ULTIPLE SCLEROSIS (MS) IS a chronic disease that often affects people in the prime of their lives. A key feature of MS is clinical progression of the disease over time manifested by the accumulation of disability. Interferon beta drugs are the most widely prescribed disease-modifying drugs (DMDs) approved by the US Food and Drug Administration for the treatment of relapsing-onset MS, the most common MS disease course. Although a substantial reduction in brain lesion development, as evidenced by magnetic resonance imaging (MRI), 1 and a one-third relative reduction in relapse frequency were demonstrated in the pivotal clinical trials of interferon beta for relapsing-remitting MS, 2 there is a lack of well-controlled longitudinal studies investigating the effect of in-terferon beta on disability progression. Typically, drug efficacy (as established through randomized clinical trials conducted under optimal conditions) is greater than drug effectiveness (as measured in "real-world" settings). 3 Patients participating in clinical trials tend to be highly selected in For editorial comment see p 290.

Objective:The relationship between relapses and long-term disability in multiple sclerosis (MS) remains to be fully elucidated. Current literature is conflicting and focused on early relapses. We investigated the effects of relapses at different stages on disability progression. Methods:We conducted a retrospective review of 2,477 patients with definite relapsing-onset MS followed until July 2003 in British Columbia, Canada. Time-dependent Cox proportional hazards models examined the effect of relapses at different time periods (0 -5; Ͼ5-10; Ͼ10 years postonset) on time to cane (Expanded Disability Status Scale [EDSS]) and secondary progressive MS (SPMS). Findings were derived from hazard ratios with 95% confidence intervals (CIs), adjusted for sex, onset age, and symptoms.Results: Mean follow-up was 20.6 years; 11,722 postonset relapses were recorded. An early relapse (within 5 years postonset) was associated with an increased hazard in disease progression over the short term, by 48%; 95% CI 37%-60% for EDSS 6 and 29%; 95% CI 20%-38% for SPMS. However, this substantially lessened to 10%; 95% CI 4%-16% (EDSS 6) and 2%; 95% CI Ϫ2%-7% (SPMS) after 10 years postonset. The impact of later relapses (Ͼ5-10 years postonset) also lessened over time. Effects were modulated by age, impact being greatest in younger (Ͻ25 years at onset) and least in older (Ն35 years) patients where relapses beyond 5-years postonset typically failed to reach significance. Relapses during SPMS had no measurable impact on time to EDSS 6 from SPMS. Conclusion:Relapses within the first 5 years of disease impacted on disease progression over the short term. However, the long-term impact was minimal, either for early or later relapses. Long-term disease progression was least affected by relapses in patients with an extended disease duration (Ͼ10 years) or already in the secondary progressive phase. Neurology Multiple sclerosis (MS) is a common cause of neurologic disability in young adults.1 The majority (85%) present with a relapsing-remitting (RR) course, with many later entering the secondary progressive (SP) phase. Disability accrual in MS is said to occur in 2 situations: incomplete recovery from a relapse or deterioration of functional ability outside of a relapse, synonymous with the progressive phase. 1The mechanisms involved in either relapses or irreversible disability are not fully understood. Relapses are associated with inflammation and demyelination; irreversible disability with axonal damage. However, destruction of axons and the myelin-oligodendrocyte complex occurs in early and late disease.

Multiple sclerosis (MS) has entered an era of immunomodulatory drug treatment, the impact of which on long-term disease progression remains controversial. The increasing use of these therapies has intensified our need to understand the true natural history of MS. The MS community is poised to establish whether the immunomodulatory drugs exhibit long-term benefits, with a suitable untreated natural history cohort likely the most practical and ethical comparator group. Thus, a thorough understanding of the natural history of MS is fundamental. In this review, we highlight recent advances in MS natural history over the last 5 years, with a focus on long-term population-based cohorts and factors associated with disease progression. Survival in MS has increased and longer times to irreversible disability have been reported in contemporary studies, indicating a slower accumulation of disability. Wide variation in the MS disease trajectory is evident within and between natural history studies, reflecting both methodologic considerations related to data collection and heterogeneity of disease activity. Recent publications have indicated that a younger age at disease onset is no longer indicative of a favorable outcome and further evidence supports the dissociation between relapses and long-term disability, although windows of opportunity may exist for some patients. We are now perhaps faced with our last chance to examine the true natural history of MS, so whether the reader is a practicing physician, health care provider, or researcher, or engaged in the pharmaceutical industry or in clinical trial design, recent advances in our understanding of the natural history of MS are of key significance.

Relapse rates were age- and time-dependent. Our observations have clinical implications: 1) any drug able to modify relapse rates has the greatest potential for a population impact in patients <40 years old and within the first few demi-decades of disease; 2) continuation of drug beyond these times may be of limited value; 3) long-term follow-up studies must consider that relapse rates probably decline at different rates over time according to the patient's onset age; 4) a relapse-quiescent period in MS is not uncommon.

The RR phase lasted on average almost two decades, being shorter for males and those older at onset of MS. However, neither were necessarily unfavorable predictors as those older at onset were typically older at SPMS and eventually males and females reached EDSS 8 at around the same age. A longer RR phase was a favorable predictor of disease progression in SPMS. Furthermore, reaching SPMS at an older age or lower EDSS did not necessarily confer a worse outcome.

Some of the longest MS survival times are reported here but the risk of death was still greater than in the age, sex and calendar year matched general population. No evidence of increased survival over time was found when improved survival in the general population was taken into consideration.

No abstract

Abstract:The authors propose two composite likelihood estimation procedures for multivariate models with regressionlunivariate and dependence parameters. One is a two-stage method based on both univariate and bivariate margins. The other estimates all the parameters simultaneously based on bivariate margins. For some special cases, the authors compare their asymptotic efficiencies with the maximum likelihood method. The performance of the two methods is reasonable, except that the first procedure is inefficient for the regression parameters under strong dependence. The second approach is generally better for the regression parameters, but less efficient for the dependence parameters under weak dependence.Estimation par vraisemblance composite pour I'analyse de donnees multivariees Rksumk : Les auteurs proposent deux procCdures d'estimation par vraisemblance composite applicables 9 des modeles multivariCs comportant des parametres de dkpendance et de kgression univarihs. La premitre mCthode consiste 9 estimer d'abord les marges univariees puis les marges bivarikes. Dans la seconde, tous les parametres sont estimCs simultanCment au moyen des marges bivarikes. Dans certains cas particuliers, les auteun comparent l'efficacitk asymptotique des deux approches 9 celle de la vraisemblance maximale. La performance des deux mCthodes s'av8re raisonnable, bien que la premiere soit inefficace pour l'estimation des paramttres de rkgression en cas de dkpendance forte. La deuxikme approche est gknkralement meilleure pour les parametres de dgression mais moins efficace pour l'estimation des parametres de dCpendance lorsque cette demiere est faible.