Efficacy of intra‐articular injections of hyaluronic acid in patients with glenohumeral joint osteoarthritis: A systematic review and meta‐analysis

Symptomatic primary glenohumeral (GH) joint osteoarthritis (OA) can be challenging to treat. Hyaluronic acid (HA) has emerged as a promising treatment for the nonsurgical management of GH‐OA. In this systematic review with meta‐analysis, we aimed to evaluate the current evidence regarding the efficacy of intra‐articular HA on pain relief in patients suffering from GH‐OA. A total of 15 studies (only randomized controlled trials providing data at the end of the intervention) were included. The relevant studies were selected based on the following PICO model: P: patients with diagnosis of shoulder OA; I: HA infiltrations as therapeutic intervention administered; C: no restriction for comparators assessed; O: pain, in terms of visual analog scale (VAS) or numeric rating scale. The risk of bias among the included studies was estimated using the PEDro scale. A total of 1023 subjects were analyzed. Comparing HA injections combined with physical therapy (PT) compared to PT alone resulted in superior scores, showing an overall effect size (ES) of 4.43 (p = 0.00006). Moreover, pooled analysis of VAS pain scores demonstrated a significant improvement in the ES of the HA in comparison with corticosteroid injections (p = 0.002). On average, we reported a PEDro score of 7.2. A total of 46.7% of studies showed probable signs of a randomization bias. The findings of this systematic review and meta‐analysis showed that IA injections of HA might be effective on pain relief with significant improvements compared to baseline and compared to corticosteroid injections in patients affected by GH‐OA.

insufficient and can be associated with well-known adverse effects, especially in elderly patients.NSAIDs have the potential to cause gastrointestinal, renal, and cardiac effects. 7,8us, hyaluronic acid (HA) has emerged as an alternative treatment for the nonsurgical management of GH-OA.HA has both analgesic and chondroprotective properties. 5The use of IA HA in patients with OA is well documented. 9HA therapies can be broadly classified as low-molecular-weight (LMW) preparations (500-730 kDa) 10 and high-molecular-weight (HMW) preparations (620-3200 kDa), whereas natural human HA is a single-chain product with a molecular weight of 5000 kDa. 11Several papers have recently investigated the efficacy of different IA HA preparations for OA. 9,12,13Concerning shoulder joint, Strauss et al. reported that HA injections are well tolerated to treat shoulder pain of various pathologies and may be an alternative to PT and CS injections. 12In 2014, Colen et al. published a systematic review of 8 studies on the effect of IA HA injections for GH-OA. 13Zhang et al. performed a systematic review and meta-analysis to evaluate the efficacy of HA to reduce pain in patients with GH-OA. 9The authors reported (1)   that intra-articular HA injection was safe and improved pain for patients with GH-OA and (2) that a significant placebo effect may have been present.
In this systematic review with meta-analysis, we aimed to evaluate the current evidence regarding the efficacy of IA HA on pain relief in patients suffering from GH-OA.

| Search strategy and selection criteria
This systematic review has been conducted according to Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines 14 and the Cochrane Handbook for Systematic Reviews of Interventions. 15The protocol of this systematic review has been registered on the International Prospective Register of Systematic Reviews (PROSPERO) with number CRD42022385161.
On January 25, 2023, two authors (MM and NM) systematically searched three different databases (PubMed, Scopus, and Web of Science).The search strategy is reported in Table 1.
After removing the duplicates, two reviewers (RR and MM) independently screened all the documents for title and abstract and then, for full-text.Then, a third author (AdS) was asked to solve any disagreement by collegial discussion.
T A B L E 1 Keyword search strategy for each Database.
Only randomized controlled trials (RCTs) providing data at the end of the intervention were included.Exclusion criteria were: (1)   patients suffering from any inflammatory disorders or rheumatic diseases (e.g., rheumatoid arthritis, psoriatic arthritis); (2) patients with fibromyalgia; (3) studies including arthrocentesis as treatment; (4) studies including local pressure pain assessment; (5) studies with a cross-over design; (6) studies written in a language different from English; (7) full-text unavailability (i.e., posters and conference abstracts); and (8) studies involving animals.All data were extracted and synthesized in a table with a qualitative synthesis performed by two authors independently from full-text documents.

| Quality assessment
We adopted the risk-of-bias checklist in the Physiotherapy Evidence Database (PEDro) scale to estimate the included studies' methodological quality. 16Two authors (RR and NM) separately evaluated each article and presented the results, and any disagreements were resolved involving a third author (AdS) The PEDRO tool consists of nine domains through which it is possible to find any bias in a study.
Each judgment consists of the following possibilities: low risk of bias, moderate risk of bias/some concerns, serious risk of bias, critical risk of bias, and no information on a 10-point scale.Domain-level reports provide the basis for an overall risk-of-bias judgment.

| Statistical analysis
A systematic summary of patient characteristics and results of the included studies were reported in an Excel spreadsheet.Summaries of intervention effects for each study were provided.The results of the included studies were reported in a qualitative manner (e.g., statistically significant results, consistency of results, or a combination thereof).Statistical analysis was performed on R 3.5.0(R Foundation) and RevMan version 5.3.A p-value < 0.05 was considered as statistical significance.The heterogeneity between the comparisons was estimated by means of the chi-squared and I² statistical tests.An I² > 50% results in significant heterogeneity across articles, legitimizing an effect size (ES) measure via a random effects model was used to determine pooled estimates with 95% CIs.

| Main characteristics of the included studies
A total of 1819 records were identified from the search process.
After the title and abstract screening step, 1752 of them were excluded.Next, out of the 67 full-text studies screened, 15 articles that satisfied the eligibility criteria were included.Further details on the identification and inclusion/exclusion of the screened studies are reported in the PRISMA 2020 flow diagram (see Figure 1 for further details).
A total of 1023 subjects were analyzed, whereas 397 subjects were included in the comparator group (undergoing no intervention, PT, corticosteroid injection, and PRP injection).The size of the study cohorts included ranged from 27 28 to 300 30 patients.Concerning the follow-up evaluations, only one study performed a follow-up at 52 weeks from baseline. 26(p < 0.001), and SST (p < 0.001) at the 6-month follow-up following a protocol of three injections of Hylan G-F 20, with an interval of 1 week between injections. 28il et al. reported the outcomes following three injections of high molecular weight hyaluronate (2.5 mL each) Euflexxa ® (1% sodium hyaluronate), at an interval of 1 week, showing improvements in pain (VAS, WOMAC), ROM, stiffness, and clinical outcome scores. 30

| HA injections combined with PT versus PT alone
In 2015, in their study, Di Giacomo et al. compared the application of five intra-articular injections with Hyalgan 20 mg/2 mL (molecular weight 500-730 kDa) at an interval of 15 days between injection combined with a specific physiotherapy (PT) program to a treatment with PT alone. 19The adjunct of HA injection showed a greater and long-lasting efficacy in terms of reduction of shoulder pain (p < 0.05) and improvement in daily activities.In 2017, Di Giacomo et al. reported a comparison of an application of a three-injection program with Hyalubrix (30 mg/2 mL, MW > 1500 kDa), with one injection every 15 days, combined with a specific physiotherapy program, to a control group who received PT only. 18ey reported greater positive effect in terms of pain reduction compared with patients who received PT only.This approach was demonstrated to be a safe and effective treatment option for the management of shoulder pain due to moderate to severe glenohumeral OA in terms of pain relief (p < 0.05) and function improvement.
In 2021, Di Giacomo et al. reported the outcomes following a single injection with HMW HA (HyalOne ® 60 mg/4 mL 1.500-2.000.000Da) in combination with PT in comparison to a PT control group. 17They reported a significantly higher decrease of shoulder pain and improvement in daily activities compared to patients treated with PT alone (p < 0.05).effective pain relief and an improving functioning in shoulder OA. 21

Tortato et al. reported the administration of a single injection of
Hylan G-F 20 (48 mg/6 mL; Synvisc One ® ) compared to a single injection of triamcinolone hexacetonide (Triancil ® , 20 mg/1 mL diluted in 5 mL saline). 29A VAS reduction was observed in both groups, especially in the cases of mild and moderate arthritis that received HA, but with a mean value from 8.1 initially to 4.9 after 6 months in HA compared to the control group (p < 0.05).

| HA versus placebo control groups
Kwon et al. reported the results following 3 weekly injections of HA in their experimental group in comparison to 3 weekly injections of phosphate-buffered saline (PBS) in a placebo control group. 27No statistically significant differences in efficacy were found between HA and PBS groups in patients with glenohumeral OA. guidance at an interval of 1 week, in comparison to a control group that did not receive a treatment. 24The HA intervention group reported a significant decrease in Constant score (62 ± 3.0 vs. 34 ± 6.5, p < 0.05), and an increase in VAS (3.3 ± 1.4 vs. 7.8 ± 3.1, p < 0.05).receiving an intervention showed a significant difference, HA injections alone compared to PBS injections, do not appear to provide an advantage; possibly due to the hydro-distensive nature of the approach.Given the low number of studies, a random-effects model was adopted.

| Quality assessment
The risk of bias among the included studies was estimated using the PEDro scale (see Table 3 for further details).All studies considered scored above 5 out of 10, on average we reported a score of 7.2.A total of 46.7% of studies showed probable signs of a randomization bias.In all the studies included the risk of missing outcome bias could be ruled out.

| DISCUSSION
The main finding of this investigation was that the intra-articular application of HA for GH-OA resulted in a significant improvement of pain and function compared to baseline at a short-term follow-up of up to 6 months.Moreover, based on the results of the meta-analysis, HA seems to be superior to alternative nonoperative treatment modalities such as corticoid injections and isolated PT for the treatment of glenohumeral OA.However, it should be noted that there could be a potential placebo effect related to the application of HA, as a superiority of intra-articular HA injection compared to placebo injections could not be demonstrated.
Although the precise cellular working mechanism of HA is not yet fully understood, it has been associated with analgesic and chondroprotective properties that are essential in comprehending | 2353 its clinical effects in a non-weight-bearing joint such as the shoulder. 312][33][34] As the concentration of native HA is reduced to approximately 50% of its physiologic concentration in OA, 35 an intraarticular injection of HA exhibits instant dual mechanical effects.As such, it functions as a lubricant during slow, low shear rate movements by increasing synovial viscosity and also provides shock absorption during rapid, high shear rate movements in its structure as an elastic solid. 31Furthermore, and potentially more relevant in a non-weightbearing joint such as the shoulder, HA injections have been linked to chondroprotective capacities.In detail, HA injections have been demonstrated to reduce chondrocyte apoptosis and increased chondrocyte proliferation. 33,34Through an immunomodulatory mechanism of action via suppression of IL-1β expression, HA has been further demonstrated to exhibit anti-inflammatory effects and thereby exhibits a beneficial impact on the osteoarthritic milieu of the affected joint. 32,36ken together, the results of this systematic review underscored the promising role of HA injections as a valid nonoperative treatment option for glenohumeral OA.
In this scenario, a previous systematic review by Colen et al.   investigated the early data regarding the effectiveness of intraarticular HA injection in joints other than the knee joint. 37The six included studies specific to the shoulder provided early statistical evidence for improvement in pain and function, but were of low level of evidence and impeded a strong conclusion at that time. 37Later systematic evaluations such as the systematic review by Zhang et al.
confirmed these initial findings. 9In their meta-analysis, they demonstrated a significant and substantial improvement in pain at 3 and 6 months after the injection as well as significant improvements in functional outcome scores. 9The meta-analysis conducted in the current study, which incorporated five additional studies, confirms these results, emphasizing the potential of intra-articular HA injections as a promising nonoperative treatment option.Given that, an improvement of 1.4 points in the VAS pain score has been defined as the minimal clinically important difference (MCID) in shoulder OA, 38 the improvement following intra-articular glenohumeral HA injection seems to be both statistically significant and clinically relevant.These findings have been confirmed by the pooled results of previous quantitative syntheses that demonstrated a mean improvement of the VAS pain score of 2.6 points at 3 months and 2.9 points at 6 months. 9Not only pain, but also the improvement in clinical function seems to be clinically relevant at short-term followup.As such, the mean improvement following glenohumeral HA injection at a cohort level exceeds the MCID values for the Constant Score (5.7 points) 39 and UCLA score (8.7 points) 39 in multiple studies. 18,21,24,28wever, to date, the evidence available in the literature is still insufficient to quantitatively investigate the optimal HA injection regimen in terms of the total number of injections as well as the F I G U R E 2 Forest plot comparing HA injections combined with PT compared to PT alone.CI, confidence interval; HA, hyaluronic acid; PT, physical therapy.
T A B L E 3 Quality assessment of the included studies according to the PEDro scale.| 2355 optimal time interval between injections.While a weekly administration is the most common injection interval, the total number of injections ranges from 1 17,20,22,23,25,26,29 to 5. 19 A total number of injections as low as 1 demonstrated significant improvement of outcome parameters compared to baseline as well as in comparison to alternative interventions such as PT 17,20,22,23,25,26,29 or corticosteroid injections. 29While there have been efforts to determine the optimal number of HA injections in knee OA, the data published are conflicting and do not clearly favor a certain range. 40,41Also, currently there is insufficient evidence specific to shoulder OA analyze the optimal formulation of HA and whether LMW or HMW preparations have a superior clinical effect.
Notably, intra-articular HA injections appear to outperform typical nonoperative treatment options that are presently regarded as the gold standard of first-line therapy for glenohumeral OA, such as PT or corticosteroid injections.The quantitative synthesis in this systematic review, which included multiple reports comparing isolated PT to PT augmented with HA injections, [17][18][19] revealed significantly superior clinical outcomes for the combined approach with an ES of 4.4.This finding expands the evidence generated by previous pooled analyses. 9,37Furthermore, in comparison to other agents such as corticosteroid formulations, 21,29 the meta-analysis conducted in this study revealed significantly greater improvements in pain following HA injections, with an ES of −1.4 on the VAS pain within the first 6 months.While both HA as well as corticosteroids exhibit analgesic and anti-inflammatory effects, these results after 6 months may reflect the sustained chondroprotective effect attributed to HA. 33,34 The effectiveness of the HA injection treatments observed must be interpreted in the context of a potential placebo effect, which is a known phenomenon related to HA injections. 42For example Kwon et al. reported in their study that the HA intervention group did not outperform a placebo group, which suggests a strong potential for a placebo effect involved in the treatment of glenohumeral OA with HA injection. 27Alternatively, the similar effect of PBS in the placebo control group compared to HA may be attributable to the effect of the hydrodistension. 43As a result-comparable to previous reviews 9 -the quantitative synthesis in this review could not demonstrate the superiority of HA compared to placebo injections.
To adequately interpret the findings of this systematic review, the results specific to HA must be benchmarked to alternative agents with differing biological mechanisms of action such as PRP or bone marrow aspirate concentrate (BMAC).While the evidence specific to the shoulder available is limited, PRP-with its demonstrated ability to reduce joint inflammation, decrease cartilage breakdown, promote tissue repair, and facilitate healing processes 44 -has been suggested as an alternative, clinically effective option for symptom alleviation in glenohumeral OA. 44,45 When comparing the efficacy of intra-articular HA to PRP injections for the management of glenohumeral OA, the study by Kirschner et al. included in this review did not show any significant differences between these two treatments. 26However, in knee joint OA, PRP was found to have a higher probability for efficacy in both a recent network meta-analysis as well as a systematic review of level 1 studies. 46As both substances may be subject to similar reimbursement categories and thus present as alternatives in nonoperative management, further evidence is needed to provide recommendations in the treatment of glenohumeral OA.
Furthermore, the application of BMAC, which contains mesenchymal stromal cells, progenitor cells, growth factors, and other biologically potent agents, has been associated with analgesic, anti-inflammatory, and anabolic effects. 47While preliminary evidence attests to superior clinical efficacy in the treatment of glenohumeral OA compared to cortisone application, a substantial increase in evidence is needed to evaluate this potential avenue. 48 summary, while in knee joint OA, there is substantial evidence that HA injections provide beneficial effects on symptom improvement, 49 the preliminary evidence for HA in shoulder OA is promising, but not yet conclusive.
It should be noted that the present is the first systematic review with meta-analysis including RCTs on this topic, evaluating the efficacy of intra-articular injections of HA to reduce pain in patients with GH-OA.
However, we are aware that our manuscript is not free from limitations.First, considering the limited number of studies reported, the authors were unable to perform any meaningful statistical analyses related to the optimal formulation of HA, the number of injections, and the injection interval.Second, confounding variables such as the products of HA utilized, the total number of injections, and the technique of injection (image-guided vs. blind) may influence the conclusions.The scarcity of available data in the current literature precluded direct group comparisons between those subgroups.Third, the follow-up times among studies included in this review varied and were largely limited to a short-term follow-up of 6 months, thus making it challenging to evaluate potential long-term benefits.
However, it should be noted that most of the treatment protocols using HA for OA consist of multiple injections that need to be repeated over time (e.g., after 6-12 months).Fourth, this review is limited to studies with a low level of evidence with their inherent limitations, precluding the formulation of strong conclusions or recommendations.Finally, it is possible that relevant subgroups of patients or related studies were inadvertently excluded from our investigation due to the nature of systematic reviews and the search criteria and strategy employed.

| CONCLUSIONS
Taken together, the findings of this systematic review and metaanalysis showed that IA injections of HA might be effective on pain relief with significant improvements compared to baseline and compared to corticosteroid injections in patients affected by GH-OA.However, to date, there is still the need for further RCTs, considering the lack of clear data on the optimal formulation, the number of injections, and the injection interval of HA for GH-OA patients.On the other hand, we are aware that this meta-analysis might be useful for improving the knowledge of his topic and for Two different authors (RR and MM) evaluated the records resulting from selection process.All relevant data were subsequently extracted independently.Then, a third author (FF) was asked to solve any disagreement by collegial discussion.The relevant data extracted were: (I) title, authors, and publication year; (II) nationality; (III) population characteristics; (IV) interventions' characteristics; (V) control characteristics; (VI) outcome measures; (VII) main findings; (VIII) follow-up evaluations; and (IX) assessment of secondary outcomes.
Merolla et al. reported the outcome following a three-injection program with Hylan G-F 20 with an interval of 1 week between injections, in comparison with a control group of three injections of methylprednisolone acetate (Depo-Medrol ® , Pfizer) 40 mg/mL with an interval of 1 week between injections.The HA group demonstrated an .2 | HA injections without comparison/control groups