When would you prioritize process robustness over maximum titer during technology transfer?

Why

A development team often delivers a process tuned to its best result. At transfer, that result meets a new plant with different bioreactor geometry, gassing, mixing times, raw material lots and operators. ICH Q10 describes technology transfer as moving product and process knowledge to achieve product realisation, forming the basis for the control strategy and process validation.

ICH Q8(R2) notes that a design space developed at small scale should be justified at production scale, with scale-up risks discussed. If titer peaks only in a narrow band of temperature shift timing or feed rate, small differences at the receiving site push batches off the peak and possibly change glycosylation or aggregates too. A robust set point, chosen in a flatter region of the response surface, gives up a little titer for consistent quality and fewer failed batches.

Robustness wins when supply depends on a new site, when PPQ is close, when the receiving site has little experience with the molecule, and when quality attributes are steep functions of the parameter. Titer may be worth pushing later, through change control, once the site has data. It is worth arguing in numbers.

A process that loses 10 percent titer but avoids a one-in-ten batch failure rate usually delivers more product and far less investigation load.