Article Text
Abstract
Objectives: To determine the long-term outcome of a cohort of 436 patients with multiple sclerosis (MS) seen in 1985 and long-term predictors of benign MS.
Methods: The initial 1985 group of 436 patients with possible MS, including 53 benign patients, were followed for 21 years. Disability was recorded using the Expanded Disability Status Scale (EDSS). Survival from disease onset was calculated. The indicators of benign MS in the initial 1985 cohort and in the survivors in 2006 were determined.
Results: Of the original 436 patients, the 21-year follow-up outcome in 397 (91%) was established. The diagnosis of MS was incorrect in 41/397 (10%) of the whole cohort and in 2/53 (4%) of the benign group. Median survival of 356 patients with MS was 43.6 years from disease onset. Of 47/51 (92%) patients with benign MS followed in 2006, 7 (15%) remained benign, 18 had died and 22 were disabled. Median survival advantage for the 47 benign patients in 1985 compared to the 88 non-benign patients, when corrected for age, was 6 years (p<0.08). In 2006, 40/356 (11%) patients had a benign outcome at a mean follow-up of 26.1 years. A benign course was significantly associated with female sex, younger age of onset and absence of motor symptoms at presentation.
Conclusions: Although designating patients as having a benign course after 10 years has a poor predictive value, three factors at presentation indicate a better prognosis.
Statistics from Altmetric.com
Multiple sclerosis (MS) is a disease that is characterised by a wide range of rates of progression.1 A number of patients with MS have a benign course and live for many years without the accumulation of disability. Benign MS is defined in terms of disability scales and disease duration; an Expanded Disability Status Scale (EDSS)2 score of ⩽3 after 10 years of disease is the most commonly used definition.3 4 It has been proposed that using the definition of benign MS as EDSS ⩽2 after 10 years has a more sensitive predictive value for long-term outcome.5
The reported frequency of benign MS has varied across studies between 5%6 and 40%7 depending on the definition used and the follow-up time. Due to a paucity of long-term studies of benign MS, the eventual outcome remains uncertain.3 5 8 Thus, the aims of our study were to follow up a large MS cohort, including a subgroup of benign MS cases, 21 years after they were originally assessed in order to ascertain the outcome and survival of the whole cohort and the original benign group. We also aimed to examine various baseline patient characteristics that may be predictive of a benign course after at least 20 years of disease.
PATIENTS AND METHODS
Study population
A total of 436 possible patients with MS attending outpatients in St. Vincent’s Hospital in Dublin were prospectively assessed in 1985.9 The diagnostic criteria of McDonald and Halliday10 were used in classifying patients as clinically definite MS (240 patients), progressive probable MS (50 patients), progressive possible MS (28 patients), early probable MS (43 patients) and suspected MS (75 patients). Demographic details, age at onset of MS, disease duration, initial presenting symptoms, subtype of MS, time between first and second relapse, and Kurtzke DSS score, were recorded.11 Mean age at onset of MS was 33 years (SD 11.6), mean DSS score in 1985 was 3.0 (SD 2.2) and mean disease duration was 8.7 years (SD 9.1). This population served as the cohort for this long-term follow-up.
Methods
Follow-up was carried out in 2006, 21 years after the initial assessment. EDSS scores were assessed either by clinical examination (15%) or by telephone (85%).12 Because health records in the Republic of Ireland are not centrally based, a wide range of agencies had to be contacted to ensure as complete ascertainment as possible. For patients who were deceased, the date and cause of death on the death certificate were recorded. Ethical approval for this study was granted by the Research and Ethics Committee of St. Vincent’s University Hospital, Dublin.
Statistics
Clinical variables were compared between groups using unpaired Student t tests and Fisher’s exact tests. Kaplan–Meier curves were used to calculate survival for the whole cohort and the benign subgroup from onset of MS symptoms. Patients who were lost to follow-up and those who were still alive at follow-up were censored. The log rank test was used to compare survival for benign versus “not benign” groups. A Cox proportional hazards model was used to correct for gender. Total survival for benign versus not benign groups was also calculated to correct for age when comparing these groups. A stepwise logistic regression model was created to examine baseline variables in relation to a benign outcome. Data were analysed using Prism 4 software (GraphPad Software Inc.; La Jolla, CA) and R Version 2.4.0 (R Development Core Team, Vienna, Austria).
RESULTS
Baseline characteristics of the 1985 cohort
Of the 436 patients seen in 1985, 141 patients had a disease duration of 10 years or more. Of these, 53 (38%) patients met the criterion for “benign MS” (EDSS ⩽3 at a disease duration of 10 years or more); 30 (21%) patients met the newer criterion for benign MS (EDSS ⩽2 at a disease duration of 10 years or more). In comparing the 53 benign patients (EDSS ⩽3 at 10 years or more) and the 88 patients who were “not benign” in 1985, there was a slightly greater proportion of women in the benign group (p = 0.07) (table 1). The age at onset of MS was earlier in the benign patients, and the mean time between first and second relapse was significantly longer in the benign group. The only significant difference between the groups in symptoms at onset was the lower proportion of benign patients who had motor involvement at the outset. The same clinical characteristics were significant when an EDSS score of ⩽2 was used to define benign disease. When a multivariate analysis of the baseline characteristics was performed, only female sex (OR 0.4, 95%CI: 0.17 to 0.91) and a longer relapse-free interval (OR 0.98, 95%CI: 0.98 to 0.99) were predictive of a benign outcome in the short term.
Follow-up of the 1985 total cohort
At review in 2006 of the 1985 cohort of 436 patients, 41 patients were found to not have MS and were excluded from analyses. These 41 patients included 17 patients with a clinically isolated syndrome and 24 with alternative diagnoses. A further 39 patients with MS were lost to follow-up, thus complete follow-up data were obtained in 356 of the original 395 patients with MS (10% lost to follow-up). Median survival of this cohort of 356 patients from onset of MS symptoms was 43.6 years.
Follow-up of the 1985 benign cohort
Of the 53 benign patients, 4 were lost to follow-up and a further 2 patients were found not to have MS; one patient had a clinically isolated syndrome and the other had cerebrovascular disease. Thus, 47 (92%) of the 51 patients who were benign in 1985 were followed for a further 21 years to 2006. Of these, seven (15%) still met the criterion for benign MS at follow-up, 22 patients (47%) had worsened disability (median EDSS 7.0, SD 1.1, range 6–9.5) and 18 (38%) had died. Of these 18 patients, 10 had died as a result of complications of MS and the other 8 had died of other causes. Similarly, of the 30 patients who had EDSS ⩽2 in 1985, 27 were followed to 2006, at which time 4r (15%) were still benign, 16 had progressive MS and 7 had died (5 MS related, 2 due to other causes).
The median survival for the group of 47 that had benign MS in 1985 was 52.1 years compared to 41.6 years for the group that were not benign (p = 0.008, HR 0.47, 95% CI: 0.30 to 0.83) (fig 1). After correction for age, there was a non-significant survival advantage of 6 years for the benign group (p = 0.08, HR 0.62, 95% CI: 0.35 to 1.1). Correction for gender did not affect the analysis.
Benign MS in 2006
Of the 1985 cohort followed to 2006, 40/356 (11%) patients with MS now met criteria for benign MS (EDSS ⩽3 after more than 20 years of having MS). Of these 40, 7 had met criteria for benign MS in 1985 and the other 33 cases had not had MS for 10 years when seen in 1985. Univariate analysis of factors recorded in 1985 significantly associated with a long-term benign outcome were female sex, younger age at onset of symptoms, and absence of motor symptoms at initial presentation. This benign group of 40 patients, however, had significantly shorter disease duration at follow-up (26.1 years) compared to the 316 patients with a non-benign outcome (30.5 years). However, when a multivariate analysis of early disease attributes in 1985, predictive of a long-term benign outcome, was conducted, the results concurred with the univariate analysis and the same baseline variables remained significant (see table 2). In 2006, 32/356 (9%) patients met criteria for benign MS using EDSS ⩽2 as a cut-off. The same baseline characteristics predicted a long-term benign outcome in this group.
The chances of a benign long-term outcome
From the logistic regression model, we can estimate the odds ratio (OR) for a long-term benign outcome based on baseline clinical variables. For example, a woman, aged 25 years at onset of disease, with no pyramidal involvement at presentation, and a disease duration of 24 years, has an OR 1.36 (95% CI: 1.10 to 1.68) of remaining benign when compared to a man, aged 35 years at onset of MS, with motor involvement at the outset and a disease duration of 36 years. This may be calculated from the multiple logistic regression. In the regression equation, the logit of the probability of benign status at follow-up is a linear combination of the covariates. The odds may be obtained by exponentiating this. As the logistic regression model with a logit link estimates the log of the odds of progression for different combinations of covariates, it is possible to estimate the OR for different combinations from the fitted model. Thus, the OR for person A versus person B is calculated as the exponential (0.09*Gender–0.0037*Age difference–0.0071*Disease duration difference–0.09*Motor Status), with the coefficients taken from the regression output. Gender was designated either 1 or 0, where one was female and zero was male. Similarly, the presence of motor involvement was designated 1 and absence of motor involvement was 0.
Cause of death
A total of 159 of the 356 patients with MS that were followed up in 2006 had died. The diagnosis of MS was recorded on 75 (47%) of death certificates as having contributed to the patient’s death. Of MS-related deaths, the most common causes were aspiration pneumonia (31 patients, 20%), urinary sepsis (9 patients, 6%), septic pressure sore (6 patients, 4%) and MS exacerbation (4 patients, 3%).
DISCUSSION
A cohort of 47 patients with benign MS in 1985 was followed for a further 21 years, making this the longest published follow-up of a benign cohort. At follow-up, only 7 patients (15%) remained benign. This rate of 15% is lower than that reported in some other longitudinal studies.5 8 This may be due, in part, to the fact that 8 benign patients followed in our study had died of other diseases and an accurate assessment of their MS disability at the time of death cannot be determined. Of the other ten deceased “benign” patients, we do know that they had significant MS-related disability that directly contributed to their deaths. Of the 22 “benign” patients who developed progressive disability, all now require ambulatory assistance and the majority are wheelchair users. Thus, a benign course after 10 years of MS has poor predictive value; significant progression will occur in the majority of these patients in the long term. This was seen to also be true when a more stringent definition of benign MS (EDSS ⩽2 at 10 years) was applied.
The frequency of benign disease in our study was 15% after a total follow-up time of 31 years (95%CI: 6% to 26%). In the Sayao et al study,8 88/169 (52%) patients remained benign after 20 years of follow-up (95%CI: 45% to 60%). In the Pittock et al study,5 34/47 (72%) remained benign after 20 years (95% CI: 59% to 84%). However, we do not feel that these studies are directly comparable to ours as different definitions of benign MS and different follow-up times were used. As our follow-up time of 31 years is longer than that of other groups, a lower proportion of patients would be expected to have benign disease, as the longer patients are followed, the less likely they are to remain benign.
Some authors have suggested13 that MS should only be defined as benign after at least 15 years of disease. From our cohort, 30/53 (57%) patients who had benign MS in 1985 had a disease duration of 15 years or more at that time. At follow-up 21 years later, only 3 patients (10%) were still benign. Thus, in our study, using a definition of 15 years or more for benign disease does not appear to be any more predictive of a long-term benign outcome than the traditional cut off of 10 years.
In addition to different diagnostic criteria for benign MS applied in different studies, different categories of patients were included in different study populations. By using the McDonald and Halliday criteria, we included a subgroup of “suspected MS” patients in our baseline cohort. As not all of this group truly had MS, this may have overestimated the proportion of patients with benign MS in our cohort at baseline. However, as patients with alternative diagnoses were excluded from further analyses at follow-up, we feel that the proportion of patients with benign disease at follow-up is accurate for this cohort.
Clinical characteristics that are weakly predictive of a long-term benign course were identified both in the 47 benign patients in 1985 with a mean disease duration of 18.6 years at baseline, and in the 40 benign patients in 2006 with a mean disease duration of 26.1 years at follow-up. In both groups, a younger age at onset was significantly associated with a benign outcome, a finding that has been observed in previous studies.7 9 14 This may relate to the observed age-dependent progression seen in MS.15 Thus, patients who are younger at onset of MS have a greater possibility of remaining benign for longer. However, it must also be borne in mind that these younger patients must cope with the burden of living with MS for longer than those who are older at disease onset.
A larger proportion of women was found in both benign groups, but this was only significant for the group followed for a mean of 26 years. In patients with benign disease in 1985, we found that the time to second relapse was significantly longer than for those whose course was not benign. However, duration of first remission was not found to be predictive of a benign course when examined in the 2006 cohort of 40 benign patients followed for a mean of 26 years. It is likely that a long first remission has an effect on disability over relatively short periods of follow-up, which is lost with longer surveillance; varying effects of this clinical parameter on outcome have been reported in the literature.16 17
The only presenting symptom that was significantly associated with a benign outcome at both time points was the absence of initial pyramidal tract involvement; this finding has been reported by other groups.9 18 19
A median survival of 43.6 years from onset of MS was found for the MS cohort as a whole, which is comparable with other studies.20 21 When we examined survival in the group of patients who had benign disease at baseline compared to those who did not, we found that the median survival was more than 10 years longer for those who had an initial benign course. This survival advantage was unaffected when corrected for gender; however, the younger age at onset of symptoms in the benign group did play a role. When we corrected for this variable, the difference between the groups was smaller and not significant at the 5% level. Survival in patients with benign MS has not previously been examined. It is interesting to note that although the majority of our benign cohort developed progressive disease with extended follow-up, an initial benign course appears to be nonetheless advantageous in terms of survival.
We acknowledge that our study is not without flaws. First, follow-up was incomplete with 10% of the original MS cohort and 8% of the benign cohort being lost to follow-up. Some 5.5% (24 patients) of the original cohort of “possible MS” cases seen were found to have alternative diagnoses at follow-up, a finding that was not unexpected given the lack of diagnostic imaging available at the time. Second, we do not have data on what proportion of this cohort received immunomodulatory therapies for MS. However, it could be argued that these agents have limited efficacy in preventing disease progression. Patients who had demonstrated a long-term benign course would have been unlikely candidates for disease-modifying therapies (DMTs), when these became available in Ireland in 1996. The mean disease duration in 1996 of the total cohort of 356 patients would have been 20 years; few of these would have been considered for treatment with DMTs. The majority of patients had EDSS assessment by telephone. Although not ideal, telephone assessment was carried out using a validated standardised questionnaire, which has been shown to correlate strongly with clinical assessment, particularly in the upper EDSS range.12 The cohort represents an ageing population and we cannot account for additional co-morbidities that might have had a bearing on the EDSS. Last, this was a hospital-based cohort and, as such, may underrepresent the number of benign patients prevalent in the community; such patients may not attend hospital and thus would not be captured. This could have skewed the number of long-term benign patients towards not benign, underestimating the overall percentage of benign patients.
We have found that, with long-term follow up, the term “benign MS” may be misleading given the inexorable progression to disability observed in the majority of our cohort. As a concept, however, it is not without clinical relevance, as those with an initially more favourable course appear to retain a survival advantage.
REFERENCES
Footnotes
Competing interests: None.