A retrospective analysis of 54 patients with paralytic scoliosis due to myelomeningocele, who underwent surgical treatment, was performed. The aim of this study was to compare different surgical techniques and to identify clinical parameters influencing primary and midterm results. Three surgical techniques were used: 1) group I, posterior fusion/instrumentation; 2) group II, anterior fusion/no instrumentation combined with posterior fusion/instrumentation; and 3) group III, anterior and posterior fusion/instrumentation. Average age at surgery was 13.1 years. A preoperative scoliosis angle of 90° [interquartile range (25th-75th percentile) (IQR), 76-106°] was primarily reduced to 38° (IQR, 30-50°). At final follow-up (mean, 3.3 years), correction deteriorated to 44° (IQR, 38-65°). The group III procedure resulted in a better midterm correction of scoliosis compared with group I (P= 0.02). The extension of anterior fusion correlated with primary and midterm correction of scoliosis (P< 0.03). Patients with a thoracic level of paralysis had a higher relative loss of correction compared with patients with a lumbar level (P< 0.06). This finding can be attributed mostly to group I patients (P= 0.011). Hardware complications occurred in 16 patients (30%). Relative loss of correction among these patients was high (P< 0.01) and relative midterm correction low (P= 0.001). We recommend anterior and posterior fusion, each with instrumentation for the treatment of paralytic scoliosis in myelomeningocele. In patients with a thoracic level of paralysis, the two-stage procedure is mandatory to reduce the risk of hardware complications and subsequent major loss of correction.