Home / Small molecules / Drug product / Late Phase Development
Late Phase Drug Product Development
The formulation that goes into Phase III is the one that will be filed. Late-phase development makes it reproducible at–scale before it is locked.
What changes at Phase III
An early-phase formulation only has to work; a late-phase formulation has to work every time, at commercial batch size, on equipment that may differ from what it was developed on, for as long as the product is on the market.
Late changes are possible but carry regulatory consequences: a formulation change after pivotal dosing usually means a bridging study, and occasionally a repeated one.
Formulation optimization
The Phase I formulation is usually the simplest one that can deliver the dose, chosen on limited API and limited stability data. Between Phase I and Phase III it is developed into the formulation that will be filed: the dosage form and strengths the clinical dose now supports, excipient grades and levels set against compatibility and process data, a manufacturing process chosen on the API’s flow, compressibility and moisture sensitivity, and a dissolution method that discriminates between variants.
Most of that work is driven by manufacturability and stability. Changes to tablet size, dosing frequency or a modified-release profile are settled as early as possible, because each can change exposure and needs a clinical bridge. The formulation is locked before pivotal dosing wherever the program allows.
Process development and scale-up
Moving from a development batch to a registration batch is where formulation problems most often appear — blending uniformity that held at 2 kg and does not at 200 kg, a granulation endpoint that shifts on a different mixer, a tablet that will not compress at line speed.
Clinical supply manufacturing
We deliver more than 3,200 oral batches a year. Across all drug product, 3,800+ batches go out in 2,000+ shipments to the United States, Europe, China and other markets.
Packaging and labeling run in the same operation, with no hand-off: bottle, blister and sachet, open-label and double-blind. On a blinded trial, a labeling error is a protocol deviation.
Registration and validation batches
Registration batches are made under the process that will be validated, on the equipment train intended for commercial manufacture, using the analytical methods that go into the filing. They generate the stability data the application rests on, so a change made after them carries a comparability obligation.
Managing change between IND and validation
Between the IND formulation and the validated one, most programs make changes. Each change carries a stability obligation and a comparability question, so the aim is to make as few as possible while still arriving at a manufacturable process.
Deciding which change justifies a bridging study, and which does not, is where prior late-phase experience is applied.