Lateral elbow tendinopathy · Comparison study · 2024
How does shockwave compare with massage or a steroid injection for tennis elbow?
In this study of 52 people with tennis elbow, pain eased in all three groups. With four shockwave sessions it went from about 7 to 5 out of 10 by 3 months, the same as with massage or a steroid injection. The injection eased pain fastest, at 2 weeks.
Similar to massage or steroid at 3 months- people with tennis elbow
- 52
- shockwave sessions over 2 weeks
- 4
- months of follow-up
- 6
Good to know: Small study at one hospital; everyone also had anti-inflammatory tablets, a splint, exercise and rest.
Çetin BV, Sepetçi Ö, Yazar İ, et al. Joint Diseases and Related Surgery, 2024 · Read the paper ↗
How this study worked
Conceptual animation: a focused front converges, then spreads. Not to scale. Tap the field to send a pulse.
What happened
Elbow pain
0–10 pain scale · medians at the start and after 3 months
StartAfter 3 months
Show the numbers
| Group | People | Start | After 3 months |
|---|---|---|---|
| Shockwave | 16 | 7 | 5 |
| Massage | 10 | 6.5 | 5 |
| Steroid injection | 14 | 7 | 5 |
The treatment
Shockwave
Shockwave, twice a week for 2 weeks
- Sessions
- 4, two a week
- Pulses
- 1,500 per session
- Strength
- 2.0bar · 8 per second
- Numbing
- None
- Device
- Not named in the paper
- Everyone also got
- Anti-inflammatory tablets, a splint, exercise and rest
Compared with: Local massage three times a week for two weeks, or one steroid injection
Settings are what each study used, not a prescription.
Who took part
52 adults with tennis elbow for more than 6 weeks, confirmed by examination and MRI, at one university hospital.
hollow squares: no final score
Side effects
The study did not report side effects in a structured way.
All the numbers and the full appraisal
Every reported result
The results, in numbers.
Adults older than 18 years with lateral epicondylitis symptoms for >6 weeks, clinical examination and MRI diagnosis, and no treatment during the prior six months; recruited at Harran University Orthopedics and Traumatology between March 2021 and June 2022. Explicit exclusion criteria are not reported.
The article calls the study randomized in the abstract, but Methods say patients entered groups in order of presentation and describe no random sequence. Enrolled 17; 14 assessed at week 2 (Figure 1 shows 16 after the losses to follow-up, before 2 massage patients switched treatment within the first 2 weeks); 10 at month 3; 9 at month 6.
Same allocation ambiguity. Enrolled 17; 15 assessed at week 2; 14 at month 3; 13 at month 6.
Same allocation ambiguity. Enrolled 18; 17 assessed at week 2; 16 at month 3 and month 6.
VAS2 reported outcomes
VAS
Visual Analog Scale score points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Local massage | 14 | 6.5 | 4.5 |
| Corticosteroid injection | 15 | 7 | 3 |
| Extracorporeal shock-wave therapy | 17 | 7 | 5 |
Values are group summaries identified in the caption; SD denotes standard deviation. CI is a confidence interval; IQR denotes the reported quartiles. A dash means no value was extracted for that cell.
At 2 weeks, the reported Visual Analog Scale medians (IQR) were Local massage: 4.5 (IQR 3.25), p=0.007 versus baseline; Corticosteroid injection: 3 (IQR 1.00), p=0.003 versus baseline; Extracorporeal shock-wave therapy: 5 (IQR 2.50), p<0.001 versus baseline. These p-values are within-group comparisons against baseline; the article provides no between-group effect estimate for this visit.
Analysis details and source
Table II reports medians and IQR. The IQR is retained in the interpretation because only its width is reported; quartile bounds are not reconstructed. The comparison p-value is within group versus baseline, not a treatment contrast. Massage n is 14, the number the Methods text gives for week 2; Figure 1 shows 16 before the 2 massage patients who switched treatment within the first 2 weeks were removed. The Results text prints p=0.03 for the injection group's week-2 decrease, whereas Table II and Figure 2 print 0.003; the Table II value is kept.
VAS
Visual Analog Scale score points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Local massage | 10 | 6.5 | 5 |
| Corticosteroid injection | 14 | 7 | 5 |
| Extracorporeal shock-wave therapy | 16 | 7 | 5 |
Values are group summaries identified in the caption; SD denotes standard deviation. CI is a confidence interval; IQR denotes the reported quartiles. A dash means no value was extracted for that cell.
At 3 months, the reported Visual Analog Scale medians (IQR) were Local massage: 5 (IQR 1.50), p=0.006 versus baseline; Corticosteroid injection: 5 (IQR 3.25), p=0.009 versus baseline; Extracorporeal shock-wave therapy: 5 (IQR 2.75), p=0.001 versus baseline. These p-values are within-group comparisons against baseline; the article provides no between-group effect estimate for this visit.
Analysis details and source
Table II reports medians and IQR. The IQR is retained in the interpretation because only its width is reported; quartile bounds are not reconstructed. The comparison p-value is within group versus baseline, not a treatment contrast.
VAS1 reported outcome
VAS
VAS score points
Table II reports 6-month VAS medians (IQR): massage 5.00 (3.00), corticosteroid injection 5.00 (3.50), and ESWT 4.00 (3.50); p versus baseline was 0.140, 0.074, and 0.001, respectively. Figure 2 instead shows ESWT VAS 5.00 at month 6. Both source locations report p=0.001 for ESWT. This is a within-group comparison, not a between-treatment effect.
Analysis details and source
Median/IQR values are from Table II; Figure 2 conflicts on ESWT month-6 median (5.00 vs 4.00). p-values are within group versus baseline.
DASH3 reported outcomes
DASH
Disabilities of the Arm, Shoulder and Hand (DASH) score points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Local massage | 14 | 29.58 | 20.83 |
| Corticosteroid injection | 15 | 50 | 13.33 |
| Extracorporeal shock-wave therapy | 17 | 49.16 | 25 |
Values are group summaries identified in the caption; SD denotes standard deviation. CI is a confidence interval; IQR denotes the reported quartiles. A dash means no value was extracted for that cell.
At 2 weeks, the reported Disabilities of the Arm, Shoulder and Hand (DASH) medians (IQR) were Local massage: 20.83 (IQR 23.13), p=0.004 versus baseline; Corticosteroid injection: 13.33 (IQR 18.70), p=0.002 versus baseline; Extracorporeal shock-wave therapy: 25 (IQR 25.25), p<0.001 versus baseline. These p-values are within-group comparisons against baseline; the article provides no between-group effect estimate for this visit.
Analysis details and source
Table II reports medians and IQR. The IQR is retained in the interpretation because only its width is reported; quartile bounds are not reconstructed. The comparison p-value is within group versus baseline, not a treatment contrast. Massage n is 14, the number the Methods text gives for week 2; Figure 1 shows 16 before the 2 massage patients who switched treatment within the first 2 weeks were removed. The Results text prints p=0.04 (massage) and p=0.02 (injection) for the week-2 decreases, whereas Table II and Figure 3 print 0.004 and 0.002; the Table II values are kept.
DASH
Disabilities of the Arm, Shoulder and Hand (DASH) score points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Local massage | 10 | 29.58 | 18.1 |
| Corticosteroid injection | 14 | 50 | 25 |
| Extracorporeal shock-wave therapy | 16 | 49.16 | 22 |
Values are group summaries identified in the caption; SD denotes standard deviation. CI is a confidence interval; IQR denotes the reported quartiles. A dash means no value was extracted for that cell.
At 3 months, the reported Disabilities of the Arm, Shoulder and Hand (DASH) medians (IQR) were Local massage: 18.1 (IQR 14.58), p=0.001 versus baseline; Corticosteroid injection: 25 (IQR 18.00), p=0.019 versus baseline; Extracorporeal shock-wave therapy: 22 (IQR 39.37), p=0.001 versus baseline. These p-values are within-group comparisons against baseline; the article provides no between-group effect estimate for this visit.
Analysis details and source
Table II reports medians and IQR. The IQR is retained in the interpretation because only its width is reported; quartile bounds are not reconstructed. The comparison p-value is within group versus baseline, not a treatment contrast.
DASH
Disabilities of the Arm, Shoulder and Hand (DASH) score points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Local massage | 9 | 29.58 | 18.1 |
| Corticosteroid injection | 13 | 50 | 25 |
| Extracorporeal shock-wave therapy | 16 | 49.16 | 17.5 |
Values are group summaries identified in the caption; SD denotes standard deviation. CI is a confidence interval; IQR denotes the reported quartiles. A dash means no value was extracted for that cell.
At 6 months, the reported Disabilities of the Arm, Shoulder and Hand (DASH) medians (IQR) were Local massage: 18.1 (IQR 10.95), p=0.072 versus baseline; Corticosteroid injection: 25 (IQR 20.45), p=0.353 versus baseline; Extracorporeal shock-wave therapy: 17.5 (IQR 39.63), p=0.001 versus baseline. These p-values are within-group comparisons against baseline; the article provides no between-group effect estimate for this visit.
Analysis details and source
Table II reports medians and IQR. The IQR is retained in the interpretation because only its width is reported; quartile bounds are not reconstructed. The comparison p-value is within group versus baseline, not a treatment contrast.
DASH-WM3 reported outcomes
DASH-WM
DASH-Work Model (DASH-WM) score points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Local massage | 14 | 50 | 25 |
| Corticosteroid injection | 15 | 68.75 | 25 |
| Extracorporeal shock-wave therapy | 17 | 75 | 37.5 |
Values are group summaries identified in the caption; SD denotes standard deviation. CI is a confidence interval; IQR denotes the reported quartiles. A dash means no value was extracted for that cell.
At 2 weeks, the reported DASH-Work Model (DASH-WM) medians (IQR) were Local massage: 25 (IQR 34.38), p=0.017 versus baseline; Corticosteroid injection: 25 (IQR 25.00), p=0.004 versus baseline; Extracorporeal shock-wave therapy: 37.5 (IQR 25.00), p=0.001 versus baseline. These p-values are within-group comparisons against baseline; the article provides no between-group effect estimate for this visit.
Analysis details and source
Table II reports medians and IQR. The IQR is retained in the interpretation because only its width is reported; quartile bounds are not reconstructed. The comparison p-value is within group versus baseline, not a treatment contrast. Massage n is 14, the number the Methods text gives for week 2; Figure 1 shows 16 before the 2 massage patients who switched treatment within the first 2 weeks were removed. The Results text prints p=0.04 for the injection group's week-2 decrease, whereas Table II and Figure 4 print 0.004; the Table II value is kept.
DASH-WM
DASH-Work Model (DASH-WM) score points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Local massage | 10 | 50 | 25 |
| Corticosteroid injection | 14 | 68.75 | 50 |
| Extracorporeal shock-wave therapy | 16 | 75 | 43.75 |
Values are group summaries identified in the caption; SD denotes standard deviation. CI is a confidence interval; IQR denotes the reported quartiles. A dash means no value was extracted for that cell.
At 3 months, the reported DASH-Work Model (DASH-WM) medians (IQR) were Local massage: 25 (IQR 25.25), p=0.07 versus baseline; Corticosteroid injection: 50 (IQR 25.00), p=0.05 versus baseline; Extracorporeal shock-wave therapy: 43.75 (IQR 43.45), p=0.001 versus baseline. These p-values are within-group comparisons against baseline; the article provides no between-group effect estimate for this visit.
Analysis details and source
Table II reports medians and IQR. The IQR is retained in the interpretation because only its width is reported; quartile bounds are not reconstructed. The comparison p-value is within group versus baseline, not a treatment contrast.
DASH-WM
DASH-Work Model (DASH-WM) score points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Local massage | 9 | 50 | 25 |
| Corticosteroid injection | 13 | 68.75 | 50 |
| Extracorporeal shock-wave therapy | 16 | 75 | 25 |
Values are group summaries identified in the caption; SD denotes standard deviation. CI is a confidence interval; IQR denotes the reported quartiles. A dash means no value was extracted for that cell.
At 6 months, the reported DASH-Work Model (DASH-WM) medians (IQR) were Local massage: 25 (IQR 31.25), p=0.452 versus baseline; Corticosteroid injection: 50 (IQR 25.00), p=0.176 versus baseline; Extracorporeal shock-wave therapy: 25 (IQR 56.25), p=0.001 versus baseline. These p-values are within-group comparisons against baseline; the article provides no between-group effect estimate for this visit.
Analysis details and source
Table II reports medians and IQR. The IQR is retained in the interpretation because only its width is reported; quartile bounds are not reconstructed. The comparison p-value is within group versus baseline, not a treatment contrast.
VAS, DASH, and DASH-WM in the additional-treatment subgroup1 reported outcome
VAS, DASH, and DASH-WM in the additional-treatment subgroup
VAS; DASH; DASH-WM score points
Figure 6 reports the 10 participants removed from the primary groups after switching treatment. Scores (baseline; week 2; month 3; month 6; the figure does not say whether they are medians or means) and p-values versus baseline: VAS 7.00; 6.50; 4.00; 3.00 (p=.317; .003; .008); DASH 47.03; 34.95; 19.55; 12.66 (p=.182; .052; .049); DASH-WM 75.00; 50.00; 25.00; 12.50 (p=.396; .071; .009). These patients received heterogeneous subsequent regimens, so the subgroup change cannot estimate any particular treatment effect.
Analysis details and source
Figure 6 values are descriptive subgroup scores; the figure does not state whether they are medians or means. Patients switched between different treatments and combinations. No randomized comparison is available.
How treatment was delivered
Age >18; symptoms >6 weeks; diagnosis by examination and MRI; no treatment in previous six months. Exclusion criteria not reported.
Abstract labels the study randomized, but Methods say group assignment followed order of presentation; sequence generation and concealment not described.
Deep transverse fingertip/skin massage over maximal tenderness, firm pressure at ECRB tendon origin, 5 min once daily, three times weekly for two weeks.
Single injection of 1 mL 40 mg methylprednisolone acetate plus 2 mL 2% prilocaine hydrochloride; needle tip touched bone at tender lateral epicondyle.
Device and focused versus radial modality are not reported.
1,500 shock waves; 2.0 bar; 8 Hz; four sessions, two per week for two weeks; each session 20–30 min.
Shoulder 30° abducted, elbow 90° flexed, forearm/wrist/hand on table; maximal tenderness marked; coupling gel used; no local anesthesia.
All groups received nonsteroidal anti-inflammatory drugs, splinting, exercise, and rest; details and adherence by arm are not reported.
VAS, DASH, and DASH-WM at baseline, 2 weeks, 3 months, and 6 months.
Shapiro–Wilk; ANOVA/Tukey for normally distributed data; Mann–Whitney, Wilcoxon, repeated-measures ANOVA, Kruskal–Wallis and Bonferroni for non-normal data; chi-square for categorical data; p<0.05; 95% CI specified but comparative CIs not reported.
The report does not provide systematic adverse-event counts or arm-specific safety monitoring. Two participants had switched to additional treatment by week 2, six more by month 3 and two more by month 6; the 10 switched participants were followed as a separate subgroup. Switching is not classified as an adverse event. The corticosteroid risks mentioned in Discussion come from cited literature.
Extractor assessment: treatment allocation is inadequately described and conflicts with the abstract claim of randomization; assignment by presentation order risks selection bias. No allocation concealment or blinding is described.
The main groups were small (17, 17, and 18 initially), follow-up denominators declined and differed by arm, and patients receiving additional treatment were excluded from main-group comparisons.
The paper reports within-group p-values but no direct comparative treatment estimates, confidence intervals, or clear prespecified primary outcome. Its conclusion of ESWT superiority therefore exceeds what these analyses establish.
Authors identify the small number of participants who received combined therapy as a limitation. The ESWT device and focused/radial modality are not stated, limiting reproducibility.
Adverse events, adherence, and co-intervention use by arm are not reported systematically. No supplementary material was identified.
Source inconsistency: ESWT month-6 VAS is 4.00 (IQR 3.50) in Table II but 5.00 in the table beneath Figure 2; both show p=.001 versus baseline. Figure 4 narrative text also says ESWT DASH-WM rose from 35.50 to 43.75 between week 2 and month 3, whereas Table II and Figure 4 table give 37.50 to 43.75. The extracted principal-arm sequence follows the outcome tables while recording the discrepancy.
Findings checked against the full source by a separate automated reviewer.
Limitations noted by the reviewers
- Extractor assessment: treatment allocation is inadequately described and conflicts with the abstract claim of randomization; assignment by presentation order risks selection bias. No allocation concealment or blinding is described.
- The main groups were small (17, 17, and 18 initially), follow-up denominators declined and differed by arm, and patients receiving additional treatment were excluded from main-group comparisons.
- The paper reports within-group p-values but no direct comparative treatment estimates, confidence intervals, or clear prespecified primary outcome. Its conclusion of ESWT superiority therefore exceeds what these analyses establish.
- Authors identify the small number of participants who received combined therapy as a limitation. The ESWT device and focused/radial modality are not stated, limiting reproducibility.
- Adverse events, adherence, and co-intervention use by arm are not reported systematically. No supplementary material was identified.
- Source inconsistency: ESWT month-6 VAS is 4.00 (IQR 3.50) in Table II but 5.00 in the table beneath Figure 2; both show p=.001 versus baseline. Figure 4 narrative text also says ESWT DASH-WM rose from 35.50 to 43.75 between week 2 and month 3, whereas Table II and Figure 4 table give 37.50 to 43.75. The extracted principal-arm sequence follows the outcome tables while recording the discrepancy.
Source-linked critical reading
Randomized comparative clinical trial
Technical source record
Complete original publisher PDF (10 pages) reviewed from abstract through references, including all methods, results, Tables I–II, Figures 1–6 and captions, discussion, conclusion, author contributions, funding and conflicts. Numeric outcomes and figure/table discrepancy checked against page images/text extraction.
Pain and function scores improved early in all three groups, while the shock-wave group retained more within-group improvement at six months. Because treatment assignment followed arrival order and follow-up was incomplete, the study cannot establish that shock waves caused better outcomes than massage or injection.
Among adults with lateral epicondylitis, how do local massage, corticosteroid injection, and ESWT compare over six months?
Prospective three-group comparative study. Although the abstract calls it randomized, Methods state that patients were assigned in order of presentation; no random sequence or concealment is described. The study therefore should not be interpreted as randomized.
Adults over 18 with symptoms longer than six weeks, examination and MRI findings consistent with lateral epicondylitis, and no treatment in the prior six months; 52 were enrolled.
Massage was applied over maximal tenderness at the ECRB tendon origin for 5 minutes, three times weekly for two weeks. The injection group received one injection of 1 mL methylprednisolone acetate 40 mg plus 2 mL prilocaine 2%. ESWT was delivered at 1,500 shocks, 2.0 bar and 8 Hz, four sessions (twice weekly for two weeks), without local anesthesia. All groups also received NSAIDs, splinting, exercise and rest.
The three active treatment strategies were compared with shared co-care; there was no sham or untreated comparator. Because assignment followed presentation order, treatment contrasts are vulnerable to confounding.
VAS, DASH and DASH-WM were assessed at baseline, two weeks, three months and six months. Table II reports medians with IQR and p-values versus baseline. The paper does not establish the native VAS range in the Methods.
All three groups improved on several measures at two weeks. Later within-group significance was reported more consistently for ESWT, but these paired comparisons do not establish comparative superiority. Outcome denominators declined over time, particularly in massage and injection groups, and ten participants who changed treatment were analyzed separately.
Harms were not systematically enumerated by group. The study excluded ten participants from the main groups after treatment changes, but it does not provide a clear adverse-event count; absence of reporting is not evidence of zero harms.
The allocation method is nonrandom despite the abstract wording. Small group sizes, differential attrition, treatment switching, shared co-interventions, no sham control, and subjective outcomes limit causal interpretation. The paper reports no adjustment for confounding. It also contains a VAS month-six discrepancy: Table II gives 4.00 for ESWT, whereas Figure 2 gives 5.00.
The findings describe this single center’s patients with symptoms lasting at least six weeks. They do not establish that ESWT is superior in a randomized comparison, and the shared co-interventions prevent attributing all changes to the assigned procedure.
A sufficiently powered randomized trial should use concealed allocation, blinded outcome assessment, balanced co-care, prespecified time-specific between-group estimates, complete follow-up accounting and prospective adverse-event recording.
Read the original article.
Comparison of local massage, steroid injection, and extracorporeal shock wave therapy efficacy in the treatment of lateral epicondylitis.
Çetin BV, Sepetçi Ö, Yazar İ, Kaptan AY, Orhan Ö, Demir M, Demir M, Altay MA. · Joint diseases and related surgery · 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 →