the case, given the stable ACES rate among DMTs found herein), it becomes imperative to validate each clinical relapse with a comprehensive brain and spinal cord MRI; this could also enhance the statistical power of future trials. In light of our findings, the negative results of the recent trial29 of evobrutinib vs teriflunomide may be explained, if most clinical events were actually ACES in both groups, by a lack of statistical power rather than a lack of efficacy on RAM. In real-life practice, it may be necessary to investigate radiological activity with a brain and spinal cord MRI when a patient presents with a clinical event to classify it as RAM, which would require a DMT change, or ACES, for which the therapeutic decision is less clear.
Limitations
The present study has several limitations. First, MRI was performed in routine practice and was not centrally reviewed; however, OFSEP proposed an MRI follow-up protocol in 2015 to harmonize practices nationally (provided in section 2 in Supplement 1), enabling more robust comparisons between different MRI results.30 Only clinical events that were identified as relapses by the neurologists were reported in the OFSEP database; thus, clinical events observed in practices that may have differed among experts and centers would be missing. Additionally, the estimated annualized rates of ACES and RAM for each DMT were calculated in a selected population without correction for indication bias; the results should therefore not be interpreted as reliable measures of DMT effectiveness.
Conclusions
In conclusion, ACES represented one-fourth of clinical events classified as relapses in patients with MS. As their impact on clinical practice and randomized clinical trials might be substantial, it is essential to identify this subgroup of clinical events, as has been done recently with the concept of PIRA, and further research is required to explore their underlying causes.
eTable 1. Comparison of included and excluded clinical events
eTable 2. Details of the longitudinal EDSS score mixed-effects model
eTable 3. Details of the longitudinal relapse mixed-effects model
eTable 4. Details of the CDA/RAW/PIRA Poisson models before and after the index event
eTable 5. Details of the time to SP transition Cox model
eTable 6. Estimation of ACES/RAM annualized rate by DMT and relative risk comparatively to overall mean rate
eTable 7. Results of the sensitivity analysis with different delay thresholds for the post-event brain and spinal cord MRI
eFigure. Flowchart
The Observatoire Français de la Sclérose en Plaques (OFSEP) investigators
Data sharing statement
References
- 1.Thompson AJ, Banwell BL, Barkhof F, et al. Diagnosis of multiple sclerosis: 2017 revisions of the McDonald criteria. Lancet Neurol. 2018;17(2):162-173. doi: 10.1016/S1474-4422(17)30470-2 [DOI] [PubMed] [Google Scholar]
- 2.Davis SL, Wilson TE, White AT, Frohman EM. Thermoregulation in multiple sclerosis. J Appl Physiol (1985). 2010;109(5):1531-1537. doi: 10.1152/japplphysiol.00460.2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Martín-Aguilar L, Presas-Rodriguez S, Rovira À, et al. Gadolinium-enhanced brain lesions in multiple sclerosis relapse. Neurologia (Engl Ed). 2022;37(7):557-563. doi: 10.1016/j.nrl.2021.10.005 [DOI] [PubMed] [Google Scholar]
- 4.Dünschede J, Ruschil C, Bender B, et al. Clinical-radiological mismatch in multiple sclerosis patients during acute relapse: discrepancy between clinical symptoms and active, topographically fitting MRI lesions. J Clin Med. 2023;12(3):739. doi: 10.3390/jcm12030739 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hauser SL, Bar-Or A, Comi G, et al. ; OPERA I and OPERA II Clinical Investigators . Ocrelizumab versus interferon beta-1a in relapsing multiple sclerosis. N Engl J Med. 2017;376(3):221-234. doi: 10.1056/NEJMoa1601277 [DOI] [PubMed] [Google Scholar]
- 6.Hauser SL, Bar-Or A, Cohen JA, et al. ; ASCLEPIOS I and ASCLEPIOS II Trial Groups . Ofatumumab versus teriflunomide in multiple sclerosis. N Engl J Med. 2020;383(6):546-557. doi: 10.1056/NEJMoa1917246 [DOI] [PubMed] [Google Scholar]
- 7.Polman CH, O’Connor PW, Havrdova E, et al. ; AFFIRM Investigators . A randomized, placebo-controlled trial of natalizumab for relapsing multiple sclerosis. N Engl J Med. 2006;354(9):899-910. doi: 10.1056/NEJMoa044397 [DOI] [PubMed] [Google Scholar]
- 8.Calabresi PA, Radue EW, Goodin D, et al. Safety and efficacy of fingolimod in patients with relapsing-remitting multiple sclerosis (FREEDOMS II): a double-blind, randomised, placebo-controlled, phase 3 trial. Lancet Neurol. 2014;13(6):545-556. doi: 10.1016/S1474-4422(14)70049-3 [DOI] [PubMed] [Google Scholar]
- 9.Rollot F, Couturier J, Casey R, et al. Comparative effectiveness of natalizumab versus anti-CD20 in highly active relapsing-remitting multiple sclerosis after fingolimod withdrawal. Neurotherapeutics. 2022;19(2):476-490. doi: 10.1007/s13311-022-01202-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Roos I, Hughes S, McDonnell G, et al. ; MSBase Study GroupDanish MS Registry Study Group . Rituximab vs ocrelizumab in relapsing-remitting multiple sclerosis. JAMA Neurol. 2023;80(8):789-797. doi: 10.1001/jamaneurol.2023.1625 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Vukusic S, Casey R, Rollot F, et al. Observatoire Français de la Sclérose en Plaques (OFSEP): a unique multimodal nationwide MS registry in France. Mult Scler. 2020;26(1):118-122. doi: 10.1177/1352458518815602 [DOI] [PubMed] [Google Scholar]
- 12.Confavreux C, Compston DA, Hommes OR, McDonald WI, Thompson AJ. EDMUS, a European database for multiple sclerosis. J Neurol Neurosurg Psychiatry. 1992;55(8):671-676. doi: 10.1136/jnnp.55.8.671 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Poser CM, Paty DW, Scheinberg L, et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol. 1983;13(3):227-231. doi: 10.1002/ana.410130302 [DOI] [PubMed] [Google Scholar]
- 14.McDonald WI, Compston A, Edan G, et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the Diagnosis of Multiple Sclerosis. Ann Neurol. 2001;50(1):121-127. doi: 10.1002/ana.1032 [DOI] [PubMed] [Google Scholar]
- 15.Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. Ann Neurol. 2011;69(2):292-302. doi: 10.1002/ana.22366 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology. 1983;33(11):1444-1452. doi: 10.1212/WNL.33.11.1444 [DOI] [PubMed] [Google Scholar]
- 17.Lublin FD, Reingold SC, Cohen JA, et al. Defining the clinical course of multiple sclerosis: the 2013 revisions. Neurology. 2014;83(3):278-286. doi: 10.1212/WNL.0000000000000560 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Kappos L, Butzkueven H, Wiendl H, et al. ; Tysabri® Observational Program (TOP) Investigators . Greater sensitivity to multiple sclerosis disability worsening and progression events using a roving versus a fixed reference value in a prospective cohort study. Mult Scler. 2018;24(7):963-973. doi: 10.1177/1352458517709619 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kappos L, Wolinsky JS, Giovannoni G, et al. 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(9):1132-1140. doi: 10.1001/jamaneurol.2020.1568 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Müller J, Cagol A, Lorscheider J, et al. Harmonizing definitions for progression independent of relapse activity in multiple sclerosis: a systematic review. JAMA Neurol. 2023;80(11):1232-1245. doi: 10.1001/jamaneurol.2023.3331 [DOI] [PubMed] [Google Scholar]
- 21.Calabrese M, Rocca MA, Atzori M, et al. A 3-year magnetic resonance imaging study of cortical lesions in relapse-onset multiple sclerosis. Ann Neurol. 2010;67(3):376-383. doi: 10.1002/ana.21906 [DOI] [PubMed] [Google Scholar]
- 22.Puthenparampil M, Poggiali D, Causin F, et al. Cortical relapses in multiple sclerosis. Mult Scler. 2016;22(9):1184-1191. doi: 10.1177/1352458514564483 [DOI] [PubMed] [Google Scholar]
- 23.Madsen MAJ, Wiggermann V, Marques MFM, et al. Linking lesions in sensorimotor cortex to contralateral hand function in multiple sclerosis: a 7T MRI study. Brain. 2022;145(10):3522-3535. doi: 10.1093/brain/awac203 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Nelson F, Poonawalla AH, Hou P, Huang F, Wolinsky JS, Narayana PA. Improved identification of intracortical lesions in multiple sclerosis with phase-sensitive inversion recovery in combination with fast double inversion recovery MR imaging. AJNR Am J Neuroradiol. 2007;28(9):1645-1649. doi: 10.3174/ajnr.A0645 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Simon B, Schmidt S, Lukas C, et al. Improved in vivo detection of cortical lesions in multiple sclerosis using double inversion recovery MR imaging at 3 Tesla. Eur Radiol. 2010;20(7):1675-1683. doi: 10.1007/s00330-009-1705-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Calvi A, Haider L, Prados F, Tur C, Chard D, Barkhof F. In vivo imaging of chronic active lesions in multiple sclerosis. Mult Scler. 2022;28(5):683-690. doi: 10.1177/1352458520958589 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Giovannoni G, Popescu V, Wuerfel J, et al. Smouldering multiple sclerosis: the ‘real MS’. Ther Adv Neurol Disord. 2022;15:17562864211066751. doi: 10.1177/17562864211066751 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Tallantyre EC, Causon EG, Harding KE, Pickersgill TP, Robertson NP. The aetiology of acute neurological decline in multiple sclerosis: experience from an open-access clinic. Mult Scler. 2015;21(1):67-75. doi: 10.1177/1352458514538333 [DOI] [PubMed] [Google Scholar]
- 29.Burger L. Merck KGaA suffers major blow as MS drug fails in late-stage trials. Reuters . Published December 6, 2023. Accessed May 31, 2024. https://www.reuters.com/business/healthcare-pharmaceuticals/merck-kgaa-says-ms-drug-fails-late-stage-trials-2023-12-05/
- 30.Cotton F, Kremer S, Hannoun S, Vukusic S, Dousset V; Imaging Working Group of the Observatoire Français de la Sclérose en Plaques . OFSEP, a nationwide cohort of people with multiple sclerosis: Consensus minimal MRI protocol. J Neuroradiol. 2015;42(3):133-140. doi: 10.1016/j.neurad.2014.12.001 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. Comparison of included and excluded clinical events
eTable 2. Details of the longitudinal EDSS score mixed-effects model
eTable 3. Details of the longitudinal relapse mixed-effects model
eTable 4. Details of the CDA/RAW/PIRA Poisson models before and after the index event
eTable 5. Details of the time to SP transition Cox model
eTable 6. Estimation of ACES/RAM annualized rate by DMT and relative risk comparatively to overall mean rate
eTable 7. Results of the sensitivity analysis with different delay thresholds for the post-event brain and spinal cord MRI
eFigure. Flowchart
The Observatoire Français de la Sclérose en Plaques (OFSEP) investigators
Data sharing statement
