Which procedure produces better results long-term?

For an active patient weighing up these two operations, the 10-year evidence points firmly to OATS/mosaicplasty as the more durable choice. The gap between the two procedures widens with time — and that trajectory matters more than any short-term snapshot.

OATS (osteochondral autograft transfer) works by transplanting small cylinders of bone and cartilage from a non-load-bearing part of the same knee into the damaged area, restoring the joint surface with true, structural cartilage. Microfracture takes a different approach: tiny holes are drilled through the bone beneath the defect to draw up marrow cells, which then form a repair tissue over the lesion.

The most robust comparison comes from a Lithuanian randomised controlled trial by Gudas and colleagues — the only published long-term RCT conducted specifically in athletes. In the 2005 three-year results (n = 60 young athletes), OATS produced excellent or good outcomes in 96% of patients against 52% for microfracture. The 2012 ten-year extension of the same trial showed OATS holding at 87.5% excellent/good outcomes, while the microfracture group had slipped to 63%. Critically, microfracture results had been declining steadily between years three and ten; the OATS group showed no such erosion.

A separate systematic review by Pareek et al. (2016), drawing on data at a mean follow-up of 10.2 years, recorded an OATS failure rate of just 6.4%, compared with microfracture failure rates of 23–47% beyond five years. The underlying reasons for this divergence — biological, mechanical, and patient-related — are explored in the sections that follow.

Why the two repairs behave differently over time

The diverging outcomes stem from a fundamental difference in the tissue each procedure creates — and how that tissue holds up under the demands of sport.

Microfracture works by breaching the bone beneath the cartilage defect, releasing marrow cells that clot over the lesion and gradually solidify into repair tissue. The problem is that this repair tissue is fibrocartilage — a softer, less organised material built primarily from type I collagen (the same protein found in tendons and scar tissue), rather than the type II collagen that makes up healthy joint cartilage. Under a microscope, native cartilage has a layered, zonal structure engineered to absorb and distribute load. Fibrocartilage lacks that architecture, which means it compresses and wears more readily.

Mithoefer and colleagues' meta-analysis of 28 studies — covering more than 3,000 patients — found that microfracture functional scores typically peak at around 18–24 months after surgery and then decline. That timeline fits the biology: the fibrocartilage patch holds up reasonably well at first, but repeated sport-level loading gradually degrades it.

Microfracture carries a further risk that is particularly relevant to active patients. The drilling process can disrupt the subchondral bone plate — the dense layer of bone immediately beneath the cartilage — and some studies suggest this disruption may complicate or limit the options for revision surgery if the initial repair fails.

OATS plugs avoid both problems. Because the transplanted cylinders carry intact hyaline cartilage together with the underlying bone, the zonal structure and subchondral interface are preserved from the outset. The graft integrates mechanically with the surrounding tissue and continues to bear load without the same degradation pathway, which is why OATS results in the Gudas trial remained stable across a decade while microfracture outcomes eroded.

What the evidence base actually shows

Three tiers of evidence build the comparative case: a randomised trial conducted specifically in athletes, a long-term systematic review, and a large real-world case series.

The Gudas randomised controlled trial

The 2005 trial (n = 60 young athletes) and its 2012 ten-year extension remain the only published long-term RCT comparing these two procedures in an exclusively athletic population. The outcome scores from both papers — cited in the section above — confirmed a decisive and durable advantage for OATS that widened progressively across the decade rather than narrowing as joints matured.

Failure rates: Pareek et al.'s systematic review

At a mean follow-up of 10.2 years, Pareek et al. (2016) recorded an OATS failure rate of 6.4%. Comparable microfracture data from the same period reported failure rates of 23–47% beyond five years, with the wide spread reflecting how sharply microfracture outcomes vary by defect size, patient age, and activity demand. Most of this body of evidence originates from specialist single-centre programmes rather than multisite trials — a limitation worth noting when interpreting the absolute figures, though the direction of the advantage is consistent across study types.

Location-dependent results: Hangody's mosaicplasty series

Hangody and Füles' series of 831 patients — the largest published mosaicplasty dataset — provided location-specific benchmarks at a mean 4.2-year follow-up: 92% good or excellent outcomes for femoral condyle defects, 87% for tibial plateau lesions, and 74% for patellofemoral cases. The lower figure for patellofemoral defects reflects the more complex loading environment behind the kneecap and directly informs surgical planning.

Return to sport

For the active patient, the most tangible measure is return to prior sport level. OATS and mosaicplasty achieve this in approximately 83–93% of cases, against 61–75% for microfracture — a gap that translates the durability difference into practical athletic terms. Taken together, these findings give patients with a focal defect within the OATS size range and a clear return-to-sport goal a consistent evidence base from which to open a specialist conversation.

Which patients are suited to OATS and which to microfracture

Defect size is the first filter. OATS suits focal chondral or osteochondral lesions of roughly 1–4 cm² — from approximately the area of a small fingernail to that of a thumbnail. Within that range, mosaicplasty (the variant using multiple small plugs, each under 6 mm in diameter) can fill irregular or slightly larger defects up to the 4 cm² ceiling. Beyond that point, the donor supply from the knee's own non-load-bearing margins runs out, making autograft transfer impractical — a situation addressed in the next section.

Size alone does not determine suitability. The Gudas ten-year trial data point to three patient-level characteristics that strengthen the case for OATS: age under 40, an acute traumatic origin for the defect rather than chronic degeneration, and a body mass index in the normal range. These factors together identify the patient most likely to gain durable benefit — someone whose joint is otherwise healthy, whose defect has not yet remodelled, and whose loading profile gives the graft a reasonable mechanical environment.

BMI warrants specific attention. Higher body weight accelerates fibrocartilage breakdown under compressive load, which is one reason microfracture deteriorates more quickly in heavier patients. For an overweight but still active patient with a focal defect within the OATS size range, this is an additional reason to favour autograft rather than assume microfracture represents a simpler equivalent.

Defect location also shapes the decision. As the Hangody series outcomes outlined in the previous section show, femoral condyle lesions respond most reliably to mosaicplasty; patellofemoral defects carry a lower success rate, reflecting the more complex loading environment behind the kneecap — a relevant consideration for runners and jumping athletes.

Microfracture may still arise in conversation for very small lesions — under 1 cm² — in patients with genuinely limited activity demands, but it is not a modern first-line choice for active individuals. Its benefits are most fragile precisely where durability matters most.

Defects larger than 4 cm² exhaust the practical autograft supply and open a separate conversation about allograft or cell-based alternatives.

Recovery, donor-site morbidity, and return to sport

Recovery from OATS follows a slower initial arc than microfracture — but a more stable long-term one. The distinction matters practically, because patients sometimes mistake early post-operative progress for overall trajectory.

Donor-site morbidity

Harvesting osteochondral plugs involves taking small cylindrical grafts from the less-loaded margins of the femoral condyle in the same knee. Most patients experience only mild, temporary discomfort at the harvest site. Hangody and Füles' series of 831 patients confirmed that donor-site morbidity is generally self-limiting and not functionally disabling, though it remains a meaningful consideration to discuss before surgery — and one that limits the total graft area that can be harvested.

Recovery arc and rehabilitation

Microfracture patients often feel functional improvement earlier — within weeks to a few months — because there is no donor site to heal and the procedure is less technically involved. That early advantage sits on a fragile base: the fibrocartilage repair tissue tends to soften under repeated athletic loading, and published series indicate that function in high-demand patients begins to decline from around two years onwards.

OATS recovery demands structured physiotherapy and staged load progression over roughly six to twelve months, varying with defect size and sport type. The return is to a repair built from true hyaline cartilage — one that has shown stable outcomes in published long-term series out to ten years.

Applying the return-to-sport data

The return-to-sport figures cited in the preceding section take on sharper meaning when applied to an individual athlete. A competitive runner or court-sport player falling into the microfracture non-return group is more likely managing progressive pain that precludes sport entirely than simply training at a reduced load. Even among successful OATS recipients, some adjustment in training volume or performance ceiling is possible at elite level — realistic pre-operative counselling should address this directly.

When a different approach may be more appropriate

OATS and mosaicplasty have a defined technical window, and not every patient falls within it. Understanding where the boundaries lie is part of the specialist consultation.

Defects above 4 cm² exhaust the practical autograft supply from a single knee. At that scale, the relevant alternatives are autologous chondrocyte implantation (ACI or its matrix-enhanced variant MACI), which use cultured cartilage cells seeded on a collagen membrane, or osteochondral allograft (OCA), which draws on donor tissue to fill larger or posttraumatic lesions.

AMIC (autologous matrix-induced chondrogenesis) is a single-stage hybrid: it combines microfracture with a scaffold designed to stabilise the marrow clot and improve the quality of repair tissue. Evidence is growing but remains shorter-term than the decade-long data available for OATS.

Malalignment of the knee — where the mechanical axis places excessive load on one compartment — can undermine any cartilage repair over time. An osteotomy (HTO to correct varus, DFO to correct valgus) may be discussed alongside a cartilage procedure to offload the treated area.

Advanced or diffuse osteoarthritis is a different clinical situation entirely: cartilage repair addresses focal defects, not widespread joint degeneration, which typically requires joint-preservation or replacement pathways.

For any active patient with knee cartilage symptoms, an MRI and specialist assessment — reviewing defect size, location, alignment, and activity level — is the appropriate starting point. Search MSK provides a directory of knee cartilage specialists across the UK, searchable by region and treatment offered.