close

Вход

Забыли?

вход по аккаунту

?

The comparative efficacy and toxicity of second-line drugs in rheumatoid arthritis results of two metaanalyses.

код для вставкиСкачать
Arthritis & Rheumatism
Official Journal of the American College of Rheumatology
THE COMPARATIVE EFFICACY AND TOXICITY OF
SECOND-LINE DRUGS IN RHEUMATOID ARTHRITIS
Results of Two Metaanalyses
DAVID T. FELSON, JENNIFER J. ANDERSON, and ROBERT F. MEENAN
We performed 2 metaanalyses of placebocontrolled and comparative clinical trials to examine the
relative efficacy and toxicity of methotrexate (MTX),
injectable gold, D-penicillamine (DP), sulfasalazine
(SSZ), auranofin (AUR), and antimalarial drugs, the
second-line drugs most commonly used to treat rheumatoid arthritis (RA). For the efficacy study, we applied a
set of inclusion criteria and focused on trials which
provided information on tender joint count, erythrocyte
sedimentation rate, or grip strength. We found 66
clinical trials that contained 117 treatment groups of
interest, and for each drug, we combined the treatment
groups. For each outcome, results showed that AUR
tended to be weaker than other second-line drugs. The
results of the 3 outcome measures were synthesized into
a composite measure of outcomes, and AUR was significantly weaker than MTX (P= 0.006), injectable gold (P
< O.OOOl), DP (P< O.OOOl), and SSZ (P= 0.009) and
was slightly, but not significantly, weaker than antimalarial agents (P = 0.11). We also found heterogeneity
among antimalarial agents, in that patients treated with
From the Boston University Arthritis Center and the Departments of Medicine, Boston City Hospital and University Hospital, Boston, Massachusetts.
Supported by Multipurpose Arthritis Center grant AR20613 from the NIH.
David T. Felson, MD, MPH: Associate Professor of Medicine; Jennifer J. Anderson, PhD: Associate Research Professor of
Medicine; Robert F. Meenan, MD, MPH: Professor of Medicine.
Address reprint requests to David T. Felson, MD, MPH,
Arthritis Center, A-203, Boston University School of Medicine, 80
East Concord Street, Boston, MA 021 18.
Submitted for publication January 3, 1990; accepted in
revised form April 11, 1990.
chloroquine did better than those treated with hydroxychloroquine. We found little difference in efficacy between MTX, injectable gold, DP, and SSZ. A power
analysis showed that a trial should contain at least 170
patients per treatment group to successfully differentiate
between more effective and less effective (e.g., AUR)
second-line drugs. None of the reported interdrug comparative trials we reviewed were this large.
For the toxicity study, our inclusion criteria
captured RA trials which reported the proportion of
patients who discontinued therapy because of drug
toxicity and the total proportion who dropped out. We
found 71 clinical trials that contained 129 treatment
groups. The average proportion who dropped out and
the average proportion who dropped out because of
drug toxicity were computed for each drug. Overall,
30.2% of the patients in these trials dropped out; 50% of
them did so because of drug toxicity. Injectable gold had
higher toxicity rates (P< 0.05) and higher total dropout
rates (P < 0.01) than any other drug; 30% of goldtreated patients dropped out because of side effects
versus 15% of all trial patients. Antimalarial drugs and
AUR had relatively low rates of toxicity; the rate for
MTX was imprecise because of discrepancies between
trials. Thus, of the commonly used second-line drugs,
AUR is the weakest, and injectable gold is the most
toxic. Agents introduced in the future will be compared
with these drugs. If a curative drug is not found, large
multicenter trials or data synthesis from multiple drug
trials may be necessary to identify new treatment regimens that have promise.
Arthritis and Rheumatism, Vol. 33, No. 10 (October 1990)
1449
1450
FELSON ET AL
Many second-line agents are now available to
treat rheumatoid arthritis (RA). All of them have been
shown in clinical trials to be more effective than
placebo. However, there is limited evidence regarding
their relative effectiveness or toxicity. Clinical trials
focusing on the comparative efficacy of second-line
drugs have, in general, failed to document differences
between these drugs (1).
While comparing an effective second-line drug
with placebo requires a small sample size, distinguishing between 2 effective drugs requires a large one.
Discussing the differences between injectable gold and
auranofin (AUR) in a large multicenter trial, Ward et a1
stated that “GST [gold sodium thiomalate] could have
had as much as a 35 percentage point advantage over
auranofin in improving joint pain/tenderness without
being statistically distinguishable by our sample sizes”
(2). Unfortunately, the failure of a comparative trial to
find significant differences does not connote equivalency (3). Type I1 errors are possible.
The choice of which slow-acting drug to prescribe for patients with RA depends on both its efficacy and its toxicity. In clinical trials, the focus is the
drug’s efficacy and not its toxicity. Although it is
commonly held that some second-line drugs are less
toxic than others, this feature has not been studied
directly. In single clinical trials, toxicity that is severe
enough to necessitate discontinuation of therapy does
not occur frequently enough to permit precise comparisons between drugs.
The main obstacle to showing differences in the
efficacy and toxicity of second-line agents appears to
be sample size. We therefore performed two metaanalyses, quantitative syntheses of studies already
performed, to evaluate the comparative effectiveness
and toxicity of second-line drugs. Metaanalysis has the
unique advantage of substantially increasing sample
size because it combines data from multiple trials.
METHODS
Identification and selection of published trials. Methods for this metaanalysis conform to the suggestions of
Sacks et a1 (4). Using MEDLINE to access reports published
from January 1966 through August 1989, we conducted a
comprehensive search for clinical trials in which second-line
agents were used to treat rheumatoid arthritis. We also
conducted bibliographic searches of trial reports and review
articles. To verify that we had obtained all relevant published trials, we searched through all issues of 2 rheumatology journals, Arthritis and Rheumatism and Scandinavian
Journal of Rheumatology, from their dates of inception to
the end of August 1989. We found no additional articles.
We obtained photocopies of all published reports of
the clinical trials identified. The methods and results sections
were blinded and then evaluated for eligibility for our
metaanalyses. Blinding the methods sections eliminated
references to the journal, the authors, the site of the study,
the year of publication, and the results of the study. Blinding
the results sections eliminated any reference to the drugs, so
that the reviewer could not determine which drug treatment
groups had fared better.
Criteria for inclusion in our metaanalyses were that
the clinical trials must have been studies of adults 2 1 8 years
old, at least 90% of whom had rheumatoid arthritis. For
studies published after 1958, the date of publication of the
American Rheumatism Association diagnostic criteria for
RA (S), subjects must have met those criteria, or sufficient
information must have been given to permit a determination
that subjects had RA. For studies published prior to 1958,
the designation that the patients had RA was deemed sufficient.
Another criterion for eligibility was that patients
must have been assigned to treatment regimens by either a
random method or a quasirandom scheme, such as sequential assignment or minimization. Trials using historical controls were excluded, as were trials in which the assignment
scheme was not specified.
The following drugs and minimum dosages were
deemed acceptable: placebo, antimalarial drugs, which included hydroxychloroquine (HCQ) at a dosage of 2200
mg/day and chloroquine at a dosage of 250 &day, injectable gold at a dosage of 50 mg/week, AUR at a dosage of at
least 6 mg/day, D-penicillamine (DP) at a dosage of at least
500 mg/day, methotrexate (MTX) either orally or parenterally at a dosage of at least 7.5 mg/week, and sulfasalazine
(SSZ) at a dosage of at least 2 gdday.
Since this was a metaanalysis of both placebocontrolled and comparative trials, we defined placebo controls as those taking either a true placebo or a trivial dose of
effective drug (55% of the minimum effective dose listed
above). A comparative trial was acceptable if at least one
treatment group received a conventionally used second-line
drug as defined above.
We also required that the duration of the clinical trial
have been at least 2 months, because some second-line
agents are not effective before this time. Thus, studies of
shorter duration might not have allowed sufficient time to
attain maximum efficacy.
Metaanalysis of drug efficacy. For the efficacy
metaanalysis, extractable results dealing with drug efficacy
must have been included in the study. Specifically, we
required numerical values for the tender joint count (or
Ritchie articular index), erythrocyte sedimentation rate
(ESR), and/or grip strength results. We have found that
these 3 outcome measures are especially sensitive and
nonduplicative in measuring improvement in RA clinical
trials (6). One other outcome measure that we have found to
be sensitive to change, the physician global assessment, was
not reported frequently enough to be useful as an outcome
measure in this study. We equated the tender joint count
with the Ritchie index because previous analyses (7,8) have
shown that they produce nearly identical results. Either the
means of the starting and ending values for these outcomes
or the mean change in these outcomes must have been
reported.
Our analysis focused on treatment arms rather than
EFFICACY AND TOXICITY METAANALYSES O F DRUGS FOR RA
individual trials. For each trial, we identified the treatment
arms of interest and extracted those data. For example, for
a placebo-controlled trial of injectable gold and AUR, there
are 3 treatment arms: injectable gold, AUR, and placebo.
For a comparative trial of injectable gold versus an experimental drug, there is 1 treatment arm for which data were
extracted, that of injectable gold. Also, when the trial
reported results over time, we chose the interval closest to 6
months.
The outcomes of interest were the tenderjoint count,
the ESR, and the grip strength. We also computed a “combined effect size,” which is the mean of the “standardized
improvement” in the tender joint count, ESR, and grip
strength. Standardized improvement in each of these outcome measures was computed as the change (improvement)
in the outcome measure in the treatment group of interest
divided by the pooled baseline standard deviation for that
particular outcome measure across trials. When data for 1 or
more of these outcome measures were absent from a clinical
trial, we computed the combined effect size based on the
outcome measure results that were presented.
We also extracted data for other variables that might
affect outcomes. These included the duration of the study,
the study protocol (whether it was placebo-controlled or
comparative or whether it was double-blind), the source of
funding (specifically, whether it was drug company funded),
and the measures of disease severity (the proportion treated
with steroids, the mean initial count of tenderjoints, and the
mean duration of disease) for each treatment group.
In addition, we independently rated the quality of
each published trial, using a conventional scoring instrument
(9), which we modified. Our quality assessment evaluated 10
different characteristics of trial reports, including information about eligibility criteria, random allocation and the
method of randomization, patient and evaluator blindedness,
treatment complications, and the number of study subjects
lost to followup. We also evaluated whether statistical
analysis was performed, whether the names of particular
techniques were reported, and whether a power analysis was
done. Each of the 10 trial characteristics was scored as
completely reported (2 points), partly reported (1 point), or
omitted (0 points), based on a priori criteria. Scores were
summed into an overall quality measure, which demonstrated high interobserver consistency.
For each drug, we combined the results across
treatment groups. Weighting each treatment group by size at
the end of trial, we generated means for each drug treatment
for each of the 4 outcomes (tender joint count, ESR, grip
strength, and combined effect size).
To compare drugs, we performed analyses of variance, weighted by treated group size, multiplied by study
quality (rescaled on a &1 scale), and adjusted for those
covariates which had a significant association (P < 0.05)
with each outcome. These important covariates were the
mean disease duration for the treatment group, the trial
length, and, for the tender joint count only, the initial tender
joint count and evaluator blindedness. This analysis used a
fixed effects model. An adjusted combined effect was computed as the mean of the separate, covariate-adjusted, standardized outcomes.
In terms of heterogeneity of the results from different
studies, the Q statistic, a chi-square measure of study
1451
heterogeneity, was computed for each outcome. For the
tender joint count unadjusted for covariates, Q = 135.9
(degrees of freedom [dfl = 83, P < 0.001), but this value
decreased to 65.3 (df = 68, P = 0.65) after adjustment for the
covariates listed above. For grip strength, the study results
showed nonsignificant heterogeneity both before (Q = 93.4,
df = 86, P = 0.30) and after (Q = 79.9, df = 77, P = 0.40)
adjustment for the covariates. For the ESR unadjusted for
covariates, Q = 96.3 (df = 95, P = 0.50), and after adjustment, this decreased to 86.6 (df = 86, P = 0.50). The Q
statistic for the composite outcome was not significant (P =
0.995). This result was expected since all of its 3 components
had nonsignificant heterogeneity.
To avoid performing multiple statistical comparisons
(and generating Type I errors), we tested interdrug differences on the composite outcome measure only, using a
2-sided a = 0.05 as our standard for significance.
Finally, we performed a number of sensitivity analyses in which especially large studies and outlier studies were
excluded so that we could determine if exclusion of these
studies would affect our results. Exclusion of individual
studies did not substantially change any of the results.
Metaanalysis of drug toxicity. For the toxicity metaanalysis, trial identification, drugs studied, and methods
inclusion criteria were the same as for the efficacy study.
In addition, information about the number of patients who
dropped out by the end of the trial andor the number
who dropped out because of drug toxicity was required.
The proportion of subjects who dropped out was
defined as the proportion of all those who entered a study
arm and who dropped out by the end of the trial. Dropouts
did not include patients who were withdrawn from the trial
immediately after randomization because they were found to
be ineligible. The proportion of subjects who dropped out
because of toxicity was defined as the proportion of all those
who entered a study arm and who had to discontinue the
drug prior to the end of trial because of side effects.
For subjects who were withdrawn from the study
because of drug toxicity, the frequency of organ-related
toxicities was tabulated. For subjects who had more than 2
side effects at the time of withdrawal, the one listed as the
most severe was recorded as the cause of dropout. When 2
concurrent side effects were listed and no mention was made
about which was the more serious, the reviewer arbitrarily
chose one, generally weighting cytopenias and renal abnormalities as more severe. Skin rash was arbitrarily judged as
less severe.
As in the efficacy study, the unit of analysis was the
treatment group. Unadjusted mean rates of total dropout and
toxicity-related dropout were computed for each second-line
drug and placebo, weighting each treatment group according
to its size at entry into the trial and its study quality score.
The homogeneity of the rates for each drug was assessed
after each rate was transformed to the arc sine of the square
root of the rate in order to stabilize variances (10).
In addition, we performed a covariate-adjusted analysis with outcomes in the transformed scale. This adjusted
analysis was also weighted by treatment group size and
study quality and was controlled in a regression format for
covariates that were significantly related to either dropout
rates or toxicity-related dropout rates. For total dropout, the
other significant predictors were patient age (higher dropout
1452
FELSON ET AL
with increased age) and year of publication (higher dropout
with later publication). For toxicity, the predictors were
disease duration (higher toxicity with longer disease duration) and length of trial (higher in longer trials).
Heterogeneity testing of the unadjusted total dropout
rate showed a Q value of 357.28 (df = 117, P < 0.05), which
fell to 287.73 (df = 11 1, P < 0.05) after covariate adjustment.
For the toxicity-related dropout rate, the Q value for the
unadjusted rate was 181.23 (df = 119, P < 0.05), dropping to
129.95 with adjustment (df = 109,O.l < P < 0.05). Because
significant heterogeneity was found for both outcomes, the
weighted mean rates and their variances were revised to
accommodate heterogeneity of outcomes among groups,
according to the method of DerSimonian and Laird (1 1). This
method increases the variance of estimates according to the
amount of heterogeneity found. With this heterogeneity
adjustment, statistical comparisons become conservative
because of the increased variances. The significance of each
pairwise comparison was tested on covariate-adjusted values in the transformed scale at the 5% level of significance
(2-tailed).
Finally, we performed sensitivity analyses, focusing
on subsets of studies identified by year of publication and
other predictors, to see if analyses of these subsets yielded
the same results as our overall results. We found no difference in the results.
We found 66 clinical trials that met our criteria for
the efficacy metaanalysis and 71 that met our criteria for the
toxicity metaanalysis. The complete references for these are
given at the end of the References section.
RESULTS
Efficacy study. Of 168 reports of clinical trials
that we reviewed, 66 met our criteria for the efficacy
study (shown at the end of the References section). In
these 66 trials, there were 117 relevant treatment
groups (see Table El). Among these treatment groups,
5,343 patients were' entered for study, and 3,957 completed the studies (finished treatment). These 3,957
constitute the analysis group for the efficacy study.
The mean length of treatment was 39.4 weeks. Our
study sample included many patients receiving AUR,
injectable gold, o r DP in clinical trials (see Table E2).
There were fewer patients receiving MTX and SSZ,
the more recently studied drugs.
Table El. Description of studies and patients in the metaanalysis
of the efficacy of second-line drugs used to treat rheumatoid arthritis
No. of trials
No. of treatment groups (mean per trial)
No. of patients entering trial
No. of patients completing trial (%)
Mean no. of patients per treatment group (range)
Mean no. of weeks of treatment (range)
Mean quality score, 0-20 scale (range)*
~~
66
1 I7 ( I .77)
5,343
3,957 (74.1)
33.8 (6490)
39.4 (12-104)
14 (9-20)
~
* See Methods for definition and calculation of quality score.
Table E2. Numbers of treatment groups and patients in the metaanalysis of the efficacy of second-line drugs used to treat rheumatoid
arthritis, according to drug treatment group
Antimalarial drugs
Auranofin
Injectable gold
Methotrexate
D-penicillamine
Sulfasalazine
Placebo
No. of
treatment groups
No. of patients
completing the trial
11
23
29
7
19
6
22
314
1,274
656
150
51 1
161
89 I
Placebo-treated patients had the least improvement in the tender joint count (see Table E3). Among
all active second-line drugs, antimalarial drugs had the
lowest score for improvement in tender joint count,
with the score for AUR being only slightly higher.
Patients treated with MTX had the greatest level of
improvement in the tender joint count. After adjustment for disease duration, trial length, initial tender
joint count, and the blindedness of the evaluator in the
trial, antimalarial agents remained the weakest of the
drugs in improving the tender joint count.
With respect t o changes in grip strength (see
Table E3), we found that, as before, the placebotreated group had the smallest level of improvement in
grip strength, with a change of <10 mm Hg over the
courses of the trials. The weakest of the second-line
drugs was AUR, with an average change of <20 mm
Hg after adjustment for covariates. The drug producing the most improvement in grip strength (after adjusting for covariates) was DP. And, after adjustment
for covariates, AUR remained weaker with respect to
grip strength improvement than the other second-line
drugs.
A similar pattern was seen in the results for the
mean improvement in the ESR (see Table E3). As
might be expected, the change among those treated
with placebo was minimal, 1 mmhour. Of the secondline drugs, AUR was the weakest, and SSZ was the
strongest. After covariate adjustment, AUR remained
the weakest second-line drug.
We combined the information from all outcome
measures and from each trial (see Figure El) into a
composite measure of treatment effect. The placebotreated patients experienced the least beneficial effect.
Of the second-line drugs, AUR was the weakest, with
DP and SSZ scoring as the strongest. After covariate
adjustment, significance testing revealed that all drugs
were stronger than placebo (P < 0.0001). In pairwise
comparisons, AUR was significantly weaker than all
-
EFFICACY AND TOXICITY METAANALYSES OF DRUGS FOR RA
1453
Table FA Comparison of the efficacy of second-line drugs used to treat rheumatoid arthritis. according to the mean level of improvement in
the tender joint count, grip strength, and erythrocyte sedimentation rate (ESR)*
Tender joint count
~
Drug
Unadjusted
Placebo
1.0
16)
7.2 t 1.4
(n = 8)
7.7 t 0.9
(n = 20)
8.7 rt 1.0
(n = 22)
12.4 2 3.3
(n = 4)
8.7 t 1.0
(n = 15)
I I .6 rt 2.8
(n = 5)
Antimalarial drugs
Auranofin
Injectable gold
Methotrexate
D-penicillamine
Sulfasalazine
Grip strength (mm Hg)
ESR (mm/hour)
~
4.8
2
(n
=
Adjusted
Unadjusted
Adjusted
Unadjusted
Adjusted
2.9
9.0 2 2.28
(n = 20)
41.7 f 5.7
(n = 8)
17.2 f 2.5
(n = 17)
36.0 f 4.8
(n = 22)
33.2 2 7.2
(n = 6)
39.4 ? 6.7
(n = 15)
26.4 f 7.3
(n = 5)
9.7 2 3.4
(n = 18)
39.0 f 6.9
(n = 8)
19.4 ? 3.6
(n = 15)
36.4 f 4.6
(n = 20)
42.3 2 8.2
(n = 5)
43.4 2 4.5
(n = 15)
28.8 f 7.8
(n = 5)
0.9 f 1.4
tn = 16)
11.9 rt 2.1
(n = 11)
8.5 t 1.4
(n = 20)
19.0 f 2.1
(n = 26)
12.2 2 4.3
(n = 5)
21.7 f 1.8
(n = 18)
23.9 t 4.2
(n = 6)
0.7 t 1.6
(n = 14)
12.7 2 2.6
(n = 1 1 )
9.2 2 1.5
(n = 18)
17.9 2 1.8
(n = 24)
12.5 2 4.1
(n = 5)
21.3 t 2.0
(n = 17)
22.7 2 2.9
(n = 6)
0.8
13)
6.9 2 1.7
(n = 6)
8.0 f 0.7
(n = 17)
9.1 f 0.9
(n = 20)
12.6 f 1.9
(n = 4)
9.4 2 1.0
(n = 14)
11.3 f 1.5
(n = 5)
(n
f
=
* Values are the mean f SEM; n = number of treatment groups. Unadjusted values were weighted by study size; adjusted values were
weighted by study size and quality score, and were adjusted for covariates (see Methods).
second-line drugs except the antimalarials (P = 0.11
versus AUR). The P values for these comparisons
were as follows: P = 0.006 for MTX versus AUR, P <
0.0001 for gold versus AUR, P < 0.0001 for DP versus
AUR, and P = 0.009 for SSZ versus AUR. Antimalarial drugs also scored relatively low in composite
efficacy and were significantly less effective than DP
(P = 0.04).
Heterogeneity on some outcomes within the
antimalarial drug-treated group prompted us to examine more closely whether the 2 antimalarial agents
Composite Treatment Effect
(in standard units)
I .oo
T
I
U
UnadJusted
05
u
Adjusted
Figure El. Comparison of the composite treatment effects of placebo and 6 second-line drugs used to treat rheumatoid arthritis. The
composite treatment effect is the mean of “standardized change” in
the tender joint count, erythrocyte sedimentation rate. and grip
strength (see Methods). Bars show the mean 2 SEM improvement
with treatment.
studied, hydroxychloroquine and chloroquine, differed in efficacy. The individual trial results are shown
in Figure E2. For all outcomes, we found that chloroquine was more effective than HCQ (for composite
outcome P = 0.014), even though the chloroquine
trials usually studied dosages of 250 mg/day, whereas
the HCQ trials often used dosages of 2400 mg/day. Of
special note, each of the chloroquine trials had a
higher score than each of the hydroxychloroquine
trials on 2 of the 4 outcomes (grip strength and
composite effect).
Because we could not detect differences between some drugs, we performed a power analysis., in
which we investigated how large a sample would be
needed to detect significant differences between drugs
(see Table E4). The differences between drugs are
expressed in “effect size” units, as used in the composite outcome measure (see Methods). The difference
between MTX, DP, injectable gold, and SSZ was -0.1
unit. To demonstrate this difference as significant, a
2-drug comparative trial with 3,000subjects would be
needed. For a comparison of 2 drugs in which 1 drug
was substantially weaker, such as an effect size difference of 0.3 units between drugs, a 2-drug comparative
trial of 170 subjects in each drug group would be
necessary. This corresponds approximately to the
difference between AUR and other second-line drugs
such as injectable gold, MTX, DP, and SSZ. The mean
size of treatment groups in the trials studied was 33.8
(see Table El).
Toxicity study. Of 168 published articles and
abstracts, 71 clinical trials met our inclusion criteria
for the toxicity study (shown at the end of the Refer-
FELSON ET AL
1454
Effect in Standard Units
1.60
Table T1. Description of studies and patients in the metaanalysis
of the toxicity of second-line drugs used to treat rheumatoid arthritis
*
1.40
T
-
I
1.20
*
*
-
1.00
**
0.80
I
No. of trials
No. of treatment groups (mean per trial)
No. of patients entering trial
Mean no. of patients per treatment
group (range)
Mean no. of weeks of treatment (range)
Mean dropout rate, % (range)
Mean toxicity-related dropout rate, %
(range)
I
*
*
*
T
0.60
0.40
0.20
0.00
'
1
Composite Effect
(p=.014)
*
I
Tender
Joint Count
(p=.382)
Chloroquine
A
Erythrocyte
Sedimentation Rate
(pi.074)
Hydroxychloroquine
Figure E2. Comparison of the efficacy of chloroquine and hydroxychloroquine, 2 second-line drugs used to treat rheumatoid arthritis.
Each point represents the results of 1 trial. Horizontal lines show
group means (weighted by study size). See Methods for explanation
of the composite treatment effect calculation.
ences section). In these trials, there were 129 treatment groups which received second-line drugs or
placebo (see Table TI). There were 6,027 patients who
were entered into these treatment groups. On average,
30.2% dropped out before the trial ended. Slightly less
than half of these patients (15% of the total) dropped
out because of drug toxicity. Many clinical trials
included injectable gold, AUR, or DP; only a few trials
studied MTX and SSZ (see Table T2).
We found that the total dropout rate (unadjusted) vaned from -17% to -39% (see Figure Tl). The
lowest total dropout rates were seen in patients assigned to receive antimalarial therapy and in those
assigned to MTX therapy. The adjusted total dropout
rates for those taking antimalarial agents and those
Table FA. Trial size needed to detect differences in efficacy between second-line drugs used to treat rheumatoid arthritis*
~
Actual difference in efficacy
(in standardized effect units)
2-drug comparative trial
No. per drug
No. per trial
3-drug comparative trial
No. per drug
No. per trial
37.5 (12-104)
30.2 (0-79.3)
15.0 (0-50)
1
I
Grip Strenglh
(p=.040)
71
129 (1.82)
6,027
46.7 (7-672)
O.IOt
0.20
0.30t
0.40
0.50
1,500
3,000
400
800
170
340
96
192
62
124
1,300
3,900
325
975
143
429
80
240
52
156
* Power = 80%, a = 0.05 (2-sided). See Methods for definition and
calculation of standardized effect units.
t Approximate difference between drugs in the stronger group of
slow-acting drugs (methotrexate, injectable gold, D-penicillamine,
sulfasalazine).
.t Approximate difference between auranofin and each drug in the
stronger group of slow-acting drugs.
taking MTX were lower than the dropout rate for
patients taking DP ( P < 0.05 and P < 0.10, respectively). The highest dropout rate was seen in patients
assigned to receive injectable gold, and the adjusted
dropout rate for this group was significantly higher (P
< 0.01) than the rates for any other drug.
Toxicity-related dropout (see Figure T2) was
the largest component of the total group of patients
who dropped out. Not surprisingly, placebo-treated
patients were very unlikely to drop out because of
toxicity (adjusted rate 4.7%; P < 0.05 versus all
drug-treatment groups). Toxicity-related withdrawal
was most common among patients given injectable
gold (adjusted rate 29.9%), and this rate was significantly higher ( P < 0.05) than the rates for all other
treatment groups. Toxicity was least common among
those treated with antimalarial agents (adjusted rate
8.5%), and AUR treatment also had a low toxicity rate
(adjusted rate 11.3%).
The specific organ systems affected by drug
toxicity, which caused the patients to drop out of the
studies, tended to reflect those commonly seen in
clinical practice (see Table T3). Skin rashes were
especially frequent among patients treated with injectable gold (13.0%) and with DP (7.1%), and rare in
those treated with antimalarial drugs (2.3%) and placebo (1.2%). Common toxicities leading to discontinuation of therapy included nausea and vomiting in
Table T2. Numbers of treatment groups and patients in the
metaanalysis of the toxicity of second-line drugs used to treat
rheumatoid arthritis, according to drug treatment group
Auranofin
Antimalarial drugs
Methotrexate
D-penicillamine
Sulfasalazine
Injectable gold
Placebo
No. of
treatment groups
No. of patients
entering the trial
25
12
7
20
7
31
27
1,887
399
195
794
253
1,046
1,453
EFFICACY AND TOXICITY METAANALYSES OF DRUGS FOR RA
t
Figure T1. Total dropout rates from rheumatoid arthritis clinical
trials of second-line agents, according to treatment group. Bars
show the mean ? SEM (confidence intervals = I.% x the SEM).
those taking SSZ (12.5%) and hepatic toxicity in those
taking MTX (10.3%).
DISCUSSION
Despite the failure of most comparative trials of
drug treatments for RA to show differences in efficacy
between auranofin and other second-line drugs, our
results strongly suggest that auranofin is less efficacious than most other second-line drugs. While auranofin also appears to be slightly weaker than the
antimalarial drugs, this comparison was not significantly different. The antimalarial group itself may be
heterogeneous with respect to efficacy. We found that
hydroxychloroquine was significantly weaker than
chloroquine. The scores for HCQ placed its efficacy at
a level close to that of AUR, whereas the efficacy of
chloroquine ranked closer to that of the other secondline drugs.
Discussions of therapeutic approaches to RA
depend on considerations of both the efficacy and the
toxicity of the drugs. The metaanalysis focusing on
drug toxicity has shown that injectable gold has a
higher rate of side effects than any other commonly
used second-line drug. Antimalarial drugs have the
lowest toxicity-related dropout rate, but the rate of
toxicity experienced by patients taking AUR was also
low. DP and SSZ had comparable rates for total
dropout,
While our efficacy study, because it synthesized
data from diverse studies, could have generated somewhat inaccurate estimates, the results (shown in Figure El) suggest that the differences between injectable
gold, MTX, DP, and SSZ are marginal. The differences are approximately 0.1 effect size units, and, if
1455
our results are correct, a trial large enough to distinguish them would have to be enormous, a 2-drug
comparative trial of 3,000 patients. This is, to our
knowledge, larger than any rheumatoid arthritis clinical trial ever undertaken. Our results strongly suggest
that, for practical purposes, MTX, injectable gold, DP,
and SSZ are equally effective and that the selection of
one over another of this class of drugs depends on
factors other than the drug's short-term efficacy. Our
power analysis speaks also to the ultimate limits of
clinical trials. In an era of limited resources, it may not
be productive to attempt to distinguish between the
short-term efficacy of these drugs.
Since these studies were metaanalyses, they
have several inherent limitations. First, like all quantitative syntheses of studies, diverse studies were
pooled. There are several unique aspects to our
metaanalyses in this regard. First, we included both
comparative and placebo-controlled studies because
we found that synthesizing placebo-controlled studies
alone did not yield sufficient statistical power for
interdrug comparisons. A more traditional metaanalysis including only placebo-controlled studies produced
the same efficacy differences, with AUR substantially
weaker than other second-line agents. We chose to
include comparative trials to enhance statistical power
and did not find a substantial difference in results of
comparative trials and placebo-controlled trials, which
suggests that our inclusion of comparative trials was
justified.
Second, for both drug efficacy and drug toxicity, we attempted to control for covariates that might
bias the study results. For both studies, we found that
several covariates significantly affected outcomes. InDrowulRate
Figure T2. Rates of dropout because of drug toxicity from rheumatoid arthritis clinical trials of second-line agents, according to
treatment group. Bars show the mean 2 SEM (confidence intervals
= I.% x the SEM).
1456
FELSON ET AL
Table T3. Percentages of patients developing organ-specific toxicity necessitating dropout, according to drug treatment group*
Skin rash
Alopecia
Mucous membranes
Alteration in taste
Blood
Decreased hematocrit
Decreased WBC
Decreased platelets
Renal
Proteinuria
Acute renal failure
Ocular
Fever
CNS
PBO
AUR
AMD
MTX
DP
SSZ
Gold
1.2
0
0.6
0.2
3.2
0
0.6
0. I
2.3
0
0.3
0
0
0
2.6
0
7.1
0
0.1
2.5
3.8
0
1.1
0
13.0
0.1
1.8
0
0.2
0.1
0.1
0.3
0
0
0.3
0
I .5
0
1 .o
0
0
0
2.5
0
0.5
1.1
0
0.9
0
1.5
1.1
0. I
0.9
0
0
0
0
0
0. I
0
0.5
0
0
5.0
0
0
0
3.7
0
0
0
0.7
0
0.3
0
0
0
0
0.5
0
0
1.1
0
0.2
0.2
0
3.3
I .3
0
0
2.I
2.I
0.5
2.0
0.3
0.4
0
12.5
0
1.6
I .3
0
0.5
0.9
0.1
0.I
0.5
0
0.3
0.6
1.1
0.1
0.I
0.1
0. I
0.5
0
0
0
0
0.1
GI tract
Nausea andor vomiting
Diarrhea
Hepatic
I .o
0.3
0.2
0.4
1.1
0.1
3.9
Lung
Breast
Other
0.1
0
0.3
0.1
0
0
I .4
0
0.4
I .O
0.5
10.3
0
0
0.5
1 .o
0
0
2.2
* PBO = placebo; AUR = auranofin; AMD = antimalarial drugs; MTX = methotrexate; DP
D-penicillamine; SSZ = sulfasalazine; Gold = injectable gold; WBC = white blood cells; CNS
central nervous system; GI = gastrointestinal.
deed, when we inserted these covariates into our
analytic equation, we found that they markedly decreased the heterogeneity of results across studies. In
the efficacy study, however, it is likely that we did not
control adequately for some covariates, especially
disease severity and disease duration. We found that
disease duration was negatively correlated with study
outcomes and that clinical trials of MTX therapy were
consistent in studying patients with long disease durations in whom other second-line drugs had failed. It is
therefore possible that we modestly underestimated
the efficacy of MTX. However, as shown in Figure E l ,
our estimate of the efficacy of MTX rose substantially
after adjustment for covariates, which suggests that
our methods of analysis appropriately took into account the nature of the RA patients in these trials.
Ironically, we may have similarly overestimated the
efficacy of AUR and the antimalarial agents because,
in general, those trials were of patients with short
disease duration. Also, our relatively small numbers of
trials of MTX and SSZ make our results for these
drugs less clear cut. Finally, because trials did not
provide intent-to-treat data (analyzing all patients entering trials), we could not perform such an analysis.
For the toxicity study, additional limitations
=
=
arose because some trials had rules that dictated
patient dropout, rules which may not be followed by
many clinicians. Furthermore, our study focused on
clinical trials that were generally short-term (mean
duration 37 weeks), and we can therefore say little
about the rate of important long-term side effects, such
as cirrhosis in patients treated with MTX, severe
proteinuria in patients taking gold or DP, or ocular
toxicities among patients treated with antimalarial
agents.
Previous comparative trials have suggested that
AUR might be slightly weaker than other drugs. For
example, although Ward et a1 could find no significant
differences in major outcomes between GST and
AUR, they stated that GST may be slightly more
effective than AUR (2). Looking at functional status
outcomes, Meenan and coworkers, in conjunction
with the Cooperative Systematic Studies of the Rheumatic Diseases (CSSRD) group, found that injectable
gold was significantly superior to AUR (12). Smith et
al (13), in a comparative trial of gold versus AUR,
found that the ESR improvement was significantly
greater in the group receiving injectable gold, although
there were no significant differences in other outcomes. Small trials have consistently found no signif-
EFFICACY AND TOXICITY METAANALYSES OF DRUGS FOR RA
icant differences in efficacy between gold and AUR.
However, when Horton quantitatively pooled trials
comparing AUR with parenteral gold (14), he found
that drug-failure rates were significantly higher among
those treated with AUR, which strongly suggests that
AUR is weaker than injectable gold.
The results of our comparison of HCQ with
chloroquine are truly surprising. It has been suggested
that 2 mg of HCQ is as effective as 1 mg of chloroquine
(15). However, our results suggest that chloroquine is
stronger, even at a dosage ratio of 1:2 (chloroquine:HCQ). These results, if borne out, would be an
important clinical insight since chloroquine is a drug
that has few serious side effects, except for retinal
toxicity, which can be anticipated. Further study of
chloroquine, a drug long favored in the treatment of
RA patients in the UK, may be indicated.
The findings from our metaanalyses have important ramifications for the planning of future clinical
trials in patients with RA. No comparative trial we
reviewed was large enough to make it likely that
clinically important differences in efficacy would be
detected. Therefore, most have not detected significant differences between drugs. Incidentally, even this
large metaanalysis sample may have lacked sufficient
statistical power to distinguish the antimalarial drugs
from the other second-line agents. It is likely that new
drugs for the treatment of RA will be compared with
the second-line drugs studied here. While it is always
hoped that a new drug will be curative and its comparative superiority will be easy to demonstrate, it is more
likely that new regimens will offer marginal but measurable improvement over traditional ones, as has
been the case in oncology. Better planning with respect to study size is necessary so that the relative
efficacy of new drugs over old ones can be demonstrated. Furthermore, determining that there is no
significant difference between a new regimen and an
old one does not mean that the two are equally
effective (3). Special power computations are necessary in any attempt to demonstrate equivalency (3).
Finally, it is clear that all possible comparisons of new
drugs for the treatment of RA will never be performed,
given the large number of study subjects necessary for
such a comparison. Trials should therefore be standardized so that different drugs can be compared post
hoc across trials (1).
The toxicity-related dropout rates for MTX
reported here may have been spuriously elevated by
idiosyncrasies of the clinical trials studied. While
elevation of liver enzyme levels accounted for no
dropouts in 5 of the 7 trials we used (16-20) and
1457
accounted for only 1 dropout in one other trial (21), it
was an important cause of dropout in the trial conducted by the CSSRD group (22). In that trial, liver
enzyme elevations that were greater than twice the
normal level dictated a patient’s withdrawal from the
study; this occurred in 18 of the 95 patients randomized to receive MTX. Since this was the largest trial of
MTX and the only trial in which a substantial proportion of cases dropped out because of abnormal findings
on liver function tests, a change in the rules in this trial
might have resulted in a substantial change in both the
organ toxicity rates (Table T3) and the toxicity-related
dropout rates (Figure T2).
While other individual clinical trials, even large
ones, have not been large enough to precisely compare
the toxicity of second-line drugs, findings from many
clinical trials and literature reviews have suggested
some of the results reported here. In the CSSRD trial
of AUR versus GST versus placebo (2), Ward et al
found that “auranofin [had] fewer significant adverse
effects than GST.” In a review of antimalarial drugs,
Mackenzie (15) stated that, “with antimalarials, the
side effects are fewer and the dropout rate is lower
than with injected gold or penicillamine.”
Mackenzie estimated that, in his experience,
the dropout rate because of side effects in patients
taking antimalarial drugs ranged from 3% to 7% (15).
That value is consistent with the -7.8% shown in our
data. Horton (14) also reviewed trials comparing AUR
and parenteral gold and found a much higher rate of
dropout because of side effects among patients treated
with parenteral gold than among those treated with
AUR. Schattenkirchner et al (23) conducted a postmarketing surveillance of AUR treatment in the Federal Republic of Germany and reported that dropout
rates due to AUR toxicity were low and closely
paralleled the rates reported in clinical trials.
In this study, we focused on toxicity severe
enough to necessitate dropping out of the clinical
trials. Some of these toxicities, such as nausea, vomiting, or skin rash, may not have been severe enough
to affect the decision to choose a particular drug
therapy. However, injectable gold, the drug with the
highest toxicity rate, tended to cause a higher frequency of worrisome side effects than the others. For
example, cytopenias and renal and pulmonary side
effects were most common among those treated with
injectable gold and DP (Table T3). Other drugs tended
to have not only a lesser rate of side effects overall, but
also a more benign spectrum of toxicity.
The rates of nausea and vomiting in patients
taking SSZ were relatively high (12.5%). Of the 7
FELSON ET AL
1458
clinical trials which we synthesized to arrive at these
rates, 6 used enteric-coated tablets of SSZ. Most of
these trials (24-28) used a maximum dosage of 2
gm/day, whereas 3 trials (28-30) allowed patients to
increase the dosage to 3 gm/day. Recent data (31)
suggest that a lower maintenance dosage of 1 gm/day
of SSZ may be associated with significantly lower
levels of gastrointestinal toxicity, although it may also
be associated with lower levels of efficacy.
Most of the studies of injectable gold, AUR,
SSZ, MTX, and antimalarial agents used almost the
same dosages. For DP, however, there was a broad
range of dosages, from 500 mg/day to 1,250 mg/day.
We analyzed the data for evidence of dose-related
toxicity in patients taking DP and found no such
association. Our findings are consistent with those of I
earlier study of dose-related toxicity (32). Jaffe, in a
still earlier study of DP treatment of RA and necrotizing vasculitis (33), noted lessened toxicity when a
graduated dosage regimen was used. This technique
had been used in all of the trials of DP assessed in the
toxicity metaanalysis.
Despite possible differences in efficacy between
the antimalarial drugs, we could find no significant
differences between the rates of toxicity-related dropout or total dropout in patients receiving these agents.
We found significant unexplained heterogeneity
across trials in the toxicity metaanalysis, and we
accommodated this feature in our analysis (see Methods). This finding suggests that different trials have
dBerent thresholds for patient dropout. Many studies
were excluded because the drug side effects were not
enumerated or discussed; many other trials failed to
mention side effects that were not severe enough to
warrant discontinuation of the drug. Since side effects
are common to all RA pharmacotherapeutics, it is
unfortunate that this information is often absent from
the reports. Specific toxicities, their severity, and their
consequences should be given in all drug trials so that
rates of toxicity can be better compared. Also, a
standardized way of grading toxicities should be
adopted. One such method has recently been proposed
by Fries et a1 (see ref. 34).
These metaanalyses have ranked second-line
drugs in terms of their efficacy and their toxicity.
Future clinical trials in rheumatoid arthritis, in which
new regimens are compared with the old regimens
studied here, should incorporate larger sample sizes
and be standardized to permit comparisons across
studies.
REFERENCES
1. Felson DT, Anderson JJ, Meenan RF: Time for changes
in the design, analysis, and reporting of rheumatoid
arthritis clinical trials. Arthritis Rheum 33: 140-149, 1990
2. Ward JR, Williams HJ, Egger MJ, Reading JC, Boyce E,
Altz-Smith M, Samuelson CO Jr, Willkens RF, Solsky
MA, Hayes SP, Blocka KL, Weinstein A, Meenan RF,
Guttadauria M, Kaplan SB, Klippel J: Comparison of
auranofin, gold sodium thiomalate, and placebo in the
treatment of rheumatoid arthritis: a controlled clinical
trial. Arthritis Rheum 26:1303-1315, 1983
3. Makuch R, Simon R: Sample size requirements for
evaluating a conservative therapy. Cancer Treat Rep
62: 1037-1040, 1978
4. Sacks HS, Berrier J, Reitman D, Ancona-Berk VA,
Chalmers TC: Meta-analysis of randomized controlled
trials. N Engl J Med 316:45M55, 1987
5 . Ropes MW, Bennett GA, Cobb S, Jacox R, Jessar RA:
1958 revision of diagnostic criteria for rheumatoid arthritis. Bull Rheum Dis 9:175-176, 1958
6. Anderson JJ, Felson DT, Meenan RF, Williams HJ:
Which traditional measures should be used in rheumatoid arthritis clinical trials? Arthritis Rheum 32: 10931099, 1989
7. Eberl DR, Fasching V, Rahlfs V, Schleyer I, Wolf R:
Repeatability and objectivity of various measurements
in rheumatoid arthritis. Arthritis Rheum 19:1278-1286,
1976
8. Ritchie DM, Boyle JA, McInnes JM, Jasani MK, Dalakos T, Grieveson P, Buchanan W: Clinical studies with
an articular index for the assessment of joint tenderness
in patients with rheumatoid arthritis. Q J Med 37:393406, 1968
9. DerSimonian R, Charette LJ, McPeek B, Mosteller F:
Reporting on methods in clinical trials. N Engl J Med
306: 1332-1337, 1982
10. Anscombe FJ: The transformation of Poisson, binomial,
and negative-binomial data. Biometrika 5:246-254, 1958
11. DerSimonian R, Laird N: Meta-analysis in clinical trials.
Controlled Clin Trials 7: 177-188, 1986
12. Meenan RF, Anderson JJ, Lewis LE, Egger MJ, AltzSmith M, Samuelson CO Jr, Willkens RF, Solsky MA,
Hayes SP, Blocka KL, Weinstein A, Guttadauria M,
Kaplan SB, Klippel J: Outcome assessment in clinical
13.
14.
15.
16.
trials: evidence for the sensitivity of a health status
measure. Arthritis Rheum 27:1344-1352, 1984
Smith PR, Brown GMM, Meyers OL: An open comparative study of auranofin vs. gold sodium thiomalate. J
Rheumatol [Suppl] 9: 190-196, 1982
Horton RJ: Comparative safety and efficacy of auranofin
and parenteral gold compounds: a review. Scand J
Rheumatol [Suppl] 5 1 : 1 ~ 1 1 0 1983
,
Mackenzie AH: Antimalarial drugs for rheumatoid arthritis. Am J Med 75 (suppl6A):48-58, 1983
Weinblatt ME, Coblyn JS, Fox DA, Fraser PA, Holdsworth DE, Glass DN, Trentham DE: Efficacy of low-
EFFICACY AND TOXICITY METAANALYSES OF DRUGS FOR RA
dose methotrexate in rheumatoid arthritis. N Engl J Med
312:818-822, 1985
17. Anderson PA, West SG, O'Dell JR, Via CS, Claypool
RG, Kotzin BL: Weekly pulse methotrexate in rheumatoid arthritis: clinical and immunologic effects in a
randomized, double-blind study. Ann Intern Med 103:
489-496, 1985
18. Suarez-Almazor ME, Fitzgerald A, Grace M. Russell
AS: A randomized controlled trial of parenteral methotrexate compared with sodium aurothiomalate (myochrysine) in the treatment of rheumatoid arthritis. J
Rheumatol 15:753-756, 1988
19. Furst DE, Koehnke R, Burmeister LF, Kohler J, Cargill
I: Increasing methotrexate effect with increasing dose in
the treatment of resistant rheumatoid arthritis. J Rheumatol 16:313-320, 1989
20. Morassut P, Goldstein R, Cyr M, Karsh J. McKendry
RJR: Gold sodium thiomalate compared to low dose
methotrexate in the treatment of rheumatoid arthritis: a
randomized double blind 26-week trial. J Rheumatol
16~302-306, 1989
21. Hamdy H, McKendry RJR, Mierins E, Liver JA: Lowdose methotrexate compared with azathioprine in the
treatment of rheumatoid arthritis: a twenty-four-week
controlled clinical trial. Arthritis Rheum 30:361-368,
1987
22. Williams HJ, Willkens RF, Samuelson CO Jr. Alarcon
GS, Guttadauria M, Yarboro C, Polisson RP, Weiner
SR, Luggen ME, Billingsley LM. Dahl SL, Egger MJ,
Reading JC, Ward JR: Comparison of low-dose oral
pulse methotrexate and placebo in the treatment of
rheumatoid arthritis: a controlled clinical trial. Arthritis
Rheum 28:721-730, 1985
23. Schattenkirchner M, Missler B, Mulz D: Postmarketing
experience with auranofin in the Federal Republic of
Germany. Scand J Rheumatol [Suppl] 51:57-66, 1983
24. Williams HJ, Ward JR, Dahl SL, Clegg DO. Willkens
RF, Oglesby T, Weisman MH, Schlegel S, Michaels
RM, Luggen ME, Polisson RP, Singer JZ, Kantor S M ,
Shiroky JB, Small RE, Gomez MI, Reading JC. Egger
MJ: A controlled trial comparing sulfasalazine, gold
sodium thiomalate, and placebo in rheumatoid arthritis.
Arthritis Rheum 31:702-713, 1988
25. Nuver-Zwart IH, van Riel PLCM, van de Putte LB.
Gribnau FWJ: A double blind comparative study of
sulphasalazine and hydroxychloroquine in rheumatoid
arthritis: evidence of an earlier effect of sulphasalazine.
Ann Rheum Dis 48:389-395, 1989
26. Neumann VC, Grindulis KA, Huball S, McConkey B,
Wright V: Comparison between penicillamine and sulphasalazine in rheumatoid arthritis: Leeds-Birmingham
trial. Br Med J 287: 1099-1 102, 1983
27. 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
1459
28. Pullar T, Hunter JA, Capell HA: Sulphasalazine in the
treatment of rheumatoid arthritis: relationship of dose
and serum levels to efficacy. Br J Rheumatol 24:269278, 1985
29. Pullar T, Hunter JA, Capell HA: Sulphasalazine in
rheumatoid arthritis: a double blind comparison of sulphasalazine with placebo and sodium aurothiomalate. Br
Med J 287: 1102-1 104, 1983
30. Pinals RS, Kaplan SB, Lawson JG, Hepburn B: Sulfasalazine in rheumatoid arthritis: a double-blind, placebocontrolled trial. Arthritis Rheum 29: 1427-1434, 1986
31. Skoskey JL: Comparison of responses to and adverse
effects of graded doses of sulfasalazine in the treatment
of rheumatoid arthritis. J Rheumatol [Suppl] 155-8,
1988
32. Hill HFH: Treatment of rheumatoid arthritis with penicillamine. Semin Arthritis Rheum 6:361-388, 1977
33. Jaffe IA: The treatment of rheumatoid arthritis and
necrotizing vasculitis with penicillamine. Arthritis
Rheum 13:436-443, 1970
34. 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 31: 121130, 1990
REFERENCES FOR CLINICAL TRIALS USED IN
AT LEAST ONE METAANALYSIS
If clinical trials were published more than once, the
most recent publication is listed. E = used in efficacy
metaanalysis: T = used in toxicity metaanalysis.
1 . Anderson PA, West SG, O'Dell JR, Via CS, Claypool RG,
2.
3.
4.
5.
Kotzin BL: Weekly pulse methotrexate in rheumatoid arthritis:
clinical and immunologic effects in a randomized, double-blind
study. Ann Intern Med 103:489496, 1985 (E,T)
Baldassare AR, Weiss TD, Arthur E, Osborn TG, Moore T L ,
Zuckner J: Auranofin in the treatment of rheumatoid arthritis: a
two-year study. Mo Med 82:711-715, 1985 (E,T)
Barraclough D. Boyden K. Brooks P, Thomas D, Brook A,
Tymms K: A comparative study of auranofin. gold sodium
thiomalate and d-penicillamine in rheumatoid arthritis: a progress report, Auranofin: Proceedings of a Smith, Kline & French
International Symposium, Amsterdam, 1982. New York, Excerpta Medica, 1982 (E,T)
Bernhard GC: Auranofin treatment for adult rheumatoid arthritis: comparison of 2 mg and 6 mg daily dose. J Rheumatol
[Suppl] 9:149-153. 1982 (E,T)
Berry H, Liyanage SP, Durance RA, Barnes CG, Berger LA,
Evans S: Azathioprine and penicillamine in treatment of rheumatoid arthritis: a controlled trial. Br Med J 1:1052-1054, 1976
6. Bird HA, Gallez PL, Dixon JS, Surrall KE. Cole DS, Goldman
MH, Wright V: A single-blind comparative study of auranofin
and hydroxychloroquine in patients with rheumatoid arthritis.
Clin Rheumatol 3 (suppl 1):57-66, 1984 (E,T)
7. Bombardier CL, Wares J , Russell IJ, Larson M, Chalmers A,
Read JL: Auranofin therapy and quality of life in patients with
rheumatoid arthritis: results of multicenter trial. Am J Med
81:565-578, 1986 (E,T)
8. Capell HA, Hunter JA, Rennie AN, Murdoch RN: Levamisole:
a possible alternative to gold and penicillamine in the longterm
1460
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
treatment of rheumatoid arthritis? J Rheumatol8:730-740, 198I
(E,T)
The Cooperating Clinics Committee of the American Rheumatism Association: A controlled trial of gold salt therapy in
rheumatoid arthritis. Arthritis Rheum 16:353-358, 1973 (E,T)
Currey HLF, H a m s J, Mason RM, Woodland J, Beveridge T ,
Roberts CJ, Vere DW, Dixon ASJ, Davies J , Owen-Smith B:
Comparison of azathioprine, cyclophosphamide, and gold in
treatment of rheumatoid arthritis. Br Med J 3:763-766, 1974
(E,T)
Davies J , Bacon PA, Hall ND, Ring EFJ: Placebo-controlled
comparison of auranofin with myocrisin in patients with rheumatoid arthritis, Auranofin: Proceedings of a Smith. Kline &
French International Symposium, Amsterdam, 1982. New
York, Excerpta Medica, 1982 (E,T)
Davis P. Menard H, Thompson J , Harth M, Beaudet F: Oneyear comparative study of gold sodium thiomalate and auranofin
in the treatment of rheumatoid arthritis. J Rheumatol 12:60-67,
1985 (E,T)
Dequeker J, Francx L, Deckers Y: Comparison of oral gold
(auranofin) and injectable gold (allochrysine, gold sodium thiopropanolsulphate) treatment in patients with rheumatoid arthritis: a single-blind study, Auranofin: Proceedings of a Smith,
Kline & French International Symposium, Amsterdam, 1982.
New York, Excerpta Medica, 1982 (E,T)
Dixon ASJ. Davies J, Dormandy T L , Hamilton EBD, Holt PJL,
Mason RM, Thompson M, Weber JCP, Zutshi DW: Synthetic
D(-) penicillamine in rheumatoid arthritis: double-blind controlled study of a high and low dosage regimen. Ann Rheum Dis
34:41W21, 1975 (E,T)
Dwosh IL, Stein HB, Urowitz MB, Smythe HA, Hunter T,
Ogryzlo MA: Azathioprine in early rheumatoid arthritis: comparison with gold and chloroquine. Arthritis Rheum 20:685492,
1977 (E,T)
El-Ghobarey AF, Mavrikakis ME, Macleod M, Reynolds PMG,
Capell HA, Spencer DG, B a h t G, Mathieu JP, McAllister T,
Cooney A, Dick WC: Clinical and laboratory studies of levamisole in patients with rheumatoid arthritis. Q J Med 47:385400,
1978 (E,T)
Felix-Davies D, Stewart A, Wilkinson BR, Bateman JR,
Delamere JP: A 12-month comparative trial of auranofin and
d-penicillamine in rheumatoid arthritis. Am J Med 75:138-141.
1983 (E,T)
Finkelstein AE, Roisman FR, Batista V, de Nudelman FG, de
Titto EM, Mizraji M, Walz DT: Oral chrysotherapy in rheumatoid arthritis: minimum effective dose. J Rheumatol 7: 160-168,
1980 (E,T)
Fowler PD, Shadforth MF, Crook PR, Lawton A: Report on
chloroquine and dapsone in the treatment of rheumatoid arthritis: a &month comparative study. Ann Rheum Dis 43:200-204,
1984 (E,T)
Freedman A: Chloroquine and rheumatoid arthritis: a shortterm controlled trial. Ann Rheum Dis 15:251-257, 1956 (T)
Freedman A, Steinberg VL: Chloroquine in rheumatoid arthritis: a double-blind trial of treatment for one year. Ann Rheum
Dis 19:243-250, 1960 (E,T)
Furst DE, Koehnke R, Burmeister LF, Kohler J , Cargill I:
Increasing methotrexate effect with increasing dose in the
treatment of resistant rheumatoid arthritis. J Rheumatol 16:313320, 1989 (E,T)
Furst DE, Levine S, Srinivasan R, Metzger AL, Bangert R,
Paulus HE: A double-blind trial of high versus conventional
dosages of gold salts for rheumatoid arthritis. Arthritis Rheum
2011473-1480, 1977 (E,T)
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 (E,T)
Halberg P. Bentzon M, Crohn 0, Gad I, Halskov 0 , Heyn J,
FELSON ET AL
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
Ingemann N , Junker P, Lorenzen 1, Moller I, Hansen TM,
Olsen N, Pedersen G, Sorensen SF, Stage P: Double-blind trial
of levamisole, penicillamine and azathioprine in rheumatoid
arthritis. Dan Med Bull 31:403409, 1984 (E,T)
Hamdy H, McKendry RJR, Mierins E, Liver JA: Low-dose
methotrexate compared with azathioprine in the treatment of
rheumatoid arthritis: a twenty-four-week controlled clinical
trial. Arthritis Rheum 30:361-368, 1987 (E,T)
Hamilton EBD, Scott JT: Hydroxychloroquine sulfate
(“Plaquenil”) in treatment of rheumatoid arthritis. Arthritis
Rheum 5502-512. 1962 (E,T)
Hill HFH,’Hill AGS, Day AT, Brown RM, Golding JR, Lyle
WH: Maintenance dose of penicillamine in rheumatoid arthritis:
a comparison between a standard and a response-related flexible
regimen. Ann Rheum Dis 3k429-433, 1979 (E,T)
Hochberg MC: Auranofin or d-penicillamine in the treatment of
rheumatoid arthritis. Ann Intern Med 105528-535, 1986 (E,T)
Huskisson EC, Gibson TJ, Balme HW, Berry H, Burry HC:
Trial comparing D-penicillamine and gold in rheumatoid arthritis. Ann Rheum Dis 33532-535, 1974 (E,T)
Huskisson EC, Scott J, Balme HW, Dieppe PA, Trapnell J ,
Willoughby DA: Immunostimulant therapy with levamisole for
rheumatoid arthritis. Lancet I:393-395, 1976 (E,T)
Jarner D, Ammitzboll F , Sinding J, Menck Thygesen H: Auranofin and myocrisin (gold sodium thiomalate): a double-blind
comparative study, Auranofin: Proceedings of a Smith, Kline &
French International Symposium, Amsterdam, 1982. New
York, Excerpta Medica, 1982 (E,T)
Kadir N . Grayson MF, Campbell ED: A double-blind one year
clinical trial comparing alclofenac with gold in rheumatoid
arthritis. Ann Clin Res 15:4-9, 1983 (E,T)
Kahn MF, Doury P, Delcambre B, Desproges-Botteron R,
Gougeon J, Kuntz D, d’Eshougues JR, Roux H, Sany J: A
multicentre double-blind comparative study of auranofin and
intramuscular gold (allochrysine), Auranofin: Proceedings of a
Smith, Kline & French International Symposium, Amsterdam,
1982. New York, Excerpta Medica, 1982 (E,T)
Kaik B, Broll H, Muller-Fassbender H, Rau R, Schattenkirchner M, Zeidler H: Auranofin and tauredon in the treatment of
chronic polyarthritis, Auranofin: Proceedings of a Smith, Kline
& French International Symposium, Amsterdam, 1982. New
York, Excerpta Medica, 1982 (E)
Katz WA, Alexander S, Bland JH, Blechman W, Bluhm GB,
Bonebrake RA, Falbo A, Greenwald RA, Hartman S, Hobbs T,
lndenbaum S, Lergier J E , Lanier BG, Lightfoot RW Jr, Phelps
P, Sheon RP, Torretti D, Wenger ME, Wilske K: The efficacy
and safety of auranofin compared to placebo in rheumatoid
arthritis. J Rheumatol [Suppl] 9:173-178, 1982 (T)
Keegan DAJ, Cole DS, Hodkinson E , Podmore P: A comparative open study of the efficacy and safety of gold sodium
thiomalate and auranofin treatment in patients with rheumatoid
arthritis, Auranofin: Proceedings of a Smith, Kline & French
International Symposium, Amsterdam, 1982. New York, Excerpta Medica, 1982 (E,T)
Lewis D, Capell HA: Is auranofin preferable to gold sodium
thiomalate in the management of rheumatoid arthritis? Curr Clin
P ~ 7~147-155,
C
1987 (T)
Lund B, Graae A, Sonne I, Andersen RB: Long-term treatment
of rheumatoid arthritis with fenclofenac or gold: a controlled,
single blind trial comparing clinical and biochemical parameters.
Eur J Rheumatol lnflamm 6:279-287, 1983 (E,T)
I
Mainland D, Sutcliffe MI: Hydroxychloroquine sulfate in rheumatoid arthritis, a six month, double-blind trial. Bull Rheum Dis
131287-290, 1962 (E,T)
Makisara P, Nissila M, Kajander A, Martio JL: Comparison of
penicillamine and gold treatment in early rheumatoid arthritis.
Scand J Rheumatol7: 166-170, 1978 (E,T)
Manthorpe R, Horbov S, Sylvest J, Vinterberg H: Auranofin
versus penicillamine in rheumatoid arthritis: one-year results
EFFICACY AND TOXICITY METAANALYSES OF DRUGS FOR RA
from a prospective clinical investigation. Scand J Rheumatol
15:13-22, 1986 (E,T)
43. Mery C, Delrieu F, Ghozlan R, Saporta L, Simon F, Amor B,
Menkes CJ, Delbarre R: Controlled trial of D-penicillamine in
rheumatoid arthritis: dose effect and the role of zinc. Scand J
Rheumatol5:241-247, 1976 (E,T)
44. Morassut P, Goldstein R, Cyr M, Karsh J , McKendry RJR:
Gold sodium thiomalate compared to low dose methotrexate in
the treatment of rheumatoid arthritis: a randomized, double
blind 26-week trial. J Rheumatol 16:302-306, 1989 (E,T)
45. The Multicentre Trial Group: Controlled trial of D(-) penicillamine in severe rheumatoid arthritis. Lancet I:275-280, 1973
(E,T)
46. Neumann VC, Grindulis KA, Huball S, McConkey B, Wright
V: Comparison between penicillamine and sulphasalazine in
rheumatoid arthritis: Leeds-Birmingham trial. Br Med J 287:
1099-1 102, 1983 (E,T)
47. Nuver-Zwart IH, van Riel PLCM, van de Putte LBA, Gribnau
FWJ: A double blind comparative study of sulphasalazine and
hydroxychloroquine in rheumatoid arthritis: evidence of an
earlier effect of sulphasalazine. Ann Rheum Dis 48:38%395,
1989 (E,T)
48. Paulus HE, Williams HJ, Ward JR, Reading JC, Egger MJ,
Coleman ML, Samuelson CO Jr, Willkens RF, Guttadauria M,
Alarc6n GS, Kaplan SB, MacLaughlin EJ, Weinstein A, Wilder
RL, Solsky MA, Meenan RF: Azathioprine versus D-penicillamine in rheumatoid arthritis patients who have been treated
unsuccessfully with gold. Arthritis Rheum 27:721-727, 1984
49. Pavelka K, Sen KP, Peliskova Z, Vacha J , Trnavsky K:
Hydroxychloroquine sulphate in the treatment of rheumatoid
arthritis: a double blind comparison of two dose regimens. Ann
Rheum Dis 48542-546, 1989 (E,T)
50. Pinals RS, Kaplan SB, Lawson JG, Hepburn B: Sulfasalazine in
rheumatoid arthritis: a double-blind, placebo-controlled trial.
Arthritis Rheum 29: 1427-1434. 1986 (E,T)
51. Popert AJ, Meijers KAE, Sharp J, Bier F: Chloroquine diphosphate in rheumatoid arthritis: a controlled trial. Ann Rheum Dis
20:18-33, 1961 (E,T)
52. Pullar T, Hunter JA, Capell HA: Sulphasalazine in rheumatoid
arthritis: a double blind comparison of sulphasalazine with
placebo and sodium aurothiomalate. Br Med J 287: 1102-1 104,
1983 (T)
53. Pullar T, Hunter JA, Capell HA: Sulphasalazine in the treatment of rheumatoid arthritis: relationship of dose and serum
levels to efficacy. Br J Rheumatol 24:26%276, 1985 (E,T)
54. Raeman F, Hanssens F, Delattre M: An European open multicentre trial with auranofin in rheumatoid arthritis. Clin Rheumatol 3 (suppl 1):33-38, 1984 (E,T)
55. Rau R, Schattenkircher M, Muller-Fassbender H, Kaik B,
Zeidler H: A three year comparative multicenter study of
auranofin (AF) and gold sodium thiomalate (GST) in the treatment of rheumatoid arthritis (RA). Clin Rheumatol 6 (suppl
2):43-52, 1987 (E,T)
56. Research Sub-committee of the Empire Rheumatism Council:
Gold therapy in rheumatoid arthritis: report of a multi-centre
controlled trial. Ann Rheum Dis 19:95-119, 1960 (E,T)
57. Richards IM, Fraser SM, Hunter JA, Capell HA: Comparison of
phenytoin and gold as second line drugs in rheumatoid arthritis.
Ann Rheum Dis 465674169, 1987 (E,T)
58. Schattenkirchner M, Kaik B, Muller-Fassbender H, Rau R,
Zeidler H: Auranofin and sodium aurothiomalate in the treatment of rheumatoid arthritis: a double-blind, comparative multicenter study. J Rheumatol [Suppl] 9:184189, 1982 (T)
59. Scott DL, Dawes PT, Tunn E, Fowler PD, Shadforth MF,
1461
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 28:12%133, 1989 (E,T)
60. Shiokowa Y, Horiuchi Y, Honma M, Kageyama T, Okada T.
Azuma T: Clinical evaluation of D-penicillamine by multicentric
double-blind comparative study in chronic rheumatoid arthritis.
Arthritis Rheum 20:1464-1472, 1977 (T)
61. Sigler JW, Bluhm GB, Duncan H, Sharp JT, Ensign DC,
McCrum WR: Gold salts in the treatment of rheumatoid arthritis: a double-blind study. Ann Intern Med 80:21-26, 1974 (E,T)
62. Suarez-Almazor ME, Fitzgerald A, Grace M, Russell AS: A
randomized controlled trial of parented methotrexate compared with sodium aurothiomalate (myochrysine) in the treatment of rheumatoid arthritis. J Rheumatol 15:753-756, 1988
(E,T)
63. 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 (E,T)
64. Tvede N , Strandberg B: The effect of auranofin compared with
myocrisin in the treatment of rheumatoid arthritis, Auranofin:
Proceedings of a Smith, Kline & French International Symposium, Amsterdam, 1982. New York, Excerpta Medica, 1982 (E)
65. Van Riel PLCM, Larsen A, van de Putte LBA, Gribnau FW:
Effects of aurothioglucose and auranofin on radiographic progression in rheumatoid arthritis. Clin Rheumatol 5:359-364,
1986 (T)
66. Vischer TL: A double blind multicentre study of OM-8980 and
auranofin in rheumatoid arthritis. Ann Rheum Dis 47582-587,
1988 (E,T)
67. Walravens M, Dequeker J: Comparison of gold and orgotein
treatment in rheumatoid arthritis. Curr Ther Res 20:6249, 1976
(E,T)
68. Ward JR, Williams HR, Egger MJ, Reading JC, Boyce E,
Altz-Smith M, Samuelson CO Jr, Willkens RF, Solsky MA,
Hayes SP, Blocka KL, Weinstein A, Meenan RF, Guttadauria
M, Kaplan SB. Klippel J: Comparison of auranofin, gold sodium
thiomalate, and placebo in the treatment of rheumatoid arthritis:
a controlled clinical trial. Arthritis Rheum 26: 1303-1315. 1983
(E,T)
69. Weinblatt ME. Coblyn JS, Fox DA, Fraser PA, Holdsworth
DR. Glass DN, Trentham DE: Efficacy of low-dose methotrexate in rheumatoid arthritis. N Engl J Med 312:818-822, 1985
(E,T)
70. Wenger ME, Alexander S. Bland JH, Blechman WJ: Auranofin
versus placebo in the treatment of rheumatoid arthritis. Am J
Med 75 (suppl 6A):123-127, 1983 (E,T)
71. Williams HJ, Ward JR, Dahl SL, Clegg DO, Willkens RF,
Oglesby T, Weisman MH, Schlegel S, Michaels RM, Luggen
ME, Polisson RP, Singer JZ. Kantor SM, Shiroky JB, Small
RE, Gomez MI, Reading JC, Egger MJ: A controlled trial
comparing sulfasalazine, gold sodium thiomalate, and placebo
in rheumatoid arthritis. Arthritis Rheum 31:702-713, 1988 (E,T)
72. Williams HJ, Ward JR, Reading JC, Egger MJ, Grandone JT,
Samuelson CO, Furst DE, Sullivan JM, Watson MA, Guttadauria M, Cathcart ES, Kaplan SB, Halla JT, Weinstein A, Plotz P:
Low-dose D-penicillarnine therapy in rheumatoid arthritis: a
controlled, double-blind clinical trial. Arthritis Rheum 26:581592, 1983 (E,T)
73. Williams HJ, Willkens RF, Samuelson CO Jr, Alarcon GS,
Guttadauria M, Yarboro C, Polisson RP, Weiner SR, Luggen
ME, Billingsley LM, Dahl SL. Egger MJ, Reading JC, Ward JR:
Comparison of low-dose oral pulse methotrexate and placebo in
the treatment of rheumatoid arthritis: a controlled clinical trial.
Arthritis Rheum 28:721-730, 1985 (E,T)
Документ
Категория
Без категории
Просмотров
0
Размер файла
1 358 Кб
Теги
second, two, efficacy, drug, toxicity, metaanalysis, arthritis, results, comparative, line, rheumatoid
1/--страниц
Пожаловаться на содержимое документа