Characterization of disease progression in DMD patients around loss of ambulation (LoA) is needed to inform selection criteria and choice of endpoints in clinical trials including these patients. We analyzed upper limb and pulmonary function in the years immediately before and after LoA (defined as inability to walk 10 meters) using data from 51 boys with DMD from the PRO-DMD-01 natural history study (data provided by CureDuchenne). Included boys were ambulatory at their first study assessment and non-ambulatory at a follow-up assessment. Mean age at LoA was 12.7 years (range: 7.1-18.6 years). Based on longitudinal mixed effects models, average annual declines for before vs. after LoA were 5.6 vs. 10.3 percentage points for forced vital capacity percent predicted (FVC-%p) and 2.3 vs. 3.8 points for Performance of Upper Limb (PUL version 1.2) total score. Neither difference was statistically significant, and there was substantial variability in function across individuals and within individuals over time. The proportion of patients with FVC-%p < 80% and PUL entry score <6 started increasing approximately two years before LoA. About 50% of patients experienced FVC-%p <80%, and ~75% experienced PUL entry score <6 by the time of LoA; all patients reached these milestones within two years after LoA. Among patients with 10 meter walk/run time > 10 seconds, median subsequent time to LoA was ~1 year and all reached LoA within 2 years. Overall, there is heterogeneity in the ordering and magnitude of deficits in upper limb and pulmonary function before and after LoA. Enriching trials for patients with declining upper limb or pulmonary function is achievable without restricting to non-ambulatory patients. Design of composite outcomes, or multi-outcome trial designs, that are sensitive to changes in multiple functional domains before and after LoA should be investigated.