Knee replacement resurfaces the joint surfaces worn away by arthritis. It is considered once pain no longer responds to conservative treatment and limits walking, sleep and daily activity. The operation depends on preoperative planning from a weight-bearing X-ray of the whole limb, because it is the alignment achieved — more than the implant itself — that determines how well the knee will function and how long it will last.
Other conditions I treat
- Meniscal tears — from a twisting injury, typical in sport, or from progressive degeneration after the age of forty. Not every tear needs surgery: the decision depends on the pattern, the location, your age and the state of the cartilage.
- Anterior cruciate ligament injury — reconstruction is mainly indicated in younger or active patients, or where instability limits daily life.
- Failed knee replacement — revision for loosening, polyethylene wear, instability, stiffness, periprosthetic fracture or infection.
Personalised alignment
CPAK and kinematic alignmentFor decades knee replacement had a single geometric goal: a perfectly straight mechanical axis for everyone. It worked — it is why the operation is one of the most reliable in orthopaedics. But a consistent share of patients report a knee that never feels like their own: stiff, functioning but not natural.
Part of the explanation lies in the starting point. Many people are slightly varus or valgus by constitution, not because of arthritis — it is their shape and their ligaments have adapted to it over a lifetime. Straightening that limb asks those ligaments to work in a position they have never known.
The CPAK classification
Published in 2021 after analysing weight-bearing radiographs of 500 healthy and 500 arthritic knees, the CPAK classification — Coronal Plane Alignment of the Knee — describes individual anatomy by combining two measurements: the constitutional alignment of the limb and the obliquity of the joint line. Nine phenotypes result. Crucially, the distribution is essentially the same in healthy and arthritic knees: the constitutional shape of a knee is not created by arthritis, it was there before.
The planning software I developed
I built a program that applies this method case by case. Eleven anatomical landmarks are placed on a weight-bearing radiograph; from these it derives the individual angles, the CPAK phenotype and — most importantly — the exact angles of the two bone cuts. Once the target alignment is set and the tibial cut chosen, the femoral cut is no longer a matter of feel: it is determined.
Ligament balance under tension remains the decisive intraoperative check — roughly 1 mm of gap asymmetry in extension corresponds to about 1° of femoral cut — but it becomes the verification of a decision made in advance, rather than the criterion by which it is improvised in theatre.
The full technical account, with tables and references, is available on the Italian page: Progettazione protesica.
What this method does not promise
Honesty- The evidence is short to medium term. Systematic reviews indicate that restricted kinematic alignment is clinically non-inferior to mechanical alignment over the short and medium term. The twenty- and thirty-year data that exist for mechanical alignment do not yet exist for these techniques.
- Not every patient benefits equally. The balancing data show clear advantages in some phenotypes and no significant difference in others.
- Software does not operate. The program calculates from landmarks I place myself: the quality of the result depends on those landmarks, on the radiograph and on the execution.
- Severe deformity remains a limit. Beyond the safe zone the alignment is corrected, and the decision returns to being a clinical compromise.
The real advantage is not the promise of a perfect knee, but a way of deciding that is transparent and reproducible: the numbers are explicit, checkable before surgery and comparable with the result obtained.