Rotator-cuff / subacromial pain: mixed or unspecified subtype · Randomized trial · 2009
Does radial shockwave help shoulder pain as much as exercise?
In this trial, shoulder pain and disability scores fell from about 45 to about 29 out of 100 over 18 weeks with radial shockwave. Supervised exercise did better, with scores falling from about 49 to about 25, and more people improved clearly with exercise.
Exercise did better through 18 weeks- adults with long-lasting shoulder pain
- 104
- shockwave sessions, one a week
- 4–6
- weeks of follow-up
- 18
Good to know: 104 adults at an Oslo clinic, followed for 18 weeks. No sham group: both groups had an active treatment.
Engebretsen K, Grotle M, Bautz-Holter E, et al. BMJ (Clinical Research Ed.), 2009 · Read the paper ↗
How this study worked
What happened
Shoulder pain and disability
Shoulder Pain and Disability Index, 0–100 · lower is better
Radial shockwaveSupervised exercise
Show the numbers
| Group | Start | Week 6 | Week 12 | Week 18 |
|---|---|---|---|---|
| Radial shockwave | 45.1 | 33.5 | 36.1 | 29.2 |
| Supervised exercise | 48.8 | 25.8 | 27 | 24.5 |
The treatment
Radial shockwave
Radial
- Pulses
- 2,000 per session · over 3–5 tender points; one session a week for 4–6 weeks
- Strength
- 2.5–4.0bar set to each person's tolerance
- Numbing
- None
- Device
- Swiss DolorClast
Compared with: Supervised exercise: two 45-minute sessions a week for up to 12 weeks, plus home exercises
Settings are what each study used, not a prescription.
Who took part
104 adults aged 18–70 with shoulder pain for at least 3 months, treated at an outpatient clinic in Oslo.
Side effects
Two people stopped shockwave early because their pain got worse; one switched to exercise. Within 18 weeks, frozen shoulder (adhesive capsulitis) was diagnosed in 2 people in the exercise group and 1 in the shockwave group.
All the numbers and the full appraisal
The results, in numbers.
Adults aged 18–70 with clinically defined subacromial shoulder pain lasting at least 3 months, recruited at an Oslo outpatient rehabilitation clinic. Rotator-cuff rupture was permitted if diagnostic criteria were met.
52 randomized per arm. The mixed model used available repeated observations and assumed data were missing at random; observed denominators vary by outcome/timepoint. At 18 weeks SPADI responder denominators were 50 per arm.
52 randomized per arm. The mixed model used available repeated observations and assumed data were missing at random; observed denominators vary by outcome/timepoint. At 18 weeks SPADI responder denominators were 50 per arm.
Shoulder Pain and Disability Index (SPADI)4 reported outcomes
Shoulder Pain and Disability Index (SPADI)
Shoulder Pain and Disability Index (SPADI) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 48.8 (SD 20.6) | 25.8 (SD 21.5) |
| Radial extracorporeal shockwave treatment | — | 45.1 (SD 22.1) | 33.5 (SD 23.3) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The 18-week treatment effect favored exercise for SPADI, but the point estimate (-8.4) was smaller than the 10-point difference used in sample-size planning. Of the 18-week secondary outcomes, only work status differed significantly (p=0.016, Table 3).
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Shoulder Pain and Disability Index (SPADI)
Shoulder Pain and Disability Index (SPADI) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 48.8 (SD 20.6) | 27 (SD 24.2) |
| Radial extracorporeal shockwave treatment | — | 45.1 (SD 22.1) | 36.1 (SD 28.4) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The 18-week treatment effect favored exercise for SPADI, but the point estimate (-8.4) was smaller than the 10-point difference used in sample-size planning. Of the 18-week secondary outcomes, only work status differed significantly (p=0.016, Table 3).
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Shoulder Pain and Disability Index (SPADI)
Shoulder Pain and Disability Index (SPADI) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 48.8 (SD 20.6) | 24.5 (SD 25.6) |
| Radial extracorporeal shockwave treatment | — | 45.1 (SD 22.1) | 29.2 (SD 25.9) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The 18-week treatment effect favored exercise for SPADI, but the point estimate (-8.4) was smaller than the 10-point difference used in sample-size planning. Of the 18-week secondary outcomes, only work status differed significantly (p=0.016, Table 3).
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Participants exceeding the 19.6-point SPADI smallest detectable difference
Shoulder Pain and Disability Index (SPADI) participants exceeding threshold
Authors classified improvement as a reduction exceeding the individual smallest detectable difference of 19.6 SPADI points.
Odds ratio as reported; no adjustment detail reported in this paragraph.
Logarithmic ratio scale.
More exercise participants exceeded the study threshold; this threshold is the smallest detectable difference, not a responder MCID.
Analysis details and source
Source reports 32/50 vs 18/50, OR 3.2 (95% CI 1.3 to 7.8), p=0.009.
Pain at rest (previous week; Likert-type scale)3 reported outcomes
Pain at rest (previous week; Likert-type scale)
Pain at rest (previous week; Likert-type scale) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 3.4 (SD 1.9) | 2.6 (SD 1.9) |
| Radial extracorporeal shockwave treatment | — | 3.5 (SD 2.1) | 2.9 (SD 2.1) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The reported overall test at 18 weeks was nonsignificant; the confidence interval includes zero.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Pain at rest (previous week; Likert-type scale)
Pain at rest (previous week; Likert-type scale) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 3.4 (SD 1.9) | 2.5 (SD 1.8) |
| Radial extracorporeal shockwave treatment | — | 3.5 (SD 2.1) | 2.9 (SD 2.1) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The reported overall test at 18 weeks was nonsignificant; the confidence interval includes zero.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Pain at rest (previous week; Likert-type scale)
Pain at rest (previous week; Likert-type scale) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 3.4 (SD 1.9) | 2.5 (SD 1.9) |
| Radial extracorporeal shockwave treatment | — | 3.5 (SD 2.1) | 2.7 (SD 2) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The reported overall test at 18 weeks was nonsignificant; the confidence interval includes zero.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Pain during activity (previous week; Likert-type scale)3 reported outcomes
Pain during activity (previous week; Likert-type scale)
Pain during activity (previous week; Likert-type scale) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 5.6 (SD 2) | 3.9 (SD 2) |
| Radial extracorporeal shockwave treatment | — | 5.4 (SD 1.9) | 4.6 (SD 2.4) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The reported overall test at 18 weeks was nonsignificant; the confidence interval includes zero.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Pain during activity (previous week; Likert-type scale)
Pain during activity (previous week; Likert-type scale) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 5.6 (SD 2) | 3.7 (SD 2.2) |
| Radial extracorporeal shockwave treatment | — | 5.4 (SD 1.9) | 4.1 (SD 2.4) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The reported overall test at 18 weeks was nonsignificant; the confidence interval includes zero.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Pain during activity (previous week; Likert-type scale)
Pain during activity (previous week; Likert-type scale) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 5.6 (SD 2) | 3.6 (SD 2.3) |
| Radial extracorporeal shockwave treatment | — | 5.4 (SD 1.9) | 4.1 (SD 2.5) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The reported overall test at 18 weeks was nonsignificant; the confidence interval includes zero.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Carry a 5-kg shopping bag (1 easy–7 impossible)3 reported outcomes
Carry a 5-kg shopping bag (1 easy–7 impossible)
Carry a 5-kg shopping bag (1 easy–7 impossible) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 4.1 (SD 1.8) | 3 (SD 1.8) |
| Radial extracorporeal shockwave treatment | — | 3.6 (SD 2) | 3.1 (SD 2.1) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The reported overall test at 18 weeks was nonsignificant; the confidence interval includes zero.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Carry a 5-kg shopping bag (1 easy–7 impossible)
Carry a 5-kg shopping bag (1 easy–7 impossible) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 4.1 (SD 1.8) | 3 (SD 1.9) |
| Radial extracorporeal shockwave treatment | — | 3.6 (SD 2) | 3.2 (SD 2) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The reported overall test at 18 weeks was nonsignificant; the confidence interval includes zero.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Carry a 5-kg shopping bag (1 easy–7 impossible)
Carry a 5-kg shopping bag (1 easy–7 impossible) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 4.1 (SD 1.8) | 2.8 (SD 1.8) |
| Radial extracorporeal shockwave treatment | — | 3.6 (SD 2) | 3 (SD 2.1) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The reported overall test at 18 weeks was nonsignificant; the confidence interval includes zero.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Take an item down from a wall cupboard (1 easy–7 impossible)3 reported outcomes
Take an item down from a wall cupboard (1 easy–7 impossible)
Take an item down from a wall cupboard (1 easy–7 impossible) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 4.9 (SD 1.2) | 3.4 (SD 1.7) |
| Radial extracorporeal shockwave treatment | — | 4.6 (SD 1.8) | 3.9 (SD 1.9) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
At 6 weeks the printed 95% confidence interval (-1.3 to -0.01) excludes zero, favoring supervised exercise, although the Discussion describes the secondary-outcome differences as not significant except work status; the overall test reported at 18 weeks was nonsignificant.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Take an item down from a wall cupboard (1 easy–7 impossible)
Take an item down from a wall cupboard (1 easy–7 impossible) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 4.9 (SD 1.2) | 3.1 (SD 1.9) |
| Radial extracorporeal shockwave treatment | — | 4.6 (SD 1.8) | 3.5 (SD 2) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The reported overall test at 18 weeks was nonsignificant; the confidence interval includes zero.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Take an item down from a wall cupboard (1 easy–7 impossible)
Take an item down from a wall cupboard (1 easy–7 impossible) scale points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Supervised therapeutic exercise | — | 4.9 (SD 1.2) | 3.2 (SD 1.8) |
| Radial extracorporeal shockwave treatment | — | 4.6 (SD 1.8) | 3.4 (SD 2) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Negative values favor supervised exercise; confidence interval and estimate are reported in Table 2.
Adjusted for baseline in a mixed repeated-measures model; p-values are given only at 18 weeks in Table 2.
The reported overall test at 18 weeks was nonsignificant; the confidence interval includes zero.
Analysis details and source
Model estimates are between-group treatment effects, distinct from within-group changes in the means. The article states ITT, baseline adjustment, and a missing-at-random assumption; 52 participants were randomized per arm, but observed sample counts are not printed per outcome/timepoint.
Working status2 reported outcomes
Working status
Work status questionnaire participants working
Working status did not differ significantly at 12 weeks (p=0.6).
Analysis details and source
Counts are working participants and denominators from Table 3; do not interpret as return-to-work events.
Working status
Work status questionnaire participants working
More participants in exercise group were working at 18 weeks; p=0.016.
Analysis details and source
Counts are working participants and denominators from Table 3; do not interpret as return-to-work events.
Drug treatment for pain, sleep problems, and depression1 reported outcome
Drug treatment for pain, sleep problems, and depression
Self-reported drug treatment use participants using drug treatment
Drug treatment use did not differ significantly at 18 weeks (p=0.26).
Analysis details and source
Includes medication use for pain, sleep problems and depression; Table 3 reports proportions.
Active shoulder range of motion1 reported outcome
Active shoulder range of motion
Active range of motion measurement degrees
The article reports no statistically significant between-group difference in active range of motion at 18 weeks; numerical estimates were not supplied in the text or tables.
Analysis details and source
Qualitative finding only because the numerical data are not available in the acquired article body.
Additional treatment between weeks 12 and 181 reported outcome
Additional treatment between weeks 12 and 18
Additional-treatment report participants receiving additional treatment
Additional treatment received after week 12 and before week 18.
As reported; no adjustment detail is given in the Results paragraph.
Logarithmic ratio scale.
More radial-shockwave participants received additional treatment; this may reflect lower satisfaction but was not a prespecified efficacy endpoint.
Analysis details and source
The article reports counts 13 vs 3, OR 5.5 (95% CI 1.3 to 26.4), p=0.014; arm-specific denominators for this comparison are not stated in the paragraph.
How treatment was delivered
Two supervised 45-minute sessions weekly for up to 12 weeks, with individualized progression and home exercises.
Swiss DolorClast (EMS), radial extracorporeal shockwave treatment.
2,000 pulses/session; frequency printed as 12–8 Hz; pressure 2.5–4.0 bar according to patient tolerance.
One session weekly for 4–6 weeks; 3–5 tender points treated, including supraspinatus insertion, dorsolateral subacromial region and up to 3 rotator-cuff trigger points.
No local anesthetic; pressure was titrated to tolerance.
Active supervised exercise; no sham group. Participants were asked to avoid additional shoulder treatment except analgesics during 18-week follow-up.
One exercise participant reported considerable pain and stiffness consistent with adhesive capsulitis and received only four sessions. Two radial-shockwave participants stopped treatment because pain worsened; one crossed over to exercise. Table 4 also lists a further crossover and surgery; the study provides no sham-controlled harm comparison.
No placebo group was included; the comparison cannot separate modality effects from differences in intervention contact, content or expectations.
The trial may be underpowered for secondary outcomes; its 18-week SPADI treatment effect was 8.4 points, smaller than the 10-point difference used for sample-size planning.
Follow-up was 18 weeks; mixed model assumes data missing at random. Fourteen participants did not return the 6-week questionnaire; additional follow-up loss occurred by weeks 12/18.
From weeks 12 to 18, additional treatment was received by 13 radial-shockwave and 3 exercise participants despite the request to avoid other care.
Findings checked against the full source by a separate automated reviewer.
Limitations noted by the reviewers
- No placebo group was included; the comparison cannot separate modality effects from differences in intervention contact, content or expectations.
- The trial may be underpowered for secondary outcomes; its 18-week SPADI treatment effect was 8.4 points, smaller than the 10-point difference used for sample-size planning.
- Follow-up was 18 weeks; mixed model assumes data missing at random. Fourteen participants did not return the 6-week questionnaire; additional follow-up loss occurred by weeks 12/18.
- From weeks 12 to 18, additional treatment was received by 13 radial-shockwave and 3 exercise participants despite the request to avoid other care.
Source-linked critical reading
Randomized comparative clinical trial
Technical source record
Complete full-text XML, all 18 article sections, Tables 1–4 and Fig. 1–2 captions read. The article describes a mixed model with baseline adjustment and missing-at-random assumption; outcome-specific observed counts vary, and Fig. 1 is the participant-flow source.
In this trial, supervised exercise improved shoulder pain and function more than the tested radial shockwave program over 6–18 weeks. Some people stopped either treatment because symptoms worsened. The comparison does not determine which option suits every person or what focused shockwaves would do.
In adults with chronic subacromial shoulder pain, did radial extracorporeal shockwave treatment (rESWT) produce better short-term outcomes than supervised exercise? The primary measure was the Shoulder Pain and Disability Index (SPADI), where higher scores mean worse pain and disability.
This single-blind randomized study enrolled 104 people aged 18–70 with symptoms for at least three months and specified clinical signs. A statistician generated block randomization stratified by sex; sealed envelopes were opened by a person outside treatment. One participant crossed from rESWT to exercise after one session. The outcome was self-reported, so treatment masking was not feasible, and the study was single-centre.
Eligibility required painful or dysfunctional abduction, preserved passive glenohumeral motion, pain on two of three resisted tests, and a positive Kennedy-Hawkins sign. Rotator-cuff rupture was allowed if criteria were met. This was a clinically defined subacromial-pain group, not a single imaging-confirmed tendon lesion.
The radial arm used a Swiss DolorClast (EMS), once weekly for four to six weeks, treating three to five tender points. Each session delivered 2,000 pulses at 12–8 Hz and 2.5–4.0 bar according to tolerance, without local anesthetic. Participants received a median of five sessions. Exercise was individualized and supervised twice weekly for 45 minutes, for up to 12 weeks, with movement retraining followed by progressively loaded endurance and functional exercises; median attendance was 14 sessions.
The comparison was between two active rehabilitation strategies, not against sham or no treatment. Exercise also involved a longer treatment period and individualized supervision, so the result does not isolate a specific biological effect of shockwaves from differences in therapeutic contact and program content.
SPADI was measured at 6, 12, and 18 weeks. Secondary measures included pain at rest and with activity, two functional tasks, shoulder motion, work status, and medication use. Follow-up was short; the study did not establish long-term recurrence or structural tendon change.
The mixed-model SPADI treatment effect favored supervised exercise at 6 weeks (-10.0, 95% CI -17.6 to -2.3), 12 weeks (-10.3, 95% CI -19.8 to -0.8), and 18 weeks (-8.4, 95% CI -16.5 to -0.6; p=0.047 at 18 weeks). At 18 weeks, 32/50 exercise participants and 18/50 shock-wave participants improved beyond the study’s 19.6-point smallest detectable difference (OR 3.2, 95% CI 1.3 to 7.8; p=0.009). The 18-week mean treatment effect was smaller than the 10-point difference used in sample-size planning.
Pain worsened enough to stop treatment in two rESWT participants; one crossed to exercise. One exercise participant developed increased pain and stiffness consistent with adhesive capsulitis and stopped after four sessions. These few events do not provide a precise comparative safety estimate.
Fourteen participants missed the six-week questionnaire; two exercise participants missed 12- and 18-week follow-up, and two rESWT participants missed 18 weeks. The primary outcome was patient-reported and unblinded to treatment. There was no sham arm, and the exercise program had more treatment sessions and a different duration. The main finding is a modest between-group advantage for exercise, not proof that rESWT is ineffective for every subgroup.
The trial applies to adults with at least three months of clinically defined subacromial pain treated in a Norwegian outpatient rehabilitation service. For this population and these schedules, supervised exercise produced better short-term SPADI results than rESWT. The study does not test focused shockwaves or determine whether a different rESWT dose would change the comparison.
Future trials should use a credible attention-matched comparator, prespecify clinically important SPADI change, retain intention-to-treat follow-up, and report adherence and co-interventions by arm. Longer follow-up should measure function, work participation, further care, and adverse events, while stratifying clinically meaningful diagnoses rather than treating all shoulder pain as one lesion.
Read the original article.
Radial extracorporeal shockwave treatment compared with supervised exercises in patients with subacromial pain syndrome: single blind randomised study.
Engebretsen K, Grotle M, Bautz-Holter E, Sandvik L, Juel NG, Ekeberg OM, Brox JI. · BMJ (Clinical research ed.) · 2009
Open the publication record or available full-text files.
Checked against the full paper · Clinical review: Dr. Marcus Yu Bin Pai, · How we review →