Background: Stable disease course may prompt consideration of disease-modifying treatment (DMT) discontinuation in relapsing-remitting multiple sclerosis (RRMS). Objective: To investigate the clinical outcome after DMT discontinuation and to identify predictive factors supporting decision-making. Methods: We included 221 RRMS patients, who discontinued DMT after â©¾12months and had documented follow-up â©¾2years after discontinuation. Hazard ratios (HRs) with 95% confidence intervals (CIs) regarding relapse and disability progression after DMT discontinuation were calculated from Cox regression models. Results: Age >45 years at discontinuation (HR=0.47, CI=0.23-0.95, p=0.038), absence of relapses for â©¾4years on DMT before discontinuation (HR=0.29, CI=0.10-0.82, p=0.020) and absence of contrast enhancing lesions (HR=0.46, CI=0.28-0.78, p=0.004) were independent predictors of absence of relapse after discontinuation. Age >45years and absence of relapses â©¾4years on DMT combined had an HR of 0.06 (CI=0.01-0.44, p45 years and longer disease duration were significantly associated with disability progression after discontinuation. Conclusion: While freedom from further disease activity is generally unpredictable, there is a subset of patients (age â©¾45years, DMT intake â©¾4years without evidence of clinical or radiological disease activity) having a high likelihood of remaining relapse-free after DMT discontinuation. However, close clinical monitoring for recurrent disease activity is mandatory after discontinuing treatment.
MS is an autoimmune, demyelinating, inflammatory neurological disease that develops from a complex interplay of both genetic and environmental factors. The mechanism of demyelination in multiple sclerosis may be activation of myelin-reactive T cells in the periphery. T cells are activated following antigen presentation by antigen-presenting cells such as macrophages and microglia, or B cells. These T cells then express adhesion molecules, allowing their entry through the blood-brain barrier (BBB). These invasive perivascular T cells can secrete proinflammatory cytokines, including interferon gamma and tumor necrosis factor alpha which contribute to the inflammatory processes in the central nervous system. Furthermore, antibodies against myelin also may be generated in the periphery or intrathecally by activated B cells. Ongoing inflammation leads to epitope spread and recruitment of other inflammatory cells (ie, bystander activation). Activated microglia may release free radicals, nitric oxide, and proteases that may contribute to tissue damage.
In summary, MS has 3 characteristic features; Inflammation leading to the infiltration of Perivascular lymphocytes into the central nervous system, demyelination of neurons and the subsequent formation of Central Nervous System lesions (Plaques)
The CNS lesions mainly affect the white matter and they are both disseminated in time (DIT) and in space (DIS). White matter help transmit information between regions of grey matter, where the processing occurs. Therefore symptoms of MS are highly dependent on the location of the lesion in the CNS.
In 1996, 4 main types of MS were defined, be the National Multiple Sclerosis Society, according to the clinical course of the disorder;
The treatments for MS are split into 3 main types that target 3 different aspects of the disease, namely treatments for MS exacerbations/attacks, treatments for specific MS symptoms and treatments to prevent relapses and disease progression. The last group of treatments are the focus of this study.
Despite having no cure for MS there are treatments that significantly reduce both the frequency and severity of relapses in some patients and slow the progress of neurological deficits in MS. These are called Disease Modifying Therapies(DMT). The goal of these therapies is to decrease the extent of damage and scarring to the myelin sheath associated with relapse and in doing so prevent the progression of disease and are especially useful for patients with RRMS.
This study focuses on the first line injectable treatments, Beta interferons and Galatiramer Acetate. Interferon beta balances the expression of pro- and anti-inflammatory agents in the brain, and reduces the number of inflammatory cells that cross the blood brain barrier therefore it also improves nerve regeneration. Galatiramer acetate resemles myelin basic protein, thus acting as a decoy for the immune system to target and in doing so protects the myelin surrounding axons in the CNS.
Despite their usefulness, DMT’s are not suitable for all MS patients due to their side effects. Thus they are only prescribed to patients with RRMS or SPMS who meet certain criteria. Specifically, Beta interferon give rise to headache, chills, fever and pain and redness at injection site while, Glatiramer acetate results in redness and hardening of skin at injection site and rarely palpitations or flushing after injection. Due to these many side effects many patients who experience few or no relapses over a long period of time and have a stable disease course frequently enquire if they can discontinue their DMT without the risk of relapses coming back and the disease and disability progression of MS getting worse. Therefore this study was designed as an observational cohort study that aimed to understand the effects of discontinuation of DMT’s on the primary and secondary endpoints; occurrences of relapses after discontinuation of DMT’s and disability progression after discontinuation of DMT’s respectively and thus and thus provide predictive criteria that may help clinicians and patients make the decision to discontinue DMT’s
0-What is the research question?
What is the Primary Outcome?
The primary outcome of the study is understanding the effects of discontinuation of DMT’s on the primary and secondary endpoints; occurrences of relapses after discontinuation of DMT’s and disability progression after discontinuation of DMT’s respectively.
What are the Secondary outcomes?
The secondary outcome of the study was to use and analyse the findings derived from outcome one to discern any factors that would minimize the adverse effect of discontinuation of DMT’s and thus provide predictive criteria that may help clinicians and patients make the decision to discontinue DMT’s.
Is the research question clear and properly addressed?
Yes it is as the paper aims to answer the question of whether discontinuation of DMT’s in patients with RRMS would largely affect the clinical outcomes in any adverse manner and through that discern any predictive factors that may help clinicians and patients make the decision to discontinue DMT’s.
1 Are the methods valid?
1a Details of Assignment of patients
Are the inclusion and exclusion criteria clear?
Patients for the study were selected form the Innsbruck MS database (IMSD), which contained 1708 patients, through a careful screening process. The exclusion criteria included patients with PPMS or SPMS, patients who received DMT for less than 12 months, patients who discontinued DMT for less than 6 months, patients with less than 2 years of follow up available and patients with a documented pregnancy during the follow-up period.
After all these exclusion criteria were applied, the patients who still remained were included in the study, thus 221 patients were included in the study
Was randomisation done properly?
Randomising before the aforementioned selection process selection process could have influenced the decisions about eligibility and introduced a source of bias and if a large number of randomly selected patients turned out to be ineligible it would have led to a very small sample size and thus significantly lowered the power of the study.
Randomisation after the aforementioned selection process was not carried out for this study since this study was designed as an observational cohort study and all the patients in the inclusion list were discontinuing DMT’s due to one of 3 reasons indicated in the study; namely adverse events, patient’s decision (including desire of pregnancy) or (3) stable disease course (subjectively defined by the treating physician and/or patient).
Randomisation would have been more essential in a case control study that compared the effects of DMT discontinuation in one group against a matched control group of MS patients who continued taking DMT’s. This is explored further in the future works section as a possible continuation from and improvement on this study.
Was assignment concealed from Drs? (at the assignment stage)
Due to the fact that the study was designed as an observational cohort study and the highly specific inclusion criteria the assignment was not concealed from Drs.
Was the sample size big enough?
The sample size for this study comprised all the 221 patients in the inclusion list. This was a relatively small sample size for a cohort study. However considering the large numbers of exclusion criteria and the fact that the study focused on a specific subset of MS patients; namely patients with RRMS who were being treated with either Interferon Beta or Galactemer Acetate for more than 12 months, who are not pregnant and had a greater than 2 year follow up, the study does have a high predictive power. For this subset of MS patients the study’s results provide a good predictive power yet this does not extend to any MS patients outside this subset.
The usefulness of this study is further undermined by the fact that just a month before this was published a similar study was done by the MSBase Registry with a case control design looking at 426 DMT stoppers, which is almost double the sample size in this study thus providing a far larger power.
1b Accounting for patients entered into the trial
How complete was the follow up?
The median follow up period was 3.8 years with maximum follow up period being 26.9 years and minimum follow up being 2 years.
How were patients lost to follow up dealt with?
As part of the exclusion criteria, patients who were lost to follow up either due to discontinuation of DMT for less than 6 months or because patients had less than 2 years of follow up available or because patients had a documented pregnancy during the follow-up period were excluded from the study and thus also excluded from the calculation on the results of the study.
How were patients not receiving assigned treatment (non-compliers) dealt with ?
During the follow up period 69 of the 221 patients restarted DMT’s after the 6 month cutoff imposed in the exclusion criteria. If DMT was restarted during this observation period, the primary endpoint was considered to be reached at the time point of reinitiation of DMT and thus the patients were still used for calculating the final results.
Is there likely to be residual bias because of any of the above?
Since data for the study was collected retrospectively at the first visit this does introduce a potential source of bias due to the lack of data completeness for some variables. Moreover, since this study was not conducted in a case control manner there is a possibility that important unmeasured variables could have had a confounding effect on the observed results.
However, since data was collected prospectively from then on it serves to minimize any sources of residual bias by attempting to correct for any confounders.
The fact that all the patients were selected from the Innsbruck MS database, which mainly represents the demographics of western Austria and its geographical catchment areas and the fact that there was no randomization carried out also introduces a potential source of selection bias and restricts the predictive power and usefulness of this study
How far was it carried out for patients, doctors, other study personnel?
In this study blinding was not carried out as all eligible patients in the inclusion list underwent the same intervention which was the discontinuation of DMT. Furthermore, the fact that this study was designed as an observational cohort study meant that blinding was not entirely necessary
1e Apart from experimental intervention – were the groups treated equally?
Apart from the discontinuation of DMT’s all the other decisions about the patients’ care was left up to their respective doctors and any MS specific interventions or healthcare concerns, such as pregnancies, were noted in the regular follow ups.
2. What are the results?
2a Were outcomes measured in a standard valid & reliable way ?
The primary and secondary endpoints of this study were the occurrences of relapses after discontinuation of DMT’s and disability progression after discontinuation of DMT’s respectively.
A relapse was defined as patient-reported symptoms or objectively observed signs typical of an acute central nervous system (CNS) inflammatory demyelinating event, current or prior to the visit, with duration of at least 24 hours in the absence of fever or infection, separated from the last relapse by at least 30 days.27 Disability progression was defined as a confirmed EDSS increase in 0.5 sustained for 6months. MRI was only included in analysis if performed within a maximum of 6months prior to discontinuation of DMT (MRI at discontinuation) and if there was another MRI available for comparison performed 1-24months prior to MRI at discontinuation (MRI before discontinuation). MRI parameters obtained were increase in T2 lesion load and presence of gadolinium-enhancing lesions. Increase in T2 lesion load was defined as â©¾1 either new or size-enlarged T2 lesion in MRI at discontinuation compared to MRI before discontinuation.
2b Are results comparable at different sites ?
These crietria for measurement and reporting of data helped standardize the measurement of the outcomes among the multiple study sites.
2c How large was the treatment effect ?
Bivariate testing showed a correlation between absence of relapses after DMT discontinuation and younger age at discontinuation (r=âˆ’0.352, p<0.001) and shorter duration of DMT intake (r=âˆ’0.177, p=0.001)
There was no association with disease duration (p=0.327).
ROC analyses indicated best possible cutoff values of 45 years for age at discontinuation (sensitivity: 65%, specificity: 85%) and 4 years for duration of DMT intake without a relapse (sensitivity: 60%, specificity: 81%)
14/56 (25%) patients aged â©¾45 years at discontinuation suffered a relapse after discontinuation of DMT compared to 84/165 (50.9%) of patients <45 years (P<0.001). Patients taking DMT longer than 4years without a relapse before discontinuation relapsed in 6/32 cases (18.8%), compared to 92/189 (48.7%) patients receiving DMT shorter than 4 years or with a relapse within 4 years before discontinuation (p=0.005).
Age >45 years and absence of gadolinium-enhancing lesion before discontinuation of DMT were found to be independent predictors of absence of relapse after discontinuation of DMT, each about cutting the risk for future relapse in half. Absence of relapses for a period of 4 years or longer on DMT reduced the risk of future relapse to under a third. Sex and EDSS at discontinuation did not have any significant impact. Patients aged >45 years with absence of relapses for a period of 4years on DMT had a cumulative hazard ratio (HR) of 0.06 (CI=0.01-0.44, p<0.001) to suffer a relapse, regardless of MRI parameters.
Bivariate analysis exhibited a statistically significant correlation between disability progression and EDSS at discontinuation (r = 0.212, p = 0.002), disease duration (r = 0.172, p = 0.004) and age at discontinuation (r = 0.123, p = 0.042), but not for duration of DMT, occurrence of relapses on DMT and presence of gadolinium-enhancing lesions.
After inclusion of these variables in a multivariate cox regression model, higher EDSS at discontinuation, age >45 years at discontinuation and longer disease duration were the only significant independent predictors of disability progression after discontinuation. Patients aged â©¾45 years with absence of relapses for a period of 4 years on DMT did not have a significant reduction in their risk for disability progression
How precisely is it measured (95% CIs)?
Categorical variables were expressed in frequencies and percentages, parametric continuous variables as mean and 95% confidence intervals (95% CIs) and nonparametric variables as median and range.
Comparisons regarding primary and secondary endpoints were made by bivariate correlations (Kendall’s tau) and Fisher’s exact or chi-square tests as appropriate. Receiver operating characteristic (ROC) analyses were conducted to define the best possible cut-off values of continuous variables for prediction of the primary endpoint. Survival analyses were performed using cox regression models for comparison of prognostic factors over time and assessment of possible confounders. A two-tailed p-value <0.05 was considered significant.
2d What is the clinical significance of the results and how precisely is it measured?Â Â
The secondary outcome of this study has a higher clinical significance than the primary outcome. The results show that
3.Â Are the results applicable to your patient(s)?
3a Are your patients similar to the study patients?
3b Were all the outcomes that are of interest to your patients considered in the trial?
3c Are the benefits worth the potential harms and costs?
In the future a different study design can be used to tackle the same research question. A further case control study, like the one done by the MSBase Registry, that compared the effects of DMT discontinuation in one group against a matched control group of MS patients who continued taking DMT’s could be carried out to minimize any effects of confounders and biases that may have affected the results of the cohort study.
A future cohort study could include a larger sample size that would be more representative of a larger proportion of the patients with MS. Choosing from a more diverse patient demographic would also provide a greater amount of predictive power over the cultural diverse population of MS patients that are found in the local London boroughs.
A similar study could be repeated with a fully prospective design so as to eliminate any selection biases that may have arisen due the retrospective nature of initial data collection. Furthermore, the retrospective nature of this study also led to only 168 of the 221 selected patients having MRI data available within 6 months before discontinuation of DMT thus reducing the predictive and representative power of many conclusions derived from MRI comparison data. Thus, more detailed and complete analysis using MRI data could have been obtained if the study was designed to be entirely prospectively conducted without the retrospective data collection.
Focus on newer DMT’s – Since this study only focused on the first line DMT’s which only account for a portion of the RRMS patients receiving DMT future work can focus on the other available newer DMTs. The data on the post-injectable DMT disease course may not be generalisable to the newer agents thus it is essential to conduct DMT discontinuation studies on other oral or intravenous DMT’s available to patients which have a larger range of side effects and arguably have more severe side effects.
Furthermore even though this study provides some predictive criteria that may help clinicians and patients make the decision to discontinue DMT’s, To definitively answer the question about safety of DMT discontinuation in this patient subset, a randomised trial is required. The first randomised DMT discontinuation trial in MS is was scheduled to start recruitment in 2016 and its findings should provide more conclusive evidence about the safety and viability of DMT discontinuation in the patient subsets identified in this study both in terms of number of relapses and disability progression.