The low contact stress patellofemoral replacement consists of a trochlear component and a modular patellar component which has a metal-backed mobile polyethylene bearing. We present the early results of the use of this prosthesis for established isolated patellofemoral arthritis in 51 consecutive patellofemoral replacements in 35 patients. The mean follow-up was 25 months (5 to 60). The estimated survival rate at three years was 63% (95% confidence interval 47 to 80) with revision as the endpoint and 46% (95% confidence interval 30 to 63) with revision and ongoing moderate or severe pain as the endpoint.
The early results of the use of the low contact stress patellofemoral replacement are disappointing with a high rate of revision. We cannot therefore recommend its use.
Isolated patellofemoral involvement is seen in 10% to 24% of knees with osteoarthritis.1 Patellar instability and trauma may predispose to isolated patellofemoral arthritis, but idiopathic causes are the most common. In patients with symptoms and signs which warrant replacement, the options are total knee or patellofemoral replacement. Although total knee replacement (TKR) has consistently given good reported outcomes,2–5 isolated patello-femoral replacement offers the potential advantage of being a less invasive procedure while preserving the tibiofemoral articulation.
The low contact stress (LCS) patellofemoral prosthesis (DePuy Orthopaedics, Warsaw, Indiana) consists of a trochlear component and a modular patellar component with a metal-backed mobile polyethylene bearing.6 The mobility of the component offers the theoretical advantages of reduced wear and of self-alignment of the patella within the trochlear groove during flexion and extension of the knee, thus aiding patellar tracking.
The early results were reported by the originating centre,6 but there has been no independent report. Our aim therefore was to present the early results of the use of this prosthesis from two independent centres.
Patients and Methods
The LCS patellofemoral replacement was used by the two senior authors (PS, RP) at two independent centres between 2004 and 2008 for patients with symptomatic isolated patellofemoral arthritis as seen on pre-operative plain radiographs and at operation. The presence of superficial chondral damage at the tibiofemoral articulation did not preclude the use of the prosthesis. Patients with inflammatory arthritis were excluded from having patellofemoral replacement. The operation was undertaken using a medial parapatellar approach. The surgical technique has been previously described.6 It involves using an appropriately sized template to outline the position of the trochlear component. All the cartilage and part of the subchondral bone within this outline is then removed to allow sitting of the implant. The patellar articular surface is resected and the patellar template peg holes are drilled. At the end of the operation, a lateral retinacular release was performed if it was thought to aid patellar tracking.
The patients were identified from the operating-theatre list records. Their notes were reviewed to identify any prosthesis which had been revised or had been listed for revision. Those patients who had not been revised were recalled for clinical and radiological assessment. Clinical evaluation consisted of the modified Oxford knee score (OKS),7 the American Knee Society (AKS) score8 and the Melbourne patellar score,9 as well as pain (classified as none, mild, moderate or severe), the perception of patellofemoral clicking and instability, determination of the ability to perform activities of daily living and satisfaction measures. We used the modified OKS,7 ranging from 0 (worst outcome) to 48 (best outcome). Radiological evaluation included anteroposterior weight-bearing, 70° lateral and 30° skyline views. The severity of osteoarthritis was assessed radiologically using the grading of Altman et al10 and the presence and progression of any radiolucent lines after patellofemoral replacement was noted from plain radiographs. Patellar tilting was assessed on the 30° skyline view by measuring the angle between a line drawn parallel to the metal tray of the patellar component and a line drawn between the two highest points of the trochlear component (Fig. 1⇓).
There were 29 women and six men (51 knees) with a mean age of 63.8 years (47 to 84). All the knees had primary patellofemoral osteoarthritis. No patients with post-traumatic arthritis were included.
Survivorship analysis was performed using the Kaplan-Meier method with 95% confidence intervals (CI). Survival times for patients who were known to have a satisfactory outcome up to a certain point in time, but whose subsequent outcome was unknown, were censored at the last known outcome. Median values and inter-quartile ranges (IQR) were calculated for the various outcome measures using SPSS software (SPSS Inc., Chicago, Illinois). The log-rank test was used to examine the effect of age (≤ 65 or > 65 years) on survival. The independent samples Kruskal-Wallis test was used to compare the patellar tilting angle and Fisher’s exact test to compare the proportion of knees showing radiolucent lines between the revised and non-revised groups. Statistical significance was set at a p-value ≤ 0.05.
The median follow-up was 25 months (5 to 60). Of the 51 knees, 17 (33.3%) were revised. Of these, 16 were revised to a TKR and one had an isolated revision of the patellar component. The median time to revision after primary surgery was 20 months (5 to 54). In 12 knees revision was performed for pain, in three for patellar instability and in two for acute dissociation of the mobile patellar bearing (Fig. 2⇓). There were no cases of infection. The median patellar tilting angle on the radiographs before revision was 7.3° (IQR 1.5° to 10.4°) in those knees which were revised. None of the revised knees had radiolucent lines and they were all found to be well fixed at revision. Of the knees revised for pain, the polyethylene bearing was found to have reduced mobility in six as a result of overgrown surrounding soft tissue as seen at revision (Fig. 3⇓). In three knees, extensive metallosis of the capsule and synovium was seen, in two the polyethylene was found to be worn and fractured (Fig. 3⇓) and in one the trochlear component had a longitudinal groove along its lateral facet (Fig. 4⇓). In two knees the arthritic process had progressed and there was advanced tibiofemoral osteoarthritis. One knee which had not been revised required lateral retinacular release for persistent patellar maltracking. A total of 28 knees which had not been revised or listed for revision were examined clinically. In addition, two knees were reviewed by completion of a postal OKS questionnaire and by reference to their last clinic letter, and two by the use of their last clinic letter to determine pain and survival. One patient with bilateral prostheses had died and the last clinic letter was used to determine the level of pain and the function of the implant. Their clinical results are shown in Tables I⇓ and II⇓.
Radiographs were unavailable for three knees. Radiological analysis showed tibiofemoral osteoarthritis in 18 of 31 (58.1%) knees (Altman grade 1 in seven, grade 2 in six, grade 3 in one, and grade 4 in four) which had not been revised. Radiolucent lines around the prosthesis were seen in five of 31 (16.1%) knees, but were not progressive when compared with earlier radiographs. The median patellar tilting angle at review was 4.9° (IQR 3.4° to 6.7°) in the knees which had not been revised. There was no significant difference in the proportion of knees showing radiolucent lines (Fisher’s exact test, p = 0.111) and in the patellar tilting angle (Kruskal-Wallis test, p = 0.281) between the revised and non-revised groups.
The estimated survival rates are presented for up to five years (Tables III⇓ and IV⇓, Figs 5⇓ and 6⇓). The estimated survival rate at three years was 63% (95% CI 47 to 80) when revision was considered to be the endpoint and 46% (95% CI 30 to 63) with revision and ongoing moderate or severe pain as the endpoint. The median survival of this prosthesis was 40 months (95% CI 26 to 54) using revision as the endpoint and 35 months (95% CI 26 to 44) with revision and ongoing moderate or severe pain as the endpoint. There was no significant difference (log-rank test, p = 0.571) in the survival of the prosthesis between patients aged up to and including 65 years and those older than 65 years at the time of surgery.
Patellofemoral replacement has been used as an alternative to TKR in patients with patellofemoral arthritis without involvement of the tibiofemoral articulation. It was first performed by McKeever in 195511 and since then several implants have been developed. Although early results were disappointing because of persistent patellar instability, polyethylene wear and progression of tibiofemoral arthritis,12–14 more recent reports have been more promising. Kooijman, Driessen and van Horn15 described the outcome of the Richards (Smith & Nephew, Memphis, Tennessee) patellofemoral prosthesis in 56 knees. Excellent or good results were obtained in 86% of knees at a mean of 17 years. Because of ongoing tibiofemoral osteoarthritis, two patients required a high tibial osteotomy and ten had a conversion to TKR after a mean of 15.6 years. Similarly, Cartier, Sanouiller and Khefacha16 using the Richards II and III implants in 79 knees reported that 75% of the pros-theses were functioning at a minimum of six years (mean 10 years). The main cause of failure was progression of tibiofemoral osteoarthritis.
More recently the use of the Avon (Stryker Howmedica Osteonics, Allendale, New Jersey) patellofemoral replacement was reported by Ackroyd et al17 from the originating centre with a survival rate at five years of 95.8% (95% CI 91.8 to 99.8) with revision as the endpoint in 109 consecutive replacements. The main complication was radiological progression of tibiofemoral arthritis which occurred in 25 of the 90 knees reviewed (27.8%). More recently, the excellent mid-term results of this prosthesis have been confirmed independently by Odumenya et al18 who described a cumulative survival rate of 100% at five years in 50 Avon patellofemoral replacements. Starks, Roberts and White19 examined 37 Avon patellofemoral replacements at a mean of two years and reported a success rate of 86%, defined as an AKS objective score ≥ 80/100. Only two knees required further surgery in that series.
The LCS patellofemoral prosthesis consists of a trochlear component and a modular patellar component which has a metal-backed mobile polyethylene bearing. The designing surgeon reported promising early results based on an activities of daily living scale.6 At a mean follow-up of 45 months, eight knees were described as having an excellent, eight a good and four a fair outcome. Our study seems to have been the first to evaluate this prosthesis independently. Our results are disappointing in that 17 (33%) of the 51 prostheses were revised and 11 knees had continuing moderate or severe pain. This gave an estimated survival rate at three years of 63% (95% CI 47 to 80) with revision as the endpoint and 46% (95% CI 30 to 63) with revision and ongoing moderate or severe pain as the endpoint.
Several reasons can be postulated for the high failure rate of the LCS prosthesis in our study, taking into account the design of this prosthesis and findings at revision. The modular mobile patellar component appears to be an intrinsic weakness which accounted for most of the complications. The polyethylene bearing rotates on its metal base in a proximal to distal direction, exposing the underlying metal base at the extremes of movement. This could allow metal-on-metal articulation of the patellar and trochlear components, the metal wear and metallosis, with resultant chronic inflammation and pain. Extensive metallosis was a prominent finding in those knees which were revised. The polyethylene bearing was intended to be mobile on the metal base so that its movement would aid patellar tracking, but at revision the polyethylene bearings were found to have reduced or no mobility because of overgrowth of soft tissue. Having a modular patellar component also raises the possibility of dissociation of components which occurred in two cases. Four other such events have been previously reported in the literature.20,21 It may be that patellar maltracking with catching of the polyethylene margin on the edge of the trochlear component precipitates this complication.
Progression of tibiofemoral arthritis has been previously cited as the main reason for failure in patellofemoral replacement.17 However, in our study only two knees were found to have progressive severe tibiofemoral osteoarthritis, suggesting that this accounted for only a minority of the revisions. Tibiofemoral arthritis was one of our exclusion criteria for undertaking this procedure. Nevertheless, we carried out isolated patellofemoral replacement in the presence of superficial chondral damage in the tibiofemoral articulation, since we felt that such chondral damage is common and is not a predictor of full-thickness lesions of the articular cartilage.
It may be argued that poor surgical technique had influenced the longevity of this prosthesis. However, all the patients included in our study were operated on by or under the direct supervision of the two senior authors (PS, RP), both of whom have extensive experience in knee surgery. The operative technique recommended for the LCS patellofemoral prosthesis6 was observed in all knees. The main weakness of our study was a lack of pre-operative scores in our patients and also of pre-revision scores in those knees which underwent revision. Nevertheless this is the first independent report of the outcome of the use of the LCS patellofemoral prosthesis with detailed intra-operative examination and determination of the possible causes of failure in those requiring revision.
In conclusion, the early results of the LCS patellofemoral replacement are disappointing with a high rate of revision. We cannot therefore recommend the use of this device.
A further opinion by Dr D. Karataglis is available with the electronic version of this article on our web-site at www.jbjs.org.uk/education/further-opinions
We would like to thank K. Howard for help with the collection of data.
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.
- Received September 14, 2010.
- Accepted January 6, 2011.
- © 2011 British Editorial Society of Bone and Joint Surgery