The efficacy of multiple versus single hyaluronic acid injections: a systematic review and meta-analysis

Knee osteoarthritis (OA) is a slowly progressive joint disorder characterized by joint pain, cartilage degeneration, and inflammation that affects approximately 250 million people worldwide [1]. Knee OA leads to negative impacts on socioeconomic factors including impaired work performance and early retirement [2, 3].

Multiple treatment options are available for knee OA, ranging from conservative management to total knee arthroplasty. Common non-operative therapies include non-steroidal anti-inflammatory drugs (NSAIDs), physical therapy, analgesics such as acetaminophen, and intra-articular (IA) therapies such as corticosteroids and hyaluronic acid (HA) [4]. IA-HA products have been used in the United States as a treatment for knee OA since 1997 and are approved by the Food and Drug Administration (FDA) [5]. The putative mechanisms of action through which IA-HA provides therapeutic effects include anti-inflammatory effects, chondroprotection, proteoglycan synthesis, and shock absorption properties [6]. IA-HA is frequently used in clinical practice to treat knee pain and suppress knee inflammation.

In 2013, the American Academy of Orthopaedic Surgeons (AAOS) recommended against the use of IA-HA for knee OA [7]. However, more recently, certain physician-specialty societies have recommended the use of IA-HA in treating knee OA. The American College of Rheumatology (ACR) recommended IA-HA for knee OA patients who had an inadequate response to initial therapy in a recent position statement [8]. The American Medical Society for Sports Medicine (AMSSM) recommended the use of HA for the appropriate patients with knee OA based on recent evidence in a network meta-analysis [9]. The European Society for Clinical and Economic Aspects of Osteoarthritis (ESCEO) task force issued a consensus statement recommending the use of IA-HA in knee OA patients with mild to moderate disease, and for more severe patients who are not good candidates for total knee replacement surgery or wishing to delay the surgical procedure [10]. The OA Research Society International (OARSI) has an uncertain recommendation for the use of IA-HA in treating knee OA based on the good level of quality of evidence available, suggesting physicians discuss the risk-benefit profile of IA-HA along with individual characteristics, comorbidities and preferences of the patient [11]. Despite the clinical considerations and availability of evidence recommending the use of IA-HA in treating knee OA, the optimal treatment regimen and patient selection criteria have yet to be determined.

IA-HA is commonly prescribed in different injection regimens, which vary from a single injection to one injection a week for 5 weeks. Multiple randomized controlled trials (RCTs) have been conducted evaluating different dosing regimens of IA-HA versus IA-Saline. We conducted a systematic review of the published literature to determine the efficacy of IA-HA vs IA-Saline in patients with knee OA, with subgroup analyses to explore the differences in levels of pain, adverse events (AEs), and serious adverse events (SAEs) across different dosing regimens.

Overall, treatment with IA-HA was observed to be more effective in treating patients with OA knee pain compared to IA-Saline, with 2–4 injections demonstrating the largest treatment effect at both 3-month and 6-month follow-ups. Single injections of IA-HA were not significantly more effective than saline at 3-month or 6-month follow ups; however there were only four published RCTs comparing single injections with one trial not reporting the effect size. The ≥5 injections subgroup demonstrated significant improvement in pain at 6 months only. These results indicate that only the 2–4 injection regimen group provided consistently significant pain relief at both 3 and 6 months follow-up, when compared to IA-Saline.

Our results showed greatest improvement at 3 months, while improvement was seen to a lesser extent at 6 months following IA-HA. This is similar to the therapeutic trajectory of IA-HA vs IA-Saline conducted by Bannuru and colleagues [47]. Effect sizes favored IA-HA by week 4 ((0.31; 95% CI 0.17, 0.45) and peak at week 8 (0.46; 0.28, 0.65), suggesting that the optimal improvement is seen around 2 months after IA-HA injection. A network meta-analysis conducted by Bannuru et al [4] found similar results when comparing IA-HA to saline controls; IA-HA was favoured over IA-Saline control (SMD = 0.429; C.I. 0.261 to 0.598, p = 0.000) on knee pain at 4 to 13 weeks of treatment. Another network meta-analysis demonstrated similar conclusions, as IA-HA was seen to have a significant effect on pain when compared to IA-Saline [48]. Our analysis has demonstrated that not only does the overall pooled estimate for HA as a class demonstrate a benefit for pain relief, there seem to be differences in effect as a result of the injection regimen provided. Previous analyses have suggested that in addition to HA class benefits in pain relief, molecular weight may also be a product characteristic that affects the potential outcomes of patients [13, 15].

The results of this meta-analysis of knee pain efficacy for 1 injection, 2–4 injections and ≥5 injections of IA-HA concur with a recently published RCT [49] in suggesting that the dosing regimen of IA-HA should be considered for development of future guidelines in treating symptomatic knee OA. Zoboli et al. published a head-to-head comparison, RCT assessing whether a single 6 ml application of HA has the same effectiveness as the three-weekly 2 ml dose [49]. Although this study showed no significant differences in efficacy between one single injection and three weekly injections of the same dose of IA-HA administered, the 3 weekly injections regimen showed statistically significant improvement from baseline pain (WOMAC pain and VAS) whereas the single injection regimen did not. Although there were subgroup differences between the numbers of injections administered within this meta-analysis, there was no subgroup difference in efficacy observed when comparing total dose of IA-HA administered to participants vs IA-Saline.

Recent evidence suggests products with an average molecular weight ≥ 3000 kDa provide favorable efficacy results when compared with products of an average molecular weight < 3000 kDa, and significantly fewer discontinuations compared with products with a molecular weight < 1500 kDa¹³. In this meta-analysis, we identified that the number of injections may also play a significant role in efficacy outcomes. Although the 2–4 injection subgroup included products within a wide range of MW’s the effect size was still significant compared to the single injection subgroup which consisted of only cross-linked high molecular weight (HMW) products. It has been suggested oxygen-derived free radicals act as a mediator in the inflammatory response, and that these radical species are responsible for increased HA degradation [50]. HMW products may achieve better efficacy due to an increased residency time of within the synovial fluid, producing a prolonged anti-inflammatory response within the joint, blocking inflammatory receptors, and a longer lasting chondroprotective effect (inhibition of metalloproteases, nitric oxide, and stimulation of proteoglycan/glycosaminoglycan synthesis) [51‐53]. Therefore, repeated exposures of HA may perpetuate improvement in the synovial fluid environment allowing subsequent IA-HA shots to provided extended effects. The results of this study help demonstrate that, while molecular weight is an important factor in the efficacy of HA products, the number of injections provided also plays a major role in optimizing the efficacy seen within knee OA patients. Thus is particularly clear given that, although all single injection HA products were of a high MW, they did not demonstrate a reduction in pain comparable to the 2–4 injections subgroup, which included several studies of LMW HA products. This study demonstrates that receiving the typical 3 injection regimen of a HA may be more effective than a single injection high molecular weight product. Future aims should investigate the mechanism of actions of high molecular weight HA in multiple injections compared to a single injection.

Significantly more treatment-related AEs were observed in participants receiving ≥5 injections vs IA-Saline; a result not seen in the single and 2–4 injection categories. Although this comparison to IA-Saline was significant, the subgroup analysis comparing the different injection regimens did not show a significant difference in treatment-related AEs between numbers of injection subgroups. Moreover, four of the five studies that reported treatment-related AEs reported a difference of only 1–3 events between treatment groups, and this analysis was largely subjugated by a single study conducted by Henderson et al.²⁵⁶ that reported a difference of 11 events, with the IA-HA treatment group experiencing more AEs. This meta-analysis observed very few serious treatment-related AEs, with little to no events reported for all injection regimens. Our findings are similar to the meta-analysis conducted by Miller and Block [54] that focused on the safety and efficacy of US approved IA-HA products in saline-controlled trials, where no SAEs were determined related to injection of HA or saline.

This review has strength in its methodological approach in systematically identifying available saline-controlled RCTs from online databases. The use of a thorough and systematic approach to article selection and data abstraction provides further strength to this report. A search of the grey literature or unpublished literature was not conducted; however, authors scanned references in articles that met the inclusion criteria for literature not captured in the search. Limitations of this review include that some studies of IA-HA treatment were excluded from the analysis due to not reporting efficacy measurements for knee pain. Another limitation is the lack of a direct comparison between the numbers of injections received, as the literature is not robust enough to permit such analysis. Additionally, the majority of these RCTs were industry funded with a moderate to high risk of bias. Further, the inconsistent reporting of pain scores, along with the variable length of follow-up time between studies provides another limitation with respect to our pooled results. Heterogeneity was also seen within some subgroups, which is an additional limitation to this study. Finally, no assessment or consideration of the effect of the placebo effect in relation to the number of injections was considered within these analyses.

This review has provided a detailed evaluation of differences between injection regimens of IA-HA for knee pain in OA. Future studies should directly compare different injection regimens of IA-HA in head-to-head RCTs. Moreover, future studies should review alternative outcome measurements such as function, stiffness, and withdrawal rates due to the different number of IA-HA injections vs saline. Further studies should also aim to further compare the different HA products and review intrinsic efficacy and safety profiles of different products based on the number of injections and their molecular weight, structure, and production method. The identification of product-specific results would also allow for greater specificity in drafting clinical guidelines for use of HA in knee OA. Injection accuracy is also a factor that may contribute to the overall efficacy and safety profile of IA-HA treatment [55], which may warrant future investigation of how accuracy improvement using ultrasound-guided injection techniques may affect clinical outcomes in trials. Additionally, the use of pre-set criteria for response, such as the OMERACT-OARSI responder criteria, is an emerging outcome measurement tool that may provide a more appropriate assessment of individual patient outcomes rather than group mean responses [56]. However, the lack of consistent reporting of this outcome within the current literature precluded the ability to conduct any analysis of OMERACT-OARSI responders between injection regimens [56]. Future studies should consider analyzing measures of individual response according to present criteria for response when designing clinical trials in knee osteoarthritis, including those addressing the impact of IA-HA.

Other factors in addition to injection regimen may contribute to the efficacy and safety of IA-HA products, such as molecular weight and production process. Altman and colleagues [6] concluded that in the available literature, IA-HA products with a molecular weight ≥ 3000 kDa and those derived from biological fermentation relate to superior efficacy and safety [6]. The low molecular weight (LMW) IA-HA pooled effect size did not meet the minimum clinically important difference (MCID) threshold, demonstrating an insignificant clinical effect for pain relief. Due to the multiple variables that may contribute to the efficacy of IA-HA products, additional investigations comparing the different types of HA products are required to fully understand the efficacy differences of IA-HA products in knee OA. Guideline development groups and clinicians should consider the injection regimen for various types of IA-HA treatments in decision-making processes regarding the appropriate use of IA-HA treatment for knee OA.

Overall, 2–4 and ≥5 injection regimens provided pain relief over IA-Saline. Intra-articular injections of HA used in a 2–4 injection treatment regimen provided the greatest benefit when compared to IA-Saline with respect to pain improvement in patients with knee OA, and were generally deemed safe with few to no treatment-related AEs reported across studies. Future research is needed to directly compare these treatment regimens, as well as further investigate the effects of other variables, such as product molecular weight, in the comparison of IA-HA injection treatment regimens.