(A-326) A Collaborative Approach to Evaluate Risk Factors for a Rare Adverse Event Among Individuals with Multiple Sclerosis Exposed to Sphingosine 1-phosphate Receptor Modulators

Background: Sphingosine 1-phosphate receptor modulators (S1PRM) have been associated with an increased risk of progressive multifocal leukoencephalopathy (PML) in individuals with multiple sclerosis (MS). Understanding risk factors for PML in this population is important to inform patient safety. Due to the rarity of the event, it is difficult to evaluate specific risk factors for PML, especially by individual S1PRM. We describe a collaborative approach implemented to address this challenge.

Objectives: To evaluate potential risk factors for PML among individuals with MS exposed to S1PRM, an observational study will be conducted by a collaboration of marketing authorization holders (MAH) to collectively assess multiple S1PRM.

Methods: PML cases identified via routine pharmacovigilance (PV) will be adjudicated and described in terms of patient age, duration of exposure to S1PRM, prior immunosuppressant use, lymphopenia, and anti-JCV antibody status. If a minimum number of confirmed PML cases have accrued among S1PRM marketed by Bristol Myers Squibb, Cycle Pharmaceuticals and Novartis (i.e., ozanimod, fingolimod, siponimod) by the end of the study period, a nested case-control study will be conducted using the confirmed PML cases identified through PV and an external control to evaluate potential risk factors.

Estimates of person-years of exposure needed to accrue over a 15-year period among individuals with MS exposed to S1PRM were calculated based on a range of potential incidence rates (IR) of PML (i.e., 3.0-9.0 per 100,000 person-years) and other assumptions.

Results: In a nested case-control study with 20 PML cases, where each PML case is matched to 100 controls, the study would have 80% power to detect an odds ratio (OR) ≥5 when the risk factor probability is ≥10% in the control group. As the size of the PML cohort increases, the detectable OR threshold decreases. However, unless the PML cohort size is >100 and the risk factor probability is >20% in the control group, a moderate OR (i.e., OR < 2) cannot be detected with 80% power.

After applying the range of potential IR and assumptions previously described, it was estimated that 15,000 to 45,000 person-years of exposure per year to S1PRM is required to accrue 20 PML cases over a 15-year period; an estimated 74,000 to 223,000 person-years of exposure per year is required to accrue 100 PML cases.

Conclusions: A collaboration of MAH to assess multiple S1PRM collectively is expected to increase the person-years of exposure to allow capture of more PML cases, increasing the likelihood of having an appropriately powered nested case-control study. Establishing similar collaborations could be considered for future studies assessing rare adverse events among products in the same class.