BackgroundThe surfaces of the bones in the knee are covered with articular cartilage, a rubber-like substance that is very smooth, allowing frictionless movement in the joint and acting as a shock absorber. The cells that form the cartilage are called chondrocytes. Natural cartilage is called hyaline cartilage. Articular cartilage has very little capacity for self-repair, so damage may be permanent. Various methods have been used to try to repair cartilage. Autologous chondrocyte implantation (ACI) involves laboratory culture of cartilage-producing cells from the knee and then implanting them into the chondral defect.ObjectiveTo assess the clinical effectiveness and cost-effectiveness of ACI in chondral defects in the knee, compared with microfracture (MF).Data sourcesA broad search was done in MEDLINE, EMBASE, The Cochrane Library, NHS Economic Evaluation Database and Web of Science, for studies published since the last Health Technology Assessment review.Review methodsSystematic review of recent reviews, trials, long-term observational studies and economic evaluations of the use of ACI and MF for repairing symptomatic articular cartilage defects of the knee. A new economic model was constructed. Submissions from two manufacturers and the ACTIVE (Autologous Chondrocyte Transplantation/Implantation Versus Existing Treatment) trial group were reviewed. Survival analysis was based on long-term observational studies.ResultsFour randomised controlled trials (RCTs) published since the last appraisal provided evidence on the efficacy of ACI. The SUMMIT (Superiority of Matrix-induced autologous chondrocyte implant versus Microfracture for Treatment of symptomatic articular cartilage defects) trial compared matrix-applied chondrocyte implantation (MACI®) against MF. The TIG/ACT/01/2000 (TIG/ACT) trial compared ACI with characterised chondrocytes against MF. The ACTIVE trial compared several forms of ACI against standard treatments, mainly MF. In the SUMMIT trial, improvements in knee injury and osteoarthritis outcome scores (KOOSs), and the proportion of responders, were greater in the MACI group than in the MF group. In the TIG/ACT trial there was improvement in the KOOS at 60 months, but no difference between ACI and MF overall. Patients with onset of symptoms < 3 years’ duration did better with ACI. Results from ACTIVE have not yet been published. Survival analysis suggests that long-term results are better with ACI than with MF. Economic modelling suggested that ACI was cost-effective compared with MF across a range of scenarios.LimitationsThe main limitation is the lack of RCT data beyond 5 years of follow-up. A second is that the techniques of ACI are evolving, so long-term data come from trials using forms of ACI that are now superseded. In the modelling, we therefore assumed that durability of cartilage repair as seen in studies of older forms of ACI could be applied in modelling of newer forms. A third is that the high list prices of chondrocytes are reduced by confidential discounting. The main research needs are for longer-term follow-up and for trials of the next generation of ACI.ConclusionsThe evidence base for ACI has improved since the last appraisal by the National Institute for Health and Care Excellence. In most analyses, the incremental cost-effectiveness ratios for ACI compared with MF appear to be within a range usually considered acceptable. Research is needed into long-term results of new forms of ACI.Study registrationThis study is registered as PROSPERO CRD42014013083.FundingThe National Institute for Health Research Health Technology Assessment programme.