EG02
Temporal Changes in the Age at Multiple Sclerosis Onset: Importance of Controlling for Equal Observation Time
Previous studies have examined whether changes in the age at multiple sclerosis (MS) onset have occurred over time, but findings have been inconsistent. If a substantial change has occurred, it could have important implications for health care planning and may indicate shifts in environmental etiological factors. However, it is challenging to assess these trends; for example, when comparing recent birth cohorts to earlier birth cohorts, spurious findings may result if insufficient time is allowed for individuals from the recent birth cohort to develop MS.
Objectives:
We investigated temporal trends in age at onset in MS patients from three Canadian provinces, and assessed the impact of controlling for equal observation time between birth cohorts.
Methods:
Data from the MS clinic databases in British Columbia (BC), Manitoba (MB) and Nova Scotia (NS) were accessed. Patients with definite MS (Poser or McDonald criteria) were grouped by birth year (1941-1980) into 5-year blocks. Trends in the age at onset of MS were assessed using linear regression adjusted for sex. Initially all MS patients were included (the broad cohort), then a restricted cohort was created to allow comparable observation times for each birth cohort, in which all patients had reached at least age 40 (80thpercentile for onset age) by study end (12/31/2007) and had MS onset by age 40.
Results:
A total of 9459 MS patients were included (BC=5423, MB=1419, and NS=2617). In the broad cohort, there was an apparent steep decline in the age at onset: from 37.0±10.8 years for the earliest birth group (1941-1945) to 28.0±6.4 years for the 1966-1970 birth group and 23.3±4.2 for the most recent birth group (1976-1980), with a mean average decrease of 2.0 years in age at onset between each consecutive birth group (p<0.001). In the restricted cohort (n=6003), a statistically significant decrease in the mean age at onset was still evident (30.0±6.7 years for the earliest birth group and 28.3±6.0 years for the 1966-1970 birth group; p<0.001), however, the decrease in average age was considerably smaller at 0.3 years between each consecutive birth group.
Conclusions:
These time restricted analyses, using three large MS clinic databases, suggest a small decrease in onset age over four decades of birth cohorts. Findings also demonstrate how apparent changes in the age at MS symptom onset between birth groups over time can be significantly inflated without due consideration of comparable observation time.