Efficacy of intra-articular injections for the treatment of osteoarthritis: A narrative review

Abstract

Osteoarthritis (OA) is a prevalent degenerative joint disease characterized by progressive cartilage loss, inflammation, and joint dysfunction. With profound effects on joint function and quality of life, OA imposes a substantial socio-economic burden. As of now, OA remains incurable, lacking approved medications, regenerative therapies, or procedures that can halt the progressive destruction of the joint. Intraarticular (IA) injections have emerged as a cornerstone in the management of knee OA, offering localized minimally invasive therapeutic options. Traditional IA therapies, including corticosteroids and hyaluronic acid (HA), primarily aim to reduce pain but lack regenerative capacity. Biologic IA therapies for knee OA including autologous blood-derived products like platelet-rich plasma (PRP), bone marrow aspirate concentrate (BMAC) and mesenchymal stromal cells (MSCs) have become more commonly used. Finally, newer IA therapies such as fibroblast growth factor 18 and gene therapy are being investigated. In this review, we highlight the current evidence around IA injections for the treatment of knee OA.

1. Introduction

Osteoarthritis (OA) is a chronic, degenerative, whole joint disease characterized by degradation of articular cartilage, formation of osteophytes, subchondral bone remodeling, and synovial inflammation [1]. OA is mediated by an imbalance of chondrocyte homeostasis, which leads to progressive degradation of cartilage extracellular matrix involving pro-inflammatory cytokines and matrix metalloproteinases [1]. Worldwide, it stands as the most prevalent joint disorder, affecting an estimated 30.8 million adults in the United States and a staggering 300 million individuals worldwide [2,3]. OA is a primary contributor to disability, inflicting joint pain, loss of function, reduced mobility, and a decline in overall quality of life (QOL) in the elderly [4].

Conservative management of knee OA includes weight loss, bracing, physiotherapy, and various medications (including acetaminophen, non-steroidal anti-inflammatory drugs (NSAIDs), and opioids). In addition, intra-articular (IA) knee injections including corticosteroids, hyaluronic acid (HA), platelet-rich plasma (PRP), bone marrow aspirate concentrate (BMAC), and mesenchymal stromal cells (MSCs), have been reported. When these conservative therapies fail, total knee arthroplasty (TKA) is the only option for end-stage knee OA. TKA is considered the most effective intervention for severe knee OA and is becoming increasingly relied upon to reduce pain, disability, and restore some patients to near normal function [5]. Despite its effectiveness, 11 ​%–19 ​% of primary TKA patients are not satisfied after surgery [6]. As well, TKA is often associated with complications, including infection, bleeding, venous thromboembolism, pneumonia, stiffness, and implant loosening and failure requiring the need for revision surgery [7].

The challenging situation is in the management of patients with mild to moderate OA who are not ready for or an ideal candidate for joint replacement. For these patients, nonsurgical therapies to help manage symptoms and potentially the disease progression are important. IA injections have become a key component in the nonsurgical management of knee OA, offering targeted, localized therapy that minimizes systemic side effects.

The purpose of this narrative review is to summarize the evidence around IA injections used for the management of knee OA. The review will emphasize the latest prospective randomized studies. By highlighting the most recent available evidence, the goal is to help direct clinical use of IA injections for knee OA and possibly identify future areas of direction.

1.1. Corticosteroid and Hyaluronic acid (HA)

Traditional IA therapies for knee OA have included corticosteroids and HA. These injections primarily aim to reduce pain and improve joint function.

The Osteoarthritis Research Society International (OARSI) guidelines have supported IA corticosteroids for the management of knee OA [8]. Corticosteroids provide rapid, short-term relief by reducing inflammation, typically for up to 4–6 weeks after injection [9]. McAlindon et al. investigated the impact of IA triamcinolone compared to saline on knee OA. The study was a double-blind, randomized clinical trial (RCT) with 140 patients with symptomatic knee OA and synovitis identified via imaging. Participants received IA injections of either triamcinolone (40 ​mg) every 12 weeks, or saline placebo every 12 weeks for the total study duration of 2 years. Primary outcomes measures included cartilage volume loss measure by MRI and pain assessed using the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain subscale. Triamcinolone showed greater loss of cartilage volume compared to the saline group (mean cartilage thickness reduction of −0.21 ​mm vs. −0.10 ​mm) and did not provide superior long-term pain relief compared to saline. This study highlights a critical trade-off in the use of IA corticosteroids for knee OA. While triamcinolone did not provide long-term pain relief compared to saline and may have accelerated cartilage loss, it remains a viable option for short-term symptom management. These findings suggest the need for cautious use of corticosteroids and further exploration of alternative treatments that balance pain relief with joint preservation (Table 1).

Table 1.

Corticosteroid clinical studies.

Study	Level of Evidence	Pathology	Number of Patients	Protocol	Follow-up	Main Findings	Conclusions
McAlindon et al. (2017)	Double blind, placebo controlled, RCT	Symptomatic, radiographic knee OA, KL grade 2 or 3, synovitis identified via ultrasound	140	Randomized to receive IA injections of either: Triamcinolone (40 ​mg) every 12 weeks, or saline placebo every 12 weeks	2 years	Triamcinolone showed greater loss of cartilage volume compared to the saline group (mean cartilage thickness reduction of −0.21 ​mm vs. −0.10 ​mm) and did not provide superior long-term pain relief compared to saline	While triamcinolone did not provide long-term pain relief compared to saline and may have accelerated cartilage loss, it remains a viable option for short-term symptom management

The use of IA HA injections, also known as viscosupplementation, for knee OA has been extensively studied. There have been mixed results regarding their efficacy and safety. In a recent systematic review and meta-analysis, Pereira et al. [10] critically evaluated the efficacy and safety of viscosupplementation in treating knee OA. An analysis of 24 large placebo-controlled trials, including 8997 randomized participants, revealed that viscosupplementation resulted in a modest reduction in pain intensity compared to placebo (SMD −0.08, 95 ​% CI −0.15 to −0.02). However, the lower limit of the confidence interval excluded the minimal clinically important difference between groups. This corresponds to a reduction of −2.0 ​mm (95 ​% CI -3.8 to −0.5 ​mm) on a 100 ​mm visual analog scale. Similar results were found for functional outcomes. Conversely, data from 15 large placebo-controlled trials involving 6462 randomized participants showed that viscosupplementation was linked to a significantly increased risk of serious adverse events compared to placebo (relative risk 1.49, 95 ​% CI 1.12 to 1.98), raising concerns about its safety profile. While HA injections have been utilized in the treatment of knee OA, current evidence and clinical guidelines suggest they offer minimal clinical benefit and may pose safety risks, leading to a general consensus against their routine use in managing knee OA (Table 2).

Table 2.

Hyaluronic acid clinical studies.

Study	Level of Evidence	Pathology	Number of Patients	Protocol	Follow-up	Main Findings	Conclusions
Pereira et al. (2022)	Systematic review and meta-analysis of RCTs	Symptomatic, radiographic knee OA	8997 in analysis of pain; 6462 in analysis of adverse events	24 large, placebo controlled trials (8997 randomized participants) included in the main analysis of pain and function; 15 large, placebo controlled trials (6462 randomized participants) in adverse events	Median total trial follow-up was 24 weeks (8.6–27)	Viscosupplementation led to a statistically significant reduction in knee OA pain compared to placebo, however, the effect was small and did not meet the threshold for minimal clinically important difference. Viscosupplementation had a higher risk of serious adverse events compared to those receiving placebo.	Despite a slight statistical improvement in pain, the clinical benefits of viscosupplementation are minimal. With an increased risk of serious adverse events, the findings do not support the widespread use of viscosupplementation for knee OA.

2. Biologic therapies

Recent advances in regenerative medicine have expanded the horizon of IA injections to include PRP, BMAC and MSCs. These biologic therapies hold promise for modifying disease progression by targeting inflammation, promoting cartilage repair, and enhancing joint homeostasis.

3. Platelet-rich plasma (PRP)

Platelet-rich plasma is a blood-derived product with high concentrations of growth factors (GFs) and other bioactive molecules [11]. Growth factors (GFs), such as bone morphogenetic proteins (BMPs)-2 and −7, have been shown to be involved in cartilage homeostasis and repair [11]. Injecting a cocktail of various bioactive GFs is thought to create a favorable environment for potential regeneration of degenerative tissue. While PRP is increasingly utilized for knee OA treatment, the available evidence supporting its clinical benefits is limited [12]. Despite the increased concentrations of growth factors and cytokines that facilitate tissue healing and suppress inflammatory processes, observable improvements in symptoms and structural benefits have not been evident [13].

There are various PRP preparations with differing concentrations of platelets, white blood cells, and growth factors. Some studies have indicated that leukocyte-rich PRP may lead to higher synoviocyte death and increased inflammation [14]. Other research has explored the impact of leukocyte concentration on clinical outcomes, with findings suggesting that leukocyte-poor PRP preparations may be more effective in providing symptomatic relief [15]. The optimal balance of these components has not yet been established and needs more investigation as it is well known that PRP preparations are heterogeneous and lack standardization, as a result, many studies may not be generalizable to other PRP preparations [13].

3.1. Clinical studies on PRP

Several systematic reviews have reported improved pain and functional outcomes associated to PRP in comparison to saline or HA [16,17]. These reviews suggested that the greatest benefits are observed in patients with mild to moderate radiographic OA [16]. Shen and colleagues [17] conducted a meta-analysis that incorporated 14 studies. In their study, the control group encompassed various treatment modalities including HA, saline, placebo, ozone, and corticosteroids compared to PRP. Their findings revealed that PRP injections were more effective in terms of both function and pain levels at the 3, 6, and 12-month follow-up intervals compared with the other injections [17]. A prospective, double-blind, RCT evaluated the efficacy of PRP injections compared to placebo in treating knee OA. The trial involved 78 patients with bilateral knee OA, who were randomly assigned to receive PRP injections in one knee and a saline placebo in the other. 156 knees with bilateral OA were divided randomly into 3 groups: group A (52 knees) received a single injection of PRP, group B (50 knees) received 2 injections of PRP 3 weeks apart, and group C (46 knees) received a single injection of normal saline. White blood cell (WBC)-filtered PRP with a platelet count 3 times that of baseline was administered in all. Clinical outcomes were evaluated using the WOMAC questionnaire, pain by Visual Analog Scale (VAS) and overall satisfaction before treatment and at 6 weeks, 3 months, and 6 months after treatment. Patients reported significant functional improvement and pain reduction in PRP-treated knees compared to placebo at 6 weeks, 3 months, and 6 months post-injection. No serious adverse events were reported [18] (Table 3).

Table 3.

PRP clinical studies.

Study	Level of Evidence	Pathology	Number of Patients	Protocol	Follow-up	Main Findings	Conclusions
Patel et al. (2013)	Prospective, double-blind, RCT, placebo controlled	Symptomatic, radiographic knee OA, Ahlback grade 1 or 2	78 patients (156 knees) with bilateral OA	156 knees with bilateral OA were divided randomly into 3 groups: Group A (52 knees) received a single injection of PRP, group B (50 knees) received 2 injections of PRP 3 weeks apart, and group C (46 knees) received a single injection of normal saline	6 weeks, 3 months, and 6 months	Patients reported significant functional improvement and pain reduction in PRP-treated knees compared to placebo at 6 weeks, 3 months, and 6 months post-injection. No serious adverse events.	PRP injections can be a viable treatment option for knee OA, offering short-term pain relief and functional improvement in knee OA patients.
Shen et al. (2017)	Systematic review and meta-analysis of RCTs	Symptomatic, radiographic knee OA	14 RCTs comprising 1423 participants	14 RCTs comprising 1423 participants, control included saline placebo, HA, ozone, and corticosteroids	Follow-up ranged from 12 weeks to 12 months	PRP significantly reduced knee pain, improved physical function, reduced disability in patients with OA, effect was more pronounced at 6 and 12 months post-injection compared to control treatments, suggesting a sustained benefit over time. When compared to HA and placebo, PRP consistently provided superior outcomes in terms of pain relief and functional improvement.	PRP can be a viable treatment option for knee OA, offering sustained pain relief and functional benefits up to 12 months post-injection.
Ahmad et al. (2018)	Prospective, single blinded, RCT	Symptomatic, radiographic knee OA, KL grade 1, 2 or 3	89 patients	Randomized into 2 groups: 45 received PRP or 44 received HA IA injections. Patients received 3 injections in the knee with a 2-week interval between injections	3 and 6 months	PRP treatment resulted in significant reductions in pain and improvements in knee function compared to HA. Ultrasound imaging revealed positive structural changes in the knee joint post-treatment, which were associated with the observed clinical improvements in the PRP group.	PRP treatment not only alleviates symptoms of knee OA but also induces favorable structural changes detectable via ultrasonography. These structural improvements may underlie the clinical benefits observed in patients receiving PRP therapy
Buendía-López et al. (2018)	Prospective, RCT	Symptomatic, radiographic knee OA, KL grade 1 or 2	106 patients	Randomized into 3 groups: 33 received daily NSAID treatment, 32 received a single HA injection, and 33 received a single leukocyte-poor PRP injection	1 year	PRP group demonstrated a 20 ​% reduction in WOMAC pain scores and a 24 ​% improvement in physical function. Both HA and NSAID groups also showed improvements in WOMAC pain and VAS scores; however, the PRP group experienced significantly better outcomes compared to the other groups (P ​< ​0.05). No significant differences were observed among the 3 groups regarding changes in KL grades or cartilage thickness.	Single leukocyte-poor PRP injection offers superior clinical improvement in pain reduction and functional enhancement compared to a single HA injection or daily NSAID administration over a one-year period. None of the treatments demonstrated a significant impact on the structural progression of the OA.
Elik et al. (2020)	Prospective, RCT	Chronic, symptomatic, radiographic knee OA	60 patients	Randomized into 2 groups: First group administered 4-mL PRP IA in three doses at one-week intervals, second group had only one dose of a 4-mL saline solution IA	1 and 6 months	At 1 and 6 months post injection, PRP treatment exhibited significant positive effects on pain, physical function, and quality of life, but it did not lead to an increase in ultrasound assessed distal femur cartilage thickness in knee OA patients.	PRP treatment had positive effects on the pain, physical function, and quality of life of patients with knee OA, but it did not increase cartilage thickness.
Bennell et al. (2021)	2-Group, placebo-controlled, participant-, injector-, and assessor-blinded, RCT	Symptomatic, radiographic medial knee OA, KL grade 2 or 3	288	Randomized to receive three weekly IA injections of either leukocyte-poor PRP or saline placebo.	1 year	After 12 months, the PRP group experienced a mean pain score reduction of 2.1 points, while the placebo group had a reduction of 1.8 points. The between-group difference was not statistically significant. The PRP group showed a mean cartilage volume decrease of 1.4 ​%, compared to a 1.2 ​% decrease in the placebo group. This difference was also not statistically significant.	IA PRP injections did not result in significant improvements in knee pain or preservation of medial tibial cartilage volume compared to placebo injections over a 12-month period.

3.2. PRP and regenerative medicine

It remains uncertain whether PRP can affect or regenerate the joint and whether clinical improvements are correlated with these joint structure changes. This highlights the need for further investigation into the mechanisms through which PRP works, and whether its benefits may be more limited to pain management rather than long-term disease modification.

In a prospective study of 22 patients with early onset knee OA, Halpern et al. [19] investigated the association of a single PRP injection on clinical outcomes and magnetic resonance imaging (MRI) structural changes by Outerbridge scores graded by 2 musculoskeletal radiologists who were blinded as to whether the MRI was performed before or after PRP treatment. At both 6 months and 1 year follow-up post injection, a PRP injection significantly decreased pain scores, while radiological scores increased at 6 and 12 months from baseline. Among the 15 knees that underwent MRI evaluation before and after PRP treatment, 12 (80 ​%) showed no significant progression of OA in the patellofemoral joint. Additionally, 83.3 ​% of cases (12 knees) exhibited no change in the appearance of OA in the lateral femoral and tibial compartments 1 year after PRP therapy compared to baseline. As well, PRP appeared to maintain MRI stability in at least 73 ​% of cases per knee of the medial compartment at 1 year follow-up [19].

In a recent randomized clinical trial involving 288 adults aged 50 years or older, Bennell et al. [13] assessed the impact of IA PRP injections on both symptoms and joint structure in individuals with symptomatic mild to moderate radiographic medial knee OA. The use of PRP injections, compared to placebo injections, showed a mean change in knee pain scores of −2.1 compared to −1.8 on an 11-point scale (range, 0–10). Additionally, there was a mean change in medial tibial cartilage volume of −1.4 ​% for PRP versus −1.2 ​% for placebo at the 12-month mark. However, neither of these comparisons reached statistical significance. The lack of a statistically significant benefit of PRP for the primary structural outcome suggests that PRP does not slow disease progression. The findings suggest that among adults with mild to moderate knee OA, there was no significant improvement in knee pain or slowing of disease progression when treated with PRP injections compared to saline injections [13] (Table 3).

In a prospective and randomized study of 98 patients exploring the impact of leukocyte-poor (LP)-PRP, HA, and oral NSAID on knee OA (33 given NSAIDs, 32 had a single HA injection, and 33 patients received a single PRP injection), the sole administration of LP-PRP exhibited clinical efficacy with reduction in pain and improvement in physical function among patients with early OA during the 52-week follow-up [20]. Across various clinical scores, including pain, stiffness, and physical function, the response to a single injection of LP-PRP surpassed that of a single HA injection and oral non-steroidal anti-inflammatory drugs [20]. In addition, X-ray and MRI evaluations were performed at baseline and at 52 weeks to monitor cartilage progression. Although a single LP-PRP injection proved effective, it did not impact Kellgren–Lawrence (KL) or cartilage thickness progression [20].

In another prospective randomized clinical trial, 60 patients with knee OA were assigned to receive PRP IA in three doses at one week intervals, while the other group received one dose of a saline solution IA [21]. At 1 and 6 months post injection, PRP treatment exhibited significant positive effects on pain, physical function, and quality of life, but it did not lead to an increase in ultrasound assessed distal femur cartilage thickness in knee OA patients [21].

In an 89 patient knee OA prospective RCT, patients either received 3 PRP (n ​= ​45) or HA (n ​= ​44) IA injections with a 2 week interval between each [22]. Both PRP and HA injections led to improvements in all clinical and ultrasonography outcome measures at both the 3 and 6 months follow-up and outcomes were notably superior in the PRP group in comparison to the HA group. In order to ascertain post injection changes to knee structural features, ultrasonography was performed at baseline, 3 and 6 months to assess the levels of synovial vascularity, synovial hypertrophy, and the presence of effusion. PRP injections resulted in significantly reduced scores for synovial vascularity, synovial hypertrophy and lower number of effusions compared to HA at all timepoints. Importantly, these positive changes in ultrasound structural appearance of PRP injected knees were correlated with clinical improvements suggesting that PRP injections are associated with better synovial hypertrophy and vascularity scores, as well as reduced effusion [22] (Table 3).

4. Bone marrow aspirate concentrate (BMAC)

Bone marrow aspirate concentrate (BMAC) is also another injection used in the treatment of knee OA. BMAC has been shown to contain high levels of white blood cells, platelets, anti-inflammatory growth factors and cytokines that decrease cell apoptosis and inflammation [23]. BMAC also contains mononuclear cells, of which about 0.001 ​% are mesenchymal stromal cells (MSCs), with chondrogenic potential and a paracrine effect in increasing growth factor and cytokine levels [24,25]. BMAC contains greater level of interleukin-1 receptor antagonist (IL-1Ra) which inhibits IL-1 proinflammatory effects as compared with PRP [26]. The theory behind the use of BMAC in the treatment of OA has focused on the presence of MSCs, although the exact mechanism of action is still unclear [24].

4.1. Clinical studies on BMAC and OA

Intra-articular injections of BMA and BMAC in the setting of knee OA have reported statistically significant benefits in improved pain scores and function [27].

Hernigou et al. [28] showed that in 30 patients who had bilateral knee OA (60 knees), subchondral injections of BMAC were more effective than TKA in terms of reducing revision surgeries (6/30 in TKA knees vs. 1/30 in BMAC knees) over a 12-year (average) period [28]. The same group reported 15-year follow-up data on 140 patients with bilateral knee OA, and concluded that BMAC has a sufficient effect on pain to postpone or avoid TKA in contralateral joints of patients with bilateral knee OA [29]. Although these studies used sub-chondral injections, the Hernigou group also compared IA injections of BMAC and showed both routes of administration were effective in reducing pain and improving function at 12-months [30].

Other studies have similarly confirmed that BMAC IA injections improve pain and function in knee OA, relative to baseline. Mautner et al. [31] also showed that 41 patients who received BMAC had 78 ​% responder rate (defined as at least a 25 ​% improvement in VAS pain scores), and Knee Osteoarthritis and Outcome Scores (KOOS) improved from baseline of 63.6 ​± ​2.94 to 79.2 ​± ​3.053. Similarly, improvements based on International Knee Documentation Committee (IKDC) and Western Ontario and McMaster Universities Arthritis Index (WOMAC) scores sustained out to 12-months was reported by Anz et al. [32] in 90 knee OA patients (KL grades 1–3, I.e., early-to-moderate OA). Kim et al. [33] reported significant improvements in 41 patients; VAS pain score decreased from 7.0 to 3.3 post-injection at 12 months while KOOS increased from 43.1 to 70.6.

However, level 1 studies failed to demonstrate superiority of BMAC to saline [34], PRP [32] or to corticosteroid [35]. The data from two RCTs conducted comparing BMAC to placebo [32,34] showed that BMAC resulted in significant improvements in pain and QOL at 12 months relative to baseline. However, these results were not statistically significant from saline injections, done in contralateral knees. The use of the contralateral knee as a placebo control limits interpretation and stresses the need for properly designed, controlled RCTs (Table 4).

Table 4.

Cell based injection clinical studies.

Study	Level of Evidence	Pathology	Number of Patients	Protocol	Follow-up	Main Findings	Conclusions
Vangsness et al. (2014)	Phase I/II, double-blind, RCT	Within 7–10 days post-surgery, partial medial meniscectomy	55	Participants were randomized to receive a single IA injection of either: 50 million allogeneic MSCs, 150 million allogeneic MSCs, placebo (sodium hyaluronate)	1 year	At 1 year, 24 ​% of patients in the 50 million MSC group and 6 ​% in the 150 million MSC group exhibited at least a 15 ​% increase in meniscal volume, as determined by MRI. No significant meniscal regeneration was observed in the placebo group. Patients receiving MSC injections reported a significant reduction in pain compared to baseline. No significant adverse events related to MSC injections were reported.	IA injection of allogeneic MSCs following partial medial meniscectomy is safe and may promote meniscal regeneration in a subset of patients, along with providing significant pain relief.
Vega et al. (2015)	Phase I/II, RCT	Chronic, symptomatic, radiographic knee OA, KL grade 2 to 4	30	Participants were randomized to 2 groups of 15 patients: Group 1 received allogeneic bone marrow MSCs by IA injection of 40 million ​cells, group 2 received IA HA (60 ​mg, single dose)	1 year	Patients receiving MSC injections reported significant pain relief compared to those treated with HA. MRI assessments showed a significant decrease in areas of poor cartilage quality in the MSC-treated group, indicating cartilage regeneration. No significant adverse events related to MSC injections were reported.	IA injection of allogeneic MSCs is a feasible and safe treatment for chronic knee OA, providing significant pain relief and improving cartilage quality.
Gupta et al. (2016)	Phase II, double-blind, multicentric, placebo-controlled, RCT	Symptomatic, radiographic knee OA, KL grade 2 to 3	60	Patients were randomized to receive different doses of cells (25, 50, 75, or 150 million cells) or placebo. Cells were adult human bone marrow-derived, cultured, pooled, allogeneic mesenchymal stromal cells (Stempeucel®)	1 year	IA administration of Stempeucel® was well-tolerated, with no significant adverse events reported, patients receiving the 25-million-cell dose of Stempeucel® exhibited a trend toward improvement in clinical outcomes, such as pain reduction and enhanced joint function, compared to baseline.	Stempeucel® is a safe therapeutic option for patients with knee OA, with evidence indicating potential benefits in pain reduction and functional improvement, particularly at the 25 million cell dose.
Shapiro et al. (2017)	Prospective, single-blind, placebo-controlled, RCT	Symptomatic, radiographic bilateral knee OA, KL grade 1 to 3	25 (50 knees)	Each patient received IA BMAC injection in one knee and IA saline injections in the contralateral knee, allowing for direct within-subject comparison	6 months	Both BMAC and placebo-treated knees showed significant improvements in pain and function compared to baseline. However, there were no significant differences between the BMAC and placebo groups in terms of clinical outcomes. No serious adverse events were reported.	BMAC injections are safe. They did not demonstrate superior efficacy over placebo in improving pain and function in patients with knee OA
Anz et al. (2020)	Prospective, RCT	Symptomatic, radiographic knee OA, KL grade 1 to 3	90	Randomized to 2 different study groups: Group 1 received a single IA injection of BMAC, group 2 received a single IA injection of PRP	1 year	Both BMAC and PRP groups showed significant improvements in pain and function compared to baseline. No significant differences were found between the 2 groups in terms of clinical outcomes at 1 year. Both treatments were well-tolerated, with no significant adverse events reported.	BMAC and PRP provide comparable clinical benefits for knee OA over 1 year. Since PRP is easier and less invasive to obtain than BMAC, it may be a more practical choice for OA treatment.
Hernigou et al. (2021)	Prospective, RCT	Symptomatic, radiographic bilateral knee OA, KL grade 1 to 4	60 (120 knees)	Each patient received 20 ​mL subchondral BMAC injections in one knee and 20 ​mL IA BMAC injections in the contralateral knee, allowing for direct within-subject comparison	2 years, up to 15 years	Knees treated with subchondral BMAC injections had a significantly longer time to TKA compared to those receiving IA injections. Patients reported greater improvements in pain relief and functional outcomes in knees treated with subchondral injections.	Subchondral administration of BMAC MSCs is more effective than IA injections in delaying the need for TKA and improving clinical outcomes in patients with knee OA over a 15-year period.
Mautner et al. (2023)	Phase 2/3, four-arm parallel, multi-center, single-blind, RCT	Symptomatic, radiographic knee OA, KL grade 2 to 4	480	Randomized to 3 different arms with a 3:1 distribution, arm 1: Autologous BMAC (n ​= ​120), CSI (n ​= ​40); arm 2: Umbilical cord tissue-derived MSCs (n ​= ​120), CSI (n ​= ​40); arm 3: Stromal vascular fraction (n ​= ​120), CSI (n ​= ​40)	1 year	Co-primary endpoints were the VAS pain score and KOOS pain score at 12 months versus baseline. Analyses of our primary endpoints, with 440 patients, revealed that at 1 year post injection, none of the 3 orthobiologic injections was superior to another, or to the CSI control. In addition, none of the 4 groups showed a significant change in MRI OA score compared to baseline.	Over a 1 year period, cell-based therapies do not provide superior outcomes compared to corticosteroid injections for knee OA.

5. Mesenchymal stromal cells (MSCs)

Cell based therapies hold significant promise for advancing regenerative medicine approaches in the treatment of OA [36]. Administering cells through direct injection can enhance the normally limited repair and mitigate destructive processes. A promising regenerative medicine therapy may be the IA injection of MSCs.

MSCs are ubiquitous in the body. They are multipotent progenitors originating from nonhematopoietic adult stem cell populations present in various tissues such as bone marrow, peripheral blood, adipose tissue, dermis, synovium, umbilical cord blood, placenta, and amniotic fluid [37]. According to the International Society for Cellular Therapy, human MSCs are defined as plastic adherent, positive for CD105, CD73, CD90 markers, negative for CD45, CD34, CD14 or CD11b, CD79alpha or CD19, HLA-DR and able to differentiate into osteoblasts, chondroblasts and adipocytes [38]. The capacity of MSCs is diverse, and include trophic, anti-inflammatory, and immunomodulatory properties [37]. It is these such properties that are thought to help with tissue repair and regeneration.

5.1. Anti-inflammatory and immunomodulatory properties of MSCs

The inflammatory state following injury or trauma can impair the natural repair processes of certain cells [37]. MSCs are able to exert anti-inflammatory and immunomodulatory responses to create an environment amenable to tissue regeneration. MSCs have shown to reduce inflammation and support other cells, thereby enhancing angiogenesis, cell survival and differentiation and preventing fibrosis [39].

5.2. Clinical trials using MSCs for OA

In a pilot study, Orozco et al. [40] examined the safety and clinical effectiveness of direct IA injection of 40 ​× ​10⁶ autologous human bone marrow derived MSCs in 12 knee OA patients. Following MSC treatment, patients reported reduced osteoarthritic knee pain, encountered no adverse effects, and displayed enhanced articular cartilage quality on MRI [40].

In a double-blinded, randomized controlled trial, Vangsness et al. [41] examined the safety and impact of IA injection of allogeneic human bone marrow-derived MSCs in OA knees post-partial medial meniscectomy. Among the 18 patients who received 50 ​× ​10⁶ MSCs and another 18 who received 150 ​× ​10⁶ MSCs, both groups exhibited evidence of meniscus regeneration and a significant reduction in knee pain compared to control patients who did not receive MSCs [41] (Table 4).

In a phase I and II proof-of-concept clinical trial, Jo and colleagues [42] explored the safety and effectiveness of IA injection of autologous human adipose tissue-derived MSCs in knee OA patients. Phase I involved groups receiving 1 ​× ​10⁷, 5 ​× ​10⁷ and 1 ​× ​10⁸ MSCs, each with 3 patients, while phase II comprised 9 patients receiving the higher dose of 1 ​× ​10⁸ MSCs. Patients treated with 1 ​× ​10⁸ MSCs experienced reduced knee pain, enhanced knee function, regeneration of osteoarthritic cartilage defects with hyaline-like cartilage observed during second-look arthroscopy, and no notable adverse events [42]. A significant observation was that the regenerated cartilage primarily concentrated in the medial femoral and tibial condyles, the areas most impacted by degeneration in the knee. This finding resonates with prior research suggesting that injected cells tend to adhere to diseased rather than healthy articular cartilage [43]. Moreover, these results underscore the homing capability of adipose-derived MSCs, underscoring their efficacy in addressing OA [42].

Vega and colleagues [44] conducted a randomized trial involving 30 patients with osteoarthritic knees, divided into two groups. Fifteen patients received IA injections of 40 ​× ​10⁶ allogeneic human bone marrow derived MSCs, while the remaining 15 patients in the control group received IA injections of HA. The MSC-treated group exhibited a notable reduction in knee pain, improved knee function, enhanced cartilage quality observed in defects on MRI, and reported no adverse events [44] (Table 4).

Gupta et al. [45] investigated the safety and effectiveness of IA injection of allogeneic human bone marrow-derived MSCs in knee OA. Sixty patients with osteoarthritis were randomized into groups receiving 25 ​× ​10⁶, 50 ​× ​10⁶, 75 ​× ​10⁶, 150 ​× ​10⁶ MSCs or no MSCs. Notably, the 25 ​× ​10⁶ MSC dose demonstrated the highest efficacy in reducing osteoarthritic knee pain, with no significant adverse events reported [45].

Chahal and colleagues [46] investigated the safety and efficacy of autologous human bone marrow derived MSCs in middle to late stage knee OA. Twelve osteoarthritic knee patients were enrolled and treated in groups of 3 with direct intra-articular injection of 1 ​× ​10⁶, 10 ​× ​10⁶, or 50 ​× ​10⁶ MSCs. The authors found that there were no local or systemic adverse events with the injections at any of the doses. With the injections, 2/3 of patients achieved the minimal clinically importance difference of 10 in the KOOS and QOL scores at 3 months. Analysis of MRI data showed that there were no changes in cartilage volume but did show improved MSC-mediated changes in synovitis.

A recent well-designed study by Mautner et al. [35] conducted a phase 2/3 multicentered randomized trial comparing various cell injection therapies to corticosteroid injection for knee OA. A total of 480 patients were randomized to three different treatment arms: (1) autologous BMAC, (2) umbilical cord tissue-derived MSCs, and (3) stromal vascular fraction, which were compared to corticosteroid injection. At 1 year post injection, the authors found no difference in VAS pain score and Knee injury and Osteoarthritis Outcome Score (KOOS) of any of the 3 biologic injections compared to corticosteroid injection. Furthermore, the MRI imaging scores showed no significant changes from baseline of any of the injection therapies indicating that, despite some clinical improvements in pain and function reported by patients, these did not correspond with observable structural changes in the knee joint as measured by MRI. The study utilized a comprehensive MRI scoring system from 0 to 69 to evaluate various aspects of knee joint health, including cartilage integrity, bone marrow lesions, meniscal status, and synovitis. Each parameter was graded to provide an overall osteoarthritis score, with higher scores indicating more severe joint degeneration [35]. (Table 4).

6. Recombinant human fibroblast growth Factor-18 (rhFGF-18)

Sprifermin, a recombinant human fibroblast growth factor 18 (rhFGF18), has been explored as a potential IA anabolic disease-modifying OA drugs (DMOAD). In preclinical models of OA, Sprifermin promotes the proliferation of articular chondrocytes responsible for producing hyaline cartilage, stimulates the synthesis of hyaline extracellular matrix in both in vitro and ex vivo settings [[47], [48], [49]] and increases cartilage thickness in the knee joint [50].

The FORWARD trial (Fibroblast Growth Factor 18 Osteoarthritis Randomized Trial) focused on injection of Sprifermin (rhFGF18) in patients with knee OA. Two pivotal studies came out of the FORWARD trial.

In the first study, Hochberg and colleagues [51] evaluated the efficacy of IA Sprifermin in increasing cartilage thickness in patients with symptomatic knee OA, with secondary outcomes assessing symptom relief. The study included 549 patients, KL grade 2 or 3 and patients were randomized to receive either 30 ​μg or 100 ​μg of Sprifermin administered IA every 6 or 12 months, or placebo (sham injections). The primary outcome was change in total femorotibial joint cartilage thickness over 2 years measured via quantitative MRI and secondary outcomes included changes in pain and function, assessed using WOMAC. Sprifermin demonstrated dose-dependent increases in cartilage thickness, with the 100 ​μg every 6 months regimen showing the largest effect. The mean change in total femorotibial joint cartilage thickness was 0.05 ​mm for the 100 ​μg every 6 months group compared to −0.02 ​mm in the placebo group (p ​< ​0.001). WOMAC scores improved modestly in all groups, including placebo, however, no statistically significant difference in symptom improvement was observed between Sprifermin and placebo groups. Sprifermin demonstrated increase cartilage thickness in patients with knee OA, but the trial did not demonstrate significant symptomatic benefits of Sprifermin over placebo. These findings may limit its clinical applicability as a stand-alone treatment and raises questions about whether cartilage restoration alone is sufficient to address OA symptoms.

In the second study [52], the effect of IA injections of Sprifermin on cartilage thickness and clinical outcomes in patients with knee OA was evaluated. With a 5-year follow-up and the primary endpoint assessed at 2 years, over 500 patients with symptomatic knee OA, KL grade 2 or 3, received IA Sprifermin injections at doses of 30 ​μg or 100 ​μg per injection, administered either every 6 months or annually. A placebo group received sham injections. Sprifermin significantly increased total femorotibial joint cartilage thickness and volume compared to placebo. The increased cartilage thickness effect was dose-dependent, with the 100 ​μg dose every 6 months showing the largest improvement. However, improvements in pain, stiffness, and physical function measured by the WOMAC were modest and did not consistently correlate with structural improvements (Table 5).

Table 5.

Sprifermin, recombinant human fibroblast growth Factor-18 (rhFGF-18) clinical studies.

Study	Level of Evidence	Pathology	Number of Patients	Protocol	Follow-up	Main Findings	Conclusions
Hochberg et al. (2019)	Multi-center, double blind, dose finding, placebo, RCT	Symptomatic, radiographic knee OA, KL grade 2 or 3	549	Randomized to receive either 30 ​μg or 100 ​μg of Sprifermin administered IA every 6 or 12 months, or placebo (sham injections)	2 years	Sprifermin demonstrated dose-dependent increases in cartilage thickness, with the 100 ​μg every 6 months regimen showing the largest effect. The mean change in total femorotibial joint cartilage thickness was 0.05 ​mm for the 100 ​μg every 6 months group compared to −0.02 ​mm in the placebo group. WOMAC scores improved modestly in all groups, including placebo, however, no statistically significant difference in symptom improvement was observed between Sprifermin and placebo groups	Sprifermin demonstrated increase cartilage thickness in patients with knee OA, but the trial did not demonstrate significant symptomatic benefits of Sprifermin over placebo
Eckstein et al. (2021)	Multi-center, double blind, dose finding, placebo, RCT	Symptomatic, radiographic knee OA, KL grade 2 or 3	378	Randomized to receive either 30 ​μg or 100 ​μg of Sprifermin administered IA every 6 or 12 months, or placebo (sham injections)	5 years	Sprifermin significantly increased total femorotibial joint cartilage thickness and volume compared to placebo. The increased cartilage thickness effect was dose-dependent, with the 100 ​μg dose every 6 months showing the largest improvement. However, improvements in pain, stiffness, and physical function measured by the WOMAC were modest and did not consistently correlate with structural improvements	Sprifermin can increase cartilage thickness, however, improvements in pain and function were modest and did not consistently correlate with structural improvements

7. Gene therapy in OA

Gene therapy offers a novel approach to OA treatment by addressing the molecular mechanisms that drive cartilage degradation, inflammation, and pain. Progress in gene delivery methods, utilizing both viral and non-viral vectors, coupled with the discovery of key therapeutic genes, has enabled significant advancements in preclinical research and clinical trials aimed at combating OA. A few landmark studies have advanced this field and stand out as pivotal in demonstrating its potential in humans.

Cherian et al. [53] evaluated the efficacy and safety of injectable genetically engineered chondrocytes virally transduced to express transforming growth factor-beta 1 (TGF-β1) in patients with grade 3 knee osteoarthritis OA. This study was a multi-center, double-blinded, placebo-controlled, randomized study with 102 patients with knee OA, randomized in a 2:1 ratio to receive a single IA injection of either the genetically engineered chondrocytes expressing TGF-β1 (GEC-TGF-β1) or a placebo. The primary outcomes were knee joint function assessed using the International Knee Documentation Committee (IKDC) score and pain measured by the VAS. Secondary outcomes included pain severity and frequency, analgesic usage, QOL assessments, adverse events and conversion to TKA post-treatment. The preliminary results suggest that IA injection of GEC-TGF-β1 were both safe and effective in improving knee joint function (IKDC scores week 12: least mean square difference (LSMD) of 10.3, week 52: LSMD of 13.6, overall: LSMD of 8.6) and reducing pain in patients (VAS scores week 12: LSMD of −13.8, week 52: LSMD of −13.1, overall: LSMD of −10.1) and less analgesic use (week 4: 27 ​% vs. 40 ​%, week 12: 27 ​% vs. 37 ​%) with grade 3 knee OA compared to placebo. This study highlights the potential of gene therapy using genetically engineered chondrocytes expressing TGF-β1 as a novel treatment approach for knee OA (Table 6).

Table 6.

Gene therapy in OA clinical studies.

Study	Level of Evidence	Pathology	Number of Patients	Protocol	Follow-up	Main Findings	Conclusions
Cherian et al. (2015)	Multi-center, double-blinded, placebo-controlled, RCT	Symptomatic, radiographic knee OA, KL grade 3	102	Randomized in a 2:1 ratio to receive a single IA injection of either the genetically engineered chondrocytes expressing TGF-β1 (GEC-TGF-β1) or a placebo	1 year	IA injection of GEC-TGF-β1 were both safe and effective in improving knee joint function (IKDC scores week 12: Least mean square difference (LSMD) of 10.3, week 52: LSMD of 13.6, overall: LSMD of 8.6) and reducing pain in patients (VAS scores week 12: LSMD of −13.8, week 52: LSMD of −13.1, overall: LSMD of −10.1) and less analgesic use (week 4: 27 ​% vs. 40 ​%, week 12: 27 ​% vs. 37 ​%) with grade 3 knee OA compared to placebo	IA injection of GEC-TGF-β1 is both safe and effective in improving knee joint function and reducing pain in patients with grade 3 knee OA. Patients receiving GEC-TGF-β1 exhibited significant improvements in IKDC and VAS scores and were less likely to require analgesics compared to the placebo group

Currently, there are two clinical trials that are completed, however, results have not yet been published. ClinicalTrial.gov ID NCT02790723 [54] focused on delivering the IL-1Ra gene using a self-complementary recombinant adeno-associated virus vector to inhibit IL-1, a pro-inflammatory cytokine implicated in OA progression. The second study, ClinicalTrial.gov ID NCT03769662 [55] focused on assessing XT-150, a novel non-viral, locally injectable plasmid DNA gene therapy designed to express IL-10 and modulate pathological inflammation, thereby alleviating pain and improving joint function in OA patients.

8. Discussion

This narrative review highlighted the evidence on the efficacy of IA injections in the management of knee OA. However, there are a number of short comings when looking at the current evidence.

Corticosteroid injections are commonly used in clinical practice currently, and have been shown to provide pain relief when compared to saline controls. However, corticosteroids do not exhibit any regenerative capacity, and may in fact accelerate cartilage loss. Ultimately, these injections do not work as a treatment for OA, but rather help to just manage symptoms.

PRP injections for knee OA have been shown to result in improvements in pain, function, and QOL up to 12 months. However, the effect of PRP on joint structure is still unclear. There are also serious concerns regarding the reported evidence on the clinical effectiveness of PRP. Many studies on PRP have a significant risk of bias, particularly with the lack blinding of patients. Additionally, the studies often lack a conservatively managed control group, with some studies having small sample sizes. Because of the very low certainty of the available evidence and the need for more robust and thorough research, leading clinical guidelines for OA, such as those from the American College of Rheumatology [56] and the Osteoarthritis Research Society International (OARSI) [8], recommend against using PRP. Future studies involving better design and methodology are needed to assess the true efficacy of PRP injections.

Studies on BMAC and MSC injections have shown they can help improve pain and function. However, many of these studies involve only small sample cohorts. In contrast, when patients are assessed in larger well-designed studies, such as the recent Level 1 RCT by Mautner et al. [35], no differences in clinical outcomes or MRI changes are seen when compared to controls. Finding such as these question the efficacy of BMAC and MSC injections.

There have also been investigations into injections targeting cartilage repair for knee OA. The FORWARD trial was one such landmark investigation. The trial provided compelling evidence that Sprifermin (rhFGF18) can increase cartilage thickness, marking a significant step toward addressing structural damage in OA. However, the investigators did not find associated improvements in patient function. These findings unfortunately raise more questions and uncertainty about if structural improvements in knee OA provide clinically significant benefits.

While there is some international effort to provide the clinical proof of concept on the potential efficacy of IA therapies in well controlled, blinded and RCTs, the most recent high-quality RCTs fail to show intra-articular injections work to treat OA. Thus, the clinical applications for injections in the treatment of OA are not supported at this time.

So where do we go from here? We should first determine if the current IA therapies actually work to treat OA. One factor to consider is that current studies may be imparting treatments too late in the disease process. Patients with symptomatic joint pain are most commonly diagnosed with OA on radiographs. However, a radiographic diagnosis of OA establishes that the disease process has already ensued. In many instances, the OA changes are advanced and past the reversible state. In order to reverse or halt the disease process related to OA, early detection is key. Multiple examples within medicine highlight the importance of early detection of disease, such as in cardiovascular disease and diabetes. Within musculoskeletal conditions, rheumatoid arthritis exemplifies this. Early detection of rheumatoid arthritis and implementation of disease modifying drugs have resulted in less advanced joint damage and arthritis [57]. MRI has improved such that early arthritic changes, not seen on radiographs, can be detected. However, the exorbitant costs needed to screen the general population for early OA and the limited healthcare resources make it such that this is not a realistic option. An alternative could be testing for the presence of serum biomarkers in early OA. Ahmed et al. [58] showed that increased levels of plasma citrullinated protein (CP) were found in patients with early OA when compared to healthy controls. Furthermore, they found higher CP concentrations in the plasma than in the synovial fluid in patients with early OA. In contrast, patients with advanced OA had higher CP concentrations in the synovial fluid. Future studies are needed to focus on better detecting early OA.

It has also been suggested that personalized medicine may be of importance and yield better results [57]. For instance, not all patients with early OA progress to advanced OA. The potential to focus injection therapies to each individual patient may be the future and needs further investigation.

Finally, future studies may need to focus on identifying new targets for intervention or discovering new injection therapies through different mechanisms of action. At the moment, gene therapy for knee OA is being investigated. While most trials are early-phase and focus on safety and feasibility, they pave the way for future innovations that could potentially modify disease progression and alleviate symptoms. Despite its promise, challenges such as vector safety, immune system reactions, and the longevity of gene expression remain significant barriers to the routine clinical use of gene therapy. To overcome these hurdles, future research should prioritize refining gene delivery systems, improving vector safety, and customizing treatments to address individual patient needs, ultimately aiming to enhance the long-term effectiveness of gene therapy for OA.

There is an increasing interest in IA injections in OA for regenerative medicine as a potential solution to overcome the clinical short comings of current treatments that produce symptomatic rather than regenerative results. The ultimate objective is to provide patients with OA an IA injection that is proven to be safe and effective, utilizing a product that is consistent, manufactured under strict regulatory guidelines and cost effective. However, the search for an IA injection that provides clinical and functional improvement, restores joint homeostasis, and has a disease modifying effect is still elusive.

Author contributions

SP, BL and CK contributed to the conception and design of the review. SP prepared the first draft of the manuscript and SP, BL, and CK contributed to the critical revision of the manuscript for important intellectual content. SP, BL and CK approved the final version of the manuscript.

Declaration of competing interest

All authors declare that they have no financial and personal relationships with other people or organisations that could inappropriately influence (bias) this manuscript.

Handling Editor: Professor H Madry