Abstract
Background:
Relapse is a characteristic clinical feature of multiple sclerosis (MS) and is commonly employed as a measure of efficacy following therapeutic intervention. However, less is known about the temporal evolution of subsequent disability or factors predicting recovery.
Objectives:
The objective of this study was to assess the pattern of recovery following relapse and identify factors which predict recovery and residual disability following relapse.
Methods:
A total of 226 relapses were studied prospectively in a cohort of 144 patients with standardised clinical assessments of physical disability including Expanded Disability Status Scale (EDSS), 10-m timed walk, 9-hole peg test and Multiple Sclerosis Impact Scale (MSIS-29) at 0, 2, 6 and 12 months. A total of 82 patients completed 12 months of follow up without further relapse.
Results:
Thirty per cent of relapses were severe (change in EDSS >2.0) of which 11% failed to recover. All measures showed significant improvement at 2 months but additional improvement was also observed in 9-hole peg test and MSIS-29 up to 12 months following initial assessment. Mean time to second relapse was 382 days. The only predictor of relapse severity in the model tested was younger age; however, increasing age and initial relapse severity were also predictors of poor outcome.
Conclusions:
This study shows that the majority of improvement in physical disability following relapse occurs by 2 months but that more subtle recovery can take place over 12 months in a small sub-group of patients. These data will aid in patient counselling and will also inform the timing of therapeutic intervention and physical support.
Introduction
Relapse is a well-established clinical feature of MS in early phase disease during which demyelination of axons occurs as a result of an ingress of activated T cells into the central nervous system (CNS) precipitating an immunological cascade which eventually results in complement mediated damage. Recovery from relapse is complex but pathological mechanisms involved include resolution of oedema and inflammation, remyelination, sodium channel re-distribution along demyelinated axons and possibly a contribution from cortical plasticity.1-3 However, the impact of relapses on long-term disability remains unclear4,5 and whilst intuitively one might expect their frequency and severity to be closely associated with longer-term outcome, this may not be the case 6 and outcome from relapse remains unpredictable.
Although various data are collected within clinical trials, these are rarely published or open to detailed scrutiny. A better knowledge of the natural history of relapse would allow clinicians to provide detailed counselling to patients, advise on treatment strategies and inform planning for clinical trials. To date, few studies have sought to assess the clinical characteristics and temporal evolution of remission following relapse. Of those that have, many have been retrospective 7 and have often only used a single rating scale, the Expanded Disability Status Scale (EDSS). Whilst this is widely used it has frequently been criticised for poor inter-rater reliability,8,9 and lack of sensitivity in relapse.10,11 In addition, the majority of published studies have assessed only patients in relapse treated with steroids12-15 and are usually of short duration.12-14 More detailed studies which follow patients over longer periods and employ a variety of clinical measures in an unselected cohort of patients in relapse are required to inform patient management and to allow patients to make informed decisions on disease modifying treatments whose principal benefit is usually to reduce frequency and severity of relapse.
In this study we assess the pattern of recovery following relapse using a widely used and strategically selected set of clinical scales of physical disability aiming to identify inter-attack interval, factors which predict recovery and residual disability following relapse and the temporal evolution of recovery.
Materials and methods
Patient identification
The University Hospital of Wales is the major tertiary referral centre for neurology in Wales serving a local population of 1.2 million and provides a regional service for patients with MS that has been described previously. 16 A self-referral clinic is available to all patients within the region and contact details of this clinic are provided to all patients on first review. Patients are initially triaged by telephone according to an established proforma, if relapse is considered possible or likely patients are offered an out-patient appointment within 7 days and assessment undertaken by a multidisciplinary clinical team. A relapse was defined as a period of at least 24 hours in which new symptoms developed, or existing ones deteriorated, with objective evidence from clinical examination for a change, and against background course of stability for at least 1 month. 17 All patients seen in the clinic that were considered to be in a relapse were consecutively recruited over a 3-year period. They were treated with steroids where appropriate and followed up prospectively at 2-, 6- and 12-month intervals. If a patient had a further relapse during the follow up period they were censored at this point and were assessed as for a new relapse.
At each visit a standardised history was obtained along with a neurological examination, EDSS, 10-m timed walk, 9-hole peg test,. All patients who were able were assessed by 10-m timed walk; performed in a standardised manner 18 by a dedicated neurophysiotherapist. The EDSS was assessed by either CH or TP and wherever possible continuity of examiner was maintained for each patient throughout the study. The 9-hole peg test was assessed by the neuro-occupational therapist at each visit. Baseline EDSS was defined as the latest EDSS prior to relapse within the last year. A severe relapse was defined as a difference of ≥2 points between baseline and relapse EDSS. For the MSIS-29 physical scale, a value ≥8 was used to define improvement; a minimal change of 8 has been suggested as being a clinically significant change. 19 Patients with clinically isolated syndromes were excluded. Current disease modifying therapy was defined as any patient who had been taking interferon or glatiramer acetate within the last month but had not been initiated within the last 3 months. Approval for the study was obtained from the local research ethics committee (05/WSE03/112).
Statistical methods
Data analysis was performed using SPSS version 16 (SPSS inc., Chicago, IL) statistical package. Data was observed for normality using histograms and scatter plots. Baseline demographic data were compared between those completing only relapse data, relapse and 2 months data, relapse 2 and 6 month data and relapse 2, 6 and 12 month data using chi square, analysis of variance (ANOVA) and Kruskal–Wallis tests. Differences between mean scores at each of the time points (relapse, 2, 6 and 12 months) were compared for the different lengths of follow-up groups (only relapse and 2 months, relapse 2 and 6 months and relapse, 2, 6 and 12 months) using Kruskal–Wallis tests. Mean EDSS, 10-m walk times, 9-hole peg test, and Multiple Sclerosis Impact Scale (MSIS-29) scores at 0, 2, 6 and 12 months were compared using the Freidman test and an increase or decrease in MSIS-29 score of ≥8 was required to define a clinically significant change. 19 Correlations were assessed using Spearman’s correlation coefficient (r).
Median time to next relapse was calculated using a Kaplan–Meier analysis. This is an established method of estimating time-to-event models in the presence of censored cases. It is based on estimating conditional probabilities at each time point when an event occurs and taking the product limit of those probabilities to estimate the survival rate at each point in time. In this study the recorded event was further relapse with cases censored at if they had not had a further relapse during the follow-up period or at the point which they were lost to follow up.
Binary logistic regression was performed for severe relapse as a dependant variable. Independent variables included in the model were sex, current disease-modifying therapy (DMT) use, disease course, disease duration and age. Age and disease duration were considered as continuous variables. A second binary logistic regression were performed for residual disability defined as any increase in EDSS score between baseline and 12 months (or 6 months if this score was lower than 12 months). In this model age, sex, current DMT use, disease course, severity of relapse and degree of early recovery (defined as return to baseline EDSS within the first 2 months) were considered as variables. Ninety-five per cent confidence intervals were calculated for the odds ratios; a p-value <0.05 was considered significant.
Results
Demographics
The study comprised 226 relapse episodes in 144 patients (mean 1.5 relapses per patient, range 1–6, median 1). A total of 116 (80.2%) were female, mean age at relapse was 38.6 (range 17 to 65) and mean disease duration was 9.7 years (range 0.3 to 32.0 years). Nineteen (13.2%) were on DMTs at the time of relapse consisting of either interferon or glatiramer acetate. A total of 122 (84.7%) had a relapsing–remitting (RR) course and the remainder secondary progressive course with well-defined relapses (SPR). Site of relapse was spinal cord in 100 (44.2%), brainstem in 51 (22.6%), cortical in 30 (13.3%), optic nerve in 20 (8.8%) or multiple sites in 25 (11.1%). In 161 relapses (71.2%) a baseline EDSS had been recorded within the previous 12 months. Steroids were used to treat 175 (77.4%) of the relapses with 52 (23.0%) receiving IV methylprednisolone (MP), 93 (41.2%) oral MP and 30 (13.3%) oral prednisolone.
During the period of the study, 226 relapse episodes were recorded, of which 182 (80.5%) completed 2-month follow up without relapse, 23 (10.2%) had a further relapse within 2 months and 21 (9.3%) were lost to follow up. 133 (58.8%) completed 6 month follow up without relapse, 42 (18.6%) had a further relapse between 2 and 6 months, 4 (1.8%) were lost to follow up and 3 (1.3%) were felt to be progressing and therefore were withdrawn. Eighty-two (36.3%) completed 12-month follow up without relapse. There was no difference in relapse severity (EDSS – baseline EDSS) between those completing and not completing 12-month follow up (p = 0.45).
Inter-relapse interval
Median time to subsequent relapse was 382 days (95% CI 335–429). Those patients who failed to complete the 12-month follow up as a result of having a further relapse within 12 months or being lost to follow up, did not differ significantly in baseline characteristics (sex, DMT usage, age, disease course, disease duration, site of relapse, EDSS, etc.) or in the degree of recovery at 2 and 6 months (data not shown) to those who did complete 12 months of follow up and therefore combined data is reported.
EDSS
Figure 1(a) shows the mean EDSS scores at relapse, 2, 6 and 12 months. There was a mean worsening of 1.6 EDSS points during relapse and mean EDSS scores had returned to baseline by 6 months (p < 0.001).
Mean (a) Expanded Disability Status Scale (EDSS), (b) 10-m walk time, (c) right and (d) left 9-hole peg test and (e) Multiple Sclerosis Impact Scale (MSIS-29) scores at baseline, relapse, 2, 6 and 12 months.
There was a significant improvement in mean EDSS scores from relapse to 2 months post-relapse (n = 182, p < 0.0001) and then further small but statistically significant improvement up to 6 months (n = 133, p = 0.006). There was no further significant improvement in mean EDSS scores between 6 and 12 months post-relapse (n = 82).
Based on an analysis of 82 patients with baseline EDSS and completed 12 months follow up: only 9.3% had no change in the EDSS score during relapse and 29.6% had a severe relapse (change in EDSS ≥ 2 points 20 ). By 2 months, 50% had returned to their baseline EDSS and 60% by 6 months. At 2 months 15% had severe residual disability (≥2 points change in EDSS from baseline) and 11.4% by 6 months. No improvement in the proportion of patients with severe residual disability was seen after this time.
Between relapse and 2 months, 23.8% had no improvement or minor worsening of EDSS and 76.2% had an improvement on EDSS score. Between 2 and 6 months, 11.3% had some minor worsening of their EDSS and 28.7% had an improvement in EDSS. Between 6 and 12 months 48.8% had no change in their EDSS level with 30% having some degree of worsening however 21.4% had an improvement in EDSS (Table 1).
Change in Expanded Disability Status Scale (EDSS) scores between relapse and 2 months, 2 and 6 months and 6 and 12 months (performed on 82 patients who completed 12-month follow up).
Ten-metre timed walk
At relapse, 70/82 patients (87.5%) who completed 12 months of follow up were able to walk 10 m. Mean time to complete 10 m at relapse was 15.7 seconds (SD 14.7, range 5–68), at 2 months 11.1 (SD 8.2, range 4–40), at 6 months 11.3 (SD 10.0, range 4–67) and at 12 months 11.5 (SD 9.8, range 4–55. Significant improvement was observed only between relapse and 2 months (p < 0.0001) (Figure 1(b)). Of the 10 patients unable to complete this measure at relapse all were able to do so by 2 months (mean 38.7s).
Nine-hole peg test
At relapse, 78/82 patients (87.5%) who completed 12 months of follow up (97.5%) were able to complete the right 9-hole peg test (mean time 35.9 s). There was a significant improvement of 5.9 seconds in the time taken to perform the test between relapse and 2 months (p < 0.001) and a further small but statistically significant improvement by 6 months (0.3 seconds, p = 0.019). A total of 77 (93.9%) patients were able to complete the left 9-hole peg test at relapse (mean time 32.8 seconds). Significant improvement of 6.5 seconds (p < 0.001) occurred only between relapse and 2 months (Figure 1(c) and (d)).
MSIS-29
There was a significant improvement in mean MSIS-29 score between relapse and 2 months (n = 182, p < 0.001) although no significant change was observed between 2 and 6 (n = 133) or 6 and 12 months (n = 82); however, the overall difference between 2 and 12 months was significant (n = 82, p = 0.006) (Figure 1(e)). A total of 15% had no significant improvement over the year, 35% had ceased improving by 2 months, 25% by 6 months and 25% still demonstrated improvement between 6 and 12 months.
Comparison of EDSS and MSIS-29
The proportion of patients having an improvement with both the EDSS and the MSIS-29 were compared. Using a minimum of +8.0 change in MSIS-29 to indicate a significant change 19 we compared the proportion of patients improving to a change in EDSS or 0.5 points or more between the time points. There was no statistical difference in the proportion of patients improving between each of the time points between the two rating scales (Table 2). The correlation between MSIS-29 at relapse and EDSS was 0.71 (p < 0.001).
Comparison of the proportion of patients improving using the Expanded Disability Status Scale (EDSS) and the Multiple Sclerosis Impact Scale (MSIS-29) at each follow-up interval.
Logistic regression
This was performed on 161 patients with baseline EDSS known. Younger age was a significant predictor (p = 0.01) of relapse severity, with younger patients having a greater risk of more severe relapses (Table 3). The only factors found predictive of severe residual disability were severity of relapse (p = 0.005) and age (p = 0.045).
Results of multivariable analysis; dependent variables: severe relapse (≥2 point change in Expanded Disability Status Scale from baseline) and presence of residual disability (RD) at 1 year.
DMT, disease-modifying therapy
Discussion
In this study of 226 relapses in 144 patients in a population-based cohort from the UK, we have demonstrated that recovery detected by the EDSS is largely complete by 6 months, although a small proportion of patients will continue to have improvement beyond this time up to 12 months. These data are largely in accord with a recent retrospective study which showed stabilisation of EDSS scores in Kaplan–Meier curves at 6 months. 7 However, the majority of the recovery seems to occur even earlier with 87.2% of patients in our study demonstrating an improvement in EDSS by 2 months and 50% recovering to baseline EDSS score, results which are comparable with those from a study by Iuliano et al. in which 78% and 55%, respectively, had improved and recovered by 3 months. 7 It should be noted that a minority of patients continued to worsen after 6 months which may be the result of an early progressive phase, inter-rater variability or time limited exposure to physical therapies.
A further study of 264 patients experiencing relapse, all of whom were receiving DMTs and were treated with either IV or IV plus oral steroids, reported that at 12 months post-relapse 17% had no change in their EDSS, 74% had improved (EDSS change ≥0.5) and 9% were worse (EDSS change ≤ -0.5). 21 Application of the same analysis to our data results in similar results with 19% demonstrating no change, 74% improving and 8% worsening. The aforementioned study of relapse recovery reported mean improvement in EDSS of 0.47, 0.60, and 0.12 improvement in EDSS between relapse and 3 months, 3 and 6 months and 6 and 12 months, respectively. 21 In comparison our study shows a mean improvement of 1.2, 0.2 and worsening of 0.1 points on the EDSS by 2, 6 and 12 months post-relapse.
In this study we have also shown that mobility largely seems to improve early post-relapse with mean 10-m walk time stabilising by 2 months post-relapse, whereas arm function assessed by the 9-hole peg test appeared to continue to improve for 6 months on the right side but only for 2 months on the left. The reasons for this are uncertain but may have been due to a higher proportion of patients included with more severe right arm dysfunction compared with left, or may simply reflect the sensitivity of the test for dominant hand function. The MSIS-29 in our cohort revealed a marked improvement from relapse to 2 months, and although the individual changes between 2 and 6 months and 6 and 12 months were too small to detect any statistical difference, a degree of improvement clearly continued for up to 12 months.
Few studies have assessed correlations between the EDSS and MSIS-29 physical score. One study assessed correlations in 53 rehabilitation patients and 76 patients receiving corticosteroids and reported correlations of 0.27 and 0.69, respectively, 22 and another study of 200 outpatients found a correlation of 0.68. 23 Correlation between EDSS and MSIS-29 physical score in this study was higher than reported in these other studies. Also comparing the proportion of patients showing improvement in EDSS and MSIS-29 using ≥8 as a minimum required for a clinically significant change, 19 then there was no difference between the two scales. The MSIS-29 physical scale has been compared to several other rating scales and has been shown to have the best efficiency to detect change. 24 Also the MSIS-29 has many advantages over the EDSS being a self-report measure it is quick and easy to perform and does not require the presence of a trained person and may be more useful in the context of relapse follow-up.
Only 82 (35.4%) patients in our study completed follow up at 2, 6 and 12 months. The main reason for patient withdrawal was that a further relapse occurred within the planned follow-up period; this may have influenced results potentially by failing to identify differences in recovery between patients with varying inter-attack intervals. However, patients who did not have a further relapse within 12 months appear to have the same demographics and pattern of recovery as those who went to on have a further relapse, suggesting that they were a representative sample. A further limitation of the study may have been the inclusion of more than one relapse from the same patient. Potentially this may have biased the results by over-representation of patients with higher relapse rates. However, on sub-analysis of only those who completed a year of follow up, and therefore a lower relapse rate, there was no difference in the pattern or degree of recovery (data not shown). The inclusion of patients with secondary progressive disease with well-defined relapses may also have influenced results by prolongation of recovery times as well as reducing reliability of the baseline EDSS; however, once again, there was no significant difference in degree or rate of recovery when compared to those with relapsing disease without a progressive component (data not shown).
We also sought to assess factors which may predict those who would have a severe relapse or have greater residual disability. The only factor that was significant in predicting a severe relapse was age.20, 25 Factors associated with severe residual disability post-relapse were severity of relapse and age.26,27 Length of deterioration during relapse has also been suggested as a predictor for poor outcome in one study, although we were did not assess this factor this in our study. 26
In summary, this study demonstrates that whilst the majority of improvement occurs early following relapse a proportion of patients continue improving for up to a year post-relapse, and this has implications for the timing and application of therapeutic interventions. We have also shown good correlation between the EDSS and the MSIS-29 and similar patterns of improvement post-relapse employing these two assessment tools, suggesting that the MSIS-29 may be a reasonable alternative for assessing relapse recovery.
Footnotes
Acknowledgements
The authors would like to acknowledge the MS nurses, Annette Pauli (clinical specialist neurophysiotherapist), Paula Cornelius (MS occupational therapist) and all of the patients who took part.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest statement
The authors declare no conflicts of interest in preparing this article.
Ethics approval
Ethical approval was obtained (05/WSE03/112)