3025 Ruptured cerebral aneurysm inpatient outcome prediction for discharge planning with machine learning: a derivation study

Abstract

Objective Explore the effects of baseline characteristics on long-term safety/efficacy of avalglucosidase alfa (AVA) in subgroups of patients with late-onset Pompe disease enrolled in COMET (Phase 3; NCT02782741).Methods COMET randomised treatment-nave patients to AVA (n=51) or alglucosidase alfa (ALG; n=49).After a 49-week primary analysis period, all patients from the AVA-arm continued and 44 patients in the ALG-arm switched to AVA (extended treatment period); total duration 145 Weeks.Subgroups and analyses included: 1) the impact of baseline age (!18-<45y, !45y) on the change in 6-minute walk test (6MWT) distance and upright forced vital capacity (FVC) % predicted, 2) the impact of baseline 6MWT (<403.5, !403.5m) on the change in FVC % predicted, and 3) the impact of baseline FVC % predicted (<55, !55%) on the change in 6MWT distance.Mean estimates (95% CI) were calculated from linear mixed effects models stratified by treatment group.Results Overall, the change from baseline to Week 145 was stable or improved for all subgroups analysed for the different outcomes.From baseline to Week 145: younger patients in the AVA-arm had a significant improvement in 6MWT distance (9.8m [4.4,15.1];p=0.0004) and AVA-arm patients with baseline FVC !55% had a significant improvement in 6MWT distance (6.8m [2.4,11.2];p=0.0026).Changes over time remained stable in all other subgroups with nonsignificant p values.Conclusions These data indicate that the positive changes seen during treatment with AVA are not driven by any subgroup and demonstrate that AVA is effective in patients with varying baseline characteristics.

Ähnliche Arbeiten