In June, CBER Division of Manufacturing and Product Quality Facilities Reviewer and Investigator Ekaterina Allen spoke at the ISPE Biopharmaceutical Manufacturing conference about the current regulatory landscape for cell and gene therapy products. Part I covered her remarks about the current manufacturing challenges. This is the second part of the series and it looks at the most common agency observations made during inspections of cell and gene therapy manufacturing facilities by the quality system.

483 observations specific topic

When an FDA inspection team performs inspection, it is done using a systems approach that has been developed by the agency. There are seven key systems that are looked at during CGT inspections:

  • Quality
  • Production
  • Facilities and equipment
  • Materials
  • Laboratory control
  • Packaging and labeling
  • Donor eligibility (this is mostly for allogeneic products)

In every system, there are three elements that are evaluated: standard operating procedures (SOPs), training, and documentation. Depending on what type of inspection it is the scope is a little different. If it is the first inspection—a prelicensing inspection—the inspection team is trying to evaluate all the key systems. If it is a Team Bio inspection, they will look at the quality system and then pick and choose depending on how in-depth an inspection was planned and what they see during the inspection.

Here is the percent of inspections that have observations related to the various systems for licensing inspections and GMP inspections for cell and gene therapy products (Figure 1).

Figure 1 Most Common 483 Observations
FIGURE 1 | Most Common 483 Observations

Regardless of the product and regardless of inspection type, the top three systems that most frequently had issues are the same—the quality system, facilities and equipment, and the production system.

Some common observations across all the systems point to issues with SOPs and documentation. For SOPs, this includes a lack of written procedures, procedures that are incomplete or deficient, or not detailed enough to allow for consistent execution, or SOPs that are not followed. Regarding documentation issues, we see documentation practices that are not followed and GMP data that is either not secured, maintained, backed up, or effectively managed.

The remainder of Allen’s presentation focused on the most common inspection observations by system. She pointed out that what is presented on a 483 are only observations and do not represent a final agency determination regarding the compliance.

Quality System Observations

The first system Allen examined was the quality system. About 80 to 90% of inspections had observations for the quality system as listed here (Figure 2).

Figure 2 Quality Systems Observations
FIGURE 2 | Quality Systems Observations

Observations were made for deviation management, corrective actions and preventive actions (CAPA), training, document control, operations oversight, change control, and product adverse event and complaint procedures, which is mostly for post licensure. The most common issue is with deviation management, noted in half the inspections.

The most common deviation management issues are due to deviations not initiated or root cause analysis not completed. Problematic CAPAs are usually either not effective or the timeline is off or just not there for implementation, evaluation of effectiveness, and closing of CAPAs.

Regarding documentation issues, we see documentation practices that are not followed

Another common issue is if a specific batch fails, or its components fail, that the investigation is just limited to that one batch, and not expanded to other batches of the same drug product that may be impacted.

There are also issues that are cell and gene therapy specific—for example, inadequate aseptic technique training is “quite common, found in probably 30% of the inspections,” Allen said.

The other one listed here is rare, but it is CGT specific: the timeline for closures of manufacturing non-conformances is not justified—for example, the firm’s procedures allow up to 45 days after the product is used to close sterility failure investigations. “This is an awfully long time because if your product fails sterility testing, the patient has already received the product,” Allen commented.

Facilities and Equipment System Observations

The next most cited system is the facilities and equipment system (Figure 3).

Figure 3 Facilities and Equipment System Observations
FIGURE 3 | Facilities and Equipment System Observations

Environmental monitoring (EM) commonly has issues. There are basically two types of observations. Either something is not monitored, like ISO5 biosafety cabinets (BSCs) and ISO-7 areas under dynamic conditions, surface monitoring of ISO7 pass-throughs, or biosafety cabinets post operations.

Another common problem area is sampling sites not being representative of the locations of critical operations. For example, if a settling plate is in the corner of the biosafety cabinet whereas the actual manufacturing takes place in the middle.

If your product fails sterility testing, the patient has already received the product

Some less frequent observations concern equipment cleaning and equipment qualification. The equipment issues mostly concern either nonspecific SOPs or the validation being incomplete. Equipment qualification issues primarily were found around BSC qualifications not being performed, or smoke studies that are not representative of what is actually happening during manufacturing inside of the BSC. Sometimes qualification of ISO5 BSC has not been performed or is deficient.

Production System Observations

The next system Allen discussed was the production system (Figure 4).

Figure 4 Production Systems Observations
FIGURE 4 | Production Systems Observations

The three most common issues in the production system are incomplete batch records, inadequate aseptic process validation (APV), and issues with general process validation. The batch record needs to document completion of each significant step as well as document that all critical process parameters were met, and often that is not the case. As far as APV is concerned, it must simulate all the critical aseptic processes that are executed during production.

Sometimes investigators see issues with growth promotion testing in that it is either not performed, or it was performed and failed yet the APV report was approved anyway.

The process validation issues commonly observed during prelicensing inspection have to do with process performance qualification (PPQ) lots.  Sometimes the process that was validated during PPQ is not the commercial process because either something else was submitted as the PPQ or there were significant changes made after the PPQ lots were manufactured. And sometimes we see in-process time limits or hold times that were not established or adequately validated.

Materials System Observations

Now, let’s look at material system observations (Figure 5).

Figure 5 Material System Observations
FIGURE 5 | Material System Observations

Most of the material system observations are related to the incoming specifications and sampling. There are also less frequent issues with inventory control and storage as well as with extractables and leachables.

Most commonly during inspections of this system investigators see that raw materials sometimes will not have specifications or the specification was not met, but the material was not rejected and was still used by the manufacturer.

Another common finding takes place “because the manufacturer has lied on the supplier certificate of analysis for components or consumables instead of periodically verifying those certificates of analysis through testing of incoming lots,” Allen pointed out.

Sometimes investigators see issues with growth promotion testing

As far as inventory control storage, there are essentially two issues that are observed. One is the segregation issue where, for example, rejects are stored together with released materials. The alternative scenario is with a warehouse being a mess with materials not in the correct place, warehouse areas overflowing with materials, nothing is labeled, and so on.

Quality Control Laboratory System Observations

The next system is the quality control laboratory (Figure 6).

Figure 6 QC Laboratory System Observations
FIGURE 6 | QC Laboratory System Observations

The bulk of the observations that were made for cell and gene therapy quality control laboratories have to do with the suitability of the testing methods that either are not verified for compendial methods or the methods are not validated with them, as is the expectation for the commercial product.

Much less frequently investigators see the lack of a documentation system in place for tracking sample storage. An even more rare observation is something that is called testing into compliance. “Testing into compliance is performing release testing on a sample and it fails. The lab analyst then retests it and it passes, and the product is released,” Allen explained. “And that was done before initiating a manufacturing investigation into the failure of the first test. I am mentioning it because it is a really big deal, not because it is very common.”

Packaging and Labeling System Observations

Then next—last but not the least—is the packaging and labeling system.

FIGURE 7 | Packaging and Labeling System Observations
FIGURE 7 | Packaging and Labeling System Observations

For this system, observations fall into three groups:

  • Visual inspection parameters not defined
  • An arbitrary visual inspection rate
  • A product-specific visual inspection issue (more details below)

Normally, if a batch fails visual inspection, the manufacturer will reinspect the batch and we expect that the acceptance criteria would be tightened. If it was, for example, a cryo-preserved product that was inspected, failed, by the time you do a reinspection it has been either released or it was cryo-preserved. 

And to reinspect, you must thaw it. “And that is just not really an option.  We understand that,” Allen explained. “But there must be an alternative strategy, one the manufacturer should plan and develop to deal with those issues.”

An even more rare observation is something that is called testing into compliance

Regarding shipping and labeling issues and shipping validation, the expectation there is twofold. First, the robustness of the shipping container must be validated. And since those products need to be shipped at a certain temperature, the temperature of your shipment must be qualified under the worst-case temperature challenge. One of the two is often not there—either the validation of the robustness of the shipping container or the temperature qualification during worst-case shipping conditions. This has been noted on inspection and also during product review.

Investigators have also seen during inspection that the shipper temperature and its integrity is not verified upon receipt.

They have also seen where the firm does not require secure access to label templates or excess printed labels that are dispositioned for destruction.  Labels returned from production are stored in a fire-rated cabinet that is unlocked. Label accountability is important.


These were the most common 483 observations that we have seen over the past 13 years.

The bottom line is that the most observations were noted for quality system, facilities and equipment, and the production system.

Quality system observations were the most common, but they are generally not unique to cell and gene therapy products. 

Another group of frequent observations have to do with stability concerns.

Special sterility considerations for CGT products reflect other frequent observations, including those for aseptic process validation, environmental monitoring, biosafety cabinet cleaning and qualification, aseptic technique, and method validation—often for sterility and endotoxin testing.

Next, Part III looks at how to cut product approval times for CGTs in half. 

483 assessment