What the headline safety figures actually show

The safety dataset behind ChondroFiller™ is unusually large for an injectable cartilage scaffold. Since 2013, more than 19,000 treatments have been performed globally — and across that entire case series, zero serious adverse device effects (SADEs) have been recorded. SADEs cover events such as organ damage, hospitalisation, or lasting injury directly attributable to the device.

The overall device complaint rate sits at approximately 0.06%, or roughly one in 1,500 procedures — a figure described as negligible in the manufacturer's Clinical Evaluation Report (CER Version 09, April 2025, Meidrix Biomedicals GmbH). The most common technical complaint is non-gelation: the liquid collagen failing to set properly during application. This accounts for around 0.01% of cases.

These numbers represent the largest real-world safety dataset available for any injectable collagen cartilage scaffold in routine orthopaedic use. They offer reasonable grounds for reassurance — but they carry an important qualifier. The 0% serious adverse event rate reflects outcomes in carefully selected patients with focal, isolated cartilage defects and healthy surrounding tissue. Patient selection is not a footnote to the safety story; it is central to it.

Where this data comes from — and how much weight it carries

The Clinical Evaluation Report Version 09, compiled by Meidrix Biomedicals GmbH and finalised in April 2025, is the single primary source for all safety figures in the dataset. It formalises data gathered through post-market clinical follow-up — the structured surveillance programme that CE-marked Class III medical devices are required to maintain once in routine use.

Post-market surveillance differs from a blinded randomised controlled trial in one important respect: it is observational rather than experimentally controlled. The data are real-world and drawn from actual clinical practice, which has genuine value — but the dataset is compiled by the manufacturer, not an independent body. That origin does not make the figures invalid. Regulatory bodies accept post-market surveillance as a legitimate evidence standard for Class III devices, and the dataset here spans more than 19,000 cases across over a decade. What it does mean is that the safety record is properly described as manufacturer-sponsored post-market cohort data rather than independently verified trial evidence.

The only available external corroboration comes from a 2025 prospective study by Matta et al., which followed 25 patients undergoing ChondroFiller treatment for cartilage defects at the wrist. It found no significant difference in complications between ChondroFiller-treated patients and matched controls — consistent with the headline surveillance figures. A 25-patient, single-joint cohort is a narrow base; it lends credibility to the broader dataset without being capable of validating it comprehensively. The honest characterisation of the overall evidence is manufacturer-sponsored post-market cohort data, corroborated by a small independent study — neither dismissed nor elevated beyond what the scale supports.

Which patients the safety record applies to — and who it doesn't

Patients considering ChondroFiller™ often want to know whether those headline figures apply to them. The answer depends on fit with a defined clinical profile.

The zero serious adverse event rate and negligible complaint rate reflect outcomes in patients with isolated, focal cartilage defects graded III or IV, covering up to 6 cm² of surface area, with structurally sound cartilage at the defect margins. Outside this profile — particularly in diffuse osteoarthritis or Kellgren-Lawrence Grade IV disease — the safety record does not extend. Those patient groups were not included in the dataset, and the figures cannot be applied to them.

Five absolute contraindication categories also place a patient outside the treated population:

• Known hypersensitivity to murine (rat or mouse-derived) Type I collagen
• Active infection within the target joint
• Active bleeding disorder, or anticoagulant therapy that cannot be paused
• Active malignancy or significant immunosuppression
• Pregnancy or breastfeeding

None of this means patients outside these criteria have no options; it means ChondroFiller's published safety evidence simply does not cover them. For patients who do fall within the indicated profile, the evidence is grounded in a large real-world dataset and consistent with independent prospective findings.

A pre-treatment assessment is the practical step that establishes which side of this boundary a patient sits on — and whether ChondroFiller represents an evidence-supported pathway for their specific defect.

Side effects most patients experience in the first few days

For most patients, the first two to three days after an injection are the most noticeable from a comfort standpoint. Localised swelling around the treated joint, a temporary flare in pain, and some stiffness are the most commonly reported effects — and all three are expected consequences of how the treatment works.

Once injected, the liquid collagen scaffold gels in situ and begins integrating with the surrounding tissue. This triggers a mild, localised inflammatory response as the body accommodates the new structure. That response is part of the integration process, not a sign that something has gone wrong, and the swelling and discomfort typically settle within 48–72 hours without specific intervention.

Beyond this early window, no pattern of systemic adverse effects has been recorded across the full surveillance dataset of more than 19,000 cases. True biological complications — as distinct from transient post-procedure discomfort — are rare when patients fall within the indicated clinical profile.

Objective confirmation of this comes from imaging data gathered in published clinical studies. MRI findings in treated patients show reduction in bone marrow oedema, diminished periarticular effusion, and visible joint space widening — markers suggesting the scaffold integrates without provoking a harmful tissue response. MOCART scores of 81.6 to 84.3 across European studies indicate more than 80% defect filling with good structural integration. These are research-level findings rather than routine monitoring for every patient, but they provide independent biological evidence that the early discomfort reflects normal adaptation rather than injury.

How complication rates compare with surgical alternatives

Published surveillance data and comparative analyses place ChondroFiller's risk figures in a meaningfully different range from the surgical alternatives used for focal cartilage defects.

Reoperation rates following the injectable scaffold treatment are reported at 3–8% in published series, against rates of up to 41% for microfracture and up to 37% for ACI/MACI. The complication rate for ChondroFiller approximates 0% within the indicated patient population; ACI/MACI carries a published complication rate of up to 17%.

These figures sit alongside an important procedural caveat. ChondroFiller is administered as an ultrasound-guided outpatient injection; microfracture and ACI/MACI are surgical procedures requiring incisions, theatre admission, and in most cases general or regional anaesthesia. Part of the more favourable complication profile for the injectable route reflects that procedural difference — the absence of surgical entry, wound healing, and perioperative risk — rather than the scaffold's biological properties alone. Reading the numbers as like-for-like comparisons would overstate what the data actually support.

The practical implication is one of pathway fit rather than simple ranking. For patients with isolated focal defects within the indicated size range, the outpatient injection route offers a substantially lower procedural burden alongside its favourable risk figures. For patients with larger or more complex disease — or whose anatomy places them outside the injectable scaffold indication — surgical pathways remain appropriate and clinically valid, and the comparison becomes less relevant to their decision.

What the evidence doesn't yet cover

Three years marks the current horizon for published follow-up data. The Jerosch et al. PMCF prospective study — the longest available in the literature — recorded an IKDC improvement of 32.4 points sustained at that point, a clinically meaningful result. What happens beyond three years remains uncharted; no published dataset currently extends there.

The evidence base also sits below randomised controlled trial level. No large-scale blinded RCT has been published — a limitation already noted in the discussion of data provenance earlier in this article — so rather than repeat that context, the more practically relevant gap is a narrower, specific one identified by independent research.

The 2025 Matta et al. prospective wrist study found that fibrous tissue formation occurred only in overfilled defects — cases where the scaffold material exceeded the depth of the cartilage defect. Flush applications, where the material sat level with the surrounding cartilage surface, were complication-free. That finding introduces a clinician-dependent variable into the safety picture: the risk is not inherent to the scaffold material itself, but to the precision of its delivery.

For patients thinking about where and by whom treatment is delivered, this distinction is the most actionable thing the current evidence offers. Accurate defect assessment before the procedure and experienced, image-guided technique at the point of injection appear to be the primary modifiable factors in avoiding the one technical complication the published literature has so far documented — making technique quality a genuine part of the safety conversation, not an afterthought.

1. [1] Controlled, randomized multicenter study to compare compatibility and safety of ChondroFiller liquid with microfracturing. (2016). https://doi.org/10.5348/VNP05-2016-1-OA-1 https://doi.org/10.5348/VNP05-2016-1-OA-1
2. [2] Cartilage reconstruction using Chondrofiller in intra-articular distal radius fractures (Matta et al., 2025). (2025). https://doi.org/10.1186/s42836-025-00333-y https://doi.org/10.1186/s42836-025-00333-y