Nonsurgical Knee Osteoarthritis Treatments for Reducing Inflammation as Measured on MRI Scans: A Systematic Review of Randomized Controlled Trials

Abstract

Background:

Knee osteoarthritis (OA) is a debilitating condition, and synovitis is a structural marker of disease progression that can be identified on magnetic resonance imaging (MRI). Nonsurgical therapies have been developed with the goal of targeting this inflammation to reduce pain and slow disease progression.

Purpose:

To review current randomized controlled trials (RCTs) that measured changes in pain outcomes and synovitis on MRI scans after nonsurgical treatment for persons with knee OA.

Study Design:

Systematic review; Level of evidence, 1.

Methods:

The PubMed database was searched using the terms “knee osteoarthritis AND (synovitis OR inflammation)” for RCTs published between 2012 and 2022. Included were studies that collected both pain outcomes and quantitative measurements of synovitis on MRI scan before and after treatment; studies that investigated surgical treatments were excluded. We calculated standardized response means (SRMs) to analyze the effect sizes of treatment on pain and synovitis outcomes; SRMs were classified as having low responsiveness to treatment if <0.5, moderate responsiveness to treatment if between 0.5 and 0.8, and large responsiveness to treatment if >0.8.

Results:

Included in the review were 9 RCTs. Vitamin D and exercise therapy were the only 2 treatments that improved both pain and synovitis outcomes in comparison with their respective control groups. Statistically significant group differences in favor of treatment were seen in changes in pain after lutikizumab (anti-interleukin-1α/β dual variable domain immunoglobulin; SRM, 0.22; P = .050), allogenic human chondrocytes transduced with retroviral vector encoding transforming growth factor-β1 (P = .0119 at 12 weeks, P = .120 at 52 weeks, and P = .0074 at 72 weeks), and Curcuma longa (turmeric; SRM, 0.35; P = .039 on the visual analog scale for pain and SRM, 0.47; P = .006 on the Western Ontario and McMaster Universities Osteoarthritis Index pain subscore). One study reported a statistically significant group difference in synovitis only after treatment with intra-articular methylprednisolone (P = .01 at 14 weeks and P = .0006 at 26 weeks).

Conclusion:

Only vitamin D and exercise therapy were found to improve pain and synovitis after treatment in this review. Further research is needed to validate these findings and investigate alternative treatments for reducing pain and synovitis in persons with knee OA.

Keywords

knee osteoarthritis inflammation MRI synovitis

Knee osteoarthritis (OA) is a debilitating condition that affects nearly 14 million persons in the United States, including >6 million people between 45 and 65 years of age who often wish to engage in sporting activity.⁸ The worldwide prevalence of knee OA is estimated at >654 million persons aged ≥40 years.⁵ Adults living with knee OA experience chronic pain and attendant physical limitations.² The pathogenesis of knee OA includes degradation of the articular cartilage, subchondral bone thickening, and formation of osteophytes.²² Mechanical factors, including malalignment of the knee joint and varus thrust, contribute to knee OA disease progression.^3,34 Although knee OA was regarded historically as a noninflammatory arthritis, synovitis (inflammation of the synovium) has also been shown to be important in the pathogenesis of OA.²⁷

Magnetic resonance imaging (MRI) has been established as a method for reliably and quantitatively measuring structural changes in knee OA,¹⁷ including synovitis.²³ While inflamed synovium is best appreciated on contrast-enhanced MRI scans, effusion volume and inflammation in the Hoffa fat pad have been used as proxy measures of synovitis that can be assessed with noncontrast MRI studies.^16,31 Studies of the associations between synovitis on MRI scans and pain in persons with knee OA have been inconsistent.^6,7,14,15,36 In a meta-analysis, Dainese et al⁶ observed considerable variability in reported correlations between pain and effusion synovitis (ES) on noncontrast-enhanced MRI scans, with correlation coefficients ranging from 0.19 to 0.61. Neogi et al³⁰ reported that extent of “synovitis and effusion” observed on noncontrast MRI scans were statistically and significantly associated with an increase in pain sensitization.

Because synovitis is an element of knee OA pathogenesis, investigators have attempted to target and diminish synovitis with therapeutic agents to slow structural progression and reduce pain.³² We are not aware of published reviews investigating the effect of knee OA treatments on synovitis as measured on MRI scans. To fill this research gap, we launched this systematic review of the effect of OA treatments on synovitis and pain in persons with knee OA. The objective of this review was to identify treatments for knee OA that reduced both pain and synovitis markers on MRI scans, and the hypothesis was that few, if any, treatments would statistically significantly reduce both pain and synovitis.

Methods

Selection of Studies

Under Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, PubMed was searched using the terms “knee osteoarthritis AND (synovitis OR inflammation)” for randomized controlled trials (RCTs) published between 2012 and 2022. The first author (J.G.) completed the search and reviewed all of the studies, applying the inclusion and exclusion criteria (Table 1). Duplicate studies were screened out in the initial search. To be included in the review, the study needed to be a clinical trial or subanalysis of a trial in which patients with knee OA were administered a treatment to improve their symptoms. The trial needed to collect both pain outcomes and quantitative measurements of synovitis (Hoffa fat pad synovitis [HS] and/or ES) on MRI scan. These pain and imaging outcomes had to be assessed before and after treatment to investigate the relationship between the treatment and both outcomes (pain and effusion or HS on MRI scan). Excluded were studies that investigated surgical therapies, were not published in English, did not include a control group, or were animal studies. We also excluded reviews and letters to the editor. After the inclusion and exclusion criteria were applied, we reviewed all references from the remaining studies for additional RCTs.

Table 1

Inclusion and Exclusion Criteria for Studies^a

Inclusion Criteria	Exclusion Criteria
1. Nonsurgical treatment administered in randomized controlled trial 2. Outcome measurements assessed include knee synovitis measured on MRI scan and patient-reported pain 3. Synovitis and pain outcomes assessed at baseline and follow-up 4. Clinical trial 5. Patients must have knee osteoarthritis	1. Not in English language 2. Synovitis measured on biomarkers other than MRI scan 3. Results did not include effect of treatment on synovitis and pain 4. No control group included in the trial

MRI, magnetic resonance imaging.

Statistical Analysis

To analyze changes in outcomes (synovitis measured on MRI scan and pain scores) for each study, we calculated the standardized response means (SRMs) for studies that reported mean changes from baseline to follow-up.²¹ SRMs were calculated by first recording the mean preintervention-to-postintervention changes in the treatment and control groups. We then divided the difference in these mean changes by the standard deviation of the mean changes. The standard deviation of the difference between mean changes was calculated by summing the square of the standard deviation from the mean change in pain from the treatment group and the square of the standard deviation from the mean change in pain from the control group and then taking the square root of this sum. When studies did not report standard deviations of the mean changes, the 95% CIs reported was used to derive these values. Standard deviations were calculated by multiplying the half-width of the 95% CI around the within-group change by the square root of the number of subjects in the group and dividing by 1.96. The square root of the total number of subjects in the respective group (treatment or control) was the numerator.¹³ If standard errors were reported instead of standard deviations, the standard deviations were derived by multiplying the standard error by the square root of the total number of subjects in the respective group. When studies reported least-square means (LSMs) and least-square mean differences (LSMDs) instead of means and mean differences, we used the LSM instead of the mean difference and LSMD instead of the difference between means to calculate the SRMs, following the same calculations described previously.

SRMs were classified as having low responsiveness to treatment if <0.5, moderate responsiveness to treatment if between 0.5 and 0.8, and large responsiveness to treatment if >0.8.⁴

Results

Search Results

The initial search yielded 189 results. After reviewing the abstracts and applying the inclusion and exclusion criteria, 166 studies were excluded. After reviewing the full text of the remaining 23 studies, 14 studies were excluded, leaving 9 studies that were included in the systematic review for analysis.^# The review of the reference lists of these selected 9 studies did not provide any additional articles that met all the inclusion criteria. A flowchart of the study-inclusion process is shown in Figure 1. Table 2 provides a summary of each study, including the total number of participants in each trial, participant characteristics, distribution of participants in the treatment and control groups. Table 3 describes the treatment interventions used in each study.

Figure 1.

Preferred reporting items for systematic reviews and meta-analyses[PRISMA] flowchart of the study-inclusion process. MRI, magnetic resonance imaging; RCT, randomized controlled trial.

Table 2

Characteristics of the Included RCTs^a

First Author (Year)	Sample Size	Age, y^b	Treatment Group, n	Control Group, n
Bandak¹ (2021)	60	65.9 (8.5)	31	29
Fleischmann⁹ (2019)^c	350	• Placebo: 59.5 (8.9) • 25 mg: 61.6 (7.5) • 100 mg: 60.2 (8.2) • 200 mg: 59.1 (10.3)	• 25 mg: 89 • 100 mg: 85 • 200 mg: 88	85
Henriksen¹² (2015)	100	63.4 (9.3)	50	50
Lee²⁰ (2020)^d	102	• Treatment: 57 • Control: 56	67	35
Mahler²⁶ (2020)	55	• Treatment: 62 (9) • Control: 68 (9)	27	28
McAlindon²⁸ (2017)	140	• Treatment: 59.1 (8.3) • Control: 57.2 (7.6)	70	70
Riis³³ (2017)	100^e	62.2 (9.4)	50	50
Wang³⁵ (2017)^f	413	63.2 (7.0)	209	204
Wang³⁷ (2020)	70	• Treatment: 61.3 (8.5) • Control: 62.4 (8.8)	36	34

MRI, magnetic resonance imaging; RCT, randomized controlled trial.

Data are reported as means ± SD.

Of the 350 participants who were randomized, 347 participants received ≥1 doses of the study drug and were included in efficacy and safety analyses.⁹

All participants had Kellgren-Lawrence grade 3 osteoarthritis of the knee.²⁰

A total of 91 patients were included in the analysis because 9 of the 100 randomized patients did not have baseline MRI scans.³³

The authors reported the results of the Vitamin D Effect on Osteoarthritis (VIDEO) study; results were also reported by Jin et al,¹⁸ not included in this review.

Table 3

Treatment Interventions^a

First Author (Year)	Treatment Group	Control Group^b	Mode of Administration^c	Dosage and Frequency of Administration
Bandak¹ (2021)	Exercise therapy	No attention control group	Supervised functional and individual exercise therapy^d	Exercise 3 times/week for 12 weeks
Fleischmann⁹ (2019)	Lutikizumab (anti-interleukin-1α/β dual variable domain immunoglobulin)	Matching placebo	Subcutaneous	25-, 100-, or 200-mg dose every 2 weeks for 52 weeks
Henriksen¹² (2015)	Methylprednisolone acetate (Depo-Medrol) and exercise	Saline and exercise	Intra-articular injection and supervised functional and individual exercise therapy^d	A single injection (40 mg dissolved in 4 mL of lidocaine hydrochloride) and exercise starting 2 weeks after injection (3 times/week for 12 weeks)
Lee²⁰ (2020)	Allogenic human chondrocytes transduced with a retroviral vector encoding TGF-β1	Saline	Intra-articular injection	A single 2-mL injection of a 3:1 mixture of nontransduced allogenic human chondrocytes expressing TGF-β
Mahler²⁶ (2020)	Low-dose radiation therapy	Sham intervention	Radiation	6 Gy (0 Gy in sham group), once every other weekday for 2 weeks
McAlindon²⁸ (2017)	Triamcinolone	Sodium chloride	Intra-articular injection	A single injection (1 mL of triamcinolone, 40 mg/mL) every 12 weeks for 2 years
Riis³³ (2017)	Methylprednisolone and exercise	Saline and exercise	Intra-articular injection	1 mL methylprednisolone or saline in 4 mL lidocaine and exercise starting 2 weeks after injection (3 times/week for 12 weeks)
Wang³⁵ (2017)	Vitamin D	Inert placebo pill	Oral	Monthly capsule (1.25 mg) for 24 months
Wang³⁷ (2020)	Curcuma longa (turmeric)	Identical placebo pill	Oral	1000 mg/dL per day for 12 weeks

TGF, transforming growth factor.

In all studies except for Bandak et al,¹ participants were blinded to their treatment group. In all studies, outcome assessors were blinded to treatment group.

In all studies, the experimental and control groups were identically administered treatments.

Therapy supervised by a physiotherapist.¹²

Two trials reported greater reduction in both pain and synovitis outcomes in the treatment group compared with the control group (Jin et al¹⁸ and Wang et al³⁵ both reported results for the same trial, the Vitamin D Effect on Osteoarthritis [VIDEO] Study)^1,18,35; 3 studies reported better pain outcomes but no difference in synovitis outcomes in the treatment group compared with the control group^9,20,37; 1 study reported better outcomes in synovitis but no difference in pain in the treatment group compared with the control group³³; and the remaining studies showed no difference in either the pain outcome or synovitis outcome.^12,26,28

A range of measures was used to report changes in synovitis on MRI scan, including the MOAKS (MRI in OA Knee Score),^1,12,33 WORMS (Whole-Organ Magnetic Resonance Imaging Score),⁹ ES volume,^28,35,37 synovial membrane thickness,⁹ BLOKS (Boston-Leeds OA Knee Score),³³ and contrast-enhanced and dynamic contrast-enhanced MRI (CE-MRI and DCE-MRI)^1,33 (Tables 4 and 5). Further, authors assessed either ES^9,28,35,37 or both ES and HS on MRI scan.^{1,12,20,26,33}

Table 4

Reported Mean Changes in Synovitis on MRI Scan^a

Change in Synovitis, Mean (SD)
Study	Follow-up	Synovitis Measures	Treatment Group	Control Group	Difference Between Groups, Mean (SD)	P (SRM)^b
Henriksen¹² (2015)	14, 26 weeks	• ES: MOAKS (0-3)^d • HS: extent of hyperintense signal changes on midsagittal fluid-sensitive sequences (0-3) ^d	• ES 14 weeks: -0.3 (0.71) 26 weeks: -0.2 (0.71) • HS 14 weeks: -0.3 (0.71) 26 weeks: -0.3 (0.71)	• ES 14 weeks: -0.2 (0.71) 26 weeks: -0.6 (0.71) • HS 14 weeks: -0.2 (0.71) 26 weeks: -0.7 (0.71)	• ES 14 weeks: 0.1 (1.00) 26 weeks: -0.4 (1.00) • HS 14 weeks: 0.1 (1.00) 26 weeks: -0.4 (1.00)	• ES 14 weeks: .54 (0.1) 26 weeks: .02 (-0.4) • HS 14 weeks: .62 (0.1) 26 weeks: .04 (-0.4)
McAlindon²⁸ (2017)^c	24 mo	ES volume, mL	-0.09 (1.45)	-0.32 (0.98)	0.23 (1.75)	0.17
Riis³³ (2017)	14,^f 26 weeks	• ES and HS on MOAKS (0-6)^e • BLOKS effusion (0-3)^d • CE-MRI (0-22)^d	• MOAKS 14 weeks: -0.38 (0.67) 26 weeks: -0.45 (0.63) • BLOKS 14 weeks: -0.09 (0.43) 26 weeks: -0.02 (0.40) • CE-MRI 14 weeks: -0.91 (2.08) 26 weeks: -0.8 (1.96)	• MOAKS 14 weeks: 0.03 (0.73) 26 weeks: 0.12 (0.62) • BLOKS 14 weeks: 0 (0.46) 26 weeks: 0.04 (0.40) • CE-MRI 14 weeks: -0.5 (2.25) 26 weeks: -0.78 (1.93)	• MOAKS 14 weeks: 0.41 (0.99) 26 weeks: 0.57 (0.89) • BLOKS 14 weeks: 0.09 (0.63) 26 weeks: 0.07 (0.57) • CE-MRI 14 weeks: 0.42 (3.06) 26 weeks: 0.02 (2.75)	• MOAKS 14 weeks: .01 (0.41) 26 weeks: .0006 (0.64) • BLOKS 14 weeks: .38 (0.14) 26 weeks: .52 (0.12) • CE-MRI 14 weeks: .40 (0.14) 26 weeks: .96 (0.01)
Wang³⁵ (2017)	24 mo	• AC in ES, mL; • PA	• AC: 0.26 (7.97); • PA: 16% (169.65)	• AC: 2.20 (8.60); • PA: 60% (211.33)	• AC: -1.94 (11.72); • PA: -45% (271)	• AC: .02 (0.17); • PA: .02 (0.17)
Wang³⁷ (2020)	12 weeks	ES volume, mL	1.1 (6.25)	-2.1 (7.22)	3.2 (9.54)	.075 (-0.34)

Boldface text indicate variables with statistically significant improvement in synovitis (P < .05). AC, absolute change; AS, anterior synovium; BLOKS, Boston-Leeds OA Knee Score; CE-MRI, contrast-enhanced MRI; ES, effusion synovitis; HS, Hoffa synovitis; MRI, magnetic resonance imaging; OA, osteoarthritis; MOAKS, MRI in OA Knee Score; PA, percentage change per annum; PS, posterior synovium; SRM, standardized response mean; TS, total synovium.

The SRM was calculated by dividing the difference between means by the SD of the difference between means. The SD of difference between means were calculated by taking the square root of the sum of the SD of the treatment group squared and the SD of the control group squared.

Changes in ES volume reported as logs.

Higher scores reflect worse synovitis.

ES and HS were graded separately (0-3) for a total summed MOAKS synovitis score (0-6), with higher scores indicating worse synovitis.

Primary timepoint for MRI outcome.

Table 5

Reported LSM Changes in Synovitis on MRI Scan^a

			LSM (SD)^b
Study	Follow-up	Synovitis Measures	Treatment Group	Control Group	Difference Between Groups (SD)^b	P (SRM)^c
Bandak¹ (2021)	12 weeks	• MOAKS (0-6)^d • CE-MRI (0-22)^e	• MOAKS: 0.26 (0.56) • CE-MRI: 0.77 (1.8)	• MOAKS: -0.18 (0.58) • CE-MRI: -0.31 (1.77)	• MOAKS: -0.44 (0.80) • CE-MRI: -1.08 (2.53)	• MOAKS: .052 (-0.55) • CE-MRI: -.122 (-0.43)
Fleischmann⁹ (2019)	26,^{f^} 52 weeks	• Synovial membrane thickness, mm • Synovial fluid volume, mL • WORMS (0-3)	• Synovial membrane thickness - 26 weeks: 25 mg, 0.01 (0.29); 100 mg, -0.08 (0.28); 200 mg: 0.01 (0.29) - 52 weeks: 5 mg, -0.04 (0.29); 100 mg, -0.05 (0.28); 200 mg, -0.02 (0.34) • Synovial fluid volume - 26 weeks: 25 mg, 0.26 (11.36); 100 mg, -1.04 (11.24); 200 mg, -1.49 (11.44) - 52 weeks: 25 mg, 1.17 (14.20); 100 mg, -0.67 (13.74); 200 mg, -1.83 (14.31) • WORMS - 26 weeks: 25 mg, -0.01 (0.58); 100 mg, -0.08 (0.56); 200 mg, -0.07 (0.57) - 52 weeks: 25 mg, -0.05 (0.63); 100 mg, -0.06 (0.66); 200 mg, -0.01 (0.62)	• Synovial membrane thickness - 26 weeks: -0.05 (0.28) - 52 weeks: -0.07 (0.28) • Synovial fluid volume - 26 weeks: 0.03 (11.90) - 52 weeks: -1.90 (14.72) • WORMS - 26 weeks: 0.07 (0.56) - 52 weeks: -0.05 (0.66)	• Synovial membrane thickness - 26 weeks: 25 mg, 0.06 (0.40); 100 mg, -0.03 (0.55); 200 mg, 0.06 (0.40) - 52 weeks: 25 mg, 0.03 (0.47); 100 mg, 0.02 (0.47); 200 mg, 0.06 (0.50) • Synovial fluid volume - 26 weeks: 25 mg, 0.22 (16.45); 100 mg, -1.07 (16.37); 200 mg, -1.52 (16.51) - 52 weeks: 25 mg, 3.08 (20.45); 100 mg, 1.23 (20.14); 200 mg, 0.07 (20.53) • WORMS - 26 weeks: 25 mg, -0.08 (0.81); 100 mg, -0.15 (0.80); 200 mg, -0.14 (0.81) - 52 weeks: 25 mg, 0.00 (0.91); 100 mg, -0.01 (0.93); 200 mg, 0.04 (0.91)	• Synovial membrane thickness - 26 weeks: 25 mg, .145 (-0.15); 100 mg, .520 (0.08); 200 mg, .159 (-0.15) - 52 weeks: 25 mg, .474 (-0.07); 100 mg, .637 (-0.05); 200 mg, .221 (-0.14) • Synovial fluid volume - 26 weeks: 25 mg, .897 (-0.01); 100 mg, .542 (0.07); 200 mg, .385 (0.09) - 52 weeks: 25 mg, .154 (-0.15); 100 mg, .569 (-0.06); 200 mg, .973 (0.00) • WORMS - 26 weeks: 25 mg, .384 (0.10); 100 mg, .095 (0.19); 200 mg, .106 (0.17) - 52 weeks: 25 mg, .967 (0.00); 100 mg, .923 (0.01); 200 mg, .692 (-0.04)

CE-MRI, contrast-enhanced MRI; CI, confidence interval; ES, effusion synovitis; HS, Hoffa synovitis; LSM, least-squares mean; LSMD, least-squares mean difference; MOAKS, MRI in OA knee score; MRI, magnetic resonance imaging; OA, osteoarthritis; SRM, standardized response mean; WORMS, whole-organ MRI score.

The SD was calculated by dividing the product of the square root of the total participants in the specific group and the half-width of the CI by 1.96 (CIs were provided in the studies). If the SE was provided instead of the SD, the SD was calculated by multiplying the square root of the total participants in the specific group and the SE.

The SRM was calculated by dividing the difference between means by the SD of the difference between means. The SD of the difference between groups were calculated by taking the square root of the sum of the SD of the treatment group squared and the SD of the control group squared.

ES and HS were graded separately (0-3) for a total summed MOAKS synovitis score (0-6), with higher scores indicating worse synovitis.

Proposed by Guermazi et al¹⁰ using postgadolinium sagittal T1-weighted turbo spin-echo MRI.

Primary timepoint for MRI outcome.

Vitamin D

The VIDEO Study, which was analyzed by Wang et al³⁵ and Jin et al¹⁸ (not included in the review) was the only study that reported a significantly greater improvement in both pain and synovitis outcomes in the treatment group compared with the control group. In post hoc analyses, Jin et al¹⁸ found that participants in the vitamin D group had a significant improvement in pain measured on visual analog scale (VAS) for pain when compared with participants in the placebo group (P = .048). Furthermore, this effect was evident only in patients with an ES score of ≥2 at baseline and not in patients without ES at baseline (ES was measured on a 0-3 scale, with higher numbers reflecting worse synovitis). Wang et al³⁵ reported significant group differences for both absolute change and percentage change per year in ES (P = .02 for both), with values remaining stable in the treatment group but increasing in the placebo group; however, the SRM value for synovitis was low (0.17) (Table 4). An SRM for pain was unable to be calculated because mean changes were not provided and only the P value was provided.³⁵ Jin and colleagues¹⁸ reported the percentage of patients that achieved either 20% or 50% improvement in Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC)-Pain subscores in the VIDEO Study (patient cohort analyzed by Wang et al³⁵ for changes in synovitis on MRI scan). At 24 months, 64% of participants in the vitamin D group and 57% of participants in the placebo group achieved a 20% improvement in WOMAC-Pain (P = .16), and 50% of participants in the vitamin D group and 39% of participants in the placebo group achieved a 50% improvement in WOMAC-Pain (P = .04, statistically significant difference).¹⁸

Exercise Therapy

Bandak et al¹ reported a significant group difference in synovitis assessed on DCE-MRI in only the anterior synovium, across all measurements. These measurements included highly perfused voxels (P = .024), the initial rate of enhancement composite score (P = .044), and the maximal enhancement composite score (P = .015). The authors also reported that the exercise group had a significantly greater improvement in Knee injury and Osteoarthritis Outcomes Score (KOOS)-Pain subscores compared with the control group, with a group mean difference of -11.7 points (P = .0075) (Table 6).¹ The calculated SRM was 0.68, indicating moderate responsiveness of the outcome measure (improvement in KOOS-Pain).⁴

Table 6

Reported LSM Changes in Pain After Treatment^a

			LSM (SD)^b
Study	Follow-up	Pain Outcomes	Treatment Group	Control Group	Difference vs Placebo (SD)^b	P (SRM)
Bandak¹ (2021)	12 weeks	Change in KOOS-Pain (0-100; 0 is more pain)	10.6 (12.4)	-1.1 (11.96)	-11.7 (17.23)	.0075 (0.68)
Fleischmann⁹ (2019)	16,^c 52 weeks	Change in WOMAC-Pain (0-50; 50 is more pain)	• 16 weeks: 25 mg, -9.2 (9.63); 100 mg, -11.8 (9.41); 200 mg, -10.1 (9.57) • 52 weeks: 25 mg, -11.0 (10.59); 100 mg, -12.1 (10.82); 200 mg, -12.2 (10.53)	• 16 weeks: -8.9 (9.41) • 52 wk: -10.0 (83.26)	• 16 weeks: 25 mg, -0.3 (13.46); 100 mg, -2.9 (13.30); 200 mg, -1.2 (13.42) • 52 weeks: 25 mg, -1.1 (15.14); 100 mg, -2.2 (15.30); 200 mg, -2.3 (15.10)	• 16 weeks: 25 mg, .834 (-0.02); 100 mg, .050 (0.22); 200 mg, .415 (0.09) • 52 weeks: 25 mg, .500 (0.01); 100 mg, .186 (0.03); 200 mg, .157 (0.03)
Lee²⁰ (2020)	12, 52, 72 weeks	VAS pain	NA^d	NA^d	• 12 weeks: -14.4 • 52 weeks: -15.5 • 72 weeks: -16.6	• 12 weeks: .0119 (NA)^d • 52 weeks: .120 (NA)^d • 72 weeks: .0074 (NA)^d

Boldface text indicate variables with statistically significant improvement in pain (P < .05). LSM, least-squares mean; NA, not available; SRM, standardized response mean.

SDs were calculated by dividing the product of the square root of the total participants in the specific group and the half-width of the CI by 1.96 (CIs were provided in the studies). If SEs were provided instead of SDs, SDs were calculated by multiplying the square root of the total participants in the specific group and the SE. SRMs were calculated by dividing the difference between means by the SD of the difference between means. The SD of the difference between groups were calculated by taking the square root of the sum of the SD of the treatment group squared and the SD of the control group squared.

Primary timepoint for pain outcome.

The authors did not provide LSMs and SEs or SDs for either of the study groups; thus, SRMs were unable to be calculated.

Lutikizumab (Anti-Interleukin-1α/β Dual Variable Domain Immunoglobulin)

Fleischman et al⁹ reported LSMDs for synovitis, with no significant improvement in synovitis after treatment (Table 5). Of the studies that reported the LSMDs in pain between treatment and control group after treatment, Fleischman et al⁹ showed that a 100 mg dose of lutikizumab, an anti-interleukin-1α/β dual variable domain immunoglobulin, led to significantly improved WOMAC-Pain scores at 16 weeks (P = .050) (Table 6). The calculated SRM for this dose and timepoint was 0.22 - a low between-group effect size.⁴

Allogenic Human Chondrocytes Transduced With Retroviral Vector Encoding TGF-β1

Lee et al²⁰ did not find significant differences between groups, although whole-knee MRI scans at 12 months showed less progression of infrapatellar fat pad synovitis and ES in the treatment group compared with placebo (9.6% vs 21.1%). Although the authors did not provide LSMs and LSMDs to calculate the SRM, they identified significant group differences in VAS pain scores at 12 and 72 weeks (P = .0119 and P = .0074, respectively) (Table 7).

Table 7

Reported Mean Changes in Pain After Treatment^a

			Mean Change (SD)
Study	Follow-up	Pain Outcomes	Treatment Group	Control Group	Difference Between Means (SD)	P (SRM)
Henriksen¹² (2015)^d	2, 14 weeks^c	Change in KOOS-Pain (0-100; 100 is less pain)	• 2 weeks: 5.0 (12.7) • 14 weeks: 13.6 (12.7)	• 2 weeks: 4.5 (12.7) • 14 weeks: 14.8 (12.7)	• 2 weeks: -0.5 (18) • 14 weeks: 1.2 (18)	• 2 weeks: .85 (0.028) • 14 weeks: .64 (-0.067)
Mahler²⁶ (2020)	12 weeks	Change in WOMAC (0-100; 100 is less pain)	8 (13)	11 (14)	-3 (19.1)	NA (-0.16)
McAlindon²⁸ (2017)	24 mo	• Change in WOMAC-Pain (0-20; 0 is no pain) • Change in VAS pain (0-100; 0 is no pain)	• WOMAC: −1.2 (2.65) • VAS: −2.7 (39.70)	• WOMAC: −1.9 (2.86) • VAS: −7.6 (33.12)	• WOMAC: −0.64 (3.90) VAS: −5.0 (51.70)	• WOMAC: .17 (-0.16) • VAS: .26 (-0.10)
Riis³³ (2017)	14, 26 weeks	Change in KOOS-Pain (0-100; 100 is no pain)	• 14 weeks: 14.4 (12.86) • 26 weeks: 14.2 (12.86)	• 14 weeks: 14.8 (13.56) • 26 weeks: 15.4 (13.89)	• 14 weeks: 0.4 (18.68) • 26 weeks: 1.2 (18.93)	• 14 weeks: .89 (-0.02) • 26 weeks: .68 (-0.06)
Wang³⁷ (2020)	12 weeks	• Change in VAS (0-100; 0 is no pain) • Change in WOMAC (0-500; 0 is no pain)	• VAS: -23.8 (18.41) • WOMAC: -84.2 (70.63)	• VAS: -14.6 (18.15) • WOMAC: -37 (71.70)	• VAS: -9.1 (25.85) • WOMAC: -47.2 (100.64)	• VAS: .039 (0.35)^b • WOMAC: .006 (0.47)^b

KOOS, Knee injury and Osteoarthritis Outcome Score; SRM, standardized response mean; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index; NA, not available.

Statistically significant improvement in pain.

Primary timepoint for pain outcome.

The authors reported SEM instead of SD, and SD were derived by dividing the SE by the number of participants in the respective treatment group.

Intra-Articular Methylprednisolone and Exercise

Among the studies reporting mean changes in synovitis,^{12,28,33,35,37} Riis et al³³ reported a significant group difference for the mean change from baseline in MOAKS for synovitis at weeks 14 and 26 (P = .01 and P = .0006, respectively) in favor of the treatment group. The authors also reported that change in synovitis on CE-MRI was significantly correlated with the change in KOOS-Pain scores at 26 weeks (r = -0.36).³³

Low-Dose Radiation Therapy

Mahler et al²⁶ did not find any group differences in ES and HS from the median and interquartile range values (both measured on scales ranging from 0 to 3, with higher scores indicating worse synovitis).

Curcuma longa Extract

Wang et al³⁷ reported significant differences in favor of the treatment group for pain on the VAS (P = .039) and WOMAC (P = .006). However, the SRMs on the VAS (SRM = 0.35) and WOMAC (SRM = 0.47) were low (Table 7).

Discussion

In this review of 9 RCTs that investigated nonsurgical treatments for knee OA,^# only 2 treatments improved both pain and synovitis outcomes in comparison with the control group: vitamin D and exercise therapy.^1,35 Of note, Bandak et al¹ showed a significant group difference for synovitis only when measured on DCE-MRI in the anterior synovium but not the total knee. Further, the VIDEO Study showed a significant improvement in VAS pain scores only in the vitamin D group in post hoc analyses.¹⁸ Three different treatments showed a significant improvement in pain after treatment compared with the control group but not in synovitis; these treatments included lutikizumab,⁹ allogenic human chondrocytes transduced with retroviral vector encoding TGF-β1,²⁰ and Curcuma longa.³⁷ One study reported a significant group difference in synovitis,³³ but no difference in pain in the treatment group (methylprednisolone followed by exercise) compared with the control group. These results are summarized in Table 8.

Table 8

Treatments Associated With Statistically Significant Improvements in Pain and Synovitis on MRI Scan^a

Significant Improvement in Pain Only	Significant Improvement in Synovitis Only	Significant Improvement in Pain and Synovitis on MRI Scan
• Lutikizumab (anti-interleukin-1α/β dual variable domain immunoglobulin) (Fleischmann et al⁹) • Allogenic human chondrocytes transduced with retroviral vector encoding TGF-β1 (Lee et al²⁰) • Curcuma longa extract (Wang et al³⁷)	Methylprednisolone and exercise (Riis et al³³)	• Vitamin D (VIDEO Study [Wang et al³⁵ and Jin et al¹⁸])^b • Exercise therapy (Bandak et al¹)^c

None of the studies included in this review reported that treatments significantly improved only synovitis on MRI scan. CDE-MRI, dynamic contrast-enhanced; MRI, magnetic resonance imaging; TGF, transforming growth factor; VAS, visual analog scale; VIDEO, Vitamin D Effect on Osteoarthritis.

Changes in VAS pain scores were only significant in the treatment group compared with the control group on post hoc analyses (P = .048).¹⁸

Bandak et al¹ reported a significant difference between groups only when synovitis was measured on DCE-MRI in the anterior synovium.

The literature on the association between vitamin D and improvement in knee pain is conflicting. Vitamin D supplementation was shown to improve WOMAC-Pain scores in a meta-analysis that reviewed 6 studies in patients with knee OA.³⁹ However, a subgroup of patients with chronic knee pain were followed annually in the VITAL study, and WOMAC-Pain scores did not differ between the experimental (vitamin D) and control groups (omega-3 fatty acids) at any point during follow-up.²⁵ We also recognize that although vitamin D was associated with a significantly greater improvement in pain and MRI markers of synovitis compared with control (P = .02), the trial was large, and the effect size was quite small (SRM = 0.17).³⁵ Further, previous literature reviews summarize the relationship between multiple markers of inflammation and pain in patients with knee OA.^6,27 In addition to measuring synovitis on MRI scan, these markers also include Baker’s cysts viewed on MRI scan, cytokines, and C-reactive protein. Kandemirli et al¹⁹ reported that Baker’s cysts found on ultrasound were found more frequently in symptomatic knees and correlated with higher pain ratings on the VAS scale compared with asymptomatic knees and lower pain ratings. In our systematic review, we reported a similar finding from Riis et al,³³ who noted that changes in synovitis measured on contrast-enhanced MRI scans correlated with an improvement in KOOS-Pain scores.³³ Examining the association between pain and synovitis is difficult because there are multiple ways to measure synovitis and inflammation. For example, when synovial fluid white blood cell count and ES measured on MRI scans using the MOAKS score were compared, the sensitivity and specificity of the synovial fluid white blood cell count in identifying ES on MRI scans were limited.²⁴

One curious finding was elucidated in our systematic review. Bandak et al¹ reported that participants in the exercise group saw a significant improvement in synovitis variables measured on DCE-MRI in the anterior synovium in comparison with the control group. However, DCE-MRI variables for synovitis were not different between groups on the total synovium. It is unclear why anterior knee inflammation decreased after treatment more than other compartments of the knee. Further, no significant differences were seen on MOAKS or CE-MRI or in systemic proinflammatory cytokines from plasma. However, the finding by Bandak et al¹ that exercise did not affect local or systemic inflammation is supported in a trial by Messier et al²⁹ of obese participants with radiographic knee OA randomized to exercise, diet, and a combined regimen of exercise and diet. Plasma interleukin (IL)-6 levels (an inflammatory cytokine) did not decrease in the exercise group.²⁹ In a similar trial of exercise in participants with knee OA, Helmark et al¹¹ found that IL-10 (an anti-inflammatory cytokine) levels from synovial fluid increased significantly in the exercise group compared with the control group immediately after exercise. However, this difference in results between Helmark et al¹¹ and Bandak et al¹ could be due to a difference in exercise regimens. This difference could also be accounted for by the difference in measuring local inflammatory changes. Bandak et al¹ measured local inflammatory changes on MRI scans, while Helmark et al¹¹ measured IL-10 in the synovial fluid directly after exercise.

Limitations

We acknowledge that our systematic review has limitations. The pain scales and measurements of synovitis and inflammation on MRI scans in each study differed, rendering it difficult to compare findings across each trial. Each trial included also had a different follow-up period, which presented a challenge for comparing changes in pain and synovitis on MRI scan after treatment. In addition, each trial had a different sample size, and some of the trials had small sample sizes.^26,37 This presented a challenge in comparing findings across studies. To make meaningful comparisons across studies, we calculated SRMs using mean changes and standard deviations of the mean changes, when presented, or by deriving the standard deviation from 95% CIs. This identical method was used when LSMs and LSMDs were presented. By calculating a comparable effect size across all studies, we were able to make meaningful comparisons. The studies included in this review were heterogeneous with respect to a number of design features including the outcome measure used, the intervention agent, and type of control, among others. We addressed the heterogeneity in outcome metrics by calculating standardized, unitless response means. To do this, we used mean changes and standard deviations of the mean changes when presented or by deriving the standard deviation from 95% CIs. To our knowledge, this is the first review to investigate the effects of treatments on pain and synovitis measured on MRI scans in persons with knee OA.

Conclusion

The findings in this review suggest the need for future research on the effect of knee OA treatments on pain and synovitis. Only vitamin D and exercise therapy were found to improve pain and synovitis after treatment.^1,18,35 More trials are needed to validate whether vitamin D and exercise can improve both pain and synovitis in persons with knee OA. Such findings would allow researchers to identify whether synovitis is a mediator of pain in this patient population. More studies are needed to compare how different measurements of synovitis change after treatment. This summary review included only studies that measured synovitis on MRI, but studies that measured inflammation on ultrasound or with different biomarkers in blood, urine, or synovial fluid could provide rich datasets for understanding how different inflammatory phenotypes in persons with knee OA effect their pain ratings. Further, patients who experience sport- or exercise-related joint injury are at a higher risk for developing posttraumatic OA, which is associated with more rapid progression.³⁸ Studies that investigate treatments that slow or prevent the mechanical components of disease progression, such as joint space narrowing, in addition to pain and inflammation, would be useful for this population and have the potential to improve how these patients prevent the progression of OA.

Footnotes

Final revision submitted January 21, 2024; accepted February 23, 2024.

One or more of the authors has declared the following potential conflict of interest or source of funding: M.H.J. has received consulting fees from Regeneron Pharmaceuticals and is on the advisory board for Regeneron and Biosplice. AOSSM checks author disclosures against the Open Payments Database (OPD). AOSSM has not conducted an independent investigation on the OPD and disclaims any liability or responsibility relating thereto.

ORCID iD

Morgan H. Jones

Notes

References

Bandak

Boesen

Bliddal

, et al. The effect of exercise therapy on inflammatory activity assessed by MRI in knee osteoarthritis: secondary outcomes from a randomized controlled trial. Knee. 2021;28:256-265.

Barbour

Helmick

Boring

Brady

TJ.

Vital signs: prevalence of doctor-diagnosed arthritis and arthritis-attributable activity limitation - United States, 2013-2015. MMWR Morb Mortal Wkly Rep. 2017;66(9):246-253.

Chang

Hayes

Dunlop

, et al. Thrust during ambulation and the progression of knee osteoarthritis. Arthritis Rheum. 2004;50(12):3897-3903.

Cornett

KMD

Menezes

Bray

, et al. Refining clinical trial inclusion criteria to optimize the standardized response mean of the CMTPedS. Ann Clin Transl Neurol. 2020;7(9):1713-1715.

Cui

Wang

, et al. Global, regional prevalence, incidence and risk factors of knee osteoarthritis in population-based studies. EClinicalMedicine. 2020;29-30:100587.

Dainese

Wyngaert

De Mits

, et al. Association between knee inflammation and knee pain in patients with knee osteoarthritis: a systematic review. Osteoarthritis Cartilage. 2022;30(4):516-534.

de Lange-Brokaar

Ioan-Facsinay

Yusuf

, et al. Association of pain in knee osteoarthritis with distinct patterns of synovitis. Arthritis Rheumatol. 2015;67(3):733-740.

Deshpande

Katz

Solomon

, et al. Number of persons with symptomatic knee osteoarthritis in the US: impact of race and ethnicity, age, sex, and obesity. Arthritis Care Res (Hoboken). 2016;68(12):1743-1750.

Fleischmann

Bliddal

Blanco

, et al. A phase II trial of lutikizumab, an anti-interleukin-1α/β dual variable domain immunoglobulin, in knee osteoarthritis patients with synovitis. Arthritis Rheumatol. 2019;71(7):1056-1069.

10.

Guermazi

Roemer

Hayashi

, et al. Assessment of synovitis with contrast-enhanced MRI using a whole-joint semiquantitative scoring system in people with, or at high risk of, knee osteoarthritis: the MOST study. Ann Rheum Dis. 2011;70(5):805-811.

11.

Helmark

Mikkelsen

Borglum

, et al. Exercise increases interleukin-10 levels both intraarticularly and peri-synovially in patients with knee osteoarthritis: a randomized controlled trial. Arthritis Res Ther. 2010;12(4):R126.

12.

Henriksen

Christensen

Klokker

, et al. Evaluation of the benefit of corticosteroid injection before exercise therapy in patients with osteoarthritis of the knee: a randomized clinical trial. JAMA Intern Med. 2015;175(6):923-930.

13.

Higgins

Green

, eds. Cochrane Handbook for Systematic Reviews of Interventions. London: The Cochrane Collaboration; 2011: Version 5.1.0.

14.

Hill

Gale

Chaisson

, et al. Knee effusions, popliteal cysts, and synovial thickening: association with knee pain in osteoarthritis. J Rheumatol. 2001;28(6):1330-1337.

15.

Hill

Hunter

Niu

, et al. Synovitis detected on magnetic resonance imaging and its relation to pain and cartilage loss in knee osteoarthritis. Ann Rheum Dis. 2007;66(12):1599-1603.

16.

Hunter

Guermazi

, et al. Evolution of semi-quantitative whole joint assessment of knee OA: MOAKS (MRI Osteoarthritis Knee Score). Osteoarthritis Cartilage. 2011;19(8):990-1002.

17.

Hunter

Zhang

Conaghan

, et al. Responsiveness and reliability of MRI in knee osteoarthritis: a meta-analysis of published evidence. Osteoarthritis Cartilage. 2011;19(5):589-605.

18.

Jin

Jones

Cicuttini

, et al. Effect of vitamin D supplementation on tibial cartilage volume and knee pain among patients with symptomatic knee osteoarthritis: a randomized clinical trial. JAMA. 2016;315(10):1005-1013.

19.

Kandemirli

Basaran

Kandemirli

Inceoglu

LA.

Assessment of knee osteoarthritis by ultrasonography and its association with knee pain. J Back Musculoskelet Rehabil. 2020;33(4):711-717.

20.

Lee

Parvizi

Bramlet

, et al. Results of a phase II study to determine the efficacy and safety of genetically engineered allogeneic human chondrocytes expressing TGF-beta1. J Knee Surg. 2020;33(2):167-172.

21.

Liang

Fossel

Larson

MG.

Comparisons of five health status instruments for orthopedic evaluation. Med Care. 1990;28(7):632-642.

22.

Loeser

Goldring

Scanzello

Goldring

MB.

Osteoarthritis: a disease of the joint as an organ. Arthritis Rheum. 2012;64(6):1697-1707.

23.

Loeuille

Rat

Goebel

, et al. Magnetic resonance imaging in osteoarthritis: which method best reflects synovial membrane inflammation? Correlations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage. 2009;17(9):1186-1192.

24.

MacFarlane

Arant

Kostic

, et al. Identifying inflammation in knee osteoarthritis: relationship of synovial fluid white blood cell count to effusion-synovitis on magnetic resonance imaging. Arthritis Care Res (Hoboken). 2023;75(8):1783-1787.

25.

MacFarlane

Cook

Kim

, et al. The effects of vitamin D and marine omega-3 fatty acid supplementation on chronic knee pain in older US adults: results from a randomized trial. Arthritis Rheumatol. 2020;72(11):1836-1844.

26.

Mahler

EAM

Minten

Leseman-Hoogenboom

, et al. Effectiveness of low-dose radiation therapy on symptoms in patients with knee osteoarthritis: a randomised, double-blinded, sham-controlled trial. Ann Rheum Dis. 2019;78(1):83-90.

27.

Mathiessen

Conaghan

PG.

Synovitis in osteoarthritis: current understanding with therapeutic implications. Arthritis Res Ther. 2017;19(1):18.

28.

McAlindon

LaValley

Harvey

, et al. Effect of intra-articular triamcinolone vs saline on knee cartilage volume and pain in patients with knee osteoarthritis: a randomized clinical trial. JAMA. 2017;317(19):1967-1975.

29.

Messier

Mihalko

Legault

, et al. Effects of intensive diet and exercise on knee joint loads, inflammation, and clinical outcomes among overweight and obese adults with knee osteoarthritis: the IDEA randomized clinical trial. JAMA. 2013;310(12):1263-1273.

30.

Neogi

Guermazi

Roemer

, et al. Association of joint inflammation with pain sensitization in knee osteoarthritis: The Multicenter Osteoarthritis Study. Arthritis Rheumatol. 2016;68(3):654-661.

31.

Oei

EHG

Hirvasniemi

van Zadelhoff

van der Heijden

. Osteoarthritis year in review 2021: imaging. Osteoarthritis Cartilage. 2022;30(2):226-236.

32.

Rezus

Burlui

Cardoneanu

, et al. From pathogenesis to therapy in knee osteoarthritis: bench-to-bedside. Int J Mol Sci. 2021;22(5):2697.

33.

Riis

RGC

Henriksen

Klokker

, et al. The effects of intra-articular glucocorticoids and exercise on pain and synovitis assessed on static and dynamic magnetic resonance imaging in knee osteoarthritis: exploratory outcomes from a randomized controlled trial. Osteoarthritis Cartilage. 2017;25(4):481-491.

34.

Tanamas

Hanna

Cicuttini

Wluka

Berry

Urquhart

DM.

Does knee malalignment increase the risk of development and progression of knee osteoarthritis? A systematic review. Arthritis Rheum. 2009;61(4):459-467.

35.

Wang

Cicuttini

Jin

, et al. Knee effusion-synovitis volume measurement and effects of vitamin D supplementation in patients with knee osteoarthritis. Osteoarthritis Cartilage. 2017;25(8):1304-1312.

36.

Wang

Jin

Han

, et al. Cross-sectional and longitudinal associations between knee joint effusion synovitis and knee pain in older adults. J Rheumatol. 2016;43(1):121-130.

37.

Wang

Jones

Winzenberg

, et al. Effectiveness of Curcuma longa extract for the treatment of symptoms and effusion-synovitis of knee osteoarthritis: a randomized trial. Ann Intern Med. 2020;173(11):861-869.

38.

Whittaker

Roos

EM.

A pragmatic approach to prevent post-traumatic osteoarthritis after sport or exercise-related joint injury. Best Pract Res Clin Rheumatol. 2019;33(1):158-171.

39.

Zhao

Peng

, et al. Does vitamin D improve symptomatic and structural outcomes in knee osteoarthritis? A systematic review and meta-analysis. Aging Clin Exp Res. 2021;33(9):2393-2403.