Juvenile idiopathic arthritis (JIA) is a wide-ranging group of diseases affecting children under the age of 16 years.1 Its prevalence varies between 0.07 and 4.01 per 1000 children.2 Enthesitis-related arthritis (ERA) is one of the JIA subtypes. It accounts for 20% of JIA children.3 Boys are more affected by this rheumatism.4 Like adult spondyloarthritis, a genetic predisposition by Human Leukocyte Antigen B27 (HLA-B27) was noted in 50–80% of ERA patients.5–7

This disease manifests by enthesitis and arthritis, occurring predominantly in the lower limbs.8 Hip involvement is the most severe articular involvement in ERA and could lead to an important functional impairment.9,10

The axial skeleton is also involved, but contrary to adults spondyloarthritis occurs in only 30% of children within 15 months of diagnosis and was often asymptomatic.8,11–13

Despite the progress in the treatment of JIA, ERA seems to have a poorer prognosis and worse health status outcomes than the other JIA subtypes.14,15 Besides, severe extra-articular manifestations, such as anterior uveitis are not rare in ERA patients.16 Hence, it is important to diagnose this disease early. Unfortunately, the diagnosis of ERA is not always obvious or easy. According to the International League of Associations for Rheumatology (ILAR) classification ERA diagnosis based on only clinical and laboratory findings.17 Yet, these criteria do not allow the diagnosis of all the ERA patients, especially with axial involvement.18 Recently, the European League Against Rheumatism (EULAR)-Pediatric Rheumatology European Society (PReS) task force has proposed a consensus-based recommendation on the use of imaging in the JIA diagnosis.19

According to this evidence, magnetic resonance imaging (MRI) of sacroiliac joints has shown increasing interest in the ERA diagnosis.18 However, unlike the adult data, the value of MRI to detect abnormalities in children remains challenging.

There is little information on the value of this imaging in the ERA diagnosis and even less in patients with silent axial involvement.

Thus, we aimed to determine the ability of sacroiliac joints MRI to diagnose ERA patients and study the predictive factors of the presence of MRI inflammatory lesions.

Sacroiliitis in MRI is frequent in ERA patients. Its prevalence varies between 20% and 88% of the patients in the literature.11,25–27 In our study, the ERA patients had a positive SI MRI scan and a painful SI in clinical examination in 59.2% (n=16) and 51.8% (n=14), respectively.

Many authors demonstrated a discrepancy between physical examination and MRI lesions in JIA. In a recent study conducted by Demir et al. including 57 ERA patients, authors showed that among the patients without axial symptoms (n=30, 53%) had MRI abnormalities (n=23, 77%) (inflammatory or erosive lesions).28 Likewise, Weiss et al. demonstrated that the positive and negative predictive value of low back pain and clinical examination were not good at detection of sacroilitiis.27 Even if most patients were asymptomatic, almost all had structural damage at the sacroiliac joints MRI.27 In a retrospective study including 160 ERA patients, authors noted that only 25% of the patients with sacroiliitis on the MRI were symptomatic.11 In our study, we found that sacroiliac joint pain in the clinical examination predicted sacroiliac bone edema in MRI with an odds ratio of 6.8. This result could be explained by the relatively long diagnosis delay in our patients. Besides, spondyloarthritis seem to be more severe in north African patients.29

Interestingly, we found that sensitivity and specificity of the sacroiliac joints MRI in the diagnosis of ERA were 61.54% and 82.35%, respectively. Positive predictive value was also good. Comparing our result with adults, the sensitivity and specificity were higher than our findings.8 This disagreement was also seen with a previous ERA study in which SI BME has a low sensitivity of 42%, but very high specificity of 96%.30 In fact, while the BME, capsulitis, ankylosis and erosions had also high specificity, the synovial enhancement and the radiologist's global diagnostic impression had the highest sensitivity for ERA diagnosis.30

Nevertheless, the MRI of SI value varied according to the definition used. According to the ASAS criteria, sacroiliac MRI is considered as “positive” if BME is present on at least two sections in the typical subchondral or peri-articular areas.24 Other MRI features presenting active inflammation are not sufficient for a “positive” MRI examination according to the criteria.

ASAS definition was extrapolated in children,31 but it is not yet valid in this population because of the lack of consideration for age-related maturational changes. As shown in the Herregods et al. study, the sensitivity and the specificity were different when the radiologist applied the ASAS definition or the global diagnosis impression.31

Recently, the MRI in JIA working group of Outcome Measures in Rheumatology and Clinical Trials (OMERACT) published the Atlas of MRI in SI joints illustrating the growth-related changes and different features of inflammation and structural damage.23

Certainly this reference will improve the sensitivity and the specificity of this tool to diagnose earlier and with more accuracy features of sacroiliac involvement even in asymptomatic patient.

Recently, Guo et al. highlighted the importance of classifying ERA into two distinctive disease subtypes, axial and peripheral forms, like adult spondyloarthritis. The authors found sacroiliitis in all 57 patients with axial ERA, and 31.58% of patients had asymptomatic inflammation in the sacroiliac joint, as demonstrated on MRI. They suggested that axial ERA was misdiagnosed for a long time due the lack of use of systematic imaging to detect axial inflammation.32

This result raises the question of whether the ILAR 2001 classification criteria,17 which use axial skeleton symptoms (sacroiliac joints pain and low back pain) in the diagnosis of ERA, are able to diagnose all ERA patients? And if the Pediatric Rheumatology International Trials Organization (PRINTO) JIA classification criteria,33 in which sacroilitiis on imaging is a major criterion, are more suitable in these patients?

Even if both of the classification criteria were shown to have great specificity,34,35 sacroiliac joints MRI should be helpful in the early diagnosis of ERA patients especially in litigious cases.

Secondly, we noted that sacroiliac joints MRI lesions were associated with clinical active enthesitis, hip joint limitation, increased CRP level, higher ESR, higher JADAS, and lower BMI. Similarly, Guo et al. showed that JIA patients who had axial involvement, tended to be older, have a longer delay in diagnosis, and exhibit significantly higher levels of inflammatory markers.32

Other studies showed that older age at the onset of ERA, female gender, the number of active joints, number of active enthesitis, HLA-B27 positivity, and increased CRP level were associated with the presence of sacroilitiis in MRI.27,36–38 These findings supported that positive sacroiliitis reflects an active and severe disease and should have an important implication in the therapeutic approach of ERA. Thus, the axial phenotype of ERA may require more aggressive treatments than the peripheral phenotype to prevent progressive damage.32

It is known that conventional synthetic DMARDs were not effective in axial involvement which requires escalating to biological treatment in NSAID refractory children.39

In sum, and despite the small number of patients and a retrospective design of our study, we can conclude that MRI of sacroiliac joints has a high specificity to detect axial involvement and should be considered in all children with peripheral enthesitis or arthritis even if asymptomatic in order to recognize the ERA diagnosis and adapt promptly the therapeutic strategy. However, MRI would be problematic since the interpretation would vary by the level of experience of the MRI reader, and the availability of this expensive tool, especially in the low incomes county. There is another limitation in our study, which is the diagnosis delay of ERA. This could be explained by the fact that this disease is not well known and the patients are late referred.

Another obstacle in MRI use, is the need for sedation in some children and the scan time.

Our study showed that sacroiliac joints MRI has a good specificity and positive predictive value in the diagnosis of ERA patients among JIA patients. We also noted an association between MRI lesions and inflammatory biomarkers and disease activity. This underlines the usefulness of sacroiliac joints MRI in the early diagnosis and the follow-up of ERA patients.

This study was approved by the ethics committee of the Hospital.

Dr. Hanene Ferjani: methodology and writing – review & editing; Dr. Lobna Kharrat: roles/writing – original draft; Dr. Dorra Ben Nessib: data curation and formal analysis; Dr. Dhia Kaffel: data analysis; Dr. Kaouther Maatallah: visualization & conceptualization; Dr. Wafa Hamdi: supervision & validation.

The authors declare that they have no funding for the research.

None declared.