Portal de Periódicos da CAPES

Artigo Acesso aberto Produção Nacional Revisado por pares

BIBTEX RIS

Investigating osteoarthritis in a subcohort of the Brazilian Longitudinal Study of Adult Health: The ELSA-Brasil Musculoskeletal Study (ELSA-Brasil MSK)

2016; Elsevier BV; Volume: 24; Linguagem: Inglês

10.1016/j.joca.2016.01.408

1522-9653

Rosa Weiss Telles, Larissa Cristina Moraes Silva, Luciana A. C. Machado, Sandhi Maria Barreto,

Musculoskeletal pain and rehabilitation

Purpose: To describe the measures and procedures used for the investigation of osteoarthritis (OA) in the ELSA-Brasil Musculoskeletal Study (ELSA-Brasil MSK) and to report preliminary findings on the quality of radiographic readings for tibiofemoral OA. Methods: Civil servants pertaining to a subcohort of the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil) have undergone additional assessments for the identification of OA and its potential risk factors through the ancillary study ELSA-Brasil MSK. At baseline (2011–14), OA was identified according to three different definitions: doctor-diagnosed OA (any joint), and radiographic and symptomatic OA (in hand and knee joints). Doctor-diagnosed OA was ascertained through participants' answer to the question "Has a doctor ever told you that you have osteoarthritis or arthrosis?". Radiographic OA was ascertained through digital radiography (PA left and right hand views, fixed-flexion knee view and left and right lateral knee views). Symptomatic OA was determined by the combination of frequent symptoms (i.e. positive answer to the question "Have you had pain, aching, or stiffness in the hand/knee on most days for at least 1 month in the past 12 months?") and evidence of radiographic OA in the respective joint. OA in hand joint groups and in the tibiofemoral (TF) joint were assessed according to Kellgren-Lawrence (K-L) scores, with scores ≥ 2 defining radiographic OA. Patellofemoral (PF) OA was assessed according to PF osteophytes (>2) and PF joint space narrowing (>2) accompanied by bony changes (osteophyte = 1, sclerosis, cysts). Readings of radiographic knee images are currently under way. These have followed a pre-defined sequence of standardized procedures, which included an initial screening by two trained radiologic technologists, confirmation of K-L scores by an experienced Radiologist (observer A), and if necessary, adjudication sessions with a Rheumatologist (observer B) and a highly experienced academically-based musculoskeletal radiologist (external observer). To ensure the quality of radiographic readings in ELSA-Brasil MSK and data comparability with other cohorts, face-to-face training and calibration sessions were conducted with the external observer before readings were initiated. Additionally, observer calibration has been continuously evaluated throughout the study by analyses on agreement of tibiofemoral K-L scores provided by the three observers in the fixed-flexion knee view, with the external observer as reference for readings. A total of 300 (∼ 5.0%) images will be used to evaluate observer calibration, which will be divided into analyses of batches of 100 images selected through a stratified random sample enriched with knee OA. Agreement will be calculated by kappa for K-L ≥ 2 and weighted kappa for K-L scores (0–4). The kappa 95% percentile confidence intervals were computed by nonparametric bootstrap method. Results: 2901 participants (52.9% women) with mean (SD) age of 55.9 (8.9) years were included at baseline of ELSA-Brasil MSK. Of these, 2830 (97.6%) have undergone at least one radiographic view. The prevalence of doctor-diagnosed OA was 18.0%. Frequent symptoms in the hand and knee were reported by 15.4% and 23.7% of the participants, respectively. By the date of submission of this abstract, initial screening of radiographic knee images had been completed in > 1700 participants. The analysis of the first batch of knee images for the evaluation of observer calibration has indicated adequate agreement between each pair of readers (Table 1).Tabled 1Table 1. Agreement for Kellgren-Lawrence (K-L) classification in 97 PA fixed-flexion knee radiographs.Weighted Kappa (95% CI)Kappa (95% CI)K-L score (0–4)K-L ≥ 2Observer A vs. B0.810 (0.721–0.885)0.773 (0.647–0.895)External observer vs. observer A0.867 (0.807–0.916)0.752 (0.617–0.874)External observer vs. observer B0.804 (0.715–0.866)0.771 (0.642–0.877)PA, posterior-anterior; CI, confidence interval.Three knee images were excluded from the analysis (01 total knee replacement and 02 images without side markers). Open table in a new tab PA, posterior-anterior; CI, confidence interval. Three knee images were excluded from the analysis (01 total knee replacement and 02 images without side markers). Conclusions: ELSA-Brasil MSK has continuously monitored a large sample of Brazilian adults for the investigation of multiple musculoskeletal disorders, including OA, and potential risk factors. Procedures used for the investigation of OA have followed widely accepted protocols and have demonstrated to comply with quality requirements. Data retrieved from subsequent cohort waves will contribute a unique repository for life course epidemiologic research on hand and knee OA in a non-high income country. Funding: Brazilian Ministry of Health (Science and Technology Department), of Science and Technology (FINEP, CNPq), and of Education (CAPES), SECTES, SEBRAE-MG, and UFMG.