Introduction The purpose of this paper was to evaluate the results on shoulder function following isolated proximal subscapularis release in children with Erb's palsy. Methods A retrospective study was conducted on 64 consecutive children with Erb's palsy who underwent a Carlioz proximal subscapularis release between 2001 and 2012. Fifty-one children with complete records and a minimum follow-up of 2 years were included for evaluation. Age at surgery ranged from 1.3 to 4,5 years (average 2.6 years). Preoperative passive and or active (depending on age at surgery) shoulder abduction anterior elevation (ABD), external and internal rotations (ER–IR) as well as the Mallet score were compared with those found at 6 and 24 months postoperatively using the o student paired t test O, with a confidence interval of 95%. The results were compared between children less than 3 years of age at surgery and those older, and between children who had an isolated C5, C6 and those with greater involvement. A P Results Abduction improved 21° at 6 months and 31° (total) at 2 years ( P P P P P = 0.013). There was no statistically significant improvement in internal rotation ( P = 0.37). We found no correlation between the child's age or the severity of involvement at surgery and the end result. Conclusion Proximal subscapularis release according to Carlioz is simple and effective in improving overall shoulder function in children with obstetrical brachial plexus palsy, mainly abduction and external rotation. Improvement tends to reach a plateau around 6 to 12 months postoperatively. The improvement of shoulder internal rotation remains the most difficult goal to achieve in these children.