Multiple tendon conditions · Review · 2024
Does shockwave ease arm and shoulder tendon pain more than placebo?
In this review of 18 trials, pain 3 months after shockwave was about 1.5 points lower than after a placebo treatment. Shoulder function scores also improved more. Results varied between tendons, wave types and strengths.
Helped more than placebo at 3 months- placebo-controlled trials, with 1,351 people
- 18
- tendon areas: shoulder, elbow, biceps, finger
- 4
- months after treatment, when pain was compared
- 3
Good to know: Most trials were in shoulder tendon pain or tennis elbow (8 each). Results shown are at about 3 months.
The paper gives two different totals: 18 trials with 1,351 people overall, and 16 trials with 1,751 people for the pain results.
Xiong Y, Wen T, Jin S, et al. Frontiers in Medicine, 2024 · Read the paper ↗
How this study worked
What happened
- trials in the pain result
- 16
- people
- 1,751
- Compared with
- Placebo (sham) shockwave
Arm and shoulder tendon pain at 3 months
Difference from placebo in pain score points; below zero favours shockwave
Show the numbers
| Difference (pain score points) | likely range | |
|---|---|---|
| Pain at 3 months, all tendons | −1.45 | −2.46 to −0.45 |
What the trials used
- Energy and session ranges
- Eighteen RCTs covered rotator cuff, lateral elbow, long-head biceps and finger tendon disorders. Table 1 lists energy from 0.006 to 0.32 mJ/mm² when reported, 1,500–6,000 pulses in examples, and varying schedules from one session to five sessions. Authors define low EFD <0.12 mJ/mm² and high EFD 0.12–0.38 mJ/mm², leaving exactly 0.12 boundary classification ambiguous. Some radial trials report no EFD or maximum tolerance.
- Control conditions and co-interventions
- Eligibility required placebo/sham control. Table 1 lists placebo comparators; protocols varied in local anesthesia or oral analgesia, and review authors state these could contribute heterogeneity.
Who took part
1,351 people with arm or shoulder tendon pain in 18 trials: 740 had shockwave and 611 a placebo. Most had shoulder tendon pain or tennis elbow.
Side effects
Some trials reported brief pain and skin reactions that usually settled after treatment. A few people could not tolerate treatment, got worse or dropped out. No serious long-term side effects were reported.
All the numbers and the full appraisal
The results, in numbers.
Eighteen placebo-controlled RCTs included 1,351 patients with upper-limb tendinopathy: 740 in ESWT groups and 611 in placebo groups. Diagnoses and outcome subgroup counts vary, and the article reports internally inconsistent denominators in its outcome sections.
Visual analogue scale (VAS)12 reported outcomes
Pain across upper-limb tendinitis
Visual analogue scale (VAS) Mean difference (VAS points; as reported)
ESWT versus placebo/sham across included upper-limb conditions
16 studies · 1,751 participants in this estimate
The review reports a pooled pain difference at 3 months favoring ESWT, with very high heterogeneity (I²=95%).
Analysis details and source
The stated 16-study/1,751-participant contribution conflicts with the review-wide 18 RCTs/1,351 people and with the 1,193 people in the later 16-study modality subgroup.
Reported test for ESWT modality subgroup differences
Visual analogue scale (VAS) Mean difference (VAS points; as reported)
Ambiguously reported subgroup-difference statistic; not verified as a direct radial-versus-focused effect
The source prints −1.45 (95% CI −2.46 to −0.45), I²=77.4%, p=.04 as a subgroup-difference result, but this duplicates the overall estimate in §3.4.1 and the same paragraph says no significant modality difference. This is not presented as a verified direct radial-versus-focused treatment effect.
Analysis details and source
Preserve the source inconsistency: exact duplicate of the overall pooled estimate, alongside contradictory prose. Do not infer a treatment contrast.
Pain by energy subgroup
Visual analogue scale (VAS) Mean difference (VAS points; as reported)
High-energy ESWT versus placebo/sham
The review reports a high-energy subgroup estimate favoring ESWT.
Analysis details and source
The same section reports low-energy MD −1.74 (95% CI −3.33 to −0.14), P=0.03, then calls it nonsignificant versus control. The text is internally contradictory; low-energy effect not charted as a resolved result.
Overall pain at six months
Visual analogue scale (VAS) Source-reported MD and interval; not charted because internally inconsistent
Results §3.4.1 prints a six-month VAS MD of −0.29 (95% CI −1.33 to −0.74), I²=0%, p=.58 (the section's 16 studies/1,751 participants describe the VAS analyses as a whole), while describing improvement at both 3 and 6 months. The interval does not contain the point estimate and conflicts with the nonsignificant p-value; exact source text is preserved and no corrected estimate is inferred.
Analysis details and source
Conflicting source statistic is retained as qualitative text and intentionally not encoded as a numeric effect chart.
VAS in calcific tendinitis
Visual analogue scale (VAS) MD
ESWT vs placebo/sham within the reported subgroup
I²=98%
Source-reported VAS estimate in calcific tendinitis: -2.31 (95% CI -6.48 to 1.86), p=.28; heterogeneity I²=98%.
Analysis details and source
Source-reported subgroup estimate. Preserve any inconsistency with surrounding article text; do not reinterpret as a direct treatment contrast unless explicitly stated. §3.4.2 pools only low-energy ESWT (EFD < 0.12 mJ/mm²) trials for both calcification subgroups (5 studies, 380 participants).
VAS in non-calcific tendinitis
Visual analogue scale (VAS) MD
ESWT vs placebo/sham within the reported subgroup
I²=67%
Source-reported VAS estimate in non-calcific tendinitis: -1.55 (95% CI -3.46 to 0.36), p=.11; heterogeneity I²=67%.
Analysis details and source
Source-reported subgroup estimate. Preserve any inconsistency with surrounding article text; do not reinterpret as a direct treatment contrast unless explicitly stated. §3.4.2 pools only low-energy ESWT (EFD < 0.12 mJ/mm²) trials for both calcification subgroups (5 studies, 380 participants).
VAS in rotator-cuff tendinitis
Visual analogue scale (VAS) MD
ESWT vs placebo/sham within the reported subgroup
I²=0%
Source-reported VAS estimate in rotator-cuff tendinitis: -0.8 (95% CI -1.47 to -0.12), p=.02; heterogeneity I²=0%.
Analysis details and source
Source-reported subgroup estimate. Preserve any inconsistency with surrounding article text; do not reinterpret as a direct treatment contrast unless explicitly stated.
VAS in lateral epicondylitis
Visual analogue scale (VAS) MD
ESWT vs placebo/sham within the reported subgroup
I²=69%
Source-reported VAS estimate in lateral epicondylitis: -0.92 (95% CI -1.57 to -0.28), p=.005; heterogeneity I²=69%.
Analysis details and source
Source-reported subgroup estimate. Preserve any inconsistency with surrounding article text; do not reinterpret as a direct treatment contrast unless explicitly stated.
VAS in trigger finger
Visual analogue scale (VAS) MD
ESWT vs placebo/sham within the reported subgroup
Source-reported VAS estimate in trigger finger: -0.47 (95% CI -1.95 to 1.281), p=.53; heterogeneity not stated.
Analysis details and source
Source-reported subgroup estimate. Preserve any inconsistency with surrounding article text; do not reinterpret as a direct treatment contrast unless explicitly stated.
VAS in focused ESWT subgroup
Visual analogue scale (VAS) MD
ESWT vs placebo/sham within the reported subgroup
I²=46%
Source-reported VAS estimate in focused ESWT subgroup: -0.73 (95% CI -1.16 to -0.29), p=.001; heterogeneity I²=46%.
Analysis details and source
Source-reported subgroup estimate. Preserve any inconsistency with surrounding article text; do not reinterpret as a direct treatment contrast unless explicitly stated.
VAS in radial ESWT subgroup
Visual analogue scale (VAS) MD
ESWT vs placebo/sham within the reported subgroup
I²=95%
Source-reported VAS estimate in radial ESWT subgroup: -2.59 (95% CI -4.27 to -0.91), p<.01; heterogeneity I²=95%.
Analysis details and source
Source-reported subgroup estimate. Preserve any inconsistency with surrounding article text; do not reinterpret as a direct treatment contrast unless explicitly stated.
VAS in low-energy ESWT subgroup
Visual analogue scale (VAS) MD
ESWT vs placebo/sham within the reported subgroup
I²=95%
Section 3.4.5 reports MD −1.74 (95% CI −3.33 to −0.14), I²=95%, p=.03, but the same passage says low-energy ESWT was “not significant compared with the control group.” The source therefore contains an explicit significance contradiction; retain both statements without resolving it.
Analysis details and source
The estimate and p value conflict with the article’s adjacent verbal statement that this subgroup was not significant versus control. Both are source-reported; do not select one as authoritative.
Constant–Murley score (CMS)1 reported outcome
Shoulder function in rotator cuff tendinitis
Constant–Murley score (CMS) Mean difference (CMS points; as reported)
ESWT versus placebo/sham
6 studies · 352 participants in this estimate
CMS favored ESWT overall. Subgroups favored ESWT for calcific cases (MD 5.76, 95% CI 0.92 to 10.61) and non-calcific cases (MD 12.20, 5.21 to 19.18).
Analysis details and source
The review also reports high-energy MD 18.47 (13.71–23.23) and low-energy MD 7.56 (3.69–11.43), with subgroup difference MD 11.91 (8.91–14.91), P<0.001; small, heterogeneous evidence base. The low-energy subgroup result is identical to the overall estimate (same MD, interval and I²=37%), which the source does not explain.
Grip strength (GS)1 reported outcome
Grip strength in lateral epicondylitis
Grip strength (GS) Mean difference (grip strength; unit not stated in the review)
ESWT versus placebo/sham
4 studies · 227 participants in this estimate
The pooled grip-strength result was not statistically significant; the interval includes no difference.
Analysis details and source
The authors rate evidence moderate because of small sample and wide confidence interval. Four trials did not report EFD.
How treatment was delivered
Eighteen RCTs covered rotator cuff, lateral elbow, long-head biceps and finger tendon disorders. Table 1 lists energy from 0.006 to 0.32 mJ/mm² when reported, 1,500–6,000 pulses in examples, and varying schedules from one session to five sessions. Authors define low EFD <0.12 mJ/mm² and high EFD 0.12–0.38 mJ/mm², leaving exactly 0.12 boundary classification ambiguous. Some radial trials report no EFD or maximum tolerance.
Eligibility required placebo/sham control. Table 1 lists placebo comparators; protocols varied in local anesthesia or oral analgesia, and review authors state these could contribute heterogeneity.
Transient pain and local reactions were reported in some studies; symptoms generally resolved after treatment. Some participants could not tolerate treatment, worsened or withdrew. No serious long-term adverse effects were reported; event rates were not pooled or fully quantified.
Pain estimates had high heterogeneity; diagnosis-specific subgroups were small, some intervention parameters were missing, and three studies had incomplete outcome data.
The article reports conflicting study and participant denominators (18/1,351 overall; 16/1,751 for VAS; 16/1,193 for modality analyses), and contradictory significance statements around the 6-month VAS, modality subgroup difference and low-energy pain. Values are retained as reported without resolving the conflict.
Forest-plot images were not available for independent visual verification; values come from the text and tables.
Findings checked against the full source by a separate automated reviewer.
Limitations noted by the reviewers
- Pain estimates had high heterogeneity; diagnosis-specific subgroups were small, some intervention parameters were missing, and three studies had incomplete outcome data.
- The article reports conflicting study and participant denominators (18/1,351 overall; 16/1,751 for VAS; 16/1,193 for modality analyses), and contradictory significance statements around the 6-month VAS, modality subgroup difference and low-energy pain. Values are retained as reported without resolving the conflict.
- Forest-plot images were not available for independent visual verification; values come from the text and tables.
Source-linked critical reading
Systematic review with meta-analysis
Technical source record
Read complete full-text XML; verified 2 raw tables and 12 figure captions; checked population counts, energy cutoffs, study-level regimens, VAS/CMS/grip subgroup estimates, harms, GRADE, denominators, internal contradictory p/CI/significance statements, limitations and end matter. Figure pixels were unavailable; the XML lists no supplementary material.
This review suggests ESWT can improve some short-term pain and shoulder-function measures, but findings differ by tendon, modality and energy. Important internal reporting inconsistencies mean it does not establish a reliable best device or dose.
In upper-limb tendon disorders, what do placebo-controlled RCT syntheses report about ESWT effects on pain, shoulder function and grip strength, and do findings differ by diagnosis, modality or energy?
Systematic review/meta-analysis of 18 RCTs across four upper-limb diagnoses, searched in five databases through December 2023 and registered CRD42023403594. The review required placebo/sham controls and rated VAS/CMS/GS evidence with GRADE.
The Results section reports 1,351 people across the 18 included RCTs, 740 assigned to ESWT and 611 to placebo. Diagnoses included eight rotator-cuff, eight lateral-epicondylitis, one long-head-biceps and one finger-tendon trial. Later outcome subsections report inconsistent larger denominators.
Focused and radial ESWT varied in energy, pulse count, sessions and schedule; local anesthesia/oral analgesia was used in some trials. Authors classify low <0.12 and high 0.12–0.38 mJ/mm². Five of 18 trials lacked precise EFD data; radial EFD was often unavailable.
The review included sham/placebo-controlled trials. Comparators and blinding details vary; pooled estimates represent ESWT versus sham across heterogeneous upper-limb conditions, not one treatment protocol.
VAS assessed pain, Constant–Murley assessed rotator-cuff function, and grip strength assessed lateral-elbow function. The review’s method says closest 1–3-month timepoints were used when exact 3-month data were unavailable. Safety was described narratively. Additional Results subgroup values: at 6 months, overall VAS MD −0.29 (95% CI −1.33 to −0.74), I²=0%, p=.58 (the section's 16 studies/1,751 participants describe all VAS analyses); the interval excludes its point estimate and conflicts with p and prose. Calcific and non-calcific subgroup estimates were −2.31 (−6.48 to 1.86), p=.28, and −1.55 (−3.46 to .36), p=.11. Diagnosis subgroups: rotator cuff −.80 (−1.47 to −.12), p=.02; lateral epicondylitis −.92 (−1.57 to −.28), p=.005; trigger finger −.47 (−1.95 to 1.281), p=.53 (upper CI bound unusual as printed). Focused and radial subgroup estimates were −.73 (−1.16 to −.29), p=.001, and −2.59 (−4.27 to −.91), p<.01. Low-energy subgroup MD −1.74 (−3.33 to −.14), p=.03, conflicts with prose calling it nonsignificant. The reported modality subgroup-difference statistic −1.45 duplicates the overall value and is not treated as a verified direct effect.
The pooled 3-month pain estimate was MD −1.45 (95% CI −2.46 to −0.45; I²=95%), but reported study/patient totals conflict. Radial and focused subgroup estimates favored treatment, while the text contradicts itself about whether their subgroup difference is significant. High-energy pain results favored ESWT; low-energy significance is internally inconsistent. CMS favored ESWT at 3 months, while grip-strength difference was not significant.
The review reports transient pain, swelling, erythema, petechiae and small hematomas in some studies, usually resolving after treatment; some participants could not tolerate treatment, had worsening symptoms or withdrew. No serious long-term adverse effects were reported, but event denominators and collection methods are not supplied in the main text.
High VAS heterogeneity (I²>80%), small condition-specific subsets, missing energy/session parameters, three trials with incomplete outcome data, and mixed diagnosis/protocols limit confidence. The source says 18 RCTs/1,351 participants overall, but 16 studies/1,751 for VAS and 16/1,193 for modality subgroups. It reports a 6-month CI entirely below zero alongside P=0.58, and contradictory significance statements for radial/focused and low-energy pain results.
The review offers hypothesis-generating findings for upper-limb tendinopathy, with stronger evidence for short-term pain/function than for finger or biceps conditions. It cannot support a universal radial-over-focused recommendation given its internally conflicting subgroup interpretation.
Trials should report arm-level denominators, exact device and EFD, pulse/session schedule, co-interventions, standardized diagnosis and complete harms. Reviews should reconcile source counts, prespecify subgroup interactions and present consistent estimates/intervals.
Read the original article.
Efficacy and safety of extracorporeal shock wave therapy for upper limb tendonitis: a systematic review and meta-analysis of randomized controlled trials.
Xiong Y, Wen T, Jin S, Lin L, Shao Q, Peng Y, Zheng Q, Li W. · Frontiers in medicine · 2024
Open the publication record or available full-text files.
Checked against the full paper · Clinical review: Dr. Marcus Yu Bin Pai, · How we review →