Knee Alignment: Why the Knees Follow the Pelvis and the Feet
Key Takeaways
- The knee is a hinge with very little rotational freedom on its own. Where it sits in space comes from the hip above (femur rotation) and the foot below (tibia rotation through the ankle).
- A 2023 ACL-rehab study found a very large correlation (r=0.78) between contralateral pelvic drop and dynamic knee valgus during single-leg loading.
- Hip muscle strength alone does not consistently predict knee valgus. A 2021 systematic review found conflicting results across seven studies in healthy adults.
- Anti-pronation foot orthoses changed hip and knee kinematics in a 2014 lab study, but a prospective study of 473 military recruits found no consistent link between foot pronation and anterior knee pain.
- Single-leg control, ankle dorsiflexion, and hip abductor work paired with task practice outperform any single-cause fix at the knee itself.
Knee alignment is rarely a knee problem. The knee is a hinge with a small amount of rotational play, and the segments above and below it (the femur driven by the hip, the tibia driven by the foot) decide where the joint sits in space. If your knees collapse inward when you squat, the question is almost never what your knee is doing wrong. It is what your pelvis is doing on top and what your foot is doing underneath. I learned this the slow way: knee work alone did almost nothing for me until I addressed the chain above and below. This piece walks through both directions of the chain and how to read your own alignment.
Why the knee is a downstream joint
The knee bends and straightens with confidence. It rotates with much less. The total available rotation at the tibiofemoral joint sits around 5 to 10 degrees in flexion and almost zero at full extension, which means the knee cannot generate dramatic rotational changes on its own. Where the joint actually points in space comes from the segments it connects to.
Above the knee, the femur turns with the hip. Hip internal rotation pulls the femur inward, which drags the knee with it. Hip abduction strength and pelvic stability decide how much the femur drifts during a single-leg load, and any drift at the hip shows up at the knee as valgus collapse.
Below the knee, the tibia turns with the foot. As the foot pronates and the arch lowers, the lower leg rotates inward in a coupled motion through the subtalar joint and ankle. The knee then has no choice but to follow the rotation of the tibia underneath it.
This is why the standard advice (do not let your knees cave) addresses the symptom, not the cause. The knee is a follower, not a driver. Coaching the knee while leaving the hip and foot alone tends to fail within a few weeks.
The knee is a follower, not a driver. Coaching it while leaving the hip and foot alone fails within a few weeks.
The hip side of the chain
Look up the chain first. The strongest single piece of evidence connecting pelvic control to knee alignment came out of a 2023 study on athletes returning to sport six months after ACL reconstruction. The researchers measured contralateral pelvic drop (the swing-leg side dropping when the standing leg is loaded) against dynamic adaptive knee valgus, and found a correlation of r=0.78, which is very large in biomechanics terms1. When the pelvis dropped, the standing-leg knee collapsed inward almost in lockstep.
What is harder to pin down is whether plain hip muscle strength is the right intervention. A 2021 systematic review of seven studies in healthy adults concluded that the relationship between hip abductor, extensor, and external rotator strength and dynamic knee valgus is conflicting2. Two studies found a negative correlation. Two found a positive one. Three found no relationship at all.
This sounds confusing until you separate strength from control. A hip can be strong in isolation and still let the pelvis collapse during single-leg loading because the timing or the cueing is off. Most of the dynamic knee valgus you see in untrained movement is a control problem, not a strength problem. The hip abductors do not need to bench press the body; they need to hold the pelvis level for the half-second the foot is on the ground.
This is the same reason single-leg walking and step-down work tend to outperform pure hip-abduction machine work for people with knee complaints. The task demand of holding the pelvis level under load is what teaches the timing.
The foot side of the chain
Now look down the chain. A 2014 lab study on healthy adults found that prefabricated anti-pronation foot orthoses reduced peak hip adduction by 1.56 degrees and knee internal rotation by 1.3 degrees during a functional step-up3. Both effects were statistically significant. The foot changed, and the hip and knee changed with it, even though the orthotic was nowhere near either joint.
That demonstrates the kinetic chain in action. But it does not mean every knee problem traces back to the foot. A prospective study following 473 infantry recruits through 14 weeks of training found that 15 percent developed anterior knee pain, and the foot pronation measurements taken at intake had no consistent association with who got hurt4. The simple version (pronation causes knee pain, fix the foot, fix the knee) is not what the long-run data shows.
The cleaner reading: the foot can change knee mechanics, and changing the foot is one available lever, but pronation is not a single-cause condition predicting knee pain. People with stiff first toes, restricted ankle dorsiflexion, or foot mobility losses tend to compensate at the knee. People with neutral, mobile feet and good single-leg control can sustain a wide range of foot postures without trouble.
Where the chain framing matters
The chain framing matters because the popular advice keeps trying to fix the knee directly. Two examples make the point.
The first is patellofemoral pain. Knee taping, quadriceps strengthening, and patella-tracking exercises were the standard for two decades. Hip abductor and external rotator work eventually outperformed all of them in head-to-head trials, but only for the subset whose dynamic valgus was traceable to the hip in the first place. If your knee pain is downstream of a stable pelvis and a normally tracking tibia, hip work will not help much.
The second is the knees-over-toes debate. The current online crowd argues knees should travel forward over the toes during squats. The previous online crowd argued knees must stay behind the toes. Both miss the same point. Knee travel is determined by ankle dorsiflexion range and hip mobility on the way down, not by a rule about the toes. If the ankle has full mobility and the hip allows the depth, the knee finds its position. If either is locked, the knee compensates and the rule does not change anything.
In practice, good posture at the knee is the natural product of mobile and controlled segments above and below. It cannot be coached at the knee itself for any length of time, and the protocols that try to do so eventually drift toward addressing the chain anyway.
How to read your own chain
Three quick self-checks anyone can do in a hallway with a mirror.
The pelvic drop test. Stand on one leg with the other knee lifted at hip height. Look down at your hips in a mirror. If the swing-leg hip drops below the standing-leg hip, you have a pelvic-drop pattern on that side. The standing-leg knee will tend to collapse inward as a result. Test both sides; differences between sides matter as much as the absolute drop. Hold for ten seconds and the drop usually shows up clearly.
The single-leg squat. Stand on one leg, holding a wall lightly with one finger for balance. Squat until the knee bends 30 to 45 degrees and watch where the knee tracks. The knee should travel over the second or third toe. If it drifts inward toward the big toe, you have dynamic valgus on that side. If it drifts outward, you have a less common varus pattern. Some people will see one knee collapse and the other track cleanly, which is the asymmetry pattern that drives a lot of unilateral knee complaints.
Your shoe wear. Pull out shoes you have walked in for six months. If the outer heel and inner forefoot are most worn, you pronate, which is the common pattern. Outer heel and outer forefoot worn means you supinate. Heavy uneven wear between left and right shoes is the strongest signal that one side of the chain is loaded differently.
What actually works
Four interventions hold up across the literature for knee alignment driven by chain issues, in roughly the order I would prioritize them.
First, restore ankle dorsiflexion. A locked ankle is the single most upstream cause of knee compensation in the lower body. Use a wall ankle test (knee to wall, foot four inches from the baseboard, see if the knee can touch). If the knee cannot reach the wall, the ankle needs work before anything above it gets prioritized.
Second, single-leg control work. Step-downs, single-leg deadlifts, and lateral step-ups loaded slowly. The point is not the load on the leg; it is the time spent holding the pelvis level while the foot is on the ground. Three sets of eight to twelve reps per side, three times a week, for six weeks tends to be enough to retrain pelvic control in most people.
Third, hip abductors with intent. The systematic review showed strength alone has mixed results, but eccentric hip abductor work paired with single-leg tasks is more reliable. Side planks, side-lying leg lifts, and banded clamshells help less than people think. Step-downs and single-leg squats with deliberate slow eccentrics help more, because they pair the strength demand with the control task.
Fourth, foot intrinsic mobility. Toe yoga, short-foot exercises, and barefoot time on uneven ground keep the foot reactive enough that it can adjust to surfaces without locking the chain. A foot that stays neutral by default takes load off the strategy of bracing the knee.
UpWise is an iOS app that analyzes your standing posture from a single side photo, including pelvic tilt and lower-body alignment patterns. Use the report to figure out which end of the chain (pelvis or foot) is showing up in your standing baseline before you commit to a six-week routine. The app also runs daily exercise routines that target single-leg control and hip-abductor work specifically, two of the four interventions above.
Frequently Asked Questions
What causes my knees to collapse inward when I squat?
Most of the time the cause is upstream at the hip or downstream at the foot, not the knee itself. Contralateral pelvic drop and dynamic knee valgus correlated very strongly (r=0.78) in a 2023 study of post-ACL athletes, which means the pelvis is doing most of the work of pulling the knee inward. Foot pronation can also rotate the tibia inward and pull the knee with it. Test the pelvic drop and the single-leg squat to figure out which end is driving.
Should my knees go past my toes when I squat?
Both rules miss the point. Knee travel during a squat is determined by ankle dorsiflexion range and hip mobility, not by toe position. If your ankle has full motion and your hip allows the depth, the knee finds its own position safely. If either is locked, the knee compensates and the rule does not change the mechanics.
Will hip strengthening fix my knee pain?
Sometimes. The 2021 systematic review found conflicting results for hip strength alone: two studies showed a negative correlation with knee valgus, two positive, and three none. Hip work paired with single-leg control practice has more consistent results, because it teaches the pelvis to stay level under load instead of just growing the muscle.
Do I need orthotics for knee alignment?
Not always. A 2014 lab study showed anti-pronation orthoses reduced hip and knee internal rotation in healthy adults during step-ups, so the chain effect is real. But a prospective study of 473 military recruits found no consistent link between foot pronation and who developed anterior knee pain. Orthoses are one available lever, not a guaranteed fix. Address ankle mobility and single-leg control first.
Can I see knee alignment problems in my standing posture?
Some patterns yes. Static standing posture shows pelvic tilt, lower-leg rotation, and foot position, all of which feed into knee alignment under load. UpWise flags these from a side photo. But dynamic knee valgus only shows up under single-leg loading, so the standing photo is the starting point and the single-leg squat is the confirmation.