What the 10-year evidence actually shows

At ten years, OATS and mosaicplasty consistently outperform microfracture where it matters most: how often the repair holds, and how long before it fails.

The largest direct dataset comes from Solheim et al.'s 2018 cohort study (n=203; 119 microfracture, 84 OAT), which tracked patients to a true long-term endpoint. Overall failure reached 62%, but the breakdown between procedures is where the picture becomes stark: microfracture failed in 66% of patients versus 51% for OAT — a statistically significant difference (P=0.01). The timing of failure separates the procedures even more sharply. Mean time to failure was 4.0 years for microfracture compared with 8.4 years for OAT (P<0.001). The Kaplan-Meier survival curves tell the same story in visual terms: OAT graft survival remained above 80% for the first seven years and above 60% at fifteen years; microfracture fell below 80% survival within twelve months and below 60% within three years.

The only prospective randomised controlled trial to follow exclusively athletic patients to a full decade — Gudas et al. (American Journal of Sports Medicine, 2012) — corroborates these findings: treatment failure occurred in 14% of OATS patients versus 38% of those treated with microfracture, a statistically significant gap (P<0.05). That a near-threefold difference in failure rates emerges in a well-controlled, single-population RCT gives the Solheim cohort data important independent support.

Muthu et al.'s 2024 network meta-analysis, drawing on a broader pool of comparative studies, reaches the same directional conclusion: mosaicplasty delivers significantly better long-term functional outcomes at ten years than microfracture. Across three distinct bodies of evidence — a large survival cohort, the only athlete RCT at ten years, and a pooled network analysis — the answer to whether the choice of procedure makes a meaningful difference at a decade is consistent: it does.

Why the two procedures produce different tissue

The difference in long-term durability comes down to biology: the two procedures grow different types of tissue, and not all cartilage is equal under load.

Microfracture works by puncturing the subchondral bone — the hard bone layer beneath the cartilage surface — with a small surgical awl. Those perforations allow marrow cells and blood to seep into the defect, forming a clot that gradually matures into fibrocartilage. Fibrocartilage is the same repair tissue the body produces when healing a skin wound; it fills the gap, but it is not the original material. Compared with native hyaline cartilage, fibrocartilage has lower stiffness, reduced ability to distribute load evenly, and poorer wear resistance under repetitive stress. Think of it as scar tissue rather than restored skin — functional in the short term, but less resilient over years of cyclical loading.

OATS and mosaicplasty take a different approach. Rather than stimulating the body to produce a substitute tissue, they transfer cylindrical plugs of osteochondral autograft — bone with its native hyaline cartilage surface still intact — harvested from a low-load-bearing area of the same knee and press-fitted into the defect. Mosaicplasty uses several smaller plugs arranged in a mosaic pattern, allowing defects up to roughly 4 cm² to be covered. The cartilage at the repair site is, structurally, the real thing.

This biological distinction is the underlying mechanism behind the survival curve divergence described above. As the knee continues to load the repair over years, fibrocartilage degrades progressively; hyaline cartilage, because it was never a compromise material, holds up far better.

The 2-to-10-year gap: why early results mislead

Evidence from shorter follow-up periods has sometimes been used to suggest the two procedures perform equally well — an interpretation the longer-term data do not support.

At two-year assessment, patient-reported outcome scores for OATS and microfracture are broadly similar. The fibrocartilage produced by marrow stimulation is still functioning reasonably at that point; the divergence has not yet emerged. Shorter-term studies reporting no significant difference between techniques are not wrong — they are simply observing patients before the critical period arrives.

That period falls between years five and ten. As repetitive loading continues, fibrocartilage progressively breaks down, and the survival gap widens — a pattern the Solheim Kaplan-Meier curves make difficult to misread. A study concluding equivalence from a two- or three-year endpoint is measuring an early snapshot, not long-term durability.

The most credible counterargument comes from Ulstein et al. (KSSTA, 2014) — a Level II randomised trial with a median 9.8-year follow-up in 25 patients — which found no statistically significant difference in Lysholm score (microfracture 69.7, OAT 62.6), KOOS subscales, or radiographic osteoarthritis grades. This deserves honest acknowledgement. The most likely explanation is that 25 patients is not enough to detect a clinically meaningful difference reliably. Even so, the Lysholm scores themselves are telling: both groups recorded results below 70, a level that reflects measurable ongoing impairment. Neither procedure reliably returned these patients to an untroubled knee at ten years.

One further consideration matters particularly for younger patients: microfracture involves perforating the subchondral bone plate, and there is evidence that this can alter the bone structure beneath the cartilage surface in ways that make a future cartilage repair procedure harder to perform and less likely to succeed. The choice of first procedure may therefore affect which options remain available later — something worth understanding before committing to a path.

Who gets the most benefit: lesion size and activity level

Two clinical factors determine whether OATS is likely to justify its added complexity for a given patient: the size of the cartilage defect and the physical demands the patient places on the knee.

Defect size as a structural threshold

Pareek et al.'s 2016 meta-analysis of prospective comparative studies found that the OAT advantage — higher activity levels and lower risk of failure — was clearest for lesions exceeding 3 cm². For smaller, contained defects below roughly 2 cm², mid-term outcomes between the two procedures are more similar, and a 2021 cartilage surgery reference text confirms that both microfracture and mosaicplasty remain reasonable options for defects in the 2–4 cm² range. The practical implication is that lesion size alone does not dictate the decision — it sets a threshold above which the case for OATS becomes considerably stronger.

Activity demand as the patient-centred modifier

Below that size threshold, what the patient actually does with their knee matters more than the anatomy. Gudas et al.'s RCT, conducted exclusively in young athletes, showed roughly double the rate of good or excellent outcomes with OATS compared with microfracture at a decade. In contrast, studies enrolling mixed-activity general populations — including Lim et al.'s Level 2 comparison — found no statistically significant difference in Lysholm or Tegner scores between the two techniques. Activity level partially accounts for this: repetitive high-impact pivoting loads expose the structural weakness of fibrocartilage in ways that light or variable activity does not.

Return to sport: an honest limitation

Even in the more favourable OATS cohorts, a gap remains between clinical improvement and full athletic recovery. Pareek et al.'s 10-year systematic review found that IKDC and Lysholm scores improved significantly from baseline, but Tegner activity scores — which capture the level of sport a patient can return to — did not change significantly. Functional gains are real; return to the patient's pre-injury sport level is not guaranteed, and this should be part of any pre-operative discussion.

Donor-site morbidity and surgical trade-offs

Every osteochondral autograft procedure introduces a second area of surgical disturbance within the same knee. To fill the damaged site, the surgeon harvests one or more cylindrical plugs from a relatively low-load region of the joint — typically the peripheral trochlear ridge or the margin of the intercondylar notch. That harvest site is a real wound, and patients should understand what it may involve.

Donor-site morbidity can include local pain, the formation of subchondral cysts, and occasional articular surface changes at the collection point. In most cases these are modest and resolve over time, but they are not trivial, and they vary between individuals. When mosaicplasty is used for a larger defect — requiring several small plugs arranged across the repair zone — donor burden increases proportionally, as does the technical complexity of the procedure.

OATS and mosaicplasty also demand considerably more surgical precision than microfracture. Matching the curvature and orientation of multiple plugs to the native joint surface requires experience with osteochondral transfer; outcomes are known to depend on the surgeon's familiarity with the technique.

Microfracture, by contrast, is arthroscopically straightforward, and that simplicity was a genuine advantage — particularly in settings where specialist expertise for autograft transfer was unavailable. The evidence reviewed at ten years does not erase that history. What it does establish is that technical simplicity alone is an insufficient basis for long-term procedure selection. Whether the added complexity and donor-site burden of OATS are acceptable trade-offs for a given patient is ultimately a conversation between patient and surgeon, informed by defect characteristics, activity goals, and access to appropriate expertise.

How this fits into a knee-preservation pathway today

For an active patient with a focal defect above 3 cm², a decade of consistent evidence points toward OATS or mosaicplasty as the more durable option. Below that threshold, and in less physically demanding patients, the picture is less definitive — and the broader cartilage repair landscape has developed considerably since these two procedures became the principal comparators.

Both sit within a joint-preservation strategy: to restore the cartilage surface of a symptomatic focal defect, relieve pain, and delay or avoid joint replacement — not to reverse diffuse disease.

Where the field has moved

Plain microfracture is no longer the default first-line choice for focal chondral defects. Matrix-augmented microfracture (AMIC) improves on it by adding a collagen or bioresorbable scaffold over the bone perforations to support better repair tissue quality. Cell-based options — autologous chondrocyte implantation (ACI) and its membrane-seeded evolution MACI — address the fibrocartilage limitation more fundamentally, though both require two surgical stages and are resource-intensive. For larger or post-traumatic defects where autograft harvest for OATS would be insufficient, fresh osteochondral allograft (OCA) provides donor-tissue coverage without creating a meaningful secondary wound in the patient's own knee.

For smaller contained defects where a non-surgical route is preferred or clinically appropriate, an injectable collagen scaffold — such as ChondroFiller injection, delivered under ultrasound guidance as an outpatient procedure — offers a matrix-induced chondrogenesis pathway that recruits the patient's own repair cells to the defect site.

Alignment as an adjunct

Limb alignment should always be assessed alongside cartilage repair planning. Varus or valgus malalignment concentrates load on the repaired compartment and can shorten the lifespan of any cartilage procedure. A corrective osteotomy — high tibial osteotomy (HTO) for varus deformity, distal femoral osteotomy (DFO) for valgus — may be recommended in combination with cartilage repair, or as a preparatory step.

Choosing the right approach depends on defect grade and size, limb alignment, activity level, and surgical history — factors that require a consultant assessment to weigh carefully. Lincolnshire Knee is part of the MSK Doctors group and accepts patients without referral; book an assessment at lincolnshireknee.co.uk.