Variations in spine shape with body size, BMD and sex in individuals entering early old age

Purpose: Spine shape varies widely between individuals and is important for load bearing, yet associations with markers of health and disease have not been explored. The purpose of this pilot study was to characterise thoracolumbar spine shape in a sample of early old-age adults from the Medical Research Council National Survey of Health and Development (NSHD) and explore how it varies with anthropometric measures, bone mineral density (BMD) and sex. Methods: Using a pilot sample of 200 individuals in the NSHD birth cohort, all born in a single week in 1946, data included height, weight, body mass index (BMI), and spine bone mineral density (BMD) acquired at 60–64 years. Sagittal spine shape between the tenth thoracic and fifth lumbar vertebrae was described using an 89 point statistical shape model (SSM) on lateral dual-energy x-ray absorptiometry (DXA) images. Points entered into the SSM identified the average spine shape and independent modes of variation in shape, using principal components analysis. Mode scores quantified deviations from the mean shape. The first 10 modes explained 88% of variation in shape and were chosen for further analysis. Modes scores were tested for normality using the Shapiro-Wilk test and the Mann-Whitney test was used to compare the median difference between men and women. Associations between potential predictive factors and shape modes were tested using Spearman’s rank correlation. Results: The sample consisted of 110 men and 90 women with a median (interquartile range) age of 63.6 (62.8–64.2) years and BMI of 27 kg/m2 (24.5–30.5). Men were significantly taller and heavier (14 cm, 13 kg, respectively) compared with women. Similar to previous studies the first two modes described the total amount (M1) and distribution (M2) of spine curvature while M3 accounted for vertebral width. Sex differences existed in 5/10 modes (P<0.01). Women were more likely to have a snaking curvature (-M2) whilst men had an evenly distributed curvature (+M2) (P<0.001) and wider vertebral bodies (-M3, P<0.0001). Women had caudally increasing vertebral width and narrower L4/L5 disc space (+M6, P<0.0001) and men had narrower thoracic vertebrae (+M8, P<0.01) with a flatter curve (+M9, P<0.01). Correlations with height, weight and BMI varied by sex. Modes 2 and 6 (Fig. 1) were positively correlated with BMD in men alone (r=0.29 and 0.26, respectively). In women only, modes 4 and 5 negatively correlated with BMD (r=−0.28 and −0.20, respectively) while M3 correlated positively with height (r=0.27) and negatively with BMI (r=−0.21, Fig. 1). Scores for M9 increased with increasing weight (r=0.32) and BMI in men (r=0.31) but decreased with greater BMD in women (r=−0.24). Conclusions: In this sample of early old-age adults, sex-differences were observed in the distribution of spinal curvature and vertebral width rather than overall curvature. Furthermore, associations between spine shape, BMD and body size varied by sex. Greater BMD was related to snakier curvatures with distinct thoracic kyphosis and lumbar lordosis in women but an even overall curvature in men. These sex differences have not been explored in detail before but may be related to morphologic variations in other joints, which requires further investigation. These data will enable prospective studies of associations between spinal morphology and incident osteoarthritis. Opens large image Figure 1 Difference in spine shape with a positive (light line) or negative (dark line) two standard deviation change in modes 2, 3, 5, and 6.