Insertional Achilles tendinopathy · Randomized trial · 2008
Does shockwave help insertional Achilles pain more than heel-drop exercises?
In this trial, three weekly radial shockwave sessions did better than a 12-week heel-drop exercise programme at four months. Pain with loading fell from about 7 to 3 out of 10 with shockwave, and from about 7 to 5 with exercise. 16 of 25 people felt recovered or much better, vs 7 of 25.
Helped more than heel-drop exercises- adults with insertional Achilles pain
- 50
- sessions, one a week
- 3
- followed up
- 4 mo
Good to know: After four months many switched treatment, so later results do not compare the groups.
Rompe JD, Furia J, Maffulli N The Journal of Bone and Joint Surgery. American Volume, 2008 · Read the paper ↗
How this study worked
What happened
Achilles pain when loading the tendon
0–10 pain scale · start and after four months
StartAfter 4 months
Show the numbers
| Group | People | Start | After 4 months |
|---|---|---|---|
| Radial shockwave | 25 | 7 | 3 |
| Heel-drop exercises | 25 | 6.8 | 5 |
The treatment
Radial shockwave
Radial, once a week for 3 weeks
- Sessions
- 3, one a week
- Pulses
- 2,000 per session
- Strength
- 2.5bar the paper equates this with 0.12 mJ/mm² for this device · 8 per second
- Numbing
- None
- Device
- EMS Swiss DolorClast
- Everyone also got
- Advice to avoid painful activity; paracetamol or naproxen allowed
Compared with: A 12-week home programme of eccentric calf exercises (slow heel drops), twice a day
Settings are what each study used, not a prescription.
Who took part
50 adults aged 18–70 at a German orthopaedic clinic with insertional Achilles pain for at least six months after other treatments, including injections, had not helped.
Side effects
Brief skin reddening in everyone treated with shockwave, without bruising; calf aching with the exercises; no tendon ruptures through 15 months.
All the numbers and the full appraisal
The results, in numbers.
Fifty adults aged 18–70 at a German orthopaedic clinic with insertional Achilles tendinopathy for at least six months after failed care that included injections, anti-inflammatory drugs and physiotherapy. Mean age 39 and 40 years; 14 and 16 women; about 60% athletic. Haglund deformity, bursitis and midportion disease were excluded by radiographs and ultrasound. Groups had 30 and 31 affected feet.
All 25 analysed at four months by intention to treat; three missing patients (two stopped for persistent pain, one gave telephone report only) had baseline values carried forward. 30 affected feet; how bilateral cases were scored is not stated.
All 25 analysed at four months; two did not attend but reported by telephone being pain-free, and their baseline values were carried forward. 31 affected feet.
VISA-A score (primary)1 reported outcome
VISA-A score (primary)
VISA-A Points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Eccentric loading | 25 | 52.7 (SD 8.4) | 63.4 (SD 12) (range 28–100) |
| Radial shockwaves | 25 | 53.2 (SD 5.8) | 79.4 (SD 10.4) (range 34–100) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Radial shockwaves versus eccentric loading
Wilcoxon test on improvement from baseline
VISA-A improved in both groups and more after shockwaves (53.2 to 79.4) than after eccentric loading (52.7 to 63.4); p=0.005 for the primary comparison.
Analysis details and source
Means ± SD with ranges; higher is better (100 = asymptomatic). Table II gives a between-group difference of 16.2 with an interval labelled "Range" (5.3 to 27.2); it is not identified as a confidence interval.
Completely recovered or much improved3 reported outcomes
Completely recovered or much improved
6-point Likert scale Patients
Radial shockwaves versus eccentric loading
Not named
Success (Likert 1 or 2) was reported by 7 of 25 (28%) after eccentric loading and 16 of 25 (64%) after shockwaves.
Analysis details and source
Denominators are all randomized patients; missing patients carried their baseline rating. The text gives p<0.020; the Methods name the Wilcoxon test for ≥50% pain improvement, which is not reported separately.
General assessment, 6-point Likert
6-point Likert scale Points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Eccentric loading | 25 | 5.4 (SD 0.6) | 3.7 (SD 1.5) (range 1–6) |
| Radial shockwaves | 25 | 4.9 (SD 0.9) | 2.8 (SD 1.6) (range 1–6) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Radial shockwaves versus eccentric loading
Not named
Mean ratings improved more after shockwaves (4.9 to 2.8) than after eccentric loading (5.4 to 3.7), p=0.043; the shockwave group started 0.5 points better.
Analysis details and source
1 = completely recovered; 1–2 counted as success and 3–6 as failure; the paper does not define the other anchors. Table II gives a between-group difference of −0.9 with an interval labelled "Range" (−1.8 to 0.03); it is not identified as a confidence interval, and it crosses zero although p<0.05, a contradiction left unresolved.
Fifteen-month course after crossover
6-point Likert scale Patients
All 18 eccentric-loading failures crossed to shockwaves (13 later successful); 8 of 9 shockwave failures crossed to eccentric loading (2 successful) and one chose surgery. All 7 early eccentric successes and 15 of 16 shockwave successes remained successful at 15 months.
Analysis details and source
Not a randomized comparison: groups were rebuilt by outcome and crossover. Several patients were contacted only by telephone and eight had last values carried forward (Figure 2). The abstract calls this "one year after completion of the initial treatment"; the Methods say 15 months from baseline.
Load-induced pain (NRS 0–10)1 reported outcome
Load-induced pain (NRS 0–10)
Numeric rating scale, load-induced pain Points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Eccentric loading | 25 | 6.8 (SD 1) | 5 (SD 2.3) (range 1–8) |
| Radial shockwaves | 25 | 7 (SD 0.8) | 3 (SD 2.3) (range 0–8) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Radial shockwaves versus eccentric loading
Wilcoxon test
Pain fell from 6.8 to 5.0 with eccentric loading and from 7.0 to 3.0 with shockwaves (p=0.004).
Analysis details and source
0 = no pain, 10 = very severe pain. Table II gives a between-group difference of −2.0 with an interval labelled "Range" (−3.3 to 0.7); it is not identified as a confidence interval, and it crosses zero although p<0.05, a contradiction left unresolved.
Pressure pain threshold1 reported outcome
Pressure pain threshold
Algometer (Pain Test Model FPK) kg
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Eccentric loading | 25 | 1.4 (SD 0.7) | 2.2 (SD 1.6) (range 1.6–4.5) |
| Radial shockwaves | 25 | 1.6 (SD 0.8) | 3.5 (SD 1.1) (range 1.5–5.2) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Radial shockwaves versus eccentric loading
Not named
The pressure needed to provoke pain rose more after shockwaves (1.6 to 3.5 kg) than after eccentric loading (1.4 to 2.2 kg), p=0.002.
Analysis details and source
Higher is better. Table II gives a between-group difference of 1.3 with an interval labelled "Range" (0.5 to 2.1); it is not identified as a confidence interval.
Tenderness at 3 kg (NRS 0–10)1 reported outcome
Tenderness at 3 kg (NRS 0–10)
Numeric rating scale at 3-kg pressure Points
| Group | n | Baseline | Follow-up |
|---|---|---|---|
| Eccentric loading | 25 | 6.2 (SD 3.7) | 4.4 (SD 3.2) (range 1–8) |
| Radial shockwaves | 25 | 6.5 (SD 3.4) | 2.4 (SD 4.2) (range 1–7) |
SD: standard deviation. CI: confidence interval. IQR: interquartile range. —: not extracted.
Radial shockwaves versus eccentric loading
Not named
Tenderness fell from 6.2 to 4.4 with eccentric loading and from 6.5 to 2.4 with shockwaves (p=0.021).
Analysis details and source
Table II gives a between-group difference of −2.0 with an interval labelled "Range" (−3.7 to 0.3); it is not identified as a confidence interval, and it crosses zero although p<0.05, a contradiction left unresolved. An SD of 4.2 with a range of 1–7 is also hard to reconcile.
Used paracetamol or naproxen1 reported outcome
Used paracetamol or naproxen
Patient diary Patients
Radial shockwaves versus eccentric loading
Not named
Nineteen of 25 after eccentric loading and 7 of 25 after shockwaves reported using rescue analgesics (p<0.001).
Analysis details and source
The text attributes this to Table II, which does not contain it; denominators assumed to be all randomized patients as the text gives no other.
How treatment was delivered
Radial device (EMS Swiss DolorClast) applied by the senior author: three sessions one week apart, 2,000 pulses at 2.5 bar and 8 Hz, without local anaesthesia, starting at the point of maximal tenderness and moving outward in a circular pattern (clinical focusing). The paper equates 2.5 bar with 0.12 mJ/mm² for this device; that is not a general pressure-energy conversion.
Twelve-week home eccentric calf program taught by the senior author: three sets of 15 slow lowering repetitions with knee straight and bent, twice daily, seven days a week, heel lowered below step level into full dorsiflexion, progressing with 5-kg backpack increments when pain-free; a medical assistant checked technique at six weeks.
Both groups avoided pain-provoking activity for 12 weeks (walking and cycling allowed; light jogging after 4–6 weeks if painless). Other treatments were discouraged; paracetamol 2–4 g or naproxen 1 g daily was allowed. Crossover was offered after four months to those not recovered.
Computer-generated permuted blocks in opaque sealed envelopes held by a medical assistant unaware of block size. Treatment was not blinded; follow-up assessment is called observer-blinded, but the authors concede the observer may have known the treatment.
All shockwave patients had transient skin reddening, without bruising or device-related complications; eccentric-loading patients reported calf aching. No Achilles rupture occurred through 15 months, and no side effects were reported at that visit. Events were recorded on standardized forms but not tabulated by severity. The authors report no outside funding and no commercial benefits.
Not blinded: the senior author delivered every shockwave session and taught every exercise program; follow-up is described as observer-blinded, but the authors concede the observer may have known the treatment.
Table II intervals are labelled "Range" and three of them (Likert, load-induced pain, tenderness) cross zero despite p<0.05, so precision cannot be judged from them. Statistical tests are not named outcome by outcome, and the sample size was based on a 3-point NRS pain difference although VISA-A was the primary endpoint.
Five patients with missing four-month data had baseline values carried forward; two eccentric-loading patients stopped because of persistent pain. Figure 1 also shows missed six-week visits (2 and 1) that the text does not mention. Bilateral disease (30 and 31 feet in 25 patients each) is not addressed in the analysis.
Crossover after four months (18 of 25 and 8 of 25) means the 15-month data do not keep the randomized comparison. Exercise adherence was self-reported (21 of 25 said they exercised twice daily). The comparator was one specific program loading the heel below step level; other insertional loading programs were not tested (extractor interpretation). Single centre, 50 patients.
The online appendix with VISA-A subscale scores and the published comment on this trial (PMID 19451774) were not read.
Findings checked against the full source by a separate automated reviewer.
Limitations noted by the reviewers
- Not blinded: the senior author delivered every shockwave session and taught every exercise program; follow-up is described as observer-blinded, but the authors concede the observer may have known the treatment.
- Table II intervals are labelled "Range" and three of them (Likert, load-induced pain, tenderness) cross zero despite p<0.05, so precision cannot be judged from them. Statistical tests are not named outcome by outcome, and the sample size was based on a 3-point NRS pain difference although VISA-A was the primary endpoint.
- Five patients with missing four-month data had baseline values carried forward; two eccentric-loading patients stopped because of persistent pain. Figure 1 also shows missed six-week visits (2 and 1) that the text does not mention. Bilateral disease (30 and 31 feet in 25 patients each) is not addressed in the analysis.
- Crossover after four months (18 of 25 and 8 of 25) means the 15-month data do not keep the randomized comparison. Exercise adherence was self-reported (21 of 25 said they exercised twice daily). The comparator was one specific program loading the heel below step level; other insertional loading programs were not tested (extractor interpretation). Single centre, 50 patients.
- The online appendix with VISA-A subscale scores and the published comment on this trial (PMID 19451774) were not read.
Source-linked critical reading
Randomized comparative clinical trial
Technical source record
Full-text appraisal from the complete publisher PDF (all 11 pages: methods, results, discussion, Tables I–II, Figures 1–2, disclosure). The online appendix with VISA-A subscales and the published comment (PMID 19451774) were not read.
In this unblinded trial of insertional Achilles pain, three weekly radial shockwave sessions gave better pain, function and recovery ratings at four months than a 12-week eccentric exercise program. Most patients without success then switched treatment, so later results do not compare the original groups.
In chronic insertional Achilles tendinopathy that has not responded to conservative care, does a 12-week eccentric loading program or three sessions of low-energy radial shockwaves lead to better function, pain and recovery at four months?
Single-centre randomized trial in Germany. Of 73 patients screened, 50 were randomized (25 per group) by computer-generated permuted blocks in sealed envelopes. Treatment was not blinded. The primary endpoint was change in VISA-A at four months, analysed by intention to treat with last (baseline) values carried forward for five patients. Crossover was allowed afterwards; a final review took place at 15 months.
Fifty adults aged 18–70 at a German orthopaedic clinic with insertional Achilles tendinopathy for at least six months after failed care that included injections, anti-inflammatory drugs and physiotherapy. Mean age 39 and 40 years; 14 and 16 women; about 60% athletic. Haglund deformity, bursitis and midportion disease were excluded by radiographs and ultrasound. Groups had 30 and 31 affected feet.
Radial device (EMS Swiss DolorClast) applied by the senior author: three sessions one week apart, 2,000 pulses at 2.5 bar and 8 Hz, without local anaesthesia, starting at the point of maximal tenderness and moving outward in a circular pattern (clinical focusing). The paper equates 2.5 bar with 0.12 mJ/mm² for this device; that is not a general pressure-energy conversion.
Twelve-week home eccentric calf program taught by the senior author: three sets of 15 slow lowering repetitions with knee straight and bent, twice daily, seven days a week, heel lowered below step level into full dorsiflexion, progressing with 5-kg backpack increments when pain-free; a medical assistant checked technique at six weeks. Paracetamol or naproxen was allowed in both groups.
VISA-A (primary), a 6-point Likert rating (1–2 counted as success), load-induced pain on a 0–10 scale, algometer pain threshold and tenderness at 3 kg, at four months from baseline; analgesic use from diaries; then a 15-month review after crossover.
At four months VISA-A rose from 52.7 to 63.4 with eccentric loading and from 53.2 to 79.4 with shockwaves (p=0.005). Success was 7/25 (28%) versus 16/25 (64%) (p<0.020). Load-induced pain fell from 6.8 to 5.0 versus 7.0 to 3.0 (p=0.004); threshold and tenderness also favoured shockwaves (p=0.002 and 0.021). Rescue analgesics: 19 versus 7 patients (p<0.001). Reported between-group intervals are labelled ranges, and three cross zero despite significant p-values.
All shockwave patients had transient skin reddening, without bruising or device-related complications; eccentric-loading patients reported calf aching. No Achilles rupture occurred through 15 months, and no side effects were reported at that visit. Events were recorded on standardized forms but not tabulated by severity. The authors report no outside funding and no commercial benefits.
Unblinded treatment delivered and taught by the senior author, with an assessor who may have known the allocation; single-centre sample; baseline values carried forward for five patients; bilateral feet not addressed; intervals in Table II that cross zero; sample size based on pain rather than the primary VISA-A; self-reported exercise adherence; and crossover that removes the randomized contrast after four months.
Applies to chronic insertional Achilles tendinopathy without Haglund deformity or bursitis after failed conservative care, compared with this particular eccentric program. It does not apply to midportion tendinopathy, does not test shockwaves against sham, and does not show long-term randomized differences.
A larger, assessor-blinded, multicentre trial should compare shockwaves with sham and with modern insertional loading programs, report confidence intervals, handle bilateral disease, keep patients in their randomized groups for long-term analysis and record harms systematically.
Read the original article.
Eccentric loading compared with shock wave treatment for chronic insertional achilles tendinopathy. A randomized, controlled trial.
Rompe JD, Furia J, Maffulli N. · The Journal of bone and joint surgery. American volume · 2008
Open the publication record or available full-text files.
Checked against the full paper · Clinical review: Dr. Marcus Yu Bin Pai, · How we review →