Jim Mencel is Senior Drug Substance Consultant at Design Space InPharmatics. With an extensive background in CMC management, Jim has a wealth of knowledge on the topic of process validation. Process validation is an integrated and mandatory process in the pharmaceutical industry to ensure all processes are in compliance with regulatory standards. In this episode, Jim provides his thoughts on the significance and science of process validation in the pharmaceutical industry as it pertains to drug substances. He discusses the evolution of batch documentation, expounds on PAR and CPP and stresses the importance of communication.
“Validation is very formal. Validation really, in a stage analogy, is Opening Night. This is where the critics are all present, everybody’s breathing on you, you’re gonna be in the headlines the next day and everything needs to go right. So, when you think about it in that context, the validation batches are very visible to regulatory authorities.”
Welcome back to CMC.Live, where we are not here just to entertain you, but also to help educate and share some of our real-life experiences and recent experiences as well. As I said before, surprises are pretty good when it comes to birthday parties and winning the lottery, but not so much when it comes to developing and manufacturing critical active pharmaceutical ingredients, or APIs as we call them. On this podcast, we're going to be speaking with Jim Mencel once again – this time on the significance and science of process validation in the pharma industry as it relates to drug substance mainly.
So, the documented verification of specific processes and systems against required specifications is known as process validation. There's some good guidance on the FDA website. Validation is, of course, an integrated process in the pharma industry as it is mandatory to comply with national and international standards at FDA and the AMA. Validation ensures that all processes are in compliance with the established cGMP standards. Any validation process requires complete documentation as well, that complies with the standards, the SOPs, and ongoing operations at the facilities. This includes the production plans, the validation protocols to development reports, and also analytical documentation. Each and every step of the manufacturing process, including subsequent processes and modifications, is irrefutably verified to ensure the complete safety for the pharma products that these people rely on out there in everyday life. Process control validation helps in bridging the gap between the ever-changing quality standards and dynamic market forces. From what I understand, today Jim will be providing us a Broadway analogy for validation. As always, I have here again both Meranda and Brian joining me as well. Hello, folks. Brian, would you like to start this conversation?
So welcome, Jim. Good to have you here. So, you know, we've talked to a few folks from the drug substance team throughout the podcast series, and one of the areas that we hadn't quite covered is the importance and timing of validation, registration lots, you know, as it relates to the advancement of the process – when do we start up and things like that. Where we left off in our story was actually with process scale-up; we had a process, we could get into clinical trials with the scale-up. And now I'd like to pick up where we move forward into late phase, possibly leading up to our filing activity. And maybe we start with, now you've got a scalable process that's manufactured Phase 2 material; maybe you can talk about some of the steps leading up to and including validation lots.
I think a place to start, Brian, let's talk just a bit about validation. So, you've got a process that you're using for late Phase 2, or early Phase 3, and you have some sense of confidence that it's reproducible. Maybe there're some improvements here and there that you need, but you may not be operating at your fullest scale. So there are typically three activities that occur that are really intended to confirm the process, and one of these is called registration lot, and there are usually three of those. Those are specifically run to generate material for very formal stability data. They also are run at a fair percentage of your intended commercial scale in the commercial equipment at the commercial site. The reason for this is because the data for your stability is tied to that site, because that's where the process was run.
Let's say that in the stage analogy, that's your dress rehearsal. Okay, that's behind closed doors, it's among friends. You can have adjustments along the way; deviations are not a problem. Obviously, you want successful batches, but it's an opportunity to really put your process through the paces. Those are typically followed by what's called an engineering batch. The engineering batch is typically one or two runs of each stage with all the changes in place. So, from your registration campaigns, you're going to learn a few things: order of addition, maybe rate of addition; and you run your engineering batches with the idea that this is going to be a further dress rehearsal. This is even a little bit more technical in some ways, because the data from the registration batches doesn't get used for anything in particular, but if they run well, they could be used commercially; we'll talk about that in a little bit.
So those are your engineering batches, then you follow with your validation. Validation’s very formal. Validation, really, in a stage analogy, is Opening Night. This is where the critics are all present, everybody's breathing on you, you're going to be in the headlines the next day, and everything needs to go right. So, when you think about it in that context, the validation batches are very visible to the regulatory authorities. When they review your filing at some point, they can, at their option, and probably will review information from your validation batches. Now, in current parlance, these are called Process Performance Qualification batches; the historical term has been validation, and the thought process for decades now, a couple of decades at least, has been three batches. There's no longer a requirement that people do three batches, because the view was that if you've done it three times, and it looks good three times, the process is confirmed. Today it really is up to the sponsor to propose how they're going to run validations and how many batches are going to run; most people will default to three batches still. In some industries like controlled substances, you can't, because the DEA only allows so many batches and so much material per year. So, there is a certain amount of latitude there, because you simply can't make three batches because you don't have DEA quota to do so. But for most general APIs, you can run three batches. That's sort of the progression. Some people do their engineering batches first, before the registration batches. The important thing is that those batches proceed your validation.
Let's talk about getting ready for validation, then. You need to have a process you feel some confidence in – a lot of confidence. The registration batches, the engineering batches should improve that confidence, identify further adjustments. At some point, your process needs to be at a state where you would consider it what's called frozen -in other words, no more major changes. That really comes as a consequence of having run it, having made some fairly substantial batches, and feeling confident that it's going to deliver what it needs to.
So again, coming from the drug product world, just trying to see if we have similarities. I know engineering batches will push our operating ranges or try to establish thorough value-added in-process data. So when you do lock in that batch record, and you go to challenge those limits and validation, you've done it before. So how important is it on the side, when you're compiling that protocol for that engineering batch, you're looking at key process parameters. And you don't want to push it, obviously, to the point of failure. But you do want to get a lot of information, not just to show you can run it optimally three times.
“Your batch record really is a document that’s going to evolve. And if you look at the batch record for even the same step of a process, what that’s going to look like when you’re in the early phases of clinical development is different from what it’s going to look like towards the end. It’s going to be much more refined.”
So, in the API world, that expansion of boundaries is going to occur in a very formal program that's often called your PAR, or your CPP – PAR being Proven Acceptable Range, or CPP (Critical Process Parameter) program. Those boundaries are expanded in the lab. So, you're coming from some history where you've got what are called center point parameters. Okay, so you've run the process, you know you need to heat it and reflux in toluene. We know you need to stir it for twelve hours. After running a number of batches, you have what I call the center point – these have worked. But what you don't know is how far left or right, above or below center point you can go. So, what you'll do in a laboratory setting is you'll challenge those numbers, and those studies are very formal. That is where you will decide what’s the highest temperature I could run this at, what's the lowest temperature I could run this at, where is my sweet spot.
What you're establishing is what is called your proven acceptable range, which is the boundaries of where you're going to run any specific operation. And then within that is what's called your normal operating range, which is where you're just a little bit left or right of your center point. So if your proven range is 10 to 30, your standard range might be 15 to 25. You have five degrees either way, for your total range. The importance of those things is that your PAR, you can run within your normal operating range and have that variability. Say if your range again is 10 to 30, and you set a normal operating range of 15 to 25. Well, if you go to 27 it becomes a deviation, but it's not serious because it's within the range of what you have said your process can tolerate. Because of the cost for running API processes, and often today because of the pace of work, there isn't really time in the plant to run whole processes at the extremes of those settings; you do that in the laboratory, and you do it with very controlled experimentation. People often use automated reactors to make sure that they have, you know, very, very reproducible performance in the experiment.
I should say that before this, you'll do a risk assessment of the process and decide where those parameters are – that if they vary, can cause a problem. You might know from history on the batch that the temperature is an issue; you might know that cooling rate has got to be carefully controlled. So you need to do a risk assessment. This is a team exercise, and there are various names for them. But at the end of the day, what it means is you sit down with the people that know the process really well, and you go through the entire batch record for the process. You look at all the stages and say, okay, so is this a step where there is a risk of failing for quality, and you score that risk; then based upon the scoring, those risks that have the highest score, where it could really threaten the quality of the batch in a way that's either unrecoverable, or at least very difficult to recover from, those are where you do your experiments to understand how wide your range can be before you get into trouble.
You said something that was pretty telling because when Dave Adams came on, and we talked about process scale-up and things like that, he did mention getting the input of the people on the floor that run this process that understand that they’re, you know, manipulating the product the entire time, and make sure you don't exclude those from your consideration when you're getting put into these effective ranges. You got equipment tolerance, things like that, that all have to be considered. And one thing you said is that you're really talking about the evolution of the batch documentation as well, right? Because what becomes your validation batch record, ultimately, your master batch record, is the information you're gleaning from the engineering batch, right? And registration lots?
That's correct. So, your batch record really is a document that's going to evolve. And if you look at the batch record for even the same step of a process, let's say you're doing hydrogenation, right, what that's going to look like when you're in the early phases of clinical development is different from what it's going to look like towards the end. It's going to be much more refined; conditions will be much more defined. And it would have moved in a direction where your reproducibility of outcome for quality, certainly, which is always the regulatory concern and quality concern, and then also for yield, which is an economic concern, right? All those things will have evolved towards the point we're running that process will give you that result every time.
You know, now, what I'm discussing here is what should occur. And in a perfect world, all of this goes exactly right. The thing to be aware of is that there are times when even in the best studies in the lab, you simply miss things, there are things you just don't see, because you can't model those well. Your validation batches may well be the biggest batches you've ever run up to that point in time. And you can have a situation where something occurs that you've not seen before, even during validation. So, it is important to try to really risk-assess the process and figure out where there are things, and really challenge yourself to look for hidden things. Because there can be things that will arise, you would never have expected, that you simply did not catch in your risk assessment. So that does occur, and it creates complications we won't get into right now, but the fact is that those things can occur. Often, they have less to do with the chemistry than they do with the physical operation of the batch. I'm aware of a case recently where during validation it was discovered that the horsepower for the motor stirring a very heavy batch was inadequate. And that's something that the chemists sitting in the room would never have known. So that's why you need the wide group.
Now at a point here, for sponsors using CMOs, there's an added challenge. You know, you need a person there on site to meet with all the people that are involved. In recent days with COVID, that's been very difficult. And it does have an impact, because you do want the immediacy of the people that know that process in a room together, if possible, to get all of the perspectives on what could be affecting that process at a given point. This issue with the horsepower on the motor, it would have been an engineer on the site that would have raised that point, and it appears as though probably that person was not involved in the discussion; whereas I think if you were on site, you'd have a more formal gathering of all the people involved. So, I think that there is a value going into validation, or into your PAR studies, your Proven Acceptable Range studies, to make sure that you’ve got all the right voices that need to be there, because you will certainly miss things that a chemist would not necessarily think of. You need to be a really experienced plant chemist to catch all this stuff. And there are people that work on the floor, essentially chemists that are just like engineers, or engineers that are really good as chemists, that will notice things that would be very important to take into account. So you're right, Brian, the conversation involving many people is just crucial.
There's nothing worse than a batch record being issued to the floor that has absolute values, and no operating ranges. I've seen that, where clearly it wasn't well thought out; and then when you start looking at your validation, which is much further down the line, and you start pushing those limits, because suddenly someone realizes you need to have operating limits. Yet, the main question is, where do you get that limit from? Does it consider all the factors? You get stuck. And that's why those kinds of thinking and as you said, inviting everybody into the process of making sure there's buy-in to those ranges, is really important, because you've invested so much by the time you get to validation. One lot’s failure could set you back significantly. So yeah, it's really setting yourself up to succeed.
Yeah, it's a hard thing to get ready for, in some ways; you need to start well in advance. I mean, let's face it, you haven't always when you're developing a product a gamble, right. And we've talked about this before, you know, the investment in CMC early on is hard to argue for, because really what people are interested in is material – because they want to get the clinical data, because that's necessary for most of the clients that we would have. They're venture-funded, or they're on the stock market; they don't have a product they sell, they have this material, right? That's a lot of the people that are out there, emerging pharma, and they need that next clinical result. And that's completely fair, just like the guy in the batter's box needs to hit the next home run. That's how you fill the stands; that clinical data is what fills the coffers, you absolutely have to do it right. And the reality is, there isn't a lot of money for CMC. But where that comes back to affect you is when you're getting ready for validation, for those registration batches, which are really your first real scale-ups, I think, in most cases, your engineering batches, all that information that you did not gather, all the development work you did not do really begins to play out because now you're getting ready to run this in a way where you really need to control it, and don't know your boundaries.
So that's where really these critical parameter studies become so important, because that's the chance to go back and try to figure out where are your boundaries. But there is a point where the process is the process. And as inefficient as it may be, it's what you have. I like to use the analogy of you know, of a mule versus a racehorse: what you need in most cases is the mule that's gonna get you over the mountain pass in a snowstorm – doesn't have to look good, doesn’t have to smell nice, but you need something that is going to be reliable and strong. So that's what you've got to work towards, because the time to develop the thoroughbred just isn't there for most companies. And you want to make sure as you're going into your registration batches that you at least know the mule’s reliable. When you get into that validation, that mule has to be 100% reliable, so that it's gonna get the same results every time.
“There is a point where the process is the process. And as inefficient as it may be, it’s what you have.”
In drug product, we say we strive for boring, boring is good.
Or it would be great! And I've never seen boring, but boring would be great. Yeah, that's something to strive for. But I have to say that on the programs that I've been on, through registration on into validation, you're not hearing any crickets. It's a pretty active discussion with a lot of things to deal with along the way.
Yeah, absolutely. Let me ask you something. When it comes to registration lots, how much do you have with the regulatory group, with the sponsor team, in terms of strategy for registration, the amount of batches that are considered registration? You know, as you said before, in the past, three’s been the magic number forever. But now you've got science behind it, and what kinds of leverage do you find, or interaction you have with the agency that maybe a number, perhaps different from three?
I've not seen a lot of latitude on that. And, you know, a lot of it's indirect. So for example, I've been involved in a project where there was a need to change manufacturers very, very late in the program. Brian, you're familiar with this program. And I think that some of the most useful feedback we got on the agency's current thinking was that they came back and asked for three batches. You know, and this was three batches where they wanted stability, they wanted an assessment of similarity versus the first across the company that was being replaced. So you know, in the FDA’s mind, at least, three seems to still be the number. I've not personally tried to make a case to make fewer registration batches. Now, you know, the thing we all need to realize is that registration batches are a recent phenomenon in the regulatory landscape. In the past, you know, people didn't do this; they did the validation batches, they put them on stability, and that was it.
So this is something new, and it's now more or less expected. They are specifically directed towards developing stability data, and generally FDA wants stability from three batches, and they want it a specific length of time before you file. So because of the way things go, if you were to wait to do your validation batches – those are typically done right before or even during the review of your filing – you won't have the stability data necessary for them to approve your product. So that's why people make their registration batches up front. So this way, they know they have the time, there is time to remake one if something goes wrong, and you're not pressing up against an NDA filing date, not having the batches in hand to get your stability data – because they will not accept your program unless you have the necessary stability data. That's one thing they won't negotiate on.
If you think about it, I mean, the month leading up to that is retesting, right? You retested the material and away you go, but you're not making formal stability claims when you're in clinic. And so it really benefits you. I mean, if you've proven your process and your product is stable, then the more stability data you can get, the sooner you can commit to registration batch number one, you know, within control, and start developing that real-time stability history, is key. And it really ties into the strategy for how you want to use that product in the market. What is your expectation for shelf life? How does it translate from what you've done in clinic and what you've seen of the profile of the product in the clinic, and now you want to submit that claim for expiry date.
So, yeah, it's important, and I think that's a really good distinction that you make, because you're right; registration batches haven't been something that we all grew up with. It's been validation. And historically, validation was done in advance of your filing, and then your launch. But now, you can push and maneuver those validation batches to possibly be even part of launch stock, pending the closing of the Validation Report. I'm thinking on the drug product side; I can't truly speak from the production side, but you're right, if you're pushing into validation lots later and later, you don't have the claim for stability, the timing of the registration batches, and locking in the process. Really that is more important.
“When you’re in Phase 3 it’s not a place to be penny-pinching. The old saying is, ‘you drive for show, you putt for dough.’ When you’re validating, you’re putting for dough.”
So you mentioned about launch stock. Validation material can be used for launch; in the eyes of the agencies, it is essentially commercial stock, because you're making it using the commercial process. Now, I mentioned this idea of the engineering batches. Those, if they are run extremely similarly to the validation batches, can also be used commercially. So, typically different companies, different sponsors, do them at different times. But if they're run right before validation, as a final dress rehearsal – let's say this is a dress rehearsal where you invite senior citizens and others who would enjoy it, can come in for free – so it's a dress rehearsal, but it's not quite among friends, it's a little bit more public, right? But the reason it's more public is because you can use the engineering batches for real, you can use those commercially, but only if they are run very, very comparably to your validation batches. And you can make a case that any small wobble is not of any consequence to quality. So there is that side of it to be aware of. Yes, your validation batches certainly count as launch stock as long as they pass all specs. That’s launch stock. Yeah, that's standard practice.
Well, it's funny, what you and I consider to be standard is oftentimes discussed and debated with some of our sponsors. So it's good to know that, for those folks that are listening in on this, that this is considered standard in industry.
Yes. And, again, you know, when you look at funding, everything has to be to a purpose, right? And certainly the end objective here is to file and to get approved. And all those batches that are made along the way, you know, they’ve got to count for something. So your registration batches typically are used clinically – and by the way, so is your engineering batch if you have clinical need; but more often than not, the registration batches will be used clinically and not commercially. But your engineering batch, should you choose to do an engineering batch, could it be used commercially, if it's close to the validation batches? Absolutely. From a cost standpoint, it is wise to aim for those being commercial material, for launch stock. Because there isn't really a lot of time, unless you have a very short synthesis. You're going to finish those probably sometime before you're filing, not long before; there isn't a lot of time after that to make your launch stock, and depending upon the size of your launch, you're gonna need all the material you can get.
Yeah, well, and then let's say you're not making that many batches a year; all the more reason to capitalize on using the validation lots and add them to launch stock, because you can be making more of the API you need, or drug substance you need. So you know, it has to be a strategy that's thought out and accepted by more than just the technical team; it has to have understanding from quality, because they're going to want to weigh in on the controls that are being used for the batch because if they're going to justify a disposition of a lot, they're going to want to see that, it's going to have to have regulatory, because you might want to discuss that strategy with the agency if it's something atypical. So yeah, it really is something that these conversations should happen sooner than later on.
“Yes, your validation batches certainly count as launch stock as long as they pass all specs.”
Now, you touched upon another phrase, Brian, the idea of the protocol. So you know, for the registration batches, it's a batch record. There's no protocol. For the engineering batches, again, it's a batch record. There's no protocol. But when you do your validation, you write what's called a protocol. And just so that people understand what a protocol really is, it's predicting what your outcome is going to be. You're saying, I'm going to do this, and it's going to give me this. And it's a bit frightening, when you think about it. You're saying, okay, we're going to run this step with these conditions, and it's going to give us these exact results, right? So you're committing to this saying, this is what we're going to see, and if you don't, you have some explaining to do. So, the protocol is a formal document. I think it's very, very important, when you are reviewing the protocol, you need to have a Quality person, a QA person involved, okay? There should be a Regulatory person involved, there should be a Quality person involved, certainly Technical.
I think something that gets overlooked by a lot of people is the importance of a QA person on a team: they're essential, they understand what roles need to be played by what people, they look at a validation protocol from a very different perspective than a guy like me who's largely technical, or understands the ICH guidances, in terms of how you defend things from data. A Quality person takes a very different look, and you need a person like that on the team, especially as you begin to move into validation in general, to begin to look at the sorts of things that need to be in place, so that your aspects of quality are being properly met. That's one thing.
The other thing I wanted to point out here, Brian, that we've missed is, when you go into validation, you've gotta have your suppliers identified. We missed that, too. And Quality plays a huge role in assessing your suppliers. And these would be for your regulatory batch starting materials. Okay, now, we won't go into what regulatory starting materials are, that's been discussed previously. But you need to be sure that Quality team, the Quality unit of the sponsor, and the Quality person that might be brought in as a contractor, are heavily involved in verifying that the sources of material that are going to be used for validation are fully vetted. Very important. I'm putting a plug in here, because I think that the role of Quality is not necessarily always understood, but I can say that when you're getting into validation, the role of Quality is essential. They are a major pillar in making sure your validation goes properly.
Absolutely, and if you don't invite them into your process, you're just going to have a tough time at the disposition; they need to know what's being done, because they're the ones that ultimately put their name on the release of things. I mean, there’s a technical review, but Quality has a big responsibility to these batches, and getting them involved, and being in the know, it just makes life easier. It really does. Because they're also going to guide you. Let's say for example, controls: I'm not always thinking of quality controls as they relate to batches, but that's just what they are thinking. That's how they're wired. So, it's good to have that input, which is the beauty of having a multi-discipline team working on a project.
Yeah, the multi-discipline is important, Brian, it does take a team. As I said, the risk assessment takes a team; the review of the protocols takes a team; to review the batch records takes a team. You're going to receive also what are called the master batch records going into a validation, and those essentially become your commercial batch records. So, what is in that batch record, what is in that protocol, is what you're committing formally to the agencies, is what you are going to be doing for your process. And if you’re going to vary from that, there is a new draft, ICH Q12, which talks about post-approval changes, and it defines for you what you would need to do during the validation phase, a pre-validation phase, to set yourself up for changes later on.
But you know, when you validate, you’re committing to run the process a certain way, at a certain facility, in a certain type of equipment. So it's very specific to site, it's specific to type of equipment. Everything is defined; that’s why I said that the validation protocol is almost like predicting the future: we're going to run it in this equipment at this site, here's our parameters, we're going to get these results. And you obviously have to have a fairly good degree of certainty before you go into this because it's expensive, it's very formal, and everything you run, whether it succeeds or not, goes into your validation report; then you've got to explain why a batch was not used.
Then those operating ranges become the operating limits that you find in your master batch record. That's for batches going forward. And you know, it's based on the work you did in validating in engineering to develop the process. And when you see those operating ranges, you may decide because of the type of equipment, let's say hypothetically, you're running at near the end of the operating range of a particular piece of equipment, you're going to want to make sure those operating ranges in the batch record reflect those controls, to make sure that you run the batch consistently each time. So, this is all in that validation report. For people that are looking at this, if they have a deviation generated, and let's say your operating ranges, at least on the drug product side are stated in the batch record, well, they may not be the same ranges that were qualified in the validation report, because you want to have a safety margin. But you'll go back to that validation report, when you try to disposition investigations or deviations. So it really is a document that's got a lot of purposes beyond just validation.
The PAR report’s important that way too, Brian, I mean, the PAR report, it's tough, because as I said, you do learn things that work for the Proven Acceptable Range, your risk analysis, there are times you actually have to go back and revisit that, because you've discovered something you didn't know before. So, you may have a signed report going into validation, and find out you now need to open that up and put an amendment to it. Now, that's not ideal at all. But it is reality; you really have to have an active conversation with the team to decide how you're going to handle these things when they come up, and how you're going to position what you're going to say so that it's acceptable to authorities. The authorities are going to come in, and they're going to read this information at some point, and they can ask for anything they want to see when they're reviewing your filing, and they do their inspection on the site of the manufacturer. You've got to have very, very good explanations for things that didn't go exactly right, or why you chose to do one thing instead of the other – that needs to be well thought out in advance.
So, what it all goes into is that preparing for the validation really is something that you should be thinking about as you're making your later clinical batches, your Phase 2 batches, your Phase 2a and b – your Phase 2 batches especially, your 2b batches – you need to start thinking, that process should be looking more and more like your commercial process. And when you're in Phase 3, you're really in practice for commercial, and you should have very, very few further major changes to your process, if any, at that point, and be locking it in while you're making your Phase 3 batches. You know, your early Phase 3 batch may have some room for change. But as you get through Phase 3, those batch records should be pretty close to final.
“When you go into validation, you’ve gotta have your suppliers identified. And Quality plays a huge role in assessing your suppliers.”
Well, I mean, we always say that the limits, everything is based on good science. Well, that's where the science is found. When you're developing your submission, and you may not have all the validation executed, but you're going to have that protocol available. And you'll be able to tell them, this is the plan at which we go forward and validate our process. But it really is important to have it well thought out and easily defended.
Yeah, a lot of it is thinking ahead. Now, you were talking at one point about the timing of when you're going to move towards validation; you know, Brian, that's really program dependent. So, in your standard program, you're going to do your full Phase 1, full Phase 2, full Phase 3, and you're going to file your morbidity and mortality results at the end of your Phase 3 – that's a classic program, right? And in that instance, you're going to want to begin planning for process validation certainly at the beginning of Phase 3. Depending on how long Phase 3 is going to be, you might have some time some runway, maybe not. But certainly you have the latitude of starting to plan for this in Phase 3. It would be during Phase 3 you would make your registration batches, okay. They would be used for your registration for they'd be used for your Phase 3 studies, your engineering batch would occur during Phase 3, your validation batches would also occur during Phase 3.
So, you know, Phase 3 can take a number of years because of the size of those studies. If you have a program, though, that is being expedited for some way, the FDA has these categories where you can submit for earlier filing based upon some agreed set of data: it could be surrogate endpoints, it could be some smaller set that doesn't include morbidity and mortality from the clinical trials; but there's something that they'll agree to. And typically, there are mechanisms by which programs are designated as expedited in some fashion. They would mean that you would be filing sooner, and the requirements for CMC don't change. So, if you're on a program that's going to get filed sooner, then everything associated with process validation needs to start sooner. It's really important in the organization to know, when do you think you're going to file, at what phase in your testing and where during your clinical program do you expect you're going to file? Because you've got to work backwards from there to give yourself the time to do the registration batches (if you're going to do them), to do the engineering batches, and then to do the validation batches. Those are the programs typically where you know, your validation might carry right into the filing review period. Those could be instances where in Phase 2, you could be looking at actually doing your validation, because you actually file at the end of Phase 2b.
Now we have quite a few clients in that bucket. But it's funny you say that, because in one of our earlier podcasts we've had Catherine Bernard on, and she was speaking to expedited drug development. And that's exactly it. That tag, that label can be misinterpreted to mean that you can skip large steps in your CMC development, and you can’t. The requirement for that is still there. So, it's really important to note – and Catherine said this – that you're just starting it sooner, you really are – you're having to start sooner because you don't have to time; the program is going to advance quickly. So considerations for shelf life, critical process parameters, all of that stuff has to be discussed sooner. So basically, what I'm saying is that what you're saying actually agrees with Catherine, from a previous podcast.
Yeah, you don't have the time. One of the problems with the expedited program is that your process is there, but you don't have the years of experience and the number of batches behind you to have begun to smooth out what I like to call “the wobble”: the odd result you can't understand, right, you have that odd result, and that's an NF one, and you don't get a chance to go back and figure it out. So it's a daunting challenge, really, if you're going to be filing earlier, because you do need to validate earlier, and validation implies that you know your process really well. And the simple fact is, being on some of these programs, you don't know it as well as you want to know it. And you might not even know it as well as you think you know it. So, you're going into validation, which is essentially predicting the future, with something where you don't necessarily know you can predict the future. So it's hard. And honestly, this whole expedited filing experience is something recent in the United States.
From our experience at DSI, it'd be interesting to see how these kinds of filings play out when everything is not perfect on CMC: you don't have a fully developed process, you don't have 1000 batches under your belt, you know, your scale-up progression has been fairly rapid from end of Phase 2, maybe during Phase 2 you might have gone to registration batches that are fairly large-scaled – and suddenly you're in validation. It's like a Doppler effect sort of thing, where you hear it in the distance, but the pitch just gets higher and higher as the validation approaches you, and it comes really fast. And it’s screaming at you from behind. Before you know it, it's at where you are. So yeah, that's where a number of our clients have been, I've got one or two that are in that situation now. And it does make it tough to be in validation, because all of a sudden, you start seeing things like, well, we never saw this before. Despite all the studies you do in the lab, you don't have that ballast of plant batches, where this particular thing came up, and this particular agitator wasn't the best, and this thing didn't drain well. You know, a lot of it isn’t even the chemistry at that point; often it's the more physical parameters that affect you, because you just don't have that much experience at scale to know: so, you’re using a powder transfer system, how fast should we transfer? Do we need to put it into a pre-loading hopper first? And how do we do this? Really, you have to take the best experience you have and extrapolate from that how you're going to run these batches, because you don't have the ballast of all these clinical batches to work from.
And then it also makes authoring the actual submission a challenge itself, right, because you're pulling from a limited batch pool of information. So really, I think the takeaway from all of this is communication. It's an understanding from the Quality and Regulatory groups of what you're up against. You're making decisions based on data that is good data – but there may not be a lot of it to show a real thorough history that you're used to seeing in a traditional timeline. So, it really is it comes down to communication with the agency or internally with the company, but you all have to be in lockstep, or the filing becomes another challenge.
It does. So, one thing that I've seen happen recently, Brian, is we did a very thorough lab assessment for parameters for a process; we identified what were clearly statistically proven – and the FDA requires statistics – statistically indicated, these are your clinical parameters. Well, it turned out that what the PARS studies told us was that these particular things were important; but we found that – and we found that even within the normal operating range – we could have a failure for a particular impurity, almost a 50% fail rate. And yet, we had a history of many plant batches for this one at fairly substantial scale and did not see a failure for that particular impurity. So the history of batches were like six batches, right, which is not a lot, but it's a fair number. And yet, we did not see a fail – anything close to fail – for those six batches.
So, we don't have enough batches for this particular material to know what we're seeing; for example, if we run another six, would all six of those fail to get out of twelve, a 50% fail rate, right? So there's something where we don't have the plant history to say either look, something's wrong with the lab data, or we haven't run enough batches to see the fail rate creep in. So that's concerning, because you're going to go into validation with something like that. And is that the time when these things pop up? Or is the lab study telling you something that's not true? So, that's a difficult thing to deal with. Again, without that plant history, which is really what matters, you have the lab data to work with, and some limited number of plant batches to work with. You have to do the best you can to extrapolate from there. Again, communication: you need a lot of people involved, a lot of voices, to look at all the circumstances you're dealing with and figure out what is the best way forward. You’re really piloting a tanker in the fog to some extent, you know. I heard that analogy once, and that's very, very accurate.
“What it all comes down to is that preparing for the validation is something that you should be thinking about as you’re making your later clinical batches, your Phase 2 batches especially.”
Well done. You're right, it needs to be well thought out and planned. And I think that's really the takeaway from this entire podcast is really, it's not just what goes into validation, it's so much more leading up to that to ensure your validation is successful. And really, it's a group effort. I can't stress that enough. But this was a really helpful podcast, I believe, because it shows that so much more goes into it than just validation. I don't think there's a person in our industry that doesn't thank you for the definition of what a validation model. But to truly understand everything that goes into it to ensure it’s successful is really where the trick is. And that's where you have to rely on people like yourself, that have been there and done that, and are willing to share that information.
Well, it's been good to be on, Brian; you know, we’ve sort of shared this in a very informal fashion, everything short of a cup of coffee or a pitcher of beer. But you know, we've both been in the field for a while. And I hope it's useful to people just to have the conversation, to make sure you've got the right people on the team. When you're in Phase 3 it’s not a place to be penny-pinching. The old saying is, “you drive for show, you putt for dough.” When you're validating, you're putting for dough. That's where every shot on that hole counts; you won't get them all in, but your chances are a whole lot better if you go prepared.
The one thing I love about this industry is you never stop learning: from your coworkers, from feedback from an agency, whatever the case may be, you never really stop learning. And so I know in the time that I've worked with you, I’ve gained a lot of insight and perspective that I didn't have before. So again, I want to thank you for your time. I think this was really helpful, and much appreciated.
My pleasure, Brian.
All right. Thanks.
Once again, thanks a lot, Jim, for visiting us here on CMC.Live. Be sure to catch Jim's previous visits to our podcast and let us know what you guys think of these topics. We'd love your feedback; we love your comments. Make us think about things in ways that we haven't thought about them, and we can discuss. Very important: subscribe to our channel if you haven't already, and you'll get notified immediately when the premiere of any future episodes of CMC.Live are out. It's like the only forum out there, and “likes” go a long way, by the way. It's an amazing thing we owe to all of you: that you like the show, and you're into it, and sharing it with others, some of your colleagues out there, spread the word. Come back next Friday for another episode. Either way, we'll have our whole Quality team on next week joining us. We actually plan to have an interesting discussion around Quality Management Systems. You'll hear all you want to know. So, Meranda and Brian, I will see you then.
FDA CMC regulations and guidance simplified through examination, real life experiences and risk-based advice. This podcast hopes to educate sponsors and individuals on agency related regulatory CMC matters. We will focus on the critical CMC issues and build programs that enhance drug development. CMC topics will include Regulatory Starting Materials, API and Drug Product Process, Formulation Development, Supply Chains, Analytical Controls. Advocating and interpreting CMC Strategy, directing CMC Operations and Quality Assurance oversight in conjunction with developing CMC submission content that represents the best interests of emerging biotech. NOT INTENDED TO BE PRESCRIPTIVE ADVICE BUT RATHER INTERPRETATION THAT IS RIGHT FOR YOU. Since 2007 we have provided our partners with innovative strategies and exceptional advice intended to enhance program development, product approval, and marketing presence.