The gap that appears two to three years after surgery

The question most patients ask at the point of choosing is not 'which procedure works better on the day?' but 'which one will still be working in three or five years?' It is the right question, and the honest answer is that the two approaches set off on different biological trajectories from the moment treatment is delivered — but that gap takes 18 to 36 months to become visible.

For much of the first year, patient-reported outcomes look reassuringly similar. In a 92-patient multicentre randomised controlled trial, IKDC score gains at 12 months were statistically equivalent between a scaffold-based implant and microfracture — 25.56 versus 27.51. A patient at their one-year review would have no reason to suspect a difference. Yet arthroscopic assessment at that same 12-month point already showed more fully regenerated cartilage in the scaffold group. The tissue quality had diverged well before the functional scores caught up.

This matters because it means short-term patient-reported outcomes are an unreliable guide to long-term durability. The biology driving the eventual gap is established on day one; the clinical consequences simply take time to surface.

The fibrocartilage problem: what microfracture actually produces

Drilling through the subchondral plate — the dense bone layer immediately beneath articular cartilage — creates channels through which bone marrow mesenchymal stem cells flow into the defect. That is exactly what microfracture is designed to do. The problem is what happens to those cells once they arrive.

Native hyaline cartilage is built around a Type II collagen scaffold interlaced with proteoglycans — a matrix that gives it resilience under compression. Bone marrow stem cells are capable of producing this architecture, but only when the local environment provides the structural cues to guide differentiation. In a bare cartilage defect, those cues are absent. Research into marrow stimulation techniques has consistently found that, without a guiding matrix, cells default to producing fibrocartilage — a tissue dominated by Type I collagen, the same collagen found in scar tissue and tendons. A 2019 review of the marrow stimulation literature described this as the consistent finding across the evidence base, not an occasional subset result.

Fibrocartilage is softer, less organised, and poorly suited to sustaining the cyclic compressive loads of everyday movement. It wears. This vulnerability explains the characteristic trajectory: genuine functional improvement in the first year or two, followed by plateau and decline as loading progressively erodes the repair tissue.

Defect size amplifies the problem. Microfracture performs comparatively better in lesions under approximately 2 cm², where fibrocartilage fill may provide adequate short-term relief. In larger defects, the greater volume of poorly supported tissue under increased biomechanical stress tends to accelerate deterioration.

The fibrocartilage outcome is a structural consequence of the biological environment microfracture creates — not a reflection of surgical technique.

What the survival and imaging data show at two to five years

Survival data from a Kaplan-Meier analysis of 203 knees traces the microfracture trajectory with unusual precision. Fewer than 80% of repairs were still functioning within 12 months; by three years, fewer than 60% remained intact. Mean time to failure was 4.0 years, and the overall long-term failure rate reached 66% — compared with 51% for osteochondral autograft, where survival stayed above 80% for seven years.

The imaging evidence reinforces this picture. MOCART — a validated MRI scoring system that measures how completely and how well a repair fills the original defect — provides an objective tissue-quality measure alongside patient-reported function. A five-year prospective RCT comparing autologous chondrocyte implantation (ACI) with microfracture found MOCART scores of 62.3 versus 26.7 (P<0.0001), Lysholm functional scores of 84.5 versus 64.9, and KOOS scores of 390.9 versus 303.0. That trial compares ACI, not ChondroFiller, with microfracture — but it documents the scale of advantage a scaffold-guided approach can accumulate over time, a principle that applies across scaffold-based techniques. The gap is not a sudden collapse; it widens progressively as the fibrocartilage layer erodes under load.

Second-look arthroscopy at a mean of 3.6 years found 36% of microfracture repairs incompletely healed, with a mean MOCART of only 67.8 — in practical terms, roughly one in three patients still had significant defect-fill problems more than three years on. A 2024 systematic review, examining medium-to-large defects at ten or more years of follow-up, confirmed limited long-term efficacy, high failure rates, and clinically meaningful functional decline.

The importance of the 12-month structural divergence — evident on arthroscopy even when patient-reported scores looked equivalent — is that it shows the gap establishing itself well before any patient would register a change.

Subchondral bone damage: a second failure mechanism

Fibrocartilage quality is only part of the failure picture. The drilling process itself can cause lasting structural damage to the subchondral bone — the dense layer that sits immediately beneath articular cartilage and provides its mechanical foundation.

When the subchondral plate is perforated, the disruption sometimes triggers abnormal bone remodelling at the repair site: cysts can form within the bone, bony spurs or overgrowth can develop at the margins, and the plate may stiffen in ways that reduce its load-distributing function. The repair tissue above — already fibrocartilage rather than hyaline cartilage — then loses mechanical support from beneath as well as being inherently inferior from within. The two problems compound each other.

The clinical implication extends beyond the initial repair. Subchondral bone damage from prior microfracture is a recognised factor that increases failure rates in subsequent cell-based procedures such as ACI, as the compromised bone environment complicates re-implantation and integration. Patients who have had microfracture may therefore find that their options for salvage procedures are narrowed, not simply because the first repair failed, but because the procedure altered the tissue bed that any second repair would depend on.

Preserving the subchondral plate is consequently not a minor procedural preference — it keeps future treatment pathways open.

ChondroFiller injection: a different biological starting point

Rather than stimulating the bone marrow to release cells into an unstructured defect, ChondroFiller injection takes a structurally different approach from the outset — and that difference begins with what it does not do to the subchondral plate.

The treatment is an ultrasound-guided outpatient injection of a cell-free, two-component collagen type I hydrogel. Once placed into the focal defect under image guidance, the scaffold gels in situ and acts as a three-dimensional matrix into which the patient's own progenitor cells migrate. Crucially, those cells encounter structural cues that guide them toward hyaline-like differentiation — a process known as matrix-induced chondrogenesis — without any drilling of the subchondral bone. The integrity of the bone layer, discussed in the previous section as a determinant of long-term repair success, is preserved.

In published series, MOCART gains of 70 to 87 have been reported across knee, hip, and small-joint applications; a prospective cohort of 17 knee patients showed statistically significant Lysholm and IKDC improvements at three, six, and twelve months, with scores stabilising between six and twelve months rather than plateauing then declining. Across all joint types, more than 20,000 implantations have now been completed over more than ten years.

Two practical caveats apply. First, technique matters considerably: overfilling the defect promotes fibrous rather than cartilage-like tissue, making precise, flush-level placement a clinical prerequisite — in one wrist series, flush application produced significantly better ICRS and Outerbridge grades than overfilled controls (P=0.006 and P=0.002 respectively). Second, the current evidence rests on smaller prospective cohorts; the kind of large, multi-year randomised trial that exists for microfracture has not yet been completed for ChondroFiller injection in matched populations. ChondroFiller is CE-marked for focal defects up to 3 cm², extendable to 6 cm² — a size range that already covers lesions where microfracture evidence is weakest.

What to weigh before choosing — and what current evidence cannot yet answer

Making sense of two evidence bases that differ in size and direction is the practical challenge facing any patient at this decision point.

Microfracture carries the larger body of long-term data — but that evidence, as the preceding sections document, consistently points toward structural deterioration, particularly for lesions above approximately 2 cm² and at follow-up beyond two to three years. A 2024 systematic review confirmed limited long-term efficacy for medium-to-large defects at ten or more years of follow-up. ChondroFiller injection, by contrast, has a mechanistically sounder biological starting point and a growing mid-term clinical record, but the large multi-year randomised trials needed to confirm durable superiority in matched populations have not yet been completed. The SISMIC trial — evaluating microfracture with and without collagen scaffold augmentation, using KOOS at 24 months as its primary endpoint — represents the field's current effort to fill that gap prospectively. Worth noting, however, is that a two-year endpoint may still fall short of the divergence arc that five-year data have already revealed.

Several individual factors moderate which approach is most appropriate: defect size, activity demands, the joint involved, and whether prior treatment has already altered the subchondral environment. Patients with small focal lesions are differently situated from those with medium or large defects, and the appropriate choice depends on integrating MRI findings, lesion characteristics, and individual goals — not on general preference.

That integration is what specialist cartilage assessment is for, and no generalised comparison of technique evidence can substitute for it. Patients looking for that evaluation in London can arrange a consultation via londoncartilage.com.

1. [1] Long-Term Survival after Microfracture and Mosaicplasty for Knee Articular Cartilage Repair. (2018). https://doi.org/10.1177/1947603518783482 https://doi.org/10.1177/1947603518783482
2. [2] Microfracture for Medium Size to Large Knee Chondral Defects Has Limited Long-Term Efficacy: A Systematic Review. (2024). https://doi.org/10.1002/jeo2.70060 https://doi.org/10.1002/jeo2.70060
3. [3] Costal Chondrocyte–Derived Pellet-Type ACI Versus Microfracture: 5-Year Follow-up of a Prospective Randomized Trial. (2024). https://doi.org/10.1177/03635465231222797 https://doi.org/10.1177/03635465231222797
4. [4] Implantation of ChondroFiller Liquid® as a Scaffold Material for Chondral Lesions of the Knee Joint. (2024). https://doi.org/10.5272/jimab.2024304.5936 https://doi.org/10.5272/jimab.2024304.5936
5. [5] Hip Chondral Defects: Arthroscopic Treatment With Needle and Curette Technique and ChondroFiller. (2021). https://doi.org/10.1016/j.eats.2021.03.011 https://doi.org/10.1016/j.eats.2021.03.011