Article Text
PDF
Abstract
Background Progression independent of relapse activity (PIRA) has been described since the early stage of relapsing multiple sclerosis (RMS). However, little is known about the relation between PIRA and inflammatory activity that is particularly important at this stage of the disease.
Method We included 110 patients in a prospective study within 18 months of RMS onset. MRI examinations and clinical visits were scheduled on the same day for months 0, 6, 12, 24, 36, 60, 84, 120, 180 and 240.
Results The mean (SD) age of patients was 30 (6.7) years at inclusion and median (range) follow-up 15 (9–20) years. Analysis of 1118 between-visit intervals revealed 93 confirmed disability accumulation events in 68 (62%) patients: 50 (54%) events related to relapse activity worsening and 43 (46%) PIRA events, including 17 (18%) with MRI activity. The risk of PIRA between two visits (stable event as the reference category) was associated with Expanded Disability Status Scale (EDSS) score (HR: 1.41; 95% CI: 1.18 to 1.69; p<0.001), disease duration (HR: 0.75; 95% CI: 0.62 to 0.90; p<0.005) and new lesions between the visits (HR: 1.09 per lesion; 95% CI: 1.01 to 1.17; p<0.05). As compared with PIRA events with MRI activity, PIRA events without such activity occurred in patients with more disability (mean EDSS score 3, p<0.05), longer disease duration (mean 11 years, p<0.001) and greater number of T2-weighted lesions (p<0.05).
Conclusion This study evidenced that inflammatory activity increases the risk of PIRA in early RMS, arguing that a significant part of PIRA is accessible to treatment targeting inflammation in these patients.
- MULTIPLE SCLEROSIS
Statistics from Altmetric.com
Introduction
Disease progression independent of relapse activity (PIRA) was described in patients with relapsing multiple sclerosis (RMS) treated with high-efficacy therapy, which suggested that PIRA could occur independent of disease activity.1 Hence, PIRA was interpreted as the clinical counterpart of the neurodegenerative component of MS. However, in patients without full control of disease activity, lesion accumulation could also contribute to PIRA.2–5 In these patients, non-active PIRA (naPIRA), defining patients free of relapse but also free of MRI activity, could better reflect the neurodegenerative component of MS as compared with those with active PIRA (aPIRA). Nevertheless, to what extent aPIRA contributes to PIRA in early RMS remains largely unknown because, in previous studies, MRI explorations were missing or not performed at the same time as clinical visits.2 3
In the present study, we analysed data from a prospective MRI study of 129 patients with RMS enrolled after their first clinical event and followed up to 20 years (median 15 years). Importantly, all clinical visits and MRI examinations (>1100) were scheduled on the same day, which allowed for qualifying each confirmed disability accumulation (CDA) event as aPIRA, naPIRA or relapse-associated worsening (RAW).
Methods
Participants
From 2001 to 2011, 129 patients were included in a prospective MRI study based on the following criteria: (1) age 18–45 years; (2) occurrence of a first inflammatory demyelinating event in the central nervous system in the last 18 months; (3) presence of oligoclonal bands in cerebral spinal fluid and (4) presence of at least two demyelinating lesions in the brain or spinal cord detected at the initial MRI performed before inclusion. All participants gave informed consent to participate in the study before taking part.6 7
Study design
Brain MRI examinations and clinical visits were scheduled on the same day for months 0, 6, 12, 24, 36, 60, 84, 120, 180 and 240. Disability level was rated with the Expanded Disability Status Scale (EDSS).8 All relapses since the last visit were recorded.
Statistics
CDA was considered according to the used three strata definitions (increase in EDSS score by 1.5 points if the last EDSS score was 0, increase by 1 point if the EDSS score was 1–5.5 or increase by 0.5 points if the EDSS score was >5.5), confirmed after at least 6 months. The existence of a PIRA event was defined at each visit considering the presence of CDA during the interval and the absence of any relapse during the same interval. With new T2-weighted lesions and PIRA during the same interval, we retained an aPIRA event. If PIRA occurred in an interval without new lesions, we retained an naPIRA event. PIRA events had to be present in all longitudinal follow-up data. With the presence of CDA and relapse during the same interval, we retained a RAW event. If no CDA occurred in an interval, we retained a STABLE event.
We compared patient characteristics at the beginning of each intervisit interval associated with an episode of CDA (RAW, aPIRA and naPIRA), including age, disease duration, EDSS score and number of brain T2 lesions, by using linear mixed-effects models accounting for recurrent measurements in patients. To explore factors associated with PIRA between two consecutive visits (with STABLE events as the reference category), we computed Cox proportional-hazards regression models for recurrent events with patient ID used as a cluster variable and first clinical event or the next visit as the end time, estimating HRs and 95% CIs. The initial Cox proportional-hazards model included age, EDSS score, disease duration, total lesion number, type of treatment (no vs high/moderate/low efficacy therapy) at the beginning of the intervisit interval and number of new T2-weighted lesions in the intervisit interval and in the previous intervisit interval. The model was then reduced by backward selection of variables. We used the Akaike Information Criterion to select variables. All analyses were performed with Statistics in R V.4.2.1.
Data availability
All data will be shared anonymised by reasonable request of a qualified investigator to the corresponding author.
Results
Characteristics of the cohort
At the time of the analysis, 110 of 129 (85%) patients were still being followed (88 women, mean (SD) age 30 (6.7) years at the first MRI/clinical visit); 14 patients dropped out after the first MRI and 5 before 5 years. At the first MRI/clinical visit, the mean (SD) time since the first MS symptoms was 0.48 (0.35) years. At the last MRI/clinical visit, the median (range) follow-up of patients was 15 (9–20) years, corresponding to 1569 patient-years. The number of combined MRI/clinical visits was 1120, corresponding to a median (range) of 10 (4–19) visits per patient.
During the same period of inclusion (2001–2011), 571 patients visited the MS centre of Marseille for a first demyelinating MS event. These patients represented 87% of the estimated number of patients with a first demyelinating MS event during this period (n=657) according to the incidence of MS in the Provence region.9 The 110 patients followed and the 571 who visited the centre did not differ in demographic characteristics (sex ratio and age) (p=0.28 and p=0.11).
CDA during follow-up
Analysis of the 1118 between-visit intervals revealed 93 CDA events in 68 of 110 (62%) patients: 50 (54%) RAW and 43 PIRA (46%) (table 1 and figure 1). PIRA events included 26 naPIRA (60%) and 17 aPIRA (40%) events. As compared with aPIRA events, naPIRA events occurred in patients with more disability (mean EDSS score 3, p<0.05), longer disease duration (mean 11 years, p<0.001) and greater number of T2-weighted lesions (mean 55, p<0.05) (tables 1 and 2). The final model included only EDSS score (HR: 1.41; 95% CI: 1.18 to 1.69; p<0.005), disease duration (HR: 0.75; 95% CI: 0.62 to 0.90; p<0.005) and number of new T2-weighted lesions during the intervisit interval (HR: 1.09 per lesion; 95% CI: 1.01 to 1.17; p<0.05) (table 3).
Patient characteristics by type of event
(A) Individual disease course for 110 patients ordered from bottom to top by their last documented Expanded Disability Status Scale (EDSS) score. The colour of the bar represents the EDSS at the beginning of each intervisit interval. Filled dots represent the timepoint of confirmed progression events classified as relapse-associated worsening (RAW), active progression independent of relapse activity (PIRA) or non-active PIRA. (B) Boxplots illustrating the difference in selected characteristics at the beginning of each intervisit interval with a confirmed progression event. Horizontal line is median, box edges are IQR and whiskers are range. DDS, Disease Durations.
Disease characteristics at the beginning of episodes with confirmed EDSS score progression labelled as RAW (the reference category), active PIRA or non-active PIRA
Predictors for a PIRA event
Discussion
Events related to disease activity (RAW and aPIRA) contribute to almost 75% of all CDA events occurring within the first 20 years of RMS. naPIRA contributed to the remaining CDA events and occurred more frequently after several years of disease evolution (mean 11 years) in patients with significant disability (mean EDSS score 3) and a high number of lesions.
In the present study, we selected patients younger than 45 years at inclusion to limit the risk of including patients not at the real first stage of MS and patients with progressive MS erroneously classified as RMS. Therefore, the association found between lesion accumulation and risk of concomitant PIRA could not be generalised to late-onset RMS.
The present study indicates that PIRA events in early RMS could not be exclusively related to diffuse inflammation and/or secondary degeneration and that the accumulation of asymptomatic demyelinating lesions in the brain was observed in 40% of all PIRA events. Crucially, the association found between lesion accumulation during the intervisit interval and risk of PIRA at the end of the intervisit interval suggests that the accumulation of demyelinating lesions independent from relapse may be involved in the pathogenesis of PIRA. The risk of PIRA at each visit was increased to a mean of 9% for each new single brain T2 lesion that occurred since the last visit. This association between MRI activity and subsequent disability worsening can partly explain the elevation of neurofilament light chain observed more than one year before disability worsening independent of clinical relapses.10 One may suppose that subclinical relapse activity could result in disability worsening as the central nervous system reserve is lost.5 Moreover, the contribution of aPIRA events was probably under-estimated here because of lack of available spinal-cord MRI images. Indeed, previous studies found that lesion accumulation in the spinal cord contributed highly to disability.11 12 Finally, we found that naPIRA events occurred generally after several years of disease in patients with a high number of lesions and more disability. All these findings suggest that many PIRA events occurring in RMS could be accessible to early treatment with high-efficacy therapy that fully controls disease activity and reduces early disability accumulation.13 14 In the present study, we did not find an association between treatment and risk of PIRA. However, the interpretation of this result is difficult because the decision to treat and the choice of therapy were based on the physician’s preference and highly depended on the level of disease activity. Thus, the design of the present study allowed for depicting an association between disease activity and risk of PIRA but not the potential effect of therapy on risk of PIRA.
The strengths of the study are its prospective design, the long-term follow-up, the high number of standardised clinical and MRI examinations performed concomitantly, the inclusion of patients right after RMS onset and the low number of patients lost to follow-up. Limitations are lack of spinal-cord MRI images, the relatively small sample size and the monocentric design. Moreover, the effect of EDSS levels on the risk of progression needs to be interpreted with caution because the non-continuous character of the EDSS scale might bias our analyses, in which EDSS score was used as a continuous variable.
Conclusions
APIRA and naPIRA may have different pathophysiologies, and full control of disease activity from disease onset may significantly reduce the risk of PIRA in RMS.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved. This study involves human participants and was approved by local ethics committee (Timone Hospital, Marseille and Comité de Protection des Personnes Sud-Méditerranée), number CCPPRB Marseille 2: 99-47 for patients explored with 1.5T MR system and CPP Sud Méditerranée 1: 08-45 for patients explored with 3T MR system. Participants gave informed consent to participate in the study before taking part.
Footnotes
Contributors AM played a major role in the acquisition of the data and the gathering of data. JPS played a major role in the acquisition of the data and conducted statistical analysis. AR, CB, SD and WZ played a major role in the acquisition of the data. J-PR and JP planned the study and played a major role in the acquisition of the data. BA planned and conducted the study, conducted statistical analysis and wrote the manuscript.
Funding This study was supported by Fondation pour l’Aide à la Recherche sur la Sclérose En Plaques (ARSEP Fondation)
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.