The first example of familial MS was brought to Charcot's attention by Duchenne. Charcot thought that the disease was caused by some sort of toxin or infection, but after reviewing Duchenne's experience with two affected sisters, he concluded that inheritance was a predisposing factor in some cases. It is worth recalling that Charcot's classic clinical descriptions of MS were based on only 34 cases, 25 females and 9 males. In the subsequent century and a quarter, overwhelming evidence for the familial aggregation of MS, and its genetic basis, has appeared. Relatively little is known, however, about the prevalence of subclinical demyelinating lesions in healthy individuals who have close relatives with known MS.1 Identification of subclinical cases could add a wealth of genetic and epidemiological information to family studies by expanding the phenotype of at-risk individuals.

In this issue of the Annals, DeStefano and colleagues have performed a remarkable MRI study of asymptomatic first-degree relatives of MS patients from the relatively homogeneous population of Sardinia, a high-risk area for MS. A mobile MRI scanner was used to study 240 asymptomatic first-degree relatives. They found that 4% of relatives from sporadic MS families, and 10% from multiple affected member families, had focal MRI changes that meet validated MRI criteria for MS. These percentages are remarkably high, even considering that 3 to 4% of first-degree relatives of MS patients also develop MS and that the risk is higher yet in multiple affected member (“familial”) cases.2 The current data suggest that roughly an equal number of additional family members may be affected in a subclinical fashion. The familial data set was somewhat “loaded” by the inclusion of four monozygotic twins (known to have a lifetime risk for MS of approximately 30%). Nonetheless, this study is by far the largest of its kind in addressing the question of the subclinical disease burden in “at risk for MS” families. It is difficult to escape the conclusion that some measures of the MS disease process occur quite commonly in close relatives of MS patients, especially in familial cases.

MS in the Sardinian population is known to have some unique characteristics. In addition to its high prevalence, there is an association with different HLA genes (DR4) and with type 1 diabetes mellitus, features not present in other ethnic groups.3 In northern European populations, the HLA association with MS is with DR2 (DRB1*1501), and although there is some aggregation of other autoimmune disorders in MS cohorts, especially in familial MS, the aggregation does not extend to type 1 diabetes, probably because DR2 is protective against diabetes. Whether the observed high prevalence of MRI changes in MS relatives is unique to Sardinia, or is applicable to other populations, remains to be determined. Because the mean age of the subjects studied by DeStefano was 41 years, most would have passed the period of highest risk for onset of MS. Longitudinal study of these asymptomatic but MRI-positive individuals will also be important to determine to what extent these individuals develop neurological symptoms of MS over subsequent years and decades.

The current results are also consistent with those from earlier autopsy studies4 and more recent MRI5 data suggesting that the prevalence of subclinical MS may be substantial. From a practical perspective, one lesson from this study is a reminder to be alert to the possibility of subclinical MS in clinical practice. In a tertiary referral center, it is not uncommon to encounter patients referred for therapy of “atypical” MS whose diagnosis was based on characteristic MRI findings, but whose true neurological problem is migraine, parkinsonism, Alzheimer's disease, or some other non-MS disorder. In these cases, it can be assumed that the MS is asymptomatic and the presenting problem is unrelated.

Another interesting finding from this study relates to magnetization transfer (MTr) data from “normal appearing white matter” in first-degree relatives of MS subjects. This technique has been previously used to identify abnormal white matter tissue, presumably demyelination, in white matter that appears normal by traditional MR sequences.6 Some studies have reported that widespread MTr changes are present even at the initial onset of clinical symptoms of MS, raising the possibility that MS lesions could develop in white matter tissue that has some undefined pre-existing abnormality.7 In contrast, DeStefano and colleagues found that MTr values were similar in asymptomatic relatives and normal controls. The tentative conclusion is that MTr changes that occur in “normal appearing white matter” of MS cases are likely a consequence of the disease process and do not arise from some pre-existing abnormality.

Clinically isolated syndromes (CISs), representing the first clinical manifestations of MS, are generally viewed as the earliest accessible time point for MS studies. However, most patients with CIS who go on to develop MS already have chronic silent lesions evident by MRI at the time of their initial presentation. Entirely asymptomatic “high risk for MS” individuals, such as the cohort reported here, are of even greater interest. Could one go yet further back in time, commencing observations before the first subclinical MRI lesions appear? One could imagine a longitudinal study of a large cohort of asymptomatic (clinically and by MRI) HLA-identical siblings of MS patients from high-risk families, with the expectation that 10% or more ultimately will develop the disease between ages 20 and 40. Clearly, such a study would have to begin in childhood. In type 1 diabetes mellitus, observational, and even primary prevention, studies of this type have been under way for several years. Diabetes is an easier target than MS, because 30% or more of HLA-identical siblings of diabetics will also develop diabetes before age 10. In MS, epidemiological, immunological, and epigenetic data collected from a prospective study might reveal the earliest events contributing to this disease in a manner that has not been possible with retrospective analyses.


  1. Top of page

Stephen L. Hauser MD*, * San Francisco, CA.