
Cross-Sectional Observational Study
Knee Osteoarthritis Is Associated With Reduced Hip but Preserved Spinal Movement Quality: An Artificial Intelligence-Based Markerless Motion Capture Study
Knee osteoarthritis is increasingly seen as a whole-chain disorder. In 177 adults with no hip osteoarthritis, hip movement quality fell steadily with radiographic knee-OA severity — while spinal movement quality stayed largely intact.
177
Participants
7.7 → 4.6
Hip-Twist, Control → KL4
5
Kellgren–Lawrence Grades
Why This Study Was Needed
A knee problem, measured up the chain.
Knee osteoarthritis has traditionally been viewed as a disease of the tibiofemoral joint, but growing evidence suggests its biomechanical consequences extend to the whole lower-limb kinematic chain — altering strategy at the hip, pelvis, trunk and spine. How radiographic knee-OA severity relates to movement quality in those proximal segments, in people without hip osteoarthritis, was poorly understood.
Does increasing knee-OA severity track with deterioration in hip and spinal movement quality — even when the hip shows no radiographic disease?
Movement quality scored with C.R.A.F.T.
Control
Within-trial steadiness of the movement. Scored for the spine; the platform did not generate a Hip-Control score for the sit-to-stand task.
Repetition
Reproducibility of the pattern across the three consecutive repetitions.
Asymmetry
Left-versus-right balance during the movement.
Flow
Temporal smoothness across the whole movement.
Twist
Transverse- and frontal-plane rotational dynamics — the dimension that fell most steeply at the hip.
How It Was Measured
One sit-to-stand, scored by AI.
A cross-sectional observational study graded 177 participants by the Kellgren–Lawrence system — controls and KL Grades 1 to 4 — each confirmed free of radiographic hip osteoarthritis. Movement quality was captured with MAI Motion, an AI-based markerless platform using a single frontal-plane RGB camera.
Step 01
Grade the knee
177 adults classified on the index knee: control (n=21), KL1 (n=32), KL2 (n=63), KL3 (n=40) and KL4 (n=21); mean age 68.4 years.
Step 02
Three-rep sit-to-stand
Each participant rose from a chair and sat back down three times at a self-selected pace — a protocol validated as equivalent to the five-repetition test.
Step 03
Score hip and spine
The C.R.A.F.T. framework generated hip and spinal movement-quality scores from 0 to 10; knee-region scores were excluded as confounded by the primary pathology.
Headline Finding
The hip gave way. The spine held.
Across Kellgren–Lawrence grades, hip movement quality deteriorated significantly — Hip-Flow, Hip-Twist and Hip-Overall all differed by grade, while Hip-Repetition and Hip-Asymmetry did not. Spinal movement quality was far more stable: only Spine-Flow reached significance, with no significant pairwise differences.
| Score (mean) | Control | KL1 | KL2 | KL3 | KL4 | p |
|---|---|---|---|---|---|---|
| Hip — Repetition | 7.6 | 6.4 | 6.6 | 5.7 | 5.8 | 0.182 |
| Hip — Asymmetry | 7.5 | 8.4 | 7.5 | 7.7 | 7.8 | 0.615 |
| Hip — Flow | 8.5 | 8.8 | 8.1 | 8.2 | 8.4 | 0.011 |
| Hip — Twist | 7.7 | 6.2 | 6.2 | 6.0 | 4.6 | 0.032 |
| Hip — Overall | 8.1 | 7.5 | 7.2 | 7.0 | 6.6 | 0.002 |
| Spine — Control | 4.5 | 4.0 | 4.9 | 3.9 | 3.5 | 0.412 |
| Spine — Repetition | 7.7 | 6.8 | 7.0 | 6.4 | 5.4 | 0.179 |
| Spine — Asymmetry | 5.9 | 6.4 | 6.6 | 8.5 | 5.7 | 0.189 |
| Spine — Flow | 8.0 | 9.0 | 8.5 | 8.4 | 8.5 | 0.024 |
| Spine — Twist | 4.5 | 4.0 | 4.8 | 4.2 | 3.7 | 0.324 |
| Spine — Overall | 5.7 | 5.7 | 6.1 | 5.8 | 5.1 | 0.205 |
Group means from Table 2 of the paper. Highlighted p-values mark a significant overall Kruskal–Wallis difference across grades (p < 0.05); higher scores indicate better movement quality. Hip-Control is absent because the platform did not generate it for the sit-to-stand task.
Finding 01
Hip-Overall fell from moderate KOA on
Overall hip movement quality was significantly lower than controls at KL Grade 2, Grade 3 and Grade 4 (p = 0.002, 0.040 and 0.032), suggesting proximal deterioration emerges relatively early in disease progression.
Finding 02
Rotational control dropped most
Hip-Twist showed the steepest gradient of any hip metric, declining from 7.7 in controls to 4.6 at KL Grade 4 — roughly a 40% reduction, and significantly lower than controls (p = 0.012).
Finding 03
The spine was largely spared
Among all spinal parameters, only Spine-Flow reached a significant overall difference (p = 0.024), with no significant pairwise differences; every other spinal score was statistically unchanged across grades.
Why It Matters Clinically
A deficit you can see before the X-ray.
Every participant was free of radiographic hip osteoarthritis, yet hip movement quality still fell as knee disease advanced. That points to functional, biomechanical change that can appear before structural degeneration becomes visible on imaging.
For a method that needs only a single conventional camera and no markers, that makes early, low-cost screening of the proximal kinematic chain realistic in routine practice — and gives rehabilitation a specific target.
Earlier
Detection
Hip movement declined despite no radiographic hip osteoarthritis — a functional signal that may precede structural change in the adjacent joint.
Targeted
Rehabilitation
A measurable hip deficit points rehabilitation toward hip strength, rotational control and frontal-plane stability rather than the knee alone.
Low-Cost
Screening
A single conventional RGB camera and no markers make this practical for routine, large-scale assessment in clinic and community settings.
Whole-Chain
View
The findings support treating knee osteoarthritis as a disorder of the whole lower-limb kinematic chain, not an isolated joint.
What This Study Does Not Show
Associative, and hypothesis-generating.
Cross-sectional, not longitudinal
The design cannot establish whether declining hip movement precedes or follows structural disease — only that the two are associated. The authors call the inferences hypothesis-generating.
No kinetics or electromyography
The C.R.A.F.T. scores are composite kinematic indices, not torque-referenced or EMG measures, so the neuromuscular mechanisms behind the deficits were not directly measured.
A single assessment platform
All movement quality came from one markerless system. Side-resolved (left/right) hip scores and a discrete Hip-Control score were not available, and knee-region scores were deliberately excluded as confounded by the primary pathology.
Index-knee grading and sex imbalance
Severity was classified on the index knee only, and the proportion of female participants rose markedly with grade (28.6% in controls to 90.5% at KL Grade 4), which future analyses should stratify for.
The work is the clinical companion to the markerless-motion programme — the three-repetition sit-to-stand validation that underpins its task, and the foundational marker-less capture study that established the method.
Knee osteoarthritis is a whole-chain disorder.
Hip movement quality deteriorated with knee-OA severity despite no hip osteoarthritis, while spinal movement stayed preserved — evidence that markerless motion analysis can detect proximal change before it is radiographically visible.
Study Identity
- Authors:
- Selin Demirel Mısırlı · Paul Lee · Tanvi Verma
- Institutions:
- Hacettepe University, Ankara · MSK Doctors Clinic, Sleaford
- Journal:
- Orthopedic Surgery and Trauma 2026;2(3), in press
- Ethics:
- Hacettepe University Non-Interventional Clinical Research Ethics Committee (2025/21-22)
- Funding:
- None declared
Citation
Demirel Mısırlı S, Lee P, Verma T. Knee Osteoarthritis Is Associated With Reduced Hip but Preserved Spinal Movement Quality: An Artificial Intelligence-Based Markerless Motion Capture Study. Orthop Surg Trauma. 2026;2(3).
doi:10.14744/ortst.2026.29887 — in press; the publisher link will be added once the article is online.
Funding and affiliations
Affiliations: Department of Sports Medicine, Hacettepe University, Ankara · MSK Doctors Clinic, Sleaford
Funding: None declared