Rare disease CMC: harder than it looks

Rare disease medicines rarely struggle with CMC because sponsors lack technical expertise. The difficulty is that conventional CMC frameworks assume things that do not hold for orphan programmes: plentiful batches, large datasets, predictable manufacturing demand and a mature process before pivotal studies begin.

An orphan programme may make fewer batches across its entire development lifecycle than a high-volume product makes in a few weeks. The CMC dossier therefore leans more heavily on scientific knowledge, prior experience, process understanding and regulatory judgement than on large statistical datasets.

Regulators can offer flexibility for products developed for small patient populations. That flexibility still has to be earned with a scientifically justified control strategy, sound risk management and transparency about what the available data can, and cannot, demonstrate.

Six areas are where this hidden complexity shows most.

1. Limited batch numbers

In a rare disease programme, every batch matters disproportionately. The same small manufacturing history may have to support clinical supply, stability studies, process understanding, validation and, eventually, comparability. That creates difficult trade-offs.

• Specifications rest on limited process data. Acceptance criteria need to draw on clinical experience, process capability, prior knowledge, platform information and the relationship between quality attributes and product performance, not just historical batch statistics.
• Process validation may need a lifecycle approach. The traditional expectation of three consecutive validation batches is not a universal regulatory rule. Depending on the product, process maturity and regulatory strategy, concurrent validation, continued process verification, justified bracketing and other risk-based strategies may be appropriate.
• Variability is harder to interpret. With only a few observations, a single out-of-trend result is difficult to tell apart from normal process variation.

A practical starting point is a batch-utilisation map prepared before manufacture begins. It identifies which batches are intended for clinical supply, stability, validation, comparability, reference purposes and retain samples.

This simple exercise prevents a costly late-stage problem: finding that the only batch representative of the intended commercial process has already been used.

2. Small patient populations

Small patient populations create a manufacturing challenge that goes beyond low volume. Annual demand may be met by one or two campaigns, so production becomes episodic, with long gaps between runs. Each campaign can need renewed attention to operators, equipment readiness, raw materials and supply-chain continuity.

A small clinical population also limits the evidence available to link product quality to clinical outcomes. A pivotal study may rely on material from only a few lots, so understanding critical quality attributes (CQAs) has to draw on several sources of evidence:

• mechanism of action;
• non-clinical studies;
• process and product characterisation;
• platform knowledge;
• analytical data; and
• clinical experience.

Paediatric rare diseases add another layer. Formulation, excipient acceptability, dose flexibility, administration volume and device considerations can all become significant CMC questions.

Supply-chain resilience matters just as much. A single CDMO, fill-finish site or critical raw-material supplier can be entirely reasonable in early development, yet become a serious continuity risk once patients depend on the product.

For rare disease programmes, manufacturing continuity belongs in the CMC strategy, not on a list of operational issues to address after approval.

3. Process changes during development

Accelerated development pathways can compress clinical timelines, but they do not automatically accelerate process maturity. The process may still be evolving when pivotal development begins. Typical changes include:

• scale-up from an academic or laboratory process to commercial GMP manufacturing;
• transfer to a commercial CDMO;
• replacement of research-grade or animal-derived raw materials;
• analytical method upgrades or changes;
• formulation optimisation; and
• changes to container closure or presentation, such as moving from a vial to a pre-filled syringe.

None of these is inherently a problem. The challenge is showing that the material used to generate pivotal clinical evidence remains representative of the product intended for commercial supply. That makes process-change planning critical.

Where possible, sponsors should settle the intended commercial process early, recognising that the right timing depends on development stage and product-specific risk. Changes should be scientifically justified and, where practical, grouped into coherent packages rather than introduced piecemeal.

Sponsors should also keep comprehensive characterisation data and retain representative samples from each meaningful process version. Those samples can prove invaluable when a later change needs a direct comparison with material used in earlier clinical studies.

4. Comparability challenges

Comparability is hardest when the manufacturing history is small. ICH Q5E provides the framework for evaluating the impact of manufacturing changes on biotechnology-derived products. The aim is not to show that analytical results are numerically similar, but to build a scientifically justified case that the change has not adversely affected quality, safety or efficacy.

In a rare disease programme the datasets may be too small to establish robust historical distributions. Comparing three pre-change lots with two post-change lots, for example, says little about the full range of process variability. That does not make comparability impossible. It makes product and process knowledge more important.

Sponsors can strengthen the assessment through:

• Deeper characterisation. Orthogonal analytical methods and well-designed stress or forced-degradation studies can reveal differences that routine release testing misses.
• Pre-defined comparability criteria. Acceptance criteria and decision rules should be set prospectively where possible, not developed after reviewing the post-change data.
• A risk-based, tiered assessment. Direct the greatest analytical sensitivity at attributes with demonstrated or plausible links to biological activity, safety or efficacy.
• Retained samples. Representative material from earlier process versions may be the only practical basis for a meaningful side-by-side comparison later.

Advanced therapies such as cell and gene therapies add further difficulty through variable starting materials, complex processes and limits in potency assays. Depending on the change and the product, regulators may also expect additional non-clinical or clinical evidence to support the bridging strategy.

For significant changes, early regulatory dialogue is usually far more efficient than building a comparability package and discovering later that the agency expected a different evidentiary approach.

5. Stability data limitations

Stability is another area where rare disease programmes meet a basic mismatch between development timelines and the real-time data available. Expectations depend on the product type, development stage and regulatory framework. ICH Q5C sets out relevant principles for biotechnology-derived products, and ICH Q1A(R2) provides the general framework for drug substances and products within its scope.

In practice, a rare disease filing may hold little long-term data from batches made with the intended commercial process, particularly when the process has recently changed or campaigns are infrequent. Stability pulls also compete directly with clinical supply. When few batches exist, allocating enough material to stability while keeping adequate clinical inventory is a genuine planning problem.

Sponsors can address gaps with scientifically justified approaches such as:

• supportive stability data from earlier or related batches;
• bracketing and matrixing where scientifically appropriate;
• trend analysis and statistical modelling;
• knowledge from representative process versions; and
• ongoing real-time stability commitments.

These approaches do not remove uncertainty. They give a structured way to manage it while real-time data accumulate.

The post-approval stability commitment should also be realistic and match the actual manufacturing cadence. Committing to annual stability batches for a product made only every 18 months, for example, creates an obligation that is hard to meet consistently.

Shelf life has practical consequences too. A short shelf life increases write-offs, limits the safety stock that can be held and complicates distribution across dispersed markets.

6. Global lifecycle management

For rare disease medicines, approval is often not where CMC complexity ends. It is where lifecycle complexity speeds up. After approval, sponsors may need to introduce:

• additional manufacturing sites;
• second-source suppliers;
• scale changes;
• analytical method improvements;
• new manufacturing technologies;
• formulation or presentation changes; and
• supply-chain resilience measures.

The regulatory classification of these changes differs across jurisdictions. A change that needs only a notification or lower reporting category in one region may need prior approval in another. That can mean parallel regulatory processes, duplicate inventory and added supply risk.

Several lifecycle tools help:

• ICH Q12. Established conditions and Product Lifecycle Management tools give a structured framework for managing certain post-approval changes.
• Post-Approval Change Management Protocols (PACMPs). These let an approach to specific changes be agreed in advance, where the relevant authority accepts it.
• Reliance and work-sharing. Where available, these reduce unnecessary divergence between regulatory assessments.
• A core Module 3 with controlled regional adaptations. This helps keep global dossiers consistent.
• Global commitment tracking. It keeps post-approval obligations visible and coordinated across markets.

Lifecycle management starts much earlier than the first post-approval variation. How the initial control strategy, process description and established conditions are designed largely determines how hard future manufacturing changes will be.

Turning limited data into a defensible dossier

Rare disease CMC is an exercise in making sound scientific decisions with less data than conventional development generates. The answer is not to pretend a small dataset gives the same statistical confidence as a large one. It is to get the most from every data point, connect analytical evidence to product and process understanding, and be clear about what is demonstrated and what remains uncertain.

Five principles help:

1. Plan batch utilisation before manufacture. Know how each batch will support clinical supply, stability, validation, comparability and retain requirements.
2. Characterise deeply and early. Strong product and process knowledge is most valuable when statistical power is limited.
3. Retain representative material from meaningful process versions. Today's clinical batch may become tomorrow's comparability reference.
4. Engage regulators on the questions that matter most. Scientific advice can be especially valuable for comparability, stability, validation and major process changes.
5. Write the CMC strategy for the whole product lifecycle. Decisions made in development can materially affect post-approval flexibility and supply continuity.

Rare disease CMC is not conventional CMC performed at a smaller scale. It is a development environment where limited batches, limited patients, evolving processes and global supply requirements all interact at once.

Sponsors who recognise this from the first batch can build a more coherent evidence strategy, avoid preventable CMC surprises and lay a stronger foundation for reliable supply after approval.