Stenosing tenosynovitis of the first dorsal compartment of the wrist was
originally described by Fritz de Quervain in
18951. Many reports
have described successful outcomes after nonoperative treatment of de Quervain
disease2-11.
The rate of a successful result after a single steroid injection has been
reported to range from 45% to
81%3-8.
In clinical practice, some physicians use nonsteroidal anti-inflammatory drugs
with or without splinting as first-line treatment of this
condition12.
However, there is uncertainty as to whether supplemental oral nonsteroidal
anti-inflammatory medications improve the effectiveness of steroid injection
in de Quervain disease. Furthermore, few studies have identified the
predictive factors for the recurrence of this
condition6,8.
Nimesulide (Eurodrug Laboratories, The Hague, The Netherlands) is a
nonsteroidal anti-inflammatory drug of the sulfonamide class, which is
marketed in European and Asian countries. It is administered orally (usually
200 mg/day) or rectally (400 mg/day) for a variety of inflammatory and pain
states. It exerts some actions in addition to inhibiting cyclooxygenase that
contribute to its anti-inflammatory effect: it inhibits neutrophil activation
and also exhibits antioxidant properties. Nimesulide has been found to be a
selective cyclooxygenase-2 inhibitor in human beings at clinically recommended
doses13. It is
associated with a low prevalence of adverse side effects, especially in the
gastrointestinal
tract13.
We conducted a randomized, double-blind, placebo-controlled trial to assess
the effectiveness and safety of triamcinolone injection and nimesulide
compared with triamcinolone injection and placebo in 160 patients with de
Quervain disease. We tested the hypothesis that there are no significant
differences in the success rates when this condition is treated with
triamcinolone injection with or without supplemental oral nimesulide. We also
studied the probability of the recurrence of symptoms after triamcinolone
injection with or without oral nimesulide and investigated whether factors
such as patient age, gender, duration of symptoms, dominant hand involvement,
previous steroid injection, and the presence of crepitation were predictive
factors for the recurrence of the disease.
Study Population
We conducted a randomized, double-blind, placebo-controlled trial at
Srinagarind Hospital, Khon Kaen University. The study was approved by the
institute's Ethics Committee, and all patients provided written informed
consent before participation. In order to establish the diagnosis of de
Quervain disease, the patients were required to have all of the following
three physical findings: radial styloid tenderness, pain over the first
extensor compartment on resisted thumb abduction or extension, and a positive
Finkelstein test. The Finkelstein test was performed by having the patient
flex his or her fingers over the involved thumb during passive ulnar deviation
of the wrist. Patients were excluded if they had contrain-dications to the use
of nonsteroidal anti-inflammatory drugs, had had a lidocaine or steroid
injection, were pregnant or lactating, had inflammatory joint disease of the
affected wrist, had a previous wrist injury, had taken nonsteroidal
anti-inflammatory drugs during the previous two weeks, or had received a
steroid injection during the previous two months.
Randomization and Treatment
The patients were randomly allocated, with use of a table of random
numbers, into the nimesulide or the placebo group. All patients in both groups
received a 1-mL (10-mg) injection of triamcinolone acetonide (Bristol-Myers
Squibb, Princeton, New Jersey) and 0.5 mL of 1% lidocaine (Astra Pharma,
Mississauga, Ontario, Canada). After the injection, a sealed opaque envelope
that contained the allocation to treatment with either 100 mg of oral
nimesulide twice daily for seven days (the nimesulide group) or placebo
tablets for seven days (the placebo group) was opened. In order to preserve
the blinding, the injection was given by a physician other than the
evaluators.
The patients were strongly encouraged to avoid gripping, grasping, and
lifting heavy objects with the injected wrist over the next three days. All
patients were advised to use ice to control the postinjection pain for
twenty-four to forty-eight hours. Splinting was not used. All patients were
discouraged from receiving other steroid injections, nonsteroidal
anti-inflammatory drugs, and analgesic drugs.
Data Collection
At the time of the first visit, demographic data on the patient, including
age, gender, the duration of symptoms before treatment, the side affected,
dominant hand involvement, new or recurrent disease, previous treatment,
crepitation, and pain intensity on performing a Finkelstein test, were
recorded. The initial follow-up examination was performed one week after the
injection to record the presence of and provide treatment for any adverse
reactions. A physical examination was performed three weeks after the
injection to evaluate the outcome. A physician who was blinded with respect to
the treatment allocation conducted all follow-up examinations.
After completion of the trial, subsequent follow-up examinations were done
every six months, or when the symptoms recurred, to evaluate the recurrence
rate. We conducted a final follow-up evaluation of 155 patients at an average
(and standard deviation) of 13.6 ± 5.5 months (range, six to
twenty-four months) after the initial treatment. Patient follow-up included a
physical examination, telephone interview, or completion of a written
questionnaire.
Outcome Measures
The primary outcome measures that were evaluated were the three physical
signs of de Quervain disease: (1) tenderness over the radial styloid, (2) pain
reproduced by resisted thumb abduction and extension, and (3) a positive
result on the Finkelstein test.
Pain aggravated by the Finkelstein test was recorded at the initial visit
on a 100-mm visual analogue
scale14, with 0
indicating no pain and 100 indicating extremely intense pain. At three weeks
after the injection, patients were asked to compare the pain with the pain
before treatment. Their responses were rated on a 200-mm-long visual analogue
scale with a 0 at the center that indicated no change in pain. Values to the
right represented deterioration or an increase in pain with an end point at
100 mm denoting the worst pain ever. Values to the left indicated improvement
or decrease in pain, with an end point at —100 mm indicating no
pain15.
The patient's response was graded as a success if all three physical signs
were completely resolved and the patient had at least 90% improvement in the
pain score. Patients with any one of the three signs and/or <90%
improvement in the pain score were graded as a failure.
The following secondary outcomes were also assessed: (1) the safety of the
drug, i.e., the adverse events reported by the patient or observed by the
investigator as recorded in the patient's medical chart; (2) the probability
of recurrence; and (3) the possible predictive factors for recurrence of the
disease.
Statistical Analysis
Before the beginning of the study, we estimated that a sample size of at
least seventy-five patients in each group would give an 80% power (ß =
0.2) of detecting a difference of 20% in the proportion of patients with a
successful outcome between the two groups at the 5% significance level
(a = 0.05) on a two-sided significance
test16,17.
Descriptive statistics (means, standard deviations, and frequencies) were
performed on all variables. The null hypothesis of no difference between the
two study groups was tested. Categorical outcomes were compared with both the
z test of proportions and a 95% confidence interval for the difference between
proportions. Continuous outcomes were compared with the Student t test. The
Kaplan-Meier method was used to analyze the risk of recurrence between the
groups within twenty-four months after the
injections18. The
log-rank test was used to determine whether there was a significant difference
between the groups. The Cox proportional-hazards model was used to adjust for
possibly confounding covariates such as patient age, gender, duration of
symptoms, dominant hand involvement, previous steroid injection, and the
presence of crepitation by means of a backward stepwise elimination procedure.
We calculated the relative risks and 95% confidence intervals associated with
the recurrence of symptoms. The alpha level for all analyses was set at 0.05.
End-point data were analyzed according to the intention-to-treat principle.
Statistical analyses were performed with use of SPSS software (version 11.5;
SPSS, Chicago, Illinois) and Stata software (version 7.0; Stata, College
Station, Texas).
Baseline Characteristics
Between April 2001 and March 2003, 168 consecutive patients were diagnosed
as having de Quervain disease. Eight of those patients were excluded for the
following reasons: refusal of consent (five patients), rheumatoid arthritis of
the affected wrist (two patients), and previous wrist injury (one patient).
Thus, a total of 160 patients (165 wrists) were included in the trial, with
eighty patients (eighty-three wrists) randomized to the nimesulide group and
eighty patients (eighty-two wrists), to the placebo group
(Fig. 1). All patients
completed the three-week trial. In the nimesulide group, seventy-seven
patients (96%) were followed for at least six months; fifty-six (70%), for
twelve months; thirty-three (41%), for eighteen months; and fifteen (19%), for
twenty-four months. In the placebo group, seventy-eight patients (98%) were
followed for at least six months; fifty-five (69%), for twelve months; thirty
(38%), for eighteen months; and twelve (15%), for twenty-four months.
Table I shows the
demographic and baseline characteristics of the participants. Both groups were
similar with respect to age, gender, duration of symptoms, dominant hand
involvement, classification of the disease (new or recurrent), previous
steroid injection, the presence of crepitation, and pretreatment pain scores
on the visual analogue scale.
Primary Outcome
Success Rate
The overall results are shown in Table
II. The success rates, assessed at three weeks after the first
injection (the primary outcome), were 67% in the treatment group and 68% in
the placebo group. The overall success rates at the final follow-up
examination after a single injection or multiple (two, three, or four)
injections were 95% in both groups. No significant differences between the two
groups were found with respect to the success rates or the pain scores on the
visual analogue scale after treatment.
Secondary Outcomes
Adverse Reactions
The overall prevalences of adverse reactions are shown in
Table III. No significant
differences were detected between the two groups with respect to adverse
reaction rates for either drug. There were no serious adverse effects of
either triamcinolone or nimesulide. The most common adverse reaction to the
triamcinolone injection was pain after the injection. Infection and tendon
rupture did not occur in our patients, even in the forty-three who had
multiple injections, with an average follow-up period of 15.3 ± 6.2
months after the initial treatment. The most common adverse reaction to
nimesulide was dyspepsia.
Probability of Recurrence
The results of a Kaplan-Meier survivorship analysis presenting the
proportion of patients who had no recurrence at each time-point are shown in
Figures 2-A and
2-B. Twenty-four patients
(twenty-seven wrists) in the nimesulide group and twenty-eight patients
(thirty wrists) in the placebo group had recurrence of the disease at the
final follow-up evaluation. The results of the log-rank test (chi-square =
0.39, p = 0.53) showed no significant association between the recurrence of
disease and the type of treatment. The median time of recurrence was five
months in the nimesulide group and four months in the placebo group.
Prediction of Recurrence
Bivariate analysis from the Cox regression model revealed that recurrence
of symptoms was significantly related to the presence of crepitation (relative
risk, 2.10; 95% confidence interval, 1.21 to 3.67) at the time of
presentation. After adjustment for the covariates of age, gender, duration of
symptoms, dominant hand involvement, and previous injection in the Cox
proportional hazards model, the relative risk of recurrence when crepitation
was present was 2.13 (95% confidence interval, 1.19 to 3.80). Other factors
were not shown to be predictive for the recurrence of disease
(Table IV).
The results of treatment of de Quervain disease with nonsteroidal
anti-inflammatory drugs have been previously
reported7,9,12.
In clinical practice, oral nonsteroidal anti-inflammatory drugs are sometimes
added after steroid injection. We could not find a report on the results of
treatment of de Quervain disease with steroid injection combined with an oral
nonsteroidal anti-inflammatory drug in the literature. As far as we know, this
study is the first randomized, placebo-controlled trial of the effectiveness
of the addition of a nonsteroidal anti-inflammatory drug to the treatment of
de Quervain disease with steroid injection.
We found no advantage with the addition of nimesulide to the steroid
injection in the present study, despite the large number of patients. Clarke
et al.19 reported
that the histopathology of de Quervain disease is not characterized by
inflammation but by thickening of the tendon sheath and by the accumulation of
mucopolysaccharides. These findings may explain why nimesulide provided no
additional value when combined with the steroid injection.
In the present study, the most common adverse event was pain after the
injection. Subcutaneous tissue atrophy or depigmentation are more likely if
large or repeated doses of a long-acting steroid are
given20, but we
rarely observed this morbidity in our Asian patients.
One of the serious adverse reactions of steroid injection in general is
tendon rupture. It has been recommended that multiple injections be avoided to
prevent this
complication21. We
found that, among the forty-three patients (fifty-seven wrists) who had
multiple injections, none had a tendon rupture. This low rate is similar to
that in another
report7 and implies
that multiple injections may be safe when direct injection into the tendon
substance is avoided.
The secondary objective of our study was to assess the probability of
recurrence and the possible predictive factors for recurrence of de Quervain
disease after injection. To our knowledge, this is the first randomized
clinical trial that reports the recurrence rate in terms of the probability of
recurrence. Several factors have consistently been associated with the
development of de Quervain tenovaginitis: female gender, perimenopausal age,
variation in the anatomy of the first extensor compartment, acute trauma, work
or repetitive trauma, and rheumatic disorders. However, these factors might
not be related to the recurrence of the disease. Witt et
al.6 reported that
age, hand dominance, and the duration of symptoms before treatment did not
affect the outcome, but the results were not adjusted for other confounding
factors. Crepitation in de Quervain disease is the feeling or sensation of
triggering when the affected thumb is moved, especially in abduction or
extension. It is usually found in patients who present with swelling and
thickening of the first extensor compartment sheath. This study demonstrated
that crepitation in the first extensor compartment was the only factor studied
that might be predictive of the recurrence of disease. This may be the case as
crepitation implies extensive stenosis of the first dorsal compartment or
nodule formation in the tendons.
The median time to recurrence in this study was five months in the
nimesulide group and four months in the placebo group. This time was shorter
than that in a previous
report7. One
possible reason for the earlier recurrence in our patients was the type and
the amount of steroid used. For instance, methylprednisolone had greater
efficacy and a more sustained clinical effect than did triamcinolone, as was
used in this study.
Our study did not investigate whether concurrent medical conditions had an
effect on the outcomes. Weiss et
al.8 observed that a
trend toward failure was noted in patients with diabetes. We did not classify
the work status of our patients, and we also excluded patients who had
rheumatic disorders. The results of treatment in these particular conditions
need further investigation.
In conclusion, steroid injection alone was safe and effective in the
treatment of de Quervain disease, but the oral administration of nimesulide
did not provide any additional benefit beyond that of the injection.
Crepitation in the first extensor compartment during thumb extension or
abduction was found to be a predictive factor for recurrence of the disease.
?