Which procedure has better long-term results?

At ten years, autologous chondrocyte implantation (ACI) outperforms mosaicplasty by a substantial margin — and the evidence comes from the only prospective randomised controlled trial with that length of follow-up. In the Bentley RCT (100 patients, published in the Journal of Bone and Joint Surgery in 2012), treatment had failed in 17% of ACI patients (10 of 58) compared with 55% of those who received mosaicplasty (23 of 42) — a statistically significant, roughly three-fold difference (p < 0.001).

That verdict carries an important caveat. The trial enrolled patients with large articular cartilage defects (mean ~440 mm² in the ACI group) and a mean of 1.5 prior surgeries — a cohort where mosaicplasty's technical size constraints are most exposed. For smaller, isolated defects, particularly below approximately 2 cm², the gap narrows and mosaicplasty remains a legitimate single-stage option for selected younger patients.

The sections that follow examine how failures accumulate over time in each group, why defect size mediates the outcome so strongly, and how current guidelines translate this evidence into clinical practice.

What the 10-year Bentley trial found

The patients enrolled in the Bentley trial were young — mean age 31.3 years — and represented a genuinely demanding clinical scenario: large defects averaging roughly 440 mm² in the ACI group and 400 mm² in the mosaicplasty group, with an average of 1.5 previous knee operations already behind them. This matters because it anchors the trial in real-world complexity rather than optimised surgical conditions.

Beyond the headline failure rates noted above, the trial captured functional quality among those whose repair survived. Measured on the modified Cincinnati knee score and the Stanmore-Bentley functional rating, ACI patients with a surviving graft scored significantly better than their mosaicplasty counterparts at ten years (p = 0.02). Durability and function moved together in the ACI group; in the mosaicplasty group, the minority whose graft did survive showed meaningfully poorer function.

The one-year interim data from the same cohort foreshadowed what followed. At that stage, 87% of ACI patients had a good or excellent outcome against 70% for mosaicplasty; more tellingly, 26% of the mosaicplasty arm had already failed clinically and arthroscopically by or around one year, with deterioration peaking at approximately two years post-surgery. The pattern was early and concentrated in the mosaicplasty group, whereas ACI failures accumulated at a low, steady rate across the full decade.

The trial authors did not qualify their conclusion: continued use of mosaicplasty for articular cartilage defects, they wrote, is "of dubious value." That judgement rests on a large, statistically significant, controlled finding — not conjecture.

How the two procedures fail — and why the timing matters

The timing of failure in each group reflects fundamentally different biological processes.

Mosaicplasty relies on cylindrical osteochondral plugs press-fitted into drilled recipient sockets. Integration either succeeds in the early remodelling window — broadly within the first two years — or it does not. Where plugs fail to bond fully to the surrounding cartilage matrix, shear stress during loading accelerates breakdown. The geometry compounds the problem: gaps between cylindrical plugs create fibrocartilage-filled dead zones lacking the load-distributing properties of hyaline cartilage. Attempting to cover larger lesions quickly hits a donor-site ceiling; harvesting from non-weight-bearing areas becomes increasingly morbid beyond roughly 2.5 cm², so large defects risk incomplete coverage from the outset.

ACI's failure pattern is biologically distinct. The implanted cells produce living repair tissue that matures gradually over years; failures therefore accumulate at a low, steady rate rather than clustering early. That maturation process does carry a specific complication profile — a 2013 systematic review by Chalmers and colleagues (20 studies, 1,375 patients) found ACI reoperations were disproportionately unplanned, driven by graft delamination and periosteal hypertrophy. These complications are generally manageable, but they require frank discussion before surgery.

One anatomical limitation stands apart. Mosaicplasty is generally contraindicated on the patella, where curved geometry and high contact pressures make cylindrical plug placement technically unreliable. ACI carries no such restriction; approximately one quarter of the Bentley trial cohort had patellar lesions, and ACI is the preferred approach for that site.

When mosaicplasty is still a reasonable choice

None of this evidence argues that mosaicplasty has no place — it argues that its place is specific.

For small, isolated, full-thickness defects in the range of 1–2 cm², and up to roughly 2.5 cm² before donor-site morbidity becomes a meaningful concern, the procedure operates within its biological limits. At that scale, the cylindrical plugs can achieve adequate coverage, integration conditions are more favourable, and the size constraints that undermined mosaicplasty in the Bentley cohort — where mean defects ran to approximately 400–440 mm² — are far less consequential. A 2016 meta-analysis by Li and colleagues, reviewing six comparison pairs of ACI versus osteochondral autograft transfer, found ACI significantly superior in only one of those six — a finding that suggests the dramatic gap seen in the Bentley trial partly reflects a patient group where mosaicplasty was already at the edge of its capability.

The single-stage pathway is also a genuine clinical advantage for certain patients. Published recovery data indicate weight-bearing is typically possible within approximately one week, with full rehabilitation around four to six months — considerably shorter than the two-stage ACI pathway. Where two-stage surgery is impractical due to health-system access, patient circumstances, or resource constraints, a single-stage procedure that performs adequately within its size envelope is a defensible choice.

The procedure is not appropriate for patellar lesions, defects exceeding roughly 4 cm², or cases where prior surgery has reduced available donor tissue — anatomical contexts in which ACI or other two-stage techniques should lead the discussion.

What NHS guidelines recommend and who qualifies

NICE guidance — specifically Technology Appraisal 477, published in 2017 — translates the clinical trial evidence into a practical eligibility framework for NHS patients.

ACI is recommended for symptomatic articular cartilage defects greater than 2 cm² in patients who have not previously undergone cartilage repair surgery, have minimal osteoarthritis damage, and are treated at a tertiary referral centre. Each criterion matters individually. The 2 cm² threshold corresponds closely to the upper boundary at which mosaicplasty can be performed without meaningful donor-site morbidity — beyond that size, ACI's capacity to achieve full biological coverage becomes most clinically relevant. The requirement for no prior cartilage repair surgery reflects a body of evidence showing that ACI success rates fall after earlier procedures such as microfracture; many patients presenting for specialist assessment have already had a first-line intervention, which is why the surgical history is one of the first things a consultant will establish.

The cost-effectiveness condition — that ACI deliver value at under £20,000 per QALY gained, with an approved maximum cell cost of £16,000 — explains why patient selection is strict rather than discretionary. Selecting appropriately chosen patients is what keeps ACI cost-effective; it is not a reason to discount eligibility before seeking assessment.

Matrix-induced ACI (MACI) is a related but distinct technique covered under a separate appraisal, NICE TA508, and involves different technical and clinical considerations. It is outside the scope of the ACI-versus-mosaicplasty comparison discussed in this article.

Getting assessed for the right cartilage procedure

Choosing between ACI and mosaicplasty depends on factors that imaging and clinical assessment must establish before any technique decision is made: defect size, anatomical location (femoral condyle versus patella), prior surgical history, patient age, and activity level. MRI characterises defect dimensions and depth; arthroscopic findings, where required, confirm grade and containment. Neither piece of information alone is sufficient — and the surgical history in particular, given what the evidence shows about failure rates after earlier procedures, is one of the first things a specialist will review.

For patients with larger defects, patellar involvement, or a prior repair that has not held, referral to a dedicated cartilage centre is the appropriate next step. In London, that specialist assessment is available at the London Cartilage Clinic on Harley Street, where imaging findings, surgical history, and patient goals are brought together to reach a considered clinical recommendation. To arrange a consultation, visit londoncartilage.com.