The study was a prospective, open-label, single-centre trial with a six-month follow-up.
All of the patients were recruited during a multidisciplinary chronic pain consultation at the Lannion-Trestel Hospital Centre, France. Patients with symptomatic TMJ OA that is not sufficiently relieved by usual first-line treatments and who had undergone a radiography showing evidence of OA (joint space narrowing and/or osteophyte), and for whom the indication of viscosupplementation had been approved by the rheumatologist or dental surgeon, could be included in the trial. Patients with other TMJ involvement (internal derangement related to disc displacement, rheumatoid arthritis and other inflammatory TMJ involvement), those who received viscosupplementation in the target joint within the last six months, or corticosteroids during the previous three months, cannot be included in the trial. Patients with contraindications to viscosupplementation (i.e. hypersensitivity to HA or mannitol) or to the IA injection procedure (infectious disease in progress, infected skin lesions next to the injection area, clotting disease, etc.) and those in whom the treatment cannot be effectively assessed, who did not speak French or who were unable to give their informed consent were not able to be included in the trial.
All patients received a single IA injection of 1 ml of HANOX-M-XL (HappyMini®, LABRHA Laboratory, Lyon, France) in the first TMJ. HANOX-M-XL is a HA viscosupplement specifically designed for small joints, made up of crosslinked high molecular weight HA (16 mg/ml) combined with 35 mg/ml of mannitol, that delays the in situ degradation of HA. These specificities allow HANOX-M-XL to be injected using a single injection regimen. Injections were performed by one single very experienced physician as per the following procedure.
The patient is positioned in the lateral decubitus position, on the healthy side. After careful disinfection with povidone, an iodinated derivative (Betadine dermique® 10%, Mylan Medical SAS, France) is applied and, if necessary, the pre-auricular zone is shaved, 1 cm below and in front of the tragus, so as to properly identify the condylar process. Then, the patient is asked to open their mouth slowly by about 2 centimetres. Under fluoroscopic control, a 25 gauge/25 mm needle is introduced in the temporomandibular joint. In the event of synovial effusion, the latter is removed before injecting HA. If no synovial fluid can be removed, 0.1 ml to 0.5 ml of ioxaglic acid, 320 mg iodine/ml (Hexabrix® 320 mg, Laboratoire GUERBET, France) is injected, allowing the visualization of a linear image of about 1 cm, which confirms the accurate positioning of the needle. The contrast agent is then removed, and 1 ml of HANOX-M-XL is slowly injected. Again, the patient is asked to open and close their mouth several times. Lastly, a dressing is placed on the injection point.
Outpatients referred to multidisciplinary orofacial pain consultations at Lannion Hospital Centre, with radiographic evidence of TMJ OA on a dental panoramic X-ray and/or Computed Tomography (CT) or Magnetic Resonance Imaging (MRI) procedure performed within the previous 12 months, were recruited by a dental surgeon algologist or a rheumatologist if experiencing the following symptoms: unilateral or bilateral TMJ pain, at rest or while chewing, impairment of jaw movements on maximum voluntary effort and/or assisted opening of the jaw.
In the present study, attempts to identify outcome predictors were focused on features that any healthcare centre can easily record on their clinical record forms (sex, age, pain levels while chewing and on TMJ palpation, pain duration, unilateral or bilateral injection, NSAIDs and analgesic use).
Because the patient’s judgement is the most used outcome measure in daily clinical practice, patient satisfaction was selected as the primary criteria, from which the predictive factors were assessed. The level of patient satisfaction was rated ‘yes’ (very satisfied/satisfied) or ‘no’ (hardly satisfied/not satisfied).
Response to treatment was also evaluated with two other tools: patient assessment of effectiveness (very effective/effective = yes, hardly effective/not effective = no) and a decrease in pain of at least 50% on the numeric scale.
Course of the study
Before enrolment, patients were required to give their informed consent to participate. During the selection visit on day 0 (D0), the following data were collected: demographics (sex, age, height, weight), duration of pain, target TMJ, previous and current treatments for the TMJ OA (NSAIDs, analgesics, specific rehabilitation of the buccal sphere, wearing an occlusal splint, previous IA corticosteroid injection, previous IA–HA injections, etc.), test for bruxism, and noises and/or crackles during chewing. TMJ pain while chewing and on palpation was assessed using a numeric scale (NS): 0–10 (0 = no pain, 10 = maximum pain). The Maximum Inter-Incisal Opening Distance (MIIOD) was measured with a vernier calliper in millimetres (mm) as the vertical distance between the maxillary and the mandibular central incisor edges.
At the end of the selection visit, the IA injection of HANOX-M-XL was performed or scheduled, mostly for organizational reasons, within a maximum of 15 days, in order to limit the risk of variation in clinical state between the evaluation and the day of injection.
- Follow-up visits at three and six months
During the visits in month 3 (D90) and month 6 (D180) following the injection, the following data were collected: TMJ pain while chewing (NS 0-10), TMJ pain on palpation (NS 0-10), MIIOD (mm), patient satisfaction with the treatment (Likert scale comprising four points, where 0 means ‘not satisfied’, 1 ‘hardly satisfied’, 2 ‘satisfied’ and 3 ‘very satisfied’), patient perception of the effectiveness of the treatment (Likert scale comprising four points, from 0, ‘not effective’, to 3 ‘very effective’), patient perception of pain reduction across six categories ([0%], [1-24%], [25-49%], [50-74%], [75%-99%], [100%]), and variation in consumption of analgesics or NSAIDs compared to the baseline ([yes/no] and if [yes], the quantification of such on a five-point Likert scale ([1-24%], [25-49 %], [50-74%], [75%-99%], [100%]). During each assessment visit, all of the adverse events were recorded.
Baseline and six-month follow-up data are provided as a number, percentage or mean value [CI 95%]. We looked for the predictive factors of response or non-response, first in univariate then in multivariate analysis, by integrating all of the factors with a p-value <0.2 in the univariate analysis, as well as possible confounding factors such as patient age, sex, NSAIDs intake and if they wear an occlusal splint.
The main analysis was intended to be carried out in the per protocol (PP) population using only patients for whom all of the data on day 0 and day 180 were available, then confirmed in the intend-to-treat (ITT) population by processing missing data with the Last Observation Carried Forward (LOCF) method. Mann-Whitney or chi-square tests were used, as appropriate, to assess the association of quantitative or qualitative factors with patient satisfaction. The regression coefficients of the multivariate models (ANCOVA and mixed model) were considered significant if they were less than 5%. Statistics were achieved using XLstats software© (Addinsoft, Paris, France).