Comparison of the effectiveness of intra-infrapatellar fat pad and intra-articular glucocorticoid injection in knee osteoarthritis patients with Hoffa’s synovitis: protocol for a multicentre randomised controlled trial

Abstract

Introduction

The infrapatellar fat pad and synovium are the sites of immune cell infiltration and the origin of proinflammation. Studies have shown that Hoffa’s synovitis may be a sign of early-stage osteoarthritis (OA). However, there have been no effective interventions specifically for Hoffa’s synovitis.

Methods and analysis

We will conduct a multicentre, multi-blind (participant, physician, outcome assessor and data analyst blinded) randomised controlled trial to compare the effectiveness of an intra-infrapatellar fat glucocorticoid versus an intra-articular injection for Hoffa’s synovitis in patients with knee OA. We will recruit 236 knee OA patients with Hoffa’s synovitis at outpatient clinics in three centres. We will randomly allocate them to two groups in a 1:1 ratio. One group will receive ultrasound-guided injection of 40 mg (1 mL) triamcinolone acetonide into the infrapatellar fat pad; the other group will receive ultrasound-guided injection of 40 mg (1 mL) triamcinolone acetonide into the knee joint cavity. All patients will be followed up at 2, 4, 8, 12 and 24 weeks after the injection. Primary outcomes are (1) Hoffa’s synovitis improvement rate, measured with the MRI Osteoarthritis Knee Score system (superiority outcome) at 24 weeks and (2) pain intensity, measured with the Western Ontario and McMasters University Osteoarthritis Index (WOMAC) at 2 weeks post-injection. Secondary outcomes include Hoffa’s synovitis score at 2 weeks post-injection, pain intensity with the numerical rating scale, WOMAC questionnaire score improvements (function, joint stiffness and total score), improvement rates in effusion synovitis at 2 and 24 weeks, articular cartilage thickness changes at 2 and 24 weeks, Intermittent and Constant Osteoarthritis Pain score, quality of life measured with the EuroQol-5D, OARSI-OMERACT response indicators, co-interventions and side effects at 2, 4, 8, 12 and 24 weeks.

Ethics and dissemination

Ethical approval has been granted by the Medical Ethics Committee of the Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (2023–178). Written informed consent will be obtained from all patients prior to data collection. The findings of this research will be shared through presentations at academic conferences and publications in peer-reviewed journals.

Trial registration number

ChiCTR2400080474.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• The multicentre randomised controlled trial design enhances generalisability and external validity.

• The multi-blind design reduces the risk of subjective influences on outcome assessments, minimises bias and ensures the reliability of results.

• Combining patient-reported subjective symptom improvement with objective MRI structural changes as outcome measures provides a comprehensive evaluation of treatment effectiveness.

• One limitation of this study is that the patients may be able to identify the injection locations, resulting in the failure of blinding.

Introduction

Osteoarthritis (OA) stands as a prominent cause of disability globally, with its prevalence expected to surge in the forthcoming years.¹,³ The knee joint is the most common site affected by OA, leading to pain and physical dysfunction, significantly reducing the quality of life for patients.⁴ Presently, no medications exist capable of reversing or halting the progression of knee OA.^{5 6} Consequently, research endeavours have primarily aimed at pain management and mitigating functional decline.^{7 8} Given the chronic nature of knee OA, exacerbating symptoms and structural degradation often necessitate joint replacement surgery,^{9 10} underscoring the urgency for a treatment strategy that addresses both symptomatic relief and structural preservation.

Intra-articular (IA) glucocorticoid injections in the knee joint are commonly used in clinical practice to ameliorate knee pain.^{11 12} These injections primarily suppress local inflammation, consequently ameliorating pain and enhancing joint function, thus augmenting patients’ quality of life. Prior investigations have indicated that such injections significantly attenuate effusion synovitis and reduce synovial swelling volume, with structural improvements correlating with symptomatic amelioration.^{13 14} However, in knee OA, synovial inflammation encompasses not only effusion synovitis but also Hoffa’s synovitis—inflammation within the infrapatellar fat pad (IPFP).¹⁵,¹⁸ Although the exact function of the IPFP remains elusive, emerging evidence underscores its pivotal biomechanical role within the knee joint.^{19 20} Moreover, recent studies suggest its involvement in knee OA pathologies, serving as a site for immune cell infiltration and inflammation.²¹ Previous clinical investigations propose Hoffa’s synovitis as an early OA marker, correlating with subsequent clinical OA development.²² Therefore, we consider whether injecting glucocorticoid into the IPFP could directly suppress Hoffa’s synovitis and simultaneously improve patients’ symptoms. Furthermore, IA glucocorticoid injections may carry the risk of joint cartilage damage and accelerate OA progression,^{23 24} whereas injecting glucocorticoid into the IPFP may potentially mitigate this risk. A recent trial in progress compares IPFP glucocorticoid and saline injections for knee OA.²⁵ While they recruit knee OA patients with both effusion and Hoffa’s synovitis, all the patients receive a background hyaluronic acid injection. This design does not allow for a thorough evaluation of the effectiveness of IPFP glucocorticoid injection in improving Hoffa’s synovitis. Besides, there is no trial comparing the effectiveness of IPFP glucocorticoid injection to the commonly used IA glucocorticoid injection.

Here, we present the protocol for a multicentre randomised controlled trial to evaluate the effectiveness and safety of IPFP glucocorticoid injections in knee OA patients with Hoffa’s synovitis, compared with IA injections. We hypothesise that IPFP glucocorticoid injection reduces Hoffa’s synovitis more effectively than IA glucocorticoid injection and shows the non-inferior pain amelioration effect. The results shall help us better understand the value of IPFP glucocorticoid injection and provide the evidence for applying IPFP glucocorticoid injections for treating Hoffa’s synovitis.

Methods and analysis

Study design

This investigation will be conducted as a multicentre, randomised controlled trial using a multi-blind design. From March 2024 to October 2028, we will enrol individuals who have been specifically chosen for their knee OA with Hoffa’s synovitis at outpatient clinics located in Shanghai Sixth People’s Hospital, Jinjiang Municipal Hospital, Shanghai Ruijin Hospital, Yancheng Third People’s Hospital and Sheyang County People’s Hospital. All trial procedures and timelines are summarised in table 1. The follow-up period will be 24 weeks. Any modifications to the protocol that could influence the study’s execution, potential patient benefits or safety considerations, including adjustments to study objectives, design, patient demographics, sample sizes, procedures or significant administrative elements, necessitate a formal protocol amendment. Ethical approval has been granted by the Medical Ethics Committee of the Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (2023–178). Written informed consent will be obtained from all patients prior to data collection (online supplemental material). This trial has been registered at the Chinese Clinical Trial Registry (ChiCTR2400080474).

Table 1. Schedule of enrolment, intervention and assessments.

	Baseline	2 weeks after injection	4 weeks after injection	8 weeks after injection	12 weeks after injection	24 weeks after injection
Demographics	x
Physical examination	x
WOMAC questionnaire (including pain, function and joint stiffness)	x	x	x	x	x	x
NRS pain questionnaire (assessing average pain levels over the past week)	x	x	x	x	x	x
ICOAP questionnaire	x	x	x	x	x	x
EQ-5D Questionnaire	x	x	x	x	x	x
OARSI-OMERACT response feedback		x	x	x	x	x
K&L grade	x					x
MOAKS Hoffa’s synovitis score	x	x				x
MOAKS effusive synovitis score	x	x				x
MRI measurement of articular cartilage thickness (mm)	x	x				x
Co-therapy	x	x	x	x	x	x
Side effects		x	x	x	x	x

EQ-5DEuroQol 5DICOAPIntermittent and Constant Osteoarthritis PainK&LKellgren and LawrenceMOAKSMRI Osteoarthritis Knee ScoreNRSNumerical Rating ScaleWOMACWestern Ontario and McMasters University Osteoarthritis Index

Inclusion criteria

The inclusion criteria are participants aged 40 years or older; patients with unilateral knee joint pain; a Western Ontario and McMasters University Osteoarthritis Index (WOMAC) pain score of 8–18 (on a scale of 0–20) for the affected knee in the past week²⁶; the knee meets the American College of Rheumatology clinical classification criteria for knee OA; Kellgren and Lawrence (K&L) grade≥2 on the knee²⁷; MRI examination of the knee revealed concomitant Hoffa’s synovitis: using the semiquantitative MRI Osteoarthritis Knee Score (MOAKS) system^{28 29} and Hoffa’s synovitis scored 1 to 3 based on reading MRI films (3.0 T, resolution 1.5, slice thickness 2 mm); there is an indication for a glucocorticoid injection for the patients assessed by the treating physician. MOAKS is a semiquantitative scoring system used to assess the severity of knee osteoarthritis through MRI, focusing on features such as cartilage damage, bone marrow lesions, synovitis and joint effusion.^{29 30} The system evaluates synovitis and effusion synovitis in the knee, including inflammation of Hoffa’s fat pad and fluid accumulation in the joint. Hoffa’s fat pad synovitis is graded on a 0–3 scale: Grade 0 indicates a normal appearance with no signal changes, Grade 1 represents mild synovitis with subtle hyperintensity or thickening, Grade 2 reflects moderate synovitis with more pronounced hyperintensity and thickening and Grade 3 denotes severe synovitis with marked hyperintensity and significant thickening. Effusion synovitis, which measures fluid and associated synovial inflammation, is also graded on a 0–3 scale: Grade 0 signifies no effusion or synovitis, Grade 1 indicates mild fluid accumulation, Grade 2 represents a moderate amount of fluid and Grade 3 reflects severe effusion with extensive fluid buildup. These scores provide a standardised way to assess inflammation and fluid changes in osteoarthritis for monitoring disease progression and treatment response. The national guideline recommends glucocorticoid injection for patients with knee OA who have a flare of knee pain and/or do not respond to other pain medications.

Exclusion criteria

Exclusion criteria were as follows: patients with rheumatoid arthritis, gout or other diseases affecting lower limb symptom scores; patients planning to undergo joint surgery in the near future; patients with a history of surgery within the affected knee joint; use of oral glucocorticoid; patients who have received IA glucocorticoid injections within the past 6 months; allergy to glucocorticoids; contraindications to MRI examination; poorly controlled type I or type II diabetes mellitus; patients with a history of peptic ulcer disease; unable to independently complete follow-up visits and questionnaire surveys; and patients unwilling to participate in the trial.

Outcome measures

The primary outcomes of this study are as follows:

1. Hoffa’s synovitis improvement rate at 24 weeks follow-up, measured with the MOAKS score system (superiority outcome). MRI films will be read by two independent radiologists for assigning MOAKs scores with blinded for patient information. Disagreements will be solved by consensus;

2. Pain intensity at 2 weeks post-injection: measured with the WOMAC pain score (non-inferiority outcome).

The secondary outcome measures are (1) Hoffa’s synovitis score at 2 weeks post-injection: to assess early changes, measured with the MOAKS score system; (2) pain intensity at multiple time points: to capture the progression and management of pain over time, measured with the numerical rating scales (NRS) at 2, 4, 8, 12 and 24 weeks; (3) WOMAC questionnaire score improvements: beyond pain, assessing function, joint stiffness and total score at 2, 4, 8, 12 and 24 weeks to evaluate the broader impact on quality of life and physical function; (4) effusion synovitis improvement rates at 2 and 24 weeks: measured with the MOAKS score system to track inflammation associated with knee OA; (5) articular cartilage thickness changes over 2 weeks and 24 weeks (MRI quantification): two radiologists will measure medial femorotibial cartilage plates at the weight-bearing region independently, and then mean cartilage thickness (mm) for the two will be calculated as the final results; (6) intermittent and Constant Osteoarthritis Pain score (ICOAP) scores at 2, 4, 8, 12 and 24 week: to differentiate between types of pain experienced by patients, providing a more nuanced understanding of pain management needs; (7) quality of life measured with the EuroQol 5D (EQ-5D) scores : to assess the overall impact of the treatment on patients’ well-being; (8) OARSI-OMERACT response indicators at 2, 4, 8, 12 and 24 weeks: to evaluate treatment response based on a standardised set of criteria; (9) co-interventions at 2, 4, 8, 12 and 24 weeks: to monitor and account for any additional treatments that may influence outcomes and (10) side effects at 2, 4, 8, 12, and 24 weeks: to ensure patient safety and to identify any adverse effects of the treatment.

Sample size calculation

According to previous literature, IA glucocorticoid injection is estimated to improve approximately 15% of Hoffa’s synovitis.³¹ There is currently no report on the improvement rate of IPFP glucocorticoid injection. We believe a 20% improvement can be considered clinically meaningful. Setting power at 90% and α=0.05, with an anticipated loss to follow-up of around 20%, each group requires approximately 118 participants, totalling 236 individuals. Based on the aforementioned sample size, the power to detect non-inferiority in WOMAC pain scores between the two groups is 86% (with a non-inferiority margin of 1.6 points, a clinically meaningful difference of 2 points, a SD of 4, assuming a true difference of 0; α=0.025).

Randomisation and blinding

An independent researcher assistant, who is not involved in the medical procedures, will prepare a computer-generated randomisation group index using 1:1 allocation with random sizes of block on an encrypted website. Nobody except this researcher has access to the randomisation list. After a patient gives written consent and finishes baseline assessments, a unique trial number will be issued for the patient. The researcher who helps patients complete baseline assessments will inform the researcher assistant of inclusion. Then, the researcher assistant will inform the ultrasound specialist and the medical professionals (who will perform the injection) of the randomisation results. The allocation outcomes will remain hidden until the data analysis has been finished. The trial has a multi-blind design, involving patients, researchers, outcome assessors and data analysts. Ultrasound specialists and injection operators will know the allocation but will be blinded for patient baseline and follow-up information, and they are instructed not to inform the patients. Patients will be asked for which injection they think they have received when they walk out of the injection room.

Interventions

Participants who have signed the informed consent will be randomly allocated to two groups by the computer randomisation software in a 1:1 ratio. One group will receive an injection of 40 mg (1 mL) of triamcinolone acetonide under ultrasound guidance into the IPFP, while the other group will receive the same dose of triamcinolone acetonide under ultrasound guidance into the knee joint cavity.

The glucocorticoid injection will be conducted in the IPFP or knee joint cavity using ultrasound as a guide (figure 1). The steps are as follows: (1) the participant assumes a seated position to flex the knee joint by 90°. (2) The ultrasound probe is placed inferiorly and laterally to the patella to visualise the infrapatellar fat pad. (3) The needle is inserted beneath the ultrasound probe. (4) In the IPFP group, triamcinolone acetonide is administered by injecting into IPFP, near the synovium, where there is clear synovial hyperplasia; in the IA group, the needle will go through the Hoffa fat pad into the joint cavity, and then triamcinolone acetonide is injected. The two groups have identical needle skin insertion points. We offer injection training courses for the injection operators and ultrasound specialists before launching the trial.

Figure 1. (A) The participant assumes a seated position to flex the knee joint by 90°. The ultrasound probe is placed inferiorly and laterally to the patella to visualise the infrapatellar fat pad. The needle is inserted beneath the ultrasound probe. (B) Ultrasound image for the localisation of the infrapatellar fat pad. Infrapatellar fat pad injection, the needle will be placed in the Hoffa fat pad; intra-articular injection, the needle will go through the Hoffa fat pad into the joint cavity.

Clinical examination procedures

Recruitment of participants and collection of baseline information

Patients with knee joint pain attending orthopaedic outpatient clinics at three centres will be selected. Patients will be screened according to inclusion and exclusion criteria and enrolled in the clinical trial after obtaining informed consent. Baseline information of enrolled patients will be recorded on enrolment, including demographics (age, gender, height, weight, smoking status, history of other diseases, duration of OA), physical examinations (including joint tenderness, patellofemoral crepitus, presence of joint redness, joint swelling), questionnaire surveys related to patient symptoms (WOMAC, NRS, ICOAP, EQ-5D questionnaires and co-therapies) and baseline knee joint X-ray and MRI examinations. Figure 2 shows a flow chart of patient recruitment.

Figure 2. Flowchart of patient recruitment.

Follow-up visits

Follow-up visits will be scheduled at the outpatient clinic at 2, 4,8,12 and 24 weeks post-injection. Follow-up examinations will include questionnaire surveys related to patient symptoms (WOMAC, NRS, ICOAP, EQ-5D questionnaires and co-therapies), assessment of side effects and knee joint X-ray and MRI examinations. Specific examination indicators required at each time point are outlined in table 1.

Statistical analysis plan

Analysis principle

The primary analysis will adhere to an intention-to-treat (as randomised) principle. All patients randomised will be included in the analysis and grouped according to randomisation results. The significance level is set at 0.05. Data analysts will be blinded to treatment allocations and strictly adhere to this analysis plan.

Baseline demographics

Baseline demographics of patients in both groups will be presented as mean (SD) or number (percentage), as appropriate. No statistical analysis will be conducted to compare these characteristics. An imbalance in baseline demographics between the two treatment arms may occur post-randomisation. Analyses will be adjusted for variables that impede baseline comparability of groups when clinically relevant and exhibiting a between-group difference exceeding 10%.

Primary outcome measures

For the MOAKS Hoffa’s synovitis score, linear mixed models with repeated measures will determine group improvement rate differences over time. The model structure with the lowest Akaike’s information criteria will be selected. Fixed effects will include time and time by group. Mean between-group rate differences and their 95% CIs will be calculated at 2 and 24 weeks, with the 24-week difference deemed primary.

Similarly, linear mixed models with repeated measures will assess group differences over time for WOMAC pain score improvements. The structure with the lowest Akaike’s information criteria will be chosen. Fixed effects will include time and time by group. Mean between-group differences and their 95% CIs will be calculated at 2, 4, 8, 12 and 24 weeks, with the 2-week difference considered primary. Non-inferiority testing will compare against the margin (1.6 points).

Secondary outcome measures

Linear mixed models with repeated measures will also analyse continuous secondary outcomes, including changes in NRS, ICOAP, EQ-5D, MOAKS effusive synovitis score and articular cartilage thickness over time. Generalised estimating equations analyses with repeated measures will be conducted for dichotomous outcomes, that is, OMERACT-OARSI responder criteria. Side effects and co-therapies will be described without statistical testing.

Missing values

Missing values can be caused by incompletely filled questionnaires or loss of follow-up. Although linear mixed models will take missing values into account, the results can be invalid in the dataset with a considerable number of missing values (>5%, rule of thumb).³² Additionally, the assumption of missing at random can be violated.³³ Therefore, to test the robustness of our results, a sensitivity analysis combined with multiple imputations will be performed.

Subgroup analysis

An explorative, predefined, subgroup analysis will be performed to assess the interaction effects between two groups regarding the baseline effusion synovitis score (stratification factor; MOAKS effusion synovitis score≥ 1 vs =0) on the primary outcome measures.

Data monitoring and auditing

The supervising inspector verified that all study participants signed informed consent prior to their involvement and will ensure adherence to the study protocol, compliance with relevant principles and the reliability and completeness of research observations and data. The data monitoring committee, consisting of four specialists in orthopaedics, anaesthetics, pharmacy and statistics, will conduct audits via frequent interviews or telephone conversations and retain the authority to examine patients at any moment. The audit process will be conducted independently of the investigators.

Strategies to improve adherence to the protocol

All investigative personnel will get comprehensive professional training to implement the recruitment process, secure informed consent, conduct follow-ups and standardise the application of evaluation scales in result assessment. The nurse and anaesthetist responsible for the preparation and administration of the ‘research agent’ will get training to adhere to the defined protocol.

Patient and public involvement

During our pilot work, two patients were asked about their experience of steroid injections into the infra-patellar fat pad, and their feedback reinforced the validity of our blinded design. The results of this trial will be shared with all participants on its completion.

Ethics and dissemination

Ethical approval has been granted by the Medical Ethics Committee of the Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine (approval no. 22 023–178). Written informed consent will be obtained from all patients who meet the inclusion criteria prior to data collection. This trial has been registered at the Chinese Clinical Trial Registry (registration number: ChiCTR2400080474). The findings of this research will be shared through presentations at academic conferences and publications in peer-reviewed journals.

Any changes to the protocol, such as adjustments to the population, sample size, procedures, inclusion criteria, interventions, outcomes or analyses, will be reported to relevant parties, including sponsors, trial registries and ethics committees. Approval from the China Ethics Committee of Registering Clinical Trials will be secured before implementation.

Discussion

The objective of this trial is to investigate the effectiveness and safety of glucocorticoid injections into the IPFP in knee OA patients with Hoffa’s synovitis. It is noteworthy that synovial inflammation is a common occurrence in knee OA progression, with Hoffa’s fat pad playing a significant role in joint inflammation.^{34 35} The IPFP, rich in nerve supply and closely associated with the highly innervated synovium, contributes to discomfort.³⁶,³⁸ When the IPFP is probed arthroscopically without IA anaesthesia, Dye et al reported that the sensation is ‘exquisitely sensitive’ and causes intense localised discomfort.³⁶ Substance P in nociceptive nerve fibres within the fat pad and surrounding synovial tissue likely contributes to this sensitivity.³⁷,³⁹ The presence of free nerve endings within the synovial lining of the fat pad confirms its sensorial capability and suggests a role in anterior knee discomfort.⁴⁰

Knee OA patients commonly undergo IA medication injections. However, studies have suggested that frequent glucocorticoid injections into the joint cavity may contribute to cartilage deterioration.^{41 42} By contrast, injecting glucocorticoids into the IPFP, situated near the synovial membrane, could potentially augment their efficacy while mitigating the risk of cartilage degradation. Moreover, reducing Hoffa’s synovitis may yield beneficial effects, supported by substantial evidence indicating it potentially facilitates OA progression.^{43 44} The present study may provide evidence for applying IPFP glucocorticoid injection to relieve Hoffa’s synovitis. Concerns about injecting glucocorticoids into adipose tissue have been raised, yet the issue remains controversial, as they have been suggested to have both adipogenic and lipolytic effects inside adipose tissue.^{45 46} In the current trial, we only inject 1 mL of triamcinolone acetonide and in our pilot attempts, no patients reported specific symptoms under the patella, such as swelling or oedema. Hence, we believe the method should be safe.

The main advantage of this study is its meticulous design, which includes specifically selecting knee OA patients with Hoffa’s synovitis, as well as multicentre and multiple-blinded designs. One limitation of this study is that the patients may feel the injection locations by themselves, causing the failure of blindness. In our pilot attempts, six patients received the injections (three for each injection) and found no significant subjective perception difference between the two injections. Furthermore, we will ask for patients’ conjectures regarding the injection location subsequent to the administration of injections. Data will be used to test the reliability of the blindness. Another limitation of the study is the lack of a sham injection group, which makes it impossible to assess the influence of the placebo effect on the outcomes observed in both injection groups. Consequently, the treatment effects observed in both groups likely include an undetermined contribution from the placebo effect.

The first patient was enrolled on 29 May 2024; enrolment is anticipated to complete by October 2028. To address slow enrolment, two additional research centres—Yancheng Third People’s Hospital and Sheyang County People’s Hospital—were added to the study. The protocol modifications required for including the two centres have been approved by the Ethics Committee of Shanghai Sixth People’s Hospital.

In conclusion, this protocol describes a high-quality multicentre, randomised controlled trial using a multi-blind design for assessing the effectiveness and safety of IPFP glucocorticoid injection. By comparing to IA glucocorticoid injections, this trial will provide evidence of whether IPFP injection is a more effective approach for knee OA patients with Hoffa’s synovitis.