Prior to patient participation, the study was approved by the ethical
committees of both hospitals involved in the study. Forty consenting patients
with chronic tennis elbow were recruited to participate. All patients had had
clinically diagnosed tennis elbow for more than six months, and all had
received at least one corticosteroid injection and a full course of
physiotherapy. Tennis elbow was diagnosed clinically on the basis of
well-localized pain over the lateral epicondyle and increased pain at this
site on resisted wrist extension. All patients had a full range of movement
with normal anatomical alignment of the elbow. Patients were excluded if they
had any neck symptoms, previous elbow surgery, systemic disease, or a negative
provocation test for tennis elbow.
The patients were block randomized into two groups. One group was to
receive an injection of 50 units of botulinum toxin type A reconstituted in 2
mL of normal saline solution (Allergan, Irvine, California), and the other was
to receive a placebo injection of 2 mL of normal saline solution only. Block
randomization is a method used to prevent unequal treatment-group sizes, since
it ensures that at certain points the numbers of participants in each group
are equal. For example, a sample of forty patients in a trial is split into
ten groups of four patients. Each subset of four patients is then randomized,
with two patients assigned to the treatment group and two assigned to the
control group. This process is repeated as the trial progresses. Therefore, if
the trial is stopped prematurely, the numbers of patients in each group will
be approximately equal. Fifty units was selected as the strength of botulinum
toxin type A to be consistent with the strength used in the only previous
studies of this kind to show an apparent therapeutic
effect6,7.
The solutions were drawn up independently, and the injection was performed by
a physician who was blinded to its content. Both injection solutions were
colorless and of equal temperature to maintain blinding. The injection was
administered 5 cm distal to the area of maximal tenderness at the lateral
epicondyle, in line with the middle of the wrist, with a 21-gauge needle
inserted deep to the forearm fascia (Fig.
1).
Prior to the injection, patients completed the Short Form-12 (SF-12) health
status
questionnaire13.
Pain was assessed on a linear visual analogue
scale14 ranging
between 0 and 10, with 0 representing no pain and 10 representing the worst
pain imaginable. Grip strength was recorded, in kilograms, as an average of
three readings with a validated Jamar dynamometer on both the affected and
unaffected sides. The patients were contacted by telephone at one week to
ensure that no adverse reactions had occurred from the injection. The patients
were examined in our clinic three months after the injection, and the same
data as had been obtained before the injection were recorded. No adjunctive
treatment, such as physiotherapy, was prescribed during the three-month
period.
Statistical analysis was performed with p < 0.05 as the level of
significance.
Of the forty patients who were recruited, nineteen were randomized to
receive botulinum toxin type A and twenty-one were randomized to receive a
placebo injection. Eighteen patients in the botulinum toxin group completed
the study, whereas one patient had an operation, performed by a different
clinician, before the three-month follow-up period was completed. Nineteen
patients in the placebo group completed the study, whereas one patient had an
operation before the completion of the three-month follow-up period and one
failed to return for follow-up. The mean age of the patients was forty-eight
years (range, thirty-five to seventy-one years). There were twenty-one men and
nineteen women.
An extensor lag was recorded when the patient was unable to fully extend a
digit actively. Twelve of the eighteen patients in the botulinum toxin group
had a transient 2-cm extensor lag of the long finger that was noticeable at
one week. All extensor lags disappeared by three months after the injection.
Two patients, one of whom repaired watches and the other of whom worked as a
typist and also played the guitar, reported that the extensor lag caused
functional problems.
Grip Strength (Table
I)
Prior to the injection, there was no significant difference in the mean
grip strength of the affected limb between the botulinum toxin group (mean,
27.6 kg) and the placebo group (mean, 30.9 kg) (p = 0.46). At three months
after the injection, there was also no significant difference between the
groups with regard to the mean grip strength (p = 0.90), the absolute change
in the mean grip strength compared with the preinjection level (p = 0.20), or
the percentage change in the mean grip strength (p = 0.08).
There was no significant difference in the mean grip strength of the
unaffected limb between the botulinum toxin group (mean, 34.6 kg) and the
placebo group (mean, 36.8 kg) prior to the injection (p = 0.56). At three
months after the injection, there was also no significant difference between
the groups with regard to the mean grip strength (p = 0.83), the mean absolute
change in grip strength (p = 0.49), or the mean percentage change in grip
strength (p = 0.56).
Pain Assessed on the Visual Analogue Scale
(Table II)
Prior to the injection, there was no significant difference in the mean
score on the visual analogue scale between the botulinum toxin group (mean,
8.80 cm) and the placebo group (mean, 9.43 cm) (p = 0.67). At three months
after the injection, there was also no significant difference in the scores
between the botulinum toxin group (mean, 11.35 cm) and the placebo group
(mean, 12.46 cm) (p = 0.54) or in the mean percentage change in the score (p =
0.80), the mean percentage fraction change in the score (p = 0.81), or the
mean change in the absolute score in centimeters (p = 0.81).
Quality of Life Measured with the SF-12
(Table III)
Prior to the injection, there was no significant difference in the mean
physical function SF-12 score between the botulinum toxin group (mean, 35.70
points) and the placebo group (mean, 41.08 points), although the difference
approached significance (p = 0.06). At three months after the injection, there
was also no significant difference between the mean physical function score of
the botulinum toxin group (mean, 32.38 points) and the placebo group (mean,
38.61 points) (p = 0.16), between the mean percentage change in the physical
function score (p = 0.84), or between the mean absolute change in the physical
function score (p = 0.83).
There was no significant difference in the mean mental function SF-12 score
between the botulinum toxin group (mean, 46.84 points) and the placebo group
(mean, 50.69 points) prior to the injection (p = 0.30). At three months after
the injection, there was also no significant difference in the mean mental
function score between the botulinum toxin group (mean, 44.61 points) and the
placebo group (mean, 48.87 points) (p = 0.42) or in the mean percentage change
(p = 0.80) or absolute change (p = 0.93) in the mental function score.
Recent reports have suggested that botulinum toxin type A may be as
effective as surgery in the treatment of chronic tennis
elbow7. However, as
surgery is required in only a minority of patients, botulinum toxin treatment
needs to be assessed as a treatment regimen on its own. With this in mind, we
investigated the effects of injection of botulinum toxin type A on tennis
elbow in a double-blind, placebo-controlled trial. We concluded that it was
not clinically effective, and therefore we accepted our null hypothesis.
Twelve of the eighteen patients who received an injection of botulinum
toxin type A had a transient extensor lag of the long finger. This did not
cause a functional problem in the majority of these patients, but it was not
tolerable to two individuals who required intricate maneuvers as part of their
hobby or work. Although six of the eighteen patients did not have an extensor
lag, they had subjective weakness in the extensor mechanism, possibly as a
result of the injection temporarily paralyzing the muscle. Keizer et
al.7 looked for
evidence of an extensor lag in their twenty patients to confirm that the
injection had been located correctly, and they repeated the injection in eight
patients who did not have a finger drop. Six of our patients did not
demonstrate a finger drop, which perhaps suggests that the injection was not
in the correct position; however, reanalysis of our results showed no
difference between the groups with and without a finger drop with regard to
any of the assessed outcome measures. The presence of an extensor lag did not
compromise the blinding of the study. The patients completed marking the
visual analogue scale and filling out the SF-12 form before they were examined
by the blinded assessor, and the patients picked up the Jamar dynamometer
before the blinded assessor observed them so that the assessor would not be
able to detect an extensor lag.
The patients treated with the botulinum toxin had an average 10%
improvement in grip strength on the affected side, whereas the preoperative
and postoperative mean grip strengths were the same in those treated with the
placebo; however, this difference did not reach significance (p = 0.08; 95%
confidence interval, —2.31 to 35.64). No substantial difference in the
grip strength on the unaffected side was seen between the groups, with a 4%
improvement in the botulinum toxin group and a 1% improvement in the placebo
group. Both groups also showed similar improvements in the scores for pain and
quality of life. However, because there was a nearly significant difference
between groups with regard to the percentage change in grip strength on the
affected side (p = 0.08), one must question the number of patients recruited
in the study. In order to show an assumed clinically relevant difference of 2
kg, with an 80% power and a type-I error of 5%, 300 patients would be needed
(150 in each treatment group). If a subsequent study were to be planned, it
would be advisable to use the mean change in grip strength as a primary
variable because the variability for this parameter was lowest in the present
study and it came close to identifying a significant difference between groups
(p = 0.196). In addition, grip strength is a quantifiable measurement of
effect, unlike the more subjective measurement of pain. ?