1. Top of page
  2. Abstract


To assess the frequency of sacroiliitis and the radiographic and clinical outcome in juvenile idiopathic arthritis (JIA) and determine patient characteristics, early disease variables, and genetic markers that predict development of sacroiliitis.


We performed a retrospective cohort study of 314 (79%) of the 400 JIA patients first admitted to the hospital between 1980 and 1985. The participants were examined after a median disease duration of 14.9 years (range 11.7–25.1). Radiographs of the sacroiliac joints, hips, ankles, and tarsi were obtained and studied in a blinded manner by 2 radiologists. The presence of HLA–DRB1 and DPB1 alleles was determined by genotyping and that of HLA–B27 by serologic testing. Variables relating to the onset and course of the disease were obtained by chart reviews.


Twenty (6%) of the JIA patients developed radiographic sacroiliitis according to the New York criteria. In 9 patients (45%), sacroiliitis had not been demonstrated before the followup examination. At followup, spinal flexion (lateral and anterior) was reduced in 70–75% of patients with sacroiliitis and in 30–35% of those without sacroiliitis. Compared with the JIA patients without sacroiliitis, those with sacroiliitis more frequently had inflammatory back pain, enthesitis, radiographic changes in the hips and calcanei, erosions of any peripheral joint, and uveitis.

Predictors of sacroiliitis were HLA–B27, absence of DPB1*02, hip joint involvement within the first 6 months, and disease onset after age 8 years. The following factors were more common among patients in whom sacroiliitis developed than in other JIA patients: DRB1*04, male sex, family history of ankylosing spondylitis, psoriasis, inflammatory back pain, and enthesitis within the first 6 months.


In the current study, radiographically evident sacroiliitis had developed in 6% of JIA patients after a median disease duration of 14.9 years. HLA–B27, absence of DPB1*02, late onset of disease, and early hip involvement were predictors of sacroiliitis.

Juvenile idiopathic arthritis (JIA) is the most frequently occurring inflammatory rheumatic condition in childhood, with an annual incidence of 10–20 cases per 100,000 children and an estimated prevalence of 90–130 cases per 100,000 children (1–4). The term JIA refers to a heterogeneous group of patients with different characteristics, outcomes, extraarticular manifestations, serologic findings, and genetic markers (5–7). Juvenile rheumatoid arthritis (JRA) and juvenile spondylarthropathy are the 2 major subcategories of JIA. Sacroiliitis is considered to be the gold standard for a diagnosis of spondylarthropathy and is characterized by clinical and radiographic signs of arthritis in the sacroiliac joints (8–10). However, many years may elapse between the time of onset of arthritis and the onset of axial involvement in children, and identification of sacroiliitis in patients with JIA requires long-term followup.

The annual incidence of juvenile spondylarthropathies is 1–69 per 100,000 children (1). The wide variation in reported frequency may be partly attributable to differences in the classification of patients. Criteria for the classification of ankylosing spondylitis (AS) in adults have been used to classify juvenile AS in persons in whom symptoms develop before age 16 years (10). However, the frequency of peripheral arthritis is higher, and symptoms and signs of spinal and sacroiliac involvement are fewer in patients with juvenile-onset AS than in those with adult-onset AS (8, 11).

Criteria have been developed to identify conditions related to juvenile AS, such as juvenile psoriatic arthritis (PsA), syndrome of seronegative enthesopathy and arthropathy (SEA), and enthesitis-related arthritis (12–14). Nonetheless, early manifestations of an evolving spondylarthropathy may be incorrectly classified as JRA (8, 15). Furthermore, accurate classification of childhood arthritides depends on the identification of clusters of characteristics, clinical features, and laboratory findings. Identification of clinical features (e.g., numbers of affected joints and the presence of enthesitis) is subjective and varies between raters (16). Therefore, use of objective measures such as radiographs and immunogenetic markers is recommended for the evaluation of classification criteria and the identification of homogeneous subtypes of JIA (17).

Radiographic assessments of sacroiliac joints have been performed in patients with selected subtypes of JIA (8, 9, 18, 19). To our knowledge, the frequency of radiographic sacroiliitis in JIA in general has not yet been assessed.

Juvenile-onset AS, like adult-onset AS, is strongly associated with the presence of HLA–B27. Furthermore, an increased prevalence of HLA-DRB1*08 in Norwegian and Mexican juvenile AS patients has been reported (20, 21), indicating that HLA-encoded genes other than B27 may also contribute to the development of sacroiliitis in children.

Male sex, disease onset in late childhood, pauciarticular subtype, enthesitis, and a family history of related disease all have been associated with juvenile spondylarthropathy (5, 13, 22, 23). However, the relative importance of patient characteristics, genetic markers, and early disease variables in the development of sacroiliitis has not yet been assessed.

Two previous studies compared the disease course in children with juvenile spondylarthropathies with that in children with JRA (5, 23). The results demonstrated differences between the 2 subgroups in joint patterns and the frequency of axial signs and symptoms. Data on the long-term outcome of patients with juvenile spondylarthropathy are scarce (18).

In the current study, we used radiographically evident sacroiliitis as an objective measure of long-term outcome in a cohort of patients with JIA who were followed up for a median of 14.9 years after disease onset. It was estimated that during this period, ∼70% of the Norwegian children with diagnosed JIA were admitted to the hospital (24). The aim of our study was to assess the frequency of sacroiliitis in JIA and to determine the role of genetic markers, patient characteristics, and early-disease variables in the development of sacroiliitis. We also sought to describe the clinical and radiographic outcome in JIA patients with and those without sacroiliitis.


  1. Top of page
  2. Abstract

Patients and controls

Between January 1980 and September 1985, 400 children with JIA were admitted for the first time to the Centre for Rheumatic Diseases, Rikshospitalet University Hospital. Three hundred fourteen (79%) of these patients were examined after a median disease duration of 14.9 years (range 11.7–25.1), and their medical records were reviewed. Rikshospitalet University Hospital has the only pediatric rheumatology clinic in Norway and thus serves the entire country (24). JIA was diagnosed before admission in 43 patients (14%), and the diagnosis was established at or after the time of the first visit in the remaining 271 patients (86%). The median duration of arthritis before admission was 2.7 months (range 0.1–92.6).

Eighty-six (21%) of the 400 patients first admitted to the hospital were not included in the study. Eight patients (2%) had died by the time of followup, 56 (14%) could not be found or chose not to participate, and 22 (5%) were excluded from the study because of missing data relating to genomic typing (n = 9) or radiographic assessments (n = 13). Patients who were not included were comparable with participants with regard to sex, type of disease at onset, age at first admission, and duration of symptoms before admission.

Two hundred ninety-five unrelated healthy individuals, randomly selected from the Norwegian Bone Marrow Donor Registry, were used as controls for the genomic typing. The study was approved by the Regional Ethics Committee for Medical Research.


JIA was defined as meeting the criteria for either JRA (14), juvenile AS (9), SEA (12), juvenile PsA (11), or arthritis associated with inflammatory bowel disease (IBD). Juvenile spondylarthropathy was defined as meeting the criteria for juvenile AS, SEA, juvenile PsA, or IBD. Onset of disease was defined as the date at which arthritis or systemic features were documented by a physician.

Clinical examination and chart reviews

The patients were examined by one of the authors (BF, GL, OV, or DS) at followup. Prior to the study, these 4 physicians had independently examined the same 8 JIA patients. Variations between the raters were assessed, followed by a meeting during which consensus on the clinical findings of each patient was achieved. Information about the onset and course of disease was obtained by reviewing medical records and conducting interviews. Information about 1–2 clinical variables relating to disease onset were missing for 6 patients. The examiners were blinded to information about the radiographic examinations at followup.

The clinical examination included an assessment of actively involved (swollen or tender and mobility-restricted) and affected (swollen or mobility-restricted) joints. Anterior lumbar flexion was assessed using the modified Schober method (25). Reduced lumbar flexion was defined as values ≤6.5 cm in boys and ≤5.5 cm in girls, based on the mean values (minus 2 SD) obtained in a study of healthy Mexican adolescents (26). Lateral dorsolumbar flexion was measured as the difference between the fingertips-to-floor distance in the upright position and at maximum lateral flexion (27). A difference in the fingertips-to-floor distance of ≤10% of body height was considered abnormal (27). A protractor was used to assess cervical rotation (range 0–180°). Reduced cervical rotation was defined as values ≤160°. Chest expansion was assessed by the standard circumferential measurement (28). Enthesitis was defined as discretely localized tenderness at the sites of insertion of ligaments, tendons, and fascia to bone (13); this definition was restricted to patients without tenderness over adjacent structures or multiple tender points. Hip joint involvement was defined as restricted mobility with or without pain or tenderness (15). A family history of AS was defined as the presence of AS in at least 1 first- or second-degree relative. Inflammatory back pain at followup was defined as low back pain of at least 3 months duration that improved with exercise and was not relieved by rest (10). Inflammatory back pain at disease onset was defined as pain in the spine at rest, with morning stiffness that improved with movement (14). IgM rheumatoid factor (IgM-RF) positivity was defined as titers ≥1:64, as measured at least twice by the Rose-Waaler test.

Radiographic examinations

Radiographs of the sacroiliac joints, hips, ankles, and tarsi of all patients were obtained at followup. The radiographs were examined independently by 2 radiologists (VJ and KD) blinded to patient identification and without access to earlier radiographic, clinical, or laboratory data. In cases of score discrepancies, the radiographs were reexamined jointly by the 2 radiologists, and a final conclusion was reached. Radiographs of other affected joints were obtained when clinically indicated.

The sacroiliac joints were examined by conventional nondigital radiographs in the anteroposterior view. Stereoscopic examinations were performed when needed. Six stages of arthritis were defined according to the following grading system: 0 = normal findings, 1 = suspicious changes, 2 = definite changes, 3 and 4 = severe arthritic changes, and 5 = total bony ankylosis (29). In this system, grades 2–5 denote definite arthritic changes and correspond to the New York criteria grades 3–4 (30).

The hips were examined in the anteroposterior view and the ankles and tarsi in the lateral view. Radiographs of peripheral joints were graded according to a radiographic classification system for JRA in which grade 0 = normal joints, grade 1 = suspicious changes, grade 2 = growth disturbances, grades 3 and 4 = joint erosions, and grade 5 = mutilating abnormalities (31).

Genomic typing for HLA

HLA–DRB1 typing was performed using the nonisotopic method, based on polymerase chain reaction (PCR) and hybridization with oligonucleotide probes labeled with biotin (32). The HLA–DRB1 associations studied included DR1 (DRB1*01), DR4 (DRB1*04), DR5 (DRB*11 and *12), and DR8 (DRB1*08) alleles. HLA–DPB1 alleles were typed by sequence-specific oligonucleotide probing using a PCR, as described previously (33). The HLA–DPB1 associations studied included DP2 (DPB1*02) and DP3 (DPB1*03) alleles. The presence of HLA–B27 was determined by serologic testing (34).

Statistical analysis

Differences between the 2 groups were tested with Student's 2-tailed unpaired t-test for continuous variables and with the chi-square test or, when appropriate, Fisher's exact test for categorical variables. Corrections for the number of comparisons were not made, because the characteristics, disease variables, and genetic markers studied have previously been reported in the literature to be associated with JIA subtypes.

Logistic regression was used to analyze factors affecting the likelihood of developing sacroiliitis. Initial tests were performed on each of the candidate factors (patient characteristics, genetic markers, and disease variables assessed during the first 6 months of disease) separately. To identify predictors of sacroiliitis, a multiple logistic regression analysis was subsequently performed, including candidate factors that were significantly associated (P < 0.05) with sacroiliitis (age, sex, HLA–B27, DR4, DP2, and hip involvement at onset). The number of candidate factors was kept low because of the limited number of patients with sacroiliitis. Enthesitis, inflammatory back pain, and psoriasis within the first 6 months of disease were not included, because few cases were positive for these factors (n = 4–10). A family history of AS was not included as a candidate factor because values were missing in 18 cases. In order to test the stability of the identified predictors, additional regression analyses were performed, with inclusion of all associated or confounding variables and exclusion of variables not included in the final model. Hosmer-Lemeshow goodness-of-fit statistics were used to assess how well the logistic regression models fit the data (P values > 0.05 were considered fit). Data on the strength of the associations were expressed as the odds ratio (OR) and 95% confidence interval (95% CI).

P values <0.05, with no overlap between the 95% CI (for comparative groups) or with 1.00 not included in the 95% CI (for the OR), were considered statistically significant. The SPSS software program was used for all analyses (35).


  1. Top of page
  2. Abstract

Frequency of sacroiliitis, patient characteristics, and radiographic outcome. At followup, radiographic sacroiliitis had developed in 20 (6%) of the 314 patients with JIA. The sacroiliitis had not been demonstrated before the followup examination in 9 patients (45%), either because radiographs had not previously been obtained (n = 6) or because they had not demonstrated sacroiliitis (n = 3). Two patients presented with radiographically evident sacroiliitis. Sacroiliitis developed in 7 patients within 5 years, in 8 patients within 10 years, in 15 patients within 15 years, and in 20 patients within 18 years of the onset of arthritis.

Of the 66 patients with juvenile spondylarthropathy, 19 (29%) had radiographic sacroiliitis at followup. This group comprised 13 patients with juvenile AS, 3 (8%) of 36 patients with juvenile PsA (2 of whom also met the criteria for juvenile AS), 3 (20%) of 14 patients with SEA, and 1 (33%) of 3 patients with IBD. One of 248 patients with JRA had radiographic sacroiliitis but did not meet the clinical criteria for the classification of juvenile AS. This female patient had IgM-RF–positive, erosive, late-onset, HLA–B27–negative polyarthritis and a family history of AS.

The characteristics and radiographic outcomes of the patients are shown in Table 1. The male-to-female ratio was 7:3 in patients with and 3:7 in patients without sacroiliitis (P < 0.001). The age at onset was 8 years or older in 19 (95%) of the patients with sacroiliitis compared with 139 (47%) of those without sacroiliitis (P = 0.003). All patients with sacroiliitis had been older than age 6 years at disease onset. AS in 1 or more first- or second-degree relatives was present in 45% of the patients with and 8% of those without sacroiliitis (P < 0.001). The onset of arthritis was pauciarticular in 60% of the patients with and 64% of those without sacroiliitis (P not significant).

Table 1. Characteristics of juvenile idiopathic arthritis (JIA) patients with and those without radiographically demonstrated sacroiliitis*
All JIA patients (n = 314)With sacroiliitis (n = 20)Without sacroiliitis (n = 294)P
  • *

    Patients were followed up after a median disease duration of 14.9 years. Values are the number (%) of patients or the mean (95% confidence interval). AS = ankylosing spondylitis.

  • With versus without sacroiliitis.

  • Changes were bilateral in all but 1 patient, who had unilateral sacroiliitis grade 5. Radiographic bilateral sacroiliitis grade 4 was found in 1 patient with IgM rheumatoid factor–positive polyarthritis.

  • §

    Destruction and/or ossification.

 Male103 (33)14 (70)89 (30)0.000
 Age at onset, years7.6 (7.2–8.1)11.3 (10.3–12.4)7.4 (6.9–7.9)0.000
 Age at followup, years22.7 (22.2–23.3)26.8 (25.6–28.0)22.5 (21.9–23.0)0.000
 Disease duration at followup, years15.1 (14.8–15.3)15.4 (14.4–16.4)15.0 (14.8–15.2)0.440
 AS in first- or second-degree relative32 (10)9 (45)23 (8)0.000
 Systemic onset25 (8)025 (9)0.351
 Pauciarticular onset200 (64)12 (60)188 (64)0.924
 Polyarticular onset88 (28)7 (35)81 (28)0.645
 Axial arthritis1 (0)1 (1)0
Radiographic changes at followup
 Sacroiliac joints
  Grade 16 (2)06 (2)1.000
  Grades 2–520 (6)20 (100)00.000
 Hips, grades 2–559 (19)8 (40)51 (17)0.012
 Ankles, grades 2–549 (16)4 (20)45 (15)0.576
 Tarsi, grades 2–559 (19)5 (25)54 (18)0.462
 Peripheral joints, grades 3–572 (23)9 (45)63 (21)0.015
  Dorsal26 (8)6 (30)20 (7)0.000
  Plantar23 (7)7 (35)16 (5)0.000

Radiographically evident sacroiliitis was bilateral grade 2 in 8 patients, grade 3 in 2 patients, grade 4 in 5 patients, and grade 5 in 4 patients, while 1 patient had unilateral grade 5 sacroiliitis. Radiographic changes in the hips (grades 2–5) and erosion of any peripheral joint (grades 3–5) were more frequent in the patients with sacroiliitis than in those without sacroiliitis (P = 0.012 and P = 0.015, respectively). Radiographic dorsal and plantar changes in the os calcanei were associated with sacroiliitis (P < 0.001).

Clinical outcome and early-disease variables. At followup, spinal anterior and lateral flexion were reduced in 15 (75%) and 14 (70%) of patients with sacroiliitis and in 87 (30%) and 90 (31%) of patients without sacroiliitis (P < 0.001), respectively. Cervical rotation was ≤160° in 12 patients with and 108 patients without sacroiliitis (60% and 38%, respectively) (P = 0.05). The mean anterior and lateral flexion was reduced in the patients with sacroiliitis compared with those without sacroiliitis (Table 2).

Table 2. Axial signs in patients with juvenile idiopathic arthritis (JIA)*
All JIA patients (n = 314)With sacroiliitis (n = 20)Without sacroiliitis (n = 294)P
  • *

    Patients were followed up after a median disease duration of 14.9 years. Values are the mean (95% confidence interval).

  • With versus without sacroiliitis.

Schober's test, cm (range 2–11)6.6 (6.5–6.8)5.5 (4.9–6.2)6.7 (6.6–6.8)0.000
Lateral flexion, cm (range 5–34)18.9 (18.4–19.4)16.2 (13.5–18.9)19.1 (18.6–19.6)0.006
Chest expansion, cm (range 1–14)6.0 (5.8–6.2)5.8 (4.6–7.1)6.0 (5.8–6.3)0.538
Cervical rotation, degrees (range 15–180)155 (151–159)139 (118–159)156 (152–160)0.056

During the followup period, patients with sacroiliitis more frequently had inflammatory back pain, enthesitis, and clinical hip joint involvement than did patients without sacroiliitis (P < 0.001, P < 0.001, and P = 0.001, respectively) (Table 3). Anterior uveitis was seen in 35% of the patients with and 16% of those without sacroiliitis (P = 0.029).

Table 3. Disease variables in patients with juvenile idiopathic arthritis (JIA) with and without sacroiliitis*
 Disease duration ≤6 monthsWithin a median disease duration of 14.9 years
All JIA patients (n = 314)With sacroiliitis (n = 20)Without sacroiliitis (n = 294)All JIA patients (n = 314)With sacroiliitis (n = 20)Without sacroiliitis (n = 294)
  • *

    Values are the number (%) of patients or the mean (95% confidence interval).

  • P < 0.05 versus patients without sacroiliitis.

  • P < 0.001 versus patients without sacroiliitis.

  • §

    P < 0.01 versus patients without sacroiliitis.

  • ¶, §

    Pain in the spine at rest with morning stiffness that improves with movement (at disease onset) or low back pain of at least 3 months' duration improved by exercise and not relieved by rest (at followup).

Enthesitis6 (2)2 (10)4 (1)22 (7)9 (45)13 (4)
Hip joint involvement, clinical30 (10)7 (35)23 (8)66 (21)10 (50)§56 (19)
Psoriasis10 (3)3 (15)§7 (2)18 (6)3 (15)15 (5)
Inflammatory back pain4 (1)2 (10)§2 (1)26 (8)13 (65)13 (4)
Cumulative no. of affected joints5.8 (5.0–6.7)9.8 (4.0–15.6)5.6 (4.7–6.4)10.7 (9.3–12.0)14.9 (7.1–22.6)10.5 (9.0–11.7)
Uveitis31 (10)1 (5)30 (10)54 (17)7 (35)47 (16)

At 6 months, enthesitis, hip joint involvement, psoriasis, and inflammatory back pain were more frequent in patients with sacroiliitis than in those without sacroiliitis (P = 0.020, P < 0.001, P = 0.007, and P = 0.007, respectively).

The numbers of patients with involvement of the upper limbs, lower limbs, tarsi, or shoulders were not significantly different among the groups with and without sacroiliitis at either followup or disease onset (data not shown).

Genetic markers. The distribution of HLA–B27, DRB1, and DPB1 alleles in the JIA patients and healthy controls is shown in Table 4. HLA–B27 was present in 75% of patients with and 22% of those without sacroiliitis (P < 0.001). The frequency of HLA–B27 was higher in JIA patients both with and without sacroiliitis compared with healthy controls (OR 17.88, P < 0.001, and OR 1.73, P < 0.001, respectively).

Table 4. Frequency of HLA alleles in patients with juvenile idiopathic arthritis (JIA)*
 All JIA patients (n = 314)With sacroiliitis (n = 20)Without sacroiliitis (n = 294)Healthy controls (n = 295)Odds ratio (95% CI), vs. controls
Patients with sacroiliitisPatients without sacroiliitis
  • *

    Values are the number (%) of patients. Controls were randomly selected from the Norwegian Bone Marrow Donor Registry. 95% CI = 95% confidence interval.

  • P values are versus patients without sacroiliitis.

  • P < 0.001.

  • §

    P < 0.01.

  • P < 0.05.

HLA–B2781 (26)15 (75)66 (22)35 (12)17.88 (5.45–62.40)1.73 (1.02–2.93)
DRB1*0166 (21)5 (25)61 (21)57 (19)1.39 (0.42–4.32)1.09 (0.72–1.67)
DRB1*0487 (28)12 (60)§75 (26)99 (34)2.97 (1.09–8.26)§0.68 (0.47–0.98)
DRB1*11/1250 (16)1 (5)49 (17)32 (11)0.43 (0.02–3.25)1.64 (0.99–2.73)
DRB1*0898 (31)6 (30)92 (31)32 (7)5.32 (1.63–16.82)5.65 (3.35–9.62)
DPB1*02106 (34)1 (5)105 (36)56 (19)0.22 (0.01–1.64)2.37 (1.60–3.52)
DPB1*0372 (23)4 (20)68 (23)59 (20)1.00 (0.27–3.35)1.20 (0.80–1.82)

The frequency of HLA–DRB1*04 was higher in patients with sacroiliitis than in those without sacroiliitis (P = 0.003). Compared with healthy controls, HLA–DRB1*04 was increased in patients with sacroiliitis (OR 2.97, P = 0.01) and decreased in patients without sacroiliitis (OR 0.68, P = 0.03). DRB1*04 was found in 30 (37%) of 81 HLA–B27–positive and 57 (25%) of 233 HLA–B27–negative JIA patients (P = 0.029).

The frequency of DRB1*08 was higher in both patient groups than in controls (OR 5.32, P = 0.02 and OR 5.65, P < 0.001). The HLA–DPB1*02 allele occurred less frequently in patients with sacroiliitis than in those without sacroiliitis (P = 0.023). In comparison with controls, DPB1*02 was increased in the patients without, but not in those with sacroiliitis (OR 2.37, P < 0.001 and OR 0.22, P = 0.10, respectively).

Among patients with JIA of pauciarticular onset (n = 201), those with sacroiliitis (n = 13) had a lower frequency of DPB1*02 and a higher frequency of DRB1*04 than did those without sacroiliitis (n = 188 [8% versus 44%, P = 0.024 and 62% versus 21%, P = 0.003, respectively]). Among patients with polyarticular-onset JIA (n = 88), there were no statistically significant differences between patients with (n = 7) and without (n = 81) sacroiliitis in the frequency of DRB1*04 (57% versus 31%, P = 0.212) or DPB1*02 (0% versus 21%, P = 0.338).

Predictors of sacroiliitis. To identify the most important early correlates of sacroiliitis in JIA, multiple logistic regression analyses were performed (Table 5). Age, sex, HLA–B27, DRB1*04, DPB1*02, and hip involvement at onset were tested as possible predictors. HLA–B27 (OR 5.29, P = 0.005), the absence of HLA–DPB1*02 (OR 0.09, P = 0.030), hip joint involvement at onset (OR 4.98, P = 0.017), and onset at age 8 years or older (OR 9.21, P = 0.039) were identified as statistically significant risk factors for the development of sacroiliitis. The regression model fit the data well (χ2 = 1.979, 8 degrees of freedom, P = 0.982). The same predictors were identified when enthesitis, inflammatory back pain at onset, family history of AS, and disease duration were included and when DRB1*04 was excluded from the multiple regression analyses. When psoriasis at onset was added as an independent variable, late onset was not a significant predictor (OR 6.98, 95% CI 0.81–60.44), and psoriasis at onset contributed significantly to the likelihood of sacroiliitis (OR 7.87, 95% CI 1.31–47.34).

Table 5. Association of patient characteristics, disease onset variables, and HLA alleles with the likelihood of the development of sacroiliitis in 314 patients with juvenile idiopathic arthritis*
 Univariate analysisMultivariate analysis
  • *

    OR = odds ratio; 95% CI = 95% confidence interval; AS = ankylosing spondylitis.

  • Results of multiple logistic regression analysis identifying predictors of sacroiliitis. Independent variables associated with the development of sacroiliitis in the univariate analyses were included except for a family history of AS (because of missing information on 18 patients) and enthesitis, inflammatory back pain, and psoriasis at onset (due to few positive cases).

Onset age ≥8 years21.122.80–159.480.0039.211.11–76.400.039
Family history of AS9.943.67–26.920.000
Enthesitis within 6 months8.061.31–46.960.020
Hip joint involvement within 6 months6.342.30–17.460.0004.981.32–18.700.017
Psoriasis within 6 months7.231.72–30.490.007
Inflammatory back pain within 6 months16.222.16–121.90.007


  1. Top of page
  2. Abstract

In the present study, the frequency of radiographic sacroiliitis in JIA was 6% after a median disease duration of 14.9 years. This frequency is higher than the 2% prevalence of juvenile AS reported in most pediatric rheumatology registers (36–39). The long duration of followup may have increased the frequency of sacroiliitis in our cohort. Furthermore, when radiographs of all JIA patients were obtained, sacroiliitis was seen almost twice as frequently as in clinical practice. Our results are consistent with the high frequency of sacroiliitis (9–92%) seen in radiographic assessments of patients with various subtypes of JIA (8, 9, 18, 19, 40).

DP2 was an independent protective factor, and HLA–B27, early hip joint involvement, and disease onset after age 8 years were risk factors for development of sacroiliitis. DP2 was statistically significantly decreased in patients with sacroiliitis and pauciarticular-onset JIA but not in those with sacroiliitis and polyarticular arthritis. This result is consistent with the fact that DP2 is a well-known correlate of pauciarticular-onset JRA (7, 20, 41, 42). The identification of HLA–B27 as a predictor of sacroiliitis is in accordance with the strong associations between HLA–B27 and juvenile AS demonstrated in several previous studies (5, 9, 43, 44). Hip joint involvement during the first 6 months of disease was an important correlate of sacroiliitis in our study. An association between hip involvement and juvenile AS has been reported by other investigators (5, 23). Several studies have shown that late onset is associated with the development of juvenile spondylarthropathy, supporting our suggestion that age at disease onset influences the risk for sacroiliitis (18, 22, 23, 45). However, the reliability of the predictive factors identified here may be influenced by the limited number of patients with sacroiliitis.

Enthesitis, male sex, and axial symptoms at onset were associated with the development of sacroiliitis but were not identified as predictors in the present study. An association between these factors and the development of juvenile AS has been found in numerous previous studies (5, 18, 22, 23, 45). Information about the prevalence of sacroiliitis in juvenile PsA is scarce. We found that sacroiliitis developed in 3 of 10 patients who had psoriasis at disease onset. Further studies are needed to determine whether axial involvement is more frequent in juvenile PsA than in other subgroups of JIA.

In our study, the frequency of DR4 was higher in patients with sacroiliitis than in either patients without sacroiliitis or healthy controls. All but 1 of the patients with sacroiliitis had peripheral arthritis. A German study previously demonstrated that DR4 was associated with adult AS complicated by peripheral arthritis (46), but this association was not seen in a Finnish patient population (47). Ploski et al reported a weak but nonsignificant increase in DR4 in patients with juvenile AS (20). The selection of patients may explain the difference in these results. An increased frequency of DR4 has consistently been seen in patients with RF-positive, late-onset polyarticular JRA (48, 49). In our study, 1 of the patients with sacroiliitis and DR4 had IgM-RF–positive, late-onset polyarticular JRA, and the possibility that this patient might have 2 conditions cannot be excluded.

In adult RA, DR4 has been associated with radiographic progression (50), and Burgos-Vargas et al suggested that sacroiliitis was a marker of severe arthritis in patients with SEA (5). The association between DR4 and sacroiliitis in our cohort might support the hypothesis that DR4 is a marker of disease severity in children with juvenile spondylarthropathy and peripheral arthritis. In the present multivariate analysis, however, DR4 was not significantly correlated with the development of sacroiliitis, showing that this factor was not an independent predictor of sacroiliitis. HLA–DR8 was associated with JIA in patients both with and without sacroiliitis in our study, as has previously been demonstrated (20, 21).

Stereoscopic examination of the sacroiliac joints in the anteroposterior view is regarded as a good, conventional approach to investigating these joints, because it provides a combination of sharp outlines and 3-dimensional views (51, 52). Conventional radiographs may be less sensitive than contrast-enhanced magnetic resonance imaging and computed tomography for the early detection of sacroiliitis (19, 53). It remains to be shown whether such early changes predict a later transition to juvenile AS (10, 19, 53). In the present study, the radiographs were evaluated by 2 independent radiologists who have worked with rheumatic diseases for many years.

Sixty-five percent of our patients with radiographic sacroiliitis had inflammatory back pain, and spinal mobility was reduced in 75% after a median disease duration of 14.9 years. The lack of Norwegian controls limited evaluation of these results. However, signs and symptoms of axial involvement were more frequent in JIA patients with sacroiliitis than in those without sacroiliitis, and our results are in accordance with those of previous studies indicating that spinal involvement is a frequent outcome of juvenile spondylarthropathy (5).

In our study, the frequency of erosions of any peripheral joint was higher in patients with sacroiliitis than in those without sacroiliitis. The reliability of this finding is influenced by the fact that radiographs of affected peripheral joints other than hips, ankles, and tarsi were obtained only when clinically indicated and were not obtained for all patients and examined in a blinded manner. Consistent with previous results, radiographic changes of the hips and the calcanei were associated with sacroiliitis (5, 11, 22, 45, 54–56).

The participants in the present study represented all new patients with JIA who were admitted to the hospital between 1980 and 1985. Patients who are admitted to hospitals tend to have more severe disease than those who are recruited from the general population. However, the public health care system in Scandinavia, which provides regular free checkups for all children of preschool and school age, probably facilitates admission of most children with chronic rheumatic diseases (4). The age, sex, and disease onset type of the present cohort of JIA patients were similar to those of patients in previous epidemiologic studies (3, 4, 57, 58), indicating that the participants were relatively representative of patients with diagnosed JIA.

In our patients, the most frequent type of onset of arthritis was pauciarticular, in those both with and without sacroiliitis, as previously demonstrated among patients with juvenile spondylarthropathy and JRA (1, 4, 5, 11, 22, 45). In contrast with the results of Burgos-Vargas et al, the pauciarticular:polyarticular ratio did not tend to be higher in patients with sacroiliitis than in those without sacroiliitis (5). Our results support the contention that, although prevalent, pauciarticular onset is not an important discriminator between juvenile spondylarthropathy and other childhood arthritides (5). However, retrospective assessment of the early disease variables may have influenced the reliability of our data concerning disease onset.

In the present long-term followup study, the prevalence of radiographic sacroiliitis in JIA was 6%. Spinal mobility was reduced in most patients with sacroiliitis. Our results indicate that, in addition to HLA–B27, DP2 can contribute to the early differentiation of juvenile spondylarthropathy from other arthritides. Furthermore, early hip involvement and late onset were related to the progression of sacroiliitis in JIA. We look forward to future studies for a determination of the reliability of early predictors of sacroiliitis in children.


  1. Top of page
  2. Abstract
  • 1
    Oen KG, Cheang M. Epidemiology of chronic arthritis in childhood. Semin Arthritis Rheum 1996; 6: 57591.
  • 2
    Peterson LS, Mason T, Nelson AM, O'Fallon WM, Gabriel SE. Juvenile rheumatoid arthritis in Rochester, Minnesota 1960–1993: is the epidemiology changing? Arthritis Rheum 1996; 39: 138590.
  • 3
    Kaipiainen-Seppanen O, Savolainen A. Incidence of chronic juvenile rheumatic diseases in Finland during 1980–1990. Clin Exp Rheumatol 1996; 14: 4414.
  • 4
    Gare BA, Fasth A. Epidemiology of juvenile chronic arthritis in southwestern Sweden: a 5-year prospective population study. Pediatrics 1992; 90: 9508.
  • 5
    Burgos-Vargas R, Vazquez-Mellado J. The early clinical recognition of juvenile-onset ankylosing spondylitis and its differentiation from juvenile rheumatoid arthritis. Arthritis Rheum 1995; 38: 83544.
  • 6
    Schaller JG. Pauciarticular arthritis of childhood. Ann Pediatr 1983; 30: 55763.
  • 7
    Ploski R, Vinje O, Ronningen KS, Spurkland A, Sorskaar D, Vartdal F, et al. HLA class II alleles and heterogeneity of juvenile rheumatoid arthritis: DRB1*0101 may define a novel subset of the disease. Arthritis Rheum 1993; 36: 46572.
  • 8
    Burgos-Vargas R, Clark P. Axial involvement in the seronegative enthesopathy and arthropathy syndrome and its progression to ankylosing spondylitis. J Rheumatol 1989; 16: 1927.
  • 9
    Cabral DA, Oen KG, Petty RE. SEA syndrome revisited: a longterm followup of children with a syndrome of seronegative enthesopathy and arthropathy. J Rheumatol 1992; 19: 12825.
  • 10
    Van der Linden S, Valkenburg HA, Cats A. Evaluation of diagnostic criteria for ankylosing spondylitis: a proposal for modification of the New York criteria. Arthritis Rheum 1984; 27: 3618.
  • 11
    Marks SH, Barnett M, Calin A. A case-control study of juvenile- and adult-onset ankylosing spondylitis. J Rheumatol 1982; 9: 73941.
  • 12
    Southwood TR, Petty RE, Malleson PN, Delgado EA, Hunt DW, Wood B, et al. Psoriatic arthritis in children. Arthritis Rheum 1989; 32: 100713.
  • 13
    Rosenberg AM, Petty RE. A syndrome of seronegative enthesopathy and arthropathy in children. Arthritis Rheum 1982; 25: 10417.
  • 14
    Petty RE, Southwood TR, Baum J, Bhettay E, Glass DN, Manners P, et al. Revision of the proposed classification criteria for juvenile idiopathic arthritis: Durban, 1997. J Rheumatol 1998; 25: 19914.
  • 15
    JRA Criteria Subcommittee of the Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Current proposed revision of the JRA criteria. Arthritis Rheum 1977; 20: Suppl 2 1959.
  • 16
    Guzman J, Burgos-Vargas R, Duarte-Salazar C, Gomez-Mora P. Reliability of the articular examination in children with juvenile rheumatoid arthritis: interobserver agreement and sources of disagreement. J Rheumatol 1995; 22: 23316.
  • 17
    Felson DT, Anderson JJ. Methodological and statistical approaches to criteria development in rheumatic diseases. Baillieres Clin Rheumatol 1995; 9: 25366.
  • 18
    Hall MA, Burgos-Vargos R, Ansell BM. Sacroiliitis in juvenile chronic arthritis: a 10-year follow-up. Clin Exp Rheumatol 1987; 5: 657.
  • 19
    Bollow M, Biedermann T, Kannenberg J, Paris S, Schauer-Petrowski C, Minden K, et al. Use of dynamic magnetic resonance imaging to detect sacroiliitis in HLA-B27 positive and negative children with juvenile arthritides. J Rheumatol 1998; 25: 55664.
  • 20
    Ploski R, Flato B, Vinje O, Maksymowych W, Forre O, Thorsby E. Association to HLA-DRB1*08, HLA-DPB1*0301 and homozygosity for an HLA-linked proteasome gene in juvenile ankylosing spondylitis. Hum Immunol 1995; 44: 8896.
  • 21
    Maksymowych WP, Gorodezky C, Olivo A, Alaez C, Wong C, Burgos-Vargas R, et al. HLA-DRB1*08 influences the development of disease in Mexican Mestizo with spondylarthropathy. J Rheumatol 1997; 24: 9047.
  • 22
    Schaller JG. Ankylosing spondylitis of childhood onset. Arthritis Rheum 1977; 20: Suppl 2 398401.
  • 23
    Hussein A, Abdul-Khaliq H, von der Hardt H. Atypical spondyloarthritis in children: proposed diagnostic criteria. Eur J Pediatr 1989; 148: 5137.
  • 24
    Vandvik IH. Psychosocial factors in juvenile rheumatic diseases with recent onset [dissertation]. Rikshospitalet University Hospital, 1992.
  • 25
    Macrae IF, Wright V. Measurement of back movement. Ann Rheum Dis 1969; 28: 5849.
  • 26
    Burgos-Vargas R, Lardizabal-Sanabria J, Katona G. Anterior spinal flexion in healthy Mexican children. J Rheumatol 1985; 12: 1235.
  • 27
    Domjan L, Nemes T, Balint G, Toth Z, Gomor B. A simple method for measuring lateral flexion of the dorsolumbar spine. J Rheumatol 1990; 17: 6635.
  • 28
    Moll JMH, Wright V. An objective clinical study of chest expansion. Ann Rheum Dis 1972; 31: 18.
  • 29
    Johnsen K, Gran JT, Dale K, Husby G. The prevalence of ankylosing spondylitis among Norwegian Samis (Lapps). J Rheumatol 1992; 19: 15914.
  • 30
    Moll JM, Wright V. New York clinical criteria for ankylosing spondylitis: a statistical evaluation. Ann Rheum Dis 1973; 32: 35463.
  • 31
    Dale K, Paus AC, Laires K. A radiographic classification system in juvenile rheumatoid arthritis applied to the knee. Eur Radiol 1994; 4: 2732.
  • 32
    Verduyn W, Doxiadis II, Anholts J, Drabbels JJ, Naipal A, D'Amaro J, et al. Biotinylated DRB sequence-specific oligonucleotides: comparison to serologic HLA-DR typing of organ donors in eurotransplant. Hum Immunol 1993; 37: 5967.
  • 33
    Ronningen KS, Spurkland A, Markussen G, Iwe T, Vartdal F, Thorsby E. Distribution of HLA class II alleles among Norwegian Caucasians. Hum Immunol 1990; 29: 27581.
  • 34
    Hill AV, Allsopp CE, Kwiatkowski D, Anstey NM, Greenwood B, McMichael AJ. HLA class 2 typing by PCR: HLA-B27 and an African B27 subtype. Lancet 1991; 337: 6402.
  • 35
    SPSS Inc. SPSS for windows user's guide. Chicago: SPSS, 1993.
  • 36
    Rosenberg AM. Analysis of a pediatric rheumatology clinic population. J Rheumatol 1990; 17: 82730.
  • 37
    Malleson PN, Fung MY, Rosenberg AM. The incidence of childhood rheumatic diseases: results from the Canadian Pediatric Rheumatology Association Disease Registry. J Rheumatol 1996; 23: 19817.
  • 38
    Bowyer S, Roettcher P, Pediatric Rheumatology Database Research Group. Pediatric rheumatology clinic populations in the United States: results of a 3 year survey. J Rheumatology 1996; 23: 196874.
  • 39
    Symmons DP, Jones M, Osborne J, Sills J, Southwood TR, Woo P. Pediatric rheumatology in the United Kingdom: data from the British Pediatric Rheumatology Group National Diagnostic Register. J Rheumatol 1996; 23: 197580.
  • 40
    Olivieri I, Foto M, Ruju GP, Gemignani G, Giustarini S, Pasero G. Low frequency of axial involvement in Caucasian pediatric patients with seronegative enthesopathy and arthropathy syndrome after 5 years of disease. J Rheumatol 1992; 19: 46975.
  • 41
    Van Kerchove C, Luyrink L, Elma MS, Maksymowych WP, Levinson JE, Larson MG, et al. HLA-DP/DR interaction in children with juvenile rheumatoid arthritis. Immunogenetics 1990; 32: 3648.
  • 42
    Fernandez-Vina MA, Fink CW, Stastny P. HLA antigens in juvenile arthritis: pauciarticular and polyarticular juvenile arthritis are immunogenetically distinct. Arthritis Rheum 1990; 33: 178794.
  • 43
    Ansell BM, Bywaters EG, Lawrence JS. Familial aggregation and twin studies in Still's disease. Juvenile chronic polyarthritis. Rheumatology 1969; 2: 3761.
  • 44
    Schaller JG, Ochs HD, Thomas ED, Nisperos B, Feigl P, Wedgwood RJ. Histocompatibility antigens in childhood-onset arthritis. J Pediatr 1976; 88: 92630.
  • 45
    Ladd JR, Cassidy JT, Martel W. Juvenile ankylosing spondylitis. Arthritis Rheum 1971; 14: 57990.
  • 46
    Miehle W, Schattenkirchner M, Albert D, Bunge M. HLA-DR4 in ankylosing spondylitis with different patterns of joint involvement. Ann Rheum Dis 1985; 44: 3944.
  • 47
    Westman P, Leirisalo-Repo M, Partanen J, Koskimies S. A comparative study of HLA genes in HLA–B27 positive ankylosing spondylitis and HLA–B27 positive peripheral reactive arthritis. Arthritis Rheum 1996; 39: 9439.
  • 48
    Murray KJ, Moroldo MB, Donnelly P, Prahalad S, Passo MH, Giannini EH, et al. Age-specific effects of juvenile rheumatoid arthritis–associated HLA alleles. Arthritis Rheum 1999; 42: 184353.
  • 49
    Førre Ø, Dobloug JH, Hoyeraal HM, Thorsby E. HLA antigens in juvenile arthritis: genetic basis for the different subtypes. Arthritis Rheum 1983; 26: 3538.
  • 50
    Listing J, Rau R, Muller B, Alten R, Czerwony G, Gromnica-Ihle E, et al. HLA-DRB1 genes, rheumatoid factor, and elevated C-reactive protein: independent risk factors of radiographic progression in early rheumatoid arthritis. J Rheumatol 2000; 27: 21009.
  • 51
    Dale K, Vinje O. Radiography of the spine and sacro-iliac joints in ankylosing spondylitis and psoriasis. Acta Radiol Diagn (Stockh) 1985; 26: 14559.
  • 52
    Azouz EM, Duffy CM. Juvenile spondyloarthropathies: clinical manifestations and medical imaging. Skeletal Radiol 1995; 24: 399408.
  • 53
    Oudjhane K, Azouz EM, Hughes S, Paquin JD. Computed tomography of the sacroiliac joints in children. Can Assoc Radiol J 1993; 44: 3134.
  • 54
    Jacobs JC, Berdon WE, Johnston AD. HLA-B27-associated spondyloarthritis and enthesopathy in childhood: clinical, pathologic, and radiographic observations in 58 patients. J Pediatr 1982; 100: 5218.
  • 55
    Burgos-Vargas R, Pineda C. New clinical and radiographic features of the seronegative spondylarthropathies. Curr Opin Rheumatol 1991; 3: 56274.
  • 56
    Cohen PA, Job-Deslandre CH, Lalande G, Adamsbaum C. Overview of the radiology of juvenile idiopathic arthritis (JIA). Eur J Radiol 2000; 33: 94101.
  • 57
    Moe N, Rygg M. Epidemiology of juvenile chronic arthritis in northern Norway: a ten-year retrospective study. Clin Exp Rheumatol 1998; 16: 99101.
  • 58
    Kunnamo I, Kallio P, Pelkonen P. Incidence of arthritis in urban Finnish children: a prospective study. Arthritis Rheum 1986; 29: 12328.