Computer-assisted total knee arthroplasty using mini midvastus or medial parapatellar approach technique

A prospective, randomized, international multicentre trial including 69 patients was performed according to a standard protocol to compare computer-assisted TKA using either a mini-midvastus (MIS group) or a medial parapatellar approach (conventional group). Patients in 3 centers (Maastricht, Zwickau, Adelaide) with end-stage osteoarthritis of knee were randomized to either an MIS group with dedicated instrumentation or a conventional group.

Ethical approval of all centres had been obtained from the local ethical committee of Maastricht (MEC 04-105), Adelaide (RGH 10/04) and Zwickau (EK-MPG-0603) as part of the research program “A prospective comparative, randomized study comparing the MIS computer navigated total knee arthroplasty vs. conventional computer navigated total knee using the Scorpio CR fixed bearing knee and the Stryker navigation system”.

Trial Registration Number: ClinicalTrials.gov NCT02625311 8 December 2015.

Patients were randomized (random permuted blocks of 4) in either the MIS group or the conventional group. A written informed consent was obtained from all participants.

Seventy six participants were included for the study. There were 4 intra-operative exclusions: 3 patients due to problems with the navigation trackers and one patient due to fracture of the tibial plateau.

Three patients were lost to follow up due to unwillingness of the participants. Sixty nine patients completed the 6 months study period.

All cases were performed by a single surgeon in each of the three centres. Before the study, the surgeons participated in training involving multiple cases of cadaver prosthetic implants using the navigation system with the MIS approach and a minimum of ten clinical cases.

Patients between 45 and 75 years of age who had an established diagnosis of knee osteoarthritis requiring primary total knee replacement. Exclusion criteria included previous cruciate ligament reconstruction, correction osteotomy of the tibia, patellectomy, BMI greater than 30, flexion contracture greater than 15°, varus or valgus deformity greater than 15°, medio-lateral instability greater than 10° and active inflammation or infection of the knee. In addition, patients were excluded if they were pre-operatively considered to require patellar surface implantation.

The Stryker (Stryker Howmedica Osteonics, Allendale, NJ USA) Navigation System II, version 3.1 was used in all cases. This is an image free, active, cordless system. All knee surgeries were performed using a tourniquet. In the MIS group a 10 cm incision with a flexed knee, mini-midvastus approach was applied per standard protocol and with dedicated instrumentation. In cases where conventional surgical technique was used, a medial parapatellar approach was applied. The navigation system trackers were then attached to the surface of the femur and the tibia. The hip joint rotation center and the center of the ankle joint were established as reference points for leg axis. The rotational position of the femoral component was determined by using the Whiteside’s line and the transepicondylar line (TEA). Tibial rotation was assessed based on the relative positions of the centre of the ankle joint and the medial one third of tibial tubercle. The navigation was accorded with the neutral mechanical axis of the extremity with tibial slope fixed at four degrees. Each resurfacing plane angle was instrumented with dedicated navigation cutting guides and checked with the navigation system following the osteotomy. With both techniques, after determining proper prosthetic size, the collateral ligaments were balanced as required based on ligament tension assessed during functional testing of the prosthetic implant. Patellar surface implantation was not performed. The femoral component was implanted without cementing, whereas the tibial component was cemented with Simplex P (Stryker Howmedica Osteonics, Allendale, NJ USA) containing antibiotics. In each case, a Scorpio (Stryker Howmedica Osteonics, Allendale, NJ USA) CR fixed bearing implant was used without patellar surface implantation.

Clinical outcomes were assessed by a blinded independent examiner. All outcome parameters were assessed preoperative and postoperative at 7 weeks, 3 and 6 months. Pain and ROM were also measured weekly between postoperative weeks two to seven.

The primary outcome was to compare postoperative pain and range-of-motion (ROM) of both groups. Pain was assessed using a Likert score [52]. Range of motion (ROM) was measured using a goniometer according to the technique described Norkin [53]. Intra-tester and inter-tester reliability was described by Brosseau [54], the reproducibility by Lenssen [55].

The secondary outcome was to compare duration of surgery, blood loss, chair rise test, quadriceps strength test, anterior knee pain, KSS and WOMAC score. Duration of surgery was measured in minutes from skin incision to closure of the wound. Blood loss was measured intra-operative and during the first 24 h postoperative. All data was recorded in ml. Chair rise test was assessed according to the description of Jones [56]. The patients were sitting on a stool with the hip and knee in 90° of flexion. The patients had to stand up from the stool without using their arms. The test was repeated five times. Quadriceps strength test (fair/good – can break/good – can’t break) [57, 58] and anterior knee pain arising from a chair (yes or no) was assessed by the method described by Insall et al. [58] Knee Society Scores [58] and WOMAC scores [59] were also measured.

The primary hypothesis was that those patients who underwent MIS would benefit from less postoperative pain and higher ROM. The secondary hypothesis was that the use of computer navigation allows MIS-TKA to be performed without increased risk of limb mal-alignment more than 2° and outliners in component positioning.

The lower limb mechanical axis was measured on long standing radiographs preoperatively and at 3 months postoperatively [60]. Outliers were defined as a coronal mechanical leg alignment of more than 2° from neutral. CT scan was performed three months postoperatively with analysis of component alignment determined by the Perth protocol [61]. The position of femoral component was determined in sagittal, coronal and transverse planes, the tibial component was determined in sagittal and coronal planes. Outliners were defined as the component were positioned more than 2° different than the planned position. Mean values were used for further analyses.

Descriptive statistics were used to summarize the data. Categorical data were analyzed using Pearson Chi square test, likelihood Chi square tests or Fisher’s Exact tests. For continuous data Student’s t-test, or two-way ANOVA was used. Analyses were performed using SPSS v19.0. P-values < 0.05 were considered statistically significant.

Boxplots represent 10, 25, 50, 75, and 90 % of data. Outliers are shown as dots. Means are not presented in the boxplots.

There was no significant difference between the two surgical groups with respect to sex, age, BMI, side of operation or primary diagnosis at p > 0.05 (Table 1).

No differences in range of motion between groups were found at the different time points (two-way ANOVA, F(1) = 0.73; p = 0.12) (Table 3).

At 7 weeks, 3 months and 6 months follow-up, no differences between types of surgery in the ability to rise from a chair (yes/no) (p = 0.56, p = 0.38, and p = 0.10 for 7 weeks, 3 months, and 6 months, respectively; Table 5), between types of surgery in quadriceps strength (fair/good – can break/good – can’t break) (p = 0.60, p = 0.30, and p = 0.69 for 7 weeks, 3 months, and 6 months, respectively; Table 6), and between types of surgery in the presence of anterior knee pain (yes/no) were found (p = 0.59, p = 0.66, and p = 0.11 for 7 weeks, 3 months, and 6 months, respectively; Table 7).

At 7 weeks, 3 months and 6 months follow-up, there were no differences between groups in functional scores (KSS score (two-way ANOVA, F(1) = 0.43; p = 0.51, Table 8) and WOMAC (Two-way ANOVA; F(1) = 0.005; p = 0.94, Table 9).

Limb alignment coronal axis was achieved within a target of 2° equally in both groups (75 % of patients in the MIS group and 78.7 % of patients in the conventional approach group). The mean mechanical leg axis in the conventional group was 0.97° ± SD 1.87°, in the MIS group 0.54° ± SD 2.53° (Fig. 1). The difference of the mean mechanical leg alignment between the groups was not statistically significant (‐0.43° (95 % CI ‐1.50 - 0.64); p = 0.43).

There was no significant difference between the two surgical groups with respect to flexion/extension (‐0.83° (95 % CI ‐2.32 - 0.67); p = 0.27), rotational alignment of the femur component (‐0.45° (95 % CI ‐1.74 - 0.83); p = 0.49) or varus/valgus (‐0.18° (95 % CI ‐0.99 - 0.62); p = 0.65) and varus/valgus alignment (‐0.16° (95 % CI ‐1.26 - 0.95); p = 0.78) or posterior slope (1.00° (95 % CI ‐0.43 - 2.42); p = 0.16) of the tibial component (Figs. 2, 3, 4, 5 and 6).

The number of patients in our study is rather low and the sample size calculation is missing, however the findings of this study is in line with the findings of previous researches.

Although all of the three surgeons were experienced, high-volume knee surgeons, they had much more experience in medial parapatellar approach than mini-midvastus approach, since the conventional approach was used routinely in all clinics.

The authors did not determine the rotational position of the tibial component since it was not a primary objective of the study and there was no statistical significant difference between groups in terms of component positions in other planes. The rotational alignment of the tibial tray might be theoretically different between groups, however is not in line with the findings of previous studies.