arthrit&

and
rheumatism
Official Journal of the American Rheumatism Association Section of the Arthritis Foundation

AZATHIOPRINE VERSUS D-PENICILLAMINE IN

RHEUMATOID ARTHRITIS PATIENTS WHO HAVE
BEEN TREATED UNSUCCESSFULLY WITH GOLD
HAROLD E . PAULUS, H. JAMES WILLIAMS, JOHN R. WARD, JAMES C. READING,
MARLENE J. EGGER, MIKI L. COLEMAN, CECIL 0. SAMUELSON, JR., ROBERT F. WILLKENS,
MARIA GUTTADAURIA, GRACIELA S. ALARCON, STANLEY B. KAPLAN,
EDMUND J . MAcLAUGHLIN, ARTHUR WEINSTEIN, RONALD L. WILDER,
MARILYN A. SOLSKY, and ROBERT F. MEENAN

Two hundred six patients were entered into a
prospective controlled, double-blind, multicenter trial
comparing azathioprine (AZA) 1.25-1.5 mg/kg/day with
D-penicillamine (DP) 10-12 mg/kg/day. One hundred
thirty-four patients completed 24 weeks of therapy.
_ _ -~
From the Cooperative Systematic Studies of Rheumatic
Diseases Program, University of Utah School of Medicine. Salt
Lake City.
Supported by NIADDK contract no. I-AM-6-2218. by
Public Health Service research grants RR-64 and RR-287 from the
Division of Research Resources, and by Burroughs Wellcome
Company.
Harold E. Paulus, MD: Division of Rheumatology, School
of Medicine, University of California, Los Angeles; H . James
Williams, MD: Department of Internal Medicine, University of Ltah
School of Medicine; John R. Ward, MD: Department of Internal
Medicine, University of Utah School of Medicine: James C. Read-
ing, PhD: Department of Family and Community Medicine, Univer-
sity of Utah School of Medicine; Marlene J. Egger. PhD: Depart-
ment of Family and Community Medicine, University of 1Jtah
School of Medicine; Miki L. Coleman, MS: Department of Family
and Community Medicine, University of Utah School of Medicine;
Cecil 0. Samuelson, Jr., MD: Department of Internal Medicine,
University of Utah School of Mydicine; Robert F. Willkcns, MD:
Division of Arthritis, Harborview Medical Center, Seattle. Wash-
ington; Maria Guttadauria, MD: Department of Medicine, State
University of New York. Downstate Medical Center, Brooklyn;
Graciela S. A l a r c h , MD: Division of Clinical Immunology and
Rheumatology, University of Alabama in Birmingham; Stanley B.
Kaplan, MD: Department of Medicine, University of Tennessee,
Memphis: Edmund J. MacLaughlin, MD: Division of Immunology
and Rheumatology, School of Medicine, University of Missouri-
Columbia; Arthur Weinstein, MD: Division of Rheumatic Diseases,
University of Connecticut, Farmington; Ronald L. Wilder. MD:
National Institute of Arthritis, Diabetes, Digestive and Kidney
Diseases, Bethesda, Maryland; Marilyn A. Solsky. MD: Arthritis
Research Unit. University of Michigan, Ann Arbor; Robert F.
Meenan, MD: Arthritis Section, Boston University Medical Center,
Boston, Massachusetts.
Address reprint requests to H. James Williams, MD, Uni-
versity of Utah Medical Center, 50 North Medical Drive. Salt Lake
City, UT 84132.
Submitted for publication October 10. 1983; accepted in
revised form February 24. 1984.
Improvement in nearly all efficacy variables was seen in
both groups. Patients taking DP demonstrated a greater
rise in hemoglobin concentration and greater fall in
erythrocyte sedimentation rate than patients receiving
AZA; these were the only efficacy variables with a
signifieant difference between the treatment groups.
Fewer withdrawals for adverse reactions occurred
among the patients receiving AZA, but the difference
was not significant. Patients receiving AZA were with-
drawn from the drug mainly for abnormal liver function
test results, nausea and gastrointestinal upset, and
leukopenia. The main reasons for withdrawal of patients
receiving DP were nausea, rash and pruritis, thrombo-
cytopenia, dysgeusia, and proteinuria.
Although patients with rheumatoid arthritis
(RA) usually respond to a conservative program of
nonsteroidal antiinflammatory drugs, rest, and physi-
cal therapy, certain patients have a more progressive
or persistcnt course and are considered candidates for
slow-acting or potentially disease-modifying antirheu-
matic drugs (SAARD). Chrysotherapy is frequently
considered as the initial choice for these second-line
drugs. Toxicity, failure to respond, or loss of efficacy
often limit its usefulness. Published reports suggest
that approximately half the patients discontinue gold
therapy during the first year, and only 10% continue it
as long as 5 years (1). Alternatives to parenteral gold,
either as the initial SAARD or for patients who are
unresponsive or intolerant to gold therapy, include D-
penicillamine (DP) and azathioprine (AZA).. Both
drugs havc been shown to be effective (2-14), but only
a single study compares their relative efficacy, toler-
ability, and safety (15).
In this paper we report the results of a prospec-

Arthritis and Rheumatism, Vol. 27, No. 7 (July 1984)

72 1
722 PAULUS ET AL

Table 1. Characteristics at entry of patients entering and patient. All patients had 6 or more potentially responsive
completing the trial, by treatment group* swollen joints and at least 2 of the following: 1) 9 or more
potentially responsive joints that were tender on pressure; 2)
Patients complet- 45 minutes or more of morning stiffness; and 3) Westergren
ing 24 weeks of
erythrocyte sedimentation rate (ESR) of 2 2 8 mm/hour.
All patients drug As background therapy, an optimal stable dose of
Variable AZA DP AZA DP aspirin or another NSAID was maintained throughout the
N 104 102 70 64
study. A stable dose of glucocorticoid not exceeding the
Mean.age 54 55 53 55 equivalent of 10 mg prednisone daily was taken by 82
Malelfemale 31/73 25/77 24/46 17/47 patients. Entry criteria were used to exclude patients with
Severity of disease contraindications to these drugs, e.g., childbearing potential,
Mild 11 13 7 8 thrombocytopenia, leukopenia, serious coexisting diseases,
Moderate 60 60 37 39 or history or presence of cancer (17). Patients had not
Severe 33 29 26 17 received gold, antimalarials, intraarticular steroids, or par-
Race enteral steroids for at least 2 months before entry. Patients
White 81 83 58 53 had not previously received AZA or DP.
Black 15 11 10 8
Other 8 8 2 3
Treatment plan. Patients randomized to the AZA
Functional class group received a daily dose of 1.25-1.5 mg/kg, provided as
I 9 8 6 5 25-mg tablets and placebo capsules. Patients assigned to the
11 61 71 39 43 DP group received a daily dose of 10-12 mg/kg, supplied as
111 33 23 24 16 125-mg capsules and placebo tablets. The dose of DP was
IV 1 0 1 0 initiated at 2 capsules daily and increased by up to 2 capsules
Mean duration of daily every 4 weeks until the required weight-adjusted dose
disease (mcnths) 136 132 133 129 was reached or until a maximum dose of 8 capsules (1 ,OOO
* P > 0.31 for all comparisons. AZA azathioprine; DP = D- mg) was achieved. Placebo tablets and capsules were identi-
penicillamine. cal to active medication in appearance.
Supplemental propoxyphene, acetaminophen, or CD-
deine was permitted and all doses were recorded in a
tive, double-blind, randomized study that compares medication record. In order to preserve the double-blind
status of the trial and to conform with reasonable clinical
the efficacy, tolerability, and safety of AZA a n d DP in management, all patients who were withdrawn for an ad-
RA patients with active synovitis w h o had been treat- verse reaction and those who withdrew for lack of response
e d unsuccessfully with gold. T h e study was conducted after 6 or more weeks of treatment were permitted to cross
by t h e 11 clinics of the Cooperative Systematic Studies over to the alternative medication and continue under dou-
of Rheumatic Diseases (CSSRD)*. ble-blind conditions. Response to the second medication was
not analyzed.
Patients were seen every 2 weeks to evaluate side
PATIENTS AND METHODS effects, to obtain complete blood cell counts and urinalyses,
Patients. A total of 206 consenting patients with and to monitor drug compliance by pill and capsule counts.
adult-onset definite or classic RA (16) of greater than 6 In addition, every 6 weeks and at the time of withdrawal a
months duration were accepted into the trial and randomized blood chemistry survey, ESR, and clinical assessment for
to receive either AZA or ISP for 24 weeks. The patients all efficacy were done. Patients were withdrawn if repeat study
had had an inadequate response to conventional antiinflam- confirmed a white blood cell count <3,500/mm3, neutrophil
matdry therapy of aspirin and/or other nonsteroidal antiin- count < 1,500/rnm~,platelet count < l O 0 , 0 ~ / m m ~urine
, pro-
flammatory drugs (NSAID) and had been treated unsuccess- tein excretion > 1 .0 gm/24 hours, or liver function test result
fully with gold injections. Unsuccessful treatment with gold (serum glutamic oxaloacetic transaminase, lactic dehydroge-
was defined as inadequate control of arthritis, the occur- nase, alkalifie phosphatase, bilirubin) more than twice the
rence of gold toxicity, or refusal of gold therapy by the upper limit of normal.
Efficacy evaluations. Efficacy was assessed in 2
~ ~. - groups-patients who completed the study and patients who
* The participating clinics and clinic directors are: Univer- entered the study-as follows.
sity of Alabama in Birmingham, Graciela S . Alarch, MD; Universi- Completed study. Seventy patients who completed
ty of California, Los Angeles, Harold E. Paulus, MD; University of 24 weeks of AZA treatment were compared with 64 patients
Connecticut, Arthur Weinstein, MD; National Institute of Arthritis, who completed 24 weeks of DP treatment, to evaluate
Diabetes, Digestive and Kidney Diseases, Arthritis and Rheumatism efficacy in patients who were able to continue the drug. The
Branch, John Klippel, MD; Boston University, Robert F. Meenan, methods used for efficacy assessment have been used by the
MD; University of Michigan, William Mikkelsen, MD; University of
Missouri-Columbia, Gordon Sharp, MD; State University of New CSSRD in previous trials (17). The same observer examined
York, Downstate Medical Center, David Kaplan, MD; University of each patient throughout the trial.
Tennessee, Stanley B. Kaplan, MD; University of Utah. Cecil 0. Observer assessments included: (a) joint count for
Samuelson, Jr., MD; University of Washington, Harborview Medi- tenderness on pressure andlor pain on motion for 68 diar-
cal Center, Robert F. Willkens. MD. throdial joints, graded on a scale of 0 = none, 1 = positive
Table 2. Changes in clinical and laboratory variables in patients completing 24 weeks of therapy, by treatment group*
Azathioprine D-penicillamine
Variable n Baseline 24 weeks Difference n Baseline 24 weeks Difference Pt
Morning stiff- 70 200 !. 212.6 105 t 126.4 95 215.3 64 210 -t 252.3 85 2 111.8 125 t 231.9 NS
ness, minutes I50 (0- I ,440) 60 (0-450) 60 (-390-1,125) 143 (0-1.440) 60 (0-530) 90 (-530-1,080)
Grip strength 70 99 5. 47.5 122 t 69.4 +
23 46.4 62 99 t 65.8 123 t 69.0 24 t 41.5 NS
(right), mm Hg 83 (3 1-237) 100 (29-300) 7 (-81-178) 80 (32-300) 107 (40-300) 16 (-61-141)
Grip strength 70 97 i 48.4 122 t 66.9 25 :+ 41.2 62 +
94 61.5 121 2 64.6 27 -t 43.9 NS
(left), mm Hg 87 (27-285) 98 (39-300) 9 (-43-137) 74 (29-300) 108 (39-297) 16 (-79-133)
Patient assessment scale, 70 3 0.7 2 t 0.7 +
I 1.0 64 3= 1.0 2 t 0.8 1 2 1.1 NS
1-5 increasing severity 3 (2-5) 2 (1-4) I(-2-2) 3 (1-5) 2 ( 1-4) 1 (-2-4)
Physician assessment 67 3 t 0.7 3 t 0.7 1 t 0.9 62 3 2 0.9 2 t 0.6 I t 0.9 NS
scale, 1-5 a s above 3 (2-5) 3 (1-4) 0 (-2-3) 3 (1-5) 2 (1-4) 1 (- 1-3)
Functional class 69 2 + 0.5 2 t 0.7 0 t 0.5 62 2 2 0.6 2 t 0.5 0 t 0.4 NS
scale, 1-4 2 (24) 2 (1-4) 0 ( 1-1) 2 (1-3) 2 (1-3) 0 ( - 1-1)
Anatomic stage 67 3 t 0.7 3 ? 0.7 0 2 0.5 60 2I 0.7 3 ? 0.7 0 i 0.5 NS
scale, 1-4 2 (1-4) 2 (2-4) 0 (-1-1) 2 (1-4) 3 (1-4) 0 (-2-1)
Painltenderness 68 32 t 13.5 22 ? 16.7 I I 2 12.6 63 32 2 14.8 21 t 16.9 I 1 t 14.0 NS
count 31 (9-60) 18 (0-58) I I ( 17-39) 31 (7-59) 16 (0-59) 10 (-23-52)
Swelling count 68 26 t 12.0 18 2 11.2 8 t- 10.7 63 25 i 12.0 15 ? 10.8 l o t 11.0 NS
25 (4-56) 15 (1-58) 9 (-14-43) 23 (6-54) 14 (0-46) 10 ( - 1 9 4 1 )
Painltenderness 68 49 t 24.6 30 2 25.8 19 ?: 18.9 63 51 + 31.0 31 _t 30.2 20 i 30.3 NS
score 44 (10-108) 21 (0-113) 20 (-31-60) 47 (9-155) 22 (0-132) 20 (-92-151)
Swelling score 68 38 t 21.9 23 2 15.0 IS ?- 19.7 63 +
37 20.6 20 -+ 17.0 16 i 17.8 NS
33 (5-121) 21 (1-64) 14 (-32-1 I I ) 32 (7-93) 14 (0-77) 14 (-16-79)
Hemoglobin (gm/dl) 60 13.1 t 1.84 13.0 2 1.36 -0.1 I 0.98 59 12.8 f- 1.42 13.4 ? 1.44 0.6 t 1.24 <O.OOI
13.0 (8.9-17.1) 12.9 (9.8-16.1) -0.1 (-2.5-1.9) 13 (9.2-15.9) 13.7 (9.6-16.8) 0.4 (- 1.7-4.3)
Platelets 68 352 t 122.2 325 _t 108.7 -27 101.5 60 359 .! 104.9 295 ? 73.2 -64 t 108.2 0.023
( X I03/mm3) 334 (143-682) 3 13 ( 145-609) - 19 (-394-281) 353 (210-650) 289(135-585) -50 (-365-195)
Westergren erythrocyte 66 46 t 30.2 38 ? 25.2 -8 2 22.7 60 45 t- 28.7 29 ? 19.8 -16 t 24.5 0.012
sedimentation rate (mm/hour) 37 (1-128) 30 (3-1 17) -3 ( 99-44) 39 (3-122) 23 (2-79) - 13 (-77-48)

* Upper line of numbers for each variable represents mean 2 SD; lower line represents median and range (in parentheses).
t By analysis of covariance, azathioprine group versus D-penicillamine group. N S = not significant (P > 0.05).

4
N
w
724
O* -AZATHIOPRINE
PAINFUL/ -
!
5
4
\.
- - -- D-PEN IC ILLAMIN E
TENDER
JOINT
SCORE
-20
SWOLLEN
-- --:-==.- - 8
JOINT
SCORE
-20 I 1 I 1
0 6 12 18 24
WEEKS
Figure 1. Mean change in joint scores over time, by treatment group.
response on questioning, 2 = spontaneous response elicited,
3 = withdrawal by patient on examination: (b)joint count for
swelling for 66 joints, graded on a scale of 0 = none, 1 =
detectable synovial thickening without loss of bony con-
tours, 2 = loss of distinctness of bony contours, 3 = bulging
synovial proliferation with cystic characteristics; (c) joint
tenderness/pain and joint swelling scores, derived by sum-
ming the severity scores for each joint; (d) grip strength,
using a mercury column sphygmomanometer with a standard
grip bag (18); and (e) physician assessment of disease
activity on a scale of 1 = asymptomatic, 2 = mild, 3 =
moderate, 4 = severe, 5 = very severe. Physician assess-
ment was not based on specific criteria but on the clinical
judgment of the observer. Patient assessment included: (a)
duration of morning stiffness on the day prior to the visit;
and (b) overall assessment of arthritis activity. graded as 1 =
very good, 2 = good, 3 = fair, 4 = poor. 5 = very poor.
Entered study. To evaluate the efficacy for patients
starting the drugs, 93 patients who received AZA were
comparcd with 99 patients who received DP. Excluded from
evaluation were 4 patients who were withdrawn before 6
weeks of drug therapy because of lack of cooperation or
worsening of arthritis, 2 patients who moved from the area, 3
who withdrew for an unrelated disease (hypernephroma,
staphylococcal sepsis, ruptured intracranial aneurysm), and
5 who were dropped for protocol violations (2 were given the
wrong dose, 1 had leukopenia at baseline, 2 had violation of
allowed background medication).
Patients who did not complete the study because of
an adverse reaction or for lack of response or lack of
cooperation after 6 weeks or more of therapy were included
in the analysis with respect to the initial treatment. It is
recognized that some of these patients had not taken the
drug long enough to experience benefit, but they are includ-
ed in this analysis to provide a more accurate assessment of
potential benefit for a patient who begins therapy. The
following classifications of response were defined, and the
percentage of all patients entering the study who met the
criteria for one of these classifications was determined for
each treatment group:
PAULUS ET AL
(a) Remission was defined using the prelimi-
nary criteria of the American Rheumatism Association
(19).
(b) Meaningful improvement in physician or
patient overall assessment of disease activity was
defined as an improvement by 2 or more gradations in
the 5-point scale or from grade 2 to grade 1.
(c) Meaningful improvement for joint swell-
ing or painltenderness was defined as a net decrease of
30% or more in the number of swollen or tenderjoints,
i.e., the number of joints that improved exceeded the
number of joints that worsened by at least 30% of
those initially involved and capablc of response. For a
single joint, improvement in tenderness or swelling
was considered to be a 2-gradation change or a change
to no involvement, using the 4-point severity scale
(17). Worsening of a joint was determined by new
involvement of a previously uninvolved joint or a
change to maximal involvement. For example, if a
patient was admitted to the trial with 10 initially
swollen joints and had improvement in 5 joints and
worsening in 2 joints, he or she would then still have 7
swollen joints and would qualify as having meaningful
improvement in joint swelling with a 30% change.
Three patients were not included in the analysis of
meaningful improvement of joint swelling, joint pain/
tenderness, or physician assessment because there
were differcnt physician observers assessing them at
the beginning and the conclusion of the trial.
Statistical analysis. Demographic data are described
by percent distributions. means, medians, ,standard devi-
ations, and ranges. The Mantel-Haenszel xL or the Mann-
Whitney procedures were used to test non-normally dis-
tributed variables. The former also permits control for
stratification. Normally distributed variables were evaluated
by analysis of covariance or the 1-test, as appropriate. The
ranked variables for the clinical response values were com-
pared by the Mantel-Haenszel xz test.
50.
W
>
0
a 7 AZATHIOPRINE
a 40
I 2 D-PENICILLAMINE
ul
I-
5
- 30 P 2.39
+
a
a
LL
0 20
w
0
I4
-
w
10
V
a
W
a
PHYSICIAN PATIENT JOINT PAIN/ JOINT
ASSESSMENT ASSESSMENT TENDERNESS SWELLING
Figure 2. Percent of patients entering study and qualifying for
analysis who experienced meaningful improvement, by treatment
group.
AZATHIOPRINE AND D-PENICILLAMINE IN RA
RESULTS
Patient population. Each clinic participating in
the study enrolled 5-36 of the 206 patients. Statistical
testing by the Mantel-Haenszel procedure demonstrat-
ed that no significant bias was introduced by the use of
patients from multiple clinics. One hundred four pa-
tients were assigned to the AZA group and 102 to the
DP group. The 2 groups did not differ significantly at
the beginning of the trial in age, sex, severity of dis-
ease, duration of disease, or functional class (Table 1).
Efficacy in patients completing 24 weeks of drug
therapy. The 70 patients who completed 24 weeks of
therapy with A Z A were similar to the 64 patients who
completed the trial with DP (Table 1). The mean and
median daily doses of AZA were 92 rng and 100 mg,
respectively, with a range of 50-125 mg. The mean and
median daily doses of D P were 740 mg and 750 mg
respectively, with a range of 500-1,000 mg. The mean
daily dose of prednisone was 5.8 mg in 33 patients in
the AZA group and 6.1 mg in 25 patients in the DP
group. Patients in both treatment groups showed im-
provement in nearly all efficacy variables (Table 2).
Improvement in AZA patients did not differ
significantly from that in DP patients, except that the
rise in hemoglobin concentration and fall in ESR were
statistically significantly better for DP patients. Hemo-
globin concentration decreased slightly in the AZA
group; however, patients initially anemic (hemoglobin
concentrations less than 12 g d d l ) improved with both
medications, with a rise of 0.76 gm/dl for 17 AZA
patients and 1.51 g d d l for 15 DP patients (P = 0.12).
In those patients with sedimentation rates greater than
28 mm/hour at baseline, values at 24 weeks decreased
by 15 mm/hour for those taking AZA (42 patients) and
24 mm/hour for those taking DP (41 patients) ( P :
0.01). Platelet counts decreased more in patients tak-
ing DP than in those receiving AZA, but it is not clear
whether this reflects efficacy or toxicity. The drop in
platelet count was significant ( P = 0.001) only in those
patients with an initial platelet count greater than
450,000/mm3, suggesting that the effect was a result of
efficacy (20).
When values for all assessments were plotted
over time (illustrated for painful joint count and score
and for swollen joint score in Figure l), maximum
improvemen! did not appear to have been reached
during the first 24 weeks, since the curves do not yet
show the development of a plateau.
Meaningful improvement in patients starting the
study. An important aspect of any therapy is the
725
Table 3. Patients withdrawn for adverse reactions, by treatment
group
Reaction Azathioprine D-penicillamine
Abnormal liver enzymes or 5 I
hepatitis
Nausea, vomiting, or 9 5
gastrointestinal pain
Dizziness 0
Rash or pruritus I o*
Leukopenia I
Thromboc ytopenia 5
Loss of taste or dysgeusia 2
Proteinuria 4*
Oral and vaginal ulcers
Herpes zoster
Total 20 29
* One patient had both rash and proteinuria.
proportion of patients started on therapy who obtain
benefit. Such an analysis requires a definition of what
constitutes an adequate response. We developed defi-
nitions for meaningful improvement for physician and
patient assessment of disease severity and joint count
variables (see Patients and Methods). These defini-
tions are arbitrary, but were considered necessary to
describe a degree of response that most physicians
would accept as meaningful.
No remissions occurred in the 192 patients who
were entered into the trial and qualified for this
analysis. The percentages of patients who experienced
meaningful improvement as assessed by physician
assessment, patient assessment, joint painknderness,
or joint swelling did not differ significantly in the 2
treatment groups (Figure 2). The response to both
drugs was modest by these criteria, with fewer than
half of the patients improving in joint count variables
and only one-sixth or fewer improving by patient or
physician assessment. The mean doses of drug taken
were similar in patients who had meaningful improve-
ment in any variable and those who did not.
Adverse reactions. During this 24-week study,
adverse reactions led to the withdrawal of 20 of the
patients assigned to the A Z A group and 29 of those
assigned to the D P group (Table 3). With AZA,
gastrointestinal upset (nausea, vomiting, pain), abnor-
mal liver function test results or hepatitis, and leuko-
penia accounted for 18 of the 20 withdrawals for
untoward effects. In the DP group, 9 patients withdrew
for rash, 5 for thrombocytopenia (22,000-81 ,OOO/
mm), 3 for proteinuria (2+ to 4+ by dipstick), and 1
for rash and proteinuria. Herpes zoster occurred in 1
patient who was receiving DP.
726
One patient who was excluded from the analy-
sis due to the occurrence of a serious unrelated disease
deserves special mention. A 60-year-old man with RA
of 4 years duration died of staphylococcal sepsis after
8 weeks of treatment with AZA. The source of the
sepsis was presumed to be pneumonia. The leukocyte
count was always greater than 10,000/mm3, and the
physicians caring for the patient did not consider the
death to be drug-related.
DISCUSSION
Azathioprine and D-penicillamine have been
approved by the Food and Drug Administration for
use in rheumatoid arthritis. Both drugs have demon-
strated efficacy, but are associated with significant
toxicity. There has been only one direct comparison
between the 2 agents (15) and a physicians choice of
agent is usually made empirically. Because the effica-
cy of both AZA and DP in RA has been documented in
previous publications (2-14), comparison with placebo
was not included in this study. The degree of improve-
ment in the various efficacy variables found in this
study was comparable with that reported previously
for these drugs in RA (4,5,8,11,17), and although
limited to patients who had had unsuccessful gold
therapy, this study supports the conclusion that these
drugs are efficacious in RA patients. Efficacy mea-
sures were continuing to improve at the end of 24
weeks (Figure l), suggesting that continued therapy is
necessary to obtain the maximum potential benefit
from these drugs in the doses given. Others have
reported no change in patients work ability until AZA
had been given for up to 20 months (9).
Among patients who completed the study, with-
in the constraints of the sample size and statistical
testing, efficacy of the 2 drugs appears to be equal.
Only laboratory variables demonstrated any statisti-
cally significant differences at the 5% level. The
changes in ESR clearly favor DP, although the actual
difference in values is not impressive. The changes in
platelet count with DP and in hemoglobin concentra-
tion with AZA may reflect the adverse effect potential
of these drugs, rather than efficacy.
With the doses used, DP was associated with
more withdrawals for adverse effects than was AZA.
Some of the untoward reactions are dose-dependent
and, in an unblinded clinical situation, the 4 AZA
patients who were withdrawn for leukopenia and 5 DP
patients who were dropped for thrombocytopenia
might have been able to continue the drugs at a lower
PAULUS ET AL
dosage. The spectrum of side effects was as anticipat-
ed for these drugs, but the prevalence of nausea,
vomiting, and hepatotoxicity emphasizes the value of
monitoring liver function in patients taking AZA. Rare
adverse reactions such as myasthenia gravis and other
induced autoimmune syndromes associated with DP
were not seen. The study design did not permit assess-
ment of the tumorigenicity of the drugs.
In a 1-year, single-blind, external observer trial
comparing 2.5 mglkglday AZA with 1,000 mglday D P
in 65 patients, Berry et a1 (15) reported significantly
greater improvement in pain relief after 6 months of
DP as compared with AZA; after 9 months of therapy,
however, 6 patients receiving DP had flares, while no
patients receiving AZA worsened. By 12 months the 2
treatments were comparable for pain relief. The au-
thors recognized that the standard deviation was large
and patient numbers were relatively small (23 AZA
patients and 27 DP patients completed the trial), which
may account for the differences seen at 6 months. The
results of our study, using smaller doses of the 2 drugs,
did not reveal any difference between the 2 treatments
at 24 weeks.
Our study confirms the observation by Berry
and coworkers (15) that the rise in hemoglobin concen-
tration and the fall in ESR were greater in the DP
group at 6 months. At the end of 12 months, only the
difference in hemoglobin concentration was still statis-
tically significant in the study by those authors. They
reported more withdrawals from the AZA group than
from the DP group, but they used a higher dose of
AZA, and most of these withdrawals were for nausea.
However, their rate of withdrawal for adverse reac-
tions of patients receiving AZA was similar to that
seen in our study and was consistent with others (8).
Conversely, their withdrawal rate for adverse reac-
tions to DP was lower than that seen in our study or
reported in other trials (1 1,14,17). The total experience
would indicate that DP is perhaps slightly less well
tolerated than AZA, but that the difference is small.
For those patients who tolerated either medication,
AZA and DP were about equally effective.
To investigate the possibility that a sizable
difference between AZA and D P exists but was not
detected in this study because of the number of
patients included and the methods used (Type I1
error), power calculations were done based on per-
centage improvements in joints from the initial visit.
This measure was selected because it was thought to
be the most meaningful. The largest difference be-
tween the 2 drugs in the population that could have
AZATHIOPRINE AND D-PENICILLAMINE IN RA
been missed with more than 10% probability at the 5%
significance level was a 33% percentage point differ-
ence in the change in the joint pain/tenderness score.
Other joint variables had smaller differences.
Requiring our statistical tests to have high pow-
er (90%) provides some confidence that an important
clinical difference in the population will be identified
statistically. Since the largest observed difference be-
tween AZA and DP was a 15 percentage point advan-
tage for DP in joint swelling scote, there could con-
ceivably be a 15-33 percentage point advantage of 1
drug over the comparison drug that was not statistical-
ly detectable in this study. For this modest amount of
difference, there was a possibility of Type fI error, but
for larger diffetences Type I1 error was unlikely.
In this study, RA patients who had had unsuc-
cessful gold therapy were perhaps more likely to
benefit from subsequent treatment with AZA than
with DP because of better tolerance to AZA. Howev-
er, for those who tolerated either medication, there
was no clear difference in the efficacy of the 2 drugs.
The decision as to which drug to prescribe first may
well be determined by the attitudes of the prescriber
and the patient regarding possible rare or delayed
adverse effects, such as tumorigenicity or induction of
autoimmune disease. Long-term surveillance studies
are needed to resolve these concerns.
ACKNOWLEDGMENTS
The assistance of staff at the Coordinating Center
was most valuable. Specific important contributions were
made by Stephen L. Dahl, PharmD and Eric Boyce, BS,
clinical pharmacists; Gerald L. Wise, data manager; and
Wanda L. Moore, administrative assistant. Placebo, azathio-
prine, and 0-penicillamine were provided by Burroughs
Wellcome Company.
REFERENCES
1. Paulus HE: An overview of benefitkisk of disease
modifying treatment of theumatoid arthritis. Ann
Rheum Dis (suppl) 41:26-29, 1982
2. Steinberg AD: Efficacy of immunosuppressive drugs in
rheumatic diseases. Arthritis Rheum 16:92-96, 1973
3. Mason M, Currey HLF, Barnes CG, Dunne JF, Hazle-
man BL, Strickland ID: Azathioprine in rheumatoid
arthritis. Br Med J 1:420-422, 1969
4. Levy J, Paulus B E , Barnett EV, Sokoloff M, Bangert R,
Pearson Chi: A double-blind controlled evaluation of
azathioprine treatment in rheumatoid arthritis and psori-
atic arthritis (abstract). Arthritis Rheum 15: 116-1 17,
1972
5. Urowitz MB, Gordon DA, Smythe HA, Pruzanski W,
Ogryzlo MA: Azathioprine in rheumatoid arthritis: a
727
double-blind, cross-over study. Arthritis Rheum 16:411-
418, 1973
6. Currey HLF, Harris J , Mason RM, Woodland J, Bever-
idge T, Roberts CJ, Vere DW, Dixon A St J, Davies J,
Owen-Smith B: Comparison of azathioprine, cyclophos-
phamide, and gold in treatment of rheumatoid arthritis.
Br Med J 1:763-766, 1974.
7. Dwosh IL, Stein HB, Urowitz MB, Smythe HA, Hunter
T, Ogryzlo MA: Azathioprine in early rheumatoid arthri-
tis. Arthritis Rheum 20:685-692, 1977
8. Urowitz MB, Hunter T, Bookman AAM, Gordon DA,
Smythe HA, Ogryzlo MA: Azathioprine in rheumatoid
arthritis: a double blind study comparing full dose to half
dose. J Rheumatol 1:274-281, 1974
9. Cade R, Stein G, Pickering M, Schlein E, Spooner G:
Low dose, Ibng-term treatment of rheumatoid arthritis
with azathioprine. South Med J 69:388-392, 1976
10. Khanna V, Woodbury JFL: Usefulness of azathioprine
in rheumatoid arthritis (abstract). Ann R Coll Phys Surg
Canada 6:60, 1973
11. Multicenter Trial Group: Controlled trial of d-penicilla-
mine in severe rheumatoid arthritis. Lancet 1:275-280,
1973
12. Hill HFH: Treatment of rheumatoid arthritis with peni-
cillamine. Semin Arthritis Rheum 6:361-388, 1977
13. Jaffe IA: D-penicillamine. Bull Rheum Dis 28:948-952,
1978
14. Day AT, Golding JR, Lee PN, Buttenvorth AD: Penicil-
lamine in rheumatoid disease: a long-term study. Br Med
J 1:180-183, 1974
15. Berry H, Liyanage SP, Durance RA, Barnes CG, Berger
LA, Evans S: Azathioprine and penicillamine in treat-
ment of rheumatoid arthritis: a controlled trial. Br Med J
1: 1052-1054, 1976
16. Ropes MW, Bennett GA, Cobb S, Jacox R, Jessar RA:
1958 revision of diagnostic criteria for rheumatoid arthri-
tis. Bull Rheum Dis 9: 175-176, 1958
17. Williams HJ, Ward JR, Reading JC, Egger MJ, Gran-
done JT, Samuelson CO, Furst DE, Sullivan JM, Wat-
son MA, Guttadauria M, Cathcart ES, Kaplan SB, Halla
JT, Weinstein A, Plotz PH: Low-dose D-penicillamine
therapy in rheumatoid arthritis: a cohtrolled, double-
blind clinical trial. Arthritis Rheum 26581-592, 1983
18. The Cooperating Clinics Committee of the American
Rheumatism Association: A seven-day variability study
of 499 patients with peripheral rheumatoid arthritis.
Arthritis Rheum 8:302-334, 1965
19. Pinals RS, Masi AT, Larsen RA, and the Subcommittee
for Criteria of Remission in Rheumatoid Arthritis of the
American Rheumatism Association Diagnostic and
Therapeutic Criteria Committee: Preliminary criteria for
clinical remission in rheumatoid arthritis. Arthritis
Rheum 24:1308-1315, 1981
20. Hutehinson RM, David P, Janson MIV: Thrombocytosis
in rheumatoid arthritis. Ann Rheum Dis 35: 138-142,
1976