CMC and regulatory3 min readBy QLogixCMC

Comparability for cell, gene and mRNA therapies: proving a process change left the product unchanged

Comparability is the evidence that a product made after a manufacturing change is still the product the clinical data describes. For cell, gene and mRNA therapies the change is almost certain: processes move from academic to commercial scale, from open to closed, from one site or vector supplier to another, often during the pivotal trial. The statistics are unforgiving. With three lots on each side, a real shift and ordinary lot-to-lot variability look alike, and the study fails for lack of data rather than lack of comparability.

Why advanced therapies change so often

  • Scale-up and automation: a manual, open process that served a phase 1 study is closed and scaled for commercial supply
  • Vector and plasmid supply: a new supplier, a new producer cell line or a new purification step changes the starting material
  • Potency assays that mature during development, so the method used for the pivotal lots is not the method used for the early ones
  • Site transfers to a CDMO or to a sponsor's own plant, with new equipment and new analysts
  • For mRNA, changes in the plasmid template, the in vitro transcription process or the lipid nanoparticle formulation
  • For autologous cell therapies, the starting material varies with the patient, so the process has to be shown robust to it

What the regulators expect

ICH Q5E sets the principles for biologics: a side-by-side comparison of pre- and post-change product on quality attributes, with extended characterization where routine release testing is not enough, and stability data where the change could affect it. FDA's draft guidance of 2023 on manufacturing changes and comparability for human cellular and gene therapy products applies the same principles to these products, and its separate draft guidance on potency assurance asks for a strategy that controls potency through the process, not a single release assay.

A comparability protocol, agreed before the change, states which attributes are compared, on how many lots, with what acceptance criteria and which statistics. Agreeing it afterwards invites the question of why the criteria fit the data.

Interactive

Does the post-change product sit inside the equivalence margin?

Comparable90% interval -1.5 to +3.9 % against ±5 %

The difference in means is +1.2 % and the whole interval sits inside the margin.

Relative potency, % of reference859095100105110115pre-changePre-change lot 1: 98.1 %Pre-change lot 2: 103.2 %Pre-change lot 3: 100.9 %Pre-change lot 4: 96.4 %Pre-change lot 5: 104.1 %Pre-change lot 6: 99.3 %Pre-change lot 7: 102.2 %Pre-change lot 8: 100.6 %post-changePost-change lot 1: 103.2 %Post-change lot 2: 99.9 %Post-change lot 3: 102.4 %Difference in means with its 90% interval, against the margin-10-50+5+10Difference 1.2 %, interval -1.5 to 3.9 %
  • Pre-change lots
  • Post-change lots and the interval
  • Equivalence margin
Illustrative lots. With three post-change lots the interval is wide enough to fail the margin even when the shift is small: the study lacks data, not comparability. More lots narrow the interval; a real shift moves it. The margin itself comes from the clinical experience and the attribute's criticality, agreed in the comparability protocol before the change.

Designing the study so it can pass

  • Choose the attributes and the margins from criticality and clinical experience, and write them into the protocol before the first post-change lot
  • Set the number of lots from the known lot-to-lot variability, not from the calendar
  • Keep retains of pre-change lots so the comparison can be run side by side on the same day with the same method
  • Characterize with orthogonal methods where the release assay alone cannot distinguish the products
  • Bridge any assay change separately, so a method difference is not mistaken for a product difference
  • Include stability, accelerated where justified, when the change could affect it

How QLogixCMC helps

QLogixCMC writes the comparability protocol and the statistical plan, assembles the pre-change data, runs the analysis and drafts the comparability report and the regulatory submission that carries it, for cell, gene and mRNA products as well as conventional biologics. The sponsor's qualified people set the margins and sign the conclusion.

Questions this post answers

How many lots does a comparability study need?

Enough that the confidence interval of the difference between pre- and post-change product fits inside the equivalence margin when the products are in fact comparable. That depends on the attribute's lot-to-lot variability and the margin. Three lots is often too few for a variable attribute; the number is calculated in the protocol, not assumed.

What if the potency assay changed as well?

Bridge the assays first: test the same lots with both methods and establish the relationship, so that a difference in the comparability data can be attributed to the product or to the method. Then run the comparability on one method.

More from the blog

Talk it through with someone who has run a plant

Bring your backlog, your submission timeline or the Form 483. We will tell you plainly what we would do first, what it would take, and what we would leave alone.

Request a meetingEmail us

In the first meeting we cover

  • What is on your mind: a backlog, a submission, a validation program, an inspection date or a system that is not working
  • What we would assess first, and what we would leave alone
  • Whether an engineer on your site, a fixed-scope assessment or a defined deliverable is the right way to start