The most common question asked by bioassayists at this juncture is “How many runs do we need to do?”  Unfortunately, there is no set and established answer to this question. The answer depends greatly on the assay precision and bias as suggested above as well as the clinical requirement for equivalence. The best advice is to work with a statistician to determine the necessary power of the bridging experiments. As a first step ask yourself and your clinical colleagues how “wrong” can I be in the potency value assignment and not impact the clinical trial? If I assign a value which might be 10% high or low – does this result in releasing  material which is no longer to be efficacious? Or is it so potent it could cause patient safety issues? If you answer yes to this question, then you know that ± 10% is too wide of a confidence interval. Continue asking yourself this question until you find an acceptable range. Give this information to your statistician along with the precision and bias characteristics of the method and they can calculate the required number of runs.

If a manufacturing change results in a change of the dose-response curve of the proposed RS it will be impossible to value assign the new reference from the old reference. (See Figure 2) In this case it is critical to demonstrate that the old and proposed RS dose-response curves are in fact different. This is typically done by running the two lots in a head-to-head study. A head-to-head study is strongly recommended, as during this time frame the potency assay is often evolving and the dose-response curves might change between assay versions. Next carefully determine that the proposed RS is representative of material to be manufactured from this point forward. Once these two items are demonstrated satisfactorily, a full characterization of the new RS should be carried out and a comparison to the prior reference material lot made. Once this is accomplished, the new batch is usually given a value of 100%. All new RS lots will now be bridged to this batch rather than the earlier RS with different dose-response curve characteristics.

It is critical that all IRS used to release CTM are bridged. This is data almost always required by regulators. It is best to involve product experts, bioassay experts and statisticians in the design of these bridging exercises.

Several key RS activities happen in late clinical trials and just prior to submitting an application/dossier for approval.  Firms need to draft final protocols for supporting the maintenance of their reference material (RFM) post commercialization. Although there is some regulatory guidance for maintaining reference material, many aspects of a good reference program are driven by good science and a thorough understanding of the patient and manufacturing risk if one does not have a rigorous RFM program in place. Typical requirements for the RFM program include the following elements:

1. Establishment of a two-tier reference program.
   1. A two-tiered system consists of a primary reference standard (PRS), sometimes called a gold standard and a working reference standard (WRS) sometimes called a secondary standard.
   2. The PRS is used to qualify all future WRS. It is not used for routine product release. It is best if a product batch which is known to be efficacious is laid down as the PRS.
   3. The WRS typically comes from the standard manufacturing stream and is qualified against the PRS. It is used on a daily basis to release product. Sometimes the very first WRS is from the same product batch as the PRS.
   4. During the development process described above, the various IRS batches are value assigned using a daisy chain, that is they are value assigning using the current IRS being used to release clinical material. This approach changes at the end of the pivotal trial.  Now all future Working Reference Standard (WRS) are qualified, and value assigned using this PRS.  See Figure 3.
2. WRS qualification protocol.
   1. This includes a description of how to select, completely characterize, and value assign the potency of the candidate WRS from the PRS.
3. Reference Stability Program.
   1. The PRS and WRS must both have ongoing real-time stability programs. The testing timepoints and details of what is to be reported is outlined in detail in the testing protocols.
4. Reference Monitoring Program
   1. Both the PRS and WRS are usually monitored. Formal trending data is typically collected and assessed semi-annually or annually.  QC samples, dose-response curves and the potency of recently manufactured lots are often part of this annual trending.