Alignment Philosophies in TKA: Mechanical, Kinematic and Functional

The question. Alignment philosophy asks one thing: what target should the components hit? The answers sit on a spectrum between two poles. Mechanical alignment references the limb's mechanical axes and accepts soft-tissue releases to balance a neutral knee; kinematic alignment references the patient's own pre-arthritic joint surfaces and balances by restoring native anatomy. Restricted kinematic and functional alignment are bounded hybrids between them.

Why it is argued about. A stubborn minority of patients remain dissatisfied after a well-performed mechanically aligned TKA. The move away from "neutral for everyone" rests on the observation that constitutional (native) lower-limb alignment is highly variable: many healthy people have constitutional varus or valgus. Mechanical alignment normalises all of these to neutral, which alters each individual's joint line and frequently requires ligament releases to balance.

Where robotics fits. Individualised philosophies try to respect native anatomy instead. Robotics and navigation are the enabling tools that make precise, reproducible execution of these individualised targets possible.

View larger

The angles that define them. Each philosophy is defined by the angles it cuts and measures, so the vocabulary comes first:

• mLDFA (mechanical lateral distal femoral angle): femoral mechanical axis to the distal femoral joint line. Normal ~85 to 90° (mean ~87.5°); a higher value means distal femoral varus.
• MPTA (medial proximal tibial angle): tibial mechanical axis to the proximal tibial joint line. Normal ~85 to 90° (mean ~87°); a lower value means tibial varus.
• JLCA (joint line convergence angle): the angle between the femoral and tibial joint lines. Normal ~0 to 2°; it reflects intra-articular wear/laxity.
• aHKA (arithmetic HKA) = MPTA − mLDFA: estimates constitutional alignment by removing the wear/JLCA. More negative is varus, more positive valgus.
• JLO (joint line obliquity) = MPTA + mLDFA: the orientation of the joint line.

aHKA and JLO are the two axes of the CPAK classification, and their neutral ranges are given with it below.

View larger

View larger

Mechanical alignment (MA). The target is a neutral hip-knee-ankle axis (HKA 180°), with the femoral and tibial components each perpendicular to their own mechanical axis (mLDFA = MPTA = 90°) and a horizontal joint line. The gaps around these neutral cuts are equalised by soft-tissue releases. Its case is reproducibility, validated long-term survivorship and even load distribution, and it is the historical and current reference standard.

Anatomic alignment. The joint line is set ~3° from the mechanical axis, restoring the historical native obliquity. It is largely historical.

Kinematic alignment (KA). The target is the patient's pre-arthritic (constitutional) joint lines, restored by resurfacing worn bone, with the components aligned to the knee's native flexion/extension and rotational axes. It recreates the native mLDFA and MPTA, often with the femur in slight valgus and the tibia in slight varus. Balance is anatomy-led: restoring native obliquity needs fewer ligament releases, and the aim is a more natural-feeling knee with less soft-tissue disruption.

The concern with KA. It may leave the tibia in varus and the joint line oblique, and the long-term effect of those positions on implant durability and survivorship is not established. That uncertainty is the reason MA is still the default.

View larger

Restricted kinematic alignment (rKA). Kinematic principles applied within defined safe boundaries: the native angles are recreated, but each cut and the overall HKA are capped within a few degrees of neutral. The aim is to gain KA's balance benefits while avoiding outlier positions.

Calipered femoral resurfacing. The distal femoral cut compensates for lost cartilage first: a shim restores the missing medial cartilage thickness, the cutting paddles contact both condyles, and the real-time display shows the resulting coronal and flexion alignment before bone is resected.

View larger

View larger

View larger

Functional alignment (FA). Typically robotically executed. The initial component position is set to native anatomy, then fine-tuned to optimise the soft-tissue envelope (gap balance) within boundaries, minimising releases. Like rKA, it is a bounded compromise between pure MA and pure KA.

View larger

View larger

Beyond the coronal plane. The philosophies are argued in the coronal plane, but an examiner expects the axial (rotational) and sagittal planes too, and malrotation is a key cause of the dissatisfied knee.

Femoral rotation sets the flexion gap and patellar tracking. Aim for one of:

• parallel to the surgical transepicondylar axis (TEA), medial sulcus to lateral epicondyle
• perpendicular to Whiteside's line, the anteroposterior trochlear axis
• about 3° of external rotation off the posterior condylar axis (the posterior condylar angle is ~3°)

KA instead sets femoral rotation to the patient's native flexion axis. Internal malrotation of the femoral component causes patellar maltracking/lateral subluxation, anterior knee pain and flexion instability/stiffness, a classic cause of the unhappy TKA.

Tibial rotation is referenced to the medial third of the tibial tubercle. Combined femoral and tibial internal rotation correlates with anterior knee pain (Berger).

The sagittal plane. Preserve an appropriate posterior tibial slope, native ~5 to 10°: too much risks flexion instability, and too little reduces flexion / over-tightens the PCL in a cruciate-retaining knee. Set the femoral component flexion to avoid anterior notching or over-flexion.

Most knees are not neutral. The Coronal Plane Alignment of the Knee (CPAK) classification was developed to describe the wide variation in constitutional limb alignment systematically. It phenotypes the knee; it is not an alignment technique. It combines two radiographic parameters: the aHKA, which estimates constitutional limb alignment (varus, neutral or valgus), and the JLO (apex distal, neutral or apex proximal).

The nine phenotypes. Three categories of each give nine CPAK phenotypes, numbered across the matrix. You need the common ones by name, because the literature now reports outcomes by type:

View larger

The boundaries. They are set at one standard deviation of the pooled 1,000-knee dataset and are examinable: neutral aHKA is 0° ± 2° (varus below -2°, valgus above +2°), and neutral JLO is 180° ± 3° (apex distal below 177°, apex proximal above 183°).

Why "apex distal" and not "varus". The vocabulary is deliberate. A "varus joint line" is ambiguous: it can arise from tibial varus in a neutrally aligned limb or from femoral valgus in a valgus limb, and reusing the limb-alignment words for an independent variable is what the classification exists to stop.

The argument in one row. CPAK Type V, neutral constitutional alignment with a neutral joint line, is precisely what mechanical alignment aims to create, and it is the natural state of only about 15% of knees. Conversely, 76% of healthy and 67% of arthritic knees have an apex-distal joint line, and around 32% and 30% respectively are constitutionally varus. It is worth saying out loud in a viva.

What that makes the case. For most patients mechanical alignment is a reconstruction toward a configuration they never had, rather than a restoration. That is the honest case for individualised alignment, and it is an anatomical case rather than an outcome one, which is exactly why it has proved easier to demonstrate in the operating theatre than in patient-reported scores.

What the validating work showed. Kinematic alignment achieved optimal intra-operative soft-tissue balance more often than mechanical across every phenotype, and the gap was widest where the theory predicts, in the constitutionally varus knees: Type I 100% with KA versus 15% with MA, and Type IV 89% versus 0%. These are intra-operative pressure measurements, not outcomes, and the same authors are careful to say the link between phenotype and patient outcome still needs establishing.

Alignment philosophy is one of the most actively debated topics in international knee arthroplasty. Practice is genuinely divided: mechanical alignment remains the most widely taught and registry-validated approach, while kinematic, restricted-kinematic and (with the spread of robotics) functional alignment have substantial and growing adoption — particularly in Australasia and Europe, where much of the constitutional-alignment and CPAK work originated.

National joint registries continue to validate the long-term survivorship of mechanically-aligned TKA, and long-term registry data for kinematic and functional alignment are still maturing. The pragmatic global position: mechanical alignment is the safe, validated default; individualised alignment is a reasonable, boundaried alternative that reduces soft-tissue releases and is increasingly executed with robotics, but its long-term superiority remains unproven and should be presented to patients as such.

The honest synthesis, which is what a viva rewards, is that the evidence is nuanced and, at the highest level, equivocal.

Short and medium term. Meta-analyses of randomised trials report better WOMAC, Knee Society Scores and flexion for kinematic versus mechanical alignment, and restricted kinematic alignment similarly improves some functional scores. Many of the differences are small, and they are not consistent across every outcome.

Long term. The best long-term randomised evidence shows equivalence. A Level I RCT with 10-year follow-up found no difference in any patient-reported outcome or in survivorship between kinematic and mechanical alignment, mirroring its own 2-year and 5-year results, and concluded that mechanical alignment remains the reference standard.

Functional alignment. In randomised data it is comparable to mechanical at 2 years while requiring fewer soft-tissue releases, with benefits that appear subgroup-specific (notably CPAK Type I knees).

Which scores move. The inconsistency is not random, and naming the pattern is what separates a good answer from a recited one. The positives in both meta-analyses sit on WOMAC and the Knee Society Score; the negatives sit on the Oxford Knee Score, the KOOS and the Forgotten Joint Score. That split is reproducible rather than noisy.

The Forgotten Joint Score is the telling one, because it measures the very thing individualised alignment claims to deliver: a knee the patient stops noticing. It is null three times over:

• in the restricted-kinematic meta-analysis
• as the designated primary endpoint of the functional-alignment randomised trial (64.4 versus 70.1, p=0.10)
• at ten years in the Level I kinematic-versus-mechanical trial

What to say about it. An intervention whose proposed mechanism is a more natural-feeling knee has repeatedly failed on the instrument built to detect a natural-feeling knee, while succeeding on older, broader scores. That does not make individualised alignment wrong, and the reduction in soft-tissue releases is a genuine and reproducible finding. It does mean the case for it currently rests on surgical process (fewer releases, easier balance, less need to correct a knee towards a target it never had) rather than on demonstrated patient benefit, and you should say so in exactly those terms.

The unhappy neutral knee. Do not blame a dissatisfied, textbook-neutral knee on its alignment by default. Work it up broadly first (infection, instability, malrotation, extensor mechanism, expectations), because alignment is only one possible contributor.