Multiple tendon conditions · Review · 2026
How does shockwave compare with other treatments for tendon pain?
In this review of 65 trials, shockwave eased pain more than placebo in tennis elbow, Achilles and shoulder tendon problems. Against exercise, results were mostly similar. When many treatments were compared together, shockwave usually ranked second, including for jumper's knee.
Ranked well in some comparisons- trials, with 3,921 people
- 65
- tendon problems: elbow, shoulder, Achilles, kneecap, outer hip
- 5
- years the trials were published
- 2002–2024
Good to know: Covers five tendon problems. For jumper's knee, results against placebo were mixed.
Guo NY, Wang SQ, Liu W, et al. Orthopaedic Surgery, 2026 · Read the paper ↗
How this study worked
What happened
- studies in the tennis-elbow comparison
- 14
- Compared with
- Placebo, exercise, ultrasound therapy, elastic taping, injections and other non-surgical treatments
Shockwave against the top-ranked treatment
Difference in score points; below zero favours shockwave
Show the numbers
| Difference (score points) | likely range | |
|---|---|---|
| Tennis elbow pain and function: vs elastic taping | +0.39 | −0.88 to 1.66 |
| Jumper's knee pain: vs platelet-rich plasma | +0.37 | −0.87 to 1.61 |
What the trials used
- Evidence base
- The authors searched PubMed, Embase, Cochrane Library, Web of Science and EBSCO, plus the U.S. clinical trial registry for registered but unpublished studies, covering studies published before July 2024. They included randomized trials of ESWT in the five specified tendon conditions and compared it with conservative controls. Only English-language studies were eligible.
- Intervention variability
- Included trials used focused or radial ESWT with heterogeneous energy-flux density, impulse counts, schedules and co-interventions. The review did not conduct formal dose or modality subgroup analyses because reporting was inconsistent and incomplete; no single regimen can be attributed to this pooled evidence.
- Analysis approach
- Pairwise random-effects meta-analyses were subgrouped by condition, measure and control. Network meta-analyses were run separately by condition and outcome scale using mean differences (MD), and reported comparative effects and SUCRA ranks; a cross-condition VAS network was explicitly exploratory.
Who took part
3,921 people in 65 trials: tennis elbow (27 trials), shoulder tendon pain (12), Achilles (11), outer hip pain (7) and jumper's knee (8).
Side effects
This review did not pool or count side effects.
All the numbers and the full appraisal
The results, in numbers.
Systematic review and pairwise/network meta-analysis of 65 randomized studies published 2002–2024, comprising 3,921 participants. The included condition counts were lateral epicondylitis 27, rotator-cuff tendinopathy 12, Achilles tendinopathy 11, GTPS 7, and patellar tendinopathy 8. Condition-specific networks and outcome datasets may overlap and have their own study/arm counts.
Lateral epicondylitis: VAS pain10 reported outcomes
Lateral epicondylitis: VAS pain
Visual analogue scale (VAS) standardized mean difference (SMD), 95% CI
Thirteen studies contributed VAS data. Results favored ESWT over ultrasound (SMD −2.53, 95% CI −3.03 to −2.03; I²=64.7%) and over placebo (SMD −0.53, 95% CI −0.76 to −0.30; I²=39.0%). Against kinesiotaping, the reported estimate was SMD 0.64 (95% CI 0.14 to 1.13; I²=96.5%); the article text inconsistently describes this comparison as having no significant effect despite a CI excluding zero, so direction/coding needs caution.
Analysis details and source
These are review-level pooled contrasts, not raw arm means. Heterogeneity (I²) and the comparator are stated for this specific estimate; non-significance does not establish equivalence. The I² values follow the Results and Abstract; the Discussion gives the placebo-comparison I² for pain and PRTEE in reverse order (55.8% and 39.0%).
Achilles tendinopathy: VAS pain
Visual analogue scale (VAS) standardized mean difference (SMD), 95% CI
Five studies contributed VAS data. Results favored ESWT over placebo (SMD −0.49, 95% CI −0.83 to −0.16; I²=0%). Versus exercise, the estimate was SMD −0.06 (95% CI −0.35 to 0.23; I²=0%), with no statistically significant difference.
Analysis details and source
These are review-level pooled contrasts, not raw arm means. Heterogeneity (I²) and the comparator are stated for this specific estimate; non-significance does not establish equivalence.
Patellar tendinopathy: VAS pain
Visual analogue scale (VAS) standardized mean difference (SMD), 95% CI
Six studies contributed VAS data. Results favored ESWT over ultrasound (SMD −3.52, 95% CI −4.14 to −2.90; I²=0%) and over placebo (SMD −0.40, 95% CI −0.69 to −0.12; I²=69.2%).
Analysis details and source
These are review-level pooled contrasts, not raw arm means. Heterogeneity (I²) and the comparator are stated for this specific estimate; non-significance does not establish equivalence. The Abstract and Conclusion say ESWT showed no significant effect over placebo in patellar tendinopathy, and the Discussion says it has no significant therapeutic advantage over placebo, although this pairwise VAS-versus-placebo estimate in the Results excludes zero; the estimate is kept as printed. The Discussion also pairs placebo with I² 0.00% and ultrasound with 69.2%, the reverse of the Results, before calling the placebo subgroup's heterogeneity moderate; the values here follow the Results.
Rotator cuff tendinopathy: VAS pain
Visual analogue scale (VAS) standardized mean difference (SMD), 95% CI
Four studies contributed VAS data. ESWT versus exercise was not significant (SMD −0.08, 95% CI −0.44 to 0.28; I²=94.2%). Versus placebo, ESWT favored treatment (SMD −1.25, 95% CI −1.61 to −0.89; I²=89.1%), with substantial heterogeneity in both comparisons.
Analysis details and source
These are review-level pooled contrasts, not raw arm means. Heterogeneity (I²) and the comparator are stated for this specific estimate; non-significance does not establish equivalence.
Exploratory cross-condition NMA: VAS
Visual analogue scale (VAS) relative treatment rank (SUCRA)
The exploratory network included 42 studies (88 arms). Among 15 interventions, laser ranked first (SUCRA 82.7%) and ESWT fourth (65.5%). The authors caution that this pooled cross-disease ranking is for hypothesis generation only and should not guide condition-specific care.
Analysis details and source
These are review-level pooled rankings, not raw arm means. The main text reports only SUCRA ranks for this network, with no comparative estimate or heterogeneity statistic (I²); a SUCRA rank does not by itself show significant superiority.
Lateral epicondylitis NMA: VAS
Visual analogue scale (VAS) mean difference (MD), 95% CI
The 21-study VAS network ranked PDRN first (SUCRA 71.0%) and ESWT second (67.8%). ESWT versus PDRN was not significant (MD 0.44, 95% CI −2.76 to 3.64).
Analysis details and source
These are review-level pooled contrasts, not raw arm means. The comparator is stated for this specific estimate; the main text reports no heterogeneity statistic (I²) for this network. Non-significance does not establish equivalence.
Achilles tendinopathy NMA: VAS
Visual analogue scale (VAS) mean difference (MD), 95% CI
The six-study network ranked laser first (SUCRA 100%) and ESWT second (75.3%). The reported ESWT-versus-laser estimate was MD 2.55 (95% CI 1.86 to 3.24), statistically favoring laser as interpreted by the authors.
Analysis details and source
These are review-level pooled contrasts, not raw arm means. The comparator is stated for this specific estimate; the main text reports no heterogeneity statistic (I²) for this network. Non-significance does not establish equivalence.
GTPS NMA: VAS pain
Visual analogue scale (VAS) mean difference (MD), 95% CI
The two-study network ranked corticosteroid injection first (SUCRA 95.6%) and ESWT second (54.4%). ESWT versus injection was not significant (MD 0.36, 95% CI −0.16 to 0.87); this sparse network cannot establish equivalence.
Analysis details and source
These are review-level pooled contrasts, not raw arm means. The comparator is stated for this specific estimate; the main text reports no heterogeneity statistic (I²) for this network. Non-significance does not establish equivalence.
Patellar tendinopathy NMA: VAS pain
Visual analogue scale (VAS) mean difference, scale points
ESWT versus platelet-rich plasma (PRP)
8 studies
The VAS network included 8 studies (16 arms). PRP ranked first (SUCRA 88.5%) and ESWT second (75.3%); their reported contrast was not statistically significant (MD 0.37, 95% CI −0.87 to 1.61).
Analysis details and source
Condition-specific NMA contrast; the full network comprised 8 studies and 16 arms. The paper reports SUCRA ranks as relative rankings, not clinical superiority. No estimate from inaccessible supporting material is inferred.
Rotator cuff NMA: VAS pain
Visual analogue scale (VAS) relative treatment rank (SUCRA)
The rotator-cuff VAS network included 7 studies (14 arms). ESWT ranked first among five conservative treatments (SUCRA 82.3%). The Results text gives no pairwise NMA estimate or interval for this rank; ranking alone does not establish superiority.
Analysis details and source
Network composition and rank are reported in Results. No missing figure or supplement values were reconstructed.
Lateral epicondylitis: PRTEE2 reported outcomes
Lateral epicondylitis: PRTEE
Patient-Rated Tennis Elbow Evaluation (PRTEE) standardized mean difference (SMD), 95% CI
Ten studies contributed PRTEE data. Results favored ESWT over placebo (SMD −0.61, 95% CI −0.85 to −0.37; I²=55.8%). Comparisons with ultrasound (SMD 0.11, 95% CI −0.36 to 0.58; I²=20.6%) and kinesiotaping (SMD 0.38, 95% CI −0.16 to 0.93; I²=87.9%) were not statistically significant.
Analysis details and source
These are review-level pooled contrasts, not raw arm means. Heterogeneity (I²) and the comparator are stated for this specific estimate; non-significance does not establish equivalence. The I² values follow the Results; the Discussion gives the placebo-comparison I² for pain and PRTEE in reverse order (55.8% and 39.0%).
Lateral epicondylitis NMA: PRTEE
Patient-Rated Tennis Elbow Evaluation (PRTEE) mean difference, score points
ESWT versus kinesiotaping (KT)
14 studies
The PRTEE network included 14 studies (28 arms). KT ranked first (SUCRA 79.1%) and ESWT second (67.5%); the reported contrast was not statistically significant (MD 0.39, 95% CI −0.88 to 1.66).
Analysis details and source
Condition-specific NMA contrast; the full network comprised 14 studies and 28 arms. The paper reports SUCRA ranks as relative rankings, not clinical superiority. No estimate from inaccessible supporting material is inferred.
Achilles tendinopathy: VISA-A1 reported outcome
Achilles tendinopathy: VISA-A
Victorian Institute of Sport Assessment–Achilles (VISA-A) standardized mean difference (SMD), 95% CI
Six studies contributed VISA-A data. Results favored ESWT over placebo (SMD 0.90, 95% CI 0.52 to 1.28; I²=0%), but ESWT did not differ significantly from exercise (SMD −0.10, 95% CI −0.34 to 0.13; I²=13.1%).
Analysis details and source
These are review-level pooled contrasts, not raw arm means. Heterogeneity (I²) and the comparator are stated for this specific estimate; non-significance does not establish equivalence. Although VISA-A is analyzed pairwise, the Results section does not report an Achilles VISA-A network meta-analysis; the condition-specific NMA reported here is for VAS.
Patellar tendinopathy: VISA-P2 reported outcomes
Patellar tendinopathy: VISA-P
Victorian Institute of Sport Assessment–Patella (VISA-P) standardized mean difference (SMD), 95% CI
Six studies contributed VISA-P data. The ESWT-placebo estimate was SMD −0.15 (95% CI −0.42 to 0.12; I²=0%); the interval includes no effect. This differs from the article’s broad narrative statements that can imply benefit, so the numerical subgroup estimate is retained.
Analysis details and source
These are review-level pooled contrasts, not raw arm means. Heterogeneity (I²) and the comparator are stated for this specific estimate; non-significance does not establish equivalence.
Patellar tendinopathy NMA: VISA-P
Victorian Institute of Sport Assessment–Patella (VISA-P) mean difference (MD), 95% CI
In the eight-study VISA-P network, PRP ranked first (SUCRA 79.2%) and ESWT second (67.4%). ESWT versus PRP was not significant (MD −0.27, 95% CI −1.46 to 0.92). A favorable rank does not establish benefit versus placebo; the text does not provide that specific estimate here.
Analysis details and source
These are review-level pooled contrasts, not raw arm means. The comparator is stated for this specific estimate; the main text reports no heterogeneity statistic (I²) for this network. Non-significance does not establish equivalence.
Rotator cuff tendinopathy: SPADI2 reported outcomes
Rotator cuff tendinopathy: SPADI
Shoulder Pain and Disability Index (SPADI) standardized mean difference (SMD), 95% CI
Three studies contributed SPADI data. Results favored ESWT over placebo (SMD −0.46, 95% CI −0.71 to −0.22; I²=0%).
Analysis details and source
These are review-level pooled contrasts, not raw arm means. Heterogeneity (I²) and the comparator are stated for this specific estimate; non-significance does not establish equivalence.
Rotator cuff NMA: SPADI
Shoulder Pain and Disability Index (SPADI) mean difference (MD), 95% CI
In the five-study SPADI network, exercise ranked first (SUCRA 84.5%) and ESWT second (61.6%); the comparison was not significant (MD 0.15, 95% CI −0.24 to 0.55).
Analysis details and source
These are review-level pooled contrasts, not raw arm means. The comparator is stated for this specific estimate; the main text reports no heterogeneity statistic (I²) for this network. Non-significance does not establish equivalence.
GTPS: NRS pain1 reported outcome
GTPS: NRS pain
Numerical Rating Scale (NRS) standardized mean difference (SMD), 95% CI
Two studies contributed NRS data. Results favored ESWT over exercise (SMD −0.41, 95% CI −0.70 to −0.11; I²=10.2%).
Analysis details and source
These are review-level pooled contrasts, not raw arm means. Heterogeneity (I²) and the comparator are stated for this specific estimate; non-significance does not establish equivalence.
How treatment was delivered
The authors searched PubMed, Embase, Cochrane Library, Web of Science and EBSCO, plus the U.S. clinical trial registry for registered but unpublished studies, covering studies published before July 2024. They included randomized trials of ESWT in the five specified tendon conditions and compared it with conservative controls. Only English-language studies were eligible.
Included trials used focused or radial ESWT with heterogeneous energy-flux density, impulse counts, schedules and co-interventions. The review did not conduct formal dose or modality subgroup analyses because reporting was inconsistent and incomplete; no single regimen can be attributed to this pooled evidence.
Pairwise random-effects meta-analyses were subgrouped by condition, measure and control. Network meta-analyses were run separately by condition and outcome scale using mean differences (MD), and reported comparative effects and SUCRA ranks; a cross-condition VAS network was explicitly exploratory.
The main text does not provide a systematic pooled harms analysis or numerical adverse-event totals. The review focuses on efficacy and cites no arm-level safety denominator; absence of a harms estimate here is not evidence that ESWT is harmless.
The authors included only English-language studies, introducing language-related information bias.
They could not perform formal subgroup analyses by focused versus radial modality or energy because intervention parameters were inconsistently or incompletely reported; residual heterogeneity remained.
Placebo, exercise and ESWT are difficult to blind, creating performance-bias concerns. Trial-level interventions, populations and controls were heterogeneous, and the GTPS network was based on only two studies.
The article references supplementary Table S4 for included-study details and Table S5 for GRADE ratings, but those files were not present in the acquired XML; unique participant totals and full certainty judgments could not be independently mapped here.
Findings checked against the full source by a separate automated reviewer.
Limitations noted by the reviewers
- The authors included only English-language studies, introducing language-related information bias.
- They could not perform formal subgroup analyses by focused versus radial modality or energy because intervention parameters were inconsistently or incompletely reported; residual heterogeneity remained.
- Placebo, exercise and ESWT are difficult to blind, creating performance-bias concerns. Trial-level interventions, populations and controls were heterogeneous, and the GTPS network was based on only two studies.
- The article references supplementary Table S4 for included-study details and Table S5 for GRADE ratings, but those files were not present in the acquired XML; unique participant totals and full certainty judgments could not be independently mapped here.
Source-linked critical reading
Systematic review with network meta-analysis
Technical source record
Full acquired main-text XML reviewed end-to-end. Numerical SMDs, confidence intervals, I² values, network estimates, study/arm counts and SUCRA rankings are transcribed from the text. Supplementary Tables S4–S5 and Figures S2–S9 are cited by the article but were not supplied; no unreported numerical values were reconstructed.
For patients, the review suggests that shockwave therapy may help some tendon problems, but the evidence differs by condition and by what it is compared with. A ranking alone cannot identify the best treatment, and the review does not settle a standard device protocol.
Among people with five common tendinopathies, what do randomized trials show about ESWT versus conservative alternatives for pain and function? The authors used both pairwise and network meta-analysis; pooled results remain condition- and comparator-specific.
This systematic review synthesized 65 randomized controlled studies. It searched five databases and the U.S. clinical trial registry for studies published before July 2024 and used pairwise and network analyses. The article reports 65 trials, while network totals differ by outcome and condition; no single independent participant denominator is established in the main text.
Results reports 65 studies and 3,921 participants across five conditions: lateral epicondylitis 27, rotator cuff 12, Achilles 11, GTPS 7, and patellar 8. Network counts are outcome-specific and may overlap.
ESWT protocols varied across trials, including focused and radial devices, energy flux density, impulses, session schedules and targets. The review could not perform formal analyses by modality or energy because reporting was incomplete. It therefore does not support one universal treatment prescription.
Comparators included placebo or no treatment, exercise, ultrasound, kinesiotaping, injections and other conservative modalities. Pairwise syntheses compared ESWT with one control category at a time; network analyses connected multiple treatments where evidence permitted. Sparse networks require caution.
For pain and function, pairwise results favored ESWT over placebo in lateral epicondylitis and Achilles tendinopathy; estimates varied by outcome and comparator. In patellar disease, VAS favored ESWT over placebo while VISA-P did not show a significant difference. Shoulder and GTPS estimates also varied, with very high shoulder-pain heterogeneity. Added source-reported NMA contrasts include LE PRTEE (14 studies/28 arms; ESWT vs KT MD 0.39, 95% CI −0.88 to 1.66; no significant difference), patellar VAS (8/16; ESWT vs PRP MD 0.37, 95% CI −0.87 to 1.61; no significant difference), and rotator-cuff VAS network rank (7/14; ESWT SUCRA 82.3%, without a comparative estimate in text). The Achilles VISA-A results are pairwise, not a separately reported network.
Network rankings are not proof of superiority. For example, ESWT ranked second for lateral-elbow VAS and patellar outcomes, with reported comparisons to top-ranked treatments crossing no effect; laser ranked above ESWT in Achilles VAS, while the GTPS injection comparison was imprecise.
The acquired main text did not provide pooled adverse-event counts or a systematic harms analysis. Safety cannot be inferred from this absence.
Important constraints include English-only eligibility, difficulty blinding treatment, heterogeneous protocols and populations, residual heterogeneity, and inability to stratify formally by shockwave type or energy. Cross-condition pooling is explicitly exploratory. Supplementary Tables S4–S5 and figures were referenced but unavailable in the acquired XML.
The findings are most useful as condition-specific comparisons. They do not establish a single ESWT dose, show that non-significant comparisons are equivalent, or justify applying the exploratory all-tendinopathy ranking to an individual diagnosis.
Future trials should report modality, energy, impulses, treatment schedule, co-interventions and follow-up completely; stratify key subtypes such as insertional/non-insertional Achilles and calcific/non-calcific shoulder disease; and prespecify clinically meaningful pain, function and harms outcomes.
Read the original article.
The Effect of Extracorporeal Shockwave Therapy in Tendinopathy: A Systematic Review and Network Meta-Analysis of Randomized Controlled Trials.
Guo NY, Wang SQ, Liu W, Mao ZM, Wang JQ, Xu BB. · Orthopaedic surgery · 2026
Open the publication record or available full-text files.
Checked against the full paper · Clinical review: Dr. Marcus Yu Bin Pai, · How we review →