Introduction

The patellofemoral (PF) joint contains the thickest articular cartilage in the human body [1]. Under physiologic conditions, the articular cartilage is able to transmit forces and disperse loads placed onto the joint. Though the PF cartilage can be considered a biphasic material similar to any other human articular cartilage, it is both more permeable and pliable than cartilage in other areas of the human body such as the tibio-femoral joint, thus causing increased susceptibility to injury [2]. Furthermore, this area is commonly prone to injury due to the high axial and shearing forces experienced by the joint [2,3,4,5]. In a retrospective analysis of 15,074 patients with chondral lesions of the knee, the patella and trochlea accounted for 36% and 8% of these lesions, respectively [6]. A major limitation to the healing of articular cartilage is its avascular quality which precludes its ability to regenerate damaged tissue [1]. With limited healing capacity, combined with the complex biomechanical environment and significant joint reactive forces, chondral defects of the PF joint are often difficult to manage [2,3,4,5].

Lesions of the PF joint are often misdiagnosed and usually recognized during treatments of other pathologies [7]. Leaving the joint untreated may predispose to secondary osteoarthritis [8, 9]. Though no gold standard exists [10], treatment options for chondral defects of the PF joint range from non-operative to various operative procedures. Surgical options range from arthroscopic debridement to cartilage restoration procedures such as conventional microfracture (MFx), autologous matrix-induced chondrogenesis (AMIC), autologous chondrocyte implantation (ACI), osteochondral autograft transfer (OAT), osteochondral allograft transplantation (OCA), and arthroplasty [11,12,13,14,15]. Three generations of ACI techniques exist and include: (1) ACI with a periosteal cover (ACI-P) [16, 17]; (2) ACI with a collagen cover (ACI-C) [17]; and (3) ACI with a collagen matrix (ACI-M) [17].

Historically, cartilage restoration procedures have been reserved for the tibio-femoral joint, but recent advances in technology have expanded the applications to the PF joint. The purpose of this study was to systematically assess the trends in surgical techniques, outcomes, and complications of cartilage restoration of the patellofemoral (PF) joint. It was hypothesized that ACI would be the most commonly reported cartilage restoration technique with short-term outcomes similar to other procedures used to treat chondral defects of the PF joint.

Materials and methods

Search strategy

The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines were followed in the development of this study [18]. Three online databases EMBASE, MEDLINE, and PubMed were searched for literature addressing the surgical techniques to treat cartilage defects of the PF joint from January 1, 2007 to April 30, 2018. Articles were searched from 2007 and onwards to determine the overall trends in cartilage restoration techniques used in the past decade for focal cartilage defects of the PF joint. The search terms included “patellofemoral joint”, “cartilage”, “trochlear groove”, and similar phrases (Supplementary Table 1). The research question, and inclusion and exclusion criteria were established a priori. Inclusion criteria were: (1) all levels of evidence; (2) surgical studies treating isolated patellofemoral chondral defects (all grades) using one or more of the following: any generation of ACI, conventional MFx, AMIC, OCA, OAT, cell-free osteochondral scaffold (COS), and De Novo NT® graft; (3) studies reporting clinical outcomes; (4) English studies; (5) human studies; and (6) patients aged ≥ 18 years. The exclusion criteria were (1) studies reporting non-surgical treatment; (2) studies with < 5 patients; and (3) arthroplasty.

Study screening

A systematic screening approach in accordance with PRISMA was employed in duplicate by two independent reviewers from title to full-text screening stages. Discrepancies at the title and abstract stages were resolved by automatic inclusion to ensure that relevant articles were not missed. Discrepancies at the full-text stage were resolved by consensus between the reviewers. The input of a third, senior reviewer was used if a consensus could not be reached. Search terms were entered onto Google Scholar and references of included studies were also screened using the same systematic approach to capture any additional relevant articles.

Quality assessment of included studies

The methodological index for non-randomized studies (MINORS) appraisal tool was used to assess the quality of the included, nonrandomized studies (e.g., case reports, case series, cohorts, etc.) by two independent reviewers with discrepancies resolved by consensus discussion [19]. A score of 0, 1, or 2 is given for each of the 12 items on the MINORS checklist with a maximum score of 16 for non-comparative studies and 24 for comparative studies. To the authors’ knowledge, there is no evidence to categorize the MINORS score. Thus, methodologic quality was categorized a priori as follows: a score of 0–8 or 0–12 was considered poor quality, 9–12 or 13–18 was considered fair quality, and 13–16 or 19–24 was considered excellent quality, for non-comparative and comparative studies, respectively.

Data abstraction

Two reviewers independently abstracted relevant data from included articles and recorded the data onto a Google Spreadsheet designed a priori. Demographic data included author, year of publication, sample size, study design, level of evidence [20], patient demographics (e.g., sex, age, etc.), and details of surgical techniques used to treat osteochondral defects. Furthermore, any information regarding rehabilitation protocols, post-operative outcomes (surgical and radiographic), concomitant lesions, associated procedures, and complications were documented.

Statistical analysis

In consultation with the study statistician, the high statistical and methodological heterogeneity amongst included studies precluded performing a meta-analysis. Descriptive statistics such as mean, range, and measures of variance (e.g., standard deviations, 95% CI) are presented where applicable. A weighted intraclass correlation coefficient (ICC) was used to evaluate inter-reviewer agreement for the MINORS score. A kappa (κ) statistic was used to evaluate inter-reviewer agreement at all screening stages. Agreement was categorized a priori as follows: ICC/κ of 0.81–0.99 was considered as almost perfect agreement; ICC/κ of 0.61–0.80 was substantial agreement; ICC/κ of 0.41–0.60 was moderate agreement; 0.21–0.40 fair agreement and a ICC/κ value of 0.20 or less was considered slight agreement [21]. A two-proportion z test was used to determine whether the differences between the proportions of cartilage restoration techniques used from 2007 to 2012 and 2013 to 2018 were statistically significant. These time frames were selected to determine if the proportion of cartilage restoration techniques for the PF joint changed over the latter half of the past decade.

Results

Study characteristics

The initial search yielded 4022 studies, of which 27 met the inclusion for this review (Fig. 1). Upon reviewing references of included studies and a search on Google Scholar, an additional study was retrieved from the references and included in the review for a total of 28 studies. Of the 28 included studies published between 2007 and 2018, there were seven prospective cohorts, one retrospective cohort, one case–control, and 19 case series. There were no studies published in 2018 (Table 1).

Fig. 1
figure 1

PRISMA flow diagram

Full size image

Patient characteristics

There were 708 patients (824 knees) with a mean age of 39.5 ± 10.5 years; two studies did not report the mean age [22, 23]. Patients had a mean follow-up of 39.1 ± 16.0 months. At the final follow-up, 92.3% (n = 653) of patients were available. Of the patients available at the final follow-up, 40.1% (n = 284) were male; sex distribution was not specified in 15.7% (n = 111) of the patients. The mean patellofemoral cartilaginous defect size was 4.2 ± 2.2 cm2, and patients were treated with either ACI (45.5%; n = 375), conventional MFx (29.6%; n = 244), AMIC (9.1%; n = 75), OAT (8.1%; n = 67), COS (4.1%; n = 34), De Novo NT® graft (1.8%; n = 15), or OCA (1.7%; n = 14). The ACI techniques used in this review were either ACI-M (22.2%; n = 183), unspecified (12.1%; n = 100), ACI-C (8.1%; n = 67), or ACI-P (3.0%; n = 25). From 2007 to 2012 and 2013 to 2018, the proportion of PF cartilage restoration techniques varied (Fig. 2). There was a significant decrease in the use of the first and second generation ACI (p < 0. 001), AMIC (p = 0.02), and conventional MFx (p < 0.001) between 2007 and 2012 and 2013 to 2018. Meanwhile, there was a significant increase in the use of ACI-M (p < 0.001), OAT (p < 0.001), COS (p < 0.001), De Novo NT® graft (p < 0.001), and OCA (p < 0.001) between 2007 and 2012 and 2013 to 2018. The defect locations included the patella (51.1%; n = 421), unspecified (33.6%; n = 277), trochlea (12.6%; n = 104), and both the patella and trochlea (i.e., bipolar/kissing lesions) (2.5%; n = 20). Only one study reported concomitant chondral defects in the tibio-femoral joint; however, only defects of the PF joint were treated.

Study quality

Seven of the included studies represented Level-II evidence, two studies represented Level-III evidence, and 19 studies represented Level-IV evidence (Table 1). There was excellent agreement among reviewers at the screening stages involving the title (k = 0.797; 95% CI 0.767–0.828), abstract (k = 0.913; 95% CI, 0.877–0.950), and full text (k = 0.914; 95% CI 0.856–0.972). There was high agreement among the quality assessment scores based on the MINORS criteria (ICC = 0.912; 95% CI 0.891–0.928). The mean MINORS score for non-comparative and comparative studies was 11.3 ± 2.3 and 16.3 ± 2.6, respectively, indicating fair quality of evidence for non-randomized studies (Table 1).

Table 1 Characteristics of included studies
Full size table
Fig. 2
figure 2

Proportion of patellofemoral cartilage restoration techniques used in 2007–2012 and 2013–2018. Asterisk indictates statistically significant differences between 2007 to 2012 and 2013 to 2018. ACI autologous chondrocyte implantation, ACI P&C ACI with a periosteal cover and collagen cover, ACI-M ACI with a collagen matrix, AMIC autologous matrix-induced chondrogenesis, MFx microfracture, OAT osteochondral autograft transfer, COS cell-free osteochondral scaffold, OCA osteochondral allograft transplantation

Full size image

Patient outcomes

ACI outcomes (first, second, and third generations)

The mean defect size for patients treated with ACI was 4.9 ± 1.9 cm2; two studies did not report the mean defect size [36, 47]. Overall, 12 out of 13 studies (n = 351) reported significant post-operative improvement on a wide range of patient important outcome measures. Specifically, of the eight studies (n = 208) reporting the Subjective International Knee Documentation Committee (IKDC) score, a significant (p < 0.05) improvement post-operatively was found in six studies (n = 114). The Tegner score was reported in three studies (n = 82), of which all found significant improvement post-operatively (p < 0.001). Seven studies (n = 168) reported the Lysholm Knee Score, of which five studies (n = 74) found a significant improvement post-operatively (p < 0.05). Finally, four studies (n = 119) reported significant improvements in the Cincinnati scores post-operatively (p < 0.05) (Table 2).

Table 2 First, second, and third generation ACI outcomes
Full size table

AMIC outcomes

The mean defect size for patients treated with AMIC was 3.6 ± 1.1 cm2. In the studies that used AMIC for PF joint cartilage restoration, four studies reported on the Visual Analog Scale (VAS), of which three studies (n = 35) found improvement post-operatively (p < 0.0001–0.01), whereas one other study (n = 5) found improvement, but did not report p value. Two studies (n = 30) reported subjective IKDC score—both these studies found improvement post-operatively (p < 0.0001–0.01). Two out of the three studies that reported the on the Knee Injury and Osteoarthritis Outcome Score (KOOS) (n = 20) found improvement post-operatively (p < 0.01–0.047), whereas the one other study (n = 5) found improvement, but did not report a p value. Two out of the three studies that reported the Tegner score (n = 30) found improvement post-operatively (p = 0.006–0.047), while the one other study (n = 5) found improvement, but did not report p value (Table 2).

Other techniques/associated procedures

The outcomes of COS, De Novo NT®, OCA, OAT, and conventional MFx are summarized in Supplementary Tables 2–6. The types of concomitant PF realignment procedures included proximal (30.6%; n = 252), distal (12.4%; n = 102), combined proximal and distal (2.3%; n = 19), and unspecified (1.0%; n = 8). Meanwhile, of the 12 studies that reported the concomitant use of PF realignment procedures in select patients, six studies did not stratify their results. Patients undergoing PF realignment (32.4%; 11/34) in addition to COS, had a slower recovery and lower functional scores (e.g., IKDC subjective and Tegner activity scores) at 12 months post-operatively, but no differences at 24 months when compared to those without realignment (Table 3).

Table 3 AMIC outcomes
Full size table
Table 4 Complications
Full size table

Complications

There were a total of 76 (9.2%) complications of which graft hypertrophy (29.7%; n = 22) was the most prevalent. Of the patients with graft hypertrophy, 36.3% (8/22) underwent ACI-P as their index procedure. Meanwhile, there were a total of eight failures in this systematic review: three patients received an unspecified generation of ACI, two patients received ACI-M, two patients received OCA, and one patient received ACI-P. The most common causes of a failure were: (1) > 25% delamination, revision cartilage repair surgery, or prosthesis implantation (37.5%; 3/8) and (2) graft delamination and/or exposed subchondral bone on MRI (25%; 2/8). Conversion to joint arthroplasty occurred in 3 patients: total knee arthroplasty (25%; 2/8) and PF arthroplasty (16%; 1/8). The distribution of complications for restoration techniques was as follows: AMIC (31.6%; 24/76), ACI-P (19.7%; 15/76), ACI-M (13.2%; 10/76), OAT (9.2%; 7/76), ACI-C (6.6%; 5/76), and OCA (2.6%; 2/76). There were no complications reported in the conventional MFx group. The complications in this systematic review are summarized in Table 4.

Discussion

Key findings

The most striking finding was the change in practice trends over the last 5 years: a rise in newer generation ACI techniques and a decrease in conventional MFx. Although conventional MFx has been considered the historical gold standard for treating cartilage defects of the tibio-femoral joint, there was limited use of this procedure for the PF joint. Only two of the included studies, both published in the 2007–2012 period (comprising 30.1% of the total sample size), used conventional MFx to treat PF cartilage defects [38, 41]. There was also a statistically significant decrease in the use of the AMIC technique, an enhanced MFx technique that uses a protecting membrane [38]. This may be in part due to the growing understanding that MFx is not suitable for larger, uncontained lesions; the mean defect size of included patients in this review was > 4 cm2. In addition, surgeons at large academic centers are more likely to adopt, investigate, and publish the results of newer techniques and technologies. Finally, the a priori hypothesis that ACI would be the most commonly reported cartilage restoration technique with short-term outcomes similar to other procedures used to treat chondral defects of the PF joint has been confirmed.

When comparing the use of the conventional MFx for different compartments in the knee, those treated for chondral defects of the femoral condyles had significantly better (p < 0.02) defect filling when confirmed by MRI at 36 months post-operatively compared to the trochlear, tibial, and retropatellar lesions [50]. Moreover, the quality of bone marrow that can be stimulated from the patella (a sesamoid bone), particularly in a patient population with a mean age of nearly 40 years, such as found in this review, may not be ideal to produce the desired effect of MFx. In an age-dependent study of patients with full-thickness, multi-compartmental chondral lesions in the knee (i.e., both tibio-femoral and PF) treated with conventional MFx, it was found that deterioration was significantly pronounced in those aged 40 years or older [51]. The authors concluded that the reduced regenerative capacity in older patients may be a result of aging chondrocytes having decreased functionality [51]. Though there are conflicting reports of the long-term durability of MFx,[52], its prevalent use comes into question given its production of fibrocartilage repair tissue as opposed to hyaline or hyaline-like repair tissue [53]. As the PF joint experiences a greater compressive force at a greater knee flexion angle compared to the tibio-femoral joint [54], intuitively, it is without surprise that MFx is not as durable in these patients. Finally, there is mounting evidence that the previous MFx can compromise outcomes of revision surgery using cell-based cartilage repair therapy due to its disruption of the subchondral plate [38, 55, 56].

Noyes and Barber-Westin conducted a systematic review of cartilage defects of the PF joint treated with arthroplasty and a limited number of restoration techniques [15]. The focus of the review was on patients under 50 years of age with patellar defects larger than 4 cm2, and studies were included from 1992 to 2012. Their findings revealed high failure rates across ACI, OAT/OCA, and arthroplasty techniques (22–60%) with no ideal technique identified. Similarly, our review did not find any one technique to be superior to others, even in the two comparative studies that were included. In contrast to the earlier article, however, our review found much lower complication and failure rates, though these were still not uncommon occurrences. This may be a reflection of improving technology, increased familiarity with novel techniques, and more awareness of the appropriate indications for the use of each technique since the publication of Noyes et al. [15]. Jungmann et al. identified risk factors associated with need for revision surgery in 413 patients after ACI (first, second, and third generations) [57]. At average follow-up of 4.4 ± 0.9 years, ACI-P was a significant predictor of revision and failure (p < 0.28) [57]. Similarly, this systematic review found higher complication rates with older generation ACI techniques, with graft hypertrophy being most common, which has been a particular concern with the periosteum patch-covered ACI technique [58].

Tibio-femoral or tricompartmental osteoarthritis is more common than isolated PF osteoarthritis, and this has been borne out in both radiographic and population-based studies [59,60,61]. Perhaps, for this reason, among others, the literature on tibio-femoral cartilage preservation and restoration is much more abundant than the literature on the use of these techniques in the PF joint. Mundi et al. conducted a systematic review and meta-analysis of level 1 studies for cartilage restoration of the knee (i.e., tibio-femoral joint) [62]. When comparing ACI to marrow stimulation (MFx), there was a trend toward favoring ACI in functional outcomes (i.e., IKDC, Lysholm and KOOS; p = n.s) and no difference between the two techniques for pain score outcomes. Since that meta-analysis, a number of new randomized controlled trials (RCTs) have been conducted on this topic all with mixed results [52, 63,64,65]. Overall, our findings regarding the PF joint are consistent with the tibio-femoral literature in that no technique is clearly superior to others and that most techniques do provide significant post-operative improvement, but those with longer term follow-up, appear to show higher rates of failure and/or revision surgery.

Finally, in a climate of rapidly growing healthcare costs [66], cost-effectiveness is an important consideration in assessing the overall value of a new technique or technology. A recent systematic review comparing the three most common surgical procedures for treatment of focal cartilage defects of the knee found that older generation ACI when compared to conventional MFx and OAT was least cost-effective. Moreover, authors of this study concluded that although it may result in smaller functional gains, the conventional MFx remains a reasonable, cost-effective option for cartilage restoration given its technical simplicity, widespread availability, and minimal invasiveness [67]. Treatment of chondral defects of the PF joint, with the various cell-based therapies, has tremendous potential for impact given the young age of patients and associated functional limitations. As the prevalence of newer generation ACI rises, the value in generating economic benefits to society relative to the direct costs of surgery and rehabilitation needs to be explored.

The strengths of this systematic review stem from its thorough methodology on a novel topic of expanding research. The systematic screening approach was employed in duplicate, thus minimizing reviewer bias. Moreover, the comprehensive search used ensured that all relevant articles were captured. There was excellent agreement between the two reviewers at all screening stages and quality assessment. This systematic review also consisted of numerous studies with moderate-to-large sample sizes, resulting in an overall large sample size with moderate length follow-up.

This review had some limitations. First, there was large variability in the outcomes used and inconsistent reporting of post-operative outcomes in patients across studies. Consequently, estimating the efficacy of cartilage restoration techniques for the PF joint and analyzing the differences between subgroups (e.g., PF realignment versus no realignment) were limited. With moderate follow-up periods across studies, long-term complications such as time to PF arthroplasty and graft failure could potentially be underreported. Moreover, the high statistical and methodological heterogeneity amongst included studies precluded the performance of a meta-analysis. Studies also did not have consistent documentation of their complications and thus may have been poorly reflected during the follow-up period. Finally, though there was a high quantity of studies available on the topic, the available quality of evidence limited the ability to make definitive conclusions on a superior restoration technique.

This manuscript provides clinicians/knee specialists insight into the changing trends of cartilage restoration of the PF joint. The advent of innovative cell-based therapies has brought new applications to the field, particularly for the PF joint, which historically has been met with unfavorable results. Future studies should better document and report their surgical and/or radiographic outcomes. By standardizing the instruments used to evaluate patient outcomes, accurate analyses can be made for each cartilage restoration technique. Future studies should also consider the use of either RCT study designs, large prospective cohorts, or patient registries with longer term follow-up to better evaluate and compare cartilage restoration techniques and assess revision rates and conversion time to PF arthroplasty. Finally, a cost–benefit analysis of each cartilage restoration technique is required to further guide clinicians on the optimal treatment for focal cartilage defects of the PF joint.

Conclusion

ACI was the most common restoration technique. The use of the third generation ACI has increased with a concurrent decline in the use of conventional MFx over the latter half of the past decade (p < 0.001). Overall, the various cartilage restoration techniques reported improvements in patient reported outcomes with low complication rates. Definitive conclusions on the optimal treatment remain elusive due to a lack of high-quality comparative studies.