KiOmedine^®CM-chitosan was obtained by controlled derivatization following the method previously described [11, 12]. A minimum of 100 g of chitosan, extracted from Agaricus bisporus, was dispersed in a reaction medium composed of isopropanol and sodium hydroxide. Monochloracetic acid was added to the chitosan suspension. The obtained carboxymethyl chitosan was subsequently reacted with acetic anhydride, and extensively purified by precipitation in ethanol. The precipitate was dried out to remove excess solvent and residual water, to yield the polymer. KiOmedine^®CM-chitosan was formulated at 1.8-2.2% (w:w) in phosphate-buffered water for injection at pH 7.2±0.2 supplemented with 3.5% sorbitol (w:w). The biomaterial was filled in a glass syringe and steam sterilized. All of the raw materials and excipients conformed to European Pharmacopoeia standards for injectable use.

This study was a first-in-human prospective, multicentre cohort with a 26-week follow-up period and was initiated in 2018. The study was single-blinded, meaning that the patient was blinded for the type of injected product. The study monitor (Factory CRO at Bilthoven, The Netherlands) ensured that this clinical investigation was conducted, recorded, and reported in accordance with good clinical practices and data protection regulations, specifically, the Clinical Investigation Plan; standard operating procedures and ISO 14155:2011; clinical investigation of medical devices for human subjects - Good clinical practice, and the requirements of the Declaration of Helsinki; or with applicable country-specific regulatory requirements - whichever afforded greater protection to the subject. Favourable opinion/approval was obtained from the National Competent Authority and Central Ethics Committees in the targeted countries. No patient was included before signing the consent form. Results are presented in line with the guideline from the International Committee of Medical Journal Editors (ICMJE) [13].

The eligibility criteria of study participants were defined in accordance with the current medical practice in viscosupplementation and were consistent with the population of patients who were enrolled in clinical trials in viscosupplementation. Patients were informed of the study design by the investigator, as per good clinical practice and ISO14155, before signing the consent form. Male and female patients were recruited who were suffering from symptomatic pain in the treatment knee for at least 26 weeks, and either were not responding or responding poorly to simple oral analgesics (non-opioid analgesics and non-steroidal anti-inflammatory drugs). Patients aged 40–85 years had a body mass index (BMI) ≤ 35 kg/m². In addition, patients had to be diagnosed with knee osteoarthritis according to the clinical and radiological criteria of the American College of Rheumatology [14] with a Kellgren & Lawrence grade II to III [15]; patients needed to have a pain score of 7-17 at baseline, when assessed with the 5-graded Likert WOMAC A after a mandatory 48-hour washout period of medication and minimal contralateral knee pain (i.e., WOMAC A ≤ 6).

Exclusion criteria were Radiological Kellgren & Lawrence grade 0, I or IV; exclusively patellofemoral osteoarthritis; chondromatosis or villonodular synovitis of the knee; clinically-apparent knee effusion, inflammation, or flare-up of the knee or abnormal synovial fluid macroscopy or significant volume upon arthrocentesis on the day of injection; history of injury or trauma in the treatment knee; significant clinically-assessed or radiographic varus or valgus deformation; inflammatory disease or pathologies interfering with the evaluation of osteoarthritis pain including homolateral coxarthrosis; hyaluronan injection, arthroscopy, or surgery in the treatment knee in the 26 weeks prior to injection; oral corticotherapy ≥5 mg/day, injection of corticosteroids or platelet-rich plasma or cell-based therapy in the 13 weeks prior to injection; change in the dosage regimen of symptomatic slow-acting drugs or in physiotherapy in the 13 weeks prior to injection; Anticipated need for any surgery or any forbidden OA treatments; history of recurrent infection or synovial infection; hypersensitivity or allergy to the product components of KIO014; history of symptomatic hip osteoarthritis; history of autoimmune disease; any severe, ongoing, and uncontrolled disease or any investigator-assessed clinically significant condition that may represent a substantial risk to the patient or impact study assessments; alcohol addiction; sever alteration of mobility; high risk of haemorrhage; participation in a therapeutic clinical trial in the last 13 weeks before injection; pregnancy or breastfeeding; being under guardianship or judicial protection.

This study was a first-in-human trial to investigate the safety and efficacy of the treatment with KiOmedine^®CM-Chitosan (KiOmed Pharma) in patients with symptomatic knee osteoarthritis. Durolane^® (Bioventus) was used as a control to support the reliability and scientific validity of the clinical evaluation procedure for the efficacy and safety of KiOmedine^®CM-Chitosan. Both Durolane^® and KiOmedine^®CM-Chitosan can be used with a single-injection of the 3-mL dose and can be injected with the same type of syringe.

The assessment schedule consisted of a screening visit, a baseline visit for injecting KiOmedine^®CM-Chitosan or Durolane^®, and four follow-up visits at 2, 6, 13, and 26 weeks post-injection. For each visit, patients were asked not to take any pain medication in the last 48 hours before the visit (i.e., 48-hour washout period). The unauthorized use of rescue medication and/or non-respect for the 48-hour washout were handled as protocol deviations in the analysis populations.

All injections and assessments were performed by qualified physicians with experience in intra-articular injections and clinical research in the relevant medical field. Accurate needle placement into the synovial joint was confirmed by slight aspiration of synovial fluid at the time of the intra-articular injection. Before intra-articular injection, the target knee was carefully examined. and synovial fluid was aspirated as needed.

The primary efficacy objective was to demonstrate that, compared to baseline, the intra-group effect size on pain is ≥ 0.8 at 13 weeks after a single injection of KiOmedine^®CM-Chitosan. An effect size of 0.8 on pain at 13 weeks post-injection is considered clinically large per Cohen’s threshold, is consistent with a clinically relevant effect on pain reported for viscosupplementation with hyaluronan [16], and is superior to the consensus intra-articular placebo effect of 0.6 [17]. In addition, the effect size is a common way of measuring the magnitude of the treatment effect in clinical studies on knee pain in patients with osteoarthritis [16, 18]. Secondary study outcomes for the different time points (baseline and 2-, 6-, 13-, and 26-weeks post-injection) included self-reported pain, self-reported functional health and well-being, responder rate, patients’ satisfaction, physician’s satisfaction from a medical and usability point-of-view, and adverse events.

Clinical assessments included a Self-administered WOMAC questionnaire for pain, stiffness and function [19-21]; a Self-administered questionnaire for pain and global assessment (scale of 0 to 10) [22]; Osteoarthritis Research Society International (OARSI) Standing Committee for Clinical Trials Response Criteria Initiative and the Outcome Measures in Rheumatology (OMERACT)” responder criteria for osteoarthritis clinical trials [22] to evaluate response to treatment; Self-administered SF-12 health survey of functional health and well-being (scale of 0 to 100) [23]; and five-point Likert scales for the usability of the investigational device. Consistent with the primary efficacy endpoint used in clinical studies reported for Durolane^® [24, 25], the responder rate was defined as a reduction in the WOMAC pain score of at least 40% with an absolute improvement of at least five points compared with baseline for the treatment knee at the different visits [23]. Pre-injection pain (baseline WOMAC score) forms the basis for showing efficacy as intended, meaning that the patient is serving as his or her own control. For OMERACT-OARSI responder criteria, an improvement in pain or in function ≥ 50% and an absolute change ≥ 20 were considered as a positive response to treatment.

Of all adverse events, the investigator recorded the incidence, severity (i.e., mild, moderate, severe), seriousness, and the relationship to the study treatment or device. Safety of the investigational medical device was evaluated at the different time points by recording the incidence, severity, and causal relationship of (serious) adverse events, (serious) adverse device effects, unanticipated serious adverse device effects and device deficiencies. Local effects such as joint pain (arthralgia), joint effusion, joint swelling, joint warmth, injection site pain, and joint stiffness are anticipated with viscosupplementation in the treatment knee. The severity and occurrence of each local effect were assessed using a four-point numerical rating scale (none, mild, moderate, serious).

Interim analysis was done after 30 patients who received a single injection of KiOmedine^®CM-Chitosan completed Visit 4 (3 months post-injection).

Trial was designed to be terminated by the Sponsor for any of the following reasons: protocol deviations, device deficiency or malfunction, safety concern, production limitation, administrative decisions. Recommendation for termination or modification of the trial for excessive adverse events was at the discretion of the Data Monitoring Committee.

Randomization in 2:1 allocation to the KiOmedine^®CM-Chitosan and Durolane^® groups, respectively, was achieved via a computer-generated program, which also stratified qualified subjects by study sites. Sealed envelopes with sequential numbers, containing randomization codes were provided to the study sites and were used in sequential order. Patients were not informed of their randomization assignment until the end of the study.

Each patient was blinded for the type of treatment at the injection visit (single blinding). Both Durolane^® and KiOmedine^®CM-Chitosan can be injected with the same type of syringe; This allows blind delivery of the product to the patient. KiOmedine^®CM-Chitosan and Durolane^® were supplied in a sterile and ready-to-use 3-mL syringe in identical packages labelled with a unique identification code. Breaking the single blinding only occurred in the case of major device deficiency or safety concerns (repeated SADEs) upon Sponsor’s decision, and/or in any circumstance that was deemed necessary for the safeguard of those patients.

Durolane^® was used as a control treatment to validate the clinical evaluation methods used in this trial. The outcomes of this control group were not intended for comparing the two treatments. The primary analysis population for efficacy and safety was the ‘Safety Population’, which was constituted of all patients who received any randomized study treatment with at least one follow-up visit. Missing values were not replaced nor extrapolated. All statistical analyses were performed using IBM SPSS Statistics (Version 21.0).

Descriptive statistical analyses were performed using mean, median, standard deviation (SD), minimum and maximum values for continuous variables, and using sample size (n), absolute (n) and relative incidence (%) for discrete variables. Inferential statistical analyses were performed with a significance level of 5% and 90% power or using non-parametric rank tests. Where appropriate, Bonferroni’s correction for multiplicity was applied to ensure the trial was conclusive for several efficacy parameters and to provide a stringent control of Type I errors (i.e., minimizing the chance of a false positive discovery).

For the primary efficacy endpoint, any change in pre-injection baseline pain at 13 weeks was evaluated using paired Student’s t-test. The intra-group effect size was calculated as the difference between mean WOMAC A (pain) at baseline and mean WOMAC A at 13 weeks divided by the pooled standard deviation from the two data sets, assuming a normal distribution of data. The over-time modification of the efficacy endpoints was assessed versus baseline by analyses of variance for repeated measures, followed by appropriate post-hoc tests, for continuous variables with a normal distribution. For discrete variables and for continuous variables for which the assumption of normality was not fulfilled, Friedman’s tests were performed, followed when significant by 2 by 2 Wilcoxon’s tests.

For safety analysis, all recorded adverse events were coded using version 21.0 of the MedDRA and were summarized by presenting the number and percentage of patients showing any sign or symptom. Post-injection adverse events were analysed in three categories based on the causality assessment: treatment-emergent adverse events irrespective of their causality with the procedure and the device; treatment-related adverse events with a possible, probable, or causal relationship with the procedure and/or the device; and adverse device effects with a possible, probable, or causal relationship with the device only. Safety analysis was subject to the independent Data Monitoring Committee’s (DMC) review scheduled per protocol. DMC is composed of one pharmacologist/ biostatistician, one orthopaedist and one general physician.

Ten (10) patients were screen failures (i.e., patient who signed the informed consent but did not meet the eligibility criteria at either the screening or injection visit). A total of 95 patients were enrolled, from 11 March 2019 to 03 June 2019, and randomly assigned to the study treatments, of which 63 patients were injected once with KiOmedine^®CM-Chitosan and 32 patients were injected once with Durolane^® (Fig. 1). The randomization ratio between KiOmedine^®CM-Chitosan and Durolane^® was 1.97 instead of 2.00. In the KiOmedine^®CM-Chitosan group, 23.8% (n=15) of patients were recruited in four centers in the Netherlands and 76.2% (n=48) in three centers in Hungary. In the Durolane^® group, 25% (n=8) of patients were recruited in the Netherlands and 75% (n=24) in Hungary. There were no protocol deviations, thus all randomized patients were included in the safety cohort for analysis. Results from the per-protocol cohort, constituting of all patients who adhered to the protocol and attended all visits up to 13 weeks post-injection, are provided in supplemental materials. The per-protocol cohort includes 51 patients in the KiOmedine^®CM-Chitosan group and 28 patients in the Durolane^® group.

There were no significant differences in demographics and baseline characteristics between the KiOmedine^®CM-Chitosan and Durolane^® group, except that more males were injected in the Durolane^® group (Fisher’s exact test, p = 0.040;) (Table 1). This difference was not considered clinically significant for the purpose of this study.

Fig. (1). CONSORT flow diagram of patients throughout the study.

The intra-articular position of the injection was anterolateral, lateral mid patellar, or lateral suprapatellar. Of note, the skin at the injection site was considered in good condition in all patients, except for one patient with poor condition of the local skin. However, the injection site was acceptable for intra-articular injection.

In the KiOmedine^®CM-Chitosan group, WOMAC A (pain) score at 13 weeks post-injection decreased by 62.5% from baseline (11.2 ±2.5 at baseline vs. 4.2 ±4.0 at 13 weeks post-injection; paired Student’s t test, p < 0.001). The effect size equalled 2.08. In the Durolane^® group, there was a 62.4% reduction in WOMAC A (pain) score at 13 weeks post-injection compared to baseline (11.7 ±2.4 at baseline vs. 4.4 ±3.8 at 13 weeks post-injection; paired Student’s t test, p < 0.001). The effects size equalled 2.28.

The overall incidence of adverse events was comparable between KiOmedine^®CM-Chitosan (n = 27, 42.9%) and Durolane^® (n = 13, 40.6%) (Table 2). There was one serious adverse event (1.6%). This SAE consisted of biliary colic, which was not related to the device and was resolved with hospitalization.

In the KiOmedine^®CM-Chitosan group, 23 patients (36.5%) experienced at least one treatment-emergent adverse event occurring in the treatment knee, 19 patients (30.2%) experienced a treatment-related adverse event, and nine patients (14.3%) experienced adverse device effect. Their severity was considered mild to moderate in most cases. Meanwhile, one adverse device effect was considered severe in one patient. No treatment-related adverse event or adverse device effect led to study withdrawal. The most commonly reported treatment-related adverse event was arthralgia (25.4%) in the treatment knee. Other treatment-related adverse events were joint effusion, joint swelling, or synovitis. A concomitant medication was prescribed in 85.7% of cases, and a concurrent procedure consisting of synovial fluid aspiration was performed in two cases. The most relevant adverse events occurred within one day after injection (mean: 0.2 day; median: 0 day) and it was resolved within 2-3 weeks (mean: 16.5 days; median: 3 days). No long-term onset adverse events were recorded.