Multiple musculoskeletal conditions · Review · 2023
How do hyaluronic acid injections compare with shockwave for tendon pain?
This review studied hyaluronic acid injections for tendon pain. In the one trial with results against shockwave (Achilles tendon pain), pain fell by about 68% with the injections and 48% with shockwave at 4 weeks; the injection group stayed ahead at 3 and 6 months.
Hyaluronic acid did better in one Achilles trial- trials of hyaluronic acid injections, with 1,629 people
- 19
- trials compared the injections with shockwave
- 2
- people per group in the Achilles trial
- 40
Good to know: A review of injections, not shockwave; shockwave was one of several comparison treatments.
Khan M, Shanmugaraj A, Prada C, et al. Sports Health, 2023 · Read the paper ↗
How this study worked
What the trials used
- Review intervention scope
- Nineteen RCTs evaluated HA injection for soft-tissue conditions with comparators including corticosteroids, placebo, PRP, shockwave therapy, physiotherapy and prolotherapy. HA products, dose and schedules varied; the review does not define one regimen.
- Direct ESWT comparison
- Table 1 lists two RCTs comparing HA with ESWT: Lynen et al. (Achilles tendinopathy; 40 participants per group) and Frizziero et al. (rotator cuff disease; 17 per group). Only Lynen has a reported result in the review text. The retrieved review text does not state ESWT type, energy, pulse count or session schedule for either trial.
Who took part
1,629 people in 19 trials with shoulder, elbow, ankle, hand, kneecap and Achilles problems.
Side effects
No serious side effects were reported. Several trials noted brief pain at the injection or tendon site.
All the numbers and the full appraisal
The results, in numbers.
Nineteen RCTs included 1,629 patients with varied soft-tissue indications; 17 trials (1,307 patients) contributed to all-indication pooled pain estimates.
Pain across all soft-tissue indications6 reported outcomes
Pain across all soft-tissue indications
VAS transformed by review to orient scores toward improvement Mean difference (VAS points)
HA versus comparators across indications
17 studies · 1,307 participants in this estimate
The review reports an average difference favoring HA, pooled across multiple conditions and comparators. Heterogeneity was very high (I²=98%).
Analysis details and source
Fixed-effect model despite extreme heterogeneity; this is not an ESWT-specific contrast. Scores were transformed so higher values indicate improvement. The Abstract and the Sensitivity Analysis paragraph print a different main estimate for this period (MD 2.48, 95% CI 2.31–2.65); the stored value is the one printed in the Results text with Figure 2.
Pain across all soft-tissue indications
VAS transformed by review to orient scores toward improvement Mean difference (VAS points)
HA versus comparators across indications
15 studies · 1,219 participants in this estimate
The pooled estimate favored HA but combines conditions and comparators and had I²=98.9%.
Analysis details and source
Fixed-effect model; not an HA-versus-ESWT estimate. Higher transformed scores indicate improvement.
Rotator cuff disease pain
VAS transformed by review to orient scores toward improvement Mean difference (VAS points)
HA versus all comparators in rotator cuff disease
For rotator cuff disease, HA was favored over comparators at 6–12 weeks; heterogeneity remained very high (I²=94%).
Analysis details and source
Fixed-effect pooled estimate. The review does not isolate an ESWT comparator in this subgroup estimate. This estimate is not restricted to placebo or physical therapy; the review reports the comparison with saline or physical therapy alone separately at 6–12 weeks (MD 1.51, 95% CI 1.20–1.81, I²=95%).
Pain across all soft-tissue indications
VAS transformed by review to orient scores toward improvement Mean difference (VAS points)
HA versus mixed comparators across the stated indications; not an ESWT-specific estimate
6 studies · 656 participants in this estimate
The pooled HA estimate favored HA by 3.57 VAS points (95% CI 3.35 to 3.78); this combines the stated indications and comparators and is not an HA-versus-ESWT effect. Heterogeneity was I²=98.9%.
Analysis details and source
Fixed-effect model; source-reported P<0.001; 6 trials, 656 participants. The review transformed pain scores so positive mean differences indicate greater improvement with HA.
Pain across all shoulder indications and comparators
VAS transformed by review to orient scores toward improvement Mean difference (VAS points)
HA versus mixed comparators across the stated indications; not an ESWT-specific estimate
10 studies
The pooled HA estimate favored HA by 1.16 VAS points (95% CI 0.88 to 1.44); this combines the stated indications and comparators and is not an HA-versus-ESWT effect. Heterogeneity was I²=92%.
Analysis details and source
Fixed-effect model; source-reported P<0.001; 10 trials; participant count not reported. The review transformed pain scores so positive mean differences indicate greater improvement with HA.
Shoulder pain at less than 6 weeks and 6–12 weeks
VAS transformed by review to orient scores toward improvement Mean difference (VAS points)
HA versus mixed comparators across the stated indications; not an ESWT-specific estimate
8 studies
The pooled HA estimate favored HA by 1.44 VAS points (95% CI 1.15 to 1.73); this combines the stated indications and comparators and is not an HA-versus-ESWT effect. Heterogeneity was I²=93%.
Analysis details and source
Fixed-effect model; source-reported P<0.001; 8 trials; participant count not reported. The review transformed pain scores so positive mean differences indicate greater improvement with HA.
Shoulder function (Constant score)1 reported outcome
Shoulder function (Constant score)
Constant score Mean difference (points)
HA versus comparators for shoulder indications
4 studies
The pooled Constant score favored HA; trial comparators and shoulder diagnoses varied.
Analysis details and source
The review cites a wide published Constant-score MCID range (8–36 points) and another threshold of 10.4; these thresholds are contextual, not trial-derived.
Achilles tendinopathy pain reduction1 reported outcome
Achilles tendinopathy pain reduction
Percent reduction reported in review; original VAS scores not given in narrative Percent reduction from baseline (as reported)
Table 1 identifies two HA-versus-ESWT RCTs: Frizziero et al. (2017), rotator-cuff disease, 17 participants per arm; and Lynen et al. (2016), Achilles tendinopathy, 40 per arm. Only Lynen has a numerical result in this review narrative: pain reductions of 68.1% with HA and 47.9% with ESWT at 4 weeks, with the HA advantage reported to persist at 3 and 6 months (P=0.0030). No specific outcome estimate for Frizziero is given in the accessible narrative.
Analysis details and source
The 40-per-arm count is the Table 1 sample size for the Lynen comparison, not a follow-up-specific analyzed denominator. The 17-per-arm count is the Table 1 Frizziero trial. No device parameters or protocol were available in the review narrative for either trial; do not attribute the Achilles effect to the Frizziero shoulder trial.
How treatment was delivered
Nineteen RCTs evaluated HA injection for soft-tissue conditions with comparators including corticosteroids, placebo, PRP, shockwave therapy, physiotherapy and prolotherapy. HA products, dose and schedules varied; the review does not define one regimen.
Table 1 lists two RCTs comparing HA with ESWT: Lynen et al. (Achilles tendinopathy; 40 participants per group) and Frizziero et al. (rotator cuff disease; 17 per group). Only Lynen has a reported result in the review text. The retrieved review text does not state ESWT type, energy, pulse count or session schedule for either trial.
No serious adverse events were reported. Several trials described transient pain at injection or tendon sites; two trials did not report complications.
The review pools varied diagnoses and comparators under fixed-effect models despite very high heterogeneity; some SDs were estimated, and many trials were small or at high risk of bias.
Table 1 lists two HA-versus-ESWT RCTs, but only one (Lynen et al., Achilles tendinopathy) has a reported result in the review narrative; its detailed group/timepoint estimates and the ESWT parameters of both trials are not reported there. The pooled all-indication effects are not HA-versus-ESWT effects.
Appendices with additional search and outcome material were referenced but not available in the retrieved XML.
Findings checked against the full source by a separate automated reviewer.
Limitations noted by the reviewers
- The review pools varied diagnoses and comparators under fixed-effect models despite very high heterogeneity; some SDs were estimated, and many trials were small or at high risk of bias.
- Table 1 lists two HA-versus-ESWT RCTs, but only one (Lynen et al., Achilles tendinopathy) has a reported result in the review narrative; its detailed group/timepoint estimates and the ESWT parameters of both trials are not reported there. The pooled all-indication effects are not HA-versus-ESWT effects.
- Appendices with additional search and outcome material were referenced but not available in the retrieved XML.
Source-linked critical reading
Systematic review with meta-analysis
Technical source record
Read the complete article, raw Table 1, the Table 2 caption and column headers and the four figure captions; the Table 2 body and Figures 1–4 are images that were not read. Checked pooled outcomes, timepoints, comparator heterogeneity, rotator-cuff and Achilles reporting in the text, risk of bias, adverse events, sensitivity analysis, discussion, limitations, conclusion and end matter. Online appendices were unavailable.
This review suggests HA may improve pain in several soft-tissue conditions, but the pooled studies are highly heterogeneous and combine different treatments. It does not show that HA is better than shockwave therapy for most conditions; of its two direct comparisons with shockwave, only the Achilles trial has a result in the review text.
Across soft-tissue injuries, what does randomized-trial evidence report about hyaluronic acid (HA) injections for pain, function and complications, including direct comparisons with shockwave therapy?
Systematic review of 19 RCTs published 2001–2021, searched through February 24, 2021 in four databases. The overall analyses included 17 RCTs (1,307 patients) for pain and function/adverse-event analyses across varied indications; study populations were deduplicated when reports reused cohorts.
The review reports 1,629 patients across 19 trials, with mean age 48.5 (SD 9.9) years. Seventeen trials and 1,307 patients contributed to the pooled all-indication pain results. Conditions included shoulder/rotator cuff, elbow, ankle, hand, patellar and Achilles disorders; these are not one clinical population.
HA formulations and schedules varied; comparators included placebo, physiotherapy, corticosteroid, PRP, prolotherapy and ESWT. Table 1 lists two distinct HA-versus-ESWT RCTs: Frizziero 2017 in rotator-cuff disease (17 per arm) and Lynen 2016 in Achilles tendinopathy (40 per arm). Only the Achilles trial has a numeric outcome in the review narrative; no device settings or time-specific analyzed denominators are supplied there.
The pooled HA estimates combine placebo and active comparators and multiple diagnoses; they are not HA-versus-ESWT effects. Some HA strategies included physiotherapy, standard care, lidocaine or other co-interventions.
Pain was assessed using VAS, with scores transformed so higher values indicated improvement. Function included Constant and ASES scores. The authors pooled short term (<6 weeks), 6–12 weeks and mid-term (>12 weeks); complications were summarized descriptively.
Across all indications, the review reports pain MDs favoring HA of 2.39 (95% CI 2.23–2.55) before 6 weeks, 2.03 (1.86–2.20) at 6–12 weeks, and 3.57 (3.35–3.78) after 12 weeks. For rotator cuff disease, the 6–12-week pain estimate across comparators was 1.50 (1.20–1.79); for shoulder indications, the 6–12-week Constant function MD was 13.45 (12.29–14.61). High heterogeneity limits pooled interpretation.
No serious adverse events were reported in included trials. Several trials reported transient injection-site or tendon pain; the review describes few minor events. Two trials did not report complications, and long-term safety surveillance was limited.
The authors used fixed-effect pooling despite very high heterogeneity (I² about 98% overall and 92–95% for shoulder/rotator-cuff subgroups), imputed or estimated SDs when missing, and included small or high-risk-of-bias trials. Indication-specific comparator effects are often sparse. In the review text, the direct HA-versus-ESWT finding comes from one Achilles trial and lacks detailed parameters.
Pooled HA results across all soft-tissue diagnoses cannot establish an HA-versus-shockwave effect for an individual condition. In the review text, the only direct ESWT comparison with a result concerns Achilles tendinopathy, not shoulder calcific tendinopathy or other conditions; the text gives no result for its second HA-versus-ESWT trial, in rotator cuff disease, and Table 2 and Figure 4, which may hold one, are images that were not read.
Larger, lower-bias trials should compare HA directly with relevant alternatives by diagnosis, report formulation and dose, provide complete group data and timepoint denominators, and monitor tendon status and adverse events over longer follow-up.
Read the original article.
The Role of Hyaluronic Acid for Soft Tissue Indications: A Systematic Review and Meta-Analysis.
Khan M, Shanmugaraj A, Prada C, Patel A, Babins E, Bhandari M. · Sports health · 2023
Open the publication record or available full-text files.
Checked against the full paper · Clinical review: Dr. Marcus Yu Bin Pai, · How we review →