Predicting multi-year changes in rise-from-floor (RFF) velocity in Duchenne muscular dystrophy (DMD)

Topic:

Other

Poster Number: 159 M

Author(s):

Francesco Muntoni, MD, Dubowitz Neuromuscular Centre, UCL and Great Ormond Street Hospital Trust, London, UK, Nathalie Goemans, MD, PhD, Child Neurology, University Hospitals Leuven, Leuven, Belgium, James Signorovitch, PhD, Analysis Group, Inc., Boston, Massachusetts, USA , Nicolae Done, PhD, Analysis Group, Boston, MA, USA, Jess Marden, BA, MPH, ScD, Analysis Group, Inc., Boston, Massachusetts, USA , Eugenio Mercuri, MD, PhD, Paediatric Neurology, Catholic University, Rome, Italy, Centro Clinico Nemo, IRCCS, Rome, Italy, Laurent Servais, MD PhD, University of Oxford, Shuang Wang, PhD, Analysis Group, Boston, MA, USA, Hanane Akbarnejad, MS, Analysis Group, Boston, MA, USA, Anyu Zhu, MS, Analysis Group, Boston, MA, USA, Craig McDonald, MD, Department of Physical Medicine and Rehabilitation, University of California Davis, CA, USA, Susan J Ward, PhD, cTAP, The cTAP Collaborative Trajectory Analysis Project, The Collaborative Trajectory Analysis Project, Cambridge, MA, USA

Prediction of functional trajectories in DMD is important for the design and interpretation of clinical trials, and for the contextualization of longer-term treatment outcomes, beyond periods studied with placebo controls. Timed rise-from-floor (RFF) is a key functional measure in DMD that tracks early disease progression. We developed a prognostic model for multi-year change in RFF velocity using harmonized data from natural history databases (UZ Leuven, PRO-DMD-01 data provided by CureDuchenne, North Star UK, and iMDEX) and clinical trial placebo arms (Lilly tadalafil trial, PTC ataluren trial, and three Biomarin drisapersen trials). The analysis included 618 ambulatory boys (2,572 post-baseline visits) meeting criteria for steroid use (≥3 months), ambulatory function, and availability of candidate prognostic factors, with ≤3.5 years of follow-up (mean 1.8, standard deviation [SD] 1.0 years). Change from baseline in RFF velocity was modeled as a function of baseline candidate predictors using generalized estimating equations. Prognostic performance was evaluated using 5-fold cross-validation and the proportion of explained variation (R-squared), considering different predictor sets and different trajectory shapes over time. Mean baseline age was 8.2 years (range 3.4-15.5) and mean RFF velocity was 0.22 s⁻¹ (SD 0.11 s⁻¹). RFF velocity declined steadily over time, with mean decreases from baseline of 0.030 s⁻¹ at year 1 and 0.144 s⁻¹ at year 3.5. A quadratic-time model incorporating baseline age explained 18% of outcome variation. Further adding functional measures (baseline RFF and timed 10-meter walk/run velocities, North Star Ambulatory Assessment [NSAA] total score) and steroid treatment characteristics (type, regimen, duration) increased R-squared to 31%. The model distinguished rapid decliners (lowest quartile) from patients with stable or improving trajectories (highest quartile). A prognostic model incorporating age, steroid use, and functional measures provides robust prediction of multi-year RFF decline and can inform trial design and contextualization of long-term treatment effects in early ambulatory DMD.