Assess the predictive value of cognitive processing speed (CPS), using the Symbol Digit Modalities Test (SDMT) score, on disability progression in secondary progressive multiple sclerosis (SPMS).

SPMS patients from the Phase 3 EXPAND study (core part [CP] and core+extension part [CP+EP]) were categorized into quartiles of baseline SDMT score (worst-WQ [Q1], intermediate [Q2-Q3], and best-BQ [Q4] quartile). The predictive value of baseline SDMT quartiles for time-to-wheelchair (T2W; i.e., Expanded Disability Status Scale [EDSS] score ≥7) sustained until end of follow-up, or 6-month confirmed disability progression (6mCDP) by EDSS, were assessed at the end of the CP (up to 37-months) and CP+EP (up to 5-years) by Cox regression (adjusted for treatment, age, gender, baseline EDSS, baseline SDMT quartile, and treatment-by-baseline SDMT quartile interaction).

Analyses included 1628/1651 patients (98.6%) randomized in EXPAND (baseline SDMT: WQ, n=435; intermediate, n=808; BQ, n=385). Risk of T2W in the CP was higher for WQ vs BQ  (HR_WQ/BQ=1.31, 95% CI:0.72–2.38; p=0.37) and increased with long-term follow-up (HR_WQ/BQ=1.81; 1.17–2.78; p=0.01). Baseline SDMT was not predictive of 6mCDP. The predictive value of baseline SDMT for T2W in the CP was weaker with siponimod (n=1088; HR_WQ/BQ=1.12, 0.55–2.29; p=0.75) vs placebo (n=540; HR_WQ/BQ=1.86, 0.73–4.78; p=0.19), possibly due to siponimod preventing relatively more T2W events in the WQ. Predictive value of on-study changes will be presented at the congress.

The results support the predictive value of CPS for long-term (up to 5-years) physical disability progression in SPMS and highlight the relevance of monitoring CPS in daily practice to help identify patients at risk of progressing.