Association of Very Early Treatment Initiation With the Risk of Long-term Disability in Patients With a First Demyelinating Event

Abstract

Background and Objectives

Early treatment is associated with better long-term outcomes in patients with a first demyelinating event and early multiple sclerosis (MS). However, magnetic resonance (MR) findings are not usually integrated to construct propensity scores (PSs) when evaluating outcomes. We assessed the association of receiving very early treatment with the risk of long-term disability including an MR score (MRS) in patients with a first demyelinating event.

Methods

We included 580 patients with a first demyelinating event prospectively collected between 1994 and 2021, who received at least 1 disease-modifying drug (DMD). Patients were classified into tertiles according to the cohort's distribution of the time from the first demyelinating event to the first DMD: first tertile (FT) or very early treatment (6 months; n = 194), second tertile (6.1–16 months, n = 192), and third tertile (TT) (16.1 months, n = 194). A 5-point MRS was built according to the sum of the following indicators: ≥9 brain lesions (1 point); ≥1 infratentorial lesion (1 point); ≥1 spinal cord (SC) lesion (1 point); ≥1 contrast-enhancing (CE) brain lesion (1 point); and ≥1 CE SC lesion (1 point). PS based on covariates and the MRS was computed for each of the outcomes. Inverse PS-weighted Cox and linear regression models assessed the risk of different outcomes between tertile groups. Finally, to confirm the role of MR in treatment decision, we studied the time elapsed from the first demyelinating event to treatment initiation according to the MRS in all patients with radiologic available information, renamed as raw-MRS.

Results

Very early treatment decreased the risk of reaching Expanded Disability Status Scale 3.0 (hazard ratio [HR] 0.55, 95% CI 0.32–0.97), secondary progressive MS (HR 0.40, 95% CI 0.19–0.85), and sustained disease progression at 12 months after treatment initiation (HR 0.50, 95% CI 0.29–0.84), when compared with patients from the TT group. Patients from the FT group had a lower disability progression rate (β estimate −0.009, 95% CI −0.016 to −0.002) and a lower severe disability measured by the Patient-Determined Disease Step (β estimate −0.52, 95% CI −0.91 to −0.13) than the TT group. Finally, there was a 62.4% reduction in the median time between the first demyelinating event and the first-ever treatment initiation from patients displaying a raw-MRS 1 to patients with a raw-MRS 5.

Discussion

Using PS models with and without MRS, we showed that treatment initiation at very early stages is associated with a reduction in the risk of long-term disability accrual in patients with a first demyelinating event.

Classification of Evidence

This study provides Class III evidence that earlier treatment of patients with MS presenting with a first demyelinating event is associated with improved clinical outcomes.

Introduction

Classical clinical trials (CTs) and their extension phases initially proved the efficacy of multiple sclerosis disease-modifying drugs (MS-DMDs) to prevent conversion from the first demyelinating event to clinically definite MS.^1-11 However, real-world observational studies in early relapsing MS have been essential to provide evidence of the effectiveness of MS-DMD on long-term disability.^12-17 Some of these studies point to a greater magnitude of the treatment effect on disability possibly because of the shorter time elapsed from the MS diagnosis to treatment initiation.^12,18-23

Despite the effectiveness displayed by real-world studies on decreasing risk of disability, it is important to underline some methodological shortcomings influencing results.^12,18-24 In multicentric studies, data are usually collected following different protocols, which makes such information difficult to harmonize under a common database. For instance, differences in inclusion criteria, clinical visit frequency, scales to measure disability, or the use of predetermined variables among centers may add inconsistency when gathering information within the same platform. It is also worth pointing out that real-world studies evaluating the effect of time to treatment initiation on disability include prospective information both from MS diagnosis (with information about the first demyelinating event recorded retrospectively) and from the first event in a proportion of patients.^19,21,24-27 Nonetheless, studies assessing this outcome and focusing exclusively in a prospective deeply phenotyped cohort (prospective standardized collection of demographic, clinical, and radiologic information and blood samples) followed up from the first demyelinating event are needed.

Finally, baseline radiologic information is very useful to predict MS conversion and prognosis^23,28 and its key role in therapeutic decision-making and treatment monitoring. Indeed, in clinical routine, decisions made by clinicians in relation to when and how to treat patients with MS largely depends on an initial magnetic resonance (MR) scan. Nonetheless, to which extent the baseline radiologic information may guide clinicians for an early treatment initiation after the first demyelinating event is not clearly defined despite being highly important for patient counseling.

The primary research question aims to assess the association of receiving very early treatment on the risk of long-term disability accrual accounting for a baseline MR score (MRS). The second endpoint is to address whether the decision of when and how to treat patients with MS depends on the initial MR, by comparing the time to treatment initiation based on the MRS in a deeply phenotyped cohort of patients with a first demyelinating event.

Methods

Patients

This is an observational retrospective study based on the deeply phenotyped Barcelona cohort of patients with a first demyelinating event from the Multiple Sclerosis Center of Catalonia (Cemcat) for which information was prospectively collected between January 1994 and November 2021.^23,29 The cohort includes patients younger than 50 years who experienced a first demyelinating event suggestive of a first demyelinating event of the CNS, which could not be attributed to other diseases, and assessed at Cemcat within 3 months from the first demyelinating event.²³

Information collected included demographics: sex and date of birth; clinical data: date of the first demyelinating event at onset, baseline topography at first attack, presence and dates of relapses, and disability according to the Expanded Disability Status Scale (EDSS) during stability periods. EDSS scores were obtained within 3 months after the first demyelinating event, then (at least) annually, and during switching treatments. Relapse was defined as the occurrence of new symptoms persisting for at least 24 hours in the absence of fever and infection. Relapses and EDSS were prospectively assessed and recorded by appropriately trained neurologists, specialized in MS.

Laboratory data included analysis of immunoglobulin G oligoclonal bands (OCBs) in the CSF and serum on agarose gel isoelectric focusing combined with immunoblotting. Treatment information included date of treatment initiation and interruption and type of treatment. DMDs were classified as moderate or high-efficacy treatments (eTable 1, links.lww.com/WNL/C982). The exclusion criteria were (1) being younger than 16 years during the first demyelinating event, (2) alternative diagnosis during follow-up, and (3) patients initiating the first-ever treatment as a part of a CT.

Radiologic Information

Baseline brain MR scans were performed within the first 3–5 months after the first demyelinating event for all patients. Spinal cord MR scans were performed systematically for all patients at study baseline (i.e., within 3–5 months from the first attack), regardless of the topography at onset, after 2007.

Then, brain MR scans were performed at 12 months after the first demyelinating event, and every 5 years thereafter. Additional new brain MR scans could be performed if new symptoms appeared or before initiating a new treatment.

Standardized brain and spinal cord MR scans were performed according to recommendations³⁰: number and topography (supratentorial, infratentorial, spinal cord) of T2 lesions and the presence of contrast gadolinium-enhancing lesions at the first demyelinating event. Sequences of MR is available in the eMethods (links.lww.com/WNL/C982).

Baseline MRS

To account for baseline radiologic information when computing propensity scores (PSs), we designed a 5-point MRS as the sum of the following MRI features: (1) ≥9 supratentorial brain T2 lesions (1 point), (2) ≥1 infratentorial brain T2 lesions (1 point), (3) ≥1 spinal cord T2 lesions (1 point), (4) ≥1 brain contrast-enhancing lesion (CEL) (1 point), and (5) ≥1 spinal cord CEL (1 point).

For missing radiologic values, the mean value of the corresponding treated group (based on tertiles of time to treatment initiation [Statistical Analysis, Descriptive Analysis]) was imputed. The final MRS was rated by the total number of analyzed variables (the 5 radiologic variables) and multiplied 100-fold (MRS = [total radiologic variable punctuation/5] × 100).

Statistical Analysis

Descriptive Analysis

Treated patients were classified into groups of tertiles (treated tertile groups) according to the time elapsed from the date of the first demyelinating event to the date of the first-ever treatment initiation: First tertile (FT) initiating treatment within 6 months (n = 194), second tertile (ST) initiating treatment between 6.1 and 16 months (n = 192), and third tertile (TT) initiating treatment from 16.1 months onward. Patients from the FT were also named as very early treated patients. Such classification was performed to have a similar number of patients in each group.

Comparison of clinical and paraclinical characteristics across treated tertile groups was performed as follows: the χ² test for categorical variables and the Mann-Whitney U test for continuous variables because they did not follow the normal distribution.

Analysis of the 2 Aims of the Study

Association of Receiving Very Early Treatment on the Risk of Long-term Disability Accrual Accounting for a Baseline MRS

First, a binary multivariable logistic regression model was performed to estimate the patient's probability of being assigned to a given treated tertile group (dependent variable) considering the following covariates: age at treatment initiation, sex, topography at first attack, EDSS at treatment initiation, presence of CSF-restricted OCB, and calendar year at treatment initiation. The MRS was included in the model, which provided the patient risk of being allocated to a given treated tertile group, according to the radiologic information at the time of the first demyelinating event (eTable 2, links.lww.com/WNL/C982).

Next, a PS method based on the inverse probability weighting (IPW) estimator was used to estimate the probability of each treatment tertile.^31,32 IPW construction is detailed in eMethods (links.lww.com/WNL/C982).

Three final statistical analyses weighted by IPW were performed to estimate the strength of the association between treatment tertile groups (FT vs ST and FT vs TT groups) and different outcomes: (1) weighted Cox proportional hazard models (hazard ratio [HR], 95% CI) to evaluate time to event outcomes using an intention-to-treat strategy. Patients were censored when reaching the event (the specified outcome) or at the last follow-up; (2) a weighted linear regression model (β estimate, 95% CI) to evaluate continuous outcome variables; and (3) a weighted logistic regression model (odds ratio [OR] estimate, 95% CI) to evaluate categorical outcome variables.

Final models were adjusted by the proportion of time on high-efficacy therapy and unbalanced variables derived from pseudopopulations,³³ and collinearity was checked among covariates. Weighted and adjusted Kaplan-Meier (KM) curves were also built to depict the cumulative probabilities of reaching different outcomes according to treated tertile groups.

The evaluated outcomes were the following: (1) time to reach confirmed EDSS 3.0 and 6.0, (2) time to reach secondary progressive MS (SPMS), (3) time to reach sustained disability progression (SDP) at 12 months after the date of the first-ever treatment initiation, (4) EDSS progression rate from treatment initiation (continuous outcome variable), and (5) the Patient-Determined Disease Step (PDDS),³⁴ a patient-reported outcome measure (PROM) of disability (continuous outcome variable).

Other categorical outcome variables were evaluated such as the multiple sclerosis performance test (MSPT), which assess the processing speed test (PST), walking speed test (WST), contrast sensitivity test (CST), and manual dexterity test (MDT).³⁵

Finally, analyses were performed to evaluate whether very early treatment could have an impact on Neuro-quality-of-life (QoL) measures.^36,37 Outcome definitions are shown in eMethods.

Time to Treatment Initiation Based on the MRS

For this analysis, we included only a subgroup of treated patients in whom all radiologic variables were available. Differently to the previous MRS, information was not imputed because there were no missing data. Thus, this MRS is an ordinal scale ranging from 1 to 5. Such MRS was renamed as raw-MRS. The time elapsed from the date of the first demyelinating event to treatment initiation date was compared according to the raw-MRS.

A p value of 0.05 was considered statistically significant. All statistical analyses were performed with STATA-12 software (64-bit; StataCorp, College Station, TX).

Standard Protocol Approvals, Registrations, and Patient Consents

Databases have been developed according to national and international standards on ethical aspects (Declaration of Helsinki and Tokyo), and these data may be used in accordance with the regulations in force regarding the protection of personal data (EU) 2016/679; April 27, 2016 (GDPR). The study protocol was evaluated by the ethics committee of Vall d'Hebron Hospital, and all patients signed the corresponding informed consent form.

Data Availability

Data that support the findings of this study are available from the corresponding author on reasonable request.

Results

Description of the Cohort

Of 1,341 patients belonging to the deeply phenotyped cohort with a first demyelinating event, 1,268 patients met the inclusion criteria to participate in this study. Of them, 688 (54.2%) and 580 (45.8%) patients were nontreated and treated at any time over disease course, respectively. According to the time from the first demyelinating event to the first-ever treatment initiation, patients were classified into treatment tertile groups: FT or very early treatment, ST, and TT (eFigure 1, links.lww.com/WNL/C982). Type of the first-ever treatment initiation is listed in eTable 1.

Among the whole treated cohort, 401/580 (69.1%) were women. The mean (SD) age at the first-ever treatment initiation was 34.0 (8.3) years, and patients from the TT group initiated treatment at a slightly older age (35.3 [8.7]) than patients from the FT (32.1 [7.8]) and ST (33.6 [8.2]) groups, p = 0.054 (Table 1).

Table 1.

Characteristics of Patients With a First Demyelinating Event According to Treatment Tertile Groups

The proportion of patients with CSF-restricted OCB showed a continuous decrease along with longer time to treatment initiation (p = 0.016) (Table 1). Finally, patients from the FT group initiated treatment in more recent years (median [interquartile range, IQR]) (2013 [2008–2018]) than the ST (median 2011 [2007–2016.5]) and TT groups (2008 [2003–2015]), p < 0.001 (Table 1).

Different MRS values were found among tertile treated groups. Patients from the FT group displayed a higher median (IQR) MRS (67.6 [60–80]) than patients from the ST (60 [42.9–75.3]) and TT groups (45.8 [34.4–53.6]), p < 0.001 (Table 1). Values of imputed radiologic variables included to construct the MRS are listed in Table 1.

The first-ever administered treatment was considered moderate in 521 (89.8%) and of high efficacy in 59 (10.2%) patients. Type of the first-ever administrated treatment and that over the whole disease course is listed in Table 2. Finally, the mean proportion (SD) of time on high-efficacy treatment over the whole treated time was higher in the FT group (27.1% [40.0]) when compared with that in the ST (16.6% [32.2]) and TT (17.9% [30.4]) groups, p = 0.034 (Table 1).

Table 2.

Type of Treatment According to Tertile Treated Groups

Descriptive Cohort Disease Course

Disability trajectories are shown in eFigure 2, A and B (links.lww.com/WNL/C982).

After a median (IQR) follow-up of 11.2 (5.5–17.1) years, the median (range) of the last available EDSS of the whole treated cohort was 2.0 (0–9). At the last follow-up, the TT group had a higher median (range) EDSS (2.0 [0–9.0]) than the ST (1.5 [0–9.0]) and FT (1.5 [0–6.5]) groups, p < 0.001. In addition, the mean (SD) of the last PDDS in the FT group was 0.48 (1.14), in contrast to 0.71 (1.61) in the ST group and 1.25 (2.01) in the TT group, p = 0.012. The last follow-up was longer in patients with a delay in treatment initiation, p < 0.001 (Table 1).

Very Early Treatment Is Associated With a Decreasing Risk of Long-term Disability

As summarized in eTable 2 (links.lww.com/WNL/C982), probabilities of belonging to a specific tertile treated group according to covariates largely depend on the MRS and, therefore, MRS was also included as an additional variable to calculate the PS. Covariate description in the pseudo-population and SMD according to tertile treated groups are listed in eTables 3.1–3.6.

The weighted Cox regression analysis adjusted by those unbalanced variables derived from pseudopopulations and the proportion of time on high-efficacy treatments showed that initiating treatment in the FT/very early treatment group (within 6 months from the first demyelinating event) decreased the risk of reaching EDSS 3.0 by 45% (HR 0.55, 95% CI 0.32–0.97), SPMS by 60% (HR 0.40, 95% CI 0.19–0.85), and SDP at 12 months after treatment initiation by 50% (HR 0.50, 95% CI 0.29–0.84), when compared with that observed in patients from the TT group (Table 3). Figure 1, A–D displays KM survival curves after introduction of the PS and variable adjustment.

Table 3.

Hazard Ratios Obtained From Cox and Linear Regression–Weighted Models Evaluating Long-term Disability, According to Different Outcomes (Propensity Scores Are Incorporated)

Figure 1. Weighted and Adjusted Kaplan-Meier Curves to Reach Time-Dependent Outcomes Between First and Third Tertile Groups.

(A) Kaplan-Meier estimation of time to reach EDSS 3.0. At 5 years after the first demyelinating event, 16.8% (95% CI 11.8–23.7) within the first tertile and 32.5% (95% CI 25.6–40.6) within the third tertile, reached EDSS 3.0 (Log-rank, p value = 0.002). (B) Kaplan-Meier estimation of time to reach EDSS 6.0. At 5 years after the first demyelinating event, 2.02% (95% CI 0.7–6.2) within the first tertile and 4.3% (95% CI 2.7–8.8) within the third tertile reached EDSS 6.0 (log rank, p value = 0.021). (C) Kaplan-Meier estimation of time to reach SPMS. At 5 years after the first demyelinating event, 3.4% (95% CI 1.4–7.9) within the first tertile and 12.9% (95% CI 8.6–19.2) within the third tertile reached SPMS (log rank, p value = 0.002). (D) Kaplan-Meier estimation of time to reach SDP after the date of the first-ever treatment initiation, according to tertile treated groups. At 5 years after the first demyelinating event, 4.5% (95% CI 2.2–9.3) within the first tertile and 22.3% (95% CI 16.5–29.8) within the third tertile reached SDP after the date of the first-ever treatment initiation (log rank, p value < 0.001). *Kaplan-Meier curves were weighted by the propensity score and adjusted by the proportion of time on high-efficacy therapy and unbalanced variables derived from pseudopopulations (MRS in time to reach EDSS 3.0, time to reach EDSS 6.0, and time to reach SPMS; and gender and MRS in time to reach SDP at 12 months after treatment initiation). EDSS = Expanded Disability Status Scale; MRS = Magnetic Resonance Score; SDP = sustained disease progression; SPMS = secondary progressive multiple sclerosis.

In addition, patients from the FT group had a lower disability progression rate (β estimate −0.009, 95% CI −0.016 to −0.002) and a lower severe disability measured by the PDDS (β estimate −0.52, 95% CI −0.91 to −0.13) than the TT group (Table 3). The study did not find evidence of differences in any of the previous studied outcomes when comparing FT and ST groups (Table 3).

The weighted binary regression analysis did not find associations when comparing MSPT categorical variables (PST, WST, CST, and MDT) between groups (eTable 4, links.lww.com/WNL/C982).

Finally, the following Neuro-QoL measures were associated with the FT group in comparison with the TT group: upper extremity function (β estimate 3.17, 95% CI 0.77–6.65, p = 0.014), lower extremity function (β estimate 51.9, 95% CI 50.32–53.58, p < 0.001), sleep disturbance (β estimate −3.64, 95% CI −6.76 to −0.53, p = 0.022), fatigue (β estimate −4.50, 95% CI −8.06 to −0.90, p = 0.013), anxiety (β estimate −3.57, 95% CI −6.77 to −0.36, p = 0.029), and cognition function (β estimate 3.64, 95% CI 0.76 to 6.53, p = 0.014). eTable 5 (links.lww.com/WNL/C982) summarizes associations between Neuro-QoL measures and different tertile groups.

Information about the association between tertile groups and the main outcomes (EDSS 3.0, EDSS 6.0, SPMS, and SDP) without including the MRS in the model is summarized in eTable 6 (links.lww.com/WNL/C982). Although the Akaike information criterion (AIC) was minimized by including the MRS (data not shown), the strength of the association and 95% CI remained similar.

Time to Treatment Initiation Depends on Baseline Radiologic Information

Among the treated cohort, 205 patients had all baseline radiologic variables available. Comparison between treated patients with all available radiologic variables and patients with at least 1 missing radiologic variable is listed in Table 4, and comparison of baseline variables among different raw-MRS is listed in eTable 7 (links.lww.com/WNL/C982).

Table 4.

Demographic and Clinical Comparison of the Treated Cohort With a First Demyelinating Event, With and Without All Available Radiologic Variables

Restricting the analysis to the 205 patients with all MRI variables, the median (IQR) time from the first demyelinating event to treatment initiation showed a sustained decrease as the raw-MRS increased: 15.7 (11–33.4) months in raw-MRS 1, 7.9 (4.3–12.3) in raw-MRS 2, 6.8 (4.8–9.2) in raw-MRS 3, 6.2 (3.9–7.5) in raw-MRS 4, and 6.0 (2.9–7.6) in raw-MRS 5. Overall, there was a 61.8% reduction in the median time between the first demyelinating event and the first-ever treatment initiation from patients displaying a raw-MRS 1 to patients with a raw-MRS 5 (Figure 2). Interaction terms between MRS and tertiles are summarized in eTable 8 (links.lww.com/WNL/C982).

Figure 2. Time From the First Demyelinating Event to Treatment Initiation According to Raw-MRS Values in the Subgroup of Patients With Available Radiologic Information.

*This figure shows a smoothed version of a histogram. This is a useful alternative to the histogram for continuous data (as time to treatment is) that comes from an underlying smooth distribution. MRS = magnetic resonance score; n (%) = number of patients (percentage) with available radiologic information included within each the raw-MRS; median (IQR) = median (interquartile range) time from the first demyelinating event to treatment initiation.

Classification of Evidence

This study provides Class III evidence that earlier treatment of patients with MS presenting with a first demyelinating event is associated with improved clinical outcomes.

Discussion

In this study, the association of time to treatment initiation with the risk of long-term disability accrual is assessed in a prospective observational, real-world, and deeply phenotyped cohort of patients with a first demyelinating event and patients with early MS, accounting for baseline radiologic information. The availability of baseline MR scans, a crucial tool for therapeutic decision-making, allowed us to integrate the MRS within the PS models to assess the risk of disability. As expected, patients with highest MRS values displayed the greatest probabilities to initiate treatment soon after the first demyelinating event. After including not only MS-related well-known covariates but also the MRS in models, patients receiving treatment at very early stages of the clinical disease onset (within the first 6 months after the first demyelinating event) had a lower risk of reaching all the disability outcomes assessed, in comparison with patients initiating treatment after 16 months.

According to classical CTs showing that a shorter time from the first demyelinating event to treatment initiation leads to lower rates of MS conversion,^2-4,38 the most recent guidelines encourage DMD initiation early after the MS diagnosis.³⁹ Nonetheless, methodological design in CTs hampers long-term follow-up, which makes difficult the demonstration of risk reduction of the most desirable endpoint in MS: long-term disability. In addition, only few observational real-world studies have shown a beneficial effect of early treatment (defined as treatment initiation between 1 and 3 years from the MS diagnosis) on disability, beyond CT extension phases.^12,18-22 This cohort offers the opportunity to evaluate treatment responses at very early phases, defined as treatment initiation within the first 6 months from the date of the first-ever clinical demyelinating event.²⁹

To tackle the issue of whether time from the first demyelinating event to treatment initiation is related to disability, this cohort was classified into tertiles according to the time elapsed between both dates. Patients initiating very early treatment after the first attack were younger at treatment initiation because of the lower delay between the first demyelinating event and treatment despite the older age at the attack.²⁹

A higher proportion of patients in the FT group had a positive OCB in the CSF and a more radiologic burden than those in the ST or TT group. Indeed, very early treated patients (the FT group) depicted a more baseline inflammatory activity in the brain and spinal cord MR scans. An early treatment is likely the consequence of a more important clinical and radiologic burden and the fact that more serious and affected patients are increasingly derived to our institution because of being recognized as a referral MS expertise center in most recent epochs.

MR is an important tool for MS diagnosis and prognostic purposes, but there is a need to follow a strict standardization of MR scan acquisition to avoid misdiagnosis and provide accurate prognostic information, as recently recommended.³⁰ Baseline MR scans in this cohort are systematically performed following such recommendations and led us to propose a comprehensive score to measure MR activity and burden. In previous real-world prognostic studies, radiologic information has been not included despite a correct use of well-known PS to balance in an equitable manner different features that could confound specific treatment responses (i.e., sex, age, disease activity, etc).^12,18-20,22

The MRS was associated with a higher probability to belong to the FT and ST groups, compared with the probability to belong to the TT group when building PS (eTable 2, links.lww.com/WNL/C982). In view of such relevance, this variable was then integrated within the models to evaluate different outcomes of disability. Overall, we found that patients who receive a very early first-ever treatment (within 6 months from the first demyelinating event) were at a lower risk of reaching long-term disability. Although we have to acknowledge that strengths of associations and 95% CI were not substantially modified in the models with or without the MRS, there was a clear improvement in the accuracy of the models measured by the AIC. In clinical practice, it is now widely accepted that individualized treatment decision relies not only on the clinical information but also considers the radiologic picture of the patients in terms of lesion load and activity. However, further studies with larger sample size in each category are probably needed to reflect the full value of incorporating baseline radiologic information to the PSs.

In addition, some important methodological differences deserve to be underlined compared with other studies. First, other observational studies have only focused on few disability outcomes, mostly time to EDSS 4.0 and 6.0.^18-20 We expanded this study to 6 different outcomes and Neuro-QoL measures and found that very early treated patients had a decreased risk of disability accrual in the 6 outcomes except for time to reach EDSS 6.0 (likely due to few events), with significant high strengths of associations. Of interest, in 6 of the 11 measures of the Neuro-QoL, very early treated patients reported an improvement in function (upper and lower extremity function and cognitive function) or a decrease in self-reported disturbances (sleep disturbance, fatigue, and anxiety) when compared with those in the TT group. These results show the effectiveness of very early treatment on PROMs. Second, time to treatment initiation is accounted for at the time of disease onset instead of at MS diagnosis.^{12,18-20,22,24} This is important because in this study, the included classical MS covariates are collected when the clinical disease starts and not in a time frame in which the disease might evolve or progress.^40,41 Other important difference is that in our cohort, patients are allowed to enter on risk of disability before the MS diagnosis, as long as treatment is initiated between the first demyelinating event and MS diagnosis. Because other studies collect clinical data during MS diagnosis, this may imply a lack of treatment information in those initiating treatment between the first demyelinating event and the MS diagnosis, especially in older periods where the MS diagnosis is delayed when compared with recent times.²⁹ Altogether, the current results support the robustness and effectiveness of a very early treatment to halt long-term disability progression and stress such implementation in the real clinical setting.

PDDS is currently a promising PROM to measure disability.³⁴ It is worth to mention that this is the first observational real-world study using PDDS as an outcome measure to evaluate treatment responses. Although PROMs must be further validated as accurate tools to measure disability,²² it seems PDDS to have good sensitivity on the bases of the high strength of association with the FT group, in contrast to other PROMs.²²

We finally assessed the relationship between raw-MRS (not imputed) and time to treatment initiation. A higher raw-MRS was related to a shorter elapsed time to initiate treatment after the first demyelinating event. This time lag showed an abrupt decrease up to 62% between patients with the lowest and the highest score (raw-MRS 1 vs 5). Our results clearly mirror our daily clinical practice indicating a promptly treatment initiation on the base of an easy score and confirm the capital role of MR in treatment decisions. However, when evaluating whether patients with higher MR burden (higher MRS) would be more likely to benefit from very early treatment than patients with less MRI burden (lower MRS), no associations were found (eTable 8, links.lww.com/WNL/C982). One interpretation would be that neurologist's treatment decision-making should not be based on the baseline radiologic information because outcomes will not be modified in very early treated patients with a high MRS (statistical interaction). However, we feel that neurologists must integrate baseline radiologic information to make treatment decisions. This study was not designed to assess this point but to evaluate whether very early treatment will change long-term outcomes accounting for several clinical and paraclinical data, including MR scans. In addition, statistical power decreases when splitting our sample to perform interactions. Finally, a collinearity might exist when including clinical and radiologic information within the same model, thus the raw effect of the radiologic information on the model could be difficult to obtain. Therefore, this estimate could be considered as an indicative purpose for neurologists, although should be further validated in other cohorts.

In recent years, the effect of high-efficacy treatments on disability has been specially proposed.^21,42-44 Thus, one could argue that the lower risk of disability found in early treated patients could arise from the higher percentage of patients initiating high-efficacy drugs in the FT group compared with the ST or TT group. Although groups differ in such proportion, the first-ever high effective DMDs were few representatives, accounting only for 10% of the patients from the whole cohort. Indeed, 83% of patients from the FT group initiated moderate-efficacy DMD as first-ever treatments. In addition, adjusted analyses were performed considering the proportion of time on high-efficacy drugs, and strengths of association were not modified. Although some effect of the high-efficacy treatments in the FT group cannot be completely ruled out, in our cohort, the time to initiate DMD is associated with the risk of long-term disability. Another point to be mentioned is the importance of radiologic information for treatment decisions, confirming that MR information should be integrated in PS analyses when evaluating studies of treatment response. A recent study showed a nonsignificant weak association between high-efficacy treatment and disability after including radiologic information to construct PS,⁴² but most studies have not integrated radiologic information.^21,43,44 In this study, we found that very early treated patients were at lower risk of disability even if this group had a higher baseline radiologic burden. For the construction of the MRS, we have imputed MRI missing data through simple imputation (in this case, the mean of the group); thus, the variance of the MRS could be underestimated. Finally, the inclusion of new brain MR lesions over the disease course was not taken into account to create the MRS because most of the patients from the FT group did not have a second MR scan before treatment initiation due to the short time elapsed from the first symptom to treatment onset. In the same line, other baseline variables such as topography during treatment initiation were not included. DMD treatments are switched because of several reasons other than relapses (i.e., adverse events, new radiologic activity, among others).

Overall, this is the first study evaluating how very early treatment initiation might affect disability in a prospective cohort of patients with a first demyelinating event. Our results also confirm the role of MR in treatment decisions by showing the strong association of baseline radiologic information with the time of treatment initiation. Moreover, PS models with and without MRS show that initiating treatment within the first 6 months after the first demyelinating event is associated with a reduction in the risk of long-term disability accrual.

Glossary

AIC

Akaike information criterion

CEL

contrast-enhancing lesion

CST

contrast sensitivity test

CT

clinical trial

DMD

disease-modifying drug

EDSS

Expanded Disability Status Scale

FT

first tertile

HR

hazard ratio

IPW

inverse probability weighting

IQR

interquartile range

KM

Kaplan-Meier

MDT

manual dexterity test

MR

magnetic resonance

MRS

MR score

MS

multiple sclerosis

MSPT

multiple sclerosis performance test

OCB

oligoclonal band

OR

odds ratio

PDDS

Patient-Determined Disease Step

PROM

patient-reported outcome measure

PS

propensity score

PST

processing speed test

QoL

quality of life

SDP

sustained disability progression

SPMS

secondary progressive MS

ST

second tertile

TT

third tertile

WST

walking speed test

Appendix. Authors

Study Funding

This study has been funded by the European Regional Development Fund and cofunded by Instituto Carlos III through the projects PI19/01606 granted to M. Tintore and S. Otero-Romero and PI20/00800 granted to A. Cobo-Calvo. It has also received support from the “Red Española de Esclerosis Múltiple (REEM),” which is sponsored by FIS, the Instituto de Salud Carlos III, the Ministry of Economy and Competitiveness in Spain, and the “Ajuts per donar Suport als Grups de Recerca de Catalunya,” which is sponsored by the “Agència de Gestió d'Ajuts Universitaris i de Recerca” (AGAUR) of the Generalitat de Catalunya in Spain. Alvaro Cobo-Calvo is supported by a grant from Instituto de Salud Carlos III, Spain; JR19/00007. P. Carbonell-Mirabent's yearly salary is supported by a grant from Biogen to Fundacio privada Cemcat for statistical analysis.

Disclosure

A. Cobo-Calvo has received grant from Instituto de Salud Carlos III, Spain, JR19/00007. C. Tur has is currently being funded by a Junior Leader La Caixa Fellowship (fellowship code is LCF/BQ/PI20/11760008), awarded by “la Caixa” Foundation (ID 100010434). She has also received the 2021 Merck's Award for the Investigation in MS, awarded by Fundación Merck Salud (Spain) and a grant awarded by the Instituto de Salud Carlos III (ISCIII), Ministerio de Ciencia e Innovación de España (PI21/01860). In 2015, she received an ECTRIMS Postdoctoral Research Fellowship and has received funding from the UK MS Society. She has also received honoraria from Roche and Novartis and is a steering committee member of the O'HAND trial and of the Consensus group on Follow-on DMTs. S. Otero-Romero has received speaking and consulting honoraria from Genzyme, Biogen-Idec, Novartis, Excemed, Bayer HealthCare Pharmaceuticals, BiogenGSK, and MSD and research support from Novartis. P. Carbonell-Mirabent has received support from traveling from Biogen, and his yearly salary is supported by a grant from Biogen to Fundacio privada Cemcat for statistical analysis. M. Ruiz and A. Papolla report no disclosures. J. Villacieros Alvarez has received grant from Instituto de Salud Carlos III, Spain; + FI21/00282. A. Vidal-Jordana reports no disclosures. G. Arrambide has received compensation for consulting services or participation in advisory boards from Sanofi, Merck, Roche, and Horizon Therapeutics; travel support for scientific meetings from Novartis, Roche, and ECTRIMS; and speaking honoraria from Sanofi, Merck, and Roche. Arrambide is a member of the executive committee of the International Women in Multiple Sclerosis (iWiMS) network and of the European Biomarkers in MS (BioMS-eu) Network. J. Castilló, I. Galan, M. Rodríguez Barranco, and L.S. Midaglia report no disclosures relevant to the manuscript. C. Nos has received consultation honoraria from Roche. B. Rodriguez Acevedo reports no disclosures relevant to the manuscript. A. Zabalza de Torres has received travel expenses for scientific meetings from Biogen-Idec and Novartis, speaking honoraria from Eisai, and a study grant from Novartis. N. Mongay has a predoctoral grant Rio Hortega from Instituto Carlos III (CM21/00018). J. Rio has received speaking honoraria and personal compensation for participating on Advisory Boards from Biogen-Idec, Genzyme, Merck-Serono, Novartis, Teva, Roche, Janssen, and Sanofi-Aventis. M. Comabella reports no disclosures relevant to the manuscript. C. Auger reports no disclosures relevant to the manuscript. J. Sastre-Garriga has received compensation for consulting services and speaking honoraria from Almirall, Bayer, Biogen, Celgene, Sanofi, Merck, Novartis, Roche, Bial, Biopass, and Teva, is member of the editorial committee of Multiple Sclerosis Journal, and director of Revista de Neurología. A. Rovira serves/served on scientific advisory boards for Novartis, Sanofi-Genzyme, SyntheticMR, Bayer, Roche, Biogen, Icometrix, and OLEA Medical and has received speaker honoraria from Bayer, Sanofi-Genzyme, Bracco, Merck-Serono, Teva Pharmaceutical Industries Ltd., Novartis, Roche, and Biogen. M. Tintore has received compensation for consulting services and speaking honoraria from Almirall, Bayer Schering Pharma, Biogen-Idec, Genzyme, Merck-Serono, Novartis, Roche, Sanofi-Aventis, Viela-Bio, and Teva Pharmaceuticals. M. Tintore is coeditor of Multiple Sclerosis Journal—Experimental, Translational and Clinical. X. Montalban has received speaking honoraria and travel expenses for scientific meetings, has been a steering committee member of clinical trials or participated in advisory boards of clinical trials in the past 3 years with Biogen Idec, Merck Serono, Genentech, Genzyme, Novartis, Sanofi-Aventis, Teva Pharmaceuticals, Roche, Celgene, Actelion, Mylan, and BMS. Go to Neurology.org/N for full disclosures.