Conventional growing rods are the most commonly used distraction-based devices in the treatment of progressive early-onset scoliosis. This technique requires repeated lengthenings with the patient anaesthetised in the operating theatre. We describe the outcomes and complications of using a non-invasive magnetically controlled growing rod (MCGR) in children with early-onset scoliosis. Lengthening is performed on an outpatient basis using an external remote control with the patient awake.Between November 2009 and March 2011, 34 children with a mean age of eight years (5 to 12) underwent treatment. The mean length of follow-up was 15 months (12 to 18). In total, 22 children were treated with dual rod constructs and 12 with a single rod. The mean number of distractions per patient was 4.8 (3 to 6). The mean pre-operative Cobb angle was 69° (46° to 108°); this was corrected to a mean 47° (28° to 91°) post-operatively. The mean Cobb angle at final review was 41° (27° to 86°). The mean pre-operative distance from T1 to S1 was 304 mm (243 to 380) and increased to 335 mm (253 to 400) in the immediate post-operative period. At final review the mean distance from T1 to S1 had increased to 348 mm (260 to 420).Two patients developed a superficial wound infection and a further two patients in the single rod group developed a loss of distraction. In the dual rod group, one patient had pull-out of a hook and one developed prominent metalwork. Two patients had a rod breakage; one patient in the single rod group and one patient in the dual rod group. Our early results show that the MCGR is safe and effective in the treatment of progressive early-onset scoliosis with the avoidance of repeated surgical lengthenings.