It has been proposed that higher knee adduction moments and associated malalignment in subjects with severe medial knee joint osteoarthritis (OA) is due to anatomical deformities as a result of OA [1, 2]. The emergence of patient-matched implants should allow for correction of any existing malalignment. Currently the plans for such surgeries are often based on three dimensional supine computed tomography (CT) scans or magnetic resonance imaging (MRI), which may not be representative of malalignment during functional loading. We investigated differences in frontal plane alignment in control subjects and subjects with severe knee joint OA who had undergone both supine imaging and gait analysis.

Fifteen subjects with severe knee OA, affecting either the medial or lateral compartment, and 18 control subjects were selected from a database established as part of a larger study. All subjects had undergone gait analysis using the Vicon motion capture system. OA subjects had undergone routine CT scans and were scheduled for knee joint replacement surgery. Control subjects had no known musculoskeletal conditions and had undergone MRI imaging of hip, knee and ankle joints. Frontal plane knee joint angles were measured from supine imaging (supine) and from motion capture during standing (static) and during gait at the first peak ground reaction force (gait).

OA subjects had a significantly higher BMI (p < 0.01) and different gender composition (13 males and 2 females vs 4 males and 5 females; p = 0.03) compared with controls. Multiple linear regression analysis indicated no significant confounding effect of these differences on frontal plane angles measured in supine, static or gait conditions.

For both OA and healthy subjects, frontal plane knee angles were significantly higher during gait compared with supine (p = 0.03 and 0.02, respectively). There were also significant differences in knee alignment between OA and healthy subjects for supine and static (p < 0.05) but not for gait, although this was approaching significance (p = 0.052). Overall there seemed to be higher variation in alignment in the OA subjects (Fig. 1).

The significantly higher frontal plane knee joint angles measured in both control and OA subjects during gait compared with supine imaging indicate that functional alignment should be taken into consideration when planning patient-specific surgeries. Higher variation in OA patients may be due to alterations in gait patterns due to pain or degree of wear in their osteoarthritic joints, and requires further investigation. In addition, methodological considerations should be taken when comparing alignment from measurements taken with imaging and motion capture to avoid systematic errors in the data. In conclusion, we believe that both supine and loadbearing imaging are insufficient to gain a full representation of functional alignment, and analysis of functional alignment should be routinely performed for optimal surgical planning.