Other issues
Asymptomatic lesions
The BASK UK39 Consensus recommends that patients should have conservative treatment with physical therapy before being considered for cartilage repair. Many will become asymptomatic and will no longer qualify for ACI according to the NICE scope. However, their cartilage defect will not recover spontaneously, and they are likely to develop OA in later years. Should they be considered for ACI?
The Dutch Orthopaedic Association recommends treatment of asymptomatic ICRS grade 5 lesions.22
Osteoarthritis
The NICE scope excludes people with ‘advanced osteoarthritis’. OA can be defined as generalised degenerative change affecting both sides of an articulation. ACI is used for isolated cartilage defects. There can be isolated defects on both surfaces (‘kissing lesions’), which could be considered for ACI if the rest of the joint is in good order. Patients with only early OA (less than grade 2 which has definite osteophytes and possible joint space narrowing) could have been included in some trials. However, no details for such a subgroup are given in the results. In the TIG/ACT trial,43 patients with advanced OA (as defined by radiographic atlas OA grade 2–3) were excluded.
A systematic review of cartilage repair in early OA by de Windt et al.166 found evidence of benefit in those having various forms of ACI, ranging from ACI-P to MACI. Early OA was defined in different ways in the nine case series, and de Windt et al.166 described the studies as being of ‘generally low methodological quality’. Nevertheless, they reported that outcomes to 9 years were good, and suggested that ACI in early OA might be used to postpone TKR, but recommended a RCT.
There may, therefore, be a place for ACI in early OA, even if only to postpone TKR until patients are older, and some of the ICERs reported earlier are within the acceptable range. However, the evidence base is much weaker than for purely chondral lesions.
However, if ACI were to be restricted based on radiological signs of OA, there are some problems to be considered. One of the difficulties in comparing the results of studies involving patients with OA is the definition of the disease and the assessment of its severity. The European League Against Rheumatism (EULAR) definition of OA emphasises the importance of pain and functional loss alongside physical changes in the joint, but this definition is hard to objectively apply in research terms, and symptoms are significantly influenced by environmental and psychosocial factors.167–170
There is a variable relationship between symptoms and structural changes in OA and it is recognised that plain radiographs, MRI and arthroscopic findings do not universally correlate with pain or physical function.171–173
The most common method for assessing structural changes in knee OA is plain radiography, graded using the Kellgren–Lawrence classification.174 Care has to be taken in interpreting plain radiographic findings, as Kellgren–Lawrence grades have moderate, but not strong, correlations with other measures of structural change, such as MRI measures of OA or operative findings.175–180
The Kellgren–Lawrence classification is a widely accepted tool in OA research and good reliability has been quoted in series in which the assessors were experienced in its use.173,178 However, it is based on a subjective assessment of structural changes and different authors often apply different criteria to define the boundaries between the grades, making comparisons across studies difficult.181
The boundary between Kellgren–Lawrence grades 2 and 3 is often difficult to define, as the interpretation of ‘possible’ and ‘definite’ joint space narrowing can be very subjective.182 However, this is not so important when considering suitability for ACI, as the Nawaz study80 showed that there was little difference in outcomes. The distinction between lower Kellgren–Lawrence grades is also difficult and is dependent on the interpretation of small osteophytes, which can variably give a score of 0, 1 or 2, depending on the exact definitions used and the radiological technique.181
The diagnosis of OA is often made based on the combination of symptoms and a Kellgren–Lawrence grade of ‘2 or more’, despite evidence that Kellgren–Lawrence grade of 1 (‘doubtful osteophytes’) has a high chance of progressing to ‘2 or more’ with time.183,184
The studies in this review have varied in terms of their reporting of the radiological assessment, and definitions were not always clearly defined in the reports, which may explain some of the variance in findings between studies. For example, relatively little detail is given in the Minas paper185 on the radiological assessment and the Kellgren–Lawrence paper is not referenced, whereas the radiological grading is reported in detail by Nawaz et al.80 A relatively high proportion of cases with Kellgren–Lawrence grade 2 or above were reported by Knutsen,67 which may explain the poor results for ACI in this series in comparison to others.
Defining OA is problematic. A big cartilage lesion with pain and some joint space loss could variably be defined as no, mild or moderate OA.
Age threshold for knee replacement
In our modelling we have assumed that TKR would not be performed for people with OA until age 55 years or later. We used that age restriction because knee replacements do not last for ever, and replacing a replacement is more difficult, more expensive and less successful than the first replacement, and may not last as long.
With increasing longevity, it may no longer be the case that a knee replacement in someone over 60 years is likely to last them all of their days. Perhaps especially in women who live longer. However, a TKR in a younger person with OA is very likely to need replacement. (This may not apply to people having knee replacement because of inflammatory arthritis because their activity, and hence the stresses put upon the prosthesis, will often be limited by problems with other joints.)
In the National Joint Registry 2015 report,157 figure 3.16 shows that the probability of a first revision after TKR is higher in people who have replacements at younger ages. Those who have TKR under the age of 55 years have a 12% probability of it being replaced by 11 years, which is more than double the risk after first TKR at older ages.
It is therefore a major decision to carry out TKR in people with OA under the age of 60 years, and very few are done. It should be noted that TKR is rarely an absolute necessity. The aim is to reduce pain, and that can be done in other ways, such as with analgesics or reducing activity.
It should also be borne in mind that TKR does not fully restore knee function. The TKR does not move like a normal knee, and younger, active patients may find function on stairs and slopes disappointing.
Autologous chondrocyte implantation can restore normal function in younger patients. In patients who are older but too young for TKR, but who do not have generalised wear and tear, ACI may help bridge the gap to TKR, even if the results are not as good as in younger patients with only an isolated chondral defect.
Body mass index
Jaiswal et al.186 from Stanmore reported a lack of benefit from ACI or MACI in patients with BMI score of over 30 kg/m2, though this was based on small numbers in the high BMI group. Their data came from the trial of MACI versus ACI. In 53 patients with BMI scores of under 25 kg/m2, 82% of patients had a good or excellent result. In the overweight group (BMI 25–30 kg/m2) 49% (22 of 45) had a good or excellent result, whereas only one of 18 patients with BMI over 30 kg/m2 had a good result.
Mithöfer et al.82 also reported worse outcomes in those with BMI score of over 30 kg/m2. Behery et al.83 reported no correlation but had data on only eight patients.
Data on the effect of high BMI on outcomes of cartilage repair is sparse. Jaiswal et al.186 reported that their literature review found few previous studies. In most studies, mean BMI scores were well below 30 kg/m2, perhaps because cartilage injuries occur largely in people active in sports. Jaiswal used the term ‘obese’ but some sportsmen with high BMI scores may be lean but very muscular.
Similar findings have been reported for MF by Asik et al.,106 with better results in those with BMI score of less than 25 kg/m2.
Research needs
Recommendations for research made in the systematic reviews.
Some of the recommendations made in the reviews are now out of date and are not included here. Other recommendations include:
High-quality clinical trials are needed, fulfilling the following criteria:
Multicentre, adequate sample size with long-term follow-up (preferably 5–10 years).
Patients in trials should be stratified based on
BMI, defect location, post-debridement defect size and previous cartilage repair.
Transparent patient enrolment with clearly stated inclusion and exclusion criteria.
Proper independently performed randomisation techniques.
No concurrent surgical interventions (anterior cruciate ligament reconstruction, realignment osteotomy, meniscal surgery, etc.), consistent surgical technique.
Use of validated, responsive and reliable patient-orientated outcome measures; clear reporting of data with a statement of both clinical relevance and significance; and use of independent assessors.
Further information is needed on the relationship between clinical, histological and radiological outcomes, and the most appropriate measure of functional outcomes that relates to a generic measure of health-related
QoL.
Cohort studies of long-term effects (≥ 10 years) are needed.
Research is needed to explain lack of return to sports by some patients.
Prospective long-term studies are needed to determine if articular cartilage repair in athletes can influence the high incidence of
OA associated with high-impact sports.
More studies should be done on the maturation process of finally formed repair tissue and on appropriate rehabilitation programmes for the different techniques.
Fourth-generation autologous chondrocyte implantation
There are several lines of investigation.
Mesenchymal cells
It has been suggested that mesenchymal stem cells from bone marrow can be used as an alternative to ACI and that their reproduction is less affected by age. (For reviews see Nakamura et al.187 and Perera et al.188)
A review of scaffold-based repair by Filardo et al.189 mentions another option, using mesenchymal stem cells and a degradable scaffold, covered with fibrin.190
The ASCOT trial191 will compare repairs with chondrocytes and bone marrow mesenchymal stem cells, and with the combination of both.
INSTRUCT
This appears to be a one-stage procedure mixing chondrocytes and bone marrow cells, without cell culture. Cells from a biopsy of cartilage are mixed with bone marrow cells then seeded into a porous scaffold, which is then implanted into the defect. Evidence comes from a poster by Hendriks et al.192 So far, only 37 patients had reached 12-month follow-up, of whom 72% had hyaline cartilage on biopsy.
Cartilage implantation
The development here is that instead of implanting cultured chondrocytes into the defect, the autologous chondrocytes are used to grow new cartilage in the laboratory, which is then implanted.193
Gel-type autologous chondrocyte implantation
Gel-type ACI appears to be a new variant without using membrane or periosteum, but using cells held in place with fibrin. Choi et al.194 report a case series with 98 patients. There do not appear to have been RCTs against standard ACI.
Single-stage procedures
Cole et al.89 report a RCT with 29 patients, comparing MF (nine patients) with a cartilage autograft implantation system (CAIS) in which chondrocytes are not sent for culture. Instead, hyaline cartilage is harvested in similar amount as for traditional ACI, but then minced and attached to a biodegradable scaffold with fibrin glue, in a single operation. Results at 24 months showed some advantages for the CAIS group, with IKDC score 83 for CAIS and 60 for MF, and KOOS also better.
Other cells
Mizuno et al.195 report that ear cartilage cells can be used, at least in dogs.
New forms of microfracture
Filardo et al.189 report five case series of autologous matrix-induced chondrogenesis, which combines MF with a collagen matrix to stabilise the blood clot. Long-term results are not yet available.
Siclari et al.196 used a combination of MF and a cell-free hyaluronan cap that had been immersed in autologous plasma in 52 patients. At 2 years, KOOS results showed good improvement. Biopsies were taken from four patients and showed hyaline or hyaline-like repair tissue.
Metal or plastic patches for knees
These were excluded by NICE as comparators, but sound sufficiently promising to be used in trials. They may not be suitable for younger patients but might be an option for the 40–60 years subgroup, perhaps in order to postpone knee replacement.
The HemiCAP® (Arthrosurface, Franklin, MA, USA)197 is used for resurfacing localised damage in femoral condyles, and is described by the manufacturer as a ‘contoured articular resurfacing implant’, and as ‘bridging the gap between biological therapies and TKR’. The evidence base seems to consist of a few case series with no RCTs:
Patello Femoral HemiCAP
®198
The BioPoly™ RS Knee System201 (Schwartz Biomedical, Fort Wayne, IN, USA) is CE marked for sale in the European Union. It is a hyaluronic and polythene implant for repairing the joint surface.
The Episealer (Episurf AB, Stockholm, Sweden)202 comes in two forms, for femoral condyle and trochlea, and is described as a small metallic button with implants tailored for each patient.
These products are said to allow rapid return to activities, unlike the long rehabilitation that is required after ACI. A recent study203 reported that after ACI or MF followed by a long period of rehabilitation, 33% and 26% of sportspeople did not regain full quadriceps power after MF and ACI, respectively. Another study204 reported good results after ACI-P, with 26 of 33 patients having good or excellent results at 10 years, but also noted that patients did not return to full pre-injury activity levels. This may be partly due to the long lay-off during the rehabilitation process. However, those who return to previous activity too early have poorer outcomes than those who wait at least 12 months.205
Conclusion
The evidence base for ACI has improved since the last appraisal by NICE. In most analyses, the ICERs for ACI compared with MF appear to be within a range usually considered acceptable.
The evidence base for ACI is much better than for MF.
