Key Points
Question
Are clinical and demographic factors associated with the rate at which disability accumulates in secondary progressive multiple sclerosis, and is immunotherapy associated with a slower accumulation of disability among patients with active secondary progressive multiple sclerosis?
Findings
In this cohort study of 1621 patients with secondary progressive multiple sclerosis, relapses during the secondary progressive disease stage were associated with a faster rate of disability accumulation. Immunotherapies were associated with reductions in disability accumulation among patients who experienced superimposed relapses during the course of secondary progressive disease, and patients with active secondary progressive multiple sclerosis who were receiving sustained immunotherapy were less likely to become wheelchair-dependent than those who were not receiving immunotherapy.
Meaning
The study’s findings suggest that disease-modifying therapies are associated with a slower rate of disability accumulation in patients with active secondary progressive multiple sclerosis.
Abstract
Importance
It is unclear whether relapses and disease-modifying therapies are associated with the rate of disability accumulation in patients with secondary progressive multiple sclerosis (SPMS).
Objective
To examine the association of relapses with the rate of disability accumulation in patients with SPMS and to assess whether treatment before or during the secondary progressive phase can slow the progression of disability accumulation.
Design, Setting, and Participants
In this observational cohort study, patient data were prospectively collected from the MSBase international registry between January 1, 1995, and February 1, 2018. Among 53 680 patients in the MSBase registry, 4997 patients with SPMS (using the Lorscheider definition) were identified. Of those, 1621 patients were eligible for study inclusion based on sufficient follow-up before and after the onset of SPMS. Data were analyzed from November 15, 2017, to January 13, 2020.
Exposures
The association between disability accumulation and several clinical and demographic variables, including relapses and exposure to immunotherapy, was evaluated.
Main Outcomes and Measures
Two outcomes were analyzed as measures of disability accumulation during SPMS: the rate of disability accumulation during the secondary progressive phase (change relative to the reference population of patients with MS and absolute change) and the risk of becoming wheelchair dependent. A third outcome, the cumulative risk of experiencing confirmed disability progression events, was used for a secondary analysis. Outcomes were evaluated using multivariable mixed models (ie, linear and Cox models).
Results
Of 1621 patients eligible for inclusion, 1103 patients (68.0%) were female, with a mean (SD) age at MS onset of 33.9 (10.6) years. A total of 661 patients (40.8%) experienced superimposed relapses during SPMS. Therapy receipt and relapses during early relapsing-remitting MS were not associated with disability accumulation during the secondary progressive phase. Higher relapse rates during the secondary progressive disease stage were associated with an increased risk of becoming wheelchair dependent (hazard ratio [HR], 1.87; 95% CI, 1.17-3.00; P = .009). Among patients who experienced superimposed relapses during SPMS, greater receipt of disease-modifying therapies was significantly associated with a reduced rate of disability progression and a lower risk of becoming wheelchair dependent.
Conclusions and Relevance
In this study, the rate of disability progression after the onset of SPMS was not associated with the early disease course and treatment decisions. Relapses during SPMS were associated with accelerated disability progression and represent an accessible treatment target. Disease-modifying therapy was associated with improvements in disability outcomes among patients with active relapses during SPMS. The study’s results suggest that inflammatory disease activity remains a substantial yet modifiable component of SPMS.
This cohort study uses data from the MSBase international registry to examine the association of demographic and clinical factors, including the presence of relapses and exposure to immunotherapy, with the rate of disability accumulation in patients with secondary progressive multiple sclerosis.
Introduction
The most common multiple sclerosis (MS) phenotype is an initially relapsing-remitting disease with clinically distinct episodes of disability exacerbation and a subsequent secondary progressive phase of gradually deteriorating neurological function. However, these phases are neither clearly demarcated nor mutually exclusive, with relapses often occurring in conjunction with progressive disease.1
To date, few studies have examined disability progression after the onset of secondary progressive MS (SPMS),2 and those studies have reported contradictory results. In 1 cohort study, a relapsing-remitting phase with a longer duration was associated with slower disease progression in patients with SPMS3; however, another study found no association between the duration of the relapsing-remitting phase and disease progression.4 Across different studies, superimposed relapses during SPMS have been reported to accelerate disability progression,5 to have minimal consequences,6 or even to be potentially associated with better outcomes.2,7 Such discrepancies may be partly explained by the lack of a standardized SPMS definition, with only 1 definition established for use in cohort studies in 2016.8 In patients with primary progressive MS (PPMS), the MSBase Study Group previously found that superimposed relapses were associated with a lower risk of disability progression.9
The role of immunotherapy during SPMS has been reexamined.10 Earlier randomized controlled clinical trials did not establish the efficacy of disease-modifying therapies for disability progression during SPMS.11,12,13 In a 2018 study, siponimod therapy was reported to reduce the risk of disability progression in patients with active SPMS.14 One may speculate that the gradual progression of disability during SPMS is a consequence of uncontrolled inflammation during relapsing-remitting MS (RRMS),15,16 with the early receipt of disease-modifying therapy in patients with RRMS associated with a reduced risk of SPMS conversion.17
In this study, we evaluated the association of disease-modifying therapies and their timing and duration with the progression of disability during SPMS in the context of underlying episodic inflammation. We hypothesized that, as in patients with PPMS, relapses in patients with SPMS represent a treatable target and are associated with a positive immunotherapy response. We also examined demographic and clinical indicators of disability progression during SPMS. To that end, we quantified the accumulation of disability as progression over time, as measured by the Expanded Disability Status Scale (EDSS), which was used to assess the rate of absolute change in disability (scale range, 20 half-steps from 0-10, with higher scores/steps indicating more severe disability); the Multiple Sclerosis Severity Score (MSSS), which was used to assess the rate of disability accumulation compared with a reference population of patients with MS (scores are assigned based on the rank of a patient’s EDSS score compared with patients with similar disease duration); and the likelihood of becoming wheelchair-dependent.
Methods
The MSBase cohort study,18 an ongoing international registry dedicated to evaluating outcomes data in MS (Australian New Zealand Clinical Trials Registry [ANZCTR] Identifier: ACTRN12605000455662), was approved by the Melbourne Health Research Ethics Committee and the local ethics committees of the participating centers (per local regulations). All enrolled patients provided written informed consent. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.
Population and Data Collection
The longitudinal data of 53 680 patients from 123 medical centers that treat patients with MS in 33 countries were extracted from the MSBase registry between January 1, 1995, and February 1, 2018. Data quality procedures were applied as described previously (eMethods 2 in the Supplement).19 Data were analyzed from November 15, 2017, to January 13, 2020.
All data were recorded as part of routine clinical practice, with most centers performing real-time data entry during clinical visits. The MSBase study protocol stipulates minimum annual updates of the data set, although patients with less frequent visits were also included. The data entry portal used was either the MSBase iMed patient record system or the MSBase online data entry system. Only prospectively acquired data were included, with the exception of the date of disease onset, which is typically determined retrospectively. The duration of prospective follow-up for each patient was defined by the dates of the first and last EDSS score entry. Disability was assessed by accredited scorers (with Neurostatus certification) using the EDSS.
Inclusion Criteria
The minimum data set for inclusion consisted of date of birth, sex, dates of first clinical presentation and MS diagnosis, and treatment and relapse information. To ensure availability of early-disease data, patients were required to have 24 months or less between the MS diagnosis and the first recorded EDSS score.
Patients in the secondary progressive stage were identified using criteria from Lorscheider et al,8 which require an increase in EDSS score (either 1.0 point or greater if the last recorded EDSS score before SPMS conversion was ≤5.5 or an increase of 0.5 point or greater if the last recorded EDSS score before SPMS conversion was ≥6) in the absence of a relapse, confirmed over 3 or more months, with a minimum postprogression EDSS score of 4.0. This objective and reproducible definition reflects clinical judgment with 85% accuracy, 89% sensitivity, and 84% specificity.8 Patients were also required to have a minimum of 24 months between MS diagnosis and conversion to SPMS and 12 or more months of prospective follow-up after SPMS conversion.
Study Outcomes
Two outcomes were evaluated as measures of disability progression after SPMS onset. The first outcome was the MSSS progression slope during SPMS. All EDSS scores during SPMS were converted into the MSSS, and the MSSS slope (decile change per year) was calculated using linear regression analysis from the date of SPMS conversion to the last available MSSS entry.20 The MSSS was selected for this purpose because it is suitable for linearization owing to its quasi-normal distribution, which mitigates fluctuations in the EDSS score. The MSSS slope represents a patient’s mean annual change in disability rank compared with a reference population of patients with MS. A positive MSSS slope indicates that a patient is accruing disability faster than most patients with MS in the reference population. The EDSS slope (ie, the change in EDSS score over time) is more immediately interpretable and was used to perform a secondary analysis. The linearized EDSS slope (ie, the mean increase in the number of EDSS steps per year) provides information about the extent of annualized change in disability.
The second outcome was the time from SPMS conversion to an EDSS score of 7.0, confirmed over 3 or more months (confirmed EDSS scores recorded within 30 days of a preceding relapse were excluded). Only patients with an EDSS score of less than 7.0 at the time of SPMS conversion were included. The cumulative risk of experiencing any EDSS progression event after SPMS conversion, confirmed over 6 or more months with no interval regression (confirmed EDSS scores recorded within 30 days of a preceding relapse were excluded), was used to perform another secondary analysis. Only patients with a minimum of 2 visits to an MSBase study center that were 6 or more months apart after SPMS onset were included. The choice of confirmation times was based on a lower probability of recovery from disability progression at higher EDSS scores.21
Patient Characteristics
Several clinical and demographic variables were collected for all patients. The annualized relapse rate and proportion of time during which patients received disease-modifying therapy were calculated for 2 disease periods (the durations of which were used as the denominators). The first disease period, defined as early RRMS, comprised the first 4 years after MS diagnosis and was censored at SPMS conversion if required. The second disease period, defined as during SPMS, comprised the period from SPMS conversion to the last recorded EDSS score. A relapse was defined as the occurrence of new symptoms or the exacerbation of existing symptoms that persisted for 24 hours or more in the absence of concurrent illness or fever and that occurred 30 days or more after a previous relapse.22
The overall proportion of time that patients received disease-modifying therapies during the 2 disease periods was stratified by estimated therapy efficacy into 3 categories: low efficacy (treatment with interferon beta preparations, glatiramer acetate, and teriflunomide), medium efficacy (treatment with fingolimod, dimethyl fumarate, cladribine, and daclizumab), and high efficacy (treatment with natalizumab, mitoxantrone, alemtuzumab, autologous hematopoietic stem cell transplantation, ocrelizumab, and rituximab).23 The proportion of time during which patients received disease-modifying therapies was defined by the recorded starting and ending dates; for disease-modifying therapies in which extended treatment effects are recognized, the estimated treatment effect duration was used to calculate the proportion of time that patients received therapy (eMethods 2 in the Supplement). For patients who switched between classes of therapy during either disease period, the proportion of time the patient received each therapy class was calculated and evaluated separately, and treatment exposure was presented as the overall proportion of time that patients received each disease-modifying therapy class during either the relapsing-remitting or secondary progressive stages.
Statistical Analysis
The associations between patient characteristics and each of the outcomes were analyzed using multivariable mixed models, with study center as a random effect. The MSSS progression slope during SPMS was analyzed using mixed linear regression analysis, weighted by the number of visits recorded during SPMS. The time from SPMS conversion to a confirmed EDSS score of 7.0 was evaluated using mixed Cox proportional hazards models. The cumulative risk of experiencing any confirmed EDSS progression event after SPMS conversion was analyzed using mixed Andersen-Gill models.24 All models were adjusted for the duration of follow-up during SPMS and the EDSS step at SPMS conversion. For all outcomes, multivariable models were applied to 2 subgroups of the study population: patients with and without relapses after the onset of SPMS.
To ensure that primary analysis results were not artifacts of short follow-up duration during SPMS, we conducted sensitivity analyses that only included patients with 5 or more years of follow-up after the onset of SPMS for both primary outcomes. All tests were 2-tailed and unpaired with a significance threshold of P = .05. Statistical analysis was performed by 2 authors (N.L. and T.K.) using R software, version 3.1.0 (R Project for Statistical Computing).
Results
Among 53 680 patients in the MSBase cohort, 4997 patients with SPMS (using the Lorscheider definition) were identified (Figure 1; eTable 1 in the Supplement). Of those, 1621 patients met the inclusion criteria (1103 women [68.0%]; mean [SD] age at MS onset, 33.9 [10.6] years). A total of 661 patients (40.8%) experienced superimposed relapses during SPMS. Most of the 48 683 patients excluded were not in the secondary progressive phase of disease during the recorded follow-up, which was expected given the continuing recruitment of new study centers and patients for the database. Demographic characteristics and clinical data of the study cohort are summarized in Table 1.
Figure 1. Study Population Flowchart.
EDSS indicates Expanded Disability Status Scale; MS, multiple sclerosis; and SPMS, secondary progressive multiple sclerosis.
Table 1. Characteristics of the Study Population.
| Characteristic | No. (%) (N = 1621) |
|---|---|
| Female sex | 1103 (68.0) |
| Age at MS onset, mean (SD), y | 33.9 (10.6) |
| Time from onset to MS diagnosis, mean (SD), y | 4.1 (5.8) |
| Time from MS diagnosis to SPMS conversion, mean (SD), y | 7.8 (4.9) |
| EDSS score/step at MS diagnosis, mean (SD) | 2.8 (1.5) |
| EDSS score/step at SPMS conversion, mean (SD) | 5.1 (1.1) |
| Follow-up during SPMS, mean (SD), y | 5.6 (3.9) |
| Any relapses during SPMS | 661 (40.8) |
| Relapse rate, No./ya | |
| Early RRMS, mean (SD)b | 0.68 (0.57) |
| SPMS, mean (SD) | 0.17 (0.29) |
| Duration of DMTc | |
| Early RRMS treatmentb | |
| Low efficacy | 1111 (68.5) |
| Proportion of time receiving treatment, mean (SD), % | 43.9 (37.5) |
| Medium efficacy | 71 (4.4) |
| Proportion of time receiving treatment, mean (SD), % | 1.6 (8.8) |
| High efficacy | 161 (9.9) |
| Proportion of time receiving treatment, mean (SD), % | 3.1 (12.0 |
| SPMS treatment | |
| Low efficacy | 902 (55.6) |
| Proportion of time receiving treatment, mean (SD), % | 35.7 (41.6) |
| Medium efficacy | 359 (22.1) |
| Proportion of time receiving treatment, mean (SD), % | 12.0 (27.7) |
| High efficacy | 376 (23.2) |
| Proportion of time receiving treatment, mean (SD), % | 11.0 (26.4) |
Abbreviations: DMT, disease-modifying therapy; EDSS, Expanded Disability Status Scale; MS, multiple sclerosis; RRMS, relapsing-remitting multiple sclerosis; SPMS, secondary progressive multiple sclerosis.
Annualized relapse rate is presented for the entire study cohort.
Early RRMS was defined as the first 4 years after MS diagnosis.
Number of patients who used DMTs of the respective class at any stage during each disease period. The mean (SD) proportion of time receiving DMT during each disease period was calculated as the percentage of the total duration of each period during which the patient received each class of DMT, allowing for multiple drug classes to be used by an individual patient. The mean value for the entire cohort is presented.
Interferon beta preparations and glatiramer acetate represented most of the low-efficacy disease-modifying therapy exposures (1112 of 1396 [79.7%] and 280 of 1396 [20.1%] exposures during early RRMS and 793 of 1163 [68.2%] and 303 of 1163 [26.1%] exposures during SPMS, respectively). Fingolimod and dimethyl fumarate represented most of the medium-efficacy exposures (66 of 74 [89.2%] and 6 of 74 exposures [8.1%] during early RRMS and 295 of 385 [76.6%] and 83 of 385 exposures [21.6%] during SPMS, respectively). Natalizumab and mitoxantrone represented most of the high-efficacy exposures (93 of 171 [54.4%] and 76 of 171 exposures [44.4%] during early RRMS and 285 of 411 [69.3%] and 105 of 411 exposures [25.5%] during SPMS, respectively).
Factors in Disability Progression
MSSS Progression Slope During SPMS
The mean (SD) MSSS progression slope during SPMS was −0.02 (0.36) points per year. We did not find data to indicate that patient demographic characteristics, early receipt of therapy, and early relapses were associated with disability progression in patients with SPMS, with the exception of an association between older age at MS onset and lower MSSS progression slope in patients with superimposed relapses (β = −0.003; 95% CI, −0.005 to −0.001; P = .003) (Table 2). For patients who experienced superimposed relapses during SPMS, a reduced MSSS progression slope during SPMS was observed among those who received disease-modifying therapies for a greater proportion of time during SPMS (per 25% increase in proportion of time which patients with SPMS spent receiving treatment, β = −0.025; 95% CI, −0.039 to −0.012; P < .001 for low-efficacy therapies; β = −0.022; 95% CI, −0.044 to −0.001; P = .06 for medium-efficacy therapies; and β = −0.034; 95% CI, −0.056 to −0.013; P = .002 for high-efficacy therapies).
Table 2. Factors in the MSSS Progression Slope During Secondary Progressive Multiple Sclerosisa.
| Variable | All patients (N = 1621) | Patients with superimposed relapse(s) (n = 661) | Patients with no relapse(s) during SPMS (n = 960) | |||
|---|---|---|---|---|---|---|
| β (95% CI) | P value | β (95% CI) | P value | β (95% CI) | P value | |
| Male sex | 0.024 (−0.004 to 0.053) | .09 | 0.012 (−0.026 to 0.051) | .52 | 0.035 (−0.007 to 0.077) | .10 |
| Age at MS onset, y | −0.001 (−0.002 to 0.000) | .11 | −0.003 (−0.005 to −0.001) | .003 | 0.001 (−0.001 to 0.003) | .42 |
| EDSS score/step at MS diagnosis | 0.001 (−0.009 to 0.012) | .75 | 0 (−0.015 to 0.014) | .97 | 0.003 (−0.012 to 0.018) | .68 |
| Relapse rate, No./y | ||||||
| Early RRMSb | 0.010 (−0.014 to 0.034) | .40 | 0.007 (−0.022 to 0.036) | .66 | 0.034 (−0.007 to 0.076) | .11 |
| SPMS | −0.029 (0.016 to 0.073) | .20 | 0.018 (−0.043 to 0.078) | .57 | NA | NA |
| Proportion of patients with early RRMS receiving DMT (per 25% increase) | ||||||
| Low efficacy | 0.006 (−0.004 to 0.016) | .26 | 0.006 (−0.007 to 0.020) | .35 | 0.004 (−0.012 to 0.019) | .64 |
| Medium efficacy | −0.046 (−0.104 to 0.013) | .12 | −0.115 (−0.213 to −0.017) | .02 | −0.015 (−0.089 to 0.059) | .68 |
| High efficacy | 0.008 (−0.025 to 0.041) | .64 | 0.028 (−0.020 to 0.076) | .25 | −0.019 (−0.066 to 0.028) | .43 |
| Proportion of patients with SPMS receiving DMT (per 25% increase) | ||||||
| Low efficacy | −0.015 (−0.025 to −0.005) | .003 | −0.025 (−0.039 to −0.012) | <.001 | −0.008 (−0.023 to 0.007) | .29 |
| Medium efficacy | −0.015 (−0.032 to 0.001) | .06 | −0.022 (−0.044 to 0.001) | .06 | −0.012 (−0.036 to 0.011) | .31 |
| High efficacy | −0.021 (−0.037 to −0.006) | .007 | −0.034 (−0.056 to −0.013) | .002 | −0.010 (−0.032 to 0.012) | .39 |
| Adjustment variables | ||||||
| Time from MS onset to diagnosis, y | 0.012 (0.009 to 0.015) | <.001 | 0.010 (0.007 to 0.015) | <.001 | 0.013 (0.009 to 0.016) | <.001 |
| Time from MS diagnosis to SPMS conversion, y | 0.013 (0.009 to 0.016) | <.001 | 0.011 (0.006 to 0.016) | <.001 | 0.014 (0.009 to 0.019) | <.001 |
| EDSS score/step at SPMS conversion | 0 (−0.015 to 0.014) | .99 | −0.003 (−0.023 to 0.017) | .77 | 0.001 (−0.020 to 0.024) | .86 |
| Follow-up duration during SPMS, y | 0.001 (−0.003 to 0.004) | .74 | −0.001 (−0.006 to 0.003) | .56 | 0.002 (−0.003 to 0.008) | .39 |
Abbreviations: DMT, disease-modifying therapy; EDSS, Expanded Disability Status Scale; MS, multiple sclerosis; MSSS, Multiple Sclerosis Severity Score; NA, not applicable; RRMS, relapsing-remitting multiple sclerosis; SPMS, secondary progressive multiple sclerosis.
Results of the multivariable linear mixed model for the entire cohort and 2 subgroups (dichotomized by the presence of any superimposed relapses during SPMS).
Early RRMS was defined as the first 4 years after MS diagnosis.
To complement the analysis of the proportion of time that patients received disease-modifying therapies during SPMS as a continuous variable, Figure 2A and B illustrates MSSS trajectories stratified by the proportion of time during SPMS for which patients received treatment. The MSSS at the time of conversion was different between the 3 groups. Notably, their trajectories continued to diverge after SPMS conversion. The trajectory of patients with high receipt of disease-modifying therapies (>90%) was distinct from the trajectory of patients with lower receipt of disease-modifying therapies (50%-90% and <50%), especially in the group who experienced superimposed relapses, in which a decreasing MSSS slope was observed; a more stable trajectory was observed in the group who did not experience relapses.
Figure 2. Disability Progression After Onset of Secondary Progressive Multiple Sclerosis (SPMS) by Treatment Persistence.
A, Patients with ≥1 relapse during SPMS. B, Patients without relapse during SPMS. C, Patients with Expanded Disability Status Scale (EDSS) score <7 and ≥1 relapse during SPMS. D, Patients with EDSS score <7 and no relapse during SPMS. DMT indicates disease-modifying therapy; MSSS, Multiple Sclerosis Severity Score.
The secondary analysis evaluating EDSS progression slopes (EDSS steps per year) revealed similar results (eTable 2 and eFigure in the Supplement). The mean (SD) EDSS progression slope was 0.16 (0.28). The reductions in MSSS slope per 25% increase in the proportion of time which patients with SPMS spent receiving treatment were as follows: for low-efficacy therapies, β = −0.022 (95% CI, −0.033 to −0.010; P < .001); for medium-efficacy therapies, β = −0.019 (95% CI, −0.037 to −0.001; P = .04); and for high-efficacy therapies, β = −0.028 (95% CI, −0.045 to −0.010; P = .002). In addition, for the combined cohort, a higher annualized relapse rate during SPMS was associated with a greater EDSS progression slope during SPMS (β = 0.036; 95% CI, 0.001-0.072; P =.05).
Progression to EDSS 7.0
Among 1494 patients with an EDSS score of less than 7.0 at SPMS conversion, 267 patients (17.9%) reached a confirmed EDSS score of 7.0 during the follow-up period. Early receipt of therapy and relapses were again not associated with the outcome (Table 3). A lower risk of reaching an EDSS score of 7.0 was observed in patients with relapses who were of an older age (hazard ratio [HR], 0.98; 95% CI, 0.96-1.00; P = .03) and had a longer disease duration (HR, 0.95; 95% CI, 0.90-1.00; P = .03) at SPMS onset.
Table 3. Factors in Progression to EDSS Score of 7.0 After Onset of Secondary Progressive Multiple Sclerosisa.
| Variable | All patients (n = 1494) | Patients with superimposed relapse(s) (n = 624) | Patients with no relapses during SPMS (n = 870) | |||
|---|---|---|---|---|---|---|
| HR (95% CI) | P value | HR (95% CI) | P value | HR (95% CI) | P value | |
| Male sex | 1.06 (0.81-1.38) | .68 | 1.15 (0.77-1.70) | .50 | 0.98 (0.67-1.43) | .90 |
| Age at MS onset, y | 0.98 (0.97-0.99) | .006 | 0.98 (0.96-1.00) | .03 | 0.98 (0.96-1.00) | .07 |
| Time from MS onset to diagnosis, y | 0.97 (0.95-1.00) | .02 | 0.97 (0.94-1.00) | .09 | 0.97 (0.94-1.00) | .09 |
| Time from MS diagnosis to SPMS conversion, y | 0.96 (0.93-0.99) | .01 | 0.95 (0.90-1.00) | .03 | 0.97 (0.93-1.02) | .19 |
| EDSS score/step at MS diagnosis | 1.05 (0.96-1.15) | .31 | 1.08 (0.95-1.22) | .24 | 1.01 (0.88-1.15) | .91 |
| Relapse rate, No./y | ||||||
| Early RRMSb | 1.02 (0.81-1.28) | .89 | 1.04 (0.76-1.41) | .82 | 1.01 (0.70-1.45) | .98 |
| SPMS | 1.87 (1.17-3.00) | .009 | 1.50 (0.76-2.97) | .24 | NA | NA |
| Proportion of patients with early RRMS receiving DMT (per 25% increase) | ||||||
| Low efficacy | 1.04 (0.94-1.15) | .44 | 1.01 (0.87-1.17) | .87 | 1.05 (0.91-1.21) | .52 |
| Medium efficacy | 0.95 (0.42-2.18) | .91 | 0.87 (0.21-3.64) | .85 | 1.02 (0.36-2.88) | .98 |
| High efficacy | 1.27 (0.94-1.72) | .12 | 1.27 (0.79-2.03) | .33 | 1.25 (0.84-1.88) | .28 |
| Proportion of patients with SPMS receiving DMT (per 25% increase) | ||||||
| Low efficacy | 0.84 (0.76-0.92) | <.001 | 0.78 (0.68-0.90) | <.001 | 0.90 (0.78-1.03) | .13 |
| Medium efficacy | 0.69 (0.56-0.85) | <.001 | 0.73 (0.55-0.96) | .03 | 0.63 (0.44-0.89) | .01 |
| High efficacy | 0.80 (0.68-0.95) | .01 | 0.73 (0.57-0.94) | .02 | 0.90 (0.71-1.13) | .36 |
| Adjustment variables | ||||||
| EDSS score/step at SPMS conversion | 2.74 (2.31-3.24) | <.001 | 2.61 (2.07-3.27) | <.001 | 3.01 (2.32-3.92) | <.001 |
| Follow-up duration during SPMS, y | 0.97 (0.93-1.00) | .07 | 0.95 (0.90-1.01) | .09 | 0.97 (0.92-1.03) | .31 |
Abbreviations: DMT, disease-modifying therapy; EDSS, Expanded Disability Status Scale; HR, hazard ratio; NA, not applicable; RRMS, relapsing-remitting multiple sclerosis; SPMS, secondary progressive multiple sclerosis.
Results of the Cox mixed survival model for the entire cohort and 2 subgroups (dichotomized by the presence of any superimposed relapses during SPMS).
Early RRMS was defined as the first 4 years after MS diagnosis.
For the complete cohort, a greater annualized relapse rate during SPMS was associated with an increased risk of reaching an EDSS score of 7.0 after SPMS onset (ie, becoming wheelchair-dependent; HR, 1.87; 95% CI, 1.17-3.00; P = .009). An association between a greater proportion of time receiving disease-modifying therapies during SPMS and a reduced risk of progression to an EDSS score of 7.0 was observed in patients who experienced superimposed relapses during SPMS (per 25% increase in proportion of SPMS, HR, 0.78; 95% CI, 0.68-0.90; P < .001 for low-efficacy therapies; HR, 0.73; 95% CI, 0.55-0.96; P = .03 for medium-efficacy therapies; and HR, 0.73; 95% CI, 0.57-0.94; P = .02 for high-efficacy therapies). In patients who did not experience relapses during SPMS, a similar association was noted with medium-efficacy disease-modifying therapies only (HR, 0.63; 95% CI, 0.44-0.89; P = .01).
Figure 2C and D shows the probability of reaching an EDSS score of 7.0, stratified by the proportion of time during which patients received any immunotherapy (complementary to the analyses that used the percentage of time that patients received treatment as a continuous variable). The difference among treatment persistence groups was more pronounced in patients who experienced superimposed relapses compared with those who did not.
Disability Progression Events
This secondary analysis included 1504 patients who had a minimum of 2 visits that were 6 or more months apart after SPMS conversion. The results of the analysis (eTable 3 in the Supplement) were largely consistent with the primary analyses.
In the combined cohort, a greater annualized relapse rate during SPMS was associated with an increased risk of experiencing EDSS progression events. The association between a greater proportion of time receiving immunotherapy during SPMS and a reduced risk of experiencing disability progression events after SPMS conversion was predominant in the group who experienced superimposed relapses.
Sensitivity Analysis
Our sensitivity analyses included only patients with at least 5 years of follow-up during SPMS. The results are shown in eTable 4 and eTable 5 in the Supplement.
The results of these analyses indicated that receipt of immunotherapy during SPMS was associated with a reduction in both the MSSS progression slope and the risk of progression to an EDSS score of 7.0. This treatment-dependent reduction was only apparent in the subgroup who experienced superimposed relapses during SPMS.
Discussion
In the present study, which used data from the international MSBase registry, we found that relapses were associated with ongoing disability accumulation even during established SPMS. Patients with superimposed relapses during SPMS benefited from continued treatment with disease-modifying therapies. We did not observe any association of relapse frequency and treatment decisions during early RRMS with later disability progression after SPMS conversion.
Quantifying disability by measuring changes in MSSS over time enabled us to compare the changes in disability in individual patients with those of a normative population of patients with MS. For instance, our SPMS cohort experienced a negligible mean (SD) decrease in MSSS of 0.2 (3.6) deciles in 10 years, indicating that their disability trajectory was comparable with that of the general population of patients with MS.20 Based on our models, the level of disability in patients with active SPMS who are continuously treated with high-efficacy immunotherapies would progress more slowly in comparison with the general population with MS by a mean (SD) of 1.56 (4.60) deciles over 10 years.
The MSSS is better suited to linearization over time than the EDSS because of its quasi-normal distribution. However, the mean annualized EDSS change (termed EDSS slope in this article) provides a more meaningful clinical interpretation, and we have therefore used it as a secondary outcome.25,26 For example, in our SPMS cohort, disability progressed by a mean (SD) of 1.6 (2.8) EDSS steps over 10 years. Based on our models, the patients who experienced superimposed relapses during SPMS and sustained treatment with high-efficacy immunotherapies without interruption would progress by a mean (SD) of only 0.48 (2.3) EDSS steps over the same period. This progression translates into a marked slowing of disability progression by a mean (SD) of 1.12 (3.60) EDSS steps (70%) over 10 years in patients who were actively receiving treatment and who had indications of episodic inflammation after SPMS conversion. These patients were also more than 70% less likely to become wheelchair-dependent if they consistently sustained treatment with medium-efficacy or high-efficacy disease-modifying therapies.
In contrast, disease-modifying therapies had little association with disability accumulation in the cohort who did not experience relapses during SPMS. Although 2 significant associations were observed in the relapse-free group (between medium-efficacy therapies and lower risk of reaching an EDSS score of 7.0 and low-efficacy therapies and lower risk of experiencing any EDSS progression event), and our graphical depictions suggest some patterns in patients without relapses, these results were not consistently observed across other disability outcomes and sensitivity analyses. The fluctuations are likely associated with the inclusion of patients with inflammatory disease whose activity had ceased owing to their immunotherapy.
Randomized controlled clinical trials, observational studies, and a systematic review found no data to support the efficacy of several immunotherapies during SPMS.10,11,12,13,27 One exception is the Exploring the Efficacy and Safety of Siponimod in Patients With Secondary Progressive Multiple Sclerosis (EXPAND) clinical trial, which reported that treatment with siponimod was associated with reductions in the risk of disability progression in patients with SPMS.14 In the EXPAND clinical trial, treatment with siponimod was associated with a 21% reduction in the risk of 3-month confirmed disability progression. This finding was associated with the subgroup of patients who experienced superimposed relapses during the 2 years before study enrollment, who had a 33% reduction in risk; in the subgroup of patients who did not experience relapses, the pattern was not statistically significant, with only a 13% risk reduction.14
The Effect of Natalizumab on Disease Progression in Secondary Progressive Multiple Sclerosis (ASCEND) clinical trial of natalizumab for the treatment of patients with SPMS enrolled a predominantly relapse-free cohort and found no indication of an association between treatment with natalizumab and EDSS progression during SPMS.12 Likewise, a previous cohort study performed by the MSBase Study Group did not find any benefits of disease-modifying therapy receipt during SPMS in an analysis adjusted for SPMS relapse rates; however, a sensitivity analysis that did not control for relapses indicated an association between treatment and a reduction in the risk of becoming wheelchair-dependent.10 Together with the present study, the existing data converge on the suggestion that relapses during SPMS provide a therapeutic target and a marker of future response to immunotherapy during SPMS.
Notably, a similar pattern has emerged in patients with PPMS. The MSBase Study Group previously reported that patients with superimposed relapses (but not patients who were relapse-free) also exhibited improvement in disability trajectories when receiving consistent treatment with the available immunotherapy.9 The Study to Evaluate the Safety and Efficacy of Rituximab in Adults With Primary Progressive Multiple Sclerosis (OLYMPUS) clinical trial of rituximab for the treatment of patients with PPMS identified a subgroup in whom treatment with rituximab was associated with a 37% reduced risk of disability progression in those younger than 51 years who had active contrast-enhancing cerebral lesions.28 In the Study of Ocrelizumab in Participants With Primary Progressive Multiple Sclerosis (ORATORIO) clinical trial, treatment with ocrelizumab was associated with a 35% reduction in confirmed disability progression in the subgroup with contrast-enhancing lesions at baseline magnetic resonance imaging compared with a 16% reduction in the subgroup with no baseline magnetic resonance imaging activity. Thus, the finding that episodic inflammation represents a treatable target can be extended to both of the progressive MS phenotypes, SPMS and PPMS.
Previous studies have reported variable significance with regard to relapses during SPMS.2,3,5,6,7 A study performed by the MSBase Study Group among patients with PPMS found that the presence of relapses (as a dichotomous variable) was associated with a lower risk of disability progression in patients receiving immunotherapy. Because the PPMS cohort was largely untreated (<30% of patients), this finding was interpreted as a consequence of relapses presenting a treatable target.9 Our present results, which indicate that a higher relapse rate is associated with faster disability accumulation, does not contradict this previous finding, as the present study analyzed relapses quantitatively rather than dichotomously, and the SPMS cohort received substantially more treatment than the PPMS cohort. In the SPMS cohort, a relapse not only is a treatable target but may signal treatment failure and therefore present an indication of worse prognosis.29 These studies converge on the concept that although the presence of relapses offers an opportunity for treatment of progressive disease, greater relapse rates, despite treatment, are associated with disability accumulation.
In the present study, we found no associations between demographic characteristics, early relapse rate, duration of the relapsing-remitting phase, and slope of MSSS and EDSS scores during SPMS. Other studies have previously indicated that future MS trajectories are difficult to estimate, in part because of rapid changes in the extent of disease activity and the rate of disability accumulation.26,30,31 Our finding of a lack of association between disease activity in the relapsing-remitting phase and disability accumulation during the secondary progressive phase is consistent with this interpretation. Although this intraindividual variability in disease course makes any individual prognosis difficult, it also offers the possibility that MS with sufficient episodic inflammatory activity remains modifiable with immunotherapy at any stage of disease.
Limitations
This study has several limitations. The observational nature of our study was its main limitation. However, randomized controlled clinical trials evaluating the association of therapy with long-term disability accumulation in patients with SPMS are impractical, have limited generalizability, and may present ethical issues.32,33 We controlled for multiple demographic and clinical variables to minimize treatment selection bias, used only prospectively acquired data to mitigate recall bias, applied a rigorous data quality control procedure19 to reduce syntactic errors, and defined minimum follow-up requirements to ensure capture of the earliest stages of clinically manifest disease as well as the clinical course after SPMS onset. We also adjusted analyses for follow-up duration to curtail detection bias, used survival models with censoring (when appropriate) to control attrition bias, used a validated definition of SPMS to maximize reproducibility, and adjusted models for several potential confounding variables. If present, any residual indication bias would be expected to dilute the results obtained owing to the receipt of preferential treatment with disease-modifying therapies among patients with more severe disease.
Results of the sensitivity analyses, which included only patients with a minimum of 5 years of follow-up after SPMS onset, were consistent with our primary results, indicating that these findings were not artifacts of short follow-up duration. Furthermore, our study was conducted using data from a large international registry of patients with MS, which maximizes the generalizability of our results given that treatment availability and practices vary across jurisdictions. To account for heterogeneity in the recorded data, we adjusted all models by MSBase study center. Our analysis was not powered to evaluate the implications of early aggressive therapy, as very few patients who reached the secondary progressive phase had access to more potent agents during their early disease stages. However, the study provided sufficient power to evaluate the consequences of immunotherapy during SPMS in cohorts with and without ongoing relapse activity, although some treatment subgroups within disease phenotypes may have been marginally powered.
Conclusions
Although early active treatment during RRMS is associated with a delay in the onset of SPMS,17 the rate of disability accumulation once the secondary progressive phase has commenced is not substantially modified by early treatment decisions. Notably, similar to PPMS,9 relapses during SPMS represent an accessible treatment target and a marker of likely response to immunotherapy. Thus, when inflammation exists in patients with progressive MS, treatment may be associated with a reduction in the rate of disability accumulation. Such an interpretation is consistent with the present disease classification, in which relapses and progression represent 2 only partially dependent characteristics of each individual disease.1
eMethods 1. List of Contributors
eMethods 2. Data Quality Procedure
eTable 1. Patients Included Per MSBase Center
eTable 2. Secondary Analysis 1: Factors in the EDSS Progression Slope During SPMS
eTable 3. Secondary Analysis 2: Factors in Any Disability Progression Event After SPMS Onset
eTable 4. Sensitivity Analysis: Factors in the MSSS Progression Slope During SPMS in Patients With at Least 5 Years of Follow-Up During SPMS
eTable 5. Sensitivity Analysis: Factors in Progression to EDSS 7 After SPMS Onset in Patients With at Least 5 Years of Follow-Up During SPMS
eFigure. EDSS Trajectories After SPMS Onset by Treatment Persistence
References
- 1.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]
- 2.Vukusic S, Confavreux C. Prognostic factors for progression of disability in the secondary progressive phase of multiple sclerosis. J Neurol Sci. 2003;206(2):135-137. doi: 10.1016/S0022-510X(02)00426-4 [DOI] [PubMed] [Google Scholar]
- 3.Tremlett H, Zhao Y, Devonshire V. Natural history of secondary-progressive multiple sclerosis. Mult Scler. 2008;14(3):314-324. doi: 10.1177/1352458507084264 [DOI] [PubMed] [Google Scholar]
- 4.Scalfari A, Neuhaus A, Daumer M, Muraro PA, Ebers GC. Onset of secondary progressive phase and long-term evolution of multiple sclerosis. J Neurol Neurosurg Psychiatry. 2014;85(1):67-75. doi: 10.1136/jnnp-2012-304333 [DOI] [PubMed] [Google Scholar]
- 5.Soldan MMP, Novotna M, Zeid NA, et al. Relapses and disability accumulation in progressive multiple sclerosis. Neurology. 2015;84(1):81-88. doi: 10.1212/WNL.0000000000001094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Tremlett H, Yousefi M, Devonshire V, Rieckmann P, Zhao Y; UBC Neurologists . Impact of multiple sclerosis relapses on progression diminishes with time. Neurology. 2009;73(20):1616-1623. doi: 10.1212/WNL.0b013e3181c1e44f [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Confavreux C, Vukusic S, Moreau T, Adeleine P. Relapses and progression of disability in multiple sclerosis. N Engl J Med. 2000;343(20):1430-1438. doi: 10.1056/NEJM200011163432001 [DOI] [PubMed] [Google Scholar]
- 8.Lorscheider J, Buzzard K, Jokubaitis V, et al. ; MSBase Study Group . Defining secondary progressive multiple sclerosis. Brain. 2016;139(Pt 9):2395-2405. doi: 10.1093/brain/aww173 [DOI] [PubMed] [Google Scholar]
- 9.Hughes J, Jokubaitis V, Lugaresi A, et al. ; MSBase Study Group . Association of inflammation and disability accrual in patients with progressive-onset multiple sclerosis. JAMA Neurol. 2018;75(11):1407-1415. doi: 10.1001/jamaneurol.2018.2109 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Lorscheider J, Jokubaitis VG, Spelman T, et al. ; MSBase Study Group . Anti-inflammatory disease-modifying treatment and short-term disability progression in SPMS. Neurology. 2017;89(10):1050-1059. doi: 10.1212/WNL.0000000000004330 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Panitch H, Miller A, Paty D, Weinshenker B; North American Study Group on Interferon Beta-1b in Secondary Progressive MS . Interferon beta-1b in secondary progressive MS: results from a 3-year controlled study. Neurology. 2004;63(10):1788-1795. doi: 10.1212/01.WNL.0000146958.77317.3E [DOI] [PubMed] [Google Scholar]
- 12.Kapoor R, Ho P-R, Campbell N, et al. ; ASCEND Investigators . Effect of natalizumab on disease progression in secondary progressive multiple sclerosis (ASCEND): a phase 3, randomised, double-blind, placebo-controlled trial with an open-label extension. Lancet Neurol. 2018;17(5):405-415. doi: 10.1016/S1474-4422(18)30069-3 [DOI] [PubMed] [Google Scholar]
- 13.Secondary Progressive Efficacy Clinical Trial of Recombinant Interferon–Beta-1a in MS (SPECTRIMS) Study Group Randomized controlled trial of interferon–beta-1a in secondary progressive MS: clinical results. Neurology. 2001;56(11):1496-1504. doi: 10.1212/WNL.56.11.1496 [DOI] [PubMed] [Google Scholar]
- 14.Kappos L, Bar-Or A, Cree BAC, et al. ; EXPAND Clinical Investigators . Siponimod versus placebo in secondary progressive multiple sclerosis (EXPAND): a double-blind, randomised, phase 3 study. Lancet. 2018;391(10127):1263-1273. doi: 10.1016/S0140-6736(18)30475-6 [DOI] [PubMed] [Google Scholar]
- 15.Frischer JM, Bramow S, Dal-Bianco A, et al. The relation between inflammation and neurodegeneration in multiple sclerosis brains. Brain. 2009;132(Pt 5):1175-1189. doi: 10.1093/brain/awp070 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Trapp BD, Nave K-A. Multiple sclerosis: an immune or neurodegenerative disorder? Annu Rev Neurosci. 2008;31(1):247-269. doi: 10.1146/annurev.neuro.30.051606.094313 [DOI] [PubMed] [Google Scholar]
- 17.Brown JWL, Coles A, Horakova D, et al. ; MSBase Study Group . Association of initial disease-modifying therapy with later conversion to secondary progressive multiple sclerosis. JAMA. 2019;321(2):175-187. doi: 10.1001/jama.2018.20588 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Butzkueven H, Chapman J, Cristiano E, et al. MSBase: an international, online registry and platform for collaborative outcomes research in multiple sclerosis. Mult Scler. 2006;12(6):769-774. doi: 10.1177/1352458506070775 [DOI] [PubMed] [Google Scholar]
- 19.Kalincik T, Spelman T, Trojano M, et al. ; MSBase Study Group . Persistence on therapy and propensity matched outcome comparison of two subcutaneous interferon beta 1a dosages for multiple sclerosis. PLoS One. 2013;8(5):e63480. doi: 10.1371/journal.pone.0063480 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Roxburgh RHSR, Seaman SR, Masterman T, et al. Multiple Sclerosis Severity Score: using disability and disease duration to rate disease severity. Neurology. 2005;64(7):1144-1151. doi: 10.1212/01.WNL.0000156155.19270.F8 [DOI] [PubMed] [Google Scholar]
- 21.Kalincik T, Cutter G, Spelman T, et al. Defining reliable disability outcomes in multiple sclerosis. Brain. 2015;138(Pt 11):3287-3298. doi: 10.1093/brain/awv258 [DOI] [PubMed] [Google Scholar]
- 22.Schumacher GA, Beebe G, Kibler RF, et al. Problems of experimental trials of therapy in multiple sclerosis: report by the panel on the evaluation of experimental trials of therapy in multiple sclerosis. Ann N Y Acad Sci. 1965;122:552-568. doi: 10.1111/j.1749-6632.1965.tb20235.x [DOI] [PubMed] [Google Scholar]
- 23.Tramacere I, Del Giovane C, Salanti G, D’Amico R, Filippini G. Immunomodulators and immunosuppressants for relapsing-remitting multiple sclerosis: a network meta-analysis. Cochrane Database Syst Rev. 2015;(9):CD011381. doi: 10.1002/14651858.CD011381.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Andersen PK, Gill RD. Cox’s regression model for counting processes: a large sample study. Ann Stat. 1982;10(4):1100-1120. doi: 10.1214/aos/1176345976 [DOI] [Google Scholar]
- 25.Fazekas F, Deisenhammer F, Strasser-Fuchs S, Nahler G, Mamoli B; Austrian Immunoglobulin in Multiple Sclerosis Study Group . Randomised placebo-controlled trial of monthly intravenous immunoglobulin therapy in relapsing-remitting multiple sclerosis. Lancet. 1997;349(9052):589-593. doi: 10.1016/S0140-6736(96)09377-4 [DOI] [PubMed] [Google Scholar]
- 26.Lizak N, Lugaresi A, Alroughani R, et al. ; MSBase Study Group . Highly active immunomodulatory therapy ameliorates accumulation of disability in moderately advanced and advanced multiple sclerosis. J Neurol Neurosurg Psychiatry. 2017;88(3):196-203. doi: 10.1136/jnnp-2016-313976 [DOI] [PubMed] [Google Scholar]
- 27.La Mantia L, Vacchi L, Di Pietrantonj C, et al. Interferon beta for secondary progressive multiple sclerosis. Cochrane Database Syst Rev. 2012;1:CD005181. doi: 10.1002/14651858.CD005181.pub3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hawker K, O’Connor P, Freedman MS, et al. ; OLYMPUS Trial Group . Rituximab in patients with primary progressive multiple sclerosis: results of a randomized double-blind placebo-controlled multicenter trial. Ann Neurol. 2009;66(4):460-471. doi: 10.1002/ana.21867 [DOI] [PubMed] [Google Scholar]
- 29.Jokubaitis VG, Spelman T, Kalincik T, et al. ; MSBase Study Group . Predictors of long-term disability accrual in relapse-onset multiple sclerosis. Ann Neurol. 2016;80(1):89-100. doi: 10.1002/ana.24682 [DOI] [PubMed] [Google Scholar]
- 30.Confavreux C, Vukusic S, Adeleine P. Early clinical predictors and progression of irreversible disability in multiple sclerosis: an amnesic process. Brain. 2003;126(Pt 4):770-782. doi: 10.1093/brain/awg081 [DOI] [PubMed] [Google Scholar]
- 31.Leray E, Yaouanq J, Le Page E, et al. Evidence for a two-stage disability progression in multiple sclerosis. Brain. 2010;133(Pt 7):1900-1913. doi: 10.1093/brain/awq076 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Lublin FD, Reingold SC. Placebo-controlled clinical trials in multiple sclerosis: ethical considerations. National Multiple Sclerosis Society (USA) Task Force on Placebo-Controlled Clinical Trials in MS. Ann Neurol. 2001;49(5):677-681. doi: 10.1002/ana.1025 [DOI] [PubMed] [Google Scholar]
- 33.Trojano M. Is it time to use observational data to estimate treatment effectiveness in multiple sclerosis? Neurology. 2007;69(15):1478-1479. doi: 10.1212/01.wnl.0000275545.08382.82 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods 1. List of Contributors
eMethods 2. Data Quality Procedure
eTable 1. Patients Included Per MSBase Center
eTable 2. Secondary Analysis 1: Factors in the EDSS Progression Slope During SPMS
eTable 3. Secondary Analysis 2: Factors in Any Disability Progression Event After SPMS Onset
eTable 4. Sensitivity Analysis: Factors in the MSSS Progression Slope During SPMS in Patients With at Least 5 Years of Follow-Up During SPMS
eTable 5. Sensitivity Analysis: Factors in Progression to EDSS 7 After SPMS Onset in Patients With at Least 5 Years of Follow-Up During SPMS
eFigure. EDSS Trajectories After SPMS Onset by Treatment Persistence