Introduction

Cartilage lesions are detected in approximately 60 % of all knee arthroscopies [8, 35]. Recently, the International Cartilage Repair Society (ICRS) grading system has become the norm [4]. Low-grade lesions (ICRS grade I and II) are characterized by softening and superficial fissures without subchondral bone contact. Deep cartilage lesions (ICRS grade III) are fissures and flakes reaching the subchondral bone. A complete defect (ICRS grade IV) is defined by complete loss of the cartilage layer with widely opened subchondral bone. The different stages of the disease are approximate orientations for different treatment options. Low-grade lesions are usually not addressed surgically. Complete defects with a small diameter (<1.5–2.0 cm) can be treated with bone marrow stimulation, such as drilling, abrasion, or microfracturing (Steadman et al. 2003). Lesions with a diameter of 1.5–2.0 cm should undergo osteochondral transplantation or chondrocyte transplantation [10, 16, 21]. However, no clear consensus has been met regarding the best treatment of deep grade III cartilage lesions, though they are a significant risk factor for the progression of osteoarthritis [6]. For many years, mechanical debridement (MD), or shaving, was considered the method of choice [28].

MD aims to produce stable and nearly smooth articular cartilage surfaces, remove loose flakes, and potentially stimulate the regeneration of articular cartilage. The most important disadvantages of MD are the inability to control ablation within the deep layer and the removal of intact tissue surrounding the lesion [9, 23]. The application of thermal energy is sufficient to smooth the articular cartilage surface within deep lesions [2].

Thermal application is possible through the use of laser devices or the application of monopolar or bipolar radiofrequency (RF). Laser or monopolar RF energy produces high temperatures within the cartilage layer (≥100 °C), smelting the matrix and causing cell death among chondrocytes [9]. Earlier investigations demonstrated that the temperature should not be >52 °C [1, 12, 14]. Though several clinical studies have reported that RF treatment can be effective in the short or middle term, no clinical data have been published regarding long-term results [3, 5, 11, 18, 19, 25, 27, 30, 32].

The present randomized study compared the effectiveness of MD and 50 °C controlled bipolar RF chondroplasty at 50 °C in patients with grade III cartilage lesions and a concomitant medial meniscus tear. Clinical outcomes, the rate of surgical revision, and the progression of knee osteoarthritis were compared between the two procedures. We hypothesized that RF treatment produces better long-term clinical outcomes than MD.

Materials and methods

Patients

Patients who presented with a medial meniscus tear and idiopathic grade III cartilage defect(s) of the medial femoral condyle were considered for participation in the study. Exclusion criteria were previous injuries (ligaments, fractures, or patellar dislocation) or previous operations. Sixty patients with knee pain (<3 months) and grade III cartilage lesions detected on MRI were enrolled. After the approval of the regional ethics committee, patients who provided informed consent were randomly assigned to undergo MD (group 1) or RF (group 2) on the morning of the day of the procedure using blind, sealed envelopes constructed using a random numbers table technique. The patients remained blinded to the chondral treatment procedure. No patient demonstrated any radiological signs of osteoarthritis in the pre-operative radiographs (varus or valgus dysalignment, osteophytes >1 mm, or subchondral sclerosis). The pre-operative data for all patients are provided in Table 1.

Table 1 Baseline data for all patients
Full size table

Operative procedure

All patients underwent partial or subtotal medial meniscectomy to treat a symptomatic medial meniscopathy. In group 1, the meniscectomy was performed using a full-radius resector (Arthrex, Naples, FL, USA). The instrument was moved over the lesions three times by permanent suction for approximately 5 s. The top of the instrument had only mild contact with the surface. No significant manual pressure was applied to avoid a deep abrasion.

In group 2, we used the Paragon T2 device (AC5531-01, Arthrocare, Austin, TX, USA). This instrument contains a temperature indicator to signal heating over 50 °C, the critical temperature for articular cartilage damage. A temperature violation was not observed in any patient. This probe also had only mild contact with the surface, and the surgeon did not apply significant manual pressure (setting 40 W).

A Redon drain was positioned for 24 h in all knees. All patients took varying doses of ibuprofen depending on the level of pain and the amount of swelling experienced. All patients completed a physical exercise programme for 6 weeks.

During the observational interval, no patient underwent additional treatments, such as slow-acting drugs in osteoarthritis (SADOA) or disease-modifying osteoarthritis drugs (DMOAD).

Self-reported assessment and radiography

The patients were evaluated using the knee injury and osteoarthritis outcome score (KOOS) before surgery (baseline) and at 1-, 4-, and 10-year follow-ups [20]. The KOSS profiles were evaluated according to the KOOS guidelines (www.koos.nu).

We used the Tegner score to assess each patient’s activity level before the onset of symptoms and at the time of follow-up [29]. To assess the score before the onset of symptoms, patients had to estimate their activity level retrospectively.

The medial joint space width and varus angle (anatomical axis) were radiologically evaluation by an independent and blinded observer using standard weight-bearing radiographs in a 30° flexion position. No patient had a pathological varus. Articular cartilage lesions were classified according to the ICRS score at the start of the study [4].

All patients meeting the inclusion criteria who agreed to participate in the study provided informed consent. This study was approved by the regional ethics committee (Medical Cabinet of Thuringia, Germany. 1026/05/111). A control group in which no chondral treatment was performed for a grade III defect was not allowed.

Statistical analysis

For validation of the study, sample size calculations were performed a priori (G Power version 3.1, Heinrich Heine University, Düsseldorf, Germany) based on a significance level of 0.05 and power of 0.80 for the measurement of clinical scores. This calculation estimated that 25 patients were needed for each group.

All statistical analyses were performed using SPSS version 20.0 (Chicago, IL, USA).

The time interval between operation and revision was calculated using the Kaplan–Meier analysis, and we tested for differences using the Mantel–Cox log-rank test.

The KOOS and joint space measurements are presented as mean (95 % CI). The Tegner score was calculated as median (range). For a comparison of dichotomous parameters between the groups, Chi-square test was used. Continuous parameters were compared between the groups using the Mann–Whitney U test. The Wilcoxon signed-rank test was used for comparisons between the pre-operative data and results obtained during follow-up. A p value ≤0.050 was considered significant.

Results

Follow-up, survival to revision, and loss to follow-up

Ten patients (one from group 1 and nine from group 2) were lost to follow-up (16.6 %, Table 2). The clinical course of these patients is unknown. The loss of these patients decreased the study power to 0.69. In addition, three patients (two from group 1 and one from group 2) died during the follow-up period. Therefore, it was only possible to evaluate the 10-year post-operative time course in 47 (78.3 %) patients. Eighteen patients (60.0 %) in group 1 underwent revision surgery: 11 were unicondylar or total replacements, 5 were osteotomies, and 2 were revision arthroscopies. The rate of revision in group 2 was 23.3 % (n = 7). The difference between the groups was not significant. The baseline data for these patients are provided in Table 3.

Table 2 Baseline data for patients lost to follow-up
Full size table
Table 3 Baseline data for patients undergoing revision surgery
Full size table

The time to a required revision in group 1 was 62.5 months (95 % CI 45.9–79.2). In group 2, the mean survival was 94.1 months (95 % CI 77.1–111.3). This difference was significant (p = 0.022, log Mantel–Cox). The Kaplan–Meier curve is shown in Fig. 1.

Fig. 1
figure 1

Survival to revision (Kaplan–Meier). The mean survival in group 2 was 94.1 (95 % CI 77.1–111.3) months. The time to a required revision in group 1 was 62.5 (95 % CI 45.9–79.2) months, p < 0.001

Full size image

Self-reported assessment

Patients who did not require revision (baseline data in Table 4) were assessed pre-operatively and at 1-, 4-, and 10-year follow-ups. The subjective outcome assessed by KOOS was better in group 2 (n = 13) than in group 1 (n = 9; Table 5). The KOOS profiles are presented in Figs. 2 and 3.

Table 4 Baseline data for patients undergoing 10-year follow-up
Full size table
Table 5 Knee injury and osteoarthritis outcome score (KOOS)
Full size table
Fig. 2
figure 2

KOOS profiles of group 1

Full size image
Fig. 3
figure 3

KOOS profiles of group 2

Full size image

Pre-operatively, patients in both groups had a moderate activity level without any significant difference between the two groups. The Tegner activity scores are presented in Table 6. The onset of disease caused a significant reduction in activity level, and patients from both groups benefitted from the operation. However, RF-treated patients achieved a higher activity level 1 and 4 years after surgery compared to MD-treated patients. The patients’ activity levels decreased in both groups over time, and activity levels were similar at the 10-year follow-up.

Table 6 Tegner activity score
Full size table

Factors (e.g. age, gender, history, diameter of the defect, and extent of meniscus surgery) that predict the outcome regarding self-reported assessment, progression of radiological osteoarthritis (JSN), or need for revision were not found in the long-term follow-up.

Radiological investigation

At the time of surgery, no patient had any radiological signs of osteoarthritis (i.e. osteophytes or subchondral sclerosis). The medial joint width was similar in both groups (group 1: 5.4 mm, 95 % CI 4.3–6.5; group 2: 5.6 mm, 95 % CI 4.9–6.3). During the follow-up period, the joint space width narrowed continuously in both groups (Table 7). The joint space narrowing tended to increase more rapidly in group 1, but the difference from group 2 was not significant.

Table 7 Joint space narrowing
Full size table

Discussion

The most important finding in the present randomized study was that RF treatment produced better clinical outcomes and lower revision rates than a simple MD over short- and middle-term follow-up. The rate of needed revisions was significantly lower after RF treatment. After the long-term follow-up, the treatments did not differ meaningfully.

Different methods are available for the treatment of grade III lesions. First, it is possible to treat conservatively by SADOA or DMOAD. Patients who suffer from a concomitant degenerative meniscus tear mostly undergo arthroscopy. In this situation, the surgeon must decide on a treatment for the cartilage lesion. Partial resection of the loose meniscus fragments often produces good short-term results because of the reduction in meniscus-related symptoms, but deep cartilage lesions (grade III) or complete defects (grade IV) must be considered as the initial phase in the development of osteoarthritis. Without treatment of the cartilage lesions, more than 50 % of these patients develop osteoarthritis within 10 years [24]. For the treatment of complete defects (grade IIIb or grade VI), the guidelines give relatively clear directions [16]. Small full-thickness cartilage defects (diameter < 1 cm) should be treated by bone marrow techniques (e.g. microfracture), whereas defects with a diameter > 1 cm should undergo autologous chondrocyte transplantation (ACT), osteochondral transplantation (OATS), or cell-free autologous matrix-induced chondrogenesis (AMIC). However, the best treatment for deep grade III cartilage lesions is controversial. Simple MD aims to smooth the joint surface and remove loose flakes or osteophytes. This procedure became very popular during the era of arthroscopy and is one of the most commonly performed arthroscopic operations [17]. Experimental studies regarding the effect of shavers have demonstrated that shaving leads to uncontrollable effects within the chondral matrix and structure [7, 31]. Spahn et al. [23] investigated the cartilage layer after shaving and monopolar RF treatment using scanning electron microscopy. Mechanical shaving produces a rough surface with grooves and open-laying collagen fibres. This causes nearly uncontrolled destruction within the ultrastructure of the chondral matrix. During a 15-month observational study in rabbits, Mitchell and Shepard observed a significant progression of the knee osteoarthritis after shaving. The grade of osteoarthritis varied in accordance with the depth of the shaving and the types of instruments used [15]. In previous randomized clinical investigations, shaving did not produce any positive effect compared to control groups. No scientific evidence supports the use of mechanical shavers to treat cartilage lesions [13]. In a study of patients who did undergo control arthroscopy after osteotomy, we observed complete defects after MD in approximately 30 % and no effects in the remaining patients [26]. In contrast, among patients who underwent RF treatment, 81.6 % demonstrated regeneration (grade III to grade I). In the current study, patients who underwent this treatment had significantly worse short- and middle-term outcomes, which may be the reason for the significantly higher revision rate in this group. Thus, MD of grade III lesions is not indicated.

RF treatment (monopolar or bipolar) produces thermal energy, which has a smelting effect within the cartilage layers. Thus, this method causes smoothing of the degenerated articular cartilage surface. After the procedure, the cartilage appears to regenerate. However, uncontrolled heating of the cartilage is associated with irreversible cell necrosis and destruction of the tissue ultrastructure. RF energy can cause the temperature within the cartilage to rise by 100° or more [34]. This effect is not controlled during surgery. After this procedure, the restoration of normal tissue is improbable [12, 22, 31]. Furthermore, thermoablation has another severe side effect; it alters the subchondral bone. Osteonecrosis occurs in a number of cases, possibly due to ischaemia of the blood vessels within the bone [5]. Clinical trials contain reports of the contradictory effects of RF treatment in regard to clinical outcome, cartilage restoration, and the progression of knee osteoarthritis. Osti et al. [18] observed a better 5-year outcome after microfracturing compared to RF treatment in a case series of ACL-deficient patients. Kang et al. [11] also reported worse results after monopolar RF treatment of cartilage lesions compared to simple debridement.

Regarding monopolar RF treatment, Barber and Iwasko [2] found that its use as an adjuvant to mechanical chondroplasty with a shaver for the treatment of grade III chondral lesions does not affect MRI findings or pain and functional outcomes compared with mechanical chondroplasty using a shaver only. Voloshin et al. reported a significant positive effect in 88 % of patients after treatment with bipolar RF energy. They performed five control arthroscopies, and more than 50 % exhibited partial or complete filling of the defect. The authors concluded that bipolar radiofrequency chondroplasty is an effective method for treating partial-thickness cartilage lesions [33].

Our study is the first long-term study in this field of investigation. Patients undergoing RF treatment benefitted more from surgery than those undergoing MD, as the clinical knee score was higher in RF-treated patients at 1 year [25].

Patients in both groups had to accept a decrease in their physical activity level (Tegner score). The better outcome regarding activity level during short- and middle-term follow-up after RF may be caused by the better clinical outcome (KOOS). After 10 years, the activity level in both groups was less than the pre-operative level. There are two possible reasons for this result. First, the progression of osteoarthritis can reduce the activity level. Second, as patients age there is a “natural” reduction in physical activity. However, the results in both groups worsened over time, and neither treatment option completely avoided the progression of osteoarthritis over the long term.

The pace of osteoarthritis progression in patients who underwent MD did not differ from the natural progression of the disease. Taking grave study-related limitations into account, we conclude that careful RF treatment is sufficient to decelerate osteoarthritis progression in a small number of patients. Among patients who underwent RF, better outcomes were observed at the short- and mid-term follow-ups, as well as at the 10-year assessment. Furthermore, the incidence of osteoarthritis, as indicated by joint space narrowing, was lower in this group. Thus, patients who underwent bipolar RF treatment benefitted during the study period.

Although this study was randomized, the relatively small number of patients in both treatment groups resulted in only moderate statistical power. Significantly more RF-treated patients than MD-treated patients were lost to the 10-year follow-up. We speculate that these patients declined a control because they were unsatisfied with the results. These patients possibly suffered from a post-arthroscopy osteonecrosis. This complication occurs more frequently after RF treatment than other procedures [30]. However, these patients were lost after the fourth post-operative year. Therefore, the RF-related osteonecrosis is more improbable, attenuating the beneficial effect of RF treatment considerably.

Another limitation is the absence of a real control group with an untreated cartilage lesion. Though we decided against this for ethical reasons, the lack of such a control group makes it difficult to appraise the effect of RF in both groups. Furthermore, all patients had a symptomatic medial meniscus tear, which often produces grave clinical symptoms. The short-term and 1-year clinical results were likely influenced by the extent of meniscus resection.

Taken together, our results indicate that the positive effect of MD is doubtful. RF treatment at 50 °C produced better clinical outcomes in the short- and middle-term follow-up, but this method is also not sufficient to stop the progression of osteoarthritis.

Conclusion

Compared to conventional MD, bipolar temperature-controlled RF treatment at 50 °C appears to be a superior method regarding short- and medium-term clinical outcomes and the progression of knee osteoarthritis. Clear predictors of the indications of different cartilage treatments are needed. Moreover, because the power of this study is limited, more randomized controlled trials in this scientific field are needed in the future.