Study Design
A prospective, two-tailed, randomized, controlled, observer-blinded
pilot trial was performed to compare the outcomes of three applications
of 1000 impulses of low-energy shock waves with those of three applications
of ten impulses of low-energy shock waves in patients with intractable
heel pain. One hundred and nineteen patients (fifty-one female and
sixty-eight male; mean age, forty-six years) who had had pain for
a mean of nine months (range, six to twenty months) were eligible
for the study. All 119 patients had been previously treated unsuccessfully.
Eighty patients had been given medication, mostly nonsteroidal anti-inflammatory
drugs; 110 had worn shock-absorbing shoe inserts; forty-two had
performed some kind of stretching exercises on a regular basis;
nineteen had used night splints; and eighty-one had been treated
with a cast for at least two weeks. An average of 1.9 corticosteroid
injections (range, one to five injections) had been given to the
119 patients, and an average of three different physical therapy regimens
(range, one to five different regimens), such as icing, ultrasound,
magnetic field therapy, iontophoresis or phonophoresis, contrast
baths, and radiation therapy, had been tried. One hundred and twelve
patients agreed to the randomization procedure, and they formed
the study sample.
Inclusion Criteria
The criterion for entry into the study was heel pain localized to
the site of the insertion of the plantar fascia and intrinsic muscles
on the medial calcaneal tuberosity on the anterior-medial aspect
of the heel for more than six months. The severity of the pain was
recorded, and a low pain score was not an exclusion criterion. The
location of the pain was tested by exerting pressure on the heel
under sonographic control. Conservative therapy had to have failed
for at least six months before referral to our hospital. In order
to allow positioning of the shock-wave focus, a plantar heel spur
had to be seen radiographically in the area of the medial calcaneal
tuberosity. The size of the spur did not play a role as an inclusion
criterion.
Exclusion Criteria
The exclusion criteria, elicited from the patient’s
medical record, included dysfunction of the knee or ankle, local
arthritis, generalized polyarthritis, rheumatoid arthritis, ankylosing spondylitis,
Reiter syndrome, neurologic abnormalities, nerve entrapment syndrome,
a previous operation on the heel, an age under eighteen years, pregnancy,
an infection, or a tumor. Thirteen patients were excluded from the
study on the basis of these criteria.
Except for previously worn shoe inserts, no additional treatmentæfor
example, nonsteroidal anti-inflammatory drugsæwas allowed
during the first three months after application of the extracorporeal
shock waves. Three patients in Group I and seven in Group II took
such drugs; this was regarded as indicating failure of the extracorporeal
shock-wave application, and the patients were withdrawn from the
study. They were instructed to use the foot but to avoid painful
stress.
Randomization
After six weeks with no treatment of any kind and after they gave
informed consent, the patients were evaluated again to make sure
that no exclusion criteria applied. Then they were randomized into
the two treatment groups with use of identical sealed envelopes.
The first application of shock waves was carried out immediately
after the identification of the treatment group.
The randomization began in 1993 and, as had been planned previously,
was stopped (in 1995) after fifty patients in one of the two groups
had not used additional treatment or drugs for three months after
shock-wave application (Fig. 1).
Group I
Group I received a total of 3000 impulses of an energy flux density
of 0.08 mJ/mm2. The group consisted
of twenty-one women and twenty-nine men, with a mean age of forty-four
years (range, twenty-six to sixty-one years). The mean duration
of pain was eight months (range, six to nineteen months).
Group II
Group II received a total of thirty impulses of an energy flux density
of 0.08 mJ/mm2. There were twenty
women and thirty men, and their mean age was forty-nine years (range,
thirty-one to sixty-three years). The mean duration of pain was
ten months (range, six to twenty months).
Method of Treatment
Extracorporeal shock waves were applied by an experimental device
(Siemens Osteostar; Siemens AG, Erlangen, Germany) characterized
by the integration of an electromagnetic shock-wave generator in
a mobile fluoroscopy unit. By means of an acoustic lens, the focus
of the shock-wave source is just at the center of the c-arm. The
typical cigar-shaped focal extent of the device, defined as the —6
dB focal contour in the x, y, and z directions around the focus
location, covers an area of 50 mm in the axis of the shock wave,
with a diameter of 7.0 mm perpendicular to the shock-wave axis.
These technical parameters are very comparable with those of modern
shock-wave units for treatment of musculoskeletal disorders.
Once the tip of the plantar heel spur was situated in the center of
the c-arm, the shock-wave unit was docked to the foot by means of
a water-filled cylinder. Common ultrasound gel (University Hospital,
Mainz, Germany) was used as a contact medium between the cylinder
and the skin. Three times, at weekly intervals, 1000 or ten impulses
of an energy flux density of 0.08 mJ/mm2 were
administered to the heel; this dose was selected on the basis of
experience in an earlier study19.
Shock waves are considered low-energy when the energy flux density
ranges from 0.05 to 0.10 mJ/mm2,
making the use of local anesthetics unnecessary, although the treatment
is unpleasant.
Method of Evaluation
All patients were assessed before and after treatment. At a mean
of twenty-four weeks (range, twenty-two to twenty-six weeks) after
the last application of the extracorporeal shock waves, follow-up
was performed by a blinded observer, an orthopaedic surgeon who
had not been involved in the selection of the patients or in the
shock-wave treatment and who did not ask the patients about the
number of impulses applied. Another blinded observer, also an orthopaedic
surgeon, performed another follow-up examination at a mean of five
years (range, fifty-four to sixty-six months). The protocols of
treatment and evaluation were closely monitored to guarantee that the
treating physician did not evaluate his or her patients at the time
of follow-up.
Primary Outcome Measure
The primary outcome measure was defined prospectively as the
pain rating at six months after shock-wave application compared
with the pretreatment condition. The rating, according to modified
criteria of the Roles and Maudsley score20,
was defined as excellent (no pain, patient satisfied with the treatment
outcome, and unlimited walking without pain), good (symptoms substantially
decreased, patient satisfied with the treatment outcome, and ability
to walk without pain for more than one hour), acceptable (symptoms
somewhat decreased, pain at a more tolerable level than before treatment,
and patient slightly satisfied with the treatment outcome), or poor
(symptoms identical or worse and patient not satisfied with the
treatment outcome). Treatment was considered successful when the
patient had an excellent or good score.
Secondary Outcome Measures
The Roles and Maudsley score20 at
five years was defined prospectively as a secondary outcome measure.
Other prospectively defined secondary outcome measures were the
extent of pain at night, at rest, and on manual pressure as specified
on a visual analog scale ranging from 0 (no pain) to 100 (worst
imaginable pain) at six months and at five years. To assess pain
on manual pressure, the physician used his or her thumb to gradually
increase pressure on the patient’s contralateral unaffected
heel until pain began; then a comparable amount of pressure was
applied to the affected heel, and the patient rated the pain that
it caused. The exact amount of pressure was not measured.
Walking ability without a need for rest to relieve pain in the heel
was rated as 0 (less than five minutes), 1 (less than fifteen minutes),
2 (less than thirty minutes), 3 (less than forty-five minutes),
4 (less than sixty minutes), or 5 (unlimited).
All patients had a radiograph made of the heel before the treatment
and at the six-month follow-up evaluation.
Statistical Analysis
The aim of this study was to assess whether there was a dose-dependent
effect of low-energy extracorporeal shock-wave therapy in the treatment
of recalcitrant heel pain. Our hypothesis was that three applications
of 1000 impulses is superior to three applications of ten impulses
with regard to the results at six months.
The methods for statistical analysis had been determined by the
local Institute for Medical Statistics and Documentation before
the study was started. Accordingly, the statistical analysis was
performed at that institute when the study was completed.
The Wilcoxon rank-sum test was applied for the comparison of
the two groups for such pseudo-continuous, non-normally distributed
variables as pain at night, pain at rest, and pain on manual pressure21. The Roles and Maudsley score20 and walking ability, categorical
variables, were compared between groups with the Fisher exact test
and its extension to 2 ¥ n contingency tables. The level
of significance was set at 95%. Differences with p values
of <5% were considered significant. Multiple adjustment
was not performed for secondary outcome parameters that were measured
in an explorative way. The primary outcome measure, the Roles and
Maudsley score at six months, was tested in a confirmatory way22.
As this was a pilot study, no sample size or power calculation could
be performed before it was started. The six-month results of this
comparative study were analyzed on the basis of the total number
of patients whom we originally intended to treatæthat is,
fifty patients in each group.
Follow-up
As had been previously planned, the randomization process was
stopped after fifty patients in either group had not used additional
treatment or drugs for three months after the shock-wave application.
To reach this goal, 112 patients were randomized to the two treatment
groups: fifty-four were assigned to Group I and fifty-eight, to
Group II. At three months, three patients in Group I and seven patients
in Group II had to be excluded from the study because, as mentioned,
they had had additional conservative therapy during that time. One
patient in each group could not be contacted, leaving fifty patients
in both groups as the basis for the current study.
At six months, forty-nine of the fifty patients in Group I could be
evaluated. One patient refused to participate in the study any longer
because the shock-wave therapy had not improved his condition. In
Group II, forty-eight of the fifty patients could be examined at
six months. Two patients stopped participating because the shock-wave
application had not improved their condition. At five years, thirty-eight
of the fifty patients in Group I could be examined. Four patients
stopped participating because the shock-wave application had not
improved their condition, and seven patients could not be contacted.
At five years, forty of the fifty patients in Group II could be
evaluated. One patient stopped participating because the shock-wave
therapy had not improved his condition, and seven patients could
not be contacted (Fig. 1).
Primary Outcome Measure
Modified Roles and Maudsley Score20 at Six Months
At six months, six (12%) of the forty-nine patients
in Group I had an excellent result, twenty-two (45%) had
a good result, twenty (41%) had an acceptable result, and
one (2%) had a poor result. In Group II, none of the forty-eight
patients had an excellent result, five (10%) had a good
result, twenty (42%) had an acceptable result, and twenty-three
(48%) had a poor result. The rate of good and excellent
outcomes (i.e., successful results) was 47% higher (95% confidence
interval, 37% to 57%) in Group I than in Group
II, and the difference between the groups was significant (p < 0.0001).
A post hoc power analysis of the primary outcome measure—with
use of the relative success rates in Group I (0.57 0.50) and Group
II (0.10 0.31), the given sample sizes in Group I (forty-nine)
and Group II (forty-eight), and the significance level of the test
to reject the null hypothesis (a = 0.05)æshowed
a statistical power of 0.9.
Secondary Outcome Measures
Modified Roles and Maudsley Score20 at Five Years
At five years, twelve (32%) of the thirty-eight patients
in Group I had an excellent result, eighteen (47%) had
a good result, seven (18%) had an acceptable result, and
one (3%) had a poor result. In Group II, fifteen (38%)
of the forty patients had an excellent result, twelve (30%)
had a good result, nine (23%) had an acceptable result,
and four (10%) had a poor result. With the numbers available,
this difference of 11% (95% confidence interval,
4% to 18%) in the success rate between the two
groups was no longer significant (p = 0.071) (Fig. 2). It should
be noted that many of the good and excellent results in Group II
followed surgery performed subsequent to the shock-wave therapy,
as discussed below.
Pain on Manual Pressure
During the five years that these patients were followed after treatment,
the mean score for pain on manual pressure gradually decreased from
77 ± 13 points (before treatment) to 19 ±
12 points (at six months) and 9 ± 11 points (at
five years) in Group I. In Group II, the mean scores were 79 ±
11 points before treatment, 77 ± 10 points at six
months, and 29 ± 25 points at five years. There
was a significant difference between Group I and Group II at both
six months (p < 0.0001) and five years (p = 0.0006)
(Fig. 3).
Night Pain and Resting Pain
Night pain in Group I was significantly less than that in Group II
at six months (p < 0.0001) and five years (p = 0.0015).
In addition, resting pain in Group I was significantly less than that
in Group II at six months (p < 0.0001) and five years (p = 0.0033)
(Table I).
Walking
The ability to walk without pain was also significantly better in
Group I than it was in Group II at six months (p < 0.0001) and
five years (p = 0.0023) (Fig. 4). In Group I, twenty-five of forty-nine
patients were able to walk completely without pain at six months
compared with zero of forty-eight patients in Group II (p < 0.0001).
Radiographic Evaluation
Radiographs made at six months after treatment did not show any
structural changes of the hindfoot.
Complications
The low-energy extracorporeal shock-wave therapy was felt to
be unpleasant by all patients, although it was not thought to be
as unpleasant as the local infiltration that all patients had experienced
during the various and unsuccessful treatment regimens prior to
the current study. No patient stopped the shock-wave procedure because
of pain. No side effects were seen at the follow-up examinations
at six months and five years. There were no hematomas, infections,
or abnormal neurologic findings.
Additional Treatment
Between three and six months: Between three
months and six months, nine of the forty-nine patients in Group
I took oral nonsteroidal anti-inflammatory drugs and had local infiltration
with corticosteroids and anesthetics and one patient had the calcaneal
spur removed surgically. In Group II, only four of the forty-eight
patients did not need any additional treatment. Patients took nonsteroidal
anti-inflammatory medication and/or had local injections,
and one had surgical release of the plantar fascia.
At five years: At an average of five years (range,
fifty-four to sixty-six months), none of the thirty-eight Group-I
patients were receiving conservative therapy on a regular basis
and five (13%) had undergone surgery. One of these five
patients had an excellent result; two, a good result; and two, an
acceptable result. In Group II, nine (23%) of the forty
patients were receiving regular conservative treatment at five years
and twenty-three (58%) had been operated on. Nine of the
twenty-three patients had an excellent result after the operation,
ten had a good result, two had an acceptable result, and two had
a poor result. There were significantly more operative procedures
in Group II than in Group I (p < 0.0001). As a consequence
of the high (83%) rate of excellent and good results after
surgery in Group II, the results in Group I and Group II were no
longer significantly different five years after shock-wave application.
In a review of the literature since 1966, Atkins et al.1 and Crawford et al.2 found only eleven randomized, controlled
trials assessing the treatment of plantar fasciitis. There was limited
evidence of the effectiveness of topical corticosteroids administered
by iontophoresis, dorsiflexion night splints, and low-energy extracorporeal
shock-wave therapy.
A satisfying clinical outcome after application of low-energy extracorporeal
shock waves was first reported in patients with chronic tendinosis
of the elbow22. We showed comparable
short-time results for patients with plantar fasciitis and a heel
spur13. Similarly positive outcomes
have been confirmed in clinical studies from various university
hospitals15-17. Maier et al.16 reported good or excellent results,
according to the modified Roles and Maudsley score, in thirty-six
of forty-eight heels at twenty-nine months. The clinical outcome
was not influenced by the duration of the follow-up period. No negative
side effects were reported. Wang et al.23 reported
that thirty-three of forty-one patients were either free of symptoms
or substantially better at twelve weeks after shock-wave therapy.
Ogden et al.24 performed a randomized,
placebo-controlled study with 119 patients in the treatment group
and 116 patients in the placebo group. Twelve weeks after a single
application of 1500 high-energy shock waves at 18 kV with the patient
under regional anesthesia, the result was successful in 47% of
the patients. The success rate after the sham treatment was only
30%. This study led the United States Food and Drug Administration
to approve shock-wave therapy for painful heels. Buch et al.25 reported the results of another
randomized, placebo-controlled study, involving 150 patients, for
the United States Food and Drug Administration. Therapy with 3800
high-energy impulses was applied once with the patient under regional
anesthesia. At three months, 70% of the patients in the
treatment group and 40% of those in the placebo group fulfilled
the success criterion, which was a change in the visual analog score
for pain while walking for the first few minutes in the morning.
Chen et al.26 studied eighty patients
treated with 1000 shock-wave impulses at 14 kV. Of fifty-four patients
who were evaluated at six months, 59% had no symptoms and
27% had substantial improvement.
In the current study, six months after low-energy shock-wave treatment,
the results of three applications of 1000 impulses were significantly
better than those of three applications of ten impulses (57% good
or excellent outcomes compared with 10% good or excellent
outcomes).
At five years, Group II had a substantial improvement in all parameters
compared with those at the six-month follow-up evaluation, and the
overall outcome, based on the four-step score, was no longer significantly
better in Group I.
It should be noted that separating the clinical results into
only four broad categories, with use of an unvalidated modified Roles
and Maudsley scale20 originally
designed for the upper extremity, may not provide a sufficiently
sensitive test. However, Group-I patients also fared better with
regard to pain on manual pressure, at night, and at rest and with
regard to walking. Five years after the shock-wave therapy, 13% of
the patients in Group I and 58% of the patients in Group
II had been operated on. Of the twenty-three patients who were operated
on in Group II, 83% had a good or excellent outcome. If
even more patients in this group had undergone surgery, the ratings
concerning pain and walking may have reached levels comparable with
those in Group I.
None of the outcome variables in our study is free from the possibility
of observer bias, although this risk was kept low by making sure
that an independent observer evaluated the patients before and after
treatment. Pain, however, may be influenced by many factors and
is difficult to measure. While we attribute the substantial improvement
in Group II at five years to the surgical procedures that the patients
had undergone during the follow-up period, the excellent long-term results
in Group I have to be regarded with caution. It is known that the
vast majority of patients with heel pain have improvement within
a few months after the onset of symptoms. Clinical evidence of the
efficacy of any treatment during this time is difficult to obtain5. The self-limiting character of the
disease therefore has to be considered as does the fact that spontaneous
improvement is difficult to distinguish from a long-lasting effect
of low-energy extracorporeal shock-wave application.
No side effects were recorded following the application of the low-energy
extracorporeal shock waves in our patients. This clinical experience
is supported by previous histological and magnetic resonance imaging-based
studies16,27. In contrast, high-energy
shock waves, which are also used for the treatment of heel pain17,24-26, may produce side effects
such as periosteal detachment and small fractures of the inner surface
of the cortex28.
In conclusion, the current pilot study revealed dose-related effects
of low-energy extracorporeal shock-wave therapy in patients with
chronic plantar fasciitis. The therapy with three applications of
1000 impulses appeared to be a useful, noninvasive treatment method
with negligible side effects that reduced the necessity for a surgical
procedure. Nevertheless, low-energy shock-wave application cannot
be recommended as a first-line procedure for chronic heel pain.
Although the United States Food and Drug Administration recently approved
a shock-wave device for therapy for heel pain29,
additional controlled studies are still needed to verify the results
of this study and to define the precise role of this new modality
in the treatment of chronic plantar fasciitis.