A 30-year follow-up of patients with clinically isolated syndrome (CIS) reports that MRI findings (lesion location, lesion number) at baseline and at one year were the most robust prognostic factors of long-term outcomes (Chung et al. Ann Neurol 2020;87:63-74). Results at 20 years were previously published as Fisniku et al. Brain 2008;131:808-817.

The prospective study enrolled 132 CIS patients at two centres during the period 1984-1987. For the present analysis, subjects were classified as CIS or MS (McDonald 2010 criteria); MS patients were further classified as RRMS or SPMS and categorized as nondisabling (EDSS <3.5) or disabling (EDSS >3.5). PASAT and BICAMS were used to evaluate cognitive impairment at 30 years. Data were obtained for 120 patients with a mean follow-up of 30.9 years. Eleven patients (14%) received treatment with a disease-modifying therapy at some point; the earliest time to DMT initiation was 10 years after diagnosis (when DMTs became available). There were 29 deaths: 19 with MS and 10 with a last known diagnosis of CIS. Mean age of survivors at follow-up was 61.6 years.

At 30 years, 80 of 120 patients (67%) were classified as MS (including deceased) and 30 remained as CIS (not including 10 patients who died with CIS at last evaluation). In the MS cohort (n=80), 16 of 19 deaths were attributed to complications of MS. In the group of 61 patients alive with MS, 35 had RRMS (including 6 patients with CIS according to pre-2010 criteria) and 26 had SPMS.

EDSS score >3.5 was found in 3 of 35 (8.5%) RRMS patients and 26 of 26 (100%) SPMS patients. In the group of RRMS patients with EDSS <3.5 (n=32), all remained employed or had retired at pension age. Age-adjusted cognitive scores did not differ significantly from those of CIS patients. In the RRMS group with EDSS >3.5 versus CIS, cognitive scores were 24% worse on SDMT (mean difference 12.90), 23% worse on PASAT (mean difference 10.68), and 21% worse on BVMTR (Brief Visuospatial Memory Test–Revised; mean difference 5.44).

Prognostic factors
The most robust predictor of worse outcomes was the presence of >1 infratentorial (IT) lesions at baseline: 19 of 22 patients (86%) with IT lesions progressed to EDSS >3.5 versus 25 of 74 (34%) without IT lesions (odds ratio 12.4). An IT lesion at baseline was also predictive of SPMS (OR 20.3). The mortality risk was 3.5-fold higher in patients with versus without a baseline IT lesion. The baseline lesion count was also a significant predictor of EDSS >3.5 at 30 years (OR 1.84). Patients presenting with a brainstem CIS had a higher risk of EDSS change at five years and of MS-related mortality compared to those presenting with optic neuritis or transverse myelitis (hazard ratio 2.87). It was noteworthy that IT lesions were more common in patients with brainstem CIS versus transverse myelitis or optic neuritis (41% vs. 19% and 11%).

At 1-year follow-up, an IT lesion (OR 11.1) and deep white-matter lesions (OR 10.7) significantly predicted long-term disability. Deep WM lesions were defined as supratentorial lesions that were not juxtacortical or periventricular.

Patient sex, age at onset and disease duration at diagnosis were not prognostic of 30-year outcomes. In contrast to other studies, baseline EDSS score was not prognostic of long-term outcomes. However, an EDSS score >2.5 at five years was prognostic of SPMS (OR 9.43). The most important metric appeared to be an early 2-point worsening of disability score in the first five years, which more strongly predicted SPMS onset (OR 40.7) than the actual EDSS score. However, the authors emphasized that MRI measures were better predictors of long-term disability than EDSS.

The study raises questions about the value of the 2010 redefinition of some CIS patients as early MS. In this cohort, six patients were diagnosed with MS based on MRI rather than clinical findings, all were untreated and none developed serious long-term disability.

The authors noted that in this cohort – likely to be the last natural history study of untreated patients – there were three key outcomes: an RRMS group with little disability; an SPMS group with significant disability; and a group at risk of early MS-related mortality. For the full cohort of untreated CIS patients (n=120), the 30-year outcomes were CIS (25%), RRMS with minimal disability (27%), RRMS with moderate disability (3%), SPMS (22%), or early death (24%). For the subgroup who converted to RRMS, the largest proportion had minimal disability (40%), but 33% developed SPMS and 24% died of MS-related complications. A recent epidemiologic study in B.C. reported a similarly high mortality rate in MS patients (21% at 27 years), with MS cited as a contributing cause in 77.9% of deaths (Kingwell et al. Neuroepidemiology 2019; epublished December 18, 2019).