Trials in rare diseases

A trial in a rare disease has to be designed for a population that is small, scattered and poorly characterised. The population, the design, the control and the endpoints have to be chosen for what so few patients can support, and the work depends on the patient community, on investigators and sites new to trials, and on a supply chain built for very small quantities.

Challenges of rare-disease trials

Trial design is complex in any context, and in a rare disease with a paediatric population it can seem insurmountable. ¹ The specific challenges of designing rare-disease trials include limited knowledge of the disease's natural history, small numbers available to take part, regulatory guidance not calibrated to the rare-disease context, manufacturing and supply problems, and safety and financial risks. ^{2, 3} The costs and other challenges of such trials are increasing, and evaluating an investigational treatment is particularly difficult when the small numbers affected limit recruitment. ³ So the randomised controlled trial, the established gold standard for evidence of efficacy and safety, can be difficult to undertake in rare diseases. ³ Industry, academia, regulators, healthcare providers and others need to collaborate to meet the need for new therapies, and the drug development and trial process is complex, with ethical, scientific, operational and regulatory considerations. ^{4, 5}

Limited research complicates finding relevant literature and developing a robust protocol. ⁶ Significant obstacles include identifying relevant endpoints, the limits on sample size and the cost of maintaining diverse clinical sites, and site inexperience and the specialised nature of rare-disease management complicate recruiting and retaining participants. ⁷ Success needs a significant upfront investment of time and money in building foundations with regulators, patient groups, key opinion leaders and clinical sites to shape an executable protocol, and developing positive, trusting relationships with patients and clinicians is a vital part of executing the trial. ⁶ Where experience is limited, every participating person or organisation is a trailblazer, and every process or system is used in an unfamiliar way. ⁶ Both the pace and the way of working need to be adjusted for rare-disease trials. ⁸

The trial in the development plan

Designing the trial starts from the bigger picture: how it fits the overall development plan, and the ultimate goal, which may be a marketing authorisation application. ⁴ The next step is to define what the trial is meant to show and why that matters, which may be superiority over standard of care, non-inferiority to it, or simply a better understanding of the disease's natural history. ⁴ Planning the design and the objectives needs an understanding of the natural history, the pathology and the competitive landscape. ² For many rare diseases little research exists, the disease is frequently not well characterised and competitors are often non-existent. ² So close collaboration with patients, patient advocacy groups, specialist healthcare professionals and subject matter experts is important to keep the objective clinically meaningful to patients in a real-life context. ² Once the objectives are defined scientifically, the company should ask the EMA to confirm that they are adequate and acceptable from a regulatory standpoint. ¹

Engagement with regulators and stakeholders

Rare-disease programmes often need numerous, iterative exchanges with the regulatory authorities. ^{9, 10} Without precedents, applicant and agency both work in uncharted territory, which can raise the likelihood of disagreement and delay. ¹⁰ Investing early in relationships with the authorities, and building their input on the protocol design and the endpoint agreements together with contributions from patient advocacy groups, key opinion leaders and clinical sites, leads to a better study design. ¹⁰ Orphan drug development is expanding greatly but is challenging, with no conventional roadmap. ¹¹ Through proactive early engagement with key opinion leaders, regulators and patient groups, a cohesive, strategic clinical development plan can be created that gives the strongest foundations for the eventual marketing application. ^{11, 2}

The study population

The population of the pivotal trials should represent the indication planned for the marketing authorisation and the product label, so getting it right from the outset is vital to the trial's potential success. ^{5, 12} The wrong population can also hurt recruitment, which in turn can affect the study's duration. ¹² Failure to identify the most appropriate target population was a key feature of failed orphan marketing authorisations. ¹³ Because the patient pool in a rare disease is limited, careful definition of the population is key. ¹³

The inclusion and exclusion criteria have to be wide enough to enrol the maximum number of patients without being so general that too much variability, or noise, dilutes the results. ¹² Setting them is hard because rare diseases are typically poorly characterised and often affect children disproportionately, and because different countries follow different national guidance on diagnosis and treatment. ¹³ A non-standard parameter in the criteria can make participation feel burdensome and so affect the recruitment of both investigators and patients. ^{13, 12}

Trial design

The EMA states that most orphan drugs and paediatric indications submitted for approval rest on randomised controlled trials that follow generally accepted rules and guidance, and that a comparative trial will usually be preferable but may not always be possible. ¹³ The randomised trial is the gold standard for an unbiased estimate of benefit and risk. ¹³ Its clear limitation in a rare disease is the number of patients: a pivotal study supporting an EMA marketing authorisation may enrol several hundred patients or fewer than 30, depending on the disease, whereas a typical trial in a disease that is not rare must enrol more for adequate statistical power. ¹³ Considerable clinical, mutational and phenotypic variability between patients with one condition can further complicate interpretation of the results. ¹³ Trials of orphan drugs sometimes enrol fewer than 15 participants. ¹⁴

A small sample can undermine statistical power and make significant outcomes hard to show, while sampling intensively enough for sufficient data points may discourage patients and investigators. ¹⁵ Sampling frequency should be balanced against the burden on patients and staff, and the statistical methods are critical for handling missing data, loss to follow-up and interpretation with a small sample. ¹⁵ Often only one pivotal phase 3 study, usually small, is possible. ¹⁶ Early studies frequently give too little data to select doses for later ones, and dose-optimisation studies can be unfeasible. ¹⁶ Dosing has to be tailored from limited existing data. ⁹

Adaptive designs, historical controls and alternative statistical approaches may be acceptable if they improve the interpretability of the results. ¹³ A crossover design is one recommended approach, but its two treatment periods lengthen the trial, which can raise significant ethical concerns in progressive irreversible diseases and can affect recruitment and retention. ¹³ Complex protocols are potentially useful for evaluating drugs in rare-disease indications, to make development more efficient. ³ Conducting a single study in several rare-disease populations, under a master protocol with disease-specific sub-protocols, can help accrue a more substantial body of safety data for the investigational treatment and overcome some of the challenges of developing drugs in very small patient populations. ^{17, 18}

The control

Standard control groups may be unavailable. ⁹ ICH E10 gives direction on the choice of control and outlines five options, the first four of them concurrent controls, in which the control and test groups are chosen from the same population and treated at the same time. ¹⁹

• Placebo is generally optimal because it shows benefit and risk most clearly. But patients with rare diseases are often gravely ill children who do not want to gamble on being randomised to an arm with no therapeutic benefit. A crossover, with placebo followed by the active drug or the reverse, or an open-label period after the placebo period, can prove highly effective, because every patient is guaranteed the active drug. ¹⁹
• No treatment, in which patients are randomised to the study drug or to nothing, can introduce bias because investigator and patient cannot be blinded and subjectivity becomes a concern. A blinded panel of assessors permits independent evaluation of outcome measures but does not remove the bias in patient-assessed outcomes. The arm can also yield useful data on the natural history and progression of the disease, which is often a relative unknown in rare diseases. ¹⁹
• A different dose or regimen of the study drug raises similar practical and ethical issues as a placebo or active-control group. ¹⁹
• A different active treatment can generally be disregarded, since only 6% of rare diseases have treatment options available. ²⁰
• An external or historical control is a virtual control formed from patients with the same disease in registry studies, medical records or the control arms of earlier trials. It lets a company compare its treatment's effects essentially against standard of care, natural disease progression or both, and it attracted a lot of attention in the rare-disease world. It needs great caution in design, analysis and interpretation, and to avoid bias the definition of which patients are included must be tightly controlled so that only those with very similar disease state, demographics and history enter it. Where the study uses a specific efficacy measure, the external patients may need data from that assessment too, which is hard when the measure is uncommon. The EMA showed willingness to accept historical controls, but the approach should be validated with it before the study, because its limitations can affect the future authorisation. ^{20, 21}

Randomised trials, in which a new treatment is compared against placebo or an existing standard of care, are widely considered the gold standard for the evidence regulators need to assess efficacy and safety, and they usually include large numbers of patients. ²² In certain areas, such as rare diseases including rare cancers, where target populations are often very small, a proportion of marketing authorisation applications reach the EMA with single-arm trials as pivotal evidence. ²² A single-arm trial has no randomised comparator, so every patient receives the experimental treatment and only the outcomes under it can be observed. ²² The EMA's reflection paper on single-arm trials submitted as pivotal evidence of efficacy in a marketing authorisation application discusses their key concepts, and it aims to stimulate scientific discussion of their challenges and to improve their design and conduct. ^{22, 23} The EMA described it as the first guidance document by an international medicine regulator on the considerations and challenges of this type of trial. ²² It was adopted by the EMA's human medicines committee with contributions from the Committee for Advanced Therapies, the Methodology Working Party and the Oncology Working Party. ²² It opened to public consultation in April 2023, with comments invited through an online form until midnight CET on 30 September 2023, and a final document was planned for 2024. ^{22, 23}

Endpoints

Safety, measured through the incidence and frequency of adverse events among other parameters, is a key endpoint for the benefit-risk assessment in any trial. ^{24, 20} The primary endpoint is the predefined main way of answering the trial's question, and in less well-characterised diseases secondary and exploratory endpoints can be crucial to showing overall benefit and should be carefully selected. ²⁰

Disease-specific clinical endpoints often do not exist, because the population is limited and natural-history data are lacking, and where they exist they are frequently unvalidated, little recognised or little used in the clinic. ²⁰ Asking patient groups which endpoints are meaningful to patients is important, and it supports the patient-benefit claim when marketing authorisation is sought. ²⁰ A rare-disease patient advocacy group, where one exists, may help determine relevant and nuanced endpoints, because a rigid interpretation of predetermined endpoints may produce an unfair study failure of an effective treatment. ¹⁰

The rarity of a disease may require novel or adapted endpoints, patient-reported outcomes among them. ¹⁰ Where no relevant endpoint exists, surrogate endpoints or biomarkers may be considered if the health authority agrees. ¹⁰ Besides direct clinical outcomes, patient-reported outcomes, surrogate endpoints and biomarkers that can be linked to clinical benefit should be considered. ²¹ Surrogate endpoints and biomarkers have both been used, at times with controversy, to gain conditional marketing authorisation for orphan drugs in the EU. ²¹

Agreement with the EMA on endpoint selection and validation should come as early as possible. ²¹ A company can request the agency's opinion on the acceptability of a novel biomarker or a surrogate endpoint, and early engagement ensures documented agreement that the selected endpoints will provide suitable efficacy data to support authorisation later. ¹⁹

The patient community and investigators

Talking with patients and with patient advocacy and alliance groups is crucial: it collects real-world information, identifies which improvements those living with the disease would count as significant, and begins to build trust. ²⁰ People with rare diseases may be wary of a healthcare system that is often ill-equipped to diagnose and treat them, and some have passed through numerous providers, procedures, misdiagnoses and treatments before a correct diagnosis. ²⁰ So companies are well advised to build a relationship with the patient community founded on trust, to maximise compliance and adherence in the trial and, later, to the approved treatment. ²⁰

To find the patients, the treating physicians have to be found first and convinced to become investigators. ²¹ Specialised investigators can be hard to find, and the Orphanet database, the European Organisation for Rare Disease and the National Organization for Rare Disorders help, as do the investigators of earlier trials, the disease's key opinion leaders, and internet and literature searches. ²¹ Compliance with global healthcare standards and good clinical practice varies around the world, trials must comply with good clinical practice, and outreach and audits to assess compliance are critical to patient safety and data veracity. ²¹ More sites bring more challenges, and local expert contract research organisations help with cultural, linguistic and procedural differences. ^{21, 24}

Selecting, funding and maintaining sites

Collaborating with clinical sites while the protocol is developed helps make it executable and reduces later amendments, but in a rare disease it is complex and far more costly than for standard trials. ¹⁰ Qualification criteria for sites are unclear or complex, patient pathways may be unknown and referral networks may not exist. ¹⁰ Prevalence and incidence data are very limited and often regional, which makes it hard to decide which countries to include. ¹⁰ Trial-naive sites often have to be included, and because eligible participants are so unlikely to be found, qualifying enough sites is crucial. ¹⁰ Funding considerations include supporting single sites across several countries rather than consolidating them in one or two, and maintaining international sites even when they recruit little. ¹⁰ The complex international approval system can also delay set-up. ¹⁰ A site recruiting no patients, or only a few a year, needs high maintenance and bespoke communication, because risk-based monitoring is insufficient at such low enrolment. ¹⁰ Trial management systems are not optimised for rare-disease studies, key performance indicators do not apply at such low numbers, and inexperienced staff must undergo extensive training and attend investigator meetings to enrol perhaps one participant, if any. ¹⁰ Given the small number of eligible participants, every site willing to recruit is indispensable, and investment in coordination, communication and the relationship with each site is crucial. ⁹ Sites may be added or closed during the trial. ²⁵

Supporting inexperienced sites

Sites chosen for rare and ultra-rare trials may be inexperienced with trials in general or with rare-disease trials, so the investment in educating their staff on trial processes and systems, and potential referral sites on the disease, should not be underestimated. ⁹ The site may have no study coordinator because it never needed one, and contracting may be outsourced in a large hospital. ⁹ The local investigator may not know how to conduct internal follow-ups and likely has a full-time role with no structured framework for trials, so any email, call or request from the sponsor could feel like an extra burden. ⁹ A single point of contact at the sponsor, working with a coordinator appointed at the site, can help the investigator navigate both new worlds, of trials and of the condition, by clarifying the acronyms for roles and the processes that may be unfamiliar to site staff. ^{9, 26} Staff may be trained months or even a year before their first patient, so continued engagement between start-up and first enrolment may avoid retraining. ²⁶ Once a patient is identified, support should remain available through that patient's whole journey in the trial. ²⁶

Inexperience can lead to deviations from study processes and to inappropriate reporting. ²⁶ Trust and open communication between site and sponsor may help keep sites from following standard care protocols where they do not apply, missing timepoints in irregular follow-up schedules, or reporting normal laboratory ranges where they are not applicable. ²⁶ Explaining things slowly, gradually and comprehensively, with no prior experience presumed of the clinical staff or their organisation, is essential. ²⁶ Connecting the participating sites with each other is critical for trial success, for enquiries and support and for sharing lessons and best practice. ²⁶ Sites with specialised knowledge may face challenges if overwhelmed by requests, and understanding the site's position is key to keeping interactions from becoming a burden or an irritation. ^{25, 9} Here too, a dedicated point of contact at the sponsor working with a coordinator appointed at the site can help prioritise tasks, reduce the burden on the local investigator and prevent frustration, and that contact may share the expert site's learnings with the other participating sites. ⁹

Participants and samples

Correct diagnosis is a challenge to screening and enrolling eligible patients, because a condition may be so hard to identify that patients often consult several specialists over many years before it is diagnosed. ²⁶ Where the condition is in a foetus, recruiting the pregnant participant is feasible only in a narrow window early in pregnancy. ²⁶

Blood sampling in very young children should be minimised while as much data as possible is collected from the few participants. ²⁶ Numerous exploratory endpoints may add burden through more blood sampling, more or longer visits or additional questionnaires, and the number or type of treatments or injections, or changes to medication patients already take, can also feel burdensome. ²⁶ Participants may drop out if the burden feels too high, shrinking an already minimal analysis set, and in a rare or ultra-rare condition recruiting replacements will likely prove impossible. ²⁶ So the protocol should be kept to what matters, and the participant's experience made positive for retention, prompt stipend payment included. ²⁶ Effective communication between the participant's primary care team and the study team is essential, so that the care team receives the study information and understands what the participant has already been through. ^{26, 25}

Every sample is valuable, and at an inexperienced site its integrity may be at considerable risk. ²⁶ Specific analyses may need shipping to a laboratory that can run the assay, and inexperienced handling can cause shipping delays and lost or unanalysable samples. ²⁶ Training local staff to analyse samples may not always be possible without the infrastructure and expertise it needs, and analysis at several local laboratories may introduce inter-assay variation that will have a considerable impact on the results, given the small sample. ²⁶ The choice is between an established central laboratory, with shipping risks and possibly high maintenance costs, and local analyses, which may cost more and bring that variation. ²⁶

Protocol amendments

The pioneering nature of these trials often makes protocol amendments frequent, before and during the trial. ²⁷ Inclusion and exclusion criteria may need to change as diagnostic methods, laboratory assessments or management guidelines improve, to fit the actual patient population. ²⁷ If recruitment targets are not met, the design or the number of participants may change, with a redesign around the available pool, or real-world data complementing the prospectively collected data, to reach the required sample size and an acceptable statistical interpretation. ²⁷ Preparing for potential amendments helps researchers navigate these trials, and strategic planning and flexibility are key. ⁸

Manufacture and supply

Drug manufacture, supply and management raise their own logistical problems, and an experienced logistics contract research organisation or partner is key. ²⁸ Orphan drugs are often extremely expensive to make because they need specialist facilities and equipment, so they are initially produced in very small quantities, sometimes at the individual patient level. ²⁸ Once safety and efficacy are first shown and phase 3 is planned, scale-up begins, possibly with a commercial phase 3 formulation that can be produced faster and more efficiently. ²⁸ That formulation in turn needs relative bioavailability studies showing that it is comparable to the earlier one. ²⁸ As the programme goes on, formulations for specific populations, such as children or patients with difficulty swallowing, may have to be developed and tested. ²⁸