The main findings of this study were that compared to controls, patients with hip OA with mild to moderate symptoms walked with significantly reduced velocity, and revealed significantly reduced hip and knee joint excursion, specifically manifested as reduced extension in the hip and knee joint during the latter 50% of the stance phase. Thus, our hypothesis that hip OA patients in this early stage of disease would reveal reduced hip excursion during stance and a reduced gait velocity compared to healthy subjects was confirmed. Our secondary hypothesis, that hip OA patients with severe ROA defined from a MJS ≤2 mm would reveal reduced hip excursion during stance and reduced gait velocity compared to patients with MJS >2 mm, was only partly confirmed; as these patients had significantly reduced hip excursion during stance, but no difference in gait velocity.
The rationale for this study was the lack of consensus on the identification of specific gait parameters in patients with hip OA, especially with regard to patients in the early stage of disease. The included patients in this study had a mean (SD) HHS score of 78.2 (8.2), defining them as non-eligible for THR in our institution, and a mean (SD) WOMAC pain subscore of 25.3 (18.1), reflecting mild to moderate pain . In our clinic, the orthopedic surgeons use ≤2 mm MJS as their clinically anchored cut-off when categorizing patients with hip OA from their level of radiographic progression. We therefore defined this as the cut-off value for subcategorizing our cohort. There is no absolute consensus on categorizing radiographic severity based on MJS in hip OA. However, several studies state ≤2.5 mm MJS as a minimum for defined ROA, and a MJS ≤1.5 mm as severe ROA [1, 2, 43–45]. Reijman and colleagues  more explicitly suggest MJS ≤2.5 mm as intermediate, ≤2 mm as moderate and ≤1.5 mm as severe ROA, respectively. The hip OA patients included in this study were, thus, overall ratified to having a mild to moderate level of disease, assessed both from symptoms and radiographs; and the subjects classified to have more severe ROA as having at least moderate radiographic changes.
The identification of reduced velocity and reduced hip and knee joint extension during the latter 50% of the stance phase as key gait features in our cohort, confirmed that the biomechanical deviations most commonly described in severe hip OA gait are manifest already at an early stage of disease. To our knowledge, no more than two other studies [27, 28] have reported 3D gait analysis data of patients at a comparable stage of hip OA, and also compared them to controls. Of these, only the study by Dujardin and colleagues  had a sample size comparable to ours. Unfortunately, their study was published in French, so we were unable to weigh our findings against theirs. In line with our findings, the other study by Watelain et al.  identified reduced velocity and reduced hip excursion as characteristics of gait in patients with early stage hip OA. However, their sample size was considerably smaller than ours. Thus, our study adds clinically relevant knowledge on biomechanics during gait in a population of hip OA patients that previously has been given little attention.
Our hip OA patients walked with significantly reduced velocity compared to the controls. Still, their walking speed was faster than what have commonly been reported, compared both to age-related normative data [47, 48] and studies including patients with severe hip OA [12, 49]. A plausible cause for the latter is the fact that we included patients with mild to moderate symptoms, in contrast to most studies that have focused on patients in the end-stage of disease. Higher walking velocity has been shown to increase joint moments  and accentuate pathological changes in patients with severe hip OA . Thus, it can be questioned whether observed biomechanical changes in hip OA patients can be explained merely from their reduced gait speed. However, reduced velocity appears inherently linked to disease progression, and can as such be regarded to be an intrinsic key feature of hip OA gait. Thus, it is not straightforward to consider velocity solely as a “disturbing” variable, that should be controlled for in a regular analysis of variance, leaving the unadjusted analyses out . From a clinical perspective, controlling for velocity can give diminutive meaning; as it is inapplicable to just tell the patients to walk faster in order to reduce their joint moments. For these reasons, we performed the main analyses unadjusted for velocity; but included a supplementary binary logistic regression model to reveal whether the observed key differences in hip and knee excursion and hip joint moments were persistent after adjusting for velocity. These analyses confirmed that the reductions in hip excursion and hip joint moments in hip OA patients compared to controls were persistent regardless of velocity; whereas knee excursion was not. The unadjusted and adjusted analyses together thus strengthened the identification of reduced hip extension during excursion and hip flexion moments as key features of gait in hip OA patients with mild to moderate symptoms.
Abnormal joint loading has been shown to contribute to detrimental shear stresses as well as disruption and loss of cartilage, and is considered to be an important mechanism of lower limb OA pathogenesis [52–59]. Accumulated inadequate loads may further play a role in disease progression, as they can facilitate enlargement of the joint surface that is worn down [29, 59, 60]. A relevant factor in this aspect is the reduced range of passive hip joint extension we found for the hip OA patients. Static contractures that increase the stiffness of the hip joint have long been suggested to limit hip extension during the late phase of stance . The hip OA patients in our study had a mean of 0.8° (8.7) passive hip ROM extension; a mean reduction of 9.1° compared to the controls (p<0.001). This was reflected during gait, where the hip OA patients had 9.6° less peak hip extension than the controls. Simonsen et al.  recently reported that walking with the upper body inclined was associated with both a significant increase in hip joint flexion angle and a significantly reduced hip flexion moment in the latter phase of stance. As our experimental set-up did not include an upper body model, we could not report the positioning of the upper body in our subjects. However, we found the exact same pattern of both reduced hip extension angle and reduced hip flexion moment during late stance in our cohort. The verification of these biomechanical deviations already in hip OA patients with mild to moderate symptoms, may reflect an unfavorable stress distribution that possibly can contribute to facilitated disease progression.
In correspondence with the hip joint, reduced extension during the latter 50% of the stance phase was found also in the adjacent knee when we compared the hip OA patients to the controls (p-values between <0.001 and 0.035). Further, a significant reduction in isokinetic knee extension strength was established between the overall group of patients and the controls (p=0.001). It is clinically important to be aware that such deviations will influence knee joint loading, and over time may contribute to the development of secondary degenerative changes also in the knee joint. As the patients in this study were excluded if they had any form of knee pain or dysfunction, including restrictions in passive knee ROM, the observed differences at the knee joint are likely to be adaptive consequences of the structural changes and pain in the hip joint. This was further supported by the finding that in contrast to the hip joint excursion, the knee joint excursion difference was no longer significant when adjusting for velocity.
To what extent the level of ROA is reflected in biomechanics during gait in hip OA patients non-eligible for THR, has never been investigated. When subcategorizing the hip OA patients from their level of ROA, the patients with ≤2 mm MJS were classified as having severe ROA (mean MJS of 1.2mm), and the patients with >2 mm MJS as having intermediate ROA (mean MJS of 2.9 mm) . The hip and knee kinematic joint excursion curves illustrated that patients with MJS ≤2 mm revealed the largest deviations. However, this was not reflected in the functional or clinical assessments. No differences were evident between patients with MJS ≤/>2 mm in either of the WOMAC subscales. It has been stated that ROM restrictions are larger in patients with more severe ROA , but we did not find statistical significant differences in hip ROM flexion or extension between the two subgroups. However, the lack of significance must be interpreted with caution, due to the relatively low number of subjects in each group. It cannot be ruled out that the 9° difference in hip flexion and the 7° difference in internal rotation may be of clinical importance. Finally, we found no differences in isokinetic hip muscle strength. However, patients with MJS ≤2 mm revealed a highly significant reduced hip flexion moment at peak hip extension (p=0.001), that still was significantly reduced when controlling for velocity (p=0.028). This may indicate non-optimal use of the hip extensor muscles that was manifested during walking; but not during the isokinetic test. In sum, our findings suggested that more severe ROA resulted in more distinct gait alterations in the sagittal plane, even though there were no significant differences in symptoms or function. This emphasizes the importance of supplementing self-reported and clinical outcome measures with assessments that quantify function and gait , when trying to understand how altered gait can influence disease progression – or vice versa.
Overall, the findings of this study are in line with previous studies suggesting reduced velocity, reduced sagittal plane joint excursion and a reduced hip flexion moment in the late stance phase as key features of hip OA gait. What is new from our study, is that these hallmarks are evident already in hip OA patients at an early stage of symptoms. These patients are not eligible for total hip replacement. Rather; they should be considered as primary candidates for exercise therapy. In this context, the identification of early stage gait alterations is clinically important, as it may contribute to enable development of targeted treatment interventions. To what extent the observed alterations were primarily caused by – or a result from – the level of pain, changes in bone and cartilage morphology, muscle weakness or reduced passive ROM, is still unknown. Thus, it remains a question whether the gait characteristics described in this study may drive further progression of either symptomatic or radiographic hip OA; or primarily reflect adaptations to the constraints posed by the degenerative changes in the joint. Hence, future studies should investigate with particular interest the relationship between these key biomechanical features and different clinical and functional outcomes.
There are some limitations to this study that need to be addressed. First, only sagittal plane gait characteristics were reported, as we wanted to emphasize hip excursion, and in particular hip extension, in this plane. A comprehensive gait analysis should however include also frontal and transversal plane kinematics and kinetics. Further, the included analyses between the two pre-defined subgroups of patients with MJS ≤/>2 mm were not statistically corrected for multiplicity. In addition, the subgroup sample sizes were smaller. Our findings related to the secondary aim of this study should therefore be regarded as an explorative supplement to the main analysis, and as such be interpreted as hypothesis-generating rather than conclusive .