Summary
Objectives
To examine the prevalence of synovial effusion, synovial hypertrophy and positive Doppler signal (DS) detected by ultrasound (US) in people with knee osteoarthritis (OA) and/or knee pain compared to that in the general population.
Method
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.
Results
Twenty four studies in people with knee pain/OA and five studies of the general population or asymptomatic controls met the inclusion criteria. The pooled prevalence of US effusion, synovial hypertrophy and positive DS in people with knee OA/pain were 51.5% (95% CI 40.2 to 62.8), 41.5% (26.3–57.5) and 32.7% (8.34–63.24), respectively, which were higher than those in the general population or asymptomatic controls (19.9% (95%CI 7.8–35.3%), 14.5% (0–58.81), and 15.8 (3.08–35.36), respectively). People with knee OA (ACR criteria or radiographic OA) had greater prevalence of US abnormalities than people with knee pain (P = 0.037, P = 0.010 and P = 0.009, respectively).
Conclusions
US detected effusion, synovial hypertrophy and DS are more common in people with knee OA/pain, compared to the general population. These abnormalities relate more to presence of OA structural changes than to pain.
Keywords: Ultrasound, Synovial, Osteoarthritis, Meta-analysis, Synovitis
Introduction
Pathologically osteoarthritis (OA) is characterised by involvement of all joint tissues, typically with focal cartilage loss, osteophyte formation, subchondral bone remodelling, and synovial and capsular thickening. Although synovial hyperplasia in knees affected by OA is focal and less marked than in knees with rheumatoid arthritis it may still play an important role in disease pathogenesis1, 2. Cohort studies have shown a positive association between synovial pathology and disease progression3, 4, 5. Therefore whether synovitis is a potential biomarker of inflammatory response and therapeutic target in OA is an important research question2.
Ultrasound (US) imaging is used widely because it is non-invasive, has no radiation burden, is relatively inexpensive, involves a short examination time, and has good patient acceptability6. US detection of synovial effusion and synovial hypertrophy in knees is reported to be more sensitive than clinical examination7, 8, correlates well with histological findings9, 10 and is equivalent to MRI in visualising effusion11, 12. Strong perfusion (i.e., Doppler signal (DS)) is associated with clinical signs of inflammation (soft tissue swelling, tenderness, increased warmth) and also with histological and laboratory markers of inflammation (e.g., serum C-reactive protein) in people with inflammatory arthritis13, 14.
A number of predominantly hospital-based studies have been undertaken in knee OA to examine US detected abnormalities. However, the normal values, thresholds and frequencies of these features in the general population and in community-based people with knee pain or OA remain largely unknown. Therefore, it is of interest to systematically review studies of synovial effusion, synovial hypertrophy and positive DS in the general population and in people with knee pain or knee OA and, if possible, the prevalence and associations of such changes.
Materials and methods
Data sources and search strategy
Two systematic literature searches were performed using computer-based literature indexes such as Medline (1946–), EMBASE (1974–), Allied and Complementary Medicine (1985–), PubMed (1960–), Web of Science, and SCOPUS (1960–) in May, 2015. Citations and abstracts retrieved from this search were downloaded to EndNote X6.0.1 (licenced to The University of Nottingham).
The first search included (a) OA of the knee, and (b) US. The search terms were ‘‘[ultrasound or sonography or ultrasonography or doppler or dopplerography or power-doppler] and [knee osteoarthritis or knee osteoarthrosis or gonarthritis or gonarthrosis or knee pain or ((osteoarthritis or osteoarthrosis or osteophyte or joint space narrowing or degenerative joint disease(s)) and knee)]’’ (Supplementary file 1).
The second search was performed for studies that have explored prevalence of synovial changes in the general population irrespective of knee pain or knee OA using terms “[knee(s) and [ultrasound or sonography or ultrasonography or doppler or dopplerography or power-doppler] and [normal or healthy or general or population-based]” (Supplementary file 1).
Selection criteria
Observational studies were included if they examined US-detected synovial effusion, synovial hypertrophy, or DS detected in people with knee pain/OA, in the general population or in normal/healthy controls. If studies were based on the same participants and same outcome measures, only one publication with the most detailed information was included in the review. There were no language restrictions.
Randomised controlled trials, studies in selected groups with synovial effusion or synovial hypertrophy, studies without clear definition of US-detected pathology (for example “synovitis” without description whether it is related to synovial hypertrophy or combined measure of effusion and hypertrophy), or studies not reporting the prevalence estimate were excluded as they cannot provide an adequate estimate of prevalence. Although reviews and conference proceedings were not included their references were cross-checked.
Data extraction and outcome measures
For each included article information on authors, year of publication, study design (cross sectional, case control), population (hospital, community), sample size, age, gender, body mass index (BMI), diagnostic criteria (e.g., American College of Rheumatology (ACR)), radiographic score (e.g., Kellgren and Lawrence score (K&L)), and US findings were systematically extracted using a specifically developed data extraction form and then transferred to a database.
The primary outcome measure was frequency/prevalence of US effusion, synovial hypertrophy and DS in people with knee pain/OA and in a control or general population derived directly or indirectly from information provided in each study. The secondary outcome measure was the association of US features with OA clinical features (pain, impaired function) and radiographic structural damage. Scores for pain intensity were standardised to a common 0 (no pain) to 100 (worse pain) scale.
Quality assessment
The Newcastle-Ottava Scales (NOS) were used for case–control and cross-sectional studies15 as recommended by the Cochrane Non-Randomized Studies Methods Working Group16. Three main criteria were assessed: participant selection and representativeness, comparability of study groups, and assessment of outcome or exposure. The quality score is based on a “star” system (range 0–9 stars for case–control studies and from 0 to 10 for cross-sectional studies) with a higher score representing better methodologic quality. The percentage of the maximum score achieved was used to present the quality of each study.
Statistical analysis
To derive a pooled estimation of prevalence across different studies, the random effects meta-analysis was undertaken using the METAPROP package (with the Freeman–Tukey double arcsine transformation and exact binominal confidence intervals for prevalence). Heterogeneity between studies was measured using the I2 and Q test17, 18, 19. 95% confidence interval (CI) and P value of 0.05 were used for a statistically significant inference. Publication bias was assessed using funnel plots and Eggers test20. If the number of studies included in the meta-analysis was too small (≤4) the Harbord test was applied to measure publication bias21. Statistical analysis was undertaken in Stata SE V13.1 (StataCorp LP, College Station, TX, USA)22, 23.
Results
Selection of studies
The first search yielded 4149 titles and abstracts, of which 65 potentially relevant publications were considered for full-text assessment. Forty-one studies were excluded by reading full-text papers, leaving a total of 24 studies which met the inclusion criteria. The second search returned 4479 citations of which only three met inclusion criteria and two additional studies were identified from the reference search (Fig. 1). All studies were published between 1990 and 2015. Three studies were translated from German, Italian and Russian24, 25, 26, other studies were written in English.
Fig. 1.
Study selections.
Characteristics of studies
Data for prevalence were derived from both cross-sectional and case–control studies. Of 24 studies reporting the prevalence of US-detected effusion, synovial hypertrophy and DS in people with knee pain/OA, 14 were case–control and 10 were cross-sectional in design. Only four studies were community-based, the rest recruited participants from hospital populations except for four studies which did not declare the setting11, 27, 28, 29. The sample size ranged from 10 to 600 with nine studies reporting a sample size of more than 100. Age varied from 36 to 74 years. There were 20 studies of people with symptomatic knee OA (defined by ACR criteria) and four studies of people with knee pain irrespective of any underlying structural change. Three studies comprised more than one study group30, 31, 32.
Four cross-sectional studies and one case–control study (in comparison with rheumatoid arthritis) explored prevalence and characteristics of US features in the general population33, 34, and pain-free volunteers24, 25, 35. None of these five studies obtained radiographic data. Three of the five studies (range 50–488) recruited more than 100 subjects33, 34, 35. Age range was from 37 to 73 years.
Ten of the 29 studies were funded from academic sources, one declared no funding, one had commercial support and others did not specify funding resources. Baseline demographic characteristics (age, gender, BMI, pain assessment and radiographic score) were generally well reported. Table I summarises the main characteristics of included studies. More details are in Supplementary file 2.
Table I.
Characteristics of the included studies
| People with knee OA/pain∗ | General/normal population | |
|---|---|---|
| Number of studies | 24 | 5 |
| Number of subjects | 3713 | 1007 |
| Mean age (years) | 61.05 | 52.74 |
| Women (%) | 75.03 | 48.93 |
| Mean BMI† (kg/m2) | 28.2 | 25.33 |
Including control groups.
BMI – body mass index.
Definitions of US pathology varied from dichotomous measures (with different thresholds) to individual scoring systems (0–3 or 0–4 scale) or summative quantitative systems (adding effusion, synovial thickness and/or DS). Supplementary file 3 provides an overview of US scoring systems used in these studies.
Study quality assessment
Of 24 studies in people with knee OA/pain 12 had a score of ≥50%. In cross-sectional studies the Newcastle-Ottava quality scores ranged from 2 to 9 stars with a median score of 5.5 (maximum 10). Three studies scored less than five28, 36, 37. In general, all samples were selected non-randomly, provided adequate definition of cases (ACR-criteria for OA diagnosis or validated tool for knee pain assessment) and blinded US assessment. The scores on each of the seven criteria and total scores for each study are presented in Supplementary file 4-1.
The quality of the case–control studies varied from 1 to 6 stars with a median score of 4 (maximum 9) (Supplementary file 4-2). Overall the majority of studies had an adequate case definition (ACR criteria or radiographic OA). The definition of controls included no history of joint disease and no OA as defined.
Prevalence of US features in people with knee OA/pain
Of the 24 included studies, 21 had data for effusion, 13 for hypertrophy and 7 for DS. The pooled prevalence was 51.5%, 41.5% and 32.7%, respectively (Table II). Studies were highly heterogeneous, but only studies involved in the meta-analysis for hypertrophy had significant publication bias (Table II).
Table II.
Prevalence of US-detected findings in people with knee OA/pain
| Number of studies | Number of subjects | Pooled prevalence (95% CI) | I2% (Pheter) | Ppub | |
|---|---|---|---|---|---|
| People with knee OA/pain | |||||
| Effusion | 21 | 3266 | 51.5 (40.2–62.8) | 97.5 (<0.0001) | 0.082∗ |
| Synovial hypertrophy | 13 | 1785 | 41.5 (26.3–57.5) | 97.6 (<0.0001) | 0.026∗ |
| DS | 7 | 538 | 32.7 (8.34–63.24) | 98.0 (<0.0001) | 0.493∗ |
| The general/normal population | |||||
| Effusion | 6 | 922 | 19.9 (7.81–35.34) | 94.7 (<0.0001) | 0.587∗ |
| Synovial hypertrophy | 4 | 601 | 14.5 (0–58.81) | 98.7 (<0.0001) | 0.118∗∗ |
| DS | 2 | 533 | 15.8 (3.08–35.36) | 93.8 (<0.001) | – |
CI: confidence interval; I2: inconsistency; Pheter: p for heterogeneity; Ppub: p for publication bias.
Egger's test.
Harbord's test.
Several subgroup analyses were undertaken according to US threshold for abnormality, sample size of study, overall quality of study and definition of OA. The results are summarised in Table III. In general larger studies (≥100) tended to give a lower prevalence than smaller studies (<100). Similarly, higher quality studies (overall score ≥50%) tended to have a lower prevalence than lower quality studies (overall score <50%). This was especially true when DS was assessed, where a clear separation was observed between higher and lower quality studies (Fig. 2). Interestingly, people with either ACR or radiographic knee OA had greater prevalence of all three US abnormalities than people with knee pain (Table III).
Table III.
Subgroup analysis in studies on people with knee OA/pain
| Subgroup analysis | Effusion |
Synovial hypertrophy |
DS |
||||||
|---|---|---|---|---|---|---|---|---|---|
| Number of studies | Prevalence (95% CI) | P∗ | Number of studies | Prevalence (95% CI) | P∗ | Number of studies | Prevalence (95% CI) | P∗ | |
| Threshold | |||||||||
| ≥4 mm | 11 | 52.5 (38–66.8) | 10 | 43.1 (26.5–60.5) | |||||
| ≥2 mm | 7 | 67.6 (55.8–78.3) | 2 | 25.5 (14.5–38.2) | |||||
| Absent or present | 6 | 32.7 (13.9–54.9) | 0.018 | 3 | 46.9 (0–99.7) | 0.234 | |||
| Sample size | |||||||||
| ≥100 subjects | 9 | 37.1 (20.8–55.2) | 5 | 21.4 (12.4–32.1) | 1 | ||||
| <100 subjects | 15 | 60.8 (48.4–72.5) | 0.034 | 10 | 52.8 (29.3–75.7) | 0.015 | 8 | ||
| Quality score | |||||||||
| <50% | 10 | 54.7 (32.1–76.4) | 5 | 47.1 (8.4–88.0) | 4 | 77 (54.5–93.7) | |||
| ≥50% | 14 | 49.2 (37.3–61.2) | 0.677 | 10 | 38.5 (24.7–53.4) | 0.726 | 5 | 6.0 (2.4–10.9) | <0.0001 |
| Case definition | |||||||||
| Knee OA | 14 | 58.7 (47–69.9) | 12 | 49 (30.5–67.6) | 7 | 43.8 (11.7–79.0) | |||
| Knee pain | 5 | 26 (5.6–54.4) | 0.037 | 3 | 15.2 (3.3–33.3) | 0.010 | 2 | 4.8 (1.9–8.7) | 0.009 |
| Study design | |||||||||
| Cross-sectional | 11 | 43.0 (28.0–58.0) | 5 | 25.0 (17.0–34.0) | 2 | 4.0 (2.0–8.0) | |||
| Case-control | 13 | 59.0 (41.0–76.0) | 0.180 | 10 | 50.0 (24.0–76.0) | 0.07 | 7 | 44.0 (13.0–78.0) | <0.0001 |
| Doppler settings | |||||||||
| Colour doppler | 2 | 87.0 (80.0–92.0) | |||||||
| Power doppler | 7 | 20.0 (4.0–42.0) | |||||||
| Not stated | 1 | 6.0 (3.0–12.0) | <0.0001 | ||||||
| Mean age | |||||||||
| ≤60 | 9 | 53.0 (42.0–63.0) | 5 | 24.0 (13.0–38.0) | 1 | ||||
| >60 | 13 | 61.0 (47.0–74.0) | 0.070 | 9 | 55.0 (28.0–80.0) | 0.050 | 7 | ||
| Women proportion | |||||||||
| ≤70 | 8 | 51.0 (29.0–73.0) | 7 | 31.0 (16.0–49.0) | 7 | ||||
| >70 | 14 | 59.0 (48.0–70.0) | 0.520 | 7 | 58.0 (30.0–84.0) | 0.110 | 1 | ||
P-test for heterogeneity between subgroups.
Fig. 2.
Forest plot showing the subgroup analysis by overall quality score for the prevalence of DS in people with knee OA/pain. P – prevalence rates, 95% CI – lower and upper confidence limits of the 95% confidence interval around the mean prevalence rate. The diamond in the forest plot denotes the summary prevalence and its edges the respective 95% CIs. Three groups from the study by Hall were included: (1) – people with symptomatic OA, (2) – people with radiographic OA, (3) – people with knee pain.
Prevalence of US features in the general/normal population
Among five studies identified from the second search, two provided data on prevalence of US detected synovial effusion in the general population33, 34. In addition, four normal (i.e., asymptomatic) control groups from the case control studies11, 27, 30, 38 reported prevalence of US synovial effusion. These made a total number of six studies in this analysis (Table II). The pooled prevalence of US synovial effusion was 19.9% (95%CI 7.8–35.3%), approximately 2–3 times lower than that in people with knee OA/pain (51.5%, 95%CI 40.2–62.8%, Table II). Similarly, four studies11, 27, 30, 34 provided data for hypertrophy and two studies30, 34 for DS. The prevalence of these findings was much lower in the general/normal population than in people with knee OA/pain. The studies were highly heterogeneous but had no evidence of publication bias (Table II).
Associations of US-detected synovial changes with pain and structural changes
Ten studies examined the relationship between knee pain and US-detected synovial changes. Overall, the most studies reported a positive association between knee effusion and pain (7 of 10 studies) and no association between synovial hypertrophy and pain (4 of 6), but there were no data for DS (Table IV).
Table IV.
Associations between effusion and synovial hypertrophy with pain
| Author, year | Sample size | Mean age (SD/range) | Proportion of women | Standardised quality score (% of the maximum score) | Association between effusion and pain | Association between synovial hypertrophy and pain |
|---|---|---|---|---|---|---|
| Bevers 2014∗ | 180 | 57 (9.2) | 66.7 | 50 | No association | No association |
| Song 2008 | 41 | 65 (6.7) | 63.4 | 22.2 | No association | No association |
| Ulasli 2014 | 86 | 56.2 (10.2) | 80.2 | 60.0 | No association | |
| Hall 2014 | 62 | 73.9 (7.8) | 67.7 | 55.6 | Positive association | Positive association |
| D'Agostino 2005∗ | 600 | 66.7 (9.8) | 72.5 | 90.0 | Positive association | No association |
| Malas 2014 | 61 | 58.88 (7.2) | 83.6 | 40.0 | Positive association | |
| Mendieta 2006 | 101 | 62.1 (9) | 70.0 | 70.0 | Positive association with pain on motion | |
| Chan 2014 | 193 | 59 (13.9) | 74.1 | 60.0 | Positive association with pain on walking, but not while sitting | Positive association with pain while sitting, but not walking |
| Wu 2012∗ | 56 | 62.9 (8.2) | 75.0 | 66.7 | Positive association with pain during movement, but not at rest | No association with pain on movement and at rest |
| Naredo 2005 | 50 | 64.3 (7.9) | 88.0 | 55.6 | Positive association with pain during movement and at rest | |
| % positivity | 7/10 | 2/6 |
Adjusted for radiographic severity.
Three studies examined knee pain on walking and at rest separately32, 38, 39. Two studies did not find any association between knee effusion and pain at rest32, 39, whereas this association was observed by Naredo (2005)38. Both studies examined synovial hypertrophy but found no association with pain on walking and indefinite results with pain at rest32, 39. Unfortunately these studies did not provide sufficient data for statistical pooling so the strength of the association between knee pain and US effusion/synovial hypertrophy remains unknown.
Only two studies examined the relationship between Doppler activity and pain, both recruited people with symptomatic knee OA with disease duration more than 6 months. Song (2009)40 found a positive correlation (r = 0.366; P = 0.020) between DS and knee pain in people with moderate to severe knee pain (mean pain score – 68.3 (SD 19.6)) and structural changes on radiographs (K&L ≥2). A study by Iagnocco (2010)41 revealed a significant association between total US score (effusion, synovial hypertrophy and DS score in both knees) and pain (P = 0.004) in participants with knee pain more than 20 mm on a 100 mm VAS scale (mean pain score 48.4 mm (SD 19.9).
Three studies examined the association between US-detected abnormalities and radiographic severity30, 32, 42. A positive association was observed in two studies which directly addressed the association between synovial changes and radiographic severity30, 42. For example, knee effusion or abnormal synovial thickness on US were associated with radiographic OA, defined as K&L ≥3 in one study with ORs of 1.91 (95% CI 1.32 to 2.77) and 2.2 (95% CI 1.33 to 3.64), respectively42. This association was independent of pain, whereas the association between US features and pain was highly dependent on the severity of radiographic changes and only significant in people without OA (K&L ≤2). These findings were supported by a recent study by Hall (2014)30, in which four groups (normal control, knee pain only, radiographic knee OA (K&L ≥2) only, and knee pain plus radiographic OA) were compared. This study found no difference between the normal control and knee pain group, but significantly higher scores in both the asymptomatic radiographic OA and symptomatic radiographic OA groups. The prevalence was 29%, 32%, 81%, and 92% for effusion (≥4 mm); 8%, 12%, 41% and 82% for hypertrophy (≥4 mm); and 2%, 3%, 6% and 16% for DS (any grade), respectively. In addition, this study followed participants for 3 months and found no association between change in pain and change in US features. The study of Wu (2012)32 did not explore directly the association between US findings and structural changes. Participants with knee OA who had bilateral equal K&L scores showed significant differences between symptomatic and asymptomatic knees (P = 0.016 for effusion and P < 0.001 for synovial hypertrophy), suggesting that synovial changes are related more to pain than structural severity. However this study does not allow us to draw a strong conclusion.
Discussion
This is the first meta-analysis of US detected synovial changes in people with and without knee OA/pain. Twenty-nine observational studies including 4720 participants from different countries were included in this study. The main findings are: [1] the prevalence of US detected effusion, synovial hypertrophy and positive DS are 2–3 times higher in people with knee OA/pain than in the general population or asymptomatic control groups; and [2] the US abnormalities relate more to presence of OA structural change than to pain.
People with knee OA had significantly higher prevalence of effusion, synovial hypertrophy and DS than people with knee pain (P = 0.037, P = 0.010 and P = 0.009, respectively) (Table II). This may be contrary to general expectation since the three US features selected are widely considered to reflect inflammation and pain in knee OA is suggested to associate with inflammation43, 44. Importantly, however, this finding suggests that US detected synovial change (effusion, hypertrophy, DS) may mainly correlate with the degree of OA structural change and pathology, which increasingly is recognised to involve all tissues that comprise the joint, rather than represents a biomarker/mechanism that links strongly with pain production.
There was significant heterogeneity between studies with respect to prevalence of all three US features. Such heterogeneity is to be expected because a systematic review brings together studies that are diverse both clinically and methodologically (e.g., thresholds of abnormality, recruitment source, sample size, age, gender proportion, BMI, disease duration, definition of knee OA/pain). For example, among studies in people with knee OA/pain the subgroup analysis revealed that studies with quality scores lower than 50% of maximum presented significantly higher prevalence of DS (P < 0.0001), and studies with sample size less than 100 reported significantly higher prevalence of effusion and synovial hypertrophy (P = 0.034 and P = 0.015, respectively). This suggests that small studies tend to inflate the results – the small study effect21, 45. Care must be taken when interpreting the results from such studies as they may overestimate the prevalence of abnormalities.
The second research question was to determine relationships between US features and knee pain. The majority of studies reported a positive association between presence of effusion and knee pain (7 out of 10) but no association between synovial hypertrophy and pain (2 out of 6) (Table IV). US-detected findings were also associated with structural changes on X-ray in two of the three studies30, 32, 42. However, our subgroup analysis according to knee pain and knee OA suggests that these three US abnormalities relate to knee OA (either ACR symptomatic or radiographic) more than knee pain. Further study is required to explain this finding.
A paucity of information was found on the prevalence of US-detected changes in the general population and no prospective community studies were identified. Considering gender differences and possible associations between normal values and changes in the musculoskeletal system and body composition with increasing age the normal values of US detected synovial characteristics are essential for the classification and diagnosis of people with knee pain and OA. It is expected that the normal values for older adults might differ from those for younger people, since age-related changes contribute to alterations in cartilage morphology, proprioception and muscle weakness even in the absence of OA. For example, in the Framingham study the prevalence of effusion/synovitis on MRI in people without knee OA was 37% if the present/absent scale was applied but only 4% if defined by WORMS grade two or more. Such synovitis was detected significantly more often in men than women (6% and 3%, respectively; P = 0.02), but there was no difference in relation to presence of knee pain or BMI46. However, at present the characteristics of US-detected synovial abnormalities especially in older age groups remains unknown.
There are several limitations to this study. Firstly, we focused only on the knee, so the results cannot be extrapolated to other joints. Secondly, there was significant heterogeneity in the results on prevalence, so the results of this review need to be interpreted with caution. For example, differences in scanning technique were common within included studies (e.g., neutral vs flexed knee position, multi-planar vs midline scan (Supplementary file 3)) which might affect the results and together with differences in participant characteristics (age, gender, disease duration, severity of structural changes) might explain some of the between-study heterogeneity40, 47, 48. Thirdly, the prevalence in the general population was obtained from just a few studies including controls from case control studies. This group is neither a random sample of the general population, nor comparable to the cases with knee OA/pain. The prevalence obtained from such an assembled “normal” control group cannot be extrapolated to the prevalence in the general population.
Our study highlights the lack of information on the presence of synovial change in the knee. Although many studies have explored this question, none has investigated the distribution of these features in the general population, hence the threshold for abnormality has yet to be established. US-detected pathology should be described in detail and studies should provide sufficient information on definition and thresholds used. The heterogeneity across studies highlights the need for a standard protocol in order to allow comparability between studies in the future.
In conclusion, US detected effusion, synovial hypertrophy and DS are more common in people with knee OA/pain, compared to the general population. These abnormalities relate more to presence of OA structural changes than to pain. Further studies on the reasons of this difference, normal values of the US features and their thresholds of abnormalities are warranted.
Contributions
Study concept and design: WZ, MD and AS. The first review (AS) ran the literature search, screened relevant full-text articles for inclusion criteria, extracted data for all studies included in the review, and assessed the study quality. MH and JM (for the1st and 2nd search, respectively) reviewed included articles, extracted data and performed the quality assessment independently from the first reviewer (AS). Any inconsistencies were solved by consensus with involvement of a further experienced reviewer (WZ). AS analysed data and wrote the first draft. All authors (AS, MH, JM, MD and WZ) contributed to data interpretation and editing of final paper. All authors approved the final version for publication.
Competing interests
None.
Role of the funding source
The sponsors of the study had no role in design and conduct of the study; collection, management, analysis and interpretation of the data; and preparation, review or approval of the manuscript.
Acknowledgements
This work was supported by the Arthritis Research UK Pain Centre (Centre Initiative grant number: 20777). The authors would like to thank Bolashak scholarship programme, offered by the Ministry of Education and Science of the Republic of Kazakhstan, for financially supporting the PhD research programme (AS) and the University of Nottingham as a sponsor and host institution of this PhD project.
Footnotes
Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.joca.2016.03.004.
Appendix A. Supplementary data
The following are the supplementary data related to this article:
References
- 1.Hayashi D., Roemer F.W., Katur A., Felson D.T., Yang S.O., Alomran F. Imaging of synovitis in osteoarthritis: current status and outlook. Semin Arthritis Rheum. 2011;41:116–130. doi: 10.1016/j.semarthrit.2010.12.003. [DOI] [PubMed] [Google Scholar]
- 2.Attur M., Samuels J., Krasnokutsky S., Abramson S.B. Targeting the synovial tissue for treating osteoarthritis (OA): where is the evidence? Best Pract Res Clin Rheumatol. 2010;24:71–79. doi: 10.1016/j.berh.2009.08.011. [DOI] [PubMed] [Google Scholar]
- 3.Roemer F.W., Guermazi A., Felson D.T., Niu J.B., Nevitt M.C., Crema M.D. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis. 2011;70:1804–1809. doi: 10.1136/ard.2011.150243. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Atukorala I., Kwoh C.K., Guermazi A., Roemer F.W., Boudreau R.M., Hannon M.J. Synovitis in knee osteoarthritis: a precursor of disease? Ann Rheum Dis. 2014 doi: 10.1136/annrheumdis-2014-205894. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Conaghan P.G., D'Agostino M.A., Le Bars M., Baron G., Schmidely N., Wakefield R. Clinical and ultrasonographic predictors of joint replacement for knee osteoarthritis: results from a large, 3-year, prospective EULAR study. Ann Rheum Dis. 2010;69:644–647. doi: 10.1136/ard.2008.099564. [DOI] [PubMed] [Google Scholar]
- 6.Iagnocco A. Imaging the joint in osteoarthritis: a place for ultrasound? Best Pract Res Clin Rheumatol. 2010;24:27–38. doi: 10.1016/j.berh.2009.08.012. [DOI] [PubMed] [Google Scholar]
- 7.Karim Z., Wakefield R.J., Quinn M., Conaghan P.G., Brown A.K., Veale D.J. Validation and reproducibility of ultrasonography in the detection of synovitis in the knee: a comparison with arthroscopy and clinical examination. Arthritis Rheum. 2004;50:387–394. doi: 10.1002/art.20054. [DOI] [PubMed] [Google Scholar]
- 8.Ulasli A.M., Yaman F., Dikici O., Karaman A., Kacar E., Demirdal U.S. Accuracy in detecting knee effusion with clinical examination and the effect of effusion, the patient's body mass index, and the clinician's experience. Clin Rheumatol. 2014;33:1139–1143. doi: 10.1007/s10067-013-2356-6. [DOI] [PubMed] [Google Scholar]
- 9.Walther M., Harms H., Krenn V., Radke S., Faehndrich T.P., Gohlke F. Correlation of power Doppler sonography with vascularity of the synovial tissue of the knee joint in patients with osteoarthritis and rheumatoid arthritis. Arthritis Rheum. 2001;44:331–338. doi: 10.1002/1529-0131(200102)44:2<331::AID-ANR50>3.0.CO;2-0. 2<331::AID-ANR50>3.0.CO;2–0. [DOI] [PubMed] [Google Scholar]
- 10.Labanauskaite G., Sarauskas V. Correlation of power Doppler sonography with vascularity of the synovial tissue. Med Kaunas Lith. 2003;39:480–483. Lithuanian. [PubMed] [Google Scholar]
- 11.Tarhan S., Unlu Z. Magnetic resonance imaging and ultrasonographic evaluation of the patients with knee osteoarthritis: a comparative study. Clin Rheumatol. 2003;22:181–188. doi: 10.1007/s10067-002-0694-x. [DOI] [PubMed] [Google Scholar]
- 12.Aleo E., Barbieri F., Sconfienza L., Zampogna G., Garlaschi G., Cimmino M.A. Ultrasound versus low-field magnetic resonance imaging in rheumatic diseases: a systematic literature review. Clin Exp Rheumatol. 2014;32:S91–S98. [PubMed] [Google Scholar]
- 13.Joshua F., Edmonds J., Lassere M. Power Doppler ultrasound in musculoskeletal disease: a systematic review. Semin Arthritis Rheum. 2006;36:99–108. doi: 10.1016/j.semarthrit.2006.04.009. [DOI] [PubMed] [Google Scholar]
- 14.Koski J.M. Doppler imaging and histology of the synovium. J Rheumatol. 2012;39:452–453. doi: 10.3899/jrheum.110814. [DOI] [PubMed] [Google Scholar]
- 15.Wells G., Shea B., O'Connell D. The Ottawa Hospital; Ottawa (ON): 2007. Newcastle-Ottawa Scale (NOS) for Assessing the Quality of Non-randomized Studies in Meta-analysis.http://www.medicine.mcgill.ca/rtamblyn/Readings/The%20Newcastle%20-%20Scale%20for%20assessing%20the%20quality%20of%20nonrandomised%20studies%20in%20meta-analyses.pdf Available from: [Google Scholar]
- 16.Reeves BC, Deeks JJ, Higgins JP, Wells TGA. Including non-randomized studies. In: Cochrane Handbook for Systematic Reviews of Interventions Higgins J, Green S Eds.: The Cochrane Collaboration 2011: chapter 13. Available from: www.cochrane-handbook.org.
- 17.Higgins J.P., Thompson S.G., Deeks J.J., Altman D.G. Measuring inconsistency in meta-analyses. BMJ. 2003;327:557–560. doi: 10.1136/bmj.327.7414.557. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Higgins J.P., Thompson S.G. Controlling the risk of spurious findings from meta-regression. Stat Med. 2004;23:1663–1682. doi: 10.1002/sim.1752. [DOI] [PubMed] [Google Scholar]
- 19.Harris R., Bradburn M., Deeks J., Harbord R., Altman D., Sterne J. METAN: fixed and random effects meta-analysis. Stata J. 2008;8:3–28. [Google Scholar]
- 20.Steichen T.J. Tests for publication bias in meta-analysis. Stata Tech Bull. 1998:9–15. [Google Scholar]
- 21.Harbord R.M., Harris R.J., Sterne J.A.C. Updated tests for small-study effects in meta-analyses. Stata J. 2009;9:197–210. [Google Scholar]
- 22.Nyaga V.N., Arbyn M., Aerts M. Metaprop: a Stata command to perform meta-analysis of binomial data. Archives Public Health. 2014;72:39. doi: 10.1186/2049-3258-72-39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Chaimani A., Mavridis D., Salanti G. A hands-on practical tutorial on performing meta-analysis with Stata. Evid Based Ment Health. 2014;17:111–116. doi: 10.1136/eb-2014-101967. [DOI] [PubMed] [Google Scholar]
- 24.Martino F., Angelelli G., Ettorre G.C., Macarini L., Patella V., Moretti B. The normal aspect of the suprapatellar bursa in echography of the knee. Radiol Med. 1992;83:43–48. Italian. [PubMed] [Google Scholar]
- 25.Mielke G., Brandruplukanow A., Bandilla K., Berg D., Higer P., Loch E.G. Sonography of the knee-joint – normal findings and changes in rheumatoid-arthritis. Ultraschall Med. 1990;11:40–43. doi: 10.1055/s-2007-1011527. German. [DOI] [PubMed] [Google Scholar]
- 26.Svetlova M.S., Vezikova N.N. Clinical, instrumental, and laboratory characteristics of early stages of gonarthrosis. Ter Arkh. 2010;82:54–58. Russian. [PubMed] [Google Scholar]
- 27.Blankstein A., Ganel A., Mirovsky Y., Chechick A., Dudkiewicz I. Early diagnosis of generalized knee pain and osteoarthritis by ultrasound. Aktuelle Traumatol. 2006;36:175–179. [Google Scholar]
- 28.Malas F.U., Kara M., Kaymak B., Akinci A., Özçakar L. Ultrasonographic evaluation in symptomatic knee osteoarthritis: clinical and radiological correlation. Int J Rheum Dis. 2014;17:536–540. doi: 10.1111/1756-185X.12190. [DOI] [PubMed] [Google Scholar]
- 29.Zivanovic S., Rackov L.P., Vucetic D., Mijuskovic Z. Arthrosonography and the biomarker cartilage oligomeric matrix protein in the detection of knee osteoarthrosis effusion. J Med Biochem. 2009;28:108–115. [Google Scholar]
- 30.Hall M., Doherty S., Courtney P., Latief K., Zhang W., Doherty M. Synovial pathology detected on ultrasound correlates with the severity of radiographic knee osteoarthritis more than with symptoms. Osteoarthritis Cartilage. 2014;22:1627–1633. doi: 10.1016/j.joca.2014.05.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.de Miguel Mendieta E., Cobo Ibanez T., Uson Jaeger J., Bonilla Hernan G., Martin Mola E. Clinical and ultrasonographic findings related to knee pain in osteoarthritis. Osteoarthritis Cartilage. 2006;14:540–544. doi: 10.1016/j.joca.2005.12.012. [DOI] [PubMed] [Google Scholar]
- 32.Wu P.T., Shao C.J., Wu K.C., Wu T.T., Chern T.C., Kuo L.C. Pain in patients with equal radiographic grades of osteoarthritis in both knees: the value of gray scale ultrasound. Osteoarthritis Cartilage. 2012;20:1507–1513. doi: 10.1016/j.joca.2012.08.021. [DOI] [PubMed] [Google Scholar]
- 33.Abraham A.M., Pearce M.S., Mann K.D., Francis R.M., Birrell F. Population prevalence of ultrasound features of osteoarthritis in the hand, knee and hip at age 63 years: the Newcastle thousand families birth cohort. BMC Musculoskelet Disord. 2014;15:162. doi: 10.1186/1471-2474-15-162. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.D'Agostino M.A., Iagnocco A., Aegerter P., Kleyer A., Zwerina J., Perricone C. Does subclinical inflammation contribute to impairment of function of knee joints in aged individuals? High prevalence of ultrasound inflammatory findings. Rheumatol Oxf. 2015;54:1622–1629. doi: 10.1093/rheumatology/kev032. [DOI] [PubMed] [Google Scholar]
- 35.Schmidt W.A., Schmidt H., Schicke B., Gromnica-Ihle E. Standard reference values for musculoskeletal ultrasonography. Ann Rheum Dis. 2004;63:988–994. doi: 10.1136/ard.2003.015081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Picerno V., Filippou G., Bertoldi I., Adinolfi A., Di Sabatino V., Galeazzi M. Prevalence of Baker's cyst in patients with knee pain: an ultrasonographic study. Reumatismo. 2013;65:264–270. doi: 10.4081/reumatismo.2013.715. [DOI] [PubMed] [Google Scholar]
- 37.Artul S., Khazin F., Hakim J., Habib G. Ultrasonographic findings in a large series of patients with knee pain. J Clin Imaging Sci. 2014;4:45. doi: 10.4103/2156-7514.139735. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Naredo E., Cabero F., Palop M.J., Collado P., Cruz A., Crespo M. Ultrasonographic findings in knee osteoarthritis: a comparative study with clinical and radiographic assessment. Osteoarthritis Cartilage. 2005;13:568–574. doi: 10.1016/j.joca.2005.02.008. [DOI] [PubMed] [Google Scholar]
- 39.Chan K.K., Sit R.W., Wu R.W., Ngai A.H. Clinical, radiological and ultrasonographic findings related to knee pain in osteoarthritis. PLoS One. 2014;9:e92901. doi: 10.1371/journal.pone.0092901. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Song I.H., Althoff C.E., Hermann K.G., Scheel A.K., Knetsch T., Burmester G.R. Contrast-enhanced ultrasound in monitoring the efficacy of a bradykinin receptor 2 antagonist in painful knee osteoarthritis compared with MRI. Ann Rheum Dis. 2009;68:75–83. doi: 10.1136/ard.2007.080382. [DOI] [PubMed] [Google Scholar]
- 41.Iagnocco A., Meenagh G., Riente L., Filippucci E., Delle Sedie A., Scire C.A. Ultrasound imaging for the rheumatologist XXIX. Sonographic assessment of the knee in patients with osteoarthritis. Clin Exp Rheumatol. 2010;28:643–646. [PubMed] [Google Scholar]
- 42.D'Agostino M.A., Conaghan P., Le Bars M., Baron G., Grassi W., Martin-Mola E. EULAR report on the use of ultrasonography in painful knee osteoarthritis. Part 1: prevalence of inflammation in osteoarthritis. Ann Rheum Dis. 2005;64:1703–1709. doi: 10.1136/ard.2005.037994. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Filippucci E., Salaffi F., Carotti M., Grassi W. Doppler ultrasound imaging techniques for assessment of synovial inflammation. Rep Med Imaging. 2013;6:83–91. [Google Scholar]
- 44.Joshua F., Lassere M., Bruyn G.A., Szkudlarek M., Naredo E., Schmidt W.A. Summary findings of a systematic review of the ultrasound assessment of synovitis. J Rheumatol. 2007;34:839–847. [PubMed] [Google Scholar]
- 45.Harbord R.M., Egger M., Sterne J.A. A modified test for small-study effects in meta-analyses of controlled trials with binary endpoints. Stat Med. 2006;25:3443–3457. doi: 10.1002/sim.2380. [DOI] [PubMed] [Google Scholar]
- 46.Guermazi A., Niu J., Hayashi D., Roemer F.W., Englund M., Neogi T. Prevalence of abnormalities in knees detected by MRI in adults without knee osteoarthritis: population based observational study (Framingham Osteoarthritis Study) BMJ. 2012;345:e5339. doi: 10.1136/bmj.e5339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Terslev L., D'Agostino M.A., Brossard M., Aegerter P., Balint P., Backhaus M. Which knee and probe position determines the final diagnosis of knee inflammation by ultrasound? Results from a European multicenter study. Ultraschall Med. 2012;33:E173–E178. doi: 10.1055/s-0031-1281973. [DOI] [PubMed] [Google Scholar]
- 48.Zivanovic S., Petrovic-Rackov L., Zivanovic A. Arthrosonography and biomarkers in the evaluation of destructive knee cartilage osteoarthrosis. Srp Arh Celok Lek. 2009;137:653–658. doi: 10.2298/sarh0912653z. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.