P.65 Consistency of changes in percent-predicted forced vital capacity between real-world data and trial placebo arms in ambulatory Duchenne muscular dystrophy

Abstract

Use of real-world and natural history data (RWD/NHD) as external controls for pulmonary outcomes is of interest for drug development in ambulatory Duchenne muscular dystrophy (DMD). However, differences across patient populations and RWD/NHD settings raise appropriate concerns about risk of bias. To assess this risk, we compared change in pulmonary function, as measured by percent-predicted forced vital capacity (FVC%p) between RWD/NHD sources and trial placebo arms in ambulatory boys with DMD. RWD/NHD data came from PRO-DMD-01, UZ Leuven, the North Star UK Clinical Network and Cincinnati Children's Hospital Medical Center. Placebo arm data came from phase 3 trials of tadalafil and drisapersen, and one phase 2 trial of drisapersen. PRO-DMD-01 and drisapersen trial data were provided by CureDuchenne. Change in FVC%p over 48-week intervals was studied in ambulatory, steroid-treated boys aged ≥5 years with baseline FVC%p > 50%. Primary analyses included 409 48-week intervals (263 patients) from RWD/NHD, and 161 intervals (161 patients) from placebo arms. At baseline, average ages in these groups were 9.5 and 8.8 years, respectively, and averages for FVC%p were 97.2% and 103.6%. Mean 48-week change in FVC%p was 2.9% greater in in RWD/NHD vs. placebo arms (95% confidence interval [CI]: -1.0%, 6.7%; p=0.14) in unadjusted analyses. After adjusting for baseline status (age, FVC%p, steroid type, height, weight, body mass index, North Star Ambulatory Assessment total score, and 10-meter walk run and rise from supine velocities) the difference in FVC%p between RWD/NHD and placebo arms decreased to -0.1% (95% CI: -3.6, 3.5; p=0.97). Similarly high consistency was observed in sensitivity analyses adjusting for fewer baseline characteristics. These findings are encouraging for use of external controls for FVC%p outcomes in this population. Additional research is needed to assess consistency of FVC%p over longer-time frames, and among older and non-ambulatory boys.