Abstract Background Chronic respiratory insufficiency from progressive muscle weakness causes morbidity and mortality in late-onset Pompe disease (LOPD). Previous Pompe Registry (NCT00231400) analyses for ≤ 5 years’ alglucosidase alfa treatment showed a single linear time trend of stable forced vital capacity (FVC) % predicted. Methods To assess longer term Pompe Registry data, piecewise linear mixed model regression analyses estimated FVC% predicted trajectories in invasive-ventilator-free patients with LOPD aged ≥ 5 years. We estimated annual FVC change 0–6 months, > 6 months–5 years, and > 5–13 years from treatment initiation, adjusting for baseline age, sex, and non-invasive ventilation. Findings Among 485 patients (4612 FVC measurements; 8.3 years median follow-up), median ages at symptom onset, diagnosis, and alglucosidase alfa initiation were 34.3, 41.1, and 44.9 years, respectively. FVC% increased during the first 6 months’ treatment (slope 1.83%/year; 95% confidence interval: 0.66, 3.01; P = 0.0023), then modestly declined −0.54%/year (−0.79, −0.30; P < 0.0001) during > 6 months–5 years, and −1.00%/year (−1.36, −0.63; P < 0.0001) during > 5–13 years. The latter two periods’ slopes were not significantly different from each other (P difference = 0.0654) and were less steep than published natural history slopes (−1% to −4.6%/year). Estimated individual slopes were ≥ 0%/year in 96.1%, 30.3%, and 13.2% of patients during the 0–6 month, > 6 month–5 year, and > 5–13 year periods, respectively. Conclusion These real-world data indicate an alglucosidase alfa benefit on FVC trajectory that persists at least 13 years compared with published natural history data. Nevertheless, unmet need remains since most individuals demonstrate lung function decline 5 years after initiating treatment. Whether altered FVC trajectory impacts respiratory failure incidence remains undetermined. Trial registration This study was registered (NCT00231400) on ClinicalTrials.gov on September 30, 2005, retrospectively registered.