Differences and similarities of SLE with onset during adulthood and cSLE

The development of clinical and laboratory manifestations with SLE and cSLE is thought to be mediated by environmental and lifestyle factors in genetically predisposed individuals. Monogenic genetic causes of cSLE are especially common among individuals with disease onset by 5 years of age.5 In those with disease onset beyond age 5 years but still early in life, there is likely a higher load of the aforementioned factors and, possibly, pathological changes of immune system development.5–8 Compared with individuals with disease onset during adulthood, children with cSLE have more commonly multiorgan disease, acute disease onset and ongoing active inflammation over time. As such, lupus nephritis is estimated to be at least 30% and neuropsychiatric involvement about 25% more common in cSLE than with adult-onset disease9; this necessitates the more frequent chronic use of corticosteroids and immunosuppressive drugs in cSLE, and there are higher rates of hospitalisations, which likely damage acquisition and mortality in cSLE, with a large variability worldwide.6 10 Despite these dissimilarities, there are no laboratory abnormalities, signs or symptoms that only occur in cSLE but not in SLE with onset during adulthood.

Epidemiology of cSLE

There are considerable differences in phenotypic presentations, incidence and prevalence rates of SLE around the world.11 12 About 10%–20% of a global lupus population has had disease onset during childhood.13 The estimated prevalence of cSLE is at 9.73 (95% CI 9.38 to 10.08) per 100 000 persons in the USA,13 which compares with the prevalence of juvenile idiopathic arthritis (JIA) at 44.7 (95% CI 39.1 to 50.2) per 100 000 persons.14 In Europe, JIA is more common at a prevalence of 70.2 (95% CI 62.9 to 78.1) per 100 000 persons,15 and cSLE is even less prevalent with estimates of 4.3 (95% CI 1.4 to 14) per 100 000 persons.16 17 Isolated subcutaneous or discoid lupus occurs rarely in children,18 19 and the disease is at least five times more common among girls.13

Regulation of drugs for children with paediatric rheumatic diseases

The Best Pharmaceuticals for Children Act (BPCA)20 and the Pediatric Research Equity Act are especially relevant to paediatric drug development in the USA.21 The latter requires new drugs and biologic therapies to be tested in children, provided there is a paediatric disease that is similar to a non-orphan disease occurring in adults.22 BPCA provides pharmaceutical companies with an additional 6 months of market exclusivity for the adult indications after the completion of drug studies in children performed at the request of FDA.23 A key document that FDA requires pharmaceutical companies to develop is the Pediatric Study Plan (PSP) on completion of phase 2 studies and before the initiation of phase 3 studies in adults.24 In the European Union (EU) the EMA is responsible for the scientific evaluation, supervision and safety monitoring of medicines in 28 EU member states, as well as the countries of the European Economic Area. In 2006, the European Parliament passed legislation relating to paediatric drug testing and approval, similar to that in place in the US pharmaceutical companies are required to submit a Paediatric Investigational Plan (PIP) to the Paediatric Committee, 25 26 irrespective of whether the adult disease is regarded an orphan disease or not. Both the US and the EU paediatric legislation has been a great success in that several medications received licensing for use in children, especially in paediatric rheumatology and infectious diseases.27 EMA and FDA regulations require that age-appropriate preparations of medication are made available.28 29 For off-patent medicines, EMA may grant a Paediatric Use Marketing Authorisation, with patent protection for 10 years, if an indication together with an appropriate dosage form or formulation specifically devised for children has been developed.30

Considerations of ethical principles and use of prevailing knowledge in cSLE drug development

For enhanced acceptance of data generated in paediatric global drug development programmes by regulatory agencies, such as EMA an FDA, and to ensure timely access to medicines for children, the International Council on Harmonisation (ICH) of Technical Requirements for Pharmaceuticals for Human Use has recently updated its E11 guideline entitled, ‘Clinical Investigation Of Medicinal Products In The Pediatric Population’. 31 32

Besides ethical principles for research in paediatrics as are summarised in figure 1, the ICH stresses that children should not be enrolled in a clinical study unless necessary to address an important paediatric public health need; and that risks and benefits of research participation need to be carefully assessed. This also implies that paediatric drug development programmes are cognisant of study feasibility and consider paediatric extrapolation (figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Ethical principles of medication studies in children.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Factors influencing paediatric extrapolation strategies. The paediatric extrapolation from adult to paediatric diseases assumed similar pathoaetiology and response to therapy are present. Existing data are appraised to determine which new data are needed to determine safety and efficacy of a medicinal product in childhood-onset systemic lupus erythematosus (cSLE). The approach to acquiring the needed information about a given medicinal product is determined accordingly, also considering feasibility and stakeholder input.

‘Paediatric extrapolation’ is defined as an approach to providing evidence in support of effective and safe use of drugs in the paediatric population when it can be assumed that the course of the disease and the expected response to a medicinal product is sufficiently similar in the paediatric and the reference (adult or other paediatric) population.32 Thus, previously acquired data from studies in adults with SLE and other diseases, information collected from other paediatric populations and preclinical data are all considered in the extrapolation concept for cSLE.33 While input from clinicians and clinical trial experts has been commonly sought in the past, the importance to also include the patient perspective and experience is now recognised.34 Based on the appraisal of the available body of knowledge, feasibility and stakeholder input, the paediatric drug development programme (PIP, PSP) will be aimed at closing a specific knowledge gap to help determine whether a medicinal product would be beneficial when used in cSLE or not.

Lupus-specific recommendations for the study of medications for children and adults by regulatory agencies

Specific guidance documents relevant to studies in SLE have been issued by FDA and EMA.35 36 Only the EMA document provides guidance for trials in lupus nephritis. The FDA guidance document states that the most feasible and ethical approach to test a new medicine in SLE is likely to add a new medicine to current standard of care therapy. Considering study size, FDA guidance also states that the most realistic approach is to conduct a superiority study.35 In general, the minimum duration of an SLE trial is 1 year for evaluating the endpoint of reduction in disease activity, complete clinical response or remission, reduction in flare/increase in time to flare and maintenance of response. FDA recommends the use of the American College of Rheumatology (ACR) Classification Criteria37 for identifying SLE subjects for trials, while the respective EMA guidance document36 also supports the use of the newer Systemic Lupus International Collaborating Clinics (SLICC) Classification Criteria for SLE.38 Both classification criteria sets have been validated for use in cSLE.39 40 Interestingly, the EMA specifically states that its guidance paper does not apply to subsets of SLE, for example, neuropsychiatric SLE and secondary antiphospholipid syndrome, in lieu of difficulties in making a diagnosis and/or the absence of validated efficacy assessment tools. Nonetheless, EMA encourages the inclusion of patients with these SLE subsets in trials.36

Similar to the EMA guidance document,36 there is a short paragraph (section 7) in the FDA Guidance document that speaks specifically to cSLE.35 Herein, EMA and FDA concur that, compared with adult-onset SLE populations, there is an increased male-to-female ratio, a higher prevalence of kidney and neuropsychiatric involvement as well as faster accrual of damage in cSLE. Biomarker use is encouraged by both agencies in studies of new medicines for SLE and cSLE.28 29 33 35 36 Further, EMA’s guidance paper points out the rarity of cSLE and that a waiver could be granted for children under the age of 5 years.36 Waivers may also be granted when a paediatric development is not needed or appears inappropriate. Regulatory agencies often provide a deferral, that is, cSLE studies are postponed until there are sufficient data to demonstrate the efficacy and safety of a medicinal product in adults with SLE.

As can be deduced from the above, the current medication development for children with cSLE relies heavily on evidence around medication safety and efficacy in adults. Notably, safety of medicines, when used in children, can never be fully extrapolated from adult data, as the impact of a medication on growth and development cannot be studied in adults. This mandates generally the conduct of long-term postauthorisation studies and establishment of patient registries in cSLE.

Trial designs found successful for the study of rare diseases such as cSLE

The cSLE drug development programme deemed necessary by EMA (PIP) and FDA (PSP) will depend on the results of paediatric extrapolation considerations, feasibility and differences in-between cSLE and SLE relevant for a given medicinal product’s method of action (figure 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Extrapolation approaches—possible design of paediatric studies based on the results of the paediatric drug development must be sufficient to support the therapeutic benefits and appropriate dosing of a medicinal product. Paediatric extrapolation assessment will inform about appropriate approaches to studying safety and efficacy of a medicinal product when used in childhood-onset systemic lupus erythematosus (cSLE). Approaches may change over time as new data (learning) become available. Feasibility of conducting medication studies in cSLE will need to be considered. PK/PD, pharmacokinetic-pharmacodynamic study.

The prototypic approach to testing medication efficacy is a parallel-arm, double-blinded, placebo-controlled randomised controlled trials with a fixed target sample size. This approach will face considerable challenges when used in cSLE as is exemplified by the double-blinded, placebo-controlled trial of belimumab in cSLE (NCT01649765). Although the concept of assembling a placebo cohort is ideal from a scientific point of view to assess drug efficacy while minimising sample, as opposed to an active arm comparison, the aforementioned trial required a 12-month exposure to placebo infusions of children as young as 5 years, and use of advanced analytic methods was not planned. As a result, enrolment was markedly delayed. The study also lacked power to assess differences in major safety events or differences in efficacy compared with adult SLE. This raises ethical concerns which, coupled with feasibility issues, have already delayed the access of children to belimumab.

Recently, much attention has been paid to adaptive clinical trials, that is, trials with prospectively planned modifications to the study design while preserving the scientific validity and integrity of that trial. Using adaptive design elements in clinical trials is expected to enhance the proficiency with which drug efficacy is shown.41 42 Indeed, adaptive design elements are often used in studies of uncommon diseases. Examples include adaptive randomisation to minimise imbalance in baseline covariates among treatment groups and/or increase the proportion of patients assigned to the presumed more effective treatment, while reducing overall trial enrolment (response-adaptive randomisation). In sequential adaptive trials, data are analysed intermittently to guide decisions on termination when safety concerns, futility, efficacy or a combination of these factors is demonstrated. Trials that are stopped early because of important interim results require fewer patients.

Other clinical trial strategies that have been favoured in medication studies of rare diseases are summarised in figure 4.43 They include seamless phase 2–3 designs where phase 2 and phase 3 studies are combined so that some patients can participate in both phases. Study designs where all study participants receive active study drug help enhance recruitment.43 The same holds true for studies that offer a high proportion of participant’s active drug over prolonged time periods. This can, for example, be achieved by using a randomised withdrawal design (RWD),43 or in observational studies where all patients receive active study drug. However, for observational studies, issues of confounding will need to be carefully addressed to delineate drug effectiveness by advanced statistical methods, such as propensity scores, or ‘new-user’ studies: inception cohorts permit investigators to establish clear temporality among study variables, that is, baseline confounders, exposures and outcome events that occur after entry to the cohort. Cross-over designs are also often used for the study of rare diseases. Although this design was successfully employed for the study of rilonacept in familial Mediterranean fever,44 it is unlikely to be useful in cSLE, among others, due to the variability of disease manifestations over time. Notably, many of the aforementioned clinical trial strategies have been successfully used in other paediatric rheumatic diseases.45–47

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Strategies for successful clinical trials in rare diseases. Successful trials of medications in rare or uncommon disease must consider limited numbers of patients who can be enrolled in a study. Access to patients may be increased through research networks and by considering the input of all stakeholders, that is, families, patients, clinicians treating patients and clinical trialists. Maximising on-treatment time will further enhance the amount of information regarding the drug effect and increases interest in research participation. Advanced trial methods can further increase the efficiency of a trial.

Validated core outcome and response measures to quantify the effects of new medicines

Another method to reduce sample size requirements in rare disease studies is through the selection of accurate outcome measures, that is, surrogate and biological markers that are highly sensitive, specific and responsive to change. Indeed, performing medication studies in children without well-validated outcome measures may be considered unethical. In the EMA and FDA guidance documents for SLE,35 36 reference is made to the cSLE core set domains which were established by the Paediatric Rheumatology International Trials Organisation (PRINTO, www.printo.it), in collaboration with the Pediatric Rheumatology Collaborative Study Group (PRCSG, www.prcsg.org). The five cSLE core set domains are disease activity, renal function, patient well-being, physician’s global assessment of cSLE activity and health status. EMA endorses the use of the PRINTO/ACR Response Criteria,48 while the FDA considers their use as exploratory, pending additional validation studies. Since the time of the publication of the FDA guidance document, additional validation of the PRINTO/ACR Response Criteria has occurred49 which may suggest that there is consensus of both agencies for the use of the PRINTO/ACR Response Criteria in cSLE clinical trials. Despite its apparently lower accuracy compared with the PRINTO/ACR Response Criteria, the Systemic Lupus Responder Index is appropriate for use in cSLE.49 There are also validated criteria for flare of global disease with cSLE.50 51 Regulatory agencies consider the avoidance of disease flares important,35 36 and flare criteria are essential for medication studies using an RWD. Table 1 summarises some of the important measures of disease activity, disease damage and patient-reported outcomes that are validated for use in cSLE.

Organ-specific assessments of cSLE response to therapy

Specific organ system involvement with cSLE differentially impacts health-related quality of life and/or importantly influences long-term prognosis. Accordingly, particular attention needs to be paid to evaluating the organ-specific impact of medicinal products.52 53 Organ-specific outcome measures for musculoskeletal manifestations, cutaneous lupus manifestation, neuropsychiatric cSLE and paediatric lupus nephritis should be collected. Consensus has been achieved among North American and European paediatric rheumatologists of how to capture changes in paediatric lupus nephritis activity.54 55 More recently, urine biomarkers of lupus nephritis have been discovered and validated for use in children and adults. The Renal Activity Index for Lupus quantifies lupus nephritis activity based on levels of several protein biomarkers measured in the urine.56–58 For capturing joint inflammation with cSLE, the assessment of number of joints with active arthritis and joints with limited range of motion, as done for JIA,59 has face validity. Despite the variability of neuropsychiatric manifestations with cSLE, cognitive ability is generally considered a global measure of brain health. A standardised battery of age-appropriate formal neuropsychiatric tests has been defined for cSLE,60 and the Pediatric Rheumatology Assessment Metrics software (PedANAM, Vista LifeSciences, Oklahoma, USA) is available for estimating cognitive ability.61–63 Notably, there is no highly specific diagnostic test for neuropsychiatric SLE in children and adolescents.

Special issues around study designs of cSLE

To avoid potential delays in drug availability for children, studies in cSLE should commence soon after efficacy and sufficient safety is demonstrated in phase 3 trials in adults for an agent in a new drug class, and after phase 2 completion for drugs within a class of medicines with proven efficacy in adults with SLE.

There is agreement among paediatric rheumatologists regarding principles of therapeutic and disease monitoring standards for cSLE.64 65 This is relevant for the design of cSLE trials, together with current principles of off-label medication use and standards of medical care.66–68 Regulatory agencies support the enrolment of adolescent patients with cSLE in adult SLE trials. However, differing standards of medical care between cSLE and adults with SLE exist,54 64 and enrolling adolescents into adult SLE trials is likely difficult, based on the experience gained in JIA.

In line with suggestions from regulators, global enrolment stratified by country of origin and race seems ideal to enable sensible subanalyses to test for potential differences in drug efficacy based on racial, ethnic or regional factors and gender effects in cSLE.8 28 29 69 Consistent access to high-quality data from well-designed geographically matched cSLE and SLE registries to define standards of care and provide evidence of similar response to current treatments across all ages of patients with lupus would much facilitate any clinical trial design of cSLE.

Enrolment of patients with cSLE into traditional double-blind, placebo-controlled parallel design studies is not favoured by paediatric rheumatologists. This is supported by the results of recent surveys regarding hypothetical future medication trials in cSLE. Survey respondents were 192 PRINTO and 161 PRCSG investigators from over 40 countries, who participated in several trials of the two networks (table 2). Over 70% of the survey responders preferred an open-label PK/PD study while only 21%–30% favoured a blinded, parallel, placebo-controlled study. Despite enthusiasm about testing new medicines for cSLE and the potential to enrol over 1000 study participants with active cSLE over a 12-month period, prolonged placebo exposures were only agreeable to a minority of the PRCSG/PRINTO survey responders.

If regulators deem blinded or controlled studies necessary to address the existing gap in scientific knowledge about a new medicine for cSLE, then trial designs seem advantageous that have proven to be successful when used in other rare diseases.70 71 These Pediatric Study Plans: Content of and Process for Submitting Initial Pediatric Study Plans and Amended Initial Pediatric Study Plans Guidance for Industry include RWD trials (see online supplementary figure 1) which provide instant open-label active study drug to all patients during the lead in phase (part 1) and limit potential placebo exposure to patients who have experienced improvement during part 1 of the study, whereas children who fail to respond to study drug are generally discontinued from the study prior to randomisation.

The primary endpoint of an RWD study is the proportion of patients with ‘disease flare’ or the time to ‘disease flare’ in part 2. Therefore, patients remain in the blinded part 2 only as long as their disease continues to demonstrate at least a similar level of cSLE control as was present at the time of randomisation. RWD trials used in paediatric rheumatology only require mild to moderate worsening during part 2 for a patient to again receive active drug in part 3 of the trial. Seemingly, the overall burden of disease activity rather than short minor to moderate global flares of cSLE carries a sizeable risk for disease damage.72–74 Nonetheless, appropriate trial discontinuation rules, provision of rescue medication and close patient follow-up will all be needed to avoid long-lasting sequelae of flare.

Despite its attractiveness, the RWD has its shortcomings which will need to be carefully considered.71 75 They include carry-over effects which may reduce the likelihood of flaring in patients switched from placebo. RWD trials are more difficult for testing medications with prolonged biological effects or drug which are only given intermittently, such as rituximab. Further, RWD trials only provide indirect evidence of drug efficacy, given that flare rather than response to therapy is the primary outcome.

Chronic use of corticosteroids remains a major concern, given the well-known detrimental effects on growth and pubertal development of children beyond the side effects encountered by adults.76 Therefore, steroid tapering needs to be integrated in clinical trial designs to address concerns about prolonged corticosteroid exposures with cSLE. This may be achieved by a sufficiently long open-label lead period of an RWD trial to enable tapering to a relatively safe dose of 0.2 mg/kg/day or 10 mg/day (whichever is lower). Corticosteroid tapering cannot be easily implemented and adequately interpreted in parallel study design.

For the study of paediatric lupus nephritis in particular, exposure to placebo seems problematic when there is high renal activity. This is because prolonged uncontrolled lupus nephritis activity and proteinuria, presumably more common with receiving placebo, is a potent risk factor for poor prognosis.77

Adaptive design elements, such as early randomisation of responders and patients who tapered corticosteroids successfully, or adjustment of sample size based on response rates should all be considered for added efficiency when studying new medicines in cSLE.

Summary and interpretation

There is a dire need to develop new medicines for the treatment of cSLE in general and lupus nephritis in particular. Based on prior experience,78 specialised paediatric research networks that offer experience in clinical trial design and execution have coordinating centres that are proficient in overseeing clinical trial operations and are familiar with the validated assessments will be essential for the successful completion of cSLE trials. Sufficient numbers of patients with cSLE are available and experienced investigative teams are in place to perform pivotal studies with the scientific rigour needed to support subsequent market authorisation for general cSLE, associated skin, joint and kidney disease. Blinded studies should omit traditional parallel-arm placebo designs. It is noted that use of extrapolation plan including data from other sources, particularly from adult trials, together with open-label and innovative design studies will be more expedient in providing children with cSLE access to new medications.