Inclusion criteria were patients 30–65 years with focal chondral defects of the femoral cartilage, International Cartilage Research Society (ICRS) grade 3 or 4 on medial or lateral femoral condyles. The cartilage lesion area was ≤ 3.2 cm2 (diameter ≤ 2 cm). The patients had previously failed conservative or surgical interventions such as abrasions, drilling or microfracture and VAS pain > 40 with activity for more than 6 months. The patients should be capable of completing self-administered questionnaires and willing to comply with the follow-up requirements of the study. Exclusion criteria were BMI > 35 kg/m2, not addressed instability or other concomitant knee injuries such as meniscus injuries, apart from small flap lesions with intact rim and intact meniscal anterior and posterior horn insertions. Further exclusion criteria were established osteoarthritis, malalignment, metabolic disorders which may impair bone formation, smokers, metal allergies, inflammatory joint diseases, administration of corticosteroids, antineoplastics, immune stimulating or immunosuppressive agents.
For this study a customized Cr–Co femoral condyle implant was manufactured by computer-aided design/computer-aided manufacturing (CAD/CAM) technique using MRI data. In addition, from the MRI data specific guide instruments were manufactured as part of the surgical kit to obtain accurate implant positioning in order to avoid damage to the opposing cartilage surface . The Cr–Co implant has a 3- to 4-mm-thick circular shape where the articulating surface was contoured to precisely reconstruct each patient’s individual cartilage lesion and condylar contour. The implant has a 15-mm-long and 4-mm-thick peg that was inserted in an undersized drill hole to provide primary fixation. For secondary or long-term fixation a double coating of hydroxyapatite-on-top-of titanium was applied on the surfaces facing bone and cartilage (Fig. 1). This coating has been shown to give firm and consistent osseochondrointegration in an animal model .
The patient-specific designs of the Episealer implants and surgical drill guides are based upon MRI imaging. The patient undergoes an MRI scan that is uploaded to a MRI processing system (μiFidelity®) through a web-based online management system. By processing MRI data, a 3-dimensional model of the damaged joint is recreated. The lesion is identified and the implant (Episealer®) and surgical drill guides (Epiguide®) are designed to enable the replacement of the damaged area with a fitting implant. The implant is available in four different diameters, 12, 15, 17 and 20 mm, depending on the condyle and lesion size.
Operative technique and description of devices
The patients were administered cloxacillin 2 g 15–45 min before surgery and 2 h postoperatively. Surgery is performed under general or spinal anesthesia. The procedure started with a diagnostic arthroscopy to evaluate the joint and confirm absence of significant concomitant injuries and to validate the size of the injury. The femoral condyle is then exposed through a medial or lateral mini-arthrotomy. After exposure of the femoral condyle, an assessment of the cartilage injury was made ensuring that the implant was sufficiently large to cover the lesion. The drill guide was matched to the unique position on the femoral condyle surface with the lesion centerd in the drill guide. At least two surgical pins should be drilled at the rim of the femoral condyle to firmly fasten the drill guide on the condyle. The rims of the cartilage defect were adjusted with a cartilage cutter and the defect was drilled and milled in order to prepare a hole in the osteochondral tissue for implant insertion. Drill depth was adjusted to counter-sunk the implant 0.5 mm as indicated by a dummy (Epidummy). The depth was fine-tuned using an adjustment ring in the tool kit allowing incremental adjustment of 0.2 mm and the final implant position was controlled by a testing device (Epidummy). By inserting the implant, the defect was sealed by the metallic resurfacing implant hence reconstructing the original contour of the femoral condyle. The drillguide and pins were removed and the wound closed in layers. Postoperative pain alleviation by intraarticular ropivacaine was given. Anti-thrombotic medication was not routinely administered. Four different experienced knee surgeons at three different centers performed the surgery and practiced the technique on saw bones prior to the surgery.
Week 1–2: crutches and non-weight-bearing, full passive and active exercises without resistance. Activation of quadriceps muscle in extension. Aims: full passive range of motion.
Week 3–6: crutches and partial weight-bearing. Cycling with light resistance. Straight leg lift. Initiation of core-stability training. Leg curls with light resistance. Aims: full active and passive range of motion. Full weight-bearing at week 6.
Week 7–12: full weight bearing, no crutches. Cycling with increasing resistance. Balance training. Squats and lunges with increasing resistance. Aims: normal walking. Good knee control. Managing ADL.
Clinical and radiological evaluation
Primary endpoints assessments were frequencies of un-anticipated side effects such as surgical tools usability and functionality, ongoing osteoarthritis, implant migration, mechanical implant loosening, implant fracture, inflammation, no pain alleviation or allergic reaction. Secondary endpoints assessments were performance compared to base-line by patient reported outcome measurements (PROMS) preoperatively, at 3 and 6 months, 1 and 2 years; Knee injuries and Osteoarthritis Outcome Score (KOOS), EuroQoL (EQ-5D), Tegner Activity Scale (Tegner Score), In addition, Visual Analogue Scale (VAS pain) and knee range of motion (ROM) measurements were recorded.
Radiostereometric analysis (RSA)
RSA is a high-precision method of assessing three-dimensional (3D) micro-movement from calibrated stereoradiographs and is a standard technique for evaluating new implants since early migration can predict loosening [9, 16, 24]. The RSA method in our study followed published guidelines for RSA  and the method developed for this particular implant . Using this method the accuracy is between 0.08 and 0.19 mm and the precision 0.12–0.33 mm. During surgery 5–6 tantalum markers (1.0 mm) were placed in the bone surrounding the implant to serve as the reference segment for the RSA analysis. The tip of the implant, in the shape of a 3-mm hemi-sphere was used as the measured point of the prosthesis. Two days postoperatively, at 6 months, 1 and 2 years after surgery, the operated knee was then placed in a biplanar calibration cage (Cage 10; RSA Biomedical AB, Umeå, Sweden). Digital radiographs (Bucky Diagnostic; Philips, Eindhoven, the Netherlands) were then taken using one fixed and one mobile X-ray source. The exposure was set to 125 kV and 2.5 mAs. The radiographs were saved in a standard dicom file format (resolution 254 dpi) and uploaded to a workstation. UmRSA 6.0 computer software (RSA Biomedical) was used for all measurements and migration analyses. The markers in the distal femur form one segment and the micromotion of the tip of the implant was then evaluated. The 3D translations of the tip in relation to the femoral bone segment were calculated at each follow-up visit and compared with the immediate postoperative measurements. We also measured the maximum total point movement (MTPM), which is the 3D translation vector of the tip. At 1 year, we performed two examinations 15 min apart on all patients with complete repositioning of the X-ray tubes and the calibration cage. We calculated the precision as the 95% confidence interval (SD 1.96) of the difference between these examinations. For translation along the x-(transverse), y-(vertical) and z-[anteroposterior (AP)] axes, this was 0.20, 0.32 and 0.30 mm, respectively, and for the MTPM it was 0.33 mm, corresponding well with previous in vitro precision determinations . For individual patients, any migrations above these threshold values means that a detectable migration has occurred of the implant in relation to the bone.
The study was approved by the local ethics committee Karolinska Institutet (2012/109-3171). Informed consent was obtained from all patients and the study was conducted in accordance with good clinical practice and the declaration of Helsinki.
All data were analyzed using IBM SPSS Statistics for Mac, version 23 (IBM Corp, Armonk, NY, USA). Data for patient demographics and patient-related outcome measures are expressed as median and range. Group differences were analyzed with Mann–Whitney rank sum test, two-tailed. A p value of less than 0.05 was considered statistically significant.