File
Carlesso L1,2, Desmeules F1, Vendittoli P-A3, Choinière M4, Fernandes J5, Arendt-Nielsen L6
1Université de Montréal, School of Rehabilitation, Montréal, Canada, 2McMaster University, School of Rehabilitation, Hamilton, Canada, 3Université de Montréal, Surgery, Montreal, Canada, 4Université de Montréal, Anesthesiology, Montréal, Canada, 5Université de Montréal, Surgery, Montréal, Canada, 6Aalborg University, Department of Health Science and Technology, Aalborg, Denmark
Background: Pain in people living with knee osteoarthritis (KOA) is known to be complex. Systematic reviews have identified significant heterogeneity in studies attempting to identify KOA pain phenotypes (PP), but they have not accounted for the multidimensional experience of pain. A standardized approach to PP assessment will allow for greater generalizability and comparison of results.
Purpose:
1) To identify PP in patients with KOA using the domains recommended by IMMPACT);
2) To identify risk factors for PP, and
3) To assess the relationship of the PPs to self-reported function and physical performance.
Methods: Participants ≥ 40 years of age having an initial consult with orthopaedic surgeons who confirmed a diagnosis of KOA were recruited from three hospitals in Montreal (Canada). Principal component analysis was used for variable reduction prior to latent profile analysis to determine PPs using the IMMPACT recommended domains:
1) pain intensity - numeric pain rating scale (NPRS);
2) pain variability reported three times a day for one week--i.e., standard deviation of the NPRS,
3) pain qualities - Short Form McGill Questionnaire 2,
4) sleep quality - Pittsburgh Sleep Quality Index,
5) anxiety/depressive symptoms - Hospital Anxiety and Depression Scale,
6) tendency to catastrophize - Pain Catastrophizing Scale,
7) somatization - Patient Health Questionnaire 15,
8) fatigue - Multidimensional Fatigue Inventory,
9) neuropathic pain - Modified Pain Detect Questionnaire, and
10) neurophysiological functioning - pressure pain thresholds (PPT), temporal summation (TS), cold and heat pain thresholds (CPT, HPT) and conditioned pain modulation (CPM).
Risk factors for the PPs which included age, sex, race, BMI, marital status, education level, income, comorbidities, optimism, chronic pain self-efficacy and Kellgren and Lawrence grade were assessed using multinomial logistic regression analyses. Linear regression was used to determine the relation of the PPs to disease specific function using the Activities of Daily Living subscale of the Knee Injury and Osteoarthritis Outcome Score questionnaire. Physical performance tests included 30 second chair stands, 40-meter fast paced walk and stair climbing.
Results: 141 participants were included (mean age 63.1 (10.2); BMI 32.4 kg/m2 (9.4), 59% women). A 3-class model was chosen based on fit criteria and current knowledge. The PPs demonstrated progressive severity of psychosocial and pain-related variables. Quantitative sensory testing values of those in PP3 were significantly different from those in PP1, having higher TS and CPTs at the forearm and lower PPTs at the patella. Risk factors for membership in PP2 and 3 compared to PP1 included age (OR 0.93 (0.88, 0.99)) chronic pain self-efficacy (OR 0.91, (0.86, 0.96)) and optimism (OR 0.73 (0.58, 0.91). In an adjusted model, participants in PP2 and 3 reported significantly worse self-reported function (ß -38.0 95%CI (-49.8, -26.2)), performed more poorly on the stair climb test (ß 8.5 95%CI (3.1, 13.9)) and the 40m walk (ß 10.8 (3.9, 17.7) compared to the PP1. No significant differences were found for chair stands.
Conclusion(s): Three PPs of increasing severity were found with the more severe having poorer functional outcomes.
Implications: Standardized PPs may be useful in determining prognosis or for stratification of treatment .
Keywords: Pain, phenotyping, knee osteoarthritis
Funding acknowledgements: Quebec Rehabilitation Research Network and the International Association for the Study of Pain.
Purpose:
1) To identify PP in patients with KOA using the domains recommended by IMMPACT);
2) To identify risk factors for PP, and
3) To assess the relationship of the PPs to self-reported function and physical performance.
Methods: Participants ≥ 40 years of age having an initial consult with orthopaedic surgeons who confirmed a diagnosis of KOA were recruited from three hospitals in Montreal (Canada). Principal component analysis was used for variable reduction prior to latent profile analysis to determine PPs using the IMMPACT recommended domains:
1) pain intensity - numeric pain rating scale (NPRS);
2) pain variability reported three times a day for one week--i.e., standard deviation of the NPRS,
3) pain qualities - Short Form McGill Questionnaire 2,
4) sleep quality - Pittsburgh Sleep Quality Index,
5) anxiety/depressive symptoms - Hospital Anxiety and Depression Scale,
6) tendency to catastrophize - Pain Catastrophizing Scale,
7) somatization - Patient Health Questionnaire 15,
8) fatigue - Multidimensional Fatigue Inventory,
9) neuropathic pain - Modified Pain Detect Questionnaire, and
10) neurophysiological functioning - pressure pain thresholds (PPT), temporal summation (TS), cold and heat pain thresholds (CPT, HPT) and conditioned pain modulation (CPM).
Risk factors for the PPs which included age, sex, race, BMI, marital status, education level, income, comorbidities, optimism, chronic pain self-efficacy and Kellgren and Lawrence grade were assessed using multinomial logistic regression analyses. Linear regression was used to determine the relation of the PPs to disease specific function using the Activities of Daily Living subscale of the Knee Injury and Osteoarthritis Outcome Score questionnaire. Physical performance tests included 30 second chair stands, 40-meter fast paced walk and stair climbing.
Results: 141 participants were included (mean age 63.1 (10.2); BMI 32.4 kg/m2 (9.4), 59% women). A 3-class model was chosen based on fit criteria and current knowledge. The PPs demonstrated progressive severity of psychosocial and pain-related variables. Quantitative sensory testing values of those in PP3 were significantly different from those in PP1, having higher TS and CPTs at the forearm and lower PPTs at the patella. Risk factors for membership in PP2 and 3 compared to PP1 included age (OR 0.93 (0.88, 0.99)) chronic pain self-efficacy (OR 0.91, (0.86, 0.96)) and optimism (OR 0.73 (0.58, 0.91). In an adjusted model, participants in PP2 and 3 reported significantly worse self-reported function (ß -38.0 95%CI (-49.8, -26.2)), performed more poorly on the stair climb test (ß 8.5 95%CI (3.1, 13.9)) and the 40m walk (ß 10.8 (3.9, 17.7) compared to the PP1. No significant differences were found for chair stands.
Conclusion(s): Three PPs of increasing severity were found with the more severe having poorer functional outcomes.
Implications: Standardized PPs may be useful in determining prognosis or for stratification of treatment .
Keywords: Pain, phenotyping, knee osteoarthritis
Funding acknowledgements: Quebec Rehabilitation Research Network and the International Association for the Study of Pain.
Topic: Pain & pain management; Rheumatology; Musculoskeletal: lower limb
Ethics approval required: Yes
Institution: Hôpital Maisonneuve Rosemont
Ethics committee: CIUSSS de l’Est-de-l’Île-de-Montréal
Ethics number: MP-12-2017-829
All authors, affiliations and abstracts have been published as submitted.
