Microfracture technique versus osteochondral autologous transplantation mosaicplasty in patients with articular chondral lesions of the knee: a prospective randomized trial with long-term follow-up

Chondral or osteochondral lesions of the knee eligible for cartilage repair surgery are diagnosed in 5–10 % of all knees subjected to knee arthroscopy [1, 20] and may contribute to disability and premature osteoarthritis (OA) [29]. Furthermore, focal chondral lesions of the knee have been shown to impair quality of life similar to patients scheduled for knee replacement, even though the chondral lesion patients are 30 years younger [18].

Various cartilage repair techniques have been developed. Resurfacing techniques include abrasion arthroplasty [24], Pridie drilling [36] and microfracture technique (MF) [3, 43]. MF procedures stimulate and recruit mesenchymal cells from the subchondral bone marrow and subsequently form a fibrin clot that eventually turns into a predominantly fibrocartilaginous regenerate with inferior biomechanical characteristics compared to native hyaline articular cartilage [11]. Despite fibrocartilage formation, several short- to mid-term follow-up studies following MF treatment of chondral lesions report significant pain relief and improvement in knee function [32, 33, 43].

Grafting and transplantation procedures, like autologous chondrocyte implantation (ACI) [6] and osteochondral autologous transplantation (OAT) mosaicplasty [16] gained popularity after introduction in the 1990s. The OAT mosaicplasty technique involves open or arthroscopic transplantation of multiple cylindrical osteochondral grafts from the relatively less weight-bearing periphery of the articular surface to the cartilage defect, thus providing a hyalinecartilage-covered resurfacing [2, 22]. Case series and comparative trials have reported 83–92 % good to excellent short- to mid-term results following OAT mosaicplasty [8, 13, 15]. Even though MF and OAT mosaicplasty have proven to be effective in short- to mid-term follow-up studies, knowledge regarding long-term outcome remains uncertain [4, 8, 14, 32, 41, 42]. To our knowledge, there is only one prospective randomized study comparing the long-term outcomes following MF and OAT mosaicplasty [12]. Due to the limited information on the long-term outcome after these two common cartilage repair techniques, patient information and decision-making regarding treatment options is challenging for the orthopaedic surgeon.

In the present prospective randomized study, the purpose was to compare long-term functional and radiological outcome following MF and OAT mosaicplasty for full-thickness chondral lesions of the knee. The null hypothesis was that there is no difference in patient-reported outcomes or radiographic OA between MF-treated patients and OAT mosaicplasty-treated patients at long-term follow-up.

Twenty-five patients [mean age 32.3 years, standard deviation (SD) 7.7] were enroled in the study between November 2000 and June 2006. Three orthopaedic cartilage repair centres participated in the study, and experienced knee surgeons performed both the selection of the patients and the surgical procedure. Informed consent was obtained from all patients.

Inclusion criteria were an arthroscopically verified chondral or osteochondral lesion of International Cartilage Repair Society (ICRS) grade 3 or 4 [7] located on the femoral condyle or trochlea, with an area between 2 and 6 cm² and depth <10 mm. Additionally, the patients had to be 18–50 years of age with Lysholm score <80 and Tegner score <6.

Exclusion criteria were radiographic osteoarthritis (OA), major malalignment, major ligament injury or instability, extension deficit >3°, flexion deficit >5° and chondral lesion(s) of ICRS grade 3 or 4 on the tibial plateau or patella. Patients were also excluded if they had contralateral impaired knee function that might influence the ability to follow the rehabilitation protocol.

Randomization between MF and OAT mosaicplasty was performed in the operating room, following arthroscopic debridement. Patients were randomized by a restricted shuffled approach [39] in blocks of 10, allocation ratio 1:1, using sequentially numbered sealed envelopes to assign treatment. The block randomization approach used ensured that all centres/surgeons performed both procedures and also ensured randomization to surgeon. Twenty-five patients were included, and in accordance with randomization, 14 patients were treated with OAT mosaicplasty and 11 patients with MF. Group characteristics at inclusion are shown in Table 1.

A total of 19 patients were excluded from the study. In most cases, this was due to findings during the arthroscopic assessment, e.g. size or localization of the chondral lesion not in accordance with the inclusion criteria or additional ICRS grade 3 or 4 chondral lesions of the tibia or patella. Two patients declined surgery due to pregnancy, and two patients withdrew their consent at the time of surgery as they insisted on being treated with one of the surgical techniques.

All outcome measures were obtained both at baseline and follow-up, except for isokinetic muscle strength measurements, which were performed only at follow-up. In addition to the outcome measures, all patients were also questioned about any additional surgical procedures to the knee during the follow-up period.

At a median follow-up of 9.8 years (range 4.9–11.4 years), all patients (25/25) reported Lysholm score and KOOS. One patient had moved abroad, and another was not available for examination in the outpatient clinic. However, these patients were contacted by postal mail and telephone, and returned their questionnaires.

Mean Lysholm score for patients treated with MF and OAT mosaicplasty at baseline and at follow-up are shown in Fig. 1. No significant differences in mean Lysholm score were detected between MF-treated patients and OAT mosaicplasty-treated patients at follow-up (n.s.), or in mean change from baseline to follow-up (Table 2). MF-treated patients scored 48.2 (95 % CI, 38.2–58.2) preoperatively and OAT mosaicplasty-treated patients 49.2 (95 % CI, 43.0–55.4). The mean Lysholm score at follow-up in the MF group was 69.7 (95 % CI, 55.1–84.4) compared to 62.6 (95 % CI, 52.6–72.6) in the OAT mosaicplasty group. The increase in Lysholm score from baseline to follow-up was significant for both groups (Table 2).

The KOOS profiles with mean scores at inclusion and at follow-up for the MF group and the OAT mosaicplasty group are shown in Fig. 2. There were no significant differences between the two groups in any of the KOOS subscales at follow-up or in the changes from baseline to follow-up (Table 2). The increase in KOOS from baseline to follow-up within the treatment groups was significant for all subscales except for ADL in the microfracture group, and pain, symptoms and ADL in the OAT mosaicplasty group (Table 2).

Isokinetic muscle strength measurements (n = 22) of the knee extensors and flexors at follow-up are shown in Table 3. There were no significant differences between the MF group and OAT mosaicplasty group in mean strength deficit of the affected knee. A significant mean extension strength deficit of the affected knee, compared to the unaffected, was detected in the OAT mosaicplasty group.

Mean body mass index (BMI) at follow-up was 28.2 (SD 4.2) for patients treated with MF and 27.9 (SD 3.8) in the OAT mosaicplasty group.

Reoperations and additional surgical procedures during follow-up are outlined in Table 4.

The main finding of the present study is that the long-term outcomes following MF and OAT mosaicplasty for treating focal chondral lesions of the knee are comparable. The evidence in this material is not sufficient to reject the study hypothesis that there is no difference between the two alternative treatments. However, the small number of included patients makes any firm conclusions regarding the hypothesis testing difficult. Due to less eligible patients for the study than expected, the duration of the inclusion period was extended up to 5 years. Still only 25 patients were enroled in the study. However, no patients were lost to follow-up.

Reoperations occurred in 6/11 patients (54 %) in the MF group and in 5/14 patients (36 %) in the OAT mosaicplasty group. Even though non-significant, all knees that underwent a second cartilage repair procedure (n = 3) or a total knee arthroplasty (n = 1) belonged to the MF group. It should also be noted that a significant reduction in extension force of the affected leg, compared to the unaffected, was found in the OAT mosaicplasty group, even though a mini-invasive arthrotomy was used when possible.

Both treatment groups reported significant improvement in Lysholm score and in several of the KOOS subscales from baseline to follow-up at 9.8 years. However, the mean Lysholm score and KOOS at follow-up were considerably lower than in the reference population [5, 35], which indicates that the long-term patient-reported outcomes are modest for both treatments. In addition, the wide confidence intervals indicate diversity among the patients, which however, is not an uncommon finding in long-term follow-up studies on cartilage repair [4, 45]. The unpredictability of these two cartilage repair methods has been found in standardized controlled animal studies as well [19].

To our knowledge, there are only two other clinical studies comparing MF and OAT mosaicplasty [12, 28]. In the only randomized trial, the OAT mosaicplasty-treated patients scored significantly higher on the ICRS outcome scores and Tegner scores compared to the MF-treated patients at a mean follow-up of 10.4 years [12]. Furthermore, the failure rate and the decrease in sports activity were significantly higher for the MF group. Although our study did not demonstrate any significant difference regarding reoperations, the trend was that reoperations occur more often in the MF group. However, comparison between the studies is difficult due to differences in study populations, surgical techniques and the use of other outcome measures. Gudas et al. included competitive or well-trained athletes, whereas the present study did not exclude non-athletes. Several studies indicate that both OAT mosaicplasty and MF provide favourable outcome in small lesions [4, 8, 25, 31, 34]. The fact that relatively small-sized lesions <2 cm² were included and that lesions >4 cm² were excluded in the Gudas study might in part explain the apparently better results at follow-up in that study compared to the present study. Another difference between these two studies is that in the Gudas study, all OAT mosaicplasty patients were treated arthroscopically, whereas in the current study an arthrotomy was performed in all mosaicplasty procedures.

In a recent retrospective, comparative study, Krych et al. [28] showed that both MF and OAT mosaicplasty-treated patients reported significant improvements in knee function and activity level at 5-year follow-up. No significant differences were detected between the two groups regarding knee function, but the mosaicplasty group maintained a superior level of activity compared to those treated with MF. The main findings of that study are in line with those of the present study, but the validity of the conclusions in the study of Krych et al. is limited by the study design, since it allows for selection bias. The unevenly distributed number of patients with previous cartilage surgery, and osteochondritis dissecans, should also be accounted for in the study by Krych et al.

There are few long-term follow-up studies following MF for treating chondral lesions of the knee. In a systematic review by Mithoefer et al. [32] only 5 studies reported a follow-up of 5 years or more, and the reports on the durability of the initial functional improvement were conflicting. The present study shows that functional improvement after MF is to be expected as long as 9.8 years after surgery.

The long-term outcome following OAT mosaicplasty in the present study supports the findings from other studies on OAT mosaicplasty, indicating acceptable long-term clinical outcome given the appropriate indication for surgery, a limitation being the defect size [12, 14, 41].

The main limitation of this study is the small number of included patients, which may lead to a false affirmation of the null hypothesis (type II error). On the other hand, the follow-up of 100 % for the main outcome (Lysholm score), and the high follow-up (88–100 %) and uniformity of comparable results between the two groups in the additional broad spectrum of outcome measures, strengthens the validity of the conclusion. Other limitations of the study are lack of a mid-term evaluation and the incompleteness of the preoperative strength measurements.

In the light of the limited information in current literature on the topic of long-term comparison between MF and OAT mosaicplasty, there is a need for further RCTs and a future cartilage repair registry in order to monitor and assess the cartilage repair procedures in use. The results from the current study might help the orthopaedic surgeon in the preoperative decision-making and in informing the patient what to expect concerning long-term outcome following these two cartilage repair techniques.

At long-term follow-up, there were no significant differences between patients treated with MF and patients treated with OAT mosaicplasty in patient-reported outcomes, muscle strength or radiological outcome. Both MF-treated as well as OAT mosaicplasty-treated patients reported improved knee function compared to the preoperative level. However, compared to a reference population, inferior patient-reported knee function was found in both treatment groups at follow-up.