Background

Paediatric inflammatory rheumatic diseases (PiRDs) are complex rare chronic inflammatory conditions with risk of chronic morbidity and mortality affecting infants, children and adolescents [1]. PiRDs include different heterogeneous disease groups, such as the juvenile idiopathic arthritis (JIA), connective tissue diseases (CTD), systemic lupus erythematosus (SLE), vasculitis, uveitis and autoinflammatory diseases (AID). JIA is one of the most common PiRD groups and can be divided into different subgroups according to the International League of Associations for Rheumatology (ILAR) [2, 3]. In paediatric rheumatology, (i) responsive and valid instruments to assess disease activity and (ii) standardized outcome measurements are important to achieve defined treatment aims, to avoid disease burden and to optimize patients care [4,5,6,7]. Treatment aims in PiRD patients include control of systemic inflammation, prevention of structural damage, avoidance of disease comorbidities and drug toxicities, improvement of physiological growth and development, increase of the quality of life and enabling participation in social life. To achieve these treatment goals, treat-to-target (T2T) strategies similar to those used in adult rheumatology have been implemented in PiRD management [8,9,10]. To reach the defined treatment targets, different levels of disease activity require different treatment selections and dose adjustments [11]. The cytokine modulating effects of biologic disease modifying antirheumatic drugs (bDMARDs) or Janus kinase (JAK) inhibitors have enabled T2T strategies, since they allow important inflammatory disease pathways to be targeted [12,13,14,15]. Over the past 15 years, bDMARDs use has become essential in paediatric rheumatology and has markedly improved clinical outcomes [13, 15,16,17,18,19]. However, off-label use in PiRD patients is still common [20,21,22,23,24,25,26].

Although some rheumatologic diseases occur in paediatric and adult patients, considerable differences in disease symptoms, disease course and disease activity might exist [27,28,29,30,31,32]. Moreover, some PiRDs are not common or well known in adulthood [33,34,35,36]. For example, the JIA associated uveitis is the most frequent and potentially the most devastating extra-articular manifestation of the JIA, commonly affecting children aged 3 to 7 years [37]. Additionally, Kawasaki Disease (KD) is an acute inflammatory febrile vasculitis of mainly medium-sized arteries that typically affects children younger than 5 years [38]. Furthermore, PiRD patients differ from adult rheumatology patients in several physiological aspects, due to age-dependent changes, maturation processes, differences in body composition and pharmacokinetic (PK) processes, such as drug absorption, distribution, metabolism, and excretion [39,40,41,42,43,44,45]. All these aspects are important factors to consider in diagnosis and treatment. This highlights that paediatric drug development cannot simply mimic development strategies for adults, but has to respect paediatric pathophysiology and specific paediatric disease characteristics [46]. Nevertheless, it is common to use the same bDMARDs and JAK inhibitors in paediatric and adult rheumatology, and most paediatric trials and dosing regimens are performed on the basis of existing adult data [47].

The goal of this review is to assess the current state of knowledge obtained from previously performed randomized controlled trials (RCTs) in PiRD patients treated with bDMARDs and JAK inhibitors. In addition, an overview of approved bDMARDs and JAK inhibitors from the Food and Drug Administration (FDA) and European Medicines Agency (EMA) in paediatric and adult rheumatology is provided.

Methods

Information sources and search

A systemic literature search was conducted on Medline via PubMed, the US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov), and the EU Clinical Trials Register (www.clinicaltrialsregister.eu). The MeSH terms used for electronic search on PubMed and ClinicalTrials.gov are detailed in the supplementary material (supplementary data S1). Similar search terms were used for the EU Clinical Trials Register. The statistical and clinical sections of the New Drug Approval (NDA) web pages of regulatory authorities in the US and Europe were reviewed for approved drugs (www.fda.gov, www.ema.europa.eu). Abstract searches were conducted after conferences, including the American College of Rheumatology (ACR), the European League Against Rheumatism (EULAR), the International Society of Systemic Auto-Inflammatory Diseases (ISSAID), Société Francophone pour la Rhumatologie et les Maladies Inflammatoires en Pédiatrie (SOFREMIP) and the Pediatric Rheumatology European Society (PRES). Additionally, relevant studies were identified by manual search of the bibliographies of references retrieved from PubMed. For all literature sources, only English articles were screened.

Eligiblity and exclusion criteria

RCTs of patients aged 20 years and younger treated with predefined bDMARDs/JAK inhibitors were included if the sample size was at least five patients, and if PiRD diagnosis had been confirmed. The PiRDs included are listed in Table 1. The drugs and drug classes reviewed included the following: (i) Anti-CD20 agents: rituximab; (ii) CD80/86 inhibitors: abatacept; (iii) IL-1 inhibitors: anakinra, canakinumab, rilonacept; (iv) IL-6 inhibitors: tocilizumab, sarilumab; (v) IL-12/23 inhibitors: ustekinumab; (vi) IL-23 inhibitors: guselkumab, risankizumab, tildrakizumab; (vii) IL-17 inhibitors: secukinumab, ixekizumab, brodalumab; (viii) Tumour necrosis factor (TNF) inhibitors: adalimumab, etanercept, golimumab, infliximab, certolizumab pegol; (ix) BAFF inhibitors: belimumab; (x) JAK inhibitors: baricitinib, tofacitinib, upadacitinib. Studies must have at least included a relevant primary or secondary efficacy endpoint/outcome as detailed in the supplementary material (supplementary data S3). Consequently, studies were excluded if (i) the indication was not relevant, (ii) the population was not relevant, including studies which enrolled both adults and children with PiRD, (iii) the study design was not relevant including observational studies, single arm studies, reviews, meta-analyses and pooled analysis of multiple RCTs, (iv) treatment was not that as predefined, (v) the endpoint/outcome was not relevant, and/or (vi) the report was a duplicate of prior published results without any additional information.

Table 1 Defined diagnoses of paediatric inflammatory rheumatic diseases (PiRD) for literature search
Full size table

Study selection and data collection process

Initial screening, based on retrieved abstracts, as well as the eligibility assessment based on full-text publications were performed by two independent review authors according to a review protocol (supplementary data S2). One scientist was responsible for the execution and documentation, and the other provided support as therapeutic area expert. Any discrepancies were resolved through discussion or consultation with a third independent reviewer. The primary reason for exclusion was documented for each excluded reference. Aggregate (summary) level data were extracted from each included trial by two independent review authors. The defined extracted variables for each RCTs included baseline demographic and clinical characteristics such as study design, location, patient population, sample size, age criteria, treatment and primary outcome/endpoint.

Results

Study selection

A systematic literature search was performed on July 26, 2020 using the predefined search criteria. A total of 608 references were screened, and 65 references for 35 uniquely identified RCTs performed in PiRD patients were selected for inclusion (Fig. 1). Of the references excluded, the large majority were due to irrelevant indication (56%, 302/543) or population (31%, 169/543). In total, the majority of RCTs were performed for TNF inhibitors (60%), IL-1 inhibitors (17%) and IL-6 inhibitors (9%). Only one RCT was available for the BAFF inhibitor belimumab, the JAK inhibitor tofacitinib, the IL-12/23 inhibitor ustekinumab, the IL-17 inhibitor ixekizumab, the TNF inhibitor golimumab and the CD 80/86 inhibitor abatacept. No RCTs with published results were identified for the anti-CD20 agent rituximab, the TNF inhibitor certolizumab pegol, the IL-6 inhibitor sarilumab, the IL-17 inhibitors brodalumab and secukinumab, the IL-23 inhibitors guselkumab, risankizumab and tildrakizumab and the JAK inhibitors baricitinib and upadacitinib (Table 2). Currently, there are no recruiting RCTs for sarilumab but there are two ongoing single-arm PK studies in sarilumab, NCT02776735 and NCT02991469 (data not shown). For the JAK inhibitors three Phase III RCTs are investigating baricitinib, and one Phase III global RCT is investigating tofacitinib in JIA patients (Table 3). Furthermore, there are ongoing studies for guselkumab, risankizumab, tildrakizumab, brodalumab, secukinumab and certolizumab pegol mainly in PiRD patients with psoriasis (Table 3). Additionally, some studies are recruiting to investigate adalimumab in JIA-associated uveitis, abatacept in OJIA, infliximab in KD and etanercept in OJIA and PJIA.

Fig. 1
figure 1

Flow chart of study selection and reasons for study exclusion after literature search

Full size image
Table 2 Overview of completed randomized controlled trials performed in paediatric rheumatology patients treated with bDMARDs/JAK inhibitors (July 2020)
Full size table
Table 3 Ongoing or recruiting studies in paediatric patients with inflammatory rheumatic diseases (July 2020)
Full size table

Study characteristics

Approximately two-thirds (25 out of 35) of the identified RCTs were conducted in JIA patients and the remaining ten RCTs were performed in non-JIA patients, including KD, plaque psoriasis, SLE, and non-infectious uveitis (Tables 4 and 5). The mean/median age of children enrolled in the JIA RCTs ranged from 8 years to 15.3 years. In contrast, the non-JIA patients included in RCTs had a mean/median age range varying between 2.2 and 15.2 years, with KD patients being younger (range 2.2 to 3.7 years). In JIA RCTs, the control was mainly placebo, and the concomitant background treatments were usually either methotrexate, NSAID or corticosteroids, whereas in non-JIA trials the control arm was a mixture of placebo or standard of care treatments and patients received mostly NSAID as background treatments (data not shown for the control arm). The primary efficacy outcome/endpoint in the JIA RCTs was mainly ACR Pedi 30/modified ACR Pedi 30 or disease flare (Table 4). Other instruments to assess the primary outcome were count of joints with active arthritis, the assessment of Spondyloarthritis International Society 40% score (ASAS 40), inactive disease, treatment failure and improvement of laser flare photometry (Table 4). In non-JIA patients, efficacy outcomes/endpoints varied due to heterogeneous subgroups. The primary efficacy outcome/endpoint of RCTs in KD was mainly related to fever, whereas for plaque psoriasis the Psoriasis Area and Severity Index (PASI 75), or the Physician Global Assessment (PGA) was used (Table 5). The RCT addressing SLE used the SLR response index (SRI 4), whereas the primary outcome/endpoint in non-infectious uveitis was assessed with uveitis disease activity using the Standardization of Uveitis Nomenclature (SUN) criteria, AC cells and vitreous haze. The majority of the JIA RCTs were global studies or otherwise conducted in either Europe or the United States, with one study (NCT00144599) located in Japan (data not shown). The non-JIA RCTs took place either in North America, Europe or globally (data not shown). In particular, KD RCTs took place mainly in Asia or the United States. Details for JIA and non-JIA RCTs are shown in Tables 4 and 5. All non-JIA studies were of a parallel study design, while in JIA studies there was a mixture of parallel and withdrawal study designs (Tables 4 and 5). The main conclusion of the majority of the studies (in terms of meeting the primary endpoint/outcome) was that the bDMARDs evaluated were more effective in comparison to placebo or standard of care (Tables 4 and 5).

Table 4 Overview and general characteristics of identified reviewed randomized controlled trials performed in JIA patients (July 2020)
Full size table
Table 5 Overview and general characteristics of identified reviewed randomized controlled trials performed in non-JIA patients (July 2020)
Full size table

Approved bDMARDs and JAK inhibitors in paediatric and adult rheumatology

In March 2020, the FDA has approved all 23 reviewed drugs, including bDMARDs and JAK inhibitors for adult rheumatology, whereas the EMA has approved 22 (Table 6). For PiRD patients, 10 bDMARDs (EMA) and 11 (FDA) have been approved (Table 6). Not surprisingly, the more recently approved bDMARDs in adult rheumatology and the JAK inhibitors have mostly not yet been approved for PiRD patients. Infliximab is approved for several rheumatologic indications in adulthood including rheumatoid arthritis (RA), PsA, ankylosing spondylitis, and plaque psoriasis, but is not approved for any PiRD indication so far. In paediatrics, infliximab is still restricted for in-label use in paediatric chronic inflammatory bowel diseases. Furthermore, there are some differences between the FDA and EMA in bDMARDs and JAK approvals. For example, the FDA has approved rilonacept for the treatment of the Cryopyrin-associated periodic syndrome (CAPS) in adults and children aged 12 years and older, whereas EMA has not. Particularly relevant for the PiRD patients are the different age limitations for different bDMARDs, and varying age restriction for different PiRD diagnoses. No bDMARDs are approved in children younger than 2 years with the exception of anakinra, which is approved by the EMA for the age ≥ 8 months. The age limitations have changed over the last couple of years, and today the common age categories are ≥2, ≥4, ≥6, or ≥ 12 years. A detailed overview about paediatric bDMARDs approvals, indications and age limitations by the EMA and FDA in March 2020 compared with adult rheumatology is given in Table 6.

Table 6 Overview of bDMARDs and JAK inhibitors approved in adult and paediatric rheumatology (March 2020)
Full size table

Discussion

This review indicates that reported data from RCTs characterizing efficacy, safety and/or PK, remains limited for several prescribed bDMARDs and JAK inhibitors in PiRD patients. As RCTs are robust research methods to determine cause-effect relationships between intervention and outcome, they are important to generate evidence in basic, translational and clinical research and can improve management of patients [112]. In the past, several clinical trials were conducted in PiRD with support of research networks in paediatric rheumatology, such as the Pediatric Rheumatology Collaborative Study Group (PRCSG) and the Paediatric Rheumatology International Trials Organisation (PRINTO) resulting in bDMARDs approval for some PiRD indications [19, 113]. This review indicates that TNF inhibitors are the most studied bDMARDs in PiRD patients, particularly in the JIA group. JIA is one of the most commonly diagnosed PiRDs with a prevalence of 16/100,000 to 150/100,000 [3]. In several JIA sub-groups, treatment with TNF inhibition is recommended, particularly when conventional disease modifying antirheumatic drugs (cDMARDs) cannot achieve the defined target [114, 115]. One of the first FDA-approved TNF inhibitors for polyarticular JIA treatment was etanercept in 1999, followed by adalimumab in 2008. This might explain why a majority of RCTs were performed for etanercept. Up to now, no JAK inhibitor is approved for PiRD patients. JAK inhibitors can be administered orally and therefore this treatment approach might be of particular interest in paediatric rheumatology, explaining why several RCTs are currently performed for JAK inhibitors. For several bDMARDs, a latency in drug approval for PiRD patients can be observed with a delay ranging between 1 year to 9 years. However, around 50% of reviewed therapeutic drugs are currently not approved for PiRD patients. Off-label and unlicensed drug use is frequent in paediatric patient populations [116, 117] and a considerable number of PiRD patients has to be treated with off-label bDMARDs or JAK inhibitors as no approved drugs are available for their age group, the PiRD indication or in general [20,21,22,23, 25]. Off-label use is often of great concern to the families of the affected children [17]. In addition, it seems that off-label and unlicensed drug use in children is associated with increased risk of medication errors and adverse events [118,119,120]. As infants and children with PiRD differ greatly from adult rheumatology patients the lack of paediatric PK data for bDMARDs and JAK inhibitors, can result in over- and under-dosing [42, 47, 121]. While under-dosing/low drug concentrations can result in drug-antibodies and drug insufficiency with uncontrolled chronic inflammation and disease burden, over-dosing can be associated with serious short- and long-term safety events [122,123,124]. There are data suggesting that based on the body weight, the clearance of several drugs is higher in paediatrics than in adults [39]. In PiRD patients, particularly in infants and younger children, there are data for bDMARDs and JAKs indicating a need for more frequent drug administration due to shorter half-life or the need for higher weight based drug dosages to achieve the defined therapy target [121, 125,126,127]. Moreover, it seems that subcutaneously administered bDMARDs are absorbed faster in young children [44]. As the therapy outcomes in PiRD patients is influenced by these age-dependent PK processes and the disease course, it is crucial to understand the developmental changes to optimize bDMARDs and JAK inhibitor dosing in paediatric rheumatology [39,40,41,42, 44, 46, 47]. As a consequence, the FDA Modernization Act stimulates the conduct of dedicated clinical studies to enhance understanding of PK, efficacy-safety balance, and optimal dosing of drugs in paediatric patients [128]. Nevertheless, concern has been raised that trial discontinuation, and nonpublication with associated risk of publication bias, seems to be common in paediatric patients [129, 130]. Slow recruitment rates in rare paediatric diseases can be challenging for paediatric trials, and poor recruitment seems to be one of the major risks for early termination or discontinuation of such studies [130]. These observations highlight the value of established research networks in paediatric rheumatology, such as PRCSG and PRINTO, in conducting clinical studies in PiRD patients as efficacy, safety and PK data obtained from PiRD patients to optimize treatment are warranted.

This review has several limitations. Despite a comprehensive search strategy and independent reviewer processes, there might be a risk of a reporting bias as unpublished RCTs were not included. Furthermore, this review does not include observational studies, single arm studies or RCTs including both children and adults. We cannot rule out that not all conducted RCTs in PiRDs were identified, despite a rigorous screening and review process. We have included RCTs with patients aged 20 years and younger, although this upper age limit of 20 years constituted the risk of having studies performed mainly in adolescents and young adults. To address this bias we have reported for each analysed RCT the age criteria and the median/mean age. As several included RCTs had an upper age criteria between 17 to 20 years, we would have missed otherwise these studies if we have limited the search to the age criteria 16 or 18 years.

Conclusion

In summary, paediatric rheumatology patients differ from adult rheumatology patients in many aspects. As therapeutic drug response is influenced by age-dependent PK processes and disease course, it is important to consider developmental changes when prescribing bDMARDs or JAK inhibitors in PiRD patients. As such, it is critical to conduct international multicentre studies in PiRD patients to enroll a sufficiently high patient number in a reasonable period of time with the goal to appropriately investigate and characterize PK, efficacy and safety for bDMARDs and JAK inhibitors. More efficacy and safety data, ideally combined with PK data from PiRD patients will optimize bDMARDs and JAK inhibitor use in paediatric rheumatology.