r e v b r a s r e u m a t o l .

2 0 1 5;5 5(6):477–484

REVISTA BRASILEIRA DE
REUMATOLOGIA
www.reumatologia.com.br

Original article

Correlation of rheumatoid arthritis activity indexes

(Disease Activity Score 28 measured with ESR and
CRP, Simplified Disease Activity Index and Clinical
Disease Activity Index) and agreement of disease
activity states with various cut-off points in a
Northeastern Brazilian population

Marta Maria das Chagas Medeiros a,∗ , Brenda Maria Gurgel Barreto de Oliveira b ,
João Victor Medeiros de Cerqueira c , Raquel Telles de Souza Quixadá b ,
Ídila Mont’Alverne Xavier de Oliveira b
a Faculdade de Medicina, Universidade Federal do Ceará, Fortaleza, CE, Brazil
b Hospital Universitário Walter Cantídio, Fortaleza, CE, Brazil
c Universidade de Fortaleza, Fortaleza, CE, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Introduction: The Disease Activity Score 28 (DAS28) and its versions have been used to mea-
Received 20 February 2014 sure rheumatoid arthritis activity, but there is no consensus about which one is the best.
Accepted 1 December 2014 Objectives: Determine the correlation among indexes (DAS28 ESR, DAS28 CRP, SDAI and CDAI)
Available online 10 March 2015 and evaluate agreement of activity strata using different cut-off points.
Methods: Rheumatoid arthritis patients were cross-sectionally evaluated with data collec-
Keywords: tion to calculate the DAS28 (ESR and CRP), SDAI and CDAI, using different cut-offs for defining
Rheumatoid arthritis remission, mild, moderate and high activity. Pearson correlations were calculated for con-
Disease activity tinuous measures and agreement (kappa test) for the strata (remission, mild, moderate and
DAS28 high activity).
Correlation Results: Of 111 patients included, 108 were women, age 55.6 years, 11-year disease duration.
Agreement DAS28 (ESR) was significantly higher than DAS28 (CRP) (4.0 vs. 3.5; p < 0.001) and the values
remained higher after stratification by age, gender, disease duration, rheumatoid factor and
HAQ. Correlations among indexes ranged from 0.84 to 0.99, with better correlation between
SDAI and CDAI. Agreements among activity strata ranged from 46.8% to 95.8%. DAS28 (CRP)
with cut-off point for the remission of 2.3 underestimated disease activity by 45.8% com-
pared with DAS28 (ESR). SDAI and CDAI showed agreement of 95.8%. The four indexes were
associated with disease duration and HAQ.

∗
Corresponding author.
E-mail: [email protected] (M.M.C. Medeiros).
http://dx.doi.org/10.1016/j.rbre.2014.12.005
2255-5021/© 2015 Elsevier Editora Ltda. All rights reserved.
478 r e v b r a s r e u m a t o l . 2 0 1 5;5 5(6):477–484

Conclusions: Although the activity indexes show good correlation, they show discrepancies
in activity strata, thus requiring more researches to define a better index and better cut-off
points.
© 2015 Elsevier Editora Ltda. All rights reserved.

Correlação dos índices de atividade da artrite reumatoide (Disease

Activity Score 28 medidos com VHS, PCR, Simplified Disease Activity
Index e Clinical Disease Activity Index) e concordância dos estados de
atividade da doença com vários pontos de corte numa população do
nordeste brasileiro

r e s u m o

Palavras-chave: Introdução: O Disease Activity Score 28 (DAS28) e versões têm sido usados para medir ativi-
Artrite reumatoide dade da artrite reumatoide (AR), mas não existe consenso sobre qual é o melhor.
Atividade doença Objetivos: Determinar a correlação entre os índices (DAS28 VHS, DAS28 PCR, SDAI e CDAI) e
DAS28 avaliar a concordância dos estratos de atividade com o uso de diferentes pontos de corte.
Correlação Métodos: Pacientes com AR foram avaliados transversalmente com coleta de dados para
Concordância cálculo do DAS28 (VHS e PCR), SDAI e CDAI, com o uso de pontos de cortes diferentes para
definição de remissão, atividade leve, moderada e alta. Correlações de Pearson foram calcu-
ladas para medidas contínuas e concordância (teste de kappa) para os estratos (remissão,
atividade leve, moderada e alta).
Resultados: De 111 pacientes incluídos, 108 foram mulheres, média de 55,6 anos, tempo de
doença de 11 anos. DAS28 (VHS) foi significantemente maior do que DAS28 (PCR) (4 vs. 3,5;
p < 0,001) e os valores permaneceram maiores após estratificação por idade, sexo, tempo
doença, fator reumatoide e HAQ. Correlações entre índices variaram de 0,84 a 0,99, com
melhor correlação entre SDAI e CDAI. Concordâncias entre estratos de atividade variaram
de 46,8% a 95,8%. DAS28 (PCR) com ponto de corte para remissão de 2,3 subestimou ativi-
dade da doença em 45,8% quando comparado com DAS28 (VHS). SDAI e CDAI apresentaram
concordância de 95,8%. Os quatro índices mostraram associação com tempo de doença e
HAQ.
Conclusões: Embora os índices de atividade apresentem boa correlação, mostram discrepân-
cias nos estratos de atividade. Tornam-se necessários mais estudos para definir melhor
índice e melhores pontos de corte.
© 2015 Elsevier Editora Ltda. Todos os direitos reservados.

or of disease activity.4 This original DAS uses Ritchie articu-

Introduction
lar index (26 joint regions) to assess the number of painful
joints, and 44 joints to evaluate the swollen ones. Later,
The main goals to be achieved during the treatment of patients the proposed DAS28 started using only 28 joints for count-
with rheumatoid arthritis (RA) are pain relief and the strict ing the swollen and tender joints5 and allowed the optional
control of the joint inflammatory process. Aiming at having a use of C-reactive Protein (CRP) instead of ESR as an inflam-
more adequate evaluation of the inflammatory activity in clin- matory marker.6 Since then, DAS28 was the most common
ical trials, the American College of Rheumatology (ACR), the measure used to evaluate the inflammatory activity, both in
European League Against Rheumatism (EULAR) and the World clinical trials and in clinical practice. However, this index
Health Organization/International League Against Rheuma- requires a complex formula including square root of Nape-
tism (WHO/ILAR) proposed a set of variables (core sets), which rian logarithm, requiring a technology tool for its calculation.
included the number of tender and swollen joints, measure- Therefore, more simple indexes were later proposed: Sim-
ment of pain, global assessment of disease activity by the plified Disease Activity Index (SDAI)7 and Clinical Disease
physician and patient, acute phase reactant and functional Activity Index (CDAI).8 SDAI is a measure proposed by Smolen
measure.1–3 et al.,7 the result of which is the simple sum of the num-
In the early 90s another index to measure RA activity ber of painful joints (28 joints), number of swollen joints (28
in clinical practice was proposed, and was called Disease joints), assessment of disease activity made by the patient
Activity Score (DAS), which considers the number of ten- in a visual analog scale from 0 to 10 cm, evaluation of dis-
der and swollen joints, erythrocyte sedimentation rate (ESR) ease activity by the physician (0–10 cm) and CRP (mg/dL). The
and assessment performed by the patient of global health calculation of the CDAI is simpler because the sum does not
 r e v b r a s r e u m a t o l . 2 0 1 5;5 5(6):477–484 479

Table 1 – Cut-off points of indexes DAS28, SDAI, and CDAI to define states of disease activity.
Indexes Disease remission Mild activity Moderate activity High activity

Original DAS28 (ESR)5 <2.6 2.6–3.2 >3.2–5.1 >5.1

DAS28 (ESR) Aletaha et al.16 <2.4 2.4–3.6 >3.6–5.5 >5.5
DAS28 (CRP) Inoue et al.17 <2.3 2.3–2.7 >2.7–4.1 >4.1
DAS28 (CRP) Castrejón et al.18 <2.3 2.3–3.8 >3.8–4.9 >4.9
DAS28 (CRP) Fujiwara et al.11 <1.72 1.72–2.98 >2.98–4.77 >4.77
SDAI7 <3.3 3.3–11 >11–26 >26
CDAI8 <2.8 2.8–10 >10–22 >22

DAS28, Disease Activity Score (28 joints); SDAI, Simplified Disease Activity Index; CDAI, Clinical Disease Activity Index; ESR, erythrocyte
sedimentation rate; CRP, C-reactive protein.

take CRP into account, only the first four measures. Although
indexes show good correlation with each other,7–11 DAS28
Methods
is the most validated index for measuring disease activity.
Another advantage is that it is possible to use both ESR and Patients with RA diagnosis according to ACR criteria17 who
CRP as an inflammatory marker, but use of this latter marker were followed in the outpatient’s of the Rheumatology service
still requires further study, since discrepancies between ESR of the University Hospital Walter Cantidio at the Federal
and CRP have been reported in some patients with RA, with University of Ceará were sequentially invited to participate
a trend toward higher values of ESR and lower values of in the study. The presence of other autoimmune diseases,
CRP.12,13 except secondary Sjogren’s syndrome, were excluded. The
All these indexes (DAS28, SDAI and CDAI) measure disease study design was cross-sectional. Data collection took place
activity on a continuous scale, and also allow categorizing from January to December 2013. Demographic data (gender,
the patient in activity strata, using different cut-off points: age, race, education level), clinical data related to RA (disease
remission, mild, moderate and high activity. With the emer- duration since diagnosis, presence of extra-articular manifes-
gence of several new drugs in the last 15 years to treat RA, tations, rheumatoid factor, medications used), were collected
disease remission is a goal that shall be sought. The indexes from medical records.
cut-off points that define disease remission vary in the litera- To calculate the activity indexes of the disease (DAS28, SDAI
ture. For example, the original DAS28 established that clinical and CDAI), the rheumatologist on the day of consultation col-
remission was defined when DAS28 (ESR) was <2.6.5 In 2005, lected the following data: count of the number of painful and
Aletaha et al. proposed to lower the cut-off point to <2.4.14 swollen joints in 28 joints (shoulders, elbows, wrists, metacar-
As to the best cut-off point to define remission when CRP pophalangeal, proximal interphalangeal, knees), global health
is used in the calculation of DAS28, a value below 2.3 was assessment (scale 0–100) by the patient, assessment of disease
already suggested.15,16 In 2013, Fujiwara and Kita’s study con- activity by the patient and physician (0–10) and inflamma-
cluded that the best index to define clinical remission was tory activity markers carried out within a maximum period
DAS28 measured by CRP with the conventional cut-off of 2.3 of 2 weeks before the consultation (ESR and CRP). If the
reduced to 1.72.10 The cut-off points for other activity strata of patient’s condition had changed after the completion of
the disease also vary among the indexes and this can lead to inflammatory markers, these were not considered and a
inconsistency in the classification of disease activity, resulting new assessment was scheduled. Patients with categorical
in different practices, also affecting the comparison of studies CRP results (positive or negative; < or >) were not considered
when using different criteria. for calculation of indexes that take CRP into account. Lab-
Another very important point is that the possibility of oratory tests were performed at the Central Laboratory of
patients’ ethnic origin influences the activity indexes, making HUWC and the methods employed were: ESR (Sedi-System
the generalization of studies’ results inadequate. Differences Automation) and CRP (immunoturbidimetry ROCHE COBAS).
in genetic polymorphism that influences CRP levels, as well DAS28 was calculated with software for specific calcula-
as other genetic and cultural factors of each population, can tion of DAS, using both ESR and CRP (mg/dl), global health
influence disease activity measures, requiring that studies on assessment by the patient, and the number of tender and
the subject are developed in different populations to estab- swollen joints (28 joints). SDAI was calculated by adding the
lish the best index. Studies comparing different versions of number of swollen joints (0–28), number of tender joints
DAS28 were performed predominantly in Europe and Asia and (0–28), evaluation of disease activity by the patient (0–10),
in some African-American and black African populations. evaluation of disease activity by the doctor (0–10) and the
The objectives of this study were to determine a correlation value of CRP (0.1–10 mg/dL). The result of the CDAI was
among the most popular indicators for measuring activity of the sum of the four previous clinical measurements without
rheumatoid arthritis (DAS28 calculated with ESR, DAS28 cal- CRP.
culated with CRP, SDAI and CDAI) and assess the agreement of The cut-off points considered to define the inflammatory
disease activity states defined by the indexes using different activity states are shown in Table 1. For the DAS28 calculated
cut-off points of DAS28 in a sample of patients in northeastern with ESR, we used the cut-off points of the original study5 and
Brazil. the study of Aletaha et al.14 For the DAS28 calculated with CRP
480 r e v b r a s r e u m a t o l . 2 0 1 5;5 5(6):477–484

we used the cut-off points proposed by Inoue et al.,15 Castre-

Table 2 – Characteristics of patients with rheumatoid
jon et al.16 and Fujiwara et al.10 The physical function was arthritis.
assessed using the Health Assessment Questionnaire (HAQ)18
Characteristics
with score of 0–3, with 0 score meaning no loss of physical
function, 3, full disability. Female (%) 97.3%
Age (mean ± SD) 55.6 ± 12.9

Skin color (%)

Statistics White 22.2%
Brown 31.5%
Paired data of continuous variables as DAS28 (ESR) and DAS28 Brunette 39.8%
(CRP) and SDAI and CDAI were compared using paired t- Black 6.5%
test. To estimate correlation between the continuous values Marital Status (%)
of DAS28 (ESR), DAS28 (CRP), SDAI, CDAI and HAQ Pearson’s Single 33.9%
correlation coefficient was used. To compare the averages, Married 50.5%
Kruskal Wallis and Mann Whitney tests were used. The Separated/widower 15.6%
agreement between the inflammatory activity states (remis- Education level (%)
sion, mild, moderate and high activity) was determined by Illiterate/literate 33.6%
kappa test. The level of statistical significance was 0.05. 1st Grade 40.0%
Statistical analysis was performed using STATA version 9.0 2nd Grade 22.7%
Graduation 3.7%
software.
Disease duration (years)
Mean ± SD 11.08 ± 7.3
Results Median (interquartile range) 10 (5.16)
Rheumatoid factor (%) 84.1%
A total of 111 patients were studied, predominantly female
Medications used (%)
(108 women and 3 men), most of them of white/brown race
Chloroquine 58.6%
(62%) with a mean age of 55.5 years (SD = 12.9) and duration Methotrexate 95.5%
of disease of 11.08 years (SD = 7.3) (Table 2). Rheumatoid fac- Leflunomide 71.2%
tor was positive in 84.1% of the sample and most made use Sulfasalazine 15.3%
of methotrexate and/or leflunomide (95.5% and 71.2%). The Biological (anti TNF-␣ and abatacept) 22.5%
mean (± SD) and median (interquartile range 25–75) values Current daily dose of prednisone (mg)
of DAS28 (ESR), DAS28 (CRP), SDAI, CDAI and HAQ are also Mean ± SD 5.1 ± 2.2
shown in Table 2. Mean DAS28 (ESR) was statistically higher ESR (mm/h)
than mean DAS28 (CRP) (p < 0.001) and also that of SDAI was Mean ± SD 29.7 ± 21.5
statistically higher than CDAI (p < 0.001). The calculation of Median (interquartile range) 24.5 (15–38)
the indexes using CRP was done with 96 of the 111 patients
CRP (mg/dL)
because 15 patients did not get the proper dosage of CRP, only Mean ± SD 0.9 ± 1.0
of ESR. Median (interquartile range) 0.6 (0.3–1.1)
The mean DAS28 (ESR), DAS28 (CRP), SDAI and CDAI
DAS28 (ESR)
were also evaluated by age strata (<40, 40 to <50, 50 to Mean ± SD 4.03 ± 1.40
<60 and ≥60 years), disease duration (≤5, >5–10, >10–20, Median (IQ) 3.99 (3.09–4.92)
>20 years), gender, rheumatoid factor (positive, negative),
DAS28 (CRP)
and HAQ (0–1, >1–2, >2–3) (Table 3). The values of the four Mean ± SD 3.55 ± 1.27
activity indexes were statistically different for each stra- Median (IQ) 3.38 (2.52–4.47)
tum of disease duration, with the highest values after 20
SDAI
years of disease, and then with ≤5 years of disease. The
Mean ± SD 16.53 ± 10.54
average values were also significantly different for HAQ stra- Median (IQ) 14.1 (9.05–22.4)
tum, with progressively increasing values as HAQ score range
CDAI
increased.
Mean ± SD 15.62 ± 10.03
Correlations between DAS28 (ESR), DAS28 (CRP), SDAI Median (IQ) 13.5 (8–21)
and CDAI were all statistically significant (p < 0.0001). Very
HAQ
strong correlations (>0.90) were observed among DAS28 (ESR)
Mean ± SD 0.98 ± 0.69
and DAS28 (CRP) (0.92), DAS28 (CRP) and SDAI (0.93), DAS28 Median (IQ) 1 (0.37–1.62)
(CRP) and CDAI (0.92) and between SDAI and CDAI (0.99).
Strong correlations (between 0.6 and 0.9) were observed ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; DAS28
when comparing DAS28 (ESR) with SDAI and CDAI (0.84). (ESR), disease activity score using ESR; DAS28 (CRP), disease activ-
ity score using CRP; SDAI, Simplified Disease Activity Index; CDAI,
Regular correlations (0.3–0.6) were observed comparing HAQ
Clinical Disease Activity Index; HAQ, health assessment question-
with DAS28 (ESR) (0.50), DAS28 (CRP) (0.48), SDAI and CDAI
naire.
(both 0.53). Comparing DAS28 (ESR) with DAS28 (CRP) and SDAI with CDAI, the
When indexes were categorized by activity strata of the level of statistical significance was <0.001.
disease (remission, mild, moderate and high activity), the
 r e v b r a s r e u m a t o l . 2 0 1 5;5 5(6):477–484 481

The agreements among activity strata (remission, mild,

Table 3 – Values (median ± SD) of DAS28 (ESR), DAS28
(CRP), CDAI and SDAI by strata of age, disease time, moderate and high activity) of different indexes and on using
rheumatoid factor and sex. different cut-off points are shown in Table 5. The best agree-
ment of categories observed was between SDAI and CDAI
DAS28 (ESR) DAS28 (CRP) SDAI CDAI
(n = 111) (n = 96) (n = 96) (n = 111) (95.8%). Only one patient was in a lower category when
CDAI definition was used. When original DAS28 (ESR) was
Age: compared with the original DAS28 (ESR) proposed by Ale-
<40 years 3.7 ± 1.6 3.6 ± 1.3 14.9 ± 11.0 14.4 ± 11.0
taha, agreement of activity strata was observed in 83.3% of
40–49 years 3.8 ± 1.4 3.6 ± 1.3 16.0 ± 10.5 15.2 ± 10.1
50–59 years 4.3 ± 1.3 3.8 ± 1.3 17.8 ± 10.7 16.8 ± 10.1
patients (80/96) with 14 patients (14.6%) falling to a lower
≥60 years 3.8 ± 1.4 3.3 ± 1.2 13.8 ± 9.3 13.0 ± 8.9 category (underestimated activity) when using Aletaha crite-
p NS NS NS NS ria (Table 6). When original DAS28 (ESR) was compared with
DAS28 (CRP) by Inoue, Castrejon and Fujiwara, the best agree-
Gender:
Female 3.9 ± 1.3 3.5 ± 1.3 15.5 ± 10.0 14.7 ± 9.6 ment was with Fujiwara criteria (67.7%). The cut-off points
Male 5.0 ± 2.3 5.1 ± 0.6 24.6 ± 17.5 22.6 ± 15.0 suggested by Castrejón underestimated the activity in almost
p NS 0.07 NS NS half of the patients (45.8%), while those by Inoue and Fujiwara
Disease time: underestimated by 16.6%. The proportions of overestimated
≤5 years 4.4 ± 1.4 4.1 ± 1.2 20.2 ± 10.7 19.0 ± 10.2 categories by Castrejon, Inoue and Fujiwara were, respectively,
>5–10 years 3.7 ± 1.3 3.4 ± 1.2 13.9 ± 10.1 13.0 ± 9.7 7.3%, 20.8% and 14.6%. Comparing indexes that use CRP to cal-
>10–20 years 3.6 ± 1.3 3.1 ± 1.2 12.7 ± 9.2 12.1 ± 8.7 culate the DAS28, the best agreement was that between what
>20 years 5.0 ± 1.2 4.2 ± 0.8 22.8 ± 5.7 22.0 ± 5.6 was proposed by Inoue and Fujiwara (71.8%), with Fujiwara
p 0.007 0.003 0.0005 0.0005
overestimating the strata in 11 patients (11.4%) and underes-
Rheumatoid factor: timating in 16 patients (16.7%). The other agreements between
Positive 3.9 ± 1.3 3.6 ± 1.2 15.7 ± 9.6 14.9 ± 9.3 multiple comparisons performed ranged from 46.8% to
Negative 3.9 ± 1.7 3.5 ± 1.6 16.6 ± 14.0 15.6 ± 13.2
67.7%.
p NS NS NS NS

HAQ:
0–1 3.5 ± 1.2 3.1 ± 1.1 11.8 ± 8.4 11.2 ± 8.1 Discussion
>1–2 4.5 ± 1.4 4.0 ± 1.3 20.8 ± 10.3 19.8 ± 9.8
>2–3 4.7 ± 1.6 4.2 ± 1.2 23.0 ± 9.7 21.6 ± 8.9
p 0.001 0.001 0.0001 0.0001
With the growing and urgent need for assessment of rheuma-
toid arthritis activity not only in clinical practice but also
ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; DAS28 in the evaluation of efficacy of new treatments in clinical
(ESR), disease activity score using ESR; DAS28 (CRP), disease activity trials, it is increasingly important to have instruments to mea-
score using CRP; SDAI, Simplified Disease Activity Index; CDAI, Clin-
sure disease activity as accurately as possible. Some indexes
ical Disease Activity Index; HAQ, health assessment questionnaire.
have been proposed in recent decades; however, they still
have properties that need better validation, since they were
tested in specific populations.4–16 There are two main ways
proportions of patients in each category according to the cut- to compare the rates between them: (1) as continuous meas-
off points adopted to calculate DAS28 using ESR (original5 ures, comparing means and medians in the same patients
and Aletaha14 ), of DAS28 using CRP (Inoue,15 Castrejón16 and and (2) as categorical measures (remission, mild, moder-
Fujiwara10 ) and SDAI and CDAI are presented in Table 4. Reduc- ate and high activity) and comparing the agreement among
ing the cut-off point from 2.6 to 2.4 in the calculation of strata.
DAS28 (ESR) proposed by Aletaha reduces the percentage of In the present study, we analyzed DAS28 calculated with
patients considered to be in remission from 15.6% to 13.5%. ESR and CRP, SDAI and CDAI as continuous variables, and
Calculating the DAS28 with CRP and cut-off of 2.3 proposed also compared the four strata to each other according to the
by Inoue and Castrejon, the percentage of patients in remis- criteria originally proposed5 and more recently by Aletaha14
sion categorized as in remission increases about 2.1% and 4.2% for the calculation of DAS28 using ESR; the criteria were
compared to original DAS28 criteria (ESR) and Aletaha, respec- proposed for DAS28 using CRP according to three studies
tively (Table 4). The reduction of the cut-off to 1.72 proposed (Inoue,15 Castrejón16 and Fujiwara10 ) and SDAI and CDAI.
by Fujiwara makes this criterion more stringent for consider- The first observation in this study, corroborated by several
ing remission because it reduces from 15.6% (DAS28 ESR) to other studies,8,13,16,19–21 is that the values of DAS28 with ESR
6.3% in this category. Of all the indexes were analyzed and are higher than the values of DAS28 by CRP, even when
presented in Table 4, the cut-off points of SDAI and CDAI to stratified by age groups, disease duration, rheumatoid fac-
define clinical remission are the most stringent, reducing from tor, gender and HAQ score. Therefore, the activity of the
15.6% through original DAS28 (ESR) calculation to 4.2%. The disease may be underestimated when using DAS28 with
absence of CRP in the calculation of CDAI does not change CRP instead of DAS28 with ESR. The excellent correlation
anything at all in terms of remission criteria when compared between the indexes observed in our study was also reg-
to the SDAI. The percentage of patients in the categories istered by other authors.13,15,21–23 Although the correlation
of mild and moderate activity has greater variation among between the two indexes has been very high (92%), the cor-
indexes than categories at the extremes (remission and high relation in the four activity strata comparing the original
activity). DAS28 (ESR) with DAS28 (CRP) with different cut-off points
482 r e v b r a s r e u m a t o l . 2 0 1 5;5 5(6):477–484

Table 4 – Proportion of patients in remission, mild, moderate and high activity according to the scores of DAS28 (ESR),
DAS28 (CRP), SDAI and CDAI.
Remission Mild activity Moderate activity High activity

Original DAS28 (ESR) 15.6% 12.5% 50.0% 22.9%

DAS28 (ESR) Aletaha et al. 13.5% 28.1% 36.5% 21.9%
DAS28 (CRP) Inoue et al. 17.7% 13.5% 37.5% 31.3%
DAS28 (CRP) Castrejón et al. 17.7% 43.7% 21.9% 16.7%
DAS28 (CRP) Fujiwara et al. 6.3% 30.2% 43.7% 19.8%
SDAI 4.2% 30.2% 48.9% 16.7%
CDAI 4.2% 31.3% 44.8% 19.7%

ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; DAS28 (ESR), disease activity score using ESR; DAS28 (CRP), disease activity score
using CRP; SDAI, Simplified Disease Activity Index; CDAI, Clinical Disease Activity Index.

Table 5 – Proportions of agreement and kappa coefficients among DAS28 (ESR), DAS28 (CRP), SDAI, and CDAI by strata of
disease activity.
Kappa coefficient Agreement

Original DAS28 (ESR) vs. DAS28 (ESR) Aletaha 0.76 83.3%

Original DAS28 (ESR) vs. DAS28 (CRP) Inoue 0.46 62.5%
DAS28 (ESR) original vs. DAS28 (CRP) Castrejón 0.31 46.8%
Original DAS28 (ESR) vs. DAS28 (CRP) Fujiwara 0.53 67.7%
Original DAS28 (ESR) vs. SDAI 0.46 63.5%
Original DAS28 (ESR) vs. CDAI 0.44 61.5%
DAS28 (CRP) Inoue vs. DAS28 (CRP) Castrejón 0.42 55.2%
DAS28 (CRP) Inoue vs. DAS28 (CRP) Fujiwara 0.61 71.8%
DAS28 (CRP) Inoue vs. SDAI 0.38 56.2%
DAS28 (CRP) Inoue vs. CDAI 0.42 58.3%
DAS28 (CRP) Castrejón vs. DAS28 (CRP) Fujiwara 0.45 60.4%
DAS28 (CRP) Castrejón vs. SDAI 0.38 55.2%
DAS28 (CRP) Castrejón vs. CDAI 0.35 53.1%
SDAI vs. CDAI 0.93 95.8%

ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; DAS28 (ESR), disease activity score using ESR; DAS28 (CRP), disease activity score
using CRP; SDAI, Simplified Disease Activity Index; CDAI, Clinical Disease Activity Index.

ranged from 46.8% to 62, 5% of patients. Therefore, an excel- cut-off for high activity, causing a greater number of patients
lent correlation does not necessarily mean good agreement to migrate from higher categories to lower ones. Impor-
between the categories of activity. For instance, the criteria tantly, Castrejón et al.16 estimated these cut-off points in a
proposed by Castrejón et al. underestimate the level of activ- population of patients with recent-onset RA (median dura-
ity in almost half of the patients. The study by Tamhane tion of disease in the first visit: 6 months; IQ 3.6–9 months).
et al., comparing original DAS28 (ESR) with DAS28 (CRP) by The population tested in this study was of patients with long-
Inoue and Castrejon, also found an underestimation of dis- term RA (mean: 11.08 ± 7.3 years; median: 10 years; IQ: 5–16
ease activity in 40% of patients when the CRP was used.13 years) and the study by Tamhane et al.13 had an average of
This can be explained by lowering of the cut-off point and 6.6 ± 9.3 years (median: 1.8 years; IQ: 0.8–9.3). When we cate-
expansion of the range of mild activity and reduction of the gorize the time of disease in groups (≤5 years, >5–10 years;

Table 6 – Agreement between original DAS28 (ESR) and DAS28 (ESR) by Aletaha et al.
Remission Mild activity Moderate activity High activity Total patients
(Aletaha) (Aletaha) (Aletaha) (Aletaha)

Remission (original) 13 2 0 0 15
Mild activity (original) 0 12 0 0 12
Moderate activity (original) 0 13 34 0 47
High activity (original) 0 0 1 21 22
Total 13 27 35 21 96

ESR, erythrocyte sedimentation rate; DAS28 (ESR), disease activity score using ESR.
 r e v b r a s r e u m a t o l . 2 0 1 5;5 5(6):477–484 483

>10–20 years and >20 years) the average values of all eval- Inoue was the most sensitive in the population tested in this
uated indexes were statistically different, with the highest study.
values at the extremes of age. We can conclude that the Another interesting finding of this study is that the mean
duration of disease can influence the disease activity indexes values of the four indexes studied increased progressively
and therefore validation of a particular index should take with the increase of HAQ and were statistically different. This
the time of the disease into consideration. The cut-off points can be explained because the physical function can influ-
proposed by Castrejon et al.16 may not apply to patients ence the global health assessment and activity made by the
with long-term disease. The cut-offs of DAS28 (CRP) by Inoue patient and doctor. The HAQ correlations with indexes were
and Fujiwara were established in a population of patients very similar to other studies.7,9,14 Age, sex and rheumatoid
with long-standing RA (about 10 years) and both underes- factor did not influence the values of the indexes stud-
timated the level of activity in only 16.6% when compared ied in the present study. Although the mean values of the
with original DAS28 (ESR), and overestimated in 20.8% and four indexes were all numerically higher in men than in
14.6%, respectively. Analyzing these properties, it appears that women, just for DAS28 (CRP) there was a trend toward statis-
the reduction of all cut-off points for the four strata pro- tical difference (p = 0.07). The small number of male patients
posed by Fujiwara is closer to the original DAS28 criteria may have influenced the statistical outcome. Some authors
(ESR). suggest that the calculation of DAS28 considers the gender
SDAI and CDAI correlated with DAS28 (ESR) of 0.84, and and age,13,21 but this has not been established and vali-
the first two with each other of 0.99. In the original study dated.
of SDAI the correlation with DAS28 (ESR) was 0.90.7 Sta- In short, while the indexes DAS28 (ESR), DAS28 (CRP), SDAI
tistical validity excluding CRP from SDAI was very much and CDAI correlated well with each other, they had many
tested in the original study of CDAI.8 In several analyses discrepancies regarding the categories/strata of inflamma-
made between SDAI and CDAI, the values correlated almost tory activity. Regarding the calculation of DAS28, the value
perfectly and the correlations between SDAI or CDAI with of disease activity as measured using CRP was underesti-
original DAS28 (ESR) ranged from 0.87 to 0.90. Although CRP mated compared to the measure using ESR. Therefore, the
showed no significant colinearity with other SDAI meas- definition of remission and activity by a particular index
ures, only 5% of the SDAI remained unexplained when it may not have the same meaning when measured by another
excluded CRP; in DAS28 (ESR), ESR only contributed to about index.
15%. Mean CDAI values are obviously smaller than the SDAI, The activity of the disease evaluated by SDAI and CDAI
since the value of CRP is suppressed in the calculation. The showed excellent correlation and agreement, stressing that
degrees of agreement of this study of original DAS28 (ESR) the CRP dosage is not indispensable for activity clinical assess-
with the original SDAI and CDAI were low, but the agree- ment. Therefore, CDAI can be used instead of SDAI because its
ment of SDAI with CDAI was almost perfect (95.8%). This simplicity allows its use at any time, and anywhere. Another
high agreement reinforces that the exclusion of CRP for the interesting fact is that all the indexes evaluated in the study
calculation of CDAI almost does not change anything in the to measure RA activity were associated with disease dura-
assessment of disease activity level compared to SDAI. Only tion and functional capacity, factors that should be taken into
4 of the 96 patients analyzed disagreed regarding the stra- account in the assessment of disease activity as measured by
tum. the indexes studied.
Defining a patient as in remission can mean reduction or The study has some limitations: small sample size, dif-
even withdrawal of some drugs in use and, therefore, the ferent evaluators to count number of tender and swollen
criteria shall prevent that a patient with residual activity joints, and patients’ cultural and ethnic background. A larger
ceases to be properly treated (underestimation of activity), and sample size would lead to more patients in the four strata
shall also prevent the patient from being overtreated (activity of activity according to the cut-off points analyzed, which
overestimation). The definition of high activity is important, would increase the power of conclusion of the study. The
especially for defining the introduction of biological agents, count of painful and swollen joints may show little variability
and also for prognostic reasons, since the probability of more when made by different evaluators. Ideally, all patients should
rapid progression is greater in patients that keep the high- be always examined by the same observer, or the examina-
est activity. The categorization of high disease activity levels tion technique should be standardized by the evaluators and
is also an important factor for eligibility of patients in clini- trained previously. Patients in the study were from a public
cal trials. Comparing the rates tested in this study, the most institution of tertiary care, with low socioeconomic status,
rigorous for the category of remission are SDAI and CDAI. low education and long-standing disease, characteristics that
While 15.6% of the patients studied were classified in this may influence the degree of response of the components to
stratum by original DAS28 (ESR), only 4.2% were at this level calculate the activity indexes, including the assessment of
by SDAI and CDAI. The cut-off point used by Fujiwara for health and activity status made by patients and counting of
the remission of 1.72, well below the cut-off points of the painful joints.
other indexes, classified 6.3% of patients as in remission. Therefore, further studies are needed to establish the best
Regarding the definition of high activity, the index covering measure of disease activity, and finer cut-off points, so that
more patients in this stratum was DAS28 (CRP) by Inoue (31.3%) the indexes more accurately express the inflammatory con-
and the lowest number was SDAI and DAS28 (CRP) by Castrejon dition, and the results of the studies are comparable, taking
(16.7%). So, for the definition of high activity, DAS28 (CRP) by demographic, cultural and clinical factors into account that
484 r e v b r a s r e u m a t o l . 2 0 1 5;5 5(6):477–484

may interfere with the assessment of disease activity. Mean- of their usefulness and validity in rheumatoid arthritis. Clin
while, the choice of index to be used will depend on: the Exp Rheumatol. 2005;23 Suppl 39:S100–8.
objective of assessment (whether for use in daily practice or 10. Fujiwara M, Kita Y. Reexamination of the assessment criteria
for rheumatoid arthritis disease activity based on comparison
in clinical trials), the desired goal (remission or therapeutic
of the Disease Activity Score 28 with other simpler
change), the practice context (availability of technology, lab- assessment methods. Mod Rheumatol. 2013;23:
oratory tests, time) and personal clinician preferences. When 260–8.
analyzing published studies, we should consider which index 11. Sing H, Gupta V, Ray S, Kumar H, Talapatra P, Kaur M, et al.
and which cut-off point were used for comparison. In clinical Evaluation of disease activity in rheumatoid arthritis by
practice, the same index with the same components for calcu- Routine Assessment of Patient Index Data 3 (RAPID3) and it is
lation should always be consistently used to allow longitudinal correlation to Disease Activity Score 28 (DAS28) and Clinical
Disease Activity Index (CDAI): an Indian experience. Clin
comparisons in decision-making.
Rheumatol. 2012;31:1663–9.
12. Rosa Neto NS, De Carvalho JF, Shoenfeld Y. Screening tests for
inflammatory activity: applications in rheumatology. Mod
Conflicts of interest
Rheumatol. 2009;19:469–77.
13. Tamhane A, Redden DT, McGwin G Jr, Brown EE, Westfall AO,
The authors declare no conflicts of interest. Reynolds RJ 4th, et al. Comparison of the disease activity
score using erythrocyte sedimentation rate and C-reactive
protein in African Americans with rheumatoid arthritis. J
references
Rheumatol. 2013;40:1812–22.
14. Aletaha D, Ward MM, Machold KP, Nell VPK, Stamm T, Smolen
JS. Remission and active disease in rheumatoid arthritis
1. Felson DT, Anderson JJ, Boers M, Bombardier C, Chernoff M, Defining criteria for disease activity states. Arthritis Rheum.
Fried B, et al. The American College of Rheumatology 2005;52:2625–36.
preliminary core set of disease activity measures for 15. Inoue E, Yamanaka H, Hara M, Tomatsu T, Kamatani N.
rheumatoid arthritis clinical trials. The Committee on Comparison of Disease Activity Score (DAS)28-erythrocyte
Outcome Measures in Rheumatoid Arthritis Clinical Trials. sedimentation rate and DAS28-C-reactive protein threshold
Arthritis Rheum. 1993;36:729–40. values. Ann Rheum Dis. 2007;66:407–9.
2. Smolen JS. The work of the Eular Standing Committee on 16. Castrejón I, Ortiz AM, Toledano E, Castañeda S, Garcia-Vadillo
International Clinical Studies Including Therapeutic Trials. Br A, Patino E, et al. Estimated cutoff points for the 28-joint
J Rheumatol. 1992;31:219–20. disease activity score based on C-reactive protein in a
3. Boers M, Tugwell P, Felson DT, Van Riel PL, Kirwan JR, longitudinal register of early arthritis. J Rheumatol.
Edmonds JP, et al. World Health Organization and 2010;37:1439–43.
International League of Associations for Rheumatology core 17. Arnett FC, Edworthy SM, Bloch DA, McShane DJ, Fries JF,
endpoints for symptom modifying antirheumatic drugs in Cooper NS, et al. The American Rheumatism Association
rheumatoid arthritis clinical trials. J Rheumatol. 1994;41 1987 revised criteria for the classification of rheumatoid
Suppl:86–9. arthritis. Arthritis Rheum. 1988;31:315–24.
4. Prevoo ML, Van’t Hof MA, Van Riel PL, Theunisse LA, Lubberts 18. Ferraz MB, Oliveira LM, Araujo PM, Atra E, Tugwell P.
EW, Van Leeuwen MA, et al. Judging disease activity in clinical Crosscultural reability of the physical ability dimension of the
practice in rheumatoid arthritis: first step in the development Health Assessment Questionnaire. J Rheumatol.
of a disease activity score. Ann Rheum Dis. 1990;49: 1990;17:813–7.
916–20. 19. Matsui T, Kuga Y, Kaneko A, Nishino J, Eto Y, Chiba N, et al.
5. Prevoo ML, Van’t Hof MA, Kuper HH, Van Leeuwen MA, Van de Disease Activity Score 28 (DAS28) using C-reactive protein
Putte LB, Van Riel PL. Modified disease activity scores that underestimates disease activity and overestimates EULAR
include twenty-eight-joint counts. Development and response criteria compared with DAS28 using erythrocyte
validation in a prospective longitudinal study of patients with sedimentation rate in a large observational cohort of
rheumatoid arthritis. Arthritis Rheum. 1995;38: rheumatoid arthritis patients in Japan. Ann Rheum Dis.
44–8. 2007;66:1221–6.
6. Wells G, Becker JC, Teng J, Dougados M, Schiff M, Smolen J, 20. Bathon JM, Martin RW, Fleischmann RM, Bingham CO,
et al. Validation of the 28-joint Disease Activity Score (DAS28) Whitmore JB, Eickenhorst T. Disease activity scores using CRP
and European League Against Rheumatism response criteria versus ESR and the relationship between EULAR and ACR
based on C-reactive protein against disease progression in responses in patients with early rheumatoid arthritis. Ann
patients with rheumatoid arthritis, and comparison with the Rheum Dis. 2005;64 Suppl 3:173 [abstract].
DAS28 based on erythrocyte sedimentation rate. Ann Rheum 21. Hensor EM, Emery P, Bingham SJ, Conaghan PG. Discrepancies
Dis. 2009;68:954–60. in categorizing rheumatoid arthritis patients by DAS-28 (ESR)
7. Smolen JS, Breedveld FC, Schiff MH, Kalden JR, Emery P, Eberl and DAS28 (CRP): can they be reduced? Rheumatology.
G, et al. A Simplified Disease Activity Index for rheumatoid 2010;49:1521–9.
arthritis for use in clinical practice. Rheumatology (Oxford). 22. Crowson CS, Rahman MU, Matteson EL. Which measure of
2003;42:244–57. inflammation to use? A comparison of erythrocyte
8. Aletaha D, Nell VP, Stamm T, Uffmann M, Pflugbeil S, Machold sedimentation rate and C-reactive protein measurements
K, et al. Acute phase reactants add little to composite disease from randomized clinical trials of golimumab in rheumatoid
activity indices for rheumatoid arthritis: validation of a arthritis. J Rheumatol. 2009;36:1606–10.
clinical activity score. Arthritis Res Ther. 2005:R796–806. 23. Rintelen B, Sautner J, Haindl PM, Andel I, Maktari A, Leeb BF.
9. Aletaha D, Smolen J. The Simplified Disease Activity Index Comparison of three rheumatoid arthritis disease activity
(SDAI) and the Clinical Disease Activity Index (CDAI): a review scores in clinical routine. Scand J Rheumatol. 2009;38:336–41.