A randomized controlled trial demonstrating sustained benefit of autologous matrix-induced chondrogenesis (AMIC^®) over microfracture: 10-year follow-up

Surgery for focal knee cartilage lesions pursues the goal of chondral repair in order to restore full, pain-free joint function and preventing, or at least delaying, the early onset of osteoarthritis. In clinical practice, a plethora of surgical interventions aim to repair cartilage tissue and overcome the poor healing potential of articular cartilage. Understandably, a better quality of the repair tissue should result in a better long-term clinical outcome [1, 2].

Microfracture (MFx) perforates the subchondral bone in order to access the bone marrow compartment, releasing mesenchymal stem cells that can differentiate but are also immunoregulatory, serving to foster regenerative microenvironments in areas of tissue injury [3]. Therefore, a more comprehensive term for this is bone marrow stimulation (BMS) [4]. As a basic BMS procedure, MFx can provide good, initial results, but unfavourable long-term results have been reported, especially in larger lesions [5]. Considering this limitation, the combination of MFx with a collagen I/III scaffold, referred to as the autologous matrix-induced chondrogenesis (AMIC^®) procedure, offers another prospective treatment option to overcome this burden. The bioresorbable membrane stabilizes the “super clot”, reduces edge loading of the surrounding cartilage and supports chondral differentiation by providing a biological chamber [6, 7] and therefore optimizes the conditions for successful cartilage repair, with the attendant clinical benefits. We therefore hypothesized that adding a collagen I/III scaffold onto a microfractured area in focal cartilage defects in the knee would result in superior outcome than with MFx alone.

Accordingly, we initiated a prospective, randomized controlled clinical trial (PRCT) to evaluate outcomes of both therapies. Short-term results had shown comparable improvements in pain and function [8]. These results, however, started to diverge by 5 years, with the MFx cohort displaying a worsening of pain and outcome scores in comparison with AMIC^®-treated patient [9]. In continuation, the aim of this current study was to evaluate the 10-year outcomes of these cohorts.

Of the 37 patients for whom we had 10-year data, MRIs were available for 32 (86%). The time between surgery and the date of the most recent MRI is shown in Table 2, with no significant difference between groups with respect to the date from surgery to the date of the MRI (p > 0.05).

The mean overall MOCART scores at 10 years were 37.7 ± 29.3 (MFx), 34.4 ± 23.2 (AMIC glued^®) and 31.0 ± 20.3 (AMIC^® sutured) points, respectively. A Kruskal–Wallis test showed no difference in the MOCART scores between the treatment groups (p = 0.879). The effusion was, at least in absolute terms, lower among the AMIC^®-treated patients (data not shown). Regarding changes in subchondral bone, which would include osteophytes, comparable proportions of patients in each group showed evidence of changes in subchondral bone. Figure 5 shows a representative case MRI from one 46-year-old male patient treated with AMIC^® glued at the medial femoral condyle.

As a result of this PRCT, treating isolated cartilage defects in the knee with an AMIC^® procedure, either glued or sutured, demonstrated significantly higher Modified Cincinnati scores and lower VAS pain compared to patients having received a MFx alone at 10 years. Thus, our hypothesis was confirmed.

Over a 10-year follow-up for the surgical repair of focal chondral defects, our results showed that the initial outcomes were comparable between the treatment groups for the first 2 years, but diverged as time went on. Both AMIC^® cohorts maintained their improvement, whether measured via the VAS or the Modified Cincinnati score, while the MFx patients exhibited a worsening of these scores at the 5- and 10-year follow-ups. MOCART scores, however, were similar between all groups at the final follow-up. Additionally, with regard to safety, adverse events and failures were similar between the patient cohorts. Another result was a notable divergence between the treatment arms in relation to the clinical responder rate in the Modified Cincinnati score. Set at a minimum of 65 points, the proportion of MFx patients exceeding this threshold was notably smaller than seen in either of the AMIC^® groups (Fig. 4).

AMIC^® is a technique within the broader category of matrix-augmented bone marrow stimulation (mBMS). To our knowledge, this is the first PRCT comparing AMIC^® procedures versus conventional MFx over a 10-year period. Additionally, other studies have shown a sustained improvement in outcomes following the AMIC^® procedure. A recent analysis, based on an ongoing registry, has shown a stable clinical improvement up to 7 years post-operative [17]. Similarly, an RCT that had compared AMIC^® to AMIC^® and bone marrow aspirate concentrate (BMAC) demonstrated clinical improvements that were maintained up to 100 months status-post for both groups [18]. While both of those publications have shown sustainable outcomes over the longer term, the registry had no comparators, while the RCT compared the standard AMIC^® procedure to an AMIC^® procedure that added BMAC to the treatment site. Thus, a limitation is the lack of a comparison with different surgical procedures.

While microfracture is a simple, single-step arthroscopic procedure that has the longest clinical history, in terms of histological outcome, MFx manifested poorer tissue quality than other cartilage repair procedures [19]. In comparison with MFx, case–control studies have indicated that AMIC^® provides superior clinical outcomes and lower rates of failure or revision. Improved IKDC, Lysholm and pain scores in the AMIC^® group along with a lower rate of failure and a trend towards a lower rate of revision were noted when compared to MFx [20].

Aside from the AMIC^® procedure, there are a number of different scaffolds used in mBMS. Examples include cell-free type-1 collagen, as well as aragonite-based, chitosan, hyaluronan-based and a biomimetic nanostructure. Recently, a meta-analysis reported a significant improvement in outcomes for scaffold-associated repair procedures compared to microfracture at 2 years for focal cartilage defects in the knee of 1699 patients at 2 years [21]. Likewise, another meta-analysis noted significantly greater improvements in 744 patients with MFx + augmentation in the Lysholm score and post-operative MOCART scores compared with an isolated MFx treatment after 26.7 (12–60) months. Here, the mean chondral defect size ranged from 1.3 to 4.8 cm² [19]. However, most of these procedures are quite limited with regard to clinical data, in contrast to the procedure that we perform which has a quantity of clinical data such that meta-analysis could be published [22].

Based on these data, the German Orthopaedic and Trauma Society (DGOU) guideline has stated that mBMS is standard of care treatment for focal chondral or osteochondral defects ranging from 1 to 4.5 cm² [23]. In the context of mBMS, the use of the Chondro-Gide^® membrane in the AMIC^® procedure is based on sound clinical data, and the DGOU consensus statement rated it having the best clinical evidence [23]. AMIC^® and other mBMS procedures, however, are not the only procedures that have shown improved outcomes for chondral repair. The results from case series concerning ACI have shown a significant clinical improvement up to 25 years [24], while the clinical superiority of ACI relative to MFx at 5 years has also been published [25]. However, an RCT that compared collagen-covered ACI to AMIC^® noted no differences in outcomes after 2 years in patients with large defects (5 cm²) [26]. Another surgical option is osteochondral autologous cylinder transfer, for which long-term results are available, but some data suggest an increased risk of failure after 2 years, thus a caveat that these procedures might be more appropriate for smaller lesions [27].

Specific to the strength of clinical evidence and level 1 studies, there are relatively few RCTs that have compared surgical techniques for the repair of focal chondral defects. Similar to our data using MFx as a control group, an RCT that compared mosaicplasty to MFx reported better patient outcomes in the mosaicplasty cohort [28]. In contrast, an RCT that compared MFx to ACI noted that both groups improved their clinical scores in the short-, medium-, and long-term evaluations, with no significant difference at the long-term follow-up [29]. Most recently, a large RCT that evaluated outcome among a variety of procedures had stated that there was no evidence of a difference in Lysholm scores between ACI and alternative techniques, among which was AMIC^® [30].

With regard to objective measures of cartilage repair, while the clinical data indicated that patients treated using the AMIC^® surgical treatment had better long-term outcomes, our radiological data were inconclusive. While 3 of the 14 AMIC^®-sutured patients showed complete filling of the treated site (21%), none of the MFx patients presented with this level of filling. With regard to joint effusion, 8/9 (88%) of the MFx patients for whom a 10-year MRI was available showed evidence of joint effusion, either mild or moderate, while this was 5/9 (55%) of the AMIC^®-glued patients and 5/14 (36%) of the AMIC^®-sutured patients. Our MRI data noted comparable rates of bony hypertrophy in the cohorts: AMIC^® 8/23 (35%) and MFx 3/9 (33%), which is comparable to what has been reported for mBMS or ACI [31]. Specific to MRI, recent research suggests that MOCART scores relate poorly to clinical outcomes [32]; therefore, the use of MOCART on assessment of surgical outcomes may have limited value.

We acknowledge our study limitations, the most obvious being low patient numbers at both enrolment and at 10 years. One of the challenges that face any PRCT is patient enrolment, and our study was no exception to this. However, in comparison with our 2-year and 5-year data [9], few patients were lost to follow-up between the 5-year and 10-year follow-up (Fig. 1). Another point worth noting in relation to the outcome scores is the lesion size. The defects for MFx (range 2–4.6 cm², mean 2.9 cm²) were significantly smaller than defects in the AMIC^® groups (range 2.4–6.3 cm², mean 3.9 cm²). While current recommendations limit MFx to defects ≤ 2 cm² [23] and certainly no surgeon would today consider MFx for such lesions, it needs to be kept in mind that this guidance was developed several years after the surgeries in this study were performed. Indeed, the lesion size may impact the decrement in outcomes that we observed among the patients in the MFx cohort. Among the 3 groups in this study, there were 5 patients whose defects were greater than 5 cm². When reviewing the case report forms, we noted that all of these patients had been randomized to an AMIC^® treatment, with 2 undergoing fixation via sutures and 3 whose repairs were secured using fibrin glue. Thus, the worst cases scenarios, with respect to defect size, had undergone an AMIC^® repair technique. Unfortunately, our sample size precludes a meaningful analysis, for example a linear regression, with regard to defect size and outcomes. The response rate is also an avenue for criticism. While there are established PASS values for IKDC and Lysholm scores, we approximated a PASS level for the Modified Cincinnati knee score, and while this is admittedly an arbitrary choice, it is close to both Lysholm and IKDC values [16].

Despite these limitations, our data provide evidence that the clinical outcomes for AMIC^®, regardless of fixation, are superior to those of MFx alone for the treatment of focal, chondral lesions in the knee over 10 years and are consistent with our 5-year results [9].

For the treatment of focal chondral lesions of the knee, the AMIC^® procedure, compared to microfracturing alone, maintains functional improvements and provides significantly better clinical outcomes at 10 years post-operatively.