Gerdesmeyer L, Wagenpfeil S, Haake M, Maier M, Loew M, Wörtler K, Lampe R, Seil R, Handle G, Gassel S, Rompe JD. Extracorporeal Shock Wave Therapy for the Treatment of Chronic Calcifying Tendonitis of the Rotator CuffA Randomized Controlled Trial. JAMA. 2003;290(19):2573-2580. doi:10.1001/jama.290.19.2573
Author Affiliations: Department of Orthopedic Surgery and Sportstraumatology (Drs Gerdesmeyer and Lampe), Institute of Medical Statistics and Epidemiology (Dr Wagenpfeil), and Department of Radiology (Dr Wörtler), Technical University Munich, Munich, Germany; Department of Orthopedic Surgery, University of Regensburg, Bad Abbach, Germany (Dr Haake); Department of Orthopedic Surgery, Ludwig Maximilian University Munich, Munich, Germany (Dr Maier); Department of Orthopedic Surgery, Ruprecht-Karls-University Heidelberg, Heidelberg, Germany (Dr Loew); Department of Orthopedic Surgery, University of Homburg, Homburg/Saar, Germany (Dr Seil); Department of Orthopedic Surgery, University of Innsbruck, Innsbruck, Austria (Dr Handle); Department of Orthopedic Surgery, University of Bonn, Bonn, Germany (Dr Gassel); and Department of Orthopedic Surgery, Johannes Gutenberg University School of Medicine, Mainz, Germany (Dr Rompe).
Context Extracorporeal shock wave therapy (ESWT) has been used to treat calcific
tendonitis of the shoulder, but trials of ESWT for this purpose have had methodological
deficiencies and thus there is limited evidence for its effectiveness.
Objective To determine whether fluoroscopy-guided ESWT improves function, reduces
pain, and diminishes the size of calcific deposits in patients with chronic
calcific tendonitis of the shoulder.
Design, Setting, and Participants Double-blind, randomized, placebo-controlled trial conducted between
February 1997 and March 2001 among 144 patients (of 164 screened) recruited
from referring primary care physicians, orthopedic surgeons, and sports physicians
in 7 orthopedic departments in Germany and Austria.
Interventions Either high-energy ESWT, low-energy ESWT, or placebo (sham treatment).
The 2 ESWT groups received the same cumulative energy dose. Patients in all
3 groups received 2 treatment sessions approximately 2 weeks apart, followed
by physical therapy.
Main Outcome Measures The primary end point was the change in the mean Constant and Murley
Scale (CMS) score from baseline to 6 months after the intervention. Secondary
end points were changes in the mean CMS scores at 3 and 12 months, as well
as changes in self-rated pain and radiographic change in size of calcific
deposits at 3, 6, and 12 months.
Results Of 144 patients enrolled, all completed treatment as randomized and
134 completed the 6-month follow-up. Both high-energy and low-energy ESWT
resulted in significant improvement in the 6-month mean (95% confidence interval
[CI]) CMS score compared with sham treatment (high-energy ESWT: 31.0 [26.7-35.3]
points; low-energy ESWT: 15.0 [10.2-19.8] points; sham treatment: 6.6 [1.4-11.8]
points; P<.001 for both comparisons). Patients
who received high-energy ESWT also had significant 6-month CMS improvements
compared with those who received low-energy ESWT (P<.001).
We found similar results for both the 3-month and 12-month CMS comparisons,
as well as for self-rated pain and radiographic changes at 3, 6, and 12 months.
Conclusions Both high-energy and low-energy ESWT appeared to provide a beneficial
effect on shoulder function, as well as on self-rated pain and diminished
size of calcifications, compared with placebo. Furthermore, high-energy ESWT
appeared to be superior to low-energy ESWT.
Calcific tendonitis of the rotator cuff is a well-known source of shoulder
pain.1 Estimates of the overall incidence vary
widely, ranging between 2.5% and 20%,1- 3 depending
on both clinical criteria and radiographic technique. The disease is usually
self-limiting but the natural course is variable.1- 5 For
instance, Gärtner6 reported that calcifications
with sharp margins and homogeneous or nonhomogeneous structure disappeared
spontaneously in 33% of patients over a period of 3 years, but that 85% of
fluffy accumulations did so during the same time period. In 1941, Bosworth1 reported that 6.4% of calcific lesions showed spontaneous
Clinically, it is important to distinguish calcific tendonitis from
a rotator cuff tear as a source of shoulder pain.7 Several
authors have found no correlation between the presence of a tendon tear and
calcific tendonitis.4,7- 10 The
treatment of patients with calcific tendonitis typically is conservative,
including use of subacromial cortisone injections, physical therapy, and systemic
nonsteroidal anti-inflammatory drugs, although evidence of efficacy is limited.11,12 For patients with chronic calcification,
surgical removal of the deposits, either with an open procedure or endoscopically,
has been reported to relieve symptoms.13- 19
Ultrasound treatment may be an alternative to surgery. Ebenbichler et
al20 reported that ultrasonic energy accelerated
functional improvement in patients with acute calcific tendonitis, although
efficacy was no better than that achieved with placebo in long-term follow-up.
Although extracorporeal shock wave therapy (ESWT) has demonstrated encouraging
results in the treatment of calcified deposits,21- 25 all
of these trials have had methodological deficiencies.12 We
compared the effectiveness of high-energy and low-energy ESWT vs placebo (ie,
sham treatment) in patients with chronic, symptomatic calcific tendonitis
of the supraspinatus tendon.
Our study was a randomized, placebo-controlled trial in 7 sites in Germany
and Austria and was conducted between February 1997 and March 2001. Participants
were assigned to receive either high-energy ESWT, low-energy ESWT, or sham
treatment (Figure 1). In designing
the trial we adhered to the standardized guidelines of good clinical practice
from the International Conference on Harmonization.26,27 All
patients provided written informed consent. The trial was approved by the
ethics committee of the Faculty of Medicine of the Technical University of
Potential participants were made aware of the trial by reports in the
press, by health insurance companies, or by orthopedic practitioners or hospitals.
They were referred to one of the participating centers in Germany and Austria.
To be eligible for the trial, participants had to have a history of at least
6 months of pain or tenderness from idiopathic calcific tendonitis, type I
or II according to Gärtner,6 that was
resistant to conservative treatment.
Participants were eligible if they were aged 18 years or older, had
calcific deposits of 5 mm in diameter or larger on radiography, and had had
symptoms for at least 6 months. Rotator cuff tears and subacromial bursitis
were ruled out in all patients by clinical and sonographic examination, and
when in doubt, by magnetic resonance imaging prior to randomization and at
all follow-up visits. Participants with type III Gärtner deposits were
excluded because of high probability of spontaneous resolution.6 We
required that all participants had had previous conservative treatments, including
both physiotherapy (eg, active and passive exercise, mobilization, manual
therapy and massage, muscle strengthening) and local anesthetic or corticosteroid
injections. We also verified that all participants had tried nonsteroidal
anti-inflammatory drugs such as ibuprofen or diclofenac. Exclusion criteria
included rheumatic disease, connective tissue disease, or diabetes; coagulation
disturbance; pregnancy; glenohumeral or acromioclavicular joint arthritis;
previous surgery for shoulder pain; bursitis, infection, or tumor of the shoulder;
instability of the shoulder or rotator cuff tear; type III calcific deposit
(by Gärtner classification); abnormal peripheral neurologic findings;
and unsuccessful prior ESWT.
Treatment allocation was determined immediately before the first treatment
by block randomization (48 per block) using a computer-generated algorithm
at a central location. Assignments were then delivered by telephone and kept
in sealed opaque envelopes. Patients, as well as the follow-up evaluators,
were blinded to treatment assignments.
Patients were assigned to receive either high-energy ESWT, low-energy
ESWT, or sham treatment. All patients had had at least a 1-month, therapy-free
period before the first treatment with ESWT. Patients in all groups were informed
that sometimes the procedure could be painful and could take up to 1 hour
per session due to the necessity to control and refocus the shock waves exactly.
Immediately after randomization, the patient was placed in the prone
position. Using fluoroscopy in an anterior-posterior view, the shoulder was
rotated until the calcific deposit was identified in a free position. For
the high-energy and low-energy groups, a shock wave head was coupled to the
shoulder with a thin sheet of polyethylene foil placed between the shock wave
head and the patient. Coupling gel was used between the shock wave head and
the foil and between the foil and shoulder.
The exact focus position was controlled using fluoroscopy during the
ESWT procedure and adjusted if necessary. After the energy level was increased
up to the assigned treatment level, the assigned number of shock waves were
applied. Patients in the high-energy group received 1500 shock waves of 0.32
mJ/mm2 per treatment, while those in the low-energy group received
6000 shock waves of 0.08 mJ/mm2. In both groups, 120 impulses were
applied per minute. Adequate intravenous analgesia and sedation were provided
as necessary. Local anesthetics were prohibited. All patients received a second
ESWT treatment at 12 to 16 days; thus, patients in each group received a cumulative
energy dose of 0.960 J/mm2. Each treatment session lasted as long
as 1 hour. Measurements with glass-fiber hydrophones in accordance with International
Electrotechnical Commission (IEC) procedures28 demonstrated
that shock waves were unaffected by the polyethylene foil when used with ultrasound
coupling gel on both sides of the foil (data not shown).
In the sham treatment, an air-chambered polyethylene foil with coupling
gel was placed against the patient's skin, but no coupling gel was applied
to the site of the shock wave head. The air-chambered polyethylene foil was
placed between the patient and the water cushion of the ESWT device in the
same technique as in the other 2 groups. In every other respect the setup
was the same. Measurements with glass-fiber hydrophones in accordance with
IEC procedures demonstrated that no shock waves could pass through the foil.
Patients in the sham treatment group received 1500 shock waves per treatment
with 120 impulses per minute after the energy level reached the assigned treatment
level of 0.32 mJ/mm2 (although a total of 0.960 J/mm2 was
emitted from the ESWT device over the 2 treatments). The patients' prone position
prevented them from seeing the device, but they could hear the typical sound
of shock waves being generated.
Patients in all 3 groups underwent 10 physiotherapy sessions after the
intervention. This included active and passive exercise mobilization techniques,
massage, and manual therapy to prevent worsening in range-of-motion, muscular
deficit, or imbalance.
Rescue medication was allowed throughout the entire study if pain became
unbearable (2 g of paracetamol or 2 g of acetaminophen per day for up to 14
days following the last treatment; thereafter, 2 g of paracetamol or 2 g of
acetaminophen per week). No other therapies (eg, chiropractic, laser, acupuncture,
ultrasound, other nonsteroidal anti-inflammatory drugs, or corticosteroids)
were allowed until after the 6-month follow-up.
The primary end point was the change in the mean Constant and Murley
Scale (CMS)29 score from baseline to 6 months
after treatment. Comparisons between the sham treatment group and the other
2 groups were prespecified, while comparisons between the groups receiving
high-energy and low-energy ESWT were performed in a post hoc fashion.
The CMS is a standardized simple clinical method of assessing shoulder
function and has a maximum score of 100 points, with both subjective (35 points)
and objective (65 points) components. The CMS has been reported to have high
interobserver and intraobserver reliability.30 The
subjective parameters assess the degree of pain perception (15 points) and
the ability to perform the normal tasks of daily living in both activity-
and position-related terms (20 points). The objective parameters include testing
of active range of motion (40 points) and shoulder power (25 points). All
observers who assessed the CMS were blinded. All were experienced and used
a goniometer to evaluate the active forward and lateral elevation and body
landmarks reached by the patient to assess the internal/external rotation.
The power in abduction was measured using a spring balance.
The 6-month interval was selected because we expected that healing would
likely be evident (although not necessarily complete) at this point. Clinically
relevant improvement was defined as a 30% increase from baseline on the CMS
score. Patients who needed additional therapies, except the allowed amount
of rescue medication and physiotherapy, were defined as failing treatment.
Secondary end points were changes in mean 3- and 12-month CMS scores,
as well as in self-rated pain at 3, 6, and 12 months as assessed by a visual
analog scale (VAS) (0 points = no pain; 10 points = unbearable pain). We also
assessed the presence and size of calcified deposits at 3, 6, and 12 months
by conventional radiography. The technique was standardized in terms of position
of the shoulder and arm, distance from the radiographic film, and exposure.31 The localization of calcifications within a specific
tendon was determined by anteroposterior radiographs of the shoulder obtained
in 45° external and 45° internal rotation.31 These
2 standard anterior-posterior views were obtained within 14 days before intervention
to exclude spontaneous healing before treatment and again at 3, 6, and 12
months after treatment and analyzed by an independent skeletal radiologist
with no knowledge of the type of treatment used. Success was defined as complete
disappearance of the deposit.
Changes in CMS scores for pain, activities of daily living, range of
motion, and power, as well changes in VAS pain scores and size of calcific
deposit were defined as the difference between the 3-month, 6-month, and 12-month
measurements and respective baseline values. These absolute changes were the
variables of interest and analysis.
All analyses were performed with SPSS release 11.5 (SPSS Inc, Chicago,
Ill). Computed P values were 2-sided, and P<.05 was used to determine statistical significance. For group
comparisons of changes we used the t test for independent
samples or the Welch test, as appropriate. Significance levels for multiple
comparisons were adjusted with the Bonferroni-Holm procedure. All analyses
of the primary outcome were performed according to the principle of intention-to-treat,
with missing values imputed with last observation carried forward. For the
secondary end points, descriptive statistics and 95% confidence intervals
We computed that a sample of 144 patients had 90% power to find a 15%
difference in the primary outcome, as compared with sham treatment, given
an α level of .025. We tested for selection bias according to the method
of Berger and Exner.32 To examine for treatment-center
effects we applied the Kruskal-Wallis test on the primary outcome variable
within each of the treatment groups separately and an analysis of covariance
with treatment-center interaction.
A total of 144 patients (48 per group) were treated as randomized according
to the study protocol (Figure 1).
The required number of pulses per treatment was achieved in all cases. Baseline
characteristics of the sample are presented in Table 1. Only 10 patients were lost to follow-up (7%) prior to the
6-month end point, but considerably more were lost to follow-up after that.
The method of Berger and Exner32 provided
strong support against selection bias; comparing baseline CMS values with
conditional probabilities that the next treatment is high energy or low energy
given knowledge of the sequence of prior allocations within the randomization
block, we obtained Pearson correlation coefficients of 0.03 and −0.01,
respectively. The 3 Kruskal-Wallis tests comparing the primary outcome measure
across the centers for each of the 3 treatment groups separately showed no
center effect (P≥.09 for all), with similar results
from analysis of covariance. Alternative evaluation of group comparisons with
a respective permutation test yielded similarly nonsignificant results.
The means of the 6-month CMS scores are presented in Table 2. In this primary analysis, both high-energy and low-energy
interventions were superior to sham treatment, and in a secondary analysis
the high-energy intervention appeared to be superior to the low-energy intervention.
The various components of the score (ie, pain, activities of daily living,
range of motion, and power) showed similar patterns of results.
Table 3 presents the results
of both the 3-month and 12-month CMS data, which generally parallel those
of the 6-month data. Use of other imputation techniques did not substantially
change the pattern of results for the 12-month results (data not shown).
Table 4 presents the 3-,
6-, and 12-month VAS pain scores as well as radiographic results. Similar
to the CMS scores, patients in the high-energy group had significantly less
pain than those in the low-energy group, but both groups reported significantly
less pain than those in the sham treatment group 6 months after intervention.
At 3 and 12 months after intervention, no significant differences in VAS score
were observed for the low-energy vs sham treatment groups.
Complete disappearance of the calcific deposit was observed in 60% of
the patients in the high-energy group after 6 months and in 86% after 12 months.
In the low-energy group, complete disappearance was observed in 21% and 37%,
respectively. In the sham treatment group, complete disappearance was observed
in 11% after 6 months and in 25% after 12 months. Finally, it appeared that
more patients in the sham treatment group used additional therapies after
6 months (Figure 1).
Adverse effects were assessed by clinical examination, ultrasound imaging,
and by patient questionnaire directly after the ESWT procedure and after every
follow-up visit. All findings were recorded on standardized forms. Patients
were explicitly asked to report any reddening of the skin, swelling, petechiae,
reaction to the anesthetic used, bleeding, acute bursitis, or syncope occurring
after the intervention. In addition, patients also were asked whether they
had experienced any other adverse effects. Unexpected or severe adverse events
were to be reported separately, but none occurred.
Pain during treatment was analyzed separately. In the group receiving
high-energy ESWT, 20 patients reported moderate pain and 16 reported severe
pain. Eight of those reporting severe pain required intravenous analgesics
during intervention. Ten patients in the high-energy group had insignificant
or no pain during the ESWT procedure. In the group receiving low-energy ESWT,
moderate pain was reported by 22 patients and severe pain by 5; 2 of those
reporting severe pain required intravenous pain medication. Twenty-one patients
in the low-energy group reported slight or no pain.
In the sham treatment group, 25 patients reported some sensation of
pain. Four had severe pain and 1 required additional intravenous pain medication.
Insignificant or no pain sensation was observed in 23 cases.
Petechiae, bleeding, hematoma, or erythema were found directly after
the treatment in 36 patients in the high-energy group, 32 patients in the
low-energy group, and 8 patients in the sham treatment group.
No clinically significant adverse effects (including neurologic disorders,
tendon rupture, infection, bone edema, aseptic necrosis, or muscle hematoma)
were observed in any of the patients at any point in time.
Shoulder pain due to calcific tendonitis is a common problem, for which
conservative therapy is sometimes ineffective.1,5 In
these cases, ESWT has been proposed as an alternative to operative procedures,21,24,33- 35 although
methodological flaws have limited the conclusions of previous studies.12 In our study, we found a significant clinical benefit
for both high- and low-energy ESWT at 6 months, with significantly better
outcomes associated with high-energy ESWT. Patients in the sham treatment
group showed a previously demonstrated spontaneous improvement.4 Nonetheless,
they required more pain medication than patients in the 2 ESWT groups and
were more likely to undergo surgery during follow up.
Some authors have stressed the importance of stone removal in the therapy
of nephrolithiasis, while others have suggested a need for complete disintegration
of the calcified deposits around joints.36 We
observed a complete disappearance of the deposit in 60% of patients 6 months
after receiving high-energy ESWT, a nearly 3-fold greater rate of complete
disintegration than that observed in those who received sham treatment. Although
some authors have discussed the potential of extracorporeal shock waves for
disintegrating calcified deposits of the rotator cuff, the mechanisms remain
Several studies have found a correlation between the applied energy
of each shock wave and the rate of disintegration,23,25 assuming
that the shock wave is carefully focused.40 At
present, however, it is unclear which parameters of shock waves are most related
to resorption of the deposit. The "energy flux density" parameter is generally
assumed to be the primary parameter for physical and biological effects.41 For instance, simply doubling the number of applied
shock waves does not appear to improve the likelihood of eliminating tendon
calcification or of improving clinical outcomes.24,25 Our
results similarly suggest that the energy level seems to be a more important
parameter. The high-energy and low-energy groups received the same total acoustic
energy but showed different clinical and radiological outcomes. In addition
to the number of shock waves and energy level, the frequency of shock waves
may have an influence. Recent studies of kidney stones found that fragmentation
efficiency, due to cavitation effects, was significantly enhanced at a delay
of between 400 and 250 microseconds between shock waves.42 These
findings support the idea that cavitation effects may be related to the disintegrating
effect of ESWT.39,43 It also seems
likely that ESWT may be more effective for calcifying tendonopathy than for
impingement syndromes that do not involve any calcified masses.44,45
We found no serious adverse effects of ESWT. As in previous studies,23,33,35,46 some
patients in our study did complain of petechial bruising, subcutaneous hematoma,
or skin reddening immediately after treatment, but in all cases these had
resolved by 3 months. It is possible that different shock wave generators
may vary in their physical parameters, and thus in their likelihood of causing
While studies in rabbits have revealed some short-term tendon pathology
associated with ESWT energy levels of at least 0.6 mJ/mm,2,47 neither
tendon nor cartilage of joints has been found to be injured by shock waves
lower than this energy level.48,49 Although
we did not perform imaging studies to detect these potential adverse effects,
neither tendon ruptures nor aseptic necrosis of the humeral head50 were
reported. Long-term observations 4 years after high-energy treatment found
neither tendon lesions nor other adverse effects due to shock waves in patients
who later underwent surgery.51,52 It
is possible that ESWT could be less expensive than surgery for treatment of
calcific tendonitis of the shoulder.53
Our results have 2 important limitations. First, our findings may be
limited by the different amounts of intravenous sedation used in the treatment
groups, which was confounded with the effects of the active therapy and the
amount of shock wave energy. It is unlikely that intravenous sedation alone,
however, may have influenced this chronic pain condition. Second, because
of the high drop-out rates after 6 months, the 12-month data should be interpreted
Our findings need to be confirmed in high-quality randomized clinical
trials with different treatment protocols and treatment parameters, including
the number of shock waves, their frequency, and their energy levels. Further
studies also are necessary to analyze the long-term prognosis, and also should
examine less-systemic forms of anesthesia, including regional nerve block
or local anesthesia.
In summary, we found evidence for a beneficial effect of high-energy
ESWT over 6 months, compared with sham treatment. High-energy ESWT appears
to be more effective than low-energy ESWT, but threshold energy has yet to