The skin incision for TKA should allow easy access to the chosen deep approaches, be extensile and, where appropriate and possible, incorporate previous incisions. A sound understanding of the blood supply of the skin around the knee joint is important and should be respected. The saphenous and lateral descending geniculate arteries are the main supply to the medial aspect of the knee [5, 6]. If more than one previous incision is present and possible to use, the more lateral one is less likely to cause skin breakdown. If previous scars are present from many years earlier, and are inappropriate for incorporation into a new incision, then an assumption can be made that a new midline incision can be safely used.

The anterior midline incision is the gold standard as it provides access to the medial and lateral aspects of the knee. It is usually centred over the medial edge of the patella, for the more common medial parapatellar arthrotomy and over the lateral margin of the patella and patellar tendon if a lateral arthrotomy is planned. This incision is extensile and lets the surgeon perform the conventional approaches. Healing rates and tensile strengths of the subsequent scars are theoretically improved if incisions are parallel to the Langer’s lines, and a more curved incision may achieve this [6]. The length of the skin incision will vary depending on the patient and surgeon and should allow clear exposure of the deep approach, without undue tension on skin edges. It is important to elevate full-thickness flaps to maintain adequate vascularity, and flaps should only be elevated sufficient to expose the deep approach. The arthrotomy is usually offset from the skin incision to avoid a direct pathway for communication with the external environment.

This is the most commonly used approach for TKA and is an excellent extensile approach that can be used in virtually all primary knee arthroplasty. Following the skin incision, the arthrotomy will divide extensor mechanism, capsule and synovium in one layer and will usually begin approximately 5 cm proximal to the superior pole of the patella, leaving a 5–10 mm cuff of tendon attached to the vastus medialis obliquus (VMO) and distally around the medial edge of the patella and patellar tendon to the level of the tibial tubercle. A small cuff of retinaculum should be left on the patella, and paratenon on the tendon, for later closure. A transverse pen marking can be made at the level of the patella prior to the arthrotomy to guide later closure. The fat pad is partially excised, sufficient for exposure, and the anterior horns of both menisci divided. A release of the anterior part of the medial collateral ligament (MCL) attachment is then done sufficient to expose the tibial plateau to the level of likely resection. The patella is then either everted or simply dislocated laterally. Studies comparing patellar eversion to dislocation without eversion have suggested that the latter approach results in reduced postoperative pain and improved quadriceps recovery. Excision of fat and synovium immediately proximal to the trochlea is done sufficient to expose the site of anterior resection. If necessary, improved exposure can be achieved by extending the medial release, particularly in cases of fixed varus deformity, or by extending the quadriceps incision proximally.

This approach is considered by proponents to be more “biological” than the medial parapatellar arthrotomy as it does not violate the quadriceps tendon and lifts the vastus medialis muscle as a whole. The skin incision is the standard midline, and the inferior border of the vastus medialis obliquus is then identified. The arthrotomy begins by incising the fascial sheath along this border from the patella medially to the intermuscular septum. The incision then continues distally along the medial patellar retinaculum along the medial border of the patella and along the medial patellar tendon to the tubercle. The vastus medialis is then peeled off the intermuscular septum, and the capsule of the suprapatellar pouch is then incised proximally to gain access to the joint and dislocate the patella.

This approach has the added advantages of preserving vascularity and causing less pain postoperatively although it may be more difficult to visualise the surgical field, particularly in larger patients or in cases of significant deformity, for whom this approach would therefore usually be unsuitable [7, 8]. Patients who have limited motion, severe flexion contracture, severe valgus alignment, patella baja or obese are not good candidates for this approach. Although this approach is technically more demanding, in appropriate patients it has been shown to have faster functional recovery, less postoperative pain and better knee flexion than the medial parapatellar approach [9, 10].

The lateral approach is for many surgeons the preferred choice in valgus deformities. It affords an excellent exposure of lateral structures that may require release for deformity correction and also preserves the medial ligament. It is particularly useful in cases of severe fixed valgus deformity, with lateral patellar subluxation. Potential difficulties include posteromedial exposure and a tendency to excessively externally rotate the tibial component. The skin incision is anterior longitudinal, just lateral to the midline and centred over the lateral margin of the patella. The arthrotomy begins proximally in the quadriceps tendon 5–10 mm medial to the vastus lateralis, from approximately 5 cm proximal to the patella distally around the lateral margin of the patella to the lateral edge of the tibial tubercle. The fat pad is partially excised for exposure, but it is beneficial to preserve as much fat pad as possible as this will assist in closing the defect lateral to the patellar tendon at the end of the procedure. The patella is dislocated medially, and the iliotibial band and lateral capsule can be released under direct vision as necessary for deformity correction.

For most implants this corresponds to 8–10 mm from the less affected surface. For example, in a varus knee the less affected side is lateral, and this serves as a guide to the correct joint line. This cut selectively influences the extension gap, and the amount resected can be adjusted depending on the surgeon’s desire to influence extension. In most cases a matched resection is performed, but additional resection may be considered in cases of fixed flexion deformity, particularly in cases with unrestricted flexion such that a preoperative flexion–extension imbalance is evident. Increased resection will raise the joint line, with implications for ligament balancing and patellar height.

This cut selectively influences the flexion gap. In most cases, a matched resection is indicated to restore posterior offset whilst avoiding overstuffing the flexion gap and restricting flexion. Increased resection can be considered in cases of relative flexion tightness, particularly after checking balancing after initial standard cuts, but excessive resection should be avoided as this will lead to flexion instability.

This cut will directly influence the anterior compartment and associated patellofemoral joint. In general the proximal limit of the cut should be flush with the anterior cortex of the femur, and whilst it is often not possible to perform a perfectly matched resection with this cut, care should be taken to remove at least as much bone as will be replaced by the prosthesis.

These cuts do not directly affect bony gaps and will vary between different prosthetic geometries.

This cut equally affects flexion and extension gaps. The depth of the cut is usually 8–10 mm, but will be varied depending on implant thickness and patient factors. In particular, if the knee has equivalent significant restriction in extension and flexion, increased tibial resection will be beneficial in restoring range of motion.

This concept prioritises the anterior cut using a point on the anterior femoral surface as a reference, in order to achieve a flush cut but avoid femoral notching; thus, the posterior resection is variable. If the assessment is between sizes, the smaller one is selected to avoid overstuffing the posterior space. This has the potential effect of a larger posterior resection, with a larger flexion gap and possible flexion instability.

This concept prioritises a matched resection posteriorly to restore posterior offset and avoid flexion instability. The posterior femoral condyles are the reference, and if between sizes the traditional recommendation is to choose the larger size in order to avoid femoral notching anteriorly. In practice however, in most cases it is usually possible to downsize and accept a small femoral notch anterolaterally. This is unlikely to increase risk of periprosthetic fracture and achieve the goals of restoring posterior offset without overstuffing anteriorly.

The primary principle of this technique is balancing flexion and extension gaps, with prosthetic alignment, particularly positioning of the femoral component secondary. After osteophyte excision the tibial cut is performed first, and referencing for positioning of the femoral cutting block is then taken from the tibial cut and directed by ligament tension. Having completed the posterior femoral cut, the resultant flexion gap is assessed. The knee is then brought into extension and the distal femoral cut completed at a depth to match the previously measured flexion gap. Having done this, the gaps are usually well balanced, and femoral finishing cuts are completed. Potential limitations of this technique are that it relies on the integrity of the collateral ligaments and an accurate tibial cut, as minor alterations in the tibial cuts could affect femoral component rotation and joint line. This should not be a significant problem for surgeons who have a good understanding of appropriate femoral component rotation and how to assess this intraoperatively.