Background: Knee pain, the main symptom of knee osteoarthritis (OA), affects one in 4 people aged over 55 years, of whom 10% have mild to moderate disability. The aetiology of knee pain is heterogeneous and its relationship with structural changes and function is unclear. An important role for synovial pathology in the initiation and progression of knee OA has been emphasised. However, the normal values of synovial changes detected on ultrasound (US) in the general population and their association with knee pain in community-based people with knee pain or OA remain largely unknown. Objectives: [1] to systematically review the literature on synovial changes detected on US in people with knee pain/OA and/or in the general population. [2] to establish the normal ranges for synovial thickness and effusion and determine an optimal cut-off associated with knee pain and radiographic osteoarthritis (ROA) in community-derived men and women over 40 years old. [3] to examine whether community-derived people with early and chronic established knee pain have different risks of having effusion, synovial hypertrophy and Power Doppler signal (PDS), and to explore whether synovial changes detected on US predict/associate with subsequent knee pain worsening. [4] to explore the role of peripheral and central risk factors of knee pain, including the role of synovial changes detected on US in different types of knee pain. Methods: A systematic literature search was undertaken in Medline, EMBASE, Allied and Complementary Medicine, PubMed Web of Science, and SCOPUS databases in May 2015. Frequencies of US abnormalities in people with knee OA/pain, in the general population or asymptomatic controls were pooled using the random effects model. Publication bias and heterogeneity between studies were examined. The source population was the Knee Pain and Related Health in the Community (KPIC, n=9506) survey in Nottingham, UK. All participants had bilateral US and radiographic examination. Synovial changes detected on US were measured by two observers (inter-observer concordance correlation was 0.8 (0.6 to 0.9) for effusion and 0.7 (0.5 to 0.9) for synovial hypertrophy). OA structural changes were measured by standardised radiographs (semi-flexed weight-bearing and flexed skyline views) using the Nottingham Line Drawing Atlas (NLDA). A cross-sectional study comprised of 299 randomly selected adults ≥40 years old (147 women, 152 men). The normal range (95% quintile) for effusion and synovial hypertrophy was calculated in the healthy sample (no current knee pain and no ROA, n=163). The optimal cut-off was established using ROC curve analysis. A case-control study compared community-derived participants with early knee pain (n=298), chronic established knee pain (n=100) and no knee pain (n=94) at baseline. 166 early knee pain participants were followed-up at one year for changes in knee pain and synovial changes detected on US. Relationships between changes in synovial changes detected on US and pain severity were examined using correlation analysis. 255 participants with early and established knee pain replied to a one-year follow-up questionnaire. Predictors of knee pain worsening were determined using logistic regression. Central and peripheral risk factors for knee pain were examined using participants from both the cross-sectional and case-control studies (n=736). The contribution of each was presented using ROC curves. Subgroup analysis was undertaken according to the presence/absence of ROA and widespread pain (WSP) for the association between synovial changes detected on US and knee pain. A within-person analysis in participants with unilateral knee pain was also undertaken. Results:Systematic review and meta-analysis: 29 studies (4720 patients) were identified from the literature. The pooled prevalence of US effusion, synovial hypertrophy and PDS in people with knee OA/pain were 51.5% (95%CI 40.2 to 62.8), 41.5% (26.3 to 57.5) and 32.7% (8.34 to 63.24), respectively, which were higher than those in the general population or asymptomatic controls (19.9% (95%CI 7.8 to 35.3), 14.5% (0 to 58.81), and 15.8% (3.08 to 35.36), respectively). People with knee OA (ACR criteria or ROA) had greater prevalence of synovial changes detected on US than people with knee pain (p=0.037, p=0.010 and p=0.009, respectively). Cross-sectional study: Synovial changes detected on US were different between men and women, therefore, gender-specific reference limits were estimated. In people without KP and structural OA the normal range for effusion was between 0 to 10.3 mm for men and between 0 to 9.8 mm for women and the normal range for synovial hypertrophy was between 0 and 6.8 mm for men and between 0 and 5.4 mm for women. The effusion cut-off able to distinguish a subgroup of people with knee pain and ROA (i.e. “symptomatic OA”) with high specificity was 8.9 mm for men and 7.8 mm for women, and for synovial hypertrophy it was 5.8 mm for men and 4.2 mm for women. Case-control study: At baseline, effusion was associated with early (OR 2.64, 95%CI 1.57 to 4.45) and established KP (OR 5.07, 95%CI 2.74 to 9.38). Synovial hypertrophy was also associated with early (OR 5.43, 95%CI 2.12 to 13.92) and established KP (OR 13.27, 95%CI 4.97 to 35.43). However, the association with effusion diminished when adjusted for ROA. PDS was uncommon (early KP 3%, established KP 2%, controls 0%). Changes in effusion or synovial hypertrophy did not correlate with changes in KP in one year. Effusion and ROA predicted worsening of knee pain at one year (aOR 1.95, 95% CI 1.05 to 3.64, and aOR 3.52 95%CI 1.37 to 9.09, respectively). Central versus peripheral risk factors: A number of central and peripheral risk factors associated with knee pain, including WSP, pain catastrophising, knee injury, ROA, effusion and synovial hypertrophy. Although 25% of knee pain was explained by peripheral risk factors, only 5% was explained by central risk factors. Knee pain was stratified into 4 subgroups according to ROA and WSP. The association between synovial changes detected on US and knee pain varied between subgroups, being strongest in people with isolated ROA (e.g., aOR for hypertrophy 9.99, 95%CI 5.06 to 19.03), moderate in people with ROA plus WSP (aOR 7.24, 95%CI 3.04 to 17.25), weak in people with neither ROA nor WSP (aOR 2.25, 95%CI 1.19 to 4.22) and statistically insignificant in people with isolated WSP (aOR 2.21 95%CI 0.99 to 4.93). This was confirmed by the “one-person two knee” analysis where WSP was fully balanced between painful knees and pain-free knees. The association between synovial changes detected on US and knee pain was stronger when the knees had underlying structure OA changes. Conclusions: Effusion and synovial hypertrophy but not PDS are common in community-derived people with knee pain. These features differ in men and women, requiring different thresholds for abnormality. Synovial changes detected on US are associated knee pain, especially in people with ROA but no WSP. However, changes in effusion and synovial hypertrophy do not correlate with changes in knee pain, and effusion but not synovial hypertrophy predicts pain progression at one year. Further study of the causality between synovial changes detected on US and structural OA, and between peripheral and central risk factors for knee pain is needed.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:748216 |
| Date | January 2018 |
| Creators | Sarmanova, A. |
| Publisher | University of Nottingham |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://eprints.nottingham.ac.uk/48392/ |
Page generated in 0.0104 seconds