ARTHRITIS & RHEUMATISM
Vol. 52, No. 11, November 2005, pp 3360–3370
© 2005, American College of Rheumatology
Low-Dose Prednisolone in Addition to the
Initial Disease-Modifying Antirheumatic Drug in
Patients With Early Active Rheumatoid Arthritis Reduces
Joint Destruction and Increases the Remission Rate
A Two-Year Randomized Trial
Bjo ¨rn Svensson,1Annelies Boonen,2Kristina Albertsson,3De ´sire ´e van der Heijde,2
Catharina Keller,4and Ingia ¨ld Hafstro ¨m,3for the BARFOT Study Group
Objective. To assess the efficacy of low-dose pred-
nisolone on joint damage and disease activity in pa-
tients with early rheumatoid arthritis (RA).
Methods. At the start of their initial treatment
(DMARD), patients with early (duration <1 year) active
RA were randomly assigned to receive either 7.5 mg/day
prednisolone or no prednisolone for 2 years. Radio-
graphs of the hands and feet were obtained at baseline
and after 1 and 2 years and scored according to the
Sharp score as modified by van der Heijde. Remission
was defined as a Disease Activity Score in 28 joints of
<2.6. Bone mineral density was measured by dual x-ray
absorptiometry at baseline and after 2 years.
Results. Of the 250 patients included, 242 com-
pleted the study and 225 had radiographs available both
at baseline and at 2 years. At 2 years, the median and
interquartile range (IQR) change in total Sharp score
was lower in the prednisolone group than in the no-
prednisolone group (1.8 [IQR 0.5–6.0] versus 3.5 [IQR
0.5–10]; P ? 0.019). In the prednisolone group, there
were fewer newly eroded joints per patient after 2 years
(median 0.5 [IQR 0–2] versus 1.25 [IQR 0–3.25]; P ?
0.007). In the prednisolone group, 25.9% of patients had
radiographic progression beyond the smallest detect-
able difference compared with 39.3% of patients in the
no-prednisolone group (P ? 0.033). At 2 years, 55.5% of
patients in the prednisolone group had achieved disease
remission, compared with 32.8% of patients in the
no-prednisolone group (P ? 0.0005). There were few
adverse events that led to withdrawal. Bone loss during
the 2-year study was similar in the 2 treatment groups.
Conclusion. Prednisolone at 7.5 mg/day added to
the initial DMARD retarded the progression of radio-
graphic damage after 2 years in patients with early RA,
provided a high remission rate, and was well tolerated.
Therefore, the data support the use of low-dose pred-
nisolone as an adjunct to DMARDs in early active RA.
Today it is generally accepted that early diagnosis
and early suppression of inflammation are important
requisites for a favorable outcome in rheumatoid arthri-
tis (RA). Glucocorticoids are powerful antiinflammatory
agents, and ever since Hench et al (1) showed a dramatic
Supported by grants from the Swedish Rheumatism Associa-
tion, the 80-Year Foundation of King Gustaf V, the Ugglas Founda-
tion, the Bo ¨rje Dahlins Foundation, the Gorthon Foundation in
Helsingborg, and Stiftelsen fo ¨r Ro ¨relsehindrade i Skåne.
1Bjo ¨rn Svensson, MD, PhD: University of Lund, Lund, Swe-
den;2Annelies Boonen, MD, PhD, De ´sire ´e van der Heijde, MD, PhD:
University Hospital Maastricht and Caphri Research Institute, Maas-
tricht, The Netherlands;3Kristina Albertsson, MD, Ingia ¨ld Hafstro ¨m,
MD, PhD: Karolinska University Hospital, Stockholm, Sweden;
4Catharina Keller, MD: Helsingborg’s Lasarett, Helsingborg, Sweden.
Members of the BARFOT Study Group include Monica
Ahlme ´n, MD, PhD, Johan Bratt, MD, PhD, Kristina Forslind, MD,
PhD, Ingia ¨ld Hafstro ¨m, MD, PhD, Catharina Keller, MD, Ido Leden,
MD, Bengt Lindell, MD, Ingemar Petersson, MD, PhD, Bjo ¨rn Svens-
son, MD, PhD, Annika Teleman, MD, and Jan Theander, MD.
Address correspondence to Bjo ¨rn Svensson, MD, PhD, Blis-
torpsva ¨gen 546, 290 38 Villands Vånga, Sweden. E-mail:
firstname.lastname@example.org. Address reprint requests to Ingia ¨ld Haf-
stro ¨m, MD, PhD, Department of Rheumatology, Karolinska Univer-
sity Hospital Huddinge, 141 86 Stockholm, Sweden. E-mail:
Submitted for publication February 17, 2005; accepted in
revised form June 23, 2005.
effect of cortisone treatment in a patient with serious
established RA, glucocorticoids have been used in the
treatment of patients with active RA. However, high
daily doses of glucocorticoids over the long term have
considerable toxicity, hampering their use. Therefore,
alternative treatment approaches have been explored,
such as short-term high-dose therapy and low-dose
treatment as monotherapy or in combination with con-
ventional disease-modifying antirheumatic drugs
(DMARDs). The pivotal question is whether these
regimens are able to reduce the progression of joint
destruction and provide sustained reduction of disease
activity with an acceptable level of toxicity (2–5).
A beneficial effect of low-dose prednisolone on
radiographic damage was not found in the first studies
evaluating radiographic outcome (6–8). However, those
studies were performed in patients with late RA, and an
effect in early disease remained possible. Therefore, the
1995 study by Kirwan and the Arthritis and Rheumatism
Council Low-Dose Glucocorticoid Study Group was met
with great interest. In this placebo-controlled study in
patients with early RA, Kirwan et al (9) found that 7.5
mg/day prednisolone, given for 2 years in addition to
optional other treatments, reduced joint destruction.
However, when prednisolone was stopped, joint destruc-
tion returned to the same level as that in the control
group (10). In a more recent randomized trial (11), it
was shown that prednisolone at a dosage of 10 mg/day
retarded the progression of joint damage over a period
of 2 years. In that study, prednisolone was compared
with placebo and was not added to DMARDs.
The short-term clinical effectiveness of low-dose
glucocorticoid treatment in patients with early RA is
widely established and supported by 2 meta-analyses
(12,13). When given for periods of ?6 months, glucocor-
ticoids are effective in reducing signs and symptoms of
RA when used as comedication with DMARDs (14).
However, the clinical response has not consistently been
shown to persist beyond that time (9,11). Since it is
increasingly accepted that early and sustained control of
the inflammatory process is crucial to the prevention of
radiographic damage and long-term disability (15–17),
we conducted a randomized study in clinical practice on
the medium-term (2 years) effects on radiographic dam-
age (primary outcome) and disease activity (secondary
outcome) of low-dose prednisolone added to the initial
DMARD in patients with early RA.
PATIENTS AND METHODS
Study design. The study was a 2-year, multicenter,
open randomized trial comparing the addition of 7.5 mg/day
prednisolone to DMARD therapy with DMARD therapy
alone. The patients had not previously been treated with
DMARDs. The primary end point was the difference in
changes in radiographic damage scores after 2 years; the
secondary end points were remission rates and differences in
disease activity and function. The patients were recruited from
the 6 centers involved in the BARFOT (Better Anti-rheumatic
Farmacotherapy) Study Group. The BARFOT study is a
long-term multicenter observational study of an inception
cohort of patients with early RA in southern Sweden (18).
No formal sample size calculation was performed.
However, according to the calculation described in the study by
Kirwan et al (9), 160 patients would be sufficient to detect a
50% reduction in the progression of joint erosion over 2 years.
Since the study by Kirwan et al and the present study have a
similar design (although the present study adds radiographs of
the feet and a more sensitive method for scoring radiographic
progression), 250 patients seemed to be sufficient to detect a
statistically significant difference in radiographic progression.
The randomization was done as block randomization
for each center according to a central randomization program
with stratification for sex. The choice of DMARD was left to
the treating physicians, who followed the recommended treat-
ment strategy in Sweden at the time of the study. Concomitant
treatment with nonsteroidal antiinflammatory drugs
(NSAIDs) was permitted, and intraarticular steroid injections
were allowed except during the 2 weeks preceding a clinical
evaluation. All patients were given 1,000 mg/day calcium
carbonate or calcium gluconate. The study was approved by
the Ethics Committee.
Patients. Patients were eligible for study if they were
diagnosed as having RA according to the 1987 revised criteria
of the American College of Rheumatology (formerly, the
American Rheumatism Association) (19), were ages 18–80
years, had a disease duration of ?1 year, had active disease
(defined as a Disease Activity Score in 28 joints [DAS28] 
of ?3.0), and were started by the treating rheumatologist on
the first DMARD. Exclusion criteria were earlier treatment
with glucocorticoids for RA or other diseases, previous treat-
ment with DMARDs, or contraindication for glucocorticoid
therapy. Furthermore, patients with previous fragility fractures
were excluded, as well as patients ages ?65 years with a T score
lower than ?2.5 on bone mineral densitometry and patients
ages ?65 years with a Z score of less than ?1.
Overall, 250 consecutive outpatients were included in
the study between September 1995 and December 1999. They
were recruited from 840 RA patients entering the BARFOT
program during the same period. Of these 840 patients, 80 did
not meet the inclusion criteria (disease activity and age), while
334 met one of the exclusion criteria (most frequently, earlier
glucocorticoid treatment [131 patients] or osteoporosis [96
patients]). Furthermore, 43 patients were considered to need
glucocorticoid therapy because of highly active disease and
were thus not included, 30 patients were missed, and 94
patients did not consent to participate. Overall, 259 patients
were randomized. Of these, 9 patients were randomized (5 to
the prednisolone group and 4 to the no-prednisolone group)
but were withdrawn since they did not have RA (3 patients) or
were not given a single dose of the stipulated drug(s) within 3
months (6 patients). All patients gave their informed consent.
LOW-DOSE PREDNISOLONE IN EARLY RHEUMATOID ARTHRITIS3361
Radiographic evaluation. Anteroposterior radiographs
of the hands and feet were obtained at each center at study
entry and after 1 and 2 years. Radiographic damage was scored
according to the Sharp method as modified by van der Heijde
(SHS) (21), which includes the hands and feet and allows for
the separate presentation of a total score (range 0–448), an
erosion score (range 0–280), and a joint space narrowing (JSN)
score (range 0–168). Two trained readers (AB and KA) who
were blinded to the patient’s treatment assignment read the
radiographs independently in chronological order. Readers
were allowed to score improvement in scores. Chronological
scoring was preferred, since this method has been shown to
have a better signal-to-noise ratio compared with paired
The interobserver intraclass correlation coefficients
(ICCs) between the 2 readers were 0.91 and 0.93 for the
erosion score, 0.90 and 0.93 for the JSN score, and 0.93 and
0.94 for the total score at baseline and after 2 years, respec-
tively. The interobserver ICC for change in the total score from
baseline to 2 years was 0.94.
Clinical evaluations. Clinical assessment took place at
the start of the study (baseline) and at 3, 6, 12, 18, and 24
months (end point). Disease activity was measured by the
DAS28. A patient’s disease was considered to be in remission
if the DAS28 was ?2.6 (23). Other laboratory assessments
included the C-reactive protein (CRP) level and rheumatoid
factor (RF) positivity according to the standard method used
at each center.
Functional disability was assessed using the Swedish
version of the Stanford Health Assessment Questionnaire
(HAQ) (24). The HAQ score ranges from 0 to 3, and a higher
score indicates a higher degree of disability.
Functional impairment was measured by the Signals of
Functional Impairment (SOFI) index, a performance test
consisting of 4 items for hand function, 3 for upper limb
function, and 4 for lower limb function. Each item is scored
according to 3 alternatives (0, 1, and 2), allowing a maximum
SOFI index of 44, and a higher score indicates a higher degree
of functional impairment (25). Adverse events were registered
at the followup visits and were either spontaneously reported
or revealed by laboratory values.
Bone mineral density (BMD) measurement. BMD was
measured by dual x-ray absorptiometry (DXA) with a densi-
tometer (Lunar, Madison, WI). Assessment of BMD was made
at the lumbar spine (L2–L4; anteroposterior view) and at the
left hip (femoral neck). BMD was expressed in gm/cm2and as
a Z score (the number of SDs from the mean in healthy age-
and sex-matched people, values for which were obtained from
Lunar’s combined European/US reference population ).
Statistical analysis. Efficacy analyses were performed
on the 250 patients treated or not treated with prednisolone at
baseline according to the intent-to-treat analysis. For the
patients who dropped out of the study, the clinical data from
the last observation were carried forward. Since the number of
missing radiographs was rather small (7.6% at baseline and 6%
at study end point), and since only 8 patients dropped out, a
completer analysis was performed for the radiographic data.
For radiographic outcome, the medians and interquar-
tile ranges (IQRs) are presented for baseline and change
scores after 1 and 2 years. The Mann-Whitney U test was used
as the primary test for the total, erosion, and JSN scores
analyzed separately. The corresponding means and SDs are
also presented in a table. In addition, cumulative probability
plots of the change scores (from end point to baseline) were
Figure 1. Distribution of the study patients, by treatment group.
treated with prednisolone*
Baseline characteristics of 250 patients treated or not
(n ? 119)
(n ? 131)
Age at inclusion, mean ? SD years
Disease duration, mean ? SD months
RF positive, %
Current or previous smoker, %
DAS28, mean ? SD (range 0–10)
HAQ score, mean ? SD (range 0–3)
SOFI index, mean ? SD (range 0–44)
CRP level, median (IQR) mg/liter
51 ? 14
6.5 ? 3.5
5.28 ? 1.11
1.01 ? 0.59
8 ? 5
59 ? 13
5.8 ? 2.9
5.42 ? 1.04
0.98 ? 0.65
9 ? 7
* RF ? rheumatoid factor; DAS28 ? Disease Activity Score in 28
joints; HAQ ? Health Assessment Questionnaire; SOFI ? Signals of
Functional Impairment; CRP ? C-reactive protein; IQR ? interquar-
3362SVENSSON ET AL
constructed. These are visual presentations of all individual
data made by plotting the observed cumulative proportion
against the corresponding change score (27). Next, we deter-
mined the number of newly eroded joints (number of joints per
patient that changed from an erosion score of 0 at baseline to
any higher score at 1 and 2 years) and the proportion of
patients with progression greater than the smallest detectable
difference (SDD) (28).
For clinical variables, group changes were presented as
the mean ? SD (or, in the case of the CRP level, as the median
[IQR] as well) at baseline and after 3, 6, 12, 18, and 24 months.
We also calculated the percentages of patients with disease in
remission (DAS28 ?2.6) in the 2 groups after 1 and 2 years.
To test for differences in change between groups, the
Mann-Whitney U test or the t-test for independent samples
was used for continuous variables, and the chi-square test was
used for proportions. Analysis of covariance was used with age
as a covariate to correct for the possible effect on outcome
variables of a statistically significant difference in mean age
between the treatment groups. All significance tests were
2-tailed and were conducted at the 5% significance level. The
statistical analyses were performed using SPSS 13.0 statistical
software (SPSS, Chicago, IL).
Demographics and patient characteristics. Table
1 presents the baseline demographics and clinical char-
acteristics of the patients. Of 250 patients who were
treated, 119 had been randomized to treatment with 7.5
mg/day prednisolone and 131 to no prednisolone ther-
apy. Sixty-four percent of patients were women, 66%
were RF-positive, and 62% were current or previous
smokers. The 2 groups of patients were very comparable
at baseline, except for age. One hundred sixteen patients
in the prednisolone group and 126 in the no-
prednisolone group completed the 2-year followup pe-
riod (Figure 1). One patient in the prednisolone group
died, 1 moved from the district, and 1 was lost to
followup; 2 patients in the no-prednisolone group moved
and 3 were lost to followup. Of the 116 completers in the
prednisolone group, 8 patients stopped prednisolone
treatment because of diabetes (n ? 1), weight gain (n ?
2), proteinuria (n ? 1), or unwillingness to continue
prednisolone therapy (n ? 4), and 16 patients received a
prednisolone dosage lower (n ? 14) or higher (n ? 2)
than the prescribed 7.5 mg/day at various times during
the study. Of the 126 completers in the no-prednisolone
group, 7 patients started prednisolone because of ex-
traarticular manifestations, allergic symptoms, or in-
creased disease activity (Figure 1).
Concomitant treatment. In accordance with the
study protocol, DMARDs were given to all patients. In
the prednisolone group, 50% started with methotrexate
(MTX) and 35% with sulfasalazine (SSZ), and the
corresponding values for the no-prednisolone group
were 53% and 37%. The mean ? SD MTX dosage was
10.1 ? 2.64 mg/week in the prednisolone group and 11 ?
3.08 mg/week in the no-prednisolone group (P ? 0.11).
The SSZ dosage was, in general, 2 gm/day. At the 2-year
visit, several patients in each group had switched from
treatment with SSZ, and at that time, 15.4% in the
prednisolone group and 16.9% in the no-prednisolone
group were no longer receiving DMARDs. There were
no statistically significant differences between the
groups for types of DMARDs, either at baseline (P ?
0.43) or after 2 years (P ? 0.76). The mean ? SD
duration of DMARD treatment was 21.5 ? 5 months in
the prednisolone group and 20.6 ? 5.1 months in the
no-prednisolone group (P ? 0.15).
modified Sharp/van der Heijde scoring method, by treatment group*
Comparison of baseline radiographic damage scores and changes in the total, erosion, and JSN scores at 1 and 2 years, as assessed by the
Variable (no. of patients)
Median (IQR)Mean ? SD
Total score at baseline (231)†
Erosion score at baseline (231)†
JSN score at baseline (231)†
Change in total score at 1 year (214)‡
Change in total score at 2 years (225)§
Change in erosion score at 1 year (214)‡
Change in erosion score at 2 years (225)§
Change in JSN score at 1 year (214)‡
Change in JSN score at 2 years (225)§
4.1 ? 9.2
1.9 ? 5.0
2.2 ? 4.6
2.4 ? 4.6
5.2 ? 9.0
0.8 ? 1.6
1.9 ? 3.6
1.5 ? 3.8
3.3 ? 6.4
4.8 ? 9.6
1.9 ? 4.0
2.9 ? 6.4
5.3 ? 9.3
9.1 ? 14.3
2.4 ? 4.0
4.0 ? 6.8
2.9 ? 6.0
5.0 ? 8.6
* P values represent the significance of the difference between medians, as determined by Mann-Whitney U test. JSN ? joint space narrowing; IQR
? interquartile range; JSN score ? joint space narrowing score.
† There were 109 and 122 patients, respectively, in the prednisolone and no-prednisolone groups.
‡ There were 99 and 115 patients, respectively, in the prednisolone and no-prednisolone groups.
§ There were 108 and 117 patients, respectively, in the prednisolone and no-prednisolone groups.
LOW-DOSE PREDNISOLONE IN EARLY RHEUMATOID ARTHRITIS3363
NSAIDs were taken by 85% and 87% of patients
in the prednisolone and no-prednisolone groups, respec-
tively, at baseline (P ? 0.62). At 2 years, however, only
44% of patients in the prednisolone group were taking
NSAIDs, compared with 65% in the no-prednisolone
group (P ? 0.001).
Patients in the prednisolone group were injected
intraarticularly with a mean ? SD of 41 ? 60 mg of
triamcinolone hexacetonide equivalents during the first
year compared with 90 ? 87 mg in the no-prednisolone
group (P ? 0.0005). During the second year, these doses
were 23 ? 40 mg and 38 ? 55 mg, respectively (P ?
Radiographic progression. Radiographs were
available for 231 patients at baseline, 224 patients at the
1-year followup visit, and 235 patients at the 2-year
followup visit. Two hundred fourteen patients had ra-
diographs both at baseline and after 1 year, and 225
patients had radiographs both at baseline and after 2
years. Baseline characteristics of the 25 patients without
available radiographs at baseline and after 2 years (see
Table 1) did not differ from those of the 225 patients
with available radiographs (data not shown). The base-
line radiographic scores were similar in the 2 groups of
patients (Table 2). As can be seen in Table 2, the change
in total score was less pronounced after 1 and 2 years in
the prednisolone group than in the no-prednisolone
group, and the difference between the groups was
already statistically significant at 1 year. The median
(IQR) change in the total score was 1.0 (0–3.0) after 1
year and 1.8 (0.5–6.0) after 2 years for the prednisolone
group and 2.0 (0–5.0) after 1 year and 3.5 (0.5–10) after
2 years for the no-prednisolone group. In addition, the
erosion score changed significantly less in the pred-
nisolone group at both time points, and there was a
similar, nonsignificant trend for the JSN score (Table 2).
The differences in change scores (from end point to
baseline) between the treatment groups are further
illustrated as probability plots in Figure 2.
The rate of joint damage before entering the
study was similar in the 2 treatment groups. Thus, the
median total score at baseline divided by the disease
duration expressed in years was 3 (IQR 0–8) for the
prednisolone group and 3.3 (IQR 0–10) for the no-
prednisolone group (P ? 0.49). The mean ? SD change
in total score in the first year was 2.4 ? 4.6 in the
prednisolone group and 5.3 ? 9.3 in the no-prednisolone
group; in the second year, these values were 2.9 ? 5.6
and 4.0 ? 6.7, respectively.
In the prednisolone group, 53.2% of the patients
had erosions at baseline, and 77.2% had erosions after 2
Figure 2. Probability plots demonstrating differences between the 2
treatment groups in change in the total score (A), erosion score (B),
and joint space narrowing (JSN) score (C) over 2 years.
3364 SVENSSON ET AL
years. In the no-prednisolone group, these values were
59.0% and 80.2%, respectively.
The development of erosions in a previously
unaffected joint differed significantly between the treat-
ment groups. Thus, in the prednisolone group, there was
a median of 0 (IQR 0–1) and 0.5 (IQR 0–2) newly
eroded joints after 1 and 2 years, respectively. The
corresponding values were 0.5 (IQR 0–2) and 1.25 (IQR
0–3.25) in the no-prednisolone group (between-group P
values were 0.004 and 0.007 at 1 and 2 years, respec-
The SDD was calculated to be 5.8. After the 2
study years, 25.9% of patients receiving prednisolone
had radiographic progression greater than the SDD
compared with 39.3% of patients not receiving pred-
nisolone (P ? 0.033).
In the prednisolone group, but not in the no-
prednisolone group, patients with disease in remission
had significantly less worsening of the total score and the
erosion score than patients with disease not in remission
at 2 years: median change in total score 1 (IQR 0–3.5)
versus 4 (IQR 1–10.5) (P ? 0.002) and median change in
erosion score 0.5 (IQR 0–1) versus 1.5 (IQR 0.5–4) (P ?
0.0005). In the no-prednisolone group, the correspond-
ing values were as follows: median change in total score
2.5 (IQR 0.5–8) versus 3.5 (IQR 0.5–11) (P ? 0.45) and
median change in erosion score 1 (IQR 0–2.5) versus 1.5
(IQR 0.5–6) (P ? 0.09). When we considered the entire
group independently of treatment assignment, patients
with disease remission had a smaller increase in the total
score than did patients without disease remission (me-
dian 1.5 [IQR 0.5–5] versus 3.5 [IQR 0.5–10.5]; P ?
0.002) and a smaller increase in the erosion score
(median 0.5 [IQR 0–2] versus 1.5 [IQR 0.5–5.5]; P ?
Clinical efficacy. Disease activity, measured as
the DAS28, decreased in both groups but significantly
more in the prednisolone group. Differences between
the 2 groups were already seen at 3 months and were
present at all time points thereafter (Figure 3A). Thus,
Figure 3. Differences between treatment groups in improvement in the Disease Activity Score in 28 joints (DAS28) (A), C-reactive protein (CRP)
level (B), Health Assessment Questionnaire (HAQ) score (C), and Signals of Functional Impairment (SOFI) index (D) over 2 years. Values are the
mean. CI ? confidence interval; ANCOVA ? analysis of covariance.
LOW-DOSE PREDNISOLONE IN EARLY RHEUMATOID ARTHRITIS3365
in the prednisolone group, the mean ? SD DAS28
decreased from 5.3 ? 1.1 at baseline to 2.7 ? 1.5 after 1
year and 2.7 ? 1.3 after 2 years. The corresponding
values in the no-prednisolone group were 5.4 ? 1.0,
3.3 ? 1.5, and 3.2 ? 1.4.
After 1 year, 51.3% of patients in the pred-
nisolone group had achieved disease remission (DAS28
?2.6) compared with 39.2% of patients in the no-
prednisolone group (P ? 0.06). After 2 years, this
difference had increased to 55.5% in the prednisolone
group compared with 32.8% in the no-prednisolone
group (P ? 0.0005).
The CRP level, not included in the DAS28, fell
rapidly in both treatment groups (Figure 3B), from a
median of 22 mg/liter (IQR 10–51) at baseline to 9
mg/liter (IQR 4–10) after 1 year and 9 mg/liter (IQR
4–11) after 2 years in the prednisolone group and from
22 mg/liter (IQR 10–54) at baseline to 10 mg/liter (IQR
4–14) after 1 year and 9 mg/liter (IQR 4–12) after 2
years in the no-prednisolone group. These differences
between the groups were not significant.
The HAQ score showed a significantly greater
reduction in the prednisolone group than in the no-
prednisolone group at all time points measured during
followup (Figure 3C). The mean ? SD HAQ score in
the prednisolone group decreased from 1.0 ? 0.6 at
baseline to 0.4 ? 0.5 at 1 year and 0.5 ? 0.5 at 2 years.
The corresponding values in the no-prednisolone group
were 1.0 ? 0.7, 0.6 ? 0.6, and 0.7 ? 0.6.
The SOFI index decreased significantly more in
the prednisolone group than in the no-prednisolone
group (Figure 3D). The mean ? SD SOFI index in the
prednisolone group decreased from 8 ? 5 at baseline to
4 ? 5 after 1 year and 4 ? 5 after 2 years. The
corresponding values in the no-prednisolone group were
9 ? 7, 6 ? 6, and 7 ? 6.
BMD findings. DXA data were available for 189
patients both at baseline and after 2 years. At baseline
and after 2 years, BMD at the lumbar spine did not
differ significantly between the prednisolone and no-
prednisolone groups (mean ? SD 1.14 ? 0.16 gm/cm2
and 1.11 ? 0.16 gm/cm2versus 1.17 ? 0.21 gm/cm2and
1.16 ? 0.21 gm/cm2, respectively). Since patients in the
prednisolone group were somewhat younger, Z scores
were also calculated. A difference was found between
the prednisolone and no-prednisolone groups both at
baseline (0.19 ? 1.19 versus 0.74 ? 1.67, respectively; P
? 0.010) and after 2 years (?0.1 ? 1.19 versus 0.58 ?
1.66, respectively; P ? 0.001). However, the decrease in
Z score during the 2 study years did not differ signifi-
cantly between the prednisolone and no-prednisolone
groups (?0.29 ? 0.51 versus ?0.16 ? 0.68, respectively;
P ? 0.13).
BMD at the femoral neck did not differ signifi-
cantly at baseline and after 2 years between the pred-
nisolone and no-prednisolone groups (0.91 ? 0.15 gm/
cm2and 0.90 ? 0.16 gm/cm2versus 0.90 ? 0.15 gm/cm2
and 0.87 ? 0.15 gm/cm2, respectively), and there was no
significant difference in Z scores (0.01 ? 1.1 and
?0.06 ? 1.13 versus 0.22 ? 1.02 and 0.06 ? 1.04,
respectively). Further, there was no difference between
the prednisolone and no-prednisolone groups in change
in Z score at the femoral neck from baseline to 2 years
(?0.07 ? 0.49 and ?0.16 ? 0.45, respectively; P ? 0.16).
Adverse events. Prednisolone was generally well
tolerated, since the frequency of adverse reactions was
small in both groups. Drug treatment was withdrawn
temporarily or permanently in 26 patients in the pred-
nisolone group and in 24 patients in the no-prednisolone
withdrawal in the 2 treatment groups*
Adverse reactions leading to temporary or permanent
Elevated liver enzymes
Various, including infections
* Disease-modifying antirheumatic drugs were the predominant cause
of adverse reactions.
† Caused by prednisolone in 1 patient.
3366 SVENSSON ET AL
group (Table 3). Most withdrawals were caused by the
DMARD given, while prednisolone was judged to be the
cause in 5 patients in the prednisolone group; these 5
patients withdrew due to diabetes, proteinuria, striae
and ecchymoses, weight gain and cushingoid appear-
ance, and weight gain.
The present study shows that low-dose pred-
nisolone added to the initial DMARD treatment in
patients with early RA in a clinical setting reduces the
progression of joint damage, controls inflammation,
improves physical functioning, and is well tolerated. The
progression of erosions and, to a lesser extent, JSN,
which represent different aspects of structural damage,
were retarded by prednisolone. Prednisolone not only
retarded progression in already affected joints, but also
prevented the development of erosions in previously
unaffected joints. Moreover, the number of patients with
progression of joint damage was lower in the pred-
nisolone group than in the no-prednisolone group when
radiographic progression (the total score) greater than
the SDD was calculated.
Only a few studies of patients with early RA have
evaluated the effects of glucocorticoid treatment on
radiographic damage. These studies have used different
designs, evaluation methods, and glucocorticoid doses.
The first study to suggest the ability of pred-
nisolone to slow the progression of radiographic damage
in early RA was a 2-year study by the Medical Research
Council (29). Investigators in that study reported that at
a mean dosage of 15 mg/day, prednisolone given for 2
years to patients with early RA resulted in less radio-
graphic damage than did aspirin alone.
Results of a placebo-controlled study of 10 mg/
day prednisolone as monotherapy in early RA were
recently reported by Van Everdingen et al (11). SSZ was
allowed as a rescue drug after the first 6 months. Similar
to the results in the present study, the change from
baseline in the SHS after 1 and 2 years was less
pronounced in the group of patients receiving pred-
nisolone. However, in the study by Van Everdingen et al,
the radiographic progression, calculated as the mean
increase in the total score per year, was higher than that
in the present study (8 versus 2.6 per year in the
prednisolone groups and 14.5 versus 4.4 per year in the
no-prednisolone groups). The lower progression rate in
the present study may likely be due to the fact that all
patients started concomitant DMARD treatment with
the goal of achieving disease remission. The addition of
prednisolone further slowed the radiographic progres-
sion in a way that may justify the use of glucocorticoids
early in the disease.
The design of the COBRA (Combinatietherapie
Bij Reumatoı ¨de Artritis) study, comparing a combina-
tion of prednisolone, MTX, and SSZ with SSZ alone
(30), differs from the present study in that a high initial
prednisolone dosage was used and 2 DMARDs were
started concomitantly. However, concordant with the
results of the present study, prednisolone had a superior
effect on radiographic progression. Because of the study
design, it cannot be concluded that the beneficial effects
were due to prednisolone, since MTX was also included
in the combination treatment arm. However, support for
a role of prednisolone in the COBRA study is provided
by the results of another study of patients with early RA,
in which the combination of MTX and SSZ, without
prednisolone, was not superior to SSZ alone with regard
to disease activity and structural changes over the 5
years of followup (31).
The study by Kirwan et al in 1995 (9) had a design
similar to that of the present study. Prednisolone at 7.5
mg/day for 2 years was compared with no prednisolone
in patients with early RA (disease duration ?2 years). In
that study, only 71% of the patients received concomi-
tant treatment with DMARDs, and only ?80% of the
randomized patients were available for evaluation. Fur-
thermore, the patients in the placebo group had more
severe disease at baseline. Nevertheless, the results
showing that prednisolone retarded joint destruction in
the hands, as measured by the Larsen score, were
consistent with the findings in the present study, in which
all patients were started on DMARD therapy, the 2
treatment groups were equal in terms of the baseline
radiographic score, and 90% of the randomized patients
were eligible for radiographic evaluation including both
the hands and feet. Thus, it is of interest to note that all
these different studies support the notion that glucocor-
ticoids may retard radiographic damage in early RA.
Two further studies on the effect of low-dose
prednisolone on radiographic damage have been per-
formed in patients with early RA. Unfortunately, both of
these studies are difficult to evaluate because of meth-
odologic difficulties. In the study by Rau et al (32), 5
mg/day prednisolone was compared with no pred-
nisolone in DMARD-treated patients with early RA.
The authors concluded that prednisolone in this setting
decreases radiographic progression. However, only 76 of
the 196 included patients completed the 2 years of the
study, making the authors’ conclusion somewhat difficult
LOW-DOSE PREDNISOLONE IN EARLY RHEUMATOID ARTHRITIS3367
A recent study by Capell et al (33) compared the
addition of 7 mg/day prednisolone to SSZ with the
addition of placebo to SSZ, but the investigators failed
to find a superior effect of the combination on radio-
graphic progression. However, as pointed out by Kirwan
and Boers (34), the results presented in the study by
Capell et al on radiographic damage cannot be ade-
As reviewed above, a limited number of studies
support the notion that glucocorticoids may retard ra-
diographic damage in early RA. More studies have
shown short-term beneficial clinical effects of low-dose
prednisolone treatment of RA (14). To our knowledge,
the present study is the first to demonstrate that low-
dose prednisolone taken in addition to DMARDs im-
proves inflammatory symptoms and signs for as long as
2 years. This finding contrasts with the results of some
other studies, which fail to demonstrate effects on
clinical symptoms for more than a few months (9,11).
This discrepancy is not likely to be due to a difference in
DMARD treatment between the groups in the present
study, since the various DMARDs used were similar in
the 2 groups and the mean number of months that each
group took DMARDs was also comparable. Further-
more, the mean dosage of MTX was similar in the 2
treatment groups. Finally, NSAID use was significantly
more frequent in the no-prednisolone group.
Other possible reasons for the divergent results
could be differences in study design, populations, and
DMARD treatment strategies. The previous studies
used a smaller number of patients (9,11), disease dura-
tion at baseline was longer, disease activity was defined
differently, and baseline disability score was higher (9).
In one study (9), only 71% of the patients were given
DMARDs at inclusion into the study and 40% of the
patients had had previous DMARD treatment. In an-
other study (11), only 53% of the patients received
DMARDs in combination with prednisolone, and then
only as a rescue drug and not before 6 months. In the
present study, all patients were started on DMARDs,
which resulted in a higher disease remission rate in the
prednisolone group, an outcome not described in re-
ports of earlier studies (9,11).
The relevance of maintained suppression of in-
flammation for retarding joint damage has been docu-
mented in several reports (35,36). The notion that early
suppression of inflammation may further influence the
rate of radiographic progression for several years is
supported by the present data showing that the DAS28
was already strongly reduced after 3 months of treat-
ment with prednisolone, which, together with a sus-
tained effect on disease activity, resulted both in a
reduced erosion score and in a strong trend toward
reduced progression of JSN. Such an impact of early
suppression of disease activity has been further sug-
gested by the results of a followup of the COBRA study,
in which the patients in the combination group contin-
ued to have a slower rate of progression after 4 years
(37). Results of the present study and of the COBRA
followup study imply that early suppression of inflam-
mation decreases the progression of both erosions and
JSN, thus emphasizing further the necessity of reducing
inflammation at an early stage of the disease (16).
Investigators in some studies (38,39) have dis-
cussed the possibility of a different pathogenesis of
synovial inflammation and joint destruction, suggesting
that the ability of drugs to retard radiographic damage
can only partly be explained by suppression of disease
activity. Of the patients with remission of disease in the
present study, only those in the prednisolone group
showed significantly reduced radiographic progression
(in spite of an almost identical DAS28 in the 2 groups).
This raises the possibility that prednisolone may reduce
joint damage by a mechanism different from that by
which it reduces inflammation, e.g., by a direct effect on
bone turnover (40).
In addition to the beneficial effects on radio-
graphic damage and disease activity, prednisolone treat-
ment resulted in better physical functioning, as proven
by larger improvement in scores on the HAQ and SOFI
index. It is recognized that, especially early in the
disease, the HAQ score is primarily influenced by dis-
ease activity and to a lesser extent by radiographic
damage (41). On the other hand, the SOFI index (25)
reflects structural damage already present in early RA,
and it may therefore be a valuable outcome variable in
the evaluation of treatment in early RA.
A concern with prednisolone treatment has been
its inhibitory effect on bone formation, at least in high
doses. However, in low doses, it has been suggested that
this inhibitory effect may be compensated by the ability
of prednisolone to suppress disease activity and the
consequent inflammation-mediated bone resorption.
This idea was strengthened by the present finding that
bone loss was not significantly more pronounced in the
group of patients receiving prednisolone. It is of note
that the measurement of BMD does not reflect all of the
aspects of bone that contribute to its resistance against
The patients participating in the present study
were recruited to the BARFOT program, whose patient
population is believed to be representative of patients
3368 SVENSSON ET AL
with early RA, since most RA patients with disease of
recent onset are referred to the rheumatology depart-
ments in these regions. However, the inclusion and
exclusion criteria for the present study limited participa-
tion to 30% of eligible patients. The issue of external
validity of study populations is important, but it is only
infrequently addressed in clinical trials. It is important to
emphasize that in the present study, the mean ? SD
DAS28 of 5.4 ? 1.1 was similar to that in the total group
of 840 patients included in the BARFOT program
during the inclusion period (5.3 ? 1.3). Therefore, it
cannot be argued that external validity was limited by a
selection of patients with very high disease activity, as
was recently suggested by Sokka and Pincus (42) when
discussing the external validity in trials of biologics. In
the present study, ineligibility was most often caused by
some exclusion criterion, of which previous low bone
mass (96 patients) was one of the most common and
required by the Ethics Committee. This high frequency
of osteoporosis in early, untreated RA was an unex-
pected finding that we have studied further and reported
elsewhere (43). Also of interest, 94 patients did not
consent to participate, which may reflect a common
negative attitude in the society toward prednisolone
Although this was a randomized controlled study,
it was performed in, and intended to reproduce, the
clinical practice setting. Therefore, blinding was not
performed and a placebo was not used. A weakness (or
strength, since it reflects clinical practice?) of this study
might be that a predefined DMARD algorithm was not
used. Instead, the choice of DMARD was left to the
treating physicians, who followed the recommended
treatment strategy in Sweden at the time of the study.
For that reason, DMARDs became similarly prescribed
in the 2 study groups. Further, the lack of blinding and of
placebo control did not seem to have a major effect on
the results, since the DMARD prescriptions were simi-
lar in the 2 treatment groups and the radiograph readers
were blinded to treatment strategies and clinical data.
Instead, the no-prednisolone group received more
NSAIDs and more intraarticular steroids, which might
underscore the differences between the 2 groups. In-
deed, such increased treatment in the control group may
mask a superior symptomatic effect of prednisolone, as
has been suggested by Van Everdingen et al (44).
In conclusion, 7.5 mg/day prednisolone in addi-
tion to the initial DMARD retards the progression of
radiographic damage after 1 and 2 years in patients with
early RA followed up in a clinical setting. Since this
regimen also showed favorable effects on the disease
remission rate and physical function (HAQ score and
SOFI index) and was well tolerated, the present data lend
strong support for the use of low-dose prednisolone as an
adjunct to the initial DMARD in patients with early RA.
We acknowledge research nurse Siv Nore ´n for skillful
1. Hench PS, Kendall EC, Slocumb CH, Polley HF. The effect of a
hormone of the adrenal cortex (17-hydroxy-11-dehydro-corticoste-
rone: compound E) and of pituitary adrenocorticotropic hormone
on rheumatoid arthritis. Proc Staff Meet Mayo Clin 1949;24:
2. Boers M. The case for corticosteroids in the treatment of early
rheumatoid arthritis [editorial]. Rheumatology (Oxford) 1999;38:
3. Morrison E, Capell HA. Corticosteroids in rheumatoid arthritis:
the case against [editorial]. Rheumatology (Oxford) 1999;38:
4. Moreland LW, O’Dell JR. Glucocorticoids and rheumatoid arthri-
tis: back to the future? [review]. Arthritis Rheum 2002;46:2553–63.
5. Bijlsma JW, Boers M, Saag KG, Furst DE. Glucocorticoids in the
treatment of early and late RA. Ann Rheum Dis 2003;62:1033–7.
6. Harris ED Jr, Emkey RD, Nichols JE, Newberg A. Low dose
prednisone therapy in rheumatoid arthritis: a double blind study.
J Rheumatol 1983;10:713–21.
7. Hansen M, Podenphant J, Florescu A, Stoltenberg M, Borch A,
Kluger E, et al. A randomised trial of differentiated prednisolone
treatment in active rheumatoid arthritis: clinical benefits and
skeletal side effects. Ann Rheum Dis 1999;58:713–8.
8. Paulus HE, DiPrimeo D, Sanda M, Lynch JM, Schwartz BA, Sharp
JT, et al, for the Long-Term Etodolac Study Investigators. Pro-
gression of radiographic joint erosion during low dose cortico-
steroid treatment of rheumatoid arthritis. J Rheumatol 2000;27:
9. Kirwan JR and the Arthritis and Rheumatism Council Low-Dose
Glucocorticoid Study Group. The effect of glucocorticoids on joint
destruction in rheumatoid arthritis. N Engl J Med 1995;333:142–6.
10. Hickling P, Jacoby RK, Kirwan JR, and the Arthritis and Rheu-
matism Council Low Dose Glucocorticoid Study Group. Joint
destruction after glucocorticoids are withdrawn in early rheuma-
toid arthritis. Br J Rheumatol 1998;37:930–6.
11. Van Everdingen AA, Jacobs JW, Siewertsz van Reesema DR,
Bijlsma JW. Low-dose prednisone therapy for patients with early
active rheumatoid arthritis: clinical efficacy, disease-modifying
properties, and side effects: a randomized, double-blind, placebo-
controlled clinical trial. Ann Intern Med 2002;136:1–12.
12. Saag KG, Criswell LA, Sems KM, Nettleman MD, Kolluri S.
Low-dose corticosteroids in rheumatoid arthritis: a meta-analysis
of their moderate-term effectiveness. Arthritis Rheum 1996;39:
13. Gotzsche PC, Johansen HK. Meta-analysis of short-term low-dose
prednisolone versus placebo and non-steroidal anti-inflammatory
drugs in rheumatoid arthritis. BMJ 1998;316:811–8.
14. Criswell LA, Saag KG, Sems KM, Welch V, Shea B, Wells G, et al.
Moderate-term, low-dose corticosteroids for rheumatoid arthritis.
Cochrane Database Rev 2000;2:CD001158.
15. Symmons DP, Jones MA, Scott DL, Prior P. Longterm mortality
outcome in patients with rheumatoid arthritis: early presenters
continue to do well. J Rheumatol 1998;25:1072–7.
LOW-DOSE PREDNISOLONE IN EARLY RHEUMATOID ARTHRITIS3369
16. Anderson JJ, Wells G, Verhoeven AC, Felson DT. Factors Download full-text
predicting response to treatment in rheumatoid arthritis: the
importance of disease duration. Arthritis Rheum 2000;43:22–9.
17. O’Dell JR. Treating rheumatoid arthritis early: a window of
opportunity? [editorial]. Arthritis Rheum 2002;46:283–5.
18. Svensson B, Schaufelberger C, Teleman A, Theander J. Remission
and response to early treatment of RA assessed by the Disease
Activity Score. Rheumatology (Oxford) 2000;39:1031–6.
19. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF,
Cooper NS, et al. The American Rheumatism Association 1987
revised criteria for the classification of rheumatoid arthritis.
Arthritis Rheum 1988;31:315–24.
20. Prevoo ML, van ’t Hof MA, Kuper HH, van Leeuwen MA, van de
Putte LB, van Riel PL. Modified disease activity scores that
include twenty-eight–joint counts: development and validation in a
prospective longitudinal study of patients with rheumatoid arthri-
tis. Arthritis Rheum 1995;38:44–8.
21. Van der Heijde DM, van Leeuwen MA, van Riel PL, Koster AM,
van ’t Hof MA, van Rijswijk MH, et al. Biannual radiographic
assessments of hands and feet in a three-year prospective followup
of patients with early rheumatoid arthritis. Arthritis Rheum 1992;
22. Van der Heijde D, Boonen A, Boers M, Kostense P, van der
Linden S. Reading radiographs in chronological order, in pairs or
as single films has important implications for the discriminative
power of rheumatoid arthritis clinical trials. Rheumatology (Ox-
23. Prevoo ML, van Gestel AM, van ’t Hof MA, van Rijswijk MH, van
de Putte LB, van Riel PL. Remission in a prospective study of
patients with rheumatoid arthritis: American Rheumatism Asso-
ciation preliminary remission criteria in relation to the disease
activity score. Br J Rheumatol 1996;35:1101–5.
24. Ekdahl C, Eberhardt K, Andersson SI, Svensson B. Assessing
disability in patients with rheumatoid arthritis: use of a Swedish
version of the Stanford Health Assessment Questionnaire. Scand
J Rheumatol 1988;17:263–71.
25. Eberhardt KB, Svensson B, Moritz U. Functional assessment of
early rheumatoid arthritis. Br J Rheumatol 1988;27:364–71.
26. Lunar Corporation. Operator manual, Expert-XL software version
1.7. Madison (WI): Lunar Corporation; 1998.
27. Landewe R, van der Heijde D. Radiographic progression depicted
by probability plots: presenting data with optimal use of individual
values. Arthritis Rheum 2004;50:699–706.
28. Bruynesteyn K, Boers M, Kostense P, van der Linden S, van der
Heijde D. Deciding on progression of joint damage in paired films
of individual patients: smallest detectable difference or change.
Ann Rheum Dis 2005:64:179–82.
29. Joint Committee of the Medical Research Council and Nuffield
Foundation. A comparison of prednisolone with aspirin or other
analgesics in the treatment of rheumatoid arthritis. Ann Rheum
30. Boers M, Verhoeven AC, Markusse HM, van de Laar MA,
Westhovens R, van Denderen JC, et al. Randomised comparison
of combined step-down prednisolone, methotrexate and sul-
phasalazine with sulphasalazine alone in early rheumatoid arthri-
tis. Lancet 1997;350:309–18.
31. Maillefert JF, Combe B, Goupille P, Cantagrel A, Dougados M.
Long term structural effects of combination therapy in patients
with early rheumatoid arthritis: five year follow up of a prospective
double blind controlled study. Ann Rheum Dis 2003;62:764–6.
32. Rau R, Wassenberg S, Zeidler H. Low dose prednisolone therapy
(LDPT) retards radiographically detectable destruction in early
rheumatoid arthritis: preliminary results of a multicentre random-
ized, parallel, double blind study. Z Rheumatol 2000;59 Suppl
33. Capell HA, Madhok R, Hunter JA, Porter D, Morrison E, Larkin
J, et al, on behalf of the WOSERACT group. Lack of radiological
and clinical benefit over two years of low dose prednisolone for
rheumatoid arthritis: results of a randomised controlled trial. Ann
Rheum Dis 2004;63:797–803.
34. Kirwan J, Boers M. Low dose prednisolone for treatment of RA.
Ann Rheum Dis 2004;63:1533–4.
35. Van Leeuwen MA, van Rijswijk MH, Sluiter WJ, van Riel PL,
Kuper IH, van de Putte LB, et al. Individual relationship between
progression of radiological damage and the acute phase response
in early rheumatoid arthritis: towards development of a decision
support system. J Rheumatol 1997;24:20–7.
36. Welsing PM, Landewe RB, van Riel PL, Boers M, van Gestel AM,
van der Linden S, et al. The relationship between disease activity
and radiologic progression in patients with rheumatoid arthritis: a
longitudinal analysis. Arthritis Rheum 2004;50:2082–93.
37. Landewe RB, Boers M, Verhoeven AC, Westhovens R, van de
Laar MA, Markusse HM, et al. COBRA combination therapy in
patients with early rheumatoid arthritis: long-term structural ben-
efits of a brief intervention. Arthritis Rheum 2002;46:347–56.
38. Mulherin D, Fitzgerald O, Bresnihan B. Clinical improvement and
radiological deterioration in rheumatoid arthritis: evidence that
the pathogenesis of synovial inflammation and articular erosion
may differ. Br J Rheumatol 1996;35:1263–8.
39. Kirwan JR. The synovium in rheumatoid arthritis: evidence for (at
least) two pathologies [editorial]. Arthritis Rheum 2004;50:1–4.
40. Landewe R, Geusens P, Boers M, van der Heijde D, Lems W, te
Koppele J, et al. Markers for type II collagen breakdown predict
the effect of disease-modifying treatment on long-term radio-
graphic progression in patients with rheumatoid arthritis. Arthritis
41. Scott DL, Pugner K, Kaarela K, Doyle A, Woolf A, Holmes J, et
al. The link between joint damage and disability in rheumatoid
arthritis. Rheumatology (Oxford) 2000;39:122–32.
42. Sokka T, Pincus T. Eligibility of patients in routine care for major
clinical trials of anti–tumor necrosis factor ? agents in rheumatoid
arthritis. Arthritis Rheum 2003;48:313–8.
43. Keller C, Hafstrom I, Svensson B. Bone mineral density in women
and men with early rheumatoid arthritis. Scand J Rheumatol
44. Van Everdingen AA, Siewertsz van Reesema DR, Jacobs JW,
Bijlsma JW. The clinical effect of glucocorticoids in patients with
rheumatoid arthritis may be masked by decreased use of additional
therapies. Arthritis Rheum 2004;51:233–8.
3370 SVENSSON ET AL