Use of dual-energy X-ray absorptiometry to evaluate variation in bone shape and alignment associated with radiographic knee osteoarthritis: Findings from a study of 19,053 individuals in UK Biobank

Abstract

Objective: Lower limb alignment may predispose to, or exacerbate, symptoms of knee osteoarthritis. To examine the role of this and other joint shape variation, we conducted a cross-sectional study investigating relationships between radiographic knee osteoarthritis (rKOA) and knee shape in dual-energy X-ray absorptiometry (DXA) images from UK Biobank (UKB).

Methods: Associations between the first ten knee shape modes (KSMs), derived from statistical shape modelling, and rKOA grade were analysed using logistic regression, adjusting for age, sex, height, weight, and ethnicity. An additional model included adjustment for hip-knee-ankle (HKA) angle, derived from total body DXA scans, to reflect knee alignment. Composite figures illustrate knee shape characteristics associated with each rKOA grade.

Results: 19,053 individuals were included (mean 63.7 years, 48 ​% males), of whom 80.7 ​%, 14.6 ​%, 3.6 ​% and 1.2 ​% had rKOA grades 0, 1, 2 and 3-4, respectively. Several KSMs were associated with rKOA in confounder-adjusted analyses, with higher grades showing stronger relationships. These associations were attenuated by adjustment for HKA. As expected, composite shape models revealed that higher rKOA grades were associated with greater varus malalignment. After HKA adjustment, composite shape models showed less varus alignment, with other shape differences, such as altered proximal tibial metaphysis and lateral patella displacement, emerging in higher-grade rKOA.

Conclusions: Our cross-sectional analyses between joint shape and DXA-derived rKOA grade showed expected relationships with varus malalignment, which were attenuated after adjusting for HKA. Other shape differences, particularly in higher-grade rKOA, emerged independently of alignment, warranting further investigation.

Keywords: Dual-energy X-ray absorptiometry; Knee alignment; Knee osteoarthritis; Statistical shape modelling.

Fig. 1

First 10 KSMs derived from knee DXA scans. Shapes +2 standard deviations (SD) from the mean are shown in blue, whilst shapes -2SD from the mean are shown in red.

Fig. 2

Landmarks used to calculate alignment measures from total body DXA scans. The tibial shaft axis (TA, blue arrow) extends from the ankle to the knee centre (intercondylar notch). The mechanical femur axis (FA, green line) runs from the centre of the femoral head to the centre of the knee joint. The hip–knee–ankle (HKA) angle is defined as the angle between the TA and the mechanical FA, while the femorotibial angle (FTA) is defined as the angle between the TA and the anatomical FA (orange arrow).

Fig. 3

Associations of KSMs with rKOA grades in combined sex analyses. Results show OR with 95 ​% confidence intervals, per SD increase in KSM. Model 1: unadjusted; Model 2: adjusted for age, sex, height, weight and ethnic group; Model 3: additionally adjusted for hip-knee-ankle (HKA) angle. Abbreviations: OR, odds ratio; KSM, knee shape mode; rKOA, radiographic knee osteoarthritis.

Fig. 4

Composite models adjusted for age, sex, height, weight, and ethnicity (sex combined). Dashed black line ​= ​normal healthy knee, red line ​= ​shape associated with rKOA. All shapes are aligned at point 17 of the SSM template, located at the distolateral corner of the femur.

Fig. 5

Composite models adjusted for age, sex, height, weight, ethnicity and HKA (sex combined). Dashed black line ​= ​normal, red line ​= ​shape associated with rKOA. Shapes are aligned at point 17 on the SSM template.