Patellofemoral arthroplasty with onlay prosthesis leads to higher rates of osteoarthritis progression than inlay design implants: a systematic review

Purpose The aim of this study was to report the clinical and functional outcomes, complication rates, implant survivorship and the progression of tibiofemoral osteoarthritis (OA), after new inlay or onlay patellofemoral arthroplasty (PFA), for isolated patellofemoral OA. Comparison of different implant types and models, where it was possible, also represented one of the objectives. Methods A systematic literature search following PRISMA guidelines was conducted on PubMed, Scopus, Embase and Cochrane databases, to identify possible relevant studies, published from the inception of these databases until 11.11.2022. Randomized control trials (RCTs), case series, case control studies and cohort studies, written in English or German, and published in peer-reviewed journals after 2010, were included. Not original studies, case reports, simulation studies, systematic reviews, or studies that included patients who underwent TKA or unicompartmental arthroplasty (UKA) of the medial or lateral compartment of the knee, were excluded. Additionally, only articles that assessed functional and/or clinical outcomes, patient-reported outcomes (PROMs), radiographic progression of OA, complication rates, implant survival rates, pain, as well as conversion to TKA rates in patients treated with PFA, using inlay or onlay trochlea designs, were included. For quality assessment, the Methodological Index for Non-Randomized Studies (MINORS) for non-comparative and comparative clinical intervention studies was used. Results The literature search identified 404 articles. 29 of them met all the inclusion criteria following the selection process. Median MINORS for non-comparative studies value was 12.5 (range 11–14), and for comparative studies 20.1 (range 17–24). In terms of clinical and functional outcomes, no difference between onlay and inlay PFA has been described. Both designs yielded satisfactory results at short, medium and long-term follow-ups. Both designs improved pain postoperatively and no difference between them in terms of postoperative VAS has been noted, although the onlay groups presented a higher preoperative VAS. When comparing the inlay to onlay trochlea designs, the inlay group displayed a lower progression of OA rate. Conclusion There is no difference in functional or clinical outcomes after PFA between the new inlay and the onlay designs, with both presenting an improvement in most of the scores that were used. A higher rate of OA progression was observed in the onlay design group. Level of evidence III. Supplementary Information The online version contains supplementary material available at 10.1007/s00167-023-07404-0.


Introduction
Patellofemoral arthroplasty (PFA) for treatment of isolated patellofemoral osteoarthritis (OA) remains until today a controversial subject due to inconsistent results found throughout the existing literature [20,26]. Patient selection, surgical technique, as well as implant choice have a direct effect on clinical outcomes. Historically, the first patellofemoral joint replacement was a vitallium cell patella cap designed by McKeever in 1955 [31]. Nowadays, PFA implant designs can be divided into two larger groups: inlay and onlay PFA.
First generation inlay designs, such as the Richard and Lubinus prosthesis, introduced back in 1979 [8], replaced only the worn cartilage, leaving the subchondral bone untouched. Short-term outcomes, were however not promising, with a low rate of patient satisfaction, but a high conversion rate to total knee arthroplasty (TKA) [9,45,47]. The second-generation, or onlay design, was introduced in the 1990s. Contrary to the first-generation inlay designs, the onlay trochlea prosthesis completely replaced the anterior compartment of the knee, providing a possibility of correcting trochlea rotation or for dysplasia [46].
Due to potential complications of onlay designs, such as patellar catching or anterior notching, overstuffing, alongside an increased bone loss when compared to inlay designs, new generation inlay trochlea implants have been introduced [18,21,28,34]. These implants aim to reproduce the complex kinematics of the patellofemoral joint with less mechanical and patellofemoral complications, increased implant stability and no alteration to the soft tissue tension or extensor mechanism [11,13,15,16,41].
Up-to-date studies, which report or compare clinical or functional outcomes, complication rates, revision or conversion rates, as well as progression of OA between different trochlea designs, are limited. Hence, the aim of this study is to report the clinical and functional outcomes, complication rates, implant survivorship and the progression of the tibiofemoral OA, after inlay or onlay PFA. Comparison of different implant types and models, where it is possible, also represents one of the objectives. The extended information provided from this systematic review will help physicians improve the patients' management, functional, clinical outcomes and, therefore, patient satisfaction.

Materials and methods
A systematic literature search following PRISMA guidelines [37] was conducted on PubMed, Scopus, Embase and Cochrane databases to identify possible relevant studies, published from the inception of these databases until 11.11.2022. The study protocol has been registered and approved by Prospero (CRD42022330285). The search strategy can be found in Additional Material 9. Randomized control trials (RCTs), case series, case control studies and cohort studies, written in English or German, and published in peer-reviewed journals after 2010, were included in the title and abstract screening of this review. Not original studies, case reports, simulation studies, systematic reviews, or studies that included patients who underwent TKA or unicompartmental arthroplasty (UKA) of the medial or lateral compartment of the knee, were excluded. In a second step, full text analysis was performed by two authors. Articles that assessed functional and/or clinical outcomes, patient-reported outcomes ( ), radiographic progression of OA, complication rates, implant survival rates, pain, as well as conversion to TKA rates in patients treated with PFA, using inlay or onlay trochlea designs, were included. Additionally, only articles presenting their results in numerical data form were considered. Finally, surgical technique studies, abstract only studies, studies reporting outcomes after PFA with additional UKA, robotic PFA, or reporting outcomes of the patellar components of TKA, or comparing PFA with TKA, as well as studies which did not report preoperative data, have been also excluded. In case of discrepancy regarding eligibility criteria a third author was consulted.

Quality assessment
In order to assess the quality of the included studies, the Methodological Index for Non-Randomized Studies (MINORS) for non-comparative and comparative clinical intervention studies was used [44]. The global ideal score for non-comparative studies was 16 and for comparative studies 24. The level of evidence of each included study war also reported. With the sole purpose of improving the systematic review's quality, articles which did not meet a score of at least 11 for non-comparative studies or at least 16 for comparative studies according to MINORS have been excluded.

Data extraction
Title, author names, study design, year and journal of publication, abstract, level of evidence, follow-up time, design of the trochlea implant, clinical outcomes, functional outcomes, revision rates, complication rates, conversion to TKA rates, progression of OA, as well as reported pain levels and PROMs were extracted into a Microsoft Excel spreadsheet (MS Microsoft, USA).

Statistical analysis
Continuous variables were described with means and standards deviations or medians and ranges. Categorical variables were reported with absolute and relative frequencies. A p < 0.05 was considered statistically significant.

Results
The literature search identified 404 publications in the initial screening process. Twenty-nine of them met all the inclusion criteria following the selection process. A detailed overview of the process is shown in Fig. 1. Median MINORS for non-comparative studies value was 12.5 (range [11][12][13][14], and for comparative studies 20.1 (range [17][18][19][20][21][22][23][24]. Results from a total number of 1,761 patients were evaluated (median age at surgery 53 years, range 22-92 years). The reported median body mass index (BMI) was 26.4 (range 20-50.  Table 1. In terms of OKS, 13 included studies, have reported improved postoperative scores, when compared to preoperative ones [1-4, 14, 19, 24, 32, 33, 36, 38, 42, 48]. No difference was observed between inlay and onlay implants, in terms of OKS [16]. Although, a couple of studies, which do not mention p values or confidence intervals do exist [14,33], the overwhelming majority of the findings qualify as statistically significant (p < 0.05). Patients have been followed at short, short to medium, medium and long terms. Collected data can be found in Table 2.
When discussing WOMAC, seven studies state that both inlay and onlay designs yield improved postoperative scores [1,16,17,22,23,32,40]. Feucht et al. also directly compared WOMAC scores, between onlay and inlay designs at a median follow-up of two years. There was no difference between the reported scores in the two groups [16]. WOMAC scores were reported at medium-and long-term follow-ups. All reported results are statistically significant (p < 0.05). Collected data can be found in Table 3.
In the case of ROM, 12 studies were identified for data extraction. The majority of the studies reported an increase in the postoperative ROM when compared to preoperative values [1,12,19,24,27,35,38,43,50]. Contrary to the majority, Al-Hadithy et al. reported no change in ROM, when comparing preoperative values to 12-months followup ones [4]. Furthermore, Ajnin et al. actually reported a decrease in ROM values at a median follow-up of 65 months (range: 30-119), when compared to preoperative values [2]. No studies were found which directly compared ROM values between onlay and inlay designs. ROM was reported preoperatively and postoperatively at short-, short-to-medium-, medium and long-term follow-ups. Collected data can be found in Table 4.
Regarding KSS, almost all of the nine analysed studies reported an increase in both postoperative clinical/objective scores and functional scores, when compared to preoperative ones [5,7,12,19,30,35,43,50]. Both currently circulating variants of KSS were used (KSS 1989 andKSS 2011). With the notable exceptions of Morris et al. [35], who did not mention the statistical significance and Bernard et al. [7], who did present his findings as statistically non-significant, the remaining majority of analysed studies reported their findings as statistically significant (p < 0.05). The KSS scores were reported preoperatively and postoperatively at short to medium, medium and long-term follow-ups. Collected data can be found in Table 5.

Discussion
The main finding of the present review was that both onlay and inlay PFA yield satisfactory clinical and functional outcomes at short-, medium-and long-term followups. No difference between designs has been described, although only one study from Feucht et al. [16] directly compared onlay and inlay designs using WOMAC and Lysholm scores, which presented a small and statistically non-significant difference in favour of the inlay design. Both designs improved pain postoperatively and no difference between them in terms of postoperative VAS has been noted, although the onlay group presented a higher preoperative VAS [16].
Regarding complication rates, implant survivorship and revision rates, the studies presented a high degree of heterogeneity between them. The most common complication described was the patella maltracking, followed closely by anterior knee pain.
One interesting finding of the study pertains to the progression of OA in the tibiofemoral compartment. When comparing the inlay and onlay trochlea designs Feucht et al. [16] found a statistically significant difference, in favour of the inlay group.

Conclusion
There is no difference in functional or clinical outcomes after PFA between the new inlay and the onlay designs, with both presenting an improvement in most of the scores that were used. A higher rate of OA progression was observed in the onlay design group.