Knee Surg Sports Traumatol Arthrosc (2012) 20:1216–1226 DOI 10.1007/s00167-012-1948-z
KNEE
Patellofemoral arthroplasty, where are we today? Se´bastien Lustig • Robert A. Magnussen Diane L. Dahm • David Parker
•
Received: 11 December 2011 / Accepted: 27 February 2012 / Published online: 10 March 2012 Ó Springer-Verlag 2012
Abstract Purpose Patellofemoral arthroplasty remains controversial, primarily due to the high failure rates reported with early implants. Numerous case series have been published over the years detailing results of various first- and secondgeneration implants. The purpose of this work is to summarize results published to date and identify common themes regarding implants, surgical techniques, and indications in order to maximize results of future procedures. Methods A comprehensive review of the MEDLINE database was carried out to identify all clinical studies related to patellofemoral arthroplasty. Results First-generation resurfacing implants were associated with relatively high failure rates in the medium term. Second-generation implants, with femoral cuts based on TKA designs have yielded more promising medium-term results. Surgical indications are specific and must be carefully followed to minimize poor results. Short-term complications are generally related to patellar maltracking, while long-term complications are generally related to progression of osteoarthritis in the tibiofemoral joint. Implant loosening and polyethylene wear are rarely reported. Short-term results are favourable for new techS. Lustig (&) D. Parker Sydney Orthopaedic Research Institute, Suite 12, Level 1, 445 Victoria Avenue, Chatswood, NSW 2067, Australia e-mail:
[email protected] R. A. Magnussen Department of Orthopaedic Surgery, The Ohio State University Medical Center, Columbus, OH, USA D. L. Dahm Department of Orthopedic Surgery, Mayo Clinic, Rochester, MN, USA
123
nology including custom implants and computer navigated surgery. Conclusions Overall, recent improvements in implant design and surgical techniques have resulted in improvements in short- and medium-term results. More work is required to assess the long-term outcomes of modern implant designs. Level of evidence IV. Keywords Results
Knee Patellofemoral joint Arthroplasty
Introduction Isolated patellofemoral arthritis is relatively uncommon [19], and various surgical options have yielded only marginal results [9]. Patellofemoral arthroplasty, although gaining in popularity [32, 41, 50], remains controversial because of inconsistent results in the literature [27, 34] with relatively high early failure rates in some reports [28, 64]. MacKeever [45] first proposed a vitallium patellar resurfacing implant in 1955, which, despite good initial results confirmed by the series of Levitt [36] and Vermeulen et al. [66], was quickly abandoned due to excessive wear of the trochlea. It was not until 1979 that the era of patellofemoral prostheses (PFP) really began with the development of the Richards prosthesis (Smith-NephewRichardsTM) [10]. Although these early implants produced mixed results, the design of implants has evolved to better reflect the native geometry of the distal femur. These design improvements have enhanced patellar tracking and reduced patellofemoral complications, resulting in better functional outcomes and implant survival.
Knee Surg Sports Traumatol Arthrosc (2012) 20:1216–1226
This article summarizes the long-term results of patellofemoral arthroplasty. It details the types of implants; surgical indications, clinical outcomes, and failure rates with various designs; and the potential implications of implant design factors on surgical results.
Patellofemoral prostheses Concept and design Patellofemoral prostheses attempt to reproduce the complex kinetics of patellar motion throughout knee flexion, allowing motion in the sagittal, axial, and coronal planes. The more constrained the design of the trochlea, the less natural motion is permitted, potentially contributing to increased stresses in the joint and increased polyethylene wear. Conversely, a less constrained patella will allow more freedom of motion but may increase the risk of patellar subluxation and dislocation. There are classically two main types of patellofemoral prosthesis based on the type of trochlear preparation: resurfacing prostheses and anterior cut prostheses. Resurfacing prostheses simply replace worn cartilage without significantly altering the shape of the subchondral bone. The positioning of these implants is dependent on the native trochlear anatomy. Anterior cut prosthesis, sometime referred to as ‘‘second-generation’’ prostheses are generally based on the femoral trochlear cuts of a total knee arthroplasty (TKA), completely replacing the patellofemoral compartment of the knee. Beyond this difference in design philosophy, one must also distinguish patellofemoral prostheses according to specific geometric criteria that are important to the fundamental biomechanics of the joint. The trochlear component can be asymmetric, with the lateral flange elevated to resist lateral forces exerted by the quadriceps [67]. Unlike symmetric trochlear designs, these implants require specific right and left components. The change in the sagittal radius of curvature of the trochlea and the amount of prosthesis coverage on the anterior femur influence the point at which the patellar button engages the trochlea [53]. The degree of flexion or extension in which the implant is placed can significant affect tracking, especially in implants with lower degrees of trochlear coverage. In contrast, some models include femoral coverage extending beyond the anatomical trochlea proximally onto the anterior femoral cortex, ensuring contact with the patellar button even in full extension [38]. Similarly, implants that extend further distally can prevent contact between the native articular cartilage of the femoral condyles and the patellar button in high degrees of flexion [4]. The shape of the trochlear groove is also highly variable among models,
1217
ranging from deep, highly constrained trochleas such as the Richards IIITM to more open, unconstrained designs such as the Avon prosthesis [2]. Amis et al. [4] showed significant variation in the trochlear angle for four models of prostheses. All currently available models utilize cobaltchrome trochlear components, the vast majority of which are cemented. Additional anchorage consists of pegs or small keels. The thickness of implants varies from 4 to 9 mm, with resurfacing models generally being the thinnest. Patellofemoral prostheses can also be distinguished by the shape of the patellar button, including domed, faceted, and asymmetrical designs. Spherical dome designs have theoretically the advantage of self-centring of the patella on the trochlea, absorbing any residual tilt [37]. These models can also generally accommodate to the trochlear shape of a subsequent TKA if necessary, eliminating the need to change the patellar button [37]. One model, the LCSTM prosthesis (De PuyTM), utilized a metal-backed patellar component [46] but was discontinued in 2009. Resurfacing prostheses The design of these prostheses is such that they simply replace the articular cartilage. They are embedded in the subchondral bone and their positioning is therefore dependent on the native anatomy of the trochlea. Some prostheses are symmetric, including the Richards III prosthesis (Smith-Nephew RichardsTM) [10] and Lubinus prosthesis (Waldemar LinkTM) [43]. The Richards III prosthesis has a symmetric trochlea and a very deep trochlear groove, with a trochlear angle of 135°. The patella button is faceted, prohibiting self-centring. There is no trochlear valgus as the groove is completely vertical. The implant is attached to the trochlea utilizing three pegs. The LubinusTM prosthesis consists of a very thin trochlear component to provide consistent coverage of the entire trochlear surface. However, the radius of curvature is quite small and therefore incompletely reproduces the trochlear shape. Its upper edge is triangular, inhibiting patellar centring early in knee flexion. The trochlear angle is near 140°, and trochlear component fixation is provided by three pegs. The patellar component is domed. The other category includes asymmetric prostheses such as the LCSTM prosthesis [48], the Spherocentric prosthesis [47], and the Autocentric prosthesis [7, 26]. The LCSTM prosthesis (De PuyTM) was a very thin double-curvature implant. The trochlear angle of 140° was fairly shallow, with an extended surface on the lateral side. The associated patellar button was constrained within a metal backing that was either press-fit or cemented into the patella. The SpherocentricTM prosthesis (FHTM) has a saddle-shaped trochlea with one side larger and more prominent and an
123
1218
upper portion designed to articulate with the patellar button in the extended position. The three square pegs providing fixation theoretically allow its use without cement. The patellar button has a spherical shape and is sized to match the trochlea. The AutocentricTM prosthesis (De PuyTM) has a cylindrical–spherical trochlea that is small medially and features a superolateral bank to resist lateral patellar motion at lower flexion angles. The patellar button is domed and attached with three pegs. Anterior cut prostheses So-called second-generation patellofemoral prostheses are generally derived from the femoral trochlea portion of TKA designs. Rather than replacing lost cartilage, they aim to completely replace the anterior compartment of the knee (Fig. 1). Some femoral prostheses feature symmetrical femoral condyles, such as the AvonTM prosthesis from StrykerTM [1]. The femoral implant has a trochlear groove aligned in valgus to coincide with the mechanical axis of the lower limb rather than the shape of the native trochlea. The trochlear angle is variable, decreasing from 150° proximally to 140° distally. This design thus offers an increased level of patellar constraint as the knee flexes that is further enhanced by the prolonged distal aspect of the implant that extends to the level of the intercondylar notch. Three pegs on the anterior surface and one distally provide fixation. The patellar button is asymmetrical with a prominent medial facet. There are also asymmetrical prostheses such as the Hermes, Vanguard, Journey Competitor, Leicester, and Gender prostheses. The HermesTM prosthesis (CeraverTM) features a trochlear groove design that attempts to reproduce native anatomy with a raised lateral side and 7° of
Fig. 1 Intraoperative view of the anterior femoral cut of a secondgeneration patellofemoral prosthesis that completely replaces the anterior compartment of the knee
123
Knee Surg Sports Traumatol Arthrosc (2012) 20:1216–1226
valgus. The trochlear angle is relatively small with a value of 136°, significantly constraining the patella. Fixation is via an intramedullary pin. The implant extends distally to the intercondylar notch, ensuring prolonged contact with the patellar implant during flexion, the patellar button being domed. The VanguardTM prosthesis (BiometTM) has a trochlea with extensive coverage (from 44 to 60 mm in height depending on size), extending to the intercondylar notch and allowing extended contact with the patellar button from full extension to over 100° of flexion but potentially impinging on the ACL. An intramedullary pin provides fixation. The patellar button is convex and fixed with three pegs. The Journey CompetitorTM prosthesis (Smith NephrewTM) has an oxinium trochlea (cobalt chrome coated with a layer of titanium oxide) with four mounting pegs. The lateral flange is quite prominent and extends more proximally on the anterior femur. The patellar button is biconvex, derived from the design of the GenesisTM TKA. The GenderTM prosthesis (ZimmerTM) has the only trochlea available in 5 sizes. It is made of wrought cobalt chrome and is relatively thin with a trochlear groove in 7° of valgus. Fixation is provided by three pegs and a distal hook that engages the front of the intercondylar notch. The patellar component is convex. Finally, the Leicester prosthesisTM (CorinTM) has the lowest trochlear constraint with a trochlear angle near 160°. It is not cemented, and fixation is provided by an intramedullary pin and two additional screws. However, these screws may impinge on the patellar button. The button is all polyethylene, convex, and uncemented, with fixation provided by press-fit of a large central peg.
Indications As with unicompartmental prostheses, the success of patellofemoral prostheses is highly dependent on selection criteria [5, 10, 35, 39]. This procedure can be offered effectively in cases of isolated patellofemoral osteoarthritis, post-traumatic patellofemoral arthritis, or for patients with patellofemoral arthritis associated with trochlear dysplasia and patellar subluxation [63] (Fig. 2). In cases of significant patellar subluxation or tilt, medialization of the tibial tubercle with or without lateral release and associated facetectomy should be considered because isolated patellofemoral arthroplasty cannot in and of itself stabilize a patellofemoral joint with severe malalignment [23]. Additionally, a distalization procedure can be considered in cases of severe patella alta if the patella does not engage with the trochlea at maximal knee extension, [23]. Contraindications are numerous and should be carefully considered. These include chondrocalcinosis, pain, cartilage defects, or evidence of significant osteoarthritis in the
Knee Surg Sports Traumatol Arthrosc (2012) 20:1216–1226
Fig. 2 Pre-operative radiograph showing patellofemoral arthritis associated with patellar subluxation. Medialization of the tibial tubercle with or without lateral release and associated facetectomy should be considered in cases of with significant malalignment because isolated patellofemoral arthroplasty often does not provide adequate stability in the patellofemoral joint in these cases
medial or lateral tibiofemoral compartments [41]. To extend indications, some authors have proposed to combine the PFP with cartilage restoration procedures in cases of a defect in the load-bearing portions of the femoral condyles [39] or unicompartmental knee arthroplasty in cases of significant degenerative disease in an additional compartment [54]. Inflammatory joint disease remains an absolute contraindication because of global involvement of the knee joint. Obesity with a BMI greater than 30 kg/m2 seems to be a relative contraindication [63] due to potential for the progression of tibiofemoral disease. There is insufficient information to determine whether chronic anterior laxity is deleterious to the function and longevity of the PFP. Patient age should also be considered. Some authors limit this procedure to patients under 60 years [42], but there is currently no definitive evidence demonstrating any influence of patient age on results.
Results The reported results of PFP vary, no doubt influenced by the gradual evolution of implant designs along with the refinement of indications and surgical techniques [41]. Lonner et al. [37] suggested that implant choice is a major driver of outcomes. They reported a revision rate of 17 % with the first-generation LubinusÒ prosthesis versus 4 % with the second-generation Avon prosthesis (Strycker Howmedica Osteonics, New Jersey). Scoring system The published series use a variety of different rating systems, which makes comparisons difficult between studies.
1219
While scores such as the Knee Society [29] or the Oxford Group [20] are very useful for evaluating treatments for knee arthritis in general, specific tools to report the results of treatment of the patellofemoral joint are essential. Several scores have been proposed including the Bristol score [38], the Bartlett and Melbourne score [24], the Fulkerson score [25], and the Lonner patellofemoral score [38]. None of these scores, however, has been validated specifically for the assessment of patellofemoral osteoarthritis. A study by Fithian and Paxton [56] compared the various clinical scores and reported that the two most suitable scores for post-operative evaluation of PFP are the Short-Form 36 (SF 36) [68] and the Knee Injury and Osteoarthritis Outcome Score (KOOS) [57]. Results of different implants: Table 1 Richards prosthesis A key feature of this implant was the highly constrained trochlear design. The first publication in 1979 by Blazina et al. [10] included 57 Richards I and II prostheses and reported a 16 % revision at 2-year follow-up thought to be primarily due to poor positioning of implants. Several series have reported short- or medium-term results of this implant. Arciero et al. [5] in 1988 in a series of 25 PFP found a number of patellofemoral joint complications (patellar catching and subluxation) and persistent pain related to poor positioning or misalignment. Three patients (20 %) were revised for tibiofemoral osteoarthritis at 5 years of follow-up. Cartier et al. [13] reported on 72 Richards I and II prostheses (many with associated distal realignment procedures, HTO, or UKA). At mean 4-year follow-up, they found lateral patellar subluxation in two knees, one case of symptomatic patella infera, and four cases of tibiofemoral osteoarthritis. Reoperation was performed in 7 % of patients. Krajci-Radcliffe et al. [31] reported 88 % good and excellent results (but with an extensor lag in 70 %) in 16 knees at 5.8-year mean followup. de Winter et al. [22] found that 7 Richards II implants in a series of 26 (27 %) had undergone repeat surgery for instability or poor positioning at a mean of 11 years postoperative. This included two distal realignment procedures, three patellectomies, and two conversions to TKA. Kooijman [30] in 2003 in a series of 56 Richards I prostheses with 15- to 21-year follow-up found 86 % satisfactory results. The leading cause of revision was progression of tibiofemoral joint osteoarthritis (23 % of patients), but additional soft tissue surgery was also required early in 18 % of patients. Case series with long follow-up have also been reported. Cartier et al. [14] reported on 79 Richards I prostheses implanted between 1975 and 1991 (average age at
123
1220
Knee Surg Sports Traumatol Arthrosc (2012) 20:1216–1226
Table 1 Clinical results of patellofemoral arthroplasty Study
Year
Implant
n
Age (years)
Follow-up (years)
Percent revised (%)
Good and excellent results (%)
Blazina [11]
1979
Richards I/II
57
39
2
35
N/A
Arciero [5]
1988
Richards II
25
62
5.3
28
72
Cartier [14]
1990
CFS-Wright Richards II/III
72
65
4
7
85
Argenson [7]
1995
Autocentric
66
57
5.5
15
84
Krajca-Radcliffe [33]
1996
Richards I/II
16
64
5.8
6
88
Mertl [50]
1997
Spherocentric
Arnbjornsson [8]
1998
Blazina
51 113
60.5
3
6
82
56
7
22
75
Lubinus Richards II Other de Cloedt [22]
1999
Autocentric
45
51
6
18
63
de Winter [23]
2001
Richards II
26
59
11
19
76
Tauro [65]
2001
Lubinus
62
66
7.5
28
45
Smith [63]
2002
Lubinus
45
72
4
19
64
Kooijman [32]
2003
Richards II
45
50
17
22
86
Board [12]
2004
Lubinus
17
66
1.5
35
53
Merchant [48] Merchant [49]
2004 2005
LCS LCS
15 16
49 47
3.75 4.5
0 0
93 94
Lonner [39]
2004
Lubinus
30
38
4
33
84
Lonner [39]
2004
Avon trochlea
25
44
6 months
0
96
Nexgen Patella Argenson [6]
2005
Autocentric
Ackroyd [1]
2005
Avon
66
57
16.2
306
62
N/A
Cartier [15]
2005
Sisto [62] Cossey [18]
42
Richards III
79
60
10
2006
Kinamatch
25
45
6
0
100
2006
Avon (navigation)
4
52
1
0
100 N/A
3.6 25
N/A N/A 77
Nicol [55]
2006
Avon
103
68
7.1
14
Ackroyd [3]
2007
Avon
109
68
5.2
15
80
Gadeyne [27]
2008
Autocentric
43
67
6
24
72
Mohammed [52]
2008
101
57
4
4
72
14 76
Avon Lubinus Femoro Patella Vialla
Butler [13] Leadbetter [34]
2009 2009
Performa Avon
22 79
48.6 58
5 3
Starks [64]
2009
Avon
37
66
2
Van Wagenberg [69]
2009
Autocentric
24
64
4.8
Gao [28]
2010
Avon
11
54
2
Van Jonbergen [67]
2010
Richard II
185
52
13.3
Odumenya [56]
2010
Avon
50
66
5.3
Charalambous [16]
2011
LCS
51
64
2
implantation 60 years). They found a survival rate of 75 % at 11 years post-operative and 77 % good and excellent results on implants still in place at that time. Thirteen reoperations (including 8 conversions to TKA) were necessary, primarily due to progression of tibiofemoral
123
N/A 84
0
86
29
30
0
100
25
N/A
4 33
N/A 33
osteoarthritis. Van Jonbergen and Werkman [63] reported on 185 Richards II prostheses performed in patients with a mean age of 52 years. They noted a 10-year survival rate of 84 % and a 20-year survival rate of 69 %. Reoperation was necessary in 38 % of the knees with the progression of
Knee Surg Sports Traumatol Arthrosc (2012) 20:1216–1226
tibiofemoral osteoarthritis being the leading cause of revision in this series (13 %). They also found a significant number of cases of implant malposition causing instability or impingement leading to reoperation (7 %) that the authors attributed to the constrained geometry of the trochlea. Loosening of the implant and polyethylene wear were rare and only described in cases in which the implant was not cemented. All published series seem to confirm the analysis of Lonner [37], which asserted that designs with high degrees of patellar constraint due to a deep trochlear groove are less able to mitigate poor patellar tracking. These implants are thus more effective in cases with good alignment or in which an associated realignment procedure accompanied implantation of the device. Lubinus prosthesisÒ [43] Tauro et al. [61] reported on 76 patients treated with this prosthesis with a mean age of 65.5 years. They reported implant survival of 65 % at 8-year follow-up. However, if patients with moderate to severe pain were also counted as surgical failures, the survival rate dropped to 48 %. They reported 45 % good and excellent results overall. Twenty patients required a change in implant including 10 revised to TKA and 10 revised to a different patellofemoral prosthesis (the Avon prosthesis, StrykerTM). The most common causes of revision were poor patellofemoral tracking (75 %) and progression of tibiofemoral osteoarthritis (25 %). Smith et al. [59] reviewed 45 patients with a mean age of 72 years at the time of treatment with the LubinusÒ prosthesis with 4-year follow-up. They reported a reoperation rate of 42 % and noted good and excellent results in 64 % of patients. Five knees were revised to TKA, including 3 for progression of tibiofemoral osteoarthritis, one for persistent diffuse pain, and one for patellar instability. Board et al. [11] reported 19-month follow-up on 17 patients treated with the LubinusÒ implant. They reported only 53 % good to very good results with a 35 % reoperation rate (including 24 % revised to TKA). The authors reported patellar instability in 18 % of patients, stiffness in 18 %, infection in 6 %, and tibiofemoral osteoarthritis progression in 12 %. They concluded by stating that the implant was very poorly tolerated and advised against its use. Several studies have noted the LubinusÒ prosthesis to be prone to revision for patellofemoral dysfunction. Lonner et al. [37] reported 17 % of 30 younger patients (mean age 38 years) suffered complications related to patellofemoral dysfunction at 4-year follow-up compared with only 4 % in a series of patients who had the Avon prosthesis implanted. At final follow-up of the Lubinus group, one-third of patients had been revised: 3 to a different type of patellofemoral prosthesis and 7 to a TKA. Finally, Hendrix et al.
1221
[28] reported on a series of 14 patients initially treated with the Lubinus prosthesis who were subsequently revised to the Avon prosthesis. They found the most common causes of failure to be initial malposition of the trochlear component, leading to patellar subluxation and excessive wear of the patellar component. In summary, the Lubinus prosthesis has been relatively poorly tolerated, with a high rate of complications and early revisions reported in all publications concerning this implant. AutocentricÒ prosthesis (Depuy) (Fig. 3) Argenson et al. [6] reported the results of 66 AutocentricÒ prostheses performed between 1972 and 1990. The average patient age at implantation was 57 years. They found satisfactory medium-term (2–10 years) results with 84 % patient satisfaction. By 16.2-year mean follow-up (range, 12–20 years), 14 knees had been revised to TKA for progression of tibiofemoral osteoarthritis at a mean of 7.3 years. Patients treated for primary patellofemoral osteoarthritis were more likely to have undergone revision for this reason than those treated for trochlear dysplasia. Eleven additional patients underwent reoperation for instability or loosening at a mean of 4.5 years after surgery. Finally, they reported one patellar fracture, 3 infections, 6 cases of post-operative stiffness (two of which required manipulation under anaesthesia), 5 lateral releases and lateral patellar facetectomy for increased patellar tilt and impingement between the patellar and lateral femoral condyle. The overall rate of implant survival at 16 years was 58 %.
Fig. 3 Autocentric prosthesis from DepuyTM. The femoral prosthesis is asymmetrical with a cylindrical–spherical trochlea
123
1222
Van Wagenberg et al. [65] reported disappointing results in 24 cases at a mean of 4.8-year follow-up (range, 2–11 years) performed primarily for primary patellofemoral osteoarthritis. They found an 87.5 % rate of reoperation and 29 % conversion rate to TKA for progression of tibiofemoral osteoarthritis or patellar maltracking. Gadeyne et al. [26] reported the results of a series of 43 Autocentric prostheses with 6-year average follow-up (range, 7 months–15 years). Considering all causes for revision, implant survival was 82 % at 5 years and 62 % at 10 years. The 11 revisions were performed for progression of tibiofemoral osteoarthritis (5 cases), loosening (2 cases), patellar maltracking (2 cases), and femoral component malposition (2 cases). Overall, the published series report a relatively high rate of reoperation, although the aetiology of osteoarthritis appears to be a predictor of outcome. Arthritis secondary to dysplasia seems to be more favourable than primary or post-traumatic osteoarthritis [6, 21]. LCSÒ prosthesis (Depuy) The cemented, metal-backed, mobile-bearing patellar component in this system is compatible with both the trochlear component in a patellofemoral arthroplasty and the femoral component of their total knee arthroplasty system. This design is advantageous because it obviates the need to change the patellar component if revision to TKA is required. However, authors have reported cases of dislocation [69] or dissociation of the polyethylene [62] with this implant. Merchant et al. [46] presented the first results of this implant in 16 cases, with 93 % excellent and good results reported at 4.5-years follow-up. Mean patient age was 47.4 years at the time of implantation. Arumilli et al. [8] reported their experience with this cemented, metal-backed patellar component. They found failure of the patellar component due to excessive wear, metal–metal contact, and fracture/dislocation of the polyethylene in the first 2 years following implantation. Charalambous et al. [15] published 2-year follow-up of their series of 51 prostheses implanted in patients with an average age of 63.8 years. Even with short follow-up, they reported 17 revisions (33.3 %), including 16 revised to TKA and one change in a patellar button. They reported several cases in which the polyethylene patella was immobile due to the development of fibrosis around the implant. They also described three cases of extensive metallosis. The survival rate at 3 years was 63 %. Only 46 % of implants were still in place and did not report moderate or severe pain. Because of high complication rates and early failures, marketing of this implant was discontinued in 2009.
123
Knee Surg Sports Traumatol Arthrosc (2012) 20:1216–1226
Kinamed Sisto et al. [58] reported the results of a custom implant in 25 patients with an average follow-up of 6 years. They found 100 % good and excellent results and no revisions in a young population (45 years on average). This implant tends to anteriorize the trochlea to increase the leverage of the extensor mechanism while preserving bone stock. Some authors have suggested that this design may overstuff the anterior compartment, leading to increased pain and limited flexion [39]. Avon prosthesis (Fig. 4) This ‘‘second-generation’’ prosthesis utilizes an anterior cut similar to that of most TKA systems. This approach has several theoretical advantages including increased reproducibility in placement and the ability to place the femoral component in external rotation to improve patellar tracking. Lonner et al. [37] were the first to report on their experience with this implant, reporting decreased complications in
Fig. 4 Intra-operative view of the AvonTM prosthesis from StrykerTM. The femoral prosthesis is symmetrical and the patellar button is asymmetrical with a prominent medial facet
Knee Surg Sports Traumatol Arthrosc (2012) 20:1216–1226
the first few months following implantation compared with the Lubinus prosthesis. Nicol et al. [51] published a report on 103 prostheses implanted in patients with a mean age of 68 years. At 7.1-years mean follow-up, they noted a revision rate of 14 %, primarily due to progression of tibiofemoral osteoarthritis. They noted excellent results in patients treated for arthritis secondary to trochlear dysplasia in their series. Ackroyd et al. [3] reported on 109 consecutive patients (average age 68 years) with mean follow-up of 5.2 years. The survival rate at 5 years was 95.8 % with 80 % satisfactory results (Bristol pain score [20). As demonstrated with many other implants, the main complication in this series was the progression of tibiofemoral osteoarthritis that occurred in 28 % of cases. Leadbetter et al. [33] completed a multicenter study of 79 knees treated with the Avon prosthesis and followed for 3 years. They noted 84 % good and excellent results and 90 % of patients reported no knee pain during activities of daily living. Starks et al. [60] also reported 100 % good and excellent results with no revisions in a series of 37 Avon prostheses. They did note radiographic evidence of tibiofemoral osteoarthritis in 22 % of patients, none of which were symptomatic. Odumenya et al. [52] reported 100 % implant survival 5 years following implantation. They did note lateral patellar tilt of more than 5 in 16 % of cases, lateral patellar subluxation in 14 % of cases, and a progression of tibiofemoral osteoarthritis in 22 % of cases. In general, the currently published results for the Avon prosthesis are encouraging, but longer follow-up is lacking. Further study is necessary to determine whether these results hold up over time. Influence of the aetiology of arthritis on outcome A number of authors have reported better clinical outcomes in patients with arthritis secondary to trochlear dysplasia than in other groups. Gadeyne et al. [26] found 68.2 % good and excellent results in the dysplasia group versus 44.4 % in the primary osteoarthritis group. Argenson et al. [6] found 73 % good results in the dysplasia group and 54 % good results in the group with primary osteoarthritis. Leadbetter et al. [32] found 83 % good to very good results and overall higher patient reported outcome scores in those with arthritis secondary to trochlear dysplasia relative to those with other diagnoses. Nicol et al. [51] reported that in their series of 103 patellofemoral prostheses, the main cause of revision by 7.1 years was the progression of tibiofemoral osteoarthritis. However, this complication was less frequent in case of patellofemoral joint arthritis secondary to trochlear dysplasia. They suggested that trochlear dysplasia is an ideal indication for patellofemoral arthroplasty.
1223
Pre-operative patella infera, regardless of aetiology, should in theory negatively impact results. Unfortunately, the results of such patients are generally not reported separately in the existing literature. However, some have argued that this condition is a contraindication to patellofemoral arthroplasty due to this concern [23]. In contrast, a recent study by Van Jonbergen et al. [63] reporting 13-year following up on a consecutive series of 185 Richards II prostheses found no influence of initial diagnosis, sex, or age at time of the intervention on revision rates. They had hypothesized that the aetiology of patellofemoral osteoarthritis (primary, post-traumatic, or secondary to trochlear dysplasia) would significantly influence implant survival. They did find that a BMI greater than 30 kg/m2 was a significant predictor of revision.
Complications Early post-operative complications Patellofemoral joint replacement is associated with a higher level of early post-operative complications than total knee arthroplasty [18, 55]. Early post-operative complications include persistent anterior knee pain, patellar catching or snapping, and extensor mechanism disruption. Increased peripatellar pain may be an early consequence of ‘‘overstuffing’’ of the patellofemoral joint through the placement of an implant that is thicker than the amount of bone and cartilage that is resected. In a recent study, Mofidi et al. found a greater increase in patellar thickness following surgery in patients with poor results relative to those with better results [48]. Implant design seems to have a direct influence on the incidence of certain early complications [39]. A larger radius of curvature may increase risk of patellar catching or maltracking, particularly when the trochlear implant is placed in flexion or in cases in which the patellar prosthesis articulates with the femur proximal to the trochlear component in full extension. This risk is increased in prostheses with relatively short proximal extension on the anterior femoral cortex. In addition, narrow implants may increase risk of patellar catching or maltracking. The risk is decreased in implants with a relatively unconstrained geometry of the trochlear component. Leadbetter et al.[32] in a study of 30 knees reported one case of femoral notching, 4 cases of arthrofibrosis requiring manipulation under anaesthesia, and two ruptures of the quadriceps tendon. The incidence of many early complications has decreased with more recent patellofemoral arthroplasty designs [23]. Ackroyd et al. [2] in a series of 306 patients treated with the Avon prosthesis noted a 4 % incidence of
123
1224
anterior knee pain, a 5 % incidence of complications related to the extensor mechanism, and arthrofibrosis requiring manipulation under anaesthesia in 1.6 % of patients. Technical errors Malpositioning of the femoral prosthesis may result in the component overhanging the femoral condyle, patellar maltracking, and patellar instability [3]. Gadeyne et al. [26] demonstrated that placing the femoral component in internal rotation is associated with a higher risk of reoperation. Component overhang was also significantly higher in the group requiring reoperation. Malalignment of the extensor mechanism as well as overstuffing of the anterior compartment is frequent causes of anterior knee pain. Both can be symptomatic quite early and lead to high rates of early revision [2, 26]. Lonner et al. [39] also reported that post-operative patellofemoral instability is frequently the result of poor soft tissue balancing. Similarly, failure to address significant patella alta with a tibial tubercle distalization also increases the risk of post-operative patellar instability [23]. Late complications
Knee Surg Sports Traumatol Arthrosc (2012) 20:1216–1226
than 0.5 % failure due to loosening at 7 years of follow-up. Most reported cases of implant loosening have involved uncemented prostheses [7]. Gadeyne et al. [26] reported a 24 % failure rate in a series of 43 Autocentric prostheses at 74 months. Only 2 of the 11 cases undergoing revision were revised for aseptic loosening, both associated with component malpositioning. Kooijman et al. [30] reported a very low rate of loosening (2 %) up to 15.6 years after surgery. Chronic effusion Leadbetter et al. [35] reported a chronic effusion rate of 33 %. He commented on the importance of avoiding placement of the femoral in internal rotation or recurvatum.
Future directions The studies reviewed above demonstrate the results of patellofemoral arthroplasty to be highly sensitive to implant design, surgical technique, and indications. Implant designs and placement techniques have progressed considerably, and further advances will continue to improve results [49].
Progression of tibiofemoral osteoarthritis Surgical navigation As recent prosthesis designs have reduced the number of complications due to patellar maltracking, progression of tibiofemoral osteoarthritis has become the main cause of revision surgery [3]. The reported percentage of prosthetic revision due to tibiofemoral osteoarthritis ranges from 0 to 22 % at 5- to 15-year follow-up in various studies [30, 31]. Degeneration of the tibiofemoral articulation is more common in patients who undergo patellofemoral arthroplasty for primary patellofemoral osteoarthritis. Nicol et al. [51] in a prospective study of 103 Avon prostheses report a 12 % revision rate due to progression of tibiofemoral osteoarthritis at 7.1-year follow-up. The average time to revision was 55 months. The rate among patients treated for primary osteoarthritis was 17 %, while no cases of progression were noted in patients treated for trochlear dysplasia (p \ 0.01). Ackroyd et al. [2] noted progression of tibiofemoral osteoarthritis to be the primary late complication in a series of 306 Avon prostheses. They noted that this complication was not predicted by the functional results or pain control achieved by the prosthesis in the first 2 years following surgery. Loosening Loosening is a relatively rare complication following patellofemoral arthroplasty. Lonner et al. [42] reported less
123
Surgical navigation has been shown to be reliable and reproducible in improving implant placement during total knee arthroplasty [44]. It may also allow for the improvement placement of patellofemoral prostheses. Cossey et al. [17] reported 1-year results of Avon prostheses placed with the aid of a navigation system using intra-operative landmarks. He reported 100 % good and excellent results, no failures, and no early post-operative instability. Other robotic systems have also appeared on the market [16, 40] that could also be used to improve the reproducibility of implant positioning. These systems rely on preoperative CT scans that allow planning of ideal implant positioning. A robotic arm then helps guide the bony resection and assists the surgeon in placing the implant in the previously defined position. Custom implants and guides New techniques allow for the manufacture of custom-made implants that match the radius of curvature of the patellofemoral joint based on three-dimensional CT reconstructions. These designs may help to optimize coverage of bone surfaces without causing overconstraint of the patella. Sisto and Sarin [58] reported encouraging results of 25 custom implants (Kinamed Custom ImplantsÒ). At a mean
Knee Surg Sports Traumatol Arthrosc (2012) 20:1216–1226
follow-up of 73 months, they found 100 % good and excellent results and no revisions. Butler et al. reported 5-year results of 22 custom implants (BiometÒ Performa Prosthesis) [12]. They reported one revision at 18 months, 2 cases of post-operative stiffness treated with arthroscopy, and no radiologic evidence of loosening of the trochlear implants—although one broken patellar button was noted. Bicompartmental arthroplasty The combination of medial unicompartmental and patellofemoral prostheses is an option that has recently garnered renewed interest [40, 54]. Some recent publications suggest that this option yields favourable long-term results and would thus extend indications for patellofemoral prosthesis, especially for young, active patients [54].
Conclusions Recent series demonstrate that improvements design and technique have led to satisfactory results in the short and medium term with ‘‘modern’’ patellofemoral implants. Still, indications remain limited and the failures of the past should serve as a warning. Long-term studies are needed to assess whether these favourable results are durable. The future for patellofemoral arthroplasty seems bright as new implants, custom guides, and surgical navigation seek to improve results. These new developments will also require longer-term study to validate their effectiveness.
References 1. Ackroyd CE, Chir B (2005) Development and early results of a new patellofemoral arthroplasty. Clin Orthop Relat Res 436:7–13 2. Ackroyd CE, Newman JH (2003) The Avon patellofemoral arthroplasty: two to five year results. J Bone Joint Surg Br 85: 162–163 3. Ackroyd CE, Newman JH, Evans R, Eldridge JD, Joslin CC (2007) The Avon patellofemoral arthroplasty: five-year survivorship and functional results. J Bone Joint Surg Br 89:310–315 4. Amis AA, Senavongse W, Darcy P (2005) Biomechanics of patellofemoral joint prostheses. Clin Orthop Relat Res 436:20–29 5. Arciero RA, Toomey HE (1988) Patellofemoral arthroplasty. A three- to nine-year follow-up study. Clin Orthop Relat Res 236: 60–71 6. Argenson JN, Flecher X, Parratte S, Aubaniac JM (2005) Patellofemoral arthroplasty: an update. Clin Orthop Relat Res 440: 50–53 7. Argenson JN, Guillaume JM, Aubaniac JM (1995) Is there a place for patellofemoral arthroplasty? Clin Orthop Relat Res 321: 162–167 8. Arumilli BR, Ng AB, Ellis DJ, Hirst P (2010) Unusual mechanical complications of unicompartmental low contact stress mobile bearing patellofemoral arthroplasty: a cause for concern? Knee 17:362–364
1225 9. Becker R, Ropke M, Krull A, Musahl V, Nebelung W (2008) Surgical treatment of isolated patellofemoral osteoarthritis. Clin Orthop Relat Res 466:443–449 10. Blazina ME, Fox JM, Del Pizzo W, Broukhim B, Ivey FM (1979) Patellofemoral replacement. Clin Orthop Relat Res 144:98–102 11. Board TN, Mahmood A, Ryan WG, Banks AJ (2004) The Lubinus patellofemoral arthroplasty: a series of 17 cases. Arch Orthop Trauma Surg 124:285–287 12. Butler JE, Shannon R (2009) Patellofemoral arthroplasty with a custom-fit femoral prosthesis. Orthopedics 32:81 13. Cartier P, Sanouiller JL, Grelsamer R (1990) Patellofemoral arthroplasty. 2–12-year follow-up study. J Arthroplasty 5:49–55 14. Cartier P, Sanouiller JL, Khefacha A (2005) Long-term results with the first patellofemoral prosthesis. Clin Orthop Relat Res 436:47–54 15. Charalambous CP, Abiddin Z, Mills SP, Rogers S, Sutton P, Parkinson R (2011) The low contact stress patellofemoral replacement: high early failure rate. J Bone Joint Surg Br 93: 484–489 16. Cobb J, Henckel J, Gomes P, Harris S, Jakopec M, Rodriguez F, Barrett A, Davies B (2006) Hands-on robotic unicompartmental knee replacement: a prospective, randomised controlled study of the acrobot system. J Bone Joint Surg Br 88:188–197 17. Cossey AJ, Spriggins AJ (2006) Computer-assisted patellofemoral arthroplasty: a mechanism for optimizing rotation. J Arthroplasty 21:420–427 18. Dahm DL, Al-Rayashi W, Dajani K, Shah JP, Levy BA, Stuart MJ (2010) Patellofemoral arthroplasty versus total knee arthroplasty in patients with isolated patellofemoral osteoarthritis. Am J Orthop (Belle Mead NJ) 39:487–491 19. Davies AP, Vince AS, Shepstone L, Donell ST, Glasgow MM (2002) The radiologic prevalence of patellofemoral osteoarthritis. Clin Orthop Relat Res 402:206–212 20. Dawson J, Fitzpatrick R, Murray D, Carr A (1998) Questionnaire on the perceptions of patients about total knee replacement. J Bone Joint Surg Br 80:63–69 21. De Cloedt P, Legaye J, Lokietek W (1999) Femoro-patellar prosthesis. A retrospective study of 45 consecutive cases with a follow-up of 3–12 years. Acta Orthop Belg 65:170–175 22. de Winter WE, Feith R, van Loon CJ (2001) The Richards type II patellofemoral arthroplasty: 26 cases followed for 1–20 years. Acta Orthop Scand 72:487–490 23. Farr J 2nd, Barrett D (2008) Optimizing patellofemoral arthroplasty. Knee 15:339–347 24. Feller JA, Bartlett RJ, Lang DM (1996) Patellar resurfacing versus retention in total knee arthroplasty. J Bone Joint Surg Br 78:226–228 25. Fulkerson JP, Shea KP (1990) Disorders of patellofemoral alignment. J Bone Joint Surg Am 72:1424–1429 26. Gadeyne S, Besse JL, Galand-Desme S, Lerat JL, Moyen B (2008) Results of self-centering patellofemoral prosthesis: a retrospective study of 57 implants. Rev Chir Orthop Reparatrice Appar Mot 94:228–240 27. Hassaballa MA, Porteous AJ, Newman JH (2004) Observed kneeling ability after total, unicompartmental and patellofemoral knee arthroplasty: perception versus reality. Knee Surg Sports Traumatol Arthrosc 12:136–139 28. Hendrix MR, Ackroyd CE, Lonner JH (2008) Revision patellofemoral arthroplasty: three- to seven-year follow-up. J Arthroplasty 23:977–983 29. Insall JN, Dorr LD, Scott RD, Scott WN (1989) Rationale of the Knee society clinical rating system. Clin Orthop Relat Res 248: 13–14 30. Kooijman HJ, Driessen AP, van Horn JR (2003) Long-term results of patellofemoral arthroplasty. A report of 56 arthroplasties with 17 years of follow-up. J Bone Joint Surg Br 85:836–840
123
1226 31. Krajca-Radcliffe JB, Coker TP (1996) Patellofemoral arthroplasty. A 2- to 18-year followup study. Clin Orthop Relat Res 330:143–151 32. Leadbetter WB, Kolisek FR, Levitt RL, Brooker AF, Zietz P, Marker DR, Bonutti PM, Mont MA (2009) Patellofemoral arthroplasty: a multi-centre study with minimum 2-year followup. Int Orthop 33:1597–1601 33. Leadbetter WB, Mont MA (2009) Patellofemoral arthroplasty: a useful option for recalcitrant symptomatic patellofemoral arthritis. Semin Arthro 20:148–160 34. Leadbetter WB, Ragland PS, Mont MA (2005) The appropriate use of patellofemoral arthroplasty: an analysis of reported indications, contraindications, and failures. Clin Orthop Relat Res 436:91–99 35. Leadbetter WB, Seyler TM, Ragland PS, Mont MA (2006) Indications, contraindications, and pitfalls of patellofemoral arthroplasty. J Bone Joint Surg Am 88(Suppl 4):122–137 36. Levitt RL (1973) A long-term evaluation of patellar prostheses. Clin Orthop Relat Res 97:153–157 37. Lonner JH (2004) Patellofemoral arthroplasty: pros, cons, and design considerations. Clin Orthop Relat Res 428:158–165 38. Lonner JH (2006) Patellofemoral arthroplasty. In: Scott N (ed) Insall and Scott surgery of the knee, 4th edn. Elsevier, Philadelphia, pp 1440–1454 39. Lonner JH (2007) Patellofemoral arthroplasty. J Am Acad Orthop Surg 15:495–506 40. Lonner JH (2009) Modular bicompartmental knee arthroplasty with robotic arm assistance. Am J Orthop (Belle Mead NJ) 38: 28–31 41. Lonner JH (2010) Patellofemoral arthroplasty. Instr Course Lect 59:67–84 42. Lonner JH, Mehta S, Booth RE Jr (2007) Ipsilateral patellofemoral arthroplasty and autogenous osteochondral femoral condylar transplantation. J Arthroplasty 22:1130–1136 43. Lubinus HH (1979) Patella glide-bearing total replacement. Orthopedics 2:119–127 44. Lustig S, Fleury C, Goy D, Neyret P, Donell ST (2011) The accuracy of acquisition of an imageless computer-assisted system and its implication for knee arthroplasty. Knee 18:15–20 45. McKeever DC (1955) Patellar prosthesis. J Bone Joint Surg Am 37-A:1074–1084 46. Merchant AC (2005) A modular prosthesis for patellofemoral arthroplasty: design and initial results. Clin Orthop Relat Res 436:40–46 47. Mertl P, Van FT, Bonhomme P, Vives P (1997) Femoropatellar osteoarthritis treated by prosthesis. Retrospective study of 50 implants. Rev Chir Orthop Reparatrice Appar Mot 83:712–718 48. Mofidi A, Bajada S, Holt MD, Davies AP (2011) Functional relevance of patellofemoral thickness before and after unicompartmental patellofemoral replacement. Knee PMID:21498901 49. Morra EA, Greenwald AS (2006) Patellofemoral replacement polymer stress during daily activities: a finite element study. J Bone Joint Surg Am 88(Suppl 4):213–216 50. Newman JH (2007) Patellofemoral arthritis and its management with isolated patellofemoral replacement: a personal experience. Orthopedics 30:58–61 51. Nicol SG, Loveridge JM, Weale AE, Ackroyd CE, Newman JH (2006) Arthritis progression after patellofemoral joint replacement. Knee 13:290–295
123
Knee Surg Sports Traumatol Arthrosc (2012) 20:1216–1226 52. Odumenya M, Costa ML, Parsons N, Achten J, Dhillon M, Krikler SJ (2010) The Avon patellofemoral joint replacement: five-year results from an independent centre. J Bone Joint Surg Br 92:56–60 53. Oni OOA (1996) The Leicester patello-femoral joint prosthesis. Knee 3:81–82 54. Parratte S, Pauly V, Aubaniac JM, Argenson JN (2010) Survival of bicompartmental knee arthroplasty at 5 to 23 years. Clin Orthop Relat Res 468:64–72 55. Parvizi J, Stuart MJ, Pagnano MW, Hanssen AD (2001) Total knee arthroplasty in patients with isolated patellofemoral arthritis. Clin Orthop Relat Res 392:147–152 56. Paxton EW, Fithian DC (2005) Outcome instruments for patellofemoral arthroplasty. Clin Orthop Relat Res 436:66–70 57. Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD (1998) Knee injury and osteoarthritis outcome score (KOOS)— development of a self-administered outcome measure. J Orthop Sports Phys Ther 28:88–96 58. Sisto DJ, Sarin VK (2006) Custom patellofemoral arthroplasty of the knee. J Bone Joint Surg Am 88:1475–1480 59. Smith AM, Peckett WR, Butler-Manuel PA, Venu KM, d’Arcy JC (2002) Treatment of patello-femoral arthritis using the Lubinus patello-femoral arthroplasty: a retrospective review. Knee 9: 27–30 60. Starks I, Roberts S, White SH (2009) The Avon patellofemoral joint replacement: independent assessment of early functional outcomes. J Bone Joint Surg Br 91:1579–1582 61. Tauro B, Ackroyd CE, Newman JH, Shah NA (2001) The Lubinus patellofemoral arthroplasty. A five- to ten-year prospective study. J Bone Joint Surg Br 83:696–701 62. van Jonbergen HP, Werkman DM, Barnaart AF (2009) Dissociation of mobile-bearing patellar component in low contact stress patellofemoral arthroplasty, its mechanism and management: two case reports. Cases J 2:7502 63. van Jonbergen HP, Werkman DM, Barnaart LF, van Kampen A (2010) Long-term outcomes of patellofemoral arthroplasty. J Arthroplasty 25:1066–1071 64. van Jonbergen HP, Werkman DM, van Kampen A (2009) Conversion of patellofemoral arthroplasty to total knee arthroplasty: A matched case-control study of 13 patients. Acta Orthop 80: 62–66 65. van Wagenberg JM, Speigner B, Gosens T, de Waal Malefijt J (2009) Midterm clinical results of the Autocentric II patellofemoral prosthesis. Int Orthop 33:1603–1608 66. Vermeulen H, De Doncker E, Watillon M (1973) The Mac Keever patellar prosthesis in femoro-patellar arthrosis. Acta Orthop Belg 39:79–90 67. Walls RJ, Eldridge JD, Mulhall KJ (2007) Patellofemoral arthroplasty: evolving indications, technique, and applications in younger patients. Semin Arthro 18:156–161 68. Ware JE Jr, Sherbourne CD (1992) The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 30:473–483 69. Witjes S, Van den Broek C, Koeter S, Van Loon C (2009) Dislocation of the mobile bearing component of a patellofemoral arthroplasty: a report of two cases. Acta Orthop Belg 75:411–416