AIMTo establish whether the initial positive effect of nusinersen (NUS) on respiratory outcomes in the first year of treatment was maintained in children with Spinal Muscular Atrophy (SMA) type 2 and to further define the effect on children with type 3 treated over 3 years.METHODSA prospective observational study of children with type 2 and 3 beginning NUS in Queensland, Australia between June 2018-December 2020 was undertaken. Investigations conducted included age-appropriate lung function and polysomnography. Lung function data for two-years preceding NUS initiation was retrospectively collected. Change in lung function/polysomnography was assessed using mixed effects linear regression.RESULTS24 of 30 children with type 2 and 3 SMA (14 males; 2.6-15.8) were included (type 2 n = 12; type 3 n = 12). No child had respiratory-related admissions during the period of study. For type 2, annual decline in FVC z-score pre-treatment was -0.75 (95 % CI: 1.14, -0.39, p < 0.001), and for the first 3 years on NUS was -0.20 ([95 % CI: 0.33, -0.06, p = 0.01] difference p = 0.008). For type 3 minimal change was seen: pre-NUS and post FVC z-scores -0.20 (95 % CI: 1.00, 0.61 p = 0.05) and -0.46 (95 % CI: 0.88, -0.04 p = 0.40) respectively (difference p = 0.46). Mean change in total apnoea-hypopnoea indices (total AHI) in type 2 tended to reduce -1.75 (95 % CI: 4.95-0.9, p = 0.24); type 3 appeared to remain stable (-0.39 [95 % CI: 1.1-0.33, p = 0.28). One child with type 2 ceased NIV due to normalisation of total AHI and gas exchange.CONCLUSIONNusinersen lung function (FVC-z-scores) stability seen in the first year was maintained over 3 years and the total AHI tended to improve in type 2, but the long-term effects in type 3 are less clear.