Plantar heel pain / plantar fasciopathy · Randomized trial · 2003
Does focused shockwave help stubborn heel pain more than a placebo?
In this large trial, 34% of people who had three focused shockwave sessions had an excellent or good result at 12 weeks without needing other treatment. With a look-alike placebo procedure, 30% did. Both groups improved.
Similar to placebo- adults with long-lasting heel pain
- 272
- sessions, two weeks apart
- 3
- weeks to the main result
- 12
Good to know: Large trial at 10 centres in Germany. Both groups had a numbing injection, and the placebo looked and sounded real.
Haake M, Buch M, Schoellner C, et al. BMJ (Clinical Research Ed.), 2003 · Read the paper ↗
How this study worked
What happened
Treatment success at 12 weeks
Excellent or good result, with no other treatment needed
Show the numbers
| Group | Count | Share |
|---|---|---|
| Shockwave | 43 / 127 | 33.9% |
| Placebo | 39 / 129 | 30.2% |
The treatment
Shockwave
Focused, every 2 weeks, 3 times
- Sessions
- 3, two weeks apart
- Pulses
- 4,000 per session
- Strength
- 0.08mJ/mm² positive energy; 0.22 total
- Numbing
- Local anaesthetic · both groups
- Device
- Dornier Epos Ultra
Compared with: A placebo procedure that looked and sounded the same, with the shock waves blocked
Settings are what each study used, not a prescription.
Who took part
272 adults with heel pain for at least six months that had not improved with other treatment, at 10 centres in Germany.
hollow squares: no final score
Side effects
Mild side effects such as skin redness, pain or swelling were more common with shockwave: 24 people (18%) vs 12 (9%). None stopped treatment. One blood clot in the placebo group was judged unrelated.
All the numbers and the full appraisal
The results, in numbers.
Adults with chronic plantar fasciitis refractory to conservative treatment for at least six months, enrolled at nine hospitals and one outpatient clinic in Germany; 272 randomized (135 ESWT, 137 placebo). Baseline Table 1 lists 135 therapy and 136 placebo participants; the paper does not explain the one-person denominator discrepancy.
Randomized n=135; 133 received assigned ESWT (one dorsal heel spur, one received placebo), primary 12-week outcome n=127, one-year R&M outcome n=113. Eight lacked the primary endpoint: dorsal spur 1, unintentional injury 1, death of a relative 1, unknown 5.
Randomized n=137; 133 received assigned placebo (one withdrew consent before treatment; three received ESWT), primary 12-week outcome n=129, one-year R&M n=115. Eight lacked the primary endpoint: withdrawal 1, thrombosis 1, other surgery 2, unknown 4. Table 1 lists baseline n=136, unexplained in the text.
Primary success rate1 reported outcome
Primary success rate
Modified Roles and Maudsley score success rule participants meeting success definition
ESWT minus placebo; outcome is Roles and Maudsley score 1 or 2 AND no additional treatment before primary endpoint.
Unadjusted two-sample comparison; centre-stratified robustness analysis also reported for primary outcome.
Success was 43/127 (34%) with ESWT and 39/129 (30%) with placebo. Absolute difference 3.6 percentage points (95% CI -8 to 15.1); p=0.5927. At one year the response includes extensive post-primary-endpoint additional care and is not a clean estimate of the initially assigned treatment alone.
Analysis details and source
The paper also reports OR 1.18 (95% CI 0.675 to 2.07); centre-stratified OR 1.20 (0.674 to 2.13). The prespecified 20-percentage-point clinically relevant difference was not approached.
Roles and Maudsley score 1 or 2 (excellent or good)3 reported outcomes
Roles and Maudsley score 1 or 2 (excellent or good)
Modified Roles and Maudsley score participants with Roles and Maudsley score 1 or 2
At 6 weeks, Roles and Maudsley score 1 or 2 occurred in 28/129 (22%) ESWT versus 33/132 (25%) placebo. Table 2 displays a 3.3-point difference with 95% CI −13.7 to 7.1, but the sign does not align with the displayed arm rates if ordered ESWT minus placebo; we preserve the source inconsistency rather than assign a direction. This is not the primary endpoint.
Analysis details and source
Table 2 does not clarify the sign convention for this interval; 3.3 percentage points is inconsistent with 22% minus 25%.
Roles and Maudsley score 1 or 2 (excellent or good)
Modified Roles and Maudsley score participants with Roles and Maudsley score 1 or 2
ESWT minus placebo; outcome is Roles and Maudsley score 1 or 2, without the additional-treatment criterion used for the primary composite.
Unadjusted two-sample comparison; centre-stratified robustness analysis also reported for primary outcome.
Roles and Maudsley score 1 or 2 was recorded in 58/127 (46%) with ESWT and 52/129 (40%) with placebo. Absolute difference 5.4 percentage points (95% CI -7 to 17.5); no p value was reported. At one year the response includes extensive post-primary-endpoint additional care and is not a clean estimate of the initially assigned treatment alone.
Analysis details and source
For the primary 12-week endpoint the paper also reports odds ratio 1.18 (95% CI 0.675 to 2.07), and centre-stratified OR 1.20 (0.674 to 2.13). No clinically relevant 20-percentage-point difference was observed.
Roles and Maudsley score 1 or 2 (excellent or good)
Modified Roles and Maudsley score participants with Roles and Maudsley score 1 or 2
ESWT minus placebo; outcome is Roles and Maudsley score 1 or 2 at one year, after substantial additional treatment and crossover.
Unadjusted two-sample comparison; centre-stratified robustness analysis also reported for primary outcome.
At one year, Roles and Maudsley score 1 or 2 was recorded in 91/113 (81%) with ESWT and 87/115 (76%) with placebo. Absolute difference 4.9 percentage points (95% CI -6 to 16); this was a secondary outcome after substantial additional care and crossover.
Analysis details and source
For the primary 12-week endpoint the paper also reports odds ratio 1.18 (95% CI 0.675 to 2.07), and centre-stratified OR 1.20 (0.674 to 2.13). No clinically relevant 20-percentage-point difference was observed.
Pain at rest, at night, on pressure, and in the morning1 reported outcome
Pain at rest, at night, on pressure, and in the morning
Numeric Rating Scale score points, mean (SD)
Table 3 means (SD), in order baseline / 6 weeks / 12 weeks / 1 year. Pain at rest: ESWT 3.9(2.5), 3.1(2.4), 2.4(2.6), 0.9(1.9); placebo 3.7(2.3), 2.7(2.5), 2.4(2.5), 0.9(1.6). Pain at night: ESWT 3.0(2.8), 2.1(2.5), 1.5(2.4), 0.8(1.7); placebo 2.8(2.8), 2.0(2.6), 1.8(2.5), 0.7(1.5). Pain at pressure: ESWT 7.0(2.6), 4.7(2.9), 4.0(3.2), 1.7(2.4); placebo 7.3(2.2), 4.3(2.6), 4.3(3.2), 1.8(2.4). Morning pain: ESWT 7.8(2.4), 5.2(3.1), 4.0(3.2), 1.5(2.6); placebo 7.7(2.3), 4.9(3.1), 4.5(3.4), 1.7(2.4). Pain measures are 0–10 numeric ratings, lower is less pain. Secondary outcomes were analyzed descriptively; no between-group inferential results are supplied for these separate pain rows.
Analysis details and source
Sample sizes vary by timepoint and outcome (Table 3); secondary pain ratings were analyzed descriptively, without reported p-values or CIs for between-group effects. Do not infer comparative benefit from within-group decline.
Pain-free walking ability1 reported outcome
Pain-free walking ability
Pain-free walking time and unlimited walking status minutes; participants
Pain-free walking time (minutes), mean (SD), baseline / 6 weeks / 12 weeks / 1 year: ESWT n=133,114,99,51: 22.1(36.1),45.4(59.1),68.5(87.1),131.0(95.5); placebo n=134,116,98,53: 25.7(46.2),50.4(65.0),53.4(77.3),115.1(102.5). Table 3 additionally lists walking category counts (unlimited pain-free walking): ESWT 1,15,28,62; placebo 2,16,31,62. The article does not report a between-group inferential test for walking ability.
Analysis details and source
Descriptive secondary outcome; follow-up denominators decline substantially by one year.
How treatment was delivered
Central randomization with computer-generated, centre-specific permuted blocks of four and six; treatment allocation communicated by telephone to the study doctor when the patient arrived for the first intervention. Patients and outcome assessors were blinded; only the intervention caregiver knew allocation and did not conduct follow-up.
Dornier Epos Ultra lithotripter with 7.5 MHz linear-array ultrasound positioning; focused shock waves targeted to the plantar fascia insertion/heel spur.
4,000 impulses per session, 3 sessions every 2 weeks (±2 days); positive energy flux density 0.08 mJ/mm²; total positive dose 0.96 J/mm²; energy flux density 0.22 mJ/mm²; positive pressure 13.7 MPa.
Both groups received 2 mL of 1% mepivacaine. Sham used an air-filled polyethylene foil before the coupling cushion to reflect shock waves; setup and device sounds were similar.
Follow-up at 6 and 12 weeks and 1 year after the last session. Primary success: R&M score 1–2 AND no additional treatment at 12 weeks. ITT stated; Fisher exact test, two-sided overall α=0.05; interim α=.005 and final α=.048; absolute risk differences, ORs, exact 95% CIs, centre-stratified robustness analysis. Secondary outcomes descriptive.
Any reported side effect: 24 (18%) with ESWT versus 12 (9%) with placebo (the paper prints no denominators for these counts); OR 2.26 (95% CI 1.02–5.18). Specific events: skin redness 16 (12%) versus 5 (4%); pain 7 (5%) versus 2 (2%); local swelling 3 (2%) versus 0. Less frequent events included haematoma, nausea, dizziness, hair loss and sleep disturbance. None stopped treatment. One deep-vein thrombosis occurred with placebo and was judged unrelated.
Sixteen of 272 randomized participants lacked the primary 12-week endpoint: eight per arm. Figure 1 lists the reasons (ESWT: dorsal spur 1, unintentional injury 1, death of a relative 1, unknown 5; placebo: withdrew consent 1, thrombosis 1, other surgery 2, unknown 4). No missing-data sensitivity analysis is reported. The study says intention-to-treat, but the exact missing-outcome handling is not detailed.
At one year, outcomes were affected by additional treatment and crossover: 41/113 (36%) in ESWT and 64/115 (56%) placebo sought further treatment; 44 placebo participants received ESWT. Thus long-term outcomes no longer isolate original assignment.
Baseline group denominators vary across abstract, Table 1 and randomized totals (135/137 versus 135/136); captured article does not explain the one-participant discrepancy.
The article refers to online eligibility details, modified Roles–Maudsley scoring Table A, and additional-treatment Table B; these supplements were not captured. Figure 1 was separately retrieved from the source-linked NCBI CDN and visually inspected.
The trial tested only one focused, ultrasound-targeted, anaesthetized high-dose protocol. Authors explicitly limit inference to the therapeutic variables applied.
The equipment was provided by Dornier Medizintechnik, although the company had no reported control over conduct or publication; study also had grant support from German research/orthopaedic organizations.
Findings checked against the full source by a separate automated reviewer.
Limitations noted by the reviewers
- Sixteen of 272 randomized participants lacked the primary 12-week endpoint: eight per arm. Figure 1 lists the reasons (ESWT: dorsal spur 1, unintentional injury 1, death of a relative 1, unknown 5; placebo: withdrew consent 1, thrombosis 1, other surgery 2, unknown 4). No missing-data sensitivity analysis is reported. The study says intention-to-treat, but the exact missing-outcome handling is not detailed.
- At one year, outcomes were affected by additional treatment and crossover: 41/113 (36%) in ESWT and 64/115 (56%) placebo sought further treatment; 44 placebo participants received ESWT. Thus long-term outcomes no longer isolate original assignment.
- Baseline group denominators vary across abstract, Table 1 and randomized totals (135/137 versus 135/136); captured article does not explain the one-participant discrepancy.
- The article refers to online eligibility details, modified Roles–Maudsley scoring Table A, and additional-treatment Table B; these supplements were not captured. Figure 1 was separately retrieved from the source-linked NCBI CDN and visually inspected.
- The trial tested only one focused, ultrasound-targeted, anaesthetized high-dose protocol. Authors explicitly limit inference to the therapeutic variables applied.
- The equipment was provided by Dornier Medizintechnik, although the company had no reported control over conduct or publication; study also had grant support from German research/orthopaedic organizations.
Source-linked critical reading
Randomized trial with sham / placebo control
Technical source record
Complete captured PMC full-article HTML read, all three article tables transcribed, and source-linked Figure 1 image visually inspected. Full text also identifies the inaccessible online BMJ eligibility criteria and Tables A/B, which remain limitations.
In this trial, focused shockwave treatment with local anaesthesia did not improve the 12-week success rate over a convincing placebo procedure. Both groups improved over time, and many later received additional care. The finding concerns this specific protocol and does not settle every shockwave approach.
In adults with chronic plantar fasciitis persisting despite at least six months of conservative care, did ultrasound-targeted focused ESWT improve outcomes over a blinded placebo procedure when both groups received identical local anaesthesia and treatment schedules?
A multicentre, randomized, blinded, parallel-group placebo-controlled trial in Germany. Central computer-generated randomization used permuted blocks of four and six within each centre. Patients and outcome assessors were blinded; the treatment caregiver knew allocation and was kept out of follow-up.
Adults with chronic plantar fasciitis refractory to conservative treatment for at least six months, recruited across nine hospitals and one outpatient clinic. The article directs readers to online criteria for full eligibility details; those supplementary criteria were not part of the captured article source.
Ultrasound-targeted focused treatment with a Dornier Epos Ultra lithotripter: 4,000 impulses per session, three sessions every two weeks (±2 days), positive energy flux density 0.08 mJ/mm², 0.22 mJ/mm² energy flux density and 13.7 MPa positive pressure; total positive dose 0.96 J/mm². Both groups received 2 mL of 1% mepivacaine local anaesthesia.
Placebo used the identical procedure and anaesthesia, with an air-filled polyethylene foil in front of the coupling cushion to reflect shock waves. Setup and device sounds were similar. This was an unusually credible sham; it controls for anaesthesia, attention and procedure ritual.
Primary endpoint at 12 weeks: success required both Roles and Maudsley score 1 or 2 and no additional treatment. Secondary endpoints included R&M scores, pain at rest/night/pressure/morning (0–10), pain-free walking, and additional treatments through one year. Primary success data were available for 256/272 (94%). Figure 1 shows 470 screened, 198 excluded, 272 randomized; eight primary outcomes were missing in each arm, with reasons including five and four unknown respectively. One-year denominators were 113 and 115.
At 12 weeks, primary success was 43/127 (34%) with ESWT versus 39/129 (30%) with placebo: absolute difference 3.6 percentage points (95% CI −8.0 to 15.1), p=.5927; OR 1.18 (95% CI .675–2.07). The study planned to detect a 20-point difference, which was not approached. Six-week and one-year R&M success contrasts also had CIs spanning no difference. Pain ratings and walking were described as secondary; no clear between-group advantage was reported.
Non-serious adverse effects were more common with ESWT: 24 (18%) versus 12 (9%); expected OR 2.26 (95% CI 1.02–5.18). Skin redness occurred in 16 (12%) versus 5 (4%), pain in 7 (5%) versus 2 (2%), and local swelling in 3 (2%) versus 0. No adverse effect stopped treatment. One deep-vein thrombosis occurred in placebo and was considered unrelated.
Sixteen participants lacked the 12-week primary outcome (256/272 available); no missing-value sensitivity analysis is reported. Later outcomes are complicated by substantial additional treatment and crossover, including 44 placebo participants receiving ESWT by one year. The study tested one specific focused, anaesthetized protocol; authors caution that findings do not resolve every protocol. Online eligibility criteria and supplemental table B were referenced but not present in the captured source. The flowchart clarifies that one ESWT-assigned patient received placebo and three placebo-assigned patients received ESWT; the article reports intention-to-treat, but this means treatment receipt was imperfect.
Directly relevant to persistent, treatment-refractory plantar fasciitis treated with ultrasound-targeted focused ESWT under local anaesthesia using this high-dose Dornier protocol. It does not establish that all radial/focused ESWT schedules are ineffective, and it does not answer whether a different dose or no anaesthesia changes outcomes.
A replication should prespecify a clinically important success difference, maintain concealed allocation and sham credibility, report all flow and adverse events, use missing-data sensitivity analyses, and separate randomized follow-up before crossovers from subsequent treatment outcomes. It should publish full eligibility criteria and protocols.
Read the original article.
Extracorporeal shock wave therapy for plantar fasciitis: randomised controlled multicentre trial.
Haake M, Buch M, Schoellner C, Goebel F, Vogel M, Mueller I, Hausdorf J, Zamzow K, Schade-Brittinger C, Mueller HH. · BMJ (Clinical research ed.) · 2003
Open the publication record or available full-text files.
Checked against the full paper · Clinical review: Dr. Marcus Yu Bin Pai, · How we review →