Introduction. Horizontal gaze palsy (HGP) in association with scoliosis has been reported in both orthopaedic and ophthalmological published work. Juvenile progressive scoliosis in combination with HGP is caused by a malfunction of the normal control mechanism for equilibrium related to the lower brain stem, mostly associated with ROBO3 gene mutation. The aim of this study is to establish the association of scoliosis and HGP. Methods. 13 cases (four families and three sporadic cases) with HGP and scoliosis were documented; other systemic and ocular associated findings were identified and genetic counselling was done. All patients had radiograph of the spine, cranial and spinal cord MRI, chromosome analysis, gene analysis, and full ophthalmological examination. Blood samples were tested for ROBO3 gene mutation at Engle Laboratory, USA. Results. Mean age at evaluation was 14·8 years (range 1–56). Scoliosis was of varying degree (Cobb angle range 10–65°, mean 38·4°), and horizontal pendularnystagmus of low amplitude and loss of conjugate horizontal eye movements were common in all patients. Scoliosis was early onset and progressive in all patients. Six of 11 patients underwent surgery (Cobb angle range 45–70°, mean 5·3°). There were four right thoracic, two left thoracic, three right thoracolumbar, and two left thoracolumbar curves. Two patients of the third family (cases 10 and 11) had mirror image thoracolumbar curves. Cranial and spinal MRI revealed cleft in medulla oblongata in all nine patients who underwent MRI. Two adult patients refused MRI and two infant MRI scans were suboptimal. Neurological examination was otherwise normal. Notably, the female patients of family 1 also had genital dysgenesis. This is the first report of this finding in association with HGPPS, and its significance is not yet known. Homozygous ROBO3 mutations were identified in affected members of all four families, despite no recognised consanguinity in two of the families. Two were homozygous nonsense mutations (G456X and W635X) and two families shared the same missense mutation (S1107R). Two of the three sporadic cases underwent mutation testing; one harboured a homozygous missense C901R mutation, whereas a mutation was not detected in the second (patient 4). There were no distinguishing features of patient 4, suggesting that she has a non-coding mutation or that there is a second HGPPS gene. Conclusions. Every child with HGP should be evaluated for a possibly associated scoliosis, which is a progressive condition. Our cases and other published work clearly indicate that even if scoliosis is not present at first evaluation, longitudinal follow-up will show the evolution and progression of scoliosis. ROBO3 mutation testing should be done

Spinal deformities are a common feature of Marfan’s syndrome and can be a significant cause of morbidity. The morphology of the scoliosis associated with this condition was previously described by Sponseller, but no correlation with the pelvic parameters has been seen. We performed a retrospective radiological study of 58 patients with scoliosis, secondary to Marfan’s syndrome and related the findings in the thoracolumbar spine to the pelvic parameters, including pelvic version (tilt), pelvic incidence and sacral slope. Our results showed marked abnormalities in the pelvic values compared with those found in the unaffected population, with increased retroversion of the pelvis in particular. In addition we found a close correlation between the different patterns of pelvic parameters and scoliosis morphology. We found that pelvic abnormalities may partially dictate the spinal disorders seen in Marfan’s syndrome. Our results supplement the well-established Sponseller classification, as well as stressing the importance of considering the orientation of the pelvis when planning surgery.