What the 12-to-36 month data actually shows

The most practical question patients ask before committing to ChondroFiller is whether the benefit holds — or whether it fades within a year or two. The short answer, based on the available evidence, is that it appears to hold, though the evidence itself comes with important caveats about study size.

Across four knee studies synthesised in the manufacturer's Clinical Evaluation Report (CER v09, April 2025), IKDC functional scores improved by roughly 30 points over the first 12 months. That comfortably exceeds the minimal clinically important difference of 16.7 points — the threshold below which patients typically cannot feel a meaningful change in their daily function.

The strongest signal for durability beyond 12 months comes from Jerosch et al., a post-market clinical follow-up study that tracked patients to 36 months. The mean IKDC gain of 32.4 points was not only sustained at three years but slightly increased, with patients reaching an absolute score of 80 — a level consistent with near-normal joint function in this context.

That is a consistently positive picture. The honest caveat is that most of the contributing cohorts are small, predominantly single-centre, and closely linked to the manufacturer's regulatory evaluation programme. No large independent randomised trial with 24–36 month follow-up has been published. The data warrants cautious optimism rather than certainty, and individual outcome will depend on factors explored in the sections that follow.

How functional scores hold up at three years

An absolute IKDC score of 80 is worth pausing on. The scale runs from zero to 100, where higher numbers reflect better joint function; scores at or above 75–80 broadly correspond to a patient who can sustain walking, cycling, and recreational activity without significant restriction. That is where the Jerosch et al. PMCF cohort sits at three years — and the trajectory to reach it was not a plateau arrived at early and then passively held.

The PMCF data indicate that the functional gain continued to consolidate between the 12-month and 36-month marks. Rather than reflecting a short-lived mechanical effect, this pattern is consistent with the collagen scaffold progressively recruiting and organising repair tissue over time — a process that, by its nature, unfolds across months rather than weeks. In available follow-up extending to three years, benefit was sustained rather than eroded.

Whether that durability extends beyond the knee is a reasonable question, and a 26-patient prospective cohort of acetabular cartilage defects larger than 2 cm² provides some cross-joint reassurance. The hip operates under different load profiles and lesion geometries than the knee, yet 17 of 21 evaluable patients in that cohort achieved good or excellent MRI-confirmed outcomes across the three-, four-, and five-year time points — mirroring the knee finding of durable multi-year benefit despite the anatomical differences.

Both datasets carry the same important caveat: the cohorts are modest in size and follow-up beyond five years has not been reported. Sustained to three years in available follow-up is the honest summary; longer-term behaviour in either joint remains to be established by larger independent studies.

What MRI shows about the repair tissue

MOCART — Magnetic Resonance Observation of Cartilage Repair Tissue — runs from 0 to 100; scores above 80 indicate good defect filling with sound integration into surrounding native cartilage. The European knee cohorts sit between 81.6 and 84.3 on this scale, which is clinically encouraging. More instructive than any single endpoint score, however, is the trajectory those numbers follow.

In the Schneider et al. randomised study, MOCART stood at 65.3 at four weeks post-injection — reflecting the fresh scaffold in situ — and climbed to 81.6 by 52 weeks. That upward arc matters mechanistically. A scaffold that simply fills a void without biological activity would not produce rising repair scores across twelve months; the progression points instead to progressive cell recruitment and tissue remodelling, with the collagen matrix guiding endogenous repair before being gradually replaced by maturing tissue. This is the process of acellular matrix-induced chondrogenesis in practice, not a passive bulking effect.

Broader imaging findings across cohorts tell a consistent structural story: reductions in bone marrow oedema, smaller periarticular effusions, and visible joint space widening have all been noted on post-injection MRI in published series.

A different kind of objective evidence comes from the wrist. In a 2025 arthroscopic second-look study of intra-articular distal radius fractures, ChondroFiller-treated patients showed significantly better cartilage quality than controls — median Outerbridge grade 1.5 versus 3 (p=0.006) and ICRS grade 1 versus 3 (p=0.002). Direct arthroscopic visualisation of cartilage quality is a more concrete tissue-level measure than a symptom score alone, and those findings align with the structural repair picture the knee imaging data suggest.

Which patients see the clearest benefit

Candidacy for ChondroFiller comes down to one anatomical distinction: a focal, isolated area of cartilage damage versus widespread joint degeneration. The treatment is designed for contained Grade III or IV lesions — meaning cartilage worn down to, or reaching, the underlying bone — surrounded by healthy cartilage edges that can anchor the scaffold in place. When damage is instead diffuse, affecting multiple surfaces or much of the joint lining, ChondroFiller is not the appropriate option; neither is it suitable where a knee or hip already shows Kellgren-Lawrence Grade IV osteoarthritis, where joint-wide wear is too advanced for a focal scaffold approach.

The hip data sharpens this further. In the 26-patient acetabular cohort, patients who had pre-existing osteoarthritis of Tönnis grade 2 or 3 — roughly equivalent to moderate-to-severe joint narrowing on X-ray — fared poorly. That finding is clinically useful: it points to disease severity, rather than defect size alone, as the key question a consultant needs to answer before recommending treatment.

Technique precision also matters. The 2025 wrist study found fibrous tissue formation occurring exclusively in defects that had been overfilled; flush, accurate application was free of that complication. This is the practical argument for experienced, image-guided placement.

Finally, the scientific basis for the standard post-procedure weight-bearing advice: early biomechanical testing shows the collagen gel takes time to stabilise within a defect, and during that initial phase it does not fully shield the opposing cartilage surface from load. Delaying full weight-bearing until the scaffold has consolidated is the clinical response to that biology — routine recovery guidance, not a cause for concern.

Where the evidence is still limited

Two specific gaps in the published record deserve naming. The randomised study comparing ChondroFiller with microfracture enrolled only 13 patients in the ChondroFiller arm — and high dropout in the microfracture group made any head-to-head comparison effectively uninterpretable. That study confirms the treatment is deliverable and that IKDC scores improve meaningfully; it cannot settle whether outcomes differ from an active comparator under controlled conditions.

The 36-month Jerosch et al. PMCF data — the primary anchor for durability at the upper end of this review window — carries a different caveat: patient-level data and the exact retained sample size are not fully detailed in the available published extracts. The direction of the result is consistent with every other dataset in the record and clinically plausible, but the figure rests on a narrower evidentiary base than the headline number alone implies.

What a properly powered independent trial — 100 to 150 patients, multi-centre, followed through 36 months — would settle is not whether ChondroFiller works, but whether the magnitude and durability of benefit hold across a broader, unselected population. The consistent directional signal across multiple small studies is itself meaningful. What is still being established is how confidently that signal generalises.

Getting a clinical opinion without a referral

For patients whose defect profile — focal Grade III or IV damage, healthy surrounding cartilage, no advanced joint-wide degeneration — fits the criteria discussed above, the practical question is what assessment looks like.

At MSK Doctors, the initial step is a consultant-led review that typically includes imaging to confirm defect size, location, and joint health before any treatment decision is made. Clinics in Sleaford, Lincolnshire and Grantham offer that assessment pathway, with Open MRI available at the Sleaford site for cases where detailed cartilage imaging is needed. No GP referral is required to book.

For patients based in London, the same specialist assessment is available through the London Cartilage Clinic.

ChondroFiller, where it is appropriate, is delivered as an ultrasound-guided outpatient injectable scaffold — no theatre admission, no general anaesthetic. Suitability is confirmed at assessment, not assumed.

Appointments can be booked directly at mskdoctors.com.

1. [1] Controlled, randomized multicenter study — ChondroFiller liquid vs microfracturing (focal knee cartilage defects). (2016). https://doi.org/10.5348/VNP05-2016-1-OA-1 https://doi.org/10.5348/VNP05-2016-1-OA-1
2. [2] Arthroscopic utilization of ChondroFiller gel for the treatment of hip articular cartilage defects: a cohort study with 12- to 60-month follow-up. (2021). https://doi.org/10.1093/jhps/hnab002 https://doi.org/10.1093/jhps/hnab002
3. [3] Development of an Ex Vivo Osteochondral Biomimetic Platform for Mechanistic Investigation of Cartilage Regeneration. (2025). https://doi.org/10.3390/ijms262311759 https://doi.org/10.3390/ijms262311759
4. [4] Influence of cartilage defects and a collagen gel on integrity of corresponding intact cartilage: a biomechanical in-vitro study. (2024). https://doi.org/10.1007/s00402-024-05530-z https://doi.org/10.1007/s00402-024-05530-z