| Samenvatting: | <p><em>Background</em>: Knee osteoarthritis (OA) is one of the leading musculoskeletal burdens in the world, causing both structural damage and pain to the joint. Radiographs are the most common imaging method for diagnosing OA, however the relationship between radiographic changes (ROA) and symptoms is not well understood. This thesis will establish the natural history of twenty-year ROA, compare diagnostic methods of ROA assessment, evaluate the cross-sectional relationship between ROA and pain, and will determine the long-term predictive validity of features of ROA with future knee replacements (TKRs).</p> <p><em>Methods</em>: Data from the Chingford women's study, a twenty-year prospective UK-based cohort was used for the analysis. Risk factors included atlas-based ROA scoring methods and quantitative joint space width (JSW), which were analysed against pain and TKR outcome measures. A novel method for assessing joint space on low-contrast x-rays was developed which had high reproducibility and validity.</p> <p><em>Results</em>: The twenty-year natural history of ROA showed relatively low levels of annual incidence (3.8%), progression (3.6%), and worsening (4.5%), and emphasised the involvement of both medial and lateral osteophytes. Severe joint space narrowing (JSN) had the best construct validity with pain, while any size of lateral tibial osteophyte had good construct validity. Medial quantitative JSW had good construct validity, but no predictive validity with future TKRs. Lateral JSN and osteophytes had the best predictive validity for future TKR.</p> <p><em>Conclusion</em>: This research demonstrates that radiographic scoring methods have strong construct and predictive validity with symptomatic knee OA. These results support the use of x-rays to identify early disease changes which indicate future joint failure.</p>
|
|---|