Développement d'une méthode pour évaluer l'impact de la suppression de la douleur sur l'activité musculaire chez des personnes arthrosiques : résultats préliminaires

Abstract

Introduction The main problem with the analysis of electromyography (EMG) signals in patients with knee osteoarthritis (OA) is the difficulty to detect muscle activation because of increased noise. There is no golden standard for analysing this type of EMG signals, and there is no study regarding how pain affects the signals. The objective of this preliminary study is to develop a method for detection of EMG signals that is sensitive enough to evaluate how suppressing pain influences muscular activity in knee OA patients. Methods The first part of this master’s thesis describes the development of a semi-automatic method to analyse OA patients’ EMG signals. The second part focuses on the results on a group of nine OA subjects. In this part, main inclusion criterion were the existence of medial knee OA with or without patellofemoral OA, and pain intensity greater than or equal to 3 on a Numeric Rating Scale (NRS) from 0 to 10. Gait trials were performed at a self-selected speed before and after pain suppression induced by an intra-articular injection of 5 ml of lidocaine (1%). EMG signals were collected on four muscles: semitendinous (ST), biceps femoris (BF), vastus lateralis (VL) and vastus medialis (VM). The raw EMG signals were assessed visually to detect potential outliers. Conditioning of the signals was performed in the following order: 1) band-pass filtering at 30-300 Hz 1, 2) rectification, 3) use of the Teager-Kaiser Energy (TKE) method known to increase the detection accuracy of muscle activation, and 4) low-pass filtering at 50 Hz and localized 6 Hz. A second visual verification was performed again to delete other potential outliers. Detection of the activations was based on the threshold method. The factor used for the threshold was determined based on the signal-to-noise ratio and was not constant over the gait cycle. Results Repeatability analysis intra-operators made on five among nine subjects who willingly participated in the study showed that the method was repeatable for both examiners who had different degrees of experience. Inter-operators reproducibility made on these five same subjects confirmed the need of visual inspection by an examiner to detect outliers, however the examiner’s experience did not have an influence on the reproducibility results. Among the nine patients who received the anesthetic injection, 28 muscles were analysed, after suppression of unusable signals. The semi-automatic method for EMG signals enabled detecting muscles activations in all patients. After suppressing pain, most of EMG signals showed modifications in muscular activations as well as a tendency towards lower intensity. Some secondary activations noticeable between 40% and 80% of the walking cycle were reduced in intensity; their frequency of occurrence was lowered and even sometimes zero. Discussion These preliminary results showed that the proposed method is repeatable and reproducible for all the analysed muscles, and allows EMG signals analysis in OA patients with minimal involvement of the examiner. Results also suggest that pain suppression has an impact on the intensity and variability of muscular activations during walking gait. This may imply that a pain protection mechanism in knee OA exists, and leads to higher contraction of muscles. Suppressing pain would therefore drive activation patterns of the knee OA patients’ muscles to evolve towards a ‘’normal’’ pattern. These results open the way for a larger study to ensure this hypothesis, and furthermore to maybe identify how pain suppression impacts on articular biomechanics.

Date	20 Aug 2018
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Nicola Hagemeister (Supervisor) & Manon Choinière (Co-supervisor)