Validade dos critérios de classificação da osteoartrite de joelho: ELSA-Brasil Musculoesquelético (ELSA-Brasil ME)

Abstract

Introduction: the classification criteria for knee osteoarthritis (OA) available in the literature are usually divided into three main groups; clinical, clinical-radiographic and radiographic. The existence of different knee OA classification criteria makes it difficult to standardize and compare results in epidemiological studies, and also to investigate risk factors and clinical manifestations associated with knee OA. It is essential to understand the extent to which these different criteria are able to discriminate between subjects with and without OA. It is also important to know the strength of association of the following items with the different criteria: risk factors, the presence of pain and the functionality Objective: To evaluate the validity of different knee osteoarthritis (OA) classification criteria: radiographic OA, symptomatic OA, clinical and clinical-radiographic criteria of the American College of Rheumatology (ACR) and the OA definition proposed by the National Institute for Health and Care Excellence (NICE). Method: A cross-sectional study of the validity of different knee OA classification criteria, radiographic OA, symptomatic OA, clinical and clinical-radiographic criteria of the American College of Rheumatology (ACR) and the OA definition proposed by the National Institute for Health and Care Excellence (NICE). The subjects were from the ELSA-Brasil Musculoskeletal cohort (ELSA-Brasil MSK), an ancillary investigation on musculoskeletal disorders of the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil). The ELSA- Brasil MSK follows 2901 active and retired civil servants at one of the six ELSA-Brasil investigations centers, located in the State of Minas Gerais, Brazil. The present study was carried out using a convenience subsample with 250 male and female subjects aged between 39 and 78 selected from ELSA-Brasil MSK. The assessment was made by a rheumatologist (RCCM) between February 2014 and June 2015. Only one knee per subject was included i.e., the one that had OA according to the rheumatologist and, when both or none of the knees were affected, one knee was randomly selected for analysis. The performance of the classification criteria of knee OA described before was evaluated, using as reference-standard the clinicalradiographic evaluation of the rheumatologist. OA prevalence, sensitivity, specificity, positive and negative predictive values and accuracy were presented (CI 95%; α = 5%). After that the construct validity of the criteria was evaluated by the relationship between the presence of knee OA identified by the criteria and the following explanatory variables: body mass index (BMI), pain and the functional limitation. The presence of current pain in knees and the subjective functional limitation were evaluated using Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) and the objective functional limitation was evaluated by the five-times sitto-stand test (FTSTS). Associations were tested by bivariate and multinomial logistic regression models (CI 95% ;α=5%). Mean age was 56.1 years (SD=8.7); 51.2% were male. The highest knee OA prevalence was observed in OA according to the rheumatologist, followed by the NICE definition (36,8%) and radiographic OA (22.0%).The sensitivity and the specificity of radiographic OA were 51% and 96.7%, respectively, while the NICE definition showed 57.0% and 76.3%, respectively. The other OA criteria showed good levels of specificity, but the levels of sensitivity were below 30%. In the convergent construct validity evaluation, BMI, current pain and subjective functional limitation were associated with knee OA identified by all criteria. The magnitude of the estimates were particularly strong between current pain and OA according to the ACR clinical criteria (OR 21.7; 95%CI 7.12-66.12) and between subjective functional limitation and OA according to NICE (OR 32.5; 95%CI 13.4- 79.0). The objective functional limitation was associated with symptomatic OA and with clinical and clinical-radiographic ACR criteria. Conclusions: the present study demonstrated that among the evaluated criteria, radiographic OA showed the best performance, followed by the NICE OA definition. The other criteria, namely symptomatic OA, clinical ACR and clinical-radiographic ACR were not adequate for studies that aim to evaluate prevalence, due to the low sensitivity of these criteria. However, they may be an alternative in longitudinal studies in which it is appropriate to use criteria with good specificity. Regarding the convergent construct validity, overall, our results offer support for all of the investigated criteria, since there was an association of BMI, pain and subjective functional limitation with all the criteria and the objective functional limitation was not radiographic OA and NICE. It is important to emphasize that the choice of criteria for a study requires deciding which the combination best meets the study objectives.