The natural history of multiple sclerosis: a geographically based study 10: relapses and long-term disability

Abstract

The relationship of relapses to long-term disability in multiple sclerosis is uncertain. Relapse reduction is a common therapeutic target but clinical trials have shown dissociation between relapse suppression and disability accumulation. We investigated relationships between relapses and disability progression for outcomes of requiring assistance to walk, being bedridden and dying from multiple sclerosis [Disability Status Scale 6, 8, 10] by analysing 28 000 patient-years of evolution in 806-bout onset patients from the London Ontario natural history cohort. Having previously shown no effect of relapse frequency among progressive multiple sclerosis subtypes, here we examined these measures in the pre-progressive or relapsing-remitting phase. Survival was compared among groups stratified by (i) early relapses--number of attacks during the first 2 years of multiple sclerosis; (ii) length of first inter-attack interval; (iii) interval between onset and Disability Status Scale 3 (moderate disability); (iv) number of attacks from the third year of disease up to onset of progression; and (v) during the entire relapsing-remitting phase. Early clinical features can predict hard disability outcomes. Frequent relapses in the first 2 years and shorter first inter-attack intervals predicted shorter times to reach hard disability endpoints. Attack frequencies, in the first 2 years, of 1 versus >or=3, gave differences of 7.6, 12.8 and 20.3 years in times from disease onset to Disability Status Scale 6, 8 and 10, respectively. Time to Disability Status Scale 3 highly and independently predicted time to Disability Status Scale 6, 8 and 10. In contrast, neither total number of relapsing-remitting phase attacks nor of relapses experienced during the relapsing-remitting phase after the second year up to onset of progression showed a deleterious effect on times from disease onset, from progression onset and from Disability Status Scale 3 to these hard endpoints. The failure of a regulatory mechanism tied to neurodegeneration is suggested. Relapse frequency beyond Year 2 does not appear to predict the key outcome of secondary progression or times to Disability Status Scale 6, 8 or 10, highlighting two distinct disease phases related to late outcome. These appear to be separated by a watershed within the relapsing-remitting phase, just a few years after clinical onset. Higher early relapse frequencies and shorter first inter-attack intervals herald more rapid deterioration via interaction with the neurodegeneration characterizing secondary progression. They increase the probability of its occurrence, its latency and influence--to a lesser degree--its slope. The prevention or delay of the progressive phase of the disease is implicated as a key therapeutic target in relapsing-remitting patients.

Figure 1

Kaplan–Meier survival curves of time from disease onset to DSS 6 in patients grouped according to (A) total number of relapses in Year 1 and Year 2 (1 relapse; 2 relapses; ≥3 relapses). The estimated mean time from disease onset to DSS 6 was significantly shorter in those patients with a larger number of attacks in Years 1 and 2: 1 relapse group = 22.7 mean years, 2 relapses group = 18.7 mean years, ≥3 relapses group = 15.1 mean years. (B) First inter-attack interval (0–2 years; 3–5 years; ≥6 years). The estimated mean time from disease onset to DSS 6 was significantly shorter in those patients with a shorter interval between the first and the second attack. 0–2 years interval group = 18.2 mean years, 3–5 years interval group = 21.0 mean years, ≥6 years interval group=25.9 mean years. (C) Time from onset to moderate disability (DSS 3) (0–2 years; 3–7 years; ≥8 years). The estimated mean time from DSS 3 to DSS 6 was significantly shorter in those patients with a shorter interval between disease onset and moderate disability (DSS 3). 0–2 years interval group = 5.4 mean years, 3–7 years interval group = 7.4 mean years, ≥8 years interval group = 8.7 mean years. (D) total number of relapses before the onset of progression (1–2 relapses; 3–4 relapses; ≥6 relapses). The estimated mean times from disease onset to DSS 6 were remarkably similar in all three groups. 1–2 relapses group = 15.6 mean years, 3–4 relapses group = 15.7 mean years, ≥5 relapses group = 15.9 mean years.

Figure 2

Cox regression univariate analysis. Risk (y-axis) of attaining DSS 6 (black), 8 (dark grey) and 10 (light grey) from disease onset according to number of relapses experienced from Year 3 up to onset of progression. Hazard ratios are obtained through comparison with zero relapses. The y-axis expresses the variation of the hazard ratio according to the number of Year 3-secondary progression relapses (x-axis). A larger number of attacks was significantly related to a lower risk and a shorter time to attain the disability endpoints from disease onset. Year 3-secondary progression is period from end of Year 2 to onset of the progressive phase (secondary progression).

Figure 3

Multiple survival Cox regression analysis. Risk (y-axis) of attaining DSS 6 from disease onset according to the combined effect of number of attacks in Years 1 and 2 (x-axis) and number of attacks from Year 3 up to onset of progression (0–1–2–3) in patients with secondary progressive multiple sclerosis. Hazard ratios are obtained through comparison with zero attacks. The y-axis shows the variation of the hazard ratio obtained by the combined effect of Years 1 and 2 relapses (x-axis) and Year 3-secondary progression relapses (each column) on the time to attain DSS 6 from disease onset. Patients at higher risk of disability have larger number of Years 1 and 2 relapses and smaller number of Year 3-secondary progression relapses. A larger number of attacks after Year 2 reduces the combined risk of attaining disability endpoints.