Which procedure fits which defect size

The size of the damaged area, measured on MRI, is the first number a surgeon will look at when deciding between these two techniques — and the thresholds are fairly clear.

Under 2 cm²: a single OATS plug, typically 8–10 mm in diameter, can cover the defect with a seamless layer of hyaline cartilage pressed directly into place. One plug means one harvest site, which keeps the surgical footprint as small as possible.

2–4 cm²: a single plug cannot resurface an area this wide. Mosaicplasty fills the gap by tiling two or more smaller cylinders — usually 2.7–6.5 mm in diameter — side by side in a mosaic pattern until the lesion is covered. More plugs are needed, but the underlying principle is the same: native cartilage transferred from a non-weight-bearing zone of the same knee.

Above 4 cm²: autograft donor capacity runs out. At this scale, surgeons typically move to cell-based techniques such as MACI or ACI, or to osteochondral allograft transplantation (OCA) using donor tissue.

This three-band framework is consistent across NICE guidance (HTG463), the 2021 Cartilage Book, and multiple clinical practice summaries.

Before either procedure is considered, the defect should be confirmed as focal and full-thickness — ICRS grade III or IV — and conservative measures such as physiotherapy or injection therapy should already have been tried without sufficient improvement. Diffuse or advanced joint-wide arthritis sits outside the scope of either technique.

How single-plug OATS works on compact lesions

During OATS, the surgeon cores out a cylindrical plug of healthy bone and cartilage — typically 8–10 mm across — from a peripheral area of the femoral condyle that bears little load during ordinary movement. A matching channel is prepared in the damaged zone, and the plug is press-fitted in: driven firmly until its surface sits flush with the surrounding joint surface. No adhesive or fixation hardware is required; the bone-to-bone interference fit holds the graft securely in place.

Hyaline cartilage distributes compressive forces evenly across the joint surface, resisting the cyclic loading a knee encounters with every step and push-off. A single plug creates a continuous, uninterrupted layer of this tissue — no gaps, no margin discontinuity, and the tightest possible plug-to-bed contact. Because only one core (or at most two) is harvested, the impact on the donor area is kept to a minimum.

The ideal candidate is generally under 40–45, with a focal, full-thickness osteochondral or traumatic defect of approximately 2 cm² or smaller, healthy surrounding cartilage on the same joint surface, and symptoms that have not resolved with conservative management such as physiotherapy or injection therapy. Patients over 50, those with generalised knee osteoarthritis (Kellgren-Lawrence grade above 2), or a BMI above 40 are less likely to be suitable.

Within the eligible group, the 2021 Cartilage Book identifies the two factors that most consistently predict a good return to the same level of sport: age below 25 and a defect below 2 cm² — the combination that maps most directly onto the compact single-plug scenario.

When mosaicplasty becomes the indicated approach

Tiling smaller cylinders across a wider defect follows the same biological logic as a single plug — each cylinder still delivers native hyaline cartilage — but the geometry of the technique creates a structural trade-off that is worth understanding clearly.

The surgeon harvests two or more cylinders, typically 2.7–6.5 mm in diameter, from the peripheral femoral condyle and presses them one by one into a cluster of prepared channels until the lesion is covered. The result resembles a mosaic floor: the tiles themselves are sound, but small gaps inevitably remain between them. Those inter-plug spaces fill not with hyaline cartilage but with fibrocartilage — a structurally inferior repair tissue. The composite surface is therefore heterogeneous, and this is the principal mechanical trade-off of the multi-plug approach compared with the seamless layer a single OATS plug provides.

Because each plug requires its own harvest, the donor-site burden also scales with the number of cylinders used. Whereas single-plug OATS creates one harvest channel in the non-weight-bearing zone, a mosaicplasty covering the upper end of the 2–4 cm² range may need three or more, increasing the potential for localised pain or bone marrow oedema at the donor area.

Efforts to address the gap problem include hexagonal osteochondral graft systems (HOGS), which use geometrically interlocking plugs designed to tile without leaving spaces. Early results — assessed by IKDC score and MOCART MRI at six months in a small case series for lesions between 1.5 and 6 cm² — are reported as encouraging, but this approach remains experimental and is not yet standard practice.

Why both beat microfracture for active patients

Choosing between OATS and mosaicplasty is partly a question of defect geometry — but the more fundamental question is why either procedure is preferable to the older, simpler alternative of microfracture, and how they compare with the cell-based techniques used for larger lesions.

Microfracture works by puncturing the subchondral bone to release marrow cells that then form a repair tissue. The problem is that this tissue is fibrocartilage — not the native hyaline cartilage that lines a healthy knee. Fibrocartilage is softer, less stiff, and carries a lower tolerance for the cyclic loading that active patients place on the joint with every stride, pivot, or jump. In published clinical series, this repair tissue has been observed to break down within two to three years in active individuals. There is a further mechanical penalty: drilling through the subchondral plate damages the bone architecture beneath the defect, which can compromise the chances of success if a more definitive restoration procedure is needed later.

Both OATS and mosaicplasty avoid this problem entirely. The grafts transferred are genuine hyaline cartilage — the same load-bearing tissue that occupied the defect site originally — with structural properties suited to long-term joint function. Gudas et al. reported good outcomes in 86–90% of OATS patients at three years; Muller et al. found that 92% of competitive athletes returned to sport at an intermediate-to-high level after a mean 42-month follow-up.

For defects too large for autograft coverage, ACI and MACI are well-established options. Both, however, require a two-stage pathway: a biopsy is taken first, chondrocytes are cultured in a laboratory over several weeks, and re-implantation is then performed in a separate operation. OATS and mosaicplasty accomplish everything in a single theatre session — a practical distinction that matters for recovery planning and for patients who want to avoid the delay and logistical burden of a staged procedure.

Donor-site morbidity: the risk both techniques share

Attention during recovery naturally gravitates toward the repair site, but the harvest zone on the peripheral femoral condyle carries its own risk profile — one that applies to both procedures.

The clearest quantified picture comes from a 2016 systematic review by Andrade et al., which pooled 21 studies and 1,726 patients undergoing knee-to-knee mosaicplasty. Mean donor-site morbidity was 5.9%. The most commonly reported complaints were crepitation (31% of symptomatic patients) and patellofemoral disturbances (22%). Importantly, the rate of morbidity did not correlate with defect size in that analysis — the per-plug risk appears relatively constant rather than scaling with the extent of the lesion being repaired. For most patients, donor-site symptoms do not persist in the long term, and the majority are asymptomatic from the harvest zone at follow-up.

Because single-plug OATS creates only one harvest channel, its absolute donor burden is inherently lower than a multi-plug mosaicplasty procedure requiring three or more cylinders — a distinction that becomes more meaningful at the upper end of the 2–4 cm² range.

One mitigation that is increasingly used regardless of plug number is back-filling: the tunnel left behind after each harvest is packed with allograft bone or a purpose-designed filler to support the surrounding tissue and reduce the mechanical irritation that can otherwise cause localised discomfort. Surgeons weigh donor-site considerations alongside defect geometry and patient activity level during planning — it is a meaningful factor, not a prohibitive one.

Recovery, outcomes, and what the evidence still cannot settle

At ten years, the Pareek et al. systematic review found IKDC and Lysholm clinical scores improved significantly following osteochondral autograft transfer — but Tegner activity scores, which reflect return to sporting activity, showed no statistically significant change. That gap between functional recovery and sport-level recovery is the honest ceiling the evidence describes: most patients gain meaningful pain relief and improved knee function, yet restoration to exactly the pre-injury sporting level cannot be assumed. Campbell et al.'s synthesis of 1,117 patients confirms that return-to-sport rates after OAT remain higher than with alternative cartilage repair strategies overall.

Within the autograft techniques, available comparisons indicate that return to previous sport takes longer after multi-plug mosaicplasty than after single-plug approaches — consistent with the greater tissue disruption when multiple harvest cylinders are required.

Beyond defect size and patient age — addressed in earlier sections — the factors most commonly associated with better outcomes include the absence of prior marrow-stimulation procedures and good surrounding cartilage quality at the time of surgery. Existing microfracture can alter subchondral bone architecture in ways that compromise plug integration, and degenerating adjacent cartilage reduces the biological stability the graft needs to consolidate.

What the evidence still cannot settle is long-term comparative durability between single- and multi-plug configurations. No large-scale head-to-head randomised trial exists; most comparative data come from systematic reviews of heterogeneous series. Retzky et al. (2024) and NICE HTG463 address initial outcomes but acknowledge that follow-up duration remains a constraint across the evidence base.

The Tegner finding and the 2 cm² threshold together frame the practical picture: a compact defect treated with a single plug offers the most favourable conditions for durable repair, but the published data — even in that scenario — support realistic functional expectations rather than a guarantee of full sporting restoration. Lincolnshire Knee's assessment pathway includes consultant-led evaluation and, where indicated, onMRI™ cartilage analysis to characterise defect size and lesion quality precisely before any procedure is planned; appointments are available without referral at lincolnshireknee.co.uk.