Natural history of MS

In an era of large computerized databases and specialty clinics it might be assumed that the natural history of chronic diseases would be well-known. The fact that it is not has much to do with ascertainment, endurance, and perseverance if not investigator longevity. The problems are not trivial in the case of a disease like multiple sclerosis (MS) which runs a clinical course over 3–4 decades or even longer. Patients move, lose interest, they may change their names, refuse to participate and in the last decade more and more frequently may take a variety of treatments. Investigators have to be persistent and systematic over much of a professional lifetime. Ideally, studies are carried out by investigators in locations where there is the least outward migration. Occasional itinerant consumers of their weather and culture have no difficulty in understanding why the French so qualify. Debouverie et al. [1] report here a large longitudinally studied cohort which highlights some previously made observations. The data from Nancy are remarkably like those from Lyons and Rennes reported by Confavreux [2] and Edan (personal communication) Similar conclusions are made notably emphasizing the anamnestic nature of progression once an EDSS of 4 is reached. These results confirm data from Lyons [2] and London [3,4] on the progression issue. However, there are differences which need to be resolved. Conceptually, it seems plausible if not certain that the clinical appearance of progression is preceded by some years by the likely biologic substrate of progression viz. axonal loss as evident pathologically. There is anecdotal evidence that axonal loss must reach a threshold in MS and ALS before clinical symptoms develop. In Parkinson s disease it has been estimated that perhaps 3=4 of substantia nigra neurons must die before symptoms appear. There are no perfect analogies to draw on but the paradigm of plasticity and duplicative reserve of nervous function is so well documented in the experimental and clinical world that making an exception for MS progression will take some convincing. There are additional issues relating to the pace of disability accumulation. The French data, internally quite consistent among Lyons, Rennes and Nancy, show striking differences from the data from London Ontario [4] and Gothenburg [5]. The time to EDSS 6 (cane requirement) is shorter, being 15 years in the Canadian and Swedish series compared with 23 years in the French studies. One would like to be entirely sure that there is no differential underascertainment of cases with the most severe disability levels as they progress to institutional care but this is ordinarily excluded by the performance of a concomitant formalized prevalence study as was the case in the London cohort. The rigour with which natural history studies pursue those lost to follow-up should be no less than expected for clinical trials seeking to establish efficacy. It may well be the case that outcome in MS does differ by geography. In spite of expectations, both the prevalence and sex ratios differ considerably by latitude. These now well-established findings have the possible implication that the additional cases which are latitude-determined account for this difference. However, even this is complicated by the increasing prevalence of MS manifested in more northerly countries. An increasing sex ratio accounts for as much as a threefold increase in overall prevalence in Canada in the last three generations [6]. Do the newer cases who would not have gotten MS had they been born several decades ago behave the same clinically? This may not be easy to determine but Debouverie et al. [1] do note a difference between males and females. Some of this could be due to the relative overrepresentation of males among the primary progressive cases and such a comparison might be best performed for RRMS and PPMS separately. However, as the increase is peculiar to females or largely so, the more benign courses reported in more recent studies could result from the more recently affected females. Even in the older studies females were overrepresented among the benign cases. These questions are amenable to study in existing datasets. There is more than academic interest in these questions. Indeed the inability to validate the measures of disability widely used in clinical trials [7] has led to increased focus on historical control data for the longer term disability measures. These have been beyond the scope of clinical trials to date in MS. The wish to compare treated groups to natural history data, which seems to be the only way to address the key questions of drug effectiveness, will have to go unrequited until there is a resolution of these issues. If MS is getting more benign, a critically important type I error could be made by comparing current treatment outcomes to older control groups. If there are real geographic differences in outcome, matching of treatment and control groups may be required for both time and place. There is one potential solution to the dilemma and it would hold only if the male rate has truly remained the same. Outcomes for males alone could be compared and the finding of no change would provide a touch-