Purpose:
Three-dimensional (3D) gait analysis is widely used in the
quantitative evaluation of gait. However, 3D gait analysis produces a
large volume of data, and simplifying such complex data into a single
measure of patients overall gait ‘quality’ would be valuable in clinical
practice. The Gait Deviation Index (GDI), used to evaluate treatment in
children with cerebral palsy, has been proposed as such a measure. The
experience with GDI in osteoarthritis (OA) patients following total hip
arthroplasty (THA) is unknown. The aim of our study was to use the GDI
to evaluate post-operative gait quality changes in patients with hip OA
following two types of THA.

Methods: A total of 38 patients (11 females and 27 males, age 56 5.6,
BMI 27.8 3.6) with unilateral end-stage primary hip osteoarthritis
were evaluated pre-operatively, two- and six-months after total hip
arthroplasty, using 3D gait analysis while walking at self-selected
speed. Upon completion of the pre-operative assessment, the patients
were randomly assigned to either resurfacing hip arthroplasty (RHA) or
conventional total hip arthroplasty(THA). All patients were allowed
early postoperative weight-bearing, and had rehabilitation supervised
by an in-hospital physiotherapist. All patients completed the gait
evaluation at each follow-up. From the entire variability in kinematic variables across a gait cycle, rather than a small number of discrete
parameters, the GDI was calculated for each limb (n ¼ 76 limbs). The
normative mean and standard deviation (mean ¼ 94.7; SD ¼ 8.4) from
our age-matched controls (n ¼ 20) were used as reference. A fixedeffects
multilevel regression model was employed to evaluate the
treatment effects.

Results: No interaction was observed between treatment and time (p ¼
0.33) or limb and time (p ¼ 0.53). The pre-operative GDI mean value
was 83.4 10.9, showing patients had a moderate deviation from
normative gait before surgical treatment. After surgical
treatment, the GDI score improved significantly by 4.9 [:95CI: 2.1 to 7.9]
equal to a 0.8 average increase in GDI per month of follow-up. Therewas
no difference in GDI scores between the two surgical treatments
groups; 1.8 [:95CI: -2.8 to 6.4]. However, the GDI score for the nonoperated
limbwas higher than the GDI score for the non-operated limb;
2.5 [:95CI: 0.1 to 4.8].

Conclusions: Our results show that, THA and RHA patients recovered
equally well from the respective treatments. The GDI increased significantly
after THA surgery, which indicates an overall improvement in
gait quality for both treatment groups. The difference between the
operated and the non-operated limb showed that asymmetrical gait
pattern do not disappear following THA. Further research is required to
establish the clinical relevant difference for the GDI score, and to
determine the association with pain and OA severity.