Sept 17, 2002
Mayo Clinic Proceedings
Nyquist, Paul A; Cascino, Gregory D; McClelland, Robyn L; Annegers, John F; Rodriguez, Moses
Previous studies have suggested that multiple sclerosis (MS) is a potentially epileptogenic neurologic disorder.1-5 Recently, a standardized incidence ratio for the occurrence of epilepsy was determined in a group of MS patients in Iceland. The risk of epilepsy for patients with MS was 3.0 times greater than that in the general population.1 This study included 188 incidence cases of MS, with 5 incidence cases of epilepsy observed in the MS population over a 25-year period. The standardized incidence ratio for this study had wide confidence intervals (CIs) (95% CI, 0.6-8.8).1
In an Italian cohort of 2353 patients with demyelinating disease, the prevalence of epilepsy, ie, recurrent seizures, was 1.7% (40 patients).4 Among 1459 patients with clinically definite MS, the prevalence was 2.3% (34 patients), which was higher than the expected rate in the general population.4 The prevalence of epilepsy was 3.5% (21 patients) in 599 patients in Finland with clinically definite MS.5
These earlier studies had several potential limitations. Multiple sclerosis is a relatively rare disease, and because of the infrequency of seizures, detecting an increase in seizure activity in such patients may be difficult. Previous series have not always excluded patients with other coexistent symptomatic neurologic disorders that may be associated with seizures. Also, patients who had seizures occurring before the diagnosis of MS may have been included. The rationale for the current study was to evaluate the occurrence of seizure activity in patients with demyelinating disorders in Olmsted County, Minnesota, by using the Rochester Epidemiology Project.6-9
PATIENTS AND METHODS
Criteria for the diagnosis of MS were based on the definition of the Poser committee.10 The first sign or symptom of MS was identified by the description of a clinical episode consistent with demyelinating disease in the patient's medical record or documentation of an appropriate abnormal neurologic finding on examination. Seizure type was classified according to the International League Against Epilepsy classification.11 A consecutive series of patients with incident MS was identified from Rochester, Minnesota, and Olmsted County with use of the Rochester Epidemiology Project computer linkage system.6-9 There were 208 Olmsted County residents who developed MS between 1935 and 1991, and they were considered incidence cases.8
The population for this study was the cohort of patients diagnosed with MS between 1935 and 1991 who were Olmsted County residents at the time of diagnosis. Within the population of incident MS cases, we estimated the incidence rate of unprovoked seizures per 100,000 personyears at risk by using an in-house SAS macro procedure developed for this purpose.12 We assumed that the incidence cases followed a Poisson distribution to estimate SEs for these rates and 2-sided P values for our comparisons.
Three variants of the incidence calculation were performed. In the first estimate, only patients with the first unprovoked seizure occurring after the date of MS diagnosis were counted in the numerator. In a second analysis, any seizures occurring after symptom onset of MS (but not necessarily before MS diagnosis) were considered. A final estimate considered all subjects with seizures and MS regardless of the time of onset. Similarly, different denominators were used in each calculation to express the personyears at risk of the eligible MS patients. In the first analysis, person-years began accumulating at MS diagnosis and continued until the date of first seizure or last follow-up. In the second analysis, person-years began accumulating at MS onset. In the third analysis, patients were at risk at any time point, and we began at date of birth. For all 3 analyses, both an unadjusted rate and a rate age-adjusted to the 1970 US population are presented.
The age-adjusted incidence of seizures in the MS population was compared to the age-adjusted rate for the general Rochester population from 1935 to 1984 and was reported by Hauser et al6 to be 61 per 100,000 person-years. From 1935 to 1991,225 incidence cases of MS were identified in Rochester and Olmsted County. Of the 225 patients, 10 were excluded because information was not available regarding the age at diagnosis of the demyelinating disease or the age at seizure onset, 6 were excluded because they experienced seizures before MS diagnosis, and I was excluded because seizures were related to another symptomatic etiology. Thus, 208 incident MS cases were included in the primary analysis. Of these, 5 patients were identified as having seizures without identifiable cause after the definite diagnosis of MS. The age-adjusted incidence of seizures in this cohort was 61 per 100,000 personyears (95% CI, 7-114). There was no statistically significant difference between the incidence of seizures in this group and the general population (P=.88).
Eight patients were identified who had seizures with no identifiable cause with onset after the first symptoms of MS. The incidence in this population was 80 per 100,000 person-years (95% CI, 24-135). There was no statistically significant difference between the incidence of seizures in this group and the general population (P=.70). The incidence of seizure activity with no identified cause at any time in the life of the patients with the diagnosis of MS was 82 per 100,000 person-years (95% CI, 41-158). Again, this was not significantly different from the general population of Olmsted County (P=.36).
The diagnosis of MS requires at least 2 neurologic deficits occurring at different sites in the central nervous system at different times, ie, dissemination in time and space.10 Often the clinical symptoms of the illness predate the definitive diagnosis of MS by several years. A populationbased study in Olmsted County documented the neurologic symptoms and signs related to MS.9 A chronic, progressive demyelinating disease may be associated with a significant impairment in the individual's quality of life.9 One important adverse effect of MS may be the emergence of seizure activity.1-5
The results of the current study provide evidence that a demyelinating disorder is not a significant risk factor for seizure activity. The study design used a strict entrance criterion for inclusion of patients with MS. This limited the recall bias inherent in identifying the time of first symptoms or signs of the disease. Patients with other potential causes of seizure activity were excluded. The conclusion of the current series did not confirm the previous Icelandic study that indicated an increased risk of seizure activity in patients with MS.1 The results of both investigations were based on a relatively small number of incident cases with seizures. With our total accumulated person-years of follow-up, we had approximately an 80% power to detect a 3-fold increase in age-adjusted incidence rates, the magnitude of the effect reported by the Icelandic study.
The potential limitations associated with populationbased epidemiological
studies include the difficulties associated with case ascertainment and
identification of associated clinical factors. Determining type and frequency
of seizures may not be possible in a retrospective study. Also, the diagnostic
evaluation of a patient with seizures has been altered substantially during
the period of this study. Finally, the small number of patients in the
current series suggests that this is a preliminary investigation. A larger
patient group may be necessary to confirm any lack of association between
MS and seizures.
(C) 2002 Mayo Clinic Proceedings