Article Text

Download PDFPDF

Extended report
Synovitis in knee osteoarthritis: a precursor of disease?
  1. I Atukorala1,2,
  2. C K Kwoh3,
  3. A Guermazi4,
  4. F W Roemer4,5,
  5. R M Boudreau6,
  6. M J Hannon6,
  7. D J Hunter1
  1. 1Institute of Bone and Joint Research, Kolling Institute, University of Sydney, Australia and Royal North Shore Hospital, St Leonards, New South Wales, Australia
  2. 2University of Colombo, Colombo, Sri Lanka
  3. 3University of Arizona, Tucson, Arizona, USA
  4. 4Boston University School of Medicine, Boston, Massachusetts, USA
  5. 5Klinikum Augsburg, Augsburg, Germany
  6. 6University of Pittsburgh, Pittsburgh, Pennsylvania, USA
  1. Correspondence to Dr Inoshi Atukorala, Department of Clinical Medicine, Faculty of Medicine, University of Colombo, PO Box 271, 25 Kynsey Road, Colombo 8 08000, Sri Lanka; inoshi.atu{at}


Objectives It is unknown whether joint inflammation precedes other articular tissue damage in osteoarthritis. Therefore, this study aims to determine if synovitis precedes the development of radiographic knee osteoarthritis (ROA).

Methods The participants in this nested case–control study were selected from persons in the Osteoarthritis Initiative with knees that had a Kellgren Lawrence grading (KLG)=0 at baseline (BL). These knees were evaluated annually with radiography and non-contrast-enhanced MRI over 4 years. MRIs were assessed for effusion-synovitis and Hoffa-synovitis. Case knees were defined by ROA (KLG≥2) on the postero-anterior knee radiographs at any assessment after BL. Radiographs were assessed at P0 (time of onset of ROA), 1 year prior to P0 (P-1) and at BL. Controls were participants who did not develop incident ROA (iROA) from BL to 48 months).

Results 133 knees of 120 persons with ROA (83 women) were matched to 133 control knees (83 women). ORs for occurrence of iROA associated with the presence of effusion-synovitis at BL, P-1 and P0 were 1.56 (95% CI 0.86 to 2.81), 3.23 (1.72 to 6.06) and 4.7 (1.10 to 2.95), respectively. The ORs for the occurrence of iROA associated with the presence of Hoffa-synovitis at BL, P-1 and P0 were 1.80 (1.1 to 2.95), 2.47 (1.45 to 4.23) and 2.40 (1.43 to 4.04), respectively.

Conclusions Effusion-synovitis and Hoffa-synovitis strongly predicted the development of iROA.

  • Knee Osteoarthritis
  • Magnetic Resonance Imaging
  • Synovitis

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Osteoarthritis (OA) is the most common joint disease worldwide and is a major cause of pain and disability. Despite increased prevalence of this disease,1 there is at present no therapy that prevents or cures knee OA. Therefore, the current focus is on developing drugs to arrest the progression of this disease. In order to facilitate this process it is imperative that the tissue targets in pathogenesis are identified.

In contrast to earlier paradigms, OA is now perceived as joint failure which is the result of complex interactive disease processes in multiple articular and periarticular structures. These tissues include synovium, cartilage, muscle, subchondral and cortical bone.2 However, there is ongoing debate as to which tissue structures change first in OA. The pathogenic role of bone, the meniscus and hyaline cartilage have been extensively examined3 previously. However, there is increasing evidence that inflammation, particularly synovitis, plays a role in the pathogenic process of OA. Therefore, synovitis is no longer perceived as an innocent bystander in OA.4–6

Previously, synovitis in OA was believed to be a consequence of underlying joint damage, with synovial inflammation resulting from macrophage phagocytosis of intra-articular cartilage or bone debris, calcium pyrophosphate dihydrate or calcium hydroxyapatite crystals with release of soluble cartilage matrix macromolecules.5 ,7

Furthermore, an overwhelming majority of persons with OA have synovitis.4 There is ample evidence that inflammation is present in the synovium in all stages of OA2 ,5 ,8 including those with early disease.9 Moreover, synovitis is associated with pain, disease severity and progression of knee OA.10 Given these findings, it is likely that synovitis has an aetio-pathogenic role in OA and that it may be a precursor rather than just a consequence of joint failure.11 ,12

Synovial inflammation manifests as synovial membrane thickening or joint effusion. However, conventional knee radiographs are unable to visualise the synovial membrane. Though ultrasonography is useful in imaging synovitis, it has not yet been used in large scale studies of OA.13 Therefore, the best imaging method to identify synovial inflammation is by magnetic resonance imaging (MRI). MRI demonstrates synovitis as thickening and enhancement after intravenous contrast administration or as an effusion.13 Accordingly, MRI is useful to investigate early ‘preclinical’ disease before radiographic changes occur. Furthermore, the MRI is invaluable to examine and delineate pathologies in all structures associated with the pathogenesis of OA. This is due to the ability of MRI to visualise deeper structures and examine the knee in its entirety in three-dimensions unlike any other imaging modality.13

Large scale epidemiological studies have established the validity of semiquantitative assessment of synovitis by non-contrast-enhanced MRI.14–16 Currently, effusion-synovitis or Hoffa-synovitis are the surrogates used to identify synovial inflammation on non-contrast-enhanced MRI.13 ,17 Effusion-synovitis is defined by the presence of fluid equivalent signal in the joint cavity on T2/IW/PDw magnetic resonance images. Hoffa-synovitis assesses the degree of hyperintensity in Hoffa's fat pad on T2/PD/Iw-w fat suppressed sequences within the fat pad.18

This nested case–control study was performed to identify whether synovitis precedes the onset of radiographic knee OA (ROA). Thus, the objective was to identify whether synovitis/effusion is more prevalent in knees that develop ROA compared with those that do not.

Patients and methods

Study design and subjects

The study participants were selected from the OA Initiative (OAI). The OAI is a multicentre, 8-year prospective observational cohort study that examines the biomarkers and risk factors for causation and progression of knee OA in 4796 individuals.

Details of subject inclusion and exclusion have been described previously.19 ,20 The OAI participants are both those with knee OA and those who are at high risk of developing knee OA.21 The participants were of all ethnicities with 59% being women and 18% being of non-white ethnicity.

Study subjects were assessed annually by radiography and MRI.

Cases and controls

Cases knees were those that developed incident ROA (iROA) on or before the 48-month OAI X-ray and that had baseline Kellgren Lawrence grades (KLG) of 0. All such knees with MRI readings at either of the visits where iROA was first read or the time prior to that were included. iROA was the first occurrence of radiographic findings compatible with OA (KLG of ≥2 on the postero-anterior view22) during the course of study. This time point was called P0 with P-1 being defined as the time point 1 year before ROA was detected on radiographs.

Each case was matched from a sample of knees that were KLG 0 at baseline, did not develop iROA during the study period, who were at risk at the time of case occurrence, had available images at the timepoints where the case had images. The case knees were matched to a single control knee on sex, age (within 5 years) and contralateral knee OA status (ie, KLG=0, 1 or 2+ in the other knee).

Included knees had to have images at baseline and either P0 or P-1, but not necessarily both. The small set of knees where effusion synovitis could be scored and Hoffa synovitis could not were included for the analysis of the former only. Because of the design, an MRI score for both parts of the case–control pair had to be available before it was included in that analysis.

History of joint surgery and previous injury to the knee was evaluated at the enrolment visit by asking the participants whether they have ever injured their knee(s) badly enough to limit their ability to walk for at least 2 days or whether they have ever had any kind of knee surgery including arthroscopy, ligament repair or meniscectomy. At the same visit, participants’ body mass index (BMI) was taken and they were asked about their knee pain that they had in the past 12 months (no pain, infrequent pain (pain not on most days of a month) or frequent pain (pain on most days of at least 1 month) and the number of knee bending activities in the past 30 days. These activities were: taking more than 10 flights of stairs, kneeling for 30+ min, squatting for 30+ min, moving a heavy (25 pounds or more) object or going into/out of a squat more than 10 times. A scale from 0 to 5 was created by taking the sum of the number of different activities to which they answered yes.


Fixed flexion radiographs acquired using the Synaflexor of both knees were performed in all subjects. The radiographs of knees were assessed for their KLG23 and were found reliable (κ=0.7 at baseline and 0.78 at 36th month follow-up). Radiographs acquired at the baseline, 12-month, 24-month, 36-month and 48-month visits were read by the OAI central readers.


MRI acquisition was done on a 3 tesla MRI (3T-MRI) machine. Non-contrast-enhanced MRIs of both cases and controls were obtained. The OAI MRI protocols and image acquisition techniques were used. Semiquantitative tissue scoring was done using coronal intermediate-weighted 2DTSE (slice thickness/in-plane area 3.0/0.37×0.46) and sagittal intermediate-weighted 2TSE images with fat suppression (slice thickness/in-plane area 3.0/0.37×0.51).20

The MRI OA Knee Score (MOAKS) instrument was used to assess the whole joint for structural changes compatible with knee OA.18 Knee inflammation was assessed by using surrogates: Hoffa-synovitis in infrapatellar and intercondylar regions and effusion-synovitis. The MRIs were scored semiquantitatively 0–3 as part of the MOAKS scoring system. Hoffa-synovitis score (on sagittal image) and effusion–synovitis (on axial) were assessed for magnitude as follows: 0=normal/ no hyperintensity, 1=mild, 2=moderate and 3=severe for Hoffa-synovitis and 0=physiological amount, 1=small, 2=medium and 3=large for effusion-synovitis (figure 1). For both, any finding (≥1) was compared with no finding.

Figure 1

Effusion-synovitis and Hoffa-synovitis on MRI. Effusion-synovitis: (A) axial dual echo at steady state (DESS) image shows grade 1 effusion-synovitis. Intra-articular fluid and synovial thickening is depicted as hyperintensity (asterisk). (B) Axial DESS image shows marked distension of the joint capsule representing grade 3 effusion-synovitis (asterisk). Hoffa-synovitis: (C) sagittal intermediate-weighted fat suppressed image shows infrapatellar areas of hyperintensity within Hoffa's fat pad representing a surrogate of synovitis on non-enhanced MRI. In this case, grade 2 Hoffa-synovitis is shown (arrows). (D) Another knee exhibits marked signal alterations within Hoffa's fat pad representing grade 3 Hoffa-synovitis (arrows). Note effusion-synovitis in addition (asterisk).

MRI readings were performed by AG and FWR with 15 and 12 years of experience, respectively, in MRI semiquantitative assessment. Scores were entered directly into an electronic web-based database. All MRIs were read sequentially, unblinded to timepoint. Inter-rater reliability on a subset of MRIs was performed for MRI reading quality control. Inter-rater reliability was weighted κ=0.68 (95% CI 0.38 to 0.99) and 0.95 (95% CI 0.61 to 1.00) for synovitis and for effusion, respectively. These were calculated using the full range of semiquantitative assessment (0–3 for synovitis, 0–3 for effusion). Subsequently, ongoing surveillance for measurement drift was carried out by AG by re-reading 5% of the MRIs.24

Statistical analysis

Conditional logistic regression models using a general estimated equations method with a robust sandwich estimator to account for correlations between case–control pairs and between knees for individuals with bilateral incident OA were employed to model the relationships between the key predictors and OA. Models with only the key predictors were run first, and unadjusted associations with OA were obtained. Then models adjusted for baseline covariates (self-reported injury, self-reported knee surgery, BMI and knee bending activities) were run to see that the findings persisted when these covariates were controlled for. Models were run at three time points: baseline (time of enrolment in the OAI), P-1 (the year prior to the finding of ROA) and P0 (concurrent with iROA).


A total of 133 knees from 120 participants with baseline KLG=0 on knee radiographs developed ROA during the first 48 months. An identical number of matched control knees were selected. The demographics and baseline clinical parameters did not differ significantly between cases and controls (table 1). All radiographs of the participants were KLG 0 at baseline. Fifty-nine subjects (44.36%) of cases and controls had contralateral OA at baseline. The majority (69.92% of cases and 78.95% of controls) were pain free or had infrequent knee pain at baseline. Forty per cent cases had a frequent knee pain compared with 28% of controls.

Table 1

Baseline characteristics of the cases and controls in the study population

The timing of the occurrence of ROA in the cases is shown in figure 2. At the onset of ROA, 57.1% of cases had frequent knee pain and 25.6% had infrequent pain. Correspondingly, 20.3% and 32.3% of controls showed frequent and infrequent pain, respectively. History of injury at baseline was reported by 20.3% of cases and 12.8% of controls. A history of knee surgery (self-reported) at baseline was similarly low: 3.76% of cases and 3.01% of controls. The KLG of cases at point of occurrence of ROA was as follows: KLG 2–75.9%; KLG 3–22.6% and KLG 4–1.5%. Only three controls (2.3%) reached a KLG 1 grade. The remainder continued to be at KLG 0 at P0.

Figure 2

Schema detailing the timing of onset of radiographic knee osteoarthritis (ROA) with number of subjects who developing ROA at each time point. P0=time of onset of ROA, P-1=1 year prior to P0, BL=Baseline, M=months, N=number (%).

At baseline, there was no significant association between effusion-synovitis and subsequent iROA (37.59% vs 30.08%, OR=1.56, CI 0.86 to 2.81). However, effusion-synovitis in the year prior to the finding of iROA (P-1) was associated with subsequent iROA (52.80% in cases vs 29.60% in controls, OR=3.23, CI 1.72 to 6.06). At the time point when iROA was read (P0), the relationship was still present (OR=4.70, CI 2.35 to 9.34).

Hoffa-synovitis showed a similar set of associations except that the difference at baseline was significant (46.21% vs 30.83%, OR 1.80, CI 1.10 to 2.96). At P-1, Hoffa-synovitis continued to be associated with subsequent iROA (53.23% in cases vs 30.40% in controls, OR=2.47, CI 1.45 to 4.23). At P0, the relationship between Hoffa-synovitis and iROA was still found (OR=2.40, CI 1.43 to 4.04) (table 2).

Table 2

Conditional logistic regression analysis for occurrence of effusion-synovitis and Hoffa-synovitis in the study population including the model adjusted for injury/surgery at baseline

The association between synovitis/effusion-synovitis with the occurrence of ROA persisted in a sensitivity analysis where we adjusted for potential covariates (table 2).


This study demonstrates that synovial inflammation and effusion is associated with the later occurrence of ROA.

The first finding is that the presence of synovial inflammation—effusion-synovitis or Hoffa-synovitis—is associated with increased odds of developing iROA. Inflammation has been postulated to play a role in the occurrence of knee OA2 and is believed to be a consequence of underlying damage in the knee joint.5 ,7 However, our study suggests that synovial membrane inflammation plays a role early in the disease. These findings are further supported by previous research on knees without OA which shows that synovitis and effusion increases risk for cartilage loss.25 Therefore, our findings concur with previous findings which indicate at least an accelerating role of inflammation in disease initiation.9 ,25 An additional strength of this study is that we used MRI, which has previously been proven to be a sensitive tool to detect synovitis.13

The second observation is that the occurrence of synovial inflammation is more marked in the year preceding the occurrence of ROA. This association was demonstrated with both surrogates of inflammation: effusion-synovitis and Hoffa-synovitis. The risk of incident OA from synovitis/effusion was the greatest when the visit was immediately proximate to the case defining time point with both parameters. This suggests that the occurrence of synovitis predates, and is also associated with, the development of ROA. However, approximately one-third of the control group had effusion-synovitis and Hoffa-synovitis without subsequent consequence. It is possible that the controls will subsequently develop ROA at a slower rate than the cases. This is best explained by the fact that all the individuals who were selected to the OAI are those at high risk of knee OA (KOA). It is likely that different individuals will develop OA at different rates.

This cohort contained relatively small numbers of participants who had injury or surgery to that knee prior to the onset of ROA. However, the association between synovial inflammation and the occurrence of ROA persisted in models adjusted for history of knee injury and history of knee surgery at baseline. The fact that majority of participants developed synovial inflammation independent of a documented insult to the knee suggesting that inflammation may occur either independently or as a result of a clinically insignificant pathology within the knee. But the current study is not able to prove whether synovitis or effusion precedes structural joint damage or is a result of early intrinsic tissue alterations such as meniscal or cartilage damage.

The next observation is that the strength of associations at each time point was greater for effusion-synovitis than for Hoffa-synovitis. This may be due to the reduced sensitivity of Hoffa-synovitis as a measure of synovial inflammation.26 Effusion-synovitis on non-contrast-enhanced MRI is perceived as a better assessment of synovial inflammation, as it circumvents the lack of specificity of using Hoffa-synovitis as an index of synovial inflammation.13

The following limitations were identified in this research project. The gold standard of identifying synovitis on imaging is contrast-enhanced MRI. However, because of possible side effects, ethical considerations and increased expense, contrast-enhanced MRI is not routinely used in the assessment of pre-symptomatic OA in large-scale clinical and epidemiological research. Instead, synovial inflammation is assessed in non-contrast-enhanced MRI using surrogate signal changes in Hoffa's fat pad (Hoffa-synovitis) and by effusion-synovitis.27 ,28 We assert that the non-contrast-enhanced MRI is a useful tool in the detection of the onset of synovial inflammation in pre-symptomatic OA, particularly in the face of ethical and practical considerations in using contrast in pre-symptomatic OA cohorts. In addition, this study retrospectively assessed subjects who did not have ROA at baseline. The retrospective study design is a disadvantage, but it was the most feasible to be used in this previously unexplored area of research.

Furthermore, Hoffa's fat pad synovitis on non-contrast-enhanced MRI has been shown to correlate histologically with synovitis in the same location.29 Intra-articular fluid signals are assessed using non-contrast-enhanced MRI according to the extent of capsular distension.30 This has been shown to be a composite of true joint effusion and synovial thickening and may underestimate synovial inflammation.31 This study used these previously validated tools to identify synovitis. However, it is possible that synovitis may be underestimated in the non-contrast-enhanced MRIs.32 Therefore, we reiterate that the use of contrast-enhanced MRI would likely strengthen rather than diminish any association detected by non-contrast-enhanced MRI in this study.27

The OAI study population contains persons at high risk of KOA, and both cases and controls in this study were predominately overweight or obese (24% of cases and 29% of controls were in normal BMI ranges). Therefore, this study is unable to identify whether Hoffa-synovitis and effusion-synovitis is associated with increasing weight.

This study is the first, to the best of our knowledge, which demonstrates the association of synovitis with incident KOA. These findings concur with previous research findings that demonstrated the pivotal role of synovitis in the development and prevalence of symptoms and progression of knee OA. It is noteworthy that these studies demonstrated the association of clinical features of joint inflammation with radiological progression in knee OA.33 Further, an MRI study has demonstrated a strong and unique association between synovitis, capsular thickening and the severity of knee pain.32 In addition, higher grades of synovitis conferred an increased risk of painful knee OA.9 Similarly, an increase in levels of synovial biomarkers associated with synovial activity, serum hyaluronan, have demonstrated an association with radiographic progression of the knee.10 Laboratory studies have established that pro-inflammatory cytokines contribute to OA pathogenesis by increasing cartilage degradation. Furthermore, inflammation produces pain and hyperalgesia in OA via multiple mechanisms.34 Several inflammatory markers including interleukin (IL)-1, IL-6, transforming growth factor (TGF)-β, tumour necrosis factor-α. IL-17 and vascular endothelial growth factor are increased in the synovium in OA.12 Increasing synovial IL-17 levels has been shown to correlate with knee OA severity and progression.35 In addition, inhibitors of the growth factors, TGF-β and bone morphogenetic protein, reduced synovial thickening and osteophyte formation.36 Therefore, it is intuitive that treatment of synovitis may improve symptoms and disease progression.

There is evidence to support the postulate that treatment of synovitis may improve symptoms. Yttrium-90 synovectomy37 used in treating painful knee OA showed a better response in early knee OA. Therefore, treating synovitis in knee OA may have a role in reducing progression of the disease and provide symptom relief. Therefore, the results of this study are useful in identifying a targeted therapy to alter the pathogenesis of knee OA.

Our study has demonstrated that synovitis is a precursor of radiographic OA. These findings, if confirmed by similar studies including those with synovial biopsies, will establish the role of inflammation in the pathogenesis of OA. The findings have the potential to be translated towards developing targeted therapies to prevent the development of knee OA.


The authors would like to thank the readers of the fixed flexion radiographs at Boston University for the central KL grading, the OAI investigators, clinic staff and OAI participants at each of the OAI clinical centres for their contributions in acquiring the publicly available clinical and imaging data, the team at the OAI coordinating centre, particularly John Lynch, Maurice Dockrell and Jason Maeda, for their help, and Stephanie Green and Hilary Peterson at Pittsburgh for administrative support.


View Abstract


  • Handling editor Tore K Kvien

  • Contributors IA and DJH conceived and designed the study, supervised its conduct, drafted the manuscript and take responsibility for the integrity of the work as a whole, from inception to finish. CKK, AG, FR, RMB and MJH were also involved in the design and conduct of the POMA study. All authors contributed to acquisition of the data and its interpretation. All authors critically revised the manuscript and gave final approval of the article for submission.

  • Funding IA was funded by the Osteoarthritis Research Society International Collaborative Scholarship. DJH is funded by an NHMRC Health Practitioner Fellowship. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication. The study and image acquisition was funded by the OAI, a public–private partnership comprised of five contracts (N01-AR-2-2258; N01-AR-2-2259; N01-AR-2-2260; N01-AR-2-2261; N01-AR-2-2262) funded by the National Institutes of Health, a branch of the Department of Health and Human Services, and conducted by the OAI Study Investigators. Private funding partners of the OAI include Merck Research Laboratories; Novartis Pharmaceuticals Corporation, GlaxoSmithKline; and Pfizer. Private sector funding for the OAI is managed by the Foundation for the National Institutes of Health. The image analysis of this study was funded by a contract with the University of Pittsburgh (Pivotal OAI MRI Analyses POMA: NIH/NHLBI Contract No. HHSN2682010000 21C) and, in part, by a vendor contract from the OAI coordinating centre at University of California, San Francisco (N01-AR-2-2258). The statistical data analysis was funded by a contract with the University of Pittsburgh (Pivotal OAI MRI Analyses POMA: NIH/NHLBI Contract No. HHSN2682010000 21C) and the University of Pittsburgh Multidisciplinary Clinical Research Center (MCRC) for Rheumatic and Musculoskeletal Diseases (P60 AR054731). None of the study sponsors had any role in data collection, storage or analysis, in manuscript writing or the decision to publish this manuscript.

  • Competing interests AG is President and co-owner of the Boston Core Imaging Lab (BICL), a company providing MRI reading services to academic researchers and to industry. He has provided consulting services to Novartis, Merck Serono, Sanofi-Aventis, TissueGene and Genzyme. FWR is CMO and co-owner of the Boston Core Imaging Lab (BICL), a company providing MRI reading services to academic researchers and to industry. He has provided consulting services to Merck Serono. CKK has provided consulting services to Novartis and has received research support from Astra-Zeneca.

  • Ethics approval This study received ethical approval from each OAI clinical site (Memorial Hospital of Rhode Island Institutional Review Board, Ohio State University's Biomedical Sciences Institutional Review Board, University of Pittsburgh Institutional Review Board, and University of Maryland Baltimore—Institutional Review Board), the OAI coordinating centre (Committee on Human Research at University of California, San Francisco), as well as the Institutional Review Board at Tufts Medical Center and Tufts University Health Sciences Campus. This manuscript has received the approval of the OAI Publications Committee based on a review of its scientific content and data interpretation.

  • Provenance and peer review Not commissioned; externally peer reviewed.