Resurfacing versus not-resurfacing the patella in one-stage bilateral total knee arthroplasty: a prospective randomized clinical trial

This prospective randomized clinical trial assessed patients with KOA undergoing first bilateral TKA in the affiliated hospital of Qingdao University, from March 2011 to August 2012. The inclusion criterion was bilateral knee OA. The exclusion criteria were previous patellectomy, inflammatory arthritis, patellar fracture, patellar instability, previous extensor mechanism procedures, high tibial osteotomy, severe valgus or varus deformity (> 20°), severe flexion contracture (> 30°), previous unicondylar knee replacement, and a history of septic arthritis or osteomyelitis.

The study protocol was approved by the ethical institutional review board of the affiliated hospital of Qingdao University. Signed informed consent was provided by each patient.

A statistical expert blinded to the research procedure generated a random number sequence using a computer. To conceal randomization outcomes, the allocated numbers were placed into concealed opaque envelopes. Randomization was accomplished by opening a randomly selected envelope in the operation room after femoral and tibial cuts, immediately prior to patellar preparation. The left knee received the treatment indicated by the envelope, while the contralateral knee (right knee) received the alternative treatment; either way, the surgery started with the left knee. Thus, all patients had one patella resurfaced and the contralateral patella non-resurfaced.

All patients were operated by a single surgeon (Kang Sun, one of the authors) using Scorpio NRG knee prostheses (Stryker, USA). The tourniquet was tied to the proximal thigh pre-operatively. A midline skin incision and the medial parapatellar approach were used, preserving the infrapatellar fat pad. Tibial bone cuts as well as distal and posterior femoral bone cuts were performed according to mechanical and anatomical axis measurements in each patient. Femoral component rotation was oriented parallel to the transepicondylar axis. The rotation reference for the tibial component was the medial half of the tibial tuberosity. The patellar cut was performed with an oscillating saw on the patellar clamp. Calipers were used to measure the patellar thickness intra-operatively in an attempt to restore the baseline composite height in all resurfacing procedures. Patellar resurfacing was undertaken with a cemented, inset domed component. When resurfacing was not performed, the so-called patelloplasty was carried out, including osteophyte removal and smoothing of fibrillated cartilage. An assessment of patellar tracking was carried out by the no thumb test; if necessary, a lateral retinaculum release was performed. Patellar thicknesses before and after resurfacing were measured.

Post-operative drains were used in all cases and removed on the first post-operative day. Pain control was achieved with epidural patient-controlled analgesia for the first 24 post-operative hours, followed by oral analgesics, as tolerated. The patients were administered antibiotics (cefazolin sodium salt 0.5 g/bid) by intravenous drips for three days to prevent infection. For thromboprophylaxis, all patients were administered subcutaneous low molecular weight heparin (LMWH) at a dose of 4000 AxaIU (0.4 ml)/day, starting 12 hours after the operation. All patients were administered physical rehabilitation therapy by the same rehabilitation technologist, as reported previously [8].

The patients were followed post-operatively at three months and annually thereafter. An evaluator not involved in study design but familiar with the assessment tools was responsible for data collection. The evaluator and patients were blinded to the surgical procedure. The primary outcome was the Knee Society System (KSS) score; the system consists of a 100-point scale for clinical status and a 100-point score for function [9].

Secondary outcomes included the Feller score, VAS score for anterior knee pain, patellar clunk and function, and patient satisfaction. Patellar function was evaluated using the Feller score [10]. Anterior knee pain rate as well as patellar clunk and crepitus were assessed. Anterior knee pain was evaluated during a simulated activity of daily living; pain intensity was rated using a visual analog scale (VAS) ranging from 0 to 10 points, with 0 being no pain and 10 representing maximum pain [11]. A score above 5 was defined as anterior knee pain. Patient satisfaction was assessed with questionnaires at each follow-up visit.

Alignment of the TKA components was evaluated by measuring the standard anteroposterior and lateral X-rays of the knee.

All patients received X-ray examination (anteroposterior, lateral, and axial views) before surgery, as well as at three months and one, two, three, four and five years after surgery. For the radiographic assessment of the patellofemoral joint, on the lateral radiograph with the knee in 30° flexion, the Insall-Salvati index was used to calculate the patellar height. The patellar tilt was measured by using the axial view of the knee, measuring the angle between the line tangent to the two femoral condyles or the femoral component, and the line joining the medial and lateral ends of the patella or the line tangent to the base of the patella component. Patellar subluxation was assessed as a percentage compared to the mid-lateral dimension of the femur or femoral component. The radiographs were analyzed by Qicai Li (associate chief physician, one of the authors).

All data analyses were performed with SPSS 19.0 (IBM, Armonk, NY, USA). The sample size was estimated according to a power analysis [12], as follows. A 10-point difference in the 100-point clinical KSS was considered the endpoint of statistical analysis. With an alpha value of 5% and a power of 95%, a sample size of 50 patients (100 knees for both groups) was required. Finally, to compensate for potential loss to follow-up, we recruited a total of 85 patients. Continuous variables were expressed as means ± standard deviation (SD) and compared using the paired t test in case of normal distribution confirmed by the Kolmogorov-Smirnov test. The Wilcoxon signed-rank test was used to assess the non-parametric data. The Greenhouse-Geisser test was performed to compare repeated measures data such as KSS and Feller scores. Categorical data, such as anterior knee pain and patellar clunk and crepitus, were compared by the McNemar-Bowker test or the Fisher exact test. P < 0.05 was considered statistically significant.

During the enrollment period, 85 patients with degenerative OA were evaluated for possible inclusion in the study. A total of 15 patients did not meet the inclusion criteria. The 70 patients who met the inclusion criteria were enrolled in the study and 66 patients provided informed consent. One patient died of cerebral haemorrhage within five years of follow-up and four were lost to follow-up. One patient developed severe dementia and was excluded. Finally, 60 patients were included in the final analysis. The study flowchart and the numbers of knees included in the current analysis were outlined in a Consolidated Standards of Reporting Trials (CONSORT) diagram (Fig. 1).

The patients in the study cohort were 65.2 ± 5.4 (range 58–70) years old. There were 38 women and 22 men. BMI was 23.8 ± 4.3 (range 18.9–28.2) kg/m². The patients’ baseline characteristics are summarized in Table 1.

The lateral retinaculum was released in no patients. The patellar resurfacing procedure affected the operative time; compared with the group without patellar replacement, TKA operation with patellar replacement showed an operation 3.48-min longer (P < 0.001). Mean peri-operative blood loss was 36.18 ± 3.46 ml in the patellar resurfacing group and 36.20 ± 3.39 ml in the non-resurfacing group, indicating no statistically significant difference between the two groups (P = 0.957). Post-operatively, mean patellar thicknesses were 21.08 ± 1.39 mm and 22.03 ± 1.68 mm in the patellar resurfacing and non-resurfacing groups, respectively. The mean difference of 0.95 mm was statistically significant (P < 0.001) (Table 2).

There were no significant differences in pre-operative KSS (clinical and function) scores between the resurfacing and non-resurfacing groups, as shown in Table 3 (P = 0.31 and P = 0.15, respectively). Both groups showed significant improvements in both KSS scores after surgery; the time effect was significant (both P < 0.001). There were significantly better scores for the resurfaced sides compared with the non-resurfaced side at annual follow-up; the main effect of group was significant (both P < 0.001).

There was no significant difference in pre-operative Feller scores between the two groups as shown in Supplementary Table 2 (P = 0.959). Both treatment groups had significant improvement in Feller scores post-operation; the time effect was significant (time effect P < 0.001). Feller scores were obviously higher in the patellar resurfacing group compared with the non-patellar resurfacing group at three months and at one, two, three, four and five years after surgery. The differences were statistically significant, indicating that patellofemoral joint function in the resurfacing group was better than that of the non-resurfacing group (main effect P < 0.001) (Supplementary Table 2).

Persistent anterior knee pain (AKP) was noted in 14 patients (23%) on the non-resurfaced side, whereas three (5%) patients reported such pain on the resurfaced side within three months of surgery. With time, the rate of AKP gradually decreased in both groups. One patient had similar pain intensity in the bilateral knee at the last follow-up. No patients with persistent pain showed symptoms severe enough to require further surgery (Supplementary Table 3).

At the final follow-up, six (10%) and 24 (40%) patients showed patellar clunk syndrome in the patellar resurfacing and non-patellar resurfacing groups, respectively, indicating a statistically significant difference (P < 0.001) (Supplementary Table 3). Two patients suffered from bilateral patellar clunk syndrome and had severe valgus alignment pre-operatively, although the mechanical axis was corrected post-operatively. In addition, post-operative patellar thickness in one patient with the patellar clunk syndrome in the patellar resurfacing group was greater than pre-operative patellar thickness (1 mm). Patellar component size in three patients with the patellar clunk syndrome in the patellar resurfacing group was below 38 mm.

Pre-operatively, the mean pre-operative mechanical axis was − 5.65° ± 1.02° (varus) on the patellar resurfacing side, compared to − 5.45° ± 0.98° (varus) on the non-patellar resurfacing side (P = 0.71). Five knees had valgus alignment > 10°, including 3 and 2 in the non-resurfaced and resurfaced groups, respectively. Post-operatively, the mechanical axis in all patients was corrected to 5.82° ± 1.13° (valgus) on the patellar resurfacing side and 5.78° ± 1.04° (valgus) on the non-patellar resurfacing side (P = 0.91). Post-operatively, alignment of the femoral and tibial prosthesis components showed no significant difference between the two groups (P = 0.91). Post-operative patellar tracking was considered to be clinically satisfactory in all cases. Mean patellar tilt was greater in knees with patellar resurfacing compared with the non-resurfacing group, but the difference was not significant (P = 0.68). The Insall-Salvati index in the two groups showed no significant difference after TKA (P = 0.93). There was no change in joint line position by > 5 mm in either group (Supplementary Table 4).

No patients at the last follow-up had required revision for patellofemoral problems. There were no cases of patellar subluxation or dislocation, rupture of the quadriceps tendon, aseptic component loosening, patellar osteonecrosis, patellar fragmentation, or periprosthetic fracture.

Regarding subjective preference, 12 (20%) patients affirmed to prefer the resurfaced side and six (10%) preferred the non-resurfaced counterpart, while 42 (70%) expressed no preference at three months after operation. However, with follow-up time, more and more patients preferred the resurfaced side (47%) and only 7% patients preferred the non-resurfaced side at the final follow-up.

Figures 2, 3, and 4 present the case of a woman of 68 years of age with bilateral KOA.