1st QbD Approval for Biologics: Gazyva Design Space

Posted On

“This will be the first OBP BLA* with an approved design space.” (CDER, *OBP: Office of Biotechology Products, BLA: Biologics License Applications).

Quality-by-Design in Biologics is challenging. Skepticism prevails. (i.e. Matej Horvat’s excellent discussion).

I outlined a few reasons why QbD in Biologics is challenging in an earlier article.

Complexity and heterogeneity of biologics. Science of biology is still evolving. Hence, difficulty in understanding CQA’s, CPP’s and their relationships.
Unclear guidance from Regulatory Agencies (FDA, EMA, etc.) on process consistency and clinical performance
Difficulty in predicting Scale-up models and technology transfers to commercial lines
Misaligned (or siloed) values within organizations: R&D must invest initially but Manufacturing will mainly reap benefits.
Lack of experience in characterizing Design Space and Multivariate Analysis

However, there’s always a first. On Nov. 1, 2013, Genentech (Roche) succeeded in receiving a full-QbD approval–including the coveted Design Space. What is more interesting is that their previous product, Perjeta used a similar approach but did not get an approval on design space. At ISPE 2013, Dr. Lynne Krummen of Genentech shared the lessons learned from Perjeta and Gazyva submissions.

One difference — FDA granted Gazyva as Breakthrough Therapy Designation. Whether this affected the Gazyva design space approval outcome or not, we do not know. In any case, the biotech industry celebrates the first success of full-QbD approval.

This article will focus on FDA’s CHEMISTRY REVIEW(S) of Gazyva. Unfortunately, this document does not reveal how much of a “design space” is approved. Although proprietary information is hidden, I’d like to highlight a few take-away’s relevant to our interest–QbD elements.

So here is the full reference: FDA’s CHEMISTRY REVIEW(S).

I bolded and/or bulleted my lessons learned. Quotes are excerpts from the document for reference.

1. Background:

Gazyva (Obinutuzumab) is a fully humanized monoclonal antibody, first originated by Glycart Biotechnology AG and developed byBiogen Idec, Chugai Pharmaceutical Co., and Hoffmann-La Roche Inc. for the treatment of chronic lymphocytic leukemia in combination with chemotherapy. (Ref) Approval Letter:

2. CQA (Control Quality Attributes)

“Design space” was acceptable.
Change control plans, quality systems is required.

” Upon review of the supporting data, the design space as proposed in STN 125486 was found to be acceptable. The Agency would like to reiterate that in addition to the information described in the application, it is our expectation that plans for implementation of the design space for the commercial process are documented within the firm’s Quality System. Such quality systems may include plans for handling movements within the design space (e.g., change control procedures, plans for updating batch records). Additionally, in accordance with ICH Q8(R2) the Agency does not expect any regulatory notification for movements within the design space; however movements outside of the design space should be reported to the Agency in concurrence with existing regulations” Critical Quality Attributes:

CQA’s came from previous knowledge, scientific literature, experience with other related products and obinutuzumab-specific data, including clinical experience.
Risk Score = Impact x Uncertainty
Impact: magnitude of the attribute’s effect on bioactivity, PK, safety, and immunogenicity
Uncertainty: the degree of confidence that the impact has been correctly assessed
If Risk Score is greater than 13, then the attribute is “Critical.”

“For product variants and process-related impurities, the sponsor used a risk based, risk ranking and filtering (RRF) tool to evaluate the impact and the uncertainty associated with each quality attribute’s potential to affect product safety and/or efficacy. The evaluation was based on previous knowledge, scientific literature, experience with other related products and obinutuzumab-specific data, including clinical experience. The impact is the magnitude of the attribute’s effect on bioactivity, PK, safety, and immunogenicity, while uncertainty is the degree of confidence that the impact has been correctly assessed. The impact and uncertainty scores were multiplied to calculate [sic] a final risk score for each attribute. Attributes with a risk score above a specified value (i.e. 13) were classified as critical quality attributes.”

Obligatory CQA’s: pH, excipient concentrations, protein concentration, osmolality, bioburden, endotoxin, leptospira, mycoplasma, potential viruses, appearance, particulates, sterility, and extractable volume, etc.

Underestimated the impact of CQA on QTPP (safety, immunogenicity, and PK)
Justification was insufficient (only based on small clinical studies)
Need a clear and logical relationship between QTPP and CQA

“While it was determined that the overall system used for identifying CQA’s was acceptable, there were concerns that there were concerns that the RRF tool used for the assessment of product variants and process-related impurities may have underestimated the impact of attributes on safety, immunogenicity, and PK. When evaluating the immunogenicity and safety impact of a number of quality attributes, the sponsor appeared to rely heavily on data from clinical studies in which the attributes were present in very small amounts. These clinical studies, therefore, were not designed to be able to detect safety or immunogenicity signals for product variants. It was not clear if the sponsor considered all of the available information with regard to the risk of a variant to impact safety and immunogenicity. In addition, there were concerns about the suitability of the in vitro FcRn assays used to assess the impact of product variants on PK.”

Any future changes to control strategy must be reported with supporting data.

“While there may have been an underestimation of the risk related to safety, immunogenicity and PK for some quality attributes, these attributes were classified as CQ A’s for other reasons. Therefore, it was not thought that any of the identified non-CQA’s should have been CQA’s. In addition, all CQA’s are appropriately included in the control strategy and the sponsor cannot change the control strategy without Agency approval. Future changes to the control strategy should include consideration for whether the impact of variants and safety, immunogenicity, and PK were underestimated.”

3. Control Strategy

Unfortunately full control strategy is not revealed. In this document, control strategy is referred to as “PALM” plan. This may have been the key in getting approval for design space. Basically, a control strategy is a comprehensive plan outlining how a company will keep their promises on their design space and how they will deal with future changes.

“5. Development of a post–approval lifecycle management (PALM) plan, which presents the sponsor’s intended overall control strategy, including how process and product monitoring will be performed post approval and how changes to the design space will be managed and reported to the Agency. The PALM plan specifies how the sponsor will monitor the process and product quality post approval, manage changes within the design space, and update the control system based on additional process and product knowledge.”

“The PALM plan includes such things as

The final testing strategies for DS and DP
A list of the relevant sections of the BLA (e.g., sections that define the design space and in-process control testing)
A description of the key process indicators (KPI’s)
A description of how CPP’s, non-CPP’s, KPI’s, in-process controls, and CQAs are routinely monitored and the potential outcomes of that monitoring
A table defined the preliminary trend limits for attributes
A description of how changes to process parameters will be managed
A description of how the design space will be verified when changes are made within the design space
A description of how changes outside of the design space will be managed

Of particular note, changes to the design space for control strategy will require Agency approval. Specifically: (hidden)”

In contrast to the Perjeta case, where PALM plan was rejected, Gazyva’s control strategy was accepted. We certainly have a win but do not have the details of how. Hopefully EMA’s assessment report (upcoming) will reveal more information.

4. Conclusion: