Associations between radiographic spinal oa assessed using lane grading and spine shape characterised using statistical shape modelling in a British birht cohort

Purpose: To investigate the associations between OA defined using Kellgren–Lawrence grading and hip shape characterised statistical shape modelling of hip DXA scans. Methods: Data were from the MRC National Survey of Health and Development, a British birth cohort born in March 1946. Between the ages of 60 and 64, n = 2856 (59% of the eligible cohort) were invited to attend a comprehensive health assessment at one of six clinical research facilities in England, Scotland and Wales, of which 1690 attended. Of these, 1601 individuals had a lumbar spine DXA scan, although 91 scans were subsequently excluded from analysis due to poor image quality leaving 1510 scans for statistical shape modelling. Scans were modelled using Shape software (University of Aberdeen). The model template consisted of 85 points and modelled T10 to L4 and the top edge of L5. Principal component analysis generated orthogonal modes of variation and in this study we included 10 modes of variation (86.5% of total variance). Images were graded using the system describe by Lane et al in 1993, with a grade of 0–3 for each vertebra and Total Lane Grade (TLG) was derived by summing the scores for all the vertebrae (T10-L4). A random sample was selected by SJW and was compared with grades previously assigned by an experienced reader (AVP). Inter-rater repeatability was determined using quadratic weighted Kappa scores in MedCalc v17.2. Correlations between age, sex, BMI, TLG and mode scores 1–10 were performed using Spearman’s rank test. One-way ANOVA with Tukey post-hoc analysis was used to assess the relationships between all spine modes (SM) [SM1] and TLG. All analyses were performed using SPSS v 23.0. [SM1]Mentions this earlier e.g. Correlations between age, sex, BMI, TLG and mode scores 1–10 were performed and spine modes associated with these factors were selected for subsequent analyses. Results: Quadratic weighted kappa scoring of repeatability between observers grading the images was good (kappa = 0.722). There were few cases of moderate or severe radiographic spinal OA identified in this cohort, with the majority of cases with a TLG of 7 or less. Most of the OA changes such as joint space narrowing and osteophytes (Lane grades 1–3) were observed in the lower lumbar vertebrae and 78% of the cohort was observed to have OA changes. TLG were lower in women and therefore more evidence of spinal OA was observed in the men. Spearman's rank correlations identified 5 modes of variation that were significantly associated with TLG - SM3, SM4, SM6, SM8 and SM9. One-way ANOVA analysis showed that SM2 (10% of total variation) was significantly negatively associated with TLG, demonstrated by a decrease in score with increasing TLG (P = 0.002; Fig 1a), although SM2 was not shown have significant associations with TLG in Spearman’s rank correlations. Negative SM2 scores describe a snaking curvature with lumbar lordosis and thoracic kyphosis centred on L1/T12 (Fig 1b). SM3 (8.6% of total variation) also showed a decrease in mode score with increasing MLG (Fig 1c), associated with greater relative anterior-posterior diameter of the vertebrae (P < 0.001; Fig 1d) as did SM4 (7.1% of total variation), which described a more S-shaped spine with greater rotation in the lower lumbar region and at T10 with smaller disc spaces caudally. Conclusions: Statistical shape modelling has demonstrated that there is a relationship between spine shape and spinal OA. In the NSHD there was limited evidence of radiographic spinal OA. There were several modes of variation identified that could be associated with features of OA in a relatively young cohort.