Arthritis Care & Research

Vol. 71, No. 7, July 2019, pp 961–969

DOI 10.1002/acr.23715
© 2018, American College of Rheumatology

Early Self-­Reported Pain in Juvenile Idiopathic Arthritis as

Related to Long-­Term Outcomes: Results From the Nordic
Juvenile Idiopathic Arthritis Cohort Study
Ellen Dalen Arnstad,1 Veronika Rypdal,2 Suvi Peltoniemi,3 Troels Herlin,4 Lillemor Berntson,5 Anders Fasth,6
Susan Nielsen,7 Mia Glerup,4 Maria Ekelund,8 Marek Zak,7 Kristiina Aalto,3 Ellen Nordal,2
Pål Richard Romundstad,9 and Marite Rygg,10 on behalf of the Nordic Study Group of Pediatric Rheumatology

Objective. To study self-­reported pain early in the disease course of juvenile idiopathic arthritis (JIA) as a predictor
of long-­term disease outcomes.
Methods. Consecutive cases of JIA with disease onset from 1997 to 2000 from defined geographical areas of
Norway, Sweden, Finland, and Denmark were prospectively enrolled in this population-­based cohort study. Self-­
reported, disease-­related pain was measured on a 10-­cm visual analog scale (VAS pain). Inclusion criteria were a
baseline visit with a pain score 6 months after disease onset, followed by an 8-­year study visit. Remission was defined
according to Wallace et al (2004) preliminary criteria. Functional disability was measured by the Childhood Health
Assessment Questionnaire and the Child Health Questionnaire Parent Form if the child was age <18 years and by the
Health Assessment Questionnaire if age ≥18 years. Damage was scored using the Juvenile Arthritis Damage Index.
Results. The final study cohort consisted of 243 participants, and 120 participants (49%) had oligoarticular onset.
At baseline, 76% reported a VAS pain score >0 compared to 57% reporting at 8 years. Half of those who reported
baseline pain also reported pain at 8 years but at a lower intensity. Compared to no pain, higher pain intensity at base-
line predicted more pain at 8 years, more functional disability, more damage, and less remission without medication.
Baseline pain predicted more use of disease-­modifying antirheumatic drugs/biologics during the disease course.
Participants with oligoarticular JIA reporting pain at baseline were more likely to develop extended oligoarticular JIA
or other JIA categories with an unfavorable prognosis.
Conclusion. Early self-­reported, disease-­related pain among children and adolescents with JIA is common and
seems to predict persistent pain and unfavorable long-­term disease outcomes.

INTRODUCTION criteria occurring during the first 6 months after disease onset
(5). Among the different categories, and within each category,
Juvenile idiopathic arthritis (JIA) is a diverse chronic disease the disease course and outcome differ markedly (6). Persistent
with onset at age <16 years. This most common rheumatic oligoarticular JIA has the best prognosis of all JIA categories (7).
disease among children is characterized by at least 6 weeks of Extended oligoarticular JIA has a more unfavorable outcome,
continuous arthritis of unknown cause in 1 or more joints (1). similar to that of polyarticular disease (8,9). Predicting outcome
The incidence rate in the Nordic countries is reported to be is challenging, and several studies have focused on associations
approximately 15–22 per 100,000 children (2–4). JIA is a hetero- between long-­term outcome and clinical characteristics and bio-
geneous disorder classified into 7 categories, based on defined markers, such as the nature of joint involvement, the intensity

1
Ellen Dalen Arnstad, MD: Levanger Hospital, Nord-Trøndelag Hospital MD: Uppsala University, Uppsala, and Ryhov County Hospital, Jonkoping,
Trust, Levanger, and Norwegian University of Science and Technology, Sweden; 9Pål Richard Romundstad, MSc, PhD: Norwegian University of
Trondheim, Norway; 2Veronika Rypdal, MD, Ellen Nordal, MD, PhD: Science and Technology, Trondheim, Norway; 10Marite Rygg, MD, PhD:
University Hospital of North Norway and Arctic University of Norway, Norwegian University of Science and Technology and St. Olavs Hospital,
Tromsø, Norway; 3Suvi Peltoniemi, MD, Kristiina Aalto, MD, PhD: Hospital for Trondheim, Norway.
Children and Adolescents, University of Helsinki, Helsinki, Finland; 4Troels No potential conflicts of interest relevant to this article were reported.
Herlin, MD, PhD, Mia Glerup, MD: Aarhus University Hospital, Aarhus, Address correspondence to Ellen Dalen Arnstad, MD, Department of
Denmark; 5Lillemor Berntson, MD, PhD: Uppsala University, Uppsala, Pediatrics, Levanger Hospital, Nord-Trøndelag Hospital Trust, Pb 333, 7601
Sweden; 6Anders Fasth, MD, PhD: University of Gothenburg, Gothenburg, Levanger, Norway. E-mail: ellen.d.arnstad@ntnu.no.
Sweden; 7Susan Nielsen, MD, Marek Zak, MD, PhD: Rigshospitalet, Submitted for publication March 24, 2018; accepted in revised form July
Copenhagen University Hospital, Copenhagen, Denmark; 8Maria Ekelund, 24, 2018.

961
962       | ARNSTAD ET AL

PATIENTS AND METHODS

SIGNIFICANCE & INNOVATIONS
Patients. The Nordic JIA cohort is a population-­ based
• Pain is a frequent symptom, tends to persist, and
affects health-related quality of life for children and cohort study. Consecutive cases of newly diagnosed JIA from
adolescents with juvenile idiopathic arthritis (JIA). defined geographical areas of Norway, Sweden, Finland, and
• In this study, we showed for the first time that Denmark with disease onset from January 1, 1997 to June 30,
an early pain report is associated with long-term 2000, were prospectively included. Disease onset was defined
­nonremission, functional impairment, more use of
as the day the child fulfilled the criteria for active arthritis accord-
disease-modifying antirheumatic drugs/biologics,
and, for those with oligoarticular JIA, development ing to information given by the parents/patient or by a physician.
into extended disease. Participants were included consecutively and as soon as possible
• The study adds to the increasing amount of evi- after the diagnosis was determined. However, the first extensive
dence establishing the importance of pain assess- baseline visit was scheduled for 6 months after disease onset.
ment in routine care of children and adolescents This time point was chosen to enable classification of the disease
with JIA.
into a JIA category, according to the International League Against
Rheumatism Edmonton criteria (5). We have no registration of
onset symptoms in the database. A detailed description of data
of acute-­phase response, and the existence of autoantibodies collection and patient enrollment has been published previously
and genetic variables (10). Because none of those predictors are (2,8). In the current study, participants were included if they had
perfect, and in order to tailor treatment to reach the target of clin- at least a baseline visit 6 months after disease onset with avail-
ical remission, there is a need for more prospective longitudinal able pain scores, and participation in the 8-­year follow-­up visit.
studies to evaluate early predictors using validated and multidi- At both visits, we had data from clinical examinations, disease
mensional measures (10). More patient-­centered ­measurements activity measures, previous and ongoing medication, and dam-
have been needed for assessment of the course and outcome age and remission status, as well as results from blood tests.
of JIA (11). Among the 6 core variables endorsed by the Ameri- Health-­related quality of life was reported by the children or by
can College of Rheumatology (ACR) (12), only the parent/patient their parents.
assessment of overall well-­being can be defined as a patient-­
reported measure. Patients, parents, and clinicians have pointed Measures. Self-­reported, disease-­related pain intensity dur-
to more specific quality-­of-­life measures, and especially pain, as ing the previous week was measured on a 10-­cm VAS for pain
important measures when evaluating the course and outcome (where 0 = no pain and 10 = worst possible pain) by the child
of JIA (11). if age ≥9 years or by the parents if the child was age <9 years.
Pain is a frequent symptom among children and adoles- VAS pain was assessed with the question “How do you rate your/
cents with JIA (13,14). Pain perception is highly subjective, your child’s pain due to your/his or her illness in the past week?”
and different self-­reported measures are used to detect pain As in previous studies on pain in JIA (15,24–26), pain analyses
frequency and intensity (15,16). Pain assessment in young were explored both by categorization of VAS pain (0, >0 to 3, >3
children is especially challenging, because pain reports are to 7, or >7 to 10), and dichotomized into 0 (no pain) or >0. We
dependent on the parents’ assumption of their child’s pain (17). performed subanalyses on VAS pain scores in participants age
Both unidimensional and multidimensional tools are available <9 years and ≥9 years to look for any discrepancies between
for parent and child/adolescent assessment of pain in JIA (16). parent-­and patient-­ reported pain (27). Self-­ reported physical
Among the unidimensional tools, the visual analog scale (VAS) disability questionnaires were the disease-­specific and validated
is a commonly used and validated scoring instrument (18,19). Childhood Health Assessment Questionnaire (C-­HAQ; where 0 =
The pathogenesis of pain in children and adolescents with JIA no difficulty and 3 = unable to do) if the child was age <18 years
is multifactorial, including both biologic and psychosocial fac- (28,29), and the Health Assessment Questionnaire (HAQ; where
tors (16,20). Pain is a distressing symptom, and several studies 0 = no difficulty and 3 = unable to do) if age ≥18 years (30). Chil-
have elucidated the relationship between pain, functional dis- dren age ≥9 years filled out the C-­HAQ, and parents filled out the
ability, and health-­related quality of life (21–24), but informa- questionnaire for those age <9 years.
tion about pain as a predictor of long-­term disease outcome For children age <18 years, the parent form of the generic
is lacking. health-­related quality-­of-­life instrument, the Child Health Question-
In our Nordic population-­based JIA cohort with comprehen- naire Parent Form (CHQ-­PF50, or simply CHQ), was answered by
sive and prospectively sampled data, we have previously studied the parents, yielding a physical summary score and a psychoso-
different aspects of JIA (2,8). In this project, we aimed to study cial summary score (range 0–100, where 0 = worst, with a mean
self-­reported pain early in the disease course and the associa- ± SD score of 50 ± 10) (28,31,32). This instrument is designed
tion with long-­term disease outcomes. to capture the child’s physical and psychosocial well-­ being
JIA LONG-­TERM OUTCOMES AS RELATED TO EARLY SELF-­REPORTED PAIN |      963

i­ndependent of his/her disease, and it is comparable to norm from children age ≥16 years and from their parents if age <16
scores from the general US population. Damage was scored by years.
experienced pediatric rheumatologists using the Juvenile Arthritis
Damage Index (JADI) assessment of articular damage (JADI-­A) Statistical analysis. We used descriptive statistics with
(range 0–72, where 0 = no damage) and extraarticular damage median and interquartile ranges (IQRs) for continuous variables,
(JADI-­E) (range 0–17, where 0 = no damage) (8,32). Damage was and absolute frequency percentage with 95% confidence inter-
defined as either JADI-­A and/or JADI-­E score >0. As in previous vals (95% CIs) for categorical variables. To evaluate the predic-
studies, C-­HAQ/HAQ and JADI scores were dichotomized into 0 tive value of pain at baseline for outcome measures after 8 years
(no disability, no damage) or >0 (8,24,33). Physical and psycho- and medication during the disease course, model-­based absolute
social summary scores of the CHQ were dichotomized into <40 risks were estimated after binominal regression using the post-­
(poor health) or ≥40 (better health) (28,31). Remission was defined estimation command lincom in Stata software, version 14. Sex
according to the preliminary criteria described by Wallace et  al adjustment was weighted 0.7 for girls to mimic the distribution in
(34). Remission status was dichotomized into remission without the population. In additional analyses, we also adjusted for age. To
medication or not in remission without medication (33). The latter estimate absolute risks, we used the mean age at disease onset
included active disease, inactive disease not yet in remission, and of 6.8 years. We used logistic regression to estimate the odds
in remission while taking medication. ratio with 95% CIs using VAS pain as a continuous variable. In
further analyses, we made receiver operator characteristic (ROC)
Ethics approval. Medical research ethics committees curves based on measures of sensitivity and specificity. The area
from each participating country gave their approval accord-
ing to national practice and regulations in accordance with the Table  1.  Clinical characteristics of the juvenile idiopathic arthritis
(JIA) study population*
Declaration of Helsinki. Written informed consent was obtained
Characteristic Total no. Values
Female 243 170 (70)
Included in the Nordic Oligoarticular JIA at onset 243 120 (49)
JIA study
n = 500 Age at disease onset, 243 6.3 (2.9–10.3)
median (IQR) years
Missing at 8-year Age at 8-­year follow-­up, 243 14.9 (11.1–18.5)
follow-up median (IQR) years
n = 60
Disease duration at baseline 243 7 (6–9)
8-year follow-up
visit, median (IQR) months
n = 440
(F = 291, M = 149) Disease duration at 8-­year 243 97 (95–102)
Finnish participants did follow-­up, median (IQR)
not fill in the pain score
months
7 months after
disease onset VAS pain >0 at baseline 243 185 (76)
8-year follow-up n = 138 visit†
(Norway, Sweden,
VAS pain >0 at 8-­year 204 117 (57)
Denmark)
follow-­up†
n = 302 Pain score missing
C-­HAQ/HAQ >0 at 8-­year 207 80 (39)
7 months after
follow-­up
disease onset
n = 59 CHQ PhS <40 at 8-­year 132 25 (19)
8-year follow-up with
pain score 7 months after follow-­up
disease onset CHQ PsS <40 at 8-­year 132 7 (5)
n = 243 follow-­up
JADI >0 at 8-­year follow-­up 203 46 (23)
Not in remission at 8-­year 236 135 (57)
follow-­up‡

All JIA * Values are the number (%) unless indicated otherwise. IQR = in-
Oligoarticular categories except terquartile range; VAS = visual analog scale; C-­HAQ = Childhood
disease oligoarticular Health Assessment Questionnaire (used for age <18 years); HAQ
at onset disease = Health Assessment Questionnaire (used for age ≥18 years); CHQ
n = 120 at onset PhS = Child Health Questionnaire physical summary score (range
n = 123 0–100); CHQ PsS = Child Health Questionnaire psychosocial sum-
mary score (range 0–100); JADI = Juvenile Arthritis Damage Index.
† Self-­reported pain was measured on a 10-­cm VAS pain scale.
Figure  1.  Flow-­ chart of the study population. JIA = juvenile ‡ Not in remission without medication according to the definition
idiopathic arthritis; F = female; M = male. by Wallace et al (ref. 34).
964       | ARNSTAD ET AL

under the curve was calculated with 95% CIs, and with the follow- ­ ssessment, ­including 21 with VAS pain scores >0, and 9 with
a
ing interpretations: an area of 0.5 or lower was considered to be VAS pain scores = 0. Participants excluded from the current
no discrimination, ≥0.7 to <0.8 as acceptable discrimination, ≥0.8 study due to lack of pain data at baseline had a lower median
to <0.9 as excellent discrimination, and ≥0.9 as outstanding dis- age (5.1 versus 6.3 years) and the proportion of males was
crimination (35). Statistical analyses were carried out using STATA slightly higher (39% versus 30%). There was no difference in
software, version 14. the proportion of oligoarticular JIA at onset and remission sta-
tus at 8 years between the included and excluded participants.

RESULTS
Pain scores. More participants reported a VAS pain score
Clinical characteristics of the study group. Of the >0 at the baseline visit (76%) than at 8 years (57%). The mean
500 patients included from the 4 Nordic countries, 440 par- pain intensity score (VAS pain) of those reporting pain was higher
ticipated in the 8-­year follow-­up. Due to lack of baseline pain at baseline, 3.0 (95% CI 2.6, 3.3) than at 8 years, 2.4 (95% CI
scores, all Finnish participants (n = 138) and an additional 59 2.0, 2.8). The distribution of pain intensity scores at the baseline
participants from the other countries were excluded. The final visit and at the 8-­year follow-­up is shown in Supplementary Figure
study population consisted of 243 children (Figure  1) with a 1, ­available on the Arthritis Care & Research web site at http://
median baseline visit at 7 months and a final follow-­up visit onlinelibrary.wiley.com/doi/10.1002/acr.23715/abstract. For partic-
at 97 months (Table 1). Among these participants 70% were ipants age <9 years at baseline, 118 of 159 (74%) had a parent-­
female, 49% had oligoarticular disease, and the median age reported pain score >0 with a mean intensity of 2.7 (95% CI 2.3,
was 6.3 years at disease onset and 14.9 years at follow-­up 3.0), while 67 of 84 participants (79%) age ≥9 years had a patient-­
(Table  1). The diagnostic delay was short, and the median reported pain score >0, with a mean intensity of 3.5 (95% CI 2.9,
interval between disease onset and diagnosis of arthritis by 4.1). Among participants with pain measures both at baseline and
a physician was 50.5 days (IQR 14–101 days). Of these 243 8 years (n = 204), 50% reported a VAS pain score >0 at both visits,
participants, intraarticular glucocorticoid injections had been and 19% reported no pain at both visits (Figure 2). We divided this
given to 91 participants, and for 34 of these the drug had been group into participants ages <9 and ≥9 years at baseline. Partici-
given within the last 3 months of the baseline visit (results not pants age <9 years had parent-­reported pain scores at their first
shown). At this baseline visit, none of the participants were visit and patient-­reported pain scores at their last visit, and 48%
taking biologics, but 20 were taking systemic steroids. Meth- reported a VAS pain score >0 at both visits. Participants age ≥9
otrexate was used by 31 of the participants, and of those, 8 years had only patient-­reported pain scores, and 55% reported a
had cumulative doses ≥100 mg. The 60 participants who did VAS pain score >0 at both visits (results not shown).
not participate in the 8-­year study did not differ significantly in
the proportion of oligoarticular JIA or with respect to sex, and Baseline pain scores and long-­ term outcome
had a median follow-­up of 47 months (range 5–83 months). ­measures. The association between baseline pain scores
At their last registered visit, 30 participants (50%) had a pain subdivided into 4 categories of pain intensity and long-­
­ term

39 VAS pain = 0
19% both at
7 months and
VAS pain >0 8 years
at 7 months n = 39
n = 150

48 102 15
24% 50% 7%

VAS pain >0

at 8 years
n = 117

Figure  2.  Venn diagram demonstrating pain persistency in the Nordic juvenile idiopathic arthritis study cohort. The cohort included 204
participants with pain measures 7 months after disease onset and at the 8-­year follow-­up. Disease-­related pain was measured on a 10-­cm
visual analog scale (VAS pain) (where 0 = no pain and 10 = worst possible pain). A continuous circle represents a VAS pain score >0 at 7 months
and a broken circle represents a VAS pain score >0 at 8 years.
JIA LONG-­TERM OUTCOMES AS RELATED TO EARLY SELF-­REPORTED PAIN |      965

Table 2.  Association between baseline pain report at 7 months after disease onset and outcomes at 8-­year follow-­up in juvenile idiopathic
arthritis*
Baseline VAS pain VAS pain >0† C-­HAQ/ HAQ >0 CHQ PhS <40 JADI >0 Not in remission‡
0§ 15/54, 9/54, 3/32, 4/53, 14/55,
28 (16, 40) 18 (8, 28) 11 (1, 21) 8 (1, 15) 26 (14, 37)
>0–3§ 53/86, 31/86, 8/62, 25/83, 61/95,
62 (51, 72) 35 (25, 45) 12 (4, 21) 30 (20, 40) 64 (54, 74)
>3–7§ 44/58, 34/60, 14/38, 15/60, 52/73,
74 (63, 85) 56 (44, 69) 37 (22, 52) 25 (14, 36) 71 (61, 81)
>7–10§ 5/6, 6/7, 0 2/7, 8/13,
90 (66, 114) 94 (80, 107) 0 28 (–5, 61) 61 (35, 88)
Continuous, OR (95% CI)⁋ 1.5 (1.2, 1.7) 1.4 (1.2, 1.6) 1.4 (1.1, 1.7) 1.1 (0.9, 1.2) 1.2 (1.1, 1.4)

* Values are the number/total number, percentage (95% confidence interval [95% CI]) unless indicated otherwise. VAS = visual analog
scale; C-­HAQ = Childhood Health Assessment Questionnaire (used for age <18 years); HAQ = Health Assessment Questionnaire (used
for age ≥18 years); CHQ PhS = Child Health Questionnaire physical summary score (range 0–100); JADI = Juvenile Arthritis Damage Index;
OR = odds ratio.
† Self-­reported pain was measured on a 10-­cm VAS pain scale.
‡ Not in remission without medication according to the definition by Wallace et al (ref. 34).
§ Self-­reported pain was measured on a 10-­cm VAS pain scale, adjusted for sex, weighted 0.7 for girls.
⁋ Self-­reported pain was measured on a 10-­cm VAS pain scale, analyzed with VAS pain as a continuous variable, adjusted for sex, weight-
ed 0.7 for girls.

­ utcome measures is shown in Table  2. Participants report-

o 8-­year follow-­up. A distinct dose-­response curve was observed
ing a VAS pain score >0 at the baseline visit more frequently with increasing pain intensity at baseline. Using VAS pain as a
reported pain and functional disability (C-­HAQ/HAQ >0) at the continuous variable, we observed an increased odds ratio for

A B

C D

Figure 3.  Receiver operator characteristic curves in the Nordic juvenile idiopathic arthritis study cohort for different disease outcomes after 8
years compared to self-­reported disease-­related pain at 7 months after disease onset, measured on a 10-­cm visual analog scale (VAS; where 0 = no
pain and 10 = worst possible pain). Remission was defined according to the preliminary criteria described by Wallace et al (34). Functional disability
was measured with the Childhood Health Assessment Questionnaire/Health Assessment Questionnaire. Damage was measured with the Juvenile
Arthritis Damage Index, articular and extraarticular. The area under the curve (AUC) values were 0.74 (95% CI 0.67, 0.80) for persistent pain, 0.68
(95% CI 0.61, 0.75) for not being in remission, 0.71 (95% CI 0.64, 0.79) for functional disability, and 0.58 (95% CI 0.50, 0.67) for joint damage.
A, Persistent pain. B, Not in remission. C, Functional disability. D, Damage.
966       | ARNSTAD ET AL

Table 3.  Association between baseline pain report at 7 months after disease onset and outcomes at 8-­year follow-­up in juvenile idiopathic
arthritis with oligoarticular onset*
C-­HAQ/ Not in
Baseline VAS pain Extend oligo/others† VAS pain >0‡ HAQ >0 JADI >0 remission§

0⁋ 12/40, 10/37, 7/37, 3/36, 11/38,

30 (16, 44)# 26 (12, 40) 16 (7, 26) 7 (0, 15) 29 (14, 43)
>0⁋ 39/80, 40/65, 26/66, 19/65, 51/78,
48 (38, 60)** 61 (50, 73) 39 (28, 51) 29 (18, 40) 65 (55, 76)
Continuous, 1.3 (1.1, 1.6) 1.4 (1.1, 1.8) 1.4 (1.1, 1.8) 1.0 (0.8, 1.3) 1.3 (1.0, 1.5)
OR (95% CI)††
* Values are the number/total number, percentage (95% confidence interval [95% CI]) unless indicated otherwise. VAS = visual analog scale;
Extend oligo = extended oligoarticular juvenile idiopathic arthritis (JIA); C-­HAQ = Childhood Health Assessment Questionnaire (used for age
<18 years); HAQ = Health Assessment Questionnaire (used for age ≥18 years); JADI = Juvenile Arthritis Damage Index; OR = odds ratio.
† Oligoarticular JIA at 6 months changed to either extended oligoarticular or other JIA categories at the 8-­year follow-­up.
‡ Self-­reported pain was measured on a 10-­cm VAS pain scale.
§ Not in remission without medication according to the definition by Wallace et al (ref. 34).
⁋ Self-­reported pain was measured on a 10-­cm VAS pain scale, adjusted for sex, weighted 0.7 for girls.
# Others were 1 with enthesitis-­associated arthritis and 2 with undifferentiated arthritis.
** Others were 2 with psoriatic arthritis, 6 with enthesitis-­associated arthritis, and 2 with undifferentiated arthritis.
†† Self-­reported pain was measured on a 10-­cm VAS pain scale, analyzed with VAS pain as a continuous variable, adjusted for sex, weighted
0.7 for girls.

the ­different long-­term outcomes. Functional disability as pre- (95% CI 16, 44) of those reporting no pain (Table 3). Also, a higher
sented by the CHQ physical summary score demonstrated proportion of participants with a VAS pain score >0 at baseline
similar results. Participants reporting pain at baseline more fre- was not in remission without medication, 65% (95% CI 55, 76),
quently were not in remission without medication at follow-­up, and reported pain, 61% (95% CI 50, 73) at 8 years, compared
compared to those reporting no baseline pain. A similar associa- to those reporting no pain at baseline. In contrast, 70% (95%
tion between increasing pain intensity at baseline and long-­term CI 56, 84) of those reporting no pain at baseline remained in the
remission status was observed, but the dose-­response relation- persistent oligoarticular JIA category at the 8-­year follow-­up. The
ship tended to level out at the most extreme pain intensities. associations were strengthened using VAS pain as a continu-
Participants reporting no pain at baseline rarely reported pain ous variable. In all analyses, we adjusted for sex, and additional
(28%) and functional disability (18%) at 8 years, and 74% were adjustment for age did not change the results.
in remission without medication. In all analyses, we adjusted
for sex, and additional adjustment for age did not change the
DISCUSSION
results. Similar to the results on long-­term remission, pain, and
functional disability, baseline pain was associated with the use Among participants in the population-­based Nordic JIA study,
of disease-­modifying antirheumatic drugs (DMARDs) and bio- a higher proportion reported disease-­specific pain 7 months after
logics during the 8-­ year disease course (see Supplementary disease onset, and the pain was of higher intensity, compared
Table 1, available on the Arthritis Care & Research web site at to the 8-­year follow-­up. Half of the participants reporting pain at
http://onlinelibrary.wiley.com/doi/10.1002/acr.23715/abstract). baseline also reported pain at 8 years. Self-­reported pain early
In contrast to the other 8-­year outcome measures, psychosocial in the disease course predicted more pain, more functional dis-
health assessed with the CHQ psychosocial summary score did ability, more damage, more use of DMARDs/biologics, and more
not show an association with the baseline pain score (results not long-­
­ term disease activity at 8 years. In addition, participants
shown). The predictive ability of baseline pain was also analyzed with oligoarticular JIA and a VAS pain score >0 at baseline more
using ROC curves, giving acceptable discrimination between often developed an extended disease or other unfavorable JIA
early pain scores and long-­term outcomes of pain, functional categories.
disability, and not being in remission at 8 years, but no clear The strength of our study is the longitudinal and
discrimination was observed for long-­term damage (Figure 3). population-­based design, a robust international cohort, and
the use of validated and multidimensional outcome meas-
Baseline pain scores and long-­term outcomes in the ures. The n ­ ovelty in looking at associations between early
oligoarticular category. Among participants with ­oligoarticular pain report and long-­term remission status, medication, and
JIA reporting a VAS pain score >0 at the baseline visit, 48% (95% development into more unfavorable disease categories is also
CI 38, 60) developed extended oligoarticular disease or other JIA a major strength. Some limitations must be recognized. The
categories during the course of the disease compared to 30% exclusion of all the Finnish participants reduced the number
JIA LONG-­TERM OUTCOMES AS RELATED TO EARLY SELF-­REPORTED PAIN |      967

of study participants but did not change the population-­based Also, to compare studies on pain, age and disease duration
design of the study. The missing early pain scores from the must be taken into account.
other countries might have skewed the remaining cohort, but Even in the biologic era with generally good disease control,
the distribution of JIA categories and the remission status were persistent pain during the course of the disease remains a con-
comparable among those with or without early pain scores. cern (26,40,41). Half of our participants reporting pain at 7 months
Even though the VAS pain instrument is disease-­specific, we after disease onset also reported pain at the 8-­year follow-­up,
cannot rule out that other musculoskeletal co-­conditions, such indicating high pain persistency. This finding is in agreement with
as generalized joint hypermobility, specific onset symptoms, results from other studies showing that a significant number of
or differences in timing of diagnosis, might have influenced children and adolescents with JIA continue to report pain during
the child’s/parent’s pain rating. However, these possibilities the course of disease and into adulthood (14,24,42). Notably, the
are a challenge to all pain research. Since pain is a subjec- proportion of pain persistence is fairly similar, whether the parents
tive descriptor, our cohort is close to population-­based, and report their child’s pain, or whether the pain is self-­reported at
because trained pediatric rheumatologists ascertained the JIA baseline. Pain persistence despite a seemingly good treatment
diagnosis, we do not think that these possibilities will seriously response supports theories that the causes of pain are multifac-
disturb the interpretation of our results. Similarly, we cannot torial (41,43). Both psychosocial and biologic factors contribute to
ascertain the nature of bodily pain that the participants scored these children’s subjective experience of pain (38,44,45).
when filling out this question in the CHQ questionnaire. How- Pain as a predictor of unfavorable health-­related quality of
ever, we only used this question in accordance with the CHQ life in children with JIA is widely studied (21,23,46,47). In a mul-
instructions, as one of many items describing a summary of tinational quality-­ of-­
life study from the Pediatric Rheumatology
physical function, and not as a pain measure. Parents report- International Trials Organization, pain was found to be a predictor
ing their child’s pain for children age <9 years constituted a of psychosocial well-­being (21). In agreement with our study, pre-
majority of pain reports at the baseline visit, but a small minor- vious studies have shown that pain at presentation was a strong
ity at the 8-­year follow-­up. We cannot rule out some element predictor of persistent pain (42,48). In accordance with our results,
of parent/child d­ iscordance, although the subanalyses on pain pain as a predictor of functional disability was also found in a small
reports according to age at baseline seem to indicate that dis- cross-­sectional study from the US (49). Except for health-­related
cordance was not a major problem. This result is in accord- quality-­of-­life outcomes and functional disability, studies that spe-
ance with a study from 2006 showing moderate agreement cifically address pain as a predictor of other long-­term outcomes,
between parent’s and child’s pain rating (17). Our results are such as remission, damage, medication, and changing of JIA
not directly comparable, because the parent/child pain reports categories, are lacking. Our results demonstrate for the first time
are not from the same visit. The early pain scores from the that early pain is associated with not achieving remission without
baseline study visit 7 months after disease onset were given medication in a long-­term perspective. We also demonstrate, for
by participants both while taking and not taking medication, the first time, that early pain reports predict a higher risk of devel-
but only a few had started DMARDs. opment into extended oligoarticular or other unfavorable JIA cate-
Consistent with previous research, we found pain as a gories during the course of the disease. This finding suggests that
frequent symptom among the participants in our study cohort early pain may be an indicator of subclinical disease activity or a
(14,16,36). We found a reduction in the number of participants marker of a more severe disease category. This possibility is also
reporting pain from baseline (76%) to the 8-­year follow-­up (57%). supported by the fact that a higher proportion of participants with
A quite similar reduction was found by Lovell and Walco (37) in an early pain report used DMARDs/biologics during the course of
1989, demonstrating pain frequency of 60% at baseline, 50% at the disease.
1-­year follow-­up, and 40% at 5-­year follow-­up. In a recent 30-­ Even though pain assessment has been highlighted as a
year follow-­up study of JIA in Norway, 66% of the participants quality measure of pediatric arthritis care (50), pain scores are
reported pain of some degree (24). infrequently used as guiding tools in daily care of these patients
In accordance with other studies, the intensity of pain was (41). Our results demonstrate that pain in children with JIA at an
mainly in the mild-­to-­moderate range (19,37,38). Our results early stage in their disease should be taken seriously, not just to
on early pain intensity with a mean VAS pain score of 3.0 are relieve ongoing discomfort, but probably also as a sign of ongoing
consistent with a recent cross-­sectional study in children and clinical or subclinical disease activity. This necessity emphasizes
adolescents with JIA from the southeastern region of the US, the importance of pain assessment in routine care of children and
showing a mean VAS pain score of 2.6 (13). Our results on adolescents with JIA. In a Canadian study where patients, par-
pain intensity at the 8-­year follow-­up appear to be lower com- ents, and clinicians were asked what matters most in the care of
pared to other studies (14,39). Those studies are, however, JIA, pain was 1 of the 5 most important factors (11). The active
skewed to the severe end of the JIA spectrum, whereas our joint count was the only 1 of these 5 factors that is included in
population-­based study included the full disease spectrum. the pediatric version of the ACR core variables for clinical care in
968       | ARNSTAD ET AL

children with JIA (12). The association between early pain reports study revealing early predictors and outcome after 14.9 years. J
and long-­term unfavorable outcome adds to the discussion on Rheumatol 2003;30:386–93.
7. Ravelli A. Toward an understanding of the long-­term outcome of
the validity of the ACR core variables, and on whether pain should
­juvenile idiopathic arthritis. Clin Exp Rheumatol 2004;22:271–5.
be included in these variables. In conclusion, early self-­reported 8. Nordal E, Zak M, Aalto K, Berntson L, Fasth A, Herlin T, et al.
pain in JIA is common, tends to persist, and seems to predict Ongoing disease activity and changing categories in a long-­term
unfavorable long-­
­ term disease outcome in several outcome Nordic cohort study of juvenile idiopathic arthritis. Arthritis Rheum
2011;63:2809–18.
dimensions.
9. Selvaag AM, Aulie HA, Lilleby V, Flato B. Disease progression into
adulthood and predictors of long-­ term active disease in juvenile
idiopathic arthritis. Ann Rheum Dis 2016;75:190–5.
ACKNOWLEDGMENTS
10. Van Dijkhuizen EH, Wulffraat NM. Early predictors of prognosis in
The authors thank the participants and their parents for juvenile idiopathic arthritis: a systematic literature review. Ann Rheum
Dis 2015;74:1996–2005.
making this study possible. We also thank the other members of
11. Guzman J, Gomez-Ramirez O, Jurencak R, Shiff NJ, Berard RA,
the Nordic Study Group of Pediatric Rheumatology: Gudmund Duffy CM, et al. What matters most for patients, parents, and
Marhaug (Trondheim), Boel Anderson-­ Gäre (Jonköping), Freddy clinicians in the course of juvenile idiopathic arthritis? A qualitative
Karup Pedersen (Copenhagen), and Pekka Lahdenne and Pirkko study. J Rheumatol 2014;41:2260–9.
12. Giannini EH, Ruperto N, Ravelli A, Lovell DJ, Felson DT, Martini A.
Pelkonen (Helsinki) for their inspiring cooperation. We are also
Preliminary definition of improvement in juvenile arthritis. Arthritis
grateful to local participating physicians for collecting data: Nina Moe Rheum 1997;40:1202–9.
and Kjell Berntzen (Trondheim), Astri Lang, Nils Thomas Songstad, 13. Bromberg MH, Connelly M, Anthony KK, Gil KM, Schanberg
and Anne Elisabeth Ross (Tromsø), Eric Ronge (Skövde), Mikael LE. Self-­ reported pain and disease symptoms persist in juvenile
idiopathic arthritis despite treatment advances: an electronic diary
Damgaard and Nils Olof Jonsson (Jönköping), Anna-­Lena Nilsson
study. Arthritis Rheumatol 2014;66:462–9.
(Östersund), Anders Berner and Hans Ekström (Karlstad), Agne Lind 14. Schanberg LE, Anthony KK, Gil KM, Maurin EC. Daily pain and
and Lars Hammarén (Borås), and Johan Robinsson (Trollhättan). symptoms in children with polyarticular arthritis. Arthritis Rheum
2003;48:1390–7.
15. Tupper SM, Rosenberg AM, Pahwa P, Stinson JN. Pain intensity
AUTHOR CONTRIBUTIONS variability and its relationship with quality of life in youths with juvenile
All authors were involved in drafting the article or revising it critically idiopathic arthritis. Arthritis Care Res (Hoboken) 2013;65:563–70.
for important intellectual content, and all authors approved the final version 16. Weiss JE, Luca NJ, Boneparth A, Stinson J. Assessment and
to be submitted for publication. Dr. Arnstad had full access to all of the management of pain in juvenile idiopathic arthritis. Paediatr Drugs
data in the study and takes responsibility for the integrity of the data and 2014;16:473–81.
the accuracy of the data analysis. 17. Garcia-Munitis P, Bandeira M, Pistorio A, Magni-Manzoni S,
Study conception and design. Herlin, Berntson, Fasth, Nielsen, Zak, Ruperto  N, Schivo A, et al. Level of agreement between children,
Aalto, Nordal, Rygg. parents, and physicians in rating pain intensity in juvenile idiopathic
Acquisition of data. Peltoniemi, Herlin, Berntson, Nielsen, Ekelund, Zak, arthritis. Arthritis Rheum 2006;55:177–83.
Aalto, Nordal, Rygg. 18. Stinson JN, Kavanagh T, Yamada J, Gill N, Stevens B. Systematic
Analysis and interpretation of data. Arnstad, Rypdal, Glerup, Nordal, review of the psychometric properties, interpretability and feasibility
Romundstad, Rygg. of self-­report pain intensity measures for use in clinical trials in
children and adolescents. Pain 2006;125:143–57.
REFERENCES 19. Shiff NJ, Tupper S, Oen K, Guzman J, Lim H, Lee CH, et al.
Trajectories of pain severity in juvenile idiopathic arthritis: results
1. Ravelli A, Martini A. Juvenile idiopathic arthritis. Lancet from the Research in Arthritis in Canadian Children Emphasizing
2007;369:767–78. Outcomes cohort. Pain 2018;159:57–66.
2. Berntson L, Andersson Gäre B, Fasth A, Herlin T, Kristinsson J, 20. Stinson JN, Luca NJ, Jibb LA. Assessment and management of pain
Lahdenne P, et al. Incidence of juvenile idiopathic arthritis in the in juvenile idiopathic arthritis. Pain Res Manag 2012;17:391–6.
Nordic countries: a population based study with special reference 21. Gutierrez-Suarez R, Pistorio A, Cespedes Cruz A, Norambuena X,
to the validity of the ILAR and EULAR criteria. J Rheumatol Flato B, Rumba I, et al. Health-­related quality of life of patients with
2003;30:2275–82. juvenile idiopathic arthritis coming from 3 different geographic areas:
3. Riise OR, Handeland KS, Cvancarova M, Wathne KO, Nakstad B, the PRINTO multinational quality of life cohort study. Rheumatology
Abrahamsen TG, et al. Incidence and characteristics of arthritis (Oxford) 2007;46:314–20.
in Norwegian children: a population-­ based study. Pediatrics 22. Shaw KL, Southwood TR, Duffy CM, McDonagh JE. Health-­related
2008;121:e299–306. quality of life in adolescents with juvenile idiopathic arthritis. Arthritis
4. Moe N, Rygg M. Epidemiology of juvenile chronic arthritis in northern Rheum 2006;55:199–207.
Norway: a ten-­ year retrospective study. Clin Exp Rheumatol 23. Klotsche J, Minden K, Thon A, Ganser G, Urban A, Horneff G.
1998;16:99–101. Improvement in health-­related quality of life for children with juvenile
5. Petty RE, Southwood TR, Manners P, Baum J, Glass DN, idiopathic arthritis after start of treatment with etanercept. Arthritis
Goldenberg J, et al. International League of Associations for Care Res (Hoboken) 2014;66:253–62.
Rheumatology classification of juvenile idiopathic arthritis: second 24. Tollisen A, Selvaag AM, Aulie HA, Lilleby V, Aasland A, Lerdal A, et al.
revision, Edmonton, 2001. J Rheumatol 2004;31:390–2. Physical functioning, pain, and health-­related quality of life in adults
6. Flato B, Lien G, Smerdel A, Vinje O, Dale K, Johnston V, et al. with juvenile idiopathic arthritis: a longitudinal 30-­year followup study.
Prognostic factors in juvenile rheumatoid arthritis: a case-­control Arthritis Care Res (Hoboken) 2018;70:741–9.
JIA LONG-­TERM OUTCOMES AS RELATED TO EARLY SELF-­REPORTED PAIN |      969

25. Consolaro A, Vitale R, Pistorio A, Lattanzi B, Ruperto N, Malattia C, 39. Guzman J, Henrey A, Loughin T, Berard RA, Shiff NJ, Jurencak R,
et al. Physicians’ and parents’ ratings of inactive disease are frequently et al. Predicting which children with juvenile idiopathic arthritis will
discordant in juvenile idiopathic arthritis. J Rheumatol 2007;34:1773–6. have a severe disease course: results from the ReACCh-­Out cohort.
26. Lomholt JJ, Thastum M, Herlin T. Pain experience in children with J Rheumatol 2017;44:230–40.
juvenile idiopathic arthritis treated with anti-­ TNF agents compared 40. Anink J, Prince FH, Dijkstra M, Otten MH, Twilt M, ten Cate R,
to non-­ biologic standard treatment. Pediatr Rheumatol Online J et al. Long-­term quality of life and functional outcome of patients
2013;11:21. with juvenile idiopathic arthritis in the biologic era: a longitudinal
27. Lal SD, McDonagh J, Baildam E, Wedderburn LR, Gardner-Medwin follow-­up study in the Dutch Arthritis and Biologicals in Children
J, Foster HE, et al. Agreement between proxy and adolescent Register. Rheumatology (Oxford) 2015;54:1964–9.
assessment of disability, pain, and well-­being in juvenile idiopathic 41. Giancane G, Alongi A, Rosina S, Calandra S, Consolaro A, Ravelli
arthritis. J Pediatr 2011;158:307–12. A. Open issues in the assessment and management of pain in
28. Ruperto N, Ravelli A, Pistorio A, Malattia C, Cavuto S, Gado-West juvenile idiopathic arthritis. Clin Exp Rheumatol 2017;35 Suppl
L, et al. Cross-­cultural adaptation and psychometric evaluation of 107:123–6.
the Childhood Health Assessment Questionnaire (CHAQ) and the 42. Thastum M, Herlin T. Pain-­ specific beliefs and pain experience
Child Health Questionnaire (CHQ) in 32 countries: review of the in children with juvenile idiopathic arthritis: a longitudinal study.
general methodology. Clin Exp Rheumatol 2001;19 Suppl 23:S1–9. J Rheumatol 2011;38:155–60.
29. Singh G, Athreya BH, Fries JF, Goldsmith DP. Measurement of health 43. Malleson PN, Oen K, Cabral DA, Petty RE, Rosenberg AM, Cheang
status in children with juvenile rheumatoid arthritis. Arthritis Rheum M. Predictors of pain in children with established juvenile rheumatoid
1994;37:1761–9. arthritis. Arthritis Rheum 2004;51:222–7.
30. Fries JF, Spitz P, Kraines RG, Holman HR. Measurement of patient 44. Thastum M, Herlin T, Zachariae R. Relationship of pain-­ coping
outcome in arthritis. Arthritis Rheum 1980;23:137–45. strategies and pain-­specific beliefs to pain experience in children
31. Landgraf JM, Abetz L, Ware JE. Child Health Questionnaire (CHQ): with juvenile idiopathic arthritis. Arthritis Rheum 2005;53:178–84.
a user’s manual. Boston: Health Institute, New England Medical 45. La Hausse de Lalouviere L, Ioannou Y, Fitzgerald M. Neural
Center; 1996. mechanisms underlying the pain of juvenile idiopathic arthritis. Nat
32. Viola S, Felici E, Magni-Manzoni S, Pistorio A, Buoncompagni A, Rev Rheumatol 2014;10:205–11.
Ruperto N, et al. Development and validation of a clinical index for 46. Oen K, Guzman J, Dufault B, Tucker LB, Shiff NJ, Duffy KW, et al.
assessment of long-­ term damage in juvenile idiopathic arthritis. Health-­related quality of life in an inception cohort of children with
­Arthritis Rheum 2005;52:2092–102. juvenile idiopathic arthritis: a longitudinal analysis. Arthritis Care Res
33. Bertilsson L, Andersson-Gare B, Fasth A, Petersson IF, Forsblad- (Hoboken) 2018;70:134–44.
D’elia H. Disease course, outcome, and predictors of outcome in a 47. Haverman L, Grootenhuis MA, van den Berg JM, van Veenendaal
population-­based juvenile chronic arthritis cohort followed for 17 years. M, Dolman KM, Swart JF, et al. Predictors of health-­related quality
J Rheumatol 2013;40:715–24. of life in children and adolescents with juvenile idiopathic arthritis:
34. Wallace CA, Ruperto N, Giannini E. Preliminary criteria for clinical results from a web-­ based survey. Arthritis Care Res (Hoboken)
remission for select categories of juvenile idiopathic arthritis. J 2012;64:694–703.
Rheumatol 2004;31:2290–4. 48. Rashid A, Cordingley L, Carrasco R, Foster HE, Baildam EM, Chieng
35. David W. Hosmer, Stanley Leweshow. Area under the ROC curve. A, et al. Patterns of pain over time among children with juvenile
In: Cressie NA, editor. Applied logistic regression. 2nd ed. New York: idiopathic arthritis. Arch Dis Child 2018;103:437–43.
John Wiley; 2000. p. 160–4. 49. Feinstein AB, Forman EM, Masuda A, Cohen LL, Herbert JD,
36. Oen K, Reed M, Malleson PN, Cabral DA, Petty RE, Rosenberg AM, Moorthy LN, et al. Pain intensity, psychological inflexibility, and
et al. Radiologic outcome and its relationship to functional disability acceptance of pain as predictors of functioning in adolescents with
in juvenile rheumatoid arthritis. J Rheumatol 2003;30:832–40. juvenile idiopathic arthritis: a preliminary investigation. J Clin Psychol
37. Lovell DJ, Walco GA. Pain associated with juvenile rheumatoid Med Settings 2011;18:291–8.
arthritis. Pediatr Clin North Am 1989;36:1015–27.
50. Lovell DJ, Passo MH, Beukelman T, Bowyer SL, Gottlieb BS,
38. Schanberg LE, Lefebvre JC, Keefe FJ, Kredich DW, Gil KM. Pain Henrickson M, et al. Measuring process of arthritis care: a proposed
coping and the pain experience in children with juvenile chronic set of quality measures for the process of care in juvenile idiopathic
­arthritis. Pain 1997;73:181–9. arthritis. Arthritis Care Res (Hoboken) 2011;63:10–6.