At the behest of Idaho (thanks for not banning me even as I got a little overzealous while voicing my despair in previous posts), I will limit my post to just one for the remainder of the findings in the 483. Just thought I’d finish what I started…
You will recall that in 483’s, the less serious findings go to the bottom of the list and that’s what I’m seeing here. My main takeaways from these less serious findings are:
-IMMU’s manufacturing facility and practices weren’t as “Good Manufacturing Practices (GMP)” as they could be. These things can be pretty fixable. It usually takes a change to procedure or the installation of some equipment. If you’ve got enough people, who know what they're doing, working on the issues, it could be done in 4 months and you would file a supplement to the FDA so they’re aware of it.
-IMMU had issues with tools getting contaminated quite frequently. I say “had” in the hopes that they’ve since fixed the sources of those issues. The problem was that they either had no procedure or no validation for some of the processes where contamination was occurring. It just looks like you’re not focused on cleanliness and quality. You’re not asking the right questions about your process…questions a commercial manufacturer concerned about its customer’s safety asks. But these things are also pretty easy to fix had you the staffing and resources.
-IMMU did not appear to take product non-conformances seriously. I feel this is the biggest bone the FDA has to pick with IMMU. They failed to sound the alarm in some instances where product was out-of-specification. If they did sound an alarm, it was usually days later. Then, when they were supposed to be investigating and correcting the root causes of these non-conformances, they just kind of just let them sit for a while with no resolution. Identifying and addressing deviations helps improve your manufacturing process and protects the patient from any harm. People that want to sell product to patients and have their best interests at heart should take deviations very seriously and want to address them IMMEDIATELY. This is why during any routine FDA inspection, they go and look at non-conformance reports and corrective actions first. This is probably one of things the FDA is holding out for…I’d think they’d want to see some proof that IMMU has gotten their act together in this department before letting them go commercial and sell product to customers.
In summary, the 483 showed us where IMMU’s head was at when Pehl stepped in and possibly where it was for another few months. I’ll have to trust he's handled these issues in light of ladyPI’s post (courtesy of fitz) about what the CRL was driven by: a manufacturing issue(s), which does not equate with a quality or impurity issue. I gave her some speculation as to what those could be (not saying this is what they are):
Here are some manufacturing issues that don't necessarily deal with product quality or impurity:
1. it could be surrounding replacement equipment or new equipment that needs qualification/validation.
2. it can be surrounding product labeling (the actual labels on the packaging) and what that reads. Labeling is pretty important in that getting it wrong leads to recalls of product since doctors have no assurance of what they're putting in patients if it's labeled improperly somehow.
3. It could be IFU(instructions for use) related. We've had to revise our IFU lately for a small piece of information as part of the agreement with FDA to allow us to conduct clinical studies. This could be possible with IMMU since they got that recent "upgrade" to being potentially used as a "second line" treament.
4. It could be packaging in and of itself. If they didn’t test the packaging’s ability to protect the integrity of the product at extreme temperatures & pressure (such as those that could be encountered on a shipment plane or in Milwaukee on a shipping dock in winter) the FDA may require them to retest at those parameters.
These are all pretty easy to address:
1. Do the validation and prove the equipment performs as well as the one that's outgoing.
2 & 3. Contact your printer and change the content of these labels/sheets of paper with instructions on them. It might take a few weeks to get the goods and then you might take a few days to re-work your existing inventory but that's not too bad.
4. Do the retest. Package some representative product, send it out to packaging testing labs to expose it to conditioning, wait 2 weeks and test your product. Write up the report and send a supplement to the FDA.
If it’s any of the above, I’ll soften my stance on the situation and keep my fingers crossed for Class 1 submission. I spoke, the other night, without the knowledge that some of these findings were self-submitted by IMMU. Also, it wasn't clear to me until today that the manufacturing issues weren't product quality issues. The 483 sure made it sound like they were. I apologize to anyone that took offense to my assertions about management but I will try to maintain a healthy skepticism as no one I know out there has perfect knowledge and perfect execution but hopefully Pehl and crew's are good enough to win us speedy approval after the CRL.
If you have free time and like details…the rest of the findings and my comments are below:
Finding #6: Differential pressure between GMP areas of different area classification is not adequately maintained and monitored. Specifically,
A. Air pressure in the GMP areas is not adequately maintained. For example, differential pressure between Rooms _____ (Class C _____ and _____ (Class D corridor) was out of action levels in 37 out of ___ measurements between July 24, 2018 and August 1, 2018.
B. Continuous monitoring of pressure in the GMP areas has been installed in July 2018 and is undergoing qualification, however not all adjacent rooms with different air classification are alarmed for low pressure differential. For example, differential pressure between the Rooms _____ (Class C _______) and ____ (Class D corridor) is not alarmed.
My comments: Differential pressure between Good Manufacturing Practices (GMP) areas are used to maintain the cleanliness of the air in those regions. There are little bits of particulate always floating in your air. A cleanroom seeks to minimize this because particulate could have bugs on them that can contaminate your product. There should always be a differential pressure that pushes clean air from the cleanroom out toward the dirty air so as to keep the air in the room clean.
A. Action levels for differential pressure are pre-determined levels of pressure that must be maintained. If the differential pressure drop below this, the company is supposed to go and take action, hence the term, and correct this. The fact that it was triggered 37 times in a week is troubling. I’d be curious what the other number is in “37 out of ____”. Regardless, there should be alarms that notify staff that action levels have been reached and they should respond promptly. A couple of action limit triggers…fine…37? That’s probably indicative of a process that needs some work.
B. So this is a good tidbit from the 483 that shows us management has been trying to upgrade the place but they missed something that experienced cleanroom designers should have put in place or notified them of…your adjacent rooms that are dirtier need to be alarmed as well so they don’t end up introducing particulate to your cleanroom.
Overall, this is not too big a deal because there are usually downstream processes or inspections in place to check for particulate or contamination before product goes out the door. I’d imagine that IMMU has that and, if nothing else, they have their bioburden tests if the particulate did cause a concerning rise in the “dirtiness” of the product.
Finding #7: The design of the facility is inadequate in that no drains are present in the ____ rooms. In addition there is no SOP for liquid containment and disposal after a catastrophic spill. All process ____ the Production Bioreactor are held in ____ bags.
My comments: This seems like more of an OSHA concern or there some details in the redacted text that make the story. If something spills, you definitely don’t use it anymore so the FDA’s not thinking about this as a product safety finding. Guess they were just trying to be nice in pointing it out?
Finding #8: There is no signed Quality Agreement between ____________ and Immunomedics Inc. _____ is the supplier of cell culture media and all buffers, solutions (including ____, and chromatography resins used for _____ of the _____.
My comments: Our company uses supplier quality agreements to constrain the freedom of suppliers and prevent them from making changes to their process that could compromise the validated state of ours. At the very least, it outlines what they can and can’t do so that they’re legally bound to that. We’ll say things like “you will notify us when you’re aware of the possibility of an Out-of-Spec situation”, “all things you provide must have traceability”, “you will ensure compliance with current FDA and government regulations at all times”, “you will test each batch of drug supplied for cleanliness from microorganisms”, etc…It’s a good thing to have when S hits the fan but it also shows suppliers you mean business and that they better run a tight ship too. IMMU SHOULD have these in place but as long as they can verify these things in-house when the product arrives, it shouldn’t be too big a deal.
Finding #9: No procedure is in place for ______ trending of results. During the process validation ___ campaign, bioburden levels in the chromatography resins were not trended and inadequately high bioburden levels were not investigated. Low level bioburden (_____CFU/___mL) was observed in the ____ resin chromatography after sanitization in ____ batches and the bioburden level increased to too numerous to count in ____ batch. No deviation was initiated.
My comments: This is not my area of expertise but I gather these chromatography resins are used to separate out some of the molecules/proteins of interest in IMMU-132 as they can be used for size exclusion, ion exchange, or other selection methods. So it stands to say that it’s important to make sure your resin is clean before you grab your “parts” out of it because they could become contaminated in the process. Being that this probably isn’t the final step in the processing, the FDA probably thought it fine to put lower on the list of priorities. You can always sterilize the product later but it’s good manufacturing practice to ensure the bioburden on things that touch your product is controlled. This finding is in the same vein as finding #4 where product-contacting tools were not tested for bioburden. There is another recurring theme as well: “No deviation was initiated”. That’s pretty bad. As I mentioned previously, the FDA wants to see you take any potential for non-conforming (dirty/dangerous) product very seriously. They want to see you initiate a report on the same day as discovery preferably. This makes IMMU look like they don’t take quality seriously and that’s a problem.
Finding #10: Deviation investigations and CAPA implementations are inadequate. For example, Deviation 18-053U was initiated after an internal audit concluded that the ____ had not been adequately tested for integrity _______. The deviation included the following deficiencies:
a. Lot number in the deviation form indicates “multiple lots” without specifying the potential lots impacted.
b. Product impact assessment includes the conclusion of a clinical Health Hazard Assessment, but no risk assessment on the presence of _______ in the product is documented.
c. The CAPA section indicates that remediation included “…purchasing additional test equipment to evaluate the ______ts use.” However, at the date of the inspection no additional equipment has been purchased and no information about the CAPA is documented in the deviation.
My comments: Finally! I was expecting to see something about deviation investigations being inadequate given the 3 times I’ve seen the FDA mention this in the 483 prior to this. So typically, the results of internal audits are handed up to management so I’m now thinking that management had knowledge of old IMMU’s dirty habit of not giving deviations enough priority. Hopefully they’ve been working on it diligently since that discovery (the date of which is not mentioned here but if I had to guess, the 18 in “Deviation 18-053U” means 2018, which I guess would coincide with about when Pehl started).
a. If you’re going to write a report about affected lots of products, it’s PROBABLY good to say which lots specifically not just “multiple lots” so you can quarantine those lots.
b. Risk assessment on the hazard is important as it guides the company on how to proceed with the investigation. I think regulators want to see high risk things go straight to CAPA where you document your corrective and preventive action to make sure you never see this deviation again. IMMU is not doing that here which speaks to the quality of its procedures and CAPA process. Needs revamping is my guess.
c. So here, the FDA is saying that “you told us you’re going to order new equipment to fix this but here we are inspecting you and you still don’t have this equipment…is this problem still out there then? IMMU is not showing proper prioritization for corrective actions. This causes auditors to lose faith in your ability to do other things as well because you’re not taking the serious stuff seriously.
Finding #11: Deviation initiation and closing times are inadequate. Specifically:
a. SOP-0152 "Deviation handling" indicates that if the deviation cannot be completed by the assigned due date, a one-time extension can be requested to the QA unit. The following deviations were not closed by the due date and did not include an extension: SEE REVERSE OF THIS PAGE
i. 18-116U: deviation due date was 7/20/2018; deviation was open at the time of the inspection
ii. 18-081U: deviation due date was 6/17/2018; deviation was open at the time of the inspection
iii. 18-079U: deviation due date was 6/16/2018; deviation was closed on 8/3/2018
iv. 18-050U: deviation due date was 5/9/20 I 8; deviation was open at the time of the inspection
b. SOP-0152 'Deviation Handling" does not specify a time limit between time/discovery of event and deviation initiation. The following deviations were initiated more than one month after event discovery:
i. 18-009U: investigation into bioburden 00S # 18-001 for _________date of event was 1/10/2018, deviation was initiated on 3/27/2017.
ii. 18-053U: _________ related discrepancy; date of event was February 2018, deviation was initiated on 4/9/2017.
My comments: the FDA has noticed a disturbing trend…IMMU has trouble initiating deviations when they should. And when they do? They don’t investigate them promptly. This inspection took place in August…one of the deviations due in May was still open. Have they resolved that issue or is it still affecting their product conformance/safety???
It also looks like their “Deviation Handling” procedure doesn’t tell them what amount of time they have to respond to these product conformance deviations. That’s a mistake because it doesn’t ensure action is taken. Our company says we have 30 days to investigate and CLOSE the report and it also says we have to initiate the report as soon as possible after discovery of the deviation. Again, this speaks to the urgency that IMMU places on a process that the FDA considers very important and indicative of the company’s readiness to be commercial. Identifying and addressing deviations helps improve your manufacturing process and protects the patient from any harm. People that want to sell product to patients and have their best interests at heart should take deviations very seriously and want to address them IMMEDIATELY. IMMU’s procedures show that they basically don’t care about the patient (I’m sure they do, but this is the optics of how this procedure is written).
Finding #12: Cleaning of _____ equipment, including _____ and product-contact parts of the ____is not validated or verified. Non-conformance report 18-009 initiated due to a ______ OOS bioburden sample includes as the primary root cause the _____ contaminated during vendor pressure testing.
My Comments: If you don’t validate, you must verify. Even when you validate, you sample (which is a less laborious version of verifying). So it seems some tool that’s tested at the vendor’s was contaminated and IMMU cleaned it and thought it was safe for use. Their cleaning process hasn’t been validated though and you can’t really “verify” cleanliness without waiting for a 14 day bioburden test but they used it anyway. That caused an out-of-spec bioburden sample. Go figure. It was bound to happen.
Finding #13: The procedure to prevent contamination of the ___ after ____ is inadequate for a product stored 2 to 8 degrees C for up to (some amount of time). Specifically, during a mock ____ and dispensing of an ____ surrogate conducted on August 9, 2018, the following was observed:
A. After ______ the surrogate was transferred first to a ___L bag ___(B?) and then from the __ to ___. The __ was removed from its container and assembled in the Biologic Safety Cabinet (BSC) used for ____ and dispensing. Multiple ____ process manipulations were conducted to prepare the ____, including cutting, closing, and connecting several lines and assembly of the ____ to the ___.
B. During the ___ preparation process, the end of the tubing ____ was observed to touch the operator’s hands, the surfaces of the BSC, and the material placed inside the BSC.
C. Prior to filling the ____ analytical samples were collected into __mL ___tubes. The diameter of the tubing (__in; __mm) used to fill the tubes and the ___ tubes (___mm) are similar. In addition, the flow of the ___surrogate was not continuous and was difficult to control. As a result, the surrogate was spilled during the sampling process.
My comments: The FDA is kind of beating a dead horse here. We know IMMU was not great at preventing contamination. The FDA is just pointing out more examples of that. It’s the last finding of the list because they simply observed instances in which product or test material could have BECOME contaminated but there’s no proof that it did unlike the previous findings where bioburden samples confirmed contamination. |
