Background/Purpose: There is a dearth of information on the incidence rate of myasthenia gravis (MG) in the US and specifically among rheumatoid arthritis (RA) patients.  RA patients have been shown to be at risk for other autoimmune diseases so we sought to describe the rate of MG among RA patients, with and without treatment.

Methods: We conducted a retrospective cohort study using an administrative claims database. There were 8 study cohorts: General population (GP); incident RA population regardless of treatment; and 6 RA sub-cohorts defined by treatment exposure. Inclusion criteria included a minimum of 2 years of continuous enrollment with pharmacy coverage between 01/01/2005 – 12/31/2011. To ensure incident RA, patients could not have any RA claims during the first 12 months of enrollment. RA patients were followed forward and assigned to exposure sub-cohorts: Untreated (RA-U), 3 cohorts of different disease modifying anti-rheumatic drugs (RA-DMARDS1, RA-DMARDS2, RA-DMARDS3); TNF-inhibitors (RA-TNF); and other biologic therapies (RA-OB).  Patients could be in ≥ 1 sub-cohort. MG diagnosis was captured using diagnostic and procedure claims.  Prevalent MG cases in the GP and RA cohorts were identified during the first 12 months of enrollment. Follow-up for GP began on day 366 following enrollment; for RA and RA-U it began on date they met the criterion for RA diagnosis, and for RA sub-cohorts it began on medication start date.  Follow-up continued until MG diagnosis, disenrollment from the database, end of study, and for the sub-cohorts, 90 days after discontinuation or switch in treatment.  Incidence rates and 95% confidence intervals per 100,000 person years were estimated for MG, excluding prevalent cases. Incidence and prevalence estimates were age- and sex- standardized to the GP.

Results: The mean age and sex distribution were 37.1 years and 52% female for GP and 56.3 years and 74% female for RA. Gender and age distributions in the RA sub-cohorts were similar to RA. Prevalence of MG was 24.3 per 100,000 in the GP cohort and 86.8 per 100,000 in the RA cohort while the incidence was 10.4 per 100,000 and 35.8 per 100,000 in the GP and RA cohorts, respectively.  MG incidence was highest in the RA DMARD3 (azathioprine, cyclosporine, cyclophosphamide, gold salts or penicillamine) sub-cohort. Due to the small number of MG cases, there was no discernable incidence pattern in the other sub-cohorts. Sensitivity analyses allowing for different follow-up times or using different algorithms to define MG showed little difference in the results.

Conclusion: As with other autoimmune diseases, subjects with RA appear to be at a higher risk from MG than the general population. We found both the prevalence and incidence of MG to be 3-4 times higher among RA patients compared to GP, however, with the exception of the RA DMARD-3 cohort, treatment did not appear to influence the incidence of MG following RA diagnosis.