вход по аккаунту


The efficacy and toxicity of combination therapy in rheumatoid arthritis. a meta-analysis

код для вставкиСкачать
Number 10, October 1994, pp 1487-1491
0 1994, American College of Rheumatology
A Meta-Analysis
Objective. To evaluate the efficacy and toxicity of
combination therapy, compared with single second-line
drug therapy, in rheumatoid arthritis.
Methods. This study was a meta-analysis of published trials that evaluated combinations of full-dose
second-line drugs and compared them with single
second-line drugs at full dose. Using a random effects
model, we summarized the difference between improvement with combination therapy and improvement with
single-drug therapy.
Results. Five trials that met inclusion criteria,
which contained 749 entering patients and 516 completing patients, were identified. The mean
SEM difference in improvement in tender joint count between
combination and single-drug therapy at end of trial
(24-52 weeks) was 2.4 f 0.7 joints (out of 60) (P C
0.001). At end of trial the difference between therapies
in swollen joint counts was 1.0 f 1.2 joints (P= 0.42).
The difference in grip strength improvement was 3.7 f
4.3 mm Hg (P = 0.40), and for erythrocyte sedimentation rate it was 3.4
3.1 mmhour (P = 0.27). In
general, the differences in efficacy between combination
and single-drug therapy were clinically marginal. Nine
percent more combination therapy-treated patients experienced side effect-related discontinuation of therapy
than patients receiving single-drug therapy (P= 0.008).
Conclusion. Combination therapy, as it has been
Supported by Multipurpose Arthritis and Musculoskeletal
Disease Center grant AR-20613 from the NIH.
David T. Felson, MD, MPH: Boston University Arthritis
Center and Boston City and University Hospitals, Boston, Massachusetts; Jennifer J. Anderson, PhD: Boston University Arthritis
Center; Robert F. Meenan, MD, MPH, MBA: Boston University
Arthritis Center and School of Public Health.
Address reprint requests to David T. Felson, MD, MPH,
Boston University Arthritis Center, 80 East Concord Street, Boston
University School of Medicine, Boston, MA 02118.
Submitted for publication December 27, 1993; accepted in
revised form April 23, 1994.
used in recent clinical trials, does not offer a substantial
improvement in efficacy, but does have higher toxicity
than single drug therapy. These combination therapy
regimens are not recommended for widespread use.
Other more aggressive regimens with additional drugs
or higher drug doses than have been studied might
be more efficacious, but with an even higher rate of
Multiple-drug therapy with nonsteroidal antiinflammatory drugs, second-line drugs, and often steroids is common in rheumatoid arthritis (RA). Furthermore, it has been suggested that therapy with only 1
second-line drug may be less effective than concurrent
therapy with more than 1 second-line drug. Combination therapy has been defined as “two or more slowacting (or second-line) antirheumatic drugs used together” (1). Since single second-line drug therapy (in
conjunction with nonsteroidal drugs) is often only
modestly effective in the treatment of RA, combination therapy has an inherent appeal. Furthermore, if 2
second-line drugs have different toxicities and target
the pathobiology of RA at different sites, their combination might theoretically offer added efficacy with
little extra toxicity. Combination chemotherapy in
oncology provides a prime example of the potential
value of this concept.
However, individual trials that have compared
combination therapy with single second-line drug therapy have not generally shown differences in efficacy,
and any differences demonstrated have often not
reached statistical significance. This may be, in part,
because comparing 2 active regimens requires a much
larger sample size than comparing a single drug with
placebo (2). We undertook the present meta-analysis
to evaluate the overall efficacy and toxicity of combination therapy in RA. The large, multi-study pooled
sample size of the meta-analysis would not only be
likely to detect any significant difference between
combination therapy and single second-line drug therapy, but would likely provide a more precise estimate
of that difference than do individual trials. Furthermore, a meta-analysis can evaluate the consistency of
individual trials, and can summarize the toxicity experience with combination therapy.
While many trials of combination interventions
have been performed, some have been uncontrolled
( 3 4 9 , and others have studied combinations in which
each drug was given at less than a therapeutic dosage.
These latter trials (7) have not shown that combination
therapy had efficacy greater than that of single secondline drug given at full dose. Therefore, we focused our
meta-analysis on trials in which the combination therapy included 2 traditional second-line drugs, each
given at full therapeutic dose.
This meta-analysis obviously is a summary of
drug comparisons which may not be equivalent. It is an
attempt to evaluate the general efficacy and toxicity of
Combination regimens studied so far in clinical trials. A
similar, but much larger, endeavor has been completed
for combination chemotherapy versus monotherapy in
breast cancer (8), in which combinations studied were
similarly heterogeneous.
To identify trials of combination therapy, we performed MEDLINE and bibliographic searches for all RA
trials published through December 1992. As part of a larger
RA meta-analysis project, we have kept an updated list of
trials in which we capture studies not only through
MEDLINE and bibliographic searches, but also through
hand searches of 5 international rheumatology journals in
which the majority are published. Furthermore, we scan all
abstracts of the American College of Rheumatology (ACR)
and British Society of Rheumatology meetings yearly to find
additional trials. Once a report of a trial is found and
photocopied, we cut out the methods and results sections,
blinding the methods section to any results reported and
blinding the results section to the individual drugs or regimens studied.
For the present study, the methods sections of reports on the above-described list were reviewed by one of us
(RFM) and the results sections by another (JJA) to determine whether they fulfilled inclusion criteria, as follows: 1)
Randomized clinical trial of RA patients at least 18 years of
age. 2) One treatment a combination of at least 2 full-dose
second-line drugs started concurrently. A full dose was
defined (2) as the accepted minimal clinically effective dosage if the drug were used as a single second-line drug. The
following drugs and minimal doses were deemed acceptable:
auranofin 6 mg/day, hydroxychloroquine 200 mg/day , chloroquine 250 mglday, gold 50 mglweek, methotrexate 7.5
mglweek, azathioprine 1 mglkglday, sulfasalazine 2 gmiday,
and D-penicillamine 500 mglday. 3) The other treatment
arm(s) had to be one of the single second-line drugs listed
above, also at full therapeutic dose. 4) Numerical data on
either the start and end values for the treated group or the
change during treatment for each treatment group had to be
provided in the published report of the trial. For efficacy
outcomes, at least one of the ACR core set of outcome
measures for RA (9) had to be reported. These include
tender and swollen joint count, patient and physician global
assessments, patient assessment of pain, self-reported disability, and measurement of an acute-phase reactant. Since
disability was not reported in any of the trials, grip strength
was deemed an acceptable proxy measure for this. We
determined trial eligibility based on the provision of efficacy
rather than toxicity data.
After determining whether the trial met inclusion
criteria, we extracted the data. For efficacy, we extracted
numerical results on the ACR core set measures (see above);
reported here are the results on measures that were reported
for at least 2 trials, tender joint count, swollen joint count,
erythrocyte sedimentation rate (ESR), patient global assessment (rescaled to loo), and grip strength. To evaluate side
effects of the combination and single second-line drugs, we
used 3 different definitions of toxicity. First, we evaluated
the proportion of all entering patients who dropped out due
to side effects. Second, for all side effects noted in the report
of the trial, we gave each a score based on the toxicity index
described by Fries et a1 (lo), modified by us after reevaluating the weights for diarrhea and with more comprehensively
scored laboratory side effects (1 1). This toxicity index ranks
each side effect on a scale of 0-10, with the side effects close
to 0 being mild and those close to 10 being severe or even life
threatening. Third, we evaluated the proportion of patients
who developed severe toxicity, which we defined as a
toxicity index score of at least 7 on the 10-point scale. Such
toxicities tended to be visceral or organ-based toxicities and
included thrombocytopenia, leukopenia, acute renal failure,
and severe elevation of the serum glutamic oxaloacetic
transaminase (aspartate aminotransferase) level, to 2 1,000
For each trial we evaluated the difference in improvement between combination therapy-treated patients
and patients treated with a single second-line drug. For trials
in which there were 2 single-drug arms, we averaged the
response of these 2 groups and compared this average
response with the response of those treated with combination therapy. We summarized study results using a random
effects approach (12). Additional analyses were performed
using the “most efficacious” single arm (with this drug
identified from our previous meta-analysis [2]); in these
analyses, the efficacy differences seen between combination
and single second-line drugs were even less than the differences reported here. Chi-square-based tests suggested no
significantheterogeneity among trials (P > 0.05,2-tailed) for
either joint count measure or for toxicity-induced dropout
rate or toxicity index score.
Of the 214 published second-line drug trials we
reviewed, most were eliminated because they did not
include a combination therapy arm. There were 13
trials with combination therapy arms but in 4, drugs
studied were not of interest based on our inclusion
criteria, and in 1 each, the study was not randomized,
drugs were not started concurrently, 1 drug in the
combination was given at a subtherapeutic dosage,
and the report provided insufficient detail on outcomes. We found 5 trials that met the inclusion criteria
(13-16) (Table I). Of 749 patients entering these trials,
only 516 completed them (68.9%). Since the reports
primarily included analyses of trial completers, we
analyzed efficacy in this way. Trials ranged from 24
weeks to 52 weeks (mean 40.4 weeks) and the patients,
on average, had 4 years of disease (range 2-8 years).
The mean tender joint count at initiation was 19.9
(Ritchie Articular Index used in 3 trials, 44 joints in 1,
and 60joints in l), and the initial ESR in the 3 trials in
which it was reported averaged 48.6 mdhour. Trial
quality scores ranged from 12 to 20 (mean 16.6) on a
scale of &20, using a published scoring method (2).
In all trials, combination therapy-treated patients had more tender joint count improvement than
patients treated with single second-line drugs (Figure
lA), although in 3 of the 5 trials the difference between
combination therapy and single drug therapy was not
statistically significant. Overall, combination therapy
led to an improvement in a mean of 2.4 joints more
than single second-line drug therapy (P < 0.001). For
ESR (Figure 13), the combination drug therapy did
not always lead to greater improvement than did single
second-line therapy, and the pooled difference between combination and single second-line drug therapy was not significantly different from 0.
In fact, when pooled differences for all the
efficacy outcomes in these trials were evaluated, tender joint count was the only one in which combination
therapy was significantly more efficacious than single
second-line drug therapy (Table 2). The pooled differTable 1. Trials meeting inclusion criteria
Slngle Drug
F yors
Gibson et al.
Taggart et al.
Scott et al.
Williams et ai.
Willkens et al.
Number of Tender Joints
Single Drug
Gibson et al.
Taggart et al.
Scott et al.
Williams et al.
- 10
ESR (mm/hr)
Figure 1. Differences between combination therapy and single-drug
therapy, in change in tenderjoint count at end of trial (A) and change
in erythrocyte sedimentation rate (ESR) at end of trial (B). Bars
represent the mean and the 95% confidence interval around that
mean. Values to the left of 0 (negative values) indicate that the
results favored single-drug therapy; values to the right of 0 indicate
that the results favored combination therapy. Trials evaluated were
those reported in references I (Scott et al), 13 (Gibson et al), 14
(Taggart et al), 15 (Williams et al), and 16 (Willkens et al).
Authors, year
Gibson et al,
1987 (13)
Taggart et al,
1987 (14)
Scott et al,
1989 (1)
Williams et al,
1992 (15)
Willkens et al,
1992 (16)
Drugs studied
C hloroquine/D-penicillaminevs.
D-penicillamine vs.
c hloroquine
Goldlhydroxychloroquine vs.
Methotrexate/auranofin vs.
methotrexate vs. auranofin
Methotrexate/azathioprine vs.
methotrexate vs. azathioprine
No. of
ence between combination therapy and single secondline drug therapy for other outcomes, such as swollen
joint count (difference of 1.0 joint at end of trial), grip
strength (difference of 3.7 mm Hg at end of trial), and
patient global assessment (difference of 6 on a scale of
100 at end of trial) were not statistically significant and
were clinically marginal.
Side effects were consistently greater in combination drug-treated patients than in patients treated
Table 2. Difference in improvement with combination treatment vs. improvement with single
second-line drug treatment
Difference at 6 months
Mean SEM
(no. of trials)
Efficacy outcome
Tender joint count, no. of joints
Swollen joint count, no. of joints
Grip strength, mm Hg
Patient global assessment, 0-100
ESR, mm/hour*
f 0.7
f 1.2
? 6.1
3.9 (2)
4.3 (2)
Difference at end of trial
Mean ? SEM
(no. of trials)
* 0.7 (5)
5 1.2
f 4.3
f 3.1
3.1 (4)
* Erythrocyte sedimentation rate.
with single second-line drugs (Table 3). In each study,
combination drug-treated patients had higher side
effect-related dropout rates than did those treated with
single second-line drugs. The pooled difference between combination therapy and single drug therapy
was 9% of all treated patients (P = 0.008). When we
weighted each side effect by its toxicity index score
(n = 5 trials), we found that the toxicity score for the
combination therapy-treated patients was 0.7 ? 0.1
higher (mean ? SEM) than that for those treated with
single second-line drugs (P < 0.001) (Table 4). However, severe side effects were no more common in the
combination therapy-treated patients than in those
treated with single second-line drugs (mean f SEM
difference in percent with severe toxicity -0.1 ?
1.2%), in part because severe side effects were generally rare in these clinical trials.
Our results suggest that the combination therapies that have been selected for study in clinical trials
are not substantially more efficacious than single
second-line drug treatments. The only significant efficacy difference we found was in the tender joint count.
Combination therapy patients experienced, on average, improvement in 2.4 more tender joints than those
treated with single second-line drugs. This difference
may be clinically meaningful, but in a meta-analysis of
the effect of single second-line drugs (1 I), we found
that second-line drugs such as methotrexate, gold, and
sulfasalazine induce an average improvement of 7 or 8
more tender joints compared with placebo.
The combination therapy regimens studied consistently had higher toxicity rates than did the single
second-line drug regimens. The increased toxicity
index score of 0.7 for combination therapy compared
with single second-line drug therapy is equivalent to
the difference in toxicity between methotrexate and
placebo or between sulfasalazine and placebo, found
in an earlier meta-analysis (11).
Even though the trials reported here assessed
combination therapy of full-dose second-line drugs,
the individual agents tended to be prescribed at the
minimal effective therapeutic dosage. For example,
the study by Willkens et a1 (16) used, as its most
aggressive combination regimen, a dosage of 100 mg of
azathioprine daily and 7.5 mg of methotrexate weekly,
compared with single-drug regimens of 150 mg of
azathioprine daily and 15 mg of methotrexate weekly.
Table 3. Percent of patients who dropped out due to toxicity:
combination vs. single-drug regimens*
Trial (ref.)
Single drug
Gold/hydroxychloroquine (1)
Methotrexate/auranofin (15)
Methotrexate/azathioprine (16)
19 (SSZ alone)
20 (gold alone)
Mean difference in rate of
dropout due to toxicity$
* All values are percents. SSZ = sulfasalazine.
't Average dropout rate for patients taking single study drug.
$: Overall difference, all patients in trials of combination therapy vs.
all patients in trials of single-drug therapy.
Table 4. Toxicity of combination treatment vs. single-drug treatment in a meta-analvsis of 5 trials. assessed bv various measures
Difference at end of trial
Toxicity measure
% who dropped out
due to toxicity
Modified Fries toxicity
score, 0-10 scale*
% with severe toxicity
* See refs.
10 and 11.
* 1.2
The Cooperative Systematic Studies of the Rheumatic
Diseases Program trial of combination therapy with
methotrexate and auranofin (15) used a combination in
which the maximal dosage of methotrexate was 7.5 mg
weekly. It is possible that higher doses of each drug in
the combination could have been more effective, but
the toxicity might also have been greater.
Our meta-analysis was limited by the inconsistent measurement of ACR core set outcomes in the
trials, all of which were published before publication
of these measures. Inclusion of data on pain and
disability would have been helpful, and a single standardized outcome, such as a measure of improvement
or an index, would have enhanced the statistical power
of each study and of our meta-analysis.
Our findings do not necessarily mean combination therapy is not a promising therapeutic modality in
rheumatoid arthritis. They mean simply that the combinations tested so far are not, for the average patient,
of great value from the perspective of offering substantial additional efficacy with little extra toxicity. Other
regimens may be better. Indeed, uncontrolled studies
have suggested that the response to combination therapy in patients whose disease has not responded to
single second-line drug therapy can be dramatic (3,6).
However, these combinations, which have included
such regimens as cyclophosphamide, azathioprine,
and hydroxychloroquine (3) and methotrexate, cyclophosphamide, and hydroxychloroquine (6), were more
aggressive than the combination therapies we studied,
and may be more toxic. These reports are consistent
with case reports suggesting that patients who have
both RA and cancer have a marked response of their
RA when their cancer is treated with combination
chemotherapeutic regimens such as cytosine arabinoside, daunorubicin, and m-AMSA (17).
In summary, in a meta-analysis of trials evaluating combination therapy, we did not show a substantial clinical difference in efficacy between combination
drug regimens and single second-line drugs. The toxicity experience of patients treated with combination
therapy was worse than that of patients treated with
single second-line drugs. Combination therapy as it
has been used in clinical trials is not a valuable
therapeutic alternative for most patients with RA.
1. Scott DL, Dawes PT, Tunn E, Fowler PD, Shadforth MF,
Fisher J, Clarke S, Collins M, Jones P, Popert AJ, Bacon PA:
Combination therapy with gold and hydroxychloroquine in
rheumatoid arthritis: a prospective, randomized placebocontrolled study. Br J Rheumatol 28:128-133, 1989
2. Felson DT, Anderson JJ, Meenan RF: The comparative efficacy
and toxicity of second-line drugs in rheumatoid arthritis: results
of two meta-analyses. Arthritis Rheum 33: 1449-1461, 1990
3. McCarty DJ, Carrera GF: Intractable rheumatoid arthritis:
treatment with combined cyclophosphamide, azathioprine, and
hydroxychloroquine. JAMA 248: 1718-1723, 1982
4. Csuka ME, Carrera GF, McCarty DJ: Treatment of intractable
rheumatoid arthritis with combined cyclophosphamide, azathioprine and hydroxychloroquine: a follow-up study. JAMA 255:
2315-2319, 1986
5 . Biro JA, Segal AM, Mackenzie H , Mazanec DJ, Wilke WS: The
combination of rnethotrexate and azathioprine for resistant
rheumatoid arthritis (abstract). Arthritis Rheum 30 (suppl 4):
S18, 1987
6. Tiliakos NA: Low-dose cytotoxic drug combination therapy in
intractable rheumatoid arthritis: two years later (abstract).
Arthritis Rheum 29 (suppl4):S79, 1986
7. Bunch TW, O’Duffy JD, Tompkins RB, O’Fallon WM: Controlled trial of hydroxychloroquine and D-penicillamine singly
and in combination in the treatment of rheumatoid arthritis.
Arthritis Rheum 27:267-276, 1984
8. Early Breast Cancer Trialists’ Collaborative Group: Effects of
adjuvant tamoxifen and of cytotoxic therapy on mortality in
early breast cancer. N Engl J Med 319:1681-1692, 1988
9. Felson DT, Anderson JJ, Boers M, Bombardier C, Chernoff M,
Fried B, Furst D, Goldsmith C, Kieszak S, Lightfoot R, Paulus
H, Tugwell P, Weinblatt M, Widmark R, Williams HJ, Wolfe F:
The American College of Rheumatology preliminary core set of
disease activity measures for rheumatoid arthritis clinical trials.
Arthritis Rheum 36:729-740, 1993
10. Fries JF, Spitz PW, Williams CA, Bloch DA, Singh G, Hubert
HB: A toxicity index for comparison of side effects among
different drugs. Arthritis Rheum 33:121-130, 1990
11. Felson DT, Anderson JJ, Meenan RF: Use of short-term
efficacyitoxicity tradeoffs to select second-line drugs in rheumatoid arthritis: a meta-analysis of published clinical trials. Arthritis Rheum 35: 1117-1 125, 1992
12. Dersimonian R, Laird N: Meta-analysis in clinical trials. Controlled Clin Trials 7:177-188, 1986
13. Gibson T, Emery P, Armstrong RD, Crisp AJ, Panayi GS:
Combined d-penicillamine and chloroquine treatment of rheumatoid arthritis: a comparative study. Br J Rheumatol 26:279284, 1987
14. Taggart AJ, Hill J, Astbury C, Dixon JS, Bird HA, Wright V:
Sulphasalazine alone or in combination with d-penicillamine in
rheumatoid arthritis. Br J Rheumatol 26:32-36, 1987
15. Williams HJ, Ward JR, Reading JC, Brooks RH, Clegg DO,
Skosey JL, Weisman MH, Willkens RF, Singer JZ, Alarc6n GS,
Field EH, Clements PJ, Russell IJ, Hochman RF, Boumpas DT,
Marble DA: Comparison of auranofin, methotrexate, and the
combination of both in the treatment of rheumatoid arthritis: a
controlled clinical trial. Arthritis Rheum 35:259-269, 1992
16. Willkens RF, Urowitz MB, Stablein DM, McKendry RJR Jr,
Berger RG, Box JH, Fiechtner JJ, Fudman EJ, Hudson NP,
Marks CR, Brooks R, Rooney TW, Rubin BR, Schmid FR,
Segal AM, Thomas JW, Goldstein AG, Yunus MB, Wortmann
RL, Sherrer YRS: Comparison of azathioprine, methotrexate,
and the combination of both in the treatment of rheumatoid
arthritis: a controlled clinical trial. Arthritis Rheum 355349-856,
17. Roubenoff R, Jones RJ, Karp JE, Stevens MB: Remission of
rheumatoid arthritis with the successful treatment of acute
myelogenous leukemia with cytosine arabinoside, daunorubicin, and m-AMSA. Arthritis Rheum 309: 1187-1 190, 1987
Без категории
Размер файла
519 Кб
efficacy, meta, toxicity, arthritis, analysis, therapy, rheumatoid, combinations
Пожаловаться на содержимое документа