Importance Patellofemoral osteoarthritis (PFOA) is a common problem in the active population, affecting 2–11% of men and 8–24% of women over 55 years of age. Conservative measures are often ineffective, and patellofemoral arthroplasty (PFA) is an appealing solution considering that it is minimally invasive procedure, without burning any bridges if conversion to a total knee arthroplasty (TKA) is required.
Objective This work aims to systematically review the current available literature on outcomes from PFA or TKA for isolated PFOA. In particular, the research questions are: (1) Is the revision surgery rate higher in PFA than in TKA? (2) Is there any difference in conversion to TKA or reoperation rates between inlay and onlay implant design? (3) What influence has tibiofemoral osteoarthritis (TFOA) progression in PFA complication rate? (4) How many PFAs are revised or converted to TKA because of TFOA progression?
Evidence review A comprehensive search was performed through PubMed, Cochrane Library, Scopus, CINAHL and Google Scholar databases to search for relevant studies on PFA. Inclusion and exclusion criteria were defined. A 20-item methodological quality tool for assessing case series studies was used.
Findings Thirty-seven papers were found to be case series, and one was a case–control study (PFA vs TKA). The general methodology level was weak. However, PFAs have a statistically significant higher complication and consequently reoperation rate, with a statistically significant lower survivorship rate compared to TKA for PFOA. When taking into consideration PFA implant design, those numbers were different. Newer PFA designs, contrary to old ones, showed comparable complication and revision rates to those of TKA. However, newer PFAs' survivorship is still significantly lower than TKAs' one. Although TFOA progression is not necessarily a complication, it represents up to one-third of the reasons for PFA requiring revision, playing a major role in implant's survivorship.
Conclusions and relevance There is a lack of high-level evidence on PFA. However, PFAs with newer implant design showed to be as reliable in terms of the complication and reoperation rate as TKA. Progression of TFOA still remains an issue, representing up to one-third of the overall complications.
Level of evidence III, systematic review of level IV and V studies.
What is already known
Patellofemoral arthroplasty (PFA) is an appealing treatment for patellofemoral osteoarthritis (PFOA), however, overall it is affected by a higher complication and reoperation rate compared to total knee arthroplasty (TKA).
There is a significant difference in the outcomes of older (inlay) design implants compared to newer (onlay) implants.
TKA represents the gold standard for arthroplasty in knee osteoarthritis (OA).
What are the new findings
Newer (onlay) implants have similar complication and reoperation rates to those of TKA.
If tibiofemoral osteoarthritis (TFOA) progression is considered a complication of PFA, it represents up to one-third of the overall complications in PFA.
As a result, onlay PFA survivorship is still significantly lower than that of TKA, speculating that TFOA progression is the main reason.
The incidence of isolated patellofemoral osteoarthritis (PFOA) is about 10% in people over the age of 40, and more commonly, it affects women than men. When considering people over 55, 2–11% of men and 8–24% of women are affected by PFOA. Despite the high prevalence, only a small percentage of patients will develop significant symptoms.1
Multiple causes of PFOA have been identified.2 Increased body weight is a risk for developing PFOA.3 ,4 Congenital or post-traumatic bony or soft-tissue disorders of the extensor apparatus have been related to higher rates of PFOA. Increased Q-angle, patella alta, femoral trochlea dysplasia, abnormal vastus medialis obliquus and lax medial patellofemoral ligament (MPFL) are all conditions that alter the biomechanics of the patellofemoral joint that may predispose to arthritic changes.2
The first line of treatment is usually conservative, though of limited effectiveness. Several surgical options are available for PFOA, either joint preservation procedures or joint replacement in the case of end-stage OA or failed previous surgeries.5 Although total knee arthroplasty (TKA) has shown to be a reliable and effective treatment for PFOA,6–9 patellofemoral unicompartmental replacement (or patellofemoral arthroplasty, PFA) provides a more conservative and less invasive approach, however with variable outcomes.10 ,11 The first generation of implants were named ‘inlay’, since the main feature of the prosthesis was to be implanted flush with the surrounding trochlear cartilage. Results from series on inlay prosthesis are highly variable.11 Subsequently, as the implant design improved, the ‘onlay’ implants have become more popular. This implant replaces the entire trochlea and is much more versatile with respect to each patient's morphology. For this implant, the percentage of good-to-excellent results has been reported to be 80% and above.12
This work aims to systematically review the current available papers reporting on outcomes from PFA or TKA for isolated PFOA. Particularly, the questions that the authors aimed to answer were:
Is the revision surgery rate higher in PFA than in TKA?
Is there any difference in conversion to TKA or reoperation rates between inlay and onlay implant design?
What influence has tibiofemoral osteoarthritis (TFOA) progression in PFA complication rate?
How many PFAs are revised or converted to TKA because of TFOA progression?
Materials and methods
The Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement have been used.13 A comprehensive search was performed through PubMed, Cochrane Library, Scopus, CINAHL and Google Scholar databases to search for relevant studies on patellofemoral replacement, published from 1975 to 2016. Additionally, the reference list from each selected study was reviewed as a potential source for further eligible studies in order to achieve the greatest number of available studies in the scientific literature. Two independent coauthors (SV and BZ) conducted the searches, and results from both were matched; any disagreement was resolved by consensus. Searches were conducted for all English published papers, without any limitation for the publishing period. The following key words were used by the authors: ‘total knee arthroplasty’ OR ‘TKA’ OR ‘knee replacement’ OR ‘knee prosthesis’ AND ‘patellofemoral arthritis’ OR ‘patellofemoral OA’ OR ‘patellofemoral osteoarthritis’. ‘Patellofemoral arthroplasty’ OR ‘patellofemoral replacement’ OR ‘patellofemoral prosthesis’ AND ‘patellofemoral arthritis’ OR ‘patellofemoral OA’ OR ‘patellofemoral osteoarthritis’.
Initially, all titles and abstracts from the selected databases were reviewed and potential eligible studies were identified to be included in the systematic review. The full text from each relevant abstract was retrieved, read and assessed for eligibility according the following criteria: (1) English language studies; (2) randomised controlled studies, prospective or retrospective case series, (3) 2 years or longer follow-up, (4) clearly stated preoperative diagnosis, (5) included 15 or more patients, (6) studies reporting outcomes about primary PFA or TKA for isolated patellofemoral arthrosis. Studies were excluded if they were: (1) reviews or meta-analyses; (2) clinical commentaries or expert opinions; (3) animal studies; (4) not clearly defining the study's inclusion/exclusion criteria; (5) studies reporting outcomes from revision PFA or TKA.
The number of patients and the number of total knees that underwent surgery were recorded, as well patient's gender and age. Previous, concomitant and further surgeries were recorded as well. The type of prosthesis used were classified as inlay or onlay according to Lonner and Bloomfield.14 The number and the type of postoperative complications were collected together with the rate to conversion to TKA, revision of the PFA implants and the reoperations for other reasons than conversion or revision. The rate of TFOA progression was collected where possible, and the survivorship rate was calculated. The studies were further classified in follow-up categories in order to have homogenous follow-up data. Two authors (SV and BZ) independently reviewed the studies and discussed any disagreement until a consensus was reached.
The data quantitative analysis was performed using Microsoft Excel 2011 for Mac, V.14.6.1. and MiniTab Express V.1.4.0. A p value of 0.05 was set for statistical significance.
The Z-Score for two population proportions, with a two-tailed hypothesis, was used to assess any statistical significant difference with regard to conversion to TKA, revision of the PFA, reoperation rate for PFA and TKA and survivorship rate for PFA and TKA.
The implant survivorship for each implant group was calculated by the following calculation: total number of implants−(number of revised implants+number of implants converted to TKA). The implant survivorship rate was calculated as the survivorship number divided by the total number of implants.
A methodological quality tool for assessing case series studies was used.15 This is a 20-item checklist rated by yes, no or partial/unclear answers. Two authors (SV and BZ) independently assessed the methodological quality of all included articles, and all disagreements were discussed until consensus was reached.
From the online database search, 419 studies were found, plus 9 from manual search from references. After reviewing the abstract and full text of each paper retrieved by the database and manual search from references, 48 studies were selected. Eight studies were excluded because of unclear inclusion criteria. Two more studies, although matching the inclusion and exclusion criteria, were successively excluded: the case series from Merchant16 was excluded because the authors did not report on reoperation, revision or conversion to TKA rates, whereas the paper from Mohammed et al17 was excluded because inlay and onlay prostheses were used without clustering results according to the implant design. Therefore, 38 studies were left for this systematic review. The selection process is illustrated in the PRISMA flow diagram (figure 1).
Description of studies
Studies were classified according to the implant type. Three main categories were identified: inlay PFA implants (15 studies),3 ,4 ,18–30 onlay PFA implants (16 studies)8 ,31–47 and TKA (6 studies).7–9 ,48 ,49 One study compared inlay PFA to TKA implants.36 Data from this study were separated and classified according to the implant category. Two studies used customised PF implants.35 ,45 According to the features of the implants, those papers were classified as ‘onlay’ studies.
Three follow-up categories were identified: first category: 20–36 months (early), second category: 48–90 months (midterm) and third category: 120 months and over (long term).
A total of 865 knees were included for the inlay PFA, with a mean follow-up time of 88.7 months (19–204). A total of 1072 knees were included for the onlay PFA, and the mean follow-up time was 45.4 months (24–84), whereas a total of 235 knees were included for the TKA implants, having a mean follow-up of 59.4 months (20–88.8).
Most of the included studies used functional rating scales to assess outcomes, although there is high variability among the tools that were used, except from the studies reporting on TKA in which the Knee Society Score (KSS) was used by all studies but one.49 Previous surgeries, concomitant procedures, postoperative complications and reoperations were reported in most of the cases. However, collected information often was not exhaustive and data reported in a heterogeneous fashion among the studies. In addition, there was very limited data on the timing for reoperations following the PFA surgery. Details are reported in tables 1⇓–3.
All but one36 of the reviewed papers were case series. These were assessed using the case series appraisal tool.15 Although most of the tool's questions were answered, the main methodological issues were related to the heterogeneity of the reported data.
The overall survivorship rate of the inlay design implants (considering the early, midterm and long-term follow-up studies) was 81.6%, with a mean follow-up of 88.7 months. The mean conversion rate to TKA, PFA revision rate and reoperation for other reasons rate were 12.1%, 6.1% and 10.2%, respectively. More details are given in table 4.
The overall survivorship rate of the onlay design implants (considering the early and midterm follow-up studies) was 95%, with a mean follow-up of 45.6 months. The mean conversion rate to TKA, PFA revision rate and reoperation for other reasons rate were 4.4%, 0.4% and 2.7%, respectively. More details are given in table 4.
Total knee arthroplasty
The overall survivorship rate of a TKA (considering the early and midterm follow-up studies) was 99.5%, at a mean follow-up of 42.7 months. The mean revision rate and reoperation rate was 0.5% and 2.0%, respectively. More details are given in table 4.
Inlay versus onlay implants
Statistical analysis revealed that the conversion rate to TKA was not significant between inlay and onlay implants at the early follow-up (Z-score: −0.29, p=NS) whereas it was significantly higher for inlay implants at midterm (48–90) (Z-score: 2.01, p=0.044). The revision rate between the inlay and onlay PFA implants was not significantly different for early follow-up (Z-score: −0.63, p=NS), but it turned into a statistically significant higher rate with the inlay PFA when considering midterm follow-up studies (Z-score: 5.54, p=0). The reoperation rate was significantly higher in inlay implants at early and midterm follow-up studies (Z-score: 13.6, p=0 and Z-score: 2.83, p=0.004, respectively). The survivorship rate difference between inlay and onlay was not significant at early follow-up studies (Z-score=0.46, p=NS); however, it was significantly lower for inlay PFAs (Z-score: −4.60, p=0) in midterm follow-up studies (48–90 months).
Revision and reoperation rate of PFA versus TKA
The overall (inlay+onlay) PFA revision rate was 3%, which is significantly higher compared to that from TKA (0.5%) (Z-score: 3.54, p=0.00248). The overall (inlay+onlay) PFA reoperation rate was 7%, significantly higher compared to that from TKA (2%) (Z-score: 4.9, p=0). When comparing the inlay PFA revision rate to that of TKAs, a significant higher rate is observed from studies with early and midterm follow-up (Z-score: not applicable, p=0 and Z-score: 2.76, p=0.005, respectively). Similar results were observed for inlay PFA versus TKA reoperation rate (Z-score: 6.17, p=0 and Z-score: 2.87, p=0.004, respectively). Reoperation and revision rate for onlay PFA versus TKA was not significantly different (p=NS) from early (Z-score: −0.45 and 0.55, respectively) and midterm (Z-score: 1.41 and 0.49, respectively) follow-up studies.
Survivorship of PFA versus TKA
The overall survivorship rate for PFA (inlay+onlay) was significantly lower compared to that from TKA from early and midterm follow-up studies (Z-score: −1.96, p=0.04 and Z-score: −3.47, p=0.00052, respectively). When considering the implant survivorship for inlay PFA versus TKA, it was not significantly different from early follow-up studies (Z-score: −1.75, p=NS), but significantly lower from midterm follow-up studies (Z-score: −4.6, p=0). The survivorship of the onlay implant compared to TKA survivorship rate was significantly lower from early and midterm follow-up studies (Z-score: −2, p=0.0455 and Z-score: −2.2, p=0.02202, respectively).
Tibiofemoral OA progression and other complications in PFA
The overall complication rate for the inlay series was 32%, whereas it was 15% for the onlay. The complication rate of the inlay compared to onlay PFAs was significantly higher at early and midterm follow-up categories (Z-score: 11.8, p=0 and Z-score: 2.37, p=0.017, respectively).
Considering the postoperative complications other than TFOA progression, inlay PFAs showed significant higher rates compared to TKA at early and midterm follow-ups (Z-score: 6.59, p=0 and Z-score: 3.43, p=0). In contrast, no significant differences were found when comparing onlay PFA rates to those of TKA at any follow times (Z-score: 0.31, p=NS and Z-score: 1.3, p=NS).
Progression of TFOA was reported on 70 out of 865 knees (8.1%) for the inlay series, and 45 out of 1072 (4.2%) for the onlay group. If TFOA progression is to be considered a complication of isolated PFA, then the analysis is significantly affected. Considering the overall postoperative complications for the early and midterm follow-ups, TFOA progression represented the 13% of complications for the inlay series and 31.3% for the onlay series. When considering the follow-up categories, TFOA progression was significantly higher in inlay implants from early follow-up studies (Z-score: 2.3, p=0.02088), however this turned to be not significant in midterm follow-up category studies (Z-score: −0.86, p=NS).
The main findings from this systematic review confirm what was previously known:50 ,51 the combined group of inlay and onlay PFAs have a statistically significant higher complication and consequently reoperation rate, with a significant lower survivorship rate compared to TKA for PFOA. However, when taking into consideration PFA implant design, those findings are different. Inlay implants, when compared to onlay, demonstrate a significantly higher rate of complications, implying a higher number of reoperations and implant revisions, leading to a significant lower survivorship rate.
The initial two research questions were therefore adequately answered, as well as the third one: TFOA progression is a common complication affecting patients underwent inlay or onlay PFA implants. It accounts for about one-third of the complications. Series reporting on PFA outcomes confirm those results showing evidence of radiological TFOA progression in up to 25%.24 ,32 ,43 The progression of TFOA is a concern and likely related to initial patient selection and second to actual disease progression. The difference in TFOA progression rate between the inlay (8.1%, earlier models) and the onlay (4.2%, later models) may be related to better patient selection initially for surgery avoiding patients who had TFOA and therefore less likely to progress with TFOA after PFA.
It was not possible to answer the final research question since all of the studies analysed did not report on how many PFA implants were revised or converted to TKA because of TFOA progression.
Functional outcome and patients' satisfaction were not analysed because this review was mainly focused on evaluating differences in complication and reoperation rate. However, while analysing the studies, it was clear that high variability in the outcome rating scales existed to assess function and satisfaction, reflecting an overall poor methodology level.
In 2012, Dy et al51 published a meta-analysis on complication after PFA versus TKA for isolated PFOA. The main finding of that study was that with onlay implants there was no significant difference in reoperation, revision, pain or mechanical complications compared to TKA, which are similar to the results from this review. However, the current review shows a significantly lower survivorship rate for inlay and onlay PFAs compared to TKAs. This data were not previously addressed in any paper, to the authors' knowledge. Although this is not a meta-analysis, the authors used some basic statistics to achieve quantitative data. Moreover to understand the influence of follow-up time on the chosen outcomes (reoperation, revision, conversion to TKA, complication and survivorship), studies analysed were classified into three follow-up time categories according to their mean follow-up time. Although this could not be able to completely avoid bias due to length of follow-up, data were compared from studies with homogenous mean duration of follow-up.
The literature on case series of inlay PFA implants reported a patellar maltracking rate varying from 17% to 36% and a revision rate ranging from 0% to 42%,10 ,19 ,23 ,27 whereas the revision rate for onlay design ranges from 0% to 12%.8 ,9 ,17 ,35–40 ,42 ,43 ,45 The Australian National Joint Registry showed a cumulative revision rate of 20% and 10% at 5 years, respectively, for inlay and onlay designs.52
Lonner and Bloomfield14 clearly summarised the reason for such differences. The inlay implants often suffered a mismatch between patient anatomy and implant morphology; prostheses were implanted with a certain degree of flexion to avoid impingement with the distal aspect of the prosthesis and anterior cruciate ligament (ACL) or tibial plateau; the intrinsic internal rotation of the native trochlea determined the internal rotation of the implant, leading to medialisation of the implant, increase in Q angle and overtightening of the lateral retinaculum; moreover, the inlay implants had a narrow width and a deep sulcus that overconstrained the patella leading to patellar maltracking and implants did not sufficiently extend proximally therefore altering the patella engagement and causing catching and subluxation. Onlay designs aimed to solve most of those issues, in fact they were provided with an anatomic radius of curvature, to better fit the trochlea and avoiding creation of off-set at the proximal aspect of the prosthesis. The rotation is decided by the surgeon, and the implants are wider and extend more proximally to favour early patella engagement.
Those significant design improvements led to lower complication, revision and reoperation rates. However, the survivorship rate of the newer implants did not improve compared to the older ones, remaining significantly lower in respect to that of TKAs. This goes along with the significant TFOA progression rate. If TFOA is considered a complication of PFA, in this analysis, it equalled up to one-third of all complications. Although it cannot explain all causes of PFA failure, it is possible to speculate that it plays a major role in PFA long-term failure likely requiring revision. Therefore, it suggests that the correct indications, that is, strictly isolated PFOA, still represent the key for success in PFA surgery.
Several limitations affect this study. First of all, no functional outcomes analysis was carried out, as data were too heterogeneous, therefore no conclusion could be made on that issue. There was only one retrospective case–control study, while the rest were case series. Analysis of low level of evidence studies intrinsically leads to low level of evidence revision. Studies did not report accurate data on the time-to-complication from the index surgery. This made difficult to assess the influence of follow-up length. Moreover, multiple prosthetic brands were used by the same study as well as different surgical techniques and postoperative cares among the studies. This prevents authors from drawing definitive conclusions. However, the strict inclusion criteria and the rigorous methodology let the authors feel that this paper gives a reasonable overview of the current status in PFA surgery.
Higher level studies, such as prospective case–control studies or better, randomised controlled studies, comparing at long-term PFA to TKA are desirable.
Taken together, inlay and onlay PFAs showed statistically significant inferior results in terms of complication, reoperation, revision and survivorship rate when compared to TKA. However, when comparing only newer (onlay) PFA designs to TKA, complication, reoperation and revision rates were not significantly different, but survivorship still remains significantly lower. The reason for this difference is likely the inclusion of TFOA progression as a potential complication of PFA. TFOA progression seems to play a significant role in causing failure of PFA in the midterm to longer term. Therefore, appropriate patient selection with isolated PFOA seems to be the key for optimising outcomes.
Contributors SV involved in article collection and selection, manuscript writing, table editing, statistics and data analysis. RP and JE-M involved in article collection and selection supervision, and manuscript revision. BZ participated in article collection and selection and table editing. AA is responsible for supervision during all the literature analysis, table editing, planning of the work, manuscript writing and revision and language editing.
Competing interests None declared.
Provenance and peer review Commissioned; externally peer reviewed.