Introduction

Childhood enthesitis-related arthritis (ERA) is a category of juvenile idiopathic arthritis that is characterized by arthritis, enthesitis (inflammation at the sites where tendons and ligaments insert into bone), risk of axial disease, and an underlying genetic predisposition. Under the International League of Associations for Rheumatology (ILAR) juvenile arthritis classification criteria, ERA is one of the three categories that is included under the umbrella term juvenile-onset spondyloarthritis (JSpA). The other two categories of JSpA, according to the ILAR criteria, are psoriatic arthritis (PsA) and undifferentiated arthritis (which includes children who have features of both ERA and PsA) [1]. In order to meet criteria for ERA, children must have arthritis and enthesitis or arthritis or enthesitis with at least two of the following: sacroiliac joint tenderness or inflammatory lumbosacral pain, human leukocyte antigen (HLA)-B27 positivity, onset of arthritis in a male patient older than 6 years of age, acute anterior uveitis, or a first degree relative with HLA-B27-associated disease (ankylosing spondylitis (AS), ERA, sacroiliitis with inflammatory bowel disease, reactive arthritis) or acute anterior uveitis. Children cannot be classified as having ERA if they have a personal or family history of psoriasis (first-degree relative), positive testing for IgM rheumatoid factor on at least two occasions 3 months apart, or systemic juvenile idiopathic arthritis (JIA) (Table 1). More recently, there has been an emphasis on categorizing adult patients based on whether or not they have axial disease. However, this is also highly relevant to the pediatric population since we do not know the consequences of untreated axial disease on axial skeleton growth. The Assessment of Spondyloarthritis (SpA) International Society (ASAS) strongly supports a simplified classification for adult disease: (1) peripheral SpA (when spine and sacroiliac joints are spared) and (2) axial SpA (when spine and sacroiliac joints are involved with or without peripheral skeletal involvement) [2–4].

Table 1 International League of Associations for Rheumatology (ILAR) criteria for enthesitis-related arthritis
Full size table

Standard treatment guidelines for ERA are lacking, and the majority of treatment recommendations for ERA are based on studies performed in adult SpA and rheumatoid arthritis or in children with the other categories of JIA [5]. However, the clinical phenotype of ERA differs significantly from both adult-onset SpA and from the other types of JIA. Pediatric patients have more peripheral arthritis and enthesitis and fewer symptoms of spinal involvement at disease onset than adults [6•]. Hip arthritis and tarsal joint arthritis (tarsitis) are more common in ERA than in adult SpA. In comparison to other categories of JIA without a predisposition to enthesitis or axial arthritis, children with ERA tend to have higher disease activity, higher pain intensity, and poorer health outcomes [7]. Children with ERA are also less likely to achieve and to sustain inactive disease than children with one of the other categories of JIA [8, 9]. The development of new disease activity scores such as the Juvenile Arthritis Disease Activity Score (JADAS) and the JSpA Disease Activity Score (JSpADA) will help better assess and address the poorer disease outcomes in children with ERA. Additionally, there is increased use of radiologic studies, specifically ultrasound with Doppler and magnetic resonance imaging, to help detect the presence of enthesitis and/or sacroiliitis in patients with ERA. These modalities will help improve the phenotyping and provision of care to these children.

Diagnosis and disease monitoring

Diagnostic procedures

  • Enthesitis and sacroiliitis are challenging to detect on physical examination, and thresholds for diagnosis vary between practitioners.

  • The application of ultrasound with Doppler and whole-body magnetic resonance imaging (MRI) to evaluate enthesitis has been studied; their applicability to routine clinical practice remains unclear.

  • MRI is becoming increasingly routine in the evaluation of sacroiliitis.

  • The standardization of the definition of sacroiliitis based on imaging findings will allow for an objective means of determining whether or not a patient has axial disease and warrants treatment with a tumor necrosis factor alpha (TNF-α) blocker.

Ultrasound with Doppler

Enthesitis in children is typically defined as localized pain, tenderness, or swelling over the entheses. However, physical examination is not perfect as evidenced by studies in adults that have shown that ultrasound can detect enthesitis that was not identified during the physical examination [10, 11]. The most common ultrasound abnormalities seen with enthesitis include increased power Doppler signal, enthesophytes, calcifications, tendon thickening, and hypoechogenicity [12]. In a pediatric study, the positive and negative predictive values of tenderness on standardized physical exam for detection of enthesitis by USD were low [13••]. Additionally, USD is useful in distinguishing enthesitis from other possible noninflammatory causes of pain. Identification of abnormalities at the entheses in children, however, mandates knowledge of the appearance of cartilage and tendons in growing children. Two studies demonstrated that tendon thickness increases with age and that a small degree of cartilage vascularity is normal, especially in younger children [14, 15]. As additional studies are performed to further our understanding of the normal appearance of pediatric cartilage and tendons on USD, we will hopefully be able to more precisely and accurately identify enthesitis using USD as part of routine clinical practice. More research is needed to determine an optimal ultrasound scoring method for enthesitis, the clinical importance of subclinical enthesitis, and the role of USD for monitoring disease activity [5].

An added benefit of USD, especially in the pediatric population, is that it is noninvasive and does not expose the patient to radiation. It is also generally cost-effective and more accessible than MRI.

Magnetic resonance imaging

The gold standard for the diagnosis of juvenile AS is the radiograph (Fig. 1); however, radiographs are not sensitive for the detection of early sacroiliitis. The presence of subchondral or periarticular bone marrow edema (BME) on MRI short-tau inversion recovery (STIR) images of the sacroiliac joints is highly suggestive of active disease (Fig. 2). Associated tendon or ligament thickening, adjacent soft-tissue swelling and edema, synovitis, and joint or bursal fluid are other important findings seen in JSpA [16]. The presence of enthesitis, synovitis, or capsulitis on MRI in the absence of BME is compatible with sacroiliitis but not sufficient for making a diagnosis of active sacroiliitis [17]. Without BME, other differential diagnoses should be considered such as infectious or oncologic processes. A recent study in children demonstrated that administration of gadolinium contrast did not add incremental value to the MRI evaluation of sacroiliitis [18]. Therefore, given the additional risks and costs associated with contrast administration, we do not recommend the use of contrast for routine evaluation of inflammatory sacroiliitis.

Fig. 1
figure 1

AP radiograph of the pelvis in a 16-year-old boy with lower back pain demonstrating sclerosis and erosive changes of the iliac side of the left sacroiliac joint (arrow) suggestive of sacroiliitis. Images courtesy of Nancy A. Chauvin, MD, Assistant Professor of Pediatric Radiology, The Children’s Hospital of Philadelphia, Philadelphia, PA.

Full size image
Fig. 2
figure 2

Coronal MR images of the sacrum. a Fluid-sensitive image and b T1-weighted image demonstrates bone marrow edema within both aspects of the left sacroiliac joint, most pronounced within the iliac bone (large arrows). There are small erosive changes within the articular surface of the left ilium (small arrows). Sclerosis is seen along the iliac side of the joint, as demonstrated by low signal intensity, extending more than 5 mm from the articular surface (dashed arrow). Images courtesy of Nancy A. Chauvin, MD, Assistant Professor of Pediatric Radiology, The Children’s Hospital of Philadelphia, Philadelphia, PA.

Full size image

In a recent study, 20 % of children with JSpA had sacroiliitis on MRI at disease onset. Of the patients with sacroiliitis, two thirds were asymptomatic and one third would have been missed if evaluated by radiograph alone [19••]. The majority of children with active inflammation also had MRI evidence of chronic damage (sclerosis and/or erosions). HLA-B27 positivity and elevated C-reactive protein levels were more prevalent in those children with active sacroiliitis. These findings suggest that there may be utility in screening JSpA patients for sacroiliitis with MRI at the time of diagnosis, especially those who are HLA-B27 positive and have elevated CRP levels.

Whole-body (WB) MRI has also been used to assess the distribution of disease activity in ERA. One study demonstrated that WB-MRI was superior to clinical exam for the detection of hip, sacroiliac, and spinal arthritis in JSpA [20]. Another study demonstrated poor agreement between clinical exam and WB-MRI for the detection of enthesitis in patients with JSpA [21]. Therefore, WB-MRI may play an important role in conjunction with clinical exam and radiography as an objective tool for assessing disease activity in children with JSpA, especially in the setting of a clinical trial [16]. Further research is still needed to evaluate the clinical scenarios in which WB-MRI might be more useful than dedicated MRI in JSpA.

MRI, especially considering sedation costs in younger children, is expensive. The ability to detect and treat early sacroiliitis, however, may be cost-effective in the long term and may prevent or diminish the consequences of axial damage in the growing child.

Measurements of disease activity

  • Children with ERA have been reported to have higher disease activity and poorer prognosis than other categories of JIA; therefore, disease activity measures that address symptoms specific to ERA are important in monitoring these patients.

  • The Juvenile Arthritis Disease Activity Scores (JADAS) and the JSpA Disease Activity Score (JSpADA) are two useful disease activity assessment tools, the latter of which is more specific to ERA.

Juvenile Arthritis Disease Activity Scores

The JADAS is a composite score consisting of four elements: the physician assessment of disease activity, parent/patient global assessment of well-being, erythrocyte sedimentation rate (ESR), and active joint count (in 10, 27, or 71 joints) [22]. The validation study for the JADAS included children with ERA, but they were a minority (<1 % of subjects). Cutoff values for defining remission, minimal disease, and acceptable symptom state with the JADAS have been validated [23]. A three-item JADAS without the sedimentation rate (JADAS3) correlated well with the original JADAS [24], suggesting that the simplified tool is sufficient for robust assessment of JIA disease activity if laboratory tests are unavailable [25]. Another version of the JADAS (JADAS-CRP), using the CRP in lieu of the ESR, was also found to be clinically effective in monitoring disease activity and correlated closely with the JADAS based on ESR [26, 27].

Juvenile Spondyloarthritis Disease Activity Score

The JSpADA is the first disease activity assessment tool developed and validated for use in JSpA (which includes ERA) [28]. It is a continuous disease activity score that was retrospectively validated in a multicenter cohort of children. This index includes eight equally weighted items: (1) active joint count, (2) tender entheses count, (3) clinical sacroiliitis, (4) morning stiffness, (5) patient assessment of pain, (6) uveitis, (7) back mobility, and (8) inflammatory markers. All items are transformed to values of 0, 0.5, or 1, and the total score ranges from 0 to 8. The JSpADA specifically includes measures of axial symptoms and enthesitis, which have been shown to independently predict poorer outcomes in JSpA [29]. The strengths of this tool include the limited number of items, inclusion of disease features specific to ERA, and the feasibility of assessing all of the items during the limited time of a routine clinic visit. This disease activity tool needs to be validated in a prospective sample, and cutoff values defining remission and minimal disease activity should be determined.

Role of the microbiome and considerations for diet modification

  • The close relationship between inflammatory bowel disease and SpA has highlighted the role of the gut microbiome in the etiopathogenesis of SpA.

  • A better understanding of the link between microbial dysbiosis and SpA may lead to the development of novel therapeutic approaches for the treatment of ERA [30].

Gut microbiome and starch

Approximately two thirds of adults with SpA have inflammatory intestinal changes similar to those detected in inflammatory bowel disease [31]. Similar prevalence of intestinal inflammation was reported in a pediatric study [32]. The true prevalence of inflammatory bowel disease in JSpA has yet to be determined, but it is likely very common as evidenced by research demonstrating subclinical IBD in JSpA [33]. The gut microbiome is the microbial community that resides in the intestines. Gut inflammation is thought to either cause or be a product of permeability of the epithelial lining of the gut, leading to loss of mucosal tolerance. Some hypothesize that HLA-B27 leads to mucosal immunodeficiency secondary to effects on intestinal permeability or alterations in the gut microbiome such as a loss of protective bacterial species [34]. Stoll et al. demonstrated that in comparison to controls, ERA patients had decreased levels of Faecalibacterium prausnitzii in their stool [35••]. This bacterium is known to have anti-inflammatory effects, and decreased levels have been demonstrated in the stool of patients with inflammatory bowel disease [36]. On the other hand, the presence of Klebsiella pneumoniae is suspected to be a causative agent in the development of SpA [37]. Klebsiella growth in the colon appears to be dependent on starch [37]; therefore, one might hypothesize a role for decreased starch consumption. However, there are no studies to date regarding diet modification in ERA. Whether it be through diet or new targeted therapies, there is future promise that recalibration of the gut microbiome may have a beneficial impact on ERA.

Treatment

Currently, most pediatric rheumatologists determine a child’s treatment regimen based on the number of affected joints and the presence of axial disease. The 2011 American College of Rheumatology (ACR) recommendations [38] do not consider the treatment of children with ERA separate from those children with other categories of JIA. According to these recommendations, an initial trial of nonsteroidal anti-inflammatory drugs (NSAIDs), with or without intraarticular corticosteroid injection(s), is recommended in patients with four or fewer affected joints [38], particularly those with predominant enthesitis. For patients with five or more active joints, the initiation of methotrexate or other traditional disease-modifying anti-rheumatic drugs (DMARDS) such as sulfasalazine is recommended. For patients with sacroiliitis, treatment with TNF-α blockade is the first-line treatment and has also been found to be of benefit in the treatment of refractory enthesitis [8, 39–43]. More recently, there have been new emerging treatments for SpA including medications that target the IL-12/23 and IL-17 axis (ustekinumab and secukinumab, respectively) as well as phosphodiesterase 4 (PDE4) inhibitors (apremilast). Therefore, clinical trials are warranted to establish the efficacy of these new treatments in the treatment of ERA. Lastly, as our understanding of the role of the microbiome becomes better elucidated, there may prove to be some utility in dietary modifications or gastrointestinal-targeted treatments in the near future.

Pharmacologic treatment

  • The goal of therapy with pharmacologic agents for ERA is to alleviate pain and decrease inflammation at the entheses and the synovial lining to preserve joint function and improve mobility.

Nonsteroidal anti-inflammatory drugs

NSAIDs are effective for the short-term, more immediate relief of pain in ERA and are commonly used for such purpose. NSAIDs are particularly effective for those children with predominantly enthesitis. NSAIDs typically do not suffice as monotherapy for the treatment of active arthritis. Some studies in adult SpA have suggested that continuous use of an NSAID not only improves symptoms but also reduces the radiographic appearance of axial inflammatory lesions and may slow spinal radiographic progression, but this remains to be demonstrated in ERA [44, 45]. A recent study in adults, however, suggested that outcomes were similar for adults with AS who received scheduled versus on demand diclofenac [46]. These medications are generally inexpensive. Commonly used NSAIDs include piroxicam, diclofenac, and meloxicam. NSAIDs are contraindicated in patients with inflammatory bowel disease, glucose-6-phosphate dehydrogenase (G6PD) deficiency, or gastrointestinal ulcers. They should be taken with food to prevent gastritis, as gastrointestinal upset is one of their main side effects. Toxicity monitoring should include a complete blood count (CBC), creatinine measurement, liver function test (LFT), and urinalysis 4–6 weeks after initiation of treatment and then every 6–12 months thereafter.

Oral corticosteroids and intraarticular corticosteroids

Oral corticosteroids (CS) were previously used more frequently for the treatment of ERA but are no longer acceptable as monotherapy for persistent, active arthritis. Oral CS can be effective in ameliorating the symptoms of an acute flare and quickly restoring mobility. They are relatively inexpensive. However, there are significant risks of oral CS use. In the immediate period, patients may experience hyperactivity, insomnia, or transient hypertension. These aforementioned symptoms, along with diabetes, glaucoma, weight gain, increased appetite, and mood changes, are potential reversible side effects from oral CS. Pediatric-specific long-term side effects include delayed puberty and short stature. There are also some significant potential irreversible side effects including cataracts, striae, osteopenia, and avascular necrosis. Included in this category of treatments are intraarticular corticosteroids (IACS), which involves injection of triamcinolone hexacetonide, or triamcinolone acetonide directly into the joint space. Triamcinolone hexacetonide is no longer commercially available. Localized injections may spare the child from exposure to systemic medication. The cost of the intervention is dependent on whether the child requires sedation services, which is typically age-dependent. If sedation is not needed, the procedure can be easily performed in the physician’s office. The beneficial effects of an IACS injection are rapid (within days). The associated risks include infection, atrophy, hypopigmentation, chemical irritation, and calcium deposits. Especially for patients who present with oligoarticular (≤4 joints) arthritis, initiation of treatment with an IACS is strongly recommended, especially before the initiation of a DMARD.

Disease-modifying anti-rheumatic drugs

Methotrexate

Methotrexate (MTX) is the most commonly used DMARD for JIA, but there are no studies to support its use specifically in the ERA population. In randomized controlled trials of MTX in the other categories of JIA, MTX has shown significant improvement in joint count, patient/physician global assessment, and erythrocyte sedimentation rate (ESR) [47–49]. Unfortunately, MTX (and DMARDS, in general) has not been proven to be effective for axial disease or enthesopathy [50, 51]. Therefore, MTX is recommended for ERA patients who demonstrate peripheral arthritis without axial involvement. The optimal dose and route of administration, despite MTX’s common use, is still uncertain, and dosing varies among practitioners. Most physicians typically initiate treatment as a single weekly dose of 10 mg/m2 and titrate as needed up to 30 mg/m2 weekly. The route of administration of MTX is not standardized. Evidence suggests that bioavailability with oral dosing often does not increase significantly beyond 20 mg/m2 per week [52]. Our practice is to start patients on subcutaneous MTX to gain initial control of arthritis as subcutaneous MTX has better bioavailability and fewer side effects and may improve patient compliance [53]. Once remission is maintained, we wean the dose and transition to oral MTX as tolerated. Patients are advised to take folic acid 0.4–1 mg/day to help ameliorate the gastrointestinal side effects of MTX including nausea and vomiting. Additional side effects include hair thinning, oral ulcers, headaches, hepatitis, cytopenias, and pneumonitis. Medication toxicity monitoring should include a CBC, aspartate aminotransferase (AST)/alanine transaminase (ALT), and creatinine drawn 4 weeks after the initiation of treatment and, if normal, every 3 months thereafter. The cost of the medication ranges from inexpensive to moderate.

Sulfasalazine

Sulfasalazine (SSZ) is another DMARD that may be beneficial in the treatment of ERA [54, 55]. In a phase II, exploratory, 26-week prospective, randomized, double-blind, placebo-controlled trial of SSZ in active JSpA, SSZ improved both doctor and patient assessments compared to placebo [54]. This study was limited by its small number of patients but suggests that SSZ may be useful in JSpA. The usual dose is 30–50 mg/kg/day, and it is titrated over the course of several weeks. Patients are expected to demonstrate clinical improvement 6–8 weeks after initiation of treatment. Potential side effects include gastrointestinal issues, cytopenias, hypogammaglobulinemia, hepatotoxicity, and Stevens-Johnson syndrome. Toxicity monitoring labs include CBC, AST/ALT, and creatinine performed 4–6 weeks after starting treatment and every 3–4 months thereafter.

Anti-tumor necrosis alpha blockers

Infliximab, etanercept, and adalimumab

The class of biologic agents that block TNF-α is useful in the treatment of enthesitis and peripheral and axial arthritis in adults. Additionally, TNF-α blockers have demonstrated efficacy for arthritis and enthesitis as well as symptomatic treatment of axial disease in JSpA [8, 39–43]. Therefore, they are considered first-line treatment for ERA patients with axial disease but should also be considered for those with enthesitis or arthritis that is refractory to NSAIDs and/or DMARDs. Studies in adults suggest that early inflammatory lesions in AS resolve following anti-TNF-α therapy and that treatment slows the development of new syndesmophytes [56]. Delay in initiation of TNF-α blockers is also associated with increased odds of structural progression in adults [57]. Given the potential detrimental consequences of untreated axial involvement in growing, developing children, the use of TNF-α blockers in patients with axial disease is likely cost-effective despite the expense of these medications.

The three recommended TNF-α blockers are etanercept, adalimumab, and infliximab. The typical etanercept dose is 0.8 mg/kg/week subcutaneously (maximum 50 mg). The standard adalimumab dose is 40 mg subcutaneously every other week in patients weighing at least 30 kg. Recently, a phase III, multicenter, randomized, double-blind, placebo-controlled study of adalimumab was performed in pediatric patients with ERA [58••]. Mean percent change from baseline in active joint count at week 12 was greater in the adalimumab group versus placebo (−62.6 versus −11.6 %, p = 0.039). Additionally, improvement in the signs and symptoms of ERA was sustained with continued adalimumab therapy through week 52. The results of this study suggest that adalimumab is efficacious and safe for the treatment of patients with active ERA who have failed conventional treatments. The intravenously administered TNF-α blocker, infliximab, is dosed at 5–10 mg/kg/dose, administered at 0, 2 weeks, and then every 4–8 weeks. Infliximab is currently not FDA-approved for JIA. The potential risks and side effects of these medications include infection, cytopenias, hypersensitivity reactions, injection site reactions, psoriasis, demyelination, and malignancy. Given the risk of activation of latent tuberculosis with these medications, tuberculosis screening should be done prior to starting treatment. This can be done with either a tuberculin skin test or interference-gamma release assays (IGRAs), the latter of which is approved for use in children 5 years of age or older. Patients should have a CBC and CMP checked 4–6 weeks after initiation of treatment and then every 6 months. We strongly support the concomitant use of MTX with infliximab to help prevent the development of human anti-chimeric antibodies (HACAs) and preserve the efficacy of this medication.

Other biologic agents

The potential use of other biologic agents, including rituximab (B cell-depleting therapy) and abatacept (T cell co-stimulation inhibitor), has been examined in small open-label studies in adults with AS [59, 60]. Rituximab was not effective in patients who had prior exposure to TNF-α blockers but was modestly effective in TNF-naïve patients. Neither rituximab nor abatacept, however, was effective as first-line therapy in adults, and neither has been studied in pediatric ERA. Lastly, there have been two randomized placebo-controlled studies examining the use of tocilizumab (IL-6 blocker) in AS that failed to show efficacy [61].

Emerging therapies

  • Recently, the role of the IL-12/23 and IL-17 axis in the pathogenesis of SpA and drugs that target this axis are being studied.

  • Drugs used for the treatment of psoriasis, including apremilast (PDE4 inhibitor), seem to be promising for the treatment of SpA [62, 63].

Ustekinumab

Ustekinumab is an anti-IL-12/23 human monoclonal antibody. In a prospective, open-label, proof-of-concept clinical trial, it effectively reduced the signs and symptoms of active AS in 20 patients, including serum C-reactive protein level and active inflammation on magnetic resonance imaging (MRI) [64••]. In this study, patients were treated with 90 mg subcutaneously at baseline, week 4, and then every 4 weeks. This medication is expensive, and additional studies regarding its efficacy in pediatric ERA are warranted. It may, however, be considered for patients who fail the aforementioned treatments. Adverse effects include infections, nausea, injection site reactions, and allergic reactions.

Secukinumab

Secukinumab is an anti-IL-17A antibody that had favorable results in a proof-of-concept trial in AS, including the open-label extension phase up to 24 months [65, 66]. Available dosing recommendations are based on its use in PsA in adults: initially, 300 mg SQ once weekly for five doses and then once every 4 weeks. Similarly to ustekinumab, further investigative studies are warranted to determine appropriate dosing for use in children with ERA. The cost and side effects not only are also similar to ustekinumab but also include upper respiratory tract involvement such as cough and pharyngitis.

Apremilast

Apremilast is an orally available, small-molecule PDE4 inhibitor, which blocks the upstream activation of cytokines important in the pathogenesis of AS. In a double-blind, placebo-controlled, single-center, phase II study, patients with symptomatic AS with active axial disease on MRI were randomized to apremilast 30 mg orally twice daily or placebo over 12 weeks [62]. Apremilast was associated with a greater improvement from baseline for all clinical assessments compared with placebo but did not achieve statistical significance. Nonetheless, this study suggests future applicability of apremilast in the treatment of axial inflammation in ERA.

Conclusions

  • The diagnosis and accurate phenotyping of ERA can be facilitated by the use of imaging modalities such as USD, conventional MRI, and WB MRI.

  • Newly developed disease activity measures, including the JADAS and JSpADA, will allow physicians to better monitor ERA patients over time.

  • Further research into the role of the microbiome in the development of SpA will hopefully provide new and targeted therapies for ERA.

  • TNF-α blockers are first-line treatment for children with axial disease.

  • New potential treatment regimens include medications used for adult psoriasis and AS patients. These include ustekinumab, secukinumab, and apremilast.