Treatments ORIGINAL RESEARCH Efficacy and safety of intra-­articular therapies in rheumatic and musculoskeletal diseases: an overview of systematic reviews Sebastián Cruz Rodriguez-­García ‍ ‍,1 Raul Castellanos-­Moreira ‍ ‍,2 Jacqueline Uson ‍ ‍,3,4 Esperanza Naredo ‍ ‍,5,6 Terence W O'Neill ‍ ‍,7 Michael Doherty ‍ ‍,8 Mikael Boesen ‍ ‍,9 Hemant Pandit ‍ ‍,10 Ingrid Möller Parera ‍ ‍,11,12 Valentina Vardanyan ‍ ‍,13 Lene Terslev ‍ ‍,14 Will Uwe Kampen ‍ ‍,15 Maria Antonieta D'Agostino ‍ ‍,16,17 Francis Berenbaum ‍ ‍,18,19 Elena Nikiphorou ‍ ‍,20,21 Irene Pitsillidou,22 Jenny de la Torre-­Aboki ‍ ‍,23 Loreto Carmona ‍ ‍24 To cite: Rodriguez-­García SC, ABSTRACT Castellanos-­Moreira R, Uson J, Objective To summarise the evidence on intra-­ Key messages et al. Efficacy and safety of articular therapies (IAT) to inform the 2020 EULAR intra-­articular therapies in recommendations. What is already known about this subject? rheumatic and musculoskeletal ►► Intra-­articular therapies are frequently used in clini- Methods An overview of systematic reviews (SR) diseases: an overview of cal practice by a wide range of health professionals systematic reviews. RMD Open including randomised-­controlled trials (RCTs) of IAT in adults with arthropathies was performed up to July 2020. from different specialties.Several compounds are 2021;7:e001658. doi:10.1136/ Pain, function, and frequency of adverse events were the currently available for intra-­articular administration, rmdopen-2021-001658 main efficacy and safety outcomes, respectively. Quality from glucocorticoids to the more recent platelet-­rich was assessed with the A MeaSurement Tool to Assess plasma or mesenchymal stem cells. Nonetheless, ►► Additional supplemental data on their efficacy in certain diseases are incon- material is published online only. Systematic Reviews (AMSTAR)-2 tool. To view, please visit the journal Results Of 184 references identified, 16 met the inclusion sistent and a matter of debate. online (http://​dx.​doi.​org/​10.​ criteria, and a search of their reference lists identified 16 What does this study add? 1136/​rmdopen-​2021-​001658). additional SRs. After quality assessment, 29 were finally ►► This overview of systematic reviews provides a included. Of these, 18 focused on knee osteoarthritis summary of the current evidence on the efficacy and (KOA), 6 on hip osteoarthritis (HOA), 3 on shoulder Received 10 March 2021 safety of most compounds commonly used for intra-­ capsulitis (SC), and 3 on rheumatoid arthritis. Overall, Accepted 12 May 2021 articular injections. hyaluronic acid showed a small effect on pain and function in KOA but not in HOA or shoulder capsulitis. Intra-­articular How might this impact on clinical practice or glucocorticoids showed a small effect in pain and function future developments? in KOA and function in HOA and SC. Platelet-­rich plasma ►► This overview of systematic reviews informed the showed benefit in pain and function in KOA but not in task force for the 2021 EULAR recommendations for HOA. Mesenchymal stem cells behaved similarly. Most SR intra-­articular therapies and constitutes an evidence results were of moderate quality and RCTs included often base for future updates presented a high risk of bias, mainly due to inadequate blinding and heterogeneous results. All interventions were well tolerated with no clear safety differences. are used in many joint disorders including © Author(s) (or their Conclusions This overview underlines that most IAT employer(s)) 2021. Re-­use osteoarthritis (OA) and rheumatoid arthritis currently used in KOA, HOA, and SC exert small effects permitted under CC BY-­NC. No (RA) and delivered by a range of health commercial re-­use. See rights and are well tolerated. However, no firm conclusions can be drawn for inflammatory arthritis due to the limited data professionals including clinicians from a and permissions. Published by BMJ. found. range of specialities and also allied health- For numbered affiliations see care professionals.2 3 However, evidence on end of article. the efficacy and safety of available therapies is not always consistent, due in part to methodo- Correspondence to INTRODUCTION logical limitations in published trials.4 5 Dr Sebastián Cruz Rodriguez-­ García; Intra-­ articular therapies (IAT) have been Currently, many compounds are available ​sebastiancruz.​rodriguez@​salud.​ widely used in clinical practice for years to as IAT from glucocorticoids (GC)—methyl- madrid.​org reduce joint pain and improve function.1 They prednisolone acetate (MPA), triamcinolone Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658    1 RMD Open acetonide (TA), and triamcinolone hexacetonide (SAE), such as local reactions or swelling for the former (TH)—radioisotopes—yttrium-90, rhenium-186, etc—or and infections in the injected joint for the latter. hyaluronic acid (HA) to more recent therapies such as platelet-­rich plasma (PRP) and mesenchymal stem cells Search strategy (MSC), mostly used for treating OA.6–10 The arrival of A search was performed in MEDLINE with the assistance the latter three products on the market was accompanied of an expert librarian, from inception to January 2019 by a vast amount of literature with contradictory results and updated in July 2020. The references of the included that are still under debate. Furthermore, intra-­articular SRs were reviewed, as well as publications provided by procedures elicit an important placebo effect, something the members of the task force. Details on the complete that adds more complexity to its efficacy assessment.5 11–13 search strategy are provided in the online supplemental As around the world life expectancy, obesity, and material. sedentary lifestyle increase,14–16 the burden of disease imposed by chronic arthropathies and their comorbidi- Study selection and data collection ties also increases, thus providing the right scenario for Two investigators (SCR-­ G and RC-­ M) independently local treatments such as IAT, while the search for disease-­ screened the titles and abstracts to ascertain eligibility. modifying osteoarthritic drugs continues. The full texts of the eligible articles were then appraised Based on all this, a task force was assembled by the using the same approach, with discrepancies solved EULAR to produce recommendations for IAT in arthrop- through consensus, including a third investigator (LC) athies. The objective of the present work was to inform if needed. Data regarding study and population char- the task force about the current state of the evidence. acteristics, inclusion/exclusion criteria, interventions, outcome definition, outcome measures, and follow-­up was extracted using a standardised form. METHODS Methodological quality assessment Study design The same two investigators performed an independent We performed an overview of systematic reviews (SR) quality assessment of the eligible SRs using the ‘A MeaS- following a prespecified protocol. The present study is urement Tool to Assess Systematic Reviews (AMSTAR)−2’ reported following the Preferred Reporting Items for tool.20 Briefly, this instrument rates the overall confi- Systematic Reviews and Meta-­ analysis (PRISMA) state- dence in the results of a given SR by thoroughly analysing ment.17 seven critical domains. The quality was used as a crite- rion for inclusion. Only SRs of high or moderate quality Eligibility criteria were included unless a low quality focused on a disease To be eligible, the SR had to include randomised clinical or intervention not covered by the already included SRs. trials (RCT) assessing IAT in adults (≥18 years old) with any arthropathy, excluding the spine and temporoman- Data analysis dibular joints. The qualitative synthesis was carried out by disease and Interventions (IAT) could be any of the following: compound. For binary variables, we extracted the ORs GC, HA, PRP, MSC, radiopharmaceuticals, anaesthetics, or risk ratios (RR) with their 95% CI. For continuous opioids or biologicals. Comparators could be any of the outcomes, data were retrieved as mean difference (MD) above mentioned, any form of intra-­articular placebo or with 95% CI. When different measurements were used drugs administered orally as the standard of care (SoC), for the same outcome, treatment effects were retrieved as such as paracetamol/acetaminophen, non-­steroidal anti-­ standardised mean difference (SMD) with CI. To inter- inflammatory drugs, pregabalin, tricyclic antidepres- pret the magnitude of the effects, we used the criteria sants. Studies evaluating botulinum toxin as intervention proposed by Cohen.21 were excluded since its use was deemed to be irrelevant to the current clinical practice of the specialities repre- sented within the task force. Surgical procedures were RESULTS also excluded as comparators since they do not represent From a total of 183 references, after removing duplicates, the SoC in most diseases covered in the current study. 62 were selected for full-­text review and 16 met inclusion SRs assessing multiple comparators, including ozone or criteria. Additionally, 16 SRs were identified through the botulinum toxin, were included as long as they presented reference lists of included studies and after an update separate comparisons for the interventions mentioned in to July 2020. Hence, 32 SRs underwent quality assess- the inclusion criteria. ment. Three SRs were rated as of ‘high confidence’, 18 All efficacy and safety outcomes were considered, as ‘moderate’, 8 as ‘low’, and 3 as ‘critically low confi- especially change in pain and function with any avail- dence’. Following the prespecified protocol, the latter able measure, such as the Visual Analogue Scale (VAS), were excluded. Those rated as of low confidence were Lequesne index18 or the Western Ontario and McMaster finally included due to the low amount of data on the Universities Osteoarthritis Index (WOMAC),19 and studied compounds. Therefore, 29 SRs were included in adverse events (AE), including serious adverse events the qualitative synthesis. A flowchart is shown in figure 1 2 Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 Treatments analysed, although pooled OMERACT-­OARSI responses reached statistical significance (RR, 1.15 (1.02 to 1.30)).30 Finally, one SR compared HA compounds and showed an increasing effect with increased molecular weight (MW).22 Of note, the number of studies included was rather low and no differences were seen in QoL. Most SRs of HA reported moderate to high hetero- geneity between studies, as well as publication bias and other biases, mostly concerning inadequate blinding, allocation concealment, and reporting. Against placebo, GC compounds showed small to moderate effect sizes for pain and function in the short-­ term (until 3 months), and no differences in QoL, stiff- ness or joint space width.23 Among GC compounds, MPA shows a faster onset of effect on pain and function than TA or TH at 6 weeks.34 No differences were detected after this time-­point as well as in OMERACT-­OARSI response and no pooled analysis was performed for this compar- ison. As with HA, authors underline inadequate blinding and allocation concealment as possible sources of bias in the included RCTs. PRP was evaluated mostly against HA and, second- Figure 1 Flow chart of the overview of systematic reviews arily, versus placebo. Compared with HA, PRP showed a (SR). small to null effect on pain, function, and stiffness. Two SRs pooled composite scores (WOMAC total score and and a list of excluded articles with reasons for exclusion is IKDC) and found better responses with PRP than HA provided in the online supplemental material. at 6 and 12 months showing large effects.27 49 Kanchan- The main features of the SRs included are summarised atawan et al25 found an improved EQ-­VAS at 12 months in table 1. Knee osteoarthritis (KOA) was analysed in 18 with PRP.25 For PRP versus placebo, no differences were SRs,4–7 22–35 hip OA in 6,36–42 shoulder adhesive capsulitis seen in the targeted outcomes, except for the composite in 3,43–45 and RA in 3.34 46 47 One SR analysed the efficacy scores, in which the pooled effect was large; this effect of IAT in both KOA and RA.34 Different HA-­containing disappeared when only high-­quality trials were pooled. compounds were assessed in 13 SR,4–7 22 30 31 33 35 37 40 44 47 Between-­trial heterogeneity was high, in terms of PRP PRP in 8,25–27 29 32 36 39 41 42 GC in 6,23 28 34 38 43 45 and MSC composition, endpoints, and comparators. Also, the SRs and yttrium synovectomy in 1 each.24 46 rated included RCTs as with moderate to high risk of bias, especially due to inadequate allocation concealment, Efficacy of intra-articular treatments blinding of participants, and outcome assessment. Knee osteoarthritis A network meta-­analysis analysed the effect of MSC The main efficacy results are shown in table 2. The most against different comparators, including placebo, HA, frequent outcomes were pain, function, OMERACT-­ or IA GC.24 The effect of MSC was moderate to large on OARSI responder index, and quality of life (QoL). An SR pain and moderate for the KOOS at 12 months, whereas included the change in joint space width and cartilage no effect was observed on the WOMAC total score at 6 volume.31 months. High-­dose adipose-­derived MSC showed a longer HA compounds were extensively analysed in compar- effect. Overall, studies included in this SR were rated as of ison mostly against IA placebo followed by IA GC. low risk of bias; nonetheless, there was evidence of publi- Compared with the former and according to Cohen’s cation bias for pain measured by VAS. Unfortunately, criteria,21 the effect sizes observed for the intervention most branches of the meta-­analysis were underpowered on pain and function were small and further reduced to draw conclusions on which strategy is better in clinical to no effect when pooling large-­blinded RCTs only. An practice. SR analysed the OMERACT-­OARSI response and found that patients treated with HA were more likely to achieve Hip osteoarthritis such a response than those receiving placebo (RR, 1.11 The main results on hip OA are summarised in table 3. (1.01 to 1.20)).30 Likewise, when compared vs IA GC, The most frequent outcomes measured were pain and the effect sizes of the intervention were small on pain function, the latter measured using the Harris Hip and function. Of note, one study favoured IA GC in the Score (HHS) and the OMERACT-­ OARSI response 1-­week to 2-­week assessment and HA from the 7–10 weeks criteria. until the 17-­week to 29-­week evaluations.48 In other SRs, PRP was the most frequent compound studied in hip there were no differences between groups in most RCTs OA, and all comparisons were against HA. Almost all Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 3 RMD Open Table 1 Main characteristics of the SR included Intervention and Study Population comparator Outcomes Quality Knee osteoarthritis Rutjes et al7 IC: RCTs HA vs sham or no Primary: pain intensity High EC: not stated. intervention Secondary: function, SAEs, withdrawal due to AEs Newberry et al22 IC: RCTs, SRs, OS, and CS* HA vs PBO or other Primary: delay or avoidance of High EC: non-­English language studies HA TKR and conference abstracts. Secondary: function, QoL, number of AE Jüni et al23 IC: RCT of patients treated with IA GC vs sham, PBO Primary: pain and function at 4–6 High GC either IA or subacromial. or SOC weeks EC: RCT including only patients Secondary: pain and function at with inflammatory arthritis subsequent time points, QoL, JSN, SAEs, withdrawals due to AEs Ding et al24 IC: RCTs reporting ≥1 of the MSC vs PBO, HA or WOMAC, KOOS, VAS, SAEs Moderate outcomes of interest. IAGC without a prespecified hierarchy EC: use of PRP or MSC+surgery or lack of a non-­cell-­based control Bannuru et al4 IC: RCTs of patients treated with HA vs HA or PBO Number of AEs, SAEs, Moderate HA with data on safety outcomes withdrawals due to AEs without a EC: non-­RCT studies prespecified hierarchy Bannuru et al6 IC: RCTs with data for ≥1 HA vs IAGC Primary: pain according to a Moderate outcome measure of pain. prespecified hierarchy at different EC: studies not including pain time-­points outcomes of interest Bannuru et al5 IC: RCTs of patients with primary HA vs PBO Primary: pain at 3 months Moderate KOA with data on ≥2 interventions HA vs IAGC according to a prespecified of interest and on ≥1 measure of IAGC vs PBO hierarchy pain, function or stiffness. Secondary: function and stiffness EC: not stated at 3 months Kanchanatawan et IC: RCTs of adults with primary PRP vs HA or PBO or WOMAC total and subscores, Moderate al25 KOA with ≥1 of the outcomes sham Lequesne score, EuroQol-­VAS, of interest and enough data to IKDC subjective scores, number extract and pool of AEs without a prespecified EC: not stated hierarchy Xu et al26 IC: RCTs with ≥30 randomised PRP vs HA, PBO Pain and function (VAS, WOMAC, Moderate patients, ≥1 month follow-­ IKDC, Lequesne) without a up, quantitative outcome prespecified hierarchy assessment, <20% of dropouts EC: not stated Dai et al27 IC: RCTs comparing PRP PRP vs HA or PBO Primary: WOMAC pain and Moderate vs controls for prespecified function scores. outcomes Secondary: WOMAC total score, EC: not stated IKDC, Lequesne, frequency of AE Arroll and Goodyear-­ IC: PBO-­controlled RCTs IAGC vs PBO Primary: improvement of Moderate Smith28 assessing the efficacy of IAGC symptoms EC: not stated Secondary: pain, response to the OA research scale Shen et al29 IC: RCT comparing any PRP vs PRP vs HA or PBO Primary: WOMAC pain, function Moderate another IAT with ≥12 w follow-­up and total at 3, 6, and 12 months EC: studies without IA control Secondary: number of patients group, other PRP or PRP+surgery with AEs Continued 4 Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 Treatments Table 1 Continued Intervention and Study Population comparator Outcomes Quality 30 Trojian et al IC: RCTs in English including HA vs PBO or IAGC OMERACT-­OARSI response Moderate outcomes of interest at ≥8 and IAGC vs PBO rates, mean change from baseline <16 weeks. in WOMAC pain, stiffness or EC: studies comparing IA GC or function, frequency of AE. Without HA vs surgical procedures hierarchy Gallagher et al31 IC: RCTs with PBO control, ≥12 HA or SOC vs PBO† Primary: changes in JSW or Moderate m follow-­up, data on structural cartilage volume. changes Secondary: WOMAC total score, EC: not stated WOMAC pain or VAS pain Di et al32 IC: English-­written RCTs PRP vs HA Primary: WOMAC, IKDC, KOOS, Low EC: unknown methodology EQ-­VAS, Tegner score. or patients with additional Secondary: frequency of AE conditions‡ between groups Trigkilidas and IC: RCTs with ≥1 outcome HA vs PBO or IAGC VAS pain, Lequesne, WOMAC Low Anand35 measure on pain or function; without a prespecified hierarchy freely available as full text from specified sources§ EC: non-­RCT and language other than English Lo et al33 IC: Blinded—RCTs comparing HA HA vs PBO Pain according to a prespecified Low (≥3 injections) vs PBO with data hierarchy on pain and 8-­week minimum follow-­up and drop-­out rate of <50% EC: not stated Hip osteoarthritis Ali et al36 IC: RCTs, with clinical and PRP vs HA VAS pain, WOMAC total, and HHS Moderate functional data with any follow-­up without a prespecified hierarchy EC: studies on animals and technical notes McCabe et al38 IC: RCTs with patients with HOA IAGC vs PBO Primary: pain according to a Moderate (clinical and radiographic) prespecified hierarchy EC: studies without a control Secondary: WOMAC function, group Lequesne Index, safety profile Liao et al37 IC: RCTs of patients with primary HA vs PBO Primary: self-­reported pain Moderate HOA according to a prespecified EC: stated as the opposite to IC hierarchy Secondary: function, OMERACT-­ OARSI responder index Medina-­Porqueres41 IC: English or Spanish-­written PRP vs IA control Primary: VAS pain, HHS, and Low studies of PRP applied in isolation (any) WOMAC function. in ≥1 arm to patients with any Secondary: growth factor’s grade of HOA as per the ACR concentration, AE and imaging criteria evaluations EC: studies including only children or animals; non-­OA injuries, OA in other joints or previous surgery 42 Ye et al IC: RCTs comparing PRP with HA PRP vs HA Primary: WOMAC total score, VAS Low EC: studies without a control pain, and Harris hip score (HHS) group, full-­text versions or Secondary: n of AE outcomes data Continued Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 5 RMD Open Table 1 Continued Intervention and Study Population comparator Outcomes Quality 40 Leite et al IC: RCT with ≥1 of the outcomes HA vs IA-­injection Primary: pain Low of interest comparators Secondary: QoL, OMERACT-­ EC: RCT comparing HOA vs other OARSI Response, frequency of sites and HA vs non-­IA controls AEs Shoulder capsulitis Sun et al45 IC: RCTs comparing IAGC vs no IAGC vs sham or Primary: VAS pain Moderate or sham injection or SOC SOC Secondary: passive external EC: injection volume >0.10 mL rotation, abduction, flexion, (classified as IAGC+distention) internal rotation,and functional scores and frequency of AEs Buchbinder et al43 IC: RCTs of shoulder pain IAGC vs PBO, other Pain, ROM, function, strength, and Moderate comparing IAGC vs PBO, another interventions return to work or school without a intervention or different IAGC prespecified hierarchy dosages EC: pain duration <3 weeks, RA, polymyalgia rheumatica, and fracture Lee et al44 IC: RCT of capsulitis (confirmed HA vs SOC Pain, ROM, and function/ Moderate clinically or by US), clearly disability scores >1 month after documenting IC and EC, administration, frequency of AEs symptom duration and follow-­up without a prespecified hierarchy >4 weeks EC: uncontrolled studies Rheumatoid arthritis Heuft-­Dorenbosch IC: RCTs of RA patients with knee Yttrium synovectomy Knee circumference, ROM, Moderate et al46 arthritis, enough quality as per the vs PBO or TH fixed flexion, pain (Likert scale), Delphi list. Language restrictions subjective change, knee effusion, applied¶ radiological assessment without EC: not stated prespecified hierarchy Silvinato and IC: RCTs of patients with RA and MPA vs TA, TH, Primary: flare time at 24 weeks, Low Bernardo34 knee arthritis prednisolone Secondary: patient-­reported pain EC: not stated and swelling, ROM, frequency of AEs Saito and Kotake47 IC: English or Japanese-­written HA vs PBO Primary: global pain measured Low RCTs of patients with RA and with Likert scale at 1 week knee arthritis including pain Secondary: inflammation assessment measured with Likert scale. EC: studies with animals or only Condition of the knee with Likert describing the injection technique scale, safety profile *Only data from RCTs were retrieved for the analyses on the present study. †Only data for the HA vs PBO comparison were retrieved. ‡Additional conditions included meniscal tears, inflammatory arthritis, among others. §Free full-­texts available from the Warwick University Library or Google Scholar. ¶Articles written in Dutch, English, French, German, or Spanish. AE, adverse events; CS, case series; EC, exclusion criteria; EQ-­VAS, Euro Quality of Life – Visual Analogue Scale; freq of AE, frequency of adverse events; GC, glucocorticoids; HA, hyaluronic acid; HHS, Harris Hip Score; HOA, hip osteoarthritis; IA, intra-­articular; IAT, intra-­ articular therapies; IC, inclusion criteria; IKDC, International Knee Documentation Committee; JSN, joint space narrowing; JSW, joint space width; KOA, Knee Osteoarthritis Index; KOOS, Knee injury and Osteoarthritis Outcome Score ; MPA, methylprednisolone acetate; MSC, mesenchymal stem cells; OS, observational studies; PBO, placebo; PRP, platelet-­rich plasma; QoL, quality of life; RCT, randomised-­ controlled trials; ROM, range of motion; SAE, serious adverse events; SoC, standard of care; TA, triamcinolone acetonide; TH, triamcinolone hexacetonide; TKR, total knee replacement; US, ultrasonography; VAS, Visual Analogue Scale; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index. RCTs showed no difference between groups at all time WOMAC function subscore or the HHS. An SR of four points except for the study by Ye et al,42 favouring PRP. RCTs with high heterogeneity and unclear or high risk of Regarding function, no differences were seen using the bias showed inconclusive results.41 50 51 6 Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 Treatments Table 2 Main efficacy results of IAT for knee osteoarthritis Study Follow-­up Outcomes Effect estimate Comments Hyaluronic acid vs placebo Rutjes et al7 3 mo Pain Overall (ES, 0.37 (0.28 to 0.46)), favouring HA Effect size defined as between-­ Large-­blinded RCTs (ES, 0.11 (0.04 to 0.18)), group differences in means divided favouring HA by the pooled SD at end of follow-­ up. Function Overall (ES, 0.33 (0.04 to 0.22)), favouring HA Minimal clinically important Large-­blinded RCTs (ES, 0.09 (0.00 to 0.17)), difference = favouring HA (−0.37 ES) Newberry et al22 1–12 mo Function SMD=0.23 (0.01 to 0.45), favouring HA (WOMAC) Consistent effect in sensitivity analysis for too short (<4 weeks) or QoL 3 RCTs—no between-­group difference (SF-36, too long (>52 weeks) RCTs EuroQol-­5D) Gallagher et al31 12–24 mo Pain 2 RCTs—no between-­group difference (VAS)   ∆ JSW 2 RCTs—no between-­group difference ∆ Cartilage volume 1 RCT—favoured HA with 2.60% (1.20–4.10) less cartilage volume lost in the medial compartment and 2.80% (0.90–4.70) less in the lateral compartment Bannuru et al5 3 mo Pain SMD, 0.34 (Cr I, 0.26 to 0.42), favouring HA MA result of a Bayesian hierarchical random-­effects model for mixed Function SMD, 0.3 (Cr I, 0.20 to 0.40), favouring HA multiple treatment comparisons Stiffness SMD, 0.23 (Cr I, 0.13 to 0.34), favouring HA Trojian et al30 2–6 mo Pain SMD, 0.19 (0.06 to 0.32), favouring HA (WOMAC) NMA. SMD refers to Hedges’ g Results obtained for the time of best Function SMD, 0.19 (0.05 to 0.32), favouring HA (WOMAC) response Stiffness SMD, 0.12 (0.03 to 0.27), favouring HA (WOMAC) No publication bias O-­O Resp RR, 1.11 (1.01 to 1.20), favouring HA Trigkilidas and 1–6 mo Pain 5 RCTs—no between-­group difference (VAS) No pooled analysis Anand35 7 RCTs—favoured HA (VAS) (small effect) Function 5 RCTs—no between-­group difference (WOMAC, Lequesne) 7 RCTs—favoured HA (WOMAC) (small effect, Lequesne) Lo et al33 2–12 mo Pain Overall, SMD=0.32 (0.17 to 0.47) Evidence of publication bias Excluding high MW, SMD=0.19 (0.10 to 0.27) Hyaluronic acid vs glucocorticoids Bannuru et al 6 1–2 wk Pain ES, 0.39 (0.12 to 0.65), favouring IAGC ES: refers to Hedges’ g corrected for small samples 3–6 wk ES, −0.01 (−0.23 to 0.21), no between-­group Effects remained consistent after difference multivariable and sensitivity analysis 7–10 wk ES, 0.22 (0.05 to 0.49), favouring HA 11–16 wk ES, 0.35 (0.03 to 0.66), favouring HA 17–29 wk ES, 0.39 (0.18 to 0.59), favouring HA Bannuru et al 5 3 mo Pain SMD, 0.02 (Cr I, −0.12 to 0.17), no between-­group NMA difference Function SMD, 0.24 (Cr I, 0.06 to 0.43), favouring HA Stiffness SMD, 0.20 (Cr I, 0.0 to 0.41), no between-­group difference Trojian et al30 4–40 mo Pain ES, −0.06 (−0.28 to 0.16), no between-­group NMA difference SMD refers to Hedges’ g Results retrieved at the time of best Function ES, −0.29 (−0.53 to −0.05), favouring HA response Stiffness ES, −0.17 (−0.50 to 0.16), no between-­group No publication bias difference O-­O Resp RR, 1.15 (1.02 to 1.30), favouring HA Trigkilidas and 1–6 mo Pain 1 RCT—favoured HA at 6 months (VAS) No pooled analysis Anand35 Function 1 RCT—no between-­group difference Hyaluronic acid compounds comparison Continued Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 7 RMD Open Table 2 Continued Study Follow-­up Outcomes Effect estimate Comments 22 Newberry et al 1–12 mo Function 1 RCT—LMW vs MMW. SMD, −0.326 (−0.52 to −0.13), favouring MMW All comparisons using the WOMAC function subscale 1 RCT—LMW vs HMW. SMD, 0.053 (−0.66 to 0.77), No pooled analysis no difference *Results of the same study54 at 2 1 RCT—LMW vs HMW. SMD, −0.882 (−1.09 to time-­points −0.68), favouring HMW 1 RCT—MMW vs HMW. SMD, −0.01 (−0.21 to 0.19), no difference 3 mo QoL 1 RCT*—LMW vs HMW, favouring LMW (EuroQol-­5D) 12 mo 1 RCT*—LMW vs HMW, favouring HMW (EuroQol-­5D) 1 RCT—LMW vs HMW. No between-­group difference (SF-36) Glucocorticoids vs placebo Jüni et al23 2 wk Pain SMD −0.48 (−0.70 to −0.27), favouring IAGC For pain and function, effects were reduced in large trials (>50 patients/ 2 mo SMD −0.41 (−0.61 to −0.21), favouring IAGC arm) 3 mo SMD −0.22 (−0.44 to 0.00), no between-­group difference 6 mo SMD −0.07 (−0.25 to 0.11), no between-­group difference 2 wk Function SMD −0.43 (−0.72 to −0.14), favouring IAGC 2 mo SMD −0.36 (−0.63 to −0.09), favouring IAGC 3 mo SMD −0.13 (−0.37 to 0.10), no between-­group difference 6 mo SMD 0.06 (−0.16 to 0.28), no between-­group difference 6 mo QoL SMD −0.01 (−0.30 to 0.28), no between-­group difference JSW SMD −0.02 (−0.49 to 0.46), no between-­group difference Arroll and Goodyear-­ 2 wk Pain WMD −16.47 (−22.92 to −10.03), favouring IAGC †Pooling studies with the highest Smith28 dose 2 wk Improvement of RR 1.66 (1.37 to 2.01), favouring IAGC symptoms 3–4 mo RR 2.09 (1.20 to 3.65), favouring IAGC† Bannuru et al 5 3 mo Pain SMD, 0.32 (Cr I, 0.16 to 0.47), favouring IAGC NMA Function SMD, 0.06 (Cr I, −0.13 to 0.26), no between-­group difference Stiffness SMD, 0.03 (Cr I, −0.19 to 0.25), no between-­group difference Glucocorticoid compounds comparison Silvinato and 1–6 mo Pain 1-­RCT—MPA vs TH. No between-­group difference *Results of the same study at 2 time-­ Bernardo34 (VAS) points ¥Results of the same study at 2 6 wk 1-­RCT*—MPA vs TA vs prednisolone, favouring MPA time-­points (VAS) No pooled analysis 3 mo 1-­RCT*—MPA vs TA vs prednisolone, no between-­ group difference 1 month 1-­RCT¥—MPA vs TH, favouring MPA (VAS) 2 mo 1-­RCT¥—MPA vs TH. No between-­group difference (VAS) 1–6 mo Function 1-­RCT—MPA vs TH. No between-­group difference (WOMAC) 1–3 mo 1-­RCT—MPA vs TA vs prednisolone. No difference (Lequesne) 2 mo 1-­RCT—MPA vs TH. No between-­group difference (Lequesne) 2 mo O-­O Response 1-­RCT—MPA vs TH. No between-­group difference Platelet-­rich plasma vs placebo Continued 8 Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 Treatments Table 2 Continued Study Follow-­up Outcomes Effect estimate Comments 49 Xu et al 6 mo Composite scores# Overall, SMD −2.13 (−3.29 to −0.98), favouring PRP #Effects of pooled results from WOMAC and IKDC scores Dai et al27 6–12 mo Pain 1 RCT—favoured PRP (WOMAC)   Function 1 RCT—favoured PRP (WOMAC) Kanchanatawan et 6–12 mo Pain No between-­group difference (WOMAC)   al25 Function No between-­group difference (WOMAC) Stiffness No between-­group difference (WOMAC) Platelet-­rich plasma vs hyaluronic acid Xu et al26 6 mo Composite scores¶ Overall, SMD = −0.85 (−1.43 to −0.28) favouring PRP ¶ Refers to observed effects when pooling results from WOMAC and High-­quality RCTs, SMD = −0.09 (−0.30 to 0.11). No IKDC scores difference Pain SMD=0.35 (−0.36 to 1.06) (VAS). No difference Function MD=−0.20 (−1.00 to 0.60) (Lequesne). No difference 3 mo WOMAC total MD=−7.10 (−17.02 to 2.82). No between-­group difference 12 mo MD=−8.93 (−27.56 to 9.71). No between group difference Shen et al29 3–12 mo Pain MD=−3.77 (−5.07 to −2.47), favouring PRP (WOMAC) Results obtained from pooling outcomes at 3, 6, and 12 months Function MD=−13.91 (−18.53 to −9.28), favouring PRP (WOMAC) WOMAC total MD=−17.39 (−22.32 to −12.46), favouring PRP Dai et al27 6 mo Pain MD=−1.54 (−4.27 to 1.20). No between-­group  §Results from pooling WOMAC difference total, IKDC,EQ and Lequesne Index 12 mo MD=−2.83 (−4.26 to −1.39), favouring PRP 6 mo Function MD=−4.39 (−10.51 to 1.74). No between-­group difference 12 mo MD=−12.53 (−14.58 to −10.47), favouring PRP 6 mo Composite scores§ SMD=0.68 (−0.04 to 1.41). No between-­group difference 12 mo SMD=1.05 (0.21 to 1.89), favouring PRP Kanchanatawan et 6–12 mo Composite scores§ MD= −15.4 (−28.6 to −2.30), favouring PRP (WOMAC §Results for WOMAC total and IKDC al25 total) reached the prespecified MCID MD=8.83 (5.88 to 11.78), favouring PRP (IKDC) Pain No between-­group difference (WOMAC) Function No between-­group difference (WOMAC) Stiffness No between-­group difference (WOMAC) QoL MD=7.37 (4.33 to 10.05), favouring PRP (EQ-­VAS) Di et al32 1–12 mo Pain 5 RCTs—favoured PRP (VAS, WOMAC) No pooled analysis 1 RCT—no between-­group difference (VAS) Function 3 RCTs—favoured PRP (WOMAC, Lequesne, KOOS) 3 RCTs—no between-­group difference (WOMAC, Lequesne, etc) Stiffness 2 RCTs—favoured PRP (WOMAC) 2 RCTs—no between-­group difference (WOMAC) O-­O Response 1 RCT—favoured PRP QoL 3 RCTs—no between-­group difference (EQ-­VAS, SF- 36) Mesenchymal stem cells vs controls Continued Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 9 RMD Open Table 2 Continued Study Follow-­up Outcomes Effect estimate Comments 24 Ding et al 6 mo Composite scores SMD=−0.36 (−0.90 to 0.18). No difference (WOMAC NMA. total) vs controls Controls include HA, PBO, and GC. High-­dosage adipose-­derived MSC 12 mo SMD=0.68 (0.07 to 1.30), favouring MSC (KOOS) vs showed a longer effect controls 12 mo Pain SMD= −1.05 (−1.46 to −0.64), favouring MSC vs controls Results are ordered by compounds and quality. The colour of the cell denotes quality: the darker the higher the quality. All effect sizes (ESs) are presented as a point estimate (95% CI) unless otherwise noted. Cr I, credible intervals; EQ-­VAS, Euro Quality of Life – Visual Analogue Scale; EuroQol-­5D, Euro Quality of Life – 5 Dimension questionnaire; GC, glucocorticoids; HA, hyaluronic acid; HMW, high molecular weight; IAGC, intra-­articular glucocorticoids; IAT, intra-­articular therapies; IKDC, International Knee Documentation Committee; ∆JSW, change in joint space width; KOOS, Knee injury and Osteoarthritis Outcome Score; LMW, low molecular weight; MCID, minimal clinically important difference; MD, mean difference; MMW, medium molecular weight; mo, months; MPA, methylprednisolone acetate; MSC, mesenchymal stem cells; NMA, network meta-­analysis; O-­O Resp, OMERACT-­OARSI Responder Index; PBO, placebo; PRP, platelet-­rich plasma; QoL, quality of life; RCT, randomised controlled trials; RR, relative risk; SF-36, Short Form 36 health survey; SMD, standardised mean difference; TA, triamcinolone acetonide; TH, triamcinolone hexacetonide; VAS, Visual Analogue Scale; wk, weeks; WMD, weighted mean difference; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index. No differences were observed for pain, function nor morning stiffness, grip strength, Ritchie articular index, OMERACT-­OARSI response between HA and placebo or or thermography index, between MPA, TA, or TH. In one MPA. McCabe et al,38 on the contrary, reported an OR=7.8 RCT, TH performed better in pain (VAS) at 1 week of (2.7–22.8) for reaching an OMERACT-­OARSI response follow-­up but there were no between-­group differences in patients treated with IA GC versus placebo. The latter at 2 to 6 weeks. Finally, Heuft-­Dorenbosch et al46 found SR included four RCTs, three of which showed better no differences in pain between yttrium synovectomy results in function (activities of daily life and WOMAC and placebo or IAGC, whereas the former performed function subscore). All studies were deemed as having a better in ROM and knee circumference (1 RCT) versus low to moderate risk of bias and no evidence of publica- placebo. Conversely, ROM was best improved in the tion bias. IAGC-­treated group (vs yttrium synovectomy). Two out of three SRs assessing treatments for RA were deemed as Shoulder capsulitis of low quality and included a very low number of RCTs Table 4 summarises the main efficacy results for shoulder with evidence of small study bias and unclear or inade- capsulitis. Pain was only measured using VAS and func- quate allocation concealment, as well as participant and tion evaluated by the range of motion (ROM). Addition- provider blinding. ally, specific composite scores such as the Shoulder Pain and Disability Index (SPADI), the American Shoulder Safety of intra-articular treatments and Elbow Surgeons score and (ASES), and the Constant Twenty-­two SRs provided data on safety (table 5). In most score were applied. HA and IAGC were the interventions cases, the outcome reported was the frequency of AEs evaluated and most comparisons were against placebo. (any), while some articles also analysed SAEs and with- One SR44 assessed the former and found no differences for pain or function. On the contrary, IAGC were evalu- drawals due to AEs. ated in two SRs and a small effect was observed favouring HA compounds were compared against placebo in a the intervention on pain, ROM, and the SPADI whereas network meta-­analysis specifically designed to assess safety no differences were seen for the ASES and the Constant in KOA.4 No between-­group differences were observed score. for any AE but local reactions and withdrawal due to AEs Overall, there was high heterogeneity between favoured placebo versus HA. Other SRs analysing HA studies regarding injection techniques dose and type of compounds reported similar results for any AEs, SAEs, compound as well as comparators. Major sources of bias and withdrawals due to AE. were inadequate blinding of participants and personnel, Of note, Rutjes et al7 found a higher risk of local reac- inadequate allocation concealment, and possible small tions, SAEs, and withdrawals with HA versus sham or no study bias. interventions. In this SR, the pooled RR of SAEs from 14 RCTs was 1.41 (1.02 to 1.97), consistent when pooling Rheumatoid arthritis only large-­blinded RCTs (RR=1.55 (1.07 to 2.24)). Said The main results of IAT in RA are also shown in table 4. SAEs consisted of 27 events in visco supplementation Outcomes varied widely and included pain, ROM, global patients versus 21 in control patients. Most frequent inflammation, number of flares, and grip strength. HA, disorders were related to the gastrointestinal system (2 IAGC, and yttrium synovectomy were the interventions vs 8), cardiovascular system (5 vs 2), cancer (6 vs 0), and assessed. Saito and Kotake47 observed better perfor- musculoskeletal system (4 vs 2). The authors underlined mance of HA over placebo for pain, global inflammation, that the poor quality of reporting safety data of the RCTs and self-­reported effectiveness. Brazilian Medical Associa- analysed made the understanding of the probable causes tion34 found no differences in the number of flares, ROM, for these observations difficult. 10 Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 Treatments Table 3 Main efficacy outcomes for hip osteoarthritis Study Comparison Follow-­up Outcomes Effect estimate Comments Hyaluronic acid Leite et al40 HA vs PBO, PRP, 1–12 months Pain No between-­group difference vs   MPA PRP (VAS) 3 months No between-­group difference vs PBO (VAS) 1–12 months O-­O Resp No between-­group difference vs MPA 3 months No between-­group difference vs PBO Liao et al37 HA vs PBO or 2 weeks Pain SMD= –0.18 (–0.47 to 0.10), no Data on pain was IAGC between-­group difference obtained as per a 4 weeks SMD= –0.14 (–0.46 to 0.18), no previously described between-­group difference hierarchy.55 Analyses use IAGC and 2–6 months SMD= –0.14 (–0.46 to 0.18), no PBO as comparators. between-­group difference 2 weeks Function SMD=−0.14 (−0.52 to 0.24), no between-­group difference 4 weeks SMD=−0.16 (−0.34 to 0.03), no between-­group difference 2–6 months SMD=−0.28 (−0.60 to 0.05), no between-­group difference Glucocorticoids McCabe et al38 IAGC vs PBO 1–3 months Pain SMD=−1.90 (−4.07 to 0.26), no Comparisons vs PBO between-­group difference 2 months O-­O Resp OR=7.8 (2.7–22.8), favouring IAGC Function 3 RCTs—favoured IAGC (ADL, WOMAC function) 1 RCT—no between-­group difference ROM 1 RCT—favoured IAGC 1 RCT—no between-­group difference Platelet-­rich plasma Medina-­ PRP vs HA 1 month Pain MD=−0.58 (−1.82 to 0.65) (VAS), All comparisons vs HA Porqueres et al41 no difference 6 months MD=0.20 (−1.36 to 1.77) (VAS), no difference 12 months MD=−0.42 (−1.80 to 0.96) (VAS), no difference 2–12 months Function 3 RCTs—no between-­group difference (HHS) 1 RCT—favoured HA (WOMAC) 1 RCT—no between-­group difference (WOMAC) Stiffness 1 RCT—favoured HA (WOMAC) 1 RCT—no between-­group difference (WOMAC) Continued Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 11 RMD Open Table 3 Continued Study Comparison Follow-­up Outcomes Effect estimate Comments Ye et al PRP vs HA 2 months Pain WMD=−0.38 (−0.61 to −0.14), All comparisons vs HA 42 favouring PRP (vs HA) 6 months WMD=−0.14 (−0.40 to 0.12), no between-­group difference 12 months WMD=−0.0 (−0.34 to 0.12), no between-­group difference 2 months Function WMD=2.07 (−2.66 to 6.79) (HHS), no difference 6 months WMD=2.78 (−6.64 to 12.20) (HHS), no difference 12 months WMD=0.71 (−6.33 to 7.75) (HHS), no difference 6 months WMD=−2.84 (−6.25 to 0.57) (WOMAC), no difference 12 months WMD=−3.13 (−6.62 to 0.36) (WOMAC), no difference Ali et al36 PRP vs HA 2–12 months Pain 1 RCT—favoured PRP (VAS) All comparisons vs HA 2 RCTs—no between-­group difference (VAS) Function 1 RCT—no between-­group difference (HHS) 1 RCT—favoured PRP (WOMAC) 1 RCT—no between-­group difference (WOMAC) All effect sizes are presented as the point estimate (95% CI) unless otherwise stated. ADL, activities of daily life; HA, hyaluronic acid; HHS, Harris Hip Score; IAGC, intra-­articular glucocorticoids; MD, mean difference; MPA, methylprednisolone acetate; O-­O Resp, OMERACT-­OARSI Responder Index; PBO, placebo; PRP, platelet-­ rich plasma; RCT, randomised-­controlled trials; SMD, standardised mean difference; VAS, Visual Analogue Scale; WMD, weighted mean difference; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index. Results on withdrawal due to AEs were obtained that found significantly more pain after injection in the after pooling 23 RCTs, but the effect disappeared when PRP group. Finally, results for MSC on KOA were in line restricting the analysis to large-­blinded RCTs. One SR with the previously described. reported significant differences between HA and IA GC, favouring the latter for joint pain after injection (17% vs 3.2%).30 DISCUSSION Safety results for HA in HOA were also reassuring, To our knowledge, this is the first overview of published with no between-­group differences observed for any of SR summarising the efficacy and safety of the most the outcomes of interest, except for an episode of septic frequently used IA treatments. Based on the available arthritis, reported in an RCT (vs placebo) included in literature, we assessed the performance of five treatment the SR by Liao et al.37 Other SRs of HA for shoulder groups in four arthropathies. Most studies evaluated the capsulitis and RA also did not report differences between effects of IAT on KOA and HOA. The average quality groups.44 47 of the SRs was moderate, and high heterogeneity was a IA GC behaved similarly to placebo without any differ- constant, prompting authors to be conservative when ences in safety outcomes in all SRs included in this over- concluding. Most compounds evaluated presented a view for KOA, HOA, shoulder capsulitis, or RA. Of note, small effect for relieving pain and improving function, Juni et al23 also did not find differences between IAGC but with inconsistent results and a high risk of bias in versus sham or no intervention, on any AEs, SAEs, or with- most cases. Regarding safety, the frequency of AEs was drawals due to AEs. Also, this trend remained consistent low, and only a few SAEs were reported, without clear when comparing different IA GC compounds and doses. differences between the different injectables assessed. In the same line, SRs on PRP for KOA and HOA showed HA compounds showed a modest effect on pain and similar safety profiles than its comparators (mostly HA), function in KOA and RA and no effect on HOA or shoulder except for an RCT in the SR by Medina-­Porqueres et al41 capsulitis. Of note, the effects seen for the former, despite 12 Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 Treatments Table 4 Main efficacy outcomes for shoulder capsulitis and rheumatoid arthritis Study Comparison Follow-­up Outcomes Effect estimate Comments Shoulder capsulitis Lee et al HA vs PBO 3–6 months Pain 1 RCT—no between-­group difference   44 (VAS) Function 1 RCT—no between-­group difference (Constant score) 1 RCT—no between-­group difference (ROM) Buchbinder et IAGC vs PBO 4 weeks Pain 1 RCT—no between-­group difference   al43 (VAS) (vs PBO) 6 weeks 1 RCT—no between-­group difference (VAS) (TA 40 mg vs 10 mg) 4 weeks Function 1 RCT—no between-­group difference (ROM) 6 weeks 1 RCT—favour higher dose (ROM) (TA 40 mg vs 10 mg) 4 weeks Success 1 RCT—no between-­group difference frequency Sun et al45 IAGC vs PBO 4–6 weeks Pain MD=1.28 cm (0.75 to 1.82) (VAS), Comparisons with favouring IAGC sham or no injection 12–16 weeks MD=1.00 cm (0.47 to 1.52) (VAS), Passive external favouring IAGC rotation and abduction were 24–26 weeks MD=0.65 cm (0.19 to 1.10), favouring significantly IAGC improved in IAGC-­ 4–6 weeks Composite MD=16.62 (11.16 to 22.09), favouring treated patients (vs scores IAGC (SPADI) PBO) at all 3 time-­ points 12–16 weeks MD=13.46 (8.15 to 18.77), favouring IAGC (SPADI) 24–26 weeks MD=9.91 (2.32 to 17.50), favouring IAGC (SPADI) 4–6 weeks MD=5.30 (–4.38 to 14.98), no difference (ASES) 12–16 weeks MD=12.20 (2.55 to 21.85), favouring IAGC (ASES) 24–26 weeks MD=7.30 (–2.02 to 16.62), no difference (ASES) 12–16 weeks MD=5.70 (–0.59 to 11.99), no difference (Constant score) 4–6 weeks Function MD=20.26° (9.70 to 30.83) favouring IAGC (ROM—Int Rotation) 12–16 weeks MD=0.81° (0.18 to 1.44) favouring IAGC (ROM—Int Rotation) 24–26 weeks MD=3.88° (0.51 to 7.25) favouring IAGC (ROM—Int Rotation) Rheumatoid arthritis Saito and HA vs PBO 1 week Pain RR=1.64 (1.14 to 2.35), favouring HA Outcomes were Kotake47 measured with Global RR=1.61 (1.34 to 1.92), favouring HA a Likert scale Inflammation ranging from ‘no Overall RR=1.50 (1.14 to 1.97), favouring HA improvement’ effectiveness to ‘marked improvement’ Continued Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 13 RMD Open Table 4 Continued Study Comparison Follow-­up Outcomes Effect estimate Comments Silvinato and MPA vs TH, TA 4–24 weeks Pain 1 RCT—MPA vs TA. No between-­ #Results of the Bernardo or prednisolone group difference (VAS) same study at 2 34 time-­points 1 week 1 RCT—MPA vs TH vs prednisolone. Favour TH (VAS)# 2–6 weeks 1 RCT—MPA vs TH vs prednisolone. No difference (VAS)# 4–24 weeks N° of flares 1 RCT—MPA vs TA. No between-­ group difference ROM 1 RCT—MPA vs TA. No between-­ group difference 1–6 weeks Morning stiffness 1 RCT—MPA vs TH vs prednisolone. No difference Grip strength 1 RCT—MPA vs TH vs prednisolone. No difference Ritchie articular 1 RCT—MPA vs TH vs prednisolone. index No difference Thermography 1 RCT—MPA vs TH vs prednisolone. index No difference Heuft-­ Yttrium 6–12 months Pain 2 RCTs—no between-­group difference No differences in Dorenbosch synovectomy vs any other outcome 6 months ROM 1 RCT—favouring yttrium et al46 PBO or TH (subjective change, synovectomy (vs PBO) knee effusion, etc) 12 months 1 RCT—favouring TA (vs yttrium synovectomy) 12 months Knee 1 RCT—favouring yttrium (vs PBO) circumference All effect sizes are presented as the point estimate (95% CI) unless otherwise stated. ASES, American Shoulder and Elbow Surgeons score; HA, hyaluronic acid; IAGC, intra-­articular glucocorticoids; MD, mean difference; MPA, methylprednisolone acetate; PBO, placebo; RCT, randomised controlled trials; ROM, range of motion; Int Rotation, internal rotation; RR, relative risk; SPADI, Shoulder Pain and Disability Index; TA, triamcinolone acetonide; TH, triamcinolone hexacetonide; VAS, Visual Analogue Scale. remaining, were reduced when pooling only large studies IAGC compounds in KOA and found no differences in with low risk of bias or longer follow-­up.5 7 22 30 31 35 HA the outcomes of interest, except for a longer effect of showed a better OMERACT-­ OARSI response in KOA MPA compared with TH.34 versus placebo and IA GC.5 6 30 35 Only one SR assessed the Although IA GC have been among the most widely effects of different HA compounds in KOA and observed used tools for managing inflammatory arthritis for years, differences in favour of those with higher MW on the our search strategy did not retrieve any SR including WOMAC, but authors acknowledge there were too few RCTs comparing them against PBO. Only one study eval- studies to conclude about the superiority of one group uated three different GC compounds in RA and found over another.22 Regarding its effect on RA, it should be no differences between them in all outcomes evaluated noted that the only SR addressing this topic included five except for pain VAS at 1 week of follow-­up in which the RCTs performed in Asian populations and efficacy was analysis favoured TH. measured using scales that are seldom used, and evidence SRs including RCTs on PRP are still limited and our of publication bias, so the results should be interpreted strategy only retrieved articles assessing its performance with caution.47 The body of evidence of IA GC in the target diseases on KOA and HOA. There were only a few RCTs included was smaller compared with that of HA, very likely due and substantial overlapping between SRs. Overall, better to greater industry support for HA. Similarly, its effect performance for pain and function was seen in KOA with versus placebo on pain and function in KOA ranged large effects reported when pooling composite scores from a small, but significant, short-­ term effect to no compared with placebo or HA.25 27 29 32 39 This trend effect. In contrast, IA GC showed a better, although was not present in HOA, with only a few RCTs showing modest, performance on HOA and shoulder capsulitis. modest effects on pain.36 39 41 42 One consistent observa- Likewise, no evidence of an effect on QoL or joint space tion between studies was that the PRP effect lasted longer narrowing was observed. One SR compared different than its comparators (mostly HA). 14 Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 Treatments Table 5 Main safety outcomes of all included compounds Study Comparison Follow-­up Outcomes Effect estimate Comments Knee osteoarthritis Bannuru et al4 HA vs PBO 4–52 weeks Any AEs No between group NMA specifically differences (vs PBO) aimed at analysing Local reactions Analyses favoured safety. PBO for 2/17 products Comparisons are assessed between PBO and all RCTs of individual Withdrawal due to AEs Analyses favoured HA products. PBO for 1/11 products No pooled analysis assessed of HA as a group was carried on. Bannuru et al5 HA vs PBO 2–6 months Any AE HA vs PBO: 16 (54.6) vs No pooled analysis HA vs IAGC 21.7 (56.0) was carried on. IAGC vs PBO HA vs IAGC: 0.0 (64.6) vs Results are median 5.5 (57.2) (IQR) of event rates, % IAGC vs PBO: No data SAEs HA vs PBO: 0 (0.9) vs 0 (0) HA vs IAGC: 0.0 (2.0) vs 0.0 (4.3) IAGC vs PBO: No data Withdrawal due to AEs HA vs PBO: 0.9 (3.9) vs 1.0 (2.6) HA vs IAGC: 1.9 (3.7) vs 2.7 (6.0) IAGC vs PBO: 0.0 (3.5) vs 0.0 (1.7) Local reactions HA vs PBO: 8.4 (14.4) vs 4.7 (16.1) HA vs IAGC: 2.2 (21.8) vs 3.0 (9.1) IAGC vs PBO: 3.3 (17.9) vs 6.9 (8.0) Septic joint HA vs PBO: 0 (0) vs 0 (0) HA vs IAGC: 0 (0) vs 0 (0) IAGC vs PBO: 0 (0) vs 0 (0) Newberry et al22 HA vs PBO 1–12 months Local reactions OR 0.70 (0.48 to 1.03).   No between-­group difference Joint pain OR 0.83 (0.60 to 1.15). No between-­group difference Serious join reactions OR 0.77 (0.25 to 2.31). No between-­group difference Other AE OR 1.26 (0.94 to 1.68). No between-­group difference Other SAE OR 0.62 (0.23 to 1.57). No between-­group difference Continued Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 15 RMD Open Table 5 Continued Study Comparison Follow-­up Outcomes Effect estimate Comments 30 Trojian et al HA vs PBO 2–6 months Joint pain 1 RCT—HA vs IAGC Some RCTs did —17% vs 3.2%, p<0.05 not report data on IAGC vs PBO 10 RCT—no between-­ withdrawal due to group difference AE HA vs IAGC Any AE 11 RCTs—no between-­ group difference   SAEs 11 RCTs—no between-­ group differences   Withdrawal due to AEs 4 RCTs—no between-­ group differences Rutjes et al7 HA vs sham or no 3 months Local reactions RR=1.34 (1.13 to 1.60) †RR for SAE intervention Any AE RR=1.04 (0.99 to 1.09). resulted from No between-­group pooling 14 RCTs. differences ¥RR for withdrawals resulted from SAEs† Overall, RR=1.41 (1.02 pooling 23 RCTs. to 1.97) The effect was Large blinded RCTs, not maintained RR=1.55 (1.07 to 2.24) when pooling large unblinded RCTs. Withdrawal due to RR=1.33 (1.01 to 1.74) AEs¥ Jüni et al23 IAGC vs sham or 2 weeks to 6 Any AE RR=0.89 (0.64 to 1.23)   no intervention months SAEs RR=0.63 (0.15 to 2.67) Withdrawal due to RR=0.33 (0.05 to 2.07) AEs¥ Brazilian Medical MPA vs TA or TH 4–24 weeks Any AE 1 RCT—o AE reported   Association34 or BP 1 RCT—no data on AE 1 RCT—no between-­ group differences Shen et al29 PRP vs HA or 3–12 months Any AE RR=1.40 (0.80 to 2.45). Comparisons were IAGC or PBO SAE No SAEs were identified mainly with HA Kanchanatawan et PRP vs HA or 6–12 months Any AE RR=0.85 (0.57 to 1.28)   al25 PBO (vs HA) RR=6.30 (0.34 to 117.48) (vs PBO) SAEs No data reported Dai et al27 PRP vs HA or 6–12 months Any AE RR=0.63 (0.20 to 1.98)   PBO (vs HA) RR=2.63 (0.04 to 158.93) (vs PBO) SAEs No data reported 32 Di et al PRP vs HA 1–12 months Any AE 1 RCT—significantly   more pain in PRP group 1 RCT—reported no AEs 1 RCT—o safety data reported 4 RCT—no between-­ group differences SAE 5 RCT—reported no SAEs Ding et al24 MSC vs PBO or 6–12 months Any AE No data reported   HA or IAGC SAE OR=1.95 (0.89 to 4.26) Continued 16 Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 Treatments Table 5 Continued Study Comparison Follow-­up Outcomes Effect estimate Comments Hip osteoarthritis Leite et al40 HA vs PBO or 1–12 months Any AE RR=1.07 (0.78 to 1.48)   MPA (vs PBO) RR=2.24 (0.24 to 20.85) (vs MPA) 3 RCTs—no between-­ group differences. (vs PBO) Liao et al37 HA vs PBO 2 weeks to 6 Any AE 4 RCTs—no between-­   months group differences SAE 1 RCT—one septic arthritis episode on the HA group Withdrawal due to AEs 1 RCT—no between-­ group differences McCabe et al38 IAGC vs PBO 1–3 months Any AE 2 RCTs—none reported   2 RCTs—no between-­ group differences Medina-­Porqueres PRP vs HA 1–12 months Any AE 1 RCT—more pain in   et al41 PRP group (p<0.05) 1 RCT—reported one sup haematoma on PRP group Ye et al42 PRP vs HA 2–12 months Any AE RR=0.95 (0.40 to 2.24) Shoulder capsulitis Lee et al44 HA vs PBO 3–6 months Any AE 2 RCTs—no AE reported   2 RCTs—no data on AE 43 Buchbinder et al TA 40 mg vs TA 4–6 weeks Any AE No between-­group   10 mg differences Sun et al45 IAGC vs PBO 4–26 weeks Any AE 3 RCTs—no between-­   group differences 5 RCTs—no data on AE Rheumatoid arthritis Saito and Kotake47 HA vs PBO 1 week Any AE RR=0.98 (0.94 to 1.02)   Silvinato MPA vs TH or TA 1–6 months Any AE 1 RCT—no AE reported   and Bernardo34 1 RCT—no data on AE   All effect sizes are presented as the point estimate (95% CI) unless otherwise stated. AE, adverse events; HA, hyaluronic acid; IAGC, intra-­articular glucocorticoids; MPA, methylprednisolone acetate; MSC, mesenchymal stem cells; NMA, Network Meta-­analysis; PBO, placebo; PRP, platelet-­rich plasma; RCT, randomised controlledl trials; RR, relative risk; SAE, serious adverse events; TA, triamcinolone acetonide; TH, triamcinolone hexacetonide. MSCs appear to be a potentially promising treatment Although we are aware that safety is best studied in large for OA, but SRs including RCTs are scarce. Our strategy long-­term observational studies, we retrieved informa- only retrieved one SR in KOA that met our inclusion tion regarding AEs from the SRs of RCTs. Of note, many criteria.24 Moderate to large effects were seen for KOOS of them did not report on this aspect.5 6 30 31 33 35–37 39 43 44 and pain, respectively, that lasted until 12 months of The SR specifically aimed at analysing this for individual follow-­up. However, the data in which to draw firm conclu- HA compounds versus different comparators found a sions were scarce. Finally, our thorough search retrieved frequency of any AE remarkably low and no increased one SR that evaluated radioisotopic synovectomy for RA risk or only for local reactions.4 7 in which a modest effect was seen over placebo, whereas Striking differences were seen regarding the number it was outperformed by IA GC for some outcomes, such of published articles for the different compounds studied as ROM.46 with HA the intervention which has been most widely Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 17 RMD Open 10 studied to date. However, this was not translated into University of Leeds, Leeds, UK 11 a better quality of evidence, preventing authors from Facultad de Medicina y Ciencias de La Salud, Universidad de Barcelona, Barcelona, Spain drawing firm conclusions regarding many of the studied 12 Rheumatology, Instituto POAL de Reumatología, Barcelona, Spain outcomes. Most of the trials included in the different SRs, 13 Rheumatology, Yerevan State Medical University after Mkhitar Heratsi, Yerevan, especially the ones of PRP and MSC, were highly hetero- Armenia 14 geneous in terms of the composition of the PRP or the Center for Rheumatology and Spine Diseases, Rigshospitalet, Copenhagen, kind of MSC and the procedures used to deliver them. Denmark 15 Nuclear Medicine, Radiologische Allianz, Hamburg, Germany The overall risk of bias within all SRs in this work was 16 Rheumatology, Fondazione Policlinico Universitario Agostino Gemelli IRCCS, high, mostly because of inadequate blinding, allocation Rome, Italy concealment, selective reporting, or publication bias. 17 Faculty of Medicine and Surgery, Catholic University of the Sacred Heart, Rome, It should be also noted that, even although all Italy 18 compounds studied presented modest effect sizes, many Rheumatology, AP-­HP Saint-­Antoine hospital, Paris, France 19 Sorbonne University - INSERM CRSA, Paris, France authors underlined the fact that a proportion of the 20 Centre for Rheumatic Diseases, King's College London, London, UK effect may be due to the placebo effect that accompanies 21 Rheumatology Department, King's College Hospital, London, UK injections5 23 35; something that should be acknowledged 22 EULAR PARE Patient Research Partner, Nicosia, Cyprus 23 when interpreting their results. Rheumatology, Hospital General Universitario de Alicante, Alicante, Spain 24 InMusc, Madrid, Spain This overview of SR has some strengths, such as the comprehensive summary of the currently available IAT Twitter Sebastián Cruz Rodriguez-­García @sdlcrodriguez, Raul Castellanos-­ including a large number of RCTs. However, it has some Moreira @raul_cast_morei, Elena Nikiphorou @ElenaNikiUK and Loreto Carmona limitations. First, including only SRs of RCTs might have @carmona_loreto precluded the analysis of more recent studies still not Acknowledgements The authors want to acknowledge the kind supervision included in said reviews, as well as a deeper evaluation of the search strategy by Maria Piedad Rosario-­Lozano, from the Agency of Technology Evaluation of Andalusia, Fundación Pública Andaluza Progreso y of some treatments, such as MSC in OA or GC in inflam- Salud.This paper presents independent research supported by the National matory arthropathies. Second, for the most frequent Institute for Health Research (NIHR) Leeds Biomedical Research Centre (BRC). diseases affecting the shoulder, SRs usually analyse both Professor Pandit is a NIHR Senior Investigator. The views expressed in this article are those of the author(s) and not necessarily those of the NIHR, or the Department IA and peri-­ articular procedures together, which fell of Health and Social Care. out of the scope of the present work, thus leading us to Contributors SCR-­G, RC-­M, LC, JU, and EN contributed to the conception and exclude them. Third, most information analysed in this study design. SCR-­G and RC-­M performed study selection and data collection. work concerned some frequently assessed outcomes, SCR-­G, RC-­M, LC, JU, and EN analysed the data. SCR-­G, RC-­M, LC, JU, EN, TWON, MD, MB, HP, IMP, VV, LT, WUK, MADA, FB, EN, IP, and JdlT contributed to the such as pain and function, but only a few studies exam- interpretation of the data. SR-­G and RC-­M wrote the first version of the manuscript ined structural outcomes like joint space narrowing or and LC revised it critically. All authors read and approved the final manuscript. cartilage volume loss, which are currently receiving more Funding This study was supported by a EULAR Task force grant CL109. attention.52 53 Finally, a more thorough search in addi- Competing interests SCR-­G reports grants from The Spanish Rheumatology tional databases would have been desirable; but given Foundation during the conduct of the study, and personal fees from Roche, Sanofi, the large amount of hits retrieved and the fact that we MSD, UCB-­Pharma, Bristol-­Myers-­Squibb and Novartis and non-­financial support from Lilly, Pfizer, Sanofi, MSD, Abbvie, UCB-­Pharma, outside the submitted work. were looking for SRs, the potential selection bias would MD has received personal fees for advisory boards from Grunenthal, Mallinckrodt be kept at a minimum. and Pfizer, and author royalties from UpToDate, and was an investigator in an In summary, the evidence shows that IAT in the most AstraZeneca-­funded, non-­drug study (the ‘Sons of Gout’ study), unrelated to this work. LT has received speakers fee from AbbVie, Janssen, Roche, Novartis, Pfizer, frequent arthropathies is well tolerated, with a very low MSD, BMS and GE. FB reports personal fees from Boehringer, Bone Therapeutics, frequency of AEs, but only marginally efficacious in the Expanscience, Galapagos, Gilead, GSK, Merck Serono, MSD, Nordic, Novartis, Pfizer, Regulaxis, Roche, Sandoz, Sanofi, Servier, UCB, Peptinov, TRB Chemedica, 4P short-­to-­medium-­term when compared with placebo. None- Pharma, outside the submitted work. LC declares that her institute receives grants theless, it should be noted that the limited data found for studies and research courses from Novartis Farmaceutica, SA, Pfizer, S.L.U., regarding the efficacy and safety of IAT in inflammatory Merck Sharp & Dohme España, S.A., Roche Farma, S.A, Sanofi Aventis, AbbVie Spain, S.L.U., and Laboratorios Gebro Pharma, SA. arthropathies prevented us from drawing firm conclusions. Patient consent for publication Not required. Author affiliations Provenance and peer review Not commissioned; externally peer reviewed. 1 Rheumatology, Hospital Universitario de la Princesa, Madrid, Spain Data availability statement Data are available upon reasonable request to the 2 Rheumatology, Hospital Clinic de Barcelona, Barcelona, Spain corresponding author. 3 Rheumatology, Hospital Universitario de Mostoles, Madrid, Spain 4 Open access This is an open access article distributed in accordance with the Medicine, Universidad Rey Juan Carlos, Madrid, Spain 5 Creative Commons Attribution Non Commercial (CC BY-­NC 4.0) license, which Rheumatology, Hospital Universitario Fundacion Jimenez Diaz, Madrid, Spain permits others to distribute, remix, adapt, build upon this work non-­commercially, 6 Rheumatology - Joint and Bone Research Unit, Hospital Universitario Fundación and license their derivative works on different terms, provided the original work is Jiménez Díaz, IIS Fundación Jiménez Díaz, Universidad Autónoma de Madrid, properly cited, appropriate credit is given, any changes made indicated, and the Madrid, Spain use is non-­commercial. See: http://​creativecommons.​org/​licenses/​by-​nc/​4.​0/. 7 Rheumatology, Centre for Epidemiology Versus Arthritis, University of Manchester & NIHR Manchester Biomedical Research Centre, Manchester University NHS ORCID iDs Sebastián Cruz Rodriguez-­García http://​orcid.​org/​0000-​0002-​7773-​151X Foundation Trust, Manchester, UK 8 Raul Castellanos-­Moreira http://​orcid.​org/​0000-​0002-​4104-​4101 Medicine, University of Nottingham, Nottingham, UK Jacqueline Uson http://​orcid.​org/​0000-​0002-​2933-​4878 9 Radiology, Copenhagen University Hospital Bispebjerg- Frederiksberg, Esperanza Naredo http://​orcid.​org/​0000-​0003-​0017-​0096 Copenhagen, Denmark Terence W O'Neill http://​orcid.​org/​0000-​0002-​8896-​4677 18 Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 Treatments Michael Doherty http://​orcid.​org/​0000-​0002-​5763-​8326 of the Knee. Rockville (MD: Agency for Healthcare Research and Mikael Boesen http://​orcid.​org/​0000-​0002-​8774-​6563 Quality (US), 2015. Hemant Pandit http://​orcid.​org/​0000-​0001-​7392-​8561 23 Jüni P, Hari R, Rutjes AWS, et al. Intra-­Articular corticosteroid for Ingrid Möller Parera http://​orcid.​org/​0000-​0002-​9225-​2568 knee osteoarthritis. Cochrane Database Syst Rev 2015:CD005328. Valentina Vardanyan http://​orcid.​org/​0000-​0002-​8582-​7837 24 Ding W, Xu Y-­Q, Zhang Y, et al. Efficacy and safety of intra-­articular Lene Terslev http://​orcid.​org/​0000-​0001-​8193-​9471 cell-­based therapy for osteoarthritis: systematic review and network Will Uwe Kampen http://​orcid.​org/​0000-​0002-​3500-​9713 meta-­analysis. Cartilage 2020:194760352094294. Maria Antonieta D'Agostino http://​orcid.​org/​0000-​0002-​5347-​0060 25 Kanchanatawan W, Arirachakaran A, Chaijenkij K, et al. Short-­ Francis Berenbaum http://​orcid.​org/​0000-​0001-​8252-​7815 Term outcomes of platelet-­rich plasma injection for treatment of Elena Nikiphorou http://​orcid.​org/​0000-​0001-​6847-​3726 osteoarthritis of the knee. Knee Surg Sports Traumatol Arthrosc 2016;24:1665–77. Jenny de la Torre-­Aboki http://​orcid.​org/​0000-​0002-​4905-​2034 26 Xu Z, Luo J, Huang X, et al. Efficacy of platelet-­rich plasma in pain Loreto Carmona http://​orcid.​org/​0000-​0002-​4401-​2551 and self-­report function in knee osteoarthritis. Am J Phys Med Rehabil 2017;96:793–800. 27 Dai W-­L, Zhou A-­G, Zhang H, et al. Efficacy of platelet-­rich plasma in the treatment of knee osteoarthritis: a meta-­analysis of randomized REFERENCES controlled trials. Arthroscopy 2017;33:659–70. 1 Miller JH, White J, Norton TH. The value of intra-­articular injections 28 Arroll B, Goodyear-­Smith F. Corticosteroid injections for in osteoarthritis of the knee. J Bone Joint Surg Br 1958;40-­ osteoarthritis of the knee: meta-­analysis. BMJ 2004;328:869. B:636–43. 29 Shen L, Yuan T, Chen S, et al. The temporal effect of platelet-­rich 2 Hochberg MC, Altman RD, April KT, et al. American College plasma on pain and physical function in the treatment of knee of rheumatology 2012 recommendations for the use of osteoarthritis: systematic review and meta-­analysis of randomized nonpharmacologic and pharmacologic therapies in osteoarthritis of controlled trials. J Orthop Surg Res 2017;12:16. the hand, hip, and knee. Arthritis Care Res 2012;64:465–74. 30 Trojian TH, Concoff AL, Joy SM, et al. AMSSM scientific statement 3 Weiss MM. Corticosteroids in rheumatoid arthritis. Semin Arthritis concerning viscosupplementation injections for knee osteoarthritis: Rheum 1989;19:9–21. importance for individual patient outcomes. Br J Sports Med 4 Bannuru RR, Osani M, Vaysbrot EE, et al. Comparative safety profile 2016;50:84–92. of hyaluronic acid products for knee osteoarthritis: a systematic 31 Gallagher B, Tjoumakaris FP, Harwood MI, et al. Chondroprotection review and network meta-­analysis. Osteoarthritis Cartilage and the prevention of osteoarthritis progression of the knee: 2016;24:2022–41. a systematic review of treatment agents. Am J Sports Med 5 Bannuru RR, Schmid CH, Kent DM, et al. Comparative effectiveness 2015;43:734–44. of pharmacologic interventions for knee osteoarthritis: a systematic 32 Di Y, Han C, Zhao L, et al. Is local platelet-­rich plasma injection review and network meta-­analysis. Ann Intern Med 2015;162:46–54. clinically superior to hyaluronic acid for treatment of knee 6 Bannuru RR, Natov NS, Obadan IE, et al. Therapeutic trajectory osteoarthritis? A systematic review of randomized controlled trials. of hyaluronic acid versus corticosteroids in the treatment of knee Arthritis Res Ther 2018;20:128. osteoarthritis: a systematic review and meta-­analysis. Arthritis 33 Lo GH, LaValley M, McAlindon T, et al. Intra-­articular hyaluronic Rheum 2009;61:1704–11. acid in treatment of knee osteoarthritis: a meta-­analysis. JAMA 7 Rutjes AWS, Jüni P, da Costa BR, et al. Viscosupplementation for 2003;290:3115–21. osteoarthritis of the knee: a systematic review and meta-­analysis. 34 Brazilian Medical Association, Silvinato A, Bernardo WM. Ann Intern Med 2012;157:180–91. Inflammatory arthritis or osteoarthritis of the knee - Efficacy of intra-­ 8 Cerza F, Carnì S, Carcangiu A, et al. Comparison between hyaluronic joint infiltration of methylprednisolone acetate versus triamcinolone acid and platelet-­rich plasma, intra-­articular infiltration in the acetonide or triamcinolone hexacetonide. Rev Assoc Med Bras treatment of gonarthrosis. Am J Sports Med 2012;40:2822–7. 2017;63:827–36. 9 McGonagle D, Baboolal TG, Jones E. Native joint-­resident 35 Trigkilidas D, Anand A. The effectiveness of hyaluronic acid intra-­ mesenchymal stem cells for cartilage repair in osteoarthritis. Nat Rev articular injections in managing osteoarthritic knee pain. Ann R Coll Rheumatol 2017;13:719–30. Surg Engl 2013;95:545–51. 10 Jiang Y, Tuan RS. Origin and function of cartilage stem/progenitor 36 Ali M, Mohamed A, Ahmed HE, et al. The use of ultrasound-­ cells in osteoarthritis. Nat Rev Rheumatol 2015;11:206–12. guided platelet-­rich plasma injections in the treatment of hip 11 Doherty M, Dieppe P. The "placebo" response in osteoarthritis osteoarthritis: a systematic review of the literature. J Ultrason and its implications for clinical practice. Osteoarthritis Cartilage 2018;18:332–7. 2009;17:1255–62. 37 Liao Y-­Y, Lin T, Zhu H-­X, et al. Intra-­Articular Viscosupplementation 12 Zhang W, Robertson J, Jones AC, et al. The placebo effect and for patients with hip osteoarthritis: a meta-­analysis and systematic its determinants in osteoarthritis: meta-­analysis of randomised review. Med Sci Monit 2019;25:6436–45. controlled trials. Ann Rheum Dis 2008;67:1716–23. 38 McCabe PS, Maricar N, Parkes MJ, et al. The efficacy of intra-­ 13 Zhang W, Zou K, Doherty M. Placebos for Knee Osteoarthritis: articular steroids in hip osteoarthritis: a systematic review. Reaffirmation of "Needle Is Better Than Pill". Ann Intern Med Osteoarthritis Cartilage 2016;24:1509–17. 2015;163:392–3. 39 Dong Y, Zhang B, Yang Q, et al. The effects of platelet-­rich 14 Lee R. The outlook for population growth. Science 2011;333:569–73. plasma injection in knee and hip osteoarthritis: a meta-­analysis of 15 Nogrady B. Childhood obesity: a growing concern. randomized controlled trials. Clin Rheumatol 2021;40:263–77. Nature2017;551:S96. 40 Leite VF, Daud Amadera JE, Buehler AM. Viscosupplementation 16 Frank J. Origins of the obesity pandemic can be analysed. Nature for hip osteoarthritis: a systematic review and meta-­analysis of 2016;532:149. 17 Moher D, Liberati A, Tetzlaff J, et al. Preferred reporting items for the efficacy on pain and disability, and the occurrence of adverse systematic reviews and meta-­analyses: the PRISMA statement. events. Arch Phys Med Rehabil 2018;99:574–83. PLoS Med 2009;6:e1000097. 41 Medina-­Porqueres I, Ortega-­Castillo M, Muriel-­Garcia A. 18 Lequesne M. Indices of severity and disease activity for Effectiveness of platelet-­rich plasma in the management of osteoarthritis. Semin Arthritis Rheum 1991;20:48–54. hip osteoarthritis: a systematic review and meta-­analysis. Clin 19 Bellamy N, Buchanan WW, Goldsmith CH, et al. Validation study Rheumatol 2021;40:53–64. of WOMAC: a health status instrument for measuring clinically 42 Ye Y, Zhou X, Mao S, et al. Platelet rich plasma versus hyaluronic important patient relevant outcomes to antirheumatic drug therapy acid in patients with hip osteoarthritis: a meta-­analysis of in patients with osteoarthritis of the hip or knee. J Rheumatol randomized controlled trials. Int J Surg 2018;53:279–87. 1988;15:1833–40. 43 Buchbinder R, Green S, Youd JM. Corticosteroid injections for 20 Shea BJ, Reeves BC, Wells G, et al. AMSTAR 2: a critical appraisal shoulder pain. Cochrane Database Syst Rev 2003:CD004016. tool for systematic reviews that include randomised or non-­ 44 Lee L-­C, Lieu F-­K, Lee H-­L, et al. Effectiveness of hyaluronic acid randomised studies of healthcare interventions, or both. BMJ administration in treating adhesive capsulitis of the shoulder: a 2017;358:j4008. systematic review of randomized controlled trials. Biomed Res Int 21 Cohen J. Statistical power analysis for the behavioral sciences. 2nd 2015;2015:1–8. edn. Hillsdale: Lawrence Erlbaum Associates, 1988. 45 Sun Y, Zhang P, Liu S, et al. Intra-­Articular steroid injection for frozen 22 Newberry SJ, Fitzgerald JD, Maglione MA. AHRQ Technology shoulder: a systematic review and meta-­analysis of randomized Assessments. Systematic Review for Effectiveness of Hyaluronic controlled trials with trial sequential analysis. Am J Sports Med Acid in the Treatment of Severe Degenerative Joint Disease (DJD) 2017;45:2171–9. Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658 19 RMD Open 46 Heuft-­Dorenbosch LL, de Vet HC, van der Linden S. Yttrium 51 Di Sante L, Villani C, Santilli V, et al. Intra-­articular hyaluronic acid radiosynoviorthesis in the treatment of knee arthritis in rheumatoid vs platelet-­rich plasma in the treatment of hip osteoarthritis. Med arthritis: a systematic review. Ann Rheum Dis 2000;59:583–6. Ultrason 2016;18:463–8. 47 Saito S, Kotake S. Is there evidence in support of the use of 52 McAlindon TE, LaValley MP, Harvey WF, et al. Effect of intra-­articular triamcinolone vs saline on knee cartilage volume and pain in intra-­articular hyaluronate in treating rheumatoid arthritis of the patients with knee osteoarthritis: a randomized clinical trial. JAMA knee? A meta-­analysis of the published literature. Mod Rheumatol 2017;317:1967–75. 2009;19:493–501. 53 Roemer FW, Kraines J, Aydemir A, et al. Evaluating the structural 48 Bannuru RR, Natov NS, Obadan IE, et al. Therapeutic trajectory effects of intra-­articular sprifermin on cartilage and non-­cartilaginous of hyaluronic acid versus corticosteroids in the treatment of knee tissue alterations, based on sqMRI assessment over 2 years. osteoarthritis: a systematic review and meta-­analysis. Arthritis Osteoarthritis Cartilage 2020;28:1229–34. Rheum 2009;61:1704–11. 54 Pavelka K, Uebelhart D. Efficacy evaluation of highly purified intra-­ 49 Xu Z, Luo J, Huang X, et al. Efficacy of platelet-­rich plasma in pain articular hyaluronic acid (Sinovial(®)) vs hylan G-­F20 (Synvisc(®)) and self-­report function in knee osteoarthritis: a Best-­Evidence in the treatment of symptomatic knee osteoarthritis. A double-­ blind, controlled, randomized, parallel-­group non-­inferiority study. synthesis. Am J Phys Med Rehabil 2017;96:793–800. Osteoarthritis Cartilage 2011;19:1294–300. 50 Dallari D, Stagni C, Rani N, et al. Ultrasound-­Guided injection 55 Juhl C, Lund H, Roos EM, et al. A hierarchy of patient-­reported of platelet-­rich plasma and hyaluronic acid, separately and in outcomes for meta-­analysis of knee osteoarthritis trials: empirical combination, for hip osteoarthritis: a randomized controlled study. evidence from a survey of high impact journals. Arthritis Am J Sports Med 2016;44:664–71. 2012;2012:1–17. 20 Rodriguez-­García SC, et al. RMD Open 2021;7:e001658. doi:10.1136/rmdopen-2021-001658